On Friday 06 February 2009 16:53:03 Peter Tyser wrote:
> On Fri, 2009-02-06 at 16:40 -0500, Mike Frysinger wrote:
> > On Friday 06 February 2009 16:22:01 Peter Tyser wrote:
> > > +uint16_t bswap_16(uint16_t __x)
> > > +{
> > > + return (__x >> 8) | (__x << 8);
> > > +}
> > > +
> > > +uint32_t bswap_32(uint32_t __x)
> > > +{
> > > + return (bswap_16(__x & 0xffff) << 16) | (bswap_16(__x >> 16));
> > > +}
> > > +
> > > +uint64_t bswap_64(uint64_t __x)
> > > +{
> > > + return (((uint64_t) bswap_32(__x & 0xffffffff)) << 32) |
> > > + (bswap_32(__x >> 32));
> > > +}
> >
> > we already have random duplicate copies of these floating around, and not
> > related to Windows. these funcs are really only reliable on Linux. like
> > the attached i'm using in my own tree.
>
> I don't follow. Are you suggesting I include a generic version in a
> header similar to the uswap.h you mentioned? Or are you suggesting the
> current implementation is prone to break? Or both?i'm saying creating yet another duplicate copy of the swap macros is wasteful. create a common version already and use it. -mike
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ U-Boot mailing list [email protected] http://lists.denx.de/mailman/listinfo/u-boot
