d/freebsd-dev
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
2b519b1707
freebsd-dev/sys/fs/nfsserver/nfs_nfsdserv.c

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

6214 lines
173 KiB
C
Raw Normal View History

Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*-
sys: further adoption of SPDX licensing ID tags. Mainly focus on files that use BSD 3-Clause license. The Software Package Data Exchange (SPDX) group provides a specification to make it easier for automated tools to detect and summarize well known opensource licenses. We are gradually adopting the specification, noting that the tags are considered only advisory and do not, in any way, superceed or replace the license texts. Special thanks to Wind River for providing access to "The Duke of Highlander" tool: an older (2014) run over FreeBSD tree was useful as a starting point.
2017-11-20 19:43:44 +00:00
* SPDX-License-Identifier: BSD-3-Clause
*
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
* Copyright (c) 1989, 1993
* The Regents of the University of California. All rights reserved.
*
* This code is derived from software contributed to Berkeley by
* Rick Macklem at The University of Guelph.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
Renumber copyright clause 4 Renumber cluase 4 to 3, per what everybody else did when BSD granted them permission to remove clause 3. My insistance on keeping the same numbering for legal reasons is too pedantic, so give up on that point. Submitted by: Jan Schaumann <jschauma@stevens.edu> Pull Request: https://github.com/freebsd/freebsd/pull/96
2017-02-28 23:42:47 +00:00
* 3. Neither the name of the University nor the names of its contributors
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
* may be used to endorse or promote products derived from this software
* without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
#include "opt_inet.h"
#include "opt_inet6.h"
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* nfs version 2, 3 and 4 server calls to vnode ops
* - these routines generally have 3 phases
* 1 - break down and validate rpc request in mbuf list
* 2 - do the vnode ops for the request, usually by calling a nfsvno_XXX()
* function in nfsd_port.c
* 3 - build the rpc reply in an mbuf list
* For nfsv4, these functions are called for each Op within the Compound RPC.
*/
#include <fs/nfs/nfsport.h>
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
#include <sys/extattr.h>
#include <sys/filio.h>
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/* Global vars */
extern u_int32_t newnfs_false, newnfs_true;
extern enum vtype nv34tov_type[8];
extern struct timeval nfsboottime;
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
extern int nfsrv_enable_crossmntpt;
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
extern int nfsrv_statehashsize;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
extern int nfsrv_layouthashsize;
extern time_t nfsdev_time;
extern volatile int nfsrv_devidcnt;
extern int nfsd_debuglevel;
extern u_long sb_max_adj;
extern int nfsrv_pnfsatime;
extern int nfsrv_maxpnfsmirror;
nfsd: Add sysctl to set maximum I/O size up to 1Mbyte Since MAXPHYS now allows the FreeBSD NFS client to do 1Mbyte I/O operations, add a sysctl called vfs.nfsd.srvmaxio so that the maximum NFS server I/O size can be set up to 1Mbyte. The Linux NFS client can also do 1Mbyte I/O operations. The default of 128Kbytes for the maximum I/O size has not been changed for two reasons: - kern.ipc.maxsockbuf must be increased to support 1Mbyte I/O - The limited benchmarking I can do actually shows a drop in I/O rate when the I/O size is above 256Kbytes. However, daveb@spectralogic.com reports seeing an increase in I/O rate for the 1Mbyte I/O size vs 128Kbytes using a Linux client. Reviewed by: asomers MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D30826
2021-07-16 22:01:03 +00:00
extern uint32_t nfs_srvmaxio;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Make it possible to force async at server side on new NFS server, similar to the old one's nfs.nfsrv.async. Please note that by enabling this option (default is disabled), the system could potentionally have silent data corruption if the server crashes before write is committed to non-volatile storage, as the client side have no way to tell if the data is already written. Submitted by: rmacklem MFC after: 2 weeks
2013-01-18 19:42:08 +00:00
static int nfs_async = 0;
SYSCTL_DECL(_vfs_nfsd);
SYSCTL_INT(_vfs_nfsd, OID_AUTO, async, CTLFLAG_RW, &nfs_async, 0,
"Tell client that writes were synced even though they were not");
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
extern int nfsrv_doflexfile;
SYSCTL_INT(_vfs_nfsd, OID_AUTO, default_flexfile, CTLFLAG_RW,
&nfsrv_doflexfile, 0, "Make Flex File Layout the default for pNFS");
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
static int nfsrv_linux42server = 1;
SYSCTL_INT(_vfs_nfsd, OID_AUTO, linux42server, CTLFLAG_RW,
&nfsrv_linux42server, 0,
"Enable Linux style NFSv4.2 server (non-RFC compliant)");
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
static bool nfsrv_openaccess = true;
SYSCTL_BOOL(_vfs_nfsd, OID_AUTO, v4openaccess, CTLFLAG_RW,
&nfsrv_openaccess, 0,
"Enable Linux style NFSv4 Open access check");
Allow setting NFS server scope and owner. By default NFS server reports as scope and owner major the host UUID value and zero for owner minor. It works good in case of standalone server. But in case of CARP-based HA cluster failover the values should remain persistent, otherwise some clients like VMware ESXi get confused by the change and fail to reconnect automatically. The patch makes server scope, major owner and minor owner values configurable via sysctls. If not set (by default) the host UUID value is still used. Reviewed by: rmacklem MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D31952
2021-09-14 18:14:30 +00:00
static char nfsrv_scope[NFSV4_OPAQUELIMIT];
SYSCTL_STRING(_vfs_nfsd, OID_AUTO, scope, CTLFLAG_RWTUN,
&nfsrv_scope, NFSV4_OPAQUELIMIT, "Server scope");
static char nfsrv_owner_major[NFSV4_OPAQUELIMIT];
SYSCTL_STRING(_vfs_nfsd, OID_AUTO, owner_major, CTLFLAG_RWTUN,
&nfsrv_owner_major, NFSV4_OPAQUELIMIT, "Server owner major");
static uint64_t nfsrv_owner_minor;
SYSCTL_U64(_vfs_nfsd, OID_AUTO, owner_minor, CTLFLAG_RWTUN,
&nfsrv_owner_minor, 0, "Server owner minor");
nfsd: Disable the NFSv4.2 Allocate operation by default Some exported file systems, such as ZFS ones, cannot do VOP_ALLOCATE(). Since an NFSv4.2 server must either support the Allocate operation for all file systems or not support it at all, define a sysctl called vfs.nfsd.enable_v42allocate to enable the Allocate operation. This sysctl is false by default and can only be set true if all exported file systems (or all DSs for a pNFS server) can perform VOP_ALLOCATE(). Unfortunately, there is no way to know if a ZFS file system will be exported once the nfsd is operational, even if there are none exported when the nfsd is started up, so enabling Allocate must be done manually for a server configuration. This problem was detected during a recent NFSv4 interoperability testing event held by the IETF working group. MFC after: 2 weeks
2021-10-11 01:46:02 +00:00
/*
* Only enable this if all your exported file systems
* (or pNFS DSs for the pNFS case) support VOP_ALLOCATE.
*/
static bool nfsrv_doallocate = false;
SYSCTL_BOOL(_vfs_nfsd, OID_AUTO, enable_v42allocate, CTLFLAG_RW,
&nfsrv_doallocate, 0,
"Enable NFSv4.2 Allocate operation");
Make it possible to force async at server side on new NFS server, similar to the old one's nfs.nfsrv.async. Please note that by enabling this option (default is disabled), the system could potentionally have silent data corruption if the server crashes before write is committed to non-volatile storage, as the client side have no way to tell if the data is already written. Submitted by: rmacklem MFC after: 2 weeks
2013-01-18 19:42:08 +00:00
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* This list defines the GSS mechanisms supported.
* (Don't ask me how you get these strings from the RFC stuff like
* iso(1), org(3)... but someone did it, so I don't need to know.)
*/
static struct nfsgss_mechlist nfsgss_mechlist[] = {
{ 9, "\052\206\110\206\367\022\001\002\002", 11 },
{ 0, "", 0 },
};
/* local functions */
static void nfsrvd_symlinksub(struct nfsrv_descript *nd, struct nameidata *ndp,
struct nfsvattr *nvap, fhandle_t *fhp, vnode_t *vpp,
vnode_t dirp, struct nfsvattr *dirforp, struct nfsvattr *diraftp,
int *diraft_retp, nfsattrbit_t *attrbitp,
NFSACL_T *aclp, NFSPROC_T *p, struct nfsexstuff *exp, char *pathcp,
int pathlen);
static void nfsrvd_mkdirsub(struct nfsrv_descript *nd, struct nameidata *ndp,
struct nfsvattr *nvap, fhandle_t *fhp, vnode_t *vpp,
vnode_t dirp, struct nfsvattr *dirforp, struct nfsvattr *diraftp,
int *diraft_retp, nfsattrbit_t *attrbitp, NFSACL_T *aclp,
NFSPROC_T *p, struct nfsexstuff *exp);
/*
* nfs access service (not a part of NFS V2)
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_access(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int getret, error = 0;
struct nfsvattr nva;
u_int32_t testmode, nfsmode, supported = 0;
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
accmode_t deletebit;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, 1, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
nfsmode = fxdr_unsigned(u_int32_t, *tl);
if ((nd->nd_flag & ND_NFSV4) &&
(nfsmode & ~(NFSACCESS_READ | NFSACCESS_LOOKUP |
NFSACCESS_MODIFY | NFSACCESS_EXTEND | NFSACCESS_DELETE |
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
NFSACCESS_EXECUTE | NFSACCESS_XAREAD | NFSACCESS_XAWRITE |
NFSACCESS_XALIST))) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = NFSERR_INVAL;
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (nfsmode & NFSACCESS_READ) {
supported |= NFSACCESS_READ;
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
if (nfsvno_accchk(vp, VREAD, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmode &= ~NFSACCESS_READ;
}
if (nfsmode & NFSACCESS_MODIFY) {
supported |= NFSACCESS_MODIFY;
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
if (nfsvno_accchk(vp, VWRITE, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmode &= ~NFSACCESS_MODIFY;
}
if (nfsmode & NFSACCESS_EXTEND) {
supported |= NFSACCESS_EXTEND;
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
if (nfsvno_accchk(vp, VWRITE | VAPPEND, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmode &= ~NFSACCESS_EXTEND;
}
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (nfsmode & NFSACCESS_XAREAD) {
supported |= NFSACCESS_XAREAD;
if (nfsvno_accchk(vp, VREAD, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
nfsmode &= ~NFSACCESS_XAREAD;
}
if (nfsmode & NFSACCESS_XAWRITE) {
supported |= NFSACCESS_XAWRITE;
if (nfsvno_accchk(vp, VWRITE, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
nfsmode &= ~NFSACCESS_XAWRITE;
}
if (nfsmode & NFSACCESS_XALIST) {
supported |= NFSACCESS_XALIST;
if (nfsvno_accchk(vp, VREAD, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
nfsmode &= ~NFSACCESS_XALIST;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nfsmode & NFSACCESS_DELETE) {
supported |= NFSACCESS_DELETE;
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
if (vp->v_type == VDIR)
deletebit = VDELETE_CHILD;
else
deletebit = VDELETE;
if (nfsvno_accchk(vp, deletebit, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmode &= ~NFSACCESS_DELETE;
}
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (vp->v_type == VDIR)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
testmode = NFSACCESS_LOOKUP;
else
testmode = NFSACCESS_EXECUTE;
if (nfsmode & testmode) {
supported |= (nfsmode & testmode);
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
if (nfsvno_accchk(vp, VEXEC, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE, NFSACCCHK_VPISLOCKED, &supported))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmode &= ~testmode;
}
nfsmode &= supported;
if (nd->nd_flag & ND_NFSV3) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_postopattr(nd, getret, &nva);
}
vput(vp);
if (nd->nd_flag & ND_NFSV4) {
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(supported);
} else
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(nfsmode);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs getattr service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_getattr(struct nfsrv_descript *nd, int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr nva;
fhandle_t fh;
Fix the experimental NFSv4 server so that it uses VOP_PATHCONF() to determine if a file system supports NFSv4 ACLs. Since VOP_PATHCONF() must be called with a locked vnode, the function is called before nfsvno_fillattr() and the result is passed in as an extra argument. MFC after: 2 weeks
2011-04-14 23:46:15 +00:00
int at_root = 0, error = 0, supports_nfsv4acls;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct nfsreferral *refp;
Fix the new NFSv4 server so that it checks for VREAD_ACL when a client does a Getattr for an ACL and not VREAD_ATTRIBUTES. This was found during the recent NFSv4 interoperability Bakeathon. MFC after: 2 weeks
2011-06-21 19:58:29 +00:00
nfsattrbit_t attrbits, tmpbits;
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
struct mount *mp;
struct vnode *tvp = NULL;
struct vattr va;
uint64_t mounted_on_fileno = 0;
Fix the new NFSv4 server so that it checks for VREAD_ACL when a client does a Getattr for an ACL and not VREAD_ATTRIBUTES. This was found during the recent NFSv4 interoperability Bakeathon. MFC after: 2 weeks
2011-06-21 19:58:29 +00:00
accmode_t accmode;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV4) {
error = nfsrv_getattrbits(nd, &attrbits, NULL, NULL);
if (error) {
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Check for a referral.
*/
refp = nfsv4root_getreferral(vp, NULL, 0);
if (refp != NULL) {
(void) nfsrv_putreferralattr(nd, &attrbits, refp, 1,
&nd->nd_repstat);
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Fix the new NFSv4 server so that it checks for VREAD_ACL when a client does a Getattr for an ACL and not VREAD_ATTRIBUTES. This was found during the recent NFSv4 interoperability Bakeathon. MFC after: 2 weeks
2011-06-21 19:58:29 +00:00
if (nd->nd_repstat == 0) {
accmode = 0;
NFSSET_ATTRBIT(&tmpbits, &attrbits);
fs: clean up empty lines in .c and .h files
2020-09-01 21:18:40 +00:00
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
/*
* GETATTR with write-only attr time_access_set and time_modify_set
* should return NFS4ERR_INVAL.
*/
if (NFSISSET_ATTRBIT(&tmpbits, NFSATTRBIT_TIMEACCESSSET) ||
NFSISSET_ATTRBIT(&tmpbits, NFSATTRBIT_TIMEMODIFYSET)){
error = NFSERR_INVAL;
vput(vp);
goto out;
}
Fix the new NFSv4 server so that it checks for VREAD_ACL when a client does a Getattr for an ACL and not VREAD_ATTRIBUTES. This was found during the recent NFSv4 interoperability Bakeathon. MFC after: 2 weeks
2011-06-21 19:58:29 +00:00
if (NFSISSET_ATTRBIT(&tmpbits, NFSATTRBIT_ACL)) {
NFSCLRBIT_ATTRBIT(&tmpbits, NFSATTRBIT_ACL);
accmode |= VREAD_ACL;
}
if (NFSNONZERO_ATTRBIT(&tmpbits))
accmode |= VREAD_ATTRIBUTES;
if (accmode != 0)
nd->nd_repstat = nfsvno_accchk(vp, accmode,
nd->nd_cred, exp, p, NFSACCCHK_NOOVERRIDE,
NFSACCCHK_VPISLOCKED, NULL);
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1, &attrbits);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV4) {
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_FILEHANDLE))
nd->nd_repstat = nfsvno_getfh(vp, &fh, p);
if (!nd->nd_repstat)
nd->nd_repstat = nfsrv_checkgetattr(nd, vp,
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
&nva, &attrbits, p);
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
if (nd->nd_repstat == 0) {
Fix the experimental NFSv4 server so that it uses VOP_PATHCONF() to determine if a file system supports NFSv4 ACLs. Since VOP_PATHCONF() must be called with a locked vnode, the function is called before nfsvno_fillattr() and the result is passed in as an extra argument. MFC after: 2 weeks
2011-04-14 23:46:15 +00:00
supports_nfsv4acls = nfs_supportsnfsv4acls(vp);
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
mp = vp->v_mount;
if (nfsrv_enable_crossmntpt != 0 &&
vp->v_type == VDIR &&
(vp->v_vflag & VV_ROOT) != 0 &&
vp != rootvnode) {
tvp = mp->mnt_vnodecovered;
VREF(tvp);
at_root = 1;
} else
at_root = 0;
vfs_ref(mp);
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
if (at_root != 0) {
if ((nd->nd_repstat =
Simple find/replace of vn_lock -> NFSVOPLOCK. This is done so that NFSVOPLOCK can be modified later to add enhanced logging and assertions. Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:05:31 +00:00
NFSVOPLOCK(tvp, LK_SHARED)) == 0) {
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
nd->nd_repstat = VOP_GETATTR(
tvp, &va, nd->nd_cred);
vput(tvp);
} else
vrele(tvp);
if (nd->nd_repstat == 0)
mounted_on_fileno = (uint64_t)
va.va_fileid;
else
at_root = 0;
}
if (nd->nd_repstat == 0)
nd->nd_repstat = vfs_busy(mp, 0);
vfs_rel(mp);
if (nd->nd_repstat == 0) {
(void)nfsvno_fillattr(nd, mp, vp, &nva,
&fh, 0, &attrbits, nd->nd_cred, p,
Fix the experimental NFSv4 server so that it uses VOP_PATHCONF() to determine if a file system supports NFSv4 ACLs. Since VOP_PATHCONF() must be called with a locked vnode, the function is called before nfsvno_fillattr() and the result is passed in as an extra argument. MFC after: 2 weeks
2011-04-14 23:46:15 +00:00
isdgram, 1, supports_nfsv4acls,
at_root, mounted_on_fileno);
Modify the experimental NFSv4 server so that it handles crossing of server mount points properly. The functions nfsvno_fillattr() and nfsv4_fillattr() were modified to take the extra arguments that are the mount point, a flag to indicate that it is a file system root and the mounted on fileno. The mount point argument needs to be busy when nfsvno_fillattr() is called, since the vp argument is not locked. Reviewed by: kib MFC after: 2 weeks
2011-04-14 21:49:52 +00:00
vfs_unbusy(mp);
}
vrele(vp);
} else
vput(vp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
nfsrv_fillattr(nd, &nva);
vput(vp);
}
} else {
vput(vp);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs setattr service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_setattr(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr nva, nva2;
u_int32_t *tl;
int preat_ret = 1, postat_ret = 1, gcheck = 0, error = 0;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
int gotproxystateid;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct timespec guard = { 0, 0 };
nfsattrbit_t attrbits, retbits;
nfsv4stateid_t stateid;
NFSACL_T *aclp = NULL;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, preat_ret, &nva2, postat_ret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
#ifdef NFS4_ACL_EXTATTR_NAME
Fix the experimental nfs subsystem so that it builds with the current NFSv4 ACLs, as defined in sys/acl.h. It still needs a way to test a mount point for NFSv4 ACL support before it will work. Until then, the NFSHASNFS4ACL() macro just always returns 0. Approved by: kib (mentor)
2009-05-26 22:21:53 +00:00
aclp = acl_alloc(M_WAITOK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
aclp->acl_cnt = 0;
#endif
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
gotproxystateid = 0;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSVNO_ATTRINIT(&nva);
if (nd->nd_flag & ND_NFSV4) {
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID);
stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
stateid.other[0] = *tl++;
stateid.other[1] = *tl++;
stateid.other[2] = *tl;
if (stateid.other[0] == 0x55555555 &&
stateid.other[1] == 0x55555555 &&
stateid.other[2] == 0x55555555 &&
stateid.seqid == 0xffffffff)
gotproxystateid = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
error = nfsrv_sattr(nd, vp, &nva, &attrbits, aclp, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (error)
goto nfsmout;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/* For NFSv4, only va_uid is used from nva2. */
NFSZERO_ATTRBIT(&retbits);
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_OWNER);
preat_ret = nfsvno_getattr(vp, &nva2, nd, p, 1, &retbits);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
nd->nd_repstat = preat_ret;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
NFSZERO_ATTRBIT(&retbits);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3) {
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
gcheck = fxdr_unsigned(int, *tl);
if (gcheck) {
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
fxdr_nfsv3time(tl, &guard);
}
if (!nd->nd_repstat && gcheck &&
(nva2.na_ctime.tv_sec != guard.tv_sec ||
nva2.na_ctime.tv_nsec != guard.tv_nsec))
nd->nd_repstat = NFSERR_NOT_SYNC;
if (nd->nd_repstat) {
vput(vp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
nfsrv_wcc(nd, preat_ret, &nva2, postat_ret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
} else if (!nd->nd_repstat && (nd->nd_flag & ND_NFSV4))
nd->nd_repstat = nfsrv_checkuidgid(nd, &nva);
/*
* Now that we have all the fields, lets do it.
* If the size is being changed write access is required, otherwise
* just check for a read only file system.
*/
if (!nd->nd_repstat) {
if (NFSVNO_NOTSETSIZE(&nva)) {
if (NFSVNO_EXRDONLY(exp) ||
nfscl: Clean up the code by removing vfs_flags() macro The vfs_flags() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so remove it to clean up the code. This commit should not result in a semantics change.
2022-06-20 20:23:04 +00:00
(vp->v_mount->mnt_flag & MNT_RDONLY))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = EROFS;
} else {
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (vp->v_type != VREG)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = EINVAL;
else if (nva2.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp))
nd->nd_repstat = nfsvno_accchk(vp,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
VWRITE, nd->nd_cred, exp, p,
NFSACCCHK_NOOVERRIDE,
NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* Proxy operations from the MDS are allowed via the all 0s special
* stateid.
*/
if (nd->nd_repstat == 0 && (nd->nd_flag & ND_NFSV4) != 0 &&
gotproxystateid == 0)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsrv_checksetattr(vp, nd, &stateid,
&nva, &attrbits, exp, p);
if (!nd->nd_repstat && (nd->nd_flag & ND_NFSV4)) {
/*
nfs: Fix common typos in source code comments - s/attrbute/attribute/ MFC after: 3 days
2022-11-08 13:58:32 +00:00
* For V4, try setting the attributes in sets, so that the
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
* reply bitmap will be correct for an error case.
*/
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_OWNER) ||
NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_OWNERGROUP)) {
NFSVNO_ATTRINIT(&nva2);
NFSVNO_SETATTRVAL(&nva2, uid, nva.na_uid);
NFSVNO_SETATTRVAL(&nva2, gid, nva.na_gid);
nd->nd_repstat = nfsvno_setattr(vp, &nva2, nd->nd_cred, p,
exp);
if (!nd->nd_repstat) {
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_OWNER))
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_OWNER);
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_OWNERGROUP))
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_OWNERGROUP);
}
}
if (!nd->nd_repstat &&
NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_SIZE)) {
NFSVNO_ATTRINIT(&nva2);
NFSVNO_SETATTRVAL(&nva2, size, nva.na_size);
nd->nd_repstat = nfsvno_setattr(vp, &nva2, nd->nd_cred, p,
exp);
if (!nd->nd_repstat)
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_SIZE);
}
if (!nd->nd_repstat &&
(NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_TIMEACCESSSET) ||
NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_TIMEMODIFYSET))) {
NFSVNO_ATTRINIT(&nva2);
NFSVNO_SETATTRVAL(&nva2, atime, nva.na_atime);
NFSVNO_SETATTRVAL(&nva2, mtime, nva.na_mtime);
if (nva.na_vaflags & VA_UTIMES_NULL) {
nva2.na_vaflags |= VA_UTIMES_NULL;
NFSVNO_SETACTIVE(&nva2, vaflags);
}
nd->nd_repstat = nfsvno_setattr(vp, &nva2, nd->nd_cred, p,
exp);
if (!nd->nd_repstat) {
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_TIMEACCESSSET))
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_TIMEACCESSSET);
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_TIMEMODIFYSET))
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_TIMEMODIFYSET);
}
}
nfscl: Add support for va_birthtime to NFSv4 There is a NFSv4 file attribute called TimeCreate that can be used for va_birthtime. r362175 added some support for use of TimeCreate. This patch completes support of va_birthtime by adding support for setting this attribute to the server. It also eanbles the client to acquire and set the attribute for a NFSv4 server that supports the attribute. Reviewed by: markj MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D30156
2021-05-08 00:30:56 +00:00
if (!nd->nd_repstat &&
NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_TIMECREATE)) {
NFSVNO_ATTRINIT(&nva2);
NFSVNO_SETATTRVAL(&nva2, btime, nva.na_btime);
nd->nd_repstat = nfsvno_setattr(vp, &nva2, nd->nd_cred, p,
exp);
if (!nd->nd_repstat)
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_TIMECREATE);
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat &&
Add support for the ModeSetMasked attribute to the NFSv4.1 server. I do not know of an extant NFSv4.1 client that currently does a Setattr operation for the ModeSetMasked, but it has been discussed on the linux-nfs mailing list. This patch adds support for doing a Setattr of ModeSetMasked, so that it will work for any future NFSv4.1 client that chooses to do so. Tested via a hacked FreeBSD NFSv4.1 client. MFC after: 2 weeks
2019-04-19 23:35:08 +00:00
(NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_MODE) ||
NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_MODESETMASKED))) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSVNO_ATTRINIT(&nva2);
NFSVNO_SETATTRVAL(&nva2, mode, nva.na_mode);
nd->nd_repstat = nfsvno_setattr(vp, &nva2, nd->nd_cred, p,
exp);
Add support for the ModeSetMasked attribute to the NFSv4.1 server. I do not know of an extant NFSv4.1 client that currently does a Setattr operation for the ModeSetMasked, but it has been discussed on the linux-nfs mailing list. This patch adds support for doing a Setattr of ModeSetMasked, so that it will work for any future NFSv4.1 client that chooses to do so. Tested via a hacked FreeBSD NFSv4.1 client. MFC after: 2 weeks
2019-04-19 23:35:08 +00:00
if (!nd->nd_repstat) {
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_MODE))
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_MODE);
if (NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_MODESETMASKED))
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_MODESETMASKED);
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
#ifdef NFS4_ACL_EXTATTR_NAME
if (!nd->nd_repstat && aclp->acl_cnt > 0 &&
NFSISSET_ATTRBIT(&attrbits, NFSATTRBIT_ACL)) {
nd->nd_repstat = nfsrv_setacl(vp, aclp, nd->nd_cred, p);
if (!nd->nd_repstat)
NFSSETBIT_ATTRBIT(&retbits, NFSATTRBIT_ACL);
}
#endif
} else if (!nd->nd_repstat) {
nd->nd_repstat = nfsvno_setattr(vp, &nva, nd->nd_cred, p,
exp);
}
if (nd->nd_flag & (ND_NFSV2 | ND_NFSV3)) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
postat_ret = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
nd->nd_repstat = postat_ret;
}
vput(vp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, preat_ret, &nva2, postat_ret, &nva);
else if (nd->nd_flag & ND_NFSV4)
(void) nfsrv_putattrbit(nd, &retbits);
else if (!nd->nd_repstat)
nfsrv_fillattr(nd, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
if (nd->nd_flag & ND_NFSV4) {
/*
* For all nd_repstat, the V4 reply includes a bitmap,
* even NFSERR_BADXDR, which is what this will end up
* returning.
*/
(void) nfsrv_putattrbit(nd, &retbits);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs lookup rpc
* (Also performs lookup parent for v4)
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_lookup(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, vnode_t *vpp, fhandle_t *fhp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nameidata named;
vnode_t vp, dirp = NULL;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
int error = 0, dattr_ret = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct nfsvattr nva, dattr;
char *bufp;
u_long *hashp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, dattr_ret, &dattr);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* For some reason, if dp is a symlink, the error
* returned is supposed to be NFSERR_SYMLINK and not NFSERR_NOTDIR.
*/
if (dp->v_type == VLNK && (nd->nd_flag & ND_NFSV4)) {
nd->nd_repstat = NFSERR_SYMLINK;
vrele(dp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, LOOKUP,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKLEAF);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (error) {
vrele(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat) {
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 0, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vrele(dp);
nfsvno_relpathbuf(&named);
}
if (nd->nd_repstat) {
if (dirp) {
if (nd->nd_flag & ND_NFSV3)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dattr_ret = nfsvno_getattr(dirp, &dattr, nd, p,
0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if (nd->nd_flag & ND_NFSV3)
nfsrv_postopattr(nd, dattr_ret, &dattr);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfsvno_relpathbuf(&named);
vp = named.ni_vp;
Fix the new NFSv4 server so that it only allows Lookup of directories and symbolic links when traversing non-exported file systems. Found during the recent NFSv4 interoperability Bakeathon. MFC after: 2 weeks
2011-06-20 22:02:01 +00:00
if ((nd->nd_flag & ND_NFSV4) != 0 && !NFSVNO_EXPORTED(exp) &&
vp->v_type != VDIR && vp->v_type != VLNK)
/*
* Only allow lookup of VDIR and VLNK for traversal of
* non-exported volumes during NFSv4 mounting.
*/
nd->nd_repstat = ENOENT;
nfsd: Handle file systems without a VOP_VPTOFH() Unlike NFSv3, the NFSv4 server follows mount points within the file system tree below the NFSv4 root directory. If there is a file system mounted within this subtree that returns EOPNOTSUPP for VOP_VPTOFH(), the NFSv4 server would return an error for the mount point entry. This resulted in an "I/O error" report from the Linux NFSv4 client. It also put an error code in the Readdir reply that is not defined in the NFSv4 RFCs. For the FreeBSD NFSv4 client, the entry with the error would be ignored, which I think is reasonable behaviour for a mounted file system that can never be exported. This patch changes the NFSv4 server behaviour to ignore the mount point entry and not send it in the Readdir reply. It also changes the behaviour of Lookup for the entry so that it replies ENOENT for the mount point directory, so that it is consistent with no entry in the Readdir reply. With these two changes, the Linux client behaviour is the same as the FreeBSD client behaviour. It also avoids putting an unknown error on the wire to the client. MFC after: 1 week
2022-12-23 23:17:34 +00:00
if (nd->nd_repstat == 0) {
Fix the new NFSv4 server so that it only allows Lookup of directories and symbolic links when traversing non-exported file systems. Found during the recent NFSv4 interoperability Bakeathon. MFC after: 2 weeks
2011-06-20 22:02:01 +00:00
nd->nd_repstat = nfsvno_getfh(vp, fhp, p);
nfsd: Handle file systems without a VOP_VPTOFH() Unlike NFSv3, the NFSv4 server follows mount points within the file system tree below the NFSv4 root directory. If there is a file system mounted within this subtree that returns EOPNOTSUPP for VOP_VPTOFH(), the NFSv4 server would return an error for the mount point entry. This resulted in an "I/O error" report from the Linux NFSv4 client. It also put an error code in the Readdir reply that is not defined in the NFSv4 RFCs. For the FreeBSD NFSv4 client, the entry with the error would be ignored, which I think is reasonable behaviour for a mounted file system that can never be exported. This patch changes the NFSv4 server behaviour to ignore the mount point entry and not send it in the Readdir reply. It also changes the behaviour of Lookup for the entry so that it replies ENOENT for the mount point directory, so that it is consistent with no entry in the Readdir reply. With these two changes, the Linux client behaviour is the same as the FreeBSD client behaviour. It also avoids putting an unknown error on the wire to the client. MFC after: 1 week
2022-12-23 23:17:34 +00:00
/*
* EOPNOTSUPP indicates the file system cannot be exported,
* so just pretend the entry does not exist.
*/
if (nd->nd_repstat == EOPNOTSUPP)
nd->nd_repstat = ENOENT;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!(nd->nd_flag & ND_NFSV4) && !nd->nd_repstat)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Modify the experimental NFSv4 server so that the lookup ops return a locked vnode. This ensures that the associated mount point will always be valid for the code that follows the operation. Also add a couple of additional checks for non-error to the other functions that create file objects. MFC after: 2 weeks
2011-01-03 00:33:32 +00:00
if (vpp != NULL && nd->nd_repstat == 0)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*vpp = vp;
Modify the experimental NFSv4 server so that the lookup ops return a locked vnode. This ensures that the associated mount point will always be valid for the code that follows the operation. Also add a couple of additional checks for non-error to the other functions that create file objects. MFC after: 2 weeks
2011-01-03 00:33:32 +00:00
else
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
if (dirp) {
if (nd->nd_flag & ND_NFSV3)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dattr_ret = nfsvno_getattr(dirp, &dattr, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if (nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV3)
nfsrv_postopattr(nd, dattr_ret, &dattr);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (nd->nd_flag & ND_NFSV2) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)fhp, 0, 0);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_fillattr(nd, &nva);
} else if (nd->nd_flag & ND_NFSV3) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)fhp, 0, 0);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_postopattr(nd, 0, &nva);
nfsrv_postopattr(nd, dattr_ret, &dattr);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs readlink service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_readlink(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
Replace all instances of the typedef mbuf_t with "struct mbuf *". The typedef mbuf_t was used for the Mac OS/X port of the code long ago. Since this port is no longer used and the use of mbuf_t obscures what the code does (and is not consistent with style(9)), it is no longer needed. This patch replaces all instances of mbuf_t with "struct mbuf *", so that it is no longer used. This patch should not result in any semantic change.
