[Bug middle-end/113750] New: [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[Bug middle-end/113750] New: [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

            Bug ID: 113750
           Summary: [14 Regression] ICE in vect building
                    gcc/m2/gm2-libs/NumberIO.mod
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: iains at gcc dot gnu.org
  Target Milestone: ---
              Host: aarch64-linux
            Target: aarch64-linux

This is on cfarm185 (aarch64-linux) with a compiler configured:
../src/configure --prefix=... --enable-languages=all  

-save-temps is not giving me any intermediate and fdump-tree-gimple gives no
output, will have figure out how to reduce this.

----

/home/iains/gcc-master/bld/./gcc/cc1gm2 -v -iprefix
/home/iains/gcc-master/bld/gcc/../lib/gcc/aarch64-unknown-linux-gnu/14.0.1/
-isystem /home/iains/gcc-master/bld/./gcc/include -isystem
/home/iains/gcc-master/bld/./gcc/include-fixed -B
/home/iains/gcc-master/bld/./gcc/ -save-temps -quiet -dumpdir .libs/ -dumpbase
NumberIO.mod -dumpbase-ext .mod -mlittle-endian -mabi=lp64 -g -g -g -O2 -O2
-Wcase-enum -Wreturn-type -version -fm2-g -fcase -fm2-prefix=m2pim -fPIC
-fdump-tree-gimple -fm2-pathname=m2pim -fm2-pathnameI.
-fm2-pathnameI/home/iains/gcc-master/src/gcc/m2/gm2-libs -fm2-pathname=m2iso
-fm2-pathnameI/home/iains/gcc-master/src/gcc/m2/gm2-libs-iso -fscaffold-dynamic
-flibs=m2cor,m2log,m2pim,m2iso -fm2-pathname=m2pim -fm2-pathnameI.
-fm2-pathnameI/home/iains/gcc-master/src/gcc/m2/gm2-libs -fm2-pathname=m2iso
-fm2-pathnameI/home/iains/gcc-master/src/gcc/m2/gm2-libs-iso
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod -c -o
.libs/NumberIO.s
GNU Modula-2 (GCC) version 14.0.1 20240203 (experimental) [master revision
r14-8771-g4b7d4d8a4af7] (aarch64-unknown-linux-gnu)
        compiled by GNU C version 14.0.1 20240203 (experimental) [master
revision r14-8771-g4b7d4d8a4af7], GMP version 6.1.2, MPFR version 3.1.6-p2, MPC
version 1.1.0, isl version none
GGC heuristics: --param ggc-min-expand=30 --param ggc-min-heapsize=4096
Compiling: ../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod: In function
‘m2pim_NumberIO_HexToStr’:
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:237:1:
error: label ‘.L873’ in the middle of basic block 3
  237 | BEGIN
      | ^~~~~
during GIMPLE pass: vect
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:237:1:
internal compiler error: verify_flow_info failed
0x91900b verify_flow_info()
        ../../src/gcc/cfghooks.cc:287
0xf0a15f checking_verify_flow_info()
        ../../src/gcc/cfghooks.h:214
0xf0a15f cleanup_tree_cfg_noloop
        ../../src/gcc/tree-cfgcleanup.cc:1154
0xf0a15f cleanup_tree_cfg(unsigned int)
        ../../src/gcc/tree-cfgcleanup.cc:1205
0xd8737b execute_function_todo
        ../../src/gcc/passes.cc:2057
0xd879ab execute_todo
        ../../src/gcc/passes.cc:2142

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu.org
   Target Milestone|---                         |14.0

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

--- Comment #1 from Iain Sandoe <iains at gcc dot gnu.org> ---
..

