Re: [PATCH 70/70] drm/crtc: Remove reset

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [High] The return value of the newly introduced `atomic_create_state` hook is ignored in `amdgpu_dm_crtc_init()`, causing a memory leak and leaving the CRTC state uninitialized.
--

--- Patch [70]: [PATCH 70/70] drm/crtc: Remove reset ---
commit de86d4e5c49ea36efff6debb8f1603e3e3c819b9
Author: Maxime Ripard <[email protected]>

drm/crtc: Remove reset
    
All crtc drivers now implement the atomic_create_state hook instead of
the reset hook for initial state creation. The reset hook conflated
initial state allocation at probe time with hardware and software reset
during suspend/resume, making error handling difficult since it is not
fallible.

Remove the reset hook from struct drm_crtc_funcs and the associated
call in drm_mode_config_reset().

> diff --git a/drivers/gpu/drm/drm_mode_config.c b/drivers/gpu/drm/drm_mode_config.c
> index 3bcc7bf0900c4..9f4eb694e1fb2 100644
> --- a/drivers/gpu/drm/drm_mode_config.c
> +++ b/drivers/gpu/drm/drm_mode_config.c
> @@ -296,9 +296,7 @@ void drm_mode_config_reset(struct drm_device *dev)
>  	}
>  
>  	drm_for_each_crtc(crtc, dev) {
> -		if (crtc->funcs->reset)
> -			crtc->funcs->reset(crtc);
> -		else if (crtc->funcs->atomic_create_state)
> +		if (crtc->funcs->atomic_create_state)
>  			drm_mode_config_crtc_reset_with_create_state(crtc);
>  	}

[Severity: High]
Does this transition leave the CRTC state uninitialized for AMD GPUs?

An earlier patch in this series switched amdgpu_dm_crtc_init() to call
atomic_create_state() instead of reset(), but the returned state pointer
is now ignored:

drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_crtc.c:amdgpu_dm_crtc_init() {
    ...
	/* Create the plane state */
	if (acrtc->base.funcs->atomic_create_state)
		acrtc->base.funcs->atomic_create_state(&acrtc->base);
    ...
}

Since the returned struct drm_crtc_state pointer is not assigned to crtc->state,
and the AMDGPU driver doesn't call drm_mode_config_reset() during probe,
could this lead to a memory leak and a NULL pointer dereference when the
DRM core later accesses the uninitialized CRTC state?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=70