broken power button

broken power button

[Scott T. Smith]

APM will let me suspend to RAM and restore, but the touchpad is broken
the whole time.

So I'm trying ACPI -- ACPI seems to suspend successfully, but it will
not wake up.  I have to unplug the battery and power supply for 10+
minutes to reset it.

I think it's related to the power button.  I notice that the power
button does not generate any events via acpid (but if I close the lid,
or press Fn-ESC, then I do see events).  However, if I hold the power
button down for 4 seconds, the machine does shut off immediately. 
(Holding it down for 4+ seconds while frozen does nothing).

In tracing this further, I noticed that the ASL lists the button 'PWRB',
while ACPI (via dmesg) lists the button 'PWRF'.  I think (can someone
clarify this?) that this is the difference between a "fixed hardware
power button" and a "control method power button."

I tried renaming all references to PWRB to PWRF and recompiling, but
that did no good.  I think (again, can someone clarify this?) that I
need to set the PWR_BUTTON flag, but I don't know how to do that.  There
is a reference to SLP_BUTTON in the dsdt.asl file, but it's in a
comment.

Am I on the right track here?

oh, specs:  2.6.2-rc2-mm1 (ACPI v20040116), Dell Lattitude X200 (P3M
800), patched DSDT to cope with battery issue.

	Scott



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

Re: broken power button

[Greg Sarjeant]

Apologies for sending this twice, Scott. I forgot to copy the list.



Hi Scott,

    I have the same issue with my power button on my Gateway 200X (which
I think is roughly the same machine as the Dell X200). It appears to be
an issue with fixed-feature power buttons. I opened bug 1920 to address
the problem. You can check it out here:

    http://bugzilla.kernel.org/show_bug.cgi?id=1920

    Thanks,
    Greg


On Thu, 2004-02-05 at 21:55, Scott T. Smith wrote:
> APM will let me suspend to RAM and restore, but the touchpad is broken
> the whole time.
> 
> So I'm trying ACPI -- ACPI seems to suspend successfully, but it will
> not wake up.  I have to unplug the battery and power supply for 10+
> minutes to reset it.
> 
> I think it's related to the power button.  I notice that the power
> button does not generate any events via acpid (but if I close the lid,
> or press Fn-ESC, then I do see events).  However, if I hold the power
> button down for 4 seconds, the machine does shut off immediately. 
> (Holding it down for 4+ seconds while frozen does nothing).
> 
> In tracing this further, I noticed that the ASL lists the button 'PWRB',
> while ACPI (via dmesg) lists the button 'PWRF'.  I think (can someone
> clarify this?) that this is the difference between a "fixed hardware
> power button" and a "control method power button."
> 
> I tried renaming all references to PWRB to PWRF and recompiling, but
> that did no good.  I think (again, can someone clarify this?) that I
> need to set the PWR_BUTTON flag, but I don't know how to do that.  There
> is a reference to SLP_BUTTON in the dsdt.asl file, but it's in a
> comment.
> 
> Am I on the right track here?
> 
> oh, specs:  2.6.2-rc2-mm1 (ACPI v20040116), Dell Lattitude X200 (P3M
> 800), patched DSDT to cope with battery issue.
> 
> 	Scott
> 
> 
> 
> -------------------------------------------------------
> The SF.Net email is sponsored by EclipseCon 2004
> Premiere Conference on Open Tools Development and Integration
> See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
> http://www.eclipsecon.org/osdn
> _______________________________________________
> Acpi-devel mailing list
> Acpi-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org
> https://lists.sourceforge.net/lists/listinfo/acpi-devel
> 



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

Re: broken power button

[Scott T. Smith]

On Thu, 2004-02-05 at 19:33, Greg Sarjeant wrote:
>     I have the same issue with my power button on my Gateway 200X (which
> I think is roughly the same machine as the Dell X200). It appears to be

the older ones are identical AFAIK; the newer 200X's have 14" screens
though and seem totally different.

> an issue with fixed-feature power buttons. I opened bug 1920 to address
> the problem. You can check it out here:

I applied the patch; I think ACPI should report a warning in this case. 
According to p63, there are 2 bits of information (4 choices) but only 2
legal possibilities.  Windows must be ignoring the PWR_BUTTON field and
just checking for the existance of PWRB.  ACPI should complain if
PWR_BUTTON is not set and PWRB exists, or if PWR_BUTTON is set and PWRB
doesn't exist.

Similarly, if SLP_BUTTON is not set and SLPB exists, ACPI should throw a
warning too.

Unfortunately, while this lets me receive PWRB notifications, it doesn't
solve the lockup on suspend.  Ugh.

	Scott



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

Re: broken power button

[Greg Sarjeant]

On Fri, 2004-02-06 at 00:07, Scott T. Smith wrote:
> I applied the patch; I think ACPI should report a warning in this case. 
> According to p63, there are 2 bits of information (4 choices) but only 2
> legal possibilities.  Windows must be ignoring the PWR_BUTTON field and
> just checking for the existance of PWRB.  ACPI should complain if
> PWR_BUTTON is not set and PWRB exists, or if PWR_BUTTON is set and PWRB
> doesn't exist.
> 
> Similarly, if SLP_BUTTON is not set and SLPB exists, ACPI should throw a
> warning too.
> 
You're probably right.  I am hopeful that we can find a more robust
solution to the problem. This is an admittedly sloppy (and hopefully
temporary) workaround that just saves me the trouble of commenting out
those three whole lines of code in future kernels. :)

> Unfortunately, while this lets me receive PWRB notifications, it doesn't
> solve the lockup on suspend.  Ugh.
> 
Ah, drat. For what it's worth, I have similar issues with suspend,
though I haven't tried suspending to RAM for a while, which was when I
saw the behavior you describe. I have been able to suspend to disk from
a console, and others have reported doing so from unaccelerated X. I've
never managed to resume successfully from X, however.

