[Bug debug/105041] New: '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[Bug debug/105041] New: '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

            Bug ID: 105041
           Summary: '-fcompare-debug' failure w/ -mcpu=power6 -O2
                    -fharden-compares -frename-registers
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Keywords: compare-debug-failure
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: asolokha at gmx dot com
  Target Milestone: ---
            Target: powerpc-*-linux-gnu

Created attachment 52678
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52678&action=edit
gkd diff

gcc 12.0.1 20220320 snapshot (g:6f30c4cd38b8385d62358065d795df590e752d6e) fails
-fcompare-debug check when compiling the following testcase w/ -mcpu=power6 -O2
-fcompare-debug -fharden-compares -frename-registers:

double m;
int n;

unsigned int
foo (unsigned int x, int y)
{
  long long int a = y, b = !a;
  int c = 0;

  if (b != x)
    while ((int) m == a)
      {
        c = a;
        a = 0;
        x = 0;
      }

  n = b = y;

  return x + c;
}

% powerpc-e300c3-linux-gnu-gcc-12.0.1 -mcpu=power6 -O2 -fcompare-debug
-fharden-compares -frename-registers -c ck9whkva.c
powerpc-e300c3-linux-gnu-gcc-12.0.1: error: ck9whkva.c: '-fcompare-debug'
failure

gkd diff attached.

I suppose it is really a duplicate of PR104771.

[Bug debug/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

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

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

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
Or the one Alex just fixed for -fharden-compares.

[Bug debug/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2022-03-29
                 CC|                            |bergner at gcc dot gnu.org
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #2 from Peter Bergner <bergner at gcc dot gnu.org> ---
Confirmed.  This only seems to fail for me on BE and using -m32.  It passes
when compiled with -mcpu={7,8,9} and fails using -mcpu={4,5,6,10}.

[Bug debug/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #3 from Peter Bergner <bergner at gcc dot gnu.org> ---
The first real change I see does occur in the rnreg dump file.

[Bug debug/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

Peter Bergner <bergner at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4

--- Comment #4 from Peter Bergner <bergner at gcc dot gnu.org> ---
Given this only affects 32-bit BE compiles that use the uncommonly used option
-frename-registers, I'm dropping the importance to P4.

[Bug debug/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[segher at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

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

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

--- Comment #5 from Segher Boessenkool <segher at gcc dot gnu.org> ---
It is not so uncommon at all: it is enabled by -funroll-loops, which is used
quite often.

[Bug debug/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[jskumari at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #6 from Surya Kumari Jangala <jskumari at gcc dot gnu.org> ---
I will be debugging the issue to figure the root cause.

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Segher Boessenkool <segher@gcc.gnu.org>:

https://gcc.gnu.org/g:3e16b4359e86b36676ed01219e6deafa95f3c16b

commit r13-1092-g3e16b4359e86b36676ed01219e6deafa95f3c16b
Author: Surya Kumari Jangala <jskumari@linux.vnet.ibm.com>
Date:   Fri Jun 10 19:52:57 2022 +0530

    regrename: Fix -fcompare-debug issue in check_new_reg_p [PR105041]

    In check_new_reg_p, the nregs of a du chain is computed by obtaining the
    MODE of the first element in the chain, and then calling
    hard_regno_nregs() with the MODE. But the first element of the chain can
    be a DEBUG_INSN whose mode need not be the same as the rest of the
    elements in the du chain. This was resulting in fcompare-debug failure
    as check_new_reg_p was returning a different result with -g for the same
    candidate register. We can instead obtain nregs from the du chain
    itself.

    2022-06-10  Surya Kumari Jangala  <jskumari@linux.ibm.com>

    gcc/
            PR rtl-optimization/105041
            * regrename.cc (check_new_reg_p): Use nregs value from du chain.

    gcc/testsuite/
            PR rtl-optimization/105041
            * gcc.target/powerpc/pr105041.c: New test.

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[jskumari at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

Surya Kumari Jangala <jskumari at gcc dot gnu.org> changed:

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

--- Comment #8 from Surya Kumari Jangala <jskumari at gcc dot gnu.org> ---
Fixed.

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[jskumari at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

Surya Kumari Jangala <jskumari at gcc dot gnu.org> changed:

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

--- Comment #9 from Surya Kumari Jangala <jskumari at gcc dot gnu.org> ---
Reopening the bug as we need to backport the fix.

