mbox series

[00/22] drm/i915: ELD precompute and readout

Message ID 20221011170011.17198-1-ville.syrjala@linux.intel.com (mailing list archive)
Headers show
Series drm/i915: ELD precompute and readout | expand

Message

Ville Syrjälä Oct. 11, 2022, 4:59 p.m. UTC
From: Ville Syrjälä <ville.syrjala@linux.intel.com>

There was some discussion around ELD precompute, so I
decided to have a quick look at hooking that up. Unfortunately
the i915 audio code is a bit of a mess so ended up with
a patchbomb of cleanups. Sorry about that.

The actually interesting stuff is at the end of the series.
The precumpute+readot+state checker is pretty self explanatory
stuff for the most part.

But I think we need to decide what to do with the hardware
ELD buffer in general. It's totally busted atm on HSW 
(and I'd expecpt BDW as well), but we had no idea since
we had no readout+state checker for it. 

So do we try to salvage it (I guess to mainly act as some
kind of "did we enable audio correctly?" canary) or do we
just stop programming it outright? And on which platforms
could we do that?

Cc: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Cc: Takashi Iwai <tiwai@suse.de>

Ville Syrjälä (22):
  drm/i915/audio: s/dev_priv/i915/
  drm/i915/audio: Nuke leftover ROUNDING_FACTOR
  drm/i915/audio: Remove CL/BLC audio stuff
  drm/i915/audio: Exract struct ilk_audio_regs
  drm/i915/audio: Use REG_BIT() & co.
  drm/i915/audio: Unify register bit naming
  drm/i915/audio: Protect singleton register with a lock
  drm/i915/audio: Nuke intel_eld_uptodate()
  drm/i915/audio: Read ELD buffer size from hardware
  drm/i915/audio: Make sure we write the whole ELD buffer
  drm/i915/audio: Use u32* for ELD
  drm/i915/audio: Use intel_de_rmw() for most audio registers
  drm/i915/audio: Split "ELD valid" vs. audio PD on hsw+
  drm/i915/audio: Do the vblank waits
  drm/i915/audio: Precompute the ELD
  drm/i915/audio: Hardware ELD readout
  drm/i915/sdvo: Extract intel_sdvo_has_audio()
  drm/i915/sdvo: Precompute the ELD
  drm/i915/sdvo: Do ELD hardware readout
  drm/i915/audio: Hook up ELD into the state checker
  drm/i915/audio: Include ELD in the state dump
  hax: drm/i915/audio: Make HSW hardware ELD buffer sort of work

 drivers/gpu/drm/i915/display/g4x_dp.c         |   2 +
 drivers/gpu/drm/i915/display/g4x_hdmi.c       |   2 +
 drivers/gpu/drm/i915/display/intel_audio.c    | 792 ++++++++++--------
 drivers/gpu/drm/i915/display/intel_audio.h    |   7 +
 .../gpu/drm/i915/display/intel_audio_regs.h   |  88 +-
 .../drm/i915/display/intel_crtc_state_dump.c  |  17 +
 drivers/gpu/drm/i915/display/intel_ddi.c      |   2 +
 drivers/gpu/drm/i915/display/intel_display.c  |  43 +
 .../drm/i915/display/intel_display_types.h    |   2 +
 drivers/gpu/drm/i915/display/intel_dp.c       |   4 +-
 drivers/gpu/drm/i915/display/intel_hdmi.c     |   4 +-
 drivers/gpu/drm/i915/display/intel_sdvo.c     |  54 +-
 12 files changed, 608 insertions(+), 409 deletions(-)

Comments

Jani Nikula Oct. 11, 2022, 5:39 p.m. UTC | #1
On Tue, 11 Oct 2022, Ville Syrjala <ville.syrjala@linux.intel.com> wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
>
> There was some discussion around ELD precompute, so I
> decided to have a quick look at hooking that up. Unfortunately
> the i915 audio code is a bit of a mess so ended up with
> a patchbomb of cleanups. Sorry about that.
>
> The actually interesting stuff is at the end of the series.
> The precumpute+readot+state checker is pretty self explanatory
> stuff for the most part.
>
> But I think we need to decide what to do with the hardware
> ELD buffer in general. It's totally busted atm on HSW 
> (and I'd expecpt BDW as well), but we had no idea since
> we had no readout+state checker for it. 
>
> So do we try to salvage it (I guess to mainly act as some
> kind of "did we enable audio correctly?" canary) or do we
> just stop programming it outright? And on which platforms
> could we do that?

