On Tue, Oct 08, 2024 at 12:59:19AM +0100, Stuart Henderson wrote: > On 2024/10/07 14:17, Stuart Henderson wrote: > > On 2024/10/07 14:04, Matthieu Herrb wrote: > > > The only file that Xorg writes to is its log file. Sometimes, when > > > input devices go bad, it can be very verbose. > > > > > > Can you check if any of the Xorg.0.log files are huge ? > > > > yes: > > > > $ ls -lh /var/log/Xorg.0.log* > > -rw-r--r-- 1 root wheel 45.6K Oct 7 13:37 /var/log/Xorg.0.log > > -rw-r--r-- 1 root wheel 433M Oct 3 11:48 /var/log/Xorg.0.log.old > > > > hmm, there are loads of modeline dumps: > > so, this is common to a bunch of machines, to the extent that some > have run out of space in /var due to the logs. > > I think I'll build an Xorg hacked up to avoid printing modelines > to buy me a bit of time for those.
That's just hiding the dirt below the carpet :-) but without a better fix If you really have 433M of modeline dumps, then there is something else wrong. in the modesetting driver afaict monitor connexion/deconnexion events come from libdrm, which gets them from the kernel drm driver. >From your log excerpt it seems to start 9300 seconds or so after the initial launch of the X server which can also make it difficult to reproduce. The start of the log file is just normal monitor suspend and resume every hour or so, probably caused by DPMS and other power saving mechanisms. I don't know if you're hitting a bug or flaky hardware to trigger this event flow (if it happens on more than one machine flaky hardware is unlikely) This is probably also slowing down X clients a lot since upon detection of a monitor a lot of churn happens inside the X server. -- Matthieu Herrb
