Bug 5107

Distribution: Suse 9.3

Hardware Environment:

Averatec 3270-EE1 Laptop
AMD Sempron 2800
Video: S3 Unichrome Pro VGA Adapter rev 01

See attached lspci -vv output


Problem Description:

Problem very similar to bug #4236.

Steps to reproduce:
# echo "3" > /proc/acpi/sleep
or
# echo "mem" > /sys/power/state
(computer falls asleep)
Press power button to wake up. Listen to fans (and probably hard disk?) spin up
but fail to do anything productive with the machine.

Usual tricks do not work 
   - tried suse debug process of booting with init=/bin/sh 
     (mentioned at www.susewiki.org/index.php/title=ACPI_suspend
   - tried using vbetool to change console on sleep then change back and
     restore state on resume
   - tried a newer kernel (in this case 2.6.12)
   
This could also be a DSDT issue so I explored that idea a bit...
I decompiled and recompiled my DSDT using iasl from Intel - it had three errors
- one of which seems related to S3 related code - that block is outside of any
scope statement or even any method. Not sure how to fix this one. The other 2
errors  are a complaint about an attempt to Store a value into DEG8 - again not
sure how to fix this. Later I retried this decompile/compile with the saved 
original dsdt.dat and for some reason only saw 1 error (shown below)


Here are the DSDT errors:

tux:/home/jperry/unpack/acpica-unix-20050624/compiler # iasl -tc dsdt.dat

Intel ACPI Component Architecture
ASL Optimizing Compiler / AML Disassembler version 20050309 [Mar 19 2005]
Copyright (C) 2000 - 2005 Intel Corporation
Supports ACPI Specification Revision 3.0

dsdt.dat     1: DSDT["
Error    1037 -    ^ syntax error, unexpected PARSEOP_NAMESEG, expecting
PARSEOP_DEFINITIONBLOCK

ASL Input:  dsdt.dat - 570 lines, 10368 bytes, 0 keywords
Compilation complete. 1 Errors, 0 Warnings, 0 Remarks, 0 Optimizations

-----------------------------------------

I suppose it could also be one of the other ACPI tables - since they
seem to be compiled using the Microsoft compiler. Here is the output of

dmesg |grep -i msft
ACPI: RSDT (v001 A M I  OEMRSDT  0x12000415 MSFT 0x00000097) @ 0x1dfd0000
ACPI: FADT (v002 A M I  OEMFACP  0x12000415 MSFT 0x00000097) @ 0x1dfd0200
ACPI: MADT (v001 A M I  OEMAPIC  0x12000415 MSFT 0x00000097) @ 0x1dfd0390
ACPI: OEMB (v001 A M I  AMI_OEM  0x12000415 MSFT 0x00000097) @ 0x1dfde040

the DSDT LOOKS (i'm guessing here) like it was compiled with the 
intel compiler....yet it still has errors

ACPI: DSDT (v001  E12K0 E12K0001 0x00000001 INTL 0x02002026) @ 0x00000000

Interesting messages from dmesg (full output of dmesg will be attached)
related to ACPI problems *probably also inidcating a bad DSDT file*

    ACPI-0071: *** Warning: Invalid 'package' argument
    ACPI-0323: *** Warning: Invalid _PSS data
    ACPI-0071: *** Warning: Invalid 'package' argument
    ACPI-0323: *** Warning: Invalid _PSS data

These warnings are also mentioned in another bug I was looking at.

I have not applied any ACPI4Linux patches to the kernel.
Am I correct in assuming these are being frequently incorporated into the 
latest Linux kernel? If they are then I'm puzzled why there is a patch at
the ACPI web site for kernel 2.6.12

Until I removed SMP support from the kernel I was seeing the reboot
on resume behaviour of bug #3586


This may be a simple "DSDT" issue...or something else....

Please let me know if you need more data to test this and isolate the issue.

Any help would be greatly appreciated - Thank you!