[Bug sanitizer/109151] New: UBsan misses a divide-by-zero

[Bug sanitizer/109151] New: UBsan misses a divide-by-zero

[shaohua.li at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

            Bug ID: 109151
           Summary: UBsan misses a divide-by-zero
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: sanitizer
          Assignee: unassigned at gcc dot gnu.org
          Reporter: shaohua.li at inf dot ethz.ch
                CC: dodji at gcc dot gnu.org, dvyukov at gcc dot gnu.org,
                    jakub at gcc dot gnu.org, kcc at gcc dot gnu.org, marxin at gcc dot gnu.org
  Target Milestone: ---

For the following code, UBsan failed to report the divide-by-zero, while
Clang's UBsan could.

Compiler explorer: https://godbolt.org/z/Y73vrf8zK

%cat a.c
int a, c;
short b;
long d;
int main() { 
    a = (short)(d == c | b > 9) / 0; 
    return a;
}
%
% gcc -fsanitize=undefined a.c &&./a.out
Floating point exception
%
% clang -fsanitize=undefined a.c &&./a.out
/a.c:5:42: runtime error: division by zero
%

[Bug c/109151] UBsan misses a divide-by-zero

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|sanitizer                   |c

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The C++ front-end can detect it ...

[Bug c/109151] UBsan misses a divide-by-zero

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |4.9.0
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2023-03-15
     Ever confirmed|0                           |1

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed
Here is the shortest testcase which shows the issue:
int d;
int main() {
    return ((short)(d == 1 | d > 9)) / 0;
}

[Bug c/109151] UBsan misses a divide-by-zero

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |jakub at gcc dot gnu.org

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 54679
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54679&action=edit
gcc13-pr109151.patch

Untested fix.

[Bug c/109151] UBsan misses a divide-by-zero

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jakub Jelinek <jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:103d423f6ce72ccb03d55b7b1dfa2dabd5854371

commit r13-6738-g103d423f6ce72ccb03d55b7b1dfa2dabd5854371
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Mar 17 16:10:14 2023 +0100

    c, ubsan: Instrument even shortened divisions [PR109151]

    On the following testcase, the C FE decides to shorten the division because
    it has a guarantee that INT_MIN / -1 division won't be encountered, the
    first operand is widened from narrower unsigned and/or the second operand
is
    a constant other than all ones (in this case both are true).
    The problem is that the narrower type in this case is _Bool and
    ubsan_instrument_division only instruments it if op0's type is INTEGER_TYPE
    or REAL_TYPE.  Strangely this doesn't happen in C++ FE.
    Anyway, we only shorten divisions if the INT_MIN / -1 case is impossible,
    so I think we should be fine even with -fstrict-enums in C++ in case it
    shortened to ENUMERAL_TYPEs.

    The following patch just instruments those on the ubsan_instrument_division
    side.  Perhaps only the first hunk and testcase might be needed because
    we shouldn't shorten if the other case could be triggered.

    2023-03-17  Jakub Jelinek  <jakub@redhat.com>

            PR c/109151
            * c-ubsan.cc (ubsan_instrument_division): Handle all scalar
integral
            types rather than just INTEGER_TYPE.

            * c-c++-common/ubsan/div-by-zero-8.c: New test.

[Bug c/109151] UBsan misses a divide-by-zero

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:ffb22189be90216e6af386aaef8b76d2dcd05b84

commit r12-9292-gffb22189be90216e6af386aaef8b76d2dcd05b84
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Mar 17 16:10:14 2023 +0100

    c, ubsan: Instrument even shortened divisions [PR109151]

    On the following testcase, the C FE decides to shorten the division because
    it has a guarantee that INT_MIN / -1 division won't be encountered, the
    first operand is widened from narrower unsigned and/or the second operand
is
    a constant other than all ones (in this case both are true).
    The problem is that the narrower type in this case is _Bool and
    ubsan_instrument_division only instruments it if op0's type is INTEGER_TYPE
    or REAL_TYPE.  Strangely this doesn't happen in C++ FE.
    Anyway, we only shorten divisions if the INT_MIN / -1 case is impossible,
    so I think we should be fine even with -fstrict-enums in C++ in case it
    shortened to ENUMERAL_TYPEs.

