UseFBDev option causes Dapper ppc live cd SIGSEGV initializing X

bootstrap.log contains a lot of sayings about broken dependencies between packages...

I just tried Xubuntu live cd in case there was an out of memory situation and booted it with the same kernel and the video=ofonly parameter. Both 'startx' and 'startxfce' yields the same SIGSEGV fault in the Rage 128 driver related to the mmio call.

The system doesn't die however so I now got a prompt to play around with and I guess 128 Mb actually is too little for Ubuntu Desktop after all. Maybe a graceful detection of this would be nice instead of a system lockup. For the iMac I continue to use Xubuntu but I will try the Ubuntu-server live cd as well.

More findings: If I edit /etc/X11/xorg.conf and do
         Option "UseFBDev" "false"
 in the Device section for "ATI Technologies, Inc. Rage 128 RL/VR AGP" instead of setting the option to "true" the bug doesn't occur. I have also found some other reports on the net for this bug in various stages of the Xfree86 and X.org development of this driver. The same behavior is found also using the the "r128" driver directly instead of indirectly through the "ati" driver.

So the bug is clearily in the FBDev part of the driver

Status now is that the driver do not display on the internal iMac CRT but I got picture through the external VGA output. However that has nothing todo with the SIGSEGV bug found, but I am not stuck at it anymore and I am not looking for support here... ;-)

After installing from the Xubuntu livecd I updated the 79(!) packages neeed to get up to date I found out that the kernel used is older than the livecd one:

X Window System Version 7.0.0
Release Date: 21 December 2005
X Protocol Version 11, Revision 0, Release 7.0
Build Operating System:Linux 2.6.15-26-powerpc64-smp ppc
Current Operating System: Linux jol-desktop 2.6.15-23-powerpc #1 Tue May 23 13:46:54 UTC 2006 ppc

The livecd kernel is not 2.6.15-23 but 2.6.15-26 (See above log) and with this older (cut -23) the Option UseFBDev "true" works!

So I think the framebuffer driver has changed between these cuts and caused the SIGSEGV crash. I am currently not at the CVS so I can't check the diff myself yet.

Ok, so I tracked down some differences from the dmesg logs between the two kernel versions and it seems apparent that a different framebuffer device driver is used in the live CD kernel. I attach the intalled kernel (2.6.15-23) log for reference against the already attached log from the live CD kernel (2.6.15-26) The interesting lines are:

< [ 0.000000] Linux version 2.6.15-23-powerpc (buildd@ross) (gcc version 4.0.3 (Ubuntu 4.0.3-1ubuntu5)) #1 Tue May 23 13:46:54 UTC 2006
> [ 0.000000] Linux version 2.6.15-26-powerpc (buildd@royal) (gcc version 4.0.3 (Ubuntu 4.0.3-1ubuntu5)) #1 Thu Aug 3 02:53:39 UTC 2006

and

< [ 32.329540] aty128fb: Invalid ROM signature 8181 should be 0xaa55
< [ 32.329568] aty128fb: BIOS not located, guessing timings.
< [ 32.329592] aty128fb: Rage128 RL AGP [chip rev 0x2] 8M 64-bit SDR SGRAM (2:1)
< [ 32.364369] Console: switching to colour frame buffer device 128x48
< [ 32.364402] fb0: ATY Rage128 frame buffer device on Rage128 RL AGP

> [ 65.884727] Using unsupported 1024x768 ATY,Rage128k at 94000000, depth=8, pitch=1024
> [ 65.919532] Console: switching to colour frame buffer device 128x48
> [ 65.919556] fb0: Open Firmware frame buffer device on /pci@f0000000/ATY,Rage128k@10

The live CD kernel uses some other device driver that isn't working with the UseFBDev option "true" as it seems.

Narrowed down the bug a bit

see https://bugs.launchpad.net/ubuntu/+bug/86066

This bug has had no activity for a considerable period. This is a check to see if there is still interest in investigating this bug report.

We are closing this bug report because it lacks the information we need to investigate the problem, as described in the previous comments. Please reopen it if you can give us the missing information, and don't hesitate to submit bug reports in the future. To reopen the bug report you can click on the current status, under the Status column, and change the Status back to "New". Thanks again!