Bug#683796: xserver-xorg-video-radeon: No 3D acceleration in iBook2 powerpc
On Don, 2012-08-09 at 05:11 -0700, Dan DeVoto wrote: --- On Wed, 8/8/12, Michel Dänzer daen...@debian.org wrote: I heard that this handover from OFfb to radeon KMS can cause problems in some cases. Maybe you can try adding video=offb:off to video=radeonfb:off . BTW, does the black screen occur only when X starts, or already when KMS takes over from OFfb, which according to the above happens about 10 seconds after the kernel starts booting? The black screen happens about 10 seconds after boot starts when the line appears onscreen: fb: conflicting fb...(it didn't give me time to read the rest, but I assume it's fb: conflicting fb hw usage radeondrmfb vs OFfb ATY,Fall_A - removing generic driver). Just to be clear about the symptoms, the screen is still lit, it's just colored black and there's nothing visible. Then when X starts in the boot process, the screen goes dark for a second as if it's powered off, then comes on again to a lit but black screen, then after another second switches to a slightly lower brightness level. When I switch to a console, something similar happens. Also, I can change the brightness of the black screen with the iBook's brightness keys. Okay, that sounds like there's no problem with the backlight, but for some reason the LCD is programmed to show only black. I added video=offb:off to video=radeonfb:off like you suggested but got the same result. Did you verify in dmesg that this resulted in OFfb being completely disabled? I.e. no messages from OFfb, in particular nothing about 'conflicting fb hw usage'? Maybe just provide the dmesg output from that again. -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ xorg-driver-ati mailing list xorg-driver-ati@lists.x.org http://lists.x.org/mailman/listinfo/xorg-driver-ati
Bug#683796: xserver-xorg-video-radeon: No 3D acceleration in iBook2 powerpc
On Mit, 2012-08-08 at 01:30 -0700, Dan DeVoto wrote: --- On Tue, 8/7/12, Michel Dänzer daen...@debian.org wrote: Also, adding video=radeonfb:off did nothing but make my screen darker and disable my brightness hotkeys, and there was no change in the error messages, so I removed the line. Note that I meant using video=radeonfb:off instead of radeon.modeset=0, to allow radeon KMS to initialize. Using video=radeonfb:off instead of radeon.modeset=0 resulted in a black screen, even when switching to a console. Can you get the corresponding dmesg output, e.g. via ssh or from /var/log/kern.log? I had to boot from an install cd in rescue mode to revert. Doesn't yaboot allow specifying kernel command line parameters at its prompt, so you don't have to make them permanent for testing? Even if it doesn't, you could use a separate testing entry in yaboot.conf. -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ xorg-driver-ati mailing list xorg-driver-ati@lists.x.org http://lists.x.org/mailman/listinfo/xorg-driver-ati
Bug#683796: xserver-xorg-video-radeon: No 3D acceleration in iBook2 powerpc
On Mit, 2012-08-08 at 01:30 -0700, Dan DeVoto wrote: --- On Tue, 8/7/12, Michel Dänzer daen...@debian.org wrote: Also, adding video=radeonfb:off did nothing but make my screen darker and disable my brightness hotkeys, and there was no change in the error messages, so I removed the line. Note that I meant using video=radeonfb:off instead of radeon.modeset=0, to allow radeon KMS to initialize. Using video=radeonfb:off instead of radeon.modeset=0 resulted in a black screen, even when switching to a console. BTW, does adding radeon.agpmode=-1 avoid the black screen problem? -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ xorg-driver-ati mailing list xorg-driver-ati@lists.x.org http://lists.x.org/mailman/listinfo/xorg-driver-ati
Bug#683796: xserver-xorg-video-radeon: No 3D acceleration in iBook2 powerpc
On Mit, 2012-08-08 at 05:00 -0700, Dan DeVoto wrote: Below is dmesg output with video=radeonfb:off in the kernel command line: [...] Aug 8 04:05:15 icebook kernel: [ 10.769815] checking generic (9c008000 c) vs hw (9800 800) Aug 8 04:05:15 icebook kernel: [ 10.769832] fb: conflicting fb hw usage radeondrmfb vs OFfb ATY,Fall_A - removing generic driver Aug 8 04:05:15 icebook kernel: [ 10.779698] Console: switching to colour dummy device 80x25 Aug 8 04:05:15 icebook kernel: [ 10.782214] checking generic (98808000 5a000) vs hw (9800 800) Aug 8 04:05:15 icebook kernel: [ 10.782235] fb: conflicting fb hw usage radeondrmfb vs OFfb ATY,Fall_B - removing generic driver I heard that this handover from OFfb to radeon KMS can cause problems in some cases. Maybe you can try adding video=offb:off to video=radeonfb:off . Aug 8 04:08:01 icebook kernel: [ 192.105636] [drm:drm_calc_timestamping_constants] *ERROR* crtc 10: Can't calculate constants, dotclock = 0! Otherwise this is the only other sign of any potential problem that I can see, though I think it shouldn't cause a black screen. Maybe you can also check what /sys/class/backlight/radeon_bl/brightness contains when the screen is black and whether writing 128 or other values between 0 and 255 to it makes a difference. -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ xorg-driver-ati mailing list xorg-driver-ati@lists.x.org http://lists.x.org/mailman/listinfo/xorg-driver-ati
Bug#683796: xserver-xorg-video-radeon: No 3D acceleration in iBook2 powerpc
On Die, 2012-08-07 at 04:38 -0700, Dan DeVoto wrote: --- On Mon, 8/6/12, Michel Dänzer daen...@debian.org wrote: [ 36.987427] [drm] radeon defaulting to kernel modesetting. [ 36.987452] [drm] radeon kernel modesetting enabled. You should probably disable KMS with radeon.modeset=0 on the kernel command line. You can also disable radeonfb with video=radeonfb:off, but KMS may not work very well with your card. After adding radeon.modeset=0 to my bootloader, still no change with glxinfo: dan@icebook:~$ glxinfo | grep render direct rendering: Yes OpenGL renderer string: Software Rasterizer GL_EXT_vertex_array_bgra, GL_NV_conditional_render, Though the error messages in my Xorg.0.log file are different, specifically: [39.505] (EE) AIGLX error: radeon does not export required DRI extension Ah, so your version of libgl1-mesa-dri doesn't have non-KMS support. Upstream Mesa dropped it in 8.0. Also, adding video=radeonfb:off did nothing but make my screen darker and disable my brightness hotkeys, and there was no change in the error messages, so I removed the line. Note that I meant using video=radeonfb:off instead of radeon.modeset=0, to allow radeon KMS to initialize. However, note that KMS doesn't support suspend/resume on PowerPC laptops yet. -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ xorg-driver-ati mailing list xorg-driver-ati@lists.x.org http://lists.x.org/mailman/listinfo/xorg-driver-ati
Bug#683796: xserver-xorg-video-radeon: No 3D acceleration in iBook2 powerpc
On Fre, 2012-08-03 at 20:15 -0700, Dan DeVoto wrote: [ 36.987427] [drm] radeon defaulting to kernel modesetting. [ 36.987452] [drm] radeon kernel modesetting enabled. You should probably disable KMS with radeon.modeset=0 on the kernel command line. You can also disable radeonfb with video=radeonfb:off, but KMS may not work very well with your card. -- Earthling Michel Dänzer | http://www.amd.com Libre software enthusiast | Debian, X and DRI developer ___ xorg-driver-ati mailing list xorg-driver-ati@lists.x.org http://lists.x.org/mailman/listinfo/xorg-driver-ati