Subject: port-i386/3259: system locks hard during (or right after) autoconf in certain cases
To: None <>
From: None <>
List: netbsd-bugs
Date: 02/25/1997 22:14:13
>Number:         3259
>Category:       port-i386
>Synopsis:       system locks hard during (or right after) autoconf in some cases
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    gnats-admin (GNATS administrator)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Tue Feb 25 19:20:01 1997
>Originator:     Rafal Boni
	what's that?
>Release:        -current from new setroot to 2/23
	NetBSD/i386 1.2C, 486DX/33, 32Meg RAM, 256K ext. cache
	1 IDE/MultiIO card (all ports/IRQs/DRQs in the standard PC places...)
	1 3C507 (0x300, IRQ15, memory 0xd0000-0xd3ffff)
	1 VGA card

	[the IDE card *is* single-channel, so it's not using IRQ15]

System: NetBSD true-faith 1.2C NetBSD 1.2C (TRUE_FAITH) #1: Mon Feb 24 04:51:33 EST 1997 rafal@doppelganger:/usr/src/sys/arch/i386/compile/TRUE_FAITH i386

	Booting any kernel since setroot changes on the 486 will lock the
	machine either during autoconf (right after probe of ie0.. the 3c507)
	or right after the ipl masks are printed but before findroot() is

	(1) If the 3c507 is removed, the machine boots OK with the same kernel
	(2) If the spl0() in autoconf() is moved further down in the function,
	    the machine dies at the new location (either in spllower() or right
	    after emerging, it seems)

	This freeze is total.. The only way out is the reset button or the big
	red switch (alright, so it's small and white... 8-)

	[FWIW, the 3c507 seems to like to generate interrupts while being 
	 probed or initialized... I wouldn't think it would be the only device
	 to do so, but it's another data point...]

	Boot any recent -current kernel on my 486 (or any machine with a