More DSI traps

[Martin Husemann <martin%duskware.de@localhost>]

I updated my NetBSD/prep RS6000 machine to -current and tried the pthread
tests. There are lots of failures, probably most of them with a common course,
related to the DSI traps already discussed here.

I have:

cpu0 at mainbus0: 604e (Revision 2.3), ID 0 (primary)
cpu0: HID0 0xc084<ICE,DCE,SGE,BHT>, powersave: 1
cpu0: 233.34 MHz

The first test failing is:

cd /usr/tests/lib/libpthread && ./t_fork fork
trap: pid 440.1 (t_fork): user write DSI trap @ 0x114 by 0xeffbc300 (DSISR 
0x42000000, err=14)
failed: /usr/src/tests/lib/libpthread/t_fork.c:94: child died wrongly

The worst case is

        ./t_kill simple

in the same directory: it hangs the whole machine, and when I break into ddb
it faults midway in traceback, and reboots.

Does this work on any other ppc cpu? Since PR 44387 explicitly talks about
t_mutex, which runs later, I suppose it works on G4 or newer. Interestingly
I have so far been unable to reproduce the t_mutex failure on this cpu.


Martin