HP XW4600 Workstation - Network card error whilst imaging

[Tom Elliott]

Is there a way we can find out what Firmware is expected for this particular nic?

I have added tigon information to the kernels and what not, but I’ve also had to add firmware to the kernels to get some things to operate.

[Sebastian Roth]

I am fairly sure this (old) NIC does not need any firmware but only the tg3 driver in the kernel. As you can see in the second screen shot the ethernet controller is using it (‘Kernel driver in use: tg3’).

@Robx64 Can you please run a debug session (Host -> Basic Tasks -> Debug) on this client, wait till you get to the shell and then see what you have in dmesg about ethernet. Command dmesg | less (you can scroll up & down and also search using ‘/’ - without the quotes) or dmesg | grep -i eth

[Robx64]

@Sebastian-Roth
Hehehe really ?
I know that, it’s quiet old and it should work but somehow unfortunately it won’t

[Robx64]

@Sebastian-Roth
I’m not sure since I left tech department many things changed. I’ll ask techs since when we have this particular model and was there any problem before.

Apparently there was no problem on Fog 0.32 at all and XW4600 worked beautifully

[Robx64]

@Sebastian-Roth said:

dmesg | grep -i eth

Is that enough ?

[Tom Elliott]

@Sebastian-Roth I think I may have figured it out.

The current drivers/net/ethernet/broadcom had the vendor ID (14e4) but it lacked the Device ID. I just added the DEVICE under TG3PCI_DEVICE_TIGON3_4 I added the line: #define TG3PCI_DEVICE_TIGON3_5755 0x167b in hopes that it would be found. I don’t know what else may be broke, but hopefully we’re kind of on to something?

[Tom Elliott]

@Robx64 I created a kernel with the patch I ‘Hope’ will work. I’m not publishing it to the normal kernel’s directory, but please download and try it.

https://mastacontrola.com/bzImageTG3

[Robx64]

@Tom-Elliott
Thank you Tom,
I’ll check and give you an answer shortly

[Robx64]

@Tom-Elliott

Once downloaded you’ll need to unpack the bz2 file using the command:
tar -xjf fog_trunk.tar.bz2
Then go into the **unextracted ** file. Follow these steps:
cd fog_trunk/bin
./installfog.sh

Then follow the onscreen prompts. After a little while, you should have a successfully installed FOG server.

[Tom Elliott]

@Robx64 ???

[Sebastian Roth]

Wow! I really wonder why those machines/NICs never had any problem in the past?!? Tom, you are right 14e4:167b does not seam to be in the code. I checked earlier versions and non of them had this ID included. How can that be??

@Robx64 I didn’t ask for the output of dmesg from the live CD but from a FOG debug session. Would be interesting to see if FOG can actually see the device. Could you please do Host -> Basic Tasks -> Debug. Then run dmesg | grep -i eth and lspci (full output of Ethernet controller so we might see if it is really using tg3). Thanks!

[george1421]

@Sebastian-Roth said:

Then run dmesg | grep -i eth and lspci (full output of Ethernet controller so we might see if it is really using tg3). Thanks!

I’m sure a few commands hard coded into a menu selection like this would be handy to have, especially if they were in the FOG advanced menu so a user could just run them upon request of a a dev.

[Sebastian Roth]

BCM 5755 support was added to the kernel back in 2006. See here. I am very sorry to say this but I kind of doubt that your patched kernel will work: http://git.kernel.org/cgit/linux/kernel/git/torvalds/linux.git/commit/?id=af36e6b6d7f4ad7a5ccfd14dfa71ec941255f93d (PCI ID is in include/linux/pci_ids.h)

[Wayne Workman]

@george1421 said:

@Sebastian-Roth said:

Then run dmesg | grep -i eth and lspci (full output of Ethernet controller so we might see if it is really using tg3). Thanks!

I’m sure a few commands hard coded into a menu selection like this would be handy to have, especially if they were in the FOG advanced menu so a user could just run them upon request of a a dev.

+1 great idea. Maybe even attempt to throw the results into /images/dev too ?

[Tom Elliott]

@george1421 the problem with doing such a feat is that it means adding a bunch of extra code to the inits. This may be more possible with fog-too as our mindset is to create a very basic set of scripts these scripts will be called via another interface that would allow this type of customization much easier. The current inits are more or less designed to work for our specific small set of tasks and not very easily adaptable.

[Robx64]

@Sebastian-Roth
Sure, I’ll do it asap