[email protected] wrote:
> The patch titled
> cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
> has been added to the -mm tree. Its filename is
> cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix.patch
>
> Before you just go and hit "reply", please:
> a) Consider who else should be cc'ed
> b) Prefer to cc a suitable mailing list as well
> c) Ideally: find the original patch on the mailing list and do a
> reply-to-all to that, adding suitable additional cc's
>
> *** Remember to use Documentation/SubmitChecklist when testing your code ***
>
> See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
> out what to do about this
>
> The current -mm tree may be found at http://userweb.kernel.org/~akpm/mmotm/
>
> ------------------------------------------------------
> Subject: cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
> From: Andres Salomon <[email protected]>
>
> force GPIO_CS553X to be defined when OLPC is selected
>
> We've begun using the generic GPIO stuff for geodes, so OLPC implicitly
> depends upon this. This patch forces selection of it.
>
> Signed-off-by: Andres Salomon <[email protected]>
> Cc: Takashi Iwai <[email protected]>
> Cc: Jeremy Katz <[email protected]>
> Cc: Jordan Crouse <[email protected]>
> Cc: David Brownell <[email protected]>
> Cc: Randy Dunlap <[email protected]>
> Signed-off-by: Andrew Morton <[email protected]>
> ---
>
> arch/x86/Kconfig | 1 +
> pci/cs5535audio/Makefile | 0
> 2 files changed, 1 insertion(+)
>
> diff -puN arch/x86/Kconfig~cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix arch/x86/Kconfig
> --- a/arch/x86/Kconfig~cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
> +++ a/arch/x86/Kconfig
> @@ -1972,6 +1972,7 @@ config GEODE_MFGPT_TIMER
>
> config OLPC
> bool "One Laptop Per Child support"
> + select GPIO_CS553X
> default n
> ---help---
> Add support for detecting the unique features of the OLPC
In mmotm-2009-0410, this patch causes build errors when GPIOLIB is not enabled
since GPIO_CS553X uses gpiolib functions & struct fields:
drivers/gpio/cs553x-gpio.c:23: error: field 'chip' has incomplete type
drivers/gpio/cs553x-gpio.c:150: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:150: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:151: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:151: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:153: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:153: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:154: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:154: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:156: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:156: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:157: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:157: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:159: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:159: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:160: error: field name not in record or union initializer
drivers/gpio/cs553x-gpio.c:160: error: (near initialization for 'cs553x_gpio_chip.chip')
drivers/gpio/cs553x-gpio.c:197: error: implicit declaration of function 'gpiochip_add'
drivers/gpio/cs553x-gpio.c:216: error: implicit declaration of function 'gpiochip_remove'
One possible patch is also to select GPIOLIB above in the OLPC config block.
--
~Randy
On Fri, 10 Apr 2009 13:00:54 -0700
Randy Dunlap <[email protected]> wrote:
> [email protected] wrote:
> > The patch titled
> > cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
> > has been added to the -mm tree. Its filename is
> > cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix.patch
> >
> > Before you just go and hit "reply", please:
> > a) Consider who else should be cc'ed
> > b) Prefer to cc a suitable mailing list as well
> > c) Ideally: find the original patch on the mailing list and do a
> > reply-to-all to that, adding suitable additional cc's
> >
> > *** Remember to use Documentation/SubmitChecklist when testing your
> > code ***
> >
> > See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
> > out what to do about this
> >
> > The current -mm tree may be found at
> > http://userweb.kernel.org/~akpm/mmotm/
> >
> > ------------------------------------------------------
> > Subject:
> > cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix From:
> > Andres Salomon <[email protected]>
> >
> > force GPIO_CS553X to be defined when OLPC is selected
> >
> > We've begun using the generic GPIO stuff for geodes, so OLPC
> > implicitly depends upon this. This patch forces selection of it.
> >
> > Signed-off-by: Andres Salomon <[email protected]>
> > Cc: Takashi Iwai <[email protected]>
> > Cc: Jeremy Katz <[email protected]>
> > Cc: Jordan Crouse <[email protected]>
> > Cc: David Brownell <[email protected]>
> > Cc: Randy Dunlap <[email protected]>
> > Signed-off-by: Andrew Morton <[email protected]>
> > ---
> >
> > arch/x86/Kconfig | 1 +
> > pci/cs5535audio/Makefile | 0
> > 2 files changed, 1 insertion(+)
> >
> > diff -puN
> > arch/x86/Kconfig~cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
> > arch/x86/Kconfig ---
> > a/arch/x86/Kconfig~cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
> > +++ a/arch/x86/Kconfig @@ -1972,6 +1972,7 @@ config
> > GEODE_MFGPT_TIMER config OLPC
> > bool "One Laptop Per Child support"
> > + select GPIO_CS553X
> > default n
> > ---help---
> > Add support for detecting the unique features of the OLPC
>
>
> In mmotm-2009-0410, this patch causes build errors when GPIOLIB is
> not enabled since GPIO_CS553X uses gpiolib functions & struct fields:
>
> drivers/gpio/cs553x-gpio.c:23: error: field 'chip' has incomplete type
[...]
