2012/5/18 cee1 fykcee1@gmail.com:
2012/5/18 cee1 fykcee1@gmail.com:
在 2012年5月18日星期五，Clemens Ladisch 写道：
cee1 wrote:
We've found a way the can reproduce the problem more quickly on our product:
- do audio playback
- alsamixer, select 'Master', press 'm' and hold for a while.
Then release 'm', playback corrupts(sounds similar to http://dev.lemote.com/files/upload/software/PA-apc/corrupted_sound.ogg).
I've written a small program[1] that will let PA mute/unmute repeatedly.
I find the problem is something to do with rewind_safeguard, If I adjust rewind_safeguard to eighth of the original value, it can also be easily reproduced through pa-mute-test.c on the AMD turion 64 x2 board in ubuntu 12.04: --- a/src/modules/alsa/alsa-sink.c +++ b/src/modules/alsa/alsa-sink.c @@ -86,8 +86,8 @@
#define VOLUME_ACCURACY (PA_VOLUME_NORM/100) /* don't require volume adjustments to be perfectly correct. do
-#define DEFAULT_REWIND_SAFEGUARD_BYTES (256U) /* 1.33ms @48kHz, we'll never rewind less than this */ -#define DEFAULT_REWIND_SAFEGUARD_USEC (1330) /* 1.33ms, depending on channels/rate/sample we may rewind more +#define DEFAULT_REWIND_SAFEGUARD_BYTES (32U) /* 1.33ms @48kHz, we'll never rewind less than this */ +#define DEFAULT_REWIND_SAFEGUARD_USEC (166) /* 1.33ms, depending on channels/rate/sample we may rewind more t
struct userdata { pa_core *core;
Should we calculate limit_nbytes considering tsched_watermark? e.g.(which works on our platform) @@ -1614,7 +1614,7 @@ static int process_rewind(struct userdata *u) { unused_nbytes = (size_t) unused * u->frame_size;
/* make sure rewind doesn't go too far, can cause issues with DMAs */ - unused_nbytes += u->rewind_safeguard; + unused_nbytes += u->tsched_watermark;
if (u->hwbuf_size > unused_nbytes) limit_nbytes = u->hwbuf_size - unused_nbytes;
---- [1]. See pa-mute-test.c in attachment, using build.sh to build it.
Regards, -- cee1