deliberations about the future of mdp5 (was: Re: [PATCH 1/2] drm/msm/dpu: fix mismatch between power and frequency)

[Konrad Dybcio]

On 3/2/26 3:27 PM, Dmitry Baryshkov wrote:
> On Mon, Mar 02, 2026 at 02:46:33PM +0100, Konrad Dybcio wrote:
>> On 3/2/26 2:28 PM, Dmitry Baryshkov wrote:
>>> On Mon, Mar 02, 2026 at 11:41:59AM +0100, Konrad Dybcio wrote:
>>>> On 2/27/26 8:05 PM, Dmitry Baryshkov wrote:
>>>>> On Fri, Feb 27, 2026 at 12:34:04PM +0100, Konrad Dybcio wrote:
>>>>>> On 2/27/26 4:48 AM, Dmitry Baryshkov wrote:
>>>>>>> On Thu, Feb 26, 2026 at 02:35:52PM +0100, Konrad Dybcio wrote:
>>>>>>>> On 1/12/26 9:25 AM, yuanjiey wrote:
>>>>>>>>> On Mon, Jan 12, 2026 at 09:38:41AM +0200, Dmitry Baryshkov wrote:
>>>>>>>>>> On Mon, 12 Jan 2026 at 08:23, yuanjiey <yuanjie.yang@xxxxxxxxxxxxxxxx> wrote:
>>>>>>>>>>>
>>>>>>>>>>> On Fri, Jan 09, 2026 at 05:22:37PM +0200, Dmitry Baryshkov wrote:
>>>>>>>>>>>> On Fri, Jan 09, 2026 at 04:38:07PM +0800, yuanjie yang wrote:
>>>>>>>>>>>>> From: Yuanjie Yang <yuanjie.yang@xxxxxxxxxxxxxxxx>
>>
>> [...]
>>
>>>
>>>>
>>>>> 1.1  - MSM8x26
>>>>>        Probably Luca can better comment on it. Should be doable, but I
>>>>>        don't see upstream devices using display on it.
>>>>
>>>> Because there's no iommu support for these
>>>
>>> I promised to put it on my todo list, but the list is very long.
>>>
>>>>
>>>>> 1.2  - MSM8974
>>>>>        I think it also had issues, no IOMMU support in upstream, etc.
>>>>> 1.3  - APQ8084
>>>>>        Had hw issues, no testing base, no MDSS in upstream DT
>>>>> 1.6  - MSM8916 / MSM8939
>>>>>        Can be done, low-hanging fruit for testing
>>>>> 1.7  - MSM8996
>>>>>        Supported in DPU
>>>>> 1.8  - MSM8936
>>>>>        No upsteram testing base
>>>>
>>>> 8936 is 39 with some CPUs fused off (unless you have info suggesting
>>>> otherwise)
>>>
>>> Hmm, you added 8x36 to mdp5_cfg.c, stating it is 1.8. See commit
>>> 81c4389e4835 ("drm/msm/mdp5: Add MDP5 configuration for MSM8x36.")
>>> Author: Konrad Dybcio <konradybcio@xxxxxxxxx>
>>>
>>> Please remove it from the mdp5_cfg to avoid confusion.
>>
>> v1.6 is strictly for 8916. 8936/39 both use v1.8.
> 
> I tend to agree with you. It's interesting that core list doesn't (and
> likely is wrong).
> 
>>
>>>>> 1.9  - MSM8994
>>>>>        No upstream testing base, no MDSS in upstream DT, normal CURSOR planes
>>>>> 1.10 - MSM8992
>>>>>        Even less testing base, no MDSS in upstream DT, normal CURSOR planes
>>>>> 1.11 - MSM8956 / 76
>>>>>        No complete display configurations upstream
>>>>
>>>> +Marijn, is your computer museum still running?
>>>
>>> Should we open a Qualcomm Virtual Museum?
>>
>> Maybe someone has a drawer full of QSD8250s!
> 
> Only APQ8060 here, but it's not even online after the reorg.
> 
> Anyway, I assume that we have something alive and kicking for:
> - 1.1, 8226 (currently unusable, no IOMMU)
> - 1.2, 8974 (currently unusable, no IOMMU)
> - 1.6, 8916
> - 1.8, 8939
> - 1.9, 8994 (almost dead, I'd say)
> - 1.10, 8992 (almost dead, I'd say)

My estimates for 8992/4 would be just a dozen or less enthusiasts, at most
a couple dozen. The platform is in a tragic/borderline unsupported state
upstream today and little effort has been seen to fix it, would probably
need to come from us anyway, with essentially a re-bring-up..

Plus those have no IOMMU either (same camp as 74) and even if they did,
I reckon any users would be happier to run the actually-maintained codebase.

I think I may have a partially-ported 94 branch somewhere on some SSD..

Definitely more of a 'weekend project' type thing and shouldn't be
considered a blocker.

> - 1.11, 8956/76
> 
> It seems we can ignore (or drop) apq8084 (no activity since Dec 2019).
> 
> Already in DPU:
> - 1.7, 8996
> - 1.14, 8937
> - 1.15, 8917
> - 1.16, 8953
> 
> Which means:
> - port 8916, copy for 8939 / 8956/76 and hope it works

I suppose it's not a high priority for you. Maybe some 8916 enthusiast
(+CC a couple) would like to try? (context: DPU1 support)

Konrad

> - add LM-cursor support
> - deprecate mdp5, possibly moving it to EXPERT
> - add 8226 / 8974 once we have IOMMU
> - drop mdp5
>