Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1543046

[PATCH V5 17/17] blk-throttle: add latency target support

Show all headers | View raw

One hard problem adding .low limit is to detect idle cgroup. If one
cgroup doesn't dispatch enough IO against its low limit, we must have a
mechanism to determine if other cgroups dispatch more IO. We added the
think time detection mechanism before, but it doesn't work for all
workloads. Here we add a latency based approach.

We already have mechanism to calculate latency threshold for each IO
size. For every IO dispatched from a cgorup, we compare its latency
against its threshold and record the info. If most IO latency is below
threshold (in the code I use 75%), the cgroup could be treated idle and
other cgroups can dispatch more IO.

Currently this latency target check is only for SSD as we can't
calcualte the latency target for hard disk. And this is only for cgroup
leaf node so far.

Signed-off-by: Shaohua Li <shli@fb.com>
---
 block/blk-throttle.c | 41 ++++++++++++++++++++++++++++++++++++-----
 1 file changed, 36 insertions(+), 5 deletions(-)

diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 1dc707a..915ebf5 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -162,6 +162,10 @@ struct throtl_grp {
 	u64 checked_last_finish_time;
 	u64 avg_ttime;
 	u64 idle_ttime_threshold;
+
+	unsigned int bio_cnt; /* total bios */
+	unsigned int bad_bio_cnt; /* bios exceeding latency threshold */
+	unsigned long bio_cnt_reset_time;
 };
 
 /* We measure latency for request size from <= 4k to >= 1M */
@@ -1688,11 +1692,14 @@ static bool throtl_tg_is_idle(struct throtl_grp *tg)
 	 * - single idle is too long, longer than a fixed value (in case user
 	 *   configure a too big threshold) or 4 times of slice
 	 * - average think time is more than threshold
+	 * - IO latency is largely below threshold
 	 */
 	u64 time = (u64)jiffies_to_usecs(4 * tg->td->throtl_slice) * 1000;
 	time = min_t(u64, MAX_IDLE_TIME, time);
 	return ktime_get_ns() - tg->last_finish_time > time ||
-	       tg->avg_ttime > tg->idle_ttime_threshold;
+	       tg->avg_ttime > tg->idle_ttime_threshold ||
+	       (tg->latency_target && tg->bio_cnt &&
+		tg->bad_bio_cnt * 5 < tg->bio_cnt);
 }
 
 static bool throtl_tg_can_upgrade(struct throtl_grp *tg)
@@ -2170,12 +2177,36 @@ void blk_throtl_bio_endio(struct bio *bio)
 
 	start_time = blk_stat_time(&bio->bi_issue_stat);
 	finish_time = __blk_stat_time(finish_time);
-	if (start_time && finish_time > start_time &&
-	    tg->td->track_bio_latency == 1 &&
-	    !(bio->bi_issue_stat.stat & SKIP_TRACK)) {
-		lat = finish_time - start_time;
+	if (!start_time || finish_time <= start_time)
+		return;
+
+	lat = finish_time - start_time;
+	if (tg->td->track_bio_latency == 1 &&
+	    !(bio->bi_issue_stat.stat & SKIP_TRACK))
 		throtl_track_latency(tg->td, blk_stat_size(&bio->bi_issue_stat),
 			bio_op(bio), lat);
+
+	if (tg->latency_target) {
+		int bucket;
+		unsigned int threshold;
+
+		bucket = request_bucket_index(
+			blk_stat_size(&bio->bi_issue_stat));
+		threshold = tg->td->avg_buckets[bucket].latency +
+			tg->latency_target;
+		if (lat > threshold)
+			tg->bad_bio_cnt++;
+		/*
+		 * Not race free, could get wrong count, which means cgroups
+		 * will be throttled
+		 */
+		tg->bio_cnt++;
+	}
+
+	if (time_after(jiffies, tg->bio_cnt_reset_time) || tg->bio_cnt > 1024) {
+		tg->bio_cnt_reset_time = tg->td->throtl_slice + jiffies;
+		tg->bio_cnt /= 2;
+		tg->bad_bio_cnt /= 2;
 	}
 }
 
-- 
2.9.3

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH V5 00/17] blk-throttle: add .low limit Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 14/17] blk-throttle: add interface for per-cgroup target latency Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 15/17] block: track request size in blk_issue_stat Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 17/17] blk-throttle: add latency target support Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 11/17] blk-throttle: add a simple idle detection Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 08/17] blk-throttle: make throtl_slice tunable Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 12/17] blk-throttle: add interface to configure idle time threshold Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 09/17] blk-throttle: detect completed idle cgroup Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 13/17] blk-throttle: ignore idle cgroup limit Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 04/17] blk-throttle: configure bps/iops limit for cgroup in low limit Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 03/17] blk-throttle: add .low interface Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 02/17] blk-throttle: prepare support multiple limits Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 01/17] blk-throttle: use U64_MAX/UINT_MAX to replace -1 Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100
  [PATCH V5 10/17] blk-throttle: make bandwidth change smooth Shaohua Li <shli@fb.com> - 2016-12-15 21:40 +0100

csiph-web