From: Hilmar P. <hi...@we...> - 2004-12-18 13:28:00
|
Hi, I guess I'm missing just a small thing. I'm trying to do networking from coLinux to my windows host. No, for the time being there is no necessity to do Internet connection share or something similar. My xml file looks like that: <!-- This allows you to modify networking parameters, see the README or website for more information --> <network index="1" type="tap" /> Inside coLinux the /etc/network/interfaces: # The loopback interface auto lo iface lo inet loopback # (network, broadcast and gateway are optional) #auto eth0 #iface eth0 inet static iface eth0 inet dhcp #address 192.168.42.26 #netmask 255.255.255.0 #network 192.168.42.0 #broadcast 192.168.42.255 #gateway 192.168.42.1 iface eth1 inet static address 192.168.0.40 gateway 192.168.0.1 netmask 255.255.255.0 #network 192.168.0.0 #broadcast 192.168.42.255 eth0 I will use, when I boot Linux natively. After booting coLinux I type ifup eth1 (this is an Debian sarge) and get the following: root@preusse:~# ifup eth1 SIOCSIFADDR: No such device eth1: ERROR while getting interface flags: No such device SIOCSIFNETMASK: No such device eth1: ERROR while getting interface flags: No such device Failed to bring up eth1. root@preusse:~# On Windows side the TUN/TAP interface is configured. Ethernet adapter TunTap for coLinux: Connection-specific DNS Suffix . : IP Address. . . . . . . . . . . . : 192.168.0.1 Subnet Mask . . . . . . . . . . . : 255.255.255.0 Default Gateway . . . . . . . . . : and yes, as soon as I boot coLinux the interface gets activated. I've taken again the Debian 3.0r0 image for testing and couldn't find a difference in configuration and yes networking works with the image. Anybody has a clue, what am I missing? Thanks and Regards, Hilmar -- "When you have to kill a man it costs nothing to be polite." -- Winston Curchill, On formal declarations of war |
From: Martin K. <ka...@po...> - 2004-12-19 10:08:25
|
What does your kernel says? I use always network index 0 and then the image had to identify which network is it on. like this: mapping eth0 script /usr/share/doc/ifupdown/examples/ping-places.sh map 192.168.0.50/24 192.168.0.1 home map 192.168.127.2/24 192.168.127.1 colinux iface home inet dhcp iface colinux inet static address 192.168.127.2 netmask 255.255.255.0 of course, the TAP on windows side has 192.168.127.1 Hope that helps. Martin |
From: Hilmar P. <hi...@we...> - 2004-12-20 12:58:33
|
On 19.12.04 Martin Kanich (ka...@po...) wrote: Hi, > What does your kernel says? > Nothing. Sorry! > I use always network index 0 and then the image had to identify > which network is it on. like this: > > mapping eth0 > script /usr/share/doc/ifupdown/examples/ping-places.sh > map 192.168.0.50/24 192.168.0.1 home > map 192.168.127.2/24 192.168.127.1 colinux > Means: you have both kind of networks (DHCP when Linux native has booted and static when coLinux is up) on the same interface and leave the decision which config to use up to some scripts? Sounds cool! Will have a look at that. > iface home inet dhcp > iface colinux inet static > address 192.168.127.2 > netmask 255.255.255.0 > > of course, the TAP on windows side has 192.168.127.1 > Meanwhile the problem is solved partially. As soon as I reconfigure my network and use the same config for eth0 it comes up smoothely. I've seen http://braindamage.alal.com/archives/debian-user/20030912/3756.html and f'ups. I guess, that could help me. Kind Regards and Thanks! Hilmar -- "That must be wonderful! I don't understand it at all." |
From: Hilmar P. <hi...@we...> - 2004-12-20 18:30:57
|
