1634610 - [Wayland][DMABUF] "Only invalid modifier specified" messages in stderr

Status: RESOLVED INVALID

(Core :: Widget: Gtk, defect)

Description

[Robert Mader [:rmader]]

I get a lot of Only invalid modifier specified when having widget.wayland-dmabuf-webgl.enabled. That being said, it appears to work just fine and the performance is much better than without.

I tested on Fedora 32 (Gnome with and without 1) with current FF (75.0-1.fc32) and mesa (20.0.5-1), as well as FF nightly (20200430155041) and mesa head (20.2.0-devel (git-d56b8c4554) which includes 2).

The error message directly originates from mesa (3, 4) and the corresponding comment states:

It's acceptable to create an image with INVALID modifier in the list, but it cannot be on the only modifier (since it will certainly fail later). While we could easily catch this after modifier creation, doing the check here is a convenient debug check likely pointing at whatever interface the client is using to build its modifier list.

1: https://gitlab.gnome.org/GNOME/mutter/-/merge_requests/804
2: https://gitlab.freedesktop.org/mesa/mesa/-/merge_requests/4294
3: https://gitlab.freedesktop.org/mesa/mesa/-/blob/6f7d94580e3b603cb036bef9a1a235ee6b910bc0/src/gbm/backends/dri/gbm_dri.c#L1164
4: https://gitlab.freedesktop.org/mesa/mesa/-/blob/6f7d94580e3b603cb036bef9a1a235ee6b910bc0/src/gbm/backends/dri/gbm_dri.c#L1289

Comment 1

[Robert Mader [:rmader]]

Attachment: about:support

Comment 2

[Robert Mader [:rmader]]

Hm, maybe a mesa bug - I get the same with weston-simple-dmabuf-egl. weston-info gives me:

interface: 'zwp_linux_dmabuf_v1', version: 3, name: 21
	formats:
	'AR4H'(0x48345241), modifier: 0x00ffffffffffffff
	'XR4H'(0x48345258), modifier: 0x00ffffffffffffff
	'XB4H'(0x48344258), modifier: 0x00ffffffffffffff
	'AB4H'(0x48344241), modifier: 0x00ffffffffffffff
	'RG16'(0x36314752), modifier: 0x00ffffffffffffff
	'XR30'(0x30335258), modifier: 0x00ffffffffffffff
	'AR30'(0x30335241), modifier: 0x00ffffffffffffff
	'XR24'(0x34325258), modifier: 0x00ffffffffffffff
	'AR24'(0x34325241), modifier: 0x00ffffffffffffff

Comment 3

[Robert Mader [:rmader]]

Na, never mind. Mutter is at fault here, should have tested directly. On Weston things are fine.