[Bug tree-optimization/113013] New: [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[Bug tree-optimization/113013] New: [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[sjames at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

            Bug ID: 113013
           Summary: [12/13/14 regression] ICE in fold_convert_loc with
                    -fsanitize=undefined
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: sjames at gcc dot gnu.org
  Target Milestone: ---

Created attachment 56876
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=56876&action=edit
reduced.i

Hit this when reducing PR113012.

```
$ gcc -c /tmp/foo.c -O2 -fsanitize=undefined
during GIMPLE pass: objsz
/tmp/foo.c: In function ‘miZeroLine_list_len’:
/tmp/foo.c:9:6: internal compiler error: in fold_convert_loc, at
fold-const.cc:2602
    9 | void miZeroLine_list_len() {
      |      ^~~~~~~~~~~~~~~~~~~
0x55eb4274eca3 fold_convert_loc(unsigned int, tree_node*, tree_node*)
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/fold-const.cc:2602
0x55eb446f7d13 alloc_object_size
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/tree-object-size.cc:811
0x55eb446f7d13 call_object_size
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/tree-object-size.cc:1339
0x55eb446f7d13 collect_object_sizes_for
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/tree-object-size.cc:1779
0x55eb4418920f compute_builtin_object_size(tree_node*, int, tree_node**)
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/tree-object-size.cc:1194
0x55eb43fab777 fold_builtin_object_size
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/builtins.cc:11556
0x55eb43fab777 fold_builtin_2
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/builtins.cc:10575
0x55eb43fab777 fold_builtin_n(unsigned int, tree_node*, tree_node*,
tree_node**, int, bool) [clone .isra.0]
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/builtins.cc:10687
0x55eb440cd22d dynamic_object_sizes_execute_one
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/tree-object-size.cc:2078
0x55eb440cd22d object_sizes_execute
       
/usr/src/debug/sys-devel/gcc-14.0.0_pre20231210/gcc-14-20231210/gcc/tree-object-size.cc:2140
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://bugs.gentoo.org/> for instructions.
```

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
           Keywords|                            |ice-on-valid-code
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2023-12-14

