This patch adds support for virtio-scsi devices. Added block variants: * virtio_scsi (default) * virtio_scsi_disk (legacy scsi) * virtio_scsi_block (only raw /dev/sd* files) * virtio_scsi_generic (only raw /dev/sg* files)
Please keep in mind that virtio-scsi is not yet upstream. To test this patch you have to: 1) have a guest OS with virtio-scsi support ( https://github.com/bonzini/virtio-scsi/commits/master ) 2) qemu support for virtio-scsi devices ( git://github.com/bonzini/qemu.git ) 3) qemu bios support (http://people.redhat.com/pbonzini/virtio-scsi/ ) NOTE: you can download testing rpm packages for all of the above on Paolo Bonzini's page http://people.redhat.com/pbonzini/virtio-scsi/ NOTE2: without modified qemu bios you won't be able to boot from virtio-scsi device NOTE3: currently only boot from lun0 is supported NOTE4: automatic hotplug that doesn't work. You have to initialize the device (echo "scsi add-single-device" 2 0 1 0 > /proc/scsi/scsi ) Also please be aware that virtio_scsi_block and virtio_scsi_generic devices are pass-through targets thus you can specify only /dev/sg (resp. /dev/sd*) devices. Also in my version of qemu it was impossible to set those targets directly from cmd-line, although hot-plug worked fine. Pull request: https://github.com/autotest/autotest/pull/170 Regards, Lukáš Doktor Signed-off-by: Lukas Doktor [email protected] _______________________________________________ Autotest mailing list [email protected] http://test.kernel.org/cgi-bin/mailman/listinfo/autotest