>
>
> One possible patch is also to select GPIOLIB above in the OLPC config
> block.
>
Does the following help? I think this is what we want (gpio-cs553x uses
gpiolib). I'm a bit unclear how the other gpio drivers get away with
including linux/gpio.h and using gpio_chip without requiring GPIOLIB.
Maybe they implicitly are built on architectures which select
CONFIG_ARCH_REQUIRE_GPIOLIB?
Force GPIO_CS553X to select GPIOLIB, which it makes use of.
Signed-off-by: Andres Salomon <[email protected]>
diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig
index 797ff45..17d7364 100644
--- a/drivers/gpio/Kconfig
+++ b/drivers/gpio/Kconfig
@@ -148,6 +148,7 @@ comment "PCI GPIO expanders:"
config GPIO_CS553X
tristate "AMD cs5535/cs5536 GPIO support"
depends on PCI && !CS5535_GPIO && !MGEODE_LX
+ select GPIOLIB
help
The AMD cs5535 and cs5536 southbridges support 28 GPIO pins that
can be used for quite a number of things. The cs553x is found on
Andres Salomon wrote:
> On Fri, 10 Apr 2009 13:00:54 -0700
> Randy Dunlap <[email protected]> wrote:
>
>> [email protected] wrote:
>>> The patch titled
>>> cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
>>> has been added to the -mm tree. Its filename is
>>> cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix.patch
>>>
>>> Before you just go and hit "reply", please:
>>> a) Consider who else should be cc'ed
>>> b) Prefer to cc a suitable mailing list as well
>>> c) Ideally: find the original patch on the mailing list and do a
>>> reply-to-all to that, adding suitable additional cc's
>>>
>>> *** Remember to use Documentation/SubmitChecklist when testing your
>>> code ***
>>>
>>> See http://userweb.kernel.org/~akpm/stuff/added-to-mm.txt to find
>>> out what to do about this
>>>
>>> The current -mm tree may be found at
>>> http://userweb.kernel.org/~akpm/mmotm/
>>>
>>> ------------------------------------------------------
>>> Subject:
>>> cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix From:
>>> Andres Salomon <[email protected]>
>>>
>>> force GPIO_CS553X to be defined when OLPC is selected
>>>
>>> We've begun using the generic GPIO stuff for geodes, so OLPC
>>> implicitly depends upon this. This patch forces selection of it.
>>>
>>> Signed-off-by: Andres Salomon <[email protected]>
>>> Cc: Takashi Iwai <[email protected]>
>>> Cc: Jeremy Katz <[email protected]>
>>> Cc: Jordan Crouse <[email protected]>
>>> Cc: David Brownell <[email protected]>
>>> Cc: Randy Dunlap <[email protected]>
>>> Signed-off-by: Andrew Morton <[email protected]>
>>> ---
>>>
>>> arch/x86/Kconfig | 1 +
>>> pci/cs5535audio/Makefile | 0
>>> 2 files changed, 1 insertion(+)
>>>
>>> diff -puN
>>> arch/x86/Kconfig~cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
>>> arch/x86/Kconfig ---
>>> a/arch/x86/Kconfig~cs553x-gpio-add-amd-cs5535-cs5536-gpio-driver-support-fix-fix
>>> +++ a/arch/x86/Kconfig @@ -1972,6 +1972,7 @@ config
>>> GEODE_MFGPT_TIMER config OLPC
>>> bool "One Laptop Per Child support"
>>> + select GPIO_CS553X
>>> default n
>>> ---help---
>>> Add support for detecting the unique features of the OLPC
>>
>> In mmotm-2009-0410, this patch causes build errors when GPIOLIB is
>> not enabled since GPIO_CS553X uses gpiolib functions & struct fields:
>>
>> drivers/gpio/cs553x-gpio.c:23: error: field 'chip' has incomplete type
> [...]
>>
>> One possible patch is also to select GPIOLIB above in the OLPC config
>> block.
>>
>
> Does the following help? I think this is what we want (gpio-cs553x uses
> gpiolib). I'm a bit unclear how the other gpio drivers get away with
> including linux/gpio.h and using gpio_chip without requiring GPIOLIB.
