Nick Bane wrote:

> By kernel loader I mean the bootldr kernel loader hence the params.h
> refers to the bootldr source not the kernel source.

Ah, okay, isn't confusion wonderful.

> It has worked for me in the past with board=aleph1 and I believe it has
> worked for David Bisset with build=hydra using my kernel builds but the
> message is not encouraging.
> 
> If I do a "set build=guralp" "params save" and "reset" I get the
> expected result using my bootldr and 2.4.25... .
> 
> What does cat /proc/balloon_board report for you?

~ $ cat /proc/balloon_board
Board type:  2.05
Board build: Red TCL dev board
CPLD version: 0x0b
~ $

Which would seem to confirm that it's broken.

I've had a close look at params.h, my version contains:

//PARAM( board, PT_STRING, PF_STRING, (long) "2.05", balloon_init );
//PARAM( build, PT_STRING, PF_STRING, (long) "tcl", balloon_init );
PARAM( board, PT_STRING, PF_STRING, (long) "2.05", NULL );
PARAM( build, PT_STRING, PF_STRING, (long) "tcl", NULL );

What did the (commented out) versions with balloon_init in them do?

I think there is still something I don't understand about how the boot
process is supposed to work. The boot.conf file has the correct build value
set, but what reads that and how is it passed to the kernel?

I did an experiment to try and narrow down where the problem is.

I rebuilt the bootldr with:

PARAM( build, PT_STRING, PF_STRING, (long) "guralp", NULL );

This works! The kernel startup messages and /proc/balloon_board both now
report that it's a "guralp" board. So it seems that the bootldr is failing
to use the parameters and is just using whatever default is set.

I've had a brief look at the code to see if I can see where it's going wrong
but didn't spot anything, do you have any ideas?

regards,

Colin

-- 
Colin Tuckley      |  colin@tuckley.org  |  PGP/GnuPG Key Id
+44(0)1903 236872  |  +44(0)7799 143369  |     0x1B3045CE

Some people drink at the fountain of knowledge, others just gargle.