d/freebsd-skq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
ecaa500c4b
freebsd-skq/sys/xdr/xdr_mem.c

276 lines
6.7 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: xdr_mem.c,v 1.15 2000/01/22 22:19:18 mycroft 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
*/
#if defined(LIBC_SCCS) && !defined(lint)
static char *sccsid2 = "@(#)xdr_mem.c 1.19 87/08/11 Copyr 1984 Sun Micro";
static char *sccsid = "@(#)xdr_mem.c 2.1 88/07/29 4.0 RPCSRC";
#endif
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*
* xdr_mem.h, XDR implementation using memory buffers.
*
* Copyright (C) 1984, Sun Microsystems, Inc.
*
* If you have some data to be interpreted as external data representation
* or to be converted to external data representation in a memory buffer,
* then this is the package for you.
*
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/malloc.h>
#include <rpc/types.h>
#include <rpc/xdr.h>
static void xdrmem_destroy(XDR *);
static bool_t xdrmem_getlong_aligned(XDR *, long *);
static bool_t xdrmem_putlong_aligned(XDR *, const long *);
static bool_t xdrmem_getlong_unaligned(XDR *, long *);
static bool_t xdrmem_putlong_unaligned(XDR *, const long *);
static bool_t xdrmem_getbytes(XDR *, char *, u_int);
static bool_t xdrmem_putbytes(XDR *, const char *, u_int);
/* XXX: w/64-bit pointers, u_int not enough! */
static u_int xdrmem_getpos(XDR *);
static bool_t xdrmem_setpos(XDR *, u_int);
static int32_t *xdrmem_inline_aligned(XDR *, u_int);
static int32_t *xdrmem_inline_unaligned(XDR *, u_int);
MFdevbranch 192944 - add FreeBSD implementation of xdrmem_control needed by zfs - have zfs define xdr_ops using FreeBSD's definition - remove solaris xdr files from zfs compile
2009-05-28 08:18:12 +00:00
static bool_t xdrmem_control(XDR *xdrs, int request, void *info);
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
static const struct xdr_ops xdrmem_ops_aligned = {
xdrmem_getlong_aligned,
xdrmem_putlong_aligned,
xdrmem_getbytes,
xdrmem_putbytes,
xdrmem_getpos,
xdrmem_setpos,
xdrmem_inline_aligned,
MFdevbranch 192944 - add FreeBSD implementation of xdrmem_control needed by zfs - have zfs define xdr_ops using FreeBSD's definition - remove solaris xdr files from zfs compile
2009-05-28 08:18:12 +00:00
xdrmem_destroy,
xdrmem_control
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
};
static const struct xdr_ops xdrmem_ops_unaligned = {
xdrmem_getlong_unaligned,
xdrmem_putlong_unaligned,
xdrmem_getbytes,
xdrmem_putbytes,
xdrmem_getpos,
xdrmem_setpos,
xdrmem_inline_unaligned,
MFdevbranch 192944 - add FreeBSD implementation of xdrmem_control needed by zfs - have zfs define xdr_ops using FreeBSD's definition - remove solaris xdr files from zfs compile
2009-05-28 08:18:12 +00:00
xdrmem_destroy,
xdrmem_control
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
};
/*
* The procedure xdrmem_create initializes a stream descriptor for a
* memory buffer.
