]> git.proxmox.com Git - mirror_qemu.git/commit
virtio-blk: always set ioeventfd during startup
authorStefan Hajnoczi <stefanha@redhat.com>
Fri, 19 Jan 2024 13:57:48 +0000 (08:57 -0500)
committerKevin Wolf <kwolf@redhat.com>
Fri, 26 Jan 2024 10:16:58 +0000 (11:16 +0100)
commitd3f6f294aeadd5f88caf0155e4360808c95b3146
tree23bc6dd87d4d830dd33ccfbff0be69d794651074
parentea0736d7f84ead109a6b701427991828f97724c3
virtio-blk: always set ioeventfd during startup

When starting ioeventfd it is common practice to set the event notifier
so that the ioeventfd handler is triggered to run immediately. There may
be no requests waiting to be processed, but the idea is that if a
request snuck in then we guarantee that it will be detected.

One scenario where self-triggering the ioeventfd is necessary is when
virtio_blk_handle_output() is called from a vCPU thread before the
VIRTIO Device Status transitions to DRIVER_OK. In that case we need to
self-trigger the ioeventfd so that the kick handled by the vCPU thread
causes the vq AioContext thread to take over handling the request(s).

Fixes: b6948ab01df0 ("virtio-blk: add iothread-vq-mapping parameter")
Reported-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Message-ID: <20240119135748.270944-7-stefanha@redhat.com>
Reviewed-by: Kevin Wolf <kwolf@redhat.com>
Signed-off-by: Kevin Wolf <kwolf@redhat.com>
hw/block/virtio-blk.c