Re: [PATCH 14/20] parisc: use the new byteorder headers

[Harvey Harrison <harvey.harrison@gmail.com>]

On Fri, 2008-08-01 at 10:06 -0600, Grant Grundler wrote:
> On Thu, Jul 17, 2008 at 05:09:43PM -0700, Harvey Harrison wrote:
> > Signed-off-by: Harvey Harrison <harvey.harrison@gmail.com>
> > ---
> >  include/asm-parisc/byteorder.h |   37 ++++++++++++-------------------------
> >  1 files changed, 12 insertions(+), 25 deletions(-)
> > 
> > diff --git a/include/asm-parisc/byteorder.h b/include/asm-parisc/byteorder.h
> > index db14831..5e56b1b 100644
> > --- a/include/asm-parisc/byteorder.h
> > +++ b/include/asm-parisc/byteorder.h
> > @@ -4,9 +4,10 @@
> >  #include <asm/types.h>
> >  #include <linux/compiler.h>
> >  
> > -#ifdef __GNUC__
> > +#define __LITTLE_ENDIAN
> 
> Sorry, this looks wrong. parisc is big endian.

Sorry, it is wrong.  It of course should be __BIG_ENDIAN

> > +#define __SWAB_64_THRU_32__
> 
> Should this only be defined if BITS_PER_LONG is <=32  ?
> 

The new byteorder headers only uses this if no __arch_swab64 is defined.

Further down in the parisc version, an __arch_swab64 is defined in the
BITS_PER_LONG >=32 case, making this moot.

>From the new linux/swab.h:

static inline __attribute_const__ __u64 ___swab64(__u64 val)
{
#ifdef __arch_swab64
	return __arch_swab64(val);
#elif defined(__arch_swab64p)
	return __arch_swab64p(&val);
#elif defined(__SWAB_64_THRU_32__)
	__u32 h = val >> 32;
	__u32 l = val & ((1ULL << 32) - 1);
	return (((__u64)___swab32(l)) << 32) | ((__u64)(___swab32(h)));
#else
	return __const_swab64(val);
#endif
}

Other than the endianness, any other concerns?

Harvey