mysterious power off problem 2.4.10-2.4.14 on laptop

mysterious power off problem 2.4.10-2.4.14 on laptop

[L A Walsh]

I haven't had time to track this down since I found it.  I'm throwing
it out in case anyone has seen anything that might figure in to this.

Machine: Dell Inspiron 8000.  
Problem: using same options, (starting with 249.config, and using make oldconfig
for new kernel version), I get a kernel that has a bad habit of turning
itself off after some period of time -- not shutting down, just turning off.
This started in 2.4.10 and, and not having the time to debug it, I just
went back to using 2.4.9 which didn't exhibit the problem.

Things tried that haven't worked (not necessarily in any particular order)
1) 2.4.14 kernel
2) disabling power management from the KDE desktop.
3) renaming the apm binary would allow apm control.
4) Turning off apm support in the kernel.
5) Turning off BIOS apm management completely.
	a) BIOS PM on AC-power was set to
		i) turn off inactive video in 10 minutes
		ii) turn off inactive hard disk in 30 minutes
		iii) Suspend: disabled.
		iv) S2D Timeout: disabled

	b) BIOS PM on battery still _is_ set to
		i) turn off inactive video in 3 minutes
		ii) turn off in hard disks in 5 minutes
		iii) Suspend: disabled
		iv) S2D Timeout: 1 hour

	common options:
		v) Smart CPU: enabled
		vi) Display lid closed: <don't shut down, remain active>
		vii) all auto 'wake'/'resume' (on lan, alarm, ring, at time)
			disabled
		viii) CPU Mode: Battery Optimized
----
Other details:
1) I noticed when the machine was on battery -- it turned off in about
   3 minutes -- it might have been 5, but seemed closer to 3.  I note this
   coincides with the Video timeout.  Previously when the timeout
   would occur on AC power -- it seemed that the inactivity timeout was
   about 10 minutes.  My estimations may be wrong, but part of the problem 
   could be related to the video timeout.  
2) testing this problem is a pain, since my disks are still the 
   primitive 'ext2' file system and the multi-gig, laptop-speed disks
   are slow to check.

So am just wondering if someone has seen this, or its a known 'feature change'
with the workaround or solution being 'X'.

I note in the changelog that ACPI changes went into 2.4.10.  I am still using
APM, not ACPI, but it it possible there is common code to both that got
changed?

Thanks for any help.  

-linda

Re: mysterious power off problem 2.4.10-2.4.14 on laptop

[L A Walsh]

"mike@morpheus" wrote:
> Give ACPI a try, for a while I've noticed APM getting mixed up
> on my home box (its a VIA chipset, I've been told that probably
> why :), doing things like not powering off and changing the instant-off
> powerbutton to a wait-5-seconds powerbutton.
> 
> I switched to ACPI and everythings been working fine :)
---
	Thanks for the suggestion.  However APM was working superbly for
my architecture in 2.4.9 and before.  It turned off the display when I
wasn't using it.  It occasionally spun down disks w/the help of noflushd,
and when suspend was indicated, it correctly suspended to RAM where it
can stay for hours using <1%/hour.  

	I did try ACPI at some point, but it didn't work as well in
providing the same features and reliability when I tested it.  The
config option says "Experimental", ACPI isn't a feature complete as
APM and ACPI was in development.  It might work differently on different
hardware, for example.  While it is to be the replacement for APM, I
don't know if I am comfortable moving to it yet -- and even so, why should
APM mysteriously break when it has been working great since the early 
2.4 series and fairly well since 2.2 (X was a problem on my hardware at one
point).

	I'd prefer not to try an unknown, where if I have a problem, I
don't know if it is my hardware, a misconfiguration on my part, or the
Experimental Hardware.  That would likely take more time than simply
staying with APM -- a known 'working configuration', and finding what
changed in 2.4.10 (and remains in 2.4.14) that lead to the new problems.

	If it worked before, then something changed in the kernel to
break it, I'd generally classify that as a bug.  Now maybe there
are new utilities needed -- however, in SuSE 7.3, they use the 2.4.10
kernel and their default setup has the same problem.

:-(
-linda

Re: mysterious power off problem 2.4.10-2.4.14 on laptop

[mike@morpheus]

On Mon, 12 Nov 2001, L A Walsh wrote:

> I haven't had time to track this down since I found it.  I'm throwing
> it out in case anyone has seen anything that might figure in to this.
>
> Machine: Dell Inspiron 8000.
> Problem: using same options, (starting with 249.config, and using make oldconfig
> for new kernel version), I get a kernel that has a bad habit of turning
> itself off after some period of time -- not shutting down, just turning off.
> This started in 2.4.10 and, and not having the time to debug it, I just
> went back to using 2.4.9 which didn't exhibit the problem.
>
> Things tried that haven't worked (not necessarily in any particular order)
> 1) 2.4.14 kernel
> 2) disabling power management from the KDE desktop.
> 3) renaming the apm binary would allow apm control.
> 4) Turning off apm support in the kernel.
> 5) Turning off BIOS apm management completely.
> 	a) BIOS PM on AC-power was set to
> 		i) turn off inactive video in 10 minutes
> 		ii) turn off inactive hard disk in 30 minutes
> 		iii) Suspend: disabled.
> 		iv) S2D Timeout: disabled
>
> 	b) BIOS PM on battery still _is_ set to
> 		i) turn off inactive video in 3 minutes
> 		ii) turn off in hard disks in 5 minutes
> 		iii) Suspend: disabled
> 		iv) S2D Timeout: 1 hour
>
> 	common options:
> 		v) Smart CPU: enabled
> 		vi) Display lid closed: <don't shut down, remain active>
> 		vii) all auto 'wake'/'resume' (on lan, alarm, ring, at time)
> 			disabled
> 		viii) CPU Mode: Battery Optimized
> ----
> Other details:
> 1) I noticed when the machine was on battery -- it turned off in about
>    3 minutes -- it might have been 5, but seemed closer to 3.  I note this
>    coincides with the Video timeout.  Previously when the timeout
>    would occur on AC power -- it seemed that the inactivity timeout was
>    about 10 minutes.  My estimations may be wrong, but part of the problem
>    could be related to the video timeout.
> 2) testing this problem is a pain, since my disks are still the
>    primitive 'ext2' file system and the multi-gig, laptop-speed disks
>    are slow to check.
>
> So am just wondering if someone has seen this, or its a known 'feature change'
> with the workaround or solution being 'X'.
>
> I note in the changelog that ACPI changes went into 2.4.10.  I am still using
> APM, not ACPI, but it it possible there is common code to both that got
> changed?
>
> Thanks for any help.
>
> -linda
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at  http://www.tux.org/lkml/
>

