[v2,0/6] ALSA: hda: Unify codec construction

A follow up to the recent HDAudio fixes series [1]. Given the recently
reported regression [2], before the page fault occurring on codec
shutdown can be fixed, codec construction procedure needs to be updated
for skylake and sof-intel drivers. Drivers: pci-hda and avs need no
changes - already making use of snd_hda_codec_device_init().

As subject touches code used by the sof-driver, additional review has
been conducted on thesofproject/linux [3].

Changes in v2:

- dropped snd_hda_ext_core <-> snd_hda_codec dependency by calling
  snd_hda_codec_device_init() directly in skylake and sof drivers probe
  enumeration routines, as suggested by Takashi
- skylake/sof portion of the change has been split into two separate
  patches

- new functions that aim to replace hdac_ext codec init & exit
  functionality are added first - for skylake and sof drivers both
- third patch in the series now combines the "field -> pointer" change
  for hdac_hda_priv->codec plus the codec-enumeration adjustments for
  skylake and sof drivers
  Both above are here to keep git bisect happy, as suggested by Pierre

[1]: https://lore.kernel.org/alsa-devel/20220706120230.427296-7-cezary.rojewski@intel.com/
[2]: https://lore.kernel.org/alsa-devel/3c40df55-3aee-1e08-493b-7b30cd84dc00@linux.intel.com/
[3]: https://github.com/thesofproject/linux/pull/3775

Cezary Rojewski (6):
  ASoC: Intel: Skylake: Introduce HDA codec init and exit routines
  ASoC: SOF: Intel: Introduce HDA codec init and exit routines
  ASoC: Intel: Drop hdac_ext usage for codec device creation
  ALSA: hda: Always free codec on the device release
  ALSA: hda: Remove codec init and exit routines
  ALSA: hda: Fix page fault in snd_hda_codec_shutdown()

 include/sound/hda_codec.h                    |  2 -
 include/sound/hdaudio_ext.h                  |  3 --
 sound/hda/ext/hdac_ext_bus.c                 | 53 -------------------
 sound/pci/hda/hda_codec.c                    | 49 ++++++++---------
 sound/soc/codecs/hdac_hda.c                  | 26 ++++-----
 sound/soc/codecs/hdac_hda.h                  |  2 +-
 sound/soc/intel/boards/hda_dsp_common.c      |  2 +-
 sound/soc/intel/boards/skl_hda_dsp_generic.c |  2 +-
 sound/soc/intel/skylake/skl.c                | 53 ++++++++++++++-----
 sound/soc/sof/intel/hda-codec.c              | 55 ++++++++++++++------
 10 files changed, 113 insertions(+), 134 deletions(-)

On Mon, Aug 15, 2022 at 07:42:24PM +0200, Cezary Rojewski wrote:
> To make snd_hda_codec_device_init() the only constructor for struct
> hda_codec instances remaining tasks are:

Acked-by: Mark Brown <broonie@kernel.org>

On 2022-08-15 7:42 PM, Cezary Rojewski wrote:
> A follow up to the recent HDAudio fixes series [1]. Given the recently
> reported regression [2], before the page fault occurring on codec
> shutdown can be fixed, codec construction procedure needs to be updated
> for skylake and sof-intel drivers. Drivers: pci-hda and avs need no
> changes - already making use of snd_hda_codec_device_init().

...

> Cezary Rojewski (6):
>    ASoC: Intel: Skylake: Introduce HDA codec init and exit routines
>    ASoC: SOF: Intel: Introduce HDA codec init and exit routines
>    ASoC: Intel: Drop hdac_ext usage for codec device creation
>    ALSA: hda: Always free codec on the device release
>    ALSA: hda: Remove codec init and exit routines
>    ALSA: hda: Fix page fault in snd_hda_codec_shutdown()


I've received a message yesterday that patches: 1/6, 3/6 and 6/6 were 
not delivered to recipient: alsa-devel@alsa-projet.org. None else is 
listed there. lore.kernel.org also shows only four letters (3 + 
cover-letter) rather than seven. Given Mark's Ack on patch 3/6 it makes 
me believe that indeed just the alsa-devel did not received some of the 
messages.

What should I do in such situation? Is the server admin about to 
"recover them from the void" or should I resend the entire series as: 
"[RESEND PATCH v2]" as see how it goes this time?


Regards,
Czarek

On 2022-08-16 11:59 AM, Takashi Iwai wrote:
> On Tue, 16 Aug 2022 10:15:44 +0200,
> Cezary Rojewski wrote:

...

>>> Cezary Rojewski (6):
>>>     ASoC: Intel: Skylake: Introduce HDA codec init and exit routines
>>>     ASoC: SOF: Intel: Introduce HDA codec init and exit routines
>>>     ASoC: Intel: Drop hdac_ext usage for codec device creation
>>>     ALSA: hda: Always free codec on the device release
>>>     ALSA: hda: Remove codec init and exit routines
>>>     ALSA: hda: Fix page fault in snd_hda_codec_shutdown()
>>
>>
>> I've received a message yesterday that patches: 1/6, 3/6 and 6/6 were
>> not delivered to recipient: alsa-devel@alsa-projet.org. None else is
>> listed there. lore.kernel.org also shows only four letters (3 +
>> cover-letter) rather than seven. Given Mark's Ack on patch 3/6 it
>> makes me believe that indeed just the alsa-devel did not received some
>> of the messages.
>>
>> What should I do in such situation? Is the server admin about to

Meant to say 'able' instead of 'about'.

>> "recover them from the void" or should I resend the entire series as:
>> "[RESEND PATCH v2]" as see how it goes this time?
> 
> I guess the full resubmission would be better.


Should I apply Mark's ACKs for patches 2/6 and 3/6 in the 
about-to-be-resent series?

On Tue, 16 Aug 2022 12:30:58 +0200,
Mark Brown wrote:
> 
> On Tue, Aug 16, 2022 at 12:08:12PM +0200, Cezary Rojewski wrote:
> > On 2022-08-16 11:59 AM, Takashi Iwai wrote:
> 
> > > > "recover them from the void" or should I resend the entire series as:
> > > > "[RESEND PATCH v2]" as see how it goes this time?
> 
> > > I guess the full resubmission would be better.
> 
> Note that if you do whatever upset the server again it'll happen again.

If this happens, we need to investigate.  alsa-devel ML is maintained
outside vger by Jaroslav personally.  Let's see...


Takashi