Hi,
unfortunately I am having trouble with Linux networking while trying to get
to the point where I can reproduce your issue. Unfortunately it's been a
while since I last tried something like this and I am a little hazy on how
this used to work.
I use Qemu and network bridging. Now I setup a bridge br0 to which the Qemu
network script adds the tun/tap interface. Optionally I add eth0 (network
card without cable plugged in) to the bridge.
I start up HelenOS and I configure an address on HelenOS side and on the
Linux side I configure an IP address on the bridge br0 interface.
I can ping in both ways and I can access HelenOS's web server and VNC server
from outside without any issue.
If I start netcat on Linux (nc -l 1234) and try to connect from HelenOS, I
get 'error connecting' from nterm, same as Agnieszka. But the tcp server
does not crash (how did you figure out it crashed, by the way?) The error is
because the Linux side sends 'RST' in response to SYN - as if nobody was
listening on the socket.
I thought maybe the bridge interface is somehow special and does not allow
listening, so I tried setting up an IP address on eth0, which I added to the
bridge. But I was not able to establish any communication between the
address on the eth0 interface and between HelenOS.
I tried this first on OpenSUSE, then on Gentoo with the same results.
Any ideas what I could be doing wrong?
Btw., Agnieszka, did you try with TCP connections from outside HelenOS to
the inside? There are plenty of options: telnet, vncviewer, or a web
browser.
Regards,
Jiri
---------- Původní zpráva ----------
Od: Agnieszka Tabaka <[email protected]>
Komu: HelenOS development mailing list <[email protected]>
Datum: 16. 6. 2014 23:18:28
Předmět: Re: [HelenOS-devel] TCP broken in mainline
"
On Mon, Jun 16, 2014 at 10:25 AM, Jakub Jermar <[email protected]
(mailto:[email protected])> wrote:
"Thanks for your report, but, I am afraid, it is not very useful like
this because it does not contain enough information for anyone to look
at that issue. Could you please elaborate on:
- exact steps needed to reproduce this (I cannot reproduce this with the
current mainline amd64 build, using remcons and e1000)
- include the compressed 'tcp' binary for reference
"
Hi Jakub,
Here's how to reproduce that issue:
1) Configure IP address on HelenOS host (either by DHCP or manually)
2) Start netcat on other PC (netcat -l 1234)
3) On HelenOS host: nterm <ip_addr_of_pc> 1234
4) nterm immediately exits due to connection error
5) ...however, connection is set up and netcat thinks that connection is
active (in Wireshark, there are SYN, SYN/ACK and ACK packets but no RST or
FIN)
6) type something in netcat and press enter
7) tcp server in HelenOS crashes
I have tested that on fresh checkout from HelenOS mainline, built with
PROFILE=amd64. I'm attaching "tcp" server binary.
Regards,
Agnieszka
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/listinfo/helenos-devel";
_______________________________________________
HelenOS-devel mailing list
[email protected]
http://lists.modry.cz/listinfo/helenos-devel
