ACPI Poweroff fails on HP OmniBook 500

ACPI Poweroff fails on HP OmniBook 500

[Adam Kessel]

I posted this recently to lkml, but got no response, so I thought I would
report it here as well.[1]

In 2.4.22-rc2, ACPI poweroff worked fine on the HP OmniBook.

Since 2.4.22, all the way up to 2.4.23-pre8, the system fails to
shutdown--the process gets to "power off" and hangs there. It is
necessary to do a hard shut down to turn the system off.

There is no problem with rebooting.

Apparently this effects a number of HP laptops, although I have only
verified it on the OmniBook 500.[2]

This problem is not present in 2.6.0-test9.

I am using force=acpi as required since I have an older BIOS.  Apparently
the problem occurs in OmniBook laptops they have newer BIOSes as well.

I expect this is related to the Sony VAIO problem discussed this month on
this list as well.[3]

[1] http://www.ussg.iu.edu/hypermail/linux/kernel/0310.1/1625.html 
[2] http://zurich.ai.mit.edu/pipermail/omnibook/2003-September/002805.html
[3] http://article.gmane.org/gmane.linux.acpi.devel/4770


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?   SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI Poweroff fails on HP OmniBook 500

[Ducrot Bruno]

On Tue, Oct 28, 2003 at 09:06:36PM -0500, Adam Kessel wrote:
> I posted this recently to lkml, but got no response, so I thought I would
> report it here as well.[1]
> 
> In 2.4.22-rc2, ACPI poweroff worked fine on the HP OmniBook.
> 
> Since 2.4.22, all the way up to 2.4.23-pre8, the system fails to
> shutdown--the process gets to "power off" and hangs there. It is
> necessary to do a hard shut down to turn the system off.
> 
> There is no problem with rebooting.
> 
> Apparently this effects a number of HP laptops, although I have only
> verified it on the OmniBook 500.[2]
> 
> This problem is not present in 2.6.0-test9.
> 
> I am using force=acpi as required since I have an older BIOS.  Apparently

You mean acpi=force, correct?

> the problem occurs in OmniBook laptops they have newer BIOSes as well.
> 
> I expect this is related to the Sony VAIO problem discussed this month on
> this list as well.[3]

Could you please test this patch?
The problem being that we power down some devices twice.

--- linux-2.4.23-pre9/drivers/acpi/system.c	2003/10/31 08:09:10	1.1
+++ linux-2.4.23-pre9/drivers/acpi/system.c	2003/10/31 10:36:35
@@ -31,6 +31,7 @@
 #include <linux/spinlock.h>
 #include <linux/poll.h>
 #include <linux/delay.h>
+#include <linux/interrupt.h>
 #include <linux/sysrq.h>
 #include <linux/compatmac.h>
 #include <linux/proc_fs.h>
@@ -92,7 +93,13 @@
 static void
 acpi_power_off (void)
 {
-	acpi_suspend(ACPI_STATE_S5);
+	if (unlikely(in_interrupt())) 
+		BUG();
+	acpi_enter_sleep_state_prep(ACPI_STATE_S5);
+	ACPI_DISABLE_IRQS();
+	acpi_enter_sleep_state(ACPI_STATE_S5);
+
+	printk(KERN_EMERG "ACPI: can not power off machine\n");
 }
 
 #endif /*CONFIG_PM*/

-- 
Ducrot Bruno

--  Which is worse:  ignorance or apathy?
--  Don't know.  Don't care.


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?   SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/

Re: ACPI Poweroff fails on HP OmniBook 500

[Adam Kessel]

On Fri, Oct 31, 2003 at 12:09:05PM +0100, Ducrot Bruno wrote:
> > I am using force=acpi as required since I have an older BIOS.  Apparently
> You mean acpi=force, correct?

Yes, typo.

> Could you please test this patch?
> The problem being that we power down some devices twice.

Yes, this patch fixes the problem 100%.  Thanks.

> --- linux-2.4.23-pre9/drivers/acpi/system.c	2003/10/31 08:09:10	1.1
> +++ linux-2.4.23-pre9/drivers/acpi/system.c	2003/10/31 10:36:35
> @@ -31,6 +31,7 @@
>  #include <linux/spinlock.h>
>  #include <linux/poll.h>
>  #include <linux/delay.h>
> +#include <linux/interrupt.h>
>  #include <linux/sysrq.h>
>  #include <linux/compatmac.h>
>  #include <linux/proc_fs.h>
> @@ -92,7 +93,13 @@
>  static void
>  acpi_power_off (void)
>  {
> -	acpi_suspend(ACPI_STATE_S5);
> +	if (unlikely(in_interrupt())) 
> +		BUG();
> +	acpi_enter_sleep_state_prep(ACPI_STATE_S5);
> +	ACPI_DISABLE_IRQS();
> +	acpi_enter_sleep_state(ACPI_STATE_S5);
> +
> +	printk(KERN_EMERG "ACPI: can not power off machine\n");
>  }
>  
>  #endif /*CONFIG_PM*/
-- 
Adam Kessel
http://bostoncoop.net/adam


-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive?  Does it
help you create better code?   SHARE THE LOVE, and help us help
YOU!  Click Here: http://sourceforge.net/donate/