Skip to contentSkip to conversation

blkcg: make blkcg_print_stat() print stats only for online blkgs

PATCH Applied Tejun Heo 2 participants 2 messages Nov 5, 2019 – Nov 7, 2019

Subsystems (3)
lore
Share & tools
lore

Tejun Heo

author maintainer
#0 lore raw

PATCH 1/1 blkcg: make blkcg_print_stat() print stats only for online blkgs

blkcg_print_stat() iterates blkgs under RCU and doesn't test whether the blkg is online. This can call into pd_stat_fn() on a pd which is still being initialized leading to an oops.

The heaviest operation - recursively summing up rwstat counters - is already done while holding the queue_lock. Expand queue_lock to cover the other operations and skip the blkg if it isn't online yet. The online state is protected by both blkcg and queue locks, so this guarantees that only online blkgs are processed.

Cc Josef Bacik <[email protected]>
Fixes 903d23f0a354 ("blk-cgroup: allow controllers to output their own stats")
Cc [email protected] # v4.19+
---
 block/blk-cgroup.c |   13 ++++++++-----
 1 file changed, 8 insertions(+), 5 deletions(-)
block/blk-cgroup.c +8 -5
@@ -934,9 +934,14 @@ static int blkcg_print_stat(struct seq_f
 		int i;
 		bool has_stats = false;
 
+		spin_lock_irq(&blkg->q->queue_lock);
+
+		if (!blkg->online)
+			goto skip;
+
 		dname = blkg_dev_name(blkg);
 		if (!dname)
-			continue;
+			goto skip;
 
 		/*
 		 * Hooray string manipulation, count is the size written NOT
@@ -946,8 +951,6 @@ static int blkcg_print_stat(struct seq_f
 		 */
 		off += scnprintf(buf+off, size-off, "%s ", dname);
 
-		spin_lock_irq(&blkg->q->queue_lock);
-
 		blkg_rwstat_recursive_sum(blkg, NULL,
 				offsetof(struct blkcg_gq, stat_bytes), &rwstat);
 		rbytes = rwstat.cnt[BLKG_RWSTAT_READ];
@@ -960,8 +963,6 @@ static int blkcg_print_stat(struct seq_f
 		wios = rwstat.cnt[BLKG_RWSTAT_WRITE];
 		dios = rwstat.cnt[BLKG_RWSTAT_DISCARD];
 
-		spin_unlock_irq(&blkg->q->queue_lock);
-
 		if (rbytes || wbytes || rios || wios) {
 			has_stats = true;
 			off += scnprintf(buf+off, size-off,
@@ -999,6 +1000,8 @@ static int blkcg_print_stat(struct seq_f
 				seq_commit(sf, -1);
 			}
 		}
+	skip:
+		spin_unlock_irq(&blkg->q->queue_lock);
 	}
 
 	rcu_read_unlock();
maintainer Applied
#1 lore raw
Quoted 9 lines · Tejun Heo

blkcg_print_stat() iterates blkgs under RCU and doesn't test whether the blkg is online. This can call into pd_stat_fn() on a pd which is still being initialized leading to an oops.

The heaviest operation - recursively summing up rwstat counters - is already done while holding the queue_lock. Expand queue_lock to cover the other operations and skip the blkg if it isn't online yet. The online state is protected by both blkcg and queue locks, so this guarantees that only online blkgs are processed.

Applied, thanks Tejun.

Signature · 1 line
Jens Axboe