At Tue, 11 May 2010 10:44:17 +0200 (CEST), Jaroslav Kysela wrote:
On Tue, 11 May 2010, Takashi Iwai wrote:
At Tue, 11 May 2010 10:28:50 +0200 (CEST), Jaroslav Kysela wrote:
On Mon, 10 May 2010, Jaroslav Kysela wrote:
Hi,
I would like to ask HDA gurus to check this patch (I will include it to my tree once acked). The patch uses WALLCLK from HDA chips (marked as required in the HDA specification) to check for bogus - too early - interrups which confuses the upper PCM layer (sound skipping issues).
More details about patch testing on problematic hardware can be found at:
No objections received, so I pushed this code to my master/devel branches.
Heh, that was quick, I had no time to review :)
The change looks good to me. Thanks for your fix (and rebasing).
One thing is that you removed bdl_pos_adj=0 check. This was there explicitly to avoid the delayed irq handling no matter whether the value is correct or not. I guess it's safer to take it back, or give any other way for user to avoid the delayed irq handling.
I see. But I need the pos % period_bytes check too. What about this change?
With bdl_pos_adj==0, it should return 1 (non-error), but otherwise this should be OK, I guess.
thanks,
Takashi
diff --git a/sound/pci/hda/hda_intel.c b/sound/pci/hda/hda_intel.c index 0a6c55b..6674841 100644 --- a/sound/pci/hda/hda_intel.c +++ b/sound/pci/hda/hda_intel.c @@ -1910,9 +1910,11 @@ static int azx_position_ok(struct azx *chip, struct azx_dev *azx_dev)
if (WARN_ONCE(!azx_dev->period_bytes, "hda-intel: zero azx_dev->period_bytes"))
return 0; /* this shouldn't happen! */
/* this shouldn't happen! */return bdl_pos_adj[chip->dev_index] ? 0 : -1;
return 0; /* NG - it's below the period boundary */
/* NG - it's below the period boundary */return bdl_pos_adj[chip->dev_index] ? 0 : -1;
Jaroslav Kysela perex@perex.cz Linux Kernel Sound Maintainer ALSA Project, Red Hat, Inc.