Bug#679312: xserver-xorg-video-radeon: Rotation fails due to Acceleration init failed, but DRI works nevertheless

2012-06-28 Thread Michel Dänzer
On Mit, 2012-06-27 at 22:24 +0200, Johannes Baiter wrote: 
 
 This morning I rebooted my machine for the first time in ~2 weeks (I usually
 just suspend to RAM).
 On rebooting with the exact same kernel, X.org will not start unless I remove
 the options to rotate my two screens from xorg.conf:
[...]
 [  5945.274] (EE) RADEON(0): Acceleration required for rotation

[...]

 Full log: http://paste.debian.net/176669/
 (Sorry for not attaching it, I can't figure out how to attach a file
 with reportbug in text mode, [...]

'a' at the prompt after writing the report, but it was included by the
package bug script anyway.


 OpenGL renderer string: Gallium 0.4 on llvmpipe (LLVM 0x209)

llvmpipe is software rendering, not hardware acceleration.


 This wouldn't be all that bad, if I could still rotate my displays via XRandR,
 but this fails, too:
 
 ~ ◆ xrandr --output HDMI-0 --rotate right
 xrandr: Configure crtc 0 failed

Same reason: Rotation requires acceleration.


 DRM Information from dmesg:
 ---
 [0.00] No AGP bridge found
 [0.00] No AGP bridge found
 [0.463471] Linux agpgart interface v0.103

Looks like something is preventing the radeon kernel driver from
loading / initializing. Is it built into the kernel or as a module? If
the latter, what happens if you try to run 'sudo modprobe radeon' from
console (ideally before X starts)?


-- 
Earthling Michel Dänzer   |   http://www.amd.com
Libre software enthusiast |  Debian, X and DRI developer



--
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#679312: xserver-xorg-video-radeon: Rotation fails due to Acceleration init failed, but DRI works nevertheless

2012-06-28 Thread terayamashuji
2012/6/28 Michel Dänzer daen...@debian.org

 On Mit, 2012-06-27 at 22:24 +0200, Johannes Baiter wrote:

 'a' at the prompt after writing the report, but it was included by the
 package bug script anyway.


The log that was auto-attached was actually the log without rotation, not
that it mattered greatly :-)




  OpenGL renderer string: Gallium 0.4 on llvmpipe (LLVM 0x209)

 llvmpipe is software rendering, not hardware acceleration.


 Ah, that explains a lot! Somehow my brain was hard-wired into associating
lack of acceleration with the MESA software driver, I haven't been using
the open source radeon driver for a long time :-)



  DRM Information from dmesg:
  ---
  [0.00] No AGP bridge found
  [0.00] No AGP bridge found
  [0.463471] Linux agpgart interface v0.103

 Looks like something is preventing the radeon kernel driver from
 loading / initializing. Is it built into the kernel or as a module? If
 the latter, what happens if you try to run 'sudo modprobe radeon' from
 console (ideally before X starts)?


That did the trick! Turns out the radeon module was not loaded at
boot-time, probably it was not in the initrd. Loading the module and
starting X.org with rotation enabled works now, as does XRandR.
Sorry for opening a silly bug like this, but maybe this could be
considered a real-world use-case for more explicit error messages ;-)

Thank you!


Bug#679312: xserver-xorg-video-radeon: Rotation fails due to Acceleration init failed, but DRI works nevertheless

2012-06-28 Thread Julien Cristau
On Thu, Jun 28, 2012 at 20:11:57 +0200, terayamashuji wrote:

 That did the trick! Turns out the radeon module was not loaded at
 boot-time, probably it was not in the initrd. Loading the module and
 starting X.org with rotation enabled works now, as does XRandR.
 Sorry for opening a silly bug like this, but maybe this could be
 considered a real-world use-case for more explicit error messages ;-)

It's not supposed to be in the initrd, udev is supposed to load it after
the system's switched to the real root.

Cheers,
Julien


signature.asc
Description: Digital signature