Re: [alsa-devel] [PATCH 0/3] ASoC: SOF: fix suspend ordering with runtime idle
[Cc: Takashi]
On 6/12/19 4:44 AM, Kai Vehmanen wrote:
Hi all, this series addresses issues with ordering of HDA codec and controller suspends in the runtime PM flows. The implemented logic for SOF is similar to what has been used by the Intel AZX HDA driver.
To implement this, first a fix is needed to hdac_hdmic ASoC codec driver. SOF framework also needs to be extended to allow SOF devices to implement a runtime_idle callback. Third, concrete implementation is in a separate patch for APL/CNL Intel hardware, for which strict ordering of codec-controller power down sequence needs to be maintained.
Looks good to me Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com
but of course we'd want feedback from Takashi and other folks more familiar with the HDaudio history. There will be additional patches in the coming days where we are catching up with issues that were fixed in the legacy driver but either missed or not known to the SOF developers.
As this extends the SOF device interface, Pierre asked me to send to the list for wider review. This series has been prereviewd at SOF github as: https://github.com/thesofproject/linux/pull/1003
Indeed we want the SOF core changes and dependencies (topology and hdaudio typically) to be reviewed on the alsa-devel mailing list. For Intel-specific fixes we still want to use GitHub and provide patches when we've hit a decent level of validation on prototype and production hardware, otherwise we'll increase the traffic on this list for no good reason.
Kai Vehmanen (3): ASoC: hdac_hdmi: report codec link up/down status to bus ASoC: SOF: add runtime idle callback ASoC: SOF: Intel: implement runtime idle for CNL/APL
sound/soc/codecs/hdac_hdmi.c | 2 ++ sound/soc/sof/intel/apl.c | 1 + sound/soc/sof/intel/cnl.c | 1 + sound/soc/sof/intel/hda-dsp.c | 13 +++++++++++++ sound/soc/sof/intel/hda.h | 1 + sound/soc/sof/pm.c | 8 ++++++++ sound/soc/sof/sof-acpi-dev.c | 2 +- sound/soc/sof/sof-pci-dev.c | 2 +- sound/soc/sof/sof-priv.h | 2 ++ 9 files changed, 30 insertions(+), 2 deletions(-)
On Wed, 12 Jun 2019 16:46:24 +0200, Pierre-Louis Bossart wrote:
[Cc: Takashi]
On 6/12/19 4:44 AM, Kai Vehmanen wrote:
Hi all, this series addresses issues with ordering of HDA codec and controller suspends in the runtime PM flows. The implemented logic for SOF is similar to what has been used by the Intel AZX HDA driver.
To implement this, first a fix is needed to hdac_hdmic ASoC codec driver. SOF framework also needs to be extended to allow SOF devices to implement a runtime_idle callback. Third, concrete implementation is in a separate patch for APL/CNL Intel hardware, for which strict ordering of codec-controller power down sequence needs to be maintained.
Looks good to me Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com
but of course we'd want feedback from Takashi and other folks more familiar with the HDaudio history. There will be additional patches in the coming days where we are catching up with issues that were fixed in the legacy driver but either missed or not known to the SOF developers.
All look good. I already replied to the first one, but feel free to take my ack for the rest, too:
Reviewed-by: Takashi Iwai tiwai@suse.de
thanks,
Takashi
Hi,
On Wed, 12 Jun 2019, Takashi Iwai wrote:
On Wed, 12 Jun 2019 16:46:24 +0200, Pierre-Louis Bossart wrote:
Looks good to me Reviewed-by: Pierre-Louis Bossart pierre-louis.bossart@linux.intel.com
All look good. I already replied to the first one, but feel free to take my ack for the rest, too:
Reviewed-by: Takashi Iwai tiwai@suse.de
thanks for the reviews!
Mark, one patch in this series conflicts with a patch in Pierre's series sent yesterday: ASoC: SOF: Intel: hda: release link DMA for paused streams during suspend -> in series "[PATCH 00/15] ASoC: SOF: Intel: HDaudio fixes and improvements"
We thought the runtime idle would take more time in review, so didn't send them in the same series, but now that the reviews seem positive, we have a potential conflict. You can drop the conflicting patch from my series, or I can rebase on top of Pierre's series and send V2 later.
Br, Kai
participants (3)
-
Kai Vehmanen
-
Pierre-Louis Bossart
-
Takashi Iwai