[vfio-users] GPU passthrough to Windows problem: BOOT_DEVICE_INACCESSIBLE & general questions on passing physical disks

Blank Field ihatethisfield at gmail.com
Wed Oct 28 04:34:39 UTC 2015


This isn't a vfio related question, but i'll answer.
Virtio scsi and virtio blk pci disk controllers do not exist in real world,
so the drivers for them is not included in any windows release.
You will need to install them manually from so called virtio.iso. There is
a git version of it somewhere, but you can use whatever you find.
The problem is - how to install it without being able to boot?
The answer is rather simple: remember that "press f2 if a special scsi
driver is required" line during the installation of windows XP? An
equivalent of that menu exists in new releases: during the setup when you
choose where to install to there is a  button which allows you to read
drivers from disk. Apply pressure to it, find your virtio.iso, load
anything you find there.
I can't tell you how to fix this issue on an already installed OS since i
don't remember all the button in windows, but the google keyword is "bsod
7b".
On Oct 28, 2015 6:54 AM, "Benjamin Slade" <slade at jnanam.net> wrote:

> I've been trying to set up a machine running Arch which allows Windows
> to be run as a virtualised process with GPU-passthrough according to the
> method laid out here:
>
> https://www.reddit.com/r/linuxmasterrace/comments/3lnorm/gpu_passthrough_revisited_an_updated_guide_on_how/
>
> The actual passthrough part I've navigated, but I'm having trouble
> figuring out how to properly pass devices through to the VM. Here's what
> I have:
>
>   #!/bin/bash
>
>   cp /usr/share/edk2.git/ovmf-x64/OVMF_VARS-pure-efi.fd /tmp/my_vars.fd
>   QEMU_PA_SAMPLES=128 QEMU_AUDIO_DRV=pa
>   qemu-system-x86_64 \
>   -enable-kvm \
>  -m 2048 \
>  -smp cores=4,threads=1 \
>  -cpu host,kvm=off \
>  -vga none \
>  -soundhw hda \
>  -usb -usbdevice host:046d:c52b -usbdevice host:0781:5530 \
>  -device vfio-pci,host=02:00.0,multifunction=on \
>  -device vfio-pci,host=02:00.1 \
>  -drive
> if=pflash,format=raw,readonly,file=/usr/share/edk2.git/ovmf-x64/OVMF_CODE-pure-efi.fd
> \
>  -drive if=pflash,format=raw,file=/tmp/my_vars.fd \
>  -device virtio-scsi-pci,id=scsi \
>  -drive file=/dev/sda,id=disk,format=raw,if=none -device ide-hd,drive=disk
> \
>  -drive
> file=/home/emacsomancer/kvm/virt-stable.iso,id=virtiocd,if=none,format=raw
> -device ide-cd,bus=ide.1,drive=virtiocd \
>  -drive file=/dev/sdc,id=disk2,format=raw,if=none -device
> scsi-hd,drive=disk2 \
>  -drive file=/dev/sdd,id=disk3,format=raw,if=none -device
> scsi-hd,drive=disk3 \
>  -boot menu=on
>
> When I try this, it ends up booting off the RedHat VirtIO .iso, and
> eventually the Windows symbol appears and the spin-y circle spins for a
> while, and then it complains "BOOT_DEVICE_INACCESSIBLE".
>
> So, It seems unable to see the other drives I've passed. These are
> physical hard drives (I'm not sure if I'm passing them correctly,
> esp. as the VM doesn't seem to be seeing them). They are indeed on
> /dev/sda, /dev/sdc, and /dev/sdd, respectively. I can properly boot off
> of /dev/sda/ where Windows is installed.
>
> I've fiddled with this for quite some time, and haven't made any
> headway.
>
> _______________________________________________
> vfio-users mailing list
> vfio-users at redhat.com
> https://www.redhat.com/mailman/listinfo/vfio-users
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://listman.redhat.com/archives/vfio-users/attachments/20151028/a4dd9659/attachment.htm>


More information about the vfio-users mailing list