IO-Uring Archive on lore.kernel.org
 help / color / Atom feed
* [PATCH] io_uring: Fix linked fixed-read/write
@ 2019-11-23 22:42 Pavel Begunkov
  2019-11-23 22:52 ` Jens Axboe
  0 siblings, 1 reply; 2+ messages in thread
From: Pavel Begunkov @ 2019-11-23 22:42 UTC (permalink / raw)
  To: Jens Axboe, io-uring

Dependant links of a fixed-read/write are always cancelled. The reason
is that io_complete_rw_common() uses req->result to decide whether to
fail links, which is set as res=io_import_iovec(). However,
io_import_fixed() doesn't return size, but error code.

Signed-off-by: Pavel Begunkov <asml.silence@gmail.com>
---
 fs/io_uring.c | 14 ++++++--------
 1 file changed, 6 insertions(+), 8 deletions(-)

diff --git a/fs/io_uring.c b/fs/io_uring.c
index 5ceec1a4faad..8119cbae4fb6 100644
--- a/fs/io_uring.c
+++ b/fs/io_uring.c
@@ -1484,9 +1484,9 @@ static void kiocb_done(struct kiocb *kiocb, ssize_t ret, struct io_kiocb **nxt,
 		io_rw_done(kiocb, ret);
 }
 
-static int io_import_fixed(struct io_ring_ctx *ctx, int rw,
-			   const struct io_uring_sqe *sqe,
-			   struct iov_iter *iter)
+static ssize_t io_import_fixed(struct io_ring_ctx *ctx, int rw,
+				const struct io_uring_sqe *sqe,
+				struct iov_iter *iter)
 {
 	size_t len = READ_ONCE(sqe->len);
 	struct io_mapped_ubuf *imu;
@@ -1555,7 +1555,7 @@ static int io_import_fixed(struct io_ring_ctx *ctx, int rw,
 		}
 	}
 
-	return 0;
+	return iter->count;
 }
 
 static ssize_t io_import_iovec(struct io_ring_ctx *ctx, int rw,
@@ -1576,11 +1576,9 @@ static ssize_t io_import_iovec(struct io_ring_ctx *ctx, int rw,
 	 * flag.
 	 */
 	opcode = READ_ONCE(sqe->opcode);
-	if (opcode == IORING_OP_READ_FIXED ||
-	    opcode == IORING_OP_WRITE_FIXED) {
-		ssize_t ret = io_import_fixed(ctx, rw, sqe, iter);
+	if (opcode == IORING_OP_READ_FIXED || opcode == IORING_OP_WRITE_FIXED) {
 		*iovec = NULL;
-		return ret;
+		return io_import_fixed(ctx, rw, sqe, iter);
 	}
 
 	if (!s->has_user)
-- 
2.24.0


^ permalink raw reply	[flat|nested] 2+ messages in thread

* Re: [PATCH] io_uring: Fix linked fixed-read/write
  2019-11-23 22:42 [PATCH] io_uring: Fix linked fixed-read/write Pavel Begunkov
@ 2019-11-23 22:52 ` Jens Axboe
  0 siblings, 0 replies; 2+ messages in thread
From: Jens Axboe @ 2019-11-23 22:52 UTC (permalink / raw)
  To: Pavel Begunkov, io-uring

On 11/23/19 3:42 PM, Pavel Begunkov wrote:
> Dependant links of a fixed-read/write are always cancelled. The reason
> is that io_complete_rw_common() uses req->result to decide whether to
> fail links, which is set as res=io_import_iovec(). However,
> io_import_fixed() doesn't return size, but error code.

See:

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=5e559561a8d7e6d4adfce6aa8fbf3daa3dec1577

Just not in the 5.5 branches. We probably do want to make it ssize_t to
be more proper, but IO in Linux is capped at 2G anyway so doesn't make
a difference right now.

-- 
Jens Axboe


^ permalink raw reply	[flat|nested] 2+ messages in thread

end of thread, back to index

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-11-23 22:42 [PATCH] io_uring: Fix linked fixed-read/write Pavel Begunkov
2019-11-23 22:52 ` Jens Axboe

IO-Uring Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/io-uring/0 io-uring/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 io-uring io-uring/ https://lore.kernel.org/io-uring \
		io-uring@vger.kernel.org
	public-inbox-index io-uring

Example config snippet for mirrors

Newsgroup available over NNTP:
	nntp://nntp.lore.kernel.org/org.kernel.vger.io-uring


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git