2020-04-17 21:17:51 +00:00
struct mbuf *mp = NULL, *mpend = NULL;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
int getret = 1, len;
struct nfsvattr nva;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
uint16_t off;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, getret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (vp->v_type != VLNK) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV2)
nd->nd_repstat = ENXIO;
else
nd->nd_repstat = EINVAL;
}
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
if (nd->nd_repstat == 0) {
if ((nd->nd_flag & ND_EXTPG) != 0)
nd->nd_repstat = nfsvno_readlink(vp, nd->nd_cred,
nd->nd_maxextsiz, p, &mp, &mpend, &len);
else
nd->nd_repstat = nfsvno_readlink(vp, nd->nd_cred,
0, p, &mp, &mpend, &len);
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
if (nd->nd_flag & ND_NFSV3)
nfsrv_postopattr(nd, getret, &nva);
if (nd->nd_repstat)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(len);
Add support for ext_pgs mbufs to nfsrv_adj(). This patch uses a slightly different algorithm for nfsrv_adj() since ext_pgs mbuf lists are not permitted to have m_len == 0 mbufs. As such, the code now frees mbufs after the adjustment in the list instead of setting their m_len field to 0. Since mbuf(s) may be trimmed off the tail of the list, the function now returns a pointer to the last mbuf in the list. This saves the caller from needing to use m_last() to find the last mbuf. It also implies that it might return a nul list, which required a check for that in nfsrvd_readlink(). This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-26 02:42:09 +00:00
if (mp != NULL) {
nd->nd_mb->m_next = mp;
nd->nd_mb = mpend;
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
if ((mpend->m_flags & M_EXTPG) != 0) {
nd->nd_bextpg = mpend->m_epg_npgs - 1;
nd->nd_bpos = (char *)(void *)
PHYS_TO_DMAP(mpend->m_epg_pa[nd->nd_bextpg]);
off = (nd->nd_bextpg == 0) ? mpend->m_epg_1st_off : 0;
nd->nd_bpos += off + mpend->m_epg_last_len;
nd->nd_bextpgsiz = PAGE_SIZE - mpend->m_epg_last_len -
off;
} else
nd->nd_bpos = mtod(mpend, char *) + mpend->m_len;
Add support for ext_pgs mbufs to nfsrv_adj(). This patch uses a slightly different algorithm for nfsrv_adj() since ext_pgs mbuf lists are not permitted to have m_len == 0 mbufs. As such, the code now frees mbufs after the adjustment in the list instead of setting their m_len field to 0. Since mbuf(s) may be trimmed off the tail of the list, the function now returns a pointer to the last mbuf in the list. This saves the caller from needing to use m_last() to find the last mbuf. It also implies that it might return a nul list, which required a check for that in nfsrvd_readlink(). This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-26 02:42:09 +00:00
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
}
/*
* nfs read service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_read(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
int error = 0, cnt, getret = 1, gotproxystateid, reqlen, eof = 0;
Replace all instances of the typedef mbuf_t with "struct mbuf *". The typedef mbuf_t was used for the Mac OS/X port of the code long ago. Since this port is no longer used and the use of mbuf_t obscures what the code does (and is not consistent with style(9)), it is no longer needed. This patch replaces all instances of mbuf_t with "struct mbuf *", so that it is no longer used. This patch should not result in any semantic change.
2020-04-17 21:17:51 +00:00
struct mbuf *m2, *m3;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct nfsvattr nva;
off_t off = 0x0;
struct nfsstate st, *stp = &st;
struct nfslock lo, *lop = &lo;
nfsv4stateid_t stateid;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
uint16_t poff;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, getret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (nd->nd_flag & ND_NFSV2) {
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
off = (off_t)fxdr_unsigned(u_int32_t, *tl++);
reqlen = fxdr_unsigned(int, *tl);
} else if (nd->nd_flag & ND_NFSV3) {
NFSM_DISSECT(tl, u_int32_t *, 3 * NFSX_UNSIGNED);
off = fxdr_hyper(tl);
tl += 2;
reqlen = fxdr_unsigned(int, *tl);
} else {
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID + 3*NFSX_UNSIGNED);
reqlen = fxdr_unsigned(int, *(tl + 6));
}
if (reqlen > NFS_SRVMAXDATA(nd)) {
reqlen = NFS_SRVMAXDATA(nd);
} else if (reqlen < 0) {
error = EBADRPC;
goto nfsmout;
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
gotproxystateid = 0;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV4) {
stp->ls_flags = (NFSLCK_CHECK | NFSLCK_READACCESS);
lop->lo_flags = NFSLCK_READ;
stp->ls_ownerlen = 0;
stp->ls_op = NULL;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
clientid.lval[0] = stp->ls_stateid.other[0] = *tl++;
clientid.lval[1] = stp->ls_stateid.other[1] = *tl++;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK1 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
stp->ls_stateid.other[2] = *tl++;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* Don't allow the client to use a special stateid for a DS op.
*/
if ((nd->nd_flag & ND_DSSERVER) != 0 &&
((stp->ls_stateid.other[0] == 0x0 &&
stp->ls_stateid.other[1] == 0x0 &&
stp->ls_stateid.other[2] == 0x0) ||
(stp->ls_stateid.other[0] == 0xffffffff &&
stp->ls_stateid.other[1] == 0xffffffff &&
stp->ls_stateid.other[2] == 0xffffffff) ||
stp->ls_stateid.seqid != 0))
nd->nd_repstat = NFSERR_BADSTATEID;
/* However, allow the proxy stateid. */
if (stp->ls_stateid.seqid == 0xffffffff &&
stp->ls_stateid.other[0] == 0x55555555 &&
stp->ls_stateid.other[1] == 0x55555555 &&
stp->ls_stateid.other[2] == 0x55555555)
gotproxystateid = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
off = fxdr_hyper(tl);
lop->lo_first = off;
tl += 2;
lop->lo_end = off + reqlen;
/*
* Paranoia, just in case it wraps around.
*/
if (lop->lo_end < off)
lop->lo_end = NFS64BITSSET;
}
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (vp->v_type != VREG) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3)
nd->nd_repstat = EINVAL;
else
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
nd->nd_repstat = (vp->v_type == VDIR) ? EISDIR :
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
EINVAL;
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
nd->nd_repstat = getret;
if (!nd->nd_repstat &&
(nva.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp))) {
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VREAD,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_cred, exp, p,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat)
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VEXEC,
nd->nd_cred, exp, p, NFSACCCHK_ALLOWOWNER,
NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* DS reads are marked by ND_DSSERVER or use the proxy special
* stateid.
*/
if (nd->nd_repstat == 0 && (nd->nd_flag & (ND_NFSV4 | ND_DSSERVER)) ==
ND_NFSV4 && gotproxystateid == 0)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, NULL, clientid,
&stateid, exp, nd, p);
if (nd->nd_repstat) {
vput(vp);
if (nd->nd_flag & ND_NFSV3)
nfsrv_postopattr(nd, getret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (off >= nva.na_size) {
cnt = 0;
eof = 1;
} else if (reqlen == 0)
cnt = 0;
Fix an NFS server issue where it was not correctly setting the eof flag when a READ had hit the end of the file. Also, clean up some cruft in the code. Approved by: re (kib) Reviewed by: rmacklem MFC after: 2 weeks
2011-08-03 18:50:19 +00:00
else if ((off + reqlen) >= nva.na_size) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
cnt = nva.na_size - off;
Fix an NFS server issue where it was not correctly setting the eof flag when a READ had hit the end of the file. Also, clean up some cruft in the code. Approved by: re (kib) Reviewed by: rmacklem MFC after: 2 weeks
2011-08-03 18:50:19 +00:00
eof = 1;
} else
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
cnt = reqlen;
m3 = NULL;
if (cnt > 0) {
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
/*
* If cnt > MCLBYTES and the reply will not be saved, use
* ext_pgs mbufs for TLS.
* For NFSv4.0, we do not know for sure if the reply will
* be saved, so do not use ext_pgs mbufs for NFSv4.0.
* Always use ext_pgs mbufs if ND_EXTPG is set.
*/
if ((nd->nd_flag & ND_EXTPG) != 0 || (cnt > MCLBYTES &&
(nd->nd_flag & (ND_TLS | ND_SAVEREPLY)) == ND_TLS &&
(nd->nd_flag & (ND_NFSV4 | ND_NFSV41)) != ND_NFSV4))
nd->nd_repstat = nfsvno_read(vp, off, cnt, nd->nd_cred,
nd->nd_maxextsiz, p, &m3, &m2);
else
nd->nd_repstat = nfsvno_read(vp, off, cnt, nd->nd_cred,
0, p, &m3, &m2);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!(nd->nd_flag & ND_NFSV4)) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
nd->nd_repstat = getret;
}
if (nd->nd_repstat) {
vput(vp);
if (m3)
Replace mbuf macros with the code they would generate in the NFS code. When the code was ported to Mac OS/X, mbuf handling functions were converted to using the Mac OS/X accessor functions. For FreeBSD, they are a simple set of macros in sys/fs/nfs/nfskpiport.h. Since porting to Mac OS/X is no longer a consideration, replacement of these macros with the code generated by them makes the code more readable. When support for external page mbufs is added as needed by the KERN_TLS, the patch becomes simpler if done without the macros. This patch should not result in any semantic change.
2020-04-11 20:57:15 +00:00
m_freem(m3);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3)
nfsrv_postopattr(nd, getret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
vput(vp);
if (nd->nd_flag & ND_NFSV2) {
nfsrv_fillattr(nd, &nva);
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
} else {
if (nd->nd_flag & ND_NFSV3) {
nfsrv_postopattr(nd, getret, &nva);
NFSM_BUILD(tl, u_int32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(cnt);
} else
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
Fix an NFS server issue where it was not correctly setting the eof flag when a READ had hit the end of the file. Also, clean up some cruft in the code. Approved by: re (kib) Reviewed by: rmacklem MFC after: 2 weeks
2011-08-03 18:50:19 +00:00
if (eof)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*tl++ = newnfs_true;
else
*tl++ = newnfs_false;
}
*tl = txdr_unsigned(cnt);
if (m3) {
Replace mbuf macros with the code they would generate in the NFS code. When the code was ported to Mac OS/X, mbuf handling functions were converted to using the Mac OS/X accessor functions. For FreeBSD, they are a simple set of macros in sys/fs/nfs/nfskpiport.h. Since porting to Mac OS/X is no longer a consideration, replacement of these macros with the code generated by them makes the code more readable. When support for external page mbufs is added as needed by the KERN_TLS, the patch becomes simpler if done without the macros. This patch should not result in any semantic change.
2020-04-11 20:57:15 +00:00
nd->nd_mb->m_next = m3;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_mb = m2;
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
if ((m2->m_flags & M_EXTPG) != 0) {
nd->nd_flag |= ND_EXTPG;
nd->nd_bextpg = m2->m_epg_npgs - 1;
nd->nd_bpos = (char *)(void *)
PHYS_TO_DMAP(m2->m_epg_pa[nd->nd_bextpg]);
poff = (nd->nd_bextpg == 0) ? m2->m_epg_1st_off : 0;
nd->nd_bpos += poff + m2->m_epg_last_len;
nd->nd_bextpgsiz = PAGE_SIZE - m2->m_epg_last_len -
poff;
} else
nd->nd_bpos = mtod(m2, char *) + m2->m_len;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs write service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_write(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
struct nfsvattr nva, forat;
int aftat_ret = 1, retlen, len, error = 0, forat_ret = 1;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
int gotproxystateid, stable = NFSWRITE_FILESYNC;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
off_t off;
struct nfsstate st, *stp = &st;
struct nfslock lo, *lop = &lo;
nfsv4stateid_t stateid;
nfsquad_t clientid;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nfsattrbit_t attrbits;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, forat_ret, &forat, aftat_ret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
gotproxystateid = 0;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV2) {
NFSM_DISSECT(tl, u_int32_t *, 4 * NFSX_UNSIGNED);
off = (off_t)fxdr_unsigned(u_int32_t, *++tl);
tl += 2;
retlen = len = fxdr_unsigned(int32_t, *tl);
} else if (nd->nd_flag & ND_NFSV3) {
NFSM_DISSECT(tl, u_int32_t *, 5 * NFSX_UNSIGNED);
off = fxdr_hyper(tl);
tl += 3;
stable = fxdr_unsigned(int, *tl++);
retlen = len = fxdr_unsigned(int32_t, *tl);
} else {
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID + 4 * NFSX_UNSIGNED);
stp->ls_flags = (NFSLCK_CHECK | NFSLCK_WRITEACCESS);
lop->lo_flags = NFSLCK_WRITE;
stp->ls_ownerlen = 0;
stp->ls_op = NULL;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
clientid.lval[0] = stp->ls_stateid.other[0] = *tl++;
clientid.lval[1] = stp->ls_stateid.other[1] = *tl++;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK2 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
stp->ls_stateid.other[2] = *tl++;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* Don't allow the client to use a special stateid for a DS op.
*/
if ((nd->nd_flag & ND_DSSERVER) != 0 &&
((stp->ls_stateid.other[0] == 0x0 &&
stp->ls_stateid.other[1] == 0x0 &&
stp->ls_stateid.other[2] == 0x0) ||
(stp->ls_stateid.other[0] == 0xffffffff &&
stp->ls_stateid.other[1] == 0xffffffff &&
stp->ls_stateid.other[2] == 0xffffffff) ||
stp->ls_stateid.seqid != 0))
nd->nd_repstat = NFSERR_BADSTATEID;
/* However, allow the proxy stateid. */
if (stp->ls_stateid.seqid == 0xffffffff &&
stp->ls_stateid.other[0] == 0x55555555 &&
stp->ls_stateid.other[1] == 0x55555555 &&
stp->ls_stateid.other[2] == 0x55555555)
gotproxystateid = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
off = fxdr_hyper(tl);
lop->lo_first = off;
tl += 2;
stable = fxdr_unsigned(int, *tl++);
retlen = len = fxdr_unsigned(int32_t, *tl);
lop->lo_end = off + len;
/*
* Paranoia, just in case it wraps around, which shouldn't
* ever happen anyhow.
*/
if (lop->lo_end < lop->lo_first)
lop->lo_end = NFS64BITSSET;
}
nfsd: Add sysctl to set maximum I/O size up to 1Mbyte Since MAXPHYS now allows the FreeBSD NFS client to do 1Mbyte I/O operations, add a sysctl called vfs.nfsd.srvmaxio so that the maximum NFS server I/O size can be set up to 1Mbyte. The Linux NFS client can also do 1Mbyte I/O operations. The default of 128Kbytes for the maximum I/O size has not been changed for two reasons: - kern.ipc.maxsockbuf must be increased to support 1Mbyte I/O - The limited benchmarking I can do actually shows a drop in I/O rate when the I/O size is above 256Kbytes. However, daveb@spectralogic.com reports seeing an increase in I/O rate for the 1Mbyte I/O size vs 128Kbytes using a Linux client. Reviewed by: asomers MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D30826
2021-07-16 22:01:03 +00:00
if (retlen > nfs_srvmaxio || retlen < 0)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = EIO;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (vp->v_type != VREG && !nd->nd_repstat) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3)
nd->nd_repstat = EINVAL;
else
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
nd->nd_repstat = (vp->v_type == VDIR) ? EISDIR :
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
EINVAL;
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_OWNER);
forat_ret = nfsvno_getattr(vp, &forat, nd, p, 1, &attrbits);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
nd->nd_repstat = forat_ret;
if (!nd->nd_repstat &&
(forat.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp)))
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VWRITE,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_cred, exp, p,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED, NULL);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* DS reads are marked by ND_DSSERVER or use the proxy special
* stateid.
*/
if (nd->nd_repstat == 0 && (nd->nd_flag & (ND_NFSV4 | ND_DSSERVER)) ==
ND_NFSV4 && gotproxystateid == 0)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, NULL, clientid,
&stateid, exp, nd, p);
if (nd->nd_repstat) {
vput(vp);
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, forat_ret, &forat, aftat_ret, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* For NFS Version 2, it is not obvious what a write of zero length
* should do, but I might as well be consistent with Version 3,
* which is to return ok so long as there are no permission problems.
*/
if (retlen > 0) {
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = nfsvno_write(vp, off, retlen, &stable,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_md, nd->nd_dpos, nd->nd_cred, p);
error = nfsm_advance(nd, NFSM_RNDUP(retlen), -1);
if (error)
Change a panic to an error return. There was a panic() in the NFS server's write operation that didn't need to be a panic() and could just be an error return. This patch makes that change. Found by code inspection during development of the pNFS service. MFC after: 2 weeks
2017-09-24 20:05:48 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (nd->nd_flag & ND_NFSV4)
aftat_ret = 0;
else
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
aftat_ret = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
if (!nd->nd_repstat)
nd->nd_repstat = aftat_ret;
if (nd->nd_flag & (ND_NFSV3 | ND_NFSV4)) {
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, forat_ret, &forat, aftat_ret, &nva);
if (nd->nd_repstat)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_BUILD(tl, u_int32_t *, 4 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(retlen);
Make it possible to force async at server side on new NFS server, similar to the old one's nfs.nfsrv.async. Please note that by enabling this option (default is disabled), the system could potentionally have silent data corruption if the server crashes before write is committed to non-volatile storage, as the client side have no way to tell if the data is already written. Submitted by: rmacklem MFC after: 2 weeks
2013-01-18 19:42:08 +00:00
/*
* If nfs_async is set, then pretend the write was FILESYNC.
* Warning: Doing this violates RFC1813 and runs a risk
* of data written by a client being lost when the server
* crashes/reboots.
*/
if (stable == NFSWRITE_UNSTABLE && nfs_async == 0)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*tl++ = txdr_unsigned(stable);
else
*tl++ = txdr_unsigned(NFSWRITE_FILESYNC);
/*
* Actually, there is no need to txdr these fields,
* but it may make the values more human readable,
* for debugging purposes.
*/
*tl++ = txdr_unsigned(nfsboottime.tv_sec);
*tl = txdr_unsigned(nfsboottime.tv_usec);
} else if (!nd->nd_repstat)
nfsrv_fillattr(nd, &nva);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs create service (creates regular files for V2 and V3. Spec. files for V2.)
* now does a truncate to 0 length via. setattr if it already exists
* The core creation routine has been extracted out into nfsrv_creatsub(),
* so it can also be used by nfsrv_open() for V4.
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_create(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr nva, dirfor, diraft;
struct nfsv2_sattr *sp;
struct nameidata named;
u_int32_t *tl;
int error = 0, tsize, dirfor_ret = 1, diraft_ret = 1;
int how = NFSCREATE_UNCHECKED, exclusive_flag = 0;
NFSDEV_T rdev = 0;
vnode_t vp = NULL, dirp = NULL;
fhandle_t fh;
char *bufp;
u_long *hashp;
enum vtype vtyp;
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
int32_t cverf[2], tverf[2] = { 0, 0 };
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, CREATE,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKPARENT | LOCKLEAF | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
NFSVNO_ATTRINIT(&nva);
if (nd->nd_flag & ND_NFSV2) {
NFSM_DISSECT(sp, struct nfsv2_sattr *, NFSX_V2SATTR);
vtyp = IFTOVT(fxdr_unsigned(u_int32_t, sp->sa_mode));
if (vtyp == VNON)
vtyp = VREG;
NFSVNO_SETATTRVAL(&nva, type, vtyp);
NFSVNO_SETATTRVAL(&nva, mode,
nfstov_mode(sp->sa_mode));
switch (nva.na_type) {
case VREG:
tsize = fxdr_unsigned(int32_t, sp->sa_size);
if (tsize != -1)
NFSVNO_SETATTRVAL(&nva, size,
(u_quad_t)tsize);
break;
case VCHR:
case VBLK:
case VFIFO:
rdev = fxdr_unsigned(NFSDEV_T, sp->sa_size);
break;
default:
break;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
how = fxdr_unsigned(int, *tl);
switch (how) {
case NFSCREATE_GUARDED:
case NFSCREATE_UNCHECKED:
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
error = nfsrv_sattr(nd, NULL, &nva, NULL, NULL, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (error)
goto nfsmout;
break;
case NFSCREATE_EXCLUSIVE:
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_VERF);
cverf[0] = *tl++;
cverf[1] = *tl;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
exclusive_flag = 1;
break;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSVNO_SETATTRVAL(&nva, type, VREG);
}
}
if (nd->nd_repstat) {
nfsvno_relpathbuf(&named);
if (nd->nd_flag & ND_NFSV3) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dp, &dirfor, nd, p, 1,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret,
&diraft);
}
vput(dp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 1, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (dirp) {
if (nd->nd_flag & ND_NFSV2) {
vrele(dirp);
dirp = NULL;
} else {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
if (nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret,
&diraft);
if (dirp)
vrele(dirp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!(nd->nd_flag & ND_NFSV2)) {
switch (how) {
case NFSCREATE_GUARDED:
if (named.ni_vp)
nd->nd_repstat = EEXIST;
break;
case NFSCREATE_UNCHECKED:
break;
case NFSCREATE_EXCLUSIVE:
if (named.ni_vp == NULL)
NFSVNO_SETATTRVAL(&nva, mode, 0);
break;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Iff doesn't exist, create it
* otherwise just truncate to 0 length
* should I set the mode too ?
*/
nd->nd_repstat = nfsvno_createsub(nd, &named, &vp, &nva,
Don't pass td to nfsvno_createsub(). MFC after: 2 weeks Sponsored by: DARPA, AFRL
2019-03-04 14:30:53 +00:00
&exclusive_flag, cverf, rdev, exp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
nd->nd_repstat = nfsvno_getfh(vp, &fh, p);
if (!nd->nd_repstat)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
if (!nd->nd_repstat) {
tverf[0] = nva.na_atime.tv_sec;
tverf[1] = nva.na_atime.tv_nsec;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (nd->nd_flag & ND_NFSV2) {
if (!nd->nd_repstat) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)&fh, 0, 0);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_fillattr(nd, &nva);
}
} else {
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
if (exclusive_flag && !nd->nd_repstat && (cverf[0] != tverf[0]
|| cverf[1] != tverf[1]))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = EEXIST;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
diraft_ret = nfsvno_getattr(dirp, &diraft, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
if (!nd->nd_repstat) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)&fh, 0, 1);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_postopattr(nd, 0, &nva);
}
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs v3 mknod service (and v4 create)
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_mknod(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, vnode_t *vpp, fhandle_t *fhp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr nva, dirfor, diraft;
u_int32_t *tl;
struct nameidata named;
int error = 0, dirfor_ret = 1, diraft_ret = 1, pathlen;
u_int32_t major, minor;
enum vtype vtyp = VNON;
nfstype nfs4type = NFNON;
vnode_t vp, dirp = NULL;
nfsattrbit_t attrbits;
char *bufp = NULL, *pathcp = NULL;
u_long *hashp, cnflags;
NFSACL_T *aclp = NULL;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSVNO_ATTRINIT(&nva);
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
cnflags = LOCKPARENT;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
#ifdef NFS4_ACL_EXTATTR_NAME
Fix the experimental nfs subsystem so that it builds with the current NFSv4 ACLs, as defined in sys/acl.h. It still needs a way to test a mount point for NFSv4 ACL support before it will work. Until then, the NFSHASNFS4ACL() macro just always returns 0. Approved by: kib (mentor)
2009-05-26 22:21:53 +00:00
aclp = acl_alloc(M_WAITOK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
aclp->acl_cnt = 0;
#endif
/*
* For V4, the creation stuff is here, Yuck!
*/
if (nd->nd_flag & ND_NFSV4) {
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
vtyp = nfsv34tov_type(*tl);
nfs4type = fxdr_unsigned(nfstype, *tl);
switch (nfs4type) {
case NFLNK:
error = nfsvno_getsymlink(nd, &nva, p, &pathcp,
&pathlen);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
break;
case NFCHR:
case NFBLK:
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
major = fxdr_unsigned(u_int32_t, *tl++);
minor = fxdr_unsigned(u_int32_t, *tl);
nva.na_rdev = NFSMAKEDEV(major, minor);
break;
case NFSOCK:
case NFFIFO:
break;
case NFDIR:
vfs: always retain path buffer after lookup This removes some of the complexity needed to maintain HASBUF and allows for removing injecting SAVENAME by filesystems. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D36542
2022-09-17 09:10:38 +00:00
cnflags = LOCKPARENT;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
break;
default:
nd->nd_repstat = NFSERR_BADTYPE;
vrele(dp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
The VOP_LOOKUP() implementations for CREATE op do not put the name into namecache, to avoid cache trashing when doing large operations. E.g., tar archive extraction is not usually followed by access to many of the files created. Right now, each VOP_LOOKUP() implementation explicitely knowns about this quirk and tests for both MAKEENTRY flag presence and op != CREATE to make the call to cache_enter(). Centralize the handling of the quirk into VFS, by deciding to cache only by MAKEENTRY flag in VOP. VFS now sets NOCACHE flag for CREATE namei() calls. Note that the change in semantic is backward-compatible and could be merged to the stable branch, and is compatible with non-changed third-party filesystems which correctly handle MAKEENTRY. Suggested by: Chris Torek <torek@pi-coral.com> Reviewed by: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 2 weeks
2014-12-18 10:01:12 +00:00
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, CREATE, cnflags | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV3) {
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
vtyp = nfsv34tov_type(*tl);
}
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
error = nfsrv_sattr(nd, NULL, &nva, &attrbits, aclp, p);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nva.na_type = vtyp;
if (!nd->nd_repstat && (nd->nd_flag & ND_NFSV3) &&
(vtyp == VCHR || vtyp == VBLK)) {
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
major = fxdr_unsigned(u_int32_t, *tl++);
minor = fxdr_unsigned(u_int32_t, *tl);
nva.na_rdev = NFSMAKEDEV(major, minor);
}
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dp, &dirfor, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat && (nd->nd_flag & ND_NFSV4)) {
if (!dirfor_ret && NFSVNO_ISSETGID(&nva) &&
dirfor.na_gid == nva.na_gid)
NFSVNO_UNSET(&nva, gid);
nd->nd_repstat = nfsrv_checkuidgid(nd, &nva);
}
if (nd->nd_repstat) {
vrele(dp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
nfsvno_relpathbuf(&named);
if (pathcp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(pathcp, M_TEMP);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret,
&diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Yuck! For V4, mkdir and link are here and some V4 clients don't fill
* in va_mode, so we'll have to set a default here.
*/
if (NFSVNO_NOTSETMODE(&nva)) {
if (vtyp == VLNK)
nva.na_mode = 0755;
else
nva.na_mode = 0400;
}
if (vtyp == VDIR)
named.ni_cnd.cn_flags |= WILLBEDIR;
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 0, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
if (dirp) {
if (nd->nd_flag & ND_NFSV3)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd,
p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret,
&diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (dirp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if ((nd->nd_flag & ND_NFSV4) && (vtyp == VDIR || vtyp == VLNK)) {
if (vtyp == VDIR) {
nfsrvd_mkdirsub(nd, &named, &nva, fhp, vpp, dirp,
&dirfor, &diraft, &diraft_ret, &attrbits, aclp, p,
exp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else if (vtyp == VLNK) {
nfsrvd_symlinksub(nd, &named, &nva, fhp, vpp, dirp,
&dirfor, &diraft, &diraft_ret, &attrbits,
aclp, p, exp, pathcp, pathlen);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(pathcp, M_TEMP);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
nd->nd_repstat = nfsvno_mknod(&named, &nva, nd->nd_cred, p);
if (!nd->nd_repstat) {
vp = named.ni_vp;
nfsrv_fixattr(nd, vp, &nva, aclp, p, &attrbits, exp);
nd->nd_repstat = nfsvno_getfh(vp, fhp, p);
if ((nd->nd_flag & ND_NFSV3) && !nd->nd_repstat)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1,
NULL);
Modify the experimental NFSv4 server so that the lookup ops return a locked vnode. This ensures that the associated mount point will always be valid for the code that follows the operation. Also add a couple of additional checks for non-error to the other functions that create file objects. MFC after: 2 weeks
2011-01-03 00:33:32 +00:00
if (vpp != NULL && nd->nd_repstat == 0) {
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*vpp = vp;
Modify the experimental NFSv4 server so that the lookup ops return a locked vnode. This ensures that the associated mount point will always be valid for the code that follows the operation. Also add a couple of additional checks for non-error to the other functions that create file objects. MFC after: 2 weeks
2011-01-03 00:33:32 +00:00
} else
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
diraft_ret = nfsvno_getattr(dirp, &diraft, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
if (!nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV3) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)fhp, 0, 1);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_postopattr(nd, 0, &nva);
} else {
NFSM_BUILD(tl, u_int32_t *, 5 * NFSX_UNSIGNED);
*tl++ = newnfs_false;
txdr_hyper(dirfor.na_filerev, tl);
tl += 2;
txdr_hyper(diraft.na_filerev, tl);
(void) nfsrv_putattrbit(nd, &attrbits);
}
}
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vrele(dp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
if (bufp)
nfsvno_relpathbuf(&named);
if (pathcp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(pathcp, M_TEMP);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs remove service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_remove(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nameidata named;
u_int32_t *tl;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
int error = 0, dirfor_ret = 1, diraft_ret = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vnode_t dirp = NULL;
struct nfsvattr dirfor, diraft;
char *bufp;
u_long *hashp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, DELETE,
LOCKPARENT | LOCKLEAF);
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (error) {
vput(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat) {
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 1, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vput(dp);
nfsvno_relpathbuf(&named);
}
if (dirp) {
if (!(nd->nd_flag & ND_NFSV2)) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vrele(dirp);
dirp = NULL;
}
}
if (!nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV4) {
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (named.ni_vp->v_type == VDIR)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsvno_rmdirsub(&named, 1,
nd->nd_cred, p, exp);
else
nd->nd_repstat = nfsvno_removesub(&named, 1,
nd->nd_cred, p, exp);
} else if (nd->nd_procnum == NFSPROC_RMDIR) {
nd->nd_repstat = nfsvno_rmdirsub(&named, 0,
nd->nd_cred, p, exp);
} else {
nd->nd_repstat = nfsvno_removesub(&named, 0,
nd->nd_cred, p, exp);
}
}
if (!(nd->nd_flag & ND_NFSV2)) {
if (dirp) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
diraft_ret = nfsvno_getattr(dirp, &diraft, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if (nd->nd_flag & ND_NFSV3) {
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret,
&diraft);
} else if (!nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, 5 * NFSX_UNSIGNED);
*tl++ = newnfs_false;
txdr_hyper(dirfor.na_filerev, tl);
tl += 2;
txdr_hyper(diraft.na_filerev, tl);
}
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs rename service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_rename(struct nfsrv_descript *nd, int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, vnode_t todp, struct nfsexstuff *exp, struct nfsexstuff *toexp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
int error = 0, fdirfor_ret = 1, fdiraft_ret = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
int tdirfor_ret = 1, tdiraft_ret = 1;
struct nameidata fromnd, tond;
vnode_t fdirp = NULL, tdirp = NULL, tdp = NULL;
struct nfsvattr fdirfor, fdiraft, tdirfor, tdiraft;
struct nfsexstuff tnes;
struct nfsrvfh tfh;
char *bufp, *tbufp = NULL;
u_long *hashp;
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
fhandle_t fh;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, fdirfor_ret, &fdirfor, fdiraft_ret, &fdiraft);
nfsrv_wcc(nd, tdirfor_ret, &tdirfor, tdiraft_ret, &tdiraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!(nd->nd_flag & ND_NFSV2))
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
fdirfor_ret = nfsvno_getattr(dp, &fdirfor, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
tond.ni_cnd.cn_nameiop = 0;
tond.ni_startdir = NULL;
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
NFSNAMEICNDSET(&fromnd.ni_cnd, nd->nd_cred, DELETE, WANTPARENT);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&fromnd, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &fromnd.ni_pathlen);
if (error) {
vput(dp);
if (todp)
vrele(todp);
nfsvno_relpathbuf(&fromnd);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Fix NFS deadlock vulnerability. [SA-14:05] Fix "Heartbleed" vulnerability and ECDSA Cache Side-channel Attack in OpenSSL. [SA-14:06]
2014-04-08 18:27:32 +00:00
/*
* Unlock dp in this code section, so it is unlocked before
* tdp gets locked. This avoids a potential LOR if tdp is the
* parent directory of dp.
