tg3 network cards

[Ryan Steele]

Ralf Utermann wrote:
> Ryan Steele wrote:
> [...]
>   
>> I should probably clarify here - obviously Broadcom takes most of the
>> blame, as this doesn't happen with nicer Intel NICs.  But I think
>>     
> Hi Ryan,
> then I don't understand why exactly the same hardware boots
> without problems from a Debian etch/FAI3.1.8 server. I should
> probably test the etch environment with an FAI kernel using a
> klibc-utils backported from lenny. Do you have a chance to
> test your system against an etch/fai3.1.8 server? 
> I would really like to understand this problem. 
>   

I don't have any Debian boxes handy.  These are all Ubuntu Hardy, but 
the packages should be pretty similar.  I'm going to try and take binary 
distributed with klibc-utils on one of my other boxes, and replace the 
one in the initrd with it.  I had assumed they are the same, but the 
checksums differ so maybe the one in the initrd generated by Ubuntu's 
update-initramfs is busted.

Just out of curiosity, what does your udev rule hack look like?  I wrote 
mine in Perl, which isn't available in the initramfs.  I'd be interested 
to see your implementation if you care to share it.

Respectfully,
Ryan