Local optimization on i386 ?

[Marc Espie <Marc dot Espie at liafa dot jussieu dot fr>]

Is this normal  behavior on a 386 ?

With egcs-current,
---- gcc -O2 -S a.c

void f(int i)
	{}

void g(int i)
	{
	f(i);
	}

---yields:

	.file	"a.c"
	.version	"01.01"
gcc2_compiled.:
___gnu_compiled_c:
.text
	.align 2,0x90
.globl _f
	.type	_f , @function
_f:
	pushl	%ebp
	movl	%esp, %ebp
	leave
	ret
	.size	_f , . - _f
	.align 2,0x90
.globl _g
	.type	_g , @function
_g:
	pushl	%ebp
	movl	%esp, %ebp
	subl	$8, %esp
	subl	$12, %esp
	pushl	8(%ebp)
	call	_f
	leave
	ret
	.size	_g , . - _g


Note the 
subl  $8, %esp
subl $12, %esp

I don't know enough about 386, but should't this be 
subl $20, %esp ?

I assume I'm missing some obvious reason why there are two separate
subs...

(gcc 2.95.2 replaces one sub with an add, but still doesn't peep-hole
anything).
-- 
	Marc Espie		
|anime, sf, juggling, unicycle, acrobatics, comics...
|AmigaOS, OpenBSD, C++, perl, Icon, PostScript...
| `real programmers don't die, they just get out of beta'