linux-kernel - Re: [RT PATCH 1/2] timers: wakeup all timer waiters

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <20160714131945.67cc8ce5@gandalf.local.home>
Date:	Thu, 14 Jul 2016 13:19:45 -0400
From:	Steven Rostedt <rostedt@...dmis.org>
To:	Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc:	linux-rt-users@...r.kernel.org, linux-kernel@...r.kernel.org,
	tglx@...utronix.de
Subject: Re: [RT PATCH 1/2] timers: wakeup all timer waiters

On Thu, 14 Jul 2016 19:14:58 +0200
Sebastian Andrzej Siewior <bigeasy@...utronix.de> wrote:

> On 07/14/2016 06:13 PM, Steven Rostedt wrote:
> >>  
> >> -# define wakeup_timer_waiters(b)	wake_up(&(b)->wait_for_running_timer)
> >> +# define wakeup_timer_waiters(b)	wake_up_all(&(b)->wait_for_running_timer)  
> > 
> > OK, I just received this patch (way after patch 2)
> > 
> > I'm assuming that patch two was done such that you don't do a
> > "wake_up_all" under a spinlock.  
> 
> No. I pulled in new timer code in and had to redo this part of RT.
> 
> While doing so I noticed that we drop the base lock during timer
> invocations and so it could be possible that we have two invocations
> of del_timer_sync() on a timer on the same "base" (one after the
> other). This is patch #1.
> 
> After that I saw that we do the wake up under the base lock but there
> is no reason for it. So here is patch #2.
> 
> Patch #1 is something that could happen in theory and I did not run in
> any problem.
>

OK, so patch 2 was just discovered by reviewing code?

-- Steve