2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* linux/include/linux/sunrpc/svc.h
|
|
|
|
*
|
|
|
|
* RPC server declarations.
|
|
|
|
*
|
|
|
|
* Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de>
|
|
|
|
*/
|
|
|
|
|
|
|
|
|
|
|
|
#ifndef SUNRPC_SVC_H
|
|
|
|
#define SUNRPC_SVC_H
|
|
|
|
|
|
|
|
#include <linux/in.h>
|
2007-02-12 15:53:35 +07:00
|
|
|
#include <linux/in6.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <linux/sunrpc/types.h>
|
|
|
|
#include <linux/sunrpc/xdr.h>
|
2006-10-04 16:15:47 +07:00
|
|
|
#include <linux/sunrpc/auth.h>
|
2005-04-17 05:20:36 +07:00
|
|
|
#include <linux/sunrpc/svcauth.h>
|
|
|
|
#include <linux/wait.h>
|
|
|
|
#include <linux/mm.h>
|
|
|
|
|
2006-10-02 16:17:59 +07:00
|
|
|
/*
|
|
|
|
* This is the RPC server thread function prototype
|
|
|
|
*/
|
2008-06-10 19:40:38 +07:00
|
|
|
typedef int (*svc_thread_fn)(void *);
|
2006-10-02 16:17:58 +07:00
|
|
|
|
2009-01-13 17:26:36 +07:00
|
|
|
/* statistics for svc_pool structures */
|
|
|
|
struct svc_pool_stats {
|
|
|
|
unsigned long packets;
|
|
|
|
unsigned long sockets_queued;
|
|
|
|
unsigned long threads_woken;
|
|
|
|
unsigned long overloads_avoided;
|
|
|
|
unsigned long threads_timedout;
|
|
|
|
};
|
|
|
|
|
2006-10-02 16:17:58 +07:00
|
|
|
/*
|
|
|
|
*
|
|
|
|
* RPC service thread pool.
|
|
|
|
*
|
|
|
|
* Pool of threads and temporary sockets. Generally there is only
|
|
|
|
* a single one of these per RPC service, but on NUMA machines those
|
|
|
|
* services that can benefit from it (i.e. nfs but not lockd) will
|
|
|
|
* have one pool per NUMA node. This optimisation reduces cross-
|
|
|
|
* node traffic on multi-node NUMA NFS servers.
|
|
|
|
*/
|
|
|
|
struct svc_pool {
|
|
|
|
unsigned int sp_id; /* pool id; also node id on NUMA */
|
|
|
|
spinlock_t sp_lock; /* protects all fields */
|
|
|
|
struct list_head sp_threads; /* idle server threads */
|
|
|
|
struct list_head sp_sockets; /* pending sockets */
|
|
|
|
unsigned int sp_nrthreads; /* # of threads in pool */
|
2006-10-02 16:17:59 +07:00
|
|
|
struct list_head sp_all_threads; /* all server threads */
|
knfsd: avoid overloading the CPU scheduler with enormous load averages
Avoid overloading the CPU scheduler with enormous load averages
when handling high call-rate NFS loads. When the knfsd bottom half
is made aware of an incoming call by the socket layer, it tries to
choose an nfsd thread and wake it up. As long as there are idle
threads, one will be woken up.
If there are lot of nfsd threads (a sensible configuration when
the server is disk-bound or is running an HSM), there will be many
more nfsd threads than CPUs to run them. Under a high call-rate
low service-time workload, the result is that almost every nfsd is
runnable, but only a handful are actually able to run. This situation
causes two significant problems:
1. The CPU scheduler takes over 10% of each CPU, which is robbing
the nfsd threads of valuable CPU time.
2. At a high enough load, the nfsd threads starve userspace threads
of CPU time, to the point where daemons like portmap and rpc.mountd
do not schedule for tens of seconds at a time. Clients attempting
to mount an NFS filesystem timeout at the very first step (opening
a TCP connection to portmap) because portmap cannot wake up from
select() and call accept() in time.
Disclaimer: these effects were observed on a SLES9 kernel, modern
kernels' schedulers may behave more gracefully.
The solution is simple: keep in each svc_pool a counter of the number
of threads which have been woken but have not yet run, and do not wake
any more if that count reaches an arbitrary small threshold.
