I've just tracked down a bug in my code for handling the DRM maps in the kernel, that was affecting nouveau (running out of vmalloc space) and radeon (hanging on drmGetLock at startup).
However fixing the bug uncovered a larger bug which is driver specific, so the initial fixes are in the kernel building as kernel-2_6_25-0_211_rc8_git6_fc9 in koji,
I'd really like feedback from nouveau, radeon, savage and mga people that have working DRI now, and to let me know via a bug if the new kernel regresses that or oopses after 211 in the drm code.
Sorry about this at this late stage, but thats the problem with nasty bugs :)
I can probably test MGA myself at some point with a PCI card, but savage hw I have none off.
Dave.
On Tue, Apr 08, 2008 at 18:19:20 +1000, Dave Airlie airlied@redhat.com wrote:
I'd really like feedback from nouveau, radeon, savage and mga people that have working DRI now, and to let me know via a bug if the new kernel regresses that or oopses after 211 in the drm code.
I tested 212 with glxgears on my 9200 at home and it seemed to work OK. I had seen some weird stuff at work (where I have an r5xx based card), but I didn't have time to look at it carefully so it might not be a problem with the kernel. I'll test it more carefully today and open a bug if it really appears to be a regression.
Dave Airlie wrote:
I've just tracked down a bug in my code for handling the DRM maps in the kernel, that was affecting nouveau (running out of vmalloc space) and radeon (hanging on drmGetLock at startup).
However fixing the bug uncovered a larger bug which is driver specific, so the initial fixes are in the kernel building as kernel-2_6_25-0_211_rc8_git6_fc9 in koji,
I'd really like feedback from nouveau, radeon, savage and mga people that have working DRI now, and to let me know via a bug if the new kernel regresses that or oopses after 211 in the drm code.
How can I get DRI support in nouveau in F9? My Xorg log says:
... (II) AIGLX: Screen 0 is not DRI2 capable drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is 12, (OK) drmOpenByBusid: Searching for BusID pci:0000:00:0d.0 drmOpenDevice: node name is /dev/dri/card0 drmOpenDevice: open result is 12, (OK) drmOpenByBusid: drmOpenMinor returns 12 drmOpenByBusid: drmGetBusid reports pci:0000:00:0d.0 (EE) AIGLX error: dlopen of /usr/lib/dri/nouveau_dri.so failed (/usr/lib/dri/nouveau_dri.so: cannot open shared object file: No such file or directory) (EE) AIGLX: reverting to software rendering ...
Regards, Dennis
Dave Airlie wrote:
I've just tracked down a bug in my code for handling the DRM maps in the kernel, that was affecting nouveau (running out of vmalloc space) and radeon (hanging on drmGetLock at startup).
However fixing the bug uncovered a larger bug which is driver specific, so the initial fixes are in the kernel building as kernel-2_6_25-0_211_rc8_git6_fc9 in koji,
I'd really like feedback from nouveau, radeon, savage and mga people that have working DRI now, and to let me know via a bug if the new kernel regresses that or oopses after 211 in the drm code.
Sorry about this at this late stage, but thats the problem with nasty bugs :)
I can probably test MGA myself at some point with a PCI card, but savage hw I have none off.
I was hoping this would fix my i865g freezing on doing any opengl apps (glxgears hard freezes it), as this seems to be a kernel issue (downgrading to 2.6.24 fixes it). Note with both kernels it is not using TTM, so that is not the cause.
See: https://bugzilla.redhat.com/show_bug.cgi?id=440384
Regards,
Hans
Dave Airlie wrote:
I've just tracked down a bug in my code for handling the DRM maps in the kernel, that was affecting nouveau (running out of vmalloc space) and radeon (hanging on drmGetLock at startup).
Could the vmalloc problem be the cause of this: https://bugzilla.redhat.com/show_bug.cgi?id=437062
This bug makes nouveau basically unusable after a while for normal 2D desktop work.
Regards, Dennis
On Tue, Apr 08, 2008 at 18:19:20 +1000, Dave Airlie airlied@redhat.com wrote:
I'd really like feedback from nouveau, radeon, savage and mga people that have working DRI now, and to let me know via a bug if the new kernel regresses that or oopses after 211 in the drm code.
My work machine's (with a FireGL V3400) oddities are still there when running 204, so they aren't problems introduced with the 212 kernel.