[Bug other/101157] New: [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[Bug other/101157] New: [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

            Bug ID: 101157
           Summary: [12 regression] ICE compiling
                    gcc.target/powerpc/stack-limit.c after r12-1702
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: other
          Assignee: unassigned at gcc dot gnu.org
          Reporter: seurer at gcc dot gnu.org
  Target Milestone: ---

g:7232f7c4c2d727431096a7ecfcf4ad4db71dcf2a, r12-1702
make  -k check-gcc RUNTESTFLAGS="powerpc.exp=gcc.target/powerpc/stack-limit.c"
FAIL: gcc.target/powerpc/stack-limit.c (internal compiler error)
FAIL: gcc.target/powerpc/stack-limit.c (test for excess errors)
# of unexpected failures        2

I saw this on a powerpc64 power 9 machine.

commit 7232f7c4c2d727431096a7ecfcf4ad4db71dcf2a (HEAD)
Author: Jason Merrill <jason@redhat.com>
Date:   Sun Jun 13 14:00:12 2021 -0400

Executing on host: /home3/seurer/gcc/git/build/gcc-test/gcc/xgcc
-B/home3/seurer/gcc/git/build/gcc-test/gcc/ exceptions_enabled146254.cc   
-fdiagnostics-plain-output  -S -o exceptions_enabled146254.s    (timeout = 300)
spawn -ignore SIGHUP /home3/seurer/gcc/git/build/gcc-test/gcc/xgcc
-B/home3/seurer/gcc/git/build/gcc-test/gcc/ exceptions_enabled146254.cc
-fdiagnostics-plain-output -S -o exceptions_enabled146254.s
FAIL: gcc.target/powerpc/stack-limit.c (test for excess errors)
Excess errors:
during RTL pass: pro_and_epilogue
/home/seurer/gcc/git/gcc-test/gcc/testsuite/gcc.target/powerpc/stack-limit.c:10:1:
internal compiler error: in df_ref_record, at df-scan.c:2598
0x1059b0df df_ref_record
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:2598
0x1059bee7 df_uses_record
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:2888
0x1059c84f df_insn_refs_collect
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:3212
0x1059f07f df_insn_rescan(rtx_insn*)
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:1087
0x106554e7 emit_insn_after_1
        /home/seurer/gcc/git/gcc-test/gcc/emit-rtl.c:4595
0x106554e7 emit_pattern_after_noloc
        /home/seurer/gcc/git/gcc-test/gcc/emit-rtl.c:4643
0x1050e14f commit_one_edge_insertion(edge_def*)
        /home/seurer/gcc/git/gcc-test/gcc/cfgrtl.c:2051
0x10513d57 commit_edge_insertions()
        /home/seurer/gcc/git/gcc-test/gcc/cfgrtl.c:2103
0x1073e6d7 thread_prologue_and_epilogue_insns()
        /home/seurer/gcc/git/gcc-test/gcc/function.c:6147
0x1073edaf rest_of_handle_thread_prologue_and_epilogue
        /home/seurer/gcc/git/gcc-test/gcc/function.c:6521
0x1073edaf execute
        /home/seurer/gcc/git/gcc-test/gcc/function.c:6597

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|other                       |middle-end
           Keywords|                            |ice-on-valid-code
   Target Milestone|---                         |12.0

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

Martin Liška <marxin at gcc dot gnu.org> changed:

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

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Likely dup of PR101160.

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Martin Liška from comment #1)
> Likely dup of PR101160.

No I doubt it.  Both might be exposed two different latent bugs.  The
df_ref_record hunk looks like it is catching latent bugs really.

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

--- Comment #3 from seurer at gcc dot gnu.org ---
OK, there is another similar error compiling a different test case traced to
the same revision.  And these are occurring on more than just power 10s.