Testing was on a 4 CPU 4 NIC Altix using 4 IRIX clients, each with 16
synthetic client threads simulating an rsync (i.e. recursive directory
listing) workload reading from an i386 RH9 install image (161480
regular files in 10841 directories) on the server. That tree is small
enough to fill in the server's RAM so no disk traffic was involved.
This setup gives a sustained call rate in excess of 60000 calls/sec
before being CPU-bound on the server. The server was running 128 nfsds.
Profiling showed schedule() taking 6.7% of every CPU, and __wake_up()
taking 5.2%. This patch drops those contributions to 3.0% and 2.2%.
Load average was over 120 before the patch, and 20.9 after.
This patch is a forward-ported version of knfsd-avoid-nfsd-overload
which has been shipping in the SGI "Enhanced NFS" product since 2006.
It has been posted before:
http://article.gmane.org/gmane.linux.nfs/10374
Signed-off-by: Greg Banks <gnb@sgi.com>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
2009-01-13 17:26:35 +07:00
|
|
|
int sp_nwaking; /* number of threads woken but not yet active */
|
2009-01-13 17:26:36 +07:00
|
|
|
struct svc_pool_stats sp_stats; /* statistics on pool operation */
|
2006-10-02 16:17:58 +07:00
|
|
|
} ____cacheline_aligned_in_smp;
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* RPC service.
|
|
|
|
*
|
|
|
|
* An RPC service is a ``daemon,'' possibly multithreaded, which
|
|
|
|
* receives and processes incoming RPC messages.
|
|
|
|
* It has one or more transport sockets associated with it, and maintains
|
|
|
|
* a list of idle threads waiting for input.
|
|
|
|
*
|
|
|
|
* We currently do not support more than one RPC program per daemon.
|
|
|
|
*/
|
|
|
|
struct svc_serv {
|
|
|
|
struct svc_program * sv_program; /* RPC program */
|
|
|
|
struct svc_stat * sv_stats; /* RPC statistics */
|
|
|
|
spinlock_t sv_lock;
|
|
|
|
unsigned int sv_nrthreads; /* # of server threads */
|
2008-10-20 22:51:57 +07:00
|
|
|
unsigned int sv_maxconn; /* max connections allowed or
|
|
|
|
* '0' causing max to be based
|
|
|
|
* on number of threads. */
|
|
|
|
|
2006-10-06 14:44:05 +07:00
|
|
|
unsigned int sv_max_payload; /* datagram payload size */
|
|
|
|
unsigned int sv_max_mesg; /* max_payload + 1 page for overheads */
|
2005-04-17 05:20:36 +07:00
|
|
|
unsigned int sv_xdrsize; /* XDR buffer size */
|
|
|
|
struct list_head sv_permsocks; /* all permanent sockets */
|
|
|
|
struct list_head sv_tempsocks; /* all temporary sockets */
|
|
|
|
int sv_tmpcnt; /* count of temporary sockets */
|
2006-10-02 16:17:54 +07:00
|
|
|
struct timer_list sv_temptimer; /* timer for aging temporary sockets */
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
char * sv_name; /* service name */
|
2006-10-02 16:17:44 +07:00
|
|
|
|
2006-10-02 16:17:58 +07:00
|
|
|
unsigned int sv_nrpools; /* number of thread pools */
|
|
|
|
struct svc_pool * sv_pools; /* array of thread pools */
|
|
|
|
|
2006-10-02 16:17:44 +07:00
|
|
|
void (*sv_shutdown)(struct svc_serv *serv);
|
|
|
|
/* Callback to use when last thread
|
|
|
|
* exits.
|
|
|
|
*/
|
2006-10-02 16:17:59 +07:00
|
|
|
|
|
|
|
struct module * sv_module; /* optional module to count when
|
|
|
|
* adding threads */
|
|
|
|
svc_thread_fn sv_function; /* main function for threads */
|
2009-04-01 20:22:58 +07:00
|
|
|
#if defined(CONFIG_NFS_V4_1)
|
|
|
|
struct list_head sv_cb_list; /* queue for callback requests
|
|
|
|
* that arrive over the same
|
|
|
|
* connection */
|
|
|
|
spinlock_t sv_cb_lock; /* protects the svc_cb_list */
|
|
|
|
wait_queue_head_t sv_cb_waitq; /* sleep here if there are no
|
|
|
|
* entries in the svc_cb_list */
|
2009-04-01 20:23:10 +07:00
|
|
|
struct svc_xprt *bc_xprt;
|
2009-04-01 20:22:58 +07:00
|
|
|
#endif /* CONFIG_NFS_V4_1 */
|
2005-04-17 05:20:36 +07:00
|
|
|
};
|
|
|
|
|
2006-10-02 16:17:58 +07:00
|
|
|
/*
|
|
|
|
* We use sv_nrthreads as a reference count. svc_destroy() drops
|
|
|
|
* this refcount, so we need to bump it up around operations that
|
|
|
|
* change the number of threads. Horrible, but there it is.
