fs/epoll: reduce the scope of wq lock in epoll_wait()
authorDavidlohr Bueso <dave@stgolabs.net>
Thu, 3 Jan 2019 23:27:15 +0000 (15:27 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Fri, 4 Jan 2019 21:13:46 +0000 (13:13 -0800)
commitc5a282e9635e9c7382821565083db5d260085e3e
tree2cd0f8fa570d493f4af033d5805bed974b0b88a6
parent21877e1a5b520132f54515f8835c963056418b4c
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@stgolabs.net: forgotten fixlets]
Link: http://lkml.kernel.org/r/20181109155258.jxcr4t2pnz6zqct3@linux-r8p5
Link: http://lkml.kernel.org/r/20181108051006.18751-6-dave@stgolabs.net
Signed-off-by: Davidlohr Bueso <dbueso@suse.de>
Suggested-by: Jason Baron <jbaron@akamai.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
fs/eventpoll.c