Re: [PATCH v2 4/6] drm/i915/psr: Merge the selective fetch area across joined pipes

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Low] drm/i915/psr: Dead code in psr2_sel_fetch_align_su_area for cursor_in_su_area
--

commit 7a64006c52d0dd79eca20d987ffe7b3c8f8b6c66
Author: Mika Kahola <mika.kahola@intel.com>
Date:   Mon Oct 12 12:41:43 2026 +0000

    drm/i915/psr: Merge the selective fetch area across joined pipes

This commit updates PSR2 selective fetch to handle joined pipes by merging
the selective fetch area across the pipes. It ensures only the primary pipe
programs the transcoder register, and falls back to a full update if a joined
pipe fails area calculation or if no joined pipe has damage.

> diff --git a/drivers/gpu/drm/i915/display/intel_psr.c b/drivers/gpu/drm/i915/display/intel_psr.c
> index 0483dad5bdd82..9a57339c004a5 100644
> --- a/drivers/gpu/drm/i915/display/intel_psr.c
> +++ b/drivers/gpu/drm/i915/display/intel_psr.c
[ ... ]
> +static void psr2_sel_fetch_align_su_area(struct intel_atomic_state *state,
> +					 u8 joined_pipes)
> +{
> +	struct intel_display *display = to_intel_display(state);
> +	struct intel_crtc *crtc;
> +	bool su_area_changed;
> +
> +	do {
> +		su_area_changed = false;
> +
> +		for_each_intel_crtc_in_pipe_mask(display, crtc, joined_pipes) {
> +			struct intel_crtc_state *crtc_state =
> +				intel_atomic_get_new_crtc_state(state, crtc);
> +			struct drm_rect display_area = {
> +				.x1 = 0,
> +				.y1 = 0,
> +				.x2 = crtc_state->hw.pipe_mode.crtc_hdisplay,
> +				.y2 = crtc_state->hw.pipe_mode.crtc_vdisplay,
> +			};
> +			struct drm_rect old_su_area = crtc_state->psr2_su_area;
> +			bool cursor_in_su_area = false;
> +
> +			/* Needs the planes already pulled into the state. */
> +			intel_psr2_sel_fetch_et_alignment(state, crtc, &display_area,
> +							  &cursor_in_su_area);
> +

[Severity: Low]
Could the cursor_in_su_area variable and its corresponding parameter in
intel_psr2_sel_fetch_et_alignment() be removed?

The previous implementation used cursor_in_su_area to conditionally break out
of the alignment loop early. Since this patch removes that break condition,
the variable is initialized and populated by intel_psr2_sel_fetch_et_alignment()
but its value is never checked, making it dead code.

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20261012124145.1960050-1-mika.kahola@intel.com?part=4