* High-level RPC service routines
*
* Copyright (C) 1995, 1996 Olaf Kirch <okir@monad.swb.de>
+ *
+ * Multiple threads pools and NUMAisation
+ * Copyright (c) 2006 Silicon Graphics, Inc.
+ * by Greg Banks <gnb@melbourne.sgi.com>
*/
#include <linux/linkage.h>
#include <linux/sunrpc/clnt.h>
#define RPCDBG_FACILITY RPCDBG_SVCDSP
-#define RPC_PARANOIA 1
+
+#define svc_serv_is_pooled(serv) ((serv)->sv_function)
+
+/*
+ * Mode for mapping cpus to pools.
+ */
+enum {
+ SVC_POOL_AUTO = -1, /* choose one of the others */
+ SVC_POOL_GLOBAL, /* no mapping, just a single global pool
+ * (legacy & UP mode) */
+ SVC_POOL_PERCPU, /* one pool per cpu */
+ SVC_POOL_PERNODE /* one pool per numa node */
+};
+#define SVC_POOL_DEFAULT SVC_POOL_GLOBAL
+
+/*
+ * Structure for mapping cpus to pools and vice versa.
+ * Setup once during sunrpc initialisation.
+ */
+static struct svc_pool_map {
+ int count; /* How many svc_servs use us */
+ int mode; /* Note: int not enum to avoid
+ * warnings about "enumeration value
+ * not handled in switch" */
+ unsigned int npools;
+ unsigned int *pool_to; /* maps pool id to cpu or node */
+ unsigned int *to_pool; /* maps cpu or node to pool id */
+} svc_pool_map = {
+ .count = 0,
+ .mode = SVC_POOL_DEFAULT
+};
+static DEFINE_MUTEX(svc_pool_map_mutex);/* protects svc_pool_map.count only */
+
+static int
+param_set_pool_mode(const char *val, struct kernel_param *kp)
+{
+ int *ip = (int *)kp->arg;
+ struct svc_pool_map *m = &svc_pool_map;
+ int err;
+
+ mutex_lock(&svc_pool_map_mutex);
+
+ err = -EBUSY;
+ if (m->count)
+ goto out;
+
+ err = 0;
+ if (!strncmp(val, "auto", 4))
+ *ip = SVC_POOL_AUTO;
+ else if (!strncmp(val, "global", 6))
+ *ip = SVC_POOL_GLOBAL;
+ else if (!strncmp(val, "percpu", 6))
+ *ip = SVC_POOL_PERCPU;
+ else if (!strncmp(val, "pernode", 7))
+ *ip = SVC_POOL_PERNODE;
+ else
+ err = -EINVAL;
+
+out:
+ mutex_unlock(&svc_pool_map_mutex);
+ return err;
+}
+
+static int
+param_get_pool_mode(char *buf, struct kernel_param *kp)
+{
+ int *ip = (int *)kp->arg;
+
+ switch (*ip)
+ {
+ case SVC_POOL_AUTO:
+ return strlcpy(buf, "auto", 20);
+ case SVC_POOL_GLOBAL:
+ return strlcpy(buf, "global", 20);
+ case SVC_POOL_PERCPU:
+ return strlcpy(buf, "percpu", 20);
+ case SVC_POOL_PERNODE:
+ return strlcpy(buf, "pernode", 20);
+ default:
+ return sprintf(buf, "%d", *ip);
+ }
+}
+
+module_param_call(pool_mode, param_set_pool_mode, param_get_pool_mode,
+ &svc_pool_map.mode, 0644);
+
+/*
+ * Detect best pool mapping mode heuristically,
+ * according to the machine's topology.
+ */
+static int
+svc_pool_map_choose_mode(void)
+{
+ unsigned int node;
+
+ if (num_online_nodes() > 1) {
+ /*
+ * Actually have multiple NUMA nodes,
+ * so split pools on NUMA node boundaries
+ */
+ return SVC_POOL_PERNODE;
+ }
+
+ node = any_online_node(node_online_map);
+ if (nr_cpus_node(node) > 2) {
+ /*
+ * Non-trivial SMP, or CONFIG_NUMA on
+ * non-NUMA hardware, e.g. with a generic
+ * x86_64 kernel on Xeons. In this case we
+ * want to divide the pools on cpu boundaries.
