[PATCH AUTOSEL 6.1 32/33] io_uring: unlock sqd->lock before sq thread release CPU

[PATCH AUTOSEL 6.1 32/33] io_uring: unlock sqd->lock before sq thread release CPU

[Sasha Levin]

From: Wenwen Chen <wenwen.chen@samsung.com>

[ Upstream commit 533ab73f5b5c95dcb4152b52d5482abcc824c690 ]

The sq thread actively releases CPU resources by calling the
cond_resched() and schedule() interfaces when it is idle. Therefore,
more resources are available for other threads to run.

There exists a problem in sq thread: it does not unlock sqd->lock before
releasing CPU resources every time. This makes other threads pending on
sqd->lock for a long time. For example, the following interfaces all
require sqd->lock: io_sq_offload_create(), io_register_iowq_max_workers()
and io_ring_exit_work().

Before the sq thread releases CPU resources, unlocking sqd->lock will
provide the user a better experience because it can respond quickly to
user requests.

Signed-off-by: Kanchan Joshi<joshi.k@samsung.com>
Signed-off-by: Wenwen Chen<wenwen.chen@samsung.com>
Link: https://lore.kernel.org/r/20230525082626.577862-1-wenwen.chen@samsung.com
Signed-off-by: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 io_uring/sqpoll.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/io_uring/sqpoll.c b/io_uring/sqpoll.c
index 559652380672c..6ffa5cf1bbb86 100644
--- a/io_uring/sqpoll.c
+++ b/io_uring/sqpoll.c
@@ -256,9 +256,13 @@ static int io_sq_thread(void *data)
 			sqt_spin = true;
 
 		if (sqt_spin || !time_after(jiffies, timeout)) {
-			cond_resched();
 			if (sqt_spin)
 				timeout = jiffies + sqd->sq_thread_idle;
+			if (unlikely(need_resched())) {
+				mutex_unlock(&sqd->lock);
+				cond_resched();
+				mutex_lock(&sqd->lock);
+			}
 			continue;
 		}
 
-- 
2.39.2