about weston behavior if DRM driver don't expose supported modifiers

Esaki Tomohito etom at igel.co.jp
Fri May 17 07:55:08 UTC 2019


A board (arm64 arch) DRM kernel driver doesn't expose supported DRM
modifiers (hardware supports linear only), so I suggested following


The driver maintainer (Laurent CC'd) replied:

> I don't think the right solution is to expose the linear modifier from
> all drivers that don't otherwise support modifiers. We should instead
> fix it either in Weston, and treat drivers that don't support the
> modifiers API as supporting the linear modifier only, or in the DRM/KMS
> core by reporting the linear modifier for drivers that don't explicitly
> support modifiers.

I would like to know if you have any opinion on this comment.
Do we need fix weston?

I want to use sprite plane. The GPU hardware is a different device from
the display DRM hardware. The GPU driver advertises support for
DRM_FORMAT_MOD_LINEAR modifier, but display driver currently doesn't
support DRM modifiers. Currently in weston, if a client attaches a
buffer with DRM_FORMAT_MOD_LINEAR modifier via linux-dmabuf protocol,
since the attached buffer modifier is different to plane modifier,
weston draws surface in primary plane not sprite plane.


Tomohito Esaki
etom at igel.co.jp

More information about the wayland-devel mailing list