diff options
author | Andreas Mohr <andi@lisas.de> | 2010-11-21 12:09:32 +0100 |
---|---|---|
committer | Takashi Iwai <tiwai@suse.de> | 2010-11-22 07:54:45 +0100 |
commit | 7974150c8524423f878e8269010e911c3cc7ddb8 (patch) | |
tree | 30b7116a13c2bd5b2252273a976c4997ec077829 /lib/bust_spinlocks.c | |
parent | 86cbbad2b6712fbd25c07a17e86b4345cee82c6d (diff) |
ALSA: azt3328: period bug fix (for PA), add missing ACK on stop timer
. Fix PulseAudio "ALSA driver bug" issue
(if we have two alternated areas within a 64k DMA buffer, then max
period size should obviously be 32k only).
Back references:
http://pulseaudio.org/wiki/AlsaIssues
http://fedoraproject.org/wiki/Features/GlitchFreeAudio
. In stop timer function, need to supply ACK in the timer control byte.
. Minor log output correction
When I did my first PA testing recently, the period size bug resulted
in quite precisely observeable half-period-based playback distortion.
PA-based operation is quite a bit more underrun-prone (despite its
zero-copy optimizations etc.) than raw ALSA with this rather spartan
sound hardware implementation on my puny Athlon.
Note that even with this patch, azt3328 still doesn't work for both
cases yet, PA tsched=0 and tsched
(on tsched=0 it will playback tiny fragments of periods, leading to tiny
stuttering sounds with some pauses in between, whereas with
timer-scheduled operation playback works fine - minus some quite increased
underrun trouble on PA vs. ALSA, that is).
Signed-off-by: Andreas Mohr <andi@lisas.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Diffstat (limited to 'lib/bust_spinlocks.c')
0 files changed, 0 insertions, 0 deletions