virtio-blk-pci: default num_queues to -smp N
commit9445e1e15e66c19e42bea942ba810db28052cd05
authorStefan Hajnoczi <stefanha@redhat.com>
Tue, 18 Aug 2020 14:33:47 +0000 (18 15:33 +0100)
committerMichael S. Tsirkin <mst@redhat.com>
Thu, 27 Aug 2020 12:29:13 +0000 (27 08:29 -0400)
tree04f90ede63d0f327c660fefbbc45f04c610ec429
parent6a558822849fab604a1fcb71e7c4e528cafa21d9
virtio-blk-pci: default num_queues to -smp N

Automatically size the number of virtio-blk-pci request virtqueues to
match the number of vCPUs.  Other transports continue to default to 1
request virtqueue.

A 1:1 virtqueue:vCPU mapping ensures that completion interrupts are
handled on the same vCPU that submitted the request.  No IPI is
necessary to complete an I/O request and performance is improved.  The
maximum number of MSI-X vectors and virtqueues limit are respected.

Performance improves from 78k to 104k IOPS on a 32 vCPU guest with 101
virtio-blk-pci devices (ioengine=libaio, iodepth=1, bs=4k, rw=randread
with NVMe storage).

Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Reviewed-by: Cornelia Huck <cohuck@redhat.com>
Reviewed-by: Pankaj Gupta <pankaj.gupta.linux@gmail.com>
Message-Id: <20200818143348.310613-7-stefanha@redhat.com>
Reviewed-by: Michael S. Tsirkin <mst@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>
hw/block/virtio-blk.c
hw/core/machine.c
hw/virtio/virtio-blk-pci.c
include/hw/virtio/virtio-blk.h