Re: lockdep trace from posix timers

[Dave Jones <davej@redhat.com>]

On Thu, Aug 16, 2012 at 08:54:31PM +0800, Ming Lei wrote:
 > On Sat, Jul 28, 2012 at 12:20 AM, Dave Jones <davej@redhat.com> wrote:
 > > On Tue, Jul 24, 2012 at 04:36:13PM -0400, Dave Jones wrote:
 > >  > Linus tree as of 5fecc9d8f59e765c2a48379dd7c6f5cf88c7d75a
 > >  >
 > >  >      Dave
 > >  >
 > >  > ======================================================
 > >  > [ INFO: HARDIRQ-safe -> HARDIRQ-unsafe lock order detected ]
 > >  > 3.5.0+ #122 Not tainted
 > >  > ------------------------------------------------------
 > >  > trinity-child2/5327 [HC0[0]:SC0[0]:HE0:SE1] is trying to acquire:
 > >  > blocked:  (tasklist_lock){.+.+..}, instance: ffffffff81c05098, at: [<ffffffff8109762b>] posix_cpu_timer_del+0x2b/0xe0
 > >  >
 > >  > and this task is already holding:
 > >  > blocked:  (&(&new_timer->it_lock)->rlock){-.-...}, instance: ffff880143bce170, at: [<ffffffff81093d49>] __lock_timer+0x89/0x1f0
 > >  > which would create a new lock dependency:
 > >  >  (&(&new_timer->it_lock)->rlock){-.-...} -> (tasklist_lock){.+.+..}
 > >  >
 > >  > but this new dependency connects a HARDIRQ-irq-safe lock:
 > >  >  (&(&new_timer->it_lock)->rlock){-.-...}
 > >  > ... which became HARDIRQ-irq-safe at:
 > >
 > > Shall I start bisecting this ? I can trigger it very easily, but if you
 > > can give me a set of commits to narrow down, it'll speed up the bisection.
 > 
 > It should a real possible deadlock, could you test the below patch to
 > see if it can fix the warning?

I've not managed to hit it in a while. It seems very dependant upon
specific builds for some reason.  Very strange.

	Dave