freebsd-nq/sys/rpc/svc_generic.c

319 lines
8.4 KiB
C
Raw Normal View History

Add the new kernel-mode NFS Lock Manager. To use it instead of the user-mode lock manager, build a kernel with the NFSLOCKD option and add '-k' to 'rpc_lockd_flags' in rc.conf. Highlights include: * Thread-safe kernel RPC client - many threads can use the same RPC client handle safely with replies being de-multiplexed at the socket upcall (typically driven directly by the NIC interrupt) and handed off to whichever thread matches the reply. For UDP sockets, many RPC clients can share the same socket. This allows the use of a single privileged UDP port number to talk to an arbitrary number of remote hosts. * Single-threaded kernel RPC server. Adding support for multi-threaded server would be relatively straightforward and would follow approximately the Solaris KPI. A single thread should be sufficient for the NLM since it should rarely block in normal operation. * Kernel mode NLM server supporting cancel requests and granted callbacks. I've tested the NLM server reasonably extensively - it passes both my own tests and the NFS Connectathon locking tests running on Solaris, Mac OS X and Ubuntu Linux. * Userland NLM client supported. While the NLM server doesn't have support for the local NFS client's locking needs, it does have to field async replies and granted callbacks from remote NLMs that the local client has contacted. We relay these replies to the userland rpc.lockd over a local domain RPC socket. * Robust deadlock detection for the local lock manager. In particular it will detect deadlocks caused by a lock request that covers more than one blocking request. As required by the NLM protocol, all deadlock detection happens synchronously - a user is guaranteed that if a lock request isn't rejected immediately, the lock will eventually be granted. The old system allowed for a 'deferred deadlock' condition where a blocked lock request could wake up and find that some other deadlock-causing lock owner had beaten them to the lock. * Since both local and remote locks are managed by the same kernel locking code, local and remote processes can safely use file locks for mutual exclusion. Local processes have no fairness advantage compared to remote processes when contending to lock a region that has just been unlocked - the local lock manager enforces a strict first-come first-served model for both local and remote lockers. Sponsored by: Isilon Systems PR: 95247 107555 115524 116679 MFC after: 2 weeks
2008-03-26 15:23:12 +00:00
/* $NetBSD: svc_generic.c,v 1.3 2000/07/06 03:10:35 christos Exp $ */
/*
* Sun RPC is a product of Sun Microsystems, Inc. and is provided for
* unrestricted use provided that this legend is included on all tape
* media and as a part of the software program in whole or part. Users
* may copy or modify Sun RPC without charge, but are not authorized
* to license or distribute it to anyone else except as part of a product or
* program developed by the user.
*
* SUN RPC IS PROVIDED AS IS WITH NO WARRANTIES OF ANY KIND INCLUDING THE
* WARRANTIES OF DESIGN, MERCHANTIBILITY AND FITNESS FOR A PARTICULAR
* PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE.
*
* Sun RPC is provided with no support and without any obligation on the
* part of Sun Microsystems, Inc. to assist in its use, correction,
* modification or enhancement.
*
* SUN MICROSYSTEMS, INC. SHALL HAVE NO LIABILITY WITH RESPECT TO THE
* INFRINGEMENT OF COPYRIGHTS, TRADE SECRETS OR ANY PATENTS BY SUN RPC
* OR ANY PART THEREOF.
*
* In no event will Sun Microsystems, Inc. be liable for any lost revenue
* or profits or other special, indirect and consequential damages, even if
* Sun has been advised of the possibility of such damages.
*
* Sun Microsystems, Inc.
* 2550 Garcia Avenue
* Mountain View, California 94043
*/
/*
* Copyright (c) 1986-1991 by Sun Microsystems Inc.
*/
#if defined(LIBC_SCCS) && !defined(lint)
#ident "@(#)svc_generic.c 1.19 94/04/24 SMI"
static char sccsid[] = "@(#)svc_generic.c 1.21 89/02/28 Copyr 1988 Sun Micro";
#endif
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*
* svc_generic.c, Server side for RPC.
