Xwayland fatal error when Wayland output disappears

[Olivier Fourdan]

Hi Pekka,

On 27 October 2017 at 08:01, Pekka Paalanen <ppaalanen at gmail.com> wrote:

> there is a known race around Wayland globals. If the Wayland server
> adds and removes a global in a very short time, it may succeed to
> remove the global (wl_output) before all clients have processed the
> add. If a client process an add after the server removed, you hit
> exactly this fatal error.
>
> It's a design flaw in Wayland, gone unnoticed for years until it was
> too late to fix properly.
>
> This issue is recorded: https://phabricator.freedesktop.org/T7722
>
> There is a suggested mitigation, but I am not aware of anyone working
> on it.
>

FWIW, it seems we have that issue with more than just wl_output, in
downstream bug 1516859 [1] this occurs with wl_seat on VT switch as well.

That's pretty easy to reproduce, issue several VT switches back and forth
(I use "chvt" from a remote connection) until the race occurs, here it
doesn't take long.

Cheers,
Olivier

[1]   https://bugzilla.redhat.com/show_bug.cgi?id=1516859
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.x.org/archives/xorg-devel/attachments/20171123/32d6b440/attachment.html>