Jeremy Fitzhardinge wrote: > I tried booting the paravirt patches on a real machine to see what would > happen. It had worked OK under qemu, so I thought it would be worth it. > > It seems to boot OK, though perhaps fairly slowly, but once it hits > usermode it gets into trouble. When starting udevd, the startup script > runs MAKEDEV, which seems to get stuck in an infinite loop in > userspace. It eventually gets past that part of the boot sequence - > very slowly - then oopses at some later point. I haven't tried > dissecting the patches to see where the problem might arise. I'll try > again on a machine with a serial console to see if I can get a better > idea of what's failing. > > The patches I'm talking about are: > > 006-paravirt_header.patch > 007-paravirt-descriptor-ops.patch > 008-paravirt-structure.patch > 009-binary-patch.patch > 011-paravirt-head.S.patch > 012-paravirt-arch-setup.patch > > applied to a 2.6.18-rc2-mm1 tree (with the initial part of the series as > well, of course). > > Any ideas? > Are the oopses you are seeing the nefarious softlockup detection oopses misfiring because of timing issues? Zach
