Subject: [Juha Nurmela: Handling of NMI]
To: None <port-i386@sun-lamp.CS.Berkeley.EDU>
From: Chris G. Demetriou <cgd@nobozo.CS.Berkeley.EDU>
List: port-i386
Date: 11/09/1993 18:36:45
somebody wanna look into this?



------- Forwarded Message

	id SAA01598; Tue, 9 Nov 1993 18:19:06 -0800
	id SAA01580; Tue, 9 Nov 1993 18:18:51 -0800
	(5.65c/IDA-1.4.4/Kim-2.2) id AA19644; Wed, 10 Nov 1993 04:19:02 +0200
From: Juha Nurmela <>
Message-Id: <>
Subject: Handling of NMI
Date: Wed, 10 Nov 93 4:19:02 EET

I am not sure, if you have already fixed this, but
the T_NMI case in trap() uses #if NISA > 0, but includes
a wrong isa.h (the one from .../isa, not the one from .../compile).
So isa_nmi() gets never called and cannot tell if the cause
was bad memory or check from bus..

Just added a button to nmi-line, printing a stacktrace on nmi
is _very_ nice, works on splhigh too :)

I found out that to acknowledge
an nmi one needs to set and clear bit 0x8 from 0x61,
otherwise it wont work again.

Juha Nurmela,

------- End of Forwarded Message