mbox series

[GIT,PULL,FOR,6.9] Various Mediatek VCodec fixes

Message ID 20240315132549.vbhui74yxjg2t36v@basti-XPS-13-9310
State New
Headers show
Series [GIT,PULL,FOR,6.9] Various Mediatek VCodec fixes | expand

Pull-request

https://gitlab.collabora.com/sebastianfricke/linux.git tags/for-6.9-mediatek-various-fixes

Message

Sebastian Fricke March 15, 2024, 1:25 p.m. UTC
Hey Hans & Mauro,

this set contains 3 fixes for two kinds of Kernel panic, one fix to be
able to access the last element of a list and one fix to be able to
access the high bits of a register.

Please pull.

The following changes since commit b14257abe7057def6127f6fb2f14f9adc8acabdb:

   media: rcar-isp: Disallow unbind of devices (2024-03-07 16:35:13 +0100)

are available in the Git repository at:

   https://gitlab.collabora.com/sebastianfricke/linux.git tags/for-6.9-mediatek-various-fixes

for you to fetch changes up to 27821568df7da25b1f454cf5461ad4b428df536e:

   media: mediatek: vcodec: support 36 bits physical address (2024-03-15 14:14:08 +0100)

----------------------------------------------------------------
Various Fixes for Mediatek VCodec

----------------------------------------------------------------
Irui Wang (1):
       media: mediatek: vcodec: Handle VP9 superframe bitstream with 8 sub-frames

Nicolas Dufresne (1):
       media: mediatek: vcodec: Fix oops when HEVC init fails

Yunfei Dong (3):
       media: mediatek: vcodec: adding lock to protect decoder context list
       media: mediatek: vcodec: adding lock to protect encoder context list
       media: mediatek: vcodec: support 36 bits physical address

  .../media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c |  8 ++++----
  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c     |  5 +++++
  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h     |  2 ++
  .../mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c     |  2 +-
  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c |  2 +-
  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c | 11 ++++++-----
  .../mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c        |  4 ++--
  drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c  |  2 ++
  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c     |  5 +++++
  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h     |  2 ++
  drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c  |  2 ++
  11 files changed, 32 insertions(+), 13 deletions(-)

Comments

Hans Verkuil March 25, 2024, 9:30 a.m. UTC | #1
Hi Sebastian,

Are these fixes for v6.9 or can these be merged for v6.10?

Not all of these patches have a 'Fixes' tag, perhaps they should have it?

Regards,

	Hans

On 15/03/2024 2:25 pm, Sebastian Fricke wrote:
> Hey Hans & Mauro,
> 
> this set contains 3 fixes for two kinds of Kernel panic, one fix to be
> able to access the last element of a list and one fix to be able to
> access the high bits of a register.
> 
> Please pull.
> 
> The following changes since commit b14257abe7057def6127f6fb2f14f9adc8acabdb:
> 
>   media: rcar-isp: Disallow unbind of devices (2024-03-07 16:35:13 +0100)
> 
> are available in the Git repository at:
> 
>   https://gitlab.collabora.com/sebastianfricke/linux.git tags/for-6.9-mediatek-various-fixes
> 
> for you to fetch changes up to 27821568df7da25b1f454cf5461ad4b428df536e:
> 
>   media: mediatek: vcodec: support 36 bits physical address (2024-03-15 14:14:08 +0100)
> 
> ----------------------------------------------------------------
> Various Fixes for Mediatek VCodec
> 
> ----------------------------------------------------------------
> Irui Wang (1):
>       media: mediatek: vcodec: Handle VP9 superframe bitstream with 8 sub-frames
> 
> Nicolas Dufresne (1):
>       media: mediatek: vcodec: Fix oops when HEVC init fails
> 
> Yunfei Dong (3):
>       media: mediatek: vcodec: adding lock to protect decoder context list
>       media: mediatek: vcodec: adding lock to protect encoder context list
>       media: mediatek: vcodec: support 36 bits physical address
> 
>  .../media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c |  8 ++++----
>  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c     |  5 +++++
>  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h     |  2 ++
>  .../mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c     |  2 +-
>  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c |  2 +-
>  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c | 11 ++++++-----
>  .../mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c        |  4 ++--
>  drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c  |  2 ++
>  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c     |  5 +++++
>  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h     |  2 ++
>  drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c  |  2 ++
>  11 files changed, 32 insertions(+), 13 deletions(-)
>
Sebastian Fricke March 25, 2024, 3:44 p.m. UTC | #2
Hey Hans,

On 25.03.2024 10:30, Hans Verkuil wrote:
>Hi Sebastian,
>
>Are these fixes for v6.9 or can these be merged for v6.10?

