[Bug c/102029] New: ice: error: type mismatch in ‘lshift_expr’

[Bug c/102029] New: ice: error: type mismatch in ‘lshift_expr’

[dcb314 at hotmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

            Bug ID: 102029
           Summary: ice:  error: type mismatch in ‘lshift_expr’
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dcb314 at hotmail dot com
  Target Milestone: ---

For this C code:

$ more bug743.c
const long fmpz_neg_f2;
void qfb_reduce(void) { __gmpz_neg((int *)(fmpz_neg_f2 << 2)); }
$ 

Recent gcc trunk does this:

$ /home/dcb/gcc/results.20210823/bin/gcc  -c -O2 bug743.c 
bug743.c: In function ‘qfb_reduce’:
bug743.c:2:25: warning: implicit declaration of function ‘__gmpz_neg’
[-Wimplici
t-function-declaration]
    2 | void qfb_reduce(void) { __gmpz_neg((int *)(fmpz_neg_f2 << 2)); }
      |                         ^~~~~~~~~~
bug743.c:2:6: error: type mismatch in ‘lshift_expr’
    2 | void qfb_reduce(void) { __gmpz_neg((int *)(fmpz_neg_f2 << 2)); }
      |      ^~~~~~~~~~
int *
int *
int
_3 = fmpz_neg_f2.1_2 << 2;
bug743.c:2:6: internal compiler error: ‘verify_gimple’ failed
0xd9532a verify_gimple_in_seq(gimple*)
        ../../trunk.git/gcc/tree-cfg.c:5183

The problem first seems to appear sometime between git hash 
e92d0ff6b5e6d4b9 and 38757aa88735ab2e, about 40 commits.

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
   Target Milestone|---                         |12.0
            Summary|ice:  error: type mismatch  |[12 Regression] ice:
                   |in ‘lshift_expr’            |error: type mismatch in
                   |                            |‘lshift_expr’

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-08-23
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Caused by r12-3073.

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

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

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

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Slightly cleaned up testcase:

int *
foo (const __PTRDIFF_TYPE__ l)
{
  return (int *) (l << 2);
}

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[roger at nextmovesoftware dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

Roger Sayle <roger at nextmovesoftware dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |roger at nextmovesoftware dot com
             Status|NEW                         |ASSIGNED
           Assignee|unassigned at gcc dot gnu.org      |roger at nextmovesoftware dot com

--- Comment #3 from Roger Sayle <roger at nextmovesoftware dot com> ---
Here's the fix that I'm currently testing...

 (simplify
  (convert (lshift:s@0 (convert:s@1 @2) INTEGER_CST@3))
- (if (tree_nop_conversion_p (type, TREE_TYPE (@0))
+ (if (INTEGRAL_TYPE_P (type)
+      && !POINTER_TYPE_P (type)
+      && tree_nop_conversion_p (type, TREE_TYPE (@0))
       && INTEGRAL_TYPE_P (TREE_TYPE (@2))
       && TYPE_PRECISION (