2) Xen appears to support spin_lock_irq{,save} and
local_irq_{enable,disable}, so I don't know what the user is actually
complaining about. I suppose I could boot my xen kernel and try to
compile/load afs on it...
Well, there is some sort of problem. osi_linux_mask triggers a GPF on xenU kernels, inside SIGMASK_LOCK(), which is the thing that calls spin_lock_irq.
There's an actual cli instruction visible in the disassembly. _that_ is the bug:
00058ad7 <osi_linux_mask>: 58ad7: 53 push %ebx 58ad8: fa cli 58ad9: bb 00 e0 ff ff mov $0xffffe000,%ebx 58ade: 21 e3 and %esp,%ebx 58ae0: 83 43 14 01 addl $0x1,0x14(%ebx)This appears to be because things think ARCH=i686 and not ARCH=xen. if I munge the MakefileProto.LINUX to say ARCH=xen when it calls into kbuild, then it works. When real xen support is added to the tree, a sane way of dealing with this will need to be thought out.
p7sX2lzdJq5U1.p7s
Description: S/MIME cryptographic signature
