From 226453d8cfbd5f3bbdcef0d71e67186715bb811d Mon Sep 17 00:00:00 2001 From: Trond Myklebust Date: Wed, 7 Oct 2015 13:54:42 -0400 Subject: SUNRPC: Use MSG_SENDPAGE_NOTLAST when calling sendpage() If we're sending more pages via kernel_sendpage(), then set MSG_SENDPAGE_NOTLAST. Signed-off-by: Trond Myklebust Signed-off-by: J. Bruce Fields --- net/sunrpc/svcsock.c | 2 +- net/sunrpc/xprtsock.c | 4 +++- 2 files changed, 4 insertions(+), 2 deletions(-) (limited to 'net') diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 0c8120229a03..48923730722d 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -181,7 +181,7 @@ int svc_send_common(struct socket *sock, struct xdr_buf *xdr, struct page **ppage = xdr->pages; size_t base = xdr->page_base; unsigned int pglen = xdr->page_len; - unsigned int flags = MSG_MORE; + unsigned int flags = MSG_MORE | MSG_SENDPAGE_NOTLAST; int slen; int len = 0; diff --git a/net/sunrpc/xprtsock.c b/net/sunrpc/xprtsock.c index 1a85e0ed0b48..f264962a1d8a 100644 --- a/net/sunrpc/xprtsock.c +++ b/net/sunrpc/xprtsock.c @@ -360,8 +360,10 @@ static int xs_send_pagedata(struct socket *sock, struct xdr_buf *xdr, unsigned i int flags = XS_SENDMSG_FLAGS; remainder -= len; - if (remainder != 0 || more) + if (more) flags |= MSG_MORE; + if (remainder != 0) + flags |= MSG_SENDPAGE_NOTLAST | MSG_MORE; err = do_sendpage(sock, *ppage, base, len, flags); if (remainder == 0 || err != len) break; -- cgit v1.2.3 From cc6a7aab5570beef884ff95f7cade6634bf815a1 Mon Sep 17 00:00:00 2001 From: Arnd Bergmann Date: Fri, 9 Oct 2015 16:13:45 +0200 Subject: sunrpc: avoid warning in gss_key_timeout The gss_key_timeout() function causes a harmless warning in some configurations, e.g. ARM imx_v6_v7_defconfig with gcc-5.2, if the compiler cannot figure out the state of the 'expire' variable across an rcu_read_unlock(): net/sunrpc/auth_gss/auth_gss.c: In function 'gss_key_timeout': net/sunrpc/auth_gss/auth_gss.c:1422:211: warning: 'expire' may be used uninitialized in this function [-Wmaybe-uninitialized] To avoid this warning without adding a bogus initialization, this rewrites the function so the comparison is done inside of the critical section. As a side-effect, it also becomes slightly easier to understand because the implementation now more closely resembles the comment above it. Signed-off-by: Arnd Bergmann Fixes: c5e6aecd034e7 ("sunrpc: fix RCU handling of gc_ctx field") Signed-off-by: J. Bruce Fields --- net/sunrpc/auth_gss/auth_gss.c | 13 ++++++------- 1 file changed, 6 insertions(+), 7 deletions(-) (limited to 'net') diff --git a/net/sunrpc/auth_gss/auth_gss.c b/net/sunrpc/auth_gss/auth_gss.c index dace13d7638e..799e65b944b9 100644 --- a/net/sunrpc/auth_gss/auth_gss.c +++ b/net/sunrpc/auth_gss/auth_gss.c @@ -1411,17 +1411,16 @@ gss_key_timeout(struct rpc_cred *rc) { struct gss_cred *gss_cred = container_of(rc, struct gss_cred, gc_base); struct gss_cl_ctx *ctx; - unsigned long now = jiffies; - unsigned long expire; + unsigned long timeout = jiffies + (gss_key_expire_timeo * HZ); + int ret = 0; rcu_read_lock(); ctx = rcu_dereference(gss_cred->gc_ctx); - if (ctx) - expire = ctx->gc_expiry - (gss_key_expire_timeo * HZ); + if (!ctx || time_after(timeout, ctx->gc_expiry)) + ret = -EACCES; rcu_read_unlock(); - if (!ctx || time_after(now, expire)) - return -EACCES; - return 0; + + return ret; } static int -- cgit v1.2.3 From 778620364ef525e83597a6edee4d0a69db67fd3d Mon Sep 17 00:00:00 2001 From: Neil Brown Date: Fri, 16 Oct 2015 08:59:08 +1100 Subject: sunrpc/cache: make cache flushing more reliable. The caches used to store sunrpc authentication information can be flushed by writing a timestamp to a file in /proc. This timestamp has a one-second resolution and any entry in cache that was last_refreshed *before* that time is treated as expired. This is problematic as it is not possible to reliably flush the cache without interrupting NFS service. If the current time is written to the "flush" file, any entry that was added since the current second started will still be treated as valid. If one second beyond than the current time is written to the file then no entries can be valid until the second ticks over. This will mean that no NFS request will be handled for up to 1 second. To resolve this issue we make two changes: 1/ treat an entry as expired if the timestamp when it was last_refreshed is before *or the same as* the expiry time. This means that current code which writes out the current time will now flush the cache reliably. 