Date: Tue, 3 Aug 1999 23:47:16 -0700 (PDT) From: Richard.Kiss@kalamalka.gizzywump.com To: FreeBSD-gnats-submit@freebsd.org Subject: ports/12949: rtsp_proxy port fixes Message-ID: <199908040647.XAA48502@kalamalka.gizzywump.com>
next in thread | raw e-mail | index | archive | help
>Number: 12949
>Category: ports
>Synopsis: Fix for rtsp_proxy port. Patches from Jonathan Hogg (jonathan@onegoodidea.com).
>Confidential: no
>Severity: serious
>Priority: medium
>Responsible: freebsd-ports
>State: open
>Quarter:
>Keywords:
>Date-Required:
>Class: sw-bug
>Submitter-Id: current-users
>Arrival-Date: Tue Aug 3 23:50:01 PDT 1999
>Closed-Date:
>Last-Modified:
>Originator: Richard Kiss
>Release: FreeBSD 3.2-STABLE i386
>Organization:
>Environment:
>Description:
Fix for rtsp_proxy port; improves reliability on various systems.
>How-To-Repeat:
>Fix:
Here's the patch:
diff -ruN rtsp_proxy/Makefile rtsp_proxy-kiss/Makefile
--- rtsp_proxy/Makefile Tue Aug 3 22:24:14 1999
+++ rtsp_proxy-kiss/Makefile Tue Aug 3 23:38:49 1999
@@ -3,7 +3,7 @@
# Date created: 8 June 1999
# Whom: Richard Kiss <richard@homemail.com>
#
-# $Id: Makefile,v 1.2 1999/08/01 02:35:05 steve Exp $
+# $Id: Makefile,v 1.1.1.1 1999/06/16 16:47:36 billf Exp $
#
DISTNAME= SS1.0.1
diff -ruN rtsp_proxy/patches/patch-ac rtsp_proxy-kiss/patches/patch-ac
--- rtsp_proxy/patches/patch-ac Wed Dec 31 16:00:00 1969
+++ rtsp_proxy-kiss/patches/patch-ac Tue Aug 3 23:26:47 1999
@@ -0,0 +1,22 @@
+--- ../proxy.c Tue Jul 6 17:42:19 1999
++++ ../proxy.c Tue Jul 20 22:26:39 1999
+@@ -1142,7 +1142,18 @@
+ }
+ //
+ // did we get a complete response yet?
+- if (! has_two_crlfs(s->sinbuf))
++ //
++
++// this used to look for a blank line, which is a dumb way to check if we
++// have a complete response since that only means we have a complete header.
++// the correct method would be to check the 'Content-Length:' header, but
++// that's too much trouble without a lot of re-writing. so this is a cheap
++// hack to look and see if the buffer ends on a 'CRLF' pair - this might get
++// broken if we happen to break on an end-of-line, but it's better than
++// nothing.
++
++ pBuf = s->sinbuf + s->amtInServerInBuffer - 2;
++ if ((pBuf[0] != '\r') || (pBuf[1] != '\n'))
+ break;
+
+ //
diff -ruN rtsp_proxy/patches/patch-ad rtsp_proxy-kiss/patches/patch-ad
--- rtsp_proxy/patches/patch-ad Wed Dec 31 16:00:00 1969
+++ rtsp_proxy-kiss/patches/patch-ad Tue Aug 3 23:26:34 1999
@@ -0,0 +1,29 @@
+--- ../shared_udp.c Tue Jul 6 17:42:26 1999
++++ ../shared_udp.c Tue Jul 20 23:54:38 1999
+@@ -259,7 +259,16 @@
+ gNextPort++;
+ if (gNextPort > gUDPPortMax)
+ gNextPort = gUDPPortMin;
+- } while (bind_socket_to_address(skt1, fromIP, port1 = gNextPort++, false) != 0);
++ } while (bind_socket_to_address(skt1, INADDR_ANY, port1 = gNextPort++, false) != 0);
++
++// i've changed the previous bind and the next one below to bind to any
++// IP number rather than using 'fromIP'. 'fromIP' is obtained from gProxyIP
++// which is in turn obtained from 'get_local_ip_address'. this function
++// gets the current hostname and then resolves that into an address. however,
++// since a proxy server runs on a bridging system with multiple IP addresses,
++// this is as likely as not to return the wrong IP address, which results in
++// the proxy missing all of the traffic from the server. not binding to a
++// specific address is cheap but gets a result.
+
+ if (withSib) {
+ retry_rtcp:
+@@ -271,7 +280,7 @@
+ else
+ goto bail_error;
+ }
+- if (bind_socket_to_address(skt2, fromIP, port2 = gNextPort++, false) != 0) {
++ if (bind_socket_to_address(skt2, INADDR_ANY, port2 = gNextPort++, false) != 0) {
+ close_socket(skt1);
+ close_socket(skt2);
+ skt1 = INVALID_SOCKET;
diff -ruN rtsp_proxy/patches/patch-ae rtsp_proxy-kiss/patches/patch-ae
--- rtsp_proxy/patches/patch-ae Wed Dec 31 16:00:00 1969
+++ rtsp_proxy-kiss/patches/patch-ae Tue Aug 3 23:26:13 1999
@@ -0,0 +1,16 @@
+--- ../proxy.h Tue Jul 6 17:42:19 1999
++++ ../proxy.h Tue Jul 20 22:26:39 1999
+@@ -80,7 +80,12 @@
+ trans_pb RTCP_C2S_tpb;
+ } track_info;
+
+-#define RTSP_SESSION_BUF_SIZE 1024
++// 1024 is too small for some server messages, so i've up'd this to 2048
++// instead as the proxy will otherwise truncate the message before sending
++// it on to the client, this will confuse the client no end.
++
++#define RTSP_SESSION_BUF_SIZE 2048
++
+ typedef struct rtsp_session {
+ struct rtsp_session *next;
+ int die;
>Release-Note:
>Audit-Trail:
>Unformatted:
To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-ports" in the body of the message
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199908040647.XAA48502>