Add a new configfs attribute for enabling pipelining of zoned writes. If
that attribute has been set, retry zoned writes that are not aligned with
the write pointer. The test script below reports 234 K IOPS with no I/O
scheduler, 5.32 K IOPS with mq-deadline and pipelining disabled and 92.2 K
IOPS with mq-deadline and pipelining enabled. This shows that pipelining
results in about 17 times more IOPS for this particular test case.
#!/bin/bash
for mode in "none 0" "mq-deadline 0" "mq-deadline 1"; do
set +e
for d in /sys/kernel/config/nullb/*; do
[ -d "$d" ] && rmdir "$d"
done
modprobe -r null_blk
set -e
read -r iosched pipelining <<<"$mode"
modprobe null_blk nr_devices=0
(
cd /sys/kernel/config/nullb
mkdir nullb0
cd nullb0
params=(
completion_nsec=100000
hw_queue_depth=64
irqmode=2 # NULL_IRQ_TIMER
max_sectors=$((4096/512))
memory_backed=1
pipeline_zoned_writes="${pipelining}"
size=1
submit_queues=1
zone_size=1
zoned=1
power=1
)
for p in "${params[@]}"; do
echo "${p//*=}" > "${p//=*}"
done
)
udevadm settle
dev=/dev/nullb0
[ -b "${dev}" ]
params=(
--direct=1
--filename="${dev}"
--iodepth=64
--iodepth_batch=16
--ioengine=io_uring
--ioscheduler="${iosched}"
--gtod_reduce=1
--hipri=0
--name=nullb0
--runtime=30
--rw=write
--time_based=1
--zonemode=zbd
)
fio "${params[@]}"
done
Cc: Damien Le Moal <damien.lemoal@xxxxxxxxxxxxxxxxxx>
Signed-off-by: Bart Van Assche <bvanassche@xxxxxxx>
---
drivers/block/null_blk/main.c | 9 +++++++++
drivers/block/null_blk/null_blk.h | 3 +++
drivers/block/null_blk/zoned.c | 13 ++++++++++++-
3 files changed, 24 insertions(+), 1 deletion(-)
diff --git a/drivers/block/null_blk/main.c b/drivers/block/null_blk/main.c
index d401230b1e20..851b55b7284f 100644
--- a/drivers/block/null_blk/main.c
+++ b/drivers/block/null_blk/main.c
@@ -424,6 +424,7 @@ NULLB_DEVICE_ATTR(zone_capacity, ulong, NULL);
NULLB_DEVICE_ATTR(zone_nr_conv, uint, NULL);
NULLB_DEVICE_ATTR(zone_max_open, uint, NULL);
NULLB_DEVICE_ATTR(zone_max_active, uint, NULL);
+NULLB_DEVICE_ATTR(pipeline_zoned_writes, bool, NULL);
NULLB_DEVICE_ATTR(virt_boundary, bool, NULL);
NULLB_DEVICE_ATTR(no_sched, bool, NULL);
NULLB_DEVICE_ATTR(shared_tag_bitmap, bool, NULL);
@@ -569,6 +570,7 @@ static struct configfs_attribute *nullb_device_attrs[] = {
&nullb_device_attr_zone_max_active,
&nullb_device_attr_zone_readonly,
&nullb_device_attr_zone_offline,
+ &nullb_device_attr_pipeline_zoned_writes,
&nullb_device_attr_virt_boundary,
&nullb_device_attr_no_sched,
&nullb_device_attr_shared_tag_bitmap,
@@ -1677,6 +1679,11 @@ static blk_status_t null_queue_rq(struct blk_mq_hw_ctx *hctx,
cmd->error = BLK_STS_OK;
cmd->nq = nq;
cmd->fake_timeout = should_timeout_request(rq);
+ if (!(rq->rq_flags & RQF_DONTPREP)) {
+ rq->rq_flags |= RQF_DONTPREP;
+ cmd->retries = 0;
+ cmd->max_attempts = rq->q->nr_requests;
+ }
blk_mq_start_request(rq);
@@ -2109,6 +2116,8 @@ static int null_add_dev(struct nullb_device *dev)
nullb->q->queuedata = nullb;
blk_queue_flag_set(QUEUE_FLAG_NONROT, nullb->q);
blk_queue_flag_clear(QUEUE_FLAG_ADD_RANDOM, nullb->q);
+ if (dev->pipeline_zoned_writes)
+ blk_queue_flag_set(QUEUE_FLAG_PIPELINE_ZONED_WRITES, nullb->q);
mutex_lock(&lock);
rv = ida_simple_get(&nullb_indexes, 0, 0, GFP_KERNEL);
diff --git a/drivers/block/null_blk/null_blk.h b/drivers/block/null_blk/null_blk.h
index eb5972c50be8..c44c3fdb1025 100644
--- a/drivers/block/null_blk/null_blk.h
+++ b/drivers/block/null_blk/null_blk.h
@@ -23,6 +23,8 @@ struct nullb_cmd {
unsigned int tag;
blk_status_t error;
bool fake_timeout;
+ u16 retries;
+ u16 max_attempts;
struct nullb_queue *nq;
struct hrtimer timer;
};
@@ -112,6 +114,7 @@ struct nullb_device {
bool memory_backed; /* if data is stored in memory */
bool discard; /* if support discard */
bool zoned; /* if device is zoned */
+ bool pipeline_zoned_writes;
bool virt_boundary; /* virtual boundary on/off for the device */
bool no_sched; /* no IO scheduler for the device */
bool shared_tag_bitmap; /* use hostwide shared tags */
diff --git a/drivers/block/null_blk/zoned.c b/drivers/block/null_blk/zoned.c
index 635ce0648133..cffc4985a7df 100644
--- a/drivers/block/null_blk/zoned.c
+++ b/drivers/block/null_blk/zoned.c
@@ -405,7 +405,18 @@ static blk_status_t null_zone_write(struct nullb_cmd *cmd, sector_t sector,
else
cmd->bio->bi_iter.bi_sector = sector;
} else if (sector != zone->wp) {
- ret = BLK_STS_IOERR;
+ /*
+ * In case of a misaligned write and if pipelining of zoned
+ * writes has been enabled, request the block layer to retry
+ * until the maximum number of attempts has been reached. If
+ * the maximum number of attempts has been reached, fail the
+ * misaligned write.
+ */
+ if (dev->pipeline_zoned_writes &&
+ ++cmd->retries < cmd->max_attempts)
+ ret = BLK_STS_DEV_RESOURCE;
+ else
+ ret = BLK_STS_IOERR;
goto unlock;
}
