1 Oct
2009
1 Oct
'09
2:09 p.m.
On Thu, Oct 01, 2009 at 02:56:24PM +0300, Eero Nurkkala wrote:
Actually, in this case: (kernel 2.6.28)
That's rather old...
But if it didn't come from snd_soc_dapm_put_volsw(), but instead: close_delayed_work() -> snd_soc_dapm_stream_event() -> dapm_power_widgets() -> dapm_set_pga() then the codec->mutex is not taken, and the bug presented in the patch is out there, no?
Right, so close_delayed_work() needs to lock the card while it runs.