Re: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression - Linux-3.x

Re: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression - Linux-3.x

[Bjorn Helgaas]

[+cc linux-acpi and Lin Ming]

On Wed, Dec 7, 2011 at 3:38 PM,  <wallak@free.fr> wrote:
> We have a regression on the ACPI stack of the last linux kernel line 3.x (3.1.4,
> 3.2-rc4...). The ACPI "_PDC" chunk is not executed on some computers (e.g. Dell
> X300; the function acpi_processor_set_pdc() is not called). This issue yield to
> an uninitialized state of some ACPI variables.
>
> A patch is available below. This patch come back to the previous linux behavior,
> and works fine.
>
> Best Regards,
> Wallak.
>
> --- linux-3.1.4-mdf/drivers/acpi/processor_core.c.orig  2011-12-07
> 23:12:57.000000000 +0100
> +++ linux-3.1.4-mdf/drivers/acpi/processor_core.c       2011-12-07
> 23:13:39.000000000 +0100
> @@ -223,8 +223,8 @@
>        type = (acpi_type == ACPI_TYPE_DEVICE) ? 1 : 0;
>        cpuid = acpi_get_cpuid(handle, type, acpi_id);
>
> -       if (cpuid == -1)
> -               return false;
> +       if ((cpuid == -1) && (num_possible_cpus() > 1))
> +               return false;
>
>        return true;
>  }
> --
> 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/
--
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: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression - Linux-3.x

[Lin Ming]

On Thu, 2011-12-08 at 13:01 +0800, Bjorn Helgaas wrote:
> [+cc linux-acpi and Lin Ming]
> 
> On Wed, Dec 7, 2011 at 3:38 PM,  <wallak@free.fr> wrote:
> > We have a regression on the ACPI stack of the last linux kernel line 3.x (3.1.4,
> > 3.2-rc4...). The ACPI "_PDC" chunk is not executed on some computers (e.g. Dell
> > X300; the function acpi_processor_set_pdc() is not called). This issue yield to
> > an uninitialized state of some ACPI variables.
> >
> > A patch is available below. This patch come back to the previous linux behavior,
> > and works fine.

Hi, Wallak

Please attach the acpidump output and kernel .config file.
And also the output of "cat /proc/cpuinfo"

(You can send these files to me offline in case they are too big for
mail list)

Thanks,
Lin Ming

> >
> > Best Regards,
> > Wallak.
> >
> > --- linux-3.1.4-mdf/drivers/acpi/processor_core.c.orig  2011-12-07
> > 23:12:57.000000000 +0100
> > +++ linux-3.1.4-mdf/drivers/acpi/processor_core.c       2011-12-07
> > 23:13:39.000000000 +0100
> > @@ -223,8 +223,8 @@
> >        type = (acpi_type == ACPI_TYPE_DEVICE) ? 1 : 0;
> >        cpuid = acpi_get_cpuid(handle, type, acpi_id);
> >
> > -       if (cpuid == -1)
> > -               return false;
> > +       if ((cpuid == -1) && (num_possible_cpus() > 1))
> > +               return false;
> >
> >        return true;
> >  }
> > --
> > 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/

Re: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression - Linux-3.x

[Lin Ming]

On Mon, 2011-12-12 at 10:26 +0800, Lin Ming wrote:
> ---------- Forwarded message ----------
> From:  <wallak@free.fr>
> Date: Thu, Dec 8, 2011 at 6:38 AM
> Subject: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression
> - Linux-3.x
> To: linux-kernel@vger.kernel.org
> 
> 
> We have a regression on the ACPI stack of the last linux kernel line 3.x (3.1.4,
> 3.2-rc4...). The ACPI "_PDC" chunk is not executed on some computers (e.g. Dell
> X300; the function acpi_processor_set_pdc() is not called). This issue yield to
> an uninitialized state of some ACPI variables.
> 
> A patch is available below. This patch come back to the previous linux behavior,
> and works fine.
> 
> Best Regards,
> Wallak.
> 
> --- linux-3.1.4-mdf/drivers/acpi/processor_core.c.orig  2011-12-07
> 23:12:57.000000000 +0100
> +++ linux-3.1.4-mdf/drivers/acpi/processor_core.c       2011-12-07
> 23:13:39.000000000 +0100
> @@ -223,8 +223,8 @@
>        type = (acpi_type == ACPI_TYPE_DEVICE) ? 1 : 0;
>        cpuid = acpi_get_cpuid(handle, type, acpi_id);
> 
> -       if (cpuid == -1)
> -               return false;
> +       if ((cpuid == -1) && (num_possible_cpus() > 1))

Hi Wallak,

BIOS may define multiple CPU handles even for UP
processor(see below).

processor_physically_present(acpi_handle handle) will be called for each
CPU handles.

