Okay, all you Linux gurus. Help me out on this one.

I am starting a new line of business for my company, selling IP PBX's, based on Asterisk. I am using the Elastix build, which is based on CentOS 5.5.

I just purchased a Shuttle XS35-701, which is an Atom powered box, which draws 40W of power. It looks like it is going to make a great Asterisk appliance. My only problem is that CentOS doesn't recognize my network card. It is a JMicron JME260. JMicron has a "driver" on its web site. There are two versions for Linux: 1.0.5, available as a tar archive, and 1.0.6, in a rar archive. I have downloaded both a tried to install both. I get the same errors on either one.

The instructions say to expand the archive, and to change to the proper directory. Then is says to "make install" I have done this multiple times. I get a series of errors:

make[1]]: Entering directory 'usr/src/kernels/2.6.18-194.3.1.el5-x86_64'
CC [m] /root/jme-1.0.6/jme.o
In file included from /root/jme-1.0.6/jme.c:41:
/root/jme-1.0.6/jme.h:415:error:redefinition of 'ip_hdr'
Include/linux/ip.h:109:error:previous definition of 'ip_hdr' was here
/root/jme-1.0.6/jme.h:420:error:redefinition of 'ipv6_hdr'
Include/linux/ipv6.h:199:error:previous definition of 'ipv6_hdr' was here
/root/jme-1.0.6/jme.h:425:error:redefinition of 'tcp_hdr'
Include/linux/tcp.h:169:error:previous definition of 'tcp_hdr' was here
make[2]: ***[/root/jme-1.0.6/jme.o] Error 1
make[1]: ***[_module_/root/jme-1.0.6] Error 2
make[1]: Leaving directory 'usr/src/kernels/2.6.18-194.3.1.el5-x86_64'
make: ***[all] Error 2


After I run this, there is no jme.ko. I have no idea why not. I don't know what to do about it. Has anyone installed this JME260 NIC? Is there anyone who can help me?

Please don't suggest that I just put another NIC in the machine. It is built onto the motherboard, and has no slots to put another one into.

Welcome to CentOS.
New users should read:
http://www.centos.org/modules/newbb/viewforum.php?forum=47

I would start by reading the FAQ at the ELRepo Project's website.

If you still have a problem after following Gerald's advice, please post the output returned by executing ./getinfo.sh network, as detailed in How To Provide Information About Your System.

To be very specific, we will need to know the Vendor:Device ID Pairing (a.k.a. the fingerprint) for the NIC in question before any other advice can be given.

Thanks for the tip, Gerald. I went to the elrepo site, and found a forum article on my NIC. I followed what it said, and downloaded the rpm file. I ran the command rpm -Uvh kmod-jme-1.0.6.1-1.el5.elrepo.x86_64.rpm. I got a whole bunch of error messages, all reading like this:

kernel(rhel5_xxxx_ga)=[whole bunch of numbers] is needed by kmod-jme-1.0.6.1-1.el5.elrepo.x86_64

There were too many to fit on one screen. I know of no way to capture this output as a file to post.

I also know of no way to get the information out of the system that Alan suggested, as I have no NIC, and no writable drives. I have tried a flash drive, but I haven't successfully gotten that to work.

I know that I am missing something important here, but I have no clue as to what it migth be. Can you help me out?

Please post the output returned by the following --


-- along with the Vendor:Device ID Pairing that you have determined for the NIC in question. (Hints as to how to determine it can be seen in FAQ #4 on the ELRepo Project's website.)

Quote:
It goes something like this:

0. Put the getinfo.sh script on a USB stick.

1. Plug in USB stick.

2. Find the device.

3. Look at the output and find the USB device - let's assume it is /dev/sde1 for this exercise.

4. Make a mount point - say /tmp/mnt - and mount the device

5. Run the script (or alternatively the commands mentioned in post #5) with output to the USB stick.

6. Release the device.

rpm -qa kernel\*
kernel-2.6.18-194.3.1.el5
kernel-devel-2.6.18-194.3.1.el5
kernel-headers-2.6.18-194.3.1.el5
kernel-module-dahdi-2.3.0.1-3_2.6.18-194.3.1.el5
kernel-module-rhino-0.99.3-2.beta2_2.6.18-194.3.1.el5
kernel-wanpipe-3.5.14-0_2.6.18-194.3.1.el5

rpm -q centos-release
centos-release-5-5.el5.centos

vendor:hardwareID
197b:0260

Quote:
O.k., so looking on the Device IDs page, I see --


So there is an appropriate kmod package for that NIC. However I cannot tell you which one you should be using because you have failed to show me the first item of information requested, the output returned by uname -rmi.

