Re: [PATCH V5 12/20] asm-generic/io.h: iounmap/ioport_unmap cleanup.h support

[David E. Box]

+Baoquan for ioremap question.

On Thu, 2023-11-23 at 16:30 +0200, Ilpo Järvinen wrote:
> On Wed, 22 Nov 2023, David E. Box wrote:
> 
> > Add auto-release cleanups for iounmap() and ioport_unmap().
> > 
> > Signed-off-by: David E. Box <david.e.box@xxxxxxxxxxxxxxx>
> > Suggested-by: Ilpo Järvinen <ilpo.jarvinen@xxxxxxxxxxxxxxx>
> > ---
> > V2 - Move from linux/io.h to asm-generic/io.h. Adds iounmap cleanup if
> >      iounmap() is defined. Adds ioport_unmap cleanup if CONFIG_IOPORT_MAP
> >      is defined.
> > 
> >  include/asm-generic/io.h | 6 ++++++
> >  1 file changed, 6 insertions(+)
> > 
> > diff --git a/include/asm-generic/io.h b/include/asm-generic/io.h
> > index bac63e874c7b..9ef0332490b1 100644
> > --- a/include/asm-generic/io.h
> > +++ b/include/asm-generic/io.h
> > @@ -8,6 +8,7 @@
> >  #define __ASM_GENERIC_IO_H
> >  
> >  #include <asm/page.h> /* I/O is all done through memory accesses */
> > +#include <linux/cleanup.h>
> >  #include <linux/string.h> /* for memset() and memcpy() */
> >  #include <linux/types.h>
> >  #include <linux/instruction_pointer.h>
> > @@ -1065,6 +1066,10 @@ static inline void __iomem *ioremap(phys_addr_t addr,
> > size_t size)
> >  #endif
> >  #endif /* !CONFIG_MMU || CONFIG_GENERIC_IOREMAP */
> >  
> > +#ifdef iounmap
> > +DEFINE_FREE(iounmap, void __iomem *, iounmap(_T));
> > +#endif

Baoquan, LKP is reporting an undeclared function 'iounmap' error with the above
change from this patch when building for s390 with PCI disabled. The ioremap
defines in arch/s390/include/asm/io.h are not wrapped under the #ifdef
CONFIG_PCI block. Shouldn't they be since the s390 Kconfig only adds
GENERIC_IOREMAP if PCI?

https://lore.kernel.org/oe-kbuild-all/202311241214.jcL84du7-lkp@xxxxxxxxx


Note that the report includes pointer arithmetic warnings that are not related
to this patch. Those warnings occur in mainline as well.

David

> > +
> >  #ifndef ioremap_wc
> >  #define ioremap_wc ioremap
> >  #endif
> > @@ -1127,6 +1132,7 @@ static inline void ioport_unmap(void __iomem *p)
> >  extern void __iomem *ioport_map(unsigned long port, unsigned int nr);
> >  extern void ioport_unmap(void __iomem *p);
> >  #endif /* CONFIG_GENERIC_IOMAP */
> > +DEFINE_FREE(ioport_unmap, void __iomem *, ioport_unmap(_T));
> >  #endif /* CONFIG_HAS_IOPORT_MAP */
> >  
> >  #ifndef CONFIG_GENERIC_IOMAP
> 
> Has this now built successfully with LKP? (I don't think we get success 
> notifications from LKP for patch submissions, only failures).
> 
> There were some odd errors last time but I think all they were unrelated 
> to this change (besides the checkpatch false positive, I mean).
>