2/ when a new entry in added to the cache - set the last_refresh timestamp to 1 second *beyond* the current flush time, when that not in the past. This ensures that newly added entries will always be valid. Now that we have a very reliable way to flush the cache, and also since we are using "since-boot" timestamps which are monotonic, change cache_purge() to set the smallest future flush_time which will work, and leave it there: don't revert to '1'. Also disable the setting of the 'flush_time' far into the future. That has never been useful and is now awkward as it would cause last_refresh times to be strange. Finally: if a request is made to set the 'flush_time' to the current second, assume the intent is to flush the cache and advance it, if necessary, to 1 second beyond the current 'flush_time' so that all active entries will be deemed to be expired. As part of this we need to add a 'cache_detail' arg to cache_init() and cache_fresh_locked() so they can find the current ->flush_time. Signed-off-by: NeilBrown Reported-by: Olaf Kirch Signed-off-by: J. Bruce Fields --- include/linux/sunrpc/cache.h | 16 ++++++++----- net/sunrpc/cache.c | 53 +++++++++++++++++++++++++++++++++----------- 2 files changed, 51 insertions(+), 18 deletions(-) (limited to 'net') diff --git a/include/linux/sunrpc/cache.h b/include/linux/sunrpc/cache.h index 03d3b4c92d9f..ed03c9f7f908 100644 --- a/include/linux/sunrpc/cache.h +++ b/include/linux/sunrpc/cache.h @@ -48,8 +48,10 @@ struct cache_head { struct hlist_node cache_list; time_t expiry_time; /* After time time, don't use the data */ - time_t last_refresh; /* If CACHE_PENDING, this is when upcall - * was sent, else this is when update was received + time_t last_refresh; /* If CACHE_PENDING, this is when upcall was + * sent, else this is when update was + * received, though it is alway set to + * be *after* ->flush_time. */ struct kref ref; unsigned long flags; @@ -105,8 +107,12 @@ struct cache_detail { /* fields below this comment are for internal use * and should not be touched by cache owners */ - time_t flush_time; /* flush all cache items with last_refresh - * earlier than this */ + time_t flush_time; /* flush all cache items with + * last_refresh at or earlier + * than this. last_refresh + * is never set at or earlier + * than this. + */ struct list_head others; time_t nextcheck; int entries; @@ -203,7 +209,7 @@ static inline void cache_put(struct cache_head *h, struct cache_detail *cd) static inline int cache_is_expired(struct cache_detail *detail, struct cache_head *h) { return (h->expiry_time < seconds_since_boot()) || - (detail->flush_time > h->last_refresh); + (detail->flush_time >= h->last_refresh); } extern int cache_check(struct cache_detail *detail, diff --git a/net/sunrpc/cache.c b/net/sunrpc/cache.c index 4a2340a54401..5e4f815c2b34 100644 --- a/net/sunrpc/cache.c +++ b/net/sunrpc/cache.c @@ -41,13 +41,16 @@ static bool cache_defer_req(struct cache_req *req, struct cache_head *item); static void cache_revisit_request(struct cache_head *item); -static void cache_init(struct cache_head *h) +static void cache_init(struct cache_head *h, struct cache_detail *detail) { time_t now = seconds_since_boot(); INIT_HLIST_NODE(&h->cache_list); h->flags = 0; kref_init(&h->ref); h->expiry_time = now + CACHE_NEW_EXPIRY; + if (now <= detail->flush_time) + /* ensure it isn't already expired */ + now = detail->flush_time + 1; h->last_refresh = now; } @@ -81,7 +84,7 @@ struct cache_head *sunrpc_cache_lookup(struct cache_detail *detail, * we might get lose if we need to * cache_put it soon. */ - cache_init(new); + cache_init(new, detail); detail->init(new, key); write_lock(&detail->hash_lock); @@ -116,10 +119,15 @@ EXPORT_SYMBOL_GPL(sunrpc_cache_lookup); static void cache_dequeue(struct cache_detail *detail, struct cache_head *ch); -static void cache_fresh_locked(struct cache_head *head, time_t expiry) +static void cache_fresh_locked(struct cache_head *head, time_t expiry, + struct cache_detail *detail) { + time_t now = seconds_since_boot(); + if (now <= detail->flush_time) + /* ensure it isn't immediately treated as expired */ + now = detail->flush_time + 1; head->expiry_time = expiry; - head->last_refresh = seconds_since_boot(); + head->last_refresh = now; smp_wmb(); /* paired with smp_rmb() in cache_is_valid() */ set_bit(CACHE_VALID, &head->flags); } @@ -149,7 +157,7 @@ struct cache_head *sunrpc_cache_update(struct cache_detail *detail, set_bit(CACHE_NEGATIVE, &old->flags); else detail->update(old, new); - cache_fresh_locked(old, new->expiry_time); + cache_fresh_locked(old, new->expiry_time, detail); write_unlock(&detail->hash_lock); cache_fresh_unlocked(old, detail); return old; @@ -162,7 +170,7 @@ struct cache_head *sunrpc_cache_update(struct cache_detail *detail, cache_put(old, detail); return NULL; } - cache_init(tmp); + cache_init(tmp, detail); detail->init(tmp, old); write_lock(&detail->hash_lock); @@ -173,8 +181,8 @@ struct cache_head *sunrpc_cache_update(struct cache_detail *detail, hlist_add_head(&tmp->cache_list, &detail->hash_table[hash]); detail->entries++; cache_get(tmp); - cache_fresh_locked(tmp, new->expiry_time); - cache_fresh_locked(old, 0); + cache_fresh_locked(tmp, new->expiry_time, detail); + cache_fresh_locked(old, 0, detail); write_unlock(&detail->hash_lock); cache_fresh_unlocked(tmp, detail); cache_fresh_unlocked(old, detail); @@ -219,7 +227,8 @@ static int try_to_negate_entry(struct cache_detail *detail, struct cache_head *h rv = cache_is_valid(h); if (rv == -EAGAIN) { set_bit(CACHE_NEGATIVE, &h->flags); - cache_fresh_locked(h, seconds_since_boot()+CACHE_NEW_EXPIRY); + cache_fresh_locked(h, seconds_since_boot()+CACHE_NEW_EXPIRY, + detail); rv = -ENOENT; } write_unlock(&detail->hash_lock); @@ -487,10 +496,13 @@ EXPORT_SYMBOL_GPL(cache_flush); void cache_purge(struct cache_detail *detail) { - detail->flush_time = LONG_MAX; + time_t now = seconds_since_boot(); + if (detail->flush_time >= now) + now = detail->flush_time + 1; + /* 'now' is the maximum value any 'last_refresh' can have */ + detail->flush_time = now; detail->nextcheck = seconds_since_boot(); cache_flush(); - detail->flush_time = 1; } EXPORT_SYMBOL_GPL(cache_purge); @@ -1436,6 +1448,7 @@ static ssize_t write_flush(struct file *file, const char __user *buf, { char tbuf[20]; char *bp, *ep; + time_t then, now; if (*ppos || count > sizeof(tbuf)-1) return -EINVAL; @@ -1447,8 +1460,22 @@ static ssize_t write_flush(struct file *file, const char __user *buf, return -EINVAL; bp = tbuf; - cd->flush_time = get_expiry(&bp); - cd->nextcheck = seconds_since_boot(); + then = get_expiry(&bp); + now = seconds_since_boot(); + cd->nextcheck = now; + /* Can only set flush_time to 1 second beyond "now", or + * possibly 1 second beyond flushtime. This is because + * flush_time never goes backwards so it mustn't get too far + * ahead of time. + */ + if (then >= now) { + /* Want to flush everything, so behave like cache_purge() */ + if (cd->flush_time >= now) + now = cd->flush_time + 1; + then = now; + } + + cd->flush_time = then; cache_flush(); *ppos += count; -- cgit v1.2.3 From ea833f5de3ab49a0aac79a16633fd510390b83a6 Mon Sep 17 00:00:00 2001 From: Stefan Hajnoczi Date: Tue, 27 Oct 2015 10:23:30 +0000 Subject: SUNRPC: drop stale comment in svc_setup_socket() The svc_setup_socket() function does set the send and receive buffer sizes, so the comment is out-of-date: Signed-off-by: Stefan Hajnoczi Signed-off-by: J. Bruce Fields --- net/sunrpc/svcsock.c | 1 - 1 file changed, 1 deletion(-) (limited