APM hosed in 2.1.28

Bob Tracy - TDS (rct@tracy-tds.wlk.com)
Wed, 5 Mar 1997 10:57:35 -0600 (CST)

I'll be tracking down and upgrading various components of my setup to
see if that helps, but somewhere in the upgrade from 2.1.27 to 2.1.28,
APM functionality left the road at a high rate of speed and is currently
mowing down weeds big-time. Specifically, running "apmd" as part of the
normal bootup process produces the following:

general protection: 0000
CPU: 0
EIP: 0010:[<c016506d>]
EFLAGS: 00010216
eax: 00000028 ebx: ffffffff ecx: 0000000a edx: 00000000
esi: ffffffff edi: 40000000 ebp: 00000000 esp: c197ef80
ds: 0018 es: 0018 ss: 0018
Process apmd (pid: 89, process nr: 5, stackpage=c197e000)
Stack: c1fff300 ffffffea 00000400 c19750f8 c01ce6b0 c194a000 ffffffff c0124ddb
c19750f8 c1fff300 40000000 00000400 c1fa6c0c 40000000 080795b8 bffffcb8
c010a438 00000003 40000000 00000400 40000000 080795b8 bffffcb8 00000003
Call Trace: [<c0124ddb>] [<c010a438>]
Code: f3 a5 89 d1 f3 a4 29 c8 75 0d 85 ed 75 33 bd f2 ff ff ff eb

Almost identical bad things happen if I try to run "apm". A simple
attempt to "cat" /proc/apm results in

Exception at [<c016506d>] (c01ab1ac)

with *lots* of "garbage" on the screen.

Somehow I don't think a point release or two difference in what "libc"
I'm using is responsible :-).

Bob Tracy		| "Friends help you move.  Real friends help
AFIWC/AFCERT		|  you move bodies."
rct@merkin.csap.af.mil	|	-- Anonymous