On Wed, 9 Nov 2022 14:24:06 +0100, Paul Menzel firstname.lastname@example.org wrote:
Thank you for your patch.
Am 09.11.22 um 14:17 schrieb Xuan Zhuo:
A small nit for the summary/title (subject), maybe use imperative mood *finalize*.
Under the standard of Virtio 1.0, the initialization process of the device must first write sub-features back to device before using device, such as finding vqs.
There are four places using vp_find_vq().
- virtio-blk.pci: put the finalizing features code in front of using device
- virtio-blk.mmio: put the finalizing features code in front of using device
- virtio-scsi.pci: is ok
- virtio-scsi.mmio: add set features before vp_find_vq()
Did you hit an error without the patch? If so, could you please document how to reproduce the issue?
qemu with patch https://email@example.com/msg920538.html
Then: $ build/qemu-system-x86_64 -M accel=kvm -m 1G -cpu host -blockdev file,node-name=drive0,filename=test.img -device virtio-blk-pci,drive=drive0 qemu: queue_enable is only suppported in devices of virtio 1.0 or later.