+ */
+ return SVC_POOL_PERCPU;
+ }
+
+ /* default: one global pool */
+ return SVC_POOL_GLOBAL;
+}
+
+/*
+ * Allocate the to_pool[] and pool_to[] arrays.
+ * Returns 0 on success or an errno.
+ */
+static int
+svc_pool_map_alloc_arrays(struct svc_pool_map *m, unsigned int maxpools)
+{
+ m->to_pool = kcalloc(maxpools, sizeof(unsigned int), GFP_KERNEL);
+ if (!m->to_pool)
+ goto fail;
+ m->pool_to = kcalloc(maxpools, sizeof(unsigned int), GFP_KERNEL);
+ if (!m->pool_to)
+ goto fail_free;
+
+ return 0;
+
+fail_free:
+ kfree(m->to_pool);
+fail:
+ return -ENOMEM;
+}
+
+/*
+ * Initialise the pool map for SVC_POOL_PERCPU mode.
+ * Returns number of pools or <0 on error.
+ */
+static int
+svc_pool_map_init_percpu(struct svc_pool_map *m)
+{
+ unsigned int maxpools = nr_cpu_ids;
+ unsigned int pidx = 0;
+ unsigned int cpu;
+ int err;
+
+ err = svc_pool_map_alloc_arrays(m, maxpools);
+ if (err)
+ return err;
+
+ for_each_online_cpu(cpu) {
+ BUG_ON(pidx > maxpools);
+ m->to_pool[cpu] = pidx;
+ m->pool_to[pidx] = cpu;
+ pidx++;
+ }
+ /* cpus brought online later all get mapped to pool0, sorry */
+
+ return pidx;
+};
+
+
+/*
+ * Initialise the pool map for SVC_POOL_PERNODE mode.
+ * Returns number of pools or <0 on error.
+ */
+static int
+svc_pool_map_init_pernode(struct svc_pool_map *m)
+{
+ unsigned int maxpools = nr_node_ids;
+ unsigned int pidx = 0;
+ unsigned int node;
+ int err;
+
+ err = svc_pool_map_alloc_arrays(m, maxpools);
+ if (err)
+ return err;
+
+ for_each_node_with_cpus(node) {
+ /* some architectures (e.g. SN2) have cpuless nodes */
+ BUG_ON(pidx > maxpools);
+ m->to_pool[node] = pidx;
+ m->pool_to[pidx] = node;
+ pidx++;
+ }
+ /* nodes brought online later all get mapped to pool0, sorry */
+
+ return pidx;
+}
+
+
+/*
+ * Add a reference to the global map of cpus to pools (and
+ * vice versa). Initialise the map if we're the first user.
+ * Returns the number of pools.
+ */
+static unsigned int
+svc_pool_map_get(void)
+{
+ struct svc_pool_map *m = &svc_pool_map;
+ int npools = -1;
+
+ mutex_lock(&svc_pool_map_mutex);
+
+ if (m->count++) {
+ mutex_unlock(&svc_pool_map_mutex);
+ return m->npools;
+ }
+
+ if (m->mode == SVC_POOL_AUTO)
+ m->mode = svc_pool_map_choose_mode();
+
+ switch (m->mode) {
+ case SVC_POOL_PERCPU:
+ npools = svc_pool_map_init_percpu(m);
+ break;
+ case SVC_POOL_PERNODE:
+ npools = svc_pool_map_init_pernode(m);
+ break;
+ }
+
+ if (npools < 0) {
+ /* default, or memory allocation failure */
+ npools = 1;
+ m->mode = SVC_POOL_GLOBAL;
+ }
+ m->npools = npools;
+
+ mutex_unlock(&svc_pool_map_mutex);
+ return m->npools;
+}
+
+
+/*
+ * Drop a reference to the global map of cpus to pools.
+ * When the last reference is dropped, the map data is
+ * freed; this allows the sysadmin to change the pool
+ * mode using the pool_mode module option without
+ * rebooting or re-loading sunrpc.ko.
