> Hi you all,
> 
> I wrote :
> > Well it isnt the bootloader i can asure you ... its somewhere in deep
> > space ... I can trace it back as to when the page0 is copied in ... or the
> > next thing that is done : setting up stuff....
> 
> Can someone with a faster machine than I have try out some positions ?
> e.g. insert a "do {} while (1);" further and further down until it breaks
> ? f.e. start after the l1 page table switch and maybe even after the
> printing ? That would give a good clue as to what sub-part goes wrong...
> 
> I think its interrupt stuff myself for that behaves strangely later on ...
> 
> Cheers,
> Reinoud
> 


I just had time last night to establish that death occurs in rpc_machdep, 
somewhere near

        /* Boot strap pmap telling it where the kernel page table is */
        printf("pmap ");
        pmap_bootstrap(kernel_l1pt.pv_va, kernel_ptpt);
        console_flush();

        /* Setup the IRQ system */
        printf("irq ");
        console_flush();

That is, the last thing I see before the screen (has someone tried to turn 
on interlace?) goes.  More precisely the last message to appear is "pmap 
".  I found this by putting in 5-second delays after each printf.

R.