We should only return valid value for CPU0 on UP processor.
With your patch, processor_physically_present will return true for all
CPU handles(CPU0, CPU1, CPU2, CPU3). This is not we want.

I think below is the correct fix.
Could you help to test it?

Thanks.

>From 5c6de7311ced7a1febf85fdcc08b6116bcfe8138 Mon Sep 17 00:00:00 2001
From: Lin Ming <ming.m.lin@intel.com>
Date: Mon, 12 Dec 2011 10:04:53 +0800
Subject: [PATCH] ACPI: processor: fix acpi_get_cpuid for UP processor

For UP processor, it is likely that no _MAT method or MADT table defined.
So currently acpi_get_cpuid(...) always return -1 for UP processor.
This is wrong. It should return valid value for CPU0.

In the other hand, BIOS may define multiple CPU handles even for UP
processor, for example

        Scope (_PR)
        {
            Processor (CPU0, 0x00, 0x00000410, 0x06) {}
            Processor (CPU1, 0x01, 0x00000410, 0x06) {}
            Processor (CPU2, 0x02, 0x00000410, 0x06) {}
            Processor (CPU3, 0x03, 0x00000410, 0x06) {}
        }

We should only return valid value for CPU0's acpi handle.
And return invalid value for others.

http://marc.info/?t=132329819900003&r=1&w=2

Reported-by: wallak@free.fr
Signed-off-by: Lin Ming <ming.m.lin@intel.com>
---
 drivers/acpi/processor_core.c |   26 ++++++++++++++++++++++++--
 1 files changed, 24 insertions(+), 2 deletions(-)

diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c
index 3a0428e..3372900 100644
--- a/drivers/acpi/processor_core.c
+++ b/drivers/acpi/processor_core.c
@@ -173,8 +173,30 @@ int acpi_get_cpuid(acpi_handle handle, int type, u32 acpi_id)
 	apic_id = map_mat_entry(handle, type, acpi_id);
 	if (apic_id == -1)
 		apic_id = map_madt_entry(type, acpi_id);
-	if (apic_id == -1)
-		return apic_id;
+	if (apic_id == -1) {
+		/*
+		 * On UP processor, there is no _MAT or MADT table.
+		 * So above apic_id is always set to -1.
+		 *
+		 * BIOS may define multiple CPU handles even for UP processor.
+		 * For example,
+		 *
+		 * Scope (_PR)
+                 * {
+		 *     Processor (CPU0, 0x00, 0x00000410, 0x06) {}
+		 *     Processor (CPU1, 0x01, 0x00000410, 0x06) {}
+		 *     Processor (CPU2, 0x02, 0x00000410, 0x06) {}
+		 *     Processor (CPU3, 0x03, 0x00000410, 0x06) {}
+		 * }
+		 *
+		 * Ignores apic_id and always return 0 for CPU0's handle.
+		 * Return -1 for other CPU's handle.
+		 */ 
+		if (acpi_id == 0)
+			return acpi_id;
+		else
+			return apic_id;
+	}
 
 #ifdef CONFIG_SMP
 	for_each_possible_cpu(i) {
-- 
1.7.2.5



> +               return false;
> 
>        return true;
>  }

Re: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression - Linux-3.x

[Wallak]

Hi Lin Ming,

This last patch, modifying the behavior of acpi_get_cpuid(), works fine 
on my non SMP enabled kernel. The test was done on a Dell X300, 
acpi_processor_set_pdc() was properly called, without the changes done 
by my previous patch.

Best Regards,
Wallak.