I was hoping for them to land in 6.9 but if that doesn't work that is
fine for me as well.

>
>Not all of these patches have a 'Fixes' tag, perhaps they should have it?

No the tags are correct, do you want me to strictly only mix patches
with fixes tags or only patches without fixes tags?

>
>Regards,
>
>	Hans

Greetings,
Sebastian

>
>On 15/03/2024 2:25 pm, Sebastian Fricke wrote:
>> Hey Hans & Mauro,
>>
>> this set contains 3 fixes for two kinds of Kernel panic, one fix to be
>> able to access the last element of a list and one fix to be able to
>> access the high bits of a register.
>>
>> Please pull.
>>
>> The following changes since commit b14257abe7057def6127f6fb2f14f9adc8acabdb:
>>
>>   media: rcar-isp: Disallow unbind of devices (2024-03-07 16:35:13 +0100)
>>
>> are available in the Git repository at:
>>
>>   https://gitlab.collabora.com/sebastianfricke/linux.git tags/for-6.9-mediatek-various-fixes
>>
>> for you to fetch changes up to 27821568df7da25b1f454cf5461ad4b428df536e:
>>
>>   media: mediatek: vcodec: support 36 bits physical address (2024-03-15 14:14:08 +0100)
>>
>> ----------------------------------------------------------------
>> Various Fixes for Mediatek VCodec
>>
>> ----------------------------------------------------------------
>> Irui Wang (1):
>>       media: mediatek: vcodec: Handle VP9 superframe bitstream with 8 sub-frames
>>
>> Nicolas Dufresne (1):
>>       media: mediatek: vcodec: Fix oops when HEVC init fails
>>
>> Yunfei Dong (3):
>>       media: mediatek: vcodec: adding lock to protect decoder context list
>>       media: mediatek: vcodec: adding lock to protect encoder context list
>>       media: mediatek: vcodec: support 36 bits physical address
>>
>>  .../media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c |  8 ++++----
>>  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c     |  5 +++++
>>  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h     |  2 ++
>>  .../mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c     |  2 +-
>>  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c |  2 +-
>>  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c | 11 ++++++-----
>>  .../mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c        |  4 ++--
>>  drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c  |  2 ++
>>  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c     |  5 +++++
>>  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h     |  2 ++
>>  drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c  |  2 ++
>>  11 files changed, 32 insertions(+), 13 deletions(-)
>>
>
Hans Verkuil March 25, 2024, 4:25 p.m. UTC | #3
Hi Sebastian,

On 25/03/2024 4:44 pm, Sebastian Fricke wrote:
> Hey Hans,
> 
> On 25.03.2024 10:30, Hans Verkuil wrote:
>> Hi Sebastian,
>>
>> Are these fixes for v6.9 or can these be merged for v6.10?
> 
> I was hoping for them to land in 6.9 but if that doesn't work that is
> fine for me as well.
> 
>>
>> Not all of these patches have a 'Fixes' tag, perhaps they should have it?
> 
> No the tags are correct, do you want me to strictly only mix patches
> with fixes tags or only patches without fixes tags?

That's why I asked: if you want that all these patches go to v6.9, then I can
apply them to our fixes branch and they will be passed on to Linus for
6.9, but in that case it would be good to have Fixes tags.

If all patches are OK to be merged for 6.10, then I can just apply them
to our staging tree.

Finally, if some are for 6.9 and some for 6.10 (presumably those patches
without the Fixes tag), then I can split it up myself.

In any case, I need to know what you want.

