Hi.

There seem to be the grave bug in that
kernel. When either process is trying
to access the cdrom, it gets stuck in
the kernel forever. Even "eject" locks
up. This started to happen somewhere
around -mm4 I think. Specifying
"pci=routeirq" makes the kernel to
deadlock at boot, right after detecting
the cdrom. Sometimes there are the
"ide1: lost interrupt" messages on console.
Attached are the different bits of info
I found usefull for that report. It
includes the Alt-PrtSc-t trace and the
random info about my cdrom.
Any suggestions on getting this fixed?
This one is really severe I suppose.