Re: [PATCH v3 1/2] drm/imx/lcdc: avoid duplicate clk_per enable

On Mon Jul 27, 2026 at 3:59 AM CST, sashiko-bot wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
>
> Pre-existing issues:
> - [High] Calls to sleeping functions (`clk_prepare_enable` and `clk_disable_unprepare`) from the plane atomic update and CRTC enable/disable callbacks, violating DRM atomic context requirements.
> --
>
> --- Patch [1]: [PATCH v3 1/2] drm/imx/lcdc: avoid duplicate clk_per enable ---
> commit 48c4f154b28e9de1d485ea56766c8a24df823739
> Author: Ze Huang <[email protected]>
>
> drm/imx/lcdc: avoid duplicate clk_per enable
>
> The simple-KMS helper calls the pipe update after enabling the CRTC,
> which causes imx_lcdc_update_hw_registers() to be called again. This
> patch avoids the duplicate clock enable by skipping the register update
> when the CRTC already needs a modeset.
>
>> diff --git a/drivers/gpu/drm/imx/lcdc/imx-lcdc.c b/drivers/gpu/drm/imx/lcdc/imx-lcdc.c
>> index c67fe80b8115e..e00e52c1c46ec 100644
>> --- a/drivers/gpu/drm/imx/lcdc/imx-lcdc.c
>> +++ b/drivers/gpu/drm/imx/lcdc/imx-lcdc.c
>> @@ -310,7 +310,8 @@ static void imx_lcdc_pipe_update(struct drm_simple_display_pipe *pipe,
>>  	else if (old_crtc != crtc)
>>  		mode_changed = true;
>>  
>> -	imx_lcdc_update_hw_registers(pipe, old_state, mode_changed);
>> +	if (!drm_atomic_crtc_needs_modeset(crtc->state))
>> +		imx_lcdc_update_hw_registers(pipe, old_state, mode_changed);
>
> [Severity: High]
> This is a pre-existing issue, but does calling imx_lcdc_update_hw_registers()
> from imx_lcdc_pipe_update() risk sleeping in an atomic context?
>
> imx_lcdc_pipe_update() acts as a plane atomic update callback, which executes
> in an atomic context where sleeping is forbidden. However, it calls
> imx_lcdc_update_hw_registers(), which in turn calls sleeping functions:
>
> imx_lcdc_update_hw_registers() {
>     ...
>     clk_disable_unprepare(lcdc->clk_per);
>     ...
>     clk_prepare_enable(lcdc->clk_per);
>     ...
> }
>
> Will this cause 'scheduling while atomic' panics when a plane format changes
> without a full modeset?

False positive