And a quick update from me as well.

On the Multia, I have modified the MILO and FMU codes to boot from the
failsafe floppy
and write either the rom0(srm) or rom2(arc) flash modules.  I can use the
failsafe SRM
floppy and the `cat rom0flash > fat:rom0.dat/dva0' to confirm that the flash
does contain
what I wrote into it.  However,  when I move the jumper from failsafe mode
back to power-up
mode, the box will hang with no output.  I have turned on a whole host of
debugging output
that will display from MILO in failsafe floppy mode, but nothing from the
flash.

I have noticed that the UDB SROM requires the last byte of the even 256K  
flash images be correct or the SROM LED repeats a long/short blink after a
few initial blinks.
Normally, in an operative state, the LED only blinks during SROM
initialization and then is quiet.

At Phil's suggestion, I added code to the keyboard init routine to turn on
the keyboard
LEDs.  This works when I boot the MILO floppy, but no LEDs come on in the
flash
MILO startup state.

I was planning on moving the keyboard init call forward into the
uncompressed stub MILO
to see if I can get any method of "seeing into the box".  Currently, I get
no LEDs, no serial
output, and have turned off the TGA support in MILO because my 21" multisync
monitor
would not cope with the mode that MILO was running it in.  Also removing TGA
allowed 
the MILO image to drop below 256K so it would fit in one flash.


The burning question in my mind is:  Why will the MILO boot from the
failsafe floppy, but
not from the flash?

Thoughts and suggestions for further study are welcome.

Bill

                -----Original Message-----
                From:   [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED]]
                Sent:   Tuesday, January 16, 2001 8:13 PM
                To:     Ronald G Minnich
                Cc:     Jim Garlick; LinuxBIOS; Jeremy Jackson; Carr, Bill;
Ezolt, Phillip
                Subject:        Re: linuxbios on DS10

                O.k. just a quick update.

                On the DS10 I now have firmware writing to the SROM debug
serial port.

                So progress is being made, and it is now debuggable!

                As soon as I get output to a normal serial port I'll let see
about
                sharing my code ;)  

                The interesting piece that I hadn't realized before is that
                When the SROM transfers control to the loaded firmware you
are
                still in palmode.  

                The other nice thing is I have the my test image booting
with
                kexec, saving a number of ROM flashes.


                Eric


                

Reply via email to