Re: WARNING in initialize_timer

From: Takashi Iwai
Date: Thu Nov 30 2017 - 04:15:11 EST


On Thu, 30 Nov 2017 06:59:01 +0100,
syzbot wrote:
>
> Hello,
>
> syzkaller hit the following crash on
> 43570f0383d6d5879ae585e6c3cf027ba321546f
> git://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/master
> compiler: gcc (GCC) 7.1.1 20170620
> .config is attached
> Raw console output is attached.
>
> Unfortunately, I don't have any reproducer for this bug yet.
>
>
> binder: 4519:4521 ioctl 40046205 2 returned -22
> binder: 4519:4521 ioctl c0106403 20008ff0 returned -22
> WARNING: CPU: 0 PID: 4528 at sound/core/seq/seq_timer.c:358
> initialize_timer+0x22d/0x290 sound/core/seq/seq_timer.c:358

This must be a spurious WARN_ON() when a slave timer is used while the
master is freed.

I'm going to queue the patch below.

#syz fix: ALSA: seq: Remove spurious WARN_ON() at timer check


thanks,

Takashi

-- 8< --
From: Takashi Iwai <tiwai@xxxxxxx>
Subject: [PATCH] ALSA: seq: Remove spurious WARN_ON() at timer check

The use of snd_BUG_ON() in ALSA sequencer timer may lead to a spurious
WARN_ON() when a slave timer is deployed as its backend and a
corresponding master timer stops meanwhile. The symptom was triggered
by syzkaller spontaneously.

Since the NULL timer is valid there, rip off snd_BUG_ON().

Reported-by: syzbot <syzkaller@xxxxxxxxxxxxxxxx>
Cc: <stable@xxxxxxxxxxxxxxx>
Signed-off-by: Takashi Iwai <tiwai@xxxxxxx>
---
sound/core/seq/seq_timer.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/sound/core/seq/seq_timer.c b/sound/core/seq/seq_timer.c
index 37d9cfbc29f9..b80985fbc334 100644
--- a/sound/core/seq/seq_timer.c
+++ b/sound/core/seq/seq_timer.c
@@ -355,7 +355,7 @@ static int initialize_timer(struct snd_seq_timer *tmr)
unsigned long freq;

t = tmr->timeri->timer;
- if (snd_BUG_ON(!t))
+ if (!t)
return -EINVAL;

freq = tmr->preferred_resolution;
--
2.15.0