PowerOff doesn't works anymore on Acer Ferrari laptop

[Chuck McKinnis]

This is a very long story started in the end of year 2005.

​http://ru.ecomstation.ru/hardware.php?action=item&id=897

System is in stable state since ACPI 2.21, while I'm still not satisfied, as I can't use Broadcom WiFi future, when 2Gb of RAM installed. ​http://comments.gmane.org/gmane.org.netlabs.genmac.user/1116
SMP/APIC mode probably may help to overcome this trouble (windows always works only in high Irq APIC mode), but I never managed to boot OS/2 in this mode (system freezed during boot). Don't ask me questions about possible system misconfiguration for this, as I have 3 desktop PCs working fine in SMP/APIC.

Recently I tried again with last ACPI build 3.07 and found that Power Off future didn't work anymore - system reboot instead of power off. Other thing didn't changed. I also tried with 3.05 build - with the same result. Latest working ACPI build is 3.03

ACPI log doesn't show anything very different from 3.03. I can provide if you need it - just tell me what ACPI boot switchs to set.

[Chuck McKinnis]

Just to not forget.

I installed debug version of ACPI.PSD. Then, according to ACPI developer instructions, switched to full screen session and start powerOff.exe

System switched back to PM, after that rebooted instead of Power Off. When I pressed Power Button in full screen session, I managed to see

PowerButton
OemHlpXHandler: CMD_SHUTDOWN 0
OemHlpXHandler: CMD_SHUTDOWN 0 (I'm not sure if it was 0 but not other digit here)

After that system rebooted.

[Chuck McKinnis]

Upgraded to ACPI 3.08

PowerOff still not work, same way as in 3.07.

[pasha]

Install debug version, boot with /DE:0xFFFFFFFF, then run full screen cmd session, run PowerOff.exe and attach photo of screen

[Chuck McKinnis]

Full screen photo of PowerOff

[pasha]

Problem not in acpi. Check other drivers, which can has problem with shutdown.

[Chuck McKinnis]

Problem is exactly in ACPI, as working versions up to 3.03 always fulfill power off fine. And nothing else was changed. If I downgrade ACPI, it start working again.

[pasha]

Replying to pasha:

Install debug version, boot with /DE:0xFFFFFFFF, then run full screen cmd session, run PowerOff.exe and attach photo of screen

If you do this, your photo say - problem outside of acpi. Check your drivers.

[Chuck McKinnis]

This is not a photo of still picture. I caught it after several attempts, as this lines appeared on screen only for a while, and the system doesn't trap or freeze, it reboot after.

other related setting:
1) os2 kernel version - debug
2) PSD switches:
/IRQ:10 /OSI="Windows 2001 SP1" /OS="Microsoft Windows NT" /DE:x0FFFFFFFF

Which drivers to check? Nothing extra installed, just standard driver set. And as I already told, exactly the same configuration works fine with ACPI up to 3.03

[pasha]

acpi01042008.zip, need photo of screen and acpi.log

[Chuck McKinnis]

First screen after PowerOff execution

[Chuck McKinnis]

Second screen after PowerOff execution. System still not reboot or Power Off

[Chuck McKinnis]

ACPI20080401.zip log

[pasha]

acpi02042008.zip, need photo of screen

[Chuck McKinnis]

C:\Poweroff.exe

Last 2 lines which, I managed to see on screen was

OemHlpXHandler: CMD_SHUTDOWN 0
OemHlpXHandler: CMD_SHUTDOWN 0

after that laptop screen switched of, while the system nether powered off, nether rebooted.

[pasha]

Problen in Intel ACPI CA, wait next version with "update Intel source" then try again

[Chuck McKinnis]

Thanks to Pasha for support.
I made more deep investigation and found the way how to return back poweroff functionality.
It seems that some GPE event activated right after writing to Power Control Registers, so power restored immidiatelly after power off.

I patch the ACPI.PSD driver the way it doesn't call AcpiHwEnableAllWakeupGpes() function in AcpiEnterSleepState() routine. After that Poweroff function start working again.
I don't think that this is a correct way to fix Poweroff issue, as this can disable wakeup from devices like LAN or modem (I don't use such future)

I also made some tests, trying to disable HW interrupts and/or disable GPE events before entering poweroff. Unfortunatelly, this doesn't help.

Hope this helps.
Dmitry.

[eco]

e-co:

ok! thank you for support!

what do you recommend for other users? is it necessary introduce switch for .PSD? /FORCEPOWEROFF or /DISABLEWAKEUPGPE or something like this..

[pasha]

From acpi.psd version 1.1 we know - Acer notebook need in fix acpi table.....

[Steven Levine]

Replying to pasha:

From acpi.psd version 1.1 we know - Acer notebook need in fix acpi table.....

This may be true, but some versions of acpi.psd worked correctly for the reporter, so there is more to the issue than just a bad acpi table.

In any case, the ticket is assigned to the feedback pending milestone. You should not be spending your time responding to tickets other than those assigned to the current milestone.

[pasha]

I don't know version of acpi ca, which can correct work with aser. Only partially work

[pasha]

Please download experimental ACPI build from Mensys site:

• Experimental build for you:

ACPI-APIC-BAT.ZIP

• Read instructions: ​http://ecomstation.ru/projects/acpitools/?action=testcase