On 19.12.04 Martin Kanich (ka...@po...) wrote: Hi, > I use always network index 0 and then the image had to identify > which network is it on. like this: > just a question: > mapping eth0 > script /usr/share/doc/ifupdown/examples/ping-places.sh > map 192.168.0.50/24 192.168.0.1 home > ^^^^^^^^^^^^ @home you're using dhcp, right? Where does this IP address come from? Is that the address your box should request and the DHCP server will hand it to the box if available? > map 192.168.127.2/24 192.168.127.1 colinux > > iface home inet dhcp > iface colinux inet static > address 192.168.127.2 > netmask 255.255.255.0 > > of course, the TAP on windows side has 192.168.127.1 > Clear so far. The usual static configuration of an interface. Thanks, Hilmar -- You know you have a small apartment when Rice Krispies echo. -- S. Rickly Christian |
From: Martin K. <ka...@po...> - 2004-12-20 19:09:59
|
Hilmar Preusse wrote: > On 19.12.04 Martin Kanich (ka...@po...) wrote: > > Hi, > > >>I use always network index 0 and then the image had to identify >>which network is it on. like this: >> > > just a question: > > >>mapping eth0 >> script /usr/share/doc/ifupdown/examples/ping-places.sh >> map 192.168.0.50/24 192.168.0.1 home >> ^^^^^^^^^^^^ > > @home you're using dhcp, right? Where does this IP address come from? > Is that the address your box should request and the DHCP server will > hand it to the box if available? Well I have DHCP server on 192.168.0.1 and the 192.168.0.50 is reserved for this purporse. > >> map 192.168.127.2/24 192.168.127.1 colinux >> >>iface home inet dhcp >>iface colinux inet static >> address 192.168.127.2 >> netmask 255.255.255.0 >> >>of course, the TAP on windows side has 192.168.127.1 >> > > Clear so far. The usual static configuration of an interface. Regards, Martin |
From: Hilmar P. <hi...@we...> - 2004-12-21 10:39:32
|
On 19.12.04 Martin Kanich (ka...@po...) wrote: Hi, > I use always network index 0 and then the image had to identify > which network is it on. like this: > <example config> > > of course, the TAP on windows side has 192.168.127.1 > > Hope that helps. > Thanks very much. works like a charm! Last question: Do you have a clue, why the TUN/TAP connection between Windows and coLinux is only 10Mbit? I'm afraid this is the bottleneck, when trying to send an xterm to the ethernet-if and then use the X server on W$ to display it. Does it make sense to play with the mii-tools within coLinux? Kind Regards, Hilmar -- Twenty Percent of Zero is Better than Nothing. -- Walt Kelly |
From: Holger K. <hol...@gm...> - 2004-12-21 11:06:12
|
> Do you have a clue, why the TUN/TAP connection between Windows and > coLinux is only 10Mbit? That's just a label. There is no 10MBit Limit. > Does it make sense to play with the mii-tools within coLinux? No. No real ethernet card there. |
From: Hilmar P. <hi...@we...> - 2004-12-22 12:16:09
|
On 21.12.04 Holger Krull (hol...@gm...) wrote: Hi, > >Do you have a clue, why the TUN/TAP connection between Windows and > >coLinux is only 10Mbit? > > That's just a label. There is no 10MBit Limit. > Hmm, did just a stupid performance measurement: 1. From outside into Linux when booted natively Script started on Wed Dec 22 10:29:02 2004 preusse@smithers:/data/> scp ENU_565_SOL_P04.ZIP hille@192.168.42.26:. Password: ENU_565_SOL_P04.ZIP 100% |*****************************| 237 MB 01:03 preusse@smithers:/data/> exit Script done on Wed Dec 22 10:31:04 2004 which is good for a 100MBit connection. 2. Booted coLinux from W$ and did the same from the W$ box preusse@preusse ~ $ scp ENU_565_SOL_P04.ZIP hille@192.168.0.2:. ENU_565_SOL_P04.ZIP 100% 238MB 689.0KB/s 05:53 preusse@preusse ~ $ which is good for 10Mbit. Is that explainable by other causes or is there possibly a workaround? Thanks, Hilmar -- I have yet to see any problem, however complicated, which, when looked at in the right way, did not become still more complicated. -- Poul Anderson |
From: Holger K. <hol...@gm...> - 2004-12-22 13:56:13
|
