May by change the heuristic? How I can help with it? чт, 28 мар. 2019 г. в 20:01, Mark Kettenis <[email protected]>:
> > Date: Thu, 28 Mar 2019 15:46:13 +0100 > > From: Frederic Cambus <[email protected]> > > > > On Tue, Mar 26, 2019 at 09:58:09PM +0500, dmitry.sensei wrote: > > > Also I have kernel loop trap.. after running X and switching to 0 > screen > > > (Ctrl-Alt-F1) > > > > Thanks for your report and for sharing your dmesg. > > > > I just reverted the other commit I previously did to efifb(4). Please > > let me know if this fixes the issue for you. > > I don't think that will help. > > > Short of being able to reproduce the issue locally, there is not much > > else I can do at the moment. > > > > There is something interesting happening on your system though, because > > your dmesg shows efifb attaches after inteldrm. > > That should *not* be happening. If inteldrm(4) attaches, efifb(4) > should not attach for the same framebuffer. Looking at the dmesg: > > https://marc.info/?l=openbsd-bugs&m=155361954332430&w=2 > > this system has both inteldrm(4) and radeondrm(4). That may > complicate things a bit and it would be good to know what graphics > hardware is used by the EFI framebuffer. > > In the end the problem probably is that the firmware doesn't really > tell us what device should be used as the console. The OpenBSD kernel > has a heuristic to figure out what device should be used but it is a > bit flawed. > -- Dmitry Orlov
