There are several scenarios where blkg_lookup_create() can fail such as
the blkcg dying, request_queue is dying, or simply being OOM. Most
handle this by simply falling back to the q->root_blkg and calling it a
day.
This patch implements the notion of closest blkg. During
blkg_lookup_create(), if it fails to create, return the closest blkg
found or the q->root_blkg. blkg_try_get_closest() is introduced and used
during association so a bio is always attached to a blkg.
Signed-off-by: Dennis Zhou <dennis@xxxxxxxxxx>
Acked-by: Tejun Heo <tj@xxxxxxxxxx>
---
block/bio.c | 17 ++++++++++-------
block/blk-cgroup.c | 23 ++++++++++++++++-------
block/blk-iolatency.c | 14 ++------------
block/blk-throttle.c | 4 +---
include/linux/blk-cgroup.h | 24 +++++++++++++++---------
5 files changed, 44 insertions(+), 38 deletions(-)
diff --git a/block/bio.c b/block/bio.c
index 7528c2324319..edc8a73b98d5 100644
--- a/block/bio.c
+++ b/block/bio.c
@@ -2009,21 +2009,24 @@ int bio_associate_blkcg(struct bio *bio, struct cgroup_subsys_state *blkcg_css)
EXPORT_SYMBOL_GPL(bio_associate_blkcg);
/**
- * bio_associate_blkg - associate a bio with the specified blkg
+ * bio_associate_blkg - associate a bio with the a blkg
* @bio: target bio
* @blkg: the blkg to associate
*
- * Associate @bio with the blkg specified by @blkg. This is the queue specific
- * blkcg information associated with the @bio, a reference will be taken on the
- * @blkg and will be freed when the bio is freed.
+ * This tries to associate @bio with the specified @blkg. Association failure
+ * is handled by walking up the blkg tree. Therefore, the blkg associated can
+ * be anything between @blkg and the root_blkg. This situation only happens
+ * when a cgroup is dying and then the remaining bios will spill to the closest
+ * alive blkg.
+ *
+ * A reference will be taken on the @blkg and will be released when @bio is
+ * freed.
*/
int bio_associate_blkg(struct bio *bio, struct blkcg_gq *blkg)
{
if (unlikely(bio->bi_blkg))
return -EBUSY;
- if (!blkg_try_get(blkg))
- return -ENODEV;
- bio->bi_blkg = blkg;
+ bio->bi_blkg = blkg_try_get_closest(blkg);
return 0;
}
diff --git a/block/blk-cgroup.c b/block/blk-cgroup.c
index b1ec98eeeae0..dfd984bbed27 100644
--- a/block/blk-cgroup.c
+++ b/block/blk-cgroup.c
@@ -258,9 +258,8 @@ static struct blkcg_gq *blkg_create(struct blkcg *blkcg,
* that all non-root blkg's have access to the parent blkg. This function
* should be called under RCU read lock and @q->queue_lock.
*
- * Returns pointer to the looked up or created blkg on success, ERR_PTR()
- * value on error. If @q is dead, returns ERR_PTR(-EINVAL). If @q is not
- * dead and bypassing, returns ERR_PTR(-EBUSY).
+ * Returns the blkg or the closest blkg if blkg_create() fails as it walks
+ * down from root.
*/
struct blkcg_gq *__blkg_lookup_create(struct blkcg *blkcg,
struct request_queue *q)
@@ -276,19 +275,29 @@ struct blkcg_gq *__blkg_lookup_create(struct blkcg *blkcg,
/*
* Create blkgs walking down from blkcg_root to @blkcg, so that all
- * non-root blkgs have access to their parents.
+ * non-root blkgs have access to their parents. Returns the closest
+ * blkg to the intended blkg should blkg_create() fail.
*/
while (true) {
struct blkcg *pos = blkcg;
struct blkcg *parent = blkcg_parent(blkcg);
-
- while (parent && !__blkg_lookup(parent, q, false)) {
+ struct blkcg_gq *ret_blkg = q->root_blkg;
+
+ while (parent) {
+ blkg = __blkg_lookup(parent, q, false);
+ if (blkg) {
+ /* remember closest blkg */
+ ret_blkg = blkg;
+ break;
+ }
pos = parent;
parent = blkcg_parent(parent);
}
blkg = blkg_create(pos, q, NULL);
- if (pos == blkcg || IS_ERR(blkg))
+ if (IS_ERR(blkg))
+ return ret_blkg;
+ if (pos == blkcg)
return blkg;
}
}
diff --git a/block/blk-iolatency.c b/block/blk-iolatency.c
index e6f68f15dee9..46e86c34cf79 100644
--- a/block/blk-iolatency.c
+++ b/block/blk-iolatency.c
@@ -483,21 +483,11 @@ static void blkcg_iolatency_throttle(struct rq_qos *rqos, struct bio *bio)
rcu_read_lock();
bio_associate_blkcg(bio, NULL);
blkcg = bio_blkcg(bio);
- blkg = blkg_lookup(blkcg, q);
- if (unlikely(!blkg)) {
- spin_lock_irq(&q->queue_lock);
- blkg = __blkg_lookup_create(blkcg, q);
- if (IS_ERR(blkg))
- blkg = NULL;
- spin_unlock_irq(&q->queue_lock);
- }
- if (!blkg)
- goto out;
-
+ blkg = blkg_lookup_create(blkcg, q);
bio_issue_init(&bio->bi_issue, bio_sectors(bio));
bio_associate_blkg(bio, blkg);
-out:
rcu_read_unlock();
+
while (blkg && blkg->parent) {
struct iolatency_grp *iolat = blkg_to_lat(blkg);
if (!iolat) {
diff --git a/block/blk-throttle.c b/block/blk-throttle.c
index 8f0a104770ee..d648d6720f46 100644
--- a/block/blk-throttle.c
+++ b/block/blk-throttle.c
@@ -2118,9 +2118,7 @@ static inline void throtl_update_latency_buckets(struct throtl_data *td)
static void blk_throtl_assoc_bio(struct throtl_grp *tg, struct bio *bio)
{
#ifdef CONFIG_BLK_DEV_THROTTLING_LOW
- /* fallback to root_blkg if we fail to get a blkg ref */
- if (bio->bi_css && (bio_associate_blkg(bio, tg_to_blkg(tg)) == -ENODEV))
- bio_associate_blkg(bio, bio->bi_disk->queue->root_blkg);
+ bio_associate_blkg(bio, tg_to_blkg(tg));
bio_issue_init(&bio->bi_issue, bio_sectors(bio));
#endif
}
diff --git a/include/linux/blk-cgroup.h b/include/linux/blk-cgroup.h
index 49f399cfb0b6..9d796a4f8ef0 100644
--- a/include/linux/blk-cgroup.h
+++ b/include/linux/blk-cgroup.h
@@ -545,6 +545,20 @@ static inline struct blkcg_gq *blkg_try_get(struct blkcg_gq *blkg)
return NULL;
}
+/**
+ * blkg_try_get_closest - try and get a blkg ref on the closet blkg
+ * @blkg: blkg to get
+ *
+ * This walks up the blkg tree to find the closest non-dying blkg and returns
+ * the blkg that it did association with as it may not be the passed in blkg.
+ */
+static inline struct blkcg_gq *blkg_try_get_closest(struct blkcg_gq *blkg)
+{
+ while (!atomic_inc_not_zero(&blkg->refcnt))
+ blkg = blkg->parent;
+
+ return blkg;
+}
void __blkg_release_rcu(struct rcu_head *rcu);
@@ -797,15 +811,7 @@ static inline bool blkcg_bio_issue_check(struct request_queue *q,
/* associate blkcg if bio hasn't attached one */
bio_associate_blkcg(bio, NULL);
blkcg = bio_blkcg(bio);
-
- blkg = blkg_lookup(blkcg, q);
- if (unlikely(!blkg)) {
- spin_lock_irq(&q->queue_lock);
- blkg = __blkg_lookup_create(blkcg, q);
- if (IS_ERR(blkg))
- blkg = NULL;
- spin_unlock_irq(&q->queue_lock);
- }
+ blkg = blkg_lookup_create(blkcg, q);
throtl = blk_throtl_bio(q, blkg, bio);
--
2.17.1