Quickly eyeballed through it all, overall looks good, but didn't do
detailed review. Ack.

Cc: Vinod too.


BR,
Jani.


>
> Cc: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
> Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
> Cc: Takashi Iwai <tiwai@suse.de>
>
> Ville Syrjälä (22):
>   drm/i915/audio: s/dev_priv/i915/
>   drm/i915/audio: Nuke leftover ROUNDING_FACTOR
>   drm/i915/audio: Remove CL/BLC audio stuff
>   drm/i915/audio: Exract struct ilk_audio_regs
>   drm/i915/audio: Use REG_BIT() & co.
>   drm/i915/audio: Unify register bit naming
>   drm/i915/audio: Protect singleton register with a lock
>   drm/i915/audio: Nuke intel_eld_uptodate()
>   drm/i915/audio: Read ELD buffer size from hardware
>   drm/i915/audio: Make sure we write the whole ELD buffer
>   drm/i915/audio: Use u32* for ELD
>   drm/i915/audio: Use intel_de_rmw() for most audio registers
>   drm/i915/audio: Split "ELD valid" vs. audio PD on hsw+
>   drm/i915/audio: Do the vblank waits
>   drm/i915/audio: Precompute the ELD
>   drm/i915/audio: Hardware ELD readout
>   drm/i915/sdvo: Extract intel_sdvo_has_audio()
>   drm/i915/sdvo: Precompute the ELD
>   drm/i915/sdvo: Do ELD hardware readout
>   drm/i915/audio: Hook up ELD into the state checker
>   drm/i915/audio: Include ELD in the state dump
>   hax: drm/i915/audio: Make HSW hardware ELD buffer sort of work
>
>  drivers/gpu/drm/i915/display/g4x_dp.c         |   2 +
>  drivers/gpu/drm/i915/display/g4x_hdmi.c       |   2 +
>  drivers/gpu/drm/i915/display/intel_audio.c    | 792 ++++++++++--------
>  drivers/gpu/drm/i915/display/intel_audio.h    |   7 +
>  .../gpu/drm/i915/display/intel_audio_regs.h   |  88 +-
>  .../drm/i915/display/intel_crtc_state_dump.c  |  17 +
>  drivers/gpu/drm/i915/display/intel_ddi.c      |   2 +
>  drivers/gpu/drm/i915/display/intel_display.c  |  43 +
>  .../drm/i915/display/intel_display_types.h    |   2 +
>  drivers/gpu/drm/i915/display/intel_dp.c       |   4 +-
>  drivers/gpu/drm/i915/display/intel_hdmi.c     |   4 +-
>  drivers/gpu/drm/i915/display/intel_sdvo.c     |  54 +-
>  12 files changed, 608 insertions(+), 409 deletions(-)
Borah, Chaitanya Kumar Oct. 14, 2022, 9:03 a.m. UTC | #2
> -----Original Message-----
> From: Ville Syrjala <ville.syrjala@linux.intel.com>
> Sent: Tuesday, October 11, 2022 10:30 PM
> To: intel-gfx@lists.freedesktop.org
> Cc: Borah, Chaitanya Kumar <chaitanya.kumar.borah@intel.com>; Kai
> Vehmanen <kai.vehmanen@linux.intel.com>; Takashi Iwai <tiwai@suse.de>
> Subject: [PATCH 00/22] drm/i915: ELD precompute and readout
> 
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> There was some discussion around ELD precompute, so I decided to have a
> quick look at hooking that up. Unfortunately the i915 audio code is a bit of a
> mess so ended up with a patchbomb of cleanups. Sorry about that.
> 
> The actually interesting stuff is at the end of the series.
> The precumpute+readot+state checker is pretty self explanatory stuff for the
> most part.
> 
> But I think we need to decide what to do with the hardware ELD buffer in
> general. It's totally busted atm on HSW (and I'd expecpt BDW as well), but
> we had no idea since we had no readout+state checker for it.
> 
> So do we try to salvage it (I guess to mainly act as some kind of "did we
> enable audio correctly?" canary) or do we just stop programming it outright?
> And on which platforms could we do that?
> 

