Andy Green wrote:
> I guess what is behind it is a race with timer resume.

It actually seems to be a little bit weirder. The call stack looks
like this:

s3c24xx_i2s_trigger
soc_pcm_trigger
snd_pcm_do_resume
snd_pcm_action_single
snd_pcm_action_lock_irq
snd_pcm_common_ioctl1

The interrupts simply get turned off in snd_pcm_action_lock_irq,
and this is seems to be the only path leading there.

Now the mystery is why we actually ever saw this work :-)

- Werner

Reply via email to