On 21 October 2014 22:15, Richard Jones <rjo...@redhat.com> wrote: > I have now also tried virtio-blk and that doesn't work either. Same > symptoms: no log messages at all (even with ignore_loglevel), and no > disks appear. > >> Yeah, "regressed with this newer kernel" sounds more like a kernel >> bug than a QEMU bug to me, especially if all the other virt devices >> still work. > > It seems like no virtio drivers work, but it's very hard to tell when > your guest has no disks and therefore no operating system. How can I > debug this further?
Oh. I assumed from the fact your commandline had other virtio devices and you were only complaining about virtio-scsi that it was a single-backend issue. Suggestions: * bisect the kernel to find out when it broke * add a bunch of debug printks to the kernel to find out why it's not finding the disks. The logging here is terrible IMHO: the kernel usually detects a specific problem and then throws all this info away in favour of just silently deciding there's no hardware there thanks -- PMM
