| Message ID | 1589776238-23877-1-git-send-email-brent.lu@intel.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | [v2] ALSA: pcm: fix incorrect hw_base increase | expand |
On Mon, 18 May 2020 06:30:38 +0200, Brent Lu wrote: > > There is a corner case that ALSA keeps increasing the hw_ptr but DMA > already stop working/updating the position for a long time. > > In following log we can see the position returned from DMA driver does > not move at all but the hw_ptr got increased at some point of time so > snd_pcm_avail() will return a large number which seems to be a buffer > underrun event from user space program point of view. The program > thinks there is space in the buffer and fill more data. > > [ 418.510086] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 4096 avail 12368 > [ 418.510149] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 6910 avail 9554 > ... > [ 418.681052] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 15102 avail 1362 > [ 418.681130] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 16464 avail 0 > [ 418.726515] sound pcmC0D5p: pos 96 hw_ptr 16464 appl_ptr 16464 avail 16368 > > This is because the hw_base will be increased by runtime->buffer_size > frames unconditionally if the hw_ptr is not updated for over half of > buffer time. As the hw_base increases, so does the hw_ptr increased > by the same number. > > The avail value returned from snd_pcm_avail() could exceed the limit > (buffer_size) easily becase the hw_ptr itself got increased by same > buffer_size samples when the corner case happens. In following log, > the buffer_size is 16368 samples but the avail is 21810 samples so > CRAS server complains about it. > > [ 418.851755] sound pcmC0D5p: pos 96 hw_ptr 16464 appl_ptr 27390 avail 5442 > [ 418.926491] sound pcmC0D5p: pos 96 hw_ptr 32832 appl_ptr 27390 avail 21810 > > cras_server[1907]: pcm_avail returned frames larger than buf_size: > sof-glkda7219max: :0,5: 21810 > 16368 > > By updating runtime->hw_ptr_jiffies each time the HWSYNC is called, > the hw_base will keep the same when buffer stall happens at long as > the interval between each HWSYNC call is shorter than half of buffer > time. > > Following is a log captured by a patched kernel. The hw_base/hw_ptr > value is fixed in this corner case and user space program should be > aware of the buffer stall and handle it. > > [ 293.525543] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 4096 avail 12368 > [ 293.525606] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 6880 avail 9584 > [ 293.525975] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 10976 avail 5488 > [ 293.611178] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 15072 avail 1392 > [ 293.696429] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 16464 avail 0 > ... > [ 381.139517] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 16464 avail 0 > > Signed-off-by: Brent Lu <brent.lu@intel.com> Thanks, applied now with Reviewed-by tag from Jaroslav. I also put Cc to stable, as it can fix the actual issues. Takashi
diff --git a/sound/core/pcm_lib.c b/sound/core/pcm_lib.c index 872a852..d531e1b 100644 --- a/sound/core/pcm_lib.c +++ b/sound/core/pcm_lib.c @@ -433,6 +433,7 @@ static int snd_pcm_update_hw_ptr0(struct snd_pcm_substream *substream, no_delta_check: if (runtime->status->hw_ptr == new_hw_ptr) { + runtime->hw_ptr_jiffies = curr_jiffies; update_audio_tstamp(substream, &curr_tstamp, &audio_tstamp); return 0; }
There is a corner case that ALSA keeps increasing the hw_ptr but DMA already stop working/updating the position for a long time. In following log we can see the position returned from DMA driver does not move at all but the hw_ptr got increased at some point of time so snd_pcm_avail() will return a large number which seems to be a buffer underrun event from user space program point of view. The program thinks there is space in the buffer and fill more data. [ 418.510086] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 4096 avail 12368 [ 418.510149] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 6910 avail 9554 ... [ 418.681052] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 15102 avail 1362 [ 418.681130] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 16464 avail 0 [ 418.726515] sound pcmC0D5p: pos 96 hw_ptr 16464 appl_ptr 16464 avail 16368 This is because the hw_base will be increased by runtime->buffer_size frames unconditionally if the hw_ptr is not updated for over half of buffer time. As the hw_base increases, so does the hw_ptr increased by the same number. The avail value returned from snd_pcm_avail() could exceed the limit (buffer_size) easily becase the hw_ptr itself got increased by same buffer_size samples when the corner case happens. In following log, the buffer_size is 16368 samples but the avail is 21810 samples so CRAS server complains about it. [ 418.851755] sound pcmC0D5p: pos 96 hw_ptr 16464 appl_ptr 27390 avail 5442 [ 418.926491] sound pcmC0D5p: pos 96 hw_ptr 32832 appl_ptr 27390 avail 21810 cras_server[1907]: pcm_avail returned frames larger than buf_size: sof-glkda7219max: :0,5: 21810 > 16368 By updating runtime->hw_ptr_jiffies each time the HWSYNC is called, the hw_base will keep the same when buffer stall happens at long as the interval between each HWSYNC call is shorter than half of buffer time. Following is a log captured by a patched kernel. The hw_base/hw_ptr value is fixed in this corner case and user space program should be aware of the buffer stall and handle it. [ 293.525543] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 4096 avail 12368 [ 293.525606] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 6880 avail 9584 [ 293.525975] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 10976 avail 5488 [ 293.611178] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 15072 avail 1392 [ 293.696429] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 16464 avail 0 ... [ 381.139517] sound pcmC0D5p: pos 96 hw_ptr 96 appl_ptr 16464 avail 0 Signed-off-by: Brent Lu <brent.lu@intel.com> --- sound/core/pcm_lib.c | 1 + 1 file changed, 1 insertion(+)