Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
From:   Pavel Begunkov <asml.silence@gmail.com>
To:     Jens Axboe <axboe@kernel.dk>, io-uring@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: [PATCH 1/4] io_uring: fix cached_sq_head in io_timeout()
Date:   Wed, 15 Apr 2020 00:39:48 +0300
Message-Id: <637d5c3f98c3303d3d1ada057414e067facf4105.1586899625.git.asml.silence@gmail.com>
In-Reply-To: <cover.1586899625.git.asml.silence@gmail.com>
References: <cover.1586899625.git.asml.silence@gmail.com>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

io_timeout() can be executed asynchronously by a worker and without
holding ctx->uring_lock

1. using ctx->cached_sq_head there is racy there
2. it should count events from a moment of timeout's submission, but
not execution

Use req->sequence.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 fs/io_uring.c | 15 ++++++++-------
 1 file changed, 8 insertions(+), 7 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index c0cf57764329..fcc320d67606 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -4673,6 +4673,7 @@ static int io_timeout(struct io_kiocb *req)
 	struct io_timeout_data *data;
 	struct list_head *entry;
 	unsigned span = 0;
+	u32 seq = req->sequence;
 
 	data = &req->io->timeout;
 
@@ -4689,7 +4690,7 @@ static int io_timeout(struct io_kiocb *req)
 		goto add;
 	}
 
-	req->sequence = ctx->cached_sq_head + count - 1;
+	req->sequence = seq + count;
 	data->seq_offset = count;
 
 	/*
@@ -4699,7 +4700,7 @@ static int io_timeout(struct io_kiocb *req)
 	spin_lock_irq(&ctx->completion_lock);
 	list_for_each_prev(entry, &ctx->timeout_list) {
 		struct io_kiocb *nxt = list_entry(entry, struct io_kiocb, list);
-		unsigned nxt_sq_head;
+		unsigned nxt_seq;
 		long long tmp, tmp_nxt;
 		u32 nxt_offset = nxt->io->timeout.seq_offset;
 
@@ -4707,18 +4708,18 @@ static int io_timeout(struct io_kiocb *req)
 			continue;
 
 		/*
-		 * Since cached_sq_head + count - 1 can overflow, use type long
+		 * Since seq + count can overflow, use type long
 		 * long to store it.
 		 */
-		tmp = (long long)ctx->cached_sq_head + count - 1;
-		nxt_sq_head = nxt->sequence - nxt_offset + 1;
-		tmp_nxt = (long long)nxt_sq_head + nxt_offset - 1;
+		tmp = (long long)seq + count;
+		nxt_seq = nxt->sequence - nxt_offset;
+		tmp_nxt = (long long)nxt_seq + nxt_offset;
 
 		/*
 		 * cached_sq_head may overflow, and it will never overflow twice
 		 * once there is some timeout req still be valid.
 		 */
-		if (ctx->cached_sq_head < nxt_sq_head)
+		if (seq < nxt_seq)
 			tmp += UINT_MAX;
 
 		if (tmp > tmp_nxt)
-- 
2.24.0