At Mon, 15 Jul 2013 15:12:07 +0200, Jaroslav Kysela wrote:
Date 15.7.2013 15:06, Takashi Iwai wrote:
Hi,
while reviewing another patch series, I stumbled on the snd_pcm_lock() call in an unlocked context. Then I started looking through the whole tree, and found a bunch of drivers doing it wrong, too. So here is a patch series to fix them.
[PATCH 01/10] ALSA: asihpi: Fix unlocked snd_pcm_stop() call [PATCH 02/10] ALSA: atiixp: Fix unlocked snd_pcm_stop() call [PATCH 03/10] ALSA: 6fire: Fix unlocked snd_pcm_stop() call [PATCH 04/10] ALSA: ua101: Fix unlocked snd_pcm_stop() call [PATCH 05/10] ALSA: usx2y: Fix unlocked snd_pcm_stop() call [PATCH 06/10] ALSA: pxa2xx: Fix unlocked snd_pcm_stop() call [PATCH 07/10] ASoC: atmel: Fix unlocked snd_pcm_stop() call [PATCH 08/10] ASoC: s6000: Fix unlocked snd_pcm_stop() call [PATCH 09/10] [media] saa7134: Fix unlocked snd_pcm_stop() call [PATCH 10/10] staging: line6: Fix unlocked snd_pcm_stop() call
Probably, it may be better to introduce 'snd_pcm_stop_dolock()' helper (inline?) in include/sound/pcm.h .
Yes, it'll be the next step. These patches are done in open-coded way so that they can be applied easily to each stable kernel.
thanks,
Takashi