Re: bswap is slow on SPARC

> For most NetBSD architectures, GCC's __builtin_bswapX inserts fast
> inline MD code.  For SPARC, SPARC64, and VAX, it falls back to a
> function call.

> When encryption algorithms expect to be able to encode little-endian
> integers in a tight loop, that adds up to being *slow*.

Um, maybe I'm just revealing my ignorance, but wouldn't a reasonable
response to this be "don't use encryption that insists on using
little-endian for integers when you're on a SPARC"?  (For that matter,
why would encryption care about integer endianness to begin with?)

Not that speeding up byte-swapping would be a bad thing, but it seems
to me it isn't the only approach here.

/~\ The ASCII				  Mouse
\ / Ribbon Campaign
 X  Against HTML		[email protected]
/ \ Email!	     7D C8 61 52 5D E7 2D 39  4E F1 31 3E E8 B3 27 4B