[PATCH v3 2/8] zbd: fix write zone accounting

Currently, zbd_convert_to_write_zones() calls io_u_quiesce() when the
number of write target zones hits one of the limits of write zones. This
wait by io_u_quiesce() significantly degrade the performance. While I
tried to remove the io_u_quiesce(), I observed that the test case 58 of
t/zbd/test-zbd-support failed with null_blk devices that have a
max_active_zones limit set.

The failure cause is an incorrect write target zone accounting in
zbd_convert_to_write_zones(). This function checks the current write
target zones, and selects one of them as the next write target zone.
After the zone selection, it locks the zone. However, when the zone is
locked, another job such as a trim workload or a write workload with the
zone_reset_threshold option might have already reset the zone and
removed it from the write target zones array. This unexpected zone
removal from the array caused an incorrect zone accounting and the test
case failure.

To avoid the incorrect zone accounting, call zbd_write_zone_get() after
the selected zone gets locked. If the zone is removed from the write
target zones array, the function adds the zone back to the array.

Signed-off-by: Shin'ichiro Kawasaki <[email protected]>
---
 zbd.c | 13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

diff --git a/zbd.c b/zbd.c
index b71f842c..c511b709 100644
--- a/zbd.c
+++ b/zbd.c
@@ -1693,8 +1693,17 @@ retry:
 
 		zone_lock(td, f, z);
 		if (zbd_zone_remainder(z) >= min_bs) {
-			need_zone_finish = false;
-			goto out;
+			/*
+			 * The zone might be already removed from
+			 * zbdi->write_zones[] by other jobs at this moment.
+			 * Even if the zone has remainder, call
+			 * zbd_write_zone_get() to ensure that it is in the
+			 * array.
+			 */
+			if (zbd_write_zone_get(td, f, z)) {
+				need_zone_finish = false;
+				goto out;
+			}
 		}
 		pthread_mutex_lock(&zbdi->mutex);
 	}
-- 
2.49.0