[PATCH] apple-gmux: Fix kconfig dependencies

[PATCH] apple-gmux: Fix kconfig dependencies

[Seth Forshee]

Fix the dependencies of apple-gmux to prevent it from being built-in
when one or more of its dependencies is built as a module. Otherwise it
can fail to build due to missing symbols.

Reported-by: Arun Raghavan <arun.raghavan@collabora.co.uk>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
---
 drivers/platform/x86/Kconfig |    4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/Kconfig b/drivers/platform/x86/Kconfig
index 2a262f5..cf65f0a 100644
--- a/drivers/platform/x86/Kconfig
+++ b/drivers/platform/x86/Kconfig
@@ -759,7 +759,9 @@ config SAMSUNG_Q10
 config APPLE_GMUX
 	tristate "Apple Gmux Driver"
 	depends on PNP
-	select BACKLIGHT_CLASS_DEVICE
+	depends on BACKLIGHT_CLASS_DEVICE
+	depends on BACKLIGHT_APPLE=n || BACKLIGHT_APPLE
+	depends on ACPI_VIDEO=n || ACPI_VIDEO
 	---help---
 	  This driver provides support for the gmux device found on many
 	  Apple laptops, which controls the display mux for the hybrid
-- 
1.7.9.5

Re: [PATCH] apple-gmux: Fix kconfig dependencies

[Seth Forshee]

On Mon, Jul 02, 2012 at 09:55:29AM -0500, Seth Forshee wrote:
> Fix the dependencies of apple-gmux to prevent it from being built-in
> when one or more of its dependencies is built as a module. Otherwise it
> can fail to build due to missing symbols.

Matthew: Are you okay with this, or is there some better way to do it?

I'm looking at Andreas's code for switcharoo support, and something I
want to do along with it is make apple-gmux not dependent on backlight
support (i.e. you can still build switcharoo support if
BACKLIGHT_CLASS_DEVICE is disabled). I'm assuming this patch as a
prerequisite. But things do get really monstrously ugly, so if there's a
cleaner way to enforce this that I haven't found I'd love to hear about
it.

Thanks,
Seth

> 
> Reported-by: Arun Raghavan <arun.raghavan@collabora.co.uk>
> Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
> ---
>  drivers/platform/x86/Kconfig |    4 +++-
>  1 file changed, 3 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/platform/x86/Kconfig b/drivers/platform/x86/Kconfig
> index 2a262f5..cf65f0a 100644
> --- a/drivers/platform/x86/Kconfig
> +++ b/drivers/platform/x86/Kconfig
> @@ -759,7 +759,9 @@ config SAMSUNG_Q10
>  config APPLE_GMUX
>  	tristate "Apple Gmux Driver"
>  	depends on PNP
> -	select BACKLIGHT_CLASS_DEVICE
> +	depends on BACKLIGHT_CLASS_DEVICE
> +	depends on BACKLIGHT_APPLE=n || BACKLIGHT_APPLE
> +	depends on ACPI_VIDEO=n || ACPI_VIDEO
>  	---help---
>  	  This driver provides support for the gmux device found on many
>  	  Apple laptops, which controls the display mux for the hybrid
> -- 
> 1.7.9.5
> 
> --
> To unsubscribe from this list: send the line "unsubscribe platform-driver-x86" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: [PATCH] apple-gmux: Fix kconfig dependencies

[Matthew Garrett]

On Mon, Jul 16, 2012 at 04:50:34PM -0700, Seth Forshee wrote:
> On Mon, Jul 02, 2012 at 09:55:29AM -0500, Seth Forshee wrote:
> > Fix the dependencies of apple-gmux to prevent it from being built-in
> > when one or more of its dependencies is built as a module. Otherwise it
> > can fail to build due to missing symbols.
> 
> Matthew: Are you okay with this, or is there some better way to do it?

I think this is probably about as good as it gets.

> I'm looking at Andreas's code for switcharoo support, and something I
> want to do along with it is make apple-gmux not dependent on backlight
> support (i.e. you can still build switcharoo support if
> BACKLIGHT_CLASS_DEVICE is disabled). I'm assuming this patch as a
> prerequisite. But things do get really monstrously ugly, so if there's a
> cleaner way to enforce this that I haven't found I'd love to hear about
> it.

Why do you want to support that case? One of the problems we have with 
the kernel at present is that it's trivial to come up with non-sensical 
configurations. I can't think of a single case where you'd want the gmux 
support without also having the backlight support code.

-- 
Matthew Garrett | mjg59@srcf.ucam.org

Re: [PATCH] apple-gmux: Fix kconfig dependencies

[Seth Forshee]

On Mon, Jul 23, 2012 at 02:30:10PM +0100, Matthew Garrett wrote:
> > I'm looking at Andreas's code for switcharoo support, and something I
> > want to do along with it is make apple-gmux not dependent on backlight
> > support (i.e. you can still build switcharoo support if
> > BACKLIGHT_CLASS_DEVICE is disabled). I'm assuming this patch as a
> > prerequisite. But things do get really monstrously ugly, so if there's a
> > cleaner way to enforce this that I haven't found I'd love to hear about
> > it.
> 
> Why do you want to support that case? One of the problems we have with 
> the kernel at present is that it's trivial to come up with non-sensical 
> configurations. I can't think of a single case where you'd want the gmux 
> support without also having the backlight support code.

Just because there isn't strictly any reason that the switching code
needs to depend on backlight support. But yes, I can't think of a good
reason why anyone would want this configuration. I won't mess with it
then.

