[lkml]   [1998]   [May]   [31]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
Messages in this thread
SubjectRe: 2.1.103 won't connect to
On Sun, May 31, 1998 at 05:32:01PM +0200, Carsten Gross wrote:
> Hello Andi and Developer list
> [Cc'ed to the List]
> Thanks for your answer.
> In article <>,
> (Andi Kleen) writes:
> [Description of my problem...and:]
> > Now it starts to be come strange. The other end floods us with ACKs.
> >
> > Can't see yet why. Is this reproducible? If yes, could you change ncftp
> > to set the SO_DEBUG option on the socket
> Yes, it is reproducible. I've fetched ncftp 2.4.3 (This version is a lot
> newer than my installed version) from a redhat mirror at
> ftp.rz.uni-karlsruhe. This server isn't a problem for me at all, I
> downloaded the 200 kbyte in one or two minutes.
> I compiled ncftp with the following line added around Row 494 (and at the
> top of the function, I set int flag = 1):
> if (setsockopt(sockfd, SOL_SOCKET, SO_DEBUG, &flag, sizeof(flag))<0) {
> perror("setsockopt");
> exit(1);
> }
> This version of ncftp has a timeout handling. It stops itselfs connecting
> after 30 seconds, but the problem remains:
> python:/home/carsten/ncftp-2.4.3#./ncftp
> (hanging, the prompt returns after 30 seconds with exit status 7)


That's very strange. Could you try to turn off RFC1323
(echo 0 > /proc/sys/net/ipv4/tcp_rfc1323), SACK (the same for
tcp_sack) and pmtu discovery (echo 1 > ..../ip_no_pmtu_disc)
It would be interesting if any combination of these flags works.

The only thing that looks strange in your packet are the timestamps,
maybe the PAWS old packet check discards them.


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to

 \ /
  Last update: 2005-03-22 13:42    [from the cache]
©2003-2014 Jasper Spaans. hosted at Digital Ocean