[Bug target/103248] New: ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[Bug target/103248] New: ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

            Bug ID: 103248
           Summary: ICE in operation_could_trap_helper_p, at
                    tree-eh.c:2479
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: marxin at gcc dot gnu.org
  Target Milestone: ---
              Host: x86_64-linux-gnu
            Target: arm-linux-gnueabi

The following fails:

$ arm-linux-gnueabi-g++
/home/marxin/Programming/gcc/gcc/testsuite/g++.dg/abi/vbase8-22.C
-ftrivial-auto-var-init=pattern
warning: NEW ICE #1: in operation_could_trap_helper_p, at tree-eh.c:2479
timeout 10 arm-linux-gnueabi-gcc  -fmax-errors=1
-I/home/marxin/Programming/llvm-project/libcxx/test/support/ -Wno-overflow -O1
/home/marxin/Programming/llvm-project/clang/test/Frontend/fixed_point_compound.c
-fipa-cp-clone --param=builtin-expect-probability=100
--param=min-size-for-stack-sharing=2147483647 -floop-interchange -Wno-varargs
-Wno-builtin-declaration-mismatch -fno-printf-return-value
--param=max-stores-to-sink=1585089593 -Wno-c99-c11-compat
--param=loop-max-datarefs-for-datadeps=1582034702 -ftree-cselim
-Wignored-qualifiers -fno-sched-pressure -fno-tree-reassoc
-fno-split-ivs-in-unroller -floop-nest-optimize
--param=uninlined-function-time=1000000 -freciprocal-math
--param=scev-max-expr-size=0 -mthumb-interwork
--param=lra-max-considered-reload-pseudos=2147483647
--param=max-rtl-if-conversion-predictable-cost=62 -Wno-stringop-overread
-fcrossjumping -Wsystem-headers --param=hot-bb-frequency-fraction=950292743
--param=builtin-expect-probability=100 -Wstrict-aliasing=1
-Wvector-operation-performance -Wmemset-elt-size -fno-ira-share-spill-slots
-mword-relocations -mtp=cp15 -finline-small-functions
--param=min-loop-cond-split-prob=0 -Wno-varargs --param=lim-expensive=633593234
--param=iv-max-considered-uses=0 -fno-move-loop-invariants -fira-loop-pressure
--param=max-isl-operations=1856947825 --param=hwasan-random-frame-tag=1
-funconstrained-commons -Wno-analyzer-write-to-const
--param=inline-heuristics-hint-percent=67117 -flifetime-dse=0
--param=loop-block-tile-size=0 --param=dse-max-object-size=729093725
-fno-tree-sink -Wmissing-field-initializers -Wno-invalid-memory-model
-Wswitch-enum -fmodulo-sched-allow-regmoves -Wuninitialized -Wno-overflow
-falign-labels -fno-stdarg-opt -ftree-cselim -fpeephole2 -Wimplicit
--param=store-merging-allow-unaligned=0 -fno-cprop-registers
-Wno-int-conversion -Wcast-align --param=max-unswitch-level=2147483647
-fpeel-loops -fno-lifetime-dse -fsel-sched-pipelining-outer-loops
--param=hwasan-instrument-writes=1 --param=max-sched-insn-conflict-delay=10
-Wtype-limits --param=ira-max-conflict-table-size=0 -Wno-enum-compare
-Wno-endif-labels -fno-tree-loop-im -Wno-invalid-memory-model
--param=parloops-min-per-thread=2 -Wno-overflow -Wsequence-point
--param=avg-loop-niter=65536 -Wno-stringop-overread
-Wno-declaration-after-statement --param=modref-max-adjustments=23
-Wformat-zero-length -Wno-analyzer-file-leak -Wfloat-conversion -Wtrampolines
--param=align-loop-iterations=0 --param=unroll-jam-max-unroll=0
-Wsign-conversion --param=switch-conversion-max-branch-ratio=33837
-frounding-math --param=max-slsr-cand-scan=825635 -Waddress
-fno-ira-share-save-slots --param=modref-max-bases=2147483647 -funswitch-loops
-Wunused-parameter --param=min-inline-recursive-probability=0
--param=ipa-cp-loop-hint-bonus=944444867 -Wno-return-local-addr
--param=max-goto-duplication-insns=2144133459 -Wno-free-nonheap-object
-Wno-switch-bool -Wno-address-of-packed-member -Wnonnull-compare
-Wno-analyzer-null-argument --param=tracer-max-code-growth=2147483647