Seth

Re: [PATCH] apple-gmux: Fix kconfig dependencies

[Randy Dunlap]

On 07/23/2012 06:44 AM, Seth Forshee wrote:

> On Mon, Jul 23, 2012 at 02:30:10PM +0100, Matthew Garrett wrote:
>>> I'm looking at Andreas's code for switcharoo support, and something I
>>> want to do along with it is make apple-gmux not dependent on backlight
>>> support (i.e. you can still build switcharoo support if
>>> BACKLIGHT_CLASS_DEVICE is disabled). I'm assuming this patch as a
>>> prerequisite. But things do get really monstrously ugly, so if there's a
>>> cleaner way to enforce this that I haven't found I'd love to hear about
>>> it.
>>
>> Why do you want to support that case? One of the problems we have with 
>> the kernel at present is that it's trivial to come up with non-sensical 
>> configurations. I can't think of a single case where you'd want the gmux 
>> support without also having the backlight support code.
> 
> Just because there isn't strictly any reason that the switching code
> needs to depend on backlight support. But yes, I can't think of a good
> reason why anyone would want this configuration. I won't mess with it
> then.


then it should be required in Kconfig file(s), not left to
create build errors.

-- 
~Randy

Re: [PATCH] apple-gmux: Fix kconfig dependencies

[Seth Forshee]

On Mon, Jul 23, 2012 at 08:28:05AM -0700, Randy Dunlap wrote:
> On 07/23/2012 06:44 AM, Seth Forshee wrote:
> 
> > On Mon, Jul 23, 2012 at 02:30:10PM +0100, Matthew Garrett wrote:
> >>> I'm looking at Andreas's code for switcharoo support, and something I
> >>> want to do along with it is make apple-gmux not dependent on backlight
> >>> support (i.e. you can still build switcharoo support if
> >>> BACKLIGHT_CLASS_DEVICE is disabled). I'm assuming this patch as a
> >>> prerequisite. But things do get really monstrously ugly, so if there's a
> >>> cleaner way to enforce this that I haven't found I'd love to hear about
> >>> it.
> >>
> >> Why do you want to support that case? One of the problems we have with 
> >> the kernel at present is that it's trivial to come up with non-sensical 
> >> configurations. I can't think of a single case where you'd want the gmux 
> >> support without also having the backlight support code.
> > 
> > Just because there isn't strictly any reason that the switching code
> > needs to depend on backlight support. But yes, I can't think of a good
> > reason why anyone would want this configuration. I won't mess with it
> > then.
> 
> 
> then it should be required in Kconfig file(s), not left to
> create build errors.

Of course. There are some build errors right now under certain
configurations with apple-gmux. That's a bug, and the patch I sent fixes
it. I'll test the patches that add switcheroo support for build failures
under relevant configurations, but I'm not expecting any new problems
since vga_switcheroo provides stubs when it's disabled.

Seth

[PATCH] apple-gmux: Fix kconfig dependencies

[Seth Forshee]

Fix the dependencies of apple-gmux to prevent it from being built-in
when one or more of its dependencies is built as a module. Otherwise it
can fail to build due to missing symbols.

v2: Add dependency on ACPI to fix build failure when ACPI=n

Reported-by: Arun Raghavan <arun.raghavan@collabora.co.uk>
Signed-off-by: Seth Forshee <seth.forshee@canonical.com>
---
 drivers/platform/x86/Kconfig |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/platform/x86/Kconfig b/drivers/platform/x86/Kconfig
index 2a262f5..637074d 100644
--- a/drivers/platform/x86/Kconfig
+++ b/drivers/platform/x86/Kconfig
@@ -758,8 +758,11 @@ config SAMSUNG_Q10
 
 config APPLE_GMUX
 	tristate "Apple Gmux Driver"
+	depends on ACPI
 	depends on PNP
-	select BACKLIGHT_CLASS_DEVICE
+	depends on BACKLIGHT_CLASS_DEVICE
+	depends on BACKLIGHT_APPLE=n || BACKLIGHT_APPLE
+	depends on ACPI_VIDEO=n || ACPI_VIDEO
 	---help---
 	  This driver provides support for the gmux device found on many
 	  Apple laptops, which controls the display mux for the hybrid
-- 
1.7.9.5

Re: [PATCH] apple-gmux: Fix kconfig dependencies

[Randy Dunlap]

On 08/02/2012 10:15 AM, Seth Forshee wrote:

> Fix the dependencies of apple-gmux to prevent it from being built-in
> when one or more of its dependencies is built as a module. Otherwise it
> can fail to build due to missing symbols.
> 
> v2: Add dependency on ACPI to fix build failure when ACPI=n
> 
> Reported-by: Arun Raghavan <arun.raghavan@collabora.co.uk>
> Signed-off-by: Seth Forshee <seth.forshee@canonical.com>


Acked-by: Randy Dunlap <rdunlap@xenotime.net>

Thanks.

> ---
>  drivers/platform/x86/Kconfig |    5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/platform/x86/Kconfig b/drivers/platform/x86/Kconfig
> index 2a262f5..637074d 100644
> --- a/drivers/platform/x86/Kconfig
> +++ b/drivers/platform/x86/Kconfig
> @@ -758,8 +758,11 @@ config SAMSUNG_Q10
>  
>  config APPLE_GMUX
>  	tristate "Apple Gmux Driver"
> +	depends on ACPI
>  	depends on PNP
> -	select BACKLIGHT_CLASS_DEVICE
> +	depends on BACKLIGHT_CLASS_DEVICE
> +	depends on BACKLIGHT_APPLE=n || BACKLIGHT_APPLE
> +	depends on ACPI_VIDEO=n || ACPI_VIDEO
>  	---help---
>  	  This driver provides support for the gmux device found on many
>  	  Apple laptops, which controls the display mux for the hybrid



-- 
~Randy