[PATCH weston 1/1] compositor: Abort on bad page flip timestamps

Michel Dänzer michel at daenzer.net
Fri Nov 7 01:56:30 PST 2014


On 07.11.2014 18:36, Pekka Paalanen wrote:
>
> If we still feel that exiting Weston on the first timestamp failure
> (zero or not increasing) is too harsh, we could simply log a warning
> and add again some heuristics on when things are too far wrong to
> continue.
>
> E.g. getting a single timestamp failure during the first 5 seconds of
> compositor uptime warrants an exit, but later probably not. We want to
> catch cases where the compositor cannot reliably work to begin with,
> but killing the user's session much later is pretty blunt if it already
> worked somewhat.
>
> Would that be good?

Yeah, that would address my main concern.


-- 
Earthling Michel Dänzer            |                  http://www.amd.com
Libre software enthusiast          |                Mesa and X developer


More information about the dri-devel mailing list