shutdown with toshiba portege

shutdown with toshiba portege

[François Boisson]

Hey 

I have a problem which seems to be related to shutdown for a Toshiba portege
(it's seems that the problem is also on other Toshiba, on VAIO cw5, you can
look at this thread on debian-user-french
http://lists.debian.org/debian-user-french/2012/08/msg00164.html
)


Symptoms (under 3.0, 3.1, 3.2, 3.3 and 3.5.2 and 3.5.4 amd64 kernels but not
with 2.6.37-486-PAE kernel) under a 
TOSHIBA PORTEGE Z830/Portable PC, BIOS Version 1.60 

If i shutdown the computer and wait some hours, the power consumption of the
computer is not null, it's about the same as the power consumtion during a
suspend of computer. In fact

Boot and halt with:
Windows 7 ---> No power consumption after
Linux 2.6.37-486-PAE --->  No power consumption after
Linux >= 3.0  ---> Power consumption (about 45-50 mW if acpitool is correct)

I try 3.0, 3.1, 3.2.0 (Debian wheezy kernel), 3.3, 3.5.2 and 3.5.4 kernels.
(System is a debian wheezy amd64 with (now) personnal kernel)

First I think it was something like Wake-On-Lan then Wake-on-WLan but I found
nothing (WLAN and WWLAN are disabled or not present)
Then I look on kernel source and I have found this in
arch/x86/kernel/apic/apic.c :

[...]
/*
 * If Linux enabled the LAPIC against the BIOS default disable it down before
 * re-entering the BIOS on shutdown.  Otherwise the BIOS may get confused and
 * not power-off.  Additionally clear all LVT entries before disable_local_APIC
 * for the case where Linux didn't enable the LAPIC.
 */
void lapic_shutdown(void)
{
[...]

I wonder if my problem is the problem describe here or something like. It
seems that computer can be shutdown but some peripherics not power off. So I
try to «trace» the shutdown process but I did not found the description a the
kernel way to shutdown the computer. It seems that it's

* Shutdown all CPUs execpt the CPU0
* Shutdown local APIC
* Shutdown computer
but perhaps the problem is before these opérations.

Where can I verify that all is really power off before shutdown (with
judicious printk for instance)

I try to look in kernel documentation but I did not found.

Thanks for any help or advices and sorry for noise if you think it is.

François Boisson

Re: shutdown with toshiba portege

[François Boisson]

Le Thu, 20 Sep 2012 20:42:34 +0200
François Boisson  <user.anti-spam@maison.homelinux.net> a écrit:

> Hey 
> 
> I have a problem which seems to be related to shutdown for a Toshiba portege
> (it's seems that the problem is also on other Toshiba, on VAIO cw5, y
[...]
> If i shutdown the computer and wait some hours, the power consumption of the
> computer is not null, it's about the same as the power consumtion during a
> suspend of computer. In fact
> 
> Boot and halt with:
> Windows 7 ---> No power consumption after
> Linux 2.6.37-486-PAE --->  No power consumption after
> Linux >= 3.0  ---> Power consumption (about 45-50 mW if acpitool is correct)
[...]

Well can anybody just tell me the sequence of shutdown and power of a x86
computer in then linux kernell, or just how to verify that a device is power
off with a printk, or where to search or even only where is the part of code
doing this. There is many documentation about use device but nothing about
power off.

Thanks

François Boisson

Re: shutdown with toshiba portege

[Dan Carpenter]

On Fri, Sep 21, 2012 at 04:57:53PM +0200, François Boisson wrote:
> Le Thu, 20 Sep 2012 20:42:34 +0200
> François Boisson  <user.anti-spam@maison.homelinux.net> a écrit:
> 
> > Hey 
> > 
> > I have a problem which seems to be related to shutdown for a Toshiba portege
> > (it's seems that the problem is also on other Toshiba, on VAIO cw5, y
> [...]
> > If i shutdown the computer and wait some hours, the power consumption of the
> > computer is not null, it's about the same as the power consumtion during a
> > suspend of computer. In fact
> > 
> > Boot and halt with:
> > Windows 7 ---> No power consumption after
> > Linux 2.6.37-486-PAE --->  No power consumption after
> > Linux >= 3.0  ---> Power consumption (about 45-50 mW if acpitool is correct)
> [...]
> 
> Well can anybody just tell me the sequence of shutdown and power of a x86
> computer in then linux kernell, or just how to verify that a device is power
> off with a printk, or where to search or even only where is the part of code
> doing this. There is many documentation about use device but nothing about
> power off.
> 

So your Toshiba is going into suspend instead of power off...

There are several ways that computers shutdown but modern systems
use ACPI.  It could be that the 2.6.37-486-PAE was using a
different method.  I've CC'd the ACPI list.

regards,
dan carpenter

Re: shutdown with toshiba portege

[François Boisson]

Le Fri, 28 Sep 2012 16:23:55 +0300
Dan Carpenter <dan.carpenter@oracle.com> a écrit:

> 
> So your Toshiba is going into suspend instead of power off...
> 
> There are several ways that computers shutdown but modern systems
> use ACPI.  It could be that the 2.6.37-486-PAE was using a
> different method.  I've CC'd the ACPI list.
> 

Thinks, first I have open a bug 
https://bugzilla.kernel.org/show_bug.cgi?id=47811
the I have compiled a 3.5.4 with some printk around the halt/poweroff. (FIles
arch/x86/kernel/reboot.c
arch/x86/kernel/apm32.c ) trying to trace call BIOS but nothing printk, either
I do not understand the power off mechanism, or the computer is in a state
where printk do not appear on screen.
In fact I don't know what I can do to find an issue to this bug.

Thanks for your answer

François Boisson
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: shutdown with toshiba portege

[François Boisson]

> Le Fri, 28 Sep 2012 16:23:55 +0300
> Dan Carpenter <dan.carpenter@oracle.com> a écrit:
> 
> > 
> > So your Toshiba is going into suspend instead of power off...
> > 
> > There are several ways that computers shutdown but modern systems
> > use ACPI.  It could be that the 2.6.37-486-PAE was using a
> > different method.  I've CC'd the ACPI list.
> > 
> 
> Thinks, first I have open a bug 
> https://bugzilla.kernel.org/show_bug.cgi?id=47811
> the I have compiled a 3.5.4 with some printk around the halt/poweroff. (FIles
> arch/x86/kernel/reboot.c
> arch/x86/kernel/apm32.c ) trying to trace call BIOS but nothing printk,
> either I do not understand the power off mechanism, or the computer is in a
> state where printk do not appear on screen.
> In fact I don't know what I can do to find an issue to this bug.
> 
> Thanks for your answer

Just to say that the cause of battery to drain is making

hwclock --systohc 

just before halt computer. Add 
HWCLOCKACCESS=no
or remove calls to hwclock --systohc
in /etc/init.d/«scripts» in «stop» part removes the bug.

So I think it's not a bug of ACPI but a bug of the RTC standard in the kernel
(file arch/x86/kernel/rtc.c). I tried to add a «sleep 1» just after the
hwclock call but no change.

This is a workaround but not the real solution of this problem.
Thanks for reading.
François Boisson
(Please Cc to me, I'm not on linux-acpi list)

François Boisson
--
To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html