Hello,

When I run 'halt' the kernel prints:

Halting.
Shutdown: hdc
ACPI: PCI interrupt for device 0000:01:05.0 disabled
ACPI: PCI interrupt for device 0000:01:04.0 disabled
ACPI: PCI interrupt for device 0000:01:03.0 disabled
ACPI: PCI interrupt for device 0000:01:02.0 disabled
Power down.
acpi_power_off called

But the system does not shut down. (The fans keep spinning, the LEDs 
keep shining, the LCD keeps displaying.) Basically, the motherboard is 
still providing power to every component, as if the power supply had 
refused to stop.

Kernel is 2.6.22.1-rt9

I followed the instructions given here:
http://bugzilla.kernel.org/show_bug.cgi?id=6431

# cat /sys/module/acpi/parameters/debug_layer
Description                     Hex        SET
ACPI_UTILITIES                  0x00000001 [*]
ACPI_HARDWARE                   0x00000002 [*]
ACPI_EVENTS                     0x00000004 [*]
ACPI_TABLES                     0x00000008 [*]
ACPI_NAMESPACE                  0x00000010 [*]
ACPI_PARSER                     0x00000020 [*]
ACPI_DISPATCHER                 0x00000040 [*]
ACPI_EXECUTER                   0x00000080 [*]
ACPI_RESOURCES                  0x00000100 [*]
ACPI_CA_DEBUGGER                0x00000200 [*]
ACPI_OS_SERVICES                0x00000400 [*]
ACPI_CA_DISASSEMBLER            0x00000800 [*]
ACPI_COMPILER                   0x00001000 [*]
ACPI_TOOLS                      0x00002000 [*]
ACPI_ALL_DRIVERS                0xFFFF0000 [*]
--
debug_layer = 0xFFFF3FFF ( * = enabled)

# cat /sys/module/acpi/parameters/debug_level
Description                     Hex        SET
ACPI_LV_ERROR                   0x00000001 [*]
ACPI_LV_WARN                    0x00000002 [*]
ACPI_LV_INIT                    0x00000004 [*]
ACPI_LV_DEBUG_OBJECT            0x00000008 [*]
ACPI_LV_INFO                    0x00000010 [*]
ACPI_LV_INIT_NAMES              0x00000020 [*]
ACPI_LV_PARSE                   0x00000040 [*]
ACPI_LV_LOAD                    0x00000080 [*]
ACPI_LV_DISPATCH      ...
[ message continues ]

If I disable the 4 integrated NICs in the BIOS, then the kernel prints:

Halting.
Shutdown: hdc
Power down.
acpi_power_off called
  hwsleep-0322 [01] enter_sleep_state     : Entering sleep state [S5]


-

[ message continues ]

I have seen this behavior earlier on a system with the SMI interrupt
disabled. I do not know if this the case here, it is just a hint.
By the way, some distros bring the CPU in a halted state on a 'halt'
command, instead of powering off (actually very logical). For real
powering off these distros require the obvious 'poweroff' command.

Some long shots, maybe it helps...

Kind Regards,

Remy
-

[ message continues ]

Hello Remy,


Thanks for the suggestion. I didn't see anything related to SMM in the 
BIOS menus. However, the system has real-time constraints. Thus, I'd 
turn SMM off if I knew how :-)

    Phoenix - AwardBIOS v6.00PG, An Energy Star Ally
    Copyright (C) 1984-2003, Phoenix Technologies, LTD
*** NAMB-3140 BIOS V1.20 ***

Good suggestion. I'm using sysvinit-2.86
http://freshmeat.net/projects/sysvinit/

AFAIU, poweroff is equivalent to halt -p

halt (no option) calls reboot(RB_HALT);
poweroff or halt -p calls reboot(RB_POWER_OFF);
man 2 reboot

Alas, when I run 'poweroff' the kernel prints the same information:

Halting.
Shutdown: hdc
ACPI: PCI interrupt for device 0000:01:05.0 disabled
ACPI: PCI interrupt for device 0000:01:04.0 disabled
ACPI: PCI interrupt for device 0000:01:03.0 disabled
ACPI: PCI interrupt for device 0000:01:02.0 disabled
Power down.
acpi_power_off called

Regards.
-

[ message continues ]

Here you can find a driver that can disable and enable the SMI
interrupt in the chipset. It supports up to the ICH5 chipsets, but by
adding the proper device/vendor IDs you can also make it support newer
chipsets.

You are right.