|
|
|
|
* Should be called with the BKL held.
|
|
|
|
*/
|
|
|
|
static inline void svc_get(struct svc_serv *serv)
|
|
|
|
{
|
|
|
|
serv->sv_nrthreads++;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Maximum payload size supported by a kernel RPC server.
|
|
|
|
* This is use to determine the max number of pages nfsd is
|
|
|
|
* willing to return in a single READ operation.
|
2006-10-04 16:15:47 +07:00
|
|
|
*
|
|
|
|
* These happen to all be powers of 2, which is not strictly
|
|
|
|
* necessary but helps enforce the real limitation, which is
|
|
|
|
* that they should be multiples of PAGE_CACHE_SIZE.
|
|
|
|
*
|
|
|
|
* For UDP transports, a block plus NFS,RPC, and UDP headers
|
|
|
|
* has to fit into the IP datagram limit of 64K. The largest
|
|
|
|
* feasible number for all known page sizes is probably 48K,
|
|
|
|
* but we choose 32K here. This is the same as the historical
|
|
|
|
* Linux limit; someone who cares more about NFS/UDP performance
|
|
|
|
* can test a larger number.
|
|
|
|
*
|
|
|
|
* For TCP transports we have more freedom. A size of 1MB is
|
|
|
|
* chosen to match the client limit. Other OSes are known to
|
|
|
|
* have larger limits, but those numbers are probably beyond
|
|
|
|
* the point of diminishing returns.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2006-10-04 16:15:47 +07:00
|
|
|
#define RPCSVC_MAXPAYLOAD (1*1024*1024u)
|
|
|
|
#define RPCSVC_MAXPAYLOAD_TCP RPCSVC_MAXPAYLOAD
|
|
|
|
#define RPCSVC_MAXPAYLOAD_UDP (32*1024u)
|
|
|
|
|
|
|
|
extern u32 svc_max_payload(const struct svc_rqst *rqstp);
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
/*
|
|
|
|
* RPC Requsts and replies are stored in one or more pages.
|
|
|
|
* We maintain an array of pages for each server thread.
|
|
|
|
* Requests are copied into these pages as they arrive. Remaining
|
|
|
|
* pages are available to write the reply into.
|
|
|
|
*
|
|
|
|
* Pages are sent using ->sendpage so each server thread needs to
|
|
|
|
* allocate more to replace those used in sending. To help keep track
|
|
|
|
* of these pages we have a receive list where all pages initialy live,
|
|
|
|
* and a send list where pages are moved to when there are to be part
|
|
|
|
* of a reply.
|
|
|
|
*
|
|
|
|
* We use xdr_buf for holding responses as it fits well with NFS
|
|
|
|
* read responses (that have a header, and some data pages, and possibly
|
|
|
|
* a tail) and means we can share some client side routines.
|
|
|
|
*
|
|
|
|
* The xdr_buf.head kvec always points to the first page in the rq_*pages
|
|
|
|
* list. The xdr_buf.pages pointer points to the second page on that
|
|
|
|
* list. xdr_buf.tail points to the end of the first page.
|
|
|
|
* This assumes that the non-page part of an rpc reply will fit
|
|
|
|
* in a page - NFSd ensures this. lockd also has no trouble.
|
|
|
|
*
|
|
|
|
* Each request/reply pair can have at most one "payload", plus two pages,
|
|
|
|
* one for the request, and one for the reply.
|
2007-01-26 15:56:59 +07:00
|
|
|
* We using ->sendfile to return read data, we might need one extra page
|
|
|
|
* if the request is not page-aligned. So add another '1'.
