It is observed on null_blk that IOPS can be improved much by simply making hw queue per NUMA node, so this patch applies the introduced .host_tagset for improving performance.
In reality, .can_queue is quite big, and NUMA node number is often small, so each hw queue's depth should be high enough to saturate device. Cc: Arun Easi <[email protected]> Cc: Omar Sandoval <[email protected]>, Cc: "Martin K. Petersen" <[email protected]>, Cc: James Bottomley <[email protected]>, Cc: Christoph Hellwig <[email protected]>, Cc: Don Brace <[email protected]> Cc: Kashyap Desai <[email protected]> Cc: Peter Rivera <[email protected]> Cc: Laurence Oberman <[email protected]> Cc: Hannes Reinecke <[email protected]> Cc: Mike Snitzer <[email protected]> Signed-off-by: Ming Lei <[email protected]> --- drivers/scsi/megaraid/megaraid_sas_base.c | 6 ++++++ 1 file changed, 6 insertions(+) diff --git a/drivers/scsi/megaraid/megaraid_sas_base.c b/drivers/scsi/megaraid/megaraid_sas_base.c index 065956cb2aeb..0b46f97cbfdb 100644 --- a/drivers/scsi/megaraid/megaraid_sas_base.c +++ b/drivers/scsi/megaraid/megaraid_sas_base.c @@ -3177,6 +3177,7 @@ static struct scsi_host_template megasas_template = { .use_clustering = ENABLE_CLUSTERING, .change_queue_depth = scsi_change_queue_depth, .no_write_same = 1, + .host_tagset = 1, }; /** @@ -5947,6 +5948,8 @@ static int megasas_start_aen(struct megasas_instance *instance) static int megasas_io_attach(struct megasas_instance *instance) { struct Scsi_Host *host = instance->host; + /* 256 tags should be high enough to saturate device */ + int max_queues = DIV_ROUND_UP(host->can_queue, 256); /* * Export parameters required by SCSI mid-layer @@ -5987,6 +5990,9 @@ static int megasas_io_attach(struct megasas_instance *instance) host->max_lun = MEGASAS_MAX_LUN; host->max_cmd_len = 16; + /* per NUMA node hw queue */ + host->nr_hw_queues = min_t(int, nr_node_ids, max_queues); + /* * Notify the mid-layer about the new controller */ -- 2.9.5
