[PATCH] drm: document better that drivers shouldn't use drm_minor directly

Maíra Canal mcanal at igalia.com
Thu Jan 5 10:21:09 UTC 2023


On 1/4/23 18:17, Daniel Vetter wrote:
> The documentation for struct drm_minor already states this, but that's
> not always that easy to find.
> 
> Also due to historical reasons we still have the minor-centric (like
> drm_debugfs_create_files), but since this is now getting fixed we can
> put a few more pointers in place as to how this should be done
> ideally.
> 
> Motvated by some discussion with Rodrigo on irc about how drm/xe
> should lay out its sysfs interfaces.
> 
> Cc: Rodrigo Vivi <rodrigo.vivi at intel.com>
> Cc: Wambui Karuga <wambui.karugax at gmail.com>
> Cc: Maíra Canal <mcanal at igalia.com>
> Cc: Maxime Ripard <maxime at cerno.tech>
> Cc: Melissa Wen <mwen at igalia.com>
> Signed-off-by: Daniel Vetter <daniel.vetter at intel.com>
> ---
>   include/drm/drm_device.h | 17 +++++++++++++++--
>   1 file changed, 15 insertions(+), 2 deletions(-)
> 
> diff --git a/include/drm/drm_device.h b/include/drm/drm_device.h
> index 9923c7a6885e..b40e07e004ee 100644
> --- a/include/drm/drm_device.h
> +++ b/include/drm/drm_device.h
> @@ -87,10 +87,23 @@ struct drm_device {
>   	 */
>   	void *dev_private;
>   
> -	/** @primary: Primary node */
> +	/**
> +	 * @primary:
> +	 *
> +	 * Primary node. Drivers should not interact with this
> +	 * directly. debugfs interface can be registered with
> +	 * drm_debugfs_add_file(), and sysfs should be directly added on the
> +	 * hardwire struct device @dev.
> +	 */
>   	struct drm_minor *primary;
>   
> -	/** @render: Render node */
> +	/**
> +	 * @render:
> +	 *
> +	 * Render node. Drivers should not interact with this directly ever.
> +	 * Drivers should not expose any additional interfaces in debugfs or
> +	 * sysfs on thise node.

I believe you meant s/thise/this.

Apart from this small nit,

Reviewed-by: Maíra Canal <mcanal at igalia.com>

Best Regards,
- Maíra Canal

> +	 */
>   	struct drm_minor *render;
>   
>   	/**


More information about the dri-devel mailing list