Regards,

	Hans

> 
>>
>> Regards,
>>
>>     Hans
> 
> Greetings,
> Sebastian
> 
>>
>> On 15/03/2024 2:25 pm, Sebastian Fricke wrote:
>>> Hey Hans & Mauro,
>>>
>>> this set contains 3 fixes for two kinds of Kernel panic, one fix to be
>>> able to access the last element of a list and one fix to be able to
>>> access the high bits of a register.
>>>
>>> Please pull.
>>>
>>> The following changes since commit b14257abe7057def6127f6fb2f14f9adc8acabdb:
>>>
>>>   media: rcar-isp: Disallow unbind of devices (2024-03-07 16:35:13 +0100)
>>>
>>> are available in the Git repository at:
>>>
>>>   https://gitlab.collabora.com/sebastianfricke/linux.git tags/for-6.9-mediatek-various-fixes
>>>
>>> for you to fetch changes up to 27821568df7da25b1f454cf5461ad4b428df536e:
>>>
>>>   media: mediatek: vcodec: support 36 bits physical address (2024-03-15 14:14:08 +0100)
>>>
>>> ----------------------------------------------------------------
>>> Various Fixes for Mediatek VCodec
>>>
>>> ----------------------------------------------------------------
>>> Irui Wang (1):
>>>       media: mediatek: vcodec: Handle VP9 superframe bitstream with 8 sub-frames
>>>
>>> Nicolas Dufresne (1):
>>>       media: mediatek: vcodec: Fix oops when HEVC init fails
>>>
>>> Yunfei Dong (3):
>>>       media: mediatek: vcodec: adding lock to protect decoder context list
>>>       media: mediatek: vcodec: adding lock to protect encoder context list
>>>       media: mediatek: vcodec: support 36 bits physical address
>>>
>>>  .../media/platform/mediatek/vcodec/common/mtk_vcodec_fw_vpu.c |  8 ++++----
>>>  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.c     |  5 +++++
>>>  .../platform/mediatek/vcodec/decoder/mtk_vcodec_dec_drv.h     |  2 ++
>>>  .../mediatek/vcodec/decoder/vdec/vdec_hevc_req_multi_if.c     |  2 +-
>>>  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp8_if.c |  2 +-
>>>  .../media/platform/mediatek/vcodec/decoder/vdec/vdec_vp9_if.c | 11 ++++++-----
>>>  .../mediatek/vcodec/decoder/vdec/vdec_vp9_req_lat_if.c        |  4 ++--
>>>  drivers/media/platform/mediatek/vcodec/decoder/vdec_vpu_if.c  |  2 ++
>>>  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.c     |  5 +++++
>>>  .../platform/mediatek/vcodec/encoder/mtk_vcodec_enc_drv.h     |  2 ++
>>>  drivers/media/platform/mediatek/vcodec/encoder/venc_vpu_if.c  |  2 ++
>>>  11 files changed, 32 insertions(+), 13 deletions(-)
>>>
>>
Sebastian Fricke March 25, 2024, 5:37 p.m. UTC | #4
Hey Hans,
>>
>> On 25.03.2024 10:30, Hans Verkuil wrote:
>>> Hi Sebastian,
>>>
>>> Are these fixes for v6.9 or can these be merged for v6.10?
>>
>> I was hoping for them to land in 6.9 but if that doesn't work that is
>> fine for me as well.
>>
>>>
>>> Not all of these patches have a 'Fixes' tag, perhaps they should have it?
>>
>> No the tags are correct, do you want me to strictly only mix patches
>> with fixes tags or only patches without fixes tags?
>
>That's why I asked: if you want that all these patches go to v6.9, then I can
>apply them to our fixes branch and they will be passed on to Linus for
>6.9, but in that case it would be good to have Fixes tags.
>
>If all patches are OK to be merged for 6.10, then I can just apply them
>to our staging tree.
>
>Finally, if some are for 6.9 and some for 6.10 (presumably those patches
>without the Fixes tag), then I can split it up myself.
>
>In any case, I need to know what you want.

Alright, then please add all of them to 6.9, the two without the fixes
tags are not critical but that enable normal functionality, so I'd call
them important enough.

>
>Regards,
>
>	Hans

Greetings,
Sebastian