-Wbool-compare --param=max-peel-branches=1838480161 -funconstrained-commons
-Wmissing-field-initializers --param=vect-inner-loop-cost-factor=1
--param=max-dse-active-local-stores=0 -Wdate-time
--param=large-stack-frame-growth=1456052770
--param=iv-consider-all-candidates-bound=0 -frerun-cse-after-loop
--param=analyzer-bb-explosion-factor=2147483647
--param=gcse-after-reload-critical-fraction=0
--param=uninlined-thunk-time=24688 -Wenum-conversion -fsched-stalled-insns
-Wjump-misses-init --param=large-function-growth=0
-Wno-analyzer-exposure-through-output-file --param=max-partial-antic-length=0
-Wwrite-strings --param=iv-max-considered-uses=831437445 -Wformat-y2k
--param=max-grow-copy-bb-insns=0 --param=max-last-value-rtl=1973541945
--param=min-nondebug-insn-uid=898 --param=sra-max-scalarization-size-Ospeed=0
-Wformat-security --param=max-last-value-rtl=1538677469
--param=unroll-jam-min-percent=100 -flifetime-dse=1
--param=inline-heuristics-hint-percent=100 -Wno-analyzer-null-dereference
-Wno-if-not-aligned --param=hwasan-instrument-reads=1
--param=vect-partial-vector-usage=0 -frerun-cse-after-loop -Wcomment
-fno-tree-coalesce-vars --param=max-tree-if-conversion-phi-args=2 -fipa-vrp
--param=sra-max-scalarization-size-Osize=2147483647
--param=analyzer-bb-explosion-factor=93678980 -Wno-cpp
--param=ipa-max-switch-predicate-bounds=0
--param=dse-max-object-size=2147483647 --param=tm-max-aggregate-size=1061172261
-flive-range-shrinkage -falign-functions --param=gcse-unrestricted-cost=0
-Wnonnull -fnon-call-exceptions
--param=jump-table-max-growth-ratio-for-size=1000 -finline-functions
--param=ipa-cp-recursion-penalty=100
--param=iv-always-prune-cand-set-bound=2147483647
--param=tracer-dynamic-coverage-feedback=3
--param=cxx-max-namespaces-for-diagnostic-help=0
--param=hwasan-random-frame-tag=1 --param=modref-max-bases=2147483647
-fno-ipa-modref -Wdangling-else -fno-dce -Wmultistatement-macros
--param=jump-table-max-growth-ratio-for-size=466 -fstack-clash-protection
-Wzero-length-bounds -Warray-bounds
--param=tracer-min-branch-probability-feedback=51 -fsimd-cost-model=very-cheap
--param=partial-inlining-entry-probability=0 -fno-tree-ch -Wno-sync-nand
-Wno-sizeof-array-argument --param=inline-heuristics-hint-percent=552091
-fno-stdarg-opt -Wsizeof-pointer-div -Wshadow=compatible-local
--param=max-inline-insns-recursive-auto=0 -fvariable-expansion-in-unroller
--param=min-spec-prob=0 --param=ira-max-conflict-table-size=1261775574
-Wduplicated-cond --param=avoid-fma-max-bits=0
-fsel-sched-pipelining-outer-loops --param=l1-cache-size=0
-fisolate-erroneous-paths-attribute -fno-sched-last-insn-heuristic -Wno-tsan
-fno-bit-tests -Wno-scalar-storage-order --param=max-stores-to-sink=2147483647
--param=analyzer-max-recursion-depth=0 -o/dev/null -S
634 | void div_ai() {
|      ^~~~~~
0xdeadbeef operation_could_trap_helper_p(tree_code, bool, bool, bool, bool,
tree_node*, bool*)
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2479
0xdeadbeef stmt_could_throw_1_p
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2891
0xdeadbeef stmt_could_throw_p(function*, gimple*)
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2941
0xdeadbeef stmt_could_throw_p(function*, gimple*)
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2913
0xdeadbeef lower_eh_constructs_2
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2058
0xdeadbeef lower_eh_constructs_1
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2153
0xdeadbeef execute
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2198

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
0xdeadbeef

Not a gc issue which a5a5a5a5 pattern.

