Bug #1974
closed
[PATCH] sound/pcm: we can wait when mallocing in feed_rate_init
Added by rumcic almost 14 years ago.
Updated almost 14 years ago.
Description
After a few days of uptime and playing movies/listening to music I was not able
to open /dev/dsp anymore (only already existing /dev/dsp0.X devices worked, no
new ones could be opened).
The problem is in M_NOWAIT flag for the kmallocs in feed_rate_init(). We can
wait at that point and my system has been running with this patch for almost a
week and still sound functioned normally.
--
Please do not CC me, since I already receive everything from these MLs.
Regards,
Rumko
Files
On Sat, Jan 29, 2011 at 12:43:25AM +0100, Rumko wrote:
After a few days of uptime and playing movies/listening to music I was not able
to open /dev/dsp anymore (only already existing /dev/dsp0.X devices worked, no
new ones could be opened).
The problem is in M_NOWAIT flag for the kmallocs in feed_rate_init(). We can
wait at that point and my system has been running with this patch for almost a
week and still sound functioned normally.
This fixes my sound problem on my laptop.
Thanks!
Peter
:On Sat, Jan 29, 2011 at 12:43:25AM +0100, Rumko wrote:
:> After a few days of uptime and playing movies/listening to music I was not able
:> to open /dev/dsp anymore (only already existing /dev/dsp0.X devices worked, no
:> new ones could be opened).
:> The problem is in M_NOWAIT flag for the kmallocs in feed_rate_init(). We can
:> wait at that point and my system has been running with this patch for almost a
:> week and still sound functioned normally.
:
:This fixes my sound problem on my laptop.
:
:Thanks!
:Peter
I have committed the patch plus I have gone through every single
kmalloc in /usr/src/sys/dev/sound and fixed the M_NOWAITs -> M_WAITOKs.
-Matt
Matthew Dillon
<dillon@backplane.com>
Fixed in 77bcca08d8bbe55e472046972138522b0652c79a.
Also available in: Atom
PDF