+ */
+static void
+svc_pool_map_put(void)
+{
+ struct svc_pool_map *m = &svc_pool_map;
+
+ mutex_lock(&svc_pool_map_mutex);
+
+ if (!--m->count) {
+ m->mode = SVC_POOL_DEFAULT;
+ kfree(m->to_pool);
+ kfree(m->pool_to);
+ m->npools = 0;
+ }
+
+ mutex_unlock(&svc_pool_map_mutex);
+}
+
+
+/*
+ * Set the current thread's cpus_allowed mask so that it
+ * will only run on cpus in the given pool.
+ *
+ * Returns 1 and fills in oldmask iff a cpumask was applied.
+ */
+static inline int
+svc_pool_map_set_cpumask(unsigned int pidx, cpumask_t *oldmask)
+{
+ struct svc_pool_map *m = &svc_pool_map;
+ unsigned int node; /* or cpu */
+
+ /*
+ * The caller checks for sv_nrpools > 1, which
+ * implies that we've been initialized.
+ */
+ BUG_ON(m->count == 0);
+
+ switch (m->mode)
+ {
+ default:
+ return 0;
+ case SVC_POOL_PERCPU:
+ node = m->pool_to[pidx];
+ *oldmask = current->cpus_allowed;
+ set_cpus_allowed(current, cpumask_of_cpu(node));
+ return 1;
+ case SVC_POOL_PERNODE:
+ node = m->pool_to[pidx];
+ *oldmask = current->cpus_allowed;
+ set_cpus_allowed(current, node_to_cpumask(node));
+ return 1;
+ }
+}
+
+/*
+ * Use the mapping mode to choose a pool for a given CPU.
+ * Used when enqueueing an incoming RPC. Always returns
+ * a non-NULL pool pointer.
+ */
+struct svc_pool *
+svc_pool_for_cpu(struct svc_serv *serv, int cpu)
+{
+ struct svc_pool_map *m = &svc_pool_map;
+ unsigned int pidx = 0;
+
+ /*
+ * An uninitialised map happens in a pure client when
+ * lockd is brought up, so silently treat it the
+ * same as SVC_POOL_GLOBAL.
+ */
+ if (svc_serv_is_pooled(serv)) {
+ switch (m->mode) {
+ case SVC_POOL_PERCPU:
+ pidx = m->to_pool[cpu];
+ break;
+ case SVC_POOL_PERNODE:
+ pidx = m->to_pool[cpu_to_node(cpu)];
+ break;
+ }
+ }
+ return &serv->sv_pools[pidx % serv->sv_nrpools];
+}
+
/*
* Create an RPC service
serv->sv_program = prog;
serv->sv_nrthreads = 1;
serv->sv_stats = prog->pg_stats;
- serv->sv_bufsz = bufsize? bufsize : 4096;
+ if (bufsize > RPCSVC_MAXPAYLOAD)
+ bufsize = RPCSVC_MAXPAYLOAD;
+ serv->sv_max_payload = bufsize? bufsize : 4096;
+ serv->sv_max_mesg = roundup(serv->sv_max_payload + PAGE_SIZE, PAGE_SIZE);
serv->sv_shutdown = shutdown;
xdrsize = 0;
while (prog) {
serv->sv_nrpools = npools;
serv->sv_pools =
- kcalloc(sizeof(struct svc_pool), serv->sv_nrpools,
+ kcalloc(serv->sv_nrpools, sizeof(struct svc_pool),
GFP_KERNEL);
if (!serv->sv_pools) {
kfree(serv);
for (i = 0; i < serv->sv_nrpools; i++) {
struct svc_pool *pool = &serv->sv_pools[i];
- dprintk("initialising pool %u for %s\n",
+ dprintk("svc: initialising pool %u for %s\n",