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #1)
> 0xdeadbeef

It's my string replacement, the original back-trace contains normal addresses.

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
With just -ftrivial-auto-var-init=pattern I see and ICE in

during GIMPLE pass: lower
vbase8-22.C: In function 'int main()':
vbase8-22.C:69:1: internal compiler error: in clear_padding_type, at
gimple-fold.c:4804
0x12d6a63 clear_padding_type
        /home/rguenther/src/trunk/gcc/gimple-fold.c:4804
0x12d69da clear_padding_type
        /home/rguenther/src/trunk/gcc/gimple-fold.c:4799
0x12d7b91 gimple_fold_builtin_clear_padding
        /home/rguenther/src/trunk/gcc/gimple-fold.c:5006
0x12d8335 gimple_fold_builtin
        /home/rguenther/src/trunk/gcc/gimple-fold.c:5180
0x12da469 gimple_fold_call
        /home/rguenther/src/trunk/gcc/gimple-fold.c:5588

cut&pasting all arguments you list doesn't change that, so ...?  How did you
configure GCC?  I have --target=arm-none-linux-gnueabihf --with-float=hard
--with-cpu=cortex-a9 --with-fpu=neon-fp16

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #4 from Martin Liška <marxin at gcc dot gnu.org> ---
arm-linux-gnueabi-gcc -v
Using built-in specs.
COLLECT_GCC=/home/marxin/BIG/bin/arm/dev/shm/buildbot/install/gcc/bin/arm-linux-gnueabi-gcc
COLLECT_LTO_WRAPPER=/home/marxin/BIG/bin/arm/dev/shm/buildbot/install/gcc/bin/../libexec/gcc/arm-linux-gnueabi/12.0.0/lto-wrapper
Target: arm-linux-gnueabi
Configured with:
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/configure
--enable-languages=c,c++,fortran --disable-bootstrap --disable-libsanitizer
--disable-multilib --enable-checking=release
--prefix=/dev/shm/buildbot/install/gcc --target=arm-linux-gnueabi
--with-as=/usr/bin/arm-suse-linux-gnueabi-as
Thread model: posix
Supported LTO compression algorithms: zlib zstd
gcc version 12.0.0 20211113 (experimental)
ecdf414bd89e6ba251f6b3f494407139b4dbae0e (GCC)

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
I can't reproduce.  Can you please post the cc1plus command-line as produced by
-v?

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #6 from Martin Liška <marxin at gcc dot gnu.org> ---
Sorry!! I copied totally wrong command line, so it's:

~/BIG/bin/arm/dev/shm/buildbot/install/gcc/bin/arm-linux-gnueabi-gcc
/home/marxin/Programming/llvm-project/clang/test/Frontend/fixed_point_compound.c
-fnon-call-exceptions
during GIMPLE pass: eh
/home/marxin/Programming/llvm-project/clang/test/Frontend/fixed_point_compound.c:
In function ‘div_ai’:
/home/marxin/Programming/llvm-project/clang/test/Frontend/fixed_point_compound.c:634:6:
internal compiler error: in operation_could_trap_helper_p, at tree-eh.c:2479
  634 | void div_ai() {
      |      ^~~~~~
0x614ff8 operation_could_trap_helper_p(tree_code, bool, bool, bool, bool,
tree_node*, bool*)
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2479
0xbd4407 stmt_could_throw_1_p
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2891
0xbd4407 stmt_could_throw_p(function*, gimple*)
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2941
0xbd628a stmt_could_throw_p(function*, gimple*)
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2913
0xbd628a lower_eh_constructs_2
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2058
0xbd8c5e lower_eh_constructs_1
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2153
0xbd8c5e execute
       