Executing on host: /home/seurer/gcc/git/build/gcc-test/gcc/xgcc
-B/home/seurer/gcc/git/build/gcc-test/gcc/ exceptions_enabled961334.cc   
-fdiagnostics-plain-output  -S -o exceptions_enabled961334.s    (timeout = 300)
spawn -ignore SIGHUP /home/seurer/gcc/git/build/gcc-test/gcc/xgcc
-B/home/seurer/gcc/git/build/gcc-test/gcc/ exceptions_enabled961334.cc
-fdiagnostics-plain-output -S -o exceptions_enabled961334.s
FAIL: gcc.target/powerpc/pr48344-1.c (test for excess errors)
Excess errors:
during RTL pass: no-opt dfinit
/home/seurer/gcc/git/gcc-test/gcc/testsuite/gcc.target/powerpc/pr48344-1.c:8:1:
internal compiler error: in df_ref_record, at df-scan.c:2598
0x1057f137 df_ref_record
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:2598
0x1057feab df_uses_record
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:3033
0x1057fd97 df_uses_record
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:2888
0x1058089f df_insn_refs_collect
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:3212
0x105849bb df_bb_refs_record(int, bool)
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:3339
0x1058520b df_scan_blocks()
        /home/seurer/gcc/git/gcc-test/gcc/df-scan.c:588
0x10563e47 rest_of_handle_df_initialize
        /home/seurer/gcc/git/gcc-test/gcc/df-core.c:715
0x10563e47 execute
        /home/seurer/gcc/git/gcc-test/gcc/df-core.c:747

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
Apparently, the same issue breaks gcc bootstrap on s390x:

[ 4191s] during RTL pass: reload
[ 4191s] ../../gcc/fortran/check.c: In function 'gfc_calculate_transfer_sizes':
[ 4191s] ../../gcc/fortran/check.c:6205: internal compiler error: in
df_ref_record, at df-scan.c:2598
[ 4191s]  6205 | }
[ 4191s]       | 
[ 4191s] 0x12c2f07 df_ref_record
[ 4191s]        ../../gcc/df-scan.c:2598
[ 4191s] 0x12c34f9 df_uses_record
[ 4191s]        ../../gcc/df-scan.c:2857
[ 4191s] 0x12c3c19 df_insn_refs_collect
[ 4191s]        ../../gcc/df-scan.c:3212
[ 4191s] 0x12c6f07 df_bb_refs_record(int, bool)
[ 4191s]        ../../gcc/df-scan.c:3339
[ 4191s] 0x12c7fc7 df_scan_blocks()
[ 4191s]        ../../gcc/df-scan.c:588
[ 4191s] 0x14c4497 do_reload
[ 4191s]        ../../gcc/ira.c:5886
[ 4191s] 0x14c4497 execute
[ 4191s]        ../../gcc/ira.c:6008
[ 4191s] Please submit a full bug report,
[ 4191s] with preprocessed source if appropriate.
[ 4191s] Please include the complete backtrace with any bug report.
[ 4191s] See <https://bugs.opensuse.org/> for instructions.
[ 4191s] make[4]: *** [/tmp/cc3oPIcv.mk:257: /tmp/ccvUgWTy.ltrans85.ltrans.o]
Error 1
[ 4191s] lto-wrapper: fatal error: make returned 2 exit status

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

--- Comment #5 from Peter Bergner <bergner at gcc dot gnu.org> ---
Are the ppc tests and s390 bootstrap still broken?

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[seurer at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

--- Comment #6 from seurer at gcc dot gnu.org ---
The stack limit test case mostly has worked since just after this was opened
(one short span where it failed again) but I don't know what fixed it.  The
other test case has worked fine since just after this was opened.  Hopefully
bootstrap on s390 hasn't been broken since June!

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[bergner at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
         Resolution|---                         |WORKSFORME

--- Comment #7 from Peter Bergner <bergner at gcc dot gnu.org> ---
(In reply to seurer from comment #6)
> The stack limit test case mostly has worked since just after this was opened
> (one short span where it failed again) but I don't know what fixed it.  The
> other test case has worked fine since just after this was opened.  Hopefully
> bootstrap on s390 hasn't been broken since June!

gcc-testresults shows s390 test results, so I'm guessing this isn't broken
anymore.  Let's close this and if it's still a problem, we can reopen it.

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|RESOLVED                    |NEW
         Resolution|WORKSFORME                  |---
   Last reconfirmed|                            |2021-11-17

--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Actually it still fails:
https://gcc.gnu.org/pipermail/gcc-testresults/2021-November/735755.html

[Bug middle-end/101157] [12 regression] ICE compiling gcc.target/powerpc/stack-limit.c after r12-1702

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101157

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

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

--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The reason for the failure is described in PR 103163 so closing as a dup.

*** This bug has been marked as a duplicate of bug 103163 ***