Hi Morimoto-san,
soc_get_playback_capture() is now handling DPCM and normal comprehensively for playback/capture stream. We can use playback/capture_only flag instead of using dpcm_playback/capture. This patch replace these.
This patch no longer applies since Mark merged "[PATCH 00/21] ASoC: Intel: boards: updates for 6.10 - part1"
https://lore.kernel.org/linux-sound/20240325221059.206042-1-pierre-louis.bos...
The odds of additional conflicts are very high due to the number of queued changes in machine drivers, I really don't know how we're going to find the time to solve those conflicts and test experimental versions, while we're also busy checking the 6.9-rc1 transition.
In addition, I have a bit of heartburn with the concept of flipping the logic from a clear description of playback or capture supported, to an implicit 'both supported unless stated otherwise'. We are going to miss some cases and have regressions, it's pretty much a given.
And then there are cases such as the following examples
@@ -496,8 +494,6 @@ static struct snd_soc_dai_link cml_rt1011_rt5682_dailink[] = { */ .name = "SSP1-Codec", .id = 6,
.dpcm_playback = 1,.dpcm_capture = 1, /* Capture stream provides Feedback */
@@ -293,8 +291,6 @@ int sof_intel_board_set_ssp_amp_link(struct device *dev,
link->id = be_id; link->no_pcm = 1;
link->dpcm_capture = 1; /* feedback stream or firmware-generated echo reference */
link->dpcm_playback = 1;
return 0;
}
diff --git a/sound/soc/intel/boards/sof_da7219.c b/sound/soc/intel/boards/sof_da7219.c index 6eb5a6144e97..c7cca3e47ed9 100644 --- a/sound/soc/intel/boards/sof_da7219.c +++ b/sound/soc/intel/boards/sof_da7219.c @@ -296,8 +296,6 @@ static struct snd_soc_dai_link jsl_dais[] = { .id = 0, .ignore_pmdown_time = 1, .no_pcm = 1,
.dpcm_playback = 1,.dpcm_capture = 1, /* IV feedback */
where we lose comments on what the capture stream is. That's not so good, we added those comments on purpose to explain that the capture stream isn't a regular PCM format, I don't see a good reason to see them go?