Subject: Re: R140, now mounting root over NFS
To: Kjetil B. Thomassen <kjetil@thomassen.priv.no>
From: Ben Harris <bjh21@netbsd.org>
List: port-arm26
Date: 12/17/2000 21:35:55
On Sun, 17 Dec 2000, Kjetil B. Thomassen wrote:

> Kernel sources updated yesterday, and new kernel made, now I get a bit
> further, but it panics:
> 
> NetBSD 1.5N (R140) #7 ...
> ...
> Avail 1920 KB
> ...
> panic: m_copym0 overrun
> Stopped in nfsio at cpu_Debugger+0x10 bl kbd_trap

Yeah, my A540's doing that too.  It usually manages to get a bit further
than yours though.  I currently have no idea why it's happening.

> > One possibility is to look at the UVM histories, making a note of the
> > timestamps of the last few entries, then type "c", wait a bit, drop back
> > into the debugger and re-run the history and see if any more things have
> > happened.  This can give you some idea of what's going on.
> 
> I did that, and at the time there were activity from time to time, but
> as I don't understand the kernel internals, I have no idea of what is
> going on.

Unfortunately, I can't really help you with the VM system, since I don't
understand most of it.

> 
> I think that if I am going to be of more help here I need to know more
> about how the kernel works internally. Any good ideas for what I should
> do to get some more insight?

Get "The Design and Implementation of the 4.4BSD Operating System" by
McKusick, Bostic, Karels and Quarterman.  It's rather out of date, but
will give you some idea of how the MI kernel is put together.  Some of the
section 9 manual pages are also worth a read -- mbuf(9) for our current
panic.

> I guess that this panic asks for a send-pr, so if you agree, then I will
> do that.

Yes.  I was planning on sending one, but got distracted by work and the
EtherH driver (which turned out to be the cause of my corrupt vector
table).

> I also looked in the source code for a driver for my HCCS IDE card, but
> there is no driver for it. That is why it is not being configured, so I
> guess I have to try to get the necessary information from somewhere and
> adapt the wd driver to work with it.

Yep.  IDE drivers aren't usually too hard.  You need to work out what
addresses the registers appear at and what the spacing between them
is.  Have a look at the dtide driver for a simple example.  Note, though,
that that driver doesn't work because I couldn't work out how to get it to
pass through interrupts.  The RISC OS driver didn't need them, but the wdc
driver depends on them rather.

> Any progress on finding the basic program to replace SetStation?

I just plugged the hard disc from my A3000 into my 540, and I've got a
copy of stationset, along with a few other toys I'd left behind.

<URL:ftp://ftp.netbsd.org/pub/NetBSD/arch/arm26/riscos/stationset,ffb>

-- 
Ben Harris                                                   <bjh21@netbsd.org>
Portmaster, NetBSD/arm26               <URL:http://www.netbsd.org/Ports/arm26/>