Re: [Etherboot-developers] UNDI driver

[Vasil V. <vas...@sy...>]

On Fri, 30 May 2003, Michael Brown wrote:

> On Thu, 29 May 2003, Georg Baum wrote:
> > I did, but it did not work (cvs from today). The PXE implementation is Intel
> > Boot Agent Version 4.0.22, the NIC is an onboard eepro100. Here are the
> > messages:
> > PXE loader for etherboot
> > 0x27EK low memory
> >  PXE    unloaded
> > ROM segment 0x000 length 0x000 reloc 0x00020000
> > Etherboot 5.1.8 (GPL) Tagged ELF for [UNDI]
> > Relocating _text from: [00007e70,00017b80) to [076f02f0,07700000)
> > Boot from (N)etwork (D)isk (F)loppy or from (L)ocal? N
> > Probing pci nic...
> > [UNDI]Hunting for PnP BIOS...found $PnP at f000:6e20...ok
> > Hunting for pixies...found !PXE at 0009de0...ok
> > UNDI API call 0x0000 failed with status 0x006a
>
> PXE chaining is failing.  I will try to sort out the pxeloader code so
> that it leaves the !PXE structure in a valid state (unless one of Peter or
> Vasil has already done this).  This looks like the same problem Marty was
> experiencing.

I am bit rusty on this and don't have much time right now to try it, but
wouldn't it be enough to define PXELOADER_KEEP_UNDI? The only thing it does is
actually not to reduce the the memory as to free the UNDI memory.

Alternatively, if "leaving !PXE structure in a valid state" means we don't
stop the the UNDI, then just don't include the whole of pxeprefix.c .

I didn't quite follow what the UNDI driver requires, but once this is clear it
shouldn't be hard to just skip whichever steps shouldn't be done. I guess that
the whole of pxeprefix shouldn't be there.

Gread job, Michael. Now, I can just specify to a vendor that I want PXE
support and not what the NIC is. If the NIC has a driver in etherboot, then
great. If not, then I can still use etherboot.

In general, etherboot is turning into a great infrastructure.

Vasil