[PATCH 5.13 078/151] io-wq: fix bug of creating io-wokers unconditionally

[Greg Kroah-Hartman]

From: Hao Xu <haoxu@xxxxxxxxxxxxxxxxx>

[ Upstream commit 49e7f0c789add1330b111af0b7caeb0e87df063e ]

The former patch to add check between nr_workers and max_workers has a
bug, which will cause unconditionally creating io-workers. That's
because the result of the check doesn't affect the call of
create_io_worker(), fix it by bringing in a boolean value for it.

Fixes: 21698274da5b ("io-wq: fix lack of acct->nr_workers < acct->max_workers judgement")
Signed-off-by: Hao Xu <haoxu@xxxxxxxxxxxxxxxxx>
Link: https://lore.kernel.org/r/20210808135434.68667-2-haoxu@xxxxxxxxxxxxxxxxx
[axboe: drop hunk that isn't strictly needed]
Signed-off-by: Jens Axboe <axboe@xxxxxxxxx>
Signed-off-by: Sasha Levin <sashal@xxxxxxxxxx>
---
 fs/io-wq.c | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/fs/io-wq.c b/fs/io-wq.c
index 77026d42cb79..2c8a9a394884 100644
--- a/fs/io-wq.c
+++ b/fs/io-wq.c
@@ -283,16 +283,24 @@ static void create_worker_cb(struct callback_head *cb)
 	struct io_wq *wq;
 	struct io_wqe *wqe;
 	struct io_wqe_acct *acct;
+	bool do_create = false;
 
 	cwd = container_of(cb, struct create_worker_data, work);
 	wqe = cwd->wqe;
 	wq = wqe->wq;
 	acct = &wqe->acct[cwd->index];
 	raw_spin_lock_irq(&wqe->lock);
-	if (acct->nr_workers < acct->max_workers)
+	if (acct->nr_workers < acct->max_workers) {
 		acct->nr_workers++;
+		do_create = true;
+	}
 	raw_spin_unlock_irq(&wqe->lock);
-	create_io_worker(wq, cwd->wqe, cwd->index);
+	if (do_create) {
+		create_io_worker(wq, cwd->wqe, cwd->index);
+	} else {
+		atomic_dec(&acct->nr_running);
+		io_worker_ref_put(wq);
+	}
 	kfree(cwd);
 }
 
-- 
2.30.2