Re: [REGRESSION] HDMI monitor not working on Radxa Rock 5B after phy rockchip samsung hdptx HDMI 2.1 FRL patchset

[Cristian Ciocaltea]

On 2/18/26 3:22 AM, Cristian Ciocaltea wrote:
> On 2/18/26 2:52 AM, Cristian Ciocaltea wrote:
>> Hi Thomas,
> [...]
> 
>>
>> Sorry, I somehow missed the following warning message, though it has been
>> already present in all the logs you've sent to me so far:
>>
>>   rockchip-hdptx-phy fed60000.phy: PLL locked by unknown consumer!
>>
>> That indicates the PHY has been preconfigured by an external component (e.g. the
>> bootloader), which is actually a scenario that I didn't verify.
>>
>> However, this just another way to expose a limitation of the current approach
>> for managing the TMDS character rate: done via the Common Clock Framework API
>> instead of the HDMI PHY configuration API.
>>
>> As a matter of fact, it was actually an item on my TODOs list for quite a while,
>> but blocked until recently due to several dependencies waiting to be merged
>> upstream.
>>
>> Hence I took the opportunity to finalize this task - please give the following
>> commits in my rk3588-hdmi-debug branch [2] a try:
> 
> I've just realized I introduced a regression while doing some cleanup work, hence
> please ignore this until further notice.

I think I got this working properly now, at least it passes all the tests I
could run.  The updated commits in [2] are:

  fa7cd1e75aaa ("phy: rockchip: samsung-hdptx: Fix rate recalculation for high bpc")
  8048db5544da ("phy: rockchip: samsung-hdptx: Exclusively use PHY config API for PLL changes")
  0085a382dfd0 ("[DEBUG] drm/rockchip: Add HDMI verbose logging")

Regards,
Cristian

[2] https://gitlab.collabora.com/cristicc/linux-next/-/commits/rk3588-hdmi-debug