All of lore.kernel.org
 help / color / mirror / Atom feed
* [PATCH v3] of: fix build failure
@ 2015-12-07 12:20 Sudip Mukherjee
  2015-12-07 12:32 ` Geert Uytterhoeven
  2015-12-07 16:55 ` Rob Herring
  0 siblings, 2 replies; 5+ messages in thread
From: Sudip Mukherjee @ 2015-12-07 12:20 UTC (permalink / raw)
  To: Rob Herring, Frank Rowand, Grant Likely
  Cc: linux-kernel, devicetree, Sudip Mukherjee, Geert Uytterhoeven

We are having build failure with linux-next for sparc allmodconfig with
the error messages:

undefined reference to 'of_io_request_and_map'

CONFIG_OF is defined for sparc so it is expected that we have a
definition of of_io_request_and_map() but of/address.c is only compiled
if it is !SPARC. In other words, CONFIG_OF_ADDRESS is not defined for
sparc so we get the build failure.

Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
---

v1: had a complicated set of #ifdefs
v2: i messed up and resulted in build failure of some other arch where
CONFIG_OF is not defined.
v3: tested with allmodconfig of x86_64, defconfig of alpha and mips.

 include/linux/of_address.h | 13 +++++++++----
 1 file changed, 9 insertions(+), 4 deletions(-)

diff --git a/include/linux/of_address.h b/include/linux/of_address.h
index 507daad..02b1265 100644
--- a/include/linux/of_address.h
+++ b/include/linux/of_address.h
@@ -112,8 +112,6 @@ static inline bool of_dma_is_coherent(struct device_node *np)
 extern int of_address_to_resource(struct device_node *dev, int index,
 				  struct resource *r);
 void __iomem *of_iomap(struct device_node *node, int index);
-void __iomem *of_io_request_and_map(struct device_node *device,
-					int index, const char *name);
 #else
 
 #include <linux/io.h>
@@ -128,13 +126,20 @@ static inline void __iomem *of_iomap(struct device_node *device, int index)
 {
 	return NULL;
 }
+#endif
+
+#ifdef CONFIG_OF_ADDRESS
+void __iomem *of_io_request_and_map(struct device_node *device,
+				    int index, const char *name);
+#else
+#include <linux/io.h>
 
 static inline void __iomem *of_io_request_and_map(struct device_node *device,
-					int index, const char *name)
+						  int index, const char *name)
 {
 	return IOMEM_ERR_PTR(-EINVAL);
 }
-#endif
+#endif /* CONFIG_OF_ADDRESS */
 
 #if defined(CONFIG_OF_ADDRESS) && defined(CONFIG_PCI)
 extern const __be32 *of_get_pci_address(struct device_node *dev, int bar_no,
-- 
1.9.1


^ permalink raw reply related	[flat|nested] 5+ messages in thread

* Re: [PATCH v3] of: fix build failure
  2015-12-07 12:20 [PATCH v3] of: fix build failure Sudip Mukherjee
@ 2015-12-07 12:32 ` Geert Uytterhoeven
  2015-12-07 16:30   ` Sudip Mukherjee
  2015-12-07 16:55 ` Rob Herring
  1 sibling, 1 reply; 5+ messages in thread
From: Geert Uytterhoeven @ 2015-12-07 12:32 UTC (permalink / raw)
  To: Sudip Mukherjee
  Cc: Rob Herring, Frank Rowand, Grant Likely, linux-kernel, devicetree

Hi Sudip,

On Mon, Dec 7, 2015 at 1:20 PM, Sudip Mukherjee
<sudipm.mukherjee@gmail.com> wrote:
> We are having build failure with linux-next for sparc allmodconfig with
> the error messages:
>
> undefined reference to 'of_io_request_and_map'
>
> CONFIG_OF is defined for sparc so it is expected that we have a
> definition of of_io_request_and_map() but of/address.c is only compiled
> if it is !SPARC. In other words, CONFIG_OF_ADDRESS is not defined for
> sparc so we get the build failure.
>
> Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
> ---
>
> v1: had a complicated set of #ifdefs
> v2: i messed up and resulted in build failure of some other arch where
> CONFIG_OF is not defined.
> v3: tested with allmodconfig of x86_64, defconfig of alpha and mips.
>
>  include/linux/of_address.h | 13 +++++++++----
>  1 file changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/include/linux/of_address.h b/include/linux/of_address.h
> index 507daad..02b1265 100644
> --- a/include/linux/of_address.h
> +++ b/include/linux/of_address.h
> @@ -112,8 +112,6 @@ static inline bool of_dma_is_coherent(struct device_node *np)
>  extern int of_address_to_resource(struct device_node *dev, int index,
>                                   struct resource *r);
>  void __iomem *of_iomap(struct device_node *node, int index);
> -void __iomem *of_io_request_and_map(struct device_node *device,
> -                                       int index, const char *name);

Upon closer look, both of_address_to_resource() and of_iomap() are
provided by drivers/of/address.c, so they should also be protected by
#ifdef CONFIG_OF_ADDRESS instead of CONFIG_OF.
Then the whole #ifdef CONFIG_OF section becomes empty.

