[PATCH 2/3] ASoC: soc-pcm: add soc_pcm_hw_update_chan()

Kuninori Morimoto kuninori.morimoto.gx at renesas.com
Tue Feb 16 08:26:04 CET 2021


Hi Kai

> I think we might still have issues if we have multiple CPU DAIs per
> runtime and dpcm_init_runtime_hw() is called multiple times. With the 
> fixup, the limits are taken from the last CPU DAI.
> 
> But if you look at original code, the same issues seems to be there
> (rate and channels taken from the stream directly with no consideration
> of already set values). Only exception is runtime->hw.formats, which
> is handled differently:

Actually, when I posted the patch, I removed 1 line which I thought not needed.
But it seems it was necessary...

Current fixup patch always initialize hw at dpcm_set_fe_runtime()'s loop,
but I guess we need is initialize once.

How about this patch ?
It reverts current fixup, and initialize hw once at dpcm_set_fe_runtime().

-----------
diff --git a/sound/soc/soc-pcm.c b/sound/soc/soc-pcm.c
index 37094aeff440..14d85ca1e435 100644
--- a/sound/soc/soc-pcm.c
+++ b/sound/soc/soc-pcm.c
@@ -1531,10 +1531,12 @@ static void dpcm_init_runtime_hw(struct snd_pcm_runtime *runtime,
 {
 	struct snd_pcm_hardware *hw = &runtime->hw;
 
-	soc_pcm_hw_init(hw);
 	soc_pcm_hw_update_rate(hw, stream);
 	soc_pcm_hw_update_chan(hw, stream);
-	soc_pcm_hw_update_format(hw, stream);
+	if (runtime->hw.formats)
+		runtime->hw.formats &= stream->formats;
+	else
+		runtime->hw.formats = stream->formats;
 }
 
 static void dpcm_runtime_merge_format(struct snd_pcm_substream *substream,
@@ -1662,10 +1664,13 @@ static void dpcm_runtime_merge_rate(struct snd_pcm_substream *substream,
 static void dpcm_set_fe_runtime(struct snd_pcm_substream *substream)
 {
 	struct snd_pcm_runtime *runtime = substream->runtime;
+	struct snd_pcm_hardware *hw = &runtime->hw;
 	struct snd_soc_pcm_runtime *rtd = asoc_substream_to_rtd(substream);
 	struct snd_soc_dai *cpu_dai;
 	int i;
 
+	soc_pcm_hw_init(hw);
+
 	for_each_rtd_cpu_dais(rtd, i, cpu_dai) {
 		/*
 		 * Skip CPUs which don't support the current stream


Thank you for your help !!

Best regards
---
Kuninori Morimoto


More information about the Alsa-devel mailing list