[PATCH v4 2/3] drm/bridge: analogix_dp: Add validation for samsung,lane-count property

Damon Ding damon.ding at rock-chips.com
Mon Jun 1 05:24:42 PDT 2026


On 6/1/2026 8:02 PM, Luca Ceresoli wrote:
> Hello Damon,
> 
> On 2026-06-01 10:59:27+08:00, Damon Ding wrote:
>> Hi Luca,
>>
>> On 5/30/2026 9:38 PM, Luca Ceresoli wrote:
>>
>>> On Sat May 30, 2026 at 3:33 PM CEST, Luca Ceresoli wrote:
>>>
>>> Meh, messed up with 'b4 review' :-/ Apologies
>>>
>>> "This sashiko report" [0] was about an enum being signed, so '== 0' could miss
>>> negative numbers coming from bogus DT values higher than 1^31.
>>
>> Ah, I agree. It would be better to add a new inline function to validate
>> that the lane count passed from DT is exactly 1, 2, or 4, just as
>> Sashiko suggested.
>>
>>>>
>>>
>>> And this was about "Additionally, does this check inadvertently allow 3,
>>> which is an invalid DisplayPort lane count?"
>>
>> Yes, the DisplayPort specification only allows lane counts of 1, 2, or
>> 4. I did miss the check for the invalid value 3 in the current code.
>>
>> (BTW: I did not find a common helper function in drm_dp_helper.h to
>> validate valid DP lane counts (1, 2, 4). I'm not sure if it would be
>> better to add a generic lane count validation function to the DP helper
>> library instead, to avoid duplicating the same functionality across
>> individual DP drivers. Perhaps other DP experts could provide some
>> advice on this.)
> 
> A bool function in common code, to be reused by all DP drivers, would be
> good.
> 
> 
> 

Yes, will do in v5.

Best regards,
Damon




More information about the linux-arm-kernel mailing list