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Surya Kumari Jangala
<jskumari@gcc.gnu.org>:

https://gcc.gnu.org/g:0380d008b1474373852fd2fc921886491304f854

commit r12-8568-g0380d008b1474373852fd2fc921886491304f854
Author: Surya Kumari Jangala <jskumari@linux.vnet.ibm.com>
Date:   Fri Jun 10 19:52:57 2022 +0530

    regrename: Fix -fcompare-debug issue in check_new_reg_p [PR105041]

    In check_new_reg_p, the nregs of a du chain is computed by obtaining the
    MODE of the first element in the chain, and then calling
    hard_regno_nregs() with the MODE. But the first element of the chain can
    be a DEBUG_INSN whose mode need not be the same as the rest of the
    elements in the du chain. This was resulting in fcompare-debug failure
    as check_new_reg_p was returning a different result with -g for the same
    candidate register. We can instead obtain nregs from the du chain
    itself.

    2022-06-10  Surya Kumari Jangala  <jskumari@linux.ibm.com>

    gcc/
            PR rtl-optimization/105041
            * regrename.cc (check_new_reg_p): Use nregs value from du chain.

    gcc/testsuite/
            PR rtl-optimization/105041
            * gcc.target/powerpc/pr105041.c: New test.

    (cherry picked from commit 3e16b4359e86b36676ed01219e6deafa95f3c16b)

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Surya Kumari Jangala
<jskumari@gcc.gnu.org>:

https://gcc.gnu.org/g:8522fab3f900d6fe0cc43be52fdd850f5c9c44db

commit r11-10156-g8522fab3f900d6fe0cc43be52fdd850f5c9c44db
Author: Surya Kumari Jangala <jskumari@linux.vnet.ibm.com>
Date:   Fri Jun 10 19:52:57 2022 +0530

    regrename: Fix -fcompare-debug issue in check_new_reg_p [PR105041]

    In check_new_reg_p, the nregs of a du chain is computed by obtaining the
    MODE of the first element in the chain, and then calling
    hard_regno_nregs() with the MODE. But the first element of the chain can
    be a DEBUG_INSN whose mode need not be the same as the rest of the
    elements in the du chain. This was resulting in fcompare-debug failure
    as check_new_reg_p was returning a different result with -g for the same
    candidate register. We can instead obtain nregs from the du chain
    itself.

    2022-06-10  Surya Kumari Jangala  <jskumari@linux.ibm.com>

    gcc/
            PR rtl-optimization/105041
            * regrename.c (check_new_reg_p): Use nregs value from du chain.

    gcc/testsuite/
            PR rtl-optimization/105041
            * gcc.target/powerpc/pr105041.c: New test.

    (cherry picked from commit 3e16b4359e86b36676ed01219e6deafa95f3c16b)

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Surya Kumari Jangala
<jskumari@gcc.gnu.org>:

https://gcc.gnu.org/g:2753c7d0cd95a3db9d97c5ee674917ad11de5246

commit r10-10907-g2753c7d0cd95a3db9d97c5ee674917ad11de5246
Author: Surya Kumari Jangala <jskumari@linux.vnet.ibm.com>
Date:   Fri Jun 10 19:52:57 2022 +0530

    regrename: Fix -fcompare-debug issue in check_new_reg_p [PR105041]

    In check_new_reg_p, the nregs of a du chain is computed by obtaining the
    MODE of the first element in the chain, and then calling
    hard_regno_nregs() with the MODE. But the first element of the chain can
    be a DEBUG_INSN whose mode need not be the same as the rest of the
    elements in the du chain. This was resulting in fcompare-debug failure
    as check_new_reg_p was returning a different result with -g for the same
    candidate register. We can instead obtain nregs from the du chain
    itself.

    2022-06-10  Surya Kumari Jangala  <jskumari@linux.ibm.com>

    gcc/
            PR rtl-optimization/105041
            * regrename.c (check_new_reg_p): Use nregs value from du chain.

    gcc/testsuite/
            PR rtl-optimization/105041
            * gcc.target/powerpc/pr105041.c: New test.

    (cherry picked from commit 3e16b4359e86b36676ed01219e6deafa95f3c16b)

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

--- Comment #13 from Arseny Solokha <asolokha at gmx dot com> ---
After the fix was backported to all open branches, should this PR be closed?

[Bug rtl-optimization/105041] '-fcompare-debug' failure w/ -mcpu=power6 -O2 -fharden-compares -frename-registers

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105041

Peter Bergner <bergner at gcc dot gnu.org> changed:

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

--- Comment #14 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to Arseny Solokha from comment #13)
> After the fix was backported to all open branches, should this PR be closed?

Yes.  Marking as fixed.  Thanks Surya!