Lin Ming wrote:
> On Mon, 2011-12-12 at 10:26 +0800, Lin Ming wrote:
>> ---------- Forwarded message ----------
>> From:<wallak@free.fr>
>> Date: Thu, Dec 8, 2011 at 6:38 AM
>> Subject: ACPI "_PDC" - acpi_processor_set_pdc()- execution regression
>> - Linux-3.x
>> To: linux-kernel@vger.kernel.org
>>
>>
>> We have a regression on the ACPI stack of the last linux kernel line 3.x (3.1.4,
>> 3.2-rc4...). The ACPI "_PDC" chunk is not executed on some computers (e.g. Dell
>> X300; the function acpi_processor_set_pdc() is not called). This issue yield to
>> an uninitialized state of some ACPI variables.
>>
>> A patch is available below. This patch come back to the previous linux behavior,
>> and works fine.
>>
>> Best Regards,
>> Wallak.
>>
>> --- linux-3.1.4-mdf/drivers/acpi/processor_core.c.orig  2011-12-07
>> 23:12:57.000000000 +0100
>> +++ linux-3.1.4-mdf/drivers/acpi/processor_core.c       2011-12-07
>> 23:13:39.000000000 +0100
>> @@ -223,8 +223,8 @@
>>         type = (acpi_type == ACPI_TYPE_DEVICE) ? 1 : 0;
>>         cpuid = acpi_get_cpuid(handle, type, acpi_id);
>>
>> -       if (cpuid == -1)
>> -               return false;
>> +       if ((cpuid == -1)&&  (num_possible_cpus()>  1))
> Hi Wallak,
>
> BIOS may define multiple CPU handles even for UP
> processor(see below).
>
> processor_physically_present(acpi_handle handle) will be called for each
> CPU handles.
>
> We should only return valid value for CPU0 on UP processor.
> With your patch, processor_physically_present will return true for all
> CPU handles(CPU0, CPU1, CPU2, CPU3). This is not we want.
>
> I think below is the correct fix.
> Could you help to test it?
>
> Thanks.
>
> > From 5c6de7311ced7a1febf85fdcc08b6116bcfe8138 Mon Sep 17 00:00:00 2001
> From: Lin Ming<ming.m.lin@intel.com>
> Date: Mon, 12 Dec 2011 10:04:53 +0800
> Subject: [PATCH] ACPI: processor: fix acpi_get_cpuid for UP processor
>
> For UP processor, it is likely that no _MAT method or MADT table defined.
> So currently acpi_get_cpuid(...) always return -1 for UP processor.
> This is wrong. It should return valid value for CPU0.
>
> In the other hand, BIOS may define multiple CPU handles even for UP
> processor, for example
>
>          Scope (_PR)
>          {
>              Processor (CPU0, 0x00, 0x00000410, 0x06) {}
>              Processor (CPU1, 0x01, 0x00000410, 0x06) {}
>              Processor (CPU2, 0x02, 0x00000410, 0x06) {}
>              Processor (CPU3, 0x03, 0x00000410, 0x06) {}
>          }
>
> We should only return valid value for CPU0's acpi handle.
> And return invalid value for others.
>
> http://marc.info/?t=132329819900003&r=1&w=2
>
> Reported-by: wallak@free.fr
> Signed-off-by: Lin Ming<ming.m.lin@intel.com>
> ---
>   drivers/acpi/processor_core.c |   26 ++++++++++++++++++++++++--
>   1 files changed, 24 insertions(+), 2 deletions(-)
>
> diff --git a/drivers/acpi/processor_core.c b/drivers/acpi/processor_core.c
> index 3a0428e..3372900 100644
> --- a/drivers/acpi/processor_core.c
> +++ b/drivers/acpi/processor_core.c
> @@ -173,8 +173,30 @@ int acpi_get_cpuid(acpi_handle handle, int type, u32 acpi_id)
>   	apic_id = map_mat_entry(handle, type, acpi_id);
>   	if (apic_id == -1)
>   		apic_id = map_madt_entry(type, acpi_id);
> -	if (apic_id == -1)
> -		return apic_id;
> +	if (apic_id == -1) {
> +		/*
> +		 * On UP processor, there is no _MAT or MADT table.
> +		 * So above apic_id is always set to -1.
> +		 *
> +		 * BIOS may define multiple CPU handles even for UP processor.
> +		 * For example,
> +		 *
> +		 * Scope (_PR)
> +                 * {
> +		 *     Processor (CPU0, 0x00, 0x00000410, 0x06) {}
> +		 *     Processor (CPU1, 0x01, 0x00000410, 0x06) {}
> +		 *     Processor (CPU2, 0x02, 0x00000410, 0x06) {}
> +		 *     Processor (CPU3, 0x03, 0x00000410, 0x06) {}
> +		 * }
> +		 *
> +		 * Ignores apic_id and always return 0 for CPU0's handle.
> +		 * Return -1 for other CPU's handle.
> +		 */
> +		if (acpi_id == 0)
> +			return acpi_id;
> +		else
> +			return apic_id;
> +	}
>
>   #ifdef CONFIG_SMP
>   	for_each_possible_cpu(i) {