For a 32-bit system there are the standard, the PAE and the xen versions.
For a 64-bit system there are the standard and the xen version.

From you earlier comments (post #4), I suspect you have a 32-bit system and have made the wrong choice between standard or PAE kmod package . . .

Thanks, Alan. That was just what I needed. I had tried all kinds of combinations of mount, with and without fs, with no success. I guess I just wasn't holding my mouth right.

Sorry about the uname stuff. I guess I am just getting old, and read past stuff now. The response on that was 2.6.18-194.3.1.el5 x86_64 x86_64. I was sure that it was 64 biit, because I had downloaded and installed the 64 bit version.

Quote:
Us too. Just noticed in post #1:
Quote:

It would be a good idea to update now that the networking is presumably working. It should be easy with the kABI tracking kmod module. Current kernel is 2.6.18-194.17.1.el5 and there have been a number of other recent security and bug-fix updates.

I'm guessing that this thread can now be marked [SOLVED] for posterity. [?]

I wish. Only the flash drive problem is solved. I thought that replying to that particular post would put the reply under the post. Obviously, I was wrong. I still get that host of errors when I try to install the kmod.

Quote:
D'oh!

As co-founder, along with Phil, of the Advanced-age Fools' Society I'll just continue to mutter to myself as I go in search of another

Quote:
You have to change "View mode" to "Threaded" near the right top of the page to see what was a reply to what. I nearly always use "Flat".

Quote:
That is Alan's territory. I'm stumped, or maybe just lost the answer in another wetware malfunction.

Still issues? That's odd.

Let me recap the facts as I understand them:

CentOS 5.5, 64-bit, with an earlier, not current, point 5 kernel-2.6.18-194.3.1.el5 installed. The NIC displays a fingerprint of 197b:0260.

The correct kmod-jme package for a standard (not xen) 64-bit el5 kernel is http://elrepo.org/linux/elrepo/el5/x86_64/RPMS/kmod-jme-1.0.6.1-1.el5.elrepo.x86_64.rpm

Currently I have two questions:

(1) If the system has no network connectivity, how did the original 5.5 installation kernel, 2.6.18-194.el5, come to be updated to 2.6.18-194.3.1.el5?
(2) And, related, from where and how were the following packages installed? --


Please show the output returned by --



Later edit: I think I have just spotted the problem. In your first post you said --

Quote:
An admission of not using true CentOS but a re-spin, thereof. I suspect that the suppliers have rebuild the kernel and have broken the el5 kABI. If that is the case, it will be proven when you post the above requested output.

And unfortunately, if that is the case, there is nothing the CentOS Community or the ELRepo Project can do to help you.

Please see the following CentOS wiki pages --

http://wiki.centos.org/AdditionalResources/OtherSpins
http://wiki.centos.org/AdditionalResources/OtherVoices

Even later edit: Could you possibly consider installing true CentOS, the kmod-jme, kmod-dahdi and dahdi-tools packages from the ELRepo Project ?

Quote:

It all came on the download of the Elastix software.

Quote:

Signature: DSA/SHA1, Wed 09 Jun 2010 01:13;39 PM EDT, Key ID 4bfe99f4d05c057c

Quote:

Linux version 2.6.18-194.3.1.el5 (mockbuilder@builder10.centos.org) (gcc version 4.1.2 20080704 (Red Hat 4.1.2-48)) #1 SMP Thu May 13 13:08;30 EDT 2010

You are probably right about the respin. The grep -ri exclude command did return something. It look innocuous, but it means that something has been changed. I will look into replacing the Centos with a true version and go from there. Thanks for all your help. You guys are the best.

Quote:
For CentOS:
It might be instructive to see all the output from
Quote:
That I don't understand. Seems to contradict the non-CentOS signing key, unless "mockbuilder@builder10.centos.org" comes from an unmodified gcc rather than from the kernel. Again, we need to await Alan's wisdom.

Quote:
Huh?

Wisdom? Huh? Phil's flattery will get him everywhere.

Quote:
That is something puzzling to me. Perhaps we should call for toracat?

Quote:
I'll second that request, please.

Equally puzzled here

Most likely the kernel is not from CentOS. But let's wait for the full output from rpm -qi.

Assume:
Wisdom >= 0
Puzzlement <= 0

Then:
sum(Wisdom) + sum(Puzzlement) ~= 0



NEEDINFO

Edit: In any case, an install of real CentOS 5.5 on the hardware would eliminate a lot of confusion.