I started having the "bha_init_ccb: can't create DMA maps" panic after
I upgraded the system memory from 40M to 96M (although NetBSD is
using only 64M of it).  After about 6 consecutive days of crashing
shortly after 2am, it stopped.  This was strangely coincidental
with my ceasing use of that system as an NFS server.  After about
7 or 8 days with no crashes, it has started crashing again, and
has done so for the past 8 consecutive days so far, still shortly
after 2am.

I tried running the test program below, and it swapped until swap
was full and never had a problem.

Other than a few idle xterms and a window manager or two, the only
thing running is the daily security script.

How do I use the kernel debugger to view variables from various
stack frames?  Is it possible to call kernel subroutines from the
debugger?

-- Chuck



Jason Thorpe writes:
>On Thu, 2 Oct 1997 19:17:43 -0400 (EDT) 
> Brad Spencer <brad@anduin.eldar.org> wrote:
>
> > Hello....
> > 
> > I'v been trying to update my Pentium from 1.2B to 1.2G and have run into
> > the panic on the 1.2G kernels:
> > 
> > panic: bha_init_ccb: can't create DMA maps
>
>This is on my list of things to look at soon'ish.
>
>I am using a Buslogic on PCI, and am not seeing these problems.  I do
>lots of disk i/o through that controller, so I'm going to need some help
>nailing this down...
>
>	(a) I need to know _exactly_ what's happening on the system
>	    when this occurs.
>
>	(b) I need to know LOTS of gory details about your hardware.
>
>Can you help me debug this?  I'd really like to get it fixed :-)
>
>I will run your test program to see if I can reproduce it...
>
> > I had someone build some -current kernels for me, in case there was
> > something odd about the way I built the 1.2G kernel and the same panic
> > occures [For a number of reasons, the GENERIC kernel won't work for me].
> > At this point, I can't help but suspect that something is wrong.
> > 
> > The following program will cause the panic around the time it starts to
> > swap:
> > 
> > #include <stdio.h>
> > 
> > main()
> > {
> > 	char *c;
> > 	int n=0;
> > 
> > 	while (c=(char *)malloc(8192)) {
> > 		*c=123;
> > 		n++;
> > 		fprintf(stderr,"%d ",n);
> > 		usleep(500);
> > 	}
> > 	sleep(2000);
> > }
> > 
> > 
> > I was unable to panic the i386 1.2B system or a sparc 1.2G system.
> > 
> > Any help or suggestions would be appreciated.
> > 
> > 
> > Thanks..
> > 
> > Brad Spencer - brad@anduin.eldar.org   http://anduin.eldar.org
>
>Jason R. Thorpe                                       thorpej@nas.nasa.gov
>NASA Ames Research Center                            Home: +1 408 866 1912
>NAS: M/S 258-6                                       Work: +1 415 604 0935
>Moffett Field, CA 94035                             Pager: +1 415 428 6939