I'm trying to boot a QEMU VM from a vhost-user-blk-pci device, which appears to be generally possible (https://github.com/spdk/spdk/issues/1728). In my case, vhost gets the image via SPDK's NVMe-oF driver. However, QEMU does not find a bootable device. What I am doing:

1. Start vhost

    bin/vhost -S /var/tmp -r /var/tmp/rpc_ctrl.sock -s 1024 -m 0x3 -A 0000:82:00.1
   

2. Connect to NVMe-oF server and create blk controller

    ./rpc.py -s /var/tmp/rpc_ctrl.sock bdev_nvme_attach_controller -t tcp -a 10.0.0.4 -s 4420 -f ipv4 -n nqn.2024-10.placeholder:bd --name placeholder
   
    ./rpc.py -s /var/tmp/rpc_ctrl.sock vhost_create_blk_controller --cpumask 0x1 vhost.0 placeholdern1
   

3. Attempt to launch QEMU with blk controller as boot device (does not find anything bootable)

    taskset -c 2,3 qemu-system-x86_64 \
        -enable-kvm \
        -m 1G \
        -smp 8 \
        -nographic \
        -object memory-backend-file,id=mem0,size=1G,mem-path=/dev/hugepages,share=on \
        -numa node,memdev=mem0 \
        -chardev socket,id=spdk_vhost_blk0,path=/var/tmp/vhost.0,reconnect=1 \
        -device vhost-user-blk-pci,chardev=spdk_vhost_blk0,bootindex=1,num-queues=2
   

Things I've checked:

• I can mount an NMVe-oF disk to the VM just fine using the same sequence of commands (giving QEMU an additional bootable drive) (just booting from it won't work)
• the image on the NVMe-oF server boots just fine if I provide it locally (via the host-kernel NVMe-oF driver that I can't use in production) and declare it in the QEMU options as a drive
• QEMU does not appear to have an NVMe-oF driver itself that I could use instead (it does have an NVMe driver)

QEMU version 7.2.15 (Debian 1:7.2+dfsg-7+deb12u12)

SPDK version SPDK v25.01-pre git sha1 8d960f1d8

I'm trying to boot a QEMU VM from a vhost-user-blk-pci device, which appears to be generally possible (https://github.com/spdk/spdk/issues/1728). In my case, vhost gets the image via SPDK's NVMe-oF driver. However, QEMU does not find a bootable device. What I am doing:

1. Start vhost

    bin/vhost -S /var/tmp -s 1024 -m 0x3 -A 0000:82:00.1
   

2. Connect to NVMe-oF server and create blk controller

    ./rpc.py bdev_nvme_attach_controller -t tcp -a 10.0.0.4 -s 4420 -f ipv4 -n nqn.2024-10.placeholder:bd --name placeholder
   
    ./rpc.py vhost_create_blk_controller --cpumask 0x1 vhost.0 placeholdern1
   

3. Attempt to launch QEMU with blk controller as boot device (does not find anything bootable)

    taskset -c 2,3 qemu-system-x86_64 \
        -enable-kvm \
        -m 1G \
        -smp 8 \
        -nographic \
        -object memory-backend-file,id=mem0,size=1G,mem-path=/dev/hugepages,share=on \
        -numa node,memdev=mem0 \
        -chardev socket,id=spdk_vhost_blk0,path=/var/tmp/vhost.0,reconnect=1 \
        -device vhost-user-blk-pci,chardev=spdk_vhost_blk0,bootindex=1,num-queues=2
   

Things I've checked:

• I can mount an NMVe-oF disk to the VM just fine using the same sequence of commands (giving QEMU an additional bootable drive) (just booting from it won't work)
• the image on the NVMe-oF server boots just fine if I provide it locally (via the host-kernel NVMe-oF driver that I can't use in production) and declare it in the QEMU options as a drive
• QEMU does not appear to have an NVMe-oF driver itself that I could use instead (it does have an NVMe driver)

QEMU version 7.2.15 (Debian 1:7.2+dfsg-7+deb12u12)

SPDK version SPDK v25.01-pre git sha1 8d960f1d8