Re: [PATCH v11 06/15] drm/panfrost: Fix PM refcnt and autosuspend issues at device probe/remove

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] pm: omitted reordering of pm_runtime_set_suspended() contrary to commit message
--

commit e5b622a5bf50c2417ee98e980c4b45ec511cb4e7
Author: Adrián Larumbe <adrian.larumbe@collabora.com>

drm/panfrost: Fix PM refcnt and autosuspend issues at device probe/remove

This commit attempts to fix PM reference count and autosuspend issues
during device probe and removal. It manually adjusts the PM reference
count, disables autosuspend in wind-down paths, and aims to reorder
PM teardown steps to mirror initialization.

> diff --git a/drivers/gpu/drm/panfrost/panfrost_drv.c b/drivers/gpu/drm/panfrost/panfrost_drv.c
> index 55fc22e8d4d4e..907d4a14a0b55 100644
> --- a/drivers/gpu/drm/panfrost/panfrost_drv.c
> +++ b/drivers/gpu/drm/panfrost/panfrost_drv.c
[ ... ]
> @@ -866,13 +871,16 @@ static int panfrost_probe(struct platform_device *pdev)
[ ... ]
>  err_out1:
> +	pm_runtime_dont_use_autosuspend(pfdev->base.dev);
>  	pm_runtime_disable(pfdev->base.dev);
>  	panfrost_device_fini(pfdev);
>  	pm_runtime_set_suspended(pfdev->base.dev);

[Severity: Medium]
The commit message states "Move pm_runtime_set_suspended() before
panfrost_device_fini() so that resource unwinding happens in the opposite
order as initialisation." However, this structural reordering appears to be
missing from the code.

Should pm_runtime_set_suspended() be moved above panfrost_device_fini() in
the panfrost_probe() error path to match the intent described in the commit
message?

> +	pm_runtime_put_noidle(pfdev->base.dev);
>  err_out0:
>  	return err;
>  }
> @@ -884,9 +892,12 @@ static void panfrost_remove(struct platform_device *pdev)
[ ... ]
>  	pm_runtime_get_sync(pfdev->base.dev);
> +	pm_runtime_dont_use_autosuspend(pfdev->base.dev);
>  	pm_runtime_disable(pfdev->base.dev);
>  	panfrost_device_fini(pfdev);
>  	pm_runtime_set_suspended(pfdev->base.dev);

[Severity: Medium]
Similarly, in the panfrost_remove() path, the reordering mentioned in the
commit message is omitted.

Should pm_runtime_set_suspended() be moved above panfrost_device_fini()
here as well?

> +	pm_runtime_put_noidle(pfdev->base.dev);
> +
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260925-claude-fixes-v11-0-0dbf5a58e7ce@collabora.com?part=6