linux-kernel - Re: [PATCH] printk: Correctly handle preemption in console

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: Windows password security audit tool. GUI, reports in PDF.

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

Date:   Wed, 18 Jan 2017 16:21:41 +0900
From:   Sergey Senozhatsky <sergey.senozhatsky.work@...il.com>
To:     Petr Mladek <pmladek@...e.com>
Cc:     Steven Rostedt <rostedt@...dmis.org>,
        Sergey Senozhatsky <sergey.senozhatsky@...il.com>,
        Tetsuo Handa <penguin-kernel@...ove.SAKURA.ne.jp>,
        Peter Zijlstra <peterz@...radead.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        Jiri Slaby <jslaby@...e.cz>, linux-fbdev@...r.kernel.org,
        linux-kernel@...r.kernel.org,
        ergey Senozhatsky <sergey.senozhatsky.work@...il.com>
Subject: Re: [PATCH] printk: Correctly handle preemption in console_unlock()

On (01/18/17 14:45), Sergey Senozhatsky wrote:
[..]
> 
> there is a function that clears @console_may_schedule out of
> console_sem scope - console_flush_on_panic().
> so I *may be* can think about a worst case scenario of race
> condition between
> 	console_flush_on_panic()->console_may_schedule = 0 on panic CPU
> and
> 	console_unlock()->console_may_schedule = 1 from CPU that panic CPU
> failed to stop (smp_send_stop() can return with secondary CPUs still being
> online).

what I mean, is that we can have, let's say, 2 CPUs spinning in
console_unlock(), both with @console_may_schedule == 1 (because secondary
CPU restores global @console_may_schedule value). now, suppose, we have
misbehaving scheduler (well, we are in panic after all). secondary CPU
will cond_resched() and may be lockup somewhere in the scheduler. which is
fine, we don't care about that secondary CPU anyway. but the same can happen
to panic CPU as well.

what do you think?

	-ss