/home/marxin/BIG/buildbot/buildworker/marxinbox-gcc-trunk-arm/build/gcc/tree-eh.c:2198
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #3)
> With just -ftrivial-auto-var-init=pattern I see and ICE in
> 
> during GIMPLE pass: lower
> vbase8-22.C: In function 'int main()':
> vbase8-22.C:69:1: internal compiler error: in clear_padding_type, at
> gimple-fold.c:4804
> 0x12d6a63 clear_padding_type
>         /home/rguenther/src/trunk/gcc/gimple-fold.c:4804
> 0x12d69da clear_padding_type
>         /home/rguenther/src/trunk/gcc/gimple-fold.c:4799
> 0x12d7b91 gimple_fold_builtin_clear_padding
>         /home/rguenther/src/trunk/gcc/gimple-fold.c:5006
> 0x12d8335 gimple_fold_builtin
>         /home/rguenther/src/trunk/gcc/gimple-fold.c:5180
> 0x12da469 gimple_fold_call
>         /home/rguenther/src/trunk/gcc/gimple-fold.c:5588
> 
> cut&pasting all arguments you list doesn't change that, so ...?  How did you
> configure GCC?  I have --target=arm-none-linux-gnueabihf --with-float=hard
> --with-cpu=cortex-a9 --with-fpu=neon-fp16

I've filed PR103249 for the above.

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2021-11-16
             Status|UNCONFIRMED                 |WAITING

--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
can you also attach fixed_point_compound.c please?

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #9 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Richard Biener from comment #8)
> can you also attach fixed_point_compound.c please?

Sorry, it's from LLVM, reduced test-case:

$ cat fixed_point_compound.c
_Accum sa;
int c;

void div_csa() { c /= sa; }

[Bug target/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|1                           |0
             Status|WAITING                     |UNCONFIRMED

[Bug middle-end/103248] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
                 CC|                            |jsm28 at gcc dot gnu.org
          Component|target                      |middle-end
     Ever confirmed|0                           |1
           Assignee|unassigned at gcc dot gnu.org      |rguenth at gcc dot gnu.org

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
So we're using RDIV_EXPR for

<fixed_point_type 0x7ffff65610a8 long _Accum>

integer_zerop does not handle FIXED_CST, so previously we considered
all fixed point divisions by constants as not trapping and by non-constants
as trapping.  I can fix that at least as far as using fixed_zerop but I'm not
sure of the fixed point behavior.

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1
            Summary|ICE in                      |[12 Regression] ICE in
                   |operation_could_trap_helper |operation_could_trap_helper
                   |_p, at tree-eh.c:2479       |_p, at tree-eh.c:2479
   Target Milestone|---                         |12.0

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[joseph at codesourcery dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #11 from joseph at codesourcery dot com <joseph at codesourcery dot com> ---
Division by zero is undefined behavior for fixed-point types the same way 
as it is for integer types (but not floating point, at least when 
infinities and NaN are supported).  Treating it like integer division in 
the compiler would seem reasonable.

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #12 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Biener <rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:fb15abdc9b61a0b7fa24a37f85d19dc449cfd5bf

commit r12-5395-gfb15abdc9b61a0b7fa24a37f85d19dc449cfd5bf
Author: Richard Biener <rguenther@suse.de>
Date:   Tue Nov 16 11:47:26 2021 +0100

    middle-end/103248 - fix RDIV_EXPR handling with fixed point

    This fixes the previous adjustment to operation_could_trap_helper_p
    where I failed to realize that RDIV_EXPR is also used for
    fixed-point types.  It also fixes that handling by properly
    checking for a fixed_zerop divisor.

    2021-11-16  Richard Biener  <rguenther@suse.de>

            PR middle-end/103248
            * tree-eh.c (operation_could_trap_helper_p): Properly handle
            fixed-point RDIV_EXPR.

