Re: kern/43565: acpi and no acpi on boot and halt or poweroff

[David Holland <dholland-bugs%netbsd.org@localhost>]

The following reply was made to PR kern/43565; it has been noted by GNATS.

From: David Holland <dholland-bugs%netbsd.org@localhost>
To: gnats-bugs%NetBSD.org@localhost
Cc: 
Subject: Re: kern/43565: acpi and no acpi on boot and halt or poweroff
Date: Sun, 4 Jul 2010 20:13:54 +0000

 On Sat, Jul 03, 2010 at 09:40:00PM +0000, 
daniel.meynen%homily-service.net@localhost wrote:
  > Jul  3 22:07:35 comete /netbsd: pms_enable: command error 35
  > Jul  3 22:07:47 comete /netbsd: pckbport: command timeout
  > Jul  3 22:07:58 comete /netbsd: pckbport: command timeout
  > Jul  3 22:07:58 comete /netbsd: pms_disable: command error
  > Jul  3 22:08:10 comete /netbsd: pckbport: command timeout
  > Jul  3 22:08:10 comete /netbsd: pms_enable: command error 35
 
 Do you have usb -> ps/2 emulation in the BIOS (for either the keyboard
 or a mouse)? If so, you may be able to get this crap to go away by
 disabling it.
 
  > if I boot without acpi, there is always a problem, if I use halt or
  > if I use poweroff to halt the system:
  > 
  > for example, if I use halt, this message is displayed:
  > 
  > ACPI Error (hwacpi-0156): No SMI_CMD in FADT, mode transition failed 
[20080321]
  > ACPI Error (evxfevnt-0221): Could not exit ACPI mode to legacy mode 
[20080321]
 
 If you've disabled ACPI it shouldn't be doing that... or so I'd think...
 
  > or if I use poweroff, this other message is displayed:
  > 
  > syncings disks... 1 done
  > unmounting file systems... done
  > uvm-fault(0xffff80005672e008, 0x0, 1) -> e
  > fatal page fault in supervisor mode
  > trap type 6 code 0 rip ffffffff806df42f cs 8 rflags 10293 cr2 0 cpl 8 rsp 
ffff800056bd6b40
  > panic: trap
  > Begin traceback...
  > uvm-fault(0xffff80005672e008, 0x0, 1) -> e
  > fatal page fault in supervisor mode
  > trap type 6 code 0 rip ffffffff804f46fc cs 8 rflags 10246 cr2 0 cpl 8 rsp 
ffff800056bd66e0
  > panic: trap
  > Faulted in mid-traceback; aborting
  > 
  > Here the system is freezed.
 
 Wonderful. :(
 
 Can you check what function 0xffffffff806df42f belongs to in your
 kernel?
 
 The easiest way to do this is
 
    % gdb /netbsd
    (gdb) x 0xffffffff806df42f
    0xffffffff806df42f <some_kernel_function+105>: 0x123456
    (gdb) q
                        ^^^^^^^^^^^^^^^^^^^^
 
 Another way is to run nm -n on the kernel image (this sorts the
 symbols in the kernel in numeric order) and use less to search for a
 prefix of the number.
 
 This information may not turn out to be very useful but it's at least
 easy to get.
 
 -- 
 David A. Holland
 dholland%netbsd.org@localhost