In one case, 2.6.0-test4 is miscompiled with 3.3.1. Should be fixed in 3.3.1-2mdk but I don't know whether this would fix your bug or not.
So is the cause for it panicking on boot? So presumably it will be fixed
by a reboot before 9.2 is released?
No, this means I was aware of one instance of a bug in gcc 3.3.1 that miscompiled kernel 2.6-test4, at least some certain usages of __builtin_expect. You can try to rebuild the current kernel SRPM with latest gcc and see how things go.
Bye, Gwenole.