Kind Regards,

Remy
-

[ message continues ]

(The original message seems to have been ignored by the mailing list 
robot, probably because the attachments made it too large. Re-send with 
links instead of attaching the documents to the message.)


http://linux.kernel.free.fr/halt/config-2.6.22.1-rt9
http://linux.kernel.free.fr/halt/acpidump.txt
http://linux.kernel.free.fr/halt/dmesg.txt
http://linux.kernel.free.fr/halt/halt.txt
http://linux.kernel.free.fr/halt/lspci.txt

Do you know what could be the problem?
(Meanwhile, I will investigate Remy Bohmer's suggestion.)

Regards.
-

[ message continues ]

Is there something else I can provide that might help in identifying
the problem?

Regards.
-

[ message continues ]

Could you please open bug at bugzilla.kernel.org and put all these files there?

Thanks,
Alex.


-

[ message continues ]

http://bugzilla.kernel.org/show_bug.cgi?id=9148

(In my browser, halt output is incorrectly displayed in UTF-8.)

Regards.
-

[ message continues ]

Writing 15361 (i.e. 0x3C01) to ACPI_REGISTER_PM1A_CONTROL appears to 
hang my system in acpi_os_write_port(). What can I do about that?

Is it a BIOS issue? a kernel issue? a hardware issue?

(All my results are attached to the bug report.)

Regards.
-

[ message continues ]

That is supposed to turn your machine off. At least we now know that ACPI 
did try to turn it off.
I think _other_ OS could turn your machine off just fine, so the issue is not HW, not BIOS.
Probably, first thing to try is 2.6.23.1 as it was just released and has some changes in
power management section...

Regards,
Alex.
-

[ message continues ]

I have the same problem in 2.6.23.1 (cf. my bug report in the database)

I'll ask the manufacturer whether they could get poweroff to work.

Regards.
-

[ message continues ]

This is an option to make. It creates .config file with some default settings,

-

[ message continues ]

Another observation: if I connect a screen to the system's VGA port, 
when I call 'poweroff' the screen goes into power saving mode. This 
seems to indicate that the integrated video card is properly shut down.

So the fans keep spinning, the LEDs keep shining, the LCD keeps 
displaying, but the video card is shut down?

Might this help pinpoint the problem?

Regards.
-

[ message continues ]

Check the BIOS to see if the "Power Button" is configured to
shut the system down. Some BIOS configure APM to do what
the power button does!

Cheers,
Dick Johnson
Penguin : Linux version 2.6.16.24 on an i686 machine (5592.59 BogoMips).
My book : http://www.AbominableFirebug.com/
_


****************************************************************
The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@analogic.com - and destroy all copies of this information, including any attachments, without reading or disclosing them.

Thank you.
-

[ message continues ]

Here is the relevant BIOS menu.

         Phoenix - AwardBIOS CMOS Setup Utility
                 Power Management Setup
+=====================================================+
|    ACPI Function             [Enabled]              |
|    MODEM Use IRQ             [NA]                   |
|    Soft-Off by PWR-BTTN      [Instant-Off]          |
|    CPU THRM-Throttling       [50.0%]                |
|    Resume by Alarm           [Disabled]             |
|  x  Date(of Month) Alarm       0                    |
|  x  Time(hh:mm:ss) Alarm       0 :  0 :  0          |
|                                                     |
|    ** Reload Global Timer Events **                 |
|    Primary IDE 0             [Disabled]             |
|    Primary IDE 1             [Disabled]             |
|    Secondary IDE 0           [Disabled]             |
|    Secondary IDE 1           [Disabled]             |
|    FDD,COM,LPT Port          [Disabled]             |
|    PCI PIRQ[A-D]#            [Disabled]             |
|                                                     |
+=====================================================+

          +===================================+
          | Soft-Off by PWR-BTTN              |
          |-----------------------------------|
          | Instant-Off  ..... [v]            |
          | Delay 4 Sec. ..... [ ]            |
          |                                   |
          |-----------------------------------|
          |  ^V:Move ENTER:Accept ESC:Abort   |
          +===================================+

'Instant-Off' is the appropriate setting, right?

In a different menu, there are two other (relevant?) options:

     PWRON After PWR-Fail      [On]
     Watch Dog Timer Select    [Disabled]

Regards.
-

[ message continues ]

Actually, default should be 4 sec delay. OS should have 
a chance to shut down the system...

-

[ message continues ]

I don't see why this setting would have an impact on the outcome
of the 'halt' and 'poweroff' commands.
-

[ message continues ]

Well, it is not possible to tell, what BIOS writer have connected to this flag...

-

[ message continues ]

(It sucks to be stuck with a closed proprietary BIOS.)

I tested the other setting, and it didn't change anything.
The system remains powered on after executing poweroff.

Len: the system is 100% Intel (Intel CPU, Intel north bridge, Intel 
south bridge, Intel integrated network controllers). Have Intel 
engineers run into the same problem on a similar platform?

http://bugzilla.kernel.org/show_bug.cgi?id=9148

Regards.
-

[ message continues ]