[PATCH] sh: make !MMU iounmap an inline function

[PATCH] sh: make !MMU iounmap an inline function

[Vincent Whitchurch]

The current definition of sh's !MMU iounmap makes it unusable in certain
kinds of code, such as this instance caught by 0day on a patch posted to
the mailing lists:

 In file included from include/linux/io.h:13,
                  from drivers/mtd/devices/phram.c:21:
 drivers/mtd/devices/phram.c: In function 'register_device':
 arch/sh/include/asm/io.h:274:33: error: expected expression before 'do'
   274 | #define iounmap(addr)           do { } while (0)
       |                                 ^~
 drivers/mtd/devices/phram.c:150:44: note: in expansion of macro 'iounmap'
   150 |         cached ? memunmap(new->mtd.priv) : iounmap(new->mtd.priv);
       |                                            ^~~~~~~

Make it an inline function as it was earlier.

Link: https://lore.kernel.org/lkml/202204131446.omJ5mC54-lkp@intel.com/
Fixes: 13f1fc870dd747131 ("sh: move the ioremap implementation out of line")
Reported-by: kernel test robot <lkp@intel.com>
Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
---
 arch/sh/include/asm/io.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/arch/sh/include/asm/io.h b/arch/sh/include/asm/io.h
index cf9a3ec32406..6bb978cf2ece 100644
--- a/arch/sh/include/asm/io.h
+++ b/arch/sh/include/asm/io.h
@@ -271,7 +271,7 @@ static inline void __iomem *ioremap_prot(phys_addr_t offset, unsigned long size,
 #endif /* CONFIG_HAVE_IOREMAP_PROT */
 
 #else /* CONFIG_MMU */
-#define iounmap(addr)		do { } while (0)
+static inline void iounmap(void __iomem *addr) {};
 #define ioremap(offset, size)	((void __iomem *)(unsigned long)(offset))
 #endif /* CONFIG_MMU */
 
-- 
2.34.1

Re: [PATCH] sh: make !MMU iounmap an inline function

[Rob Landley]

On 4/14/22 03:12, Vincent Whitchurch wrote:
> The current definition of sh's !MMU iounmap makes it unusable in certain
> kinds of code, such as this instance caught by 0day on a patch posted to
> the mailing lists:
> 
>  In file included from include/linux/io.h:13,
>                   from drivers/mtd/devices/phram.c:21:
>  drivers/mtd/devices/phram.c: In function 'register_device':
>  arch/sh/include/asm/io.h:274:33: error: expected expression before 'do'
>    274 | #define iounmap(addr)           do { } while (0)
>        |                                 ^~
>  drivers/mtd/devices/phram.c:150:44: note: in expansion of macro 'iounmap'
>    150 |         cached ? memunmap(new->mtd.priv) : iounmap(new->mtd.priv);
>        |                                            ^~~~~~~
> 
> Make it an inline function as it was earlier.
> 
> Link: https://lore.kernel.org/lkml/202204131446.omJ5mC54-lkp@intel.com/
> Fixes: 13f1fc870dd747131 ("sh: move the ioremap implementation out of line")
> Reported-by: kernel test robot <lkp@intel.com>
> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>

Tested-by: Rob Landley <rob@landley.net>

Worked For Me.

Rob

Re: [PATCH] sh: make !MMU iounmap an inline function

[Geert Uytterhoeven]

