On Tue, Feb 20, 2018 at 11:29 AM, Thierry Reding
<thierry.reding@xxxxxxxxx> wrote:
> From: Thierry Reding <treding@xxxxxxxxxx>
>
> DRM_DUMB_VGA_DAC is a user-visible symbol. Selecting it can cause unmet
> direct dependencies such as this (on i386, randconfig):
>
> warning: (DRM_PL111) selects DRM_DUMB_VGA_DAC which has unmet direct dependencies (HAS_IOMEM && DRM && DRM_BRIDGE && OF)
>
> This is because DRM_DUMB_VGA_DAC depends on OF while DRM_PL111 does not.
> It does indirectly depend on OF via the ARM and ARM64 dependencies, but
> since it can also be enabled under COMPILE_TEST, randconfig can find a
> case where DRM_PL111 is selected without pulling in OF and not meeting
> the dependency for DRM_DUMB_VGA_DAC.
>
> Since select is "heavy handed", DRM_DUMB_VGA_DAC is going to be enabled
> regardless of the above warning and causes the following build error:
>
> ../drivers/gpu/drm/bridge/dumb-vga-dac.c: In function 'dumb_vga_probe':
> ../drivers/gpu/drm/bridge/dumb-vga-dac.c:207:13: error: 'struct drm_bridge' has no member named 'of_node'
> vga->bridge.of_node = pdev->dev.of_node;
>
> See Documentation/kbuild/kconfig-language.txt, "reverse dependencies".
>
> Fixes: 49f81d80ab84 ("drm/pl111: Support handling bridge timings")
> Reported-by: Randy Dunlap <rdunlap@xxxxxxxxxxxxx>
> Cc: Laurent Pinchart <laurent.pinchart@xxxxxxxxxxxxxxxx>
> Cc: Archit Taneja <architt@xxxxxxxxxxxxxx>
> Cc: Eric Anholt <eric@xxxxxxxxxx>
> Signed-off-by: Thierry Reding <treding@xxxxxxxxxx>
OK sorry about that, I just get to let the platforms that need the bridge
select it explicitly.
Reviewed-by: Linus Walleij <linus.walleij@xxxxxxxxxx>
Yours,
Linus Walleij
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel
