I tried this query on linux-usb-users, with no luck... can my local linux community help, please?
I've been following the instructions at http://www.linux-usb.org/USB-guide/x555.html to set up a two-machine USB network with a Linux box at one end.
The command
ifconfig plusb0 192.168.0.1 pointopoint 192.168.0.2
fails with the error message
SIOCSIFADDR: No such device plusb0: ERROR while getting interface flags: No such device SIOCSIFDSTADDR: No such device plusb0: ERROR while getting interface flags: No such device
Afterwards,
ping 192.168.0.2
gets no response (i.e. 100% packet loss.) Any ideas how to fix this, please?
plusb is loaded as a kernel module. The kernel version is 2.2.20.
The cable between the two machines is a Vivanco "USB Network Cable," part number 10758. It shows up in usbview as
Unknown Device Manufacturer: Prolific Technology Inc. Speed: 12Mb/s (full) USB Version: 1.00 Device Class: 00(>ifc ) Device Subclass: 00 Device Protocol: 00 Maximum Default Endpoint Size: 8 Number of Configurations: 1 Vendor Id: 067b Product Id: 0001 Revision Number: 0.01
Config Number: 1 Number of Interfaces: 1 Attributes: a0 MaxPower Needed: 100mA
Interface Number: 0 Name: (none) Alternate Number: 0 Class: ff(vend.) Sub Class: 0 Protocol: 0 Number of Endpoints: 3
Endpoint Address: 81 Direction: in Attribute: 3 Type: Int. Max Packet Size: 1 Interval: 1ms
Endpoint Address: 02 Direction: out Attribute: 2 Type: Bulk Max Packet Size: 64 Interval: 0ms
Endpoint Address: 83 Direction: in Attribute: 2 Type: Bulk Max Packet Size: 64 Interval: 0ms
At the far end is a Windows 2000 machine, configured to have IP address 192.168.0.2, and Default gateway 192.168.0.1.
When the Linux box is rebooted into Windows 98, the USB LAN works, i.e. I can share files across it using Network Neighborhood/My Network Places.
On Tue, Jul 15, 2003 at 11:04:04PM +0100, Dan Hatton wrote:
I tried this query on linux-usb-users, with no luck... can my local linux community help, please?
I've been following the instructions at http://www.linux-usb.org/USB-guide/x555.html to set up a two-machine USB network with a Linux box at one end.
The command
ifconfig plusb0 192.168.0.1 pointopoint 192.168.0.2
fails with the error message
SIOCSIFADDR: No such device plusb0: ERROR while getting interface flags: No such device SIOCSIFDSTADDR: No such device plusb0: ERROR while getting interface flags: No such device
Afterwards,
ping 192.168.0.2
gets no response (i.e. 100% packet loss.) Any ideas how to fix this, please?
plusb is loaded as a kernel module. The kernel version is 2.2.20.
Didn't they change the interface names? Have you tried "usb0" or "usbnet0" (both from vague memory) instead of plusb0?
J.
On Tue, 15 Jul 2003, Jonathan McDowell wrote:
Didn't they change the interface names? Have you tried "usb0" or "usbnet0" (both from vague memory) instead of plusb0?
Thnaks. Just had a crack at these, and plusb1, usb1, and usbnet1 for good measure. No joy. The error messages are slightly different now, for example:
SIOCSIFADDR: No such device plusb0: ERROR while getting interface flags: No such device pointtopoint: Unknown host ifconfig: `--help' gives usage information.
BTW, no relevant devices seem to exist in /dev. I have a vague idea that there exists a program called mkdev that maight help. Anyone know anything about this?
Dan Hatton dan.hatton@btinternet.com wrote:
pointtopoint: Unknown host
ITYM pointopoint
BTW, no relevant devices seem to exist in /dev. I have a vague idea that there exists a program called mkdev that maight help. Anyone know anything about this?
sh MAKEDEV and it's normally in /dev, I think.
On Wed, Jul 16, 2003 at 12:15:32AM +0100, Dan Hatton wrote:
On Tue, 15 Jul 2003, Jonathan McDowell wrote:
Didn't they change the interface names? Have you tried "usb0" or "usbnet0" (both from vague memory) instead of plusb0?
Thnaks. Just had a crack at these, and plusb1, usb1, and usbnet1 for good measure. No joy.
