freebsd-dev/sys/nfsserver/nfs_srvcache.c
Mohan Srinivasan 5943f1a781 Size the NFS server dupreq cache on the basis of nmbclusters. On servers
with low nmbclusters, we tie up too many mbclusters in the NFS duplicate
request cache. This change limits the size of the dupreq cache to 1/2
the nmbclusters (and flaots in a range of [64, 2048]).

MFC after 2 weeks.

Reported by: Steve Kargl, David O'Brien
Tested by:   Steve Kargl
2006-06-23 00:42:26 +00:00

375 lines
9.8 KiB
C

/*-
* 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.
* 4. Neither the name of the University nor the names of its contributors
* 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.
*
* @(#)nfs_srvcache.c 8.3 (Berkeley) 3/30/95
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
/*
* Reference: Chet Juszczak, "Improving the Performance and Correctness
* of an NFS Server", in Proc. Winter 1989 USENIX Conference,
* pages 53-63. San Diego, February 1989.
*/
#include <sys/param.h>
#include <sys/malloc.h>
#include <sys/mount.h>
#include <sys/systm.h>
#include <sys/lock.h>
#include <sys/mbuf.h>
#include <sys/mutex.h>
#include <sys/socket.h>
#include <sys/socketvar.h> /* for sodupsockaddr */
#include <sys/eventhandler.h>
#include <netinet/in.h>
#include <nfs/rpcv2.h>
#include <nfs/nfsproto.h>
#include <nfsserver/nfs.h>
#include <nfsserver/nfsrvcache.h>
static long numnfsrvcache;
static long desirednfsrvcache;
#define NFSRCHASH(xid) \
(&nfsrvhashtbl[((xid) + ((xid) >> 24)) & nfsrvhash])
static LIST_HEAD(nfsrvhash, nfsrvcache) *nfsrvhashtbl;
static TAILQ_HEAD(nfsrvlru, nfsrvcache) nfsrvlruhead;
static u_long nfsrvhash;
#define TRUE 1
#define FALSE 0
#define NETFAMILY(rp) \
(((rp)->rc_flag & RC_NAM) ? (rp)->rc_nam->sa_family : AF_INET)
/*
* Static array that defines which nfs rpc's are nonidempotent
*/
static const int nonidempotent[NFS_NPROCS] = {
FALSE,
FALSE,
TRUE,
FALSE,
FALSE,
FALSE,
FALSE,
TRUE,
TRUE,
TRUE,
TRUE,
TRUE,
TRUE,
TRUE,
TRUE,
TRUE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
};
/* True iff the rpc reply is an nfs status ONLY! */
static const int nfsv2_repstat[NFS_NPROCS] = {
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
FALSE,
TRUE,
TRUE,
TRUE,
TRUE,
FALSE,
TRUE,
FALSE,
FALSE,
};
/*
* Size the NFS server's duplicate request cache at 1/2 the nmbclsters, floating
* within a (64, 2048) range. This is to prevent all mbuf clusters being tied up
* in the NFS dupreq cache for small values of nmbclusters.
*/
static void
nfsrvcache_size_change(void *tag)
{
desirednfsrvcache = nmbclusters /2;
if (desirednfsrvcache > NFSRVCACHE_MAX_SIZE)
desirednfsrvcache = NFSRVCACHE_MAX_SIZE;
if (desirednfsrvcache < NFSRVCACHE_MIN_SIZE)
desirednfsrvcache = NFSRVCACHE_MIN_SIZE;
}
/*
* Initialize the server request cache list
*/
void
nfsrv_initcache(void)
{
nfsrvcache_size_change(NULL);
nfsrvhashtbl = hashinit(desirednfsrvcache, M_NFSD, &nfsrvhash);
TAILQ_INIT(&nfsrvlruhead);
EVENTHANDLER_REGISTER(nmbclusters_change, nfsrvcache_size_change, NULL,
EVENTHANDLER_PRI_FIRST);
}
/*
* Look for the request in the cache
* If found then
* return action and optionally reply
* else
* insert it in the cache
*
* The rules are as follows:
* - if in progress, return DROP request
* - if completed within DELAY of the current time, return DROP it
* - if completed a longer time ago return REPLY if the reply was cached or
* return DOIT
* Update/add new request at end of lru list
*/
int
nfsrv_getcache(struct nfsrv_descript *nd, struct mbuf **repp)
{
struct nfsrvcache *rp;
struct mbuf *mb;
struct sockaddr_in *saddr;
caddr_t bpos;
int ret;
NFSD_LOCK_ASSERT();
/*
* Don't cache recent requests for reliable transport protocols.
