search for: set_nodelay

...address, followed by an IPv4 address. The loop is designed to try connections until it finds one that works. Unfortunately, on my system, the IPv6 connection returns EINPROGRESS, which is interpreted as success. This causes the retry loop to exit, and the next thing that happens is a call to set_nodelay which fails when the IPv6 connection is not available. I think it would be better to put the call to set_nodelay inside the retry loop. Then if the call fails for IPv6, ssh can revert to IPv4. Of course, the error report in the call to set_nodelay would have to be suppressed is the connection...

...packet_set_interactive() in packet.c appears to have changed from 3.5p1. It appears there may be a typographical error. Specifically, 3.6.1p1 shows: /* Only set socket options if using a socket. */ if (!packet_connection_is_on_socket()) if (interactive) set_nodelay(connection_in); when it is probably supposed to say: /* Only set socket options if using a socket. */ if (!packet_connection_is_on_socket()) return; if (interactive) set_nodelay(connection_in); In other words, it appears somebody accidentily...

...9;s make it a debug message only... Index: misc.c =================================================================== RCS file: /cvs/openssh/misc.c,v retrieving revision 1.37 diff -u -p -r1.37 misc.c --- misc.c 22 Sep 2003 11:04:23 -0000 1.37 +++ misc.c 26 Oct 2003 10:48:35 -0000 @@ -97,7 +97,7 @@ set_nodelay(int fd) optlen = sizeof opt; if (getsockopt(fd, IPPROTO_TCP, TCP_NODELAY, &opt, &optlen) == -1) { - error("getsockopt TCP_NODELAY: %.100s", strerror(errno)); + debug("getsockopt TCP_NODELAY: %.100s", strerror(errno)); return; } if (opt == 1) { -- dwmw2

Hi, I'm observing a nasty and strange behaviour with OpenSSH (SSH-2.0-OpenSSH_3.7.1p2) on Solaris 8 (Sparc). I searched the FAQ and list archive and I didn't find anything about it. The problem is that uploading through sftp is tremendously slow (~ 0.2KB/s) while downloading is ok (~ 200-300 KB/s), so I'm quite surprised. The machines I tested (client & server) are all in the

...de <netinet/in_systm.h> #include <netinet/ip.h> #include <netinet/tcp.h> +#ifdef USE_SCTP +# include <netinet/sctp.h> +#endif #include <ctype.h> #include <errno.h> @@ -134,25 +137,42 @@ ssh_gai_strerror(int gaierr) } /* disable nagle on socket */ -void -set_nodelay(int fd) +static int +set_nodelay_proto(int fd, int proto, int optno, const char *pname) { int opt; socklen_t optlen; optlen = sizeof opt; - if (getsockopt(fd, IPPROTO_TCP, TCP_NODELAY, &opt, &optlen) == -1) { - debug("getsockopt TCP_NODELAY: %.100s", strerror(errno)); -...

...-socket Then things work normally. The problem seems to be that a "return;" has been deleted from the function "packet_set_interactive" in packet.c. The code reads: /* Only set socket options if using a socket. */ if (!packet_connection_is_on_socket()) if (interactive) set_nodelay(connection_in); The odd indentation suggests that something has gone wrong. The logic doens't make any sense, either: if the connection is not on a socket, then one shouldn't try to set socket options on it. Finally, in 3.4p1, the code reads: /* Only set socket options if using a socke...

...n using a proxy script to connect (I'm using the connect.c code found thru google) I get an error message trying to set TCP_NODELAY on a non-socket. I silenced the message by skipping the call to error only if errno == ENOTSOCK. There is probably a better way to handle this, maybe not calling set_nodelay when a proxy is in use? I just figured reporting it was better than keeping silent. Apply, mangle, or ignore as you see fit. See attached. -- Lee Eakin - leakin at dfw.nostrum.com I think our coffee machine is networked -- I keep seeing these dropped sugar packets all around it. --------...

