[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-Id: <9d549ed7-69ce-428b-a622-c3a43909937e@app.fastmail.com>
Date: Thu, 27 Jul 2023 16:47:13 +0200
From: "Arnd Bergmann" <arnd@...db.de>
To: "Javier Martinez Canillas" <javierm@...hat.com>,
"Arnd Bergmann" <arnd@...nel.org>,
"Dave Airlie" <airlied@...il.com>,
"Daniel Vetter" <daniel@...ll.ch>,
"Maxime Ripard" <mripard@...nel.org>
Cc: "Thomas Zimmermann" <tzimmermann@...e.de>,
"Sam Ravnborg" <sam@...nborg.org>,
"Geert Uytterhoeven" <geert@...ux-m68k.org>,
dri-devel@...ts.freedesktop.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] drm/ssd130x: drop DRM_GEM_SHADOW_PLANE_FUNCS override
On Thu, Jul 27, 2023, at 16:07, Javier Martinez Canillas wrote:
> Arnd Bergmann <arnd@...nel.org> writes:
>
> Hello Arnd,
>
> Thanks a lot for your patch.
>
>> From: Arnd Bergmann <arnd@...db.de>
>>
>> It is not possible for ssd130x_primary_plane_funcs to use both
>> its own reset/duplicate/destroy callbacks and the ones from
>> DRM_GEM_SHADOW_PLANE_FUNCS:
>>
>> In file included from drivers/gpu/drm/solomon/ssd130x.c:29:
>> include/drm/drm_gem_atomic_helper.h:100:18: error: initialized field overwritten [-Werror=override-init]
>> 100 | .reset = drm_gem_reset_shadow_plane, \
>> 101 | .atomic_duplicate_state = drm_gem_duplicate_shadow_plane_state, \
>> 102 | .atomic_destroy_state = drm_gem_destroy_shadow_plane_state
>> | ^~~~~~~~~~~~~~~~~~~~~~~~~~
>> drivers/gpu/drm/solomon/ssd130x.c:744:9: note: in expansion of macro 'DRM_GEM_SHADOW_PLANE_FUNCS'
>> 744 | DRM_GEM_SHADOW_PLANE_FUNCS,
>> | ^~~~~~~~~~~~~~~~~~~~~~~~~~
>>
>
> Gah, I missed that. I wonder how this compile warning doesn't show up for me.
The warning is normally disabled, I have it turned on in my local
patch series on top of linux-next.
>> Since the custom callbacks were just added, I assume these are the
>> ones it is meant to use, so remvoe the generic overrides.
>>
>> Fixes: 45b58669e532b ("drm/ssd130x: Allocate buffer in the plane's .atomic_check() callback")
>
> Unfortunately that's not enough, because the plane atomic state handlers
> that were introduced in that commit are broken so removing the macro will
> regress the driver.
>
> I've posted a patch that drops the macro but also fixes the handlers for
> the driver to work with the custom callbacks:
>
> https://lists.freedesktop.org/archives/dri-devel/2023-July/415897.html
Ok, makes sense.
Arnd
Powered by blists - more mailing lists