After my IRC proxy VM was updated to kernel 2.6.40-4, the next reboot was a horror show. After several attempts I just couldn't make the system boot at all; it'd get stuck trying to mount a CIFS share, with a gpf trace showing. Looking through the console output I saw some stuff about selinux, so after a few failed attempts at rebooting, I used virt-edit to disable selinux and set the share to noauto. After that the system booted, but hit another gpf when I brought up a connection to a VPN, and was unresponsive after that. I then flipped it back to 2.6.38.8-35.fc15.x86_64; with that kernel it seems to be running fine. I'll attach the gpfs I was able to get out of the logs, I think there were more but they didn't all get logged.
Created attachment 516957 [details] first gpf (this is during boot, I think, with cifs mount and selinux enabled)
Created attachment 516958 [details] second gpf (I think this is on bringing up the VPN after booting with selinux and cifs share disabled)
I have a hunch. Can you try booting with vdso=0 (this is for the gpf thing, the cifs bug is likely a dupe of another bug).
will try to check this soon, but it's below Alpha stuff on my todo list - just for info. i have to find a bit of a quiet time to do it as while i'm screwing with this, i'm not on IRC :)
Guh. Sorry, I suck. Will re-test this right now just to knock it off my list.
well, after some fun libvirt drama on the host, I booted with 2.6.40-5 with vdso=0 and it worked fine...dunno if it's -5, vdso=0, or something else. even mounting the CIFS share doesn't seem to cause any trouble.
Pretty sure Adam isn't on F15 anymore, and it's not clear if an update or vdso=0 fixed things. Closing this out as WORKSFORME since F15 is going EOL.