>  #else
>
>  #include <linux/io.h>
> @@ -128,13 +126,20 @@ static inline void __iomem *of_iomap(struct device_node *device, int index)
>  {
>         return NULL;
>  }
> +#endif
> +
> +#ifdef CONFIG_OF_ADDRESS
> +void __iomem *of_io_request_and_map(struct device_node *device,
> +                                   int index, const char *name);
> +#else
> +#include <linux/io.h>
>
>  static inline void __iomem *of_io_request_and_map(struct device_node *device,
> -                                       int index, const char *name)
> +                                                 int index, const char *name)
>  {
>         return IOMEM_ERR_PTR(-EINVAL);
>  }
> -#endif
> +#endif /* CONFIG_OF_ADDRESS */
>
>  #if defined(CONFIG_OF_ADDRESS) && defined(CONFIG_PCI)
>  extern const __be32 *of_get_pci_address(struct device_node *dev, int bar_no,

Hence can you please move all of them to the existing
#ifdef CONFIG_OF_ADDRESS section, and move the #include <linux/io.h>
to the top?

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v3] of: fix build failure
  2015-12-07 12:32 ` Geert Uytterhoeven
@ 2015-12-07 16:30   ` Sudip Mukherjee
  2015-12-07 18:12     ` Geert Uytterhoeven
  0 siblings, 1 reply; 5+ messages in thread
From: Sudip Mukherjee @ 2015-12-07 16:30 UTC (permalink / raw)
  To: Geert Uytterhoeven
  Cc: Rob Herring, Frank Rowand, Grant Likely, linux-kernel, devicetree

On Mon, Dec 07, 2015 at 01:32:30PM +0100, Geert Uytterhoeven wrote:
> Hi Sudip,
> 
> On Mon, Dec 7, 2015 at 1:20 PM, Sudip Mukherjee
> <sudipm.mukherjee@gmail.com> wrote:
> > We are having build failure with linux-next for sparc allmodconfig with
> > the error messages:
> >
> > undefined reference to 'of_io_request_and_map'
> >
> > CONFIG_OF is defined for sparc so it is expected that we have a
> > definition of of_io_request_and_map() but of/address.c is only compiled
> > if it is !SPARC. In other words, CONFIG_OF_ADDRESS is not defined for
> > sparc so we get the build failure.
> >
> > Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
> > ---
> >
<snip>
> > diff --git a/include/linux/of_address.h b/include/linux/of_address.h
> > index 507daad..02b1265 100644
> > --- a/include/linux/of_address.h
> > +++ b/include/linux/of_address.h
> > @@ -112,8 +112,6 @@ static inline bool of_dma_is_coherent(struct device_node *np)
> >  extern int of_address_to_resource(struct device_node *dev, int index,
> >                                   struct resource *r);
> >  void __iomem *of_iomap(struct device_node *node, int index);
> > -void __iomem *of_io_request_and_map(struct device_node *device,
> > -                                       int index, const char *name);
> 
> Upon closer look, both of_address_to_resource() and of_iomap() are
> provided by drivers/of/address.c, so they should also be protected by
> #ifdef CONFIG_OF_ADDRESS instead of CONFIG_OF.
> Then the whole #ifdef CONFIG_OF section becomes empty.
> 
> >  #else
<snip>
> >
> >  #if defined(CONFIG_OF_ADDRESS) && defined(CONFIG_PCI)
> >  extern const __be32 *of_get_pci_address(struct device_node *dev, int bar_no,
> 
> Hence can you please move all of them to the existing
> #ifdef CONFIG_OF_ADDRESS section, and move the #include <linux/io.h>
> to the top?

oops, no. I think you have missed seeing few more places.
of_address_to_resource() of_iomap() are both defined in
arch/sparc/kernel/of_device_common.c
So if it is moved under #ifdef CONFIG_OF_ADDRESS then we again
get a build failure for duplicate definition of them.

regards
sudip

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v3] of: fix build failure
  2015-12-07 12:20 [PATCH v3] of: fix build failure Sudip Mukherjee
  2015-12-07 12:32 ` Geert Uytterhoeven
@ 2015-12-07 16:55 ` Rob Herring
  1 sibling, 0 replies; 5+ messages in thread
From: Rob Herring @ 2015-12-07 16:55 UTC (permalink / raw)
  To: Sudip Mukherjee
  Cc: Frank Rowand, Grant Likely, linux-kernel, devicetree, Geert Uytterhoeven

On Mon, Dec 7, 2015 at 6:20 AM, Sudip Mukherjee
<sudipm.mukherjee@gmail.com> wrote:
> We are having build failure with linux-next for sparc allmodconfig with
> the error messages:

Please come up with a subject more specific than applicable to any
build failure.

> undefined reference to 'of_io_request_and_map'

Specifying on which driver and or commit would be helpful here. Then I
know if I need to flag this for stable or not without having to figure
that out myself. I'm assume something new got enabled for sparc?