* (Maybe we should for the case of a reconnect, but..)
*/
if (!nd->nd_nam2)
return (RC_DOIT);
loop:
LIST_FOREACH(rp, NFSRCHASH(nd->nd_retxid), rc_hash) {
if (nd->nd_retxid == rp->rc_xid && nd->nd_procnum == rp->rc_proc &&
netaddr_match(NETFAMILY(rp), &rp->rc_haddr, nd->nd_nam)) {
NFS_DPF(RC, ("H%03x", rp->rc_xid & 0xfff));
if ((rp->rc_flag & RC_LOCKED) != 0) {
rp->rc_flag |= RC_WANTED;
(void) msleep(rp, &nfsd_mtx, PZERO-1,
"nfsrc", 0);
goto loop;
}
rp->rc_flag |= RC_LOCKED;
/* If not at end of LRU chain, move it there */
if (TAILQ_NEXT(rp, rc_lru)) {
TAILQ_REMOVE(&nfsrvlruhead, rp, rc_lru);
TAILQ_INSERT_TAIL(&nfsrvlruhead, rp, rc_lru);
}
if (rp->rc_state == RC_UNUSED)
panic("nfsrv cache");
if (rp->rc_state == RC_INPROG) {
nfsrvstats.srvcache_inproghits++;
ret = RC_DROPIT;
} else if (rp->rc_flag & RC_REPSTATUS) {
nfsrvstats.srvcache_nonidemdonehits++;
*repp = nfs_rephead(0, nd, rp->rc_status,
&mb, &bpos);
ret = RC_REPLY;
} else if (rp->rc_flag & RC_REPMBUF) {
nfsrvstats.srvcache_nonidemdonehits++;
NFSD_UNLOCK();
*repp = m_copym(rp->rc_reply, 0, M_COPYALL,
M_TRYWAIT);
NFSD_LOCK();
ret = RC_REPLY;
} else {
nfsrvstats.srvcache_idemdonehits++;
rp->rc_state = RC_INPROG;
ret = RC_DOIT;
}
rp->rc_flag &= ~RC_LOCKED;
if (rp->rc_flag & RC_WANTED) {
rp->rc_flag &= ~RC_WANTED;
wakeup(rp);
}
return (ret);
}
}
nfsrvstats.srvcache_misses++;
NFS_DPF(RC, ("M%03x", nd->nd_retxid & 0xfff));
if (numnfsrvcache < desirednfsrvcache) {
NFSD_UNLOCK();
rp = (struct nfsrvcache *)malloc((u_long)sizeof *rp,
M_NFSD, M_WAITOK | M_ZERO);
NFSD_LOCK();
numnfsrvcache++;
rp->rc_flag = RC_LOCKED;
} else {
rp = TAILQ_FIRST(&nfsrvlruhead);
while ((rp->rc_flag & RC_LOCKED) != 0) {
rp->rc_flag |= RC_WANTED;
(void) msleep(rp, &nfsd_mtx, PZERO-1, "nfsrc", 0);
rp = TAILQ_FIRST(&nfsrvlruhead);
}
rp->rc_flag |= RC_LOCKED;
LIST_REMOVE(rp, rc_hash);
TAILQ_REMOVE(&nfsrvlruhead, rp, rc_lru);
if (rp->rc_flag & RC_REPMBUF)
m_freem(rp->rc_reply);
if (rp->rc_flag & RC_NAM)
FREE(rp->rc_nam, M_SONAME);
rp->rc_flag &= (RC_LOCKED | RC_WANTED);
}
TAILQ_INSERT_TAIL(&nfsrvlruhead, rp, rc_lru);
rp->rc_state = RC_INPROG;
rp->rc_xid = nd->nd_retxid;
saddr = (struct sockaddr_in *)nd->nd_nam;
switch (saddr->sin_family) {
case AF_INET:
rp->rc_flag |= RC_INETADDR;
rp->rc_inetaddr = saddr->sin_addr.s_addr;
break;
/* case AF_INET6: */
/* case AF_ISO: */
default:
/*
* XXXRW: Seems like we should only set RC_NAM if we
* actually manage to set rc_nam to something non-NULL.