Greg



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

RE: broken power button

[Yu, Luming]

> > I applied the patch; I think ACPI should report a warning 
> in this case. 

I suspect that patch is not necessary.

IAccording to the description of ACPI spec section 4.7.2.2.1.1
there should have SCI interrupt triggered by Fixed Power Button.

To verify whether Fixed Power Button works as expected, you 
can just hit power button , and check /proc/interrupts to see whether
ACPI interrupt number get increased. 

--Luming


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

RE: broken power button

[Scott T. Smith]

On Fri, 2004-02-06 at 09:17, Yu, Luming wrote: 
> > > I applied the patch; I think ACPI should report a warning 
> > in this case. 
> 
> I suspect that patch is not necessary.
> 
> IAccording to the description of ACPI spec section 4.7.2.2.1.1
> there should have SCI interrupt triggered by Fixed Power Button.
> 
> To verify whether Fixed Power Button works as expected, you 
> can just hit power button , and check /proc/interrupts to see whether
> ACPI interrupt number get increased. 

My ACPI interrupt number climbs by 8 about once every two seconds (step
function).

if I press the power button with Greg's patch (i.e. force PWRB not
PWRF), it jumps by 2 when I press the button (then another 8 less than
one second later).  I also get this in /var/log/acpid:

[Fri Feb  6 09:26:22 2004] received event "button/power PWRB 00000080 00000006"
[Fri Feb  6 09:26:22 2004] executing action "/etc/acpi/default.sh button/power PWRB 00000080 00000006"
[Fri Feb  6 09:26:22 2004] BEGIN HANDLER MESSAGES
[Fri Feb  6 09:26:22 2004] END HANDLER MESSAGES
[Fri Feb  6 09:26:22 2004] action exited with status 0
[Fri Feb  6 09:26:22 2004] completed event "button/power PWRB 00000080 00000006"

without the patch, the power button also causes the # of interrupts to
jump by 2, but I get nothing in /var/log/acpid.

	Scott



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

RE: broken power button

[greg]

---------- Original Message -----------
From: "Scott T. Smith" <scott-j3vAvQ9dNB9ByuSxxbvQtw@public.gmane.org>
To: "Yu, Luming" <luming.yu-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org>
Sent: Fri, 06 Feb 2004 09:38:48 -0800
Subject: RE: [ACPI] broken power button

> On Fri, 2004-02-06 at 09:17, Yu, Luming wrote: 
> > > > I applied the patch; I think ACPI should report a warning 
> > > in this case. 
> > 
> > I suspect that patch is not necessary.
> > 
> > IAccording to the description of ACPI spec section 4.7.2.2.1.1
> > there should have SCI interrupt triggered by Fixed Power Button.
> > 
> > To verify whether Fixed Power Button works as expected, you 
> > can just hit power button , and check /proc/interrupts to see whether
> > ACPI interrupt number get increased. 
> 
> My ACPI interrupt number climbs by 8 about once every two seconds (step
> function).
> 
> if I press the power button with Greg's patch (i.e. force PWRB not
> PWRF), it jumps by 2 when I press the button (then another 8 less 
> than one second later).  I also get this in /var/log/acpid:
> 
> [Fri Feb  6 09:26:22 2004] received event "button/power PWRB 
> 00000080 00000006"
> [Fri Feb  6 09:26:22 2004] executing action "/etc/acpi/default.sh 
> button/power PWRB 00000080 00000006"
> [Fri Feb  6 09:26:22 2004] BEGIN HANDLER MESSAGES
> [Fri Feb  6 09:26:22 2004] END HANDLER MESSAGES
> [Fri Feb  6 09:26:22 2004] action exited with status 0
> [Fri Feb  6 09:26:22 2004] completed event "button/power PWRB 
> 00000080 00000006"
> 
> without the patch, the power button also causes the # of interrupts 
> to jump by 2, but I get nothing in /var/log/acpid.
> 
> 	Scott
------- End of Original Message -------


I can confirm this behavior on my machine as well.


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

RE: broken power button

[Scott T. Smith]

On Fri, 2004-02-06 at 09:17, Yu, Luming wrote:
> > > I applied the patch; I think ACPI should report a warning 
> > in this case. 
> 
> I suspect that patch is not necessary.
> 
> IAccording to the description of ACPI spec section 4.7.2.2.1.1
> there should have SCI interrupt triggered by Fixed Power Button.

perhaps, but according the the ACPI spec (4.7.2.2.1, table 4-8, p63 of
ACPI spec 2.0b), the situation should not exist (i.e. you should not
have a definition of PWRB in the AML code, AND assume a Fixed Power
Button).

	Scott



-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn

RE: broken power button

[Yu, Luming]

> 
> I can confirm this behavior on my machine as well.

You need to check:
1. Whether a fixed event can be detected  . (Please refer
acpi_ev_fixed_event_detect in drivers/acpi/events/evevent.c)
2. try to comment  out below codes in acpi_button_notify
(drivers/acpi/button.c)
	if (!button || !button->device)
		return_VOID.

--Luming


-------------------------------------------------------
The SF.Net email is sponsored by EclipseCon 2004
Premiere Conference on Open Tools Development and Integration
See the breadth of Eclipse activity. February 3-5 in Anaheim, CA.
http://www.eclipsecon.org/osdn