i, serv->sv_name);
pool->sp_id = i;
svc_thread_fn func, int sig, struct module *mod)
{
struct svc_serv *serv;
+ unsigned int npools = svc_pool_map_get();
- serv = __svc_create(prog, bufsize, /*npools*/1, shutdown);
+ serv = __svc_create(prog, bufsize, npools, shutdown);
if (serv != NULL) {
serv->sv_function = func;
svc_destroy(struct svc_serv *serv)
{
struct svc_sock *svsk;
+ struct svc_sock *tmp;
- dprintk("RPC: svc_destroy(%s, %d)\n",
+ dprintk("svc: svc_destroy(%s, %d)\n",
serv->sv_program->pg_name,
serv->sv_nrthreads);
del_timer_sync(&serv->sv_temptimer);
- while (!list_empty(&serv->sv_tempsocks)) {
- svsk = list_entry(serv->sv_tempsocks.next,
- struct svc_sock,
- sk_list);
- svc_delete_socket(svsk);
- }
+ list_for_each_entry_safe(svsk, tmp, &serv->sv_tempsocks, sk_list)
+ svc_force_close_socket(svsk);
+
if (serv->sv_shutdown)
serv->sv_shutdown(serv);
- while (!list_empty(&serv->sv_permsocks)) {
- svsk = list_entry(serv->sv_permsocks.next,
- struct svc_sock,
- sk_list);
- svc_delete_socket(svsk);
- }
-
+ list_for_each_entry_safe(svsk, tmp, &serv->sv_permsocks, sk_list)
+ svc_force_close_socket(svsk);
+
+ BUG_ON(!list_empty(&serv->sv_permsocks));
+ BUG_ON(!list_empty(&serv->sv_tempsocks));
+
cache_clean_deferred(serv);
+ if (svc_serv_is_pooled(serv))
+ svc_pool_map_put();
+
/* Unregister service with the portmapper */
svc_register(serv, 0, 0);
kfree(serv->sv_pools);
{
int pages;
int arghi;
-
- if (size > RPCSVC_MAXPAYLOAD)
- size = RPCSVC_MAXPAYLOAD;
- pages = 2 + (size+ PAGE_SIZE -1) / PAGE_SIZE;
- rqstp->rq_argused = 0;
- rqstp->rq_resused = 0;
+
+ pages = size / PAGE_SIZE + 1; /* extra page as we hold both request and reply.
+ * We assume one is at most one page
+ */
arghi = 0;
BUG_ON(pages > RPCSVC_MAXPAGES);
while (pages) {
struct page *p = alloc_page(GFP_KERNEL);
if (!p)
break;
- rqstp->rq_argpages[arghi++] = p;
+ rqstp->rq_pages[arghi++] = p;
pages--;
}
- rqstp->rq_arghi = arghi;
return ! pages;
}
static void
svc_release_buffer(struct svc_rqst *rqstp)
{
- while (rqstp->rq_arghi)
- put_page(rqstp->rq_argpages[--rqstp->rq_arghi]);
- while (rqstp->rq_resused) {
- if (rqstp->rq_respages[--rqstp->rq_resused] == NULL)
- continue;
- put_page(rqstp->rq_respages[rqstp->rq_resused]);
- }
- rqstp->rq_argused = 0;
+ int i;
+ for (i=0; i<ARRAY_SIZE(rqstp->rq_pages); i++)
+ if (rqstp->rq_pages[i])
+ put_page(rqstp->rq_pages[i]);
}
/*
* Create a thread in the given pool. Caller must hold BKL.
+ * On a NUMA or SMP machine, with a multi-pool serv, the thread
+ * will be restricted to run on the cpus belonging to the pool.