> CONFIG_OF is defined for sparc so it is expected that we have a
> definition of of_io_request_and_map() but of/address.c is only compiled
> if it is !SPARC. In other words, CONFIG_OF_ADDRESS is not defined for
> sparc so we get the build failure.
>
> Signed-off-by: Sudip Mukherjee <sudip@vectorindia.org>
> ---
>
> v1: had a complicated set of #ifdefs
> v2: i messed up and resulted in build failure of some other arch where
> CONFIG_OF is not defined.
> v3: tested with allmodconfig of x86_64, defconfig of alpha and mips.
>
>  include/linux/of_address.h | 13 +++++++++----
>  1 file changed, 9 insertions(+), 4 deletions(-)
>
> diff --git a/include/linux/of_address.h b/include/linux/of_address.h
> index 507daad..02b1265 100644
> --- a/include/linux/of_address.h
> +++ b/include/linux/of_address.h
> @@ -112,8 +112,6 @@ static inline bool of_dma_is_coherent(struct device_node *np)
>  extern int of_address_to_resource(struct device_node *dev, int index,
>                                   struct resource *r);
>  void __iomem *of_iomap(struct device_node *node, int index);
> -void __iomem *of_io_request_and_map(struct device_node *device,
> -                                       int index, const char *name);
>  #else
>
>  #include <linux/io.h>
> @@ -128,13 +126,20 @@ static inline void __iomem *of_iomap(struct device_node *device, int index)
>  {
>         return NULL;
>  }
> +#endif
> +
> +#ifdef CONFIG_OF_ADDRESS

This is the correct protection (Geert is wrong), but please move it to
the existing "#ifdef CONFIG_OF_ADDRESS" section.

> +void __iomem *of_io_request_and_map(struct device_node *device,
> +                                   int index, const char *name);
> +#else
> +#include <linux/io.h>
>
>  static inline void __iomem *of_io_request_and_map(struct device_node *device,
> -                                       int index, const char *name)
> +                                                 int index, const char *name)
>  {
>         return IOMEM_ERR_PTR(-EINVAL);
>  }
> -#endif
> +#endif /* CONFIG_OF_ADDRESS */
>
>  #if defined(CONFIG_OF_ADDRESS) && defined(CONFIG_PCI)
>  extern const __be32 *of_get_pci_address(struct device_node *dev, int bar_no,
> --
> 1.9.1
>

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v3] of: fix build failure
  2015-12-07 16:30   ` Sudip Mukherjee
@ 2015-12-07 18:12     ` Geert Uytterhoeven
  0 siblings, 0 replies; 5+ messages in thread
From: Geert Uytterhoeven @ 2015-12-07 18:12 UTC (permalink / raw)
  To: Sudip Mukherjee
  Cc: Rob Herring, Frank Rowand, Grant Likely, linux-kernel, devicetree

On Mon, Dec 7, 2015 at 5:30 PM, Sudip Mukherjee
<sudipm.mukherjee@gmail.com> wrote:
>> > --- a/include/linux/of_address.h
>> > +++ b/include/linux/of_address.h
>> > @@ -112,8 +112,6 @@ static inline bool of_dma_is_coherent(struct device_node *np)
>> >  extern int of_address_to_resource(struct device_node *dev, int index,
>> >                                   struct resource *r);
>> >  void __iomem *of_iomap(struct device_node *node, int index);
>> > -void __iomem *of_io_request_and_map(struct device_node *device,
>> > -                                       int index, const char *name);
>>
>> Upon closer look, both of_address_to_resource() and of_iomap() are
>> provided by drivers/of/address.c, so they should also be protected by
>> #ifdef CONFIG_OF_ADDRESS instead of CONFIG_OF.
>> Then the whole #ifdef CONFIG_OF section becomes empty.
>>
>> >  #else
> <snip>
>> >
>> >  #if defined(CONFIG_OF_ADDRESS) && defined(CONFIG_PCI)
>> >  extern const __be32 *of_get_pci_address(struct device_node *dev, int bar_no,
>>
>> Hence can you please move all of them to the existing
>> #ifdef CONFIG_OF_ADDRESS section, and move the #include <linux/io.h>
>> to the top?
>
> oops, no. I think you have missed seeing few more places.
> of_address_to_resource() of_iomap() are both defined in
> arch/sparc/kernel/of_device_common.c
> So if it is moved under #ifdef CONFIG_OF_ADDRESS then we again
> get a build failure for duplicate definition of them.

Oops, you're right.

Sorry, I never noticed before sparc provides its own definitions.

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2015-12-07 18:12 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2015-12-07 12:20 [PATCH v3] of: fix build failure Sudip Mukherjee
2015-12-07 12:32 ` Geert Uytterhoeven
2015-12-07 16:30   ` Sudip Mukherjee
2015-12-07 18:12     ` Geert Uytterhoeven
2015-12-07 16:55 ` Rob Herring

This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.