Bug#983466: Is a mesa-bug

From	Salvatore Bonaccorso <carnil@debian.org>
Newsgroups	linux.debian.bugs.dist, linux.debian.kernel
Subject	Bug#983466: Is a mesa-bug
Date	2021-03-13 14:10 +0100
Message-ID	<BScbL-6iD-1@gated-at.bofh.it> (permalink)
References	<BM5G1-as-1@gated-at.bofh.it> <BQvix-7yL-3@gated-at.bofh.it> <BM5G1-as-1@gated-at.bofh.it> <BQvix-7yL-3@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Cross-posted to 2 groups.

Show all headers | View raw

Control: retitle -1 Black screen at dual monitor with warn from i915_gem_ww_ctx_backoff
Control: reassign -1 src:mesa 20.3.4-1
Control: forwarded -1 https://gitlab.freedesktop.org/mesa/mesa/-/issues/790

Hi,

On Mon, Mar 08, 2021 at 09:42:04PM +0100, Gert van de Kraats wrote:
> because wayland itself needs 2 fence registers for 2 momitors  i
> and X11 only 1.
> A trace without using wayland shows the 15th register is used without
> giving any problem:
> 
> Feb 24 20:10:59 debian systemd[1270]: Started GNOME Shell on X11.
> 
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 0 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 1 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 2 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 3 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 4 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 5 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 6 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 7 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 8 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 9 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 10 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 11 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 12 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 13 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 14 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 15 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 1 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 2 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 3 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 4 0
> Feb 24 20:11:29 debian gnome-shell[1505]: GK_intelClearWithBlit_fences 5 0
> .....
> ....
> So this seems to be  a mesa bug and an exact duplicate
> of https://gitlab.freedesktop.org/mesa/mesa/-/issues/790 .
> 
> Only the kernel-driver-error has changed and is misleading now.
> 
> I add the mesa-2.3.4-patch I currently use, which solves the problem.
> 
> Testing for register-overflow is done after updating the batch, If overflow
> occurs
> 
> the batch is restored to the situation before the last update, the buffer is
> flushed
> 
> and the batch again is updated with the last update.

So reassigning this for now to src:mesa and updating the metadata.

Regards,
Salvatore

Thread

Bug#983466: Is a mesa-bug Gert van de Kraats <vd.Kraats@hccnet.nl> - 2021-03-08 22:10 +0100
  Bug#983466: Is a mesa-bug Salvatore Bonaccorso <carnil@debian.org> - 2021-03-13 14:10 +0100
    Processed: Re: Bug#983466: Is a mesa-bug "Debian Bug Tracking System" <owner@bugs.debian.org> - 2021-03-13 14:10 +0100

csiph-web