[RFC] arm: use built-in byte swap function

[Kim Phillips]

On Wed, 6 Feb 2013 09:02:04 +0000
"Woodhouse, David" <david.woodhouse at intel.com> wrote:

> On Tue, 2013-02-05 at 21:04 -0600, Kim Phillips wrote:
> > gcc -Os emits calls to __bswapsi2 on those platforms to save space
> > because they don't have the single rev byte swap instruction.
> 
> Is that the right thing for GCC to do in that situation?

if it saves space, why wouldn't it be?

"Many of these functions are only optimized in certain cases; if they
are not optimized in a particular case, a call to the library
function is emitted." [1]

I see "(arm_arch6 || !optimize_size)" in gcc's define_expand
"bswapsi2" source, so GCC considers size optimization as a
legitimate one of those cases.

> If so, perhaps we should be *providing* __bswap[sd]i2 functions for it
> to use?

either that, or link with libgcc - why does arch/arm64 do this and
arch/arm not?  It's not obvious from git log.

> If not, perhaps there should be a PR filed?
> 
> Or is our use case justifiably different to the general case of '-Os'?
> If so, why?

shouldn't be - a patch, such as this, that claims to reduce code
size, and that only turns on the new built-in when
CC_OPTIMIZE_FOR_SIZE is off, is generally not good :)

OTOH, the target here is armv6+ performance - not armv4,5 code
density - the OPTIMIZE_FOR_SIZE protection prevents armv4,5 build
breakage.

Kim

[1] http://gcc.gnu.org/onlinedocs/gcc/Other-Builtins.html#Other-Builtins