[Bug ada/109817] New: -O2 and ./gnat.dg/sync_iface_call_pkg2.adb don't mix

[Bug ada/109817] New: -O2 and ./gnat.dg/sync_iface_call_pkg2.adb don't mix

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

            Bug ID: 109817
           Summary: -O2 and ./gnat.dg/sync_iface_call_pkg2.adb don't mix
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: ada
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
                CC: dkm at gcc dot gnu.org
  Target Milestone: ---

$ ~/gcc/results.20230511.others/bin/gcc -c -O2
./gnat.dg/sync_iface_call_pkg2.adb
during IPA pass: icf
+===========================GNAT BUG DETECTED==============================+
| 14.0.0 20230511 (experimental) (x86_64-pc-linux-gnu) GCC error:          |
| in former_thunk_p, at cgraph.cc:4229                                     |
| Error detected around
/home/dcb38/gcc/results.20230511.others/lib/gcc/x86_64-pc-linux-gnu/14.0.0/adainclude/s-atacco.ads:39:16|
| Compiling ./gnat.dg/sync_iface_call_pkg2.adb                             |
| Please submit a bug report; see https://gcc.gnu.org/bugs/ .   

-O1 seems fine:

$ ~/gcc/results.20230511.others/bin/gcc -c -O1
./gnat.dg/sync_iface_call_pkg2.adb
$

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
            Version|unknown                     |14.0
          Component|ada                         |ipa
                 CC|                            |ebotcazou at gcc dot gnu.org,
                   |                            |marxin at gcc dot gnu.org
   Last reconfirmed|                            |2023-05-13
            Summary|-O2 and                     |internal error in ICF pass
                   |./gnat.dg/sync_iface_call_p |on Ada interfaces
                   |kg2.adb don't mix           |

--- Comment #1 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
Yes, the ICF pass is not run at -O2 and -fno-ipa-icf is a workaround.

[Bug ipa/109817] [14 Regression] internal error in ICF pass on Ada interfaces

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|internal error in ICF pass  |[14 Regression] internal
                   |on Ada interfaces           |error in ICF pass on Ada
                   |                            |interfaces
   Target Milestone|---                         |14.0
                 CC|                            |pinskia at gcc dot gnu.org
           Keywords|                            |ice-on-valid-code

[Bug ipa/109817] [14 Regression] internal error in ICF pass on Ada interfaces

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-checking

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
  gcc_checking_assert (i->fixed_offset || i->virtual_offset_p
                       || i->indirect_offset);

That assert was added in GCC 11: r11-4329-g67f3791f7d1332

So this might be an older regression but it is definitely a regression.

[Bug ipa/109817] [14 regression] internal error in ICF pass on Ada interfaces

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
            Summary|[14 Regression] internal    |[14 regression] internal
                   |error in ICF pass on Ada    |error in ICF pass on Ada
                   |interfaces                  |interfaces
           Assignee|unassigned at gcc dot gnu.org      |ebotcazou at gcc dot gnu.org

--- Comment #3 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
I'll have a look.

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[14 regression] internal    |internal error in ICF pass
                   |error in ICF pass on Ada    |on Ada interfaces
                   |interfaces                  |

--- Comment #4 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
The Ada front-end has generated null thunks, i.e. mere forwarders, for the last
couple of decades.  Nothing wrong with that, but this is clearly inefficient.

Not a (recent) regression in any case.

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[hubicka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

Jan Hubicka <hubicka at gcc dot gnu.org> changed:

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

--- Comment #5 from Jan Hubicka <hubicka at gcc dot gnu.org> ---
That check was added to verify that we do not lose the thunk annotations.  Now
when datastructure is stable, i think we can simply drop it, if that makes Ada
to work.

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

--- Comment #6 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
It's not visible in release builds and testing shows that it's a very rare
situation in practice, so no real need IMO.

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|14.0                        |14.2

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 14.1 is being released, retargeting bugs to GCC 14.2.

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

--- Comment #8 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Marc Poulhi?s <dkm@gcc.gnu.org>:

https://gcc.gnu.org/g:55ceb87a72fee45a9a3e547a4e688f31d376a95a

commit r15-1322-g55ceb87a72fee45a9a3e547a4e688f31d376a95a
Author: Eric Botcazou <ebotcazou@adacore.com>
Date:   Mon Apr 29 09:15:13 2024 +0200

    ada: Do not create null GCC thunks

    This prevents Gigi from creating null GCC thunks, i.e. thunks that have all
    their internal parameters set to zero, replacing them with aliases.  They
    can arise in degenerate cases and null thunks would trip on an assertion in
    former_thunk_p when they are later optimized.

    gcc/ada/

            PR ada/109817
            * gcc-interface/trans.cc (maybe_make_gnu_thunk): Create an alias
            instead of a null thunk.

[Bug ipa/109817] internal error in ICF pass on Ada interfaces

[ebotcazou at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109817

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|14.2                        |15.0
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #9 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
Fixed on the mainline.