Adam Agnew <[EMAIL PROTECTED]> writes:

> 
> Hrmm, interesting. What approach did you use with this? Did you make an
> attempt at reading the file off of a filesystem on the floppy or are you
> assuming a raw mapped elf image (ie dd if=kernel.ebi of=/dev/fd0)?

The latter.   Actually now that I have fail safe booting, and flashing
from linux going it is ironic but it was faster to test this as part of
an image flashed into my ROM then to network boot it or similiar.

My round trip time was under a minute from typing make to having my code
start to boot up :)

And it was real fun everytime I hosed the firmware (about 100 times
last night) all I had to do was power off and power back on and the
system came right back up :)  

So I had real fun testing my fallback booting work :)

> Considering the latter is the same proceedure we'll use to chainload to
> yes another bootloader in our later scheme, I anticipate your code
> greatly.

I reached the put up or shut up stage.  It probably took me 30 hours with some
small naps to put up, but hey floppies drivers are much easier then
memory.  Especially if you have some one elses code you can look at
and can get inspired by. 

> As a status update, I have grub's code grafted into etherboot now and am
> now replacing it's int13 based disk access with polled. cross your
> fingers.

It will be interesting to see what comes out.  I'm probably going to look
at seeing if I can build a minimal bootloader out of the fill_inbuf
methods that we have in the rom.

Eric

Reply via email to