Hello Ville,

I have gone through the patches and they look good. However, there is one aspect that is still not clear for me(may be I have missed something!)
The changes does not touch the callback i915_audio_component_get_eld() which is actually used by the audio driver to access the ELD. So we are still sending out an "non-precomputed" eld stored in the connector
structure. Should we be passing eld data from crtc_state instead here?

Regards,

Chaitanya

> Cc: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
> Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
> Cc: Takashi Iwai <tiwai@suse.de>
> 
> Ville Syrjälä (22):
>   drm/i915/audio: s/dev_priv/i915/
>   drm/i915/audio: Nuke leftover ROUNDING_FACTOR
>   drm/i915/audio: Remove CL/BLC audio stuff
>   drm/i915/audio: Exract struct ilk_audio_regs
>   drm/i915/audio: Use REG_BIT() & co.
>   drm/i915/audio: Unify register bit naming
>   drm/i915/audio: Protect singleton register with a lock
>   drm/i915/audio: Nuke intel_eld_uptodate()
>   drm/i915/audio: Read ELD buffer size from hardware
>   drm/i915/audio: Make sure we write the whole ELD buffer
>   drm/i915/audio: Use u32* for ELD
>   drm/i915/audio: Use intel_de_rmw() for most audio registers
>   drm/i915/audio: Split "ELD valid" vs. audio PD on hsw+
>   drm/i915/audio: Do the vblank waits
>   drm/i915/audio: Precompute the ELD
>   drm/i915/audio: Hardware ELD readout
>   drm/i915/sdvo: Extract intel_sdvo_has_audio()
>   drm/i915/sdvo: Precompute the ELD
>   drm/i915/sdvo: Do ELD hardware readout
>   drm/i915/audio: Hook up ELD into the state checker
>   drm/i915/audio: Include ELD in the state dump
>   hax: drm/i915/audio: Make HSW hardware ELD buffer sort of work
> 
>  drivers/gpu/drm/i915/display/g4x_dp.c         |   2 +
>  drivers/gpu/drm/i915/display/g4x_hdmi.c       |   2 +
>  drivers/gpu/drm/i915/display/intel_audio.c    | 792 ++++++++++--------
>  drivers/gpu/drm/i915/display/intel_audio.h    |   7 +
>  .../gpu/drm/i915/display/intel_audio_regs.h   |  88 +-
>  .../drm/i915/display/intel_crtc_state_dump.c  |  17 +
>  drivers/gpu/drm/i915/display/intel_ddi.c      |   2 +
>  drivers/gpu/drm/i915/display/intel_display.c  |  43 +
>  .../drm/i915/display/intel_display_types.h    |   2 +
>  drivers/gpu/drm/i915/display/intel_dp.c       |   4 +-
>  drivers/gpu/drm/i915/display/intel_hdmi.c     |   4 +-
>  drivers/gpu/drm/i915/display/intel_sdvo.c     |  54 +-
>  12 files changed, 608 insertions(+), 409 deletions(-)
> 
> --
> 2.35.1
Jani Nikula Oct. 14, 2022, 9:13 a.m. UTC | #3
On Fri, 14 Oct 2022, "Borah, Chaitanya Kumar" <chaitanya.kumar.borah@intel.com> wrote:
>> -----Original Message-----
>> From: Ville Syrjala <ville.syrjala@linux.intel.com>
>> Sent: Tuesday, October 11, 2022 10:30 PM
>> To: intel-gfx@lists.freedesktop.org
>> Cc: Borah, Chaitanya Kumar <chaitanya.kumar.borah@intel.com>; Kai
>> Vehmanen <kai.vehmanen@linux.intel.com>; Takashi Iwai <tiwai@suse.de>
>> Subject: [PATCH 00/22] drm/i915: ELD precompute and readout
>> 
>> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
>> 
>> There was some discussion around ELD precompute, so I decided to have a
>> quick look at hooking that up. Unfortunately the i915 audio code is a bit of a
>> mess so ended up with a patchbomb of cleanups. Sorry about that.
>> 
>> The actually interesting stuff is at the end of the series.
>> The precumpute+readot+state checker is pretty self explanatory stuff for the
>> most part.
>> 
>> But I think we need to decide what to do with the hardware ELD buffer in
>> general. It's totally busted atm on HSW (and I'd expecpt BDW as well), but
>> we had no idea since we had no readout+state checker for it.
>> 
>> So do we try to salvage it (I guess to mainly act as some kind of "did we
>> enable audio correctly?" canary) or do we just stop programming it outright?
>> And on which platforms could we do that?
>> 
>
> Hello Ville,
>
> I have gone through the patches and they look good. However, there is one aspect that is still not clear for me(may be I have missed something!)
> The changes does not touch the callback i915_audio_component_get_eld() which is actually used by the audio driver to access the ELD. So we are still sending out an "non-precomputed" eld stored in the connector
> structure. Should we be passing eld data from crtc_state instead here?

