[PATCH v7 6/7] drm/msm/hdmi: also send the SPD and HDMI Vendor Specific InfoFrames

[Abhinav Kumar]

On 2/7/2025 6:04 PM, Dmitry Baryshkov wrote:
> On Fri, Feb 07, 2025 at 05:31:20PM -0800, Abhinav Kumar wrote:
>>
>>
>> On 2/7/2025 4:27 PM, Dmitry Baryshkov wrote:
>>> Extend the driver to send SPD and HDMI Vendor Specific InfoFrames.
>>>
>>> While the HDMI block has special block to send HVS InfoFrame, use
>>> GENERIC0 block instead. VENSPEC_INFO registers pack frame data in a way
>>> that requires manual repacking in the driver, while GENERIC0 doesn't
>>> have such format requirements. The msm-4.4 kernel uses GENERIC0 to send
>>> HDR InfoFrame which we do not at this point anyway.
>>>
>>
>> True that GENERIC_0/1 packets can be used for any infoframe. But because we
>> have so many of them, thats why when there are dedicated registers for some
>> of them, we use them to save the GENERIC0 ones for others.
> 
> True
> 
>> Lets take a case where we want to send HVSIF, SPD and HDR together for the
>> same frame, then we run out as there are no HDR specific infoframe registers
>> we can use. Is the expectation that we will migrate to VENSPEC_INFO regs for
>> HVSIF when we add HDR support?
> 
> Most likely, yes. That would be a part of the HDR support. At the same
> time note, we can use GENERIC0 to send new HFVS InfoFrames defined in
> HDMI 2.x (once Linux gets support for that). At the same time we can not
> use VENSPEC_INFO to send those.
> 
> I can imagine that the driver will have to switch GENERIC1 between HDR
> (if required) and SPD (in all other cases).
> 

We dont have to use GENERIC0 for HFVS infoframes. We have dedicated 
HFVS_INFO registers for those.

>>
>> Also from a validation standpoint, I guess to really validate this change
>> you need an analyzer which decodes the HVSIF. So was this mostly sanity
>> tested at this pointed to make sure that the sink just comes up?
> 
> Vertex 2 dumps received HVS InfoFrame, so the InfoFrame contents has
> been validated (validated SPD, AUD, HVS and AVI frames).
> 

Yup, vertex2 validation is perfect for this!

Overall, I am fine with this,

Reviewed-by: Abhinav Kumar <quic_abhinavk at quicinc.com>