*/
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV4) {
tdp = todp;
tnes = *toexp;
Fix NFS deadlock vulnerability. [SA-14:05] Fix "Heartbleed" vulnerability and ECDSA Cache Side-channel Attack in OpenSSL. [SA-14:06]
2014-04-08 18:27:32 +00:00
if (dp != tdp) {
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(dp);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/* Might lock tdp. */
tdirfor_ret = nfsvno_getattr(tdp, &tdirfor, nd, p, 0,
NULL);
Fix NFS deadlock vulnerability. [SA-14:05] Fix "Heartbleed" vulnerability and ECDSA Cache Side-channel Attack in OpenSSL. [SA-14:06]
2014-04-08 18:27:32 +00:00
} else {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
tdirfor_ret = nfsvno_getattr(tdp, &tdirfor, nd, p, 1,
NULL);
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(dp);
Fix NFS deadlock vulnerability. [SA-14:05] Fix "Heartbleed" vulnerability and ECDSA Cache Side-channel Attack in OpenSSL. [SA-14:06]
2014-04-08 18:27:32 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
tfh.nfsrvfh_len = 0;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
error = nfsrv_mtofh(nd, &tfh);
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
if (error == 0)
error = nfsvno_getfh(dp, &fh, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (error) {
vput(dp);
/* todp is always NULL except NFSv4 */
nfsvno_relpathbuf(&fromnd);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
/* If this is the same file handle, just VREF() the vnode. */
if (tfh.nfsrvfh_len == NFSX_MYFH &&
!NFSBCMP(tfh.nfsrvfh_data, &fh, NFSX_MYFH)) {
VREF(dp);
tdp = dp;
tnes = *exp;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
tdirfor_ret = nfsvno_getattr(tdp, &tdirfor, nd, p, 1,
NULL);
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(dp);
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
} else {
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(dp);
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
nd->nd_cred->cr_uid = nd->nd_saveduid;
nfsd_fhtovp(nd, &tfh, LK_EXCLUSIVE, &tdp, &tnes, NULL,
nfsd: Fix when NFSERR_WRONGSEC may be replied to NFSv4 clients Commit d224f05fcfc1 pre-parsed the next operation number for the put file handle operations. This patch uses this next operation number, plus the type of the file handle being set by the put file handle operation, to implement the rules in RFC5661 Sec. 2.6 with respect to replying NFSERR_WRONGSEC. This patch also adds a check to see if NFSERR_WRONGSEC should be replied when about to perform Lookup, Lookupp or Open with a file name component, so that the NFSERR_WRONGSEC reply is done for these operations, as required by RFC5661 Sec. 2.6. This patch does not have any practical effect for the FreeBSD NFSv4 client and I believe that the same is true for the Linux client, since NFSERR_WRONGSEC is considered a fatal error at this time. MFC after: 2 weeks
2021-06-05 23:53:07 +00:00
0, -1); /* Locks tdp. */
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
if (tdp) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
tdirfor_ret = nfsvno_getattr(tdp, &tdirfor, nd,
p, 1, NULL);
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(tdp);
Fix rename in the new NFS server so that it does not require a recursive vnode lock on the directory for the case where the new file name is in the same directory as the old one. The patch handles this as a special case, recognized by the new directory having the same file handle as the old one and just VREF()s the old dir vnode for this case, instead of doing a second VFS_FHTOVP() to get it. This is required so that the server will work for file systems like msdosfs, that do not support recursive vnode locking. This problem was discovered during recent testing by pho@ when exporting an msdosfs file system via the new NFS server. Tested by: pho Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-07-31 20:06:11 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
NFSNAMEICNDSET(&tond.ni_cnd, nd->nd_cred, RENAME, LOCKPARENT | LOCKLEAF | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&tond, &tbufp, &hashp);
if (!nd->nd_repstat) {
error = nfsrv_parsename(nd, tbufp, hashp, &tond.ni_pathlen);
if (error) {
Since the VFS_LOCK_GIANT() code in the experimental NFS server is broken and the major file systems are now all mpsafe, modify the server so that it will only export mpsafe file systems. This was discussed on freebsd-fs@ and removes a fair bit of crufty code. MFC after: 12 days
2011-01-06 19:50:11 +00:00
if (tdp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(tdp);
Fix NFS deadlock vulnerability. [SA-14:05] Fix "Heartbleed" vulnerability and ECDSA Cache Side-channel Attack in OpenSSL. [SA-14:06]
2014-04-08 18:27:32 +00:00
vrele(dp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_relpathbuf(&fromnd);
nfsvno_relpathbuf(&tond);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
if (nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV3) {
nfsrv_wcc(nd, fdirfor_ret, &fdirfor, fdiraft_ret,
&fdiraft);
nfsrv_wcc(nd, tdirfor_ret, &tdirfor, tdiraft_ret,
&tdiraft);
}
Since the VFS_LOCK_GIANT() code in the experimental NFS server is broken and the major file systems are now all mpsafe, modify the server so that it will only export mpsafe file systems. This was discussed on freebsd-fs@ and removes a fair bit of crufty code. MFC after: 12 days
2011-01-06 19:50:11 +00:00
if (tdp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(tdp);
Fix NFS deadlock vulnerability. [SA-14:05] Fix "Heartbleed" vulnerability and ECDSA Cache Side-channel Attack in OpenSSL. [SA-14:06]
2014-04-08 18:27:32 +00:00
vrele(dp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_relpathbuf(&fromnd);
nfsvno_relpathbuf(&tond);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Done parsing, now down to business.
*/
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &fromnd, dp, 0, exp, &fdirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV3) {
nfsrv_wcc(nd, fdirfor_ret, &fdirfor, fdiraft_ret,
&fdiraft);
nfsrv_wcc(nd, tdirfor_ret, &tdirfor, tdiraft_ret,
&tdiraft);
}
if (fdirp)
vrele(fdirp);
Since the VFS_LOCK_GIANT() code in the experimental NFS server is broken and the major file systems are now all mpsafe, modify the server so that it will only export mpsafe file systems. This was discussed on freebsd-fs@ and removes a fair bit of crufty code. MFC after: 12 days
2011-01-06 19:50:11 +00:00
if (tdp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(tdp);
nfsvno_relpathbuf(&tond);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (fromnd.ni_vp->v_type == VDIR)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
tond.ni_cnd.cn_flags |= WILLBEDIR;
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &tond, tdp, 0, &tnes, &tdirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsvno_rename(&fromnd, &tond, nd->nd_repstat,
nd->nd_flag, nd->nd_cred, p);
if (fdirp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
fdiraft_ret = nfsvno_getattr(fdirp, &fdiraft, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (tdirp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
tdiraft_ret = nfsvno_getattr(tdirp, &tdiraft, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (fdirp)
vrele(fdirp);
if (tdirp)
vrele(tdirp);
if (nd->nd_flag & ND_NFSV3) {
nfsrv_wcc(nd, fdirfor_ret, &fdirfor, fdiraft_ret, &fdiraft);
nfsrv_wcc(nd, tdirfor_ret, &tdirfor, tdiraft_ret, &tdiraft);
} else if ((nd->nd_flag & ND_NFSV4) && !nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, 10 * NFSX_UNSIGNED);
*tl++ = newnfs_false;
txdr_hyper(fdirfor.na_filerev, tl);
tl += 2;
txdr_hyper(fdiraft.na_filerev, tl);
tl += 2;
*tl++ = newnfs_false;
txdr_hyper(tdirfor.na_filerev, tl);
tl += 2;
txdr_hyper(tdiraft.na_filerev, tl);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs link service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_link(struct nfsrv_descript *nd, int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, vnode_t tovp, struct nfsexstuff *exp, struct nfsexstuff *toexp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nameidata named;
u_int32_t *tl;
int error = 0, dirfor_ret = 1, diraft_ret = 1, getret = 1;
vnode_t dirp = NULL, dp = NULL;
struct nfsvattr dirfor, diraft, at;
struct nfsexstuff tnes;
struct nfsrvfh dfh;
char *bufp;
u_long *hashp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, getret, &at);
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (vp->v_type == VDIR) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV4)
nd->nd_repstat = NFSERR_ISDIR;
else
nd->nd_repstat = NFSERR_INVAL;
if (tovp)
vrele(tovp);
}
if (!nd->nd_repstat) {
if (nd->nd_flag & ND_NFSV4) {
dp = tovp;
tnes = *toexp;
} else {
error = nfsrv_mtofh(nd, &dfh);
if (error) {
vrele(vp);
/* tovp is always NULL unless NFSv4 */
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfsd: Fix when NFSERR_WRONGSEC may be replied to NFSv4 clients Commit d224f05fcfc1 pre-parsed the next operation number for the put file handle operations. This patch uses this next operation number, plus the type of the file handle being set by the put file handle operation, to implement the rules in RFC5661 Sec. 2.6 with respect to replying NFSERR_WRONGSEC. This patch also adds a check to see if NFSERR_WRONGSEC should be replied when about to perform Lookup, Lookupp or Open with a file name component, so that the NFSERR_WRONGSEC reply is done for these operations, as required by RFC5661 Sec. 2.6. This patch does not have any practical effect for the FreeBSD NFSv4 client and I believe that the same is true for the Linux client, since NFSERR_WRONGSEC is considered a fatal error at this time. MFC after: 2 weeks
2021-06-05 23:53:07 +00:00
nfsd_fhtovp(nd, &dfh, LK_EXCLUSIVE, &dp, &tnes, NULL,
0, -1);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (dp)
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(dp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
vfs: always retain path buffer after lookup This removes some of the complexity needed to maintain HASBUF and allows for removing injecting SAVENAME by filesystems. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D36542
2022-09-17 09:10:38 +00:00
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, CREATE, LOCKPARENT | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (error) {
vrele(vp);
Since the VFS_LOCK_GIANT() code in the experimental NFS server is broken and the major file systems are now all mpsafe, modify the server so that it will only export mpsafe file systems. This was discussed on freebsd-fs@ and removes a fair bit of crufty code. MFC after: 12 days
2011-01-06 19:50:11 +00:00
if (dp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat) {
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 0, &tnes,
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
&dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
if (dp)
vrele(dp);
nfsvno_relpathbuf(&named);
}
}
if (dirp) {
if (nd->nd_flag & ND_NFSV2) {
vrele(dirp);
dirp = NULL;
} else {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
if (!nd->nd_repstat)
nd->nd_repstat = nfsvno_link(&named, vp, nd->nd_cred, p, exp);
if (nd->nd_flag & ND_NFSV3)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &at, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (dirp) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
diraft_ret = nfsvno_getattr(dirp, &diraft, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
vrele(vp);
if (nd->nd_flag & ND_NFSV3) {
nfsrv_postopattr(nd, getret, &at);
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
} else if ((nd->nd_flag & ND_NFSV4) && !nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, 5 * NFSX_UNSIGNED);
*tl++ = newnfs_false;
txdr_hyper(dirfor.na_filerev, tl);
tl += 2;
txdr_hyper(diraft.na_filerev, tl);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs symbolic link service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_symlink(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, vnode_t *vpp, fhandle_t *fhp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr nva, dirfor, diraft;
struct nameidata named;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
int error = 0, dirfor_ret = 1, diraft_ret = 1, pathlen;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vnode_t dirp = NULL;
char *bufp, *pathcp = NULL;
u_long *hashp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (vpp)
*vpp = NULL;
NFSVNO_ATTRINIT(&nva);
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, CREATE,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKPARENT | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (!error && !nd->nd_repstat)
error = nfsvno_getsymlink(nd, &nva, p, &pathcp, &pathlen);
if (error) {
vrele(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat) {
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 0, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vrele(dp);
nfsvno_relpathbuf(&named);
}
if (dirp != NULL && !(nd->nd_flag & ND_NFSV3)) {
vrele(dirp);
dirp = NULL;
}
/*
* And call nfsrvd_symlinksub() to do the common code. It will
* return EBADRPC upon a parsing error, 0 otherwise.
*/
if (!nd->nd_repstat) {
if (dirp != NULL)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_symlinksub(nd, &named, &nva, fhp, vpp, dirp,
&dirfor, &diraft, &diraft_ret, NULL, NULL, p, exp,
pathcp, pathlen);
} else if (dirp != NULL) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if (pathcp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(pathcp, M_TEMP);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV3) {
if (!nd->nd_repstat) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)fhp, 0, 1);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_postopattr(nd, 0, &nva);
}
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Common code for creating a symbolic link.
*/
static void
nfsrvd_symlinksub(struct nfsrv_descript *nd, struct nameidata *ndp,
struct nfsvattr *nvap, fhandle_t *fhp, vnode_t *vpp,
vnode_t dirp, struct nfsvattr *dirforp, struct nfsvattr *diraftp,
int *diraft_retp, nfsattrbit_t *attrbitp,
NFSACL_T *aclp, NFSPROC_T *p, struct nfsexstuff *exp, char *pathcp,
int pathlen)
{
u_int32_t *tl;
nd->nd_repstat = nfsvno_symlink(ndp, nvap, pathcp, pathlen,
!(nd->nd_flag & ND_NFSV2), nd->nd_saveduid, nd->nd_cred, p, exp);
if (!nd->nd_repstat && !(nd->nd_flag & ND_NFSV2)) {
nfsrv_fixattr(nd, ndp->ni_vp, nvap, aclp, p, attrbitp, exp);
if (nd->nd_flag & ND_NFSV3) {
nd->nd_repstat = nfsvno_getfh(ndp->ni_vp, fhp, p);
if (!nd->nd_repstat)
nd->nd_repstat = nfsvno_getattr(ndp->ni_vp,
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nvap, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Modify the experimental NFSv4 server so that the lookup ops return a locked vnode. This ensures that the associated mount point will always be valid for the code that follows the operation. Also add a couple of additional checks for non-error to the other functions that create file objects. MFC after: 2 weeks
2011-01-03 00:33:32 +00:00
if (vpp != NULL && nd->nd_repstat == 0) {
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(ndp->ni_vp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*vpp = ndp->ni_vp;
Modify the experimental NFSv4 server so that the lookup ops return a locked vnode. This ensures that the associated mount point will always be valid for the code that follows the operation. Also add a couple of additional checks for non-error to the other functions that create file objects. MFC after: 2 weeks
2011-01-03 00:33:32 +00:00
} else
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(ndp->ni_vp);
}
if (dirp) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
*diraft_retp = nfsvno_getattr(dirp, diraftp, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if ((nd->nd_flag & ND_NFSV4) && !nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, 5 * NFSX_UNSIGNED);
*tl++ = newnfs_false;
txdr_hyper(dirforp->na_filerev, tl);
tl += 2;
txdr_hyper(diraftp->na_filerev, tl);
(void) nfsrv_putattrbit(nd, attrbitp);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs mkdir service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_mkdir(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, vnode_t *vpp, fhandle_t *fhp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr nva, dirfor, diraft;
struct nameidata named;
u_int32_t *tl;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
int error = 0, dirfor_ret = 1, diraft_ret = 1;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vnode_t dirp = NULL;
char *bufp;
u_long *hashp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
vfs: always retain path buffer after lookup This removes some of the complexity needed to maintain HASBUF and allows for removing injecting SAVENAME by filesystems. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D36542
2022-09-17 09:10:38 +00:00
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, CREATE, LOCKPARENT | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
NFSVNO_ATTRINIT(&nva);
if (nd->nd_flag & ND_NFSV3) {
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
error = nfsrv_sattr(nd, NULL, &nva, NULL, NULL, p);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
nva.na_mode = nfstov_mode(*tl++);
}
}
if (!nd->nd_repstat) {
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 0, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vrele(dp);
nfsvno_relpathbuf(&named);
}
if (dirp != NULL && !(nd->nd_flag & ND_NFSV3)) {
vrele(dirp);
dirp = NULL;
}
if (nd->nd_repstat) {
if (dirp != NULL) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if (nd->nd_flag & ND_NFSV3)
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret,
&diraft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (dirp != NULL)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
dirfor_ret = nfsvno_getattr(dirp, &dirfor, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* Call nfsrvd_mkdirsub() for the code common to V4 as well.
*/
nfsrvd_mkdirsub(nd, &named, &nva, fhp, vpp, dirp, &dirfor, &diraft,
&diraft_ret, NULL, NULL, p, exp);
if (nd->nd_flag & ND_NFSV3) {
if (!nd->nd_repstat) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)fhp, 0, 1);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_postopattr(nd, 0, &nva);
}
nfsrv_wcc(nd, dirfor_ret, &dirfor, diraft_ret, &diraft);
} else if (!nd->nd_repstat) {
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)fhp, 0, 0);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_fillattr(nd, &nva);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vrele(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* Code common to mkdir for V2,3 and 4.
*/
static void
nfsrvd_mkdirsub(struct nfsrv_descript *nd, struct nameidata *ndp,
struct nfsvattr *nvap, fhandle_t *fhp, vnode_t *vpp,
vnode_t dirp, struct nfsvattr *dirforp, struct nfsvattr *diraftp,
int *diraft_retp, nfsattrbit_t *attrbitp, NFSACL_T *aclp,
NFSPROC_T *p, struct nfsexstuff *exp)
{
vnode_t vp;
u_int32_t *tl;
NFSVNO_SETATTRVAL(nvap, type, VDIR);
nd->nd_repstat = nfsvno_mkdir(ndp, nvap, nd->nd_saveduid,
nd->nd_cred, p, exp);
if (!nd->nd_repstat) {
vp = ndp->ni_vp;
nfsrv_fixattr(nd, vp, nvap, aclp, p, attrbitp, exp);
nd->nd_repstat = nfsvno_getfh(vp, fhp, p);
if (!(nd->nd_flag & ND_NFSV4) && !nd->nd_repstat)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, nvap, nd, p, 1,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (vpp && !nd->nd_repstat) {
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*vpp = vp;
} else {
vput(vp);
}
}
if (dirp) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
*diraft_retp = nfsvno_getattr(dirp, diraftp, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vrele(dirp);
}
if ((nd->nd_flag & ND_NFSV4) && !nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, 5 * NFSX_UNSIGNED);
*tl++ = newnfs_false;
txdr_hyper(dirforp->na_filerev, tl);
tl += 2;
txdr_hyper(diraftp->na_filerev, tl);
(void) nfsrv_putattrbit(nd, attrbitp);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* nfs commit service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_commit(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsvattr bfor, aft;
u_int32_t *tl;
int error = 0, for_ret = 1, aft_ret = 1, cnt;
u_int64_t off;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
if (nd->nd_repstat) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_wcc(nd, for_ret, &bfor, aft_ret, &aft);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
/* Return NFSERR_ISDIR in NFSv4 when commit on a directory. */
if (vp->v_type != VREG) {
if (nd->nd_flag & ND_NFSV3)
error = NFSERR_NOTSUPP;
else
error = (vp->v_type == VDIR) ? NFSERR_ISDIR : NFSERR_INVAL;
goto nfsmout;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 3 * NFSX_UNSIGNED);
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* XXX At this time VOP_FSYNC() does not accept offset and byte
* count parameters, so these arguments are useless (someday maybe).
*/
off = fxdr_hyper(tl);
tl += 2;
cnt = fxdr_unsigned(int, *tl);
if (nd->nd_flag & ND_NFSV3)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
for_ret = nfsvno_getattr(vp, &bfor, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsvno_fsync(vp, off, cnt, nd->nd_cred, p);
if (nd->nd_flag & ND_NFSV3) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
aft_ret = nfsvno_getattr(vp, &aft, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrv_wcc(nd, for_ret, &bfor, aft_ret, &aft);
}
vput(vp);
if (!nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, NFSX_VERF);
*tl++ = txdr_unsigned(nfsboottime.tv_sec);
*tl = txdr_unsigned(nfsboottime.tv_usec);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs statfs service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_statfs(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct statfs *sf;
u_int32_t *tl;
int getret = 1;
struct nfsvattr at;
u_quad_t tval;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
sf = NULL;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, getret, &at);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
sf = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK);
Remove the thread argument from the FSD (File-System Dependent) parts of the VFS. Now all the VFS_* functions and relating parts don't want the context as long as it always refers to curthread. In some points, in particular when dealing with VOPs and functions living in the same namespace (eg. vflush) which still need to be converted, pass curthread explicitly in order to retain the old behaviour. Such loose ends will be fixed ASAP. While here fix a bug: now, UFS_EXTATTR can be compiled alone without the UFS_EXTATTR_AUTOSTART option. VFS KPI is heavilly changed by this commit so thirdy parts modules needs to be recompiled. Bump __FreeBSD_version in order to signal such situation.
2009-05-11 15:33:26 +00:00
nd->nd_repstat = nfsvno_statfs(vp, sf);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &at, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
if (nd->nd_flag & ND_NFSV3)
nfsrv_postopattr(nd, getret, &at);
if (nd->nd_repstat)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_flag & ND_NFSV2) {
NFSM_BUILD(tl, u_int32_t *, NFSX_V2STATFS);
*tl++ = txdr_unsigned(NFS_V2MAXDATA);
*tl++ = txdr_unsigned(sf->f_bsize);
*tl++ = txdr_unsigned(sf->f_blocks);
*tl++ = txdr_unsigned(sf->f_bfree);
*tl = txdr_unsigned(sf->f_bavail);
} else {
NFSM_BUILD(tl, u_int32_t *, NFSX_V3STATFS);
tval = (u_quad_t)sf->f_blocks;
tval *= (u_quad_t)sf->f_bsize;
txdr_hyper(tval, tl); tl += 2;
tval = (u_quad_t)sf->f_bfree;
tval *= (u_quad_t)sf->f_bsize;
txdr_hyper(tval, tl); tl += 2;
tval = (u_quad_t)sf->f_bavail;
tval *= (u_quad_t)sf->f_bsize;
txdr_hyper(tval, tl); tl += 2;
tval = (u_quad_t)sf->f_files;
txdr_hyper(tval, tl); tl += 2;
tval = (u_quad_t)sf->f_ffree;
txdr_hyper(tval, tl); tl += 2;
tval = (u_quad_t)sf->f_ffree;
txdr_hyper(tval, tl); tl += 2;
*tl = 0;
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
free(sf, M_STATFS);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
}
/*
* nfs fsinfo service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_fsinfo(struct nfsrv_descript *nd, int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
struct nfsfsinfo fs;
int getret = 1;
struct nfsvattr at;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, getret, &at);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &at, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_getfs(&fs, isdgram);
vput(vp);
nfsrv_postopattr(nd, getret, &at);
NFSM_BUILD(tl, u_int32_t *, NFSX_V3FSINFO);
*tl++ = txdr_unsigned(fs.fs_rtmax);
*tl++ = txdr_unsigned(fs.fs_rtpref);
*tl++ = txdr_unsigned(fs.fs_rtmult);
*tl++ = txdr_unsigned(fs.fs_wtmax);
*tl++ = txdr_unsigned(fs.fs_wtpref);
*tl++ = txdr_unsigned(fs.fs_wtmult);
*tl++ = txdr_unsigned(fs.fs_dtpref);
txdr_hyper(fs.fs_maxfilesize, tl);
tl += 2;
txdr_nfsv3time(&fs.fs_timedelta, tl);
tl += 2;
*tl = txdr_unsigned(fs.fs_properties);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
}
/*
* nfs pathconf service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_pathconf(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
struct nfsv3_pathconf *pc;
int getret = 1;
Use long for the last argument to VOP_PATHCONF rather than a register_t. pathconf(2) and fpathconf(2) both return a long. The kern_[f]pathconf() functions now accept a pointer to a long value rather than modifying td_retval directly. Instead, the system calls explicitly store the returned long value in td_retval[0]. Requested by: bde Reviewed by: kib Sponsored by: Chelsio Communications
2018-01-17 22:36:58 +00:00
long linkmax, namemax, chownres, notrunc;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct nfsvattr at;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat) {
nfsrv_postopattr(nd, getret, &at);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nd->nd_repstat = nfsvno_pathconf(vp, _PC_LINK_MAX, &linkmax,
nd->nd_cred, p);
if (!nd->nd_repstat)
nd->nd_repstat = nfsvno_pathconf(vp, _PC_NAME_MAX, &namemax,
nd->nd_cred, p);
if (!nd->nd_repstat)
nd->nd_repstat=nfsvno_pathconf(vp, _PC_CHOWN_RESTRICTED,
&chownres, nd->nd_cred, p);
if (!nd->nd_repstat)
nd->nd_repstat = nfsvno_pathconf(vp, _PC_NO_TRUNC, &notrunc,
nd->nd_cred, p);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
getret = nfsvno_getattr(vp, &at, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
nfsrv_postopattr(nd, getret, &at);
if (!nd->nd_repstat) {
NFSM_BUILD(pc, struct nfsv3_pathconf *, NFSX_V3PATHCONF);
pc->pc_linkmax = txdr_unsigned(linkmax);
pc->pc_namemax = txdr_unsigned(namemax);
pc->pc_notrunc = txdr_unsigned(notrunc);
pc->pc_chownrestricted = txdr_unsigned(chownres);
/*
* These should probably be supported by VOP_PATHCONF(), but
* until msdosfs is exportable (why would you want to?), the
* Unix defaults should be ok.
*/
pc->pc_caseinsensitive = newnfs_false;
pc->pc_casepreserving = newnfs_true;
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
}
/*
* nfsv4 lock service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_lock(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int i;
struct nfsstate *stp = NULL;
struct nfslock *lop;
struct nfslockconflict cf;
int error = 0;
u_short flags = NFSLCK_LOCK, lflags;
u_int64_t offset, len;
nfsv4stateid_t stateid;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 7 * NFSX_UNSIGNED);
i = fxdr_unsigned(int, *tl++);
switch (i) {
case NFSV4LOCKT_READW:
flags |= NFSLCK_BLOCKING;
case NFSV4LOCKT_READ:
lflags = NFSLCK_READ;
break;
case NFSV4LOCKT_WRITEW:
flags |= NFSLCK_BLOCKING;
case NFSV4LOCKT_WRITE:
lflags = NFSLCK_WRITE;
break;
default:
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (*tl++ == newnfs_true)
flags |= NFSLCK_RECLAIM;
offset = fxdr_hyper(tl);
tl += 2;
len = fxdr_hyper(tl);
tl += 2;
if (*tl == newnfs_true)
flags |= NFSLCK_OPENTOLOCK;
if (flags & NFSLCK_OPENTOLOCK) {
NFSM_DISSECT(tl, u_int32_t *, 5 * NFSX_UNSIGNED + NFSX_STATEID);
i = fxdr_unsigned(int, *(tl+4+(NFSX_STATEID / NFSX_UNSIGNED)));
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
if (i <= 0 || i > NFSV4_OPAQUELIMIT) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
stp = malloc(sizeof (struct nfsstate) + i,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDSTATE, M_WAITOK);
stp->ls_ownerlen = i;
stp->ls_op = nd->nd_rp;
stp->ls_seq = fxdr_unsigned(int, *tl++);
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stp->ls_stateid.other,
NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* For the special stateid of other all 0s and seqid == 1, set
* the stateid to the current stateid, if it is set.
*/
if ((nd->nd_flag & ND_NFSV41) != 0 &&
stp->ls_stateid.seqid == 1 &&
stp->ls_stateid.other[0] == 0 &&
stp->ls_stateid.other[1] == 0 &&
stp->ls_stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stp->ls_stateid = nd->nd_curstateid;
stp->ls_stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
stp->ls_opentolockseq = fxdr_unsigned(int, *tl++);
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl++;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK3 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
error = nfsrv_mtostr(nd, stp->ls_owner, stp->ls_ownerlen);
if (error)
goto nfsmout;
} else {
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID + NFSX_UNSIGNED);
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
stp = malloc(sizeof (struct nfsstate),
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDSTATE, M_WAITOK);
stp->ls_ownerlen = 0;
stp->ls_op = nd->nd_rp;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stp->ls_stateid.other,
NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* For the special stateid of other all 0s and seqid == 1, set
* the stateid to the current stateid, if it is set.
*/
if ((nd->nd_flag & ND_NFSV41) != 0 &&
stp->ls_stateid.seqid == 1 &&
stp->ls_stateid.other[0] == 0 &&
stp->ls_stateid.other[1] == 0 &&
stp->ls_stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stp->ls_stateid = nd->nd_curstateid;
stp->ls_stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
stp->ls_seq = fxdr_unsigned(int, *tl);
clientid.lval[0] = stp->ls_stateid.other[0];
clientid.lval[1] = stp->ls_stateid.other[1];
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK4 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
}
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
lop = malloc(sizeof (struct nfslock),
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDLOCK, M_WAITOK);
lop->lo_first = offset;
if (len == NFS64BITSSET) {
lop->lo_end = NFS64BITSSET;
} else {
lop->lo_end = offset + len;
if (lop->lo_end <= lop->lo_first)
nd->nd_repstat = NFSERR_INVAL;
}
lop->lo_flags = lflags;
stp->ls_flags = flags;
stp->ls_uid = nd->nd_cred->cr_uid;
/*
* Do basic access checking.
*/
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (!nd->nd_repstat && vp->v_type != VREG) {
if (vp->v_type == VDIR)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = NFSERR_ISDIR;
else
nd->nd_repstat = NFSERR_INVAL;
}
if (!nd->nd_repstat) {
if (lflags & NFSLCK_WRITE) {
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VWRITE,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_cred, exp, p, NFSACCCHK_ALLOWOWNER,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VREAD,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_cred, exp, p, NFSACCCHK_ALLOWOWNER,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat)
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VEXEC,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_cred, exp, p, NFSACCCHK_ALLOWOWNER,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
}
/*
* We call nfsrv_lockctrl() even if nd_repstat set, so that the
* seqid# gets updated. nfsrv_lockctrl() will return the value
* of nd_repstat, if it gets that far.
*/
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, &cf, clientid,
&stateid, exp, nd, p);
if (lop)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(lop, M_NFSDLOCK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/* For NFSv4.1, set the Current StateID. */
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_curstateid = stateid;
nd->nd_flag |= ND_CURSTATEID;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY((caddr_t)stateid.other,(caddr_t)tl,NFSX_STATEIDOTHER);
} else if (nd->nd_repstat == NFSERR_DENIED) {
NFSM_BUILD(tl, u_int32_t *, 7 * NFSX_UNSIGNED);
txdr_hyper(cf.cl_first, tl);
tl += 2;
if (cf.cl_end == NFS64BITSSET)
len = NFS64BITSSET;
else
len = cf.cl_end - cf.cl_first;
txdr_hyper(len, tl);
tl += 2;
if (cf.cl_flags == NFSLCK_WRITE)
*tl++ = txdr_unsigned(NFSV4LOCKT_WRITE);
else
*tl++ = txdr_unsigned(NFSV4LOCKT_READ);
*tl++ = stateid.other[0];
*tl = stateid.other[1];
(void) nfsm_strtom(nd, cf.cl_owner, cf.cl_ownerlen);
}
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 lock test service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_lockt(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int i;
struct nfsstate *stp = NULL;
struct nfslock lo, *lop = &lo;
struct nfslockconflict cf;
int error = 0;
nfsv4stateid_t stateid;
nfsquad_t clientid;
u_int64_t len;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 8 * NFSX_UNSIGNED);
i = fxdr_unsigned(int, *(tl + 7));
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
if (i <= 0 || i > NFSV4_OPAQUELIMIT) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
stp = malloc(sizeof (struct nfsstate) + i,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDSTATE, M_WAITOK);
stp->ls_ownerlen = i;
stp->ls_op = NULL;
stp->ls_flags = NFSLCK_TEST;
stp->ls_uid = nd->nd_cred->cr_uid;
i = fxdr_unsigned(int, *tl++);
switch (i) {
case NFSV4LOCKT_READW:
stp->ls_flags |= NFSLCK_BLOCKING;
case NFSV4LOCKT_READ:
lo.lo_flags = NFSLCK_READ;
break;
case NFSV4LOCKT_WRITEW:
stp->ls_flags |= NFSLCK_BLOCKING;
case NFSV4LOCKT_WRITE:
lo.lo_flags = NFSLCK_WRITE;
break;
default:
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
lo.lo_first = fxdr_hyper(tl);
tl += 2;
len = fxdr_hyper(tl);
if (len == NFS64BITSSET) {
lo.lo_end = NFS64BITSSET;
} else {
lo.lo_end = lo.lo_first + len;
if (lo.lo_end <= lo.lo_first)
nd->nd_repstat = NFSERR_INVAL;
}
tl += 2;
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK5 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
error = nfsrv_mtostr(nd, stp->ls_owner, stp->ls_ownerlen);
if (error)
goto nfsmout;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (!nd->nd_repstat && vp->v_type != VREG) {
if (vp->v_type == VDIR)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = NFSERR_ISDIR;
else
nd->nd_repstat = NFSERR_INVAL;
}
if (!nd->nd_repstat)
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, &cf, clientid,
&stateid, exp, nd, p);
if (nd->nd_repstat) {
if (nd->nd_repstat == NFSERR_DENIED) {
NFSM_BUILD(tl, u_int32_t *, 7 * NFSX_UNSIGNED);
txdr_hyper(cf.cl_first, tl);
tl += 2;
if (cf.cl_end == NFS64BITSSET)
len = NFS64BITSSET;
else
len = cf.cl_end - cf.cl_first;
txdr_hyper(len, tl);
tl += 2;
if (cf.cl_flags == NFSLCK_WRITE)
*tl++ = txdr_unsigned(NFSV4LOCKT_WRITE);
else
*tl++ = txdr_unsigned(NFSV4LOCKT_READ);
*tl++ = stp->ls_stateid.other[0];
*tl = stp->ls_stateid.other[1];
(void) nfsm_strtom(nd, cf.cl_owner, cf.cl_ownerlen);
}
}
vput(vp);
nfsd: Fix use-after-free in NFS4 lock test service Trivial use-after-free where stp was freed too soon in the non-error path. To fix, simply move its release to the end of the routine. Reported by: Coverity CID: 1006105 Sponsored by: EMC / Isilon Storage Division
2016-05-12 05:03:12 +00:00
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 unlock service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_locku(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int i;
struct nfsstate *stp;
struct nfslock *lop;
int error = 0;
nfsv4stateid_t stateid;
nfsquad_t clientid;
u_int64_t len;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 6 * NFSX_UNSIGNED + NFSX_STATEID);
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
stp = malloc(sizeof (struct nfsstate),
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDSTATE, M_WAITOK);
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
lop = malloc(sizeof (struct nfslock),
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDLOCK, M_WAITOK);
stp->ls_flags = NFSLCK_UNLOCK;
lop->lo_flags = NFSLCK_UNLOCK;
stp->ls_op = nd->nd_rp;
i = fxdr_unsigned(int, *tl++);
switch (i) {
case NFSV4LOCKT_READW:
stp->ls_flags |= NFSLCK_BLOCKING;
case NFSV4LOCKT_READ:
break;
case NFSV4LOCKT_WRITEW:
stp->ls_flags |= NFSLCK_BLOCKING;
case NFSV4LOCKT_WRITE:
break;
default:
nd->nd_repstat = NFSERR_BADXDR;
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
free(stp, M_NFSDSTATE);
free(lop, M_NFSDLOCK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
goto nfsmout;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
stp->ls_ownerlen = 0;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_seq = fxdr_unsigned(int, *tl++);
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stp->ls_stateid.other,
NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* For the special stateid of other all 0s and seqid == 1, set the
* stateid to the current stateid, if it is set.
*/
if ((nd->nd_flag & ND_NFSV41) != 0 && stp->ls_stateid.seqid == 1 &&
stp->ls_stateid.other[0] == 0 && stp->ls_stateid.other[1] == 0 &&
stp->ls_stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stp->ls_stateid = nd->nd_curstateid;
stp->ls_stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
Fix nfsrvd_locku memory leak Coverity detected memory leak fix. Submitted by: bret_ketchum@dell.com Reported by: Coverity Reviewed by: rmacklem MFC after: 2 weeks Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D26231
2020-08-31 15:31:17 +00:00
free(stp, M_NFSDSTATE);
free(lop, M_NFSDLOCK);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
goto nfsmout;
}
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
lop->lo_first = fxdr_hyper(tl);
tl += 2;
len = fxdr_hyper(tl);
if (len == NFS64BITSSET) {
lop->lo_end = NFS64BITSSET;
} else {
lop->lo_end = lop->lo_first + len;
if (lop->lo_end <= lop->lo_first)
nd->nd_repstat = NFSERR_INVAL;
}
clientid.lval[0] = stp->ls_stateid.other[0];
clientid.lval[1] = stp->ls_stateid.other[1];
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK6 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (!nd->nd_repstat && vp->v_type != VREG) {
if (vp->v_type == VDIR)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = NFSERR_ISDIR;
else
nd->nd_repstat = NFSERR_INVAL;
}
/*
* Call nfsrv_lockctrl() even if nd_repstat is set, so that the
* seqid# gets incremented. nfsrv_lockctrl() will return the
* value of nd_repstat, if it gets that far.