*
*/
#include "opt_inet6.h"
#include <sys/param.h>
#include <sys/lock.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
#include <sys/mutex.h>
#include <sys/protosw.h>
#include <sys/queue.h>
#include <sys/socket.h>
#include <sys/socketvar.h>
#include <sys/systm.h>
#include <sys/sx.h>
#include <sys/ucred.h>
#include <rpc/rpc.h>
#include <rpc/rpcb_clnt.h>
#include <rpc/nettype.h>
#include <rpc/rpc_com.h>
Add the new kernel-mode NFS Lock Manager. To use it instead of the user-mode lock manager, build a kernel with the NFSLOCKD option and add '-k' to 'rpc_lockd_flags' in rc.conf. Highlights include: * Thread-safe kernel RPC client - many threads can use the same RPC client handle safely with replies being de-multiplexed at the socket upcall (typically driven directly by the NIC interrupt) and handed off to whichever thread matches the reply. For UDP sockets, many RPC clients can share the same socket. This allows the use of a single privileged UDP port number to talk to an arbitrary number of remote hosts. * Single-threaded kernel RPC server. Adding support for multi-threaded server would be relatively straightforward and would follow approximately the Solaris KPI. A single thread should be sufficient for the NLM since it should rarely block in normal operation. * Kernel mode NLM server supporting cancel requests and granted callbacks. I've tested the NLM server reasonably extensively - it passes both my own tests and the NFS Connectathon locking tests running on Solaris, Mac OS X and Ubuntu Linux. * Userland NLM client supported. While the NLM server doesn't have support for the local NFS client's locking needs, it does have to field async replies and granted callbacks from remote NLMs that the local client has contacted. We relay these replies to the userland rpc.lockd over a local domain RPC socket. * Robust deadlock detection for the local lock manager. In particular it will detect deadlocks caused by a lock request that covers more than one blocking request. As required by the NLM protocol, all deadlock detection happens synchronously - a user is guaranteed that if a lock request isn't rejected immediately, the lock will eventually be granted. The old system allowed for a 'deferred deadlock' condition where a blocked lock request could wake up and find that some other deadlock-causing lock owner had beaten them to the lock. * Since both local and remote locks are managed by the same kernel locking code, local and remote processes can safely use file locks for mutual exclusion. Local processes have no fairness advantage compared to remote processes when contending to lock a region that has just been unlocked - the local lock manager enforces a strict first-come first-served model for both local and remote lockers. Sponsored by: Isilon Systems PR: 95247 107555 115524 116679 MFC after: 2 weeks
2008-03-26 15:23:12 +00:00
extern int __svc_vc_setflag(SVCXPRT *, int);
/*
* The highest level interface for server creation.
* It tries for all the nettokens in that particular class of token
* and returns the number of handles it can create and/or find.
*
* It creates a link list of all the handles it could create.
* If svc_create() is called multiple times, it uses the handle
* created earlier instead of creating a new handle every time.
*/
int
svc_create(
SVCPOOL *pool,
void (*dispatch)(struct svc_req *, SVCXPRT *),
rpcprog_t prognum, /* Program number */
rpcvers_t versnum, /* Version number */
const char *nettype) /* Networktype token */
{
int num = 0;
SVCXPRT *xprt;
struct netconfig *nconf;
void *handle;
if ((handle = __rpc_setconf(nettype)) == NULL) {
printf("svc_create: unknown protocol");
return (0);
}
while ((nconf = __rpc_getconf(handle)) != NULL) {
mtx_lock(&pool->sp_lock);
TAILQ_FOREACH(xprt, &pool->sp_xlist, xp_link) {
if (strcmp(xprt->xp_netid, nconf->nc_netid) == 0) {
/* Found an old one, use it */
mtx_unlock(&pool->sp_lock);
(void) rpcb_unset(prognum, versnum, nconf);
if (svc_reg(xprt, prognum, versnum,
dispatch, nconf) == FALSE) {
printf(
"svc_create: could not register prog %u vers %u on %s\n",
(unsigned)prognum, (unsigned)versnum,
nconf->nc_netid);
mtx_lock(&pool->sp_lock);
} else {
num++;
mtx_lock(&pool->sp_lock);
break;
}
}
}
mtx_unlock(&pool->sp_lock);
if (xprt == NULL) {
/* It was not found. Now create a new one */
xprt = svc_tp_create(pool, dispatch, prognum, versnum,
NULL, nconf);
if (xprt)
num++;
}
}
__rpc_endconf(handle);
/*
* In case of num == 0; the error messages are generated by the
* underlying layers; and hence not needed here.
*/
return (num);
}
/*
* The high level interface to svc_tli_create().