(gdb) p debug_tree(current_function_decl)
 <function_decl 0xfffff5aad100 m2pim_NumberIO_HexToStr
    type <function_type 0xfffff5b5d008
        type <void_type 0xfffff5b50f18 PROC asm_written VOID
            align:8 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0xfffff5b50f18
            pointer_to_this <pointer_type 0xfffff5b50fc0 ADDRESS>>
        SI
        size <integer_cst 0xfffff5a60fc0 constant 32>
        unit-size <integer_cst 0xfffff5a60fd8 constant 4>
        align:32 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type
0xfffff5b5d008
        arg-types <tree_list 0xfffff5ad9538 value <integer_type 0xfffff5b50690
CARDINAL>
            chain <tree_list 0xfffff5ad9510 value <integer_type 0xfffff5b50690
CARDINAL>
                chain <tree_list 0xfffff5ad94e8 value <record_type
0xfffff5b5b868>
                    chain <tree_list 0xfffff5ad94c0 value <void_type
0xfffff5b50f18 PROC>>>>>
        pointer_to_this <pointer_type 0xfffff5b54ad0>>
    addressable used public static DI
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:237:1
align:32 warn_if_not_align:0 context <translation_unit_decl 0xfffff5b90000
D.56> initial <block 0xfffff5be03c0>
    result <result_decl 0xfffff5b90c30 D.321 type <void_type 0xfffff5b50f18
PROC>
        VOID
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:270:1
        align:8 warn_if_not_align:0 context <function_decl 0xfffff5aad100
m2pim_NumberIO_HexToStr>>
    arguments <parm_decl 0xfffff7a51380 x
        type <integer_type 0xfffff5b50690 CARDINAL sizes-gimplified asm_written
public unsigned SI size <integer_cst 0xfffff5a60fc0 32> unit-size <integer_cst
0xfffff5a60fd8 4>
            align:32 warn_if_not_align:0 symtab:-172293664 alias-set 2
canonical-type 0xfffff5b50690 precision:32 min <integer_cst 0xfffff5a60ff0 0>
max <integer_cst 0xfffff5a60fa8 4294967295>
            pointer_to_this <pointer_type 0xfffff5b5bbb0> reference_to_this
<reference_type 0xfffff5b5cac8>>
        used unsigned SI
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:232:21 size
<integer_cst 0xfffff5a60fc0 32> unit-size <integer_cst 0xfffff5a60fd8 4>
        align:32 warn_if_not_align:0 context <function_decl 0xfffff5aad100
m2pim_NumberIO_HexToStr> arg-type <integer_type 0xfffff5b50690 CARDINAL>
        chain <parm_decl 0xfffff7a51300 n type <integer_type 0xfffff5b50690
CARDINAL>
            used unsigned SI
../../../../src/libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:232:24 size
<integer_cst 0xfffff5a60fc0 32> unit-size <integer_cst 0xfffff5a60fd8 4>
            align:32 warn_if_not_align:0 context <function_decl 0xfffff5aad100
m2pim_NumberIO_HexToStr> arg-type <integer_type 0xfffff5b50690 CARDINAL> chain
<parm_decl 0xfffff7a51280 a>>>
    struct-function 0xfffff5bd0730 chain <function_decl 0xfffff5aad000
m2pim_NumberIO_StrToCard>>
$1 = void
(gdb) p debug_tree(current_function_decl->function_decl.saved_tree)

$2 = void

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[gaius at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

Gaius Mulley <gaius at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |gaius at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2024-02-04

--- Comment #2 from Gaius Mulley <gaius at gcc dot gnu.org> ---
Confirmed seen it fail on aarch64 (non bootstrap)

/bin/bash ../libtool --tag=CC   \
                   --mode=compile
/home/gaius/GCC/gcc-read-write/build/./gcc/gm2
-B/home/gaius/GCC/gcc-read-write/build/./gcc/ \
                   -c -g -O2 -g -O2 -fm2-pathname=m2pim -I.
-I/home/gaius/GCC/gcc-read-write/gcc/m2/gm2-libs -fm2-pathname=m2iso
-I/home/gaius/GCC/gcc-read-write/gcc/m2/gm2-libs-iso -fm2-g -g -Wcase-enum
-Wreturn-type -fcase -fm2-prefix=m2pim
../../../../libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod -o NumberIO.lo
libtool: compile:  /home/gaius/GCC/gcc-read-write/build/./gcc/gm2
-B/home/gaius/GCC/gcc-read-write/build/./gcc/ -c -g -O2 -g -O2
-fm2-pathname=m2pim -I. -I/home/gaius/GCC/gcc-read-write/gcc/m2/gm2-libs
-fm2-pathname=m2iso -I/home/gaius/GCC/gcc-read-write/gcc/m2/gm2-libs-iso -fm2-g
-g -Wcase-enum -Wreturn-type -fcase -fm2-prefix=m2pim
../../../../libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod  -fPIC -DPIC -o
.libs/NumberIO.o
../../../../libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod: In function
'm2pim_NumberIO_HexToStr':
../../../../libgm2/libm2pim/../../gcc/m2/gm2-libs/NumberIO.mod:237:1: error:
label '.L873' in the middle of basic block 3

It passes prior to git commit 64b0130bb6702c67a13caefaae9facef23d6ac60

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

--- Comment #3 from Iain Sandoe <iains at gcc dot gnu.org> ---
Created attachment 57317
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57317&action=edit
gimple for 179t.ifcvt

the label seems OK here.

