Thinkpad T60 hangs from time to time and sometimes even completely...

[ecsnl]

With the ACPI315.zip

Build Level Display Facility Version 6.12.675 Sep 25 2001 (C) Copyright IBM Corporation 1993-2001 Signature: @#netlabs dot org:3.15#@##1## 29 Apr 2009 10:17:52 pasha::: :0::@@ ACPI core PSD Driver. (c) netlabs.org 2005-2009 Vendor: netlabs dot org Revision: 3.15 Date/Time: 29 Apr 2009 10:17:52 Build Machine: pasha File Version: 3.15 Description: ACPI core PSD Driver. (c) netlabs.org 2005-2009

Sometimes my Thinkpad T60 pauzes for a couple of seconds. This means the all programs are frozen and when playing audio file it loops. This can also be triggered by pulling out the power plug or plugging it back in.

As from the email Steve wrote:

This does not occur for 3.14, so there is some sort of regression that needs to be corrected.

One sequence that appeared several times was

0178:f9aca6fa 00000000 00000046 00000008 f9009a84 AcpiEvAddressSpaceDispatch + 24a 0178:f9ab35f1 f9a980c8 00000000 00000046 00000008 AcpiExAccessRegion + 141 0178:f9ab3ac9 f8df3118 00000000 f9009a84 00000000 AcpiExAccessRegion + 619 0178:f9ab2f78 f8df3118 00000000 f9009a84 00000000 AcpiExExtractFromField + d4 0178:f9ab140c f8df3118 f9a93a74 00000004 00000011 AcpiExReadDataFromField + 29c 0178:f9ace483 f8d84e90 f8df3118 f9009af4 f9a9b511 AcpiExResolveNodeToValue + 353 0178:f9ad11c7 f8d85014 f8d84e90 f8d85014 f8d85014 AcpiExResolveToValue + 10e 0178:f9abb291 f8d85014 f8d84e90 00000000 00000000 AcpiDsEvaluateNamePath + a1 0178:f9ab70aa f8d84e90 f8d84e90 00000000 f8d84eb8 AcpiDsExecEndOp + ca 0178:f9adb734 f8d84e90 f8d84e90 00000000 f8d84e90 AcpiPsParseLoop + 414 0178:f9ab992a f8d84e90 f8d84368 ffc6a4a2 f8d84c80 AcpiPsParseAml + 153 0178:f9ab2b61 f8d84c80 00000000 f8d84368 f8d84368 AcpiPsExecuteMethod + 2af 0178:f9ad53f0 f8d84368 00000000 f8dfecbd f8d84368 AcpiNsEvaluate + 210 0178:f9acdbc4 f8d84368 00000000 f8dfecbd f8dfecb8 AcpiEvaluateObject + 1bd 0178:f9aa83ff f9a98258 f8dfecbd 00000000 00000000 ECEvaluate + 1c 0178:f9aa3ace f8dfecb8 00001000 26cb0010 ffffffff CheckEvent + 4d 0178:f9aa2aa8 00000000 00001000 26cb0010 f9009ccc _PSD_APP_COMM + 7b0 0178:f9aa1c5e f9009ccc 00000000 f9ae001e 26cb0010 OemHlpXHandler + 1d4 0178:f9aa5b0f 00108896 007c5fd0 007e0000 10001000 OemHlpX_Stub32 + f

I didn't figure out which event this was for.

We also saw

0178:f9aca6fa 00000000 00000046 00000008 f9068ad8 AcpiEvAddressSpaceDispatch + 24a 0178:f9ab35f1 f9a980c8 00000000 00000046 00000008 AcpiExAccessRegion + 141 0178:f9ab3ac9 f8df3118 00000000 f9068ad8 00000000 AcpiExAccessRegion + 619 0178:f9ab2f78 f8df3118 00000000 f9068ad8 00000000 AcpiExExtractFromField + d4 0178:f9ab140c f8df3118 f9a93a74 00000004 00000011 AcpiExReadDataFromField + 29c 0178:f9ace483 f8d84ae0 f8df3118 f9068b48 f9a9b511 AcpiExResolveNodeToValue + 353 0178:f9ad11c7 f8d84c64 f8d84ae0 f8d84c64 f8d84c64 AcpiExResolveToValue + 10e 0178:f9abb291 f8d84c64 f8d84ae0 00000000 00000000 AcpiDsEvaluateNamePath + a1 0178:f9ab70aa f8d84ae0 f8d84ae0 00000000 f8d84b08 AcpiDsExecEndOp + ca 0178:f9adb734 f8d84ae0 f8d84ae0 00000000 f8d84ae0 AcpiPsParseLoop + 414 0178:f9ab992a f8d84ae0 f8d841c8 ffc6ae2a f8d84ae0 AcpiPsParseAml + 153 0178:f9ab2b61 f8d84ae0 00000000 f8d841c8 f8d841c8 AcpiPsExecuteMethod + 2af 0178:f9ad53f0 f8d841c8 00000004 00f0f924 f8d841c8 AcpiNsEvaluate + 210 0178:f9acdbc4 f8d841c8 00000004 00f0f924 00f0f924 AcpiEvaluateObject + 1bd 0178:ffc4aaff f9a9d2d8 ffc4c234 00000000 f9068ca0 SendPowerEventFun + 114 0178:ffc4af77 00000000 00f0f924 f9068d00 ffc49db3 GetACStatus 0178:ffc49db3 00f0f930 00000008 00000004 00000cd2 IOCtl + 1dc 0178:ffc4915e f8dc1cd2 5440d483 000000f0 069004b8 SendDDNotify + 45 0178:04b80000 00400690 0cd20000 01800000 f95b0000

If I repeat pulling the power plugin and reinserting it after some time my system will freeze completely. Its frozen in the sense that there is no more activity in PM, mouse and keyboard are not responding. Steve did find out the debug kernel was still alive. Not even a memory dump could be made with CTRL ALT F10.

The file short_machine_freeze.txt is output of the PSD (buffer machine empty on purpose). This "Emm" entries at the end could point to what is happening. This dump from acpica$ was made directly after such a short lockup was heard in the audio.

But even when playing no audio Steve and I have also observed this type of freeze and when we went inside with the kernel debugger to find it in this loop. It can be UNIAUD, that is true, we need todo more debugging later today. However this malfunctions did not occur with ACPI 3.14 or earlier.

When I had these short freezes without AUDIO playing I could not type for a few seconds.