* It tries to create a server for "nconf" and registers the service
* with the rpcbind. It calls svc_tli_create();
*/
SVCXPRT *
svc_tp_create(
SVCPOOL *pool,
void (*dispatch)(struct svc_req *, SVCXPRT *),
rpcprog_t prognum, /* Program number */
rpcvers_t versnum, /* Version number */
const char *uaddr, /* Address (or null for default) */
const struct netconfig *nconf) /* Netconfig structure for the network */
{
struct netconfig nconfcopy;
struct netbuf *taddr;
struct t_bind bind;
SVCXPRT *xprt;
if (nconf == NULL) {
printf(
"svc_tp_create: invalid netconfig structure for prog %u vers %u\n",
(unsigned)prognum, (unsigned)versnum);
return (NULL);
}
if (uaddr) {
taddr = uaddr2taddr(nconf, uaddr);
bind.addr = *taddr;
free(taddr, M_RPC);
bind.qlen = SOMAXCONN;
xprt = svc_tli_create(pool, NULL, nconf, &bind, 0, 0);
free(bind.addr.buf, M_RPC);
} else {
xprt = svc_tli_create(pool, NULL, nconf, NULL, 0, 0);
}
if (xprt == NULL) {
return (NULL);
}
/*LINTED const castaway*/
nconfcopy = *nconf;
(void) rpcb_unset(prognum, versnum, &nconfcopy);
if (svc_reg(xprt, prognum, versnum, dispatch, nconf) == FALSE) {
printf(
"svc_tp_create: Could not register prog %u vers %u on %s\n",
(unsigned)prognum, (unsigned)versnum,
nconf->nc_netid);
Implement support for RPCSEC_GSS authentication to both the NFS client and server. This replaces the RPC implementation of the NFS client and server with the newer RPC implementation originally developed (actually ported from the userland sunrpc code) to support the NFS Lock Manager. I have tested this code extensively and I believe it is stable and that performance is at least equal to the legacy RPC implementation. The NFS code currently contains support for both the new RPC implementation and the older legacy implementation inherited from the original NFS codebase. The default is to use the new implementation - add the NFS_LEGACYRPC option to fall back to the old code. When I merge this support back to RELENG_7, I will probably change this so that users have to 'opt in' to get the new code. To use RPCSEC_GSS on either client or server, you must build a kernel which includes the KGSSAPI option and the crypto device. On the userland side, you must build at least a new libc, mountd, mount_nfs and gssd. You must install new versions of /etc/rc.d/gssd and /etc/rc.d/nfsd and add 'gssd_enable=YES' to /etc/rc.conf. As long as gssd is running, you should be able to mount an NFS filesystem from a server that requires RPCSEC_GSS authentication. The mount itself can happen without any kerberos credentials but all access to the filesystem will be denied unless the accessing user has a valid ticket file in the standard place (/tmp/krb5cc_<uid>). There is currently no support for situations where the ticket file is in a different place, such as when the user logged in via SSH and has delegated credentials from that login. This restriction is also present in Solaris and Linux. In theory, we could improve this in future, possibly using Brooks Davis' implementation of variant symlinks. Supporting RPCSEC_GSS on a server is nearly as simple. You must create service creds for the server in the form 'nfs/<fqdn>@<REALM>' and install them in /etc/krb5.keytab. The standard heimdal utility ktutil makes this fairly easy. After the service creds have been created, you can add a '-sec=krb5' option to /etc/exports and restart both mountd and nfsd. The only other difference an administrator should notice is that nfsd doesn't fork to create service threads any more. In normal operation, there will be two nfsd processes, one in userland waiting for TCP connections and one in the kernel handling requests. The latter process will create as many kthreads as required - these should be visible via 'top -H'. The code has some support for varying the number of service threads according to load but initially at least, nfsd uses a fixed number of threads according to the value supplied to its '-n' option. Sponsored by: Isilon Systems MFC after: 1 month
2008-11-03 10:38:00 +00:00
xprt_unregister(xprt);
Add the new kernel-mode NFS Lock Manager. To use it instead of the user-mode lock manager, build a kernel with the NFSLOCKD option and add '-k' to 'rpc_lockd_flags' in rc.conf. Highlights include: * Thread-safe kernel RPC client - many threads can use the same RPC client handle safely with replies being de-multiplexed at the socket upcall (typically driven directly by the NIC interrupt) and handed off to whichever thread matches the reply. For UDP sockets, many RPC clients can share the same socket. This allows the use of a single privileged UDP port number to talk to an arbitrary number of remote hosts. * Single-threaded kernel RPC server. Adding support for multi-threaded server would be relatively straightforward and would follow approximately the Solaris KPI. A single thread should be sufficient for the NLM since it should rarely block in normal operation. * Kernel mode NLM server supporting cancel requests and granted callbacks. I've tested the NLM server reasonably extensively - it passes both my own tests and the NFS Connectathon locking tests running on Solaris, Mac OS X and Ubuntu Linux. * Userland NLM client supported. While the NLM server doesn't have support for the local NFS client's locking needs, it does have to field async replies and granted callbacks from remote NLMs that the local client has contacted. We relay these replies to the userland rpc.lockd over a local domain RPC socket. * Robust deadlock detection for the local lock manager. In particular it will detect deadlocks caused by a lock request that covers more than one blocking request. As required by the NLM protocol, all deadlock detection happens synchronously - a user is guaranteed that if a lock request isn't rejected immediately, the lock will eventually be granted. The old system allowed for a 'deferred deadlock' condition where a blocked lock request could wake up and find that some other deadlock-causing lock owner had beaten them to the lock. * Since both local and remote locks are managed by the same kernel locking code, local and remote processes can safely use file locks for mutual exclusion. Local processes have no fairness advantage compared to remote processes when contending to lock a region that has just been unlocked - the local lock manager enforces a strict first-come first-served model for both local and remote lockers. Sponsored by: Isilon Systems PR: 95247 107555 115524 116679 MFC after: 2 weeks
2008-03-26 15:23:12 +00:00
return (NULL);
}
return (xprt);
}
/*
* If so is NULL, then it opens a socket for the given transport
* provider (nconf cannot be NULL then). If the t_state is T_UNBND and
* bindaddr is NON-NULL, it performs a t_bind using the bindaddr. For
* NULL bindadr and Connection oriented transports, the value of qlen
* is set to 8.
