Merge remote branch 'trond/bugfixes' into for-2.6.37
Without some client-side fixes, server testing is currently difficult.
diff --git a/fs/nfs/dns_resolve.c b/fs/nfs/dns_resolve.c
index dba50a5..a6e711a 100644
--- a/fs/nfs/dns_resolve.c
+++ b/fs/nfs/dns_resolve.c
@@ -167,7 +167,7 @@
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 @@
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 @@
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/export.c b/fs/nfsd/export.c
index c2a4f71..e56827b 100644
--- a/fs/nfsd/export.c
+++ b/fs/nfsd/export.c
@@ -935,10 +935,9 @@
ek = exp_get_fsid_key(exp->ex_client, exp->ex_fsid);
if (!IS_ERR(ek)) {
- ek->h.expiry_time = get_seconds()-1;
+ sunrpc_invalidate(&ek->h, &svc_expkey_cache);
cache_put(&ek->h, &svc_expkey_cache);
}
- svc_expkey_cache.nextcheck = get_seconds();
}
static int exp_fsid_hash(svc_client *clp, struct svc_export *exp)
@@ -973,10 +972,9 @@
ek = exp_get_key(exp->ex_client, inode->i_sb->s_dev, inode->i_ino);
if (!IS_ERR(ek)) {
- ek->h.expiry_time = get_seconds()-1;
+ sunrpc_invalidate(&ek->h, &svc_expkey_cache);
cache_put(&ek->h, &svc_expkey_cache);
}
- svc_expkey_cache.nextcheck = get_seconds();
}
/*
@@ -1097,8 +1095,7 @@
static void
exp_do_unexport(svc_export *unexp)
{
- unexp->h.expiry_time = get_seconds()-1;
- svc_export_cache.nextcheck = get_seconds();
+ sunrpc_invalidate(&unexp->h, &svc_export_cache);
exp_unhash(unexp);
exp_fsid_unhash(unexp);
}
diff --git a/fs/nfsd/nfs4callback.c b/fs/nfsd/nfs4callback.c
index 988cbb3..014482c 100644
--- a/fs/nfsd/nfs4callback.c
+++ b/fs/nfsd/nfs4callback.c
@@ -41,7 +41,6 @@
#define NFSPROC4_CB_NULL 0
#define NFSPROC4_CB_COMPOUND 1
-#define NFS4_STATEID_SIZE 16
/* Index of predefined Linux callback client operations */
diff --git a/fs/nfsd/nfs4idmap.c b/fs/nfsd/nfs4idmap.c
index c78dbf4..808b33a4 100644
--- a/fs/nfsd/nfs4idmap.c
+++ b/fs/nfsd/nfs4idmap.c
@@ -550,7 +550,7 @@
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 7bf3e84..52a7d72 100644
--- a/include/linux/sunrpc/cache.h
+++ b/include/linux/sunrpc/cache.h
@@ -125,6 +125,9 @@
*/
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
@@ -218,14 +221,42 @@
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 = 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 2b06410..da872f9 100644
--- a/net/sunrpc/cache.c
+++ b/net/sunrpc/cache.c
@@ -42,7 +42,7 @@
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 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_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 @@
/* 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 @@
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 @@
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 @@
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;
}
@@ -509,40 +509,39 @@
static struct list_head cache_defer_hash[DFR_HASHSIZE];
static int cache_defer_cnt;
-static int cache_defer_req(struct cache_req *req, struct cache_head *item)
+static void __unhash_deferred_req(struct cache_deferred_req *dreq)
{
- struct cache_deferred_req *dreq, *discard;
+ list_del_init(&dreq->recent);
+ list_del_init(&dreq->hash);
+ cache_defer_cnt--;
+}
+
+static void __hash_deferred_req(struct cache_deferred_req *dreq, struct cache_head *item)
+{
int hash = DFR_HASH(item);
- if (cache_defer_cnt >= DFR_MAX) {
- /* too much in the cache, randomly drop this one,
- * or continue and drop the oldest below
- */
- if (net_random()&1)
- return -ENOMEM;
- }
- dreq = req->defer(req);
- if (dreq == NULL)
- return -ENOMEM;
+ 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]);
+}
+
+static int setup_deferral(struct cache_deferred_req *dreq, struct cache_head *item)
+{
+ struct cache_deferred_req *discard;
dreq->item = item;
spin_lock(&cache_defer_lock);
- 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]);
+ __hash_deferred_req(dreq, item);
/* it is in, now maybe clean up */
discard = NULL;
if (++cache_defer_cnt > DFR_MAX) {
discard = list_entry(cache_defer_list.prev,
struct cache_deferred_req, recent);
- list_del_init(&discard->recent);
- list_del_init(&discard->hash);
- cache_defer_cnt--;
+ __unhash_deferred_req(discard);
}
spin_unlock(&cache_defer_lock);
@@ -558,6 +557,87 @@
return 0;
}
+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_wait_req(struct cache_req *req, struct cache_head *item)
+{
+ struct thread_deferred_req sleeper;
+ struct cache_deferred_req *dreq = &sleeper.handle;
+ int ret;
+
+ sleeper.completion = COMPLETION_INITIALIZER_ONSTACK(sleeper.completion);
+ dreq->revisit = cache_restart_thread;
+
+ ret = setup_deferral(dreq, item);
+ if (ret)
+ return ret;
+
+ 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)) {
+ __unhash_deferred_req(&sleeper.handle);
+ 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
+ */
+ return -ETIMEDOUT;
+ }
+ /* 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;
+}
+
+static int cache_defer_req(struct cache_req *req, struct cache_head *item)
+{
+ struct cache_deferred_req *dreq;
+ int ret;
+
+ if (cache_defer_cnt >= DFR_MAX) {
+ /* too much in the cache, randomly drop this one,
+ * or continue and drop the oldest
+ */
+ if (net_random()&1)
+ return -ENOMEM;
+ }
+ if (req->thread_wait) {
+ ret = cache_wait_req(req, item);
+ if (ret != -ETIMEDOUT)
+ return ret;
+ }
+ dreq = req->defer(req);
+ if (dreq == NULL)
+ return -ENOMEM;
+ return setup_deferral(dreq, item);
+}
+
static void cache_revisit_request(struct cache_head *item)
{
struct cache_deferred_req *dreq;
@@ -575,9 +655,8 @@
dreq = list_entry(lp, struct cache_deferred_req, hash);
lp = lp->next;
if (dreq->item == item) {
- list_del_init(&dreq->hash);
- list_move(&dreq->recent, &pending);
- cache_defer_cnt--;
+ __unhash_deferred_req(dreq);
+ list_add(&dreq->recent, &pending);
}
}
}
@@ -600,11 +679,8 @@
spin_lock(&cache_defer_lock);
list_for_each_entry_safe(dreq, tmp, &cache_defer_list, recent) {
- if (dreq->owner == owner) {
- list_del_init(&dreq->hash);
- list_move(&dreq->recent, &pending);
- cache_defer_cnt--;
- }
+ if (dreq->owner == owner)
+ __unhash_deferred_req(dreq);
}
spin_unlock(&cache_defer_lock);
@@ -902,7 +978,7 @@
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 +1110,7 @@
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 +1295,8 @@
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 +1362,7 @@
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 +1380,20 @@
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;
diff --git a/net/sunrpc/svc_xprt.c b/net/sunrpc/svc_xprt.c
index cbc0849..95fc3e8 100644
--- a/net/sunrpc/svc_xprt.c
+++ b/net/sunrpc/svc_xprt.c
@@ -651,6 +651,11 @@
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 @@
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 queued, 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);