...uo at isl.rdc.toshiba.co.jp> diff -u packet.c.dist packet.c --- packet.c.dist Mon Dec 23 11:42:53 2002 +++ packet.c Thu Apr 3 18:06:12 2003 @@ -1344,6 +1344,7 @@ /* Only set socket options if using a socket. */ if (!packet_connection_is_on_socket()) + return; if (interactive) set_nodelay(connection_in); #if defined(IP_TOS) && !defined(IP_TOS_IS_BROKEN)

...ot set on the main connection. This can cause buffering of data flowing from the server, but not the other direction. The fix that we have proven to work is to add the following code of the most recent source release: In the "main" function of "sshd.c": - add a call to "set_nodelay(newsock)" immediately after the "accept()" call. This will ensure that data travelling from this socket will not buffer and cause a bursting effect for small ammounts of data being sent at a fast rate. Thank you, Brian Genisio Oasis Advanced Engineering

...dvancedengineering.com When port forwarding is set up, TCP_NODELAY is set on the ports, in order to prevent buffering. This flag is not set in the actual SSH connection. This causes data that flows from the server to the client to be buffered, causing a bursing effect. The solution is to add set_nodelay(newsock) after the accept call in the main function. This solves the problem, but requires TCP_NODELAY to be set on all connections, regardless of port forwards. ------- You are receiving this mail because: ------- You are the assignee for the bug, or are watching the assignee.

...-r3.6.1.2 -r3.6.1.2.0.1 --- packet.c 2003/04/01 11:43:39 3.6.1.2 +++ packet.c 2003/05/10 21:45:22 3.6.1.2.0.1 @@ -1344,6 +1344,7 @@ packet_set_interactive(int interactive) /* Only set socket options if using a socket. */ if (!packet_connection_is_on_socket()) + return; if (interactive) set_nodelay(connection_in); #if defined(IP_TOS) && !defined(IP_TOS_IS_BROKEN)

I know that there was a discussion on this about a year back, and there is a bug 556 this week that mentions TCP_NODELAY. However, when I use ssh through a pipe (e.g., to tunnel through an HTTP proxy using CONNECT) I see: getsockopt TCP_NODELAY: Socket operation on non-socket How do I tell which end is generating this (I'm assuming the local side, which is running through the pipe). Also,

https://bugzilla.mindrot.org/show_bug.cgi?id=2016 openssh at ml.breakpoint.cc changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |openssh at ml.breakpoint.cc --- Comment #5 from openssh at ml.breakpoint.cc --- The link local address (fe80::) require the

Hello SSH developers, I am sorry if this is not really a bug, and I am missing something, but I am running into an issue with port forwarding in SSH. I am using 3.4p1 of SSH on both sides. I am running the ssh daemon on a Slackware Linux 8.0 machine, and the ssh client is running in Cywgin version 1.3.13. The ssh client is creating about 7 port forwards in a mix of local and remote forwarding

http://bugzilla.mindrot.org/show_bug.cgi?id=602 Summary: enormous bitching about netdb.h Product: Portable OpenSSH Version: 3.6.1p2 Platform: All OS/Version: All Status: NEW Severity: normal Priority: P2 Component: Build system AssignedTo: openssh-bugs at mindrot.org ReportedBy: tedm at

.../misc.h --- openssh-3.8p1/misc.h 2003-08-24 18:16:21.000000000 -0700 +++ openssh-3.8p1-localbind/misc.h 2004-08-05 11:02:08.916639344 -0700 @@ -14,6 +14,7 @@ char *chop(char *); char *strdelim(char **); +int strcount( char *, char ); void set_nonblock(int); void unset_nonblock(int); void set_nodelay(int); diff -u -r openssh-3.8p1/readconf.c openssh-3.8p1-localbind/readconf.c --- openssh-3.8p1/readconf.c 2004-08-05 08:59:24.775159000 -0700 +++ openssh-3.8p1-localbind/readconf.c 2004-08-05 12:53:24.203841440 -0700 @@ -203,8 +203,8 @@ */ void -add_local_forward(Options *options, u_short port...