[PATCH 1/1] io_uring: fix async close()

[PATCH 1/1] io_uring: fix async close()

[Pavel Begunkov]

First, io_close() misses filp_close() and io_cqring_add_event(), when
f_op->flush is defined. That's because in this case it will
io_queue_async_work() itself not grabbing files, so the corresponding
chunk in io_close_finish() won't be executed.

Second, when submitted through io_wq_submit_work(), it will do
filp_close() and *_add_event() twice: first inline in io_close(),
and the second one in call to io_close_finish() from io_close().
The second one will also fire, because it was submitted async through
generic path, and so have grabbed files.

And the last nice thing is to remove this weird pilgrimage with checking
work/old_work and casting it to nxt. Just use a helper instead.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 fs/io_uring.c | 44 ++++++++++++++++----------------------------
 1 file changed, 16 insertions(+), 28 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index c3bac9d850a5..b18022b0c273 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -2854,23 +2854,25 @@ static int io_close_prep(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 	return 0;
 }
 
+/* only called when __close_fd_get_file() is done */
+static void __io_close_finish(struct io_kiocb *req, struct io_kiocb **nxt)
+{
+	int ret;
+
+	ret = filp_close(req->close.put_file, req->work.files);
+	if (ret < 0)
+		req_set_fail_links(req);
+	io_cqring_add_event(req, ret);
+	fput(req->close.put_file);
+	io_put_req_find_next(req, nxt);
+}
+
 static void io_close_finish(struct io_wq_work **workptr)
 {
 	struct io_kiocb *req = container_of(*workptr, struct io_kiocb, work);
 	struct io_kiocb *nxt = NULL;
 
-	/* Invoked with files, we need to do the close */
-	if (req->work.files) {
-		int ret;
-
-		ret = filp_close(req->close.put_file, req->work.files);
-		if (ret < 0)
-			req_set_fail_links(req);
-		io_cqring_add_event(req, ret);
-	}
-
-	fput(req->close.put_file);
-	io_put_req_find_next(req, &nxt);
+	__io_close_finish(req, &nxt);
 	if (nxt)
 		io_wq_assign_next(workptr, nxt);
 }
@@ -2893,22 +2895,8 @@ static int io_close(struct io_kiocb *req, struct io_kiocb **nxt,
 	 * No ->flush(), safely close from here and just punt the
 	 * fput() to async context.
 	 */
-	ret = filp_close(req->close.put_file, current->files);
-
-	if (ret < 0)
-		req_set_fail_links(req);
-	io_cqring_add_event(req, ret);
-
-	if (io_wq_current_is_worker()) {
-		struct io_wq_work *old_work, *work;
-
-		old_work = work = &req->work;
-		io_close_finish(&work);
-		if (work && work != old_work)
-			*nxt = container_of(work, struct io_kiocb, work);
-		return 0;
-	}
-
+	__io_close_finish(req, nxt);
+	return 0;
 eagain:
 	req->work.func = io_close_finish;
 	/*
-- 
2.24.0

Re: [PATCH 1/1] io_uring: fix async close()

[Jens Axboe]

On 2/8/20 4:04 AM, Pavel Begunkov wrote:
> First, io_close() misses filp_close() and io_cqring_add_event(), when
> f_op->flush is defined. That's because in this case it will
> io_queue_async_work() itself not grabbing files, so the corresponding
> chunk in io_close_finish() won't be executed.
> 
> Second, when submitted through io_wq_submit_work(), it will do
> filp_close() and *_add_event() twice: first inline in io_close(),
> and the second one in call to io_close_finish() from io_close().
> The second one will also fire, because it was submitted async through
> generic path, and so have grabbed files.
> 
> And the last nice thing is to remove this weird pilgrimage with checking
> work/old_work and casting it to nxt. Just use a helper instead.

Thanks, applied. Nice cleanup, too!

-- 
Jens Axboe