Re: Endianness macros capitalization

[Christian Franke]

Pavel Roskin wrote:
> On Tue, 2008-07-08 at 20:04 +0200, Christian Franke wrote:
> 
> > With old gcc versions without the "rol" optimization, even the 16
> > bit swap should be a function:
> > 
> 
> Or better yet, an asm statement.
> 
> We should consider optimized assembly vs function call.  Even the
> 32-bit swap could be shorter:
> 
> a:   86 c4                   xchg   %al,%ah
> c:   c1 c0 10                rol    $0x10,%eax
> f:   86 c4                   xchg   %al,%ah
> 11:
> 
> That's 7 bytes!  ...

But the function call in the 32-bit case requires only 5 bytes :-)
(The 16-bit case requires more due to short->int propagation)

Overall size from inline asm would only be smaller if there is any
benefit from additional optimizations.

Here is a possibly working draft for include/grub/i386/types.h (requires
a #ifdef in include/grub/types.h):

#define grub_swap_bytes32(x) \
({ \
   grub_uint32_t _x = (x), _y;
   asm ( \
        "xchg %%al,%%ah\n" \
        "roll $0x10,%%eax\n" \
        "xchg %%al,%%ah\n" \
        : "=a"(_y) : "0"(_x) \
    ); \
    _y; \
})

Result with the test script from my last mail:

Debian gcc 4.1.2-7:
inline (portable)=357, inline (asm)=126, function=104

Cygwin gcc 3.4.4:
inline (portable)=340, inline (asm)=124, function=96

Function call is still better. The only candidate for inline is probably
 grub_swap_bytes16().

> .... And if written properly, it could work with any of
> the registers that allow access to the lower two bytes (%eax, %ebx,
> %ecx and %edx), thus giving more flexibility to the optimizer.
> 

This would require support to access the Rl and Rh parts of eRx for each
R in [a-d]. Something like:

   asm (
        "xchg %0:l,%0:h\n"
        "roll $0x10,%0\n"
        "xchg %0:l,%0:h\n"
        : "=r"(_y) : "0"(_x) \
    );

Do gcc or gas provide a syntax to do this?


Christian