Re: vax -current kernel trouble

[Henry Bent <hbent%cs.oberlin.edu@localhost>]

David Brownlee wrote:
> On Mon, 18 Feb 2008, Henry Bent wrote:
>
> > Hi all,
> >
> > I'm aware that vax may not be considered very stable right now, but at 
> > least a kernel builds again.  This looks to be the work of Matt Thomas 
> > so Matt, if you're reading this, thank you very much!
> >
> > On to the problem.  A -current kernel will begin booting, but as soon 
> > as it prints the memory sizes it drops right back to the ROM monitor.  
> > It looks like:
> >
> > ...
> > VAXstation 3100/m{30,40}
> > cpu: KA41/42
> > cpu: Enabling primary cache, secondary cache
> > total memory = 8060 KB
> > avail memory = 6076 KB
> >
> > and then I immediately get returned to the monitor.  The next line 
> > should be
> >
> > mainbus0 (root)
> >
> > I compiled a kernel with DEBUG on but didn't see any change at all.  
> > How should I go about debugging this?
>
>     The brute force option is to start dropping printf() lines in to
>     see exactly where its failing. I assume your familiar with
>     crosscompiling just the kernel to test changes? Its unfortunate
>     its not ending up in DDB which would make debugging this much
>     easier :/

I've been compiling only the tools and the kernel recently, on the 
assumption that if the kernel doesn't work there is certainly no need to 
bother with a userland.

Here's a pseudo-backtrace of what I found with printf():

uvm_uarea_alloc()
kthread_create()
create_idle_lwp()
mi_cpu_attach()

At this point I'm in a bit over my head with respect to debugging 
further.  Any suggestions would be welcome.

-Henry Bent