[PATCH 08/10] io_uring: add support for IORING_OP_CLOSE

[Jens Axboe <axboe@kernel.dk>]

This works just like close(2), unsurprisingly. We remove the file
descriptor and post the completion inline, then offload the actual
(potential) last file put to async context.

Mark the async part of this work as uncancellable, as we really must
guarantee that the latter part of the close is run.

Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 fs/io_uring.c                 | 70 +++++++++++++++++++++++++++++++++++
 include/uapi/linux/io_uring.h |  1 +
 2 files changed, 71 insertions(+)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 132f887ef18d..927f28112f0e 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -298,6 +298,11 @@ struct io_poll_iocb {
 	struct wait_queue_entry		wait;
 };
 
+struct io_close {
+	struct file			*file;
+	struct file			*put_file;
+};
+
 struct io_timeout_data {
 	struct io_kiocb			*req;
 	struct hrtimer			timer;
@@ -350,6 +355,7 @@ struct io_kiocb {
 		struct file		*file;
 		struct kiocb		rw;
 		struct io_poll_iocb	poll;
+		struct io_close		close;
 	};
 
 	const struct io_uring_sqe	*sqe;
@@ -2093,6 +2099,64 @@ static int io_openat(struct io_kiocb *req, struct io_kiocb **nxt,
 	return 0;
 }
 
+static int io_close(struct io_kiocb *req, struct io_kiocb **nxt,
+		    bool force_nonblock)
+{
+	const struct io_uring_sqe *sqe = req->sqe;
+	int ret, fd;
+
+	if (sqe->ioprio || sqe->off || sqe->addr || sqe->len ||
+	    sqe->rw_flags || sqe->buf_index)
+		return -EINVAL;
+
+	fd = READ_ONCE(sqe->fd);
+	if (req->file->f_op == &io_uring_fops || fd == req->ring_fd)
+		return -EBADF;
+
+	/*
+	 * If we queue this for async, it must not be cancellable. That would
+	 * leave the 'file' in an undeterminate state.
+	 */
+	req->work.flags |= IO_WQ_WORK_NO_CANCEL;
+
+	ret = 0;
+	if (force_nonblock) {
+		req->close.put_file = NULL;
+		ret = __close_fd_get_file(fd, &req->close.put_file);
+		if (ret < 0)
+			return ret;
+
+		/* if the file has a flush method, be safe and punt to async */
+		if (req->close.put_file->f_op->flush) {
+			req->work.flags |= IO_WQ_WORK_NEEDS_FILES;
+			return -EAGAIN;
+		}
+
+		/*
+		 * 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);
+		return -EAGAIN;
+	} else {
+		/* Invoked with files, we need to do the close */
+		if (req->work.files) {
+			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);
+	return ret;
+}
+
 static int io_prep_sfr(struct io_kiocb *req, const struct io_uring_sqe *sqe)
 {
 	struct io_ring_ctx *ctx = req->ctx;
@@ -3116,6 +3180,9 @@ static int io_issue_sqe(struct io_kiocb *req, struct io_kiocb **nxt,
 	case IORING_OP_OPENAT:
 		ret = io_openat(req, nxt, force_nonblock);
 		break;
+	case IORING_OP_CLOSE:
+		ret = io_close(req, nxt, force_nonblock);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
@@ -3275,6 +3342,9 @@ static int io_grab_files(struct io_kiocb *req)
 	int ret = -EBADF;
 	struct io_ring_ctx *ctx = req->ctx;
 
+	if (!req->ring_file)
+		return -EBADF;
+
 	rcu_read_lock();
 	spin_lock_irq(&ctx->inflight_lock);
 	/*
diff --git a/include/uapi/linux/io_uring.h b/include/uapi/linux/io_uring.h
index 02af580754ce..42a7f0e8dee3 100644
--- a/include/uapi/linux/io_uring.h
+++ b/include/uapi/linux/io_uring.h
@@ -79,6 +79,7 @@ enum {
 	IORING_OP_CONNECT,
 	IORING_OP_FALLOCATE,
 	IORING_OP_OPENAT,
+	IORING_OP_CLOSE,
 
 	/* this goes last, obviously */
 	IORING_OP_LAST,
-- 
2.24.1