[PATCH 0/3] send/recv bundle fixes

[PATCH 0/3] send/recv bundle fixes

[Jens Axboe]

Hi,

Just three minor fixes for 6.11, all marked for stable as well. Two
of them fix a missing REQ_F_NEED_CLEANUP being set for recv and send,
and the last one fixes non-bundle sends potentially picking more than
one buffer. Only bundle sends should do that.

-- 
Jens Axboe

[PATCH 1/3] io_uring/net: ensure expanded bundle recv gets marked for cleanup

[Jens Axboe]

If the iovec inside the kmsg isn't already allocated AND one gets
expanded beyond the fixed size, then the request may not already have
been marked for cleanup. Ensure that it is.

Cc: stable@vger.kernel.org
Fixes: 2f9c9515bdfd ("io_uring/net: support bundles for recv")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 io_uring/net.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/io_uring/net.c b/io_uring/net.c
index 594490a1389b..97a48408cec3 100644
--- a/io_uring/net.c
+++ b/io_uring/net.c
@@ -1094,6 +1094,7 @@ static int io_recv_buf_select(struct io_kiocb *req, struct io_async_msghdr *kmsg
 		if (arg.iovs != &kmsg->fast_iov && arg.iovs != kmsg->free_iov) {
 			kmsg->free_iov_nr = ret;
 			kmsg->free_iov = arg.iovs;
+			req->flags |= REQ_F_NEED_CLEANUP;
 		}
 	} else {
 		void __user *buf;
-- 
2.43.0

[PATCH 2/3] io_uring/net: ensure expanded bundle send gets marked for cleanup

[Jens Axboe]

If the iovec inside the kmsg isn't already allocated AND one gets
expanded beyond the fixed size, then the request may not already have
been marked for cleanup. Ensure that it is.

Cc: stable@vger.kernel.org
Fixes: a05d1f625c7a ("io_uring/net: support bundles for send")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 io_uring/net.c | 1 +
 1 file changed, 1 insertion(+)

diff --git a/io_uring/net.c b/io_uring/net.c
index 97a48408cec3..050bea5e7256 100644
--- a/io_uring/net.c
+++ b/io_uring/net.c
@@ -623,6 +623,7 @@ int io_send(struct io_kiocb *req, unsigned int issue_flags)
 		if (arg.iovs != &kmsg->fast_iov && arg.iovs != kmsg->free_iov) {
 			kmsg->free_iov_nr = ret;
 			kmsg->free_iov = arg.iovs;
+			req->flags |= REQ_F_NEED_CLEANUP;
 		}
 	}
 
-- 
2.43.0

[PATCH 3/3] io_uring/net: don't pick multiple buffers for non-bundle send

[Jens Axboe]

If a send is issued marked with IOSQE_BUFFER_SELECT for selecting a
buffer, unless it's a bundle, it should not select multiple buffers.

Cc: stable@vger.kernel.org
Fixes: a05d1f625c7a ("io_uring/net: support bundles for send")
Signed-off-by: Jens Axboe <axboe@kernel.dk>
---
 io_uring/net.c | 5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/io_uring/net.c b/io_uring/net.c
index 050bea5e7256..d08abcca89cc 100644
--- a/io_uring/net.c
+++ b/io_uring/net.c
@@ -601,17 +601,18 @@ int io_send(struct io_kiocb *req, unsigned int issue_flags)
 			.iovs = &kmsg->fast_iov,
 			.max_len = INT_MAX,
 			.nr_iovs = 1,
-			.mode = KBUF_MODE_EXPAND,
 		};
 
 		if (kmsg->free_iov) {
 			arg.nr_iovs = kmsg->free_iov_nr;
 			arg.iovs = kmsg->free_iov;
-			arg.mode |= KBUF_MODE_FREE;
+			arg.mode = KBUF_MODE_FREE;
 		}
 
 		if (!(sr->flags & IORING_RECVSEND_BUNDLE))
 			arg.nr_iovs = 1;
+		else
+			arg.mode |= KBUF_MODE_EXPAND;
 
 		ret = io_buffers_select(req, &arg, issue_flags);
 		if (unlikely(ret < 0))
-- 
2.43.0