[Bug tree-optimization/98953] New: Failure to optimize two reads from adjacent addresses into one

[Bug tree-optimization/98953] New: Failure to optimize two reads from adjacent addresses into one

[gabravier at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

            Bug ID: 98953
           Summary: Failure to optimize two reads from adjacent addresses
                    into one
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gabravier at gmail dot com
  Target Milestone: ---

int f(uint8_t *read, size_t abs)
{
    return read[abs] + 0x100 * read[abs + 1];;
}

This can be optimized to (ignoring aliasing) `return *(uint16_t *)(read +
abs)`. This transformation is done by LLVM, but not by GCC.

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2021-02-04
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Confirmed.  The bswap would be the place that detects similar cases but I think
it doesn't consider this case because the number is "incomplete" and because
the pass doesnt' consider PLUS_EXPR or MULT_EXPR AFAIR.  Also it doesn't
handle non-obviously same bases gracefully.

The following works:

#include <stdint.h>
#include <stdlib.h>
int f(uint8_t *read, size_t abs)
{
    return read[0] + (read[1] << 8);
}

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Severity|normal                      |enhancement
   Last reconfirmed|2021-02-04 00:00:00         |2021-7-28

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one

[roger at nextmovesoftware dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

Roger Sayle <roger at nextmovesoftware dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |roger at nextmovesoftware dot com
             Status|NEW                         |ASSIGNED
                 CC|                            |roger at nextmovesoftware dot com

--- Comment #2 from Roger Sayle <roger at nextmovesoftware dot com> ---
Patch proposed
https://gcc.gnu.org/pipermail/gcc-patches/2021-November/585117.html

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Roger Sayle <sayle@gcc.gnu.org>:

https://gcc.gnu.org/g:a944b5dec3adb28ed199234d2116145ca9010d6a

commit r12-5453-ga944b5dec3adb28ed199234d2116145ca9010d6a
Author: Roger Sayle <roger@nextmovesoftware.com>
Date:   Mon Nov 22 18:15:36 2021 +0000

    tree-optimization/103345: Improved load merging.

    This patch implements PR tree-optimization/103345 to merge adjacent
    loads when combined with addition or bitwise xor.  The current code
    in gimple-ssa-store-merging.c's find_bswap_or_nop alreay handles ior,
    so that all that's required is to treat PLUS_EXPR and BIT_XOR_EXPR in
    the same way at BIT_IOR_EXPR.  Many thanks to Andrew Pinski for
    pointing out that this also resolves PR target/98953.

    2021-11-22  Roger Sayle  <roger@nextmovesoftware.com>

    gcc/ChangeLog
            PR tree-optimization/98953
            PR tree-optimization/103345
            * gimple-ssa-store-merging.c (find_bswap_or_nop_1): Handle
            BIT_XOR_EXPR and PLUS_EXPR the same as BIT_IOR_EXPR.
            (pass_optimize_bswap::execute): Likewise.

    gcc/testsuite/ChangeLog
            PR tree-optimization/98953
            PR tree-optimization/103345
            * gcc.dg/tree-ssa/pr98953.c: New test case.
            * gcc.dg/tree-ssa/pr103345.c: New test case.

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one

[roger at nextmovesoftware dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

Roger Sayle <roger at nextmovesoftware dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |NEW
           Assignee|roger at nextmovesoftware dot com  |unassigned at gcc dot gnu.org

--- Comment #4 from Roger Sayle <roger at nextmovesoftware dot com> ---
The MULT_EXPR and PLUS_EXPR aspects of this PR are now resolved (i.e. the case
in comment #1), but unfortunately the abs-based indexing used in the original
report still causes problems.  The bswap pass doesn't yet handle memory
accesses of the form read[abs]/read[abs+1] (but does handle read[0]/read[1]).

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 102391 has been marked as a duplicate of this bug. ***

[Bug tree-optimization/98953] Failure to optimize two reads from adjacent addresses into one due to having an offset (index)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98953

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |llvm at rifkin dot dev

--- Comment #6 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
*** Bug 113392 has been marked as a duplicate of this bug. ***