From 17cebf658e088935d4bdebfc7ad9800e9fc4a0b2 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Thu, 12 Aug 2010 16:55:22 +1000 Subject: sunrpc: extract some common sunrpc_cache code from nfsd Rather can duplicating this idiom twice, put it in an inline function. This reduces the usage of 'expiry_time' out side the sunrpc/cache.c code and thus the impact of a change that is about to be made to that field. Signed-off-by: NeilBrown Signed-off-by: J. Bruce Fields --- include/linux/sunrpc/cache.h | 6 ++++++ 1 file changed, 6 insertions(+) (limited to 'include/linux/sunrpc/cache.h') diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index 7bf3e84b92f4..0e1febf4e5bc 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -228,4 +228,10 @@ static inline time_t get_expiry(char **bpp) return rv; } +static inline void sunrpc_invalidate(struct cache_head *h, + struct cache_detail *detail) +{ + h->expiry_time = get_seconds() - 1; + detail->nextcheck = get_seconds(); +} #endif /* _LINUX_SUNRPC_CACHE_H_ */ -- cgit v1.2.3 From c5b29f885afe890f953f7f23424045cdad31d3e4 Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Thu, 12 Aug 2010 16:55:22 +1000 Subject: sunrpc: use seconds since boot in expiry cache This protects us from confusion when the wallclock time changes. We convert to and from wallclock when setting or reading expiry times. Also use seconds since boot for last_clost time. Signed-off-by: NeilBrown Signed-off-by: J. Bruce Fields --- fs/nfs/dns_resolve.c | 6 +++--- fs/nfsd/nfs4idmap.c | 2 +- include/linux/sunrpc/cache.h | 28 +++++++++++++++++++++++++--- net/sunrpc/cache.c | 36 +++++++++++++++++++----------------- 4 files changed, 48 insertions(+), 24 deletions(-) (limited to 'include/linux/sunrpc/cache.h') diff --git a/fs/nfs/dns_resolve.c b/fs/nfs/dns_resolve.c index dba50a5625db..a6e711ad130f 100644 --- a/fs/nfs/dns_resolve.c +++ b/fs/nfs/dns_resolve.c @@ -167,7 +167,7 @@ static int nfs_dns_show(struct seq_file *m, struct cache_detail *cd, return 0; } item = container_of(h, struct nfs_dns_ent, h); - ttl = (long)item->h.expiry_time - (long)get_seconds(); + ttl = item->h.expiry_time - seconds_since_boot(); if (ttl < 0) ttl = 0; @@ -239,7 +239,7 @@ static int nfs_dns_parse(struct cache_detail *cd, char *buf, int buflen) ttl = get_expiry(&buf); if (ttl == 0) goto out; - key.h.expiry_time = ttl + get_seconds(); + key.h.expiry_time = ttl + seconds_since_boot(); ret = -ENOMEM; item = nfs_dns_lookup(cd, &key); @@ -301,7 +301,7 @@ static int do_cache_lookup_nowait(struct cache_detail *cd, goto out_err; ret = -ETIMEDOUT; if (!test_bit(CACHE_VALID, &(*item)->h.flags) - || (*item)->h.expiry_time < get_seconds() + || (*item)->h.expiry_time < seconds_since_boot() || cd->flush_time > (*item)->h.last_refresh) goto out_put; ret = -ENOENT; diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c index c78dbf493424..808b33a4a090 100644 --- a/fs/nfsd/nfs4idmap.c +++ b/fs/nfsd/nfs4idmap.c @@ -550,7 +550,7 @@ do_idmap_lookup_nowait(struct ent *(*lookup_fn)(struct ent *), goto out_err; ret = -ETIMEDOUT; if (!test_bit(CACHE_VALID, &(*item)->h.flags) - || (*item)->h.expiry_time < get_seconds() + || (*item)->h.expiry_time < seconds_since_boot() || detail->flush_time > (*item)->h.last_refresh) goto out_put; ret = -ENOENT; diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index 0e1febf4e5bc..ece432b7f87f 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -218,20 +218,42 @@ static inline int get_int(char **bpp, int *anint) return 0; } +/* + * timestamps kept in the cache are expressed in seconds + * since boot. This is the best for measuring differences in + * real time. + */ +static inline time_t seconds_since_boot(void) +{ + struct timespec boot; + getboottime(&boot); + return get_seconds() - boot.tv_sec; +} + +static inline time_t convert_to_wallclock(time_t sinceboot) +{ + struct timespec boot; + getboottime(&boot); + return boot.tv_sec + sinceboot; +} + static inline time_t get_expiry(char **bpp) { int rv; + struct timespec boot; + if (get_int(bpp, &rv)) return 0; if (rv < 0) return 0; - return rv; + getboottime(&boot); + return rv - boot.tv_sec; } static inline void sunrpc_invalidate(struct