Re: [GIT PULL FOR v6.18] TI and Cadence CSI2RX drivers changes

[Hans Verkuil <hverkuil+cisco@kernel.org>]

On 15/09/2025 10:44, Laurent Pinchart wrote:
> On Mon, Sep 15, 2025 at 10:24:34AM +0200, Hans Verkuil wrote:
>> On 12/09/2025 13:21, Jai Luthra wrote:
>>> On Fri, Sep 12, 2025 at 12:17:17PM +0200, Hans Verkuil wrote:
>>>> On 12/09/2025 11:55, Jai Luthra wrote:
>>>>> Hi Hans, Mauro,
>>>>>
>>>>> The following changes since commit ecba852dc9f4993f4f894ea1f352564560e19a3e:
>>>>>
>>>>>   media: redrat3: use int type to store negative error codes (2025-09-11 08:46:10 +0200)
>>>>>
>>>>> are available in the Git repository at:
>>>>>
>>>>>   https://gitlab.freedesktop.org/linux-media/users/jluthra.git tags/for-6.18-ti-j721e-csi2rx-signed
>>>>>
>>>>> for you to fetch changes up to c694e74c651e79838e817a8c6644dc72cf80540d:
>>>>>
>>>>>   media: ti: j721e-csi2rx: Support system suspend using pm_notifier (2025-09-12 14:37:56 +0530)
>>>>>
>>>>> ----------------------------------------------------------------
>>>>> TI and Cadence CSI2RX driver changes:
>>>>>
>>>>> - Multistream support for Cadence and TI CSI2RX
>>>>> - Runtime PM support for Cadence and TI CSI2RX
>>>>> - System suspend and resume support for TI CSI2RX
>>>>> - Misc cleanups
>>>>
>>>> This should go through Sakari most likely, possibly Laurent, seeing that this is CSI
>>>> related. Several patches do not even have a Reviewed-by from Sakari or Laurent.
>>>>
>>>> I can't take this, I'm afraid.
>>>>
>>>> I'm not sure why we get a PR from you, was there some misunderstanding?
>>>
>>> Apologies. I had a look at other "GIT PULL" messages on linux-media and 
>>> inferred that any driver maintainer can send a pull request for media tree.
>>
>> Ah, no. It's media maintainers that do that. For small subsystems you would post
>> patches and the subsystem maintainer will (when ready) merge them and send a PR
>> to Linus.
>>
>> For large subsystems like media you have multiple media maintainers, each responsible
>> for a part of the subsystem. They will review and prep a PR for the media maintainers
>> (me and Mauro) to pick up. The media subsystem is far too big for a single maintainer.
>>
>> Hopefully once we get the multi-committer up and running (pending doc changes) we can
>> do away with most of the PRs.
> 
> Is there an objection against pull requests from driver maintainers ?
> Assuming all the other rules and requirements are met (passing CI,
> having the expected R-b tags, ...), is there a reason why I would need
> to collect a series posted by Jai and then send a pull request, instead
> of having Jai send the pull request directly ?

If I get a PR from you, then I know that you, with your expertise, judged the
patches in the PR to be ready to be merged. Since it's from you I also can
trust that all the right tags are in place, that all the right people have
been pulled in if that was needed, etc.

If the PR would be from a driver maintainer, then I don't have that additional
level of trust and I will have to pay closer attention to the patches in the PR,
checking tags, checking that nothing was forgotten, etc. Precisely what happened
in this case where tags were missing.

As a general rule we don't accept PRs from non-media maintainers. It's something
we can discuss, though.

BTW, when I talk about 'trust', it's about trusting the PR author to know all the
procedures, and having the expertise to know when something is ready to be merged.
It's not about malicious code or anything like that.

So a driver maintainer will certainly know the driver code, and is expected to review
patches for that driver. But it takes time to gain the expertise to become a
media maintainer.

I hope this helps.

Regards,

	Hans

> 
>>> I looked at 
>>> https://docs.kernel.org/driver-api/media/maintainer-entry-profile.html but 
>>> that was not very informative.
>>>
>>> But yes, happy to wait for more reviews from Sakari and Laurent (or Tomi given 
>>> this is about streams support).
>>>
>>> Could the process and who-owns-what sub-subsystem be better documented in 
>>> docs.kernel.org or linuxtv.org? I am aware of the proposed multi-committer 
>>> model, but until that is finalized even the current process is hard to 
>>> understand for a newbie :-)
>>
>> It's here (link below), but I see that it is out-of-date: it's missing Nicolas Dufresne for
>> codecs and Bryan O'Donoghue for Qualcomm drivers.
>>
>> https://linuxtv.org/downloads/v4l-dvb-apis-new/driver-api/maintainer-entry-profile.html#media-maintainers
>>
>> In any case, your PR came in too late. After rc6 is released we only merge fixes as a
>> general rule. It's good to keep that in mind for the next kernel cycle.
>>
>> If your PR contains fixes that you believe are important for v6.18, then discuss that
>> with Sakari and he can post a PR with those fixes.
>>
>>>>> ----------------------------------------------------------------
>>>>> Changhuang Liang (1):
>>>>>       media: cadence: csi2rx: Support runtime PM
>>>>>
>>>>> Jai Luthra (9):
>>>>>       dt-bindings: media: ti,j721e-csi2rx-shim: Support 32 dma chans
>>>>>       media: ti: j721e-csi2rx: separate out device and context
>>>>>       media: ti: j721e-csi2rx: add a subdev for the core device
>>>>>       media: ti: j721e-csi2rx: add support for processing virtual channels
>>>>>       media: cadence: csi2rx: add multistream support
>>>>>       media: ti: j721e-csi2rx: add multistream support
>>>>>       media: ti: j721e-csi2rx: Submit all available buffers
>>>>>       media: ti: j721e-csi2rx: Support runtime suspend
>>>>>       media: ti: j721e-csi2rx: Support system suspend using pm_notifier
>>>>>
>>>>> Pratyush Yadav (4):
>>>>>       media: ti: j721e-csi2rx: prepare SHIM code for multiple contexts
>>>>>       media: ti: j721e-csi2rx: allocate DMA channel based on context index
>>>>>       media: ti: j721e-csi2rx: get number of contexts from device tree
>>>>>       media: cadence: csi2rx: add get_frame_desc wrapper
>>>>>
>>>>> Rishikesh Donadkar (2):
>>>>>       media: ti: j721e-csi2rx: Remove word size alignment on frame width
>>>>>       media: ti: j721e-csi2rx: Change the drain architecture for multistream
>>>>>
>>>>>  Documentation/devicetree/bindings/media/ti,j721e-csi2rx-shim.yaml |   39 ++++++-
>>>>>  drivers/media/platform/cadence/Kconfig                            |    1 +
>>>>>  drivers/media/platform/cadence/cdns-csi2rx.c                      |  491 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++---------------------
>>>>>  drivers/media/platform/ti/Kconfig                                 |    1 +
>>>>>  drivers/media/platform/ti/j721e-csi2rx/j721e-csi2rx.c             | 1167 +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++--------------------------------------------------
>>>>>  5 files changed, 1300 insertions(+), 399 deletions(-)
>