            * gcc.dg/pr103248.c: New testcase.

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

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

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

--- Comment #13 from Richard Biener <rguenth at gcc dot gnu.org> ---
Fixed.

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #14 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:2f0c8f74daef93c0c7b33294213e7db6df58c4d1

commit r11-9259-g2f0c8f74daef93c0c7b33294213e7db6df58c4d1
Author: Richard Biener <rguenther@suse.de>
Date:   Thu Nov 11 09:40:36 2021 +0100

    middle-end/103181 - fix operation_could_trap_p for vector division

    For integer vector division we only checked for all zero vector
    constants rather than checking whether any element in the constant
    vector is zero.

    It also fixes the adjustment to operation_could_trap_helper_p
    where I failed to realize that RDIV_EXPR is also used for
    fixed-point types.  It also fixes that handling by properly
    checking for a fixed_zerop divisor.

    2021-11-11  Richard Biener  <rguenther@suse.de>

            PR middle-end/103181
            PR middle-end/103248
            * tree-eh.c (operation_could_trap_helper_p): Properly
            check vector constants for a zero element for integer
            division.  Separate floating point and integer division code.
            Properly handle fixed-point RDIV_EXPR.

            * gcc.dg/torture/pr103181.c: New testcase.
            * gcc.dg/pr103248.c: Likewise.

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #15 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:5f5c26a2f02572996efd2b2de0dd5600ad7e91e7

commit r10-10464-g5f5c26a2f02572996efd2b2de0dd5600ad7e91e7
Author: Richard Biener <rguenther@suse.de>
Date:   Thu Nov 11 09:40:36 2021 +0100

    middle-end/103181 - fix operation_could_trap_p for vector division

    For integer vector division we only checked for all zero vector
    constants rather than checking whether any element in the constant
    vector is zero.

    It also fixes the adjustment to operation_could_trap_helper_p
    where I failed to realize that RDIV_EXPR is also used for
    fixed-point types.  It also fixes that handling by properly
    checking for a fixed_zerop divisor.

    2021-11-11  Richard Biener  <rguenther@suse.de>

            PR middle-end/103181
            PR middle-end/103248
            * tree-eh.c (operation_could_trap_helper_p): Properly
            check vector constants for a zero element for integer
            division.  Separate floating point and integer division code.
            Properly handle fixed-point RDIV_EXPR.

            * gcc.dg/torture/pr103181.c: New testcase.
            * gcc.dg/pr103248.c: Likewise.

    (cherry picked from commit 2f0c8f74daef93c0c7b33294213e7db6df58c4d1)

[Bug middle-end/103248] [12 Regression] ICE in operation_could_trap_helper_p, at tree-eh.c:2479

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103248

--- Comment #16 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Richard Biener
<rguenth@gcc.gnu.org>:

https://gcc.gnu.org/g:1ac5fbea476e4ebd6a0086ccfbf92e648768be7b

commit r9-9955-g1ac5fbea476e4ebd6a0086ccfbf92e648768be7b
Author: Richard Biener <rguenther@suse.de>
Date:   Thu Nov 11 09:40:36 2021 +0100

    middle-end/103181 - fix operation_could_trap_p for vector division

    For integer vector division we only checked for all zero vector
    constants rather than checking whether any element in the constant
    vector is zero.

    It also fixes the adjustment to operation_could_trap_helper_p
    where I failed to realize that RDIV_EXPR is also used for
    fixed-point types.  It also fixes that handling by properly
    checking for a fixed_zerop divisor.

    2021-11-11  Richard Biener  <rguenther@suse.de>

            PR middle-end/103181
            PR middle-end/103248
            * tree-eh.c (operation_could_trap_helper_p): Properly
            check vector constants for a zero element for integer
            division.  Separate floating point and integer division code.
            Properly handle fixed-point RDIV_EXPR.

            * gcc.dg/torture/pr103181.c: New testcase.
            * gcc.dg/pr103248.c: Likewise.

    (cherry picked from commit 2f0c8f74daef93c0c7b33294213e7db6df58c4d1)