Steven James <[EMAIL PROTECTED]> writes:
> For slave nodes, this seems to fit the bill. With some debugging, it should be
> able to directly fetch a stage 2 boot image. Right now, it works great fetching
> the stage 1 boot.
>
> For the server situation, I'm not sure what is going to be the best solution
> when DoC isn't available (due to PLCC flash socket or whatever). At the least,
> the netboot hacks allowed me to get an 810LMR up and running. A simple netboot
> is nice, but not always appropriate.
Agreed. It does seem to make a good fallback piece of code.
Could we please hold off a day or two before putting the etherboot
derived code into the linuxBIOS source tree?
I've just written a preliminary implementation of reusing etherboot
whole. For long term maintenance and multiple NIC support reusing all
of etherboot looks to me like the better solution. Whole it is only
about 15KB uncompressed.
We definentily need to tackle the case where network booting isn't
appropriate but it definentily fixes the development case.
Actually I think this is cool I can now boot linux, etherboot, and
memtest86 directly from linuxBIOS. A generic ELF bootloader is pretty
simple to natively support. If for no other reason the tools for
making ELF images are already used in the build process.
Eric