cache_head *h, struct cache_detail *detail) { - h->expiry_time = get_seconds() - 1; - detail->nextcheck = get_seconds(); + h->expiry_time = seconds_since_boot() - 1; + detail->nextcheck = seconds_since_boot(); } #endif /* _LINUX_SUNRPC_CACHE_H_ */ diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index 2b06410e584e..8dc121955fdc 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -42,7 +42,7 @@ static void cache_revisit_request(struct cache_head *item); static void cache_init(struct cache_head *h) { - time_t now = get_seconds(); + time_t now = seconds_since_boot(); h->next = NULL; h->flags = 0; kref_init(&h->ref); @@ -52,7 +52,7 @@ static void cache_init(struct cache_head *h) static inline int cache_is_expired(struct cache_detail *detail, struct cache_head *h) { - return (h->expiry_time < get_seconds()) || + return (h->expiry_time < seconds_since_boot()) || (detail->flush_time > h->last_refresh); } @@ -127,7 +127,7 @@ static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch); static void cache_fresh_locked(struct cache_head *head, time_t expiry) { head->expiry_time = expiry; - head->last_refresh = get_seconds(); + head->last_refresh = seconds_since_boot(); set_bit(CACHE_VALID, &head->flags); } @@ -238,7 +238,7 @@ int cache_check(struct cache_detail *detail, /* now see if we want to start an upcall */ refresh_age = (h->expiry_time - h->last_refresh); - age = get_seconds() - h->last_refresh; + age = seconds_since_boot() - h->last_refresh; if (rqstp == NULL) { if (rv == -EAGAIN) @@ -253,7 +253,7 @@ int cache_check(struct cache_detail *detail, cache_revisit_request(h); if (rv == -EAGAIN) { set_bit(CACHE_NEGATIVE, &h->flags); - cache_fresh_locked(h, get_seconds()+CACHE_NEW_EXPIRY); + cache_fresh_locked(h, seconds_since_boot()+CACHE_NEW_EXPIRY); cache_fresh_unlocked(h, detail); rv = -ENOENT; } @@ -388,11 +388,11 @@ static int cache_clean(void) return -1; } current_detail = list_entry(next, struct cache_detail, others); - if (current_detail->nextcheck > get_seconds()) + if (current_detail->nextcheck > seconds_since_boot()) current_index = current_detail->hash_size; else { current_index = 0; - current_detail->nextcheck = get_seconds()+30*60; + current_detail->nextcheck = seconds_since_boot()+30*60; } } @@ -477,7 +477,7 @@ EXPORT_SYMBOL_GPL(cache_flush); void cache_purge(struct cache_detail *detail) { detail->flush_time = LONG_MAX; - detail->nextcheck = get_seconds(); + detail->nextcheck = seconds_since_boot(); cache_flush(); detail->flush_time = 1; } @@ -902,7 +902,7 @@ static int cache_release(struct inode *inode, struct file *filp, filp->private_data = NULL; kfree(rp); - cd->last_close = get_seconds(); + cd->last_close = seconds_since_boot(); atomic_dec(&cd->readers); } module_put(cd->owner); @@ -1034,7 +1034,7 @@ int sunrpc_cache_pipe_upcall(struct cache_detail *detail, struct cache_head *h, int len; if (atomic_read(&detail->readers) == 0 && - detail->last_close < get_seconds() - 30) { + detail->last_close < seconds_since_boot() - 30) { warn_no_listener(detail); return -EINVAL; } @@ -1219,7 +1219,8 @@ static int c_show(struct seq_file *m, void *p) ifdebug(CACHE) seq_printf(m, "# expiry=%ld refcnt=%d flags=%lx\n", - cp->expiry_time, atomic_read(&cp->ref.refcount), cp->flags); + convert_to_wallclock(cp->expiry_time), + atomic_read(&cp->ref.refcount), cp->flags); cache_get(cp); if (cache_check(cd, cp, NULL)) /* cache_check does a cache_put on failure */ @@ -1285,7 +1286,7 @@ static ssize_t read_flush(struct file *file, char __user *buf, unsigned long p = *ppos; size_t len; - sprintf(tbuf, "%lu\n", cd->flush_time); + sprintf(tbuf, "%lu\n", convert_to_wallclock(cd->flush_time)); len = strlen(tbuf); if (p >= len) return 0; @@ -1303,19 +1304,20 @@ static ssize_t write_flush(struct file *file, const char __user *buf, struct cache_detail *cd) { char tbuf[20]; - char *ep; - long flushtime; + char *bp, *ep; + if (*ppos || count > sizeof(tbuf)-1) return -EINVAL; if (copy_from_user(tbuf, buf, count)) return -EFAULT; tbuf[count] = 0; - flushtime = simple_strtoul(tbuf, &ep, 0); + simple_strtoul(tbuf, &ep, 0); if (*ep && *ep != '\n') return -EINVAL; - cd->flush_time = flushtime; - cd->nextcheck = get_seconds(); + bp = tbuf; + cd->flush_time = get_expiry(&bp); + cd->nextcheck = seconds_since_boot(); cache_flush(); *ppos += count; -- cgit v1.2.3 From f16b6e8d838b2e2bb4561201311c66ac02ad67df Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Thu, 12 Aug 2010 17:04:06 +1000 Subject: sunrpc/cache: allow threads to block while waiting for cache update. The current practice of waiting for cache updates by queueing the whole request to be retried has (at least) two problems. 