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
2007-01-26 15:56:59 +07:00
|
|
|
#define RPCSVC_MAXPAGES ((RPCSVC_MAXPAYLOAD+PAGE_SIZE-1)/PAGE_SIZE \
|
|
|
|
+ 2 + 1)
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2006-09-27 12:28:46 +07:00
|
|
|
static inline u32 svc_getnl(struct kvec *iov)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2006-09-27 12:28:46 +07:00
|
|
|
__be32 val, *vp;
|
2005-04-17 05:20:36 +07:00
|
|
|
vp = iov->iov_base;
|
|
|
|
val = *vp++;
|
|
|
|
iov->iov_base = (void*)vp;
|
2006-09-27 12:28:46 +07:00
|
|
|
iov->iov_len -= sizeof(__be32);
|
|
|
|
return ntohl(val);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void svc_putnl(struct kvec *iov, u32 val)
|
|
|
|
{
|
|
|
|
__be32 *vp = iov->iov_base + iov->iov_len;
|
|
|
|
*vp = htonl(val);
|
|
|
|
iov->iov_len += sizeof(__be32);
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline __be32 svc_getu32(struct kvec *iov)
|
|
|
|
{
|
|
|
|
__be32 val, *vp;
|
|
|
|
vp = iov->iov_base;
|
|
|
|
val = *vp++;
|
|
|
|
iov->iov_base = (void*)vp;
|
|
|
|
iov->iov_len -= sizeof(__be32);
|
2005-04-17 05:20:36 +07:00
|
|
|
return val;
|
|
|
|
}
|
|
|
|
|
|
|
|
static inline void svc_ungetu32(struct kvec *iov)
|
|
|
|
{
|
2006-09-27 12:28:46 +07:00
|
|
|
__be32 *vp = (__be32 *)iov->iov_base;
|
2005-04-17 05:20:36 +07:00
|
|
|
iov->iov_base = (void *)(vp - 1);
|
|
|
|
iov->iov_len += sizeof(*vp);
|
|
|
|
}
|
|
|
|
|
2006-09-27 12:28:46 +07:00
|
|
|
static inline void svc_putu32(struct kvec *iov, __be32 val)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
2006-09-27 12:28:46 +07:00
|
|
|
__be32 *vp = iov->iov_base + iov->iov_len;
|
2005-04-17 05:20:36 +07:00
|
|
|
*vp = val;
|
2006-09-27 12:28:46 +07:00
|
|
|
iov->iov_len += sizeof(__be32);
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
2007-02-12 15:53:35 +07:00
|
|
|
union svc_addr_u {
|
|
|
|
struct in_addr addr;
|
|
|
|
struct in6_addr addr6;
|
|
|
|
};
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* The context of a single thread, including the request currently being
|
|
|
|
* processed.
|
|
|
|
*/
|
|
|
|
struct svc_rqst {
|
|
|
|
struct list_head rq_list; /* idle list */
|
2006-10-02 16:17:59 +07:00
|
|
|
struct list_head rq_all; /* all threads list */
|
2007-12-31 10:08:22 +07:00
|
|
|
struct svc_xprt * rq_xprt; /* transport ptr */
|
2007-02-12 15:53:34 +07:00
|
|
|
struct sockaddr_storage rq_addr; /* peer address */
|
|
|
|
size_t rq_addrlen;
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
struct svc_serv * rq_server; /* RPC service definition */
|
2006-10-02 16:17:58 +07:00
|
|
|
struct svc_pool * rq_pool; /* thread pool */
|
2005-04-17 05:20:36 +07:00
|
|
|
struct svc_procedure * rq_procinfo; /* procedure info */
|
|
|
|
struct auth_ops * rq_authop; /* authentication flavour */
|
2007-07-17 18:04:42 +07:00
|
|
|
u32 rq_flavor; /* pseudoflavor */
|
2005-04-17 05:20:36 +07:00
|
|
|
struct svc_cred rq_cred; /* auth info */
|
2007-12-31 10:07:25 +07:00
|
|
|
void * rq_xprt_ctxt; /* transport specific context ptr */
|
2005-04-17 05:20:36 +07:00
|
|
|
struct svc_deferred_req*rq_deferred; /* deferred request we are replaying */
|
2009-04-03 12:27:32 +07:00
|
|
|
int rq_usedeferral; /* use deferral */
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2007-12-31 10:08:29 +07:00
|
|
|
size_t rq_xprt_hlen; /* xprt header len */
|
2005-04-17 05:20:36 +07:00
|
|
|
struct xdr_buf rq_arg;
|
|
|
|
struct xdr_buf rq_res;
|
2006-10-04 16:15:46 +07:00
|
|
|
struct page * rq_pages[RPCSVC_MAXPAGES];
|
|
|
|
struct page * *rq_respages; /* points into rq_pages */
|
|
|
|
int rq_resused; /* number of pages used for result */
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2006-10-04 16:15:47 +07:00
|
