The patch titled
Subject: fs/epoll: reduce the scope of wq lock in epoll_wait()
has been removed from the -mm tree. Its filename was
fs-epoll-reduce-the-scope-of-wq-lock-in-epoll_wait.patch
This patch was dropped because it was merged into mainline or a subsystem tree
------------------------------------------------------
From: Davidlohr Bueso <dave@xxxxxxxxxxxx>
Subject: fs/epoll: reduce the scope of wq lock in epoll_wait()
This patch aims at reducing ep wq.lock hold times in epoll_wait(2). For
the blocking case, there is no need to constantly take and drop the
spinlock, which is only needed to manipulate the waitqueue.
The call to ep_events_available() is now lockless, and only exposed to
benign races. Here, if false positive (returns available events and does
not see another thread deleting an epi from the list) we call into
send_events and then the list's state is correctly seen. Otoh, if a false
negative and we don't see a list_add_tail(), for example, from irq
callback, then it is rechecked again before blocking, which will see the
correct state.
In order for more accuracy to see concurrent list_del_init(), use the
list_empty_careful() variant -- of course, this won't be safe against
insertions from wakeup.
For the overflow list we obviously need to prevent load/store tearing as
we don't want to see partial values while the ready list is disabled.
[dave@xxxxxxxxxxxx: forgotten fixlets]
Link: http://lkml.kernel.org/r/20181109155258.jxcr4t2pnz6zqct3@linux-r8p5
Link: http://lkml.kernel.org/r/20181108051006.18751-6-dave@xxxxxxxxxxxx
Signed-off-by: Davidlohr Bueso <dbueso@xxxxxxx>
Suggested-by: Jason Baron <jbaron@xxxxxxxxxx>
Cc: Al Viro <viro@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Andrew Morton <akpm@xxxxxxxxxxxxxxxxxxxx>
---
fs/eventpoll.c | 114 ++++++++++++++++++++++++-----------------------
1 file changed, 60 insertions(+), 54 deletions(-)
--- a/fs/eventpoll.c~fs-epoll-reduce-the-scope-of-wq-lock-in-epoll_wait
+++ a/fs/eventpoll.c
@@ -381,7 +381,8 @@ static void ep_nested_calls_init(struct
*/
static inline int ep_events_available(struct eventpoll *ep)
{
- return !list_empty(&ep->rdllist) || ep->ovflist != EP_UNACTIVE_PTR;
+ return !list_empty_careful(&ep->rdllist) ||
+ READ_ONCE(ep->ovflist) != EP_UNACTIVE_PTR;
}
#ifdef CONFIG_NET_RX_BUSY_POLL
@@ -698,7 +699,7 @@ static __poll_t ep_scan_ready_list(struc
*/
spin_lock_irq(&ep->wq.lock);
list_splice_init(&ep->rdllist, &txlist);
- ep->ovflist = NULL;
+ WRITE_ONCE(ep->ovflist, NULL);
spin_unlock_irq(&ep->wq.lock);
/*
@@ -712,7 +713,7 @@ static __poll_t ep_scan_ready_list(struc
* other events might have been queued by the poll callback.
* We re-insert them inside the main ready-list here.
*/
- for (nepi = ep->ovflist; (epi = nepi) != NULL;
+ for (nepi = READ_ONCE(ep->ovflist); (epi = nepi) != NULL;
nepi = epi->next, epi->next = EP_UNACTIVE_PTR) {
/*
* We need to check if the item is already in the list.
@@ -730,7 +731,7 @@ static __poll_t ep_scan_ready_list(struc
* releasing the lock, events will be queued in the normal way inside
* ep->rdllist.
*/
- ep->ovflist = EP_UNACTIVE_PTR;
+ WRITE_ONCE(ep->ovflist, EP_UNACTIVE_PTR);
/*
* Quickly re-inject items left on "txlist".
@@ -1153,10 +1154,10 @@ static int ep_poll_callback(wait_queue_e
* semantics). All the events that happen during that period of time are
* chained in ep->ovflist and requeued later on.
*/
- if (ep->ovflist != EP_UNACTIVE_PTR) {
+ if (READ_ONCE(ep->ovflist) != EP_UNACTIVE_PTR) {
if (epi->next == EP_UNACTIVE_PTR) {
- epi->next = ep->ovflist;
- ep->ovflist = epi;
+ epi->next = READ_ONCE(ep->ovflist);
+ WRITE_ONCE(ep->ovflist, epi);
if (epi->ws) {
/*
* Activate ep->ws since epi->ws may get
@@ -1762,10 +1763,17 @@ static int ep_poll(struct eventpoll *ep,
} else if (timeout == 0) {
/*
* Avoid the unnecessary trip to the wait queue loop, if the
- * caller specified a non blocking operation.
