From: Dongli Zhang <dongli.zhang@xxxxxxxxxx> [ Upstream commit 1978f30a87732d4d9072a20abeded9fe17884f1b ] When tag_set->nr_maps is 1, the block layer limits the number of hw queues by nr_cpu_ids. No matter how many hw queues are used by virtio-scsi, as it has (tag_set->nr_maps == 1), it can use at most nr_cpu_ids hw queues. In addition, specifically for pci scenario, when the 'num_queues' specified by qemu is more than maxcpus, virtio-scsi would not be able to allocate more than maxcpus vectors in order to have a vector for each queue. As a result, it falls back into MSI-X with one vector for config and one shared for queues. Considering above reasons, this patch limits the number of hw queues used by virtio-scsi by nr_cpu_ids. Reviewed-by: Stefan Hajnoczi <stefanha@xxxxxxxxxx> Signed-off-by: Dongli Zhang <dongli.zhang@xxxxxxxxxx> Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> Signed-off-by: Kunkun Jiang <jiangkunkun@xxxxxxxxxx> --- drivers/scsi/virtio_scsi.c | 1 + 1 file changed, 1 insertion(+) diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c index 50e87823baab..0b45424e7ca5 100644 --- a/drivers/scsi/virtio_scsi.c +++ b/drivers/scsi/virtio_scsi.c @@ -853,6 +853,7 @@ static int virtscsi_probe(struct virtio_device *vdev) /* We need to know how many queues before we allocate. */ num_queues = virtscsi_config_get(vdev, num_queues) ? : 1; + num_queues = min_t(unsigned int, nr_cpu_ids, num_queues); num_targets = virtscsi_config_get(vdev, max_target) + 1; -- 2.33.0
