Re: [PATCH v2] drm/dp: move intel_dp_vsc_sdp_pack() to generic helper

[Abhinav Kumar <quic_abhinavk@quicinc.com>]

On 2/20/2024 11:20 AM, Dmitry Baryshkov wrote:
> On Tue, 20 Feb 2024 at 21:05, Dmitry Baryshkov
> <dmitry.baryshkov@linaro.org> wrote:
>>
>> On Tue, 20 Feb 2024 at 20:53, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
>>>
>>>
>>>
>>> On 2/20/2024 10:49 AM, Dmitry Baryshkov wrote:
>>>> On Thu, 15 Feb 2024 at 21:08, Abhinav Kumar <quic_abhinavk@quicinc.com> wrote:
>>>>>
>>>>> intel_dp_vsc_sdp_pack() can be re-used by other DRM drivers as well.
>>>>> Lets move this to drm_dp_helper to achieve this.
>>>>>
>>>>> changes in v2:
>>>>>           - rebased on top of drm-tip
>>>>>
>>>>> Acked-by: Dmitry Baryshkov <dmitry.baryshkov@linaro.org>
>>>>
>>>> v1 had an explicit comment before the ack:
>>>>
>>>
>>> Yes, I remember the comment. I did not make any changes to v2 other than
>>> just rebasing it on drm-tip to get the ack from i915 folks.
>>>
>>>>>      From my side, with the promise of the size fixup.
>>>>
>>>> However I observe neither a second patch removing the size argument
>>>> nor it being dropped as a part of this patch.
>>>>
>>>
>>> Yes, now that in v2 we got the ack for this patch, I can spin a v3 with
>>> the addition of the next patch to remove the size OR as another series
>>> so as to not block the main series which needs this patch.
>>>
>>> I would prefer the latter.
>>
>> It doesn't work this way. The comment should have been fixed for v2.
> 
> This probably deserves some explanation. Currently there is only one
> user of this function. So it is easy to fix it. Once there are several
> users, you have to fix all of them at the same time, patching
> different drm subtrees. That complicates the life of maintainers.
> 

Yes, understood. Its easier to fix it now if its really needed.

Actually, I think the reason the size was passed was to make sure
a valid struct dp_sdp *sdp was being passed.

If we drop the size, we need to have a if (!sdp) check as there is a 
memset followed by dereference.

So maybe, if we want to keep this API protected, its not too bad to have?