|
|
struct kvec rq_vec[RPCSVC_MAXPAGES]; /* generally useful.. */
|
|
|
|
|
2006-09-27 12:29:38 +07:00
|
|
|
__be32 rq_xid; /* transmission id */
|
2005-04-17 05:20:36 +07:00
|
|
|
u32 rq_prog; /* program number */
|
|
|
|
u32 rq_vers; /* program version */
|
|
|
|
u32 rq_proc; /* procedure number */
|
|
|
|
u32 rq_prot; /* IP protocol */
|
|
|
|
unsigned short
|
|
|
|
rq_secure : 1; /* secure port */
|
|
|
|
|
2007-02-12 15:53:35 +07:00
|
|
|
union svc_addr_u rq_daddr; /* dest addr of request
|
|
|
|
* - reply from here */
|
2005-04-17 05:20:36 +07:00
|
|
|
|
|
|
|
void * rq_argp; /* decoded arguments */
|
|
|
|
void * rq_resp; /* xdr'd results */
|
|
|
|
void * rq_auth_data; /* flavor-specific data */
|
|
|
|
|
|
|
|
int rq_reserved; /* space on socket outq
|
|
|
|
* reserved for this request
|
|
|
|
*/
|
|
|
|
|
|
|
|
struct cache_req rq_chandle; /* handle passed to caches for
|
|
|
|
* request delaying
|
|
|
|
*/
|
|
|
|
/* Catering to nfsd */
|
|
|
|
struct auth_domain * rq_client; /* RPC peer info */
|
knfsd: nfsd: set rq_client to ip-address-determined-domain
We want it to be possible for users to restrict exports both by IP address and
by pseudoflavor. The pseudoflavor information has previously been passed
using special auth_domains stored in the rq_client field. After the preceding
patch that stored the pseudoflavor in rq_pflavor, that's now superfluous; so
now we use rq_client for the ip information, as auth_null and auth_unix do.
However, we keep around the special auth_domain in the rq_gssclient field for
backwards compatibility purposes, so we can still do upcalls using the old
"gss/pseudoflavor" auth_domain if upcalls using the unix domain to give us an
appropriate export. This allows us to continue supporting old mountd.
In fact, for this first patch, we always use the "gss/pseudoflavor"
auth_domain (and only it) if it is available; thus rq_client is ignored in the
auth_gss case, and this patch on its own makes no change in behavior; that
will be left to later patches.
Note on idmap: I'm almost tempted to just replace the auth_domain in the idmap
upcall by a dummy value--no version of idmapd has ever used it, and it's
unlikely anyone really wants to perform idmapping differently depending on the
where the client is (they may want to perform *credential* mapping
differently, but that's a different matter--the idmapper just handles id's
used in getattr and setattr). But I'm updating the idmapd code anyway, just
out of general backwards-compatibility paranoia.
Signed-off-by: "J. Bruce Fields" <bfields@citi.umich.edu>
Signed-off-by: Neil Brown <neilb@suse.de>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2007-07-17 18:04:46 +07:00
|
|
|
struct auth_domain * rq_gssclient; /* "gss/"-style peer info */
|
2005-04-17 05:20:36 +07:00
|
|
|
struct svc_cacherep * rq_cacherep; /* cache info */
|
|
|
|
struct knfsd_fh * rq_reffh; /* Referrence filehandle, used to
|
|
|
|
* determine what device number
|
|
|
|
* to report (real or virtual)
|
|
|
|
*/
|
2007-06-13 02:22:14 +07:00
|
|
|
int rq_splice_ok; /* turned off in gss privacy
|
2006-06-30 15:56:19 +07:00
|
|
|
* to prevent encrypting page
|
|
|
|
* cache pages */
|
2005-04-17 05:20:36 +07:00
|
|
|
wait_queue_head_t rq_wait; /* synchronization */
|
2006-10-02 16:17:59 +07:00
|
|
|
struct task_struct *rq_task; /* service thread */
|
knfsd: avoid overloading the CPU scheduler with enormous load averages
Avoid overloading the CPU scheduler with enormous load averages
when handling high call-rate NFS loads. When the knfsd bottom half
is made aware of an incoming call by the socket layer, it tries to
choose an nfsd thread and wake it up. As long as there are idle
threads, one will be woken up.
