[OpenBIOS] OpenBIOS/LinuxBIOS on a NIC ROM? Or am I mad?

[David Christensen]

Peter,

An interesting idea but I don't think it's possible.  On RESET the CPU will
begin executing instructions from a fixed address of F000:FFF0h (being x86
compatible).  At this point the PCI bus structure has not been initialized
and the OPROM on the PCI NIC has not been enabled.  Thus, you can't replace
the BIOS this way.  If you could somehow modify the card such that it comes
up enabled and decoding it's ROM in the F000h segment, and assuming it's on
PCI bus 0, you might be able to do such a thing since the PCI card should
claim the memory accesses rather than letting them filter down to the ISA
bus where the ROM exists.

Dave

-----Original Message-----
From: owner-openbios at elvis.informatik.uni-freiburg.de
[mailto:owner-openbios at elvis.informatik.uni-freiburg.de]On Behalf Of
Peter Lister
Sent: Wednesday, May 17, 2000 2:07 PM
To: openbios at elvis.informatik.uni-freiburg.de
Cc: P.Lister at sychron.com
Subject: [OpenBIOS] OpenBIOS/LinuxBIOS on a NIC ROM? Or am I mad?



First, my belated congratulations to Ron for his multiuser boot. I've been
lurking on the openBIOS list for a bit. It's interesting read, but I've had
nothing to contribute, other than vague encouragement.

Sychron is developing software for clusters (currently on Linux). Direct
control of cluster hosts without commercial M$ obsessed BIOSs getting in the
way would be extremely valuable to us. I'd *really* like a
serial-line-console
only boot; no kb or VGA (goodbye KVM switch). Currently, we only need
support
for serial lines, 905B/C NICs and IDE disks.

Be gentle with me if the following is a totally dumb idea. IANABH *

Our software has a driver for 3com 905B/C NICs; like many NICs, BIOS ROMs
can
be fitted for diskless network boots (this is of some interest, but has
always
seemd a bit too much effort). But, on reading (for completely unrelated
reasons) the 3Com data book, I see that these NICs have the facility for
drivers to *write* to a 64K or 128K Atmel PEROM (AT29C512 or AT29C010).

So I think to myself... is it possible / practical to write openbios /
linuxbios into a NIC PEROM and then boot from it, overriding the motherboard
BIOS? If so, the advantages seem great: programming is independent of (and
non-destructive to) the motherboard and its BIOS; if it all goes wrong, all
I
do is blank the PEROM and my system is no longer a brick. :-)

We have Linux driver code to initialise and control 905B/C NICs, so we could
probably contribute a software Linux 3Com 905B/C PEROM "blower" quite
easily.
Other NICs may be able to do this, but I only have the 3Com book. We don't
have
linuxbios supported motherboards here, but if this seems likely to work, I
can
make a case for buying SiS boards. BTW, if anyone from SiS is listening,
Sychron may soon be purchasing (or influencing the purchase of) some large
clusters, so if this works, your co-operation may pay back quite well.

* I Am Not A BIOS Hacker.
--
Peter Lister         P.Lister at sychron.com    PGP (RSA): 0xE4D85541
Sychron Ltd          http://www.sychron.com  PGP (DSS): 0xBC1D7258
1 Cambridge Terrace  Voice: +44 1865 200211
Oxford OX1 1UR  UK   FAX:   +44 1865 249666


-
To unsubscribe: send mail to majordomo at freiburg.linux.de
with 'unsubscribe openbios' in the body of the message

-
To unsubscribe: send mail to majordomo at freiburg.linux.de
with 'unsubscribe openbios' in the body of the message