Date: Sun, 19 Oct 2003 21:25:26 +0200
   From: Frank van der Linden <fvdl@netbsd.org>

   On Sun, Oct 19, 2003 at 09:25:11PM +0200, Mark Kettenis wrote:
   > Anyway, I committed an
   > NetBSD/amd64 port yesterday.  It doesn't do core-dumps yet, but I have
   > some pending changes that should make those work.  Backtraces through
   > signal trampolines are working AFAICT.  However, there is some strange
   > interaction between ptrace(2) and pending signals; in particular
   > SIGALRM.  If you try to single-step a program with a SIGALRM pending,
   > GDB doesn't see the SIGALRM and doesn't step either.  This smells like
   > a kernel bug to me.  Haven't looked into it yet though.

   The problem I have with backtracing through signal handlers is that
   the frame that's "above" the signal handler isn't correctly seen.
   I need to look at that.. I'll look at your code. It's probably
   too late to pull gdb 6 into NetBSD in order to get it into
   NetBSD 2.0, so I'm going to have to stick to gdb 5.3 for now.

Fixing backtraces through signal handlers is rather difficult in GDB
5.3.  The architectural changes that were made this spring make it
much easier to do it right in GDB 6.0.

   I did fix a signal problem today in the kernel, perhaps it
   was related to your problem.

The behaviour has changed.  The SIGALRM test pass now, but
unfortunately there is a test further down in signal.exp that hangs
the machine; I can't ping the machine and the console is dead.  I'll
need to investigate this, since I don't even know exactlt what test is
failing.

Mark