*/
rp->rc_flag |= RC_NAM;
rp->rc_nam = sodupsockaddr(nd->nd_nam, M_NOWAIT);
break;
};
rp->rc_proc = nd->nd_procnum;
LIST_INSERT_HEAD(NFSRCHASH(nd->nd_retxid), rp, rc_hash);
rp->rc_flag &= ~RC_LOCKED;
if (rp->rc_flag & RC_WANTED) {
rp->rc_flag &= ~RC_WANTED;
wakeup(rp);
}
return (RC_DOIT);
}
/*
* Update a request cache entry after the rpc has been done
*/
void
nfsrv_updatecache(struct nfsrv_descript *nd, int repvalid, struct mbuf *repmbuf)
{
struct nfsrvcache *rp;
NFSD_LOCK_ASSERT();
if (!nd->nd_nam2)
return;
loop:
LIST_FOREACH(rp, NFSRCHASH(nd->nd_retxid), rc_hash) {
if (nd->nd_retxid == rp->rc_xid && nd->nd_procnum == rp->rc_proc &&
netaddr_match(NETFAMILY(rp), &rp->rc_haddr, nd->nd_nam)) {
NFS_DPF(RC, ("U%03x", rp->rc_xid & 0xfff));
if ((rp->rc_flag & RC_LOCKED) != 0) {
rp->rc_flag |= RC_WANTED;
(void) msleep(rp, &nfsd_mtx, PZERO-1,
"nfsrc", 0);
goto loop;
}
rp->rc_flag |= RC_LOCKED;
if (rp->rc_state == RC_DONE) {
/*
* This can occur if the cache is too small.
* Retransmits of the same request aren't
* dropped so we may see the operation
* complete more then once.
*/
if (rp->rc_flag & RC_REPMBUF) {
m_freem(rp->rc_reply);
rp->rc_flag &= ~RC_REPMBUF;
}
}
rp->rc_state = RC_DONE;
/*
* If we have a valid reply update status and save
* the reply for non-idempotent rpc's.
*/
if (repvalid && nonidempotent[nd->nd_procnum]) {
if ((nd->nd_flag & ND_NFSV3) == 0 &&
nfsv2_repstat[
nfsrvv2_procid[nd->nd_procnum]]) {
rp->rc_status = nd->nd_repstat;
rp->rc_flag |= RC_REPSTATUS;
} else {
NFSD_UNLOCK();
rp->rc_reply = m_copym(repmbuf,
0, M_COPYALL, M_TRYWAIT);
NFSD_LOCK();
rp->rc_flag |= RC_REPMBUF;
}
}
rp->rc_flag &= ~RC_LOCKED;
if (rp->rc_flag & RC_WANTED) {
rp->rc_flag &= ~RC_WANTED;
wakeup(rp);
}
return;
}
}
NFS_DPF(RC, ("L%03x", nd->nd_retxid & 0xfff));
}
/*
* Clean out the cache. Called when the last nfsd terminates.
*/
void
nfsrv_cleancache(void)
{
struct nfsrvcache *rp, *nextrp;
NFSD_LOCK_ASSERT();
for (rp = TAILQ_FIRST(&nfsrvlruhead); rp != 0; rp = nextrp) {
nextrp = TAILQ_NEXT(rp, rc_lru);
LIST_REMOVE(rp, rc_hash);
TAILQ_REMOVE(&nfsrvlruhead, rp, rc_lru);
if (rp->rc_flag & RC_REPMBUF)
m_freem(rp->rc_reply);
if (rp->rc_flag & RC_NAM)
free(rp->rc_nam, M_SONAME);
free(rp, M_NFSD);
}
numnfsrvcache = 0;
}