*
* If sendsz or recvsz are zero, their default values are chosen.
*/
SVCXPRT *
svc_tli_create(
SVCPOOL *pool,
struct socket *so, /* Connection end point */
const struct netconfig *nconf, /* Netconfig struct for nettoken */
const struct t_bind *bindaddr, /* Local bind address */
size_t sendsz, /* Max sendsize */
size_t recvsz) /* Max recvsize */
{
SVCXPRT *xprt = NULL; /* service handle */
bool_t madeso = FALSE; /* whether so opened here */
struct __rpc_sockinfo si;
struct sockaddr_storage ss;
if (!so) {
if (nconf == NULL) {
printf("svc_tli_create: invalid netconfig\n");
return (NULL);
}
so = __rpc_nconf2socket(nconf);
if (!so) {
printf(
"svc_tli_create: could not open connection for %s\n",
nconf->nc_netid);
return (NULL);
}
__rpc_nconf2sockinfo(nconf, &si);
madeso = TRUE;
} else {
/*
* It is an open socket. Get the transport info.
*/
if (!__rpc_socket2sockinfo(so, &si)) {
printf(
"svc_tli_create: could not get transport information\n");
return (NULL);
}
}
/*
* If the socket is unbound, try to bind it.
*/
if (madeso || !__rpc_sockisbound(so)) {
if (bindaddr == NULL) {
if (bindresvport(so, NULL)) {
memset(&ss, 0, sizeof ss);
ss.ss_family = si.si_af;
ss.ss_len = si.si_alen;
if (sobind(so, (struct sockaddr *)&ss,
curthread)) {
printf(
"svc_tli_create: could not bind to anonymous port\n");
goto freedata;
}
}
solisten(so, SOMAXCONN, curthread);
} else {
if (bindresvport(so,
(struct sockaddr *)bindaddr->addr.buf)) {
printf(
"svc_tli_create: could not bind to requested address\n");
goto freedata;
}
solisten(so, (int)bindaddr->qlen, curthread);
}
}
/*
* call transport specific function.
*/
switch (si.si_socktype) {
case SOCK_STREAM:
#if 0
slen = sizeof ss;
if (_getpeername(fd, (struct sockaddr *)(void *)&ss, &slen)
== 0) {
/* accepted socket */
xprt = svc_fd_create(fd, sendsz, recvsz);
} else
#endif
xprt = svc_vc_create(pool, so, sendsz, recvsz);
if (!nconf || !xprt)
break;
#if 0
/* XXX fvdl */
if (strcmp(nconf->nc_protofmly, "inet") == 0 ||
strcmp(nconf->nc_protofmly, "inet6") == 0)
(void) __svc_vc_setflag(xprt, TRUE);
#endif
break;
case SOCK_DGRAM:
xprt = svc_dg_create(pool, so, sendsz, recvsz);
break;
default:
printf("svc_tli_create: bad service type");
goto freedata;
}
if (xprt == NULL)
/*
* The error messages here are spitted out by the lower layers:
* svc_vc_create(), svc_fd_create() and svc_dg_create().
