[PATCH 2/4] md: Compute average flush time from component devices

From: Darrick J. Wong
Date: Mon Nov 29 2010 - 17:06:47 EST


For md devices which are composed of other block devices, a flush is spread out
to those other block devices. Therefore, the average flush time can be
computed as the average flush time of whichever device flushes most slowly.

Signed-off-by: Darrick J. Wong <djwong@xxxxxxxxxx>
---
drivers/md/md.c | 23 +++++++++++++++++++++++
1 files changed, 23 insertions(+), 0 deletions(-)


diff --git a/drivers/md/md.c b/drivers/md/md.c
index 43243a4..af25c96 100644
--- a/drivers/md/md.c
+++ b/drivers/md/md.c
@@ -357,6 +357,28 @@ EXPORT_SYMBOL(mddev_congested);
* Generic flush handling for md
*/

+static void measure_flushes(mddev_t *mddev)
+{
+ mdk_rdev_t *rdev;
+ u64 max = 0, samples = 0;
+
+ rcu_read_lock();
+ list_for_each_entry_rcu(rdev, &mddev->disks, same_set)
+ if (rdev->raid_disk >= 0 &&
+ !test_bit(Faulty, &rdev->flags)) {
+ if (rdev->bdev->bd_disk->avg_flush_time_ns <= max)
+ continue;
+ max = rdev->bdev->bd_disk->avg_flush_time_ns;
+ samples = rdev->bdev->bd_disk->flush_samples;
+ }
+ rcu_read_unlock();
+
+ spin_lock(&mddev->gendisk->flush_time_lock);
+ mddev->gendisk->avg_flush_time_ns = max;
+ mddev->gendisk->flush_samples = samples;
+ spin_unlock(&mddev->gendisk->flush_time_lock);
+}
+
static void md_end_flush(struct bio *bio, int err)
{
mdk_rdev_t *rdev = bio->bi_private;
@@ -365,6 +387,7 @@ static void md_end_flush(struct bio *bio, int err)
rdev_dec_pending(rdev, mddev);

if (atomic_dec_and_test(&mddev->flush_pending)) {
+ measure_flushes(mddev);
/* The pre-request flush has finished */
queue_work(md_wq, &mddev->flush_work);
}

--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/