From: Willy Tarreau <w@1wt.eu>
Date: Sat, 1 Sep 2012 15:59:22 +0000 (+0200)
Subject: MEDIUM: connection: only call tcp_connect_probe when nothing was attempted yet
X-Git-Tag: v1.5-dev12~47
X-Git-Url: http://git.ipfire.org/?a=commitdiff_plain;h=f8deb0cfa8126e6132f6ff033f785e13da607e8a;p=thirdparty%2Fhaproxy.git

MEDIUM: connection: only call tcp_connect_probe when nothing was attempted yet

It was observed that after a failed send() on EAGAIN, a second connect()
would still be attempted in tcp_connect_probe() because there was no way
to know that a send() had failed.

By checking the WANT_WR status flag, we know if a previous write attempt
failed on EAGAIN, so we don't try to connect again if we know this has
already failed.

With this simple change, the second connect() has disappeared.
---

diff --git a/src/connection.c b/src/connection.c
index 6ed44080c9..15b94856ff 100644
--- a/src/connection.c
+++ b/src/connection.c
@@ -89,9 +89,10 @@ int conn_fd_handler(int fd)
 	if (unlikely(conn->flags & CO_FL_HANDSHAKE))
 		goto process_handshake;
 
-	if (unlikely(conn->flags & CO_FL_WAIT_L4_CONN)) {
-		/* still waiting for a connection to establish and no data to
-		 * send in order to probe it ? Then let's retry the connect().
+	if (unlikely(conn->flags & CO_FL_WAIT_L4_CONN) && !(conn->flags & CO_FL_WAIT_WR)) {
+		/* still waiting for a connection to establish and nothing was
+		 * attempted yet to probe the connection. Then let's retry the
+		 * connect().
 		 */
 		if (!tcp_connect_probe(conn))
 			goto leave;