*/
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, NULL, clientid,
&stateid, exp, nd, p);
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (lop)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(lop, M_NFSDLOCK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY((caddr_t)stateid.other,(caddr_t)tl,NFSX_STATEIDOTHER);
}
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 open service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_open(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, vnode_t *vpp, __unused fhandle_t *fhp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
int i, retext;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct nfsstate *stp = NULL;
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
int error = 0, create, claim, exclusive_flag = 0, override;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
u_int32_t rflags = NFSV4OPEN_LOCKTYPEPOSIX, acemask;
int how = NFSCREATE_UNCHECKED;
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
int32_t cverf[2], tverf[2] = { 0, 0 };
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vnode_t vp = NULL, dirp = NULL;
struct nfsvattr nva, dirfor, diraft;
struct nameidata named;
nfsv4stateid_t stateid, delegstateid;
nfsattrbit_t attrbits;
nfsquad_t clientid;
char *bufp = NULL;
u_long *hashp;
NFSACL_T *aclp = NULL;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
nfsserver: Fix vrele() panic in nfsvno_open() Commit 65127e982b94 removed a check for ni_startdir != NULL. This allowed the vrele(ndp->ni_dvp) to be called with a NULL argument. This patch adds a new boolean argument to nfsvno_open() that can be checked instead of ni_startdir, since mjg@ requested that ni_startdir not be used. (Discussed in PR#268828.) PR: 268828 Reviewed by: mjg Differential Revision: https://reviews.freebsd.org/D38032
2023-01-13 00:45:26 +00:00
bool done_namei;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
#ifdef NFS4_ACL_EXTATTR_NAME
Fix the experimental nfs subsystem so that it builds with the current NFSv4 ACLs, as defined in sys/acl.h. It still needs a way to test a mount point for NFSv4 ACL support before it will work. Until then, the NFSHASNFS4ACL() macro just always returns 0. Approved by: kib (mentor)
2009-05-26 22:21:53 +00:00
aclp = acl_alloc(M_WAITOK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
aclp->acl_cnt = 0;
#endif
NFSZERO_ATTRBIT(&attrbits);
nfsserver: Fix vrele() panic in nfsvno_open() Commit 65127e982b94 removed a check for ni_startdir != NULL. This allowed the vrele(ndp->ni_dvp) to be called with a NULL argument. This patch adds a new boolean argument to nfsvno_open() that can be checked instead of ni_startdir, since mjg@ requested that ni_startdir not be used. (Discussed in PR#268828.) PR: 268828 Reviewed by: mjg Differential Revision: https://reviews.freebsd.org/D38032
2023-01-13 00:45:26 +00:00
done_namei = false;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
named.ni_cnd.cn_nameiop = 0;
NFSM_DISSECT(tl, u_int32_t *, 6 * NFSX_UNSIGNED);
i = fxdr_unsigned(int, *(tl + 5));
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
if (i <= 0 || i > NFSV4_OPAQUELIMIT) {
nd->nd_repstat = NFSERR_BADXDR;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
}
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
stp = malloc(sizeof (struct nfsstate) + i,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDSTATE, M_WAITOK);
stp->ls_ownerlen = i;
stp->ls_op = nd->nd_rp;
stp->ls_flags = NFSLCK_OPEN;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_seq = fxdr_unsigned(u_int32_t, *tl++);
i = fxdr_unsigned(int, *tl++);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
retext = 0;
if ((i & (NFSV4OPEN_WANTDELEGMASK | NFSV4OPEN_WANTSIGNALDELEG |
NFSV4OPEN_WANTPUSHDELEG)) != 0 && (nd->nd_flag & ND_NFSV41) != 0) {
retext = 1;
/* For now, ignore these. */
i &= ~(NFSV4OPEN_WANTPUSHDELEG | NFSV4OPEN_WANTSIGNALDELEG);
switch (i & NFSV4OPEN_WANTDELEGMASK) {
case NFSV4OPEN_WANTANYDELEG:
stp->ls_flags |= (NFSLCK_WANTRDELEG |
NFSLCK_WANTWDELEG);
i &= ~NFSV4OPEN_WANTDELEGMASK;
break;
case NFSV4OPEN_WANTREADDELEG:
stp->ls_flags |= NFSLCK_WANTRDELEG;
i &= ~NFSV4OPEN_WANTDELEGMASK;
break;
case NFSV4OPEN_WANTWRITEDELEG:
stp->ls_flags |= NFSLCK_WANTWDELEG;
i &= ~NFSV4OPEN_WANTDELEGMASK;
break;
case NFSV4OPEN_WANTNODELEG:
stp->ls_flags |= NFSLCK_WANTNODELEG;
i &= ~NFSV4OPEN_WANTDELEGMASK;
break;
case NFSV4OPEN_WANTCANCEL:
printf("NFSv4: ignore Open WantCancel\n");
i &= ~NFSV4OPEN_WANTDELEGMASK;
break;
default:
/* nd_repstat will be set to NFSERR_INVAL below. */
break;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
switch (i) {
case NFSV4OPEN_ACCESSREAD:
stp->ls_flags |= NFSLCK_READACCESS;
break;
case NFSV4OPEN_ACCESSWRITE:
stp->ls_flags |= NFSLCK_WRITEACCESS;
break;
case NFSV4OPEN_ACCESSBOTH:
stp->ls_flags |= (NFSLCK_READACCESS | NFSLCK_WRITEACCESS);
break;
default:
nd->nd_repstat = NFSERR_INVAL;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
i = fxdr_unsigned(int, *tl++);
switch (i) {
case NFSV4OPEN_DENYNONE:
break;
case NFSV4OPEN_DENYREAD:
stp->ls_flags |= NFSLCK_READDENY;
break;
case NFSV4OPEN_DENYWRITE:
stp->ls_flags |= NFSLCK_WRITEDENY;
break;
case NFSV4OPEN_DENYBOTH:
stp->ls_flags |= (NFSLCK_READDENY | NFSLCK_WRITEDENY);
break;
default:
nd->nd_repstat = NFSERR_INVAL;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK7 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
error = nfsrv_mtostr(nd, stp->ls_owner, stp->ls_ownerlen);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSVNO_ATTRINIT(&nva);
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
create = fxdr_unsigned(int, *tl);
if (!nd->nd_repstat)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(dp, &dirfor, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (create == NFSV4OPEN_CREATE) {
nva.na_type = VREG;
nva.na_mode = 0;
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
how = fxdr_unsigned(int, *tl);
switch (how) {
case NFSCREATE_UNCHECKED:
case NFSCREATE_GUARDED:
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
error = nfsv4_sattr(nd, NULL, &nva, &attrbits, aclp, p);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
if (error)
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* If the na_gid being set is the same as that of
* the directory it is going in, clear it, since
* that is what will be set by default. This allows
* a user that isn't in that group to do the create.
*/
if (!nd->nd_repstat && NFSVNO_ISSETGID(&nva) &&
nva.na_gid == dirfor.na_gid)
NFSVNO_UNSET(&nva, gid);
if (!nd->nd_repstat)
nd->nd_repstat = nfsrv_checkuidgid(nd, &nva);
break;
case NFSCREATE_EXCLUSIVE:
NFSM_DISSECT(tl, u_int32_t *, NFSX_VERF);
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
cverf[0] = *tl++;
cverf[1] = *tl;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
break;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
case NFSCREATE_EXCLUSIVE41:
NFSM_DISSECT(tl, u_int32_t *, NFSX_VERF);
cverf[0] = *tl++;
cverf[1] = *tl;
Replace "vp" with NULL to make the code more readable. At the time of this nfsv4_sattr() call, "vp == NULL", so this patch doesn't change the semantics, but I think it makes the code more readable. It also makes it consistent with the nfsv4_sattr() call a few lines above this one. Found during code inspection. MFC after: 2 weeks
2019-04-19 23:27:23 +00:00
error = nfsv4_sattr(nd, NULL, &nva, &attrbits, aclp, p);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if (error != 0)
goto nfsmout;
if (NFSISSET_ATTRBIT(&attrbits,
NFSATTRBIT_TIMEACCESSSET))
nd->nd_repstat = NFSERR_INVAL;
/*
* If the na_gid being set is the same as that of
* the directory it is going in, clear it, since
* that is what will be set by default. This allows
* a user that isn't in that group to do the create.
*/
if (nd->nd_repstat == 0 && NFSVNO_ISSETGID(&nva) &&
nva.na_gid == dirfor.na_gid)
NFSVNO_UNSET(&nva, gid);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsrv_checkuidgid(nd, &nva);
break;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
default:
nd->nd_repstat = NFSERR_BADXDR;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else if (create != NFSV4OPEN_NOCREATE) {
nd->nd_repstat = NFSERR_BADXDR;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Now, handle the claim, which usually includes looking up a
* name in the directory referenced by dp. The exception is
* NFSV4OPEN_CLAIMPREVIOUS.
*/
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
claim = fxdr_unsigned(int, *tl);
nfsd: Add support for CLAIM_DELEG_CUR_FH to the NFSv4.1/4.2 Open The Linux NFSv4.1/4.2 client now uses the CLAIM_DELEG_CUR_FH variant of the Open operation when delegations are recalled and the client has a local open of the file. This patch adds support for this variant of Open to the NFSv4.1/4.2 server. This patch only affects mounts from Linux clients when delegations are enabled on the server. MFC after: 2 weeks
2021-05-18 22:53:54 +00:00
if (claim == NFSV4OPEN_CLAIMDELEGATECUR || claim ==
NFSV4OPEN_CLAIMDELEGATECURFH) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID);
stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl,(caddr_t)stateid.other,NFSX_STATEIDOTHER);
stp->ls_flags |= NFSLCK_DELEGCUR;
nfsd: Add support for CLAIM_DELEG_PREV_FH to the NFSv4.1/4.2 Open Commit b3d4c70dc60f added support for CLAIM_DELEG_CUR_FH to Open. While doing this, I noticed that CLAIM_DELEG_PREV_FH support could be added the same way. Although I am not aware of any extant NFSv4.1/4.2 client that uses this claim type, it seems prudent to add support for this variant of Open to the NFSv4.1/4.2 server. This patch does not affect mounts from extant NFSv4.1/4.2 clients, as far as I know. MFC after: 2 weeks
2021-05-21 01:37:40 +00:00
} else if (claim == NFSV4OPEN_CLAIMDELEGATEPREV || claim ==
NFSV4OPEN_CLAIMDELEGATEPREVFH) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
stp->ls_flags |= NFSLCK_DELEGPREV;
}
if (claim == NFSV4OPEN_CLAIMNULL || claim == NFSV4OPEN_CLAIMDELEGATECUR
|| claim == NFSV4OPEN_CLAIMDELEGATEPREV) {
if (!nd->nd_repstat && create == NFSV4OPEN_CREATE &&
claim != NFSV4OPEN_CLAIMNULL)
nd->nd_repstat = NFSERR_INVAL;
if (nd->nd_repstat) {
nd->nd_repstat = nfsrv_opencheck(clientid,
&stateid, stp, NULL, nd, p, nd->nd_repstat);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (create == NFSV4OPEN_CREATE)
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, CREATE,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKPARENT | LOCKLEAF | NOCACHE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
else
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, LOOKUP,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKLEAF);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (error) {
vrele(dp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
if (!nd->nd_repstat) {
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 0, exp,
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
&dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vrele(dp);
nfsvno_relpathbuf(&named);
}
if (create == NFSV4OPEN_CREATE) {
switch (how) {
case NFSCREATE_UNCHECKED:
nfsd: Fix handling of the error case for nfsvno_open Using done_namei instead of ni_startdir did not fix the crashes reported in the PR. Upon looking more closely at the code, the only case where the code near the end of nfsvno_open() needs to be executed is when nfsvno_namei() has succeeded, but a subsequent error was detected. This patch uses done_namei to indicate this case. Also, nfsvno_relpathbuf() should only be called for this case and not whenever nfsvno_open() is called with nd_repstat != 0. A bug was introduced here when the HASBUF flag was deleted. Reviewed by: mjg PR: 268971 Tested by: ish@amail.plala.or.jp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D38430
2023-02-08 21:06:07 +00:00
if (nd->nd_repstat == 0 && named.ni_vp != NULL) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* Clear the setable attribute bits, except
* for Size, if it is being truncated.
*/
NFSZERO_ATTRBIT(&attrbits);
if (NFSVNO_ISSETSIZE(&nva))
NFSSETBIT_ATTRBIT(&attrbits,
NFSATTRBIT_SIZE);
}
break;
case NFSCREATE_GUARDED:
nfsd: Fix handling of the error case for nfsvno_open Using done_namei instead of ni_startdir did not fix the crashes reported in the PR. Upon looking more closely at the code, the only case where the code near the end of nfsvno_open() needs to be executed is when nfsvno_namei() has succeeded, but a subsequent error was detected. This patch uses done_namei to indicate this case. Also, nfsvno_relpathbuf() should only be called for this case and not whenever nfsvno_open() is called with nd_repstat != 0. A bug was introduced here when the HASBUF flag was deleted. Reviewed by: mjg PR: 268971 Tested by: ish@amail.plala.or.jp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D38430
2023-02-08 21:06:07 +00:00
if (nd->nd_repstat == 0 && named.ni_vp != NULL) {
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = EEXIST;
nfsd: Fix handling of the error case for nfsvno_open Using done_namei instead of ni_startdir did not fix the crashes reported in the PR. Upon looking more closely at the code, the only case where the code near the end of nfsvno_open() needs to be executed is when nfsvno_namei() has succeeded, but a subsequent error was detected. This patch uses done_namei to indicate this case. Also, nfsvno_relpathbuf() should only be called for this case and not whenever nfsvno_open() is called with nd_repstat != 0. A bug was introduced here when the HASBUF flag was deleted. Reviewed by: mjg PR: 268971 Tested by: ish@amail.plala.or.jp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D38430
2023-02-08 21:06:07 +00:00
done_namei = true;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
break;
case NFSCREATE_EXCLUSIVE:
exclusive_flag = 1;
nfsd: Fix handling of the error case for nfsvno_open Using done_namei instead of ni_startdir did not fix the crashes reported in the PR. Upon looking more closely at the code, the only case where the code near the end of nfsvno_open() needs to be executed is when nfsvno_namei() has succeeded, but a subsequent error was detected. This patch uses done_namei to indicate this case. Also, nfsvno_relpathbuf() should only be called for this case and not whenever nfsvno_open() is called with nd_repstat != 0. A bug was introduced here when the HASBUF flag was deleted. Reviewed by: mjg PR: 268971 Tested by: ish@amail.plala.or.jp MFC after: 1 week Differential Revision: https://reviews.freebsd.org/D38430
2023-02-08 21:06:07 +00:00
if (nd->nd_repstat == 0 && named.ni_vp == NULL)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nva.na_mode = 0;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
break;
case NFSCREATE_EXCLUSIVE41:
exclusive_flag = 1;
break;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfsvno_open(nd, &named, clientid, &stateid, stp,
&exclusive_flag, &nva, cverf, create, aclp, &attrbits,
nfsserver: Fix vrele() panic in nfsvno_open() Commit 65127e982b94 removed a check for ni_startdir != NULL. This allowed the vrele(ndp->ni_dvp) to be called with a NULL argument. This patch adds a new boolean argument to nfsvno_open() that can be checked instead of ni_startdir, since mjg@ requested that ni_startdir not be used. (Discussed in PR#268828.) PR: 268828 Reviewed by: mjg Differential Revision: https://reviews.freebsd.org/D38032
2023-01-13 00:45:26 +00:00
nd->nd_cred, done_namei, exp, &vp);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
} else if (claim == NFSV4OPEN_CLAIMPREVIOUS || claim ==
nfsd: Add support for CLAIM_DELEG_PREV_FH to the NFSv4.1/4.2 Open Commit b3d4c70dc60f added support for CLAIM_DELEG_CUR_FH to Open. While doing this, I noticed that CLAIM_DELEG_PREV_FH support could be added the same way. Although I am not aware of any extant NFSv4.1/4.2 client that uses this claim type, it seems prudent to add support for this variant of Open to the NFSv4.1/4.2 server. This patch does not affect mounts from extant NFSv4.1/4.2 clients, as far as I know. MFC after: 2 weeks
2021-05-21 01:37:40 +00:00
NFSV4OPEN_CLAIMFH || claim == NFSV4OPEN_CLAIMDELEGATECURFH ||
claim == NFSV4OPEN_CLAIMDELEGATEPREVFH) {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if (claim == NFSV4OPEN_CLAIMPREVIOUS) {
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
i = fxdr_unsigned(int, *tl);
switch (i) {
case NFSV4OPEN_DELEGATEREAD:
stp->ls_flags |= NFSLCK_DELEGREAD;
break;
case NFSV4OPEN_DELEGATEWRITE:
stp->ls_flags |= NFSLCK_DELEGWRITE;
case NFSV4OPEN_DELEGATENONE:
break;
default:
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
stp->ls_flags |= NFSLCK_RECLAIM;
} else {
if (nd->nd_repstat == 0 && create == NFSV4OPEN_CREATE)
nd->nd_repstat = NFSERR_INVAL;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vp = dp;
Simple find/replace of vn_lock -> NFSVOPLOCK. This is done so that NFSVOPLOCK can be modified later to add enhanced logging and assertions. Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:05:31 +00:00
NFSVOPLOCK(vp, LK_EXCLUSIVE | LK_RETRY);
vfs: introduce v_irflag and make v_type smaller The current vnode layout is not smp-friendly by having frequently read data avoidably sharing cachelines with very frequently modified fields. In particular v_iflag inspected for VI_DOOMED can be found in the same line with v_usecount. Instead make it available in the same cacheline as the v_op, v_data and v_type which all get read all the time. v_type is avoidably 4 bytes while the necessary data will easily fit in 1. Shrinking it frees up 3 bytes, 2 of which get used here to introduce a new flag field with a new value: VIRF_DOOMED. Reviewed by: kib, jeff Differential Revision: https://reviews.freebsd.org/D22715
2019-12-08 21:30:04 +00:00
if (!VN_IS_DOOMED(vp))
Add checks for VI_DOOMED and vn_lock() failures to the experimental NFS server, to handle the case where an exported file system is forced dismounted while an RPC is in progress. Further commits will fix the cases where a mount point is used when the associated vnode isn't locked. Reviewed by: kib MFC after: 2 weeks
2011-01-02 19:58:39 +00:00
nd->nd_repstat = nfsrv_opencheck(clientid, &stateid,
stp, vp, nd, p, nd->nd_repstat);
else
nd->nd_repstat = NFSERR_PERM;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
nd->nd_repstat = NFSERR_BADXDR;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
/*
* Do basic access checking.
*/
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (!nd->nd_repstat && vp->v_type != VREG) {
Fix the NFSv4 server so that it returns NFSERR_SYMLINK when an attempt to do an Open operation on any type of file other than VREG is done. A recent discussion on the IETF working group's mailing list (nfsv4@ietf.org) decided that NFSERR_SYMLINK should be returned for all non-regular files and not just symlinks, so that the Linux client would work correctly. This change does not affect the FreeBSD NFSv4 client and is not believed to have a negative effect on other NFSv4 clients. Reviewed by: zkirsch Approved by: re (kib) MFC after: 2 weeks
2011-08-20 21:26:35 +00:00
/*
* The IETF working group decided that this is the correct
* error return for all non-regular files.
*/
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
nd->nd_repstat = (vp->v_type == VDIR) ? NFSERR_ISDIR : NFSERR_SYMLINK;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
/*
* If the Open is being done for a file that already exists, apply
* normal permission checking including for the file owner, if
* vfs.nfsd.v4openaccess is set.
* Previously, the owner was always allowed to open the file to
* be consistent with the NFS tradition of always allowing the
* owner of the file to write to the file regardless of permissions.
* It now appears that the Linux client expects the owner
* permissions to be checked for opens that are not creating the
* file. I believe the correct approach is to use the Access
* operation's results to be consistent with NFSv3, but that is
* not what the current Linux client appears to be doing.
* Since both the Linux and OpenSolaris NFSv4 servers do this check,
nfsd: Allow file owners to perform Open(Delegate_cur) Commit b0b7d978b6a8 changed the NFSv4 server's default behaviour to check the file's mode or ACL for permission to open the file, to be Linux and Solaris compatible. However, it turns out that Linux makes an exception for the case of Claim_delegate_cur(_fh). When a NFSv4 client is returning a delegation, it must acquire Opens against the server to replace the ones done locally in the client. The client does this via an Open operation with Claim_delegate_cur(_fh). If this operation fails, due to a change to the file's mode or ACL after the delegation was issued, the client does not have any way to retain the open. As such, the Linux client allows the file's owner to perform an Open with Claim_delegate_cur(_fh) no matter what the mode or ACL allows. This patch makes the FreeBSD server allow this case, to be Linux compatible. This patch only affects the case where delegations are enabled, which is not the default. MFC after: 2 weeks
2022-02-02 22:10:16 +00:00
* I have enabled it by default. Since Linux does not apply this
* check for claim_delegate_cur, this code does the same.
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
* If this semantic change causes a problem, it can be disabled by
* setting the sysctl vfs.nfsd.v4openaccess to 0 to re-enable the
* previous semantics.
*/
nfsd: Allow file owners to perform Open(Delegate_cur) Commit b0b7d978b6a8 changed the NFSv4 server's default behaviour to check the file's mode or ACL for permission to open the file, to be Linux and Solaris compatible. However, it turns out that Linux makes an exception for the case of Claim_delegate_cur(_fh). When a NFSv4 client is returning a delegation, it must acquire Opens against the server to replace the ones done locally in the client. The client does this via an Open operation with Claim_delegate_cur(_fh). If this operation fails, due to a change to the file's mode or ACL after the delegation was issued, the client does not have any way to retain the open. As such, the Linux client allows the file's owner to perform an Open with Claim_delegate_cur(_fh) no matter what the mode or ACL allows. This patch makes the FreeBSD server allow this case, to be Linux compatible. This patch only affects the case where delegations are enabled, which is not the default. MFC after: 2 weeks
2022-02-02 22:10:16 +00:00
if (nfsrv_openaccess && create == NFSV4OPEN_NOCREATE &&
(stp->ls_flags & NFSLCK_DELEGCUR) == 0)
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
override = NFSACCCHK_NOOVERRIDE;
else
override = NFSACCCHK_ALLOWOWNER;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat && (stp->ls_flags & NFSLCK_WRITEACCESS))
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VWRITE, nd->nd_cred,
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
exp, p, override, NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat && (stp->ls_flags & NFSLCK_READACCESS)) {
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VREAD, nd->nd_cred,
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
exp, p, override, NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (nd->nd_repstat)
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
nd->nd_repstat = nfsvno_accchk(vp, VEXEC,
Fix an interoperability issue w.r.t. the Linux client and the NFSv4 server. Luoqi Chen reported a problem on freebsd-fs@ where a Linux NFSv4 client was able to open and write to a file when the file's permissions were not set to allow the owner write access. Since NFS servers check file permissions on every write RPC, it is standard practice to allow the owner of the file to do writes, regardless of file permissions. This provides POSIX like behaviour, since POSIX only checks permissions upon open(2). The traditional way NFS clients handle this is to check access via the Access operation/RPC and use that to determine if an open(2) on the client is allowed. It appears that, for NFSv4, the Linux client expects the NFSv4 Open (not a POSIX open) operation to fail with NFSERR_ACCES if the file is not being created and file permissions do not allow owner access, unlike NFSv3. Since both the Linux and OpenSolaris NFSv4 servers seem to exhibit this behaviour, this patch changes the FreeBSD NFSv4 server to do the same. A sysctl called vfs.nfsd.v4openaccess can be set to 0 to return the NFSv4 server to its previous behaviour. Since both the Linux and FreeBSD NFSv4 clients seem to exhibit correct behaviour with the access check for file owner in Open enabled, it is enabled by default. Reported by: luoqi.chen@gmail.com MFC after: 2 weeks
2020-04-08 01:12:54 +00:00
nd->nd_cred, exp, p, override,
Modify the experimental server so that it uses VOP_ACCESSX(). This is necessary in order to enable NFSv4 ACL support. The argument to nfsvno_accchk() was changed to an accmode_t and the function nfsrv_aclaccess() was no longer needed and, therefore, deleted. Reviewed by: trasz MFC after: 2 weeks
2009-12-25 20:44:19 +00:00
NFSACCCHK_VPISLOCKED, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
if (!nd->nd_repstat) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Patch the experimental NFS server is a manner analagous to r197525, so that the creation verifier is handled correctly in va_atime for 64bit architectures. There were two problems. One was that the code incorrectly assumed that sizeof (struct timespec) == 8 and the other was that the tv_sec field needs to be assigned from a signed 32bit integer, so that sign extension occurs on 64bit architectures. This is required for correct operation when exporting ZFS volumes. Reviewed by: pjd MFC after: 2 weeks
2009-11-20 21:21:13 +00:00
if (!nd->nd_repstat) {
tverf[0] = nva.na_atime.tv_sec;
tverf[1] = nva.na_atime.tv_nsec;
}
}
if (!nd->nd_repstat && exclusive_flag && (cverf[0] != tverf[0] ||
cverf[1] != tverf[1]))
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = EEXIST;
/*
* Do the open locking/delegation stuff.
*/
if (!nd->nd_repstat)
nd->nd_repstat = nfsrv_openctrl(nd, vp, &stp, clientid, &stateid,
&delegstateid, &rflags, exp, p, nva.na_filerev);
/*
* vp must be unlocked before the call to nfsvno_getattr(dirp,...)
* below, to avoid a deadlock with the lookup in nfsvno_namei() above.
* (ie: Leave the NFSVOPUNLOCK() about here.)
*/
if (vp)
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat && dirp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(dirp, &diraft, nd, p, 0, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/* For NFSv4.1, set the Current StateID. */
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_curstateid = stateid;
nd->nd_flag |= ND_CURSTATEID;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID + 6 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY((caddr_t)stateid.other,(caddr_t)tl,NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
if (claim == NFSV4OPEN_CLAIMPREVIOUS) {
*tl++ = newnfs_true;
*tl++ = 0;
*tl++ = 0;
*tl++ = 0;
*tl++ = 0;
} else {
*tl++ = newnfs_false; /* Since dirp is not locked */
txdr_hyper(dirfor.na_filerev, tl);
tl += 2;
txdr_hyper(diraft.na_filerev, tl);
tl += 2;
}
*tl = txdr_unsigned(rflags & NFSV4OPEN_RFLAGS);
(void) nfsrv_putattrbit(nd, &attrbits);
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
if (rflags & NFSV4OPEN_READDELEGATE)
*tl = txdr_unsigned(NFSV4OPEN_DELEGATEREAD);
else if (rflags & NFSV4OPEN_WRITEDELEGATE)
*tl = txdr_unsigned(NFSV4OPEN_DELEGATEWRITE);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
else if (retext != 0) {
*tl = txdr_unsigned(NFSV4OPEN_DELEGATENONEEXT);
Modify the reasons for not issuing a delegation in the NFSv4.1 server. The ESXi NFSv4.1 client will generate warning messages when the reason for not issuing a delegation is two. Two refers to a resource limit and I do not see why it would be considered invalid. However it probably was not the best choice of reason for not issuing a delegation. This patch changes the reasons used to ones that the ESXi client doesn't complain about. This change does not affect the FreeBSD client and does not appear to affect behaviour of the Linux NFSv4.1 client. RFC5661 defines these "reasons" but does not give any guidance w.r.t. which ones are more appropriate to return to a client. Tested by: andreas.nagy@frequentis.com PR: 226650 MFC after: 2 weeks
2018-07-16 21:32:50 +00:00
if ((rflags & NFSV4OPEN_WDNOTWANTED) != 0) {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(NFSV4OPEN_NOTWANTED);
} else if ((rflags & NFSV4OPEN_WDSUPPFTYPE) != 0) {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(NFSV4OPEN_NOTSUPPFTYPE);
} else if ((rflags & NFSV4OPEN_WDCONTENTION) != 0) {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(NFSV4OPEN_CONTENTION);
*tl = newnfs_false;
} else if ((rflags & NFSV4OPEN_WDRESOURCE) != 0) {
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(NFSV4OPEN_RESOURCE);
*tl = newnfs_false;
} else {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(NFSV4OPEN_NOTWANTED);
}
} else
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
*tl = txdr_unsigned(NFSV4OPEN_DELEGATENONE);
if (rflags & (NFSV4OPEN_READDELEGATE|NFSV4OPEN_WRITEDELEGATE)) {
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID+NFSX_UNSIGNED);
*tl++ = txdr_unsigned(delegstateid.seqid);
NFSBCOPY((caddr_t)delegstateid.other, (caddr_t)tl,
NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
if (rflags & NFSV4OPEN_RECALL)
*tl = newnfs_true;
else
*tl = newnfs_false;
if (rflags & NFSV4OPEN_WRITEDELEGATE) {
NFSM_BUILD(tl, u_int32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(NFSV4OPEN_LIMITSIZE);
txdr_hyper(nva.na_size, tl);
}
NFSM_BUILD(tl, u_int32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(NFSV4ACE_ALLOWEDTYPE);
*tl++ = txdr_unsigned(0x0);
acemask = NFSV4ACE_ALLFILESMASK;
if (nva.na_mode & S_IRUSR)
acemask |= NFSV4ACE_READMASK;
if (nva.na_mode & S_IWUSR)
acemask |= NFSV4ACE_WRITEMASK;
if (nva.na_mode & S_IXUSR)
acemask |= NFSV4ACE_EXECUTEMASK;
*tl = txdr_unsigned(acemask);
(void) nfsm_strtom(nd, "OWNER@", 6);
}
*vpp = vp;
} else if (vp) {
vrele(vp);
}
if (dirp)
vrele(dirp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vrele(dp);
#ifdef NFS4_ACL_EXTATTR_NAME
acl_free(aclp);
#endif
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 close service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_close(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
struct nfsstate st, *stp = &st;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
int error = 0, writeacc;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsv4stateid_t stateid;
nfsquad_t clientid;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
struct nfsvattr na;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED + NFSX_STATEID);
stp->ls_seq = fxdr_unsigned(u_int32_t, *tl++);
stp->ls_ownerlen = 0;
stp->ls_op = nd->nd_rp;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stp->ls_stateid.other,
NFSX_STATEIDOTHER);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* For the special stateid of other all 0s and seqid == 1, set the
* stateid to the current stateid, if it is set.
*/
if ((nd->nd_flag & ND_NFSV41) != 0 && stp->ls_stateid.seqid == 1 &&
stp->ls_stateid.other[0] == 0 && stp->ls_stateid.other[1] == 0 &&
stp->ls_stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0)
stp->ls_stateid = nd->nd_curstateid;
else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
stp->ls_flags = NFSLCK_CLOSE;
clientid.lval[0] = stp->ls_stateid.other[0];
clientid.lval[1] = stp->ls_stateid.other[1];
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK8 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsrv_openupdate(vp, stp, clientid, &stateid, nd, p,
&writeacc);
/* For pNFS, update the attributes. */
if (writeacc != 0 || nfsrv_pnfsatime != 0)
nfsrv_updatemdsattr(vp, &na, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
vput(vp);
if (!nd->nd_repstat) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* If the stateid that has been closed is the current stateid,
* unset it.
*/
if ((nd->nd_flag & ND_CURSTATEID) != 0 &&
stateid.other[0] == nd->nd_curstateid.other[0] &&
stateid.other[1] == nd->nd_curstateid.other[1] &&
stateid.other[2] == nd->nd_curstateid.other[2])
nd->nd_flag &= ~ND_CURSTATEID;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY((caddr_t)stateid.other,(caddr_t)tl,NFSX_STATEIDOTHER);
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 delegpurge service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_delegpurge(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int error = 0;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK9 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nd->nd_repstat = nfsrv_delegupdate(nd, clientid, NULL, NULL,
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
NFSV4OP_DELEGPURGE, nd->nd_cred, p, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmout:
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 delegreturn service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_delegreturn(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
int error = 0, writeacc;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsv4stateid_t stateid;
nfsquad_t clientid;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
struct nfsvattr na;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID);
stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stateid.other, NFSX_STATEIDOTHER);
clientid.lval[0] = stateid.other[0];
clientid.lval[1] = stateid.other[1];
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK10 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nd->nd_repstat = nfsrv_delegupdate(nd, clientid, &stateid, vp,
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
NFSV4OP_DELEGRETURN, nd->nd_cred, p, &writeacc);
/* For pNFS, update the attributes. */
if (writeacc != 0 || nfsrv_pnfsatime != 0)
nfsrv_updatemdsattr(vp, &na, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 get file handle service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_getfh(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
fhandle_t fh;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_repstat = nfsvno_getfh(vp, &fh, p);
vput(vp);
if (!nd->nd_repstat)
nfscl: Make coverity happy Coverity does not like code that checks a function's return value sometimes. Add "(void)" in front of the function when the return value does not matter to try and make it happy. A recent commit deleted "(void)"s in front of nfsm_fhtom(). This commit puts them back in. Reported by: emaste MFC after: 3 months
2023-03-29 00:08:45 +00:00
(void)nfsm_fhtom(NULL, nd, (u_int8_t *)&fh, 0, 0);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
}
/*
* nfsv4 open confirm service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_openconfirm(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
struct nfsstate st, *stp = &st;
int error = 0;
nfsv4stateid_t stateid;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID + NFSX_UNSIGNED);
stp->ls_ownerlen = 0;
stp->ls_op = nd->nd_rp;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stp->ls_stateid.other,
NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
stp->ls_seq = fxdr_unsigned(u_int32_t, *tl);
stp->ls_flags = NFSLCK_CONFIRM;
clientid.lval[0] = stp->ls_stateid.other[0];
clientid.lval[1] = stp->ls_stateid.other[1];
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK11 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsrv_openupdate(vp, stp, clientid, &stateid, nd, p,
NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY((caddr_t)stateid.other,(caddr_t)tl,NFSX_STATEIDOTHER);
}
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 open downgrade service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_opendowngrade(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int i;
struct nfsstate st, *stp = &st;
int error = 0;
nfsv4stateid_t stateid;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Fix failures and warnings reported by newpynfs20090424 test tool. This fix addresses only issues with the pynfs reports, none of these issues are know to create problems for extant real clients. Submitted by: Bart Hsiao <bart.hsiao@gmail.com> Reworked by: myself Reviewed by: rmacklem Approved by: rmacklem Sponsored by: QNAP Systems Inc.