Good catch, I missed that in my review.

BR,
Jani.


>
> Regards,
>
> Chaitanya
>
>> Cc: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
>> Cc: Kai Vehmanen <kai.vehmanen@linux.intel.com>
>> Cc: Takashi Iwai <tiwai@suse.de>
>> 
>> Ville Syrjälä (22):
>>   drm/i915/audio: s/dev_priv/i915/
>>   drm/i915/audio: Nuke leftover ROUNDING_FACTOR
>>   drm/i915/audio: Remove CL/BLC audio stuff
>>   drm/i915/audio: Exract struct ilk_audio_regs
>>   drm/i915/audio: Use REG_BIT() & co.
>>   drm/i915/audio: Unify register bit naming
>>   drm/i915/audio: Protect singleton register with a lock
>>   drm/i915/audio: Nuke intel_eld_uptodate()
>>   drm/i915/audio: Read ELD buffer size from hardware
>>   drm/i915/audio: Make sure we write the whole ELD buffer
>>   drm/i915/audio: Use u32* for ELD
>>   drm/i915/audio: Use intel_de_rmw() for most audio registers
>>   drm/i915/audio: Split "ELD valid" vs. audio PD on hsw+
>>   drm/i915/audio: Do the vblank waits
>>   drm/i915/audio: Precompute the ELD
>>   drm/i915/audio: Hardware ELD readout
>>   drm/i915/sdvo: Extract intel_sdvo_has_audio()
>>   drm/i915/sdvo: Precompute the ELD
>>   drm/i915/sdvo: Do ELD hardware readout
>>   drm/i915/audio: Hook up ELD into the state checker
>>   drm/i915/audio: Include ELD in the state dump
>>   hax: drm/i915/audio: Make HSW hardware ELD buffer sort of work
>> 
>>  drivers/gpu/drm/i915/display/g4x_dp.c         |   2 +
>>  drivers/gpu/drm/i915/display/g4x_hdmi.c       |   2 +
>>  drivers/gpu/drm/i915/display/intel_audio.c    | 792 ++++++++++--------
>>  drivers/gpu/drm/i915/display/intel_audio.h    |   7 +
>>  .../gpu/drm/i915/display/intel_audio_regs.h   |  88 +-
>>  .../drm/i915/display/intel_crtc_state_dump.c  |  17 +
>>  drivers/gpu/drm/i915/display/intel_ddi.c      |   2 +
>>  drivers/gpu/drm/i915/display/intel_display.c  |  43 +
>>  .../drm/i915/display/intel_display_types.h    |   2 +
>>  drivers/gpu/drm/i915/display/intel_dp.c       |   4 +-
>>  drivers/gpu/drm/i915/display/intel_hdmi.c     |   4 +-
>>  drivers/gpu/drm/i915/display/intel_sdvo.c     |  54 +-
>>  12 files changed, 608 insertions(+), 409 deletions(-)
>> 
>> --
>> 2.35.1
>