[Bug rtl-optimization/23684] New: Combine stores for non strict alignment targets

[Bug rtl-optimization/23684] New: Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

Take the following code:
#define OSSwapHostToBigInt32
void foo(char *input) {
  input[0] = 'H';
  input[1] = 'e';
  input[2] = 'l';
  input[3] = 'l';
  input[4] = 'o';
  input[5] = ' ';
  input[6] = 'w';
  input[7] = 'o';
  input[8] = 'r';
  input[9] = 'l';
  input[10] = 'd';
  input[11] = '\0';
}

void bar(char *input) {
  ((unsigned int *)input)[0] = OSSwapHostToBigInt32('Hell');
  ((unsigned int *)input)[1] = OSSwapHostToBigInt32('o wo');
  ((unsigned int *)input)[2] = OSSwapHostToBigInt32('rld\0');
}

They should generate the same asm but currently don't.  Note change OSSwapHostToBigInt32 to correct 
for little endian.

-- 
           Summary: Combine stores for non strict alignment targets
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Keywords: missed-optimization
          Severity: enhancement
          Priority: P2
         Component: rtl-optimization
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: pinskia at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
 GCC target triplet|                            |powerpc-*-*


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[amodra at bigpond dot net dot au]

------- Additional Comments From amodra at bigpond dot net dot au  2005-09-07 01:42 -------
Also even when -mstrict-align if using
typedef char align_char __attribute__ ((aligned (4)));
void foo (align_char *input) ...


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-09-15 20:14 -------
Confirmed.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2005-09-15 20:14:53
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |justin at fathomdb dot com

--- Comment #13 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 50246 has been marked as a duplicate of this bug. ***

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[msharov at users dot sourceforge.net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

--- Comment #12 from msharov at users dot sourceforge.net ---
I'd like to add that this is not some corner case; this is a very common issue.
In my own projects, the compiler's inability to combine stores is the single
largest reason for using inline assembly and raw casts. Pretty much every time
I have an object 8 or 16 bytes in size, I end up writing a zeroing ctor, copy
ctor, and operator= that use full-object memory access. That's cast to uint64_t
for 8 bytes, and movups/movaps for 16 bytes. It also shows up when writing raw
protocol data, such as X calls, where it is very common to write several
constants in succession. The last time I checked, forcing whole-object moves in
these cases results in projectwide code size reduction ~10%. Unfortunately, it
also causes a variety of aliasing pessimizations, so I also have to test
including or not including each of the above functions to get the smallest code
size. I would be a very big deal if the optimizer could do this.

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
This is related to PR22141 .  But it caused performance regressions on ppc,
which is why it hasn't been installed.  If there are ideas how to tune that, I
can try to ressurrect the patch.

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |msharov at users dot sourceforge.n
                   |                            |et

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
*** Bug 57302 has been marked as a duplicate of this bug. ***

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

Richard Guenther <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |plasmahh at gmx dot net

--- Comment #9 from Richard Guenther <rguenth at gcc dot gnu.org> 2012-07-18 09:35:42 UTC ---
*** Bug 54011 has been marked as a duplicate of this bug. ***

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

------- Comment #8 from pinskia at gcc dot gnu dot org  2007-04-03 17:54 -------
*** Bug 19726 has been marked as a duplicate of this bug. ***


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |yuri at tsoft dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

------- Comment #7 from pinskia at gcc dot gnu dot org  2007-04-03 17:53 -------
*** Bug 14613 has been marked as a duplicate of this bug. ***


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |cesarb at cesarb dot net


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[acahalan at gmail dot com]

------- Comment #6 from acahalan at gmail dot com  2006-11-29 05:11 -------
This hits it too. (example is PowerPC)

///////////////////
#include <string.h>

char *foo(char *buf)
{
 short temp;
 int temp1;

 *buf++=42;

 temp = 0xfeed;
 memcpy(buf, &temp, sizeof(temp));
 buf+=sizeof(temp);

 temp1 = 0x12345678;
 memcpy(buf, &temp1, sizeof(temp1));
 buf+=sizeof(temp1);

 temp1 = 0x12345678;
 memcpy(buf, &temp1, sizeof(temp1));
 buf+=sizeof(temp1);

 *buf++=42;

 return buf;
}
///////////////
foo:
        stwu 1,-32(1)
        lis 9,0x1234
        li 11,42
        li 0,-275
        ori 9,9,22136
        stb 11,11(3)
        stb 11,0(3)
        sth 0,1(3)
        stw 9,7(3)
        stw 9,3(3)
        addi 3,3,12
        addi 1,1,32
        blr


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

------- Comment #5 from pinskia at gcc dot gnu dot org  2006-11-29 05:05 -------
(In reply to comment #4)
> Target powerpc-*-* is wrong. See bug 29158 (my dupe) for the i386 assembly
> produced by a slightly more recent compiler. It's x86_64 too.

I just used the target as powerpc as I was also showing asm code.

> It's not just (char*) either. It's (short*) and even struct members.

The problems are all the same, really.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[acahalan at gmail dot com]

------- Comment #4 from acahalan at gmail dot com  2006-09-21 05:12 -------
Target powerpc-*-* is wrong. See bug 29158 (my dupe) for the i386 assembly
produced by a slightly more recent compiler. It's x86_64 too.

It's not just (char*) either. It's (short*) and even struct members.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684

[Bug rtl-optimization/23684] Combine stores for non strict alignment targets

[pinskia at gcc dot gnu dot org]

------- Comment #3 from pinskia at gcc dot gnu dot org  2006-09-21 04:39 -------
*** Bug 29158 has been marked as a duplicate of this bug. ***


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |acahalan at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=23684