Re: [Qemu-devel] Bug: SCSI/VirtIO: VM unable to reboot from non-IDE controller

[Ryan Harper]

* Alexey Eremenko <address@hidden> [2009-04-20 04:56]:
> Hi All !
> 
> Most Linux VMs (Ubuntu and Red Hat-based systems -- RHEL, Fedora) have
> failed during automated tests
> on SCSI and VirtIO disks on Qemu/KVM.
> It turned out to be a problem of Qemu/KVM, where after a software reboot
> (i.e. initiated by guest OS), the system is unable to boot from SCSI
> controller.
> 
> To reboot successfully you need either soft reboot+IDE or Hard reboot (Qemu
> cold boot)+SCSI.
> 
> The Command sent to Qemu/KVM:
> /usr/local/bin/qemu-system-x86_64 -m 512 -monitor
> tcp:localhost:4602,server,nowait -cdrom /isos/linux/Fedora-8-i386-DVD.iso
> -drive file=/vm/fedora8-32.qcow2,if=scsi,boot=on -name fedora8-32

You're using extboot (boot=on) which is part of kvm.  I sent a patch[1] 
to since the upstream kvm-userspace bits had the wrong option rom reset
address.  This should be fixed in upstream kvm-userspace.git[2].

1. http://kerneltrap.org/mailarchive/linux-kvm/2009/4/6/5418914
2. 
http://git.kernel.org/?p=linux/kernel/git/avi/kvm-userspace.git;a=commit;h=a1075de527f309850df278484f2ef4127827c6f4

-- 
Ryan Harper
Software Engineer; Linux Technology Center
IBM Corp., Austin, Tx
address@hidden