Hi Rich, Sato-san,

On Sat, Apr 16, 2022 at 12:01 AM Rob Landley <rob@landley.net> wrote:
> On 4/14/22 03:12, Vincent Whitchurch wrote:
> > The current definition of sh's !MMU iounmap makes it unusable in certain
> > kinds of code, such as this instance caught by 0day on a patch posted to
> > the mailing lists:
> >
> >  In file included from include/linux/io.h:13,
> >                   from drivers/mtd/devices/phram.c:21:
> >  drivers/mtd/devices/phram.c: In function 'register_device':
> >  arch/sh/include/asm/io.h:274:33: error: expected expression before 'do'
> >    274 | #define iounmap(addr)           do { } while (0)
> >        |                                 ^~
> >  drivers/mtd/devices/phram.c:150:44: note: in expansion of macro 'iounmap'
> >    150 |         cached ? memunmap(new->mtd.priv) : iounmap(new->mtd.priv);
> >        |                                            ^~~~~~~
> >
> > Make it an inline function as it was earlier.
> >
> > Link: https://lore.kernel.org/lkml/202204131446.omJ5mC54-lkp@intel.com/
> > Fixes: 13f1fc870dd747131 ("sh: move the ioremap implementation out of line")
> > Reported-by: kernel test robot <lkp@intel.com>
> > Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
>
> Tested-by: Rob Landley <rob@landley.net>
>
> Worked For Me.

Can we please get any of [1], [2], [3], or [4] applied and sent
upstream, so people can stop wasting time on recreating and
resubmitting similar patches all the time?

Thank you!

[1] PATCH] sh: Fix set but not used warnings with !CONFIG_MMU
    https://lore.kernel.org/r/20201208182117.163642-2-sam@ravnborg.org
[2] [PATCH] sh: make iounmap() a static inline again
    https://lore.kernel.org/r/20220105085746.1116726-1-mkl@pengutronix.de
[3] [PATCH] sh: Convert nommu io{re,un}map() to static inline functions
    https://lore.kernel.org/r/4ed0a7a0d3fa912a5b44c451884818f2c138ef42.1644914600.git.geert+renesas@glider.be/
[4] [PATCH] sh: make !MMU iounmap an inline function
    https://lore.kernel.org/r/20220414081257.1487499-1-vincent.whitchurch@axis.com/

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

Re: [PATCH] sh: make !MMU iounmap an inline function

[John Paul Adrian Glaubitz]

Hi Geert!

On 4/19/22 11:42, Geert Uytterhoeven wrote:
> Hi Rich, Sato-san,
> 
> On Sat, Apr 16, 2022 at 12:01 AM Rob Landley <rob@landley.net> wrote:
>> On 4/14/22 03:12, Vincent Whitchurch wrote:
>>> The current definition of sh's !MMU iounmap makes it unusable in certain
>>> kinds of code, such as this instance caught by 0day on a patch posted to
>>> the mailing lists:
>>>
>>>  In file included from include/linux/io.h:13,
>>>                   from drivers/mtd/devices/phram.c:21:
>>>  drivers/mtd/devices/phram.c: In function 'register_device':
>>>  arch/sh/include/asm/io.h:274:33: error: expected expression before 'do'
>>>    274 | #define iounmap(addr)           do { } while (0)
>>>        |                                 ^~
>>>  drivers/mtd/devices/phram.c:150:44: note: in expansion of macro 'iounmap'
>>>    150 |         cached ? memunmap(new->mtd.priv) : iounmap(new->mtd.priv);
>>>        |                                            ^~~~~~~
>>>
>>> Make it an inline function as it was earlier.
>>>
>>> Link: https://lore.kernel.org/lkml/202204131446.omJ5mC54-lkp@intel.com/
>>> Fixes: 13f1fc870dd747131 ("sh: move the ioremap implementation out of line")
>>> Reported-by: kernel test robot <lkp@intel.com>
>>> Signed-off-by: Vincent Whitchurch <vincent.whitchurch@axis.com>
>>
>> Tested-by: Rob Landley <rob@landley.net>
>>
>> Worked For Me.
> 
> Can we please get any of [1], [2], [3], or [4] applied and sent
> upstream, so people can stop wasting time on recreating and
> resubmitting similar patches all the time?

I agree. Those patches shouldn't be sent in vain. I very much appreciate everyone
sending in patches to improve Linux support for SH which is why I find it saddening
that the maintainers don't find enough time to review the patches.

Adrian

-- 
 .''`.  John Paul Adrian Glaubitz
: :' :  Debian Developer - glaubitz@debian.org
`. `'   Freie Universitaet Berlin - glaubitz@physik.fu-berlin.de
  `-    GPG: 62FF 8A75 84E0 2956 9546  0006 7426 3B37 F5B5 F913