2014-10-03 02:24:41 +00:00
/* opendowngrade can only work on a file object.*/
if (vp->v_type != VREG) {
error = NFSERR_INVAL;
goto nfsmout;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_STATEID + 3 * NFSX_UNSIGNED);
stp->ls_ownerlen = 0;
stp->ls_op = nd->nd_rp;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
NFSBCOPY((caddr_t)tl, (caddr_t)stp->ls_stateid.other,
NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* For the special stateid of other all 0s and seqid == 1, set the
* stateid to the current stateid, if it is set.
*/
if ((nd->nd_flag & ND_NFSV41) != 0 && stp->ls_stateid.seqid == 1 &&
stp->ls_stateid.other[0] == 0 && stp->ls_stateid.other[1] == 0 &&
stp->ls_stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0)
stp->ls_stateid = nd->nd_curstateid;
else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
stp->ls_seq = fxdr_unsigned(u_int32_t, *tl++);
i = fxdr_unsigned(int, *tl++);
Fix OpenDowngrade for NFSv4.1 if a client sets the OPEN_SHARE_ACCESS_WANT* bits. The NFSv4.1 RFC specifies that the OPEN_SHARE_ACCESS_WANT bits can be set in the OpenDowngrade share_access argument and are basically ignored. I do not know of a extant NFSv4.1 client that does this, but this little patch fixes it just in case. It also changes the error from NFSERR_BADXDR to NFSERR_INVAL since the NFSv4.1 RFC specifies this as the error to be returned if bogus bits are set. (The NFSv4.0 RFC didn't specify any error for this, so the error reply can be changed for NFSv4.0 as well.) Found by inspection while looking at a problem with OpenDowngrade reported for the ESXi 6.5 NFSv4.1 client. Reported by: andreas.nagy@frequentis.com PR: 227214 MFC after: 1 week
2018-04-19 20:30:33 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
i &= ~NFSV4OPEN_WANTDELEGMASK;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
switch (i) {
case NFSV4OPEN_ACCESSREAD:
stp->ls_flags = (NFSLCK_READACCESS | NFSLCK_DOWNGRADE);
break;
case NFSV4OPEN_ACCESSWRITE:
stp->ls_flags = (NFSLCK_WRITEACCESS | NFSLCK_DOWNGRADE);
break;
case NFSV4OPEN_ACCESSBOTH:
stp->ls_flags = (NFSLCK_READACCESS | NFSLCK_WRITEACCESS |
NFSLCK_DOWNGRADE);
break;
default:
Fix OpenDowngrade for NFSv4.1 if a client sets the OPEN_SHARE_ACCESS_WANT* bits. The NFSv4.1 RFC specifies that the OPEN_SHARE_ACCESS_WANT bits can be set in the OpenDowngrade share_access argument and are basically ignored. I do not know of a extant NFSv4.1 client that does this, but this little patch fixes it just in case. It also changes the error from NFSERR_BADXDR to NFSERR_INVAL since the NFSv4.1 RFC specifies this as the error to be returned if bogus bits are set. (The NFSv4.0 RFC didn't specify any error for this, so the error reply can be changed for NFSv4.0 as well.) Found by inspection while looking at a problem with OpenDowngrade reported for the ESXi 6.5 NFSv4.1 client. Reported by: andreas.nagy@frequentis.com PR: 227214 MFC after: 1 week
2018-04-19 20:30:33 +00:00
nd->nd_repstat = NFSERR_INVAL;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
i = fxdr_unsigned(int, *tl);
switch (i) {
case NFSV4OPEN_DENYNONE:
break;
case NFSV4OPEN_DENYREAD:
stp->ls_flags |= NFSLCK_READDENY;
break;
case NFSV4OPEN_DENYWRITE:
stp->ls_flags |= NFSLCK_WRITEDENY;
break;
case NFSV4OPEN_DENYBOTH:
stp->ls_flags |= (NFSLCK_READDENY | NFSLCK_WRITEDENY);
break;
default:
Fix OpenDowngrade for NFSv4.1 if a client sets the OPEN_SHARE_ACCESS_WANT* bits. The NFSv4.1 RFC specifies that the OPEN_SHARE_ACCESS_WANT bits can be set in the OpenDowngrade share_access argument and are basically ignored. I do not know of a extant NFSv4.1 client that does this, but this little patch fixes it just in case. It also changes the error from NFSERR_BADXDR to NFSERR_INVAL since the NFSv4.1 RFC specifies this as the error to be returned if bogus bits are set. (The NFSv4.0 RFC didn't specify any error for this, so the error reply can be changed for NFSv4.0 as well.) Found by inspection while looking at a problem with OpenDowngrade reported for the ESXi 6.5 NFSv4.1 client. Reported by: andreas.nagy@frequentis.com PR: 227214 MFC after: 1 week
2018-04-19 20:30:33 +00:00
nd->nd_repstat = NFSERR_INVAL;
Cleanup unnecessary semicolons from the kernel. Found with devel/coccinelle.
2016-04-10 23:07:00 +00:00
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
clientid.lval[0] = stp->ls_stateid.other[0];
clientid.lval[1] = stp->ls_stateid.other[1];
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK12 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
if (!nd->nd_repstat)
nd->nd_repstat = nfsrv_openupdate(vp, stp, clientid, &stateid,
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd, p, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat) {
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/* For NFSv4.1, set the Current StateID. */
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_curstateid = stateid;
nd->nd_flag |= ND_CURSTATEID;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_BUILD(tl, u_int32_t *, NFSX_STATEID);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY((caddr_t)stateid.other,(caddr_t)tl,NFSX_STATEIDOTHER);
}
nfsmout:
vput(vp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 renew lease service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_renew(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int error = 0;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_HYPER);
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK13 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
nd->nd_repstat = nfsrv_getclient(clientid, (CLOPS_RENEWOP|CLOPS_RENEW),
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NULL, NULL, (nfsquad_t)((u_quad_t)0), 0, nd, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmout:
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 security info service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_secinfo(struct nfsrv_descript *nd, int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t dp, struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int len;
struct nameidata named;
vnode_t dirp = NULL, vp;
struct nfsrvfh fh;
struct nfsexstuff retnes;
u_int32_t *sizp;
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
int error = 0, i;
uint64_t savflag;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
char *bufp;
u_long *hashp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* All this just to get the export flags for the name.
*/
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, LOOKUP,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKLEAF);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (error) {
vput(dp);
nfsvno_relpathbuf(&named);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat) {
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 1, exp, &dirp);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
vput(dp);
nfsvno_relpathbuf(&named);
}
if (dirp)
vrele(dirp);
if (nd->nd_repstat)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsvno_relpathbuf(&named);
fh.nfsrvfh_len = NFSX_MYFH;
vp = named.ni_vp;
nd->nd_repstat = nfsvno_getfh(vp, (fhandle_t *)fh.nfsrvfh_data, p);
vput(vp);
savflag = nd->nd_flag;
if (!nd->nd_repstat) {
nfsd: Fix when NFSERR_WRONGSEC may be replied to NFSv4 clients Commit d224f05fcfc1 pre-parsed the next operation number for the put file handle operations. This patch uses this next operation number, plus the type of the file handle being set by the put file handle operation, to implement the rules in RFC5661 Sec. 2.6 with respect to replying NFSERR_WRONGSEC. This patch also adds a check to see if NFSERR_WRONGSEC should be replied when about to perform Lookup, Lookupp or Open with a file name component, so that the NFSERR_WRONGSEC reply is done for these operations, as required by RFC5661 Sec. 2.6. This patch does not have any practical effect for the FreeBSD NFSv4 client and I believe that the same is true for the Linux client, since NFSERR_WRONGSEC is considered a fatal error at this time. MFC after: 2 weeks
2021-06-05 23:53:07 +00:00
/*
* Pretend the next op is Secinfo, so that no wrongsec
* test will be done.
*/
nfsd_fhtovp(nd, &fh, LK_SHARED, &vp, &retnes, NULL, 0,
NFSV4OP_SECINFO);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (vp)
vput(vp);
}
nd->nd_flag = savflag;
if (nd->nd_repstat)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* Finally have the export flags for name, so we can create
* the security info.
*/
len = 0;
NFSM_BUILD(sizp, u_int32_t *, NFSX_UNSIGNED);
nfsd: Fix NFSv4.1/4.2 Secinfo_no_name when security flavors empty Commit 947bd2479ba9 added support for the Secinfo_no_name operation. When a non-exported file system is being traversed, the list of security flavors is empty. It turns out that the Linux client mount attempt fails when the security flavors list in the Secinfo_no_name reply is empty. This patch modifies Secinfo/Secinfo_no_name so that it replies with all four security flavors when the list is empty. This fixes Linux NFSv4.1/4.2 mounts when the file system at the NFSv4 root (as specified on a V4: exports(5) line) is not exported. MFC after: 2 weeks
2021-06-05 03:31:20 +00:00
/* If nes_numsecflavor == 0, all are allowed. */
if (retnes.nes_numsecflavor == 0) {
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(RPCAUTH_UNIX);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl++ = txdr_unsigned(RPCAUTHGSS_SVCNONE);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl++ = txdr_unsigned(RPCAUTHGSS_SVCINTEGRITY);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCPRIVACY);
len = 4;
}
Apply changes to the experimental nfs server so that it uses the security flavors as exported in FreeBSD-CURRENT. This allows it to use a slightly modified mountd.c instead of a different utility. Approved by: kib (mentor)
2009-05-14 21:39:08 +00:00
for (i = 0; i < retnes.nes_numsecflavor; i++) {
if (retnes.nes_secflavors[i] == AUTH_SYS) {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(RPCAUTH_UNIX);
len++;
} else if (retnes.nes_secflavors[i] == RPCSEC_GSS_KRB5) {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl++ = txdr_unsigned(RPCAUTH_GSS);
(void) nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCNONE);
len++;
} else if (retnes.nes_secflavors[i] == RPCSEC_GSS_KRB5I) {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl++ = txdr_unsigned(RPCAUTH_GSS);
(void) nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCINTEGRITY);
len++;
} else if (retnes.nes_secflavors[i] == RPCSEC_GSS_KRB5P) {
NFSM_BUILD(tl, u_int32_t *, NFSX_UNSIGNED);
*tl++ = txdr_unsigned(RPCAUTH_GSS);
(void) nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCPRIVACY);
len++;
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
*sizp = txdr_unsigned(len);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(error, nd);
return (error);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
/*
* nfsv4 security info no name service
*/
int
nfsrvd_secinfononame(struct nfsrv_descript *nd, int isdgram,
vnode_t dp, struct nfsexstuff *exp)
{
uint32_t *tl, *sizp;
struct nameidata named;
vnode_t dirp = NULL, vp;
struct nfsrvfh fh;
struct nfsexstuff retnes;
int error = 0, fhstyle, i, len;
uint64_t savflag;
char *bufp;
u_long *hashp;
struct thread *p = curthread;
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
fhstyle = fxdr_unsigned(int, *tl);
switch (fhstyle) {
case NFSSECINFONONAME_PARENT:
nfsd: Add a sanity check to SecinfoNoname for file type Robert Morris reported that, for the case of SecinfoNoname with the Parent option, providing a non-directory could cause a crash. This patch adds a sanity check for v_type == VDIR for this case, to avoid the crash. Reported by: rtm@lcs.mit.edu PR: 260300 MFC after: 2 weeks
2022-05-01 20:41:31 +00:00
if (dp->v_type != VDIR) {
vput(dp);
nd->nd_repstat = NFSERR_NOTDIR;
goto nfsmout;
}
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
NFSNAMEICNDSET(&named.ni_cnd, nd->nd_cred, LOOKUP,
nfs: stop using SAVESTART Only the name is wanted which is already always provided. Reviewed by: rmacklem Tested by: pho, rmacklem Differential Revision: https://reviews.freebsd.org/D34470
2022-11-10 02:15:04 +00:00
LOCKLEAF);
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
nfsvno_setpathbuf(&named, &bufp, &hashp);
error = nfsrv_parsename(nd, bufp, hashp, &named.ni_pathlen);
if (error != 0) {
vput(dp);
nfsvno_relpathbuf(&named);
goto nfsmout;
}
if (nd->nd_repstat == 0)
nfs: eliminate thread argument from nfsvno_namei This is a step towards retiring struct componentname cn_thread Reviewed by: rmacklem Differential Revision: https://reviews.freebsd.org/D32267
2021-10-01 21:20:15 +00:00
nd->nd_repstat = nfsvno_namei(nd, &named, dp, 1, exp, &dirp);
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
else
vput(dp);
if (dirp != NULL)
vrele(dirp);
nfsvno_relpathbuf(&named);
vp = named.ni_vp;
break;
case NFSSECINFONONAME_CURFH:
vp = dp;
break;
default:
nd->nd_repstat = NFSERR_INVAL;
vput(dp);
}
if (nd->nd_repstat != 0)
goto nfsmout;
fh.nfsrvfh_len = NFSX_MYFH;
nd->nd_repstat = nfsvno_getfh(vp, (fhandle_t *)fh.nfsrvfh_data, p);
vput(vp);
savflag = nd->nd_flag;
if (nd->nd_repstat == 0) {
nfsd: Fix when NFSERR_WRONGSEC may be replied to NFSv4 clients Commit d224f05fcfc1 pre-parsed the next operation number for the put file handle operations. This patch uses this next operation number, plus the type of the file handle being set by the put file handle operation, to implement the rules in RFC5661 Sec. 2.6 with respect to replying NFSERR_WRONGSEC. This patch also adds a check to see if NFSERR_WRONGSEC should be replied when about to perform Lookup, Lookupp or Open with a file name component, so that the NFSERR_WRONGSEC reply is done for these operations, as required by RFC5661 Sec. 2.6. This patch does not have any practical effect for the FreeBSD NFSv4 client and I believe that the same is true for the Linux client, since NFSERR_WRONGSEC is considered a fatal error at this time. MFC after: 2 weeks
2021-06-05 23:53:07 +00:00
/*
* Pretend the next op is Secinfo, so that no wrongsec
* test will be done.
*/
nfsd_fhtovp(nd, &fh, LK_SHARED, &vp, &retnes, NULL, 0,
NFSV4OP_SECINFO);
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
if (vp != NULL)
vput(vp);
}
nd->nd_flag = savflag;
if (nd->nd_repstat != 0)
goto nfsmout;
/*
* Finally have the export flags for fh/parent, so we can create
* the security info.
*/
len = 0;
NFSM_BUILD(sizp, uint32_t *, NFSX_UNSIGNED);
nfsd: Fix NFSv4.1/4.2 Secinfo_no_name when security flavors empty Commit 947bd2479ba9 added support for the Secinfo_no_name operation. When a non-exported file system is being traversed, the list of security flavors is empty. It turns out that the Linux client mount attempt fails when the security flavors list in the Secinfo_no_name reply is empty. This patch modifies Secinfo/Secinfo_no_name so that it replies with all four security flavors when the list is empty. This fixes Linux NFSv4.1/4.2 mounts when the file system at the NFSv4 root (as specified on a V4: exports(5) line) is not exported. MFC after: 2 weeks
2021-06-05 03:31:20 +00:00
/* If nes_numsecflavor == 0, all are allowed. */
if (retnes.nes_numsecflavor == 0) {
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(RPCAUTH_UNIX);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl++ = txdr_unsigned(RPCAUTHGSS_SVCNONE);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 3 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl++ = txdr_unsigned(RPCAUTHGSS_SVCINTEGRITY);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCPRIVACY);
len = 4;
}
nfsd: Add support for the NFSv4.1/4.2 Secinfo_no_name operation The Linux client is now attempting to use the Secinfo_no_name operation for NFSv4.1/4.2 mounts. Although it does not seem to mind the NFSERR_NOTSUPP reply, adding support for it seems reasonable. I also noticed that "savflag" needed to be 64bits in nfsrvd_secinfo() since nd_flag in now 64bits, so I changed the declaration of it there. I also added code to set "vp" NULL after performing Secinfo/Secinfo_no_name, since these operations consume the current FH, which is represented by "vp" in nfsrvd_compound(). Fixing when the server replies NFSERR_WRONGSEC so that it conforms to RFC5661 Sec. 2.6 still needs to be done in a future commit. MFC after: 2 weeks
2021-05-31 00:52:43 +00:00
for (i = 0; i < retnes.nes_numsecflavor; i++) {
if (retnes.nes_secflavors[i] == AUTH_SYS) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(RPCAUTH_UNIX);
len++;
} else if (retnes.nes_secflavors[i] == RPCSEC_GSS_KRB5) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCNONE);
len++;
} else if (retnes.nes_secflavors[i] == RPCSEC_GSS_KRB5I) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCINTEGRITY);
len++;
} else if (retnes.nes_secflavors[i] == RPCSEC_GSS_KRB5P) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(RPCAUTH_GSS);
nfsm_strtom(nd, nfsgss_mechlist[KERBV_MECH].str,
nfsgss_mechlist[KERBV_MECH].len);
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(GSS_KERBV_QOP);
*tl = txdr_unsigned(RPCAUTHGSS_SVCPRIVACY);
len++;
}
}
*sizp = txdr_unsigned(len);
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
/*
* nfsv4 set client id service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_setclientid(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
int i;
int error = 0, idlen;
struct nfsclient *clp = NULL;
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
#ifdef INET
struct sockaddr_in *rin;
#endif
#ifdef INET6
struct sockaddr_in6 *rin6;
#endif
#if defined(INET) || defined(INET6)
u_char *ucp, *ucp2;
#endif
u_char *verf, *addrbuf;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsquad_t clientid, confirm;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto out;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_VERF + NFSX_UNSIGNED);
verf = (u_char *)tl;
tl += (NFSX_VERF / NFSX_UNSIGNED);
i = fxdr_unsigned(int, *tl);
if (i > NFSV4_OPAQUELIMIT || i <= 0) {
nd->nd_repstat = NFSERR_BADXDR;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
idlen = i;
if (nd->nd_flag & ND_GSS)
i += nd->nd_princlen;
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
clp = malloc(sizeof(struct nfsclient) + i, M_NFSDCLIENT, M_WAITOK |
M_ZERO);
clp->lc_stateid = malloc(sizeof(struct nfsstatehead) *
nfsrv_statehashsize, M_NFSDCLIENT, M_WAITOK);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSINITSOCKMUTEX(&clp->lc_req.nr_mtx);
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
/* Allocated large enough for an AF_INET or AF_INET6 socket. */
clp->lc_req.nr_nam = malloc(sizeof(struct sockaddr_in6), M_SONAME,
nfs: Remove NFSSOCKADDRALLOC, NFSSOCKADDRFREE macros They were just thin wrappers over malloc(9) w/ M_ZERO and free(9). Discussed with: rmacklem, markj Sponsored by: Dell EMC Isilon
2018-01-25 22:38:39 +00:00
M_WAITOK | M_ZERO);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
clp->lc_req.nr_cred = NULL;
NFSBCOPY(verf, clp->lc_verf, NFSX_VERF);
clp->lc_idlen = idlen;
error = nfsrv_mtostr(nd, clp->lc_id, idlen);
if (error)
goto nfsmout;
if (nd->nd_flag & ND_GSS) {
clp->lc_flags = LCL_GSS;
if (nd->nd_flag & ND_GSSINTEGRITY)
clp->lc_flags |= LCL_GSSINTEGRITY;
else if (nd->nd_flag & ND_GSSPRIVACY)
clp->lc_flags |= LCL_GSSPRIVACY;
} else {
clp->lc_flags = 0;
}
if ((nd->nd_flag & ND_GSS) && nd->nd_princlen > 0) {
clp->lc_flags |= LCL_NAME;
clp->lc_namelen = nd->nd_princlen;
clp->lc_name = &clp->lc_id[idlen];
NFSBCOPY(nd->nd_principal, clp->lc_name, clp->lc_namelen);
} else {
clp->lc_uid = nd->nd_cred->cr_uid;
clp->lc_gid = nd->nd_cred->cr_gid;
}
Add flags to enable NFS over TLS to the NFS client and server. An Internet Draft titled "Towards Remote Procedure Call Encryption By Default" (soon to be an RFC I think) describes how Sun RPC is to use TLS with NFS as a specific application case. Various commits prepared the NFS code to use KERN_TLS, mainly enabling use of ext_pgs mbufs for large RPC messages. r364475 added TLS support to the kernel RPC. This commit (which is the final one for kernel changes required to do NFS over TLS) adds support for three export flags: MNT_EXTLS - Requires a TLS connection. MNT_EXTLSCERT - Requires a TLS connection where the client presents a valid X.509 certificate during TLS handshake. MNT_EXTLSCERTUSER - Requires a TLS connection where the client presents a valid X.509 certificate with "user@domain" in the otherName field of the SubjectAltName during TLS handshake. Without these export options, clients are permitted, but not required, to use TLS. For the client, a new nmount(2) option called "tls" makes the client do a STARTTLS Null RPC and TLS handshake for all TCP connections used for the mount. The CLSET_TLS client control option is used to indicate to the kernel RPC that this should be done. Unless the above export flags or "tls" option is used, semantics should not change for the NFS client nor server. For NFS over TLS to work, the userspace daemons rpctlscd(8) { for client } or rpctlssd(8) daemon { for server } must be running.
2020-08-27 23:57:30 +00:00
/* If the client is using TLS, do so for the callback connection. */
if (nd->nd_flag & ND_TLS)
clp->lc_flags |= LCL_TLSCB;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
clp->lc_program = fxdr_unsigned(u_int32_t, *tl);
error = nfsrv_getclientipaddr(nd, clp);
if (error)
goto nfsmout;
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
clp->lc_callback = fxdr_unsigned(u_int32_t, *tl);
/*
* nfsrv_setclient() does the actual work of adding it to the
* client list. If there is no error, the structure has been
* linked into the client list and clp should no longer be used
* here. When an error is returned, it has not been linked in,
* so it should be free'd.
*/
nd->nd_repstat = nfsrv_setclient(nd, &clp, &clientid, &confirm, p);
if (nd->nd_repstat == NFSERR_CLIDINUSE) {
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
/*
* 8 is the maximum length of the port# string.
*/
addrbuf = malloc(INET6_ADDRSTRLEN + 8, M_TEMP, M_WAITOK);
switch (clp->lc_req.nr_nam->sa_family) {
#ifdef INET
case AF_INET:
if (clp->lc_flags & LCL_TCPCALLBACK)
(void) nfsm_strtom(nd, "tcp", 3);
else
(void) nfsm_strtom(nd, "udp", 3);
rin = (struct sockaddr_in *)clp->lc_req.nr_nam;
ucp = (u_char *)&rin->sin_addr.s_addr;
ucp2 = (u_char *)&rin->sin_port;
sprintf(addrbuf, "%d.%d.%d.%d.%d.%d", ucp[0] & 0xff,
ucp[1] & 0xff, ucp[2] & 0xff, ucp[3] & 0xff,
ucp2[0] & 0xff, ucp2[1] & 0xff);
break;
#endif
#ifdef INET6
case AF_INET6:
if (clp->lc_flags & LCL_TCPCALLBACK)
(void) nfsm_strtom(nd, "tcp6", 4);
else
(void) nfsm_strtom(nd, "udp6", 4);
rin6 = (struct sockaddr_in6 *)clp->lc_req.nr_nam;
ucp = inet_ntop(AF_INET6, &rin6->sin6_addr, addrbuf,
INET6_ADDRSTRLEN);
if (ucp != NULL)
i = strlen(ucp);
else
i = 0;
ucp2 = (u_char *)&rin6->sin6_port;
sprintf(&addrbuf[i], ".%d.%d", ucp2[0] & 0xff,
ucp2[1] & 0xff);
break;
#endif
}
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
(void) nfsm_strtom(nd, addrbuf, strlen(addrbuf));
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
free(addrbuf, M_TEMP);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (clp) {
nfs: Remove NFSSOCKADDRALLOC, NFSSOCKADDRFREE macros They were just thin wrappers over malloc(9) w/ M_ZERO and free(9). Discussed with: rmacklem, markj Sponsored by: Dell EMC Isilon
2018-01-25 22:38:39 +00:00
free(clp->lc_req.nr_nam, M_SONAME);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSFREEMUTEX(&clp->lc_req.nr_mtx);
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
free(clp->lc_stateid, M_NFSDCLIENT);
free(clp, M_NFSDCLIENT);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
if (!nd->nd_repstat) {
NFSM_BUILD(tl, u_int32_t *, 2 * NFSX_HYPER);
*tl++ = clientid.lval[0];
*tl++ = clientid.lval[1];
*tl++ = confirm.lval[0];
*tl = confirm.lval[1];
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
out:
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
if (clp) {
nfs: Remove NFSSOCKADDRALLOC, NFSSOCKADDRFREE macros They were just thin wrappers over malloc(9) w/ M_ZERO and free(9). Discussed with: rmacklem, markj Sponsored by: Dell EMC Isilon
2018-01-25 22:38:39 +00:00
free(clp->lc_req.nr_nam, M_SONAME);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSFREEMUTEX(&clp->lc_req.nr_mtx);
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
free(clp->lc_stateid, M_NFSDCLIENT);
free(clp, M_NFSDCLIENT);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
}
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 set client id confirm service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_setclientidcfrm(struct nfsrv_descript *nd,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused int isdgram, __unused vnode_t vp,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
__unused struct nfsexstuff *exp)
{
u_int32_t *tl;
int error = 0;
nfsquad_t clientid, confirm;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_HYPER);
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl++;
confirm.lval[0] = *tl++;
confirm.lval[1] = *tl;
/*
* nfsrv_getclient() searches the client list for a match and
* returns the appropriate NFSERR status.
*/
nd->nd_repstat = nfsrv_getclient(clientid, (CLOPS_CONFIRM|CLOPS_RENEW),
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NULL, NULL, confirm, 0, nd, p);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsmout:
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 verify service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_verify(struct nfsrv_descript *nd, int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
int error = 0, ret, fhsize = NFSX_MYFH;
struct nfsvattr nva;
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
struct statfs *sf;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
struct nfsfsinfo fs;
fhandle_t fh;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
sf = malloc(sizeof(struct statfs), M_STATFS, M_WAITOK);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1, NULL);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
nd->nd_repstat = nfsvno_statfs(vp, sf);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!nd->nd_repstat)
nd->nd_repstat = nfsvno_getfh(vp, &fh, p);
if (!nd->nd_repstat) {
nfsvno_getfs(&fs, isdgram);
error = nfsv4_loadattr(nd, vp, &nva, NULL, &fh, fhsize, NULL,
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
sf, NULL, &fs, NULL, 1, &ret, NULL, NULL, p, nd->nd_cred);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
if (!error) {
if (nd->nd_procnum == NFSV4OP_NVERIFY) {
if (ret == 0)
nd->nd_repstat = NFSERR_SAME;
else if (ret != NFSERR_NOTSAME)
nd->nd_repstat = ret;
} else if (ret)
nd->nd_repstat = ret;
}
}
vput(vp);
Do not allocate struct statfs on kernel stack. Right now size of the structure is 472 bytes on amd64, which is already large and stack allocations are indesirable. With the ino64 work, MNAMELEN is increased to 1024, which will make it impossible to have struct statfs on the stack. Extracted from: ino64 work by gleb Discussed with: mckusick Tested by: pho Sponsored by: The FreeBSD Foundation MFC after: 1 week
2017-01-05 17:19:26 +00:00
free(sf, M_STATFS);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfs openattr rpc
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_openattr(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t dp, __unused vnode_t *vpp, __unused fhandle_t *fhp,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
Silence newer gcc warnings. Newer versions of gcc generate "set, but not used" warnings in the NFS server. Add __unused macros to silence these warnings. Requested by: mmacy
2018-07-29 21:51:17 +00:00
int error = 0, createdir __unused;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, NFSX_UNSIGNED);
createdir = fxdr_unsigned(int, *tl);
nd->nd_repstat = NFSERR_NOTSUPP;
nfsmout:
vrele(dp);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
/*
* nfsv4 release lock owner service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nfsrvd_releaselckown(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
{
u_int32_t *tl;
struct nfsstate *stp = NULL;
int error = 0, len;
nfsquad_t clientid;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
NFSM_DISSECT(tl, u_int32_t *, 3 * NFSX_UNSIGNED);
len = fxdr_unsigned(int, *(tl + 2));
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
if (len <= 0 || len > NFSV4_OPAQUELIMIT) {
nd->nd_repstat = NFSERR_BADXDR;
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
goto nfsmout;
Harden the experimental NFS server a little, by adding range checks on the length of the client's open/lock owner name. Also, add free()'s for one case where they were missing and would have caused a leak if NFSERR_BADXDR had been replied. Probably never happens, but the leak is now plugged, just in case. MFC after: 2 weeks
2010-04-06 01:14:49 +00:00
}
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
stp = malloc(sizeof (struct nfsstate) + len,
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
M_NFSDSTATE, M_WAITOK);
stp->ls_ownerlen = len;
stp->ls_op = NULL;
stp->ls_flags = NFSLCK_RELEASE;
stp->ls_uid = nd->nd_cred->cr_uid;
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK14 multiple clids\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
} else {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
error = nfsrv_mtostr(nd, stp->ls_owner, len);
if (error)
goto nfsmout;
nd->nd_repstat = nfsrv_releaselckown(stp, clientid, p);
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(0, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (0);
nfsmout:
if (stp)
style: Remove remaining deprecated MALLOC/FREE macros Mechanically replace uses of MALLOC/FREE with appropriate invocations of malloc(9) / free(9) (a series of sed expressions). Something like: * MALLOC(a, b, ... -> a = malloc(... * FREE( -> free( * free((caddr_t) -> free( No functional change. For now, punt on modifying contrib ipfilter code, leaving a definition of the macro in its KMALLOC(). Reported by: jhb Reviewed by: cy, imp, markj, rmacklem Sponsored by: Dell EMC Isilon Differential Revision: https://reviews.freebsd.org/D14035
2018-01-25 22:25:13 +00:00
free(stp, M_NFSDSTATE);
Add DEXITCODE plumbing to NFS. Isilon has the concept of an in-memory exit-code ring that saves the last exit code of a function and allows for stack tracing. This is very helpful when debugging tough issues. This patch is essentially a no-op for BSD at this point, until we upstream the dexitcode logic itself. The patch adds DEXITCODE calls to every NFS function that returns an errno error code. A number of code paths were also reorganized to have single exit paths, to reduce code duplication. Submitted by: David Kwan <dkwan@isilon.com> Reviewed by: rmacklem Approved by: zml (mentor) MFC after: 2 weeks
2011-07-16 08:51:09 +00:00
NFSEXITCODE2(error, nd);
Add the experimental nfs subtree to the kernel, that includes support for NFSv4 as well as NFSv2 and 3. It lives in 3 subdirs under sys/fs: nfs - functions that are common to the client and server nfsclient - a mutation of sys/nfsclient that call generic functions to do RPCs and handle state. As such, it retains the buffer cache handling characteristics and vnode semantics that are found in sys/nfsclient, for the most part. nfsserver - the server. It includes a DRC designed specifically for NFSv4, that is used instead of the generic DRC in sys/rpc. The build glue will be checked in later, so at this point, it consists of 3 new subdirs that should not affect kernel building. Approved by: kib (mentor)
2009-05-04 15:23:58 +00:00
return (error);
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
/*
* nfsv4 exchange_id service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_exchangeid(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint32_t *tl;
int error = 0, i, idlen;
struct nfsclient *clp = NULL;
nfsquad_t clientid, confirm;
uint8_t *verf;
uint32_t sp4type, v41flags;
struct timespec verstime;
nfsd: Add support for the SP4_MACH_CRED case in ExchangeID Commit f4179ad46fa4 added support for operation bitmaps for NFSv4.1/4.2. This commit uses those to implement the SP4_MACH_CRED case for the NFSv4.1/4.2 ExchangeID operation since the Linux NFSv4.1/4.2 client is now using this for Kerberized mounts. The Linux Kerberized NFSv4.1/4.2 mounts currently work without support for this because Linux will fall back to SP4_NONE, but there is no guarantee this fallback will work forever. This commit only affects Kerberized NFSv4.1/4.2 mounts from Linux at this time. MFC after: 3 months
2023-04-07 19:49:23 +00:00
nfsopbit_t mustops, allowops;
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
#ifdef INET
struct sockaddr_in *sin, *rin;
#endif
#ifdef INET6
struct sockaddr_in6 *sin6, *rin6;
#endif
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Allow setting NFS server scope and owner. By default NFS server reports as scope and owner major the host UUID value and zero for owner minor. It works good in case of standalone server. But in case of CARP-based HA cluster failover the values should remain persistent, otherwise some clients like VMware ESXi get confused by the change and fail to reconnect automatically. The patch makes server scope, major owner and minor owner values configurable via sysctls. If not set (by default) the host UUID value is still used. Reviewed by: rmacklem MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D31952
2021-09-14 18:14:30 +00:00
char *s;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
goto nfsmout;
NFSM_DISSECT(tl, u_int32_t *, NFSX_VERF + NFSX_UNSIGNED);
verf = (uint8_t *)tl;
tl += (NFSX_VERF / NFSX_UNSIGNED);
i = fxdr_unsigned(int, *tl);
if (i > NFSV4_OPAQUELIMIT || i <= 0) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
idlen = i;
if (nd->nd_flag & ND_GSS)
i += nd->nd_princlen;
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
clp = malloc(sizeof(struct nfsclient) + i, M_NFSDCLIENT, M_WAITOK |
M_ZERO);
clp->lc_stateid = malloc(sizeof(struct nfsstatehead) *
nfsrv_statehashsize, M_NFSDCLIENT, M_WAITOK);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NFSINITSOCKMUTEX(&clp->lc_req.nr_mtx);
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
/* Allocated large enough for an AF_INET or AF_INET6 socket. */
clp->lc_req.nr_nam = malloc(sizeof(struct sockaddr_in6), M_SONAME,
nfs: Remove NFSSOCKADDRALLOC, NFSSOCKADDRFREE macros They were just thin wrappers over malloc(9) w/ M_ZERO and free(9). Discussed with: rmacklem, markj Sponsored by: Dell EMC Isilon
2018-01-25 22:38:39 +00:00
M_WAITOK | M_ZERO);
Add support for INET6 addresses to the kernel code that dumps open/lock state. PR#223036 reported that INET6 callback addresses were not printed by nfsdumpstate(8). This kernel patch adds INET6 addresses to the dump structure, so that nfsdumpstate(8) can print them out, post-r346190. The patch also includes the addition of #ifdef INET, INET6 as requested by bz@. PR: 223036 Reviewed by: bz, rgrimes MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D19839
2019-04-13 22:00:09 +00:00
switch (nd->nd_nam->sa_family) {
#ifdef INET
case AF_INET:
rin = (struct sockaddr_in *)clp->lc_req.nr_nam;
sin = (struct sockaddr_in *)nd->nd_nam;
rin->sin_family = AF_INET;
rin->sin_len = sizeof(struct sockaddr_in);
rin->sin_port = 0;
rin->sin_addr.s_addr = sin->sin_addr.s_addr;
break;
#endif
#ifdef INET6
case AF_INET6:
rin6 = (struct sockaddr_in6 *)clp->lc_req.nr_nam;
sin6 = (struct sockaddr_in6 *)nd->nd_nam;
rin6->sin6_family = AF_INET6;
rin6->sin6_len = sizeof(struct sockaddr_in6);
rin6->sin6_port = 0;
rin6->sin6_addr = sin6->sin6_addr;
break;
#endif
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
clp->lc_req.nr_cred = NULL;
NFSBCOPY(verf, clp->lc_verf, NFSX_VERF);
clp->lc_idlen = idlen;
error = nfsrv_mtostr(nd, clp->lc_id, idlen);
if (error != 0)
goto nfsmout;
if ((nd->nd_flag & ND_GSS) != 0) {
clp->lc_flags = LCL_GSS | LCL_NFSV41;
if ((nd->nd_flag & ND_GSSINTEGRITY) != 0)
clp->lc_flags |= LCL_GSSINTEGRITY;
else if ((nd->nd_flag & ND_GSSPRIVACY) != 0)
clp->lc_flags |= LCL_GSSPRIVACY;
} else
clp->lc_flags = LCL_NFSV41;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if ((nd->nd_flag & ND_NFSV42) != 0)
clp->lc_flags |= LCL_NFSV42;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if ((nd->nd_flag & ND_GSS) != 0 && nd->nd_princlen > 0) {
clp->lc_flags |= LCL_NAME;
clp->lc_namelen = nd->nd_princlen;
clp->lc_name = &clp->lc_id[idlen];
NFSBCOPY(nd->nd_principal, clp->lc_name, clp->lc_namelen);
} else {
clp->lc_uid = nd->nd_cred->cr_uid;
clp->lc_gid = nd->nd_cred->cr_gid;
}
NFSM_DISSECT(tl, u_int32_t *, 2 * NFSX_UNSIGNED);
v41flags = fxdr_unsigned(uint32_t, *tl++);
if ((v41flags & ~(NFSV4EXCH_SUPPMOVEDREFER | NFSV4EXCH_SUPPMOVEDMIGR |
NFSV4EXCH_BINDPRINCSTATEID | NFSV4EXCH_MASKPNFS |
NFSV4EXCH_UPDCONFIRMEDRECA)) != 0) {
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
if ((v41flags & NFSV4EXCH_UPDCONFIRMEDRECA) != 0)
confirm.lval[1] = 1;
else
confirm.lval[1] = 0;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
if (nfsrv_devidcnt == 0)
v41flags = NFSV4EXCH_USENONPNFS | NFSV4EXCH_USEPNFSDS;
else
v41flags = NFSV4EXCH_USEPNFSMDS;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
sp4type = fxdr_unsigned(uint32_t, *tl);
nfsd: Add support for the SP4_MACH_CRED case in ExchangeID Commit f4179ad46fa4 added support for operation bitmaps for NFSv4.1/4.2. This commit uses those to implement the SP4_MACH_CRED case for the NFSv4.1/4.2 ExchangeID operation since the Linux NFSv4.1/4.2 client is now using this for Kerberized mounts. The Linux Kerberized NFSv4.1/4.2 mounts currently work without support for this because Linux will fall back to SP4_NONE, but there is no guarantee this fallback will work forever. This commit only affects Kerberized NFSv4.1/4.2 mounts from Linux at this time. MFC after: 3 months
2023-04-07 19:49:23 +00:00
if (sp4type == NFSV4EXCH_SP4MACHCRED) {
if ((nd->nd_flag & (ND_GSSINTEGRITY | ND_GSSPRIVACY)) == 0 ||
nd->nd_princlen == 0)
nd->nd_repstat = (NFSERR_AUTHERR | AUTH_TOOWEAK);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsrv_getopbits(nd, &mustops, NULL);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsrv_getopbits(nd, &allowops, NULL);
if (nd->nd_repstat != 0)
goto nfsmout;
NFSOPBIT_CLRNOTMUST(&mustops);
NFSSET_OPBIT(&clp->lc_mustops, &mustops);
NFSOPBIT_CLRNOTALLOWED(&allowops);
NFSSET_OPBIT(&clp->lc_allowops, &allowops);
clp->lc_flags |= LCL_MACHCRED;
} else if (sp4type != NFSV4EXCH_SP4NONE) {
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
/*
* nfsrv_setclient() does the actual work of adding it to the
* client list. If there is no error, the structure has been
* linked into the client list and clp should no longer be used
* here. When an error is returned, it has not been linked in,
* so it should be free'd.