*/
void
xdrmem_create(XDR *xdrs, char *addr, u_int size, enum xdr_op op)
{
xdrs->x_op = op;
xdrs->x_ops = ((unsigned long)addr & (sizeof(int32_t) - 1))
? &xdrmem_ops_unaligned : &xdrmem_ops_aligned;
xdrs->x_private = xdrs->x_base = addr;
xdrs->x_handy = size;
}
/*ARGSUSED*/
static void
xdrmem_destroy(XDR *xdrs)
{
}
static bool_t
xdrmem_getlong_aligned(XDR *xdrs, long *lp)
{
if (xdrs->x_handy < sizeof(int32_t))
return (FALSE);
xdrs->x_handy -= sizeof(int32_t);
*lp = ntohl(*(u_int32_t *)xdrs->x_private);
xdrs->x_private = (char *)xdrs->x_private + sizeof(int32_t);
return (TRUE);
}
static bool_t
xdrmem_putlong_aligned(XDR *xdrs, const long *lp)
{
if (xdrs->x_handy < sizeof(int32_t))
return (FALSE);
xdrs->x_handy -= sizeof(int32_t);
*(u_int32_t *)xdrs->x_private = htonl((u_int32_t)*lp);
xdrs->x_private = (char *)xdrs->x_private + sizeof(int32_t);
return (TRUE);
}
static bool_t
xdrmem_getlong_unaligned(XDR *xdrs, long *lp)
{
u_int32_t l;
if (xdrs->x_handy < sizeof(int32_t))
return (FALSE);
xdrs->x_handy -= sizeof(int32_t);
memmove(&l, xdrs->x_private, sizeof(int32_t));
*lp = ntohl(l);
xdrs->x_private = (char *)xdrs->x_private + sizeof(int32_t);
return (TRUE);
}
static bool_t
xdrmem_putlong_unaligned(XDR *xdrs, const long *lp)
{
u_int32_t l;
if (xdrs->x_handy < sizeof(int32_t))
return (FALSE);
xdrs->x_handy -= sizeof(int32_t);
l = htonl((u_int32_t)*lp);
memmove(xdrs->x_private, &l, sizeof(int32_t));
xdrs->x_private = (char *)xdrs->x_private + sizeof(int32_t);
return (TRUE);
}
static bool_t
xdrmem_getbytes(XDR *xdrs, char *addr, u_int len)
{
if (xdrs->x_handy < len)
return (FALSE);
xdrs->x_handy -= len;
memmove(addr, xdrs->x_private, len);
xdrs->x_private = (char *)xdrs->x_private + len;
return (TRUE);
}
static bool_t
xdrmem_putbytes(XDR *xdrs, const char *addr, u_int len)
{
if (xdrs->x_handy < len)
return (FALSE);
xdrs->x_handy -= len;
memmove(xdrs->x_private, addr, len);
xdrs->x_private = (char *)xdrs->x_private + len;
return (TRUE);
}
static u_int
xdrmem_getpos(XDR *xdrs)
{
/* XXX w/64-bit pointers, u_int not enough! */
return (u_int)((u_long)xdrs->x_private - (u_long)xdrs->x_base);
}
static bool_t
xdrmem_setpos(XDR *xdrs, u_int pos)
{
char *newaddr = xdrs->x_base + pos;
char *lastaddr = (char *)xdrs->x_private + xdrs->x_handy;
if (newaddr > lastaddr)
return (FALSE);
xdrs->x_private = newaddr;
xdrs->x_handy = (u_int)(lastaddr - newaddr); /* XXX sizeof(u_int) <? sizeof(ptrdiff_t) */
return (TRUE);
}
static int32_t *
xdrmem_inline_aligned(XDR *xdrs, u_int len)
{
RPC: for pointers replace 0 with NULL. These are mostly cosmetical, no functional change. Found with devel/coccinelle.
2016-04-14 17:06:37 +00:00
int32_t *buf = NULL;
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
if (xdrs->x_handy >= len) {
xdrs->x_handy -= len;
buf = (int32_t *)xdrs->x_private;
xdrs->x_private = (char *)xdrs->x_private + len;
}
return (buf);
}
/* ARGSUSED */
static int32_t *
xdrmem_inline_unaligned(XDR *xdrs, u_int len)
{
return (0);
}
MFdevbranch 192944 - add FreeBSD implementation of xdrmem_control needed by zfs - have zfs define xdr_ops using FreeBSD's definition - remove solaris xdr files from zfs compile
2009-05-28 08:18:12 +00:00
static bool_t
xdrmem_control(XDR *xdrs, int request, void *info)
{
xdr_bytesrec *xptr;
int32_t *l;
int len;
switch (request) {
case XDR_GET_BYTES_AVAIL:
xptr = (xdr_bytesrec *)info;
xptr->xc_is_last_record = TRUE;
xptr->xc_num_avail = xdrs->x_handy;
return (TRUE);
case XDR_PEEK:
/*
* Return the next 4 byte unit in the XDR stream.
*/
if (xdrs->x_handy < sizeof (int32_t))
return (FALSE);
l = (int32_t *)info;
*l = (int32_t)ntohl((uint32_t)
(*((int32_t *)(xdrs->x_private))));
return (TRUE);
case XDR_SKIPBYTES:
/*
* Skip the next N bytes in the XDR stream.
*/
l = (int32_t *)info;
len = RNDUP((int)(*l));
if (xdrs->x_handy < len)
return (FALSE);
xdrs->x_handy -= len;
xdrs->x_private = (char *)xdrs->x_private + len;
return (TRUE);
}
return (FALSE);
}
Reference in New Issue Copy Permalink