You have loaded the plusb module, right (or have hotplug running and configured to do so)? What do the logs (or the last half dozen lines of "dmesg") say?
J.
On Tue, Jul 15, 2003 at 11:04:04PM +0100, Dan Hatton wrote:
plusb is loaded as a kernel module. The kernel version is 2.2.20.
have you considered using a 2.4 series kernel? usb networking has recently had some updates in the 2.4 series.
Also what does the command ifconfig -a say when you have the module loaded? that should give you the device name.
Adam
On Wed, 16 Jul 2003 abower@thebowery.co.uk wrote:
have you considered using a 2.4 series kernel? usb networking has recently had some updates in the 2.4 series.
I'm slightly (only slightly) reluctant to go to 2.4.x, because it will involve upgrading my boot-loader to something other than the version in Debian stable (see Debian bugs #142421 and #163098.) I've had bad experiences before with mixing more recent bits and bobs into stable, although I think it would almost certainly be alright in this case.
On Wed, Jul 16, 2003 at 12:49:38PM +0100, Dan Hatton wrote:
On Wed, 16 Jul 2003 abower@thebowery.co.uk wrote:
have you considered using a 2.4 series kernel? usb networking has recently had some updates in the 2.4 series.
I'm slightly (only slightly) reluctant to go to 2.4.x, because it will involve upgrading my boot-loader to something other than the version in Debian stable (see Debian bugs #142421 and #163098.) I've had bad experiences before with mixing more recent bits and bobs into stable, although I think it would almost certainly be alright in this case.
I presume that you are using loadlin from those bugs, could you not use a bootloader like grub or lilo? Then you can use things that are part of Debian stable and not have to mix and match from testing/unstable although I would build a kpkg kernel using 2.4.21 kernel source (as that seems to fix many usb networking bugs for me, although YMMV) and just replace the bootloader (if you can)
Adam
On Wed, 16 Jul 2003 abower@thebowery.co.uk wrote:
have you considered using a 2.4 series kernel? usb networking has recently had some updates in the 2.4 series.
OK. I'm now the proud owner of a machine running kernel 2.4.18 through loadlin 1.6c.
Also what does the command ifconfig -a say when you have the module loaded? that should give you the device name.
Looks like it. The relevant bit of the output seems to be
usb0 Link encap:Ethernet HWaddr 7A:37:C8:F1:F9:5A BROADCAST MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:100 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
Therefore, I tried
ifconfig usb0 192.168.0.1 pointopoint 192.168.0.2
which resulted in the error message
SIOCSIFFLAGS: Broken pipe
This appears to be a known issue. Someone anonymous says "I think the root cause of the problem is flakey hardware from Prolific." http://www.geocrawler.com/mail/msg.php3?msg_id=8713512&list=4563, and the general opinion in the linux-usb-users list seems to be that the best solution is to get a different cable. Anyone have other ideas?
Need more help, please, folks.
I bought another USB network cable, this time one from Maplins, which says "USB NetLink" on the box, and appears to have a Genesys chip-set. With this one, and with help from http://www.linux-usb.org/usbnet/, the link is now up, at least as far as the Linux box (10.0.1.1) on one end is concerned. I've also, with help from http://www.linux.org/docs/ldp/howto/Net-HOWTO/index.html, got the USB cable to be the default route from 10.0.1.1 to 10.0.1.2, which is the Windows 2000 machine on the other end. Now I'd like to be able to do something with the link. However,
10.0.1.2 thinks the link is down ("network cable unplugged," in Windows parlance,) and
trying to ping either machine, from the other, results in 100% packet loss.
If I reboot 10.0.1.1 into Windows 98, 10.0.1.2 sees the link go up, and I'm able to manipulate files across it happily, using SMB. I reckon this indicates that the hardware is working correctly.
How do I get to do networky things across this link, with Linux at the 10.0.1.1 end, please?
