Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
From:   Stefano Garzarella <sgarzare@redhat.com>
To:     Jens Axboe <axboe@kernel.dk>
Cc:     linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        Hannes Reinecke <hare@suse.com>, io-uring@vger.kernel.org
Subject: [PATCH] io_uring: prevent sq_thread from spinning when it should stop
Date:   Fri, 21 Feb 2020 16:42:16 +0100
Message-Id: <20200221154216.206367-1-sgarzare@redhat.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

This patch drops 'cur_mm' before calling cond_resched(), to prevent
the sq_thread from spinning even when the user process is finished.

Before this patch, if the user process ended without closing the
io_uring fd, the sq_thread continues to spin until the
'sq_thread_idle' timeout ends.

In the worst case where the 'sq_thread_idle' parameter is bigger than
INT_MAX, the sq_thread will spin forever.

Fixes: 6c271ce2f1d5 ("io_uring: add submission polling")
Signed-off-by: Stefano Garzarella <sgarzare@redhat.com>
---

Hi Jens,
I'm also sending a test to liburing for this case.

Cheers,
Stefano
---
 fs/io_uring.c | 24 ++++++++++++------------
 1 file changed, 12 insertions(+), 12 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 5a826017ebb8..f902f77964ef 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -5138,6 +5138,18 @@ static int io_sq_thread(void *data)
 		 * to enter the kernel to reap and flush events.
 		 */
 		if (!to_submit || ret == -EBUSY) {
+			/*
+			 * Drop cur_mm before scheduling, we can't hold it for
+			 * long periods (or over schedule()). Do this before
+			 * adding ourselves to the waitqueue, as the unuse/drop
+			 * may sleep.
+			 */
+			if (cur_mm) {
+				unuse_mm(cur_mm);
+				mmput(cur_mm);
+				cur_mm = NULL;
+			}
+
 			/*
 			 * We're polling. If we're within the defined idle
 			 * period, then let us spin without work before going
@@ -5152,18 +5164,6 @@ static int io_sq_thread(void *data)
 				continue;
 			}
 
-			/*
-			 * Drop cur_mm before scheduling, we can't hold it for
-			 * long periods (or over schedule()). Do this before
-			 * adding ourselves to the waitqueue, as the unuse/drop
-			 * may sleep.
-			 */
-			if (cur_mm) {
-				unuse_mm(cur_mm);
-				mmput(cur_mm);
-				cur_mm = NULL;
-			}
-
 			prepare_to_wait(&ctx->sqo_wait, &wait,
 						TASK_INTERRUPTIBLE);
 
-- 
2.24.1