Subject: Debugging interrupt hangs
To: None <tech-ports@netbsd.org>
From: Peter Seebach <seebs@plethora.net>
List: tech-ports
Date: 02/28/2006 01:10:28
So, I'm trying to get some interrupt routing stuff worked out on my TAMS 3011,
and ... Suddenly the machine wedges during boot.  Specifically, it wedges
at the tail end of cpu_configure, when it tries to enable hardware interrupts.

The only difference in dmesg output, and I can't explain it, is that the
kernel used to say:

biomask 1c3c netmask 1c3c ttymask 1c3c

and now it says:

biomask 1c1c netmask 1c3c ttymask 1c3c

Unfortunately, I don't particularly understand what these numbers measure;
they're 16-bit values on a platform where many interrupts have numbers well
over 16... In fact, eliminating the casts to u_short gets me:

biomask 1c1c netmask 7f1c3c ttymask 7f1c3c

which is at least a little more plausible.

... But it doesn't explain what happened to 0x20 (that's, what, irq 6?), or
why the machine locks up now when it used to work.  Any suggestions?

-s