[PATCH v3 00/14] ASoC: qcom: q6dsp: few fixes and enhancements

Srinivas Kandagatla posted 14 patches 1 month, 1 week ago
There is a newer version of this series
.../sound/qcom,q6dsp-lpass-ports.h            |  12 ++
sound/soc/qcom/common.c                       |   6 +
sound/soc/qcom/common.h                       |   3 +
sound/soc/qcom/lpass.h                        |   3 +-
sound/soc/qcom/qdsp6/audioreach.c             |  60 ------
sound/soc/qcom/qdsp6/audioreach.h             |   5 +-
sound/soc/qcom/qdsp6/q6afe-dai.c              |   6 +
sound/soc/qcom/qdsp6/q6afe.c                  |   6 +
sound/soc/qcom/qdsp6/q6afe.h                  |   3 +-
sound/soc/qcom/qdsp6/q6apm-dai.c              | 101 ++++++++--
sound/soc/qcom/qdsp6/q6apm-lpass-dais.c       |  48 +++--
sound/soc/qcom/qdsp6/q6apm.c                  | 183 ++++++++++++------
sound/soc/qcom/qdsp6/q6apm.h                  |  21 +-
sound/soc/qcom/qdsp6/q6dsp-lpass-ports.c      |  76 ++++++++
sound/soc/qcom/qdsp6/topology.c               |   8 +-
15 files changed, 380 insertions(+), 161 deletions(-)
[PATCH v3 00/14] ASoC: qcom: q6dsp: few fixes and enhancements
Posted by Srinivas Kandagatla 1 month, 1 week ago
This patchset contains few fixes for the bugs hit during testing with
Monza EVK platform
- around array out of bounds access on dai ids which keep extending but
  the drivers seems to have hardcoded some numbers, fix this and clean
the mess up
- fix few issues discovered while trying to shut down dsp.
- flooding rpmsg with write requests due to not resetting queue pointer,
  fix this resetting the pointer in trigger stop.
- possible multiple graph opens which can result in open failures.

Apart from this few new enhancements to the dsp side
- add new LPI MI2S and senary dai entries
- handle pipewire and Displayport issues by moving graph start to
  trigger level, which should fix outstanding pipewire and DP issues on
Qualcomm SoCs.
- remove some unnessary loops in hot path
- support early memory map on DSP.


Thanks,
Srini

Changes since v2:
	- removed non-dt header update in bindings patch.
	- removed some leftover debug statements
	- updated function to void where it did not make sense to
	  return anything.

Changes since v1:
	- removed LPASS MAX PORT define from bindings
	- added few fixes found while testing dsp shutdown usecase on
	  Monaco.
	- replaced tab with space
	- added port range check in place at runtime to catch array out
	  of bounds in future.

Mohammad Rafi Shaik (2):
  ASoC: dt-bindings: qcom,q6dsp-lpass-ports: Add Senary MI2S port
  ASoC: qcom: q6dsp: Add Senary MI2S audio interface support

Srinivas Kandagatla (12):
  ASoC: qcom: q6apm: move component registration to unmanaged version
  ASoC: qcom: q6apm: remove child devices when apm is removed
  ASoC: qcom: qdsp6: topology: check widget type before accessing data
  ASoC: qcom: q6apm-lpass-dai: Fix multiple graph opens
  ASoC: qcom: q6apm-dai: reset queue ptr on trigger stop
  ASoC: dt-bindings: qcom: add LPASS LPI MI2S dai ids
  ASoC: qcom: qdsp6: lpass-ports: add support for LPASS LPI MI2S dais
  ASoC: qcom: common: validate cpu dai id during parsing
  ASoC: qcom: qdapm-lpass-dai: correct the error message
  ASoC: qcom: q6apm-lpass-dai: move graph start to trigger
  ASoC: qcom: qdsp6: remove search for module iid in hot path
  ASoC: qcom: q6apm: Add support for early buffer mapping on DSP

 .../sound/qcom,q6dsp-lpass-ports.h            |  12 ++
 sound/soc/qcom/common.c                       |   6 +
 sound/soc/qcom/common.h                       |   3 +
 sound/soc/qcom/lpass.h                        |   3 +-
 sound/soc/qcom/qdsp6/audioreach.c             |  60 ------
 sound/soc/qcom/qdsp6/audioreach.h             |   5 +-
 sound/soc/qcom/qdsp6/q6afe-dai.c              |   6 +
 sound/soc/qcom/qdsp6/q6afe.c                  |   6 +
 sound/soc/qcom/qdsp6/q6afe.h                  |   3 +-
 sound/soc/qcom/qdsp6/q6apm-dai.c              | 101 ++++++++--
 sound/soc/qcom/qdsp6/q6apm-lpass-dais.c       |  48 +++--
 sound/soc/qcom/qdsp6/q6apm.c                  | 183 ++++++++++++------
 sound/soc/qcom/qdsp6/q6apm.h                  |  21 +-
 sound/soc/qcom/qdsp6/q6dsp-lpass-ports.c      |  76 ++++++++
 sound/soc/qcom/qdsp6/topology.c               |   8 +-
 15 files changed, 380 insertions(+), 161 deletions(-)

-- 
2.47.3
Re: [PATCH v3 00/14] ASoC: qcom: q6dsp: few fixes and enhancements
Posted by Richard Acayan 1 month, 1 week ago
On Wed, Mar 04, 2026 at 01:06:58PM +0000, Srinivas Kandagatla wrote:
> This patchset contains few fixes for the bugs hit during testing with
> Monza EVK platform
> - around array out of bounds access on dai ids which keep extending but
>   the drivers seems to have hardcoded some numbers, fix this and clean
> the mess up
> - fix few issues discovered while trying to shut down dsp.
> - flooding rpmsg with write requests due to not resetting queue pointer,
>   fix this resetting the pointer in trigger stop.
> - possible multiple graph opens which can result in open failures.
> 
> Apart from this few new enhancements to the dsp side
> - add new LPI MI2S and senary dai entries

Is the LPI MI2S on Q6AFE? I don't see the AFE port ID.

It sounds different from INT_MI2S so I can rebase my patches to depend
on this series.
Re: [PATCH v3 00/14] ASoC: qcom: q6dsp: few fixes and enhancements
Posted by Srinivas Kandagatla 1 month, 1 week ago
On 3/5/26 1:48 PM, Richard Acayan wrote:
> On Wed, Mar 04, 2026 at 01:06:58PM +0000, Srinivas Kandagatla wrote:
>> This patchset contains few fixes for the bugs hit during testing with
>> Monza EVK platform
>> - around array out of bounds access on dai ids which keep extending but
>>   the drivers seems to have hardcoded some numbers, fix this and clean
>> the mess up
>> - fix few issues discovered while trying to shut down dsp.
>> - flooding rpmsg with write requests due to not resetting queue pointer,
>>   fix this resetting the pointer in trigger stop.
>> - possible multiple graph opens which can result in open failures.
>>
>> Apart from this few new enhancements to the dsp side
>> - add new LPI MI2S and senary dai entries
> 
> Is the LPI MI2S on Q6AFE? I don't see the AFE port ID.
I think the older SOCs did not have this I guess,

> 
> It sounds different from INT_MI2S so I can rebase my patches to depend
> on this series.
That would be wonderful
Sending out v4 soon!

--srini