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Reduced slightly more:
```
struct _xPoint {
  short x;
} xPoint;
void *reallocarray() __attribute__((__alloc_size__(1)));
struct _xPoint * miZeroLine_pspanInit;
void miZeroLine_list_len(void) {
  miZeroLine_pspanInit = reallocarray(xPoint);
  miZeroLine_pspanInit->x = 0;
}
``

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |12.4

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Most likely r12-6479-g404c787e2bfe8cae666b075ed903990ea452220e or
r12-6482-g06bc1b0c539e3a60692d7432d15e701c38610f80 .

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
The use of __alloc_size__ here is definitely not well defined but GCC should
NOT crash.

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[siddhesh at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

Siddhesh Poyarekar <siddhesh at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|ice-on-valid-code           |ice-on-invalid-code
                 CC|                            |siddhesh at gcc dot gnu.org

--- Comment #4 from Siddhesh Poyarekar <siddhesh at gcc dot gnu.org> ---
Agreed, the attribute is invalid and maybe the frontend needs to flag it early
and flag an error.

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

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 #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Created attachment 56885
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=56885&action=edit
gcc14-pr113013.patch

I don't see how FEs could do anything, especially on declarations like this
where the argument types aren't specified.
For attributes like this, it is generally responsibility of the function
using those to use it only if they are sane, and after all, even the
tree-object-size.cc code does that (e.g. verify that such argument exists at
all).  Just doesn't check that it is convertible to sizetype.

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[siddhesh at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

--- Comment #6 from Siddhesh Poyarekar <siddhesh at gcc dot gnu.org> ---
Sorry I misread the reproducer as void *reallocarray(void)
__attribute__((__alloc_size__(1)));

Your fix looks fine to me, thanks.

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

--- Comment #7 from GCC 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:5347263b347d02e875879ca40ca6e289ac178919

commit r14-6654-g5347263b347d02e875879ca40ca6e289ac178919
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Dec 18 09:59:28 2023 +0100

    tree-object-size: Robustify alloc_size attribute handling [PR113013]

    The following testcase ICEs because we aren't careful enough with
    alloc_size attribute.  We do check that such an argument exists
    (although wouldn't handle correctly functions with more than INT_MAX
    arguments), but didn't check that it is scalar integer, the ICE is
    trying to fold_convert a structure to sizetype.

    Given that the attribute can also appear on non-prototyped functions
    where the arguments aren't known, I don't see how the FE could diagnose
    that and because we already handle the case where argument doesn't exist,
    I think we should also verify the argument is scalar integer convertible
    to sizetype.  Furthermore, given this is not just in diagnostics but
    used for code generation, I think it is better to punt on arguments with
    larger precision then sizetype, the upper bits are then truncated.

    The patch also fixes some formatting issues and avoids duplication of the
    fold_convert, plus removes unnecessary check for if (arg1 >= 0), that is
    always the case after if (arg1 < 0) return ...;

    2023-12-18  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/113013
            * tree-object-size.cc (alloc_object_size): Return size_unknown if
            corresponding argument(s) don't have integral type or have integral
            type with higher precision than sizetype.  Don't check arg1 >= 0
            uselessly.  Compare argument indexes against gimple_call_num_args
            in unsigned type rather than int.  Formatting fixes.

            * gcc.dg/pr113013.c: New test.

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

--- Comment #8 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-13 branch has been updated by Jakub Jelinek
<jakub@gcc.gnu.org>:

https://gcc.gnu.org/g:2ea5a22b3e594cdf88e916b3df774883ce9dbc3a

commit r13-8165-g2ea5a22b3e594cdf88e916b3df774883ce9dbc3a
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Dec 18 09:59:28 2023 +0100

    tree-object-size: Robustify alloc_size attribute handling [PR113013]

    The following testcase ICEs because we aren't careful enough with
    alloc_size attribute.  We do check that such an argument exists
    (although wouldn't handle correctly functions with more than INT_MAX
    arguments), but didn't check that it is scalar integer, the ICE is
    trying to fold_convert a structure to sizetype.

    Given that the attribute can also appear on non-prototyped functions
    where the arguments aren't known, I don't see how the FE could diagnose
    that and because we already handle the case where argument doesn't exist,
    I think we should also verify the argument is scalar integer convertible
    to sizetype.  Furthermore, given this is not just in diagnostics but
    used for code generation, I think it is better to punt on arguments with
    larger precision then sizetype, the upper bits are then truncated.

    The patch also fixes some formatting issues and avoids duplication of the
    fold_convert, plus removes unnecessary check for if (arg1 >= 0), that is
    always the case after if (arg1 < 0) return ...;

    2023-12-18  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/113013
            * tree-object-size.cc (alloc_object_size): Return size_unknown if
            corresponding argument(s) don't have integral type or have integral
            type with higher precision than sizetype.  Don't check arg1 >= 0
            uselessly.  Compare argument indexes against gimple_call_num_args
            in unsigned type rather than int.  Formatting fixes.

            * gcc.dg/pr113013.c: New test.

    (cherry picked from commit 5347263b347d02e875879ca40ca6e289ac178919)

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

--- Comment #9 from GCC 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:febb86b366047e0ea48da85c9b52ee03878cb5bf

commit r12-10059-gfebb86b366047e0ea48da85c9b52ee03878cb5bf
Author: Jakub Jelinek <jakub@redhat.com>
Date:   Mon Dec 18 09:59:28 2023 +0100

    tree-object-size: Robustify alloc_size attribute handling [PR113013]

    The following testcase ICEs because we aren't careful enough with
    alloc_size attribute.  We do check that such an argument exists
    (although wouldn't handle correctly functions with more than INT_MAX
    arguments), but didn't check that it is scalar integer, the ICE is
    trying to fold_convert a structure to sizetype.

    Given that the attribute can also appear on non-prototyped functions
    where the arguments aren't known, I don't see how the FE could diagnose
    that and because we already handle the case where argument doesn't exist,
    I think we should also verify the argument is scalar integer convertible
    to sizetype.  Furthermore, given this is not just in diagnostics but
    used for code generation, I think it is better to punt on arguments with
    larger precision then sizetype, the upper bits are then truncated.

    The patch also fixes some formatting issues and avoids duplication of the
    fold_convert, plus removes unnecessary check for if (arg1 >= 0), that is
    always the case after if (arg1 < 0) return ...;

    2023-12-18  Jakub Jelinek  <jakub@redhat.com>

            PR tree-optimization/113013
            * tree-object-size.cc (alloc_object_size): Return size_unknown if
            corresponding argument(s) don't have integral type or have integral
            type with higher precision than sizetype.  Don't check arg1 >= 0
            uselessly.  Compare argument indexes against gimple_call_num_args
            in unsigned type rather than int.  Formatting fixes.

            * gcc.dg/pr113013.c: New test.

    (cherry picked from commit 5347263b347d02e875879ca40ca6e289ac178919)

[Bug tree-optimization/113013] [12/13/14 regression] ICE in fold_convert_loc with -fsanitize=undefined

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113013

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

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

--- Comment #10 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Should be fixed now.