[Bug c/67925] New: docs lie about being unable to inline function call before definition

[Bug c/67925] New: docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

            Bug ID: 67925
           Summary: docs lie about being unable to inline function call
                    before definition
           Product: gcc
           Version: 6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: arkadiusz at drabczyk dot org
  Target Milestone: ---

Here https://gcc.gnu.org/onlinedocs/gcc-5.2.0/gcc/Inline.html it says:

"Some calls cannot be integrated for various reasons (in particular, calls that
precede the function's definition cannot be integrated (...)).". However, since
several releases of GCC this in no longer true as shown in the example:

$ ./gcc --version
gcc (GCC) 6.0.0 20151004 (experimental)
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

$ cat bug.c
#include <stdio.h>
#include <stdlib.h>

inline static void def(void);

int main(void)
{
        def();
        exit(0);
}

static void def(void)
{
        puts("blah blah");
}
$ ./gcc bug.c -O2 -S
$ cat bug.s
        .file   "bug.c"
        .section        .rodata.str1.1,"aMS",@progbits,1
.LC0:
        .string "blah blah"
        .section        .text.startup,"ax",@progbits
        .p2align 4,,15
        .globl  main
        .type   main, @function
main:
.LFB21:
        .cfi_startproc
        movl    $.LC0, %edi
        subq    $8, %rsp
        .cfi_def_cfa_offset 16
        call    puts
        xorl    %edi, %edi
        call    exit
        .cfi_endproc
.LFE21:
        .size   main, .-main
        .ident  "GCC: (GNU) 6.0.0 20151004 (experimental)"
        .section        .note.GNU-stack,"",@progbits

The same happens with -O1 and on older versions of GCC such as 4.9.2. I think
that this outdated piece of documentation should be removed.

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #1 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
BTW, the same incorrect wording is repeated in doc/extend.texi in GCC source
tarball.

[Bug c/67925] docs lie about being unable to inline function call before definition

[miyuki at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

Mikhail Maltsev <miyuki at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |documentation
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-10-11
                 CC|                            |miyuki at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #2 from Mikhail Maltsev <miyuki at gcc dot gnu.org> ---
OMG. IIUC, this limitation was fixed in GCC 3.x (early 2000s):
http://hubicka.blogspot.ru/2014/04/linktime-optimization-in-gcc-1-brief.html

[Bug c/67925] docs lie about being unable to inline function call before definition

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Patches welcome ;)  I think refering to -Winline here might make most sense.

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #4 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
Created attachment 36488
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36488&action=edit
suggested patch

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #5 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
-Winline is mentioned in the next paragraph. The whole sentence I posted in the
first comment is:

"Some calls cannot be integrated for various reasons (in particular, calls that
precede the function's definition cannot be integrated, and neither can
recursive calls within the definition)."

I think it's completely wrong - recursive functions can also be inlined:

$ ./gcc --version
gcc (GCC) 6.0.0 20151004 (experimental)
Copyright (C) 2015 Free Software Foundation, Inc.
This is free software; see the source for copying conditions.  There is NO
warranty; not even for MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.

$ cat bug1.c
#include <stdio.h>
#include <stdlib.h>

inline static int factorial(unsigned int i)
{
   if(i <= 1)
   {
      return 1;
   }
   return i * factorial(i - 1);
}

int main(void)
{
        factorial(12);
        exit(0);
}
$ ./gcc bug1.c -O2 -S
$ cat bug1.s
        .file   "bug1.c"
        .section        .text.startup,"ax",@progbits
        .p2align 4,,15
        .globl  main
        .type   main, @function
main:
.LFB22:
        .cfi_startproc
        subq    $8, %rsp
        .cfi_def_cfa_offset 16
        xorl    %edi, %edi
        call    exit
        .cfi_endproc
.LFE22:
        .size   main, .-main
        .ident  "GCC: (GNU) 6.0.0 20151004 (experimental)"
        .section        .note.GNU-stack,"",@progbits

I attach a suggested patch.

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #6 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
Sorry, this code is wrong of course, value returned by factorial() must be used
to generate an actual code:

$ cat bug1.c
#include <stdio.h>
#include <stdlib.h>

inline static int factorial(unsigned int i)
{
   if(i <= 1)
   {
      return 1;
   }
   return i * factorial(i - 1);
}

int main(void)
{
        printf("%d\n", factorial(12));
        exit(0);
}
$ cat bug1.s
        .file   "bug1.c"
        .section        .rodata.str1.1,"aMS",@progbits,1
.LC0:
        .string "%d\n"
        .section        .text.startup,"ax",@progbits
        .p2align 4,,15
        .globl  main
        .type   main, @function
main:
.LFB22:
        .cfi_startproc
        subq    $8, %rsp
        .cfi_def_cfa_offset 16
        movl    $1, %esi
        movl    $12, %eax
        .p2align 4,,10
        .p2align 3
.L2:
        imull   %eax, %esi
        subl    $1, %eax
        cmpl    $1, %eax
        jne     .L2
        movl    $.LC0, %edi
        xorl    %eax, %eax
        call    printf
        xorl    %edi, %edi
        call    exit
        .cfi_endproc
.LFE22:
        .size   main, .-main
        .ident  "GCC: (GNU) 6.0.0 20151004 (experimental)"
        .section        .note.GNU-stack,"",@progbits

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #36488|0                           |1
        is obsolete|                            |

--- Comment #7 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
Created attachment 36502
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=36502&action=edit
better suggested patch

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #8 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
I attach a new patch with an extra whitespace after dot to keep up with an
existing convention.

[Bug c/67925] docs lie about being unable to inline function call before definition

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

Segher Boessenkool <segher at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |segher at gcc dot gnu.org

--- Comment #9 from Segher Boessenkool <segher at gcc dot gnu.org> ---
Please send the patch to gcc-patches@gcc.gnu.org as described at
<http://gcc.gnu.org/contribute.html>.  Thanks!

[Bug c/67925] docs lie about being unable to inline function call before definition

[arkadiusz at drabczyk dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

--- Comment #10 from Arkadiusz Drabczyk <arkadiusz at drabczyk dot org> ---
Patch sent for review here:
https://gcc.gnu.org/ml/gcc-patches/2015-10/msg01303.html

[Bug c/67925] docs lie about being unable to inline function call before definition

[law at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=67925

Jeffrey A. Law <law at redhat dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
                 CC|                            |law at redhat dot com
         Resolution|---                         |FIXED

--- Comment #11 from Jeffrey A. Law <law at redhat dot com> ---
Fixed on trunk.