Greetings Peter, > Sent: Tuesday, October 20, 2020 at 3:22 PM > From: "Peter Krempa" <pkrempa@xxxxxxxxxx> > To: "daggs" <daggs@xxxxxxx> > Cc: libvir-list@xxxxxxxxxx > Subject: Re: [PATCH 0/6] qemu: Fix cdrom as SCSI hostdev via -blockdev > > Oh, so your guest OS doesn't seem to recognize virtio-scsi apparently! > > On my test VM I've got the following: > > $ ls -lia /sys/bus/scsi/devices > total 0 > 5659 drwxr-xr-x. 2 root root 0 Oct 19 17:34 . > 5657 drwxr-xr-x. 4 root root 0 Oct 19 17:34 .. > 18418 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 0:0:0:0 -> ../../../devices/pci0000:00/0000:00:1f.2/ata1/host0/target0:0:0/0:0:0:0 > 20322 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 6:0:0:1 -> ../../../devices/pci0000:00/0000:00:02.7/0000:08:00.0/virtio6/host6/target6:0:0/6:0:0:1 > 19773 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 6:0:0:3 -> ../../../devices/pci0000:00/0000:00:02.7/0000:08:00.0/virtio6/host6/target6:0:0/6:0:0:3 > 15523 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host0 -> ../../../devices/pci0000:00/0000:00:1f.2/ata1/host0 > 15571 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host1 -> ../../../devices/pci0000:00/0000:00:1f.2/ata2/host1 > 15619 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host2 -> ../../../devices/pci0000:00/0000:00:1f.2/ata3/host2 > 15667 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host3 -> ../../../devices/pci0000:00/0000:00:1f.2/ata4/host3 > 15715 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host4 -> ../../../devices/pci0000:00/0000:00:1f.2/ata5/host4 > 15763 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host5 -> ../../../devices/pci0000:00/0000:00:1f.2/ata6/host5 > 19403 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 host6 -> ../../../devices/pci0000:00/0000:00:02.7/0000:08:00.0/virtio6/host6 > 18376 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 target0:0:0 -> ../../../devices/pci0000:00/0000:00:1f.2/ata1/host0/target0:0:0 > 19733 lrwxrwxrwx. 1 root root 0 Oct 19 17:34 target6:0:0 -> ../../../devices/pci0000:00/0000:00:02.7/0000:08:00.0/virtio6/host6/target6:0:0 > > You can see that host6 is a virtio-scsi controller and it has targets. > > You need to figure out why your guest doesn't have the driver. > Alternative would be to use a different scsi controller model, but any > emulated will result in overhead, so the best solution will be to get > drivers. > > Sorry I've mislead you with the initial red herring, but your report > really came after I've modified how we do hostdevs so I thought I messed > up something. > > that might consistent with the fact that I had to use e1000e for the nic because virtio didn't worked as should. the docs states the model one of 'auto', 'buslogic', 'ibmvscsi', 'lsilogic', 'lsisas1068', 'lsisas1078', 'virtio-scsi', 'vmpvscsi', 'virtio-transitional', 'virtio-non-transitional'. any recommendation on which to use? DAgg.
