Hi Jonathan,
Hi Mike, Mike Hore wrote:It was simpler than I thought -- I've got those files for you now, so I'll attach the zipped archive.Nice. From inittab: co:2345:respawn:/sbin/getty hvc0 9600 linuxFrom the installer log:finish-install: Setting up virtualized serial console on /dev/hvc0 finish-install: adding console to /etc/inittab That is bug #636040 (finish-install: hvc0 console added to inittab even when not available), which seems to be mostly harmless. It should be safe to remove the "co:" line to remove the message about spawning getty there.
Are you suggesting I do that? Sorry, I have a software development background but not on Linux, so I don't know where that line is to remove it.
From syslog:gdm3[1777]: WARNING: GdmDisplay: display lasted 0.038594 seconds gdm3[1777]: WARNING: GdmDisplay: display lasted 0.039019 seconds gdm3[1777]: WARNING: GdmDisplay: display lasted 0.038234 seconds [...] That is probably gdm or the X server coping poorly with the radeon driver failing to load. Please add "1" or "single" (as a separate word, no quotes) to the kernel command line at bootup so we can see if a non-graphical bootup works.
OK, same problem. Where and what is that "kernel command line" so I can add to it?
If it works, then it is probably worth trying to log in on the console and running "modprobe radeon" and then sending logs (/var/log/dmesg) and a photograph of the artifacts.
Yes, I can do that if I get that far.
netconsole[1] or a serial console[2] could help to recover messages if they aren't getting logged properly before rebooting. [1] http://www.kernel.org/doc/Documentation/networking/netconsole.txt [2] http://www.kernel.org/doc/Documentation/serial-console.txt .
Thanks -- if you can point me in the right direction I'll try your suggestions.
Cheers, Mike.
---------------------------------------------------------------
Mike Hore mike_hore@aapt.net.au
---------------------------------------------------------------