summaryrefslogtreecommitdiffstats
path: root/sound/isa
diff options
context:
space:
mode:
authorSteve Soule <sts11dbxr@gmail.com>2009-12-14 11:06:03 -0700
committerJaroslav Kysela <perex@perex.cz>2009-12-15 09:31:31 +0100
commitf74890277a196949e4004fe2955e1d4fb3930f98 (patch)
treef6681a24fc8536dd0b62e87ec3779308079b82af /sound/isa
parentf40542532e96dda5506eb76badea322f2ae4731c (diff)
downloadlinux-f74890277a196949e4004fe2955e1d4fb3930f98.tar.gz
linux-f74890277a196949e4004fe2955e1d4fb3930f98.tar.bz2
linux-f74890277a196949e4004fe2955e1d4fb3930f98.zip
ALSA: ac97_codec - increase timeout for analog sections to 5 second
I have a Soundblaster 16PCI. For many years, alsa has had a bug where not all of the card's controls are detected (many alsa versions, many kernel versions). In particular, Master Playback Volume is usually not detected, and so I get no sound or extremely faint sound. The problem has always been inconsistent: sometimes all of the controls are detected correctly, and sometimes a partial set is detected. It works correctly about 10% of the time. Finally, I got around to tracking down the problem. When the driver fails, it prints the kernel message "AC'97 0 analog subsections not ready". This message is generated from the function snd_ac97_mixer() in ac97_codec.c. The message indicates that the card failed to come back after reset within the time limit. The time limit is 120 milliseconds. I tried increasing the time limit to 1 second, and found that this made the driver work about 70% of the time. I tried increasing it to 5 seconds, and it now seems to work 100% of the time. I expect that this change would be completely harmless for existing cards that work, and would only introduce additional delay for cards that do not work. ALSA bug#4032. Signed-off-by: Steve Soule <sts11dbxr@gmail.com> Signed-off-by: Jaroslav Kysela <perex@perex.cz>
Diffstat (limited to 'sound/isa')
0 files changed, 0 insertions, 0 deletions