*/
nd->nd_repstat = nfsrv_setclient(nd, &clp, &clientid, &confirm, p);
if (clp != NULL) {
nfs: Remove NFSSOCKADDRALLOC, NFSSOCKADDRFREE macros They were just thin wrappers over malloc(9) w/ M_ZERO and free(9). Discussed with: rmacklem, markj Sponsored by: Dell EMC Isilon
2018-01-25 22:38:39 +00:00
free(clp->lc_req.nr_nam, M_SONAME);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NFSFREEMUTEX(&clp->lc_req.nr_mtx);
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
free(clp->lc_stateid, M_NFSDCLIENT);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
free(clp, M_NFSDCLIENT);
}
if (nd->nd_repstat == 0) {
if (confirm.lval[1] != 0)
v41flags |= NFSV4EXCH_CONFIRMEDR;
nfsd: Add support for the SP4_MACH_CRED case in ExchangeID Commit f4179ad46fa4 added support for operation bitmaps for NFSv4.1/4.2. This commit uses those to implement the SP4_MACH_CRED case for the NFSv4.1/4.2 ExchangeID operation since the Linux NFSv4.1/4.2 client is now using this for Kerberized mounts. The Linux Kerberized NFSv4.1/4.2 mounts currently work without support for this because Linux will fall back to SP4_NONE, but there is no guarantee this fallback will work forever. This commit only affects Kerberized NFSv4.1/4.2 mounts from Linux at this time. MFC after: 3 months
2023-04-07 19:49:23 +00:00
NFSM_BUILD(tl, uint32_t *, NFSX_HYPER + 3 * NFSX_UNSIGNED);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
*tl++ = clientid.lval[0]; /* ClientID */
*tl++ = clientid.lval[1];
*tl++ = txdr_unsigned(confirm.lval[0]); /* SequenceID */
*tl++ = txdr_unsigned(v41flags); /* Exch flags */
nfsd: Add support for the SP4_MACH_CRED case in ExchangeID Commit f4179ad46fa4 added support for operation bitmaps for NFSv4.1/4.2. This commit uses those to implement the SP4_MACH_CRED case for the NFSv4.1/4.2 ExchangeID operation since the Linux NFSv4.1/4.2 client is now using this for Kerberized mounts. The Linux Kerberized NFSv4.1/4.2 mounts currently work without support for this because Linux will fall back to SP4_NONE, but there is no guarantee this fallback will work forever. This commit only affects Kerberized NFSv4.1/4.2 mounts from Linux at this time. MFC after: 3 months
2023-04-07 19:49:23 +00:00
*tl = txdr_unsigned(sp4type); /* No SSV */
if (sp4type == NFSV4EXCH_SP4MACHCRED) {
nfsrv_putopbit(nd, &mustops);
nfsrv_putopbit(nd, &allowops);
}
NFSM_BUILD(tl, uint32_t *, NFSX_HYPER);
Allow setting NFS server scope and owner. By default NFS server reports as scope and owner major the host UUID value and zero for owner minor. It works good in case of standalone server. But in case of CARP-based HA cluster failover the values should remain persistent, otherwise some clients like VMware ESXi get confused by the change and fail to reconnect automatically. The patch makes server scope, major owner and minor owner values configurable via sysctls. If not set (by default) the host UUID value is still used. Reviewed by: rmacklem MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D31952
2021-09-14 18:14:30 +00:00
txdr_hyper(nfsrv_owner_minor, tl); /* Owner Minor */
if (nfsrv_owner_major[0] != 0)
s = nfsrv_owner_major;
else
s = nd->nd_cred->cr_prison->pr_hostuuid;
nfsm_strtom(nd, s, strlen(s)); /* Owner Major */
if (nfsrv_scope[0] != 0)
s = nfsrv_scope;
else
s = nd->nd_cred->cr_prison->pr_hostuuid;
nfsm_strtom(nd, s, strlen(s) ); /* Scope */
Fix the eir_server_scope reply argument for NFSv4.1 ExchangeID. In the reply to an ExchangeID operation, the NFSv4.1 server returns a "scope" value (eir_server_scope). If this value is the same, it indicates that two servers share state, which is never the case for FreeBSD servers. As such, the value needs to be unique and it was without this patch. However, I just found out that it is not supposed to change when the server reboots and without this patch, it did change. This patch fixes eir_server_scope so that it does not change when the server is rebooted. The only affect not having this patch has is that Linux clients don't reclaim opens and locks after a server reboot, which meant they lost any byte range locks held before the server rebooted. It only affects NFSv4.1 mounts and the FreeBSD NFSv4.1 client was not affected by this bug. MFC after: 1 week
2018-05-13 23:38:01 +00:00
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
*tl = txdr_unsigned(1);
(void)nfsm_strtom(nd, "freebsd.org", strlen("freebsd.org"));
(void)nfsm_strtom(nd, version, strlen(version));
NFSM_BUILD(tl, uint32_t *, NFSX_V4TIME);
verstime.tv_sec = 1293840000; /* Jan 1, 2011 */
verstime.tv_nsec = 0;
txdr_nfsv4time(&verstime, tl);
}
NFSEXITCODE2(0, nd);
return (0);
nfsmout:
if (clp != NULL) {
nfs: Remove NFSSOCKADDRALLOC, NFSSOCKADDRFREE macros They were just thin wrappers over malloc(9) w/ M_ZERO and free(9). Discussed with: rmacklem, markj Sponsored by: Dell EMC Isilon
2018-01-25 22:38:39 +00:00
free(clp->lc_req.nr_nam, M_SONAME);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NFSFREEMUTEX(&clp->lc_req.nr_mtx);
Make the size of the hash tables used by the NFSv4 server tunable. No appreciable change in performance was observed after increasing the sizes of these tables and then testing with a single client. However, there was an email that indicated high CPU overheads for a heavily loaded NFSv4 and it is hoped that increasing the sizes of the hash tables via these tunables might help. The tables remain the same size by default. Differential Revision: https://reviews.freebsd.org/D2596 MFC after: 2 weeks
2015-05-27 22:00:05 +00:00
free(clp->lc_stateid, M_NFSDCLIENT);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
free(clp, M_NFSDCLIENT);
}
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 create session service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_createsession(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint32_t *tl;
int error = 0;
nfsquad_t clientid, confirm;
struct nfsdsession *sep = NULL;
uint32_t rdmacnt;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
nfsd: Add sysctl to set maximum I/O size up to 1Mbyte Since MAXPHYS now allows the FreeBSD NFS client to do 1Mbyte I/O operations, add a sysctl called vfs.nfsd.srvmaxio so that the maximum NFS server I/O size can be set up to 1Mbyte. The Linux NFS client can also do 1Mbyte I/O operations. The default of 128Kbytes for the maximum I/O size has not been changed for two reasons: - kern.ipc.maxsockbuf must be increased to support 1Mbyte I/O - The limited benchmarking I can do actually shows a drop in I/O rate when the I/O size is above 256Kbytes. However, daveb@spectralogic.com reports seeing an increase in I/O rate for the 1Mbyte I/O size vs 128Kbytes using a Linux client. Reviewed by: asomers MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D30826
2021-07-16 22:01:03 +00:00
static bool do_printf = true;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
goto nfsmout;
sep = (struct nfsdsession *)malloc(sizeof(struct nfsdsession),
M_NFSDSESSION, M_WAITOK | M_ZERO);
sep->sess_refcnt = 1;
mtx_init(&sep->sess_cbsess.nfsess_mtx, "nfscbsession", NULL, MTX_DEF);
NFSM_DISSECT(tl, uint32_t *, NFSX_HYPER + 2 * NFSX_UNSIGNED);
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl++;
confirm.lval[0] = fxdr_unsigned(uint32_t, *tl++);
sep->sess_crflags = fxdr_unsigned(uint32_t, *tl);
/* Persistent sessions and RDMA are not supported. */
sep->sess_crflags &= NFSV4CRSESS_CONNBACKCHAN;
/* Fore channel attributes. */
NFSM_DISSECT(tl, uint32_t *, 7 * NFSX_UNSIGNED);
tl++; /* Header pad always 0. */
sep->sess_maxreq = fxdr_unsigned(uint32_t, *tl++);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
if (sep->sess_maxreq > sb_max_adj - NFS_MAXXDR) {
sep->sess_maxreq = sb_max_adj - NFS_MAXXDR;
nfsd: Add sysctl to set maximum I/O size up to 1Mbyte Since MAXPHYS now allows the FreeBSD NFS client to do 1Mbyte I/O operations, add a sysctl called vfs.nfsd.srvmaxio so that the maximum NFS server I/O size can be set up to 1Mbyte. The Linux NFS client can also do 1Mbyte I/O operations. The default of 128Kbytes for the maximum I/O size has not been changed for two reasons: - kern.ipc.maxsockbuf must be increased to support 1Mbyte I/O - The limited benchmarking I can do actually shows a drop in I/O rate when the I/O size is above 256Kbytes. However, daveb@spectralogic.com reports seeing an increase in I/O rate for the 1Mbyte I/O size vs 128Kbytes using a Linux client. Reviewed by: asomers MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D30826
2021-07-16 22:01:03 +00:00
if (do_printf)
printf("Consider increasing kern.ipc.maxsockbuf\n");
do_printf = false;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
sep->sess_maxresp = fxdr_unsigned(uint32_t, *tl++);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
if (sep->sess_maxresp > sb_max_adj - NFS_MAXXDR) {
sep->sess_maxresp = sb_max_adj - NFS_MAXXDR;
nfsd: Add sysctl to set maximum I/O size up to 1Mbyte Since MAXPHYS now allows the FreeBSD NFS client to do 1Mbyte I/O operations, add a sysctl called vfs.nfsd.srvmaxio so that the maximum NFS server I/O size can be set up to 1Mbyte. The Linux NFS client can also do 1Mbyte I/O operations. The default of 128Kbytes for the maximum I/O size has not been changed for two reasons: - kern.ipc.maxsockbuf must be increased to support 1Mbyte I/O - The limited benchmarking I can do actually shows a drop in I/O rate when the I/O size is above 256Kbytes. However, daveb@spectralogic.com reports seeing an increase in I/O rate for the 1Mbyte I/O size vs 128Kbytes using a Linux client. Reviewed by: asomers MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D30826
2021-07-16 22:01:03 +00:00
if (do_printf)
printf("Consider increasing kern.ipc.maxsockbuf\n");
do_printf = false;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
sep->sess_maxrespcached = fxdr_unsigned(uint32_t, *tl++);
sep->sess_maxops = fxdr_unsigned(uint32_t, *tl++);
sep->sess_maxslots = fxdr_unsigned(uint32_t, *tl++);
if (sep->sess_maxslots > NFSV4_SLOTS)
sep->sess_maxslots = NFSV4_SLOTS;
rdmacnt = fxdr_unsigned(uint32_t, *tl);
if (rdmacnt > 1) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
} else if (rdmacnt == 1)
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
/* Back channel attributes. */
NFSM_DISSECT(tl, uint32_t *, 7 * NFSX_UNSIGNED);
tl++; /* Header pad always 0. */
sep->sess_cbmaxreq = fxdr_unsigned(uint32_t, *tl++);
sep->sess_cbmaxresp = fxdr_unsigned(uint32_t, *tl++);
sep->sess_cbmaxrespcached = fxdr_unsigned(uint32_t, *tl++);
sep->sess_cbmaxops = fxdr_unsigned(uint32_t, *tl++);
sep->sess_cbsess.nfsess_foreslots = fxdr_unsigned(uint32_t, *tl++);
rdmacnt = fxdr_unsigned(uint32_t, *tl);
if (rdmacnt > 1) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
} else if (rdmacnt == 1)
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
sep->sess_cbprogram = fxdr_unsigned(uint32_t, *tl);
/*
* nfsrv_getclient() searches the client list for a match and
* returns the appropriate NFSERR status.
*/
nd->nd_repstat = nfsrv_getclient(clientid, CLOPS_CONFIRM | CLOPS_RENEW,
NULL, sep, confirm, sep->sess_cbprogram, nd, p);
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_V4SESSIONID);
NFSBCOPY(sep->sess_sessionid, tl, NFSX_V4SESSIONID);
NFSM_BUILD(tl, uint32_t *, 18 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(confirm.lval[0]); /* sequenceid */
*tl++ = txdr_unsigned(sep->sess_crflags);
/* Fore channel attributes. */
*tl++ = 0;
*tl++ = txdr_unsigned(sep->sess_maxreq);
*tl++ = txdr_unsigned(sep->sess_maxresp);
*tl++ = txdr_unsigned(sep->sess_maxrespcached);
*tl++ = txdr_unsigned(sep->sess_maxops);
*tl++ = txdr_unsigned(sep->sess_maxslots);
*tl++ = txdr_unsigned(1);
*tl++ = txdr_unsigned(0); /* No RDMA. */
/* Back channel attributes. */
*tl++ = 0;
*tl++ = txdr_unsigned(sep->sess_cbmaxreq);
*tl++ = txdr_unsigned(sep->sess_cbmaxresp);
*tl++ = txdr_unsigned(sep->sess_cbmaxrespcached);
*tl++ = txdr_unsigned(sep->sess_cbmaxops);
*tl++ = txdr_unsigned(sep->sess_cbsess.nfsess_foreslots);
*tl++ = txdr_unsigned(1);
*tl = txdr_unsigned(0); /* No RDMA. */
}
nfsmout:
if (nd->nd_repstat != 0 && sep != NULL)
free(sep, M_NFSDSESSION);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 sequence service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_sequence(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint32_t *tl;
uint32_t highest_slotid, sequenceid, sflags, target_highest_slotid;
int cache_this, error = 0;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
goto nfsmout;
NFSM_DISSECT(tl, uint32_t *, NFSX_V4SESSIONID);
NFSBCOPY(tl, nd->nd_sessionid, NFSX_V4SESSIONID);
NFSM_DISSECT(tl, uint32_t *, 4 * NFSX_UNSIGNED);
sequenceid = fxdr_unsigned(uint32_t, *tl++);
nd->nd_slotid = fxdr_unsigned(uint32_t, *tl++);
highest_slotid = fxdr_unsigned(uint32_t, *tl++);
if (*tl == newnfs_true)
cache_this = 1;
else
cache_this = 0;
nd->nd_repstat = nfsrv_checksequence(nd, sequenceid, &highest_slotid,
&target_highest_slotid, cache_this, &sflags, p);
Revert "wpa: Import wpa_supplicant/hostapd commit 14ab4a816" This reverts commit 266f97b5e9a7958e365e78288616a459b40d924a, reversing changes made to a10253cffea84c0c980a36ba6776b00ed96c3e3b. A mismerge of a merge to catch up to main resulted in files being committed which should not have been.
2021-12-02 22:45:04 +00:00
if (nd->nd_repstat != NFSERR_BADSLOT)
nd->nd_flag |= ND_HASSEQUENCE;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_V4SESSIONID);
NFSBCOPY(nd->nd_sessionid, tl, NFSX_V4SESSIONID);
NFSM_BUILD(tl, uint32_t *, 5 * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(sequenceid);
*tl++ = txdr_unsigned(nd->nd_slotid);
*tl++ = txdr_unsigned(highest_slotid);
*tl++ = txdr_unsigned(target_highest_slotid);
*tl = txdr_unsigned(sflags);
}
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 reclaim complete service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_reclaimcomplete(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint32_t *tl;
Modify the NFSv4.1 server so that it allows ReclaimComplete as done by ESXi 6.7. I believe that a ReclaimComplete with rca_one_fs == TRUE is only to be used after a file system has been transferred to a different file server. However, RFC5661 is somewhat vague w.r.t. this and the ESXi 6.7 client does both a ReclaimComplete with rca_one_fs == TRUE and one with ReclaimComplete with rca_one_fs == FALSE. Therefore, just ignore the rca_one_fs == TRUE operation and return NFS_OK without doing anything instead of replying NFS4ERR_NOTSUPP. This allows the ESXi 6.7 NFSv4.1 client to do a mount. After discussion on the NFSv4 IETF working group mailing list, doing this along with setting a flag to note that a ReclaimComplete with rca_one_fs TRUE was an appropriate way to handle this. The flag that indicates that a ReclaimComplete with rca_one_fs == TRUE was done may be used to disable replies of NFS4ERR_GRACE for non-reclaim state operations in a future commit. This patch along with r332790, r334492 and r336357 allow ESXi 6.7 NFSv4.1 mounts work ok. ESX 6.5 NFSv4.1 mounts do not work well, due to what I believe are violations of RFC-5661 and should not be used. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com, daniel@ftml.net (earlier version) MFC after: 2 weeks Relnotes: yes
2018-07-28 20:21:04 +00:00
int error = 0, onefs;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
Modify the NFSv4.1 server so that it allows ReclaimComplete as done by ESXi 6.7. I believe that a ReclaimComplete with rca_one_fs == TRUE is only to be used after a file system has been transferred to a different file server. However, RFC5661 is somewhat vague w.r.t. this and the ESXi 6.7 client does both a ReclaimComplete with rca_one_fs == TRUE and one with ReclaimComplete with rca_one_fs == FALSE. Therefore, just ignore the rca_one_fs == TRUE operation and return NFS_OK without doing anything instead of replying NFS4ERR_NOTSUPP. This allows the ESXi 6.7 NFSv4.1 client to do a mount. After discussion on the NFSv4 IETF working group mailing list, doing this along with setting a flag to note that a ReclaimComplete with rca_one_fs TRUE was an appropriate way to handle this. The flag that indicates that a ReclaimComplete with rca_one_fs == TRUE was done may be used to disable replies of NFS4ERR_GRACE for non-reclaim state operations in a future commit. This patch along with r332790, r334492 and r336357 allow ESXi 6.7 NFSv4.1 mounts work ok. ESX 6.5 NFSv4.1 mounts do not work well, due to what I believe are violations of RFC-5661 and should not be used. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com, daniel@ftml.net (earlier version) MFC after: 2 weeks Relnotes: yes
2018-07-28 20:21:04 +00:00
/*
* I believe that a ReclaimComplete with rca_one_fs == TRUE is only
* to be used after a file system has been transferred to a different
* file server. However, RFC5661 is somewhat vague w.r.t. this and
* the ESXi 6.7 client does both a ReclaimComplete with rca_one_fs
* == TRUE and one with ReclaimComplete with rca_one_fs == FALSE.
* Therefore, just ignore the rca_one_fs == TRUE operation and return
* NFS_OK without doing anything.
*/
onefs = 0;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
if (*tl == newnfs_true)
Modify the NFSv4.1 server so that it allows ReclaimComplete as done by ESXi 6.7. I believe that a ReclaimComplete with rca_one_fs == TRUE is only to be used after a file system has been transferred to a different file server. However, RFC5661 is somewhat vague w.r.t. this and the ESXi 6.7 client does both a ReclaimComplete with rca_one_fs == TRUE and one with ReclaimComplete with rca_one_fs == FALSE. Therefore, just ignore the rca_one_fs == TRUE operation and return NFS_OK without doing anything instead of replying NFS4ERR_NOTSUPP. This allows the ESXi 6.7 NFSv4.1 client to do a mount. After discussion on the NFSv4 IETF working group mailing list, doing this along with setting a flag to note that a ReclaimComplete with rca_one_fs TRUE was an appropriate way to handle this. The flag that indicates that a ReclaimComplete with rca_one_fs == TRUE was done may be used to disable replies of NFS4ERR_GRACE for non-reclaim state operations in a future commit. This patch along with r332790, r334492 and r336357 allow ESXi 6.7 NFSv4.1 mounts work ok. ESX 6.5 NFSv4.1 mounts do not work well, due to what I believe are violations of RFC-5661 and should not be used. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com, daniel@ftml.net (earlier version) MFC after: 2 weeks Relnotes: yes
2018-07-28 20:21:04 +00:00
onefs = 1;
nd->nd_repstat = nfsrv_checkreclaimcomplete(nd, onefs);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 destroy clientid service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_destroyclientid(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint32_t *tl;
nfsquad_t clientid;
int error = 0;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
goto nfsmout;
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_UNSIGNED);
clientid.lval[0] = *tl++;
clientid.lval[1] = *tl;
nfsd: Add support for the SP4_MACH_CRED case in ExchangeID Commit f4179ad46fa4 added support for operation bitmaps for NFSv4.1/4.2. This commit uses those to implement the SP4_MACH_CRED case for the NFSv4.1/4.2 ExchangeID operation since the Linux NFSv4.1/4.2 client is now using this for Kerberized mounts. The Linux Kerberized NFSv4.1/4.2 mounts currently work without support for this because Linux will fall back to SP4_NONE, but there is no guarantee this fallback will work forever. This commit only affects Kerberized NFSv4.1/4.2 mounts from Linux at this time. MFC after: 3 months
2023-04-07 19:49:23 +00:00
nd->nd_repstat = nfsrv_destroyclient(nd, clientid, p);
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
Add the BindConnectiontoSession operation to the NFSv4.1 server. Under some fairly unusual circumstances, the Linux NFSv4.1 client is doing a BindConnectiontoSession operation for TCP connections. It is also used by the ESXi6.5 NFSv4.1 client. This patch adds this operation to the NFSv4.1 server. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com MFC after: 2 weeks
2018-06-01 19:47:41 +00:00
/*
* nfsv4 bind connection to session service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add the BindConnectiontoSession operation to the NFSv4.1 server. Under some fairly unusual circumstances, the Linux NFSv4.1 client is doing a BindConnectiontoSession operation for TCP connections. It is also used by the ESXi6.5 NFSv4.1 client. This patch adds this operation to the NFSv4.1 server. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com MFC after: 2 weeks
2018-06-01 19:47:41 +00:00
nfsrvd_bindconnsess(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Add the BindConnectiontoSession operation to the NFSv4.1 server. Under some fairly unusual circumstances, the Linux NFSv4.1 client is doing a BindConnectiontoSession operation for TCP connections. It is also used by the ESXi6.5 NFSv4.1 client. This patch adds this operation to the NFSv4.1 server. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com MFC after: 2 weeks
2018-06-01 19:47:41 +00:00
{
uint32_t *tl;
uint8_t sessid[NFSX_V4SESSIONID];
int error = 0, foreaft;
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Add the BindConnectiontoSession operation to the NFSv4.1 server. Under some fairly unusual circumstances, the Linux NFSv4.1 client is doing a BindConnectiontoSession operation for TCP connections. It is also used by the ESXi6.5 NFSv4.1 client. This patch adds this operation to the NFSv4.1 server. Reported by: andreas.nagy@frequentis.com Tested by: andreas.nagy@frequentis.com MFC after: 2 weeks
2018-06-01 19:47:41 +00:00
goto nfsmout;
NFSM_DISSECT(tl, uint32_t *, NFSX_V4SESSIONID + 2 * NFSX_UNSIGNED);
NFSBCOPY(tl, sessid, NFSX_V4SESSIONID);
tl += (NFSX_V4SESSIONID / NFSX_UNSIGNED);
foreaft = fxdr_unsigned(int, *tl++);
if (*tl == newnfs_true) {
/* RDMA is not supported. */
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
nd->nd_repstat = nfsrv_bindconnsess(nd, sessid, &foreaft);
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_V4SESSIONID + 2 *
NFSX_UNSIGNED);
NFSBCOPY(sessid, tl, NFSX_V4SESSIONID);
tl += (NFSX_V4SESSIONID / NFSX_UNSIGNED);
*tl++ = txdr_unsigned(foreaft);
*tl = newnfs_false;
}
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
/*
* nfsv4 destroy session service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_destroysession(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint8_t *cp, sessid[NFSX_V4SESSIONID];
int error = 0;
nfsd: Fix the failure return for non-fh NFSv4 operations Without this patch, nfsd_checkrootexp() returns failure and then the NFSv4 operation would reply NFSERR_WRONGSEC. RFC5661 Sec. 2.6 only allows a few NFSv4 operations, none of which call nfsv4_checktootexp(), to return NFSERR_WRONGSEC. This patch modifies nfsd_checkrootexp() to return the error instead of a boolean and sets the returned error to an RPC layer AUTH_ERR, as discussed on nfsv4@ietf.org. The patch also fixes nfsd_errmap() so that the pseudo error NFSERR_AUTHERR is handled correctly such that an RPC layer AUTH_ERR is replied to the NFSv4 client. The two new "enum auth_stat" values have not yet been assigned by IANA, but are the expected next two values. The effect on extant NFSv4 clients of this change appears limited to reporting a different failure error when a mount that does not use adequate security is attempted. MFC after: 2 weeks
2021-06-02 22:28:07 +00:00
if ((nd->nd_repstat = nfsd_checkrootexp(nd)) != 0)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
goto nfsmout;
NFSM_DISSECT(cp, uint8_t *, NFSX_V4SESSIONID);
NFSBCOPY(cp, sessid, NFSX_V4SESSIONID);
nd->nd_repstat = nfsrv_destroysession(nd, sessid);
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 free stateid service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_freestateid(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
uint32_t *tl;
nfsv4stateid_t stateid;
int error = 0;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
NFSM_DISSECT(tl, uint32_t *, NFSX_STATEID);
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* For the special stateid of other all 0s and seqid == 1, set the
* stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nd->nd_repstat = nfsrv_freestateid(nd, &stateid, p);
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/* If the current stateid has been free'd, unset it. */
if (nd->nd_repstat == 0 && (nd->nd_flag & ND_CURSTATEID) != 0 &&
stateid.other[0] == nd->nd_curstateid.other[0] &&
stateid.other[1] == nd->nd_curstateid.other[1] &&
stateid.other[2] == nd->nd_curstateid.other[2])
nd->nd_flag &= ~ND_CURSTATEID;
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 layoutget service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nfsrvd_layoutget(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
{
uint32_t *tl;
nfsv4stateid_t stateid;
int error = 0, layoutlen, layouttype, iomode, maxcnt, retonclose;
uint64_t offset, len, minlen;
char *layp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
NFSM_DISSECT(tl, uint32_t *, 4 * NFSX_UNSIGNED + 3 * NFSX_HYPER +
NFSX_STATEID);
tl++; /* Signal layout available. Ignore for now. */
layouttype = fxdr_unsigned(int, *tl++);
iomode = fxdr_unsigned(int, *tl++);
offset = fxdr_hyper(tl); tl += 2;
len = fxdr_hyper(tl); tl += 2;
minlen = fxdr_hyper(tl); tl += 2;
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
maxcnt = fxdr_unsigned(int, *tl);
NFSD_DEBUG(4, "layoutget ltyp=%d iom=%d off=%ju len=%ju mlen=%ju\n",
layouttype, iomode, (uintmax_t)offset, (uintmax_t)len,
(uintmax_t)minlen);
if (len < minlen ||
(minlen != UINT64_MAX && offset + minlen < offset) ||
(len != UINT64_MAX && offset + len < offset)) {
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
/*
* For the special stateid of other all 0s and seqid == 1, set the
* stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
layp = NULL;
if (layouttype == NFSLAYOUT_NFSV4_1_FILES && nfsrv_maxpnfsmirror == 1)
layp = malloc(NFSX_V4FILELAYOUT, M_TEMP, M_WAITOK);
else if (layouttype == NFSLAYOUT_FLEXFILE)
layp = malloc(NFSX_V4FLEXLAYOUT(nfsrv_maxpnfsmirror), M_TEMP,
M_WAITOK);
else
nd->nd_repstat = NFSERR_UNKNLAYOUTTYPE;
if (layp != NULL)
nd->nd_repstat = nfsrv_layoutget(nd, vp, exp, layouttype,
&iomode, &offset, &len, minlen, &stateid, maxcnt,
&retonclose, &layoutlen, layp, nd->nd_cred, p);
NFSD_DEBUG(4, "nfsrv_layoutget stat=%u layoutlen=%d\n", nd->nd_repstat,
layoutlen);
if (nd->nd_repstat == 0) {
/* For NFSv4.1, set the Current StateID. */
if ((nd->nd_flag & ND_NFSV41) != 0) {
nd->nd_curstateid = stateid;
nd->nd_flag |= ND_CURSTATEID;
}
NFSM_BUILD(tl, uint32_t *, 4 * NFSX_UNSIGNED + NFSX_STATEID +
2 * NFSX_HYPER);
*tl++ = txdr_unsigned(retonclose);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY(stateid.other, tl, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
*tl++ = txdr_unsigned(1); /* Only returns one layout. */
txdr_hyper(offset, tl); tl += 2;
txdr_hyper(len, tl); tl += 2;
*tl++ = txdr_unsigned(iomode);
*tl = txdr_unsigned(layouttype);
nfsm_strtom(nd, layp, layoutlen);
} else if (nd->nd_repstat == NFSERR_LAYOUTTRYLATER) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = newnfs_false;
}
free(layp, M_TEMP);
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 layoutcommit service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nfsrvd_layoutcommit(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
{
uint32_t *tl;
nfsv4stateid_t stateid;
int error = 0, hasnewoff, hasnewmtime, layouttype, maxcnt, reclaim;
int hasnewsize;
Silence some "might not be initialized" warnings for riscv64. None of these case were actually using the variable(s) uninitialized, but I figured that silencing the warnings via initializing them made sense. Some of these predated r355677.