+ * caller specified a non blocking operation. We still need
+ * lock because we could race and not see an epi being added
+ * to the ready list while in irq callback. Thus incorrectly
+ * returning 0 back to userspace.
*/
timed_out = 1;
+
spin_lock_irq(&ep->wq.lock);
+ eavail = ep_events_available(ep);
+ spin_unlock_irq(&ep->wq.lock);
+
goto check_events;
}
@@ -1774,64 +1782,62 @@ fetch_events:
if (!ep_events_available(ep))
ep_busy_loop(ep, timed_out);
+ eavail = ep_events_available(ep);
+ if (eavail)
+ goto check_events;
+
+ /*
+ * Busy poll timed out. Drop NAPI ID for now, we can add
+ * it back in when we have moved a socket with a valid NAPI
+ * ID onto the ready list.
+ */
+ ep_reset_busy_poll_napi_id(ep);
+
+ /*
+ * We don't have any available event to return to the caller.
+ * We need to sleep here, and we will be wake up by
+ * ep_poll_callback() when events will become available.
+ */
+ init_waitqueue_entry(&wait, current);
spin_lock_irq(&ep->wq.lock);
+ __add_wait_queue_exclusive(&ep->wq, &wait);
+ spin_unlock_irq(&ep->wq.lock);
- if (!ep_events_available(ep)) {
+ for (;;) {
/*
- * Busy poll timed out. Drop NAPI ID for now, we can add
- * it back in when we have moved a socket with a valid NAPI
- * ID onto the ready list.
+ * We don't want to sleep if the ep_poll_callback() sends us
+ * a wakeup in between. That's why we set the task state
+ * to TASK_INTERRUPTIBLE before doing the checks.
*/
- ep_reset_busy_poll_napi_id(ep);
-
+ set_current_state(TASK_INTERRUPTIBLE);
/*
- * We don't have any available event to return to the caller.
- * We need to sleep here, and we will be wake up by
- * ep_poll_callback() when events will become available.
+ * Always short-circuit for fatal signals to allow
+ * threads to make a timely exit without the chance of
+ * finding more events available and fetching
+ * repeatedly.
*/
- init_waitqueue_entry(&wait, current);
- __add_wait_queue_exclusive(&ep->wq, &wait);
-
- for (;;) {
- /*
- * We don't want to sleep if the ep_poll_callback() sends us
- * a wakeup in between. That's why we set the task state
- * to TASK_INTERRUPTIBLE before doing the checks.
- */
- set_current_state(TASK_INTERRUPTIBLE);
- /*
- * Always short-circuit for fatal signals to allow
- * threads to make a timely exit without the chance of
- * finding more events available and fetching
- * repeatedly.
- */
- if (fatal_signal_pending(current)) {
- res = -EINTR;
- break;
- }
- if (ep_events_available(ep) || timed_out)
- break;
- if (signal_pending(current)) {
- res = -EINTR;
- break;
- }
-
- spin_unlock_irq(&ep->wq.lock);
- if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS))
- timed_out = 1;
-
- spin_lock_irq(&ep->wq.lock);
+ if (fatal_signal_pending(current)) {
+ res = -EINTR;
+ break;
+ }
+ if (ep_events_available(ep) || timed_out)
+ break;
+ if (signal_pending(current)) {
+ res = -EINTR;
+ break;
}
- __remove_wait_queue(&ep->wq, &wait);
- __set_current_state(TASK_RUNNING);
+ if (!schedule_hrtimeout_range(to, slack, HRTIMER_MODE_ABS))
+ timed_out = 1;
}
-check_events:
- /* Is it worth to try to dig for events ? */
- eavail = ep_events_available(ep);
+ __set_current_state(TASK_RUNNING);
+
+ spin_lock_irq(&ep->wq.lock);
+ __remove_wait_queue(&ep->wq, &wait);
spin_unlock_irq(&ep->wq.lock);
+check_events:
/*
* Try to transfer events to user space. In case we get 0 events and
* there's still timeout left over, we go trying again in search of
_
Patches currently in -mm which might be from dave@xxxxxxxxxxxx are
