diff mbox series

[4/4] block, bfq: push up injection only after setting service time

Message ID 20190822152037.15413-5-paolo.valente@linaro.org
State Accepted
Commit 58494c980f40274c465ebfdece02d401def088bf
Headers show
Series block, bfq: series of improvements and small fixes of the injection mechanism | expand

Commit Message

Paolo Valente Aug. 22, 2019, 3:20 p.m. UTC 
If equal to 0, the injection limit for a bfq_queue is pushed to 1
after a first sample of the total service time of the I/O requests of
the queue is computed (to allow injection to start). Yet, because of a
mistake in the branch that performs this action, the push may happen
also in some other case. This commit fixes this issue.

Signed-off-by: Paolo Valente <paolo.valente@linaro.org>

---
 block/bfq-iosched.c | 12 +++++++-----
 1 file changed, 7 insertions(+), 5 deletions(-)

-- 
2.20.1
diff mbox series

Patch

diff --git a/block/bfq-iosched.c b/block/bfq-iosched.c
index ddac93e910fa..0319d6339822 100644
--- a/block/bfq-iosched.c
+++ b/block/bfq-iosched.c
@@ -5823,12 +5823,14 @@  static void bfq_update_inject_limit(struct bfq_data *bfqd,
 	 */
 	if ((bfqq->last_serv_time_ns == 0 && bfqd->rq_in_driver == 1) ||
 	    tot_time_ns < bfqq->last_serv_time_ns) {
+		if (bfqq->last_serv_time_ns == 0) {
+			/*
+			 * Now we certainly have a base value: make sure we
+			 * start trying injection.
+			 */
+			bfqq->inject_limit = max_t(unsigned int, 1, old_limit);
+		}
 		bfqq->last_serv_time_ns = tot_time_ns;
-		/*
-		 * Now we certainly have a base value: make sure we
-		 * start trying injection.
-		 */
-		bfqq->inject_limit = max_t(unsigned int, 1, old_limit);
 	} else if (!bfqd->rqs_injected && bfqd->rq_in_driver == 1)
 		/*
 		 * No I/O injected and no request still in service in