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod

[iains at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

--- Comment #4 from Iain Sandoe <iains at gcc dot gnu.org> ---
Created attachment 57318
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57318&action=edit
gimple for 179t.vect

now there's a mem at the start of bb3 with the label following

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod since r14-8769-g64b0130bb6702c

[sjames at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

Sam James <sjames at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[14 Regression] ICE in vect |[14 Regression] ICE in vect
                   |building                    |building
                   |gcc/m2/gm2-libs/NumberIO.mo |gcc/m2/gm2-libs/NumberIO.mo
                   |d                           |d since
                   |                            |r14-8769-g64b0130bb6702c
           Assignee|unassigned at gcc dot gnu.org      |gaius at gcc dot gnu.org

--- Comment #5 from Sam James <sjames at gcc dot gnu.org> ---
(assigning to gaius@gcc.gnu.org as assuming that was the intention, feel free
to fix if not)

[Bug middle-end/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod since r14-8769-g64b0130bb6702c

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|gaius at gcc dot gnu.org           |tnfchris at gcc dot gnu.org

--- Comment #6 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
I assumed that labels could only still be present at this point of the CFG only
to deal with control flow from e.g. switch statements which we reject earlier.

It seems that these labels are an artifact of the front-end and are dead by the
time they reach vect but still there in the IL.

I'll add a quick check for if they are and can set the stores before the last
non-debug statement.

Testing a patch.

[Bug tree-optimization/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod since r14-8769-g64b0130bb6702c

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

--- Comment #7 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Tamar Christina <tnfchris@gcc.gnu.org>:

https://gcc.gnu.org/g:5c3ba60024fedc6b3d374ebb071bcf5b3e27cd62

commit r14-8840-g5c3ba60024fedc6b3d374ebb071bcf5b3e27cd62
Author: Tamar Christina <tamar.christina@arm.com>
Date:   Wed Feb 7 10:59:32 2024 +0000

    middle-end: fix ICE when destination BB for stores starts with a label
[PR113750]

    The report shows that if the FE leaves a label as the first thing in the
dest
    BB then we ICE because we move the stores before the label.

    This is easy to fix if we know that there's still only one way into the BB.
    We would have already rejected the loop if there was multiple paths into
the BB
    however I added an additional check just for early break in case the other
    constraints are relaxed later with an explanation.

    After that we fix the issue just by getting the GSI after the labels and I
add
    a bunch of testcases for different positions the label can be added.  Only
the
    vect-early-break_112-pr113750.c one results in the label being kept.

    gcc/ChangeLog:

            PR tree-optimization/113750
            * tree-vect-data-refs.cc (vect_analyze_early_break_dependences):
Check
            for single predecessor when doing early break vect.
            * tree-vect-loop.cc (move_early_exit_stmts): Get gsi at the start
but
            after labels.

    gcc/testsuite/ChangeLog:

            PR tree-optimization/113750
            * gcc.dg/vect/vect-early-break_112-pr113750.c: New test.
            * gcc.dg/vect/vect-early-break_113-pr113750.c: New test.
            * gcc.dg/vect/vect-early-break_114-pr113750.c: New test.
            * gcc.dg/vect/vect-early-break_115-pr113750.c: New test.
            * gcc.dg/vect/vect-early-break_116-pr113750.c: New test.

[Bug tree-optimization/113750] [14 Regression] ICE in vect building gcc/m2/gm2-libs/NumberIO.mod since r14-8769-g64b0130bb6702c

[tnfchris at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113750

Tamar Christina <tnfchris at gcc dot gnu.org> changed:

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

--- Comment #8 from Tamar Christina <tnfchris at gcc dot gnu.org> ---
Fixed.