Technical details:
output of ifconfig on 10.0.1.1 is
ppp0 Link encap:Point-to-Point Protocol inet addr:213.122.68.17 P-t-P:213.120.208.144 Mask:255.255.255.255 UP POINTOPOINT RUNNING NOARP MULTICAST MTU:1500 Metric:1 RX packets:697 errors:15 dropped:0 overruns:0 frame:0 TX packets:809 errors:0 dropped:57 overruns:0 carrier:0 collisions:0 txqueuelen:3 RX bytes:531067 (518.6 KiB) TX bytes:113456 (110.7 KiB)
usb0 Link encap:Ethernet HWaddr B6:A8:47:26:3B:A3 inet addr:10.0.1.1 Bcast:10.0.1.255 Mask:255.255.255.0 UP BROADCAST RUNNING MULTICAST MTU:1500 Metric:1 RX packets:0 errors:0 dropped:0 overruns:0 frame:0 TX packets:0 errors:0 dropped:0 overruns:0 carrier:0 collisions:0 txqueuelen:100 RX bytes:0 (0.0 b) TX bytes:0 (0.0 b)
Output of route on 10.0.1.1 is
Kernel IP routing table Destination Gateway Genmask Flags Metric Ref Use Iface 10.0.1.2 * 255.255.255.255 UH 0 0 0 usb0 213.120.208.144 * 255.255.255.255 UH 0 0 0 ppp0 10.0.1.0 * 255.255.255.0 U 0 0 0 usb0 default 213.120.208.144 0.0.0.0 UG 0 0 0 ppp0
The TCP/IP properties for this link, on 10.0.1.2, are
IP address: 10.0.1.2 Subnet mask: 255.255.255.0 Default gateway: 10.0.1.1 Preferred DNS server: 194.73.73.94 Alternate DNS server: 194.73.73.95
On Thu, 17 Jul 2003, Dan Hatton wrote:
Therefore, I tried
ifconfig usb0 192.168.0.1 pointopoint 192.168.0.2
which resulted in the error message
SIOCSIFFLAGS: Broken pipe
This appears to be a known issue. Someone anonymous says "I think the root cause of the problem is flakey hardware from Prolific." http://www.geocrawler.com/mail/msg.php3?msg_id=8713512&list=4563, and the general opinion in the linux-usb-users list seems to be that the best solution is to get a different cable. Anyone have other ideas?
Somewhere in the Network configuration for the USB connection in Win2000 should be an option to detect the link state.
Can't remember exactly where this is in 2000 or what it says, but uncheck it so that 2000 always thinks the link is up.
I've had problems like this before where Windows thinks the link is down when it is fine in reality
On Sunday 27 July 2003 21:19, Dan Hatton wrote:
got the USB cable to be the default route from 10.0.1.1 to 10.0.1.2, which is the Windows 2000 machine on the other end. Now I'd like to be able to do something with the link. However,
10.0.1.2 thinks the link is down ("network cable unplugged," in Windows parlance,) and trying to ping either machine, from the other, results in 100% packet loss.
If I reboot 10.0.1.1 into Windows 98, 10.0.1.2 sees the link go up, and I'm able to manipulate files across it happily, using SMB. I reckon this indicates that the hardware is working correctly.
Did you mean the registry hack at http://www.jsifaq.com/SUBE/Tip2400/rh2444.htm? I just tried that, to no avail.
On Sun, 27 Jul 2003, Wayne Stallwood wrote:
Somewhere in the Network configuration for the USB connection in Win2000 should be an option to detect the link state.
Can't remember exactly where this is in 2000 or what it says, but uncheck it so that 2000 always thinks the link is up.
I've had problems like this before where Windows thinks the link is down when it is fine in reality
On Monday 28 July 2003 21:33, Dan Hatton wrote:
Did you mean the registry hack at http://www.jsifaq.com/SUBE/Tip2400/rh2444.htm? I just tried that, to no avail.
Not sure if that hack does everything I was referring to or not. I thought there was a dialogue option somewhere maybe I am thinking of WinXP.
If I get a chance I will fire up a Win2000 box tomorrow and have a look.
On Mon, 28 Jul 2003, Dan Hatton wrote:
Did you mean the registry hack at http://www.jsifaq.com/SUBE/Tip2400/rh2444.htm? I just tried that, to no avail.
With a little more playing (to be precise, uninstalling all clients/services/protocols except TCP/IP on the Windows 2000 box,) I got the Windows 2000 machine to recognize that the link to the Linux machine was up, and both machines to ping each other successfully, using this registry hack. It even continued to work after I re-installed "File and printer sharing for Microsoft networks."
Now I need to go learn about iptables and smbd :-P.