All of lore.kernel.org
 help / color / mirror / Atom feed
* Re: [tip:x86/headers] x86: truncate ISA addresses to unsigned int
       [not found] <tip-a7eb518998529c08cc53fef17756d9fe433b0c23@kernel.org>
@ 2009-02-17 21:24 ` James Bottomley
  2009-02-17 21:25   ` H. Peter Anvin
  2009-02-17 21:55   ` Ingo Molnar
  0 siblings, 2 replies; 3+ messages in thread
From: James Bottomley @ 2009-02-17 21:24 UTC (permalink / raw)
  To: hpa, mingo, tglx, hpa, linux-kernel; +Cc: linux-tip-commits

On Tue, 2009-02-17 at 21:12 +0000, H. Peter Anvin wrote:
> Author:     H. Peter Anvin <hpa@linux.intel.com>
> AuthorDate: Tue, 17 Feb 2009 13:01:51 -0800
> Commit:     H. Peter Anvin <hpa@linux.intel.com>
> CommitDate: Tue, 17 Feb 2009 13:01:51 -0800
> 
> x86: truncate ISA addresses to unsigned int
> 
> Impact: Cleanup; fix inappropriate macro use
> 
> ISA addresses on x86 are mapped 1:1 with the physical address space.
> Since the ISA address space is only 24 bits (32 for VLB or LPC) it
> will always fit in an unsigned int, and at least in the aha1542 driver
> using a wider type would cause an undesirable promotion.  Hence
> explicitly cast the ISA bus addresses to unsigned int.
> 
> Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
> Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
> 
> 
> ---
>  arch/x86/include/asm/io.h |   11 ++++++++---
>  1 files changed, 8 insertions(+), 3 deletions(-)
> 
> diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
> index 4f8e820..683d0b4 100644
> --- a/arch/x86/include/asm/io.h
> +++ b/arch/x86/include/asm/io.h
> @@ -124,10 +124,15 @@ static inline void *phys_to_virt(phys_addr_t address)
>  
>  /*
>   * ISA I/O bus memory addresses are 1:1 with the physical address.
> + * However, we truncate the address to unsigned int to avoid undesirable
> + * promitions in legacy drivers.
>   */
> -#define isa_virt_to_bus (unsigned long)virt_to_phys
> -#define isa_page_to_bus page_to_phys
> -#define isa_bus_to_virt phys_to_virt
> +static inline unsigned int isa_virt_to_bus(volatile void *address)

The rest looks fine, but why is the address volatile?  The macro below
never actually dereferences it.

> +{
> +	return (unsigned int)virt_to_phys(address);
> +}
> +#define isa_page_to_bus(page)	((unsigned int)page_to_phys(page))
> +#define isa_bus_to_virt		phys_to_virt
>  
>  /*
>   * However PCI ones are not necessarily 1:1 and therefore these interfaces

James



^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [tip:x86/headers] x86: truncate ISA addresses to unsigned int
  2009-02-17 21:24 ` [tip:x86/headers] x86: truncate ISA addresses to unsigned int James Bottomley
@ 2009-02-17 21:25   ` H. Peter Anvin
  2009-02-17 21:55   ` Ingo Molnar
  1 sibling, 0 replies; 3+ messages in thread
From: H. Peter Anvin @ 2009-02-17 21:25 UTC (permalink / raw)
  To: James Bottomley; +Cc: hpa, mingo, tglx, linux-kernel, linux-tip-commits

James Bottomley wrote:
>> +static inline unsigned int isa_virt_to_bus(volatile void *address)
> 
> The rest looks fine, but why is the address volatile?  The macro below
> never actually dereferences it.
> 

Because a volatile is permitted.  virt_to_phys() takes a volatile 
pointer for that reason.

	-hpa

^ permalink raw reply	[flat|nested] 3+ messages in thread

* Re: [tip:x86/headers] x86: truncate ISA addresses to unsigned int
  2009-02-17 21:24 ` [tip:x86/headers] x86: truncate ISA addresses to unsigned int James Bottomley
  2009-02-17 21:25   ` H. Peter Anvin
@ 2009-02-17 21:55   ` Ingo Molnar
  1 sibling, 0 replies; 3+ messages in thread
From: Ingo Molnar @ 2009-02-17 21:55 UTC (permalink / raw)
  To: James Bottomley; +Cc: hpa, mingo, tglx, hpa, linux-kernel, linux-tip-commits


* James Bottomley <James.Bottomley@HansenPartnership.com> wrote:

> On Tue, 2009-02-17 at 21:12 +0000, H. Peter Anvin wrote:
> > Author:     H. Peter Anvin <hpa@linux.intel.com>
> > AuthorDate: Tue, 17 Feb 2009 13:01:51 -0800
> > Commit:     H. Peter Anvin <hpa@linux.intel.com>
> > CommitDate: Tue, 17 Feb 2009 13:01:51 -0800
> > 
> > x86: truncate ISA addresses to unsigned int
> > 
> > Impact: Cleanup; fix inappropriate macro use
> > 
> > ISA addresses on x86 are mapped 1:1 with the physical address space.
> > Since the ISA address space is only 24 bits (32 for VLB or LPC) it
> > will always fit in an unsigned int, and at least in the aha1542 driver
> > using a wider type would cause an undesirable promotion.  Hence
> > explicitly cast the ISA bus addresses to unsigned int.
> > 
> > Signed-off-by: H. Peter Anvin <hpa@linux.intel.com>
> > Cc: James Bottomley <James.Bottomley@hansenpartnership.com>
> > 
> > 
> > ---
> >  arch/x86/include/asm/io.h |   11 ++++++++---
> >  1 files changed, 8 insertions(+), 3 deletions(-)
> > 
> > diff --git a/arch/x86/include/asm/io.h b/arch/x86/include/asm/io.h
> > index 4f8e820..683d0b4 100644
> > --- a/arch/x86/include/asm/io.h
> > +++ b/arch/x86/include/asm/io.h
> > @@ -124,10 +124,15 @@ static inline void *phys_to_virt(phys_addr_t address)
> >  
> >  /*
> >   * ISA I/O bus memory addresses are 1:1 with the physical address.
> > + * However, we truncate the address to unsigned int to avoid undesirable
> > + * promitions in legacy drivers.
> >   */
> > -#define isa_virt_to_bus (unsigned long)virt_to_phys
> > -#define isa_page_to_bus page_to_phys
> > -#define isa_bus_to_virt phys_to_virt
> > +static inline unsigned int isa_virt_to_bus(volatile void *address)
> 
> The rest looks fine, but why is the address volatile?  The macro below
> never actually dereferences it.

That is because the volatile keyword here broadens the permitted 
_input_ type.

	Ingo

^ permalink raw reply	[flat|nested] 3+ messages in thread

end of thread, other threads:[~2009-02-17 21:55 UTC | newest]

Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <tip-a7eb518998529c08cc53fef17756d9fe433b0c23@kernel.org>
2009-02-17 21:24 ` [tip:x86/headers] x86: truncate ISA addresses to unsigned int James Bottomley
2009-02-17 21:25   ` H. Peter Anvin
2009-02-17 21:55   ` Ingo Molnar

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.