If there are lot of nfsd threads (a sensible configuration when
the server is disk-bound or is running an HSM), there will be many
more nfsd threads than CPUs to run them. Under a high call-rate
low service-time workload, the result is that almost every nfsd is
runnable, but only a handful are actually able to run. This situation
causes two significant problems:
1. The CPU scheduler takes over 10% of each CPU, which is robbing
the nfsd threads of valuable CPU time.
2. At a high enough load, the nfsd threads starve userspace threads
of CPU time, to the point where daemons like portmap and rpc.mountd
do not schedule for tens of seconds at a time. Clients attempting
to mount an NFS filesystem timeout at the very first step (opening
a TCP connection to portmap) because portmap cannot wake up from
select() and call accept() in time.
Disclaimer: these effects were observed on a SLES9 kernel, modern
kernels' schedulers may behave more gracefully.
The solution is simple: keep in each svc_pool a counter of the number
of threads which have been woken but have not yet run, and do not wake
any more if that count reaches an arbitrary small threshold.
Testing was on a 4 CPU 4 NIC Altix using 4 IRIX clients, each with 16
synthetic client threads simulating an rsync (i.e. recursive directory
listing) workload reading from an i386 RH9 install image (161480
regular files in 10841 directories) on the server. That tree is small
enough to fill in the server's RAM so no disk traffic was involved.
This setup gives a sustained call rate in excess of 60000 calls/sec
before being CPU-bound on the server. The server was running 128 nfsds.
Profiling showed schedule() taking 6.7% of every CPU, and __wake_up()
taking 5.2%. This patch drops those contributions to 3.0% and 2.2%.
Load average was over 120 before the patch, and 20.9 after.
This patch is a forward-ported version of knfsd-avoid-nfsd-overload
which has been shipping in the SGI "Enhanced NFS" product since 2006.
It has been posted before:
http://article.gmane.org/gmane.linux.nfs/10374
Signed-off-by: Greg Banks <gnb@sgi.com>
Signed-off-by: J. Bruce Fields <bfields@citi.umich.edu>
2009-01-13 17:26:35 +07:00
|
|
|
int rq_waking; /* 1 if thread is being woken */
|
2005-04-17 05:20:36 +07:00
|
|
|
};
|
|
|
|
|
2007-02-12 15:53:34 +07:00
|
|
|
/*
|
|
|
|
* Rigorous type checking on sockaddr type conversions
|
|
|
|
*/
|
2008-08-28 03:57:46 +07:00
|
|
|
static inline struct sockaddr_in *svc_addr_in(const struct svc_rqst *rqst)
|
2007-02-12 15:53:34 +07:00
|
|
|
{
|
|
|
|
return (struct sockaddr_in *) &rqst->rq_addr;
|
|
|
|
}
|
|
|
|
|
2008-08-28 03:57:46 +07:00
|
|
|
static inline struct sockaddr_in6 *svc_addr_in6(const struct svc_rqst *rqst)
|
2007-02-12 15:53:38 +07:00
|
|
|
{
|
|
|
|
return (struct sockaddr_in6 *) &rqst->rq_addr;
|
|
|
|
}
|
|
|
|
|
2008-08-28 03:57:46 +07:00
|
|
|
static inline struct sockaddr *svc_addr(const struct svc_rqst *rqst)
|
2007-02-12 15:53:34 +07:00
|
|
|
{
|
|
|
|
return (struct sockaddr *) &rqst->rq_addr;
|
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/*
|
|
|
|
* Check buffer bounds after decoding arguments
|
|
|
|
*/
|
|
|
|
static inline int
|
2006-09-27 12:29:38 +07:00
|
|
|
