From: Jens Axboe <jaxboe@fusionio.com>
Date: Tue, 30 Aug 2011 21:34:14 +0000 (-0600)
Subject: Fix bad latency reporting for rated IO jobs
X-Git-Tag: fio-1.58~19
X-Git-Url: https://git.kernel.dk/?p=fio.git;a=commitdiff_plain;h=78c1eda5486f586ca73e9cbe7e40b93fccdbd218;hp=5f739e0e2912b809635fac635a8c1140155577e6

Fix bad latency reporting for rated IO jobs

If we are going to sleep to satisfy a given rate limit, then
ensure that we flush pending IO first. Otherwise they end up
including the sleep time in the calculated completion latencies.

Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
---

diff --git a/io_u.c b/io_u.c
index 6a53bdaa..7709df5f 100644
--- a/io_u.c
+++ b/io_u.c
@@ -486,6 +486,16 @@ static enum fio_ddir rate_ddir(struct thread_data *td, enum fio_ddir ddir)
 	} else
 		usec = td->rate_pending_usleep[ddir];
 
+	/*
+	 * We are going to sleep, ensure that we flush anything pending as
+	 * not to skew our latency numbers
+	 */
+	if (td->cur_depth) {
+		int fio_unused ret;
+
+		ret = io_u_queued_complete(td, td->cur_depth, NULL);
+	}
+
 	fio_gettime(&t, NULL);
 	usec_sleep(td, usec);
 	usec = utime_since_now(&t);