futex: avoid wake_futex() for a PI futex_q
authorDarren Hart <dvhart@linux.intel.com>
Tue, 27 Nov 2012 00:29:56 +0000 (16:29 -0800)
committerLinus Torvalds <torvalds@linux-foundation.org>
Tue, 27 Nov 2012 01:41:24 +0000 (17:41 -0800)
commitaa10990e028cac3d5e255711fb9fb47e00700e35
treedc39cb7cf8f193fb266bf3da618b77ea76a89c2d
parent8ffeb9b0e6369135bf03a073514f571ef10606b9
futex: avoid wake_futex() for a PI futex_q

Dave Jones reported a bug with futex_lock_pi() that his trinity test
exposed.  Sometime between queue_me() and taking the q.lock_ptr, the
lock_ptr became NULL, resulting in a crash.

While futex_wake() is careful to not call wake_futex() on futex_q's with
a pi_state or an rt_waiter (which are either waiting for a
futex_unlock_pi() or a PI futex_requeue()), futex_wake_op() and
futex_requeue() do not perform the same test.

Update futex_wake_op() and futex_requeue() to test for q.pi_state and
q.rt_waiter and abort with -EINVAL if detected.  To ensure any future
breakage is caught, add a WARN() to wake_futex() if the same condition
is true.

This fix has seen 3 hours of testing with "trinity -c futex" on an
x86_64 VM with 4 CPUS.

[akpm@linux-foundation.org: tidy up the WARN()]
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Reported-by: Dave Jones <davej@redat.com>
Cc: Thomas Gleixner <tglx@linutronix.de>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: John Kacur <jkacur@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
kernel/futex.c