*/
static int
__svc_create_thread(svc_thread_fn func, struct svc_serv *serv,
{
struct svc_rqst *rqstp;
int error = -ENOMEM;
+ int have_oldmask = 0;
+ cpumask_t oldmask;
rqstp = kzalloc(sizeof(*rqstp), GFP_KERNEL);
if (!rqstp)
if (!(rqstp->rq_argp = kmalloc(serv->sv_xdrsize, GFP_KERNEL))
|| !(rqstp->rq_resp = kmalloc(serv->sv_xdrsize, GFP_KERNEL))
- || !svc_init_buffer(rqstp, serv->sv_bufsz))
+ || !svc_init_buffer(rqstp, serv->sv_max_mesg))
goto out_thread;
serv->sv_nrthreads++;
spin_unlock_bh(&pool->sp_lock);
rqstp->rq_server = serv;
rqstp->rq_pool = pool;
+
+ if (serv->sv_nrpools > 1)
+ have_oldmask = svc_pool_map_set_cpumask(pool->sp_id, &oldmask);
+
error = kernel_thread((int (*)(void *)) func, rqstp, 0);
+
+ if (have_oldmask)
+ set_cpus_allowed(current, oldmask);
+
if (error < 0)
goto out_thread;
svc_sock_update_bufs(serv);
if (pool != NULL)
return pool;
- return &serv->sv_pools[(*state)++ % serv->sv_nrpools];
+ return &serv->sv_pools[(*state)++ % serv->sv_nrpools];
}
/*
spin_lock_bh(&pool->sp_lock);
} else {
/* choose a pool in round-robin fashion */
- for (i = 0; i < serv->sv_nrpools; i++) {
- pool = &serv->sv_pools[--(*state) % serv->sv_nrpools];
+ for (i = 0; i < serv->sv_nrpools; i++) {
+ pool = &serv->sv_pools[--(*state) % serv->sv_nrpools];
spin_lock_bh(&pool->sp_lock);
- if (!list_empty(&pool->sp_all_threads))
- goto found_pool;
+ if (!list_empty(&pool->sp_all_threads))
+ goto found_pool;
spin_unlock_bh(&pool->sp_lock);
- }
+ }
return NULL;
}
rqstp = list_entry(pool->sp_all_threads.next, struct svc_rqst, rq_all);
list_del_init(&rqstp->rq_all);
task = rqstp->rq_task;
- }
+ }
spin_unlock_bh(&pool->sp_lock);
return task;
/*
* Register an RPC service with the local portmapper.
- * To unregister a service, call this routine with
+ * To unregister a service, call this routine with
* proto and port == 0.
*/
int
unsigned long flags;
int i, error = 0, dummy;
- progp = serv->sv_program;
-
- dprintk("RPC: svc_register(%s, %s, %d)\n",
- progp->pg_name, proto == IPPROTO_UDP? "udp" : "tcp", port);
-
if (!port)
clear_thread_flag(TIF_SIGPENDING);
- for (i = 0; i < progp->pg_nvers; i++) {
- if (progp->pg_vers[i] == NULL)
- continue;
- error = rpc_register(progp->pg_prog, i, proto, port, &dummy);
- if (error < 0)
- break;
- if (port && !dummy) {
- error = -EACCES;
- break;
+ for (progp = serv->sv_program; progp; progp = progp->pg_next) {
+ for (i = 0; i < progp->pg_nvers; i++) {
+ if (progp->pg_vers[i] == NULL)
+ continue;
+
+ dprintk("svc: svc_register(%s, %s, %d, %d)%s\n",
+ progp->pg_name,
+ proto == IPPROTO_UDP? "udp" : "tcp",
+ port,
+ i,
+ progp->pg_vers[i]->vs_hidden?
+ " (but not telling portmap)" : "");
+
+ if (progp->pg_vers[i]->vs_hidden)
+ continue;
+
+ error = rpcb_register(progp->pg_prog, i, proto, port, &dummy);
+ if (error < 0)
+ break;
+ if (port && !dummy) {
+ error = -EACCES;
+ break;
+ }
}
}
u32 dir, prog, vers, proc;
__be32 auth_stat, rpc_stat;
int auth_res;
- __be32 *accept_statp;
+ __be32 *reply_statp;
rpc_stat = rpc_success;
goto err_short_len;
/* setup response xdr_buf.
