Netdev Archive on lore.kernel.org
help / color / mirror / Atom feed
* [PATCH net-next] tipc: keep the skb in rcv queue until the whole data is read
@ 2021-07-16 21:44 Xin Long
  2021-07-17  0:30 ` patchwork-bot+netdevbpf
  0 siblings, 1 reply; 2+ messages in thread
From: Xin Long @ 2021-07-16 21:44 UTC (permalink / raw)
  To: network dev, davem, kuba, Jon Maloy, tipc-discussion; +Cc: Erin Shepherd

Currently, when userspace reads a datagram with a buffer that is
smaller than this datagram, the data will be truncated and only
part of it can be received by users. It doesn't seem right that
users don't know the datagram size and have to use a huge buffer
to read it to avoid the truncation.

This patch to fix it by keeping the skb in rcv queue until the
whole data is read by users. Only the last msg of the datagram
will be marked with MSG_EOR, just as TCP/SCTP does.

Note that this will work as above only when MSG_EOR is set in the
flags parameter of recvmsg(), so that it won't break any old user
applications.

Signed-off-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Jon Maloy <jmaloy@redhat.com>
---
 net/tipc/socket.c | 36 +++++++++++++++++++++++++++---------
 1 file changed, 27 insertions(+), 9 deletions(-)

diff --git a/net/tipc/socket.c b/net/tipc/socket.c
index 34a97ea36cc8..9b0b311c7ec1 100644
--- a/net/tipc/socket.c
+++ b/net/tipc/socket.c
@@ -1880,6 +1880,7 @@ static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
 	bool connected = !tipc_sk_type_connectionless(sk);
 	struct tipc_sock *tsk = tipc_sk(sk);
 	int rc, err, hlen, dlen, copy;
+	struct tipc_skb_cb *skb_cb;
 	struct sk_buff_head xmitq;
 	struct tipc_msg *hdr;
 	struct sk_buff *skb;
@@ -1903,6 +1904,7 @@ static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
 		if (unlikely(rc))
 			goto exit;
 		skb = skb_peek(&sk->sk_receive_queue);
+		skb_cb = TIPC_SKB_CB(skb);
 		hdr = buf_msg(skb);
 		dlen = msg_data_sz(hdr);
 		hlen = msg_hdr_sz(hdr);
@@ -1922,18 +1924,33 @@ static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
 
 	/* Capture data if non-error msg, otherwise just set return value */
 	if (likely(!err)) {
-		copy = min_t(int, dlen, buflen);
-		if (unlikely(copy != dlen))
-			m->msg_flags |= MSG_TRUNC;
-		rc = skb_copy_datagram_msg(skb, hlen, m, copy);
+		int offset = skb_cb->bytes_read;
+
+		copy = min_t(int, dlen - offset, buflen);
+		rc = skb_copy_datagram_msg(skb, hlen + offset, m, copy);
+		if (unlikely(rc))
+			goto exit;
+		if (unlikely(offset + copy < dlen)) {
+			if (flags & MSG_EOR) {
+				if (!(flags & MSG_PEEK))
+					skb_cb->bytes_read = offset + copy;
+			} else {
+				m->msg_flags |= MSG_TRUNC;
+				skb_cb->bytes_read = 0;
+			}
+		} else {
+			if (flags & MSG_EOR)
+				m->msg_flags |= MSG_EOR;
+			skb_cb->bytes_read = 0;
+		}
 	} else {
 		copy = 0;
 		rc = 0;
-		if (err != TIPC_CONN_SHUTDOWN && connected && !m->msg_control)
+		if (err != TIPC_CONN_SHUTDOWN && connected && !m->msg_control) {
 			rc = -ECONNRESET;
+			goto exit;
+		}
 	}
-	if (unlikely(rc))
-		goto exit;
 
 	/* Mark message as group event if applicable */
 	if (unlikely(grp_evt)) {
@@ -1956,9 +1973,10 @@ static int tipc_recvmsg(struct socket *sock, struct msghdr *m,
 		tipc_node_distr_xmit(sock_net(sk), &xmitq);
 	}
 
-	tsk_advance_rx_queue(sk);
+	if (!skb_cb->bytes_read)
+		tsk_advance_rx_queue(sk);
 
-	if (likely(!connected))
+	if (likely(!connected) || skb_cb->bytes_read)
 		goto exit;
 
 	/* Send connection flow control advertisement when applicable */
-- 
2.27.0


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

* Re: [PATCH net-next] tipc: keep the skb in rcv queue until the whole data is read
  2021-07-16 21:44 [PATCH net-next] tipc: keep the skb in rcv queue until the whole data is read Xin Long
@ 2021-07-17  0:30 ` patchwork-bot+netdevbpf
  0 siblings, 0 replies; 2+ messages in thread
From: patchwork-bot+netdevbpf @ 2021-07-17  0:30 UTC (permalink / raw)
  To: Xin Long; +Cc: netdev, davem, kuba, jmaloy, tipc-discussion, erin.shepherd

Hello:

This patch was applied to netdev/net-next.git (refs/heads/master):

On Fri, 16 Jul 2021 17:44:07 -0400 you wrote:
> Currently, when userspace reads a datagram with a buffer that is
> smaller than this datagram, the data will be truncated and only
> part of it can be received by users. It doesn't seem right that
> users don't know the datagram size and have to use a huge buffer
> to read it to avoid the truncation.
> 
> This patch to fix it by keeping the skb in rcv queue until the
> whole data is read by users. Only the last msg of the datagram
> will be marked with MSG_EOR, just as TCP/SCTP does.
> 
> [...]

Here is the summary with links:
  - [net-next] tipc: keep the skb in rcv queue until the whole data is read
    https://git.kernel.org/netdev/net-next/c/f4919ff59c28

You are awesome, thank you!
--
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html



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

end of thread, other threads:[~2021-07-17  0:30 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-07-16 21:44 [PATCH net-next] tipc: keep the skb in rcv queue until the whole data is read Xin Long
2021-07-17  0:30 ` patchwork-bot+netdevbpf

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).