*/
goto freedata;
/* Fill in type of service */
xprt->xp_type = __rpc_socktype2seman(si.si_socktype);
if (nconf) {
xprt->xp_netid = strdup(nconf->nc_netid, M_RPC);
}
return (xprt);
freedata:
if (madeso)
(void)soclose(so);
if (xprt) {
if (!madeso) /* so that svc_destroy doesnt close fd */
xprt->xp_socket = NULL;
Implement support for RPCSEC_GSS authentication to both the NFS client and server. This replaces the RPC implementation of the NFS client and server with the newer RPC implementation originally developed (actually ported from the userland sunrpc code) to support the NFS Lock Manager. I have tested this code extensively and I believe it is stable and that performance is at least equal to the legacy RPC implementation. The NFS code currently contains support for both the new RPC implementation and the older legacy implementation inherited from the original NFS codebase. The default is to use the new implementation - add the NFS_LEGACYRPC option to fall back to the old code. When I merge this support back to RELENG_7, I will probably change this so that users have to 'opt in' to get the new code. To use RPCSEC_GSS on either client or server, you must build a kernel which includes the KGSSAPI option and the crypto device. On the userland side, you must build at least a new libc, mountd, mount_nfs and gssd. You must install new versions of /etc/rc.d/gssd and /etc/rc.d/nfsd and add 'gssd_enable=YES' to /etc/rc.conf. As long as gssd is running, you should be able to mount an NFS filesystem from a server that requires RPCSEC_GSS authentication. The mount itself can happen without any kerberos credentials but all access to the filesystem will be denied unless the accessing user has a valid ticket file in the standard place (/tmp/krb5cc_<uid>). There is currently no support for situations where the ticket file is in a different place, such as when the user logged in via SSH and has delegated credentials from that login. This restriction is also present in Solaris and Linux. In theory, we could improve this in future, possibly using Brooks Davis' implementation of variant symlinks. Supporting RPCSEC_GSS on a server is nearly as simple. You must create service creds for the server in the form 'nfs/<fqdn>@<REALM>' and install them in /etc/krb5.keytab. The standard heimdal utility ktutil makes this fairly easy. After the service creds have been created, you can add a '-sec=krb5' option to /etc/exports and restart both mountd and nfsd. The only other difference an administrator should notice is that nfsd doesn't fork to create service threads any more. In normal operation, there will be two nfsd processes, one in userland waiting for TCP connections and one in the kernel handling requests. The latter process will create as many kthreads as required - these should be visible via 'top -H'. The code has some support for varying the number of service threads according to load but initially at least, nfsd uses a fixed number of threads according to the value supplied to its '-n' option. Sponsored by: Isilon Systems MFC after: 1 month
2008-11-03 10:38:00 +00:00
xprt_unregister(xprt);
Add the new kernel-mode NFS Lock Manager. To use it instead of the user-mode lock manager, build a kernel with the NFSLOCKD option and add '-k' to 'rpc_lockd_flags' in rc.conf. Highlights include: * Thread-safe kernel RPC client - many threads can use the same RPC client handle safely with replies being de-multiplexed at the socket upcall (typically driven directly by the NIC interrupt) and handed off to whichever thread matches the reply. For UDP sockets, many RPC clients can share the same socket. This allows the use of a single privileged UDP port number to talk to an arbitrary number of remote hosts. * Single-threaded kernel RPC server. Adding support for multi-threaded server would be relatively straightforward and would follow approximately the Solaris KPI. A single thread should be sufficient for the NLM since it should rarely block in normal operation. * Kernel mode NLM server supporting cancel requests and granted callbacks. I've tested the NLM server reasonably extensively - it passes both my own tests and the NFS Connectathon locking tests running on Solaris, Mac OS X and Ubuntu Linux. * Userland NLM client supported. While the NLM server doesn't have support for the local NFS client's locking needs, it does have to field async replies and granted callbacks from remote NLMs that the local client has contacted. We relay these replies to the userland rpc.lockd over a local domain RPC socket. * Robust deadlock detection for the local lock manager. In particular it will detect deadlocks caused by a lock request that covers more than one blocking request. As required by the NLM protocol, all deadlock detection happens synchronously - a user is guaranteed that if a lock request isn't rejected immediately, the lock will eventually be granted. The old system allowed for a 'deferred deadlock' condition where a blocked lock request could wake up and find that some other deadlock-causing lock owner had beaten them to the lock. * Since both local and remote locks are managed by the same kernel locking code, local and remote processes can safely use file locks for mutual exclusion. Local processes have no fairness advantage compared to remote processes when contending to lock a region that has just been unlocked - the local lock manager enforces a strict first-come first-served model for both local and remote lockers. Sponsored by: Isilon Systems PR: 95247 107555 115524 116679 MFC after: 2 weeks
2008-03-26 15:23:12 +00:00
}
return (NULL);
}