According to Simon J. Gerraty:
>
>We never got as far as calling sema_p() before the watchdog reset
>occurred.  Probably because we are still using only the sp cache flush
>routines up to this point.  Without the printfs, the boot reliably
>gets to:
>

This may sound dumb but could it be that because you have put the
printf's in that your kernel is taking longer to boot and getting
clunked by the watchdog much sooner in the boot process simply because
the boot process is taking longer?  I mean, really, the whole idea of
the watchdog is to clunk the processor if it does not get reset
regularly on the basis that Something Bad has happened.

BTW anyone know where I can get a cheap hypersparc module?  I have one
but two would be v. cool :-)

-- 
===============================================================================
Brett Lymn, Computer Systems Administrator, BAE SYSTEMS
===============================================================================