- * Initially it has just one page
+ * Initially it has just one page
*/
- svc_take_page(rqstp); /* must succeed */
+ rqstp->rq_resused = 1;
resv->iov_base = page_address(rqstp->rq_respages[0]);
resv->iov_len = 0;
- rqstp->rq_res.pages = rqstp->rq_respages+1;
+ rqstp->rq_res.pages = rqstp->rq_respages + 1;
rqstp->rq_res.len = 0;
rqstp->rq_res.page_base = 0;
rqstp->rq_res.page_len = 0;
rqstp->rq_res.tail[0].iov_base = NULL;
rqstp->rq_res.tail[0].iov_len = 0;
/* Will be turned off only in gss privacy case: */
- rqstp->rq_sendfile_ok = 1;
+ rqstp->rq_splice_ok = 1;
/* tcp needs a space for the record length... */
if (rqstp->rq_prot == IPPROTO_TCP)
svc_putnl(resv, 0);
goto err_bad_rpc;
/* Save position in case we later decide to reject: */
- accept_statp = resv->iov_base + resv->iov_len;
+ reply_statp = resv->iov_base + resv->iov_len;
svc_putnl(resv, 0); /* ACCEPT */
memset(rqstp->rq_argp, 0, procp->pc_argsize);
memset(rqstp->rq_resp, 0, procp->pc_ressize);
- /* un-reserve some of the out-queue now that we have a
+ /* un-reserve some of the out-queue now that we have a
* better idea of reply size
*/
if (procp->pc_xdrressize)
- svc_reserve(rqstp, procp->pc_xdrressize<<2);
+ svc_reserve_auth(rqstp, procp->pc_xdrressize<<2);
/* Call the function that processes the request. */
if (!versp->vs_dispatch) {
*statp = procp->pc_func(rqstp, rqstp->rq_argp, rqstp->rq_resp);
/* Encode reply */
+ if (*statp == rpc_drop_reply) {
+ if (procp->pc_release)
+ procp->pc_release(rqstp, NULL, rqstp->rq_resp);
+ goto dropit;
+ }
if (*statp == rpc_success && (xdr = procp->pc_encode)
&& !xdr(rqstp, resv->iov_base+resv->iov_len, rqstp->rq_resp)) {
dprintk("svc: failed to encode reply\n");
return 0;
err_short_len:
-#ifdef RPC_PARANOIA
- printk("svc: short len %Zd, dropping request\n", argv->iov_len);
-#endif
+ if (net_ratelimit())
+ printk("svc: short len %Zd, dropping request\n", argv->iov_len);
+
goto dropit; /* drop request */
err_bad_dir:
-#ifdef RPC_PARANOIA
- printk("svc: bad direction %d, dropping request\n", dir);
-#endif
+ if (net_ratelimit())
+ printk("svc: bad direction %d, dropping request\n", dir);
+
serv->sv_stats->rpcbadfmt++;
goto dropit; /* drop request */
dprintk("svc: authentication failed (%d)\n", ntohl(auth_stat));
serv->sv_stats->rpcbadauth++;
/* Restore write pointer to location of accept status: */
- xdr_ressize_check(rqstp, accept_statp);
+ xdr_ressize_check(rqstp, reply_statp);
svc_putnl(resv, 1); /* REJECT */
svc_putnl(resv, 1); /* AUTH_ERROR */
svc_putnl(resv, ntohl(auth_stat)); /* status */
goto sendit;
err_bad_vers:
-#ifdef RPC_PARANOIA
- printk("svc: unknown version (%d)\n", vers);
-#endif
+ if (net_ratelimit())
+ printk("svc: unknown version (%d for prog %d, %s)\n",
+ vers, prog, progp->pg_name);
+
serv->sv_stats->rpcbadfmt++;
svc_putnl(resv, RPC_PROG_MISMATCH);
svc_putnl(resv, progp->pg_lovers);
goto sendit;
err_bad_proc:
-#ifdef RPC_PARANOIA
- printk("svc: unknown procedure (%d)\n", proc);
-#endif
+ if (net_ratelimit())
+ printk("svc: unknown procedure (%d)\n", proc);
+
serv->sv_stats->rpcbadfmt++;
svc_putnl(resv, RPC_PROC_UNAVAIL);
goto sendit;
err_garbage:
-#ifdef RPC_PARANOIA
- printk("svc: failed to decode args\n");
-#endif
+ if (net_ratelimit())
+ printk("svc: failed to decode args\n");
+
rpc_stat = rpc_garbage_args;
err_bad:
serv->sv_stats->rpcbadfmt++;
svc_putnl(resv, ntohl(rpc_stat));
goto sendit;
}
+
+/*
+ * Return (transport-specific) limit on the rpc payload.
+ */
+u32 svc_max_payload(const struct svc_rqst *rqstp)
+{
+ int max = RPCSVC_MAXPAYLOAD_TCP;
+
+ if (rqstp->rq_sock->sk_sock->type == SOCK_DGRAM)
+ max = RPCSVC_MAXPAYLOAD_UDP;
+ if (rqstp->rq_server->sv_max_payload < max)
+ max = rqstp->rq_server->sv_max_payload;
+ return max;
+}
+EXPORT_SYMBOL_GPL(svc_max_payload);
git://ftp.safe.ca / safe / jmp / linux-2.6 / blobdiff