On Mon, Jan 19, 2026 at 4:31 AM matthew green <mrg%eterna23.net@localhost> wrote: > > Kevin Bowling writes: > > I get the following with 10.1 or NetBSD 11.0_BETA (GENERIC) #0: Thu > > Jan 8 10:53:08 UTC 2026 > > when starting X: > > > > [ 104.512784] radeon0: autoconfiguration error: error: ring 0 > > stalled for more than 10000msec > > [ 104.512784] radeon0: warn: GPU lockup (current fence id > > 0x0000000000000003 last fence id 0x0000000000000006 on ring 0) > > > > I'm able to get a working session with this file added to > > /etc/X11/xorg.conf.d/20-radeon.conf: > > Section "Device" > > Identifier "Radeon" > > Driver "modesetting" > > Option "AccelMethod" "none" > > EndSection > > > > With the modesetting default of AccelMethod "glamor", X will not start with: > > glamor requires at least 128 instructions (64 reported) > > > > So it seems the older EXA radeon might be preferable, but something is > > going wrong in the kernel. Has anyone looked into this drm stuff > > recently? > > i haven't tried an r300 based card for a while. what's the specific model? > can you post the dmesg bits? i'll try to find one of my old cards, though > i feel like i had problems with it in my newer test PC last time i tried. > (as in, the bios failed.) > > thanks. Hi Matthew, It is an IBM ThinkPad T42p, with a Radeon 9000 series. My outcome is similar as your [*1] in this file https://www.netbsd.org/~mrg/mrg-tested-murder-drm.txt except these days console drm seems to automatically attach fine. Using drm2, xorg modesetting driver and disabling xorg acceleration with that (no glamor) works plausibly well but something large like Firefox52 is more lethargic than it should be. > > .mrg