Hi Florian, "pelzflorian (Florian Pelz)" <pelzflor...@pelzflorian.de> skribis:
> On Tue, Apr 14, 2020 at 12:28:38PM +0200, pelzflorian (Florian Pelz) wrote: >> On Tue, Apr 14, 2020 at 12:09:12PM +0200, Ludovic Courtès wrote: >> > Or better yet, is there a way to check whether KMS is already using >> > another driver, and to not load uvesafb in that case? Probably there’s >> > some info in /sys. >> >> Hmm I will try if checking for the presence of /dev/fb0 helps. > > I have tested and pushed a check for /dev/fb0 as > 0ad60b2a89d6d387236466e0bcdd61ac489fca37 to the version-1.1.0 branch. > > All my machines that need uvesafb use it while those that don’t need > it do not modprobe the kernel module anymore. Possibly there remain > problems should the normal Kernel Mode Setting be very slow to create > /dev/fb0, but otherwise the logs should no longer show uvesafb errors. > > They do show that the uvesafb shepherd service was started, even when > it did not load uvesafb. Well done! I’ll send an image to Niels in case they can double-check that the /dev/fb0 check works for them. > Maybe the shepherd service should be renamed. To ‘maybe-uvesafb’? Doesn’t matter much to me. :-) Thank you! Ludo’.