Give ACPI a try, for a while I've noticed APM getting mixed up
on my home box (its a VIA chipset, I've been told that probably
why :), doing things like not powering off and changing the instant-off
powerbutton to a wait-5-seconds powerbutton.

I switched to ACPI and everythings been working fine :)

mike

Re: mysterious power off problem 2.4.10-2.4.14 on laptop

[Stuart Young]

At 01:06 PM 12/11/01 -0800, L A Walsh wrote:
>don't know if I am comfortable moving to it yet -- and even so, why should
>APM mysteriously break when it has been working great since the early
>2.4 series and fairly well since 2.2 (X was a problem on my hardware at one
>point).
>
>         I'd prefer not to try an unknown, where if I have a problem, I
>don't know if it is my hardware, a misconfiguration on my part, or the
>Experimental Hardware.  That would likely take more time than simply
>staying with APM -- a known 'working configuration', and finding what
>changed in 2.4.10 (and remains in 2.4.14) that lead to the new problems.

There was mention recently about some changes to APM to fix some "broken" 
behavior regarding the way it set/reset values. Some of these were cases 
where APM seemed to blindly set values instead of leaving them as is. I 
don't remember just how long ago, but you might try looking over the archives.

If this is the case, then it's possible that the APM code was checking 
status, and then blindly setting things to suit that status instead of what 
was read. Makes sense, but potentially broken. If the BIOS is broken and 
doesn't set these values correctly, then this could produce the symptoms 
you describe. Check for a BIOS upgrade anyway - usually a good manufacturer 
will have some sort of changelog with the BIOS, so if there is a new 
version, you might find APM was broken and fixed.

It's also possible that this never made it into those kernels, and they've 
been using the broken behavior since 2.4.10 which is causing your problems. 
I'd also suggest checking out the latest of the 2.4.14pre* kernels. My 
memory of the list (I may be wrong) makes me think the discussion was quite 
recent.

Good luck.

AMC Enterprises P/L    - Stuart Young
First Floor            - Network and Systems Admin
3 Chesterville Rd      - sgy@amc.com.au
Cheltenham Vic 3192    - Ph:  (03) 9584-2700
http://www.amc.com.au/ - Fax: (03) 9584-2755

Re: mysterious power off problem 2.4.10-2.4.14 on laptop

[Pavel Machek]

Hi!

> > Give ACPI a try, for a while I've noticed APM getting mixed up
> > on my home box (its a VIA chipset, I've been told that probably
> > why :), doing things like not powering off and changing the instant-off
> > powerbutton to a wait-5-seconds powerbutton.
> > 
> > I switched to ACPI and everythings been working fine :)
> ---
> 	Thanks for the suggestion.  However APM was working superbly for
> my architecture in 2.4.9 and before.  It turned off the display when I

You don't want to switch to acpi.... Di binary search of which patch broke 
it....
-- 
Philips Velo 1: 1"x4"x8", 300gram, 60, 12MB, 40bogomips, linux, mutt,
details at http://atrey.karlin.mff.cuni.cz/~pavel/velo/index.html.

Re: mysterious power off problem 2.4.10-2.4.14 on laptop

[Thomas Winischhofer]

Linda,

it could be a bad BIOS.

I don't know if this helps, but anyway:

I had a similar problem, but in ma case the machine just went to suspend
mode - annoying, but somewhat easier to track. Just like in your case,
nothing worked; switching off APM in BIOS, disabling APM in kernel, etc.
no matter what I did (typing, using the mouse), the machine suspended.

Well:

Question 1: Are you running Windows on that machine, too? If no, you
have a different problem; as said, most probably caused by a bad BIOS.
That it worked under a previous 2.4.x kernel doesn't mean a lot; apm
could have interpreted bad (malformed) apm messages in a different way
(just a guess).

Question 2: If yes: Did you ever happen to try to boot (ie cold boot =
switch ON) the machine with AC power chord connected after a Windows
session?

In my case, it was a bug in the BIOS. This piece of sh** - right after a
Windows session - reported a battery charging status of _more_ than
100%, leading into miscomprehending APM messages from the BIOS - leading
into APM confusion. Strange enough, this _only_ happened when _switching
on_ the machine on battery. It never happens when booting with AC
connected.

Even more strange, in my case it is sufficient to cold boot the machine
_once_ after running Windows on AC to avoid the effect in the future
(however, until running windows again).

Good luck,

Thomas

-- 
Thomas Winischhofer
Vienna/Austria                  Check it out:         
mailto:tw@webit.com              *** http://www.webit.com/tw