diff options
author | Andreas Mohr <andim2@users.sourceforge.net> | 2009-07-12 22:17:54 +0200 |
---|---|---|
committer | Takashi Iwai <tiwai@suse.de> | 2009-07-15 12:03:26 +0200 |
commit | 78df617acf83745908ae71f322e084284054ea66 (patch) | |
tree | f32d64d18cc18de52440740ce91be6f5ca5379dd /drivers/md/raid10.h | |
parent | dfbf9511155d3584b8747c935216077f46eb9a4f (diff) |
ALSA: azt3328: fix previous breakage, improve suspend, cleanups
- fix my previous codec activity breakage (_non-warned_ variable assignment
issue)
- convert suspend/resume to 32bit I/O access (I/O is painful; to improve
suspend/resume performance)
- change DEBUG_PLAY_REC to DEBUG_CODEC for consistency
- printk cleanup
- some logging improvements
- minor cleanup/improvements
The variable assignment issue above was a conditional assignment to the
call_function variable (this ended with the non-preinitialized variable
not getting assigned in some cases, thus a dangling stack value, yet gcc 4.3.3
unbelievably did _NOT_ warn about it in this case!!),
needed to change this into _always_ assigning the check result.
Practical result of this bug was that when shutting down
_either_ playback or capture, _both_ streams dropped dead :P
Tested, working (plus resume) and checkpatch.pl:ed on 2.6.30-rc5,
applies cleanly to 2.6.30 proper with my previous (committed)
patches applied.
Signed-off-by: Andreas Mohr <andi@lisas.de>
Signed-off-by: Takashi Iwai <tiwai@suse.de>
Diffstat (limited to 'drivers/md/raid10.h')
0 files changed, 0 insertions, 0 deletions