1/ With NFSv4, requests can be quite complex and re-trying a whole request when a later part fails should only be a last-resort, not a normal practice. 2/ Large requests, and in particular any 'write' request, will not be queued by the current code and doing so would be undesirable. In many cases only a very sort wait is needed before the cache gets valid data. So, providing the underlying transport permits it by setting ->thread_wait, arrange to wait briefly for an upcall to be completed (as reflected in the clearing of CACHE_PENDING). If the short wait was not long enough and CACHE_PENDING is still set, fall back on the old approach. The 'thread_wait' value is set to 5 seconds when there are spare threads, and 1 second when there are no spare threads. These values are probably much higher than needed, but will ensure some forward progress. Note that as we only request an update for a non-valid item, and as non-valid items are updated in place it is extremely unlikely that cache_check will return -ETIMEDOUT. Normally cache_defer_req will sleep for a short while and then find that the item is_valid. Signed-off-by: NeilBrown Signed-off-by: J. Bruce Fields --- include/linux/sunrpc/cache.h | 3 +++ net/sunrpc/cache.c | 59 +++++++++++++++++++++++++++++++++++++++++++- net/sunrpc/svc_xprt.c | 11 +++++++++ 3 files changed, 72 insertions(+), 1 deletion(-) (limited to 'include/linux/sunrpc/cache.h') diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index ece432b7f87f..52a7d7224e90 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -125,6 +125,9 @@ struct cache_detail { */ struct cache_req { struct cache_deferred_req *(*defer)(struct cache_req *req); + int thread_wait; /* How long (jiffies) we can block the + * current thread to wait for updates. + */ }; /* this must be embedded in a deferred_request that is being * delayed awaiting cache-fill diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index 8dc121955fdc..2c5297f245b4 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -509,10 +509,22 @@ static LIST_HEAD(cache_defer_list); static struct list_head cache_defer_hash[DFR_HASHSIZE]; static int cache_defer_cnt; +struct thread_deferred_req { + struct cache_deferred_req handle; + struct completion completion; +}; +static void cache_restart_thread(struct cache_deferred_req *dreq, int too_many) +{ + struct thread_deferred_req *dr = + container_of(dreq, struct thread_deferred_req, handle); + complete(&dr->completion); +} + static int cache_defer_req(struct cache_req *req, struct cache_head *item) { struct cache_deferred_req *dreq, *discard; int hash = DFR_HASH(item); + struct thread_deferred_req sleeper; if (cache_defer_cnt >= DFR_MAX) { /* too much in the cache, randomly drop this one, @@ -521,7 +533,15 @@ static int cache_defer_req(struct cache_req *req, struct cache_head *item) if (net_random()&1) return -ENOMEM; } - dreq = req->defer(req); + if (req->thread_wait) { + dreq = &sleeper.handle; + sleeper.completion = + COMPLETION_INITIALIZER_ONSTACK(sleeper.completion); + dreq->revisit = cache_restart_thread; + } else + dreq = req->defer(req); + + retry: if (dreq == NULL) return -ENOMEM; @@ -555,6 +575,43 @@ static int cache_defer_req(struct cache_req *req, struct cache_head *item) cache_revisit_request(item); return -EAGAIN; } + + if (dreq == &sleeper.handle) { + if (wait_for_completion_interruptible_timeout( + &sleeper.completion, req->thread_wait) <= 0) { + /* The completion wasn't completed, so we need + * to clean up + */ + spin_lock(&cache_defer_lock); + if (!list_empty(&sleeper.handle.hash)) { + list_del_init(&sleeper.handle.recent); + list_del_init(&sleeper.handle.hash); + cache_defer_cnt--; + spin_unlock(&cache_defer_lock); + } else { + /* cache_revisit_request already removed + * this from the hash table, but hasn't + * called ->revisit yet. It will very soon + * and we need to wait for it. + */ + spin_unlock(&cache_defer_lock); + wait_for_completion(&sleeper.completion); + } + } + if (test_bit(CACHE_PENDING, &item->flags)) { + /* item is still pending, try request + * deferral + */ + dreq = req->defer(req); + goto retry; + } + /* only return success if we actually deferred the + * request. In this case we waited until it was + * answered so no deferral has happened - rather + * an answer already exists. + */ + return -EEXIST; + } return 0; } diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c index cbc084939dd8..8ff6840866fa 100644 --- a/net/sunrpc/svc_xprt.c +++ b/net/sunrpc/svc_xprt.c @@ -651,6 +651,11 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) if (signalled() || kthread_should_stop()) return -EINTR; + /* Normally we will wait up to 5 seconds for any required + * cache information to be provided. + */ + rqstp->rq_chandle.thread_wait = 5*HZ; + spin_lock_bh(&pool->sp_lock); xprt = svc_xprt_dequeue(pool); if (xprt) { @@ -658,6 +663,12 @@ int svc_recv(struct svc_rqst *rqstp, long timeout) svc_xprt_get(xprt); rqstp->rq_reserved = serv->sv_max_mesg; atomic_add(rqstp->rq_reserved, &xprt->xpt_reserved); + + /* As there is a shortage of threads and this request + * had to be queue, don't allow the thread to wait so + * long for cache updates. + */ + rqstp->rq_chandle.thread_wait = 1*HZ; } else { /* No data pending. Go to sleep */ svc_thread_enqueue(pool, rqstp); -- cgit v1.2.3 From 1117449276bb909b029ed0b9ba13f53e4784db9d Mon Sep 17 00:00:00 2001 From: NeilBrown Date: Thu, 12 Aug 2010 17:04:08 +1000 Subject: sunrpc/cache: change deferred-request hash table to use hlist. Being a hash table, hlist is the best option. There is currently some ugliness were we treat "->next == NULL" as a special case to avoid having to initialise the whole array. This change nicely gets rid of that case. Signed-off-by: NeilBrown Signed-off-by: J. Bruce Fields --- include/linux/sunrpc/cache.h | 2 +- net/sunrpc/cache.c | 28 ++++++++++------------------ 2 files changed, 11 insertions(+), 19 deletions(-) (limited to 'include/linux/sunrpc/cache.h') diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index 52a7d7224e90..03496357f455 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -133,7 +133,7 @@ struct cache_req { * delayed awaiting cache-fill */ struct cache_deferred_req { - struct list_head hash; /* on hash chain */ + struct hlist_node hash; /* on hash chain */ struct list_head recent; /* on fifo */ struct cache_head *item; /* cache item we wait on */ void *owner; /* we might need to discard all defered requests diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index ca7c621cd975..2a8405194056 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -506,13 +506,13 @@ EXPORT_SYMBOL_GPL(cache_purge); static DEFINE_SPINLOCK(cache_defer_lock); static LIST_HEAD(cache_defer_list); -static struct list_head cache_defer_hash[DFR_HASHSIZE]; +static struct hlist_head cache_defer_hash[DFR_HASHSIZE]; static int cache_defer_cnt; static void __unhash_deferred_req(struct cache_deferred_req *dreq) { list_del_init(&dreq->recent); - list_del_init(&dreq->hash); + hlist_del_init(&dreq->hash); cache_defer_cnt--; } @@ -521,9 +521,7 @@ static void __hash_deferred_req(struct cache_deferred_req *dreq, struct cache_he int hash = DFR_HASH(item); list_add(&dreq->recent, &cache_defer_list); - if (cache_defer_hash[hash].next == NULL) - INIT_LIST_HEAD(&cache_defer_hash[hash]); - list_add(&dreq->hash, &cache_defer_hash[hash]); + hlist_add_head(&dreq->hash, &cache_defer_hash[hash]); } static int setup_deferral(struct cache_deferred_req *dreq, struct cache_head *item) @@ -588,7 +586,7 @@ static int cache_wait_req(struct cache_req *req, struct cache_head *item) * to clean up */ spin_lock(&cache_defer_lock); - if (!list_empty(&sleeper.handle.hash)) { + if (!hlist_unhashed(&sleeper.handle.hash)) { __unhash_deferred_req(&sleeper.handle); spin_unlock(&cache_defer_lock); } else { @@ -642,24 +640,18 @@ static void cache_revisit_request(struct cache_head *item) { struct cache_deferred_req *dreq; struct list_head pending; - - struct list_head *lp; + struct hlist_node *lp, *tmp; int hash = DFR_HASH(item); INIT_LIST_HEAD(&pending); spin_lock(&cache_defer_lock); - lp = cache_defer_hash[hash].next; - if (lp) { - while (lp != &cache_defer_hash[hash]) { - dreq = list_entry(lp, struct cache_deferred_req, hash); - lp = lp->next; - if (dreq->item == item) { - __unhash_deferred_req(dreq); - list_add(&dreq->recent, &pending); - } + hlist_for_each_entry_safe(dreq, lp, tmp, &cache_defer_hash[hash], hash) + if (dreq->item == item) { + __unhash_deferred_req(dreq); + list_add(&dreq->recent, &pending); } - } + spin_unlock(&cache_defer_lock); while (!list_empty(&pending)) { -- cgit v1.2.3 From 593ce16b943ea37d4ec62c377b32d7f3f4085e84 Mon Sep 17 00:00:00 2001 From: Pavel Emelyanov Date: Mon, 27 Sep 2010 14:00:15 +0400 Subject: sunrpc: Add routines that allow registering per-net caches Existing calls do the same, but for the init_net. Signed-off-by: Pavel Emelyanov Signed-off-by: J. Bruce Fields --- include/linux/sunrpc/cache.h | 2 ++ net/sunrpc/cache.c | 27 +++++++++++++++++++-------- 2 files changed, 21 insertions(+), 8 deletions(-) (limited to 'include/linux/sunrpc/cache.h') diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index 03496357f455..6950c981882d 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -197,7 +197,9 @@ extern void cache_purge(struct cache_detail *detail); #define NEVER (0x7FFFFFFF) extern void __init cache_initialize(void); extern int cache_register(struct cache_detail *cd); +extern int cache_register_net(struct cache_detail *cd, struct net *net); extern void cache_unregister(struct cache_detail *cd); +extern void cache_unregister_net(struct cache_detail *cd, struct net *net); extern int sunrpc_cache_register_pipefs(struct dentry *parent, const char *, mode_t, struct cache_detail *); diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index ff733dfef3b8..e84e7ddeecd4 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -34,6 +34,7 @@ #include #include #include +#include #define RPCDBG_FACILITY RPCDBG_CACHE @@ -1537,7 +1538,7 @@ static const struct file_operations cache_flush_operations_procfs = { .release = release_flush_procfs, }; -static void remove_cache_proc_entries(struct cache_detail *cd) +static void remove_cache_proc_entries(struct cache_detail *cd, struct net *net) { if (cd->u.procfs.proc_ent == NULL) return; @@ -1552,7 +1553,7 @@ static void remove_cache_proc_entries(struct cache_detail *cd) } #ifdef CONFIG_PROC_FS -static int create_cache_proc_entries(struct cache_detail *cd) +static int create_cache_proc_entries(struct cache_detail *cd, struct net *net) { struct proc_dir_entry *p; @@ -1587,11 +1588,11 @@ static int create_cache_proc_entries(struct cache_detail *cd) } return 0; out_nomem: - remove_cache_proc_entries(cd); + remove_cache_proc_entries(cd, net); return -ENOMEM; } #else /* CONFIG_PROC_FS */ -static int create_cache_proc_entries(struct cache_detail *cd) +static int create_cache_proc_entries(struct cache_detail *cd, struct net *net) { return 0; } @@ -1602,23 +1603,33 @@ void __init cache_initialize(void) INIT_DELAYED_WORK_DEFERRABLE(&cache_cleaner, do_cache_clean); } -int cache_register(struct cache_detail *cd) +int cache_register_net(struct cache_detail *cd, struct net *net) { int ret; sunrpc_init_cache_detail(cd); - ret = create_cache_proc_entries(cd); + ret = create_cache_proc_entries(cd, net); if (ret) sunrpc_destroy_cache_detail(cd); return ret; } + +int cache_register(struct cache_detail *cd) +{ + return cache_register_net(cd, &init_net); +} EXPORT_SYMBOL_GPL(cache_register); -void cache_unregister(struct cache_detail *cd) +void cache_unregister_net(struct cache_detail *cd, struct net *net) { - remove_cache_proc_entries(cd); + remove_cache_proc_entries(cd, net); sunrpc_destroy_cache_detail(cd); } + +void cache_unregister(struct cache_detail *cd) +{ + cache_unregister_net(cd, &init_net); +} EXPORT_SYMBOL_GPL(cache_unregister); static ssize_t cache_read_pipefs(struct file *filp, char __user *buf, -- cgit v1.2.3