Thank you for the answer.
Yet, I agree this example has no real meaning, but my question is,
as v is already an address (global variable), why GCC prefers to
copy the value pointed by v into the stack and give the new pointer
rather than just give v directly?
Is it to ensure the correctness of the code or just because nobody
have met this strange configuration and so no optimisation was
developed?
________________________________________
From: Segher Boessenkool [segher@xxxxxxxxxxxxxxxxxxx]
Sent: Tuesday, January 14, 2020 1:33 PM
To: RECOULES Frederic
Cc: gcc-help@xxxxxxxxxxx
Subject: Re: Interactions between function inlining and inline assembly
Hi!
On Tue, Jan 14, 2020 at 11:54:00AM +0000, RECOULES Frederic wrote:
> Let use the following (dumb) snippet as an example:
>
> static int foo (int x)
> {
> int r;
> __asm__ ("movl %1, %0" : "=r" (r) : "m" (x));
> return r;
> }
>
> int v;
>
> int main (int argc, char *argv[])
> {
> return foo(v);
> }
>
> Which when compiled with GCC 9.2 (as long as I know, version does not matter)
> -m32 -O3 produce the following code:
>
> main:
> subl $16, %esp
> movl v, %eax
> movl %eax, 12(%esp)
> #APP
> movl 12(%esp), %eax
> #NO_APP
> addl $16, %esp
> ret
>
>
> So, even if the function call is inlined, it continues to pass the argument v
> by the stack while, in fact, I would have expected it to forward the
> address like this:
>
> main:
> #APP
> movl v, %eax
> #NO_APP
> ret
But you said that assembler argument has to be in memory! Try this:
static int foo (int x)
{
int r;
__asm__ ("movl %1, %0" : "=r" (r) : "rm" (x));
return r;
}
You could also use "rmi" (which can be written "g"), here.
Segher