xdr_argsize_check(struct svc_rqst *rqstp, __be32 *p)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
char *cp = (char *)p;
|
|
|
|
struct kvec *vec = &rqstp->rq_arg.head[0];
|
2005-11-07 16:00:26 +07:00
|
|
|
return cp >= (char*)vec->iov_base
|
|
|
|
&& cp <= (char*)vec->iov_base + vec->iov_len;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
|
|
|
|
static inline int
|
2006-09-27 12:29:38 +07:00
|
|
|
xdr_ressize_check(struct svc_rqst *rqstp, __be32 *p)
|
2005-04-17 05:20:36 +07:00
|
|
|
{
|
|
|
|
struct kvec *vec = &rqstp->rq_res.head[0];
|
|
|
|
char *cp = (char*)p;
|
|
|
|
|
|
|
|
vec->iov_len = cp - (char*)vec->iov_base;
|
|
|
|
|
|
|
|
return vec->iov_len <= PAGE_SIZE;
|
|
|
|
}
|
|
|
|
|
2006-10-04 16:15:46 +07:00
|
|
|
static inline void svc_free_res_pages(struct svc_rqst *rqstp)
|
2005-06-23 00:16:26 +07:00
|
|
|
{
|
2006-10-04 16:15:46 +07:00
|
|
|
while (rqstp->rq_resused) {
|
|
|
|
struct page **pp = (rqstp->rq_respages +
|
|
|
|
--rqstp->rq_resused);
|
|
|
|
if (*pp) {
|
|
|
|
put_page(*pp);
|
|
|
|
*pp = NULL;
|
2005-04-17 05:20:36 +07:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
struct svc_deferred_req {
|
|
|
|
u32 prot; /* protocol (UDP or TCP) */
|
2007-12-31 10:08:10 +07:00
|
|
|
struct svc_xprt *xprt;
|
2007-02-12 15:53:33 +07:00
|
|
|
struct sockaddr_storage addr; /* where reply must go */
|
|
|
|
size_t addrlen;
|
2007-02-12 15:53:35 +07:00
|
|
|
union svc_addr_u daddr; /* where reply must come from */
|
2005-04-17 05:20:36 +07:00
|
|
|
struct cache_deferred_req handle;
|
2007-12-31 10:08:29 +07:00
|
|
|
size_t xprt_hlen;
|
2005-04-17 05:20:36 +07:00
|
|
|
int argslen;
|
2006-09-27 12:29:38 +07:00
|
|
|
__be32 args[0];
|
2005-04-17 05:20:36 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
2005-06-23 00:16:24 +07:00
|
|
|
* List of RPC programs on the same transport endpoint
|
2005-04-17 05:20:36 +07:00
|
|
|
*/
|
|
|
|
struct svc_program {
|
2005-06-23 00:16:24 +07:00
|
|
|
struct svc_program * pg_next; /* other programs (same xprt) */
|
2005-04-17 05:20:36 +07:00
|
|
|
u32 pg_prog; /* program number */
|
|
|
|
unsigned int pg_lovers; /* lowest version */
|
|
|
|
unsigned int pg_hivers; /* lowest version */
|
|
|
|
unsigned int pg_nvers; /* number of versions */
|
|
|
|
struct svc_version ** pg_vers; /* version array */
|
|
|
|
char * pg_name; /* service name */
|
|
|
|
char * pg_class; /* class name: services sharing authentication */
|
|
|
|
struct svc_stat * pg_stats; /* rpc statistics */
|
|
|
|
int (*pg_authenticate)(struct svc_rqst *);
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* RPC program version
|
|
|
|
*/
|
|
|
|
struct svc_version {
|
|
|
|
u32 vs_vers; /* version number */
|
|
|
|
u32 vs_nproc; /* number of procedures */
|
|
|
|
struct svc_procedure * vs_proc; /* per-procedure info */
|
|
|
|
u32 vs_xdrsize; /* xdrsize needed for this version */
|
|
|
|
|
2006-10-04 16:16:05 +07:00
|
|
|
unsigned int vs_hidden : 1; /* Don't register with portmapper.
|
|
|
|
* Only used for nfsacl so far. */
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
/* Override dispatch function (e.g. when caching replies).
|
|
|
|
* A return value of 0 means drop the request.
|
|
|
|
* vs_dispatch == NULL means use default dispatcher.
