[RFC PATCH v2 7/8] bfq: avoid blkg lookup from locked cgroup update

From: Yu Kuai <[email protected]>

bfq_bio_bfqg() is called while bfqd->lock is held from the merge and
request insertion paths.  It walks bio->bi_blkg and its parent chain to
find the closest online BFQ group, and re-associates the bio by looking
the chosen blkg up again through bio_associate_blkg_from_css().

Now that blkg creation runs under q->blkcg_mutex from the sleepable submit
path, bio_associate_blkg_from_css() can sleep on a lookup miss, which BFQ
must not do while holding bfqd->lock.  The blkg BFQ wants is already in
hand from the ancestry walk, so update bio->bi_blkg by swapping references
to that existing blkg directly instead of looking it up again by css.

Signed-off-by: Yu Kuai <[email protected]>
---
 block/bfq-cgroup.c | 16 +++++++++++++---
 1 file changed, 13 insertions(+), 3 deletions(-)

diff --git a/block/bfq-cgroup.c b/block/bfq-cgroup.c
index 42614aa78cd4..8a3ff9510386 100644
--- a/block/bfq-cgroup.c
+++ b/block/bfq-cgroup.c
@@ -604,6 +604,16 @@ static void bfq_link_bfqg(struct bfq_data *bfqd, struct bfq_group *bfqg)
 	}
 }
 
+static void bfq_bio_update_blkg(struct bio *bio, struct blkcg_gq *blkg)
+{
+	if (bio->bi_blkg == blkg)
+		return;
+
+	blkg_get(blkg);
+	blkg_put(bio->bi_blkg);
+	bio->bi_blkg = blkg;
+}
+
 struct bfq_group *bfq_bio_bfqg(struct bfq_data *bfqd, struct bio *bio)
 {
 	struct blkcg_gq *blkg = bio->bi_blkg;
@@ -616,13 +626,13 @@ struct bfq_group *bfq_bio_bfqg(struct bfq_data *bfqd, struct bio *bio)
 		}
 		bfqg = blkg_to_bfqg(blkg);
 		if (bfqg->pd.online) {
-			bio_associate_blkg_from_css(bio, &blkg->blkcg->css);
+			bfq_bio_update_blkg(bio, blkg);
 			return bfqg;
 		}
 		blkg = blkg->parent;
 	}
-	bio_associate_blkg_from_css(bio,
-				&bfqg_to_blkg(bfqd->root_group)->blkcg->css);
+	blkg = bfqg_to_blkg(bfqd->root_group);
+	bfq_bio_update_blkg(bio, blkg);
 	return bfqd->root_group;
 }
 
-- 
2.51.0