RE: [patch 5/9] ACPI should depend on, not select PCI

RE: [patch 5/9] ACPI should depend on, not select PCI

[Brown, Len]

No, I can't apply this -- it allows
Kconfig to create IA64 configs without PCI,
which do not build.

there must be a better way.

-Len 

>-----Original Message-----
>From: akpm@osdl.org [mailto:akpm@osdl.org] 
>Sent: Friday, October 28, 2005 6:00 PM
>To: Brown, Len
>Cc: acpi-devel@lists.sourceforge.net; akpm@osdl.org; bunk@stusta.de
>Subject: [patch 5/9] ACPI should depend on, not select PCI
>
>
>From: Adrian Bunk <bunk@stusta.de>
>
>ACPI should depend on, not select PCI.
>
>The practical differences should be nearly zero except that it avoids 
>the illegal configuration PCI=y, X86_VOYAGER=y.
>
>Signed-off-by: Adrian Bunk <bunk@stusta.de>
>Signed-off-by: Andrew Morton <akpm@osdl.org>
>---
>
> drivers/acpi/Kconfig |    2 +-
> 1 files changed, 1 insertion(+), 1 deletion(-)
>
>diff -puN 
>drivers/acpi/Kconfig~acpi-should-depend-on-not-select-pci 
>drivers/acpi/Kconfig
>--- 
>25/drivers/acpi/Kconfig~acpi-should-depend-on-not-select-pci	
>Fri Oct 28 14:59:08 2005
>+++ 25-akpm/drivers/acpi/Kconfig	Fri Oct 28 14:59:21 2005
>@@ -10,8 +10,8 @@ menu "ACPI (Advanced Configuration and P
> config ACPI
> 	bool "ACPI Support"
> 	depends on IA64 || X86
>+	depends on PCI
> 	select PM
>-	select PCI
> 
> 	default y
> 	---help---
>_
>

RE: [patch 5/9] ACPI should depend on, not select PCI

[Roman Zippel]

Hi

On Wed, 30 Nov 2005, Brown, Len wrote:

> No, I can't apply this -- it allows
> Kconfig to create IA64 configs without PCI,
> which do not build.

Why? Simply enable the PCI in IA64 like everyone else, what's the problem?

> > config ACPI
> > 	bool "ACPI Support"
> > 	depends on IA64 || X86
> >+	depends on PCI
> > 	select PM
> >-	select PCI
> > 
> > 	default y

This default should also go.

bye, Roman

Re: [patch 5/9] ACPI should depend on, not select PCI

[Adrian Bunk]

On Wed, Nov 30, 2005 at 08:58:22PM -0500, Brown, Len wrote:

> No, I can't apply this -- it allows
> Kconfig to create IA64 configs without PCI,
> which do not build.
> 
> there must be a better way.

According to arch/ia64/Kconfig, PCI support is optional on ia64.

Either arch/ia64/Kconfig should be fixed or the PCI=n builds must be 
fixed on ia64.

Looking closer, is the problem you are talking about caused by the fact 
that IA64_GENERIC select's options without ensuring that the 
dependencies of what it is select'ing are fulfilled?

> -Len 

cu
Adrian

-- 

       "Is there not promise of rain?" Ling Tan asked suddenly out
        of the darkness. There had been need of rain for many days.
       "Only a promise," Lao Er said.
                                       Pearl S. Buck - Dragon Seed