Bug 1935019

Is anyone in cc: able to tell if this feature is/should be supported?

(In reply to qing.wang from comment #0)

> 
> Actual results:
> guest boot failed
> Expected results:
> if this feature not support it should removed from the options or give user
> friendly message.

If this is just a documentation/usage info issue, it shouldn't be set to high severity/high priority imho.

Kevin, anything from your side?

The bug on the QEMU side is that it didn't check whether the backend (in the external process) actually supports the flag. This is fixed as of upstream commit 04ceb61a. I'd suggest backporting this along with other similar missing checks for a nicer failure mode, though as vhost-user-blk is still tech preview, in theory waiting for the next rebase wouldn't be a big problem either.

I don't know whether we want to support the flag in the vhost-user-blk export eventually, but it's not a priority and would be an RFE.

QA_ACK, please?

QE bot(pre verify): Set 'Verified:Tested,SanityOnly' as gating/tier1 test pass.

Tested on
Red Hat Enterprise Linux release 8.5 Beta (Ootpa)
4.18.0-315.el8.x86_64
qemu-kvm-common-6.0.0-24.module+el8.5.0+11844+1e3017bd.x86_64

The issue get fixed, no crash happened.

1.boot qsd
qemu-storage-daemon \
  --chardev socket,path=/tmp/qmp.sock,server=on,wait=off,id=char1 \
  --monitor chardev=char1 \
  --object iothread,id=iothread0 \
  --blockdev driver=file,node-name=file1,filename=/home/kvm_autotest_root/images/disk1.qcow2 \
  --blockdev driver=qcow2,node-name=fmt1,file=file1 \
  --export type=vhost-user-blk,id=export1,addr.type=unix,addr.path=/tmp/vhost-user-blk1.sock,node-name=fmt1,writable=on,iothread=iothread0  \
  --blockdev driver=file,node-name=file2,filename=/home/kvm_autotest_root/images/disk2.qcow2 \
  --blockdev driver=qcow2,node-name=fmt2,file=file2 \
  --export type=vhost-user-blk,id=export2,addr.type=unix,addr.path=/tmp/vhost-user-blk2.sock,node-name=fmt2,writable=on,iothread=iothread0  \



2.boot vm with host-user-blk-pci device

/usr/libexec/qemu-kvm -enable-kvm \
  -m 4G -M accel=kvm,memory-backend=mem \
  -nodefaults \
  -vga qxl \
  -smp 4 \
  -object memory-backend-memfd,id=mem,size=4G,share=on \
  -device pcie-root-port,id=pcie-root-port-0,multifunction=on,bus=pci.0,addr=0x3,chassis=1 \
  -device pcie-root-port,id=pcie-root-port-1,port=0x1,addr=0x3.0x1,bus=pci.0,chassis=2 \
  -device pcie-root-port,id=pcie-root-port-2,port=0x2,addr=0x3.0x2,bus=pci.0,chassis=3 \
  -device pcie-root-port,id=pcie-root-port-3,port=0x3,addr=0x3.0x3,bus=pci.0,chassis=4 \
  -device pcie-root-port,id=pcie-root-port-4,port=0x4,addr=0x3.0x4,bus=pci.0,chassis=5 \
  -device pcie-root-port,id=pcie-root-port-5,port=0x5,addr=0x3.0x5,bus=pci.0,chassis=6 \
  -device pcie-root-port,id=pcie-root-port-6,port=0x6,addr=0x3.0x6,bus=pci.0,chassis=7 \
  -device pcie-root-port,id=pcie-root-port-7,port=0x7,addr=0x3.0x7,bus=pci.0,chassis=8 \
  -device qemu-xhci,id=usb1,bus=pcie-root-port-1,addr=0x0 \
  -device usb-tablet,id=usb-tablet1,bus=usb1.0,port=1 \
  -device virtio-scsi-pci,id=scsi0,bus=pcie-root-port-5 \
  -device virtio-scsi-pci,id=scsi1,bus=pcie-root-port-6 \
  -blockdev driver=qcow2,file.driver=file,file.filename=/home/kvm_autotest_root/images/rhel840-64-virtio-scsi.qcow2,node-name=os_image1 \
  -device virtio-blk-pci,id=blk0,drive=os_image1,bootindex=0 \
  \
  -chardev socket,path=/tmp/vhost-user-blk1.sock,id=vhost1 \
  -device vhost-user-blk-pci,chardev=vhost1,id=blk1,bootindex=1,num-queues=1,iommu_platform=off \
  \
  -vnc :5 \
  -monitor stdio \
  -qmp tcp:0:5955,server,nowait

3.login guest, check the disk exist and io 
dd if=/dev/zero of=/dev/vdb count=100 bs=1M oflag=direct

if we set iommu_platform=on boot failed :
iommu_platform=true is not supported by the device.


but the help message 
/usr/libexec/qemu-kvm -device vhost-user-blk-pci,?
vhost-user-blk-pci options:
  acpi-index=<uint32>    -  (default: 0)
  addr=<int32>           - Slot and optional function number, example: 06.0 or 06 (default: -1)
  aer=<bool>             - on/off (default: false)
  any_layout=<bool>      - on/off (default: true)
  ats=<bool>             - on/off (default: false)
  bootindex=<int32>
  chardev=<str>          - ID of a chardev to use as a backend
  class=<uint32>         -  (default: 0)
  config-wce=<bool>      - on/off (default: true)
  disable-legacy=<OnOffAuto> - on/off/auto (default: "auto")
  disable-modern=<bool>  -  (default: false)
  event_idx=<bool>       - on/off (default: true)
  failover_pair_id=<str>
  indirect_desc=<bool>   - on/off (default: true)
  iommu_platform=<bool>  - on/off (default: false)

What do you think fire a new bug to request hide  "iommu_platform=<bool>  - on/off (default: false)" since it is not supported.

(In reply to qing.wang from comment #11)
> What do you think fire a new bug to request hide  "iommu_platform=<bool>  -
> on/off (default: false)" since it is not supported.

You are right that there is currently no way to make iommu_platform=on work with qemu-storage-daemon. Though in theory, you could connect to other vhost-user-blk backends that do support the option, so I think keeping it visible on the QEMU side is more correct.

Also passed on
Red Hat Enterprise Linux release 9.0 Beta (Plow)
5.14.0-0.rc6.46.el9.x86_64
qemu-kvm-common-6.0.0-12.el9.x86_64

Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.

For information on the advisory (virt:av bug fix and enhancement update), and where to find the updated
files, follow the link below.

If the solution does not work for you, open a new bug report.

https://access.redhat.com/errata/RHBA-2021:4684