Hi MST and paolo, mq-deadline is good for slow media, and none is good for high-speed media. It depends on how the community views this issue. When virtio-blk adopts multi-queue,it automatically changes from deadline to none, which is not uniform here. I don't have ideas right now to answer Christoph/Paolo's question. Thanks, Li > On Dec 25, 2023, at 22:20, Michael S. Tsirkin <mst@xxxxxxxxxx> wrote: > > On Thu, Dec 07, 2023 at 12:31:05PM +0800, Li Feng wrote: >> virtio-blk is generally used in cloud computing scenarios, where the >> performance of virtual disks is very important. The mq-deadline scheduler >> has a big performance drop compared to none with single queue. In my tests, >> mq-deadline 4k readread iops were 270k compared to 450k for none. So here >> the default scheduler of virtio-blk is set to "none". >> >> Signed-off-by: Li Feng <fengli@xxxxxxxxxx> > > I dropped this for now, pls try to address comments by Christoph/Paolo > if it's still needed > >> --- >> drivers/block/virtio_blk.c | 2 +- >> 1 file changed, 1 insertion(+), 1 deletion(-) >> >> diff --git a/drivers/block/virtio_blk.c b/drivers/block/virtio_blk.c >> index d53d6aa8ee69..5183ec8e00be 100644 >> --- a/drivers/block/virtio_blk.c >> +++ b/drivers/block/virtio_blk.c >> @@ -1367,7 +1367,7 @@ static int virtblk_probe(struct virtio_device *vdev) >> vblk->tag_set.ops = &virtio_mq_ops; >> vblk->tag_set.queue_depth = queue_depth; >> vblk->tag_set.numa_node = NUMA_NO_NODE; >> - vblk->tag_set.flags = BLK_MQ_F_SHOULD_MERGE; >> + vblk->tag_set.flags = BLK_MQ_F_SHOULD_MERGE | BLK_MQ_F_NO_SCHED_BY_DEFAULT; >> vblk->tag_set.cmd_size = >> sizeof(struct virtblk_req) + >> sizeof(struct scatterlist) * VIRTIO_BLK_INLINE_SG_CNT; >> -- >> 2.42.0 >
