> I've played around with Feng Shuo's patch and
>
> - removed some gcc warnings from netboot/*.c
> - translated etherboot printf() stuff to grub printf()
> - tried to reproduce when undi fails
>
> Sometimes the kernel gets transferred via tftp
> and gets started, and sometimes this starting fails.
>
> Whenever I started tcpdump at the same time on the
> dhcp/tftp server this starting did not fail once.
> Whatever this means.
>
>
> Is there someone else working on this strange
> starting behaviour?
>
> I'm using debian/unstable as server and
> an eepro100 on the client.
>
> You can find my diff on:
>
> http://student.physik.uni-mainz.de/~reiffert/grub-0.95.undi.patch
> http://student.physik.uni-mainz.de/~reiffert/grub-0.95.undi.patch.gz
>
>
> Please give me feedback.

I tested a similar implementation of the undi driver with grub and it 
works only with 3com nics. Other nics doesn't work. So, I think the 
difference between the cards are: the 3com code is most stable and the 
bootix code on the other cards is very buggy. 

I hope undi will work with grub2.

regards


_______________________________________________
Bug-grub mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/bug-grub

Reply via email to