Using volume buttons on Thinkpad T60 traps machine

[ecsnl]

While playing music on my T60 I had closed/opened the screen a couple of times, exact order not known. Its very difficult to find a pattern in the events what leads to the TRAP (like how many times the screen was opened and closed). But its always related "ACPI events" like the trap occurring EXACTLY on the moment when you hit the volume button on the keyboard again.

On the 17th of May with ACPI 3.15 from the 11th of may , I got a TRAP 0008 (module unknown) with CS:EIP 168:F96D16AC. Because laptop was running with the switches /APIC /SMP I could make no memory dump. In ACPI.CFG the embedded controller was set to Polling. See picture attached trap0008_acpi315.jpg

How Steve and I got to the point of finding this out I can't remember. But at one moment we ran a test with the SMP kernel 104a (retail) and ACPI 3.15 with the /SMP only. It would always trap when the network logon would place on the domain controller or when there was network traffic (with a few seconds). The trap in this case was at CS:EIP 168:f96d16ac. This was exactly the same as the TRAP with trap0008_acpi315.jpg when hitting the volume button.

We also have a memory dump of this ACPI 3.15 logon trap.

I tried to reproduce this TRAP with ACPI 3.16: Signature: @#netlabs dot org:3.16#@##1## 29 May 2009 04:33:40 :0::@@ ACPI core PSD Driver. (c) netlabs.org 2005-2009 Vendor: netlabs dot org Revision: 3.16 Date/Time: 29 May 2009 04:33:40 Build Machine: pasha File Version: 3.16 Description: ACPI core PSD Driver. (c) netlabs.org 2005-2009

I also attached an acpica$ log to this ticket. This was created with ACPI 3.16 from the 29th of May.

I try to reproduce the TRAP with the volume buttons with ACPI 3.16. But this did not work. The logon TRAP, with ACPI.PSD started only with /SMP, I was able to reproduce again. I don't know if its useful a picture is attached to this bug. The CS:EIP was different.

[eco]

e-co:

Please use other stress test:

start 5-10 copies of this .cmd:

/* ACPI Battery Stress test             */
/* the goal is cause heavy load on ACPI */

/*   Run 5 copies of this script and    */
/*  attach/detach AC, adjust brigthness */ 
  
/*  If no traps then embedded controller works perfectly */
/*  else the notebook is not ready for every day usage   */

:1
acpibattery
goto 1

[ecsnl]

Replying to eco:

e-co:

Please use other stress test:

start 5-10 copies of this .cmd:

/* ACPI Battery Stress test             */
/* the goal is cause heavy load on ACPI */

/*   Run 5 copies of this script and    */
/*  attach/detach AC, adjust brigthness */ 
  
/*  If no traps then embedded controller works perfectly */
/*  else the notebook is not ready for every day usage   */

:1
acpibattery
goto 1

Start 10 copies and ran for 2 minutes with music running changing brightness and volume while playing music (test done with ACPI 3.16). Nothing happened.