> 2. Booted coLinux from W$ and did the same from the W$ box > > preusse@preusse ~ $ scp ENU_565_SOL_P04.ZIP hille@192.168.0.2:. > ENU_565_SOL_P04.ZIP 100% 238MB 689.0KB/s 05:53 > preusse@preusse ~ $ > > which is good for 10Mbit. > > Is that explainable by other causes or is there possibly a workaround? I didn't notice it so far, but yes the speed for scp transfers from colinux to outside (800KB/s) are low compared to the speed outside to colinux (5MB/s). Strangely the speed rises to 1.3MB/s if i reduce the mtu on the colinux box (ifconfig eth0 mtu 1300). Still slow. I have no idea why this is asymetric. |
From: Hilmar P. <hi...@we...> - 2004-12-22 14:54:05
|
On 22.12.04 Holger Krull (hol...@gm...) wrote: Hi, > >2. Booted coLinux from W$ and did the same from the W$ box > > > >preusse@preusse ~ $ scp ENU_565_SOL_P04.ZIP hille@192.168.0.2:. > >ENU_565_SOL_P04.ZIP 100% 238MB 689.0KB/s 05:53 > >preusse@preusse ~ $ > > > >which is good for 10Mbit. > > > >Is that explainable by other causes or is there possibly a workaround? > > I didn't notice it so far, but yes the speed for scp transfers from > colinux to outside (800KB/s) are low compared to the speed outside > to colinux (5MB/s). > Strangely the speed rises to 1.3MB/s if i reduce the mtu on the > colinux box (ifconfig eth0 mtu 1300). Still slow. I have no idea > why this is asymetric. > Well, in both cases I transferred from outside to (co)Linux. In the first case Linux was natively booted and the transfer came really from outside. In the latter case I just wanted to show you, that the connection between Windows and Linux is either: 1. a 10 MBit connection or 2. a badly configured 100 Mbit connection. However networking works now and that thing I'm describing here doesn't matter that much. Thanks, Hilmar -- After a number of decimal places, nobody gives a damn. |
From: Hilmar P. <hi...@we...> - 2004-12-28 13:10:07
|
On 22.12.04 Holger Krull (hol...@gm...) wrote: Hi, > >preusse@preusse ~ $ scp ENU_565_SOL_P04.ZIP hille@192.168.0.2:. > >ENU_565_SOL_P04.ZIP 100% 238MB 689.0KB/s 05:53 > >preusse@preusse ~ $ > > > >which is good for 10Mbit. > > > >Is that explainable by other causes or is there possibly a workaround? > > I didn't notice it so far, but yes the speed for scp transfers from > colinux to outside (800KB/s) are low compared to the speed outside > to colinux (5MB/s). > To coLinux: preusse@preusse ~ $ scp huge_file.wmv hille@192.168.0.2:. huge_file.wmv 5% 14MB 820.9KB/s 05:15 ETA Killed by signal 2. lost connection From coLinux: preusse@preusse ~ $ scp hille@192.168.0.2:teTeX-beta/tetex-texmf-2.99.6.20041211-beta.tar.gz . tetex-texmf-2.99.6.20041211-beta.tar.gz 9% 8188KB 404.6KB/s 03:18 ETA Killed by signal 2. preusse@preusse ~ $ Asymmetric indeed, however that was not the point. Rather the fact, that this speed is not acceptable for an 100MBit connection. Maybe for a 10MBit connection. > Strangely the speed rises to 1.3MB/s if i reduce the mtu on the > colinux box (ifconfig eth0 mtu 1300). Still slow. I have no idea > why this is asymetric. > Doesn't help me. Reduces the the speed, when running scp. My uname string in coLinux: root@preusse:~# uname -a Linux preusse 2.6.8.1-co-0.6.2-pre6 #58 Sun Oct 24 21:48:46 IST 2004 i686 GNU/Linux The TUN/TAP is TAP-Win32 Adapter V8 (coLinux) Driver version is 8.0.0.1, MTU is 1500 root@preusse:~# ifconfig eth0 eth0 Link encap:Ethernet HWaddr 00:FF:9E:F3:20:00 inet addr:192.168.0.2 Bcast:192.168.0.255 Mask:255.255.255.0 inet6 addr: fe80::2ff:9eff:fef3:2000/64 Scope:Link UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:56999 errors:0 dropped:0 overruns:0 frame:0 TX packets:68505 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:1000 RX bytes:3158398 (3.0 MiB) TX bytes:97666770 (93.1 MiB) Interrupt:2 Running X applications over that connection is quite slow. Maybe there are others reasons I'm not aware of. Regards, Hilmar -- "All flesh is grass" -- Isiah Smoke a friend today. |