|
|
|
|
*/
|
2006-09-27 12:29:38 +07:00
|
|
|
int (*vs_dispatch)(struct svc_rqst *, __be32 *);
|
2005-04-17 05:20:36 +07:00
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* RPC procedure info
|
|
|
|
*/
|
2006-10-20 13:28:45 +07:00
|
|
|
typedef __be32 (*svc_procfunc)(struct svc_rqst *, void *argp, void *resp);
|
2005-04-17 05:20:36 +07:00
|
|
|
struct svc_procedure {
|
|
|
|
svc_procfunc pc_func; /* process the request */
|
|
|
|
kxdrproc_t pc_decode; /* XDR decode args */
|
|
|
|
kxdrproc_t pc_encode; /* XDR encode result */
|
|
|
|
kxdrproc_t pc_release; /* XDR free result */
|
|
|
|
unsigned int pc_argsize; /* argument struct size */
|
|
|
|
unsigned int pc_ressize; /* result struct size */
|
|
|
|
unsigned int pc_count; /* call count */
|
|
|
|
unsigned int pc_cachetype; /* cache info (NFS) */
|
|
|
|
unsigned int pc_xdrressize; /* maximum size of XDR reply */
|
|
|
|
};
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Function prototypes.
|
|
|
|
*/
|
2009-03-19 07:46:29 +07:00
|
|
|
struct svc_serv *svc_create(struct svc_program *, unsigned int,
|
2008-07-01 05:45:30 +07:00
|
|
|
void (*shutdown)(struct svc_serv *));
|
2008-01-29 22:30:54 +07:00
|
|
|
struct svc_rqst *svc_prepare_thread(struct svc_serv *serv,
|
|
|
|
struct svc_pool *pool);
|
2005-04-17 05:20:36 +07:00
|
|
|
void svc_exit_thread(struct svc_rqst *);
|
2006-10-02 16:17:59 +07:00
|
|
|
struct svc_serv * svc_create_pooled(struct svc_program *, unsigned int,
|
2009-03-19 07:46:29 +07:00
|
|
|
void (*shutdown)(struct svc_serv *),
|
2008-07-01 05:45:30 +07:00
|
|
|
svc_thread_fn, struct module *);
|
2006-10-02 16:17:59 +07:00
|
|
|
int svc_set_num_threads(struct svc_serv *, struct svc_pool *, int);
|
2009-01-13 17:26:36 +07:00
|
|
|
int svc_pool_stats_open(struct svc_serv *serv, struct file *file);
|
2005-04-17 05:20:36 +07:00
|
|
|
void svc_destroy(struct svc_serv *);
|
2006-10-02 16:17:50 +07:00
|
|
|
int svc_process(struct svc_rqst *);
|
2009-04-01 20:23:07 +07:00
|
|
|
int bc_svc_process(struct svc_serv *, struct rpc_rqst *,
|
|
|
|
struct svc_rqst *);
|
2009-03-19 07:46:06 +07:00
|
|
|
int svc_register(const struct svc_serv *, const int,
|
|
|
|
const unsigned short, const unsigned short);
|
2008-08-19 06:34:16 +07:00
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
void svc_wake_up(struct svc_serv *);
|
|
|
|
void svc_reserve(struct svc_rqst *rqstp, int space);
|
2006-10-02 16:18:01 +07:00
|
|
|
struct svc_pool * svc_pool_for_cpu(struct svc_serv *serv, int cpu);
|
2007-02-12 15:53:32 +07:00
|
|
|
char * svc_print_addr(struct svc_rqst *, char *, size_t);
|
|
|
|
|
|
|
|
#define RPC_MAX_ADDRBUFLEN (63U)
|
2005-04-17 05:20:36 +07:00
|
|
|
|
2007-05-09 16:34:50 +07:00
|
|
|
/*
|
|
|
|
* When we want to reduce the size of the reserved space in the response
|
|
|
|
* buffer, we need to take into account the size of any checksum data that
|
|
|
|
* may be at the end of the packet. This is difficult to determine exactly
|
|
|
|
* for all cases without actually generating the checksum, so we just use a
|
|
|
|
* static value.
|
|
|
|
*/
|
2008-02-08 11:10:21 +07:00
|
|
|
static inline void svc_reserve_auth(struct svc_rqst *rqstp, int space)
|
2007-05-09 16:34:50 +07:00
|
|
|
{
|
2008-02-08 11:10:21 +07:00
|
|
|
int added_space = 0;
|
2007-05-09 16:34:50 +07:00
|
|
|
|
2008-02-08 11:10:21 +07:00
|
|
|
if (rqstp->rq_authop->flavour)
|
|
|
|
added_space = RPC_MAX_AUTH_SIZE;
|
|
|
|
svc_reserve(rqstp, space + added_space);
|
2007-05-09 16:34:50 +07:00
|
|
|
}
|
|
|
|
|
2005-04-17 05:20:36 +07:00
|
|
|
#endif /* SUNRPC_SVC_H */
|