    The following patch just instruments those on the ubsan_instrument_division
    side.  Perhaps only the first hunk and testcase might be needed because
    we shouldn't shorten if the other case could be triggered.

    2023-03-17  Jakub Jelinek  <jakub@redhat.com>

            PR c/109151
            * c-ubsan.cc (ubsan_instrument_division): Handle all scalar
integral
            types rather than just INTEGER_TYPE.

            * c-c++-common/ubsan/div-by-zero-8.c: New test.

    (cherry picked from commit 103d423f6ce72ccb03d55b7b1dfa2dabd5854371)

[Bug c/109151] UBsan misses a divide-by-zero

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 12.3 too.

[Bug c/109151] UBsan misses a divide-by-zero

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:7d9f6140f0fdd59be002a2efb1d61939b29b450f

commit r11-10726-g7d9f6140f0fdd59be002a2efb1d61939b29b450f
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Mar 17 16:10:14 2023 +0100

    c, ubsan: Instrument even shortened divisions [PR109151]

    On the following testcase, the C FE decides to shorten the division because
    it has a guarantee that INT_MIN / -1 division won't be encountered, the
    first operand is widened from narrower unsigned and/or the second operand
is
    a constant other than all ones (in this case both are true).
    The problem is that the narrower type in this case is _Bool and
    ubsan_instrument_division only instruments it if op0's type is INTEGER_TYPE
    or REAL_TYPE.  Strangely this doesn't happen in C++ FE.
    Anyway, we only shorten divisions if the INT_MIN / -1 case is impossible,
    so I think we should be fine even with -fstrict-enums in C++ in case it
    shortened to ENUMERAL_TYPEs.

    The following patch just instruments those on the ubsan_instrument_division
    side.  Perhaps only the first hunk and testcase might be needed because
    we shouldn't shorten if the other case could be triggered.

    2023-03-17  Jakub Jelinek  <jakub@redhat.com>

            PR c/109151
            * c-ubsan.c (ubsan_instrument_division): Handle all scalar integral
            types rather than just INTEGER_TYPE.

            * c-c++-common/ubsan/div-by-zero-8.c: New test.

    (cherry picked from commit 103d423f6ce72ccb03d55b7b1dfa2dabd5854371)

[Bug c/109151] UBsan misses a divide-by-zero

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

--- Comment #8 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 11.4 as well.

[Bug c/109151] UBsan misses a divide-by-zero

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:bdd543db33706643993a468740487147c89a4bd2

commit r10-11379-gbdd543db33706643993a468740487147c89a4bd2
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Fri Mar 17 16:10:14 2023 +0100

    c, ubsan: Instrument even shortened divisions [PR109151]

    On the following testcase, the C FE decides to shorten the division because
    it has a guarantee that INT_MIN / -1 division won't be encountered, the
    first operand is widened from narrower unsigned and/or the second operand
is
    a constant other than all ones (in this case both are true).
    The problem is that the narrower type in this case is _Bool and
    ubsan_instrument_division only instruments it if op0's type is INTEGER_TYPE
    or REAL_TYPE.  Strangely this doesn't happen in C++ FE.
    Anyway, we only shorten divisions if the INT_MIN / -1 case is impossible,
    so I think we should be fine even with -fstrict-enums in C++ in case it
    shortened to ENUMERAL_TYPEs.

    The following patch just instruments those on the ubsan_instrument_division
    side.  Perhaps only the first hunk and testcase might be needed because
    we shouldn't shorten if the other case could be triggered.

    2023-03-17  Jakub Jelinek  <jakub@redhat.com>

            PR c/109151
            * c-ubsan.c (ubsan_instrument_division): Handle all scalar integral
            types rather than just INTEGER_TYPE.

            * c-c++-common/ubsan/div-by-zero-8.c: New test.

    (cherry picked from commit 103d423f6ce72ccb03d55b7b1dfa2dabd5854371)

[Bug c/109151] UBsan misses a divide-by-zero

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109151

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

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

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Fixed for 10.5 too.