to 'net') diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index 48923730722d..e0c7b3355495 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -1367,7 +1367,6 @@ EXPORT_SYMBOL_GPL(svc_sock_update_bufs); /* * Initialize socket for RPC use and create svc_sock struct - * XXX: May want to setsockopt SO_SNDBUF and SO_RCVBUF. */ static struct svc_sock *svc_setup_socket(struct svc_serv *serv, struct socket *sock, -- cgit v1.2.3 From 0442f14b15f8e7a8b3778a9f8cf640ef89b2df26 Mon Sep 17 00:00:00 2001 From: "J. Bruce Fields" Date: Fri, 9 Oct 2015 01:44:07 +0000 Subject: svcrpc: document lack of some memory barriers We're missing memory barriers in net/sunrpc/svcsock.c in some spots we'd expect them. But it doesn't appear they're necessary in our case, and this is likely a hot path--for now just document the odd behavior. Kosuke Tatsukawa found this issue while looking through the linux source code for places calling waitqueue_active() before wake_up*(), but without preceding memory barriers, after sending a patch to fix a similar issue in drivers/tty/n_tty.c (Details about the original issue can be found here: https://lkml.org/lkml/2015/9/28/849). Reported-by: Kosuke Tatsukawa Signed-off-by: J. Bruce Fields --- net/sunrpc/svcsock.c | 37 +++++++++++++++++++++++++++++++------ 1 file changed, 31 insertions(+), 6 deletions(-) (limited to 'net') diff --git a/net/sunrpc/svcsock.c b/net/sunrpc/svcsock.c index e0c7b3355495..1413cdcc131c 100644 --- a/net/sunrpc/svcsock.c +++ b/net/sunrpc/svcsock.c @@ -399,6 +399,31 @@ static int svc_sock_secure_port(struct svc_rqst *rqstp) return svc_port_is_privileged(svc_addr(rqstp)); } +static bool sunrpc_waitqueue_active(wait_queue_head_t *wq) +{ + if (!wq) + return false; + /* + * There should normally be a memory * barrier here--see + * wq_has_sleeper(). + * + * It appears that isn't currently necessary, though, basically + * because callers all appear to have sufficient memory barriers + * between the time the relevant change is made and the + * time they call these callbacks. + * + * The nfsd code itself doesn't actually explicitly wait on + * these waitqueues, but it may wait on them for example in + * sendpage() or sendmsg() calls. (And those may be the only + * places, since it it uses nonblocking reads.) + * + * Maybe we should add the memory barriers anyway, but these are + * hot paths so we'd need to be convinced there's no sigificant + * penalty. + */ + return waitqueue_active(wq); +} + /* * INET callback when data has been received on the socket. */ @@ -414,7 +439,7 @@ static void svc_udp_data_ready(struct sock *sk) set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags); svc_xprt_enqueue(&svsk->sk_xprt); } - if (wq && waitqueue_active(wq)) + if (sunrpc_waitqueue_active(wq)) wake_up_interruptible(wq); } @@ -432,7 +457,7 @@ static void svc_write_space(struct sock *sk) svc_xprt_enqueue(&svsk->sk_xprt); } - if (wq && waitqueue_active(wq)) { + if (sunrpc_waitqueue_active(wq)) { dprintk("RPC svc_write_space: someone sleeping on %p\n", svsk); wake_up_interruptible(wq); @@ -787,7 +812,7 @@ static void svc_tcp_listen_data_ready(struct sock *sk) } wq = sk_sleep(sk); - if (wq && waitqueue_active(wq)) + if (sunrpc_waitqueue_active(wq)) wake_up_interruptible_all(wq); } @@ -808,7 +833,7 @@ static void svc_tcp_state_change(struct sock *sk) set_bit(XPT_CLOSE, &svsk->sk_xprt.xpt_flags); svc_xprt_enqueue(&svsk->sk_xprt); } - if (wq && waitqueue_active(wq)) + if (sunrpc_waitqueue_active(wq)) wake_up_interruptible_all(wq); } @@ -823,7 +848,7 @@ static void svc_tcp_data_ready(struct sock *sk) set_bit(XPT_DATA, &svsk->sk_xprt.xpt_flags); svc_xprt_enqueue(&svsk->sk_xprt); } - if (wq && waitqueue_active(wq)) + if (sunrpc_waitqueue_active(wq)) wake_up_interruptible(wq); } @@ -1593,7 +1618,7 @@ static void svc_sock_detach(struct svc_xprt *xprt) sk->sk_write_space = svsk->sk_owspace; wq = sk_sleep(sk); - if (wq && waitqueue_active(wq)) + if (sunrpc_waitqueue_active(wq)) wake_up_interruptible(wq); } -- cgit v1.2.3