Re: [PATCH] drm/i915/dp: ALPM init to be done after DPCD init

["Murthy, Arun R" <arun.r.murthy@intel.com>]

On 02-03-2026 13:30, Hogander, Jouni wrote:
> On Mon, 2026-03-02 at 12:50 +0530, Murthy, Arun R wrote:
>> On 02-03-2026 12:14, Hogander, Jouni wrote:
>>> On Mon, 2026-03-02 at 09:06 +0530, Arun R Murthy wrote:
>>>> In ALPM init, the DPCD register ALPM_CAPABILITIES are to be read.
>>>> This
>>>> read has to happen after the DPCD init.
>>>>
>>>> Signed-off-by: Arun R Murthy <arun.r.murthy@intel.com>
>>>> ---
>>>>    drivers/gpu/drm/i915/display/intel_dp.c | 4 ++--
>>>>    1 file changed, 2 insertions(+), 2 deletions(-)
>>>>
>>>> diff --git a/drivers/gpu/drm/i915/display/intel_dp.c
>>>> b/drivers/gpu/drm/i915/display/intel_dp.c
>>>> index 613559d11133..e60802b0b8c9 100644
>>>> --- a/drivers/gpu/drm/i915/display/intel_dp.c
>>>> +++ b/drivers/gpu/drm/i915/display/intel_dp.c
>>>> @@ -6948,8 +6948,6 @@ static bool intel_edp_init_connector(struct
>>>> intel_dp *intel_dp,
>>>>    	 */
>>>>    	intel_hpd_enable_detection(encoder);
>>>>    
>>>> -	intel_alpm_init(intel_dp);
>>>> -
>>>>    	/* Cache DPCD and EDID for edp. */
>>>>    	has_dpcd = intel_edp_init_dpcd(intel_dp, connector);
>>>>    
>>>> @@ -6961,6 +6959,8 @@ static bool intel_edp_init_connector(struct
>>>> intel_dp *intel_dp,
>>>>    		goto out_vdd_off;
>>>>    	}
>>>>    
>>>> +	intel_alpm_init(intel_dp);
>>> This will cause problems to PSR/Panel Replay. intel_psr_init_dpcd
>>> is
>>> relying on ALPM dpcd init is done first.
>>>
>>> BR,
>>> Jouni Högander
>> I think this will also have to be moved down.
>>
>> For eDP intel_psr_init_dpcd() is called in intel_edp_init_dpcd.
>> So can
>> we move intel_alpm_init inside intel_edp_init_dpcd() ? Just before
>> calling intel_psr_init_dpcd !
>> Also rename intel_alpm_init() to intel_alpm_init_dpcd() ?
> Animesh has some ongoing work related:
>
> https://patchwork.freedesktop.org/patch/687711/?series=156417&rev=4
I think we can still go ahead with those. The one that I am trying to 
fix is move alpm_init inside intel_edp_init_dpcd() just before 
intel_psr_init_dpcd().

Will float a patch with the said changes.

Thanks for the review.


Thanks and Regards,
Arun R Murthy
--------------------

>
> I think it should be ok to move and rename it. That mutex_init within
> should be handled somehow.
>
> BR,
> Jouni Högander
>
>> Thanks and Regards,
>> Arun R Murthy
>> --------------------
>>
>>>> +
>>>>    	/*
>>>>    	 * VBT and straps are liars. Also check HPD as that
>>>> seems
>>>>    	 * to be the most reliable piece of information
>>>> available.