[Bug tree-optimization/101755] New: [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[Bug tree-optimization/101755] New: [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

            Bug ID: 101755
           Summary: [12 regression] gcc.target/arm/reg_equal_test.c fails
                    on arm since r12-2637
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: clyon at gcc dot gnu.org
  Target Milestone: ---

Since r12-2637, I've noticed a regression on arm:
FAIL: gcc.target/arm/reg_equal_test.c scan-rtl-dump expand "expr_list:REG_EQUAL
\\(const_int -942519458"

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[clyon at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

Christophe Lyon <clyon at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |arm
   Target Milestone|---                         |12.0
                 CC|                            |amacleod at redhat dot com

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[amacleod at redhat dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

--- Comment #1 from Andrew Macleod <amacleod at redhat dot com> ---
What is that suppose to test?

The range ecosystem has gotten smarter and we basically fold that function away
to return 0 now.  between EVRP, VRP and threading, you'd have to turn off a lot
of optimizations to return to the original state.. I can't find a sequence.

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
  d = a | -2;
  b = (d == 0) ? c : (c % d);

d can never be 0, VRP figures this out now.
so we end up with:
d = a | -2;
b = 0xc7d24b5e % d;
if (b != 0xc7d24b5e)
  abort();

and b cannot be 0xc7d24b5e because d is negative :).

So VRP is optimizing this all the way out.

The test was trying to make sure the value 0xc7d24b5e was emitted as a
REG_EQUAL note when splitting up the constant (for the forming) in RTL.

https://gcc.gnu.org/pipermail/gcc-patches/2015-June/422430.html
https://gcc.gnu.org/pipermail/gcc-patches/2015-June/422432.html

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |testsuite-fail

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
Maybe add -fno-thread-jumps?

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[rearnsha at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2022-04-12
     Ever confirmed|0                           |1
           Assignee|unassigned at gcc dot gnu.org      |rearnsha at gcc dot gnu.org

--- Comment #4 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
Mine

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Earnshaw <rearnsha@gcc.gnu.org>:

https://gcc.gnu.org/g:62082d278d1fc9773539244194061ab32556db4d

commit r12-8108-g62082d278d1fc9773539244194061ab32556db4d
Author: Richard Earnshaw <rearnsha@arm.com>
Date:   Tue Apr 12 13:35:27 2022 +0100

    arm: fix testsuite failure of reg_equal_test.c [PR101755]

    The test failure in PR101755 is due to the gimple optimizers getting
    smarter.  But really we are just testing that RTL expansion is doing
    the right thing and annotating a constant accordingly.  So rework the
    test to use GIMPLE input and simplify the code entirely.  Also, this
    test only ever worked on architecture versions with the MOVW/MOVT
    instructions, so check for this before running.

    gcc/testsuite/ChangeLog:

            PR target/101755
            * gcc.target/arm/reg_equal_test.c: Convert to gimple test. 
Restrict
            to architectures with MOVW/MOVT.

[Bug tree-optimization/101755] [12 regression] gcc.target/arm/reg_equal_test.c fails on arm since r12-2637

[rearnsha at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101755

Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #6 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
Fixed.