patches/kernel/0014-sched-wait-Fix-add_wait_queue-behavioral-change.patch

   1 From 0000000000000000000000000000000000000000 Mon Sep 17 00:00:00 2001
   2 From: Omar Sandoval <osandov@fb.com>
   3 Date: Tue, 5 Dec 2017 23:15:31 -0800
   4 Subject: [PATCH] sched/wait: Fix add_wait_queue() behavioral change
   5 MIME-Version: 1.0
   6 Content-Type: text/plain; charset=UTF-8
   7 Content-Transfer-Encoding: 8bit
   8
   9 The following cleanup commit:
  10
  11   50816c48997a ("sched/wait: Standardize internal naming of wait-queue entries")
  12
  13 ... unintentionally changed the behavior of add_wait_queue() from
  14 inserting the wait entry at the head of the wait queue to the tail
  15 of the wait queue.
  16
  17 Beyond a negative performance impact this change in behavior
  18 theoretically also breaks wait queues which mix exclusive and
  19 non-exclusive waiters, as non-exclusive waiters will not be
  20 woken up if they are queued behind enough exclusive waiters.
  21
  22 Signed-off-by: Omar Sandoval <osandov@fb.com>
  23 Reviewed-by: Jens Axboe <axboe@kernel.dk>
  24 Acked-by: Peter Zijlstra <peterz@infradead.org>
  25 Cc: Linus Torvalds <torvalds@linux-foundation.org>
  26 Cc: Thomas Gleixner <tglx@linutronix.de>
  27 Cc: kernel-team@fb.com
  28 Fixes: ("sched/wait: Standardize internal naming of wait-queue entries")
  29 Link: http://lkml.kernel.org/r/a16c8ccffd39bd08fdaa45a5192294c784b803a7.1512544324.git.osandov@fb.com
  30 Signed-off-by: Ingo Molnar <mingo@kernel.org>
  31 (cherry picked from commit c6b9d9a33029014446bd9ed84c1688f6d3d4eab9)
  32 Signed-off-by: Fabian Grünbichler <f.gruenbichler@proxmox.com>
  33 ---
  34  kernel/sched/wait.c | 2 +-
  35  1 file changed, 1 insertion(+), 1 deletion(-)
  36
  37 diff --git a/kernel/sched/wait.c b/kernel/sched/wait.c
  38 index d6afed6d0752..c09ebe92a40a 100644
  39 --- a/kernel/sched/wait.c
  40 +++ b/kernel/sched/wait.c
  41 @@ -27,7 +27,7 @@ void add_wait_queue(struct wait_queue_head *wq_head, struct wait_queue_entry *wq
  42
  43         wq_entry->flags &= ~WQ_FLAG_EXCLUSIVE;
  44         spin_lock_irqsave(&wq_head->lock, flags);
  45 -       __add_wait_queue_entry_tail(wq_head, wq_entry);
  46 +       __add_wait_queue(wq_head, wq_entry);
  47         spin_unlock_irqrestore(&wq_head->lock, flags);
  48  }
  49  EXPORT_SYMBOL(add_wait_queue);
  50 --
  51 2.14.2
  52