Re: [PATCH RFT 0/5] drm/msm: Attach a driver to GMU

[Akhil P Oommen]

On 10/28/2026 11:23 PM, Akisaira wrote:
> On Wed, May 13, 2026 at 02:22:55AM +0530, Akhil P Oommen wrote:
>> With the introduction of sync_state in the genpd framework, any consumer
>> device of GCC and GPUCC which is not bound to a driver will result in
>> bootup warnings like below:
> 
> Tested on SC8280XP (Huawei MateBook E Go, gaokun3, Adreno 690) running
> Android, on top of v7.2.9. Patch 5 needed a small fixup there:
> a6xx_gpu_init() declares the GMU node with __free(device_node), so the
> explicit of_node_put(node) at the end goes away.
> 
> Without the series, GCC and GPUCC report
> 
>   gcc-sc8280xp 100000.clock-controller: sync_state() pending due to 3d6a000.gmu
>   gpu_cc-sc8280xp 3d90000.clock-controller: sync_state() pending due to 3d6a000.gmu
> 
> and six unused GDSCs (ufs_phy, ufs_card and the four
> hlos1_vote_mmnoc_mmu_tbu_* ones) stay on unless
> fw_devlink.sync_state=timeout is given.
> 
> With the series, 3d6a000.gmu binds to adreno_gmu, both sync_states
> complete without the timeout, and the six GDSCs are off. The GPU works
> (ANGLE on Turnip), GPU and GMU runtime suspend when idle, and a game
> session followed by five s2idle cycles gave no GPU or GMU errors.
> 
> Tested-by: Akisaira <akisaira@xxxxxxxxxxx> # sc8280xp-huawei-gaokun3

Thanks. I am testing the rev2 with some updates. Will post it soon.

-Akhil.
> 
> Thanks!