> Maybe they implicitly are built on architectures which select
> CONFIG_ARCH_REQUIRE_GPIOLIB?
>
It helps some (and is needed), but there is another problem.
When PCI is not enabled:
drivers/gpio/cs553x-gpio.c: In function 'cs553x_gpio_probe':
drivers/gpio/cs553x-gpio.c:176: error: implicit declaration of function 'pci_enable_device_io'
drivers/gpio/cs553x-gpio.c:182: error: implicit declaration of function 'pci_request_region'
drivers/gpio/cs553x-gpio.c:207: error: implicit declaration of function 'pci_release_region'
I guess that OLPC needs an additional dependency on PCI ??
>
>
> Force GPIO_CS553X to select GPIOLIB, which it makes use of.
>
> Signed-off-by: Andres Salomon <[email protected]>
>
> diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig
> index 797ff45..17d7364 100644
> --- a/drivers/gpio/Kconfig
> +++ b/drivers/gpio/Kconfig
> @@ -148,6 +148,7 @@ comment "PCI GPIO expanders:"
> config GPIO_CS553X
> tristate "AMD cs5535/cs5536 GPIO support"
> depends on PCI && !CS5535_GPIO && !MGEODE_LX
> + select GPIOLIB
> help
> The AMD cs5535 and cs5536 southbridges support 28 GPIO pins that
> can be used for quite a number of things. The cs553x is found on
--
~Randy
On Sun, 12 Apr 2009 21:52:03 -0400
Andres Salomon <[email protected]> wrote:
> Force GPIO_CS553X to select GPIOLIB, which it makes use of.
>
> Signed-off-by: Andres Salomon <[email protected]>
>
> diff --git a/drivers/gpio/Kconfig b/drivers/gpio/Kconfig
> index 797ff45..17d7364 100644
> --- a/drivers/gpio/Kconfig
> +++ b/drivers/gpio/Kconfig
> @@ -148,6 +148,7 @@ comment "PCI GPIO expanders:"
> config GPIO_CS553X
> tristate "AMD cs5535/cs5536 GPIO support"
> depends on PCI && !CS5535_GPIO && !MGEODE_LX
> + select GPIOLIB
> help
> The AMD cs5535 and cs5536 southbridges support 28 GPIO pins that
> can be used for quite a number of things. The cs553x is found on
x86_64 allmodconfig:
drivers/gpio/Kconfig:28:error: found recursive dependency: GPIOLIB -> GPIO_CS553X -> GPIOLIB
Andres, how does this look to you?
It fixes the problems that I know about.
From: Randy Dunlap <[email protected]>
Fix multiple build errors due to cs553x-gpio using GPIO APIs
and using PCI APIs.
Signed-off-by: Randy Dunlap <[email protected]>
---
arch/x86/Kconfig | 3 ++-
1 file changed, 2 insertions(+), 1 deletion(-)
--- mmotm-2009-0417-1519.orig/arch/x86/Kconfig
+++ mmotm-2009-0417-1519/arch/x86/Kconfig
@@ -1969,7 +1969,8 @@ config GEODE_MFGPT_TIMER
config OLPC
bool "One Laptop Per Child support"
- select GPIO_CS553X
+ select GPIO_CS553X if PCI
+ select GPIOLIB if PCI
default n
---help---
Add support for detecting the unique features of the OLPC
On Sat, 18 Apr 2009 18:01:18 -0700
Randy Dunlap <[email protected]> wrote:
> Andres, how does this look to you?
>
> It fixes the problems that I know about.
>
>
>
This may be okay as a separate patch, but doesn't really fix the problem.
cs553x-gpio isn't just for OLPC, it can be used by plain Geodes. I've
lost track of the state of the patches at this point, but the solution
would be to simply have CONFIG_GPIO_CS553X depend upon PCI.
> From: Randy Dunlap <[email protected]>
>
> Fix multiple build errors due to cs553x-gpio using GPIO APIs
> and using PCI APIs.
>
> Signed-off-by: Randy Dunlap <[email protected]>
> ---
> arch/x86/Kconfig | 3 ++-
> 1 file changed, 2 insertions(+), 1 deletion(-)
>
> --- mmotm-2009-0417-1519.orig/arch/x86/Kconfig
> +++ mmotm-2009-0417-1519/arch/x86/Kconfig
> @@ -1969,7 +1969,8 @@ config GEODE_MFGPT_TIMER
>
> config OLPC
> bool "One Laptop Per Child support"
> - select GPIO_CS553X
> + select GPIO_CS553X if PCI
> + select GPIOLIB if PCI
> default n
> ---help---
> Add support for detecting the unique features of the OLPC