[PATCH] ALSA: hda: fix trigger_tstamp_latched

[PATCH] ALSA: hda: fix trigger_tstamp_latched

[Jaroslav Kysela]

When the trigger_tstamp_latched flag is set, the PCM core code assumes that
the low-level driver handles the trigger timestamping itself. Ensure that
runtime->trigger_tstamp is always updated.

Buglink: https://github.com/alsa-project/alsa-lib/issues/387
Reported-by: Zeno Endemann <zeno.endemann@mailbox.org>
Signed-off-by: Jaroslav Kysela <perex@perex.cz>
---
 include/sound/hdaudio.h        | 2 +-
 sound/hda/hdac_stream.c        | 6 +++++-
 sound/pci/hda/hda_controller.c | 3 +--
 3 files changed, 7 insertions(+), 4 deletions(-)

diff --git a/include/sound/hdaudio.h b/include/sound/hdaudio.h
index 7e39d486374a..b098ceadbe74 100644
--- a/include/sound/hdaudio.h
+++ b/include/sound/hdaudio.h
@@ -590,7 +590,7 @@ void snd_hdac_stream_sync_trigger(struct hdac_stream *azx_dev, bool set,
 void snd_hdac_stream_sync(struct hdac_stream *azx_dev, bool start,
 			  unsigned int streams);
 void snd_hdac_stream_timecounter_init(struct hdac_stream *azx_dev,
-				      unsigned int streams);
+				      unsigned int streams, bool start);
 int snd_hdac_get_stream_stripe_ctl(struct hdac_bus *bus,
 				struct snd_pcm_substream *substream);
 
diff --git a/sound/hda/hdac_stream.c b/sound/hda/hdac_stream.c
index b53de020309f..0411a8fe9d6f 100644
--- a/sound/hda/hdac_stream.c
+++ b/sound/hda/hdac_stream.c
@@ -664,7 +664,7 @@ static void azx_timecounter_init(struct hdac_stream *azx_dev,
  * updated accordingly, too.
  */
 void snd_hdac_stream_timecounter_init(struct hdac_stream *azx_dev,
-				      unsigned int streams)
+				      unsigned int streams, bool start)
 {
 	struct hdac_bus *bus = azx_dev->bus;
 	struct snd_pcm_runtime *runtime = azx_dev->substream->runtime;
@@ -672,6 +672,9 @@ void snd_hdac_stream_timecounter_init(struct hdac_stream *azx_dev,
 	bool inited = false;
 	u64 cycle_last = 0;
 
+	if (!start)
+		goto skip;
+
 	list_for_each_entry(s, &bus->stream_list, list) {
 		if ((streams & (1 << s->index))) {
 			azx_timecounter_init(s, inited, cycle_last);
@@ -682,6 +685,7 @@ void snd_hdac_stream_timecounter_init(struct hdac_stream *azx_dev,
 		}
 	}
 
+skip:
 	snd_pcm_gettime(runtime, &runtime->trigger_tstamp);
 	runtime->trigger_tstamp_latched = true;
 }
diff --git a/sound/pci/hda/hda_controller.c b/sound/pci/hda/hda_controller.c
index 5d86e5a9c814..f3330b7e0fcf 100644
--- a/sound/pci/hda/hda_controller.c
+++ b/sound/pci/hda/hda_controller.c
@@ -275,8 +275,7 @@ static int azx_pcm_trigger(struct snd_pcm_substream *substream, int cmd)
 	spin_lock(&bus->reg_lock);
 	/* reset SYNC bits */
 	snd_hdac_stream_sync_trigger(hstr, false, sbits, sync_reg);
-	if (start)
-		snd_hdac_stream_timecounter_init(hstr, sbits);
+	snd_hdac_stream_timecounter_init(hstr, sbits, start);
 	spin_unlock(&bus->reg_lock);
 	return 0;
 }
-- 
2.46.0

Re: [PATCH] ALSA: hda: fix trigger_tstamp_latched

[Takashi Iwai]

On Wed, 02 Oct 2024 10:13:06 +0200,
Jaroslav Kysela wrote:
> 
> When the trigger_tstamp_latched flag is set, the PCM core code assumes that
> the low-level driver handles the trigger timestamping itself. Ensure that
> runtime->trigger_tstamp is always updated.
> 
> Buglink: https://github.com/alsa-project/alsa-lib/issues/387
> Reported-by: Zeno Endemann <zeno.endemann@mailbox.org>
> Signed-off-by: Jaroslav Kysela <perex@perex.cz>

Thanks, applied now.


Takashi