2019-12-13 21:38:08 +00:00
uint64_t offset, len, newoff = 0, newsize;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
struct timespec newmtime;
char *layp;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
layp = NULL;
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_UNSIGNED + 2 * NFSX_HYPER +
NFSX_STATEID);
offset = fxdr_hyper(tl); tl += 2;
len = fxdr_hyper(tl); tl += 2;
reclaim = fxdr_unsigned(int, *tl++);
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
/*
* For the special stateid of other all 0s and seqid == 1, set the
* stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
hasnewoff = fxdr_unsigned(int, *tl);
if (hasnewoff != 0) {
NFSM_DISSECT(tl, uint32_t *, NFSX_HYPER + NFSX_UNSIGNED);
newoff = fxdr_hyper(tl); tl += 2;
} else
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
hasnewmtime = fxdr_unsigned(int, *tl);
if (hasnewmtime != 0) {
NFSM_DISSECT(tl, uint32_t *, NFSX_V4TIME + 2 * NFSX_UNSIGNED);
fxdr_nfsv4time(tl, &newmtime);
tl += (NFSX_V4TIME / NFSX_UNSIGNED);
} else
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_UNSIGNED);
layouttype = fxdr_unsigned(int, *tl++);
maxcnt = fxdr_unsigned(int, *tl);
if (maxcnt > 0) {
layp = malloc(maxcnt + 1, M_TEMP, M_WAITOK);
error = nfsrv_mtostr(nd, layp, maxcnt);
if (error != 0)
goto nfsmout;
}
nd->nd_repstat = nfsrv_layoutcommit(nd, vp, layouttype, hasnewoff,
newoff, offset, len, hasnewmtime, &newmtime, reclaim, &stateid,
maxcnt, layp, &hasnewsize, &newsize, nd->nd_cred, p);
NFSD_DEBUG(4, "nfsrv_layoutcommit stat=%u\n", nd->nd_repstat);
if (nd->nd_repstat == 0) {
if (hasnewsize != 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED + NFSX_HYPER);
*tl++ = newnfs_true;
txdr_hyper(newsize, tl);
} else {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = newnfs_false;
}
}
nfsmout:
free(layp, M_TEMP);
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 layoutreturn service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nfsrvd_layoutreturn(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
vnode_t vp, struct nfsexstuff *exp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
{
uint32_t *tl, *layp;
nfsv4stateid_t stateid;
int error = 0, fnd, kind, layouttype, iomode, maxcnt, reclaim;
uint64_t offset, len;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
layp = NULL;
NFSM_DISSECT(tl, uint32_t *, 4 * NFSX_UNSIGNED);
reclaim = *tl++;
layouttype = fxdr_unsigned(int, *tl++);
iomode = fxdr_unsigned(int, *tl++);
kind = fxdr_unsigned(int, *tl);
NFSD_DEBUG(4, "layoutreturn recl=%d ltyp=%d iom=%d kind=%d\n", reclaim,
layouttype, iomode, kind);
if (kind == NFSV4LAYOUTRET_FILE) {
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_HYPER + NFSX_STATEID +
NFSX_UNSIGNED);
offset = fxdr_hyper(tl); tl += 2;
len = fxdr_hyper(tl); tl += 2;
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
/*
* For the special stateid of other all 0s and seqid == 1, set
* the stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
maxcnt = fxdr_unsigned(int, *tl);
nfsd: Add checks for layout errors in LayoutReturn For a LayoutReturn when using the Flexible File Layout, error reports may be provided in the request. Sanity check the size of these error reports and check that they exist before calling nfsrv_flexlayouterr(). Reported by: rtm@lcs.mit.edu Tested by: rtm@lcs.mit.edu PR: 260012 MFC after: 2 weeks
2021-11-26 23:42:32 +00:00
/*
* There is no fixed upper bound defined in the RFCs,
* but 128Kbytes should be more than sufficient.
*/
if (maxcnt < 0 || maxcnt > 131072)
maxcnt = 0;
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
if (maxcnt > 0) {
layp = malloc(maxcnt + 1, M_TEMP, M_WAITOK);
error = nfsrv_mtostr(nd, (char *)layp, maxcnt);
if (error != 0)
goto nfsmout;
}
} else {
if (reclaim == newnfs_true) {
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
offset = len = 0;
maxcnt = 0;
}
nd->nd_repstat = nfsrv_layoutreturn(nd, vp, layouttype, iomode,
offset, len, reclaim, kind, &stateid, maxcnt, layp, &fnd,
nd->nd_cred, p);
NFSD_DEBUG(4, "nfsrv_layoutreturn stat=%u fnd=%d\n", nd->nd_repstat,
fnd);
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
if (fnd != 0) {
*tl = newnfs_true;
NFSM_BUILD(tl, uint32_t *, NFSX_STATEID);
*tl++ = txdr_unsigned(stateid.seqid);
NFSBCOPY(stateid.other, tl, NFSX_STATEIDOTHER);
} else
*tl = newnfs_false;
}
nfsmout:
free(layp, M_TEMP);
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
/*
* nfsv4 layout error service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_layouterror(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, struct nfsexstuff *exp)
{
uint32_t *tl;
nfsv4stateid_t stateid;
int cnt, error = 0, i, stat;
int opnum __unused;
char devid[NFSX_V4DEVICEID];
uint64_t offset, len;
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_HYPER + NFSX_STATEID +
NFSX_UNSIGNED);
offset = fxdr_hyper(tl); tl += 2;
len = fxdr_hyper(tl); tl += 2;
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
cnt = fxdr_unsigned(int, *tl);
NFSD_DEBUG(4, "layouterror off=%ju len=%ju cnt=%d\n", (uintmax_t)offset,
(uintmax_t)len, cnt);
/*
* For the special stateid of other all 0s and seqid == 1, set
* the stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
/*
* Ignore offset, len and stateid for now.
*/
for (i = 0; i < cnt; i++) {
NFSM_DISSECT(tl, uint32_t *, NFSX_V4DEVICEID + 2 *
NFSX_UNSIGNED);
NFSBCOPY(tl, devid, NFSX_V4DEVICEID);
tl += (NFSX_V4DEVICEID / NFSX_UNSIGNED);
stat = fxdr_unsigned(int, *tl++);
opnum = fxdr_unsigned(int, *tl);
NFSD_DEBUG(4, "nfsrvd_layouterr op=%d stat=%d\n", opnum, stat);
/*
nfsd: Fix the NFSv4 pNFS MDS server for DS NFSERR_NOSPC If a pNFS server's DS runs out of disk space, it replies NFSERR_NOSPC to the client doing writing. For the Linux client, it then sends a LayoutError RPC to the server to tell it about the error and keeps retrying, doing repeated LayoutGet and Write RPCs to the DS. The Linux client is "stuck" until disk space on the DS is free'd up. For a mirrored server configuration, the first mirror that ran out of space was taken offline. This does not make much sense, since the other mirror(s) will run out of space soon and the fix is a manual cleanup up disk space. This patch changes the pNFS server to not disable a mirror for the mirrored case when this occurs. Further work is needed, since the Linux client expects the MDS to reply NFSERR_NOSPC to LayoutGets once the DS is out of space. Without this further change, the above mentioned looping occurs. Found during a recent IEFT NFSv4 working group testing event. MFC after: 2 weeks
2021-11-07 19:43:03 +00:00
* Except for NFSERR_ACCES, NFSERR_STALE and NFSERR_NOSPC
* errors, disable the mirror.
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
*/
nfsd: Fix the NFSv4 pNFS MDS server for DS NFSERR_NOSPC If a pNFS server's DS runs out of disk space, it replies NFSERR_NOSPC to the client doing writing. For the Linux client, it then sends a LayoutError RPC to the server to tell it about the error and keeps retrying, doing repeated LayoutGet and Write RPCs to the DS. The Linux client is "stuck" until disk space on the DS is free'd up. For a mirrored server configuration, the first mirror that ran out of space was taken offline. This does not make much sense, since the other mirror(s) will run out of space soon and the fix is a manual cleanup up disk space. This patch changes the pNFS server to not disable a mirror for the mirrored case when this occurs. Further work is needed, since the Linux client expects the MDS to reply NFSERR_NOSPC to LayoutGets once the DS is out of space. Without this further change, the above mentioned looping occurs. Found during a recent IEFT NFSv4 working group testing event. MFC after: 2 weeks
2021-11-07 19:43:03 +00:00
if (stat != NFSERR_ACCES && stat != NFSERR_STALE &&
stat != NFSERR_NOSPC)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrv_delds(devid, curthread);
nfsd: Fix the NFSv4.2 pNFS MDS server for NFSERR_NOSPC via LayoutError If a pNFS server's DS runs out of disk space, it replies NFSERR_NOSPC to the client doing writing. For the Linux client, it then sends a LayoutError RPC to the MDS server to tell it about the error and keeps retrying, doing repeated LayoutGets to the MDS and Write RPCs to the DS. The Linux client is "stuck" until disk space on the DS is free'd up unless a subsequent LayoutGet request is sent a NFSERR_NOSPC reply. The looping problem still occurs for NFSv4.1 mounts, but no fix for this is known at this time. This patch changes the pNFS MDS server to reply to LayoutGet operations with NFSERR_NOSPC once a LayoutError reports the problem, until the DS has available space. This keeps the Linux NFSv4.2 from looping. Found during recent testing because of issues w.r.t. a DS being out of space found during a recent IEFT NFSv4 working group testing event. MFC after: 2 weeks
2021-11-08 23:58:00 +00:00
/* For NFSERR_NOSPC, mark all deviceids and layouts. */
if (stat == NFSERR_NOSPC)
nfsrv_marknospc(devid, true);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
}
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 layout stats service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_layoutstats(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, struct nfsexstuff *exp)
{
uint32_t *tl;
nfsv4stateid_t stateid;
int cnt, error = 0;
int layouttype __unused;
char devid[NFSX_V4DEVICEID] __unused;
nfs: Quiet a few "unused" warnings For most of these warnings, the variable is loaded with data parsed out of an RPC messages. In case the data is useful in the future, I just marked these with __unused.
2021-11-28 23:48:51 +00:00
uint64_t offset __unused, len __unused, readcount __unused;
uint64_t readbytes __unused, writecount __unused, writebytes __unused;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
NFSM_DISSECT(tl, uint32_t *, 6 * NFSX_HYPER + NFSX_STATEID +
NFSX_V4DEVICEID + 2 * NFSX_UNSIGNED);
offset = fxdr_hyper(tl); tl += 2;
len = fxdr_hyper(tl); tl += 2;
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
readcount = fxdr_hyper(tl); tl += 2;
readbytes = fxdr_hyper(tl); tl += 2;
writecount = fxdr_hyper(tl); tl += 2;
writebytes = fxdr_hyper(tl); tl += 2;
NFSBCOPY(tl, devid, NFSX_V4DEVICEID);
tl += (NFSX_V4DEVICEID / NFSX_UNSIGNED);
layouttype = fxdr_unsigned(int, *tl++);
cnt = fxdr_unsigned(int, *tl);
error = nfsm_advance(nd, NFSM_RNDUP(cnt), -1);
if (error != 0)
goto nfsmout;
NFSD_DEBUG(4, "layoutstats cnt=%d\n", cnt);
/*
* For the special stateid of other all 0s and seqid == 1, set
* the stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
/*
* No use for the stats for now.
*/
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfsv4 io_advise service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_ioadvise(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, struct nfsexstuff *exp)
{
uint32_t *tl;
nfsv4stateid_t stateid;
nfsattrbit_t hints;
int error = 0, ret;
off_t offset, len;
NFSM_DISSECT(tl, uint32_t *, NFSX_STATEID + 2 * NFSX_HYPER);
stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, stateid.other, NFSX_STATEIDOTHER);
tl += (NFSX_STATEIDOTHER / NFSX_UNSIGNED);
offset = fxdr_hyper(tl); tl += 2;
len = fxdr_hyper(tl);
error = nfsrv_getattrbits(nd, &hints, NULL, NULL);
if (error != 0)
goto nfsmout;
/*
* For the special stateid of other all 0s and seqid == 1, set
* the stateid to the current stateid, if it is set.
*/
if (stateid.seqid == 1 && stateid.other[0] == 0 &&
stateid.other[1] == 0 && stateid.other[2] == 0) {
if ((nd->nd_flag & ND_CURSTATEID) != 0) {
stateid = nd->nd_curstateid;
stateid.seqid = 0;
} else {
nd->nd_repstat = NFSERR_BADSTATEID;
goto nfsmout;
}
}
if (offset < 0) {
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
if (len < 0)
len = 0;
if (vp->v_type != VREG) {
if (vp->v_type == VDIR)
nd->nd_repstat = NFSERR_ISDIR;
else
nd->nd_repstat = NFSERR_WRONGTYPE;
goto nfsmout;
}
/*
* For now, we can only handle WILLNEED and DONTNEED and don't use
* the stateid.
*/
if ((NFSISSET_ATTRBIT(&hints, NFSV4IOHINT_WILLNEED) &&
!NFSISSET_ATTRBIT(&hints, NFSV4IOHINT_DONTNEED)) ||
(NFSISSET_ATTRBIT(&hints, NFSV4IOHINT_DONTNEED) &&
!NFSISSET_ATTRBIT(&hints, NFSV4IOHINT_WILLNEED))) {
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (NFSISSET_ATTRBIT(&hints, NFSV4IOHINT_WILLNEED)) {
ret = VOP_ADVISE(vp, offset, len, POSIX_FADV_WILLNEED);
NFSZERO_ATTRBIT(&hints);
if (ret == 0)
NFSSETBIT_ATTRBIT(&hints, NFSV4IOHINT_WILLNEED);
else
NFSSETBIT_ATTRBIT(&hints, NFSV4IOHINT_NORMAL);
} else {
ret = VOP_ADVISE(vp, offset, len, POSIX_FADV_DONTNEED);
NFSZERO_ATTRBIT(&hints);
if (ret == 0)
NFSSETBIT_ATTRBIT(&hints, NFSV4IOHINT_DONTNEED);
else
NFSSETBIT_ATTRBIT(&hints, NFSV4IOHINT_NORMAL);
}
vrele(vp);
} else {
NFSZERO_ATTRBIT(&hints);
NFSSETBIT_ATTRBIT(&hints, NFSV4IOHINT_NORMAL);
vput(vp);
}
nfsrv_putattrbit(nd, &hints);
NFSEXITCODE2(error, nd);
return (error);
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
/*
* nfsv4 getdeviceinfo service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
nfsrvd_getdevinfo(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the pNFS server code from projects/pnfs-planb-server into head. This code merge adds a pNFS service to the NFSv4.1 server. Although it is a large commit it should not affect behaviour for a non-pNFS NFS server. Some documentation on how this works can be found at: http://people.freebsd.org/~rmacklem/pnfs-planb-setup.txt and will hopefully be turned into a proper document soon. This is a merge of the kernel code. Userland and man page changes will come soon, once the dust settles on this merge. It has passed a "make universe", so I hope it will not cause build problems. It also adds NFSv4.1 server support for the "current stateid". Here is a brief overview of the pNFS service: A pNFS service separates the Read/Write oeprations from all the other NFSv4.1 Metadata operations. It is hoped that this separation allows a pNFS service to be configured that exceeds the limits of a single NFS server for either storage capacity and/or I/O bandwidth. It is possible to configure mirroring within the data servers (DSs) so that the data storage file for an MDS file will be mirrored on two or more of the DSs. When this is used, failure of a DS will not stop the pNFS service and a failed DS can be recovered once repaired while the pNFS service continues to operate. Although two way mirroring would be the norm, it is possible to set a mirroring level of up to four or the number of DSs, whichever is less. The Metadata server will always be a single point of failure, just as a single NFS server is. A Plan B pNFS service consists of a single MetaData Server (MDS) and K Data Servers (DS), all of which are recent FreeBSD systems. Clients will mount the MDS as they would a single NFS server. When files are created, the MDS creates a file tree identical to what a single NFS server creates, except that all the regular (VREG) files will be empty. As such, if you look at the exported tree on the MDS directly on the MDS server (not via an NFS mount), the files will all be of size 0. Each of these files will also have two extended attributes in the system attribute name space: pnfsd.dsfile - This extended attrbute stores the information that the MDS needs to find the data storage file(s) on DS(s) for this file. pnfsd.dsattr - This extended attribute stores the Size, AccessTime, ModifyTime and Change attributes for the file, so that the MDS doesn't need to acquire the attributes from the DS for every Getattr operation. For each regular (VREG) file, the MDS creates a data storage file on one (or more if mirroring is enabled) of the DSs in one of the "dsNN" subdirectories. The name of this file is the file handle of the file on the MDS in hexadecimal so that the name is unique. The DSs use subdirectories named "ds0" to "dsN" so that no one directory gets too large. The value of "N" is set via the sysctl vfs.nfsd.dsdirsize on the MDS, with the default being 20. For production servers that will store a lot of files, this value should probably be much larger. It can be increased when the "nfsd" daemon is not running on the MDS, once the "dsK" directories are created. For pNFS aware NFSv4.1 clients, the FreeBSD server will return two pieces of information to the client that allows it to do I/O directly to the DS. DeviceInfo - This is relatively static information that defines what a DS is. The critical bits of information returned by the FreeBSD server is the IP address of the DS and, for the Flexible File layout, that NFSv4.1 is to be used and that it is "tightly coupled". There is a "deviceid" which identifies the DeviceInfo. Layout - This is per file and can be recalled by the server when it is no longer valid. For the FreeBSD server, there is support for two types of layout, call File and Flexible File layout. Both allow the client to do I/O on the DS via NFSv4.1 I/O operations. The Flexible File layout is a more recent variant that allows specification of mirrors, where the client is expected to do writes to all mirrors to maintain them in a consistent state. The Flexible File layout also allows the client to report I/O errors for a DS back to the MDS. The Flexible File layout supports two variants referred to as "tightly coupled" vs "loosely coupled". The FreeBSD server always uses the "tightly coupled" variant where the client uses the same credentials to do I/O on the DS as it would on the MDS. For the "loosely coupled" variant, the layout specifies a synthetic user/group that the client uses to do I/O on the DS. The FreeBSD server does not do striping and always returns layouts for the entire file. The critical information in a layout is Read vs Read/Writea and DeviceID(s) that identify which DS(s) the data is stored on. At this time, the MDS generates File Layout layouts to NFSv4.1 clients that know how to do pNFS for the non-mirrored DS case unless the sysctl vfs.nfsd.default_flexfile is set non-zero, in which case Flexible File layouts are generated. The mirrored DS configuration always generates Flexible File layouts. For NFS clients that do not support NFSv4.1 pNFS, all I/O operations are done against the MDS which acts as a proxy for the appropriate DS(s). When the MDS receives an I/O RPC, it will do the RPC on the DS as a proxy. If the DS is on the same machine, the MDS/DS will do the RPC on the DS as a proxy and so on, until the machine runs out of some resource, such as session slots or mbufs. As such, DSs must be separate systems from the MDS. Tested by: james.rose@framestore.com Relnotes: yes
2018-06-12 19:36:32 +00:00
{
uint32_t *tl, maxcnt, notify[NFSV4_NOTIFYBITMAP];
int cnt, devaddrlen, error = 0, i, layouttype;
char devid[NFSX_V4DEVICEID], *devaddr;
time_t dev_time;
NFSM_DISSECT(tl, uint32_t *, 3 * NFSX_UNSIGNED + NFSX_V4DEVICEID);
NFSBCOPY(tl, devid, NFSX_V4DEVICEID);
tl += (NFSX_V4DEVICEID / NFSX_UNSIGNED);
layouttype = fxdr_unsigned(int, *tl++);
maxcnt = fxdr_unsigned(uint32_t, *tl++);
cnt = fxdr_unsigned(int, *tl);
NFSD_DEBUG(4, "getdevinfo ltyp=%d maxcnt=%u bitcnt=%d\n", layouttype,
maxcnt, cnt);
if (cnt > NFSV4_NOTIFYBITMAP || cnt < 0) {
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
if (cnt > 0) {
NFSM_DISSECT(tl, uint32_t *, cnt * NFSX_UNSIGNED);
for (i = 0; i < cnt; i++)
notify[i] = fxdr_unsigned(uint32_t, *tl++);
}
for (i = cnt; i < NFSV4_NOTIFYBITMAP; i++)
notify[i] = 0;
/*
* Check that the device id is not stale. Device ids are recreated
* each time the nfsd threads are restarted.
*/
NFSBCOPY(devid, &dev_time, sizeof(dev_time));
if (dev_time != nfsdev_time) {
nd->nd_repstat = NFSERR_NOENT;
goto nfsmout;
}
/* Look for the device id. */
nd->nd_repstat = nfsrv_getdevinfo(devid, layouttype, &maxcnt,
notify, &devaddrlen, &devaddr);
NFSD_DEBUG(4, "nfsrv_getdevinfo stat=%u\n", nd->nd_repstat);
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(layouttype);
nfsm_strtom(nd, devaddr, devaddrlen);
cnt = 0;
for (i = 0; i < NFSV4_NOTIFYBITMAP; i++) {
if (notify[i] != 0)
cnt = i + 1;
}
NFSM_BUILD(tl, uint32_t *, (cnt + 1) * NFSX_UNSIGNED);
*tl++ = txdr_unsigned(cnt);
for (i = 0; i < cnt; i++)
*tl++ = txdr_unsigned(notify[i]);
} else if (nd->nd_repstat == NFSERR_TOOSMALL) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(maxcnt);
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsmout:
NFSEXITCODE2(error, nd);
return (error);
}
Add support for the TestStateID operation to the NFSv4.1 server. The Linux client now uses the TestStateID operation, so this patch adds support for it to the NFSv4.1 server. The FreeBSD client never uses this operation, so it should not be affected. MFC after: 2 months
2018-05-11 22:16:23 +00:00
/*
* nfsv4 test stateid service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for the TestStateID operation to the NFSv4.1 server. The Linux client now uses the TestStateID operation, so this patch adds support for it to the NFSv4.1 server. The FreeBSD client never uses this operation, so it should not be affected. MFC after: 2 months
2018-05-11 22:16:23 +00:00
nfsrvd_teststateid(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Add support for the TestStateID operation to the NFSv4.1 server. The Linux client now uses the TestStateID operation, so this patch adds support for it to the NFSv4.1 server. The FreeBSD client never uses this operation, so it should not be affected. MFC after: 2 months
2018-05-11 22:16:23 +00:00
{
uint32_t *tl;
nfsv4stateid_t *stateidp = NULL, *tstateidp;
int cnt, error = 0, i, ret;
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
struct thread *p = curthread;
Add support for the TestStateID operation to the NFSv4.1 server. The Linux client now uses the TestStateID operation, so this patch adds support for it to the NFSv4.1 server. The FreeBSD client never uses this operation, so it should not be affected. MFC after: 2 months
2018-05-11 22:16:23 +00:00
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
cnt = fxdr_unsigned(int, *tl);
if (cnt <= 0 || cnt > 1024) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
stateidp = mallocarray(cnt, sizeof(nfsv4stateid_t), M_TEMP, M_WAITOK);
tstateidp = stateidp;
for (i = 0; i < cnt; i++) {
NFSM_DISSECT(tl, uint32_t *, NFSX_STATEID);
tstateidp->seqid = fxdr_unsigned(uint32_t, *tl++);
NFSBCOPY(tl, tstateidp->other, NFSX_STATEIDOTHER);
tstateidp++;
}
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(cnt);
tstateidp = stateidp;
for (i = 0; i < cnt; i++) {
ret = nfsrv_teststateid(nd, tstateidp, p);
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(ret);
tstateidp++;
}
nfsmout:
free(stateidp, M_TEMP);
NFSEXITCODE2(error, nd);
return (error);
}
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
/*
* nfs allocate service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_allocate(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, struct nfsexstuff *exp)
{
uint32_t *tl;
struct nfsvattr forat;
int error = 0, forat_ret = 1, gotproxystateid;
off_t off, len;
struct nfsstate st, *stp = &st;
struct nfslock lo, *lop = &lo;
nfsv4stateid_t stateid;
nfsquad_t clientid;
nfsattrbit_t attrbits;
nfsd: Disable the NFSv4.2 Allocate operation by default Some exported file systems, such as ZFS ones, cannot do VOP_ALLOCATE(). Since an NFSv4.2 server must either support the Allocate operation for all file systems or not support it at all, define a sysctl called vfs.nfsd.enable_v42allocate to enable the Allocate operation. This sysctl is false by default and can only be set true if all exported file systems (or all DSs for a pNFS server) can perform VOP_ALLOCATE(). Unfortunately, there is no way to know if a ZFS file system will be exported once the nfsd is operational, even if there are none exported when the nfsd is started up, so enabling Allocate must be done manually for a server configuration. This problem was detected during a recent NFSv4 interoperability testing event held by the IETF working group. MFC after: 2 weeks
2021-10-11 01:46:02 +00:00
if (!nfsrv_doallocate) {
/*
* If any exported file system, such as a ZFS one, cannot
* do VOP_ALLOCATE(), this operation cannot be supported
* for NFSv4.2. This cannot be done 'per filesystem', but
* must be for the entire nfsd NFSv4.2 service.
*/
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
gotproxystateid = 0;
NFSM_DISSECT(tl, uint32_t *, NFSX_STATEID + 2 * NFSX_HYPER);
stp->ls_flags = (NFSLCK_CHECK | NFSLCK_WRITEACCESS);
lop->lo_flags = NFSLCK_WRITE;
stp->ls_ownerlen = 0;
stp->ls_op = NULL;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
clientid.lval[0] = stp->ls_stateid.other[0] = *tl++;
clientid.lval[1] = stp->ls_stateid.other[1] = *tl++;
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK2 multiple clids\n");
} else {
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
stp->ls_stateid.other[2] = *tl++;
/*
* Don't allow this to be done for a DS.
*/
if ((nd->nd_flag & ND_DSSERVER) != 0)
nd->nd_repstat = NFSERR_NOTSUPP;
/* However, allow the proxy stateid. */
if (stp->ls_stateid.seqid == 0xffffffff &&
stp->ls_stateid.other[0] == 0x55555555 &&
stp->ls_stateid.other[1] == 0x55555555 &&
stp->ls_stateid.other[2] == 0x55555555)
gotproxystateid = 1;
off = fxdr_hyper(tl); tl += 2;
lop->lo_first = off;
len = fxdr_hyper(tl);
nfsd: Fix sanity check for NFSv4.2 Allocate operations The NFSv4.2 Allocate operation sanity checks the aa_offset and aa_length arguments. Since they are assigned to variables of type off_t (signed) it was possible for them to be negative. It was also possible for aa_offset+aa_length to exceed OFF_MAX when stored in lo_end, which is uint64_t. This patch adds checks for these cases to the sanity check. Reviewed by: kib MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D31511
2021-08-12 23:48:28 +00:00
lop->lo_end = lop->lo_first + len;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
/*
nfsd: Fix sanity check for NFSv4.2 Allocate operations The NFSv4.2 Allocate operation sanity checks the aa_offset and aa_length arguments. Since they are assigned to variables of type off_t (signed) it was possible for them to be negative. It was also possible for aa_offset+aa_length to exceed OFF_MAX when stored in lo_end, which is uint64_t. This patch adds checks for these cases to the sanity check. Reviewed by: kib MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D31511
2021-08-12 23:48:28 +00:00
* Sanity check the offset and length.
* off and len are off_t (signed int64_t) whereas
* lo_first and lo_end are uint64_t and, as such,
* if off >= 0 && len > 0, lo_end cannot overflow
* unless off_t is changed to something other than
* int64_t. Check lo_end < lo_first in case that
* is someday the case.
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
*/
nfsd: Fix sanity check for NFSv4.2 Allocate operations The NFSv4.2 Allocate operation sanity checks the aa_offset and aa_length arguments. Since they are assigned to variables of type off_t (signed) it was possible for them to be negative. It was also possible for aa_offset+aa_length to exceed OFF_MAX when stored in lo_end, which is uint64_t. This patch adds checks for these cases to the sanity check. Reviewed by: kib MFC after: 2 weeks Differential Revision: https://reviews.freebsd.org/D31511
2021-08-12 23:48:28 +00:00
if (nd->nd_repstat == 0 && (len <= 0 || off < 0 || lop->lo_end >
OFF_MAX || lop->lo_end < lop->lo_first))
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_INVAL;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (nd->nd_repstat == 0 && vp->v_type != VREG)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_WRONGTYPE;
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_OWNER);
forat_ret = nfsvno_getattr(vp, &forat, nd, curthread, 1, &attrbits);
if (nd->nd_repstat == 0)
nd->nd_repstat = forat_ret;
if (nd->nd_repstat == 0 && (forat.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp)))
nd->nd_repstat = nfsvno_accchk(vp, VWRITE, nd->nd_cred, exp,
curthread, NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED,
NULL);
if (nd->nd_repstat == 0 && gotproxystateid == 0)
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, NULL, clientid,
&stateid, exp, nd, curthread);
nfsd: Disable the NFSv4.2 Allocate operation by default Some exported file systems, such as ZFS ones, cannot do VOP_ALLOCATE(). Since an NFSv4.2 server must either support the Allocate operation for all file systems or not support it at all, define a sysctl called vfs.nfsd.enable_v42allocate to enable the Allocate operation. This sysctl is false by default and can only be set true if all exported file systems (or all DSs for a pNFS server) can perform VOP_ALLOCATE(). Unfortunately, there is no way to know if a ZFS file system will be exported once the nfsd is operational, even if there are none exported when the nfsd is started up, so enabling Allocate must be done manually for a server configuration. This problem was detected during a recent NFSv4 interoperability testing event held by the IETF working group. MFC after: 2 weeks
2021-10-11 01:46:02 +00:00
NFSD_DEBUG(4, "nfsrvd_allocate: off=%jd len=%jd stat=%d\n",
(intmax_t)off, (intmax_t)len, nd->nd_repstat);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsvno_allocate(vp, off, len, nd->nd_cred,
curthread);
nfsd: Disable the NFSv4.2 Allocate operation by default Some exported file systems, such as ZFS ones, cannot do VOP_ALLOCATE(). Since an NFSv4.2 server must either support the Allocate operation for all file systems or not support it at all, define a sysctl called vfs.nfsd.enable_v42allocate to enable the Allocate operation. This sysctl is false by default and can only be set true if all exported file systems (or all DSs for a pNFS server) can perform VOP_ALLOCATE(). Unfortunately, there is no way to know if a ZFS file system will be exported once the nfsd is operational, even if there are none exported when the nfsd is started up, so enabling Allocate must be done manually for a server configuration. This problem was detected during a recent NFSv4 interoperability testing event held by the IETF working group. MFC after: 2 weeks
2021-10-11 01:46:02 +00:00
NFSD_DEBUG(4, "nfsrvd_allocate: aft nfsvno_allocate=%d\n",
nd->nd_repstat);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
vput(vp);
NFSEXITCODE2(0, nd);
return (0);
nfsd: Add support for the NFSv4.2 Deallocate operation The recently added VOP_DEALLOCATE(9) VOP call allows implementation of the Deallocate NFSv4.2 operation. Since the Deallocate operation is a single succeed/fail operation, the call to VOP_DEALLOCATE(9) loops so long as progress is being made. It calls maybe_yield() between loop iterations to allow other processes to preempt it. Where RFC 7862 underspecifies behaviour, the code is written to be Linux NFSv4.2 server compatible. Reviewed by: khng Differential Revision: https://reviews.freebsd.org/D31624
2021-08-27 01:14:11 +00:00
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfs deallocate service
*/
int
nfsrvd_deallocate(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, struct nfsexstuff *exp)
{
uint32_t *tl;
struct nfsvattr forat;
int error = 0, forat_ret = 1, gotproxystateid;
off_t off, len;
struct nfsstate st, *stp = &st;
struct nfslock lo, *lop = &lo;
nfsv4stateid_t stateid;
nfsquad_t clientid;
nfsattrbit_t attrbits;
gotproxystateid = 0;
NFSM_DISSECT(tl, uint32_t *, NFSX_STATEID + 2 * NFSX_HYPER);
stp->ls_flags = (NFSLCK_CHECK | NFSLCK_WRITEACCESS);
lop->lo_flags = NFSLCK_WRITE;
stp->ls_ownerlen = 0;
stp->ls_op = NULL;
stp->ls_uid = nd->nd_cred->cr_uid;
stp->ls_stateid.seqid = fxdr_unsigned(u_int32_t, *tl++);
clientid.lval[0] = stp->ls_stateid.other[0] = *tl++;
clientid.lval[1] = stp->ls_stateid.other[1] = *tl++;
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0) {
if ((nd->nd_flag & ND_NFSV41) != 0)
clientid.qval = nd->nd_clientid.qval;
else if (nd->nd_clientid.qval != clientid.qval)
printf("EEK2 multiple clids\n");
} else {
if ((nd->nd_flag & ND_NFSV41) != 0)
printf("EEK! no clientid from session\n");
nd->nd_flag |= ND_IMPLIEDCLID;
nd->nd_clientid.qval = clientid.qval;
}
stp->ls_stateid.other[2] = *tl++;
/*
* Don't allow this to be done for a DS.
*/
if ((nd->nd_flag & ND_DSSERVER) != 0)
nd->nd_repstat = NFSERR_NOTSUPP;
/* However, allow the proxy stateid. */
if (stp->ls_stateid.seqid == 0xffffffff &&
stp->ls_stateid.other[0] == 0x55555555 &&
stp->ls_stateid.other[1] == 0x55555555 &&
stp->ls_stateid.other[2] == 0x55555555)
gotproxystateid = 1;
off = fxdr_hyper(tl); tl += 2;
lop->lo_first = off;
len = fxdr_hyper(tl);
if (len < 0)
len = OFF_MAX;
NFSD_DEBUG(4, "dealloc: off=%jd len=%jd\n", (intmax_t)off,
(intmax_t)len);
lop->lo_end = lop->lo_first + len;
/*
* Sanity check the offset and length.
* off and len are off_t (signed int64_t) whereas
* lo_first and lo_end are uint64_t and, as such,
* if off >= 0 && len > 0, lo_end cannot overflow
* unless off_t is changed to something other than
* int64_t. Check lo_end < lo_first in case that
* is someday the case.
* The error to return is not specified by RFC 7862 so I
* made this compatible with the Linux knfsd.
*/
if (nd->nd_repstat == 0) {
if (off < 0 || lop->lo_end > NFSRV_MAXFILESIZE)
nd->nd_repstat = NFSERR_FBIG;
else if (len == 0 || lop->lo_end < lop->lo_first)
nd->nd_repstat = NFSERR_INVAL;
}
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (nd->nd_repstat == 0 && vp->v_type != VREG)
nfsd: Add support for the NFSv4.2 Deallocate operation The recently added VOP_DEALLOCATE(9) VOP call allows implementation of the Deallocate NFSv4.2 operation. Since the Deallocate operation is a single succeed/fail operation, the call to VOP_DEALLOCATE(9) loops so long as progress is being made. It calls maybe_yield() between loop iterations to allow other processes to preempt it. Where RFC 7862 underspecifies behaviour, the code is written to be Linux NFSv4.2 server compatible. Reviewed by: khng Differential Revision: https://reviews.freebsd.org/D31624
2021-08-27 01:14:11 +00:00
nd->nd_repstat = NFSERR_WRONGTYPE;
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_OWNER);
forat_ret = nfsvno_getattr(vp, &forat, nd, curthread, 1, &attrbits);
if (nd->nd_repstat == 0)
nd->nd_repstat = forat_ret;
if (nd->nd_repstat == 0 && (forat.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp)))
nd->nd_repstat = nfsvno_accchk(vp, VWRITE, nd->nd_cred, exp,
curthread, NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED,
NULL);
if (nd->nd_repstat == 0 && gotproxystateid == 0)
nd->nd_repstat = nfsrv_lockctrl(vp, &stp, &lop, NULL, clientid,
&stateid, exp, nd, curthread);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsvno_deallocate(vp, off, len, nd->nd_cred,
curthread);
vput(vp);
NFSD_DEBUG(4, "eo deallocate=%d\n", nd->nd_repstat);
NFSEXITCODE2(0, nd);
return (0);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfs copy service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_copy_file_range(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, vnode_t tovp, struct nfsexstuff *exp, struct nfsexstuff *toexp)
{
uint32_t *tl;
struct nfsvattr at;
int cnt, error = 0, ret;
off_t inoff, outoff;
uint64_t len;
nfsd: Fix build after commit 103b207536f9 for 32bit arches MFC after: 2 weeks
2021-09-09 01:55:06 +00:00
size_t xfer;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
struct nfsstate inst, outst, *instp = &inst, *outstp = &outst;
struct nfslock inlo, outlo, *inlop = &inlo, *outlop = &outlo;
nfsquad_t clientid;
nfsv4stateid_t stateid;
nfsattrbit_t attrbits;
void *rl_rcookie, *rl_wcookie;
rl_rcookie = rl_wcookie = NULL;
if (nfsrv_devidcnt > 0) {
/*
* For a pNFS server, reply NFSERR_NOTSUPP so that the client
* will do the copy via I/O on the DS(s).
*/
nd->nd_repstat = NFSERR_NOTSUPP;
goto nfsmout;
}
if (vp == tovp) {
/* Copying a byte range within the same file is not allowed. */
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_STATEID + 3 * NFSX_HYPER +
3 * NFSX_UNSIGNED);
instp->ls_flags = (NFSLCK_CHECK | NFSLCK_READACCESS);
inlop->lo_flags = NFSLCK_READ;
instp->ls_ownerlen = 0;
instp->ls_op = NULL;
instp->ls_uid = nd->nd_cred->cr_uid;
instp->ls_stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
clientid.lval[0] = instp->ls_stateid.other[0] = *tl++;
clientid.lval[1] = instp->ls_stateid.other[1] = *tl++;
if ((nd->nd_flag & ND_IMPLIEDCLID) != 0)
clientid.qval = nd->nd_clientid.qval;
instp->ls_stateid.other[2] = *tl++;
outstp->ls_flags = (NFSLCK_CHECK | NFSLCK_WRITEACCESS);
outlop->lo_flags = NFSLCK_WRITE;
outstp->ls_ownerlen = 0;
outstp->ls_op = NULL;
outstp->ls_uid = nd->nd_cred->cr_uid;
outstp->ls_stateid.seqid = fxdr_unsigned(uint32_t, *tl++);
outstp->ls_stateid.other[0] = *tl++;
outstp->ls_stateid.other[1] = *tl++;
outstp->ls_stateid.other[2] = *tl++;
inoff = fxdr_hyper(tl); tl += 2;
inlop->lo_first = inoff;
outoff = fxdr_hyper(tl); tl += 2;
outlop->lo_first = outoff;
len = fxdr_hyper(tl); tl += 2;
if (len == 0) {
/* len == 0 means to EOF. */
inlop->lo_end = OFF_MAX;
outlop->lo_end = OFF_MAX;
} else {
inlop->lo_end = inlop->lo_first + len;
outlop->lo_end = outlop->lo_first + len;
}
/*
* At this time only consecutive, synchronous copy is supported,
* so ca_consecutive and ca_synchronous can be ignored.
*/
tl += 2;
cnt = fxdr_unsigned(int, *tl);
if ((nd->nd_flag & ND_DSSERVER) != 0 || cnt != 0)
nd->nd_repstat = NFSERR_NOTSUPP;
if (nd->nd_repstat == 0 && (inoff > OFF_MAX || outoff > OFF_MAX ||
inlop->lo_end > OFF_MAX || outlop->lo_end > OFF_MAX ||
inlop->lo_end < inlop->lo_first || outlop->lo_end <
outlop->lo_first))
nd->nd_repstat = NFSERR_INVAL;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (nd->nd_repstat == 0 && vp->v_type != VREG)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_WRONGTYPE;
/* Check permissions for the input file. */
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_OWNER);
ret = nfsvno_getattr(vp, &at, nd, curthread, 1, &attrbits);
if (nd->nd_repstat == 0)
nd->nd_repstat = ret;
if (nd->nd_repstat == 0 && (at.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp)))
nd->nd_repstat = nfsvno_accchk(vp, VREAD, nd->nd_cred, exp,
curthread, NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED,
NULL);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsrv_lockctrl(vp, &instp, &inlop, NULL,
clientid, &stateid, exp, nd, curthread);
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (nd->nd_repstat != 0)
goto out;
error = NFSVOPLOCK(tovp, LK_SHARED);
if (error != 0)
goto out;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (tovp->v_type != VREG)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_WRONGTYPE;
/* For the output file, we only need the Owner attribute. */
ret = nfsvno_getattr(tovp, &at, nd, curthread, 1, &attrbits);
if (nd->nd_repstat == 0)
nd->nd_repstat = ret;
if (nd->nd_repstat == 0 && (at.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp)))
nd->nd_repstat = nfsvno_accchk(tovp, VWRITE, nd->nd_cred, toexp,
curthread, NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED,
NULL);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsrv_lockctrl(tovp, &outstp, &outlop, NULL,
clientid, &stateid, toexp, nd, curthread);
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(tovp);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
/* Range lock the byte ranges for both invp and outvp. */
if (nd->nd_repstat == 0) {
for (;;) {
if (len == 0) {
rl_wcookie = vn_rangelock_wlock(tovp, outoff,
OFF_MAX);
rl_rcookie = vn_rangelock_tryrlock(vp, inoff,
OFF_MAX);
} else {
rl_wcookie = vn_rangelock_wlock(tovp, outoff,
outoff + len);
rl_rcookie = vn_rangelock_tryrlock(vp, inoff,
inoff + len);
}
if (rl_rcookie != NULL)
break;
vn_rangelock_unlock(tovp, rl_wcookie);
if (len == 0)
rl_rcookie = vn_rangelock_rlock(vp, inoff,
OFF_MAX);
else
rl_rcookie = vn_rangelock_rlock(vp, inoff,
inoff + len);
vn_rangelock_unlock(vp, rl_rcookie);
}
error = NFSVOPLOCK(vp, LK_SHARED);
if (error == 0) {
ret = nfsvno_getattr(vp, &at, nd, curthread, 1, NULL);
if (ret == 0) {
/*
* Since invp is range locked, na_size should
* not change.
*/
if (len == 0 && at.na_size > inoff) {
/*
* If len == 0, set it based on invp's
* size. If offset is past EOF, just
* leave len == 0.
*/
len = at.na_size - inoff;
} else if (nfsrv_linux42server == 0 &&
inoff + len > at.na_size) {
/*
* RFC-7862 says that NFSERR_INVAL must
* be returned when inoff + len exceeds
* the file size, however the NFSv4.2
* Linux client likes to do this, so
* only check if nfsrv_linux42server
* is not set.
*/
nd->nd_repstat = NFSERR_INVAL;
}
}
vfs: drop the mostly unused flags argument from VOP_UNLOCK Filesystems which want to use it in limited capacity can employ the VOP_UNLOCK_FLAGS macro. Reviewed by: kib (previous version) Differential Revision: https://reviews.freebsd.org/D21427
2020-01-03 22:29:58 +00:00
NFSVOPUNLOCK(vp);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (ret != 0 && nd->nd_repstat == 0)
nd->nd_repstat = ret;
} else if (nd->nd_repstat == 0)
nd->nd_repstat = error;
}
nfsd: Fix build after commit 103b207536f9 for 32bit arches MFC after: 2 weeks
2021-09-09 01:55:06 +00:00
xfer = len;
if (nd->nd_repstat == 0) {
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = vn_copy_file_range(vp, &inoff, tovp, &outoff,
nfsd: Fix build after commit 103b207536f9 for 32bit arches MFC after: 2 weeks
2021-09-09 01:55:06 +00:00
&xfer, COPY_FILE_RANGE_TIMEO1SEC, nd->nd_cred, nd->nd_cred,
nfsd: Use the COPY_FILE_RANGE_TIMEO1SEC flag Although it is not specified in the RFCs, the concept that the NFSv4 server should reply to an RPC request within a reasonable time is accepted practice within the NFSv4 community. Without this patch, the NFSv4.2 server attempts to reply to a Copy operation within 1 second by limiting the copy to vfs.nfs.maxcopyrange bytes (default 10Mbytes). This is crude at best, given the large variation in I/O subsystem performance. This patch uses the COPY_FILE_RANGE_TIMEO1SEC flag added by commit c5128c48df3c to limit the reply time for a Copy operation to approximately 1 second. MFC after: 2 weeks
2021-09-08 21:29:20 +00:00
NULL);
nfsd: Fix build after commit 103b207536f9 for 32bit arches MFC after: 2 weeks
2021-09-09 01:55:06 +00:00
if (nd->nd_repstat == 0)
len = xfer;
}
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
/* Unlock the ranges. */
if (rl_rcookie != NULL)
vn_rangelock_unlock(vp, rl_rcookie);
if (rl_wcookie != NULL)
vn_rangelock_unlock(tovp, rl_wcookie);
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, 4 * NFSX_UNSIGNED + NFSX_HYPER +
NFSX_VERF);
*tl++ = txdr_unsigned(0); /* No callback ids. */
txdr_hyper(len, tl); tl += 2;
*tl++ = txdr_unsigned(NFSWRITE_UNSTABLE);
*tl++ = txdr_unsigned(nfsboottime.tv_sec);
*tl++ = txdr_unsigned(nfsboottime.tv_usec);
*tl++ = newnfs_true;
*tl = newnfs_true;
}
out:
vrele(vp);
vrele(tovp);
NFSEXITCODE2(error, nd);
return (error);
nfsmout:
vput(vp);
vrele(tovp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfs seek service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_seek(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, struct nfsexstuff *exp)
{
uint32_t *tl;
struct nfsvattr at;
int content, error = 0;
off_t off;
u_long cmd;
nfsattrbit_t attrbits;
bool eof;
NFSM_DISSECT(tl, uint32_t *, NFSX_STATEID + NFSX_HYPER + NFSX_UNSIGNED);
/* Ignore the stateid for now. */
tl += (NFSX_STATEID / NFSX_UNSIGNED);
off = fxdr_hyper(tl); tl += 2;
content = fxdr_unsigned(int, *tl);
if (content == NFSV4CONTENT_DATA)
cmd = FIOSEEKDATA;
else if (content == NFSV4CONTENT_HOLE)
cmd = FIOSEEKHOLE;
else
nd->nd_repstat = NFSERR_BADXDR;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (nd->nd_repstat == 0 && vp->v_type == VDIR)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_ISDIR;
nfsd: Clean up the code by not using the vnode_vtype() macro The vnode_vtype() macro was used to make the code compatible with Mac OSX, for the Mac OSX port. For FreeBSD, this macro just obscured the code, so avoid using it to clean up the code. This commit should not result in a semantics change.
2022-06-22 20:20:32 +00:00
if (nd->nd_repstat == 0 && vp->v_type != VREG)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_WRONGTYPE;
if (nd->nd_repstat == 0 && off < 0)
nd->nd_repstat = NFSERR_NXIO;
if (nd->nd_repstat == 0) {
/* Check permissions for the input file. */
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_OWNER);
nd->nd_repstat = nfsvno_getattr(vp, &at, nd, curthread, 1,
&attrbits);
}
if (nd->nd_repstat == 0 && (at.na_uid != nd->nd_cred->cr_uid ||
NFSVNO_EXSTRICTACCESS(exp)))
nd->nd_repstat = nfsvno_accchk(vp, VREAD, nd->nd_cred, exp,
curthread, NFSACCCHK_ALLOWOWNER, NFSACCCHK_VPISLOCKED,
NULL);
if (nd->nd_repstat != 0)
goto nfsmout;
/* nfsvno_seek() unlocks and vrele()s the vp. */
nd->nd_repstat = nfsvno_seek(nd, vp, cmd, &off, content, &eof,
nd->nd_cred, curthread);
if (nd->nd_repstat == 0 && eof && content == NFSV4CONTENT_DATA &&
nfsrv_linux42server != 0)
nd->nd_repstat = NFSERR_NXIO;
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED + NFSX_HYPER);
if (eof)
*tl++ = newnfs_true;
else
*tl++ = newnfs_false;
txdr_hyper(off, tl);
}
NFSEXITCODE2(error, nd);
return (error);
nfsmout:
vput(vp);
NFSEXITCODE2(error, nd);
return (error);
}
/*
* nfs get extended attribute service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_getxattr(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, __unused struct nfsexstuff *exp)
{
uint32_t *tl;
Replace all instances of the typedef mbuf_t with "struct mbuf *". The typedef mbuf_t was used for the Mac OS/X port of the code long ago. Since this port is no longer used and the use of mbuf_t obscures what the code does (and is not consistent with style(9)), it is no longer needed. This patch replaces all instances of mbuf_t with "struct mbuf *", so that it is no longer used. This patch should not result in any semantic change.
2020-04-17 21:17:51 +00:00
struct mbuf *mp = NULL, *mpend = NULL;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
int error, len;
char *name;
struct thread *p = curthread;
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
uint16_t off;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
error = 0;
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
len = fxdr_unsigned(int, *tl);
if (len <= 0) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
if (len > EXTATTR_MAXNAMELEN) {
nd->nd_repstat = NFSERR_NOXATTR;
goto nfsmout;
}
name = malloc(len + 1, M_TEMP, M_WAITOK);
nd->nd_repstat = nfsrv_mtostr(nd, name, len);
if (nd->nd_repstat == 0)
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
nd->nd_repstat = nfsvno_getxattr(vp, name,
nd->nd_maxresp, nd->nd_cred, nd->nd_flag,
nd->nd_maxextsiz, p, &mp, &mpend, &len);
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (nd->nd_repstat == ENOATTR)
nd->nd_repstat = NFSERR_NOXATTR;
else if (nd->nd_repstat == EOPNOTSUPP)
nd->nd_repstat = NFSERR_NOTSUPP;
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
*tl = txdr_unsigned(len);
Fix the NFSv2 extended attribute support to handle 0 length attributes. I did not realize that zero length attributes are allowed, but they are. This patch fixes the NFSv4.2 client and server to handle zero length extended attributes correctly. Submitted by: Frank van der Linden <fllinden@amazon.com> (earlier version) Reported by: Frank van der Linden <fllinder@amazon.com>
2020-04-14 22:57:21 +00:00
if (len > 0) {
nd->nd_mb->m_next = mp;
nd->nd_mb = mpend;
Add optional support for ext_pgs mbufs to the NFS server's read, readlink and getxattr operations. This patch optionally enables generation of read, readlink and getxattr replies in ext_pgs mbufs. Since neither of ND_EXTPG or ND_TLS are currently ever set, there is no change in semantics at this time. It also corrects the message in a couple of panic()s that should never occur. This is another in the series of commits that add support to the NFS client and server for building RPC messages in ext_pgs mbufs with anonymous pages. This is useful so that the entire mbuf list does not need to be copied before calling sosend() when NFS over TLS is enabled. Use of ext_pgs mbufs will not be enabled until the kernel RPC is updated to handle TLS.
2020-07-31 23:35:49 +00:00
if ((mpend->m_flags & M_EXTPG) != 0) {
nd->nd_flag |= ND_EXTPG;
nd->nd_bextpg = mpend->m_epg_npgs - 1;
nd->nd_bpos = (char *)(void *)
PHYS_TO_DMAP(mpend->m_epg_pa[nd->nd_bextpg]);
off = (nd->nd_bextpg == 0) ?
mpend->m_epg_1st_off : 0;
nd->nd_bpos += off + mpend->m_epg_last_len;
nd->nd_bextpgsiz = PAGE_SIZE -
mpend->m_epg_last_len - off;
} else
nd->nd_bpos = mtod(mpend, char *) +
mpend->m_len;
Fix the NFSv2 extended attribute support to handle 0 length attributes. I did not realize that zero length attributes are allowed, but they are. This patch fixes the NFSv4.2 client and server to handle zero length extended attributes correctly. Submitted by: Frank van der Linden <fllinden@amazon.com> (earlier version) Reported by: Frank van der Linden <fllinder@amazon.com>
2020-04-14 22:57:21 +00:00
}
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
}
free(name, M_TEMP);
nfsmout:
if (nd->nd_repstat == 0)
nd->nd_repstat = error;
vput(vp);
NFSEXITCODE2(0, nd);
return (0);
}
/*
* nfs set extended attribute service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_setxattr(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, __unused struct nfsexstuff *exp)
{
uint32_t *tl;
struct nfsvattr ova, nva;
nfsattrbit_t attrbits;
int error, len, opt;
char *name;
size_t siz;
struct thread *p = curthread;
error = 0;
name = NULL;
NFSM_DISSECT(tl, uint32_t *, 2 * NFSX_UNSIGNED);
opt = fxdr_unsigned(int, *tl++);
len = fxdr_unsigned(int, *tl);
if (len <= 0) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
if (len > EXTATTR_MAXNAMELEN) {
nd->nd_repstat = NFSERR_NOXATTR;
goto nfsmout;
}
name = malloc(len + 1, M_TEMP, M_WAITOK);
error = nfsrv_mtostr(nd, name, len);
if (error != 0)
goto nfsmout;
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
len = fxdr_unsigned(int, *tl);
Fix the NFSv2 extended attribute support to handle 0 length attributes. I did not realize that zero length attributes are allowed, but they are. This patch fixes the NFSv4.2 client and server to handle zero length extended attributes correctly. Submitted by: Frank van der Linden <fllinden@amazon.com> (earlier version) Reported by: Frank van der Linden <fllinder@amazon.com>
2020-04-14 22:57:21 +00:00
if (len < 0 || len > IOSIZE_MAX) {
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = NFSERR_XATTR2BIG;
goto nfsmout;
}
switch (opt) {
case NFSV4SXATTR_CREATE:
error = VOP_GETEXTATTR(vp, EXTATTR_NAMESPACE_USER, name, NULL,
&siz, nd->nd_cred, p);
if (error != ENOATTR)
nd->nd_repstat = NFSERR_EXIST;
error = 0;
break;
case NFSV4SXATTR_REPLACE:
error = VOP_GETEXTATTR(vp, EXTATTR_NAMESPACE_USER, name, NULL,
&siz, nd->nd_cred, p);
if (error != 0)
nd->nd_repstat = NFSERR_NOXATTR;
break;
case NFSV4SXATTR_EITHER:
break;
default:
nd->nd_repstat = NFSERR_BADXDR;
}
if (nd->nd_repstat != 0)
goto nfsmout;
/* Now, do the Set Extended attribute, with Change before and after. */
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_CHANGE);
nd->nd_repstat = nfsvno_getattr(vp, &ova, nd, p, 1, &attrbits);
if (nd->nd_repstat == 0) {
nd->nd_repstat = nfsvno_setxattr(vp, name, len, nd->nd_md,
nd->nd_dpos, nd->nd_cred, p);
if (nd->nd_repstat == ENXIO)
nd->nd_repstat = NFSERR_XATTR2BIG;
}
Fix the NFSv2 extended attribute support to handle 0 length attributes. I did not realize that zero length attributes are allowed, but they are. This patch fixes the NFSv4.2 client and server to handle zero length extended attributes correctly. Submitted by: Frank van der Linden <fllinden@amazon.com> (earlier version) Reported by: Frank van der Linden <fllinder@amazon.com>
2020-04-14 22:57:21 +00:00
if (nd->nd_repstat == 0 && len > 0)
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nd->nd_repstat = nfsm_advance(nd, NFSM_RNDUP(len), -1);
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1, &attrbits);
if (nd->nd_repstat == 0) {
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_HYPER + NFSX_UNSIGNED);
*tl++ = newnfs_true;
txdr_hyper(ova.na_filerev, tl); tl += 2;
txdr_hyper(nva.na_filerev, tl);
}
nfsmout:
free(name, M_TEMP);
if (nd->nd_repstat == 0)
nd->nd_repstat = error;
vput(vp);
NFSEXITCODE2(0, nd);
return (0);
}
/*
* nfs remove extended attribute service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_rmxattr(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, __unused struct nfsexstuff *exp)
{
uint32_t *tl;
struct nfsvattr ova, nva;
nfsattrbit_t attrbits;
int error, len;
char *name;
struct thread *p = curthread;
error = 0;
name = NULL;
NFSM_DISSECT(tl, uint32_t *, NFSX_UNSIGNED);
len = fxdr_unsigned(int, *tl);
if (len <= 0) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
if (len > EXTATTR_MAXNAMELEN) {
nd->nd_repstat = NFSERR_NOXATTR;
goto nfsmout;
}
name = malloc(len + 1, M_TEMP, M_WAITOK);
error = nfsrv_mtostr(nd, name, len);
if (error != 0)
goto nfsmout;
if ((nd->nd_flag & ND_IMPLIEDCLID) == 0) {
printf("EEK! nfsrvd_rmxattr: no implied clientid\n");
error = NFSERR_NOXATTR;
goto nfsmout;
}
/*
* Now, do the Remove Extended attribute, with Change before and
* after.
*/
NFSZERO_ATTRBIT(&attrbits);
NFSSETBIT_ATTRBIT(&attrbits, NFSATTRBIT_CHANGE);
nd->nd_repstat = nfsvno_getattr(vp, &ova, nd, p, 1, &attrbits);
if (nd->nd_repstat == 0) {
nd->nd_repstat = nfsvno_rmxattr(nd, vp, name, nd->nd_cred, p);
if (nd->nd_repstat == ENOATTR)
nd->nd_repstat = NFSERR_NOXATTR;
}
if (nd->nd_repstat == 0)
nd->nd_repstat = nfsvno_getattr(vp, &nva, nd, p, 1, &attrbits);
if (nd->nd_repstat == 0) {
Fix the NFSv4.2 extended attribute support for remove extended attrbute. I missed the "atomic" field of the RemoveExtendedAttribute operation's reply when I implemented it. It worked between FreeBSD client and server, since it was missed for both, but it did not conform to RFC 8276. This patch adds the field for both client and server. Thanks go to Frank for doing interoperability testing of the extended attribute support against patches for Linux. Submitted by: Frank van der Linden <fllinden@amazon.com> Reported by: Frank van der Linden <fllinden@amazon.com>
2020-04-15 21:27:52 +00:00
NFSM_BUILD(tl, uint32_t *, 2 * NFSX_HYPER + NFSX_UNSIGNED);
*tl++ = newnfs_true;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
txdr_hyper(ova.na_filerev, tl); tl += 2;
txdr_hyper(nva.na_filerev, tl);
}
nfsmout:
free(name, M_TEMP);
if (nd->nd_repstat == 0)
nd->nd_repstat = error;
vput(vp);
NFSEXITCODE2(0, nd);
return (0);
}
/*
* nfs list extended attribute service
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
nfsrvd_listxattr(struct nfsrv_descript *nd, __unused int isdgram,
vnode_t vp, __unused struct nfsexstuff *exp)
{
uint32_t cnt, *tl, len, len2, i, pos, retlen;
int error;
uint64_t cookie, cookie2;
u_char *buf;
bool eof;
struct thread *p = curthread;
error = 0;
buf = NULL;
NFSM_DISSECT(tl, uint32_t *, NFSX_HYPER + NFSX_UNSIGNED);
/*
* The cookie doesn't need to be in net byte order, but FreeBSD
* does so to make it more readable in packet traces.
*/
cookie = fxdr_hyper(tl); tl += 2;
len = fxdr_unsigned(uint32_t, *tl);
if (len == 0 || cookie >= IOSIZE_MAX) {
nd->nd_repstat = NFSERR_BADXDR;
goto nfsmout;
}
if (len > nd->nd_maxresp - NFS_MAXXDR)
len = nd->nd_maxresp - NFS_MAXXDR;
len2 = len;
nd->nd_repstat = nfsvno_listxattr(vp, cookie, nd->nd_cred, p, &buf,
&len, &eof);
if (nd->nd_repstat == EOPNOTSUPP)
nd->nd_repstat = NFSERR_NOTSUPP;
if (nd->nd_repstat == 0) {
cookie2 = cookie + len;
if (cookie2 < cookie)
nd->nd_repstat = NFSERR_BADXDR;
}
nfsd: Sanity check the len argument for ListXattr The check for the original len being >= retlen needs to be done before the "if (nd->nd_repstat == 0)" code, so that it can be reported as too small. Reported by: rtm@lcs.mit.edu Tested by: rtm@lcs.mit.edu PR: 260046 MFC after: 2 weeks
2021-11-26 23:56:29 +00:00
retlen = NFSX_HYPER + 2 * NFSX_UNSIGNED;
if (nd->nd_repstat == 0 && len2 < retlen)
nd->nd_repstat = NFSERR_TOOSMALL;
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
if (nd->nd_repstat == 0) {
/* Now copy the entries out. */
nfsd: Sanity check the len argument for ListXattr The check for the original len being >= retlen needs to be done before the "if (nd->nd_repstat == 0)" code, so that it can be reported as too small. Reported by: rtm@lcs.mit.edu Tested by: rtm@lcs.mit.edu PR: 260046 MFC after: 2 weeks
2021-11-26 23:56:29 +00:00
if (len == 0) {
Add support for NFSv4.2 to the NFS client and server. This patch adds support for NFSv4.2 (RFC-7862) and Extended Attributes (RFC-8276) to the NFS client and server. NFSv4.2 is comprised of several optional features that can be supported in addition to NFSv4.1. This patch adds the following optional features: - posix_fadvise(POSIX_FADV_WILLNEED/POSIX_FADV_DONTNEED) - posix_fallocate() - intra server file range copying via the copy_file_range(2) syscall --> Avoiding data tranfer over the wire to/from the NFS client. - lseek(SEEK_DATA/SEEK_HOLE) - Extended attribute syscalls for "user" namespace attributes as defined by RFC-8276. Although this patch is fairly large, it should not affect support for the other versions of NFS. However it does add two new sysctls that allow a sysadmin to limit which minor versions of NFSv4 a server supports, allowing a sysadmin to disable NFSv4.2. Unfortunately, when the NFS stats structure was last revised, it was assumed that there would be no additional operations added beyond what was specified in RFC-7862. However RFC-8276 did add additional operations, forcing the NFS stats structure to revised again. It now has extra unused entries in all arrays, so that future extensions to NFSv4.2 can be accomodated without revising this structure again. A future commit will update nfsstat(1) to report counts for the new NFSv4.2 specific operations/procedures. This patch affects the internal interface between the nfscommon, nfscl and nfsd modules and, as such, they all must be upgraded simultaneously. I will do a version bump (although arguably not needed), due to this. This code has survived a "make universe" but has not been built with a recent GCC. If you encounter build problems, please email me. Relnotes: yes
2019-12-12 23:22:55 +00:00
/* The cookie was at eof. */
NFSM_BUILD(tl, uint32_t *, NFSX_HYPER + 2 *
NFSX_UNSIGNED);
txdr_hyper(cookie2, tl); tl += 2;
*tl++ = txdr_unsigned(0);
*tl = newnfs_true;
goto nfsmout;
}
/* Sanity check the cookie. */
for (pos = 0; pos < len; pos += (i + 1)) {
if (pos == cookie)
break;
i = buf[pos];
}
if (pos != cookie) {
nd->nd_repstat = NFSERR_INVAL;
goto nfsmout;
}
/* Loop around copying the entrie(s) out. */
cnt = 0;
len -= cookie;
i = buf[pos];
while (i < len && len2 >= retlen + NFSM_RNDUP(i) +
NFSX_UNSIGNED) {
if (cnt == 0) {
NFSM_BUILD(tl, uint32_t *, NFSX_HYPER +
NFSX_UNSIGNED);
txdr_hyper(cookie2, tl); tl += 2;
}
retlen += nfsm_strtom(nd, &buf[pos + 1], i);
len -= (i + 1);
pos += (i + 1);
i = buf[pos];
cnt++;
}
/*
* eof is set true/false by nfsvno_listxattr(), but if we
* can't copy all entries returned by nfsvno_listxattr(),
* we are not at eof.
*/
if (len > 0)
eof = false;
if (cnt > 0) {
/* *tl is set above. */
*tl = txdr_unsigned(cnt);
NFSM_BUILD(tl, uint32_t *, NFSX_UNSIGNED);
if (eof)
*tl = newnfs_true;
else
*tl = newnfs_false;
} else
nd->nd_repstat = NFSERR_TOOSMALL;
}
nfsmout:
free(buf, M_TEMP);
if (nd->nd_repstat == 0)
nd->nd_repstat = error;
vput(vp);
NFSEXITCODE2(0, nd);
return (0);
}
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
/*
* nfsv4 service not supported
*/
nfs: Remove APPLESTATIC macro It is no longer useful. Reviewed by: rmacklem Approved by: mav (mentor) MFC after: 1 week Sponsored by: iXsystems, Inc. Differential Revision: https://reviews.freebsd.org/D24811
2020-05-12 13:23:25 +00:00
int
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
nfsrvd_notsupp(struct nfsrv_descript *nd, __unused int isdgram,
Push down the thread argument in NFS server code, using curthread instead of passing it explicitly. No functional changes Reviewed by: rmacklem (earlier version) MFC after: 2 weeks Sponsored by: DARPA, AFRL Differential Revision: https://reviews.freebsd.org/D19419
2019-03-04 13:12:23 +00:00
__unused vnode_t vp, __unused struct nfsexstuff *exp)
Merge the NFSv4.1 server code in projects/nfsv4.1-server over into head. The code is not believed to have any effect on the semantics of non-NFSv4.1 server behaviour. It is a rather large merge, but I am hoping that there will not be any regressions for the NFS server. MFC after: 1 month
2014-07-01 20:47:16 +00:00
{
nd->nd_repstat = NFSERR_NOTSUPP;
NFSEXITCODE2(0, nd);
return (0);
}
Reference in New Issue Copy Permalink