[RFC PATCH] drm: bridge: cdns-mhdp8546: Fix possible null pointer dereference

[Jayesh Choudhary]

Hello Tomi,

On 28/01/25 11:27, Jayesh Choudhary wrote:
> Hello Tomi, Alexander,
> 
> On 24/01/25 13:38, Alexander Stein wrote:
>> Hi,
>>
>> Am Donnerstag, 23. Januar 2025, 17:20:34 CET schrieb Tomi Valkeinen:
>>> Hi,
>>>
>>> On 16/01/2025 13:16, Jayesh Choudhary wrote:
>>>> For the cases we have DRM_BRIDGE_ATTACH_NO_CONNECTOR flag set,
>>>
>>> Any idea if any other platform than K3 is using this driver? tidss
>>> supports DRM_BRIDGE_ATTACH_NO_CONNECTOR, so if K3 is the only user, we
>>> could drop the legacy !DRM_BRIDGE_ATTACH_NO_CONNECTOR case. Which would
>>> remove quite a bit of code, I think, and make the driver a bit more easy
>>> to understand (although I think it could use a major cleanup...).
>>

I cannot access the atomic state in this worker thread so I cannot reach
the connector using drm_atomic_get_new_connector_for_encoder.
I can't find any other way to do this. So I am rolling v2 for RFC while
addressing your other comments and fixing it in a different manner.
Please let me know if that is acceptable.

v2: https://lore.kernel.org/all/20250521073237.366463-1-j-choudhary@ti.com/

Warm Regards,
Jayesh



>> FYI: Not directly using it, but patch series [1] is at least touching
>> this file.
> 
> I can't see any other platform using this.
> No one uses compatible "cdns,mhdp8546" and only K3 devices uses the
> wrapper compatible "ti,j721e-mhdp8546"
> Let me post next RFC version with legacy !DRM_BRIDGE_ATTACH_NO_CONNECTOR
> dropped.
> 
> And the mentioned series only touches mailbox access function to move
> then to a common helper file. (So independent change)
> 
> 
>>
>> Best regards,
>> Alexander
>>
>> [1] https://lore.kernel.org/all/cover.1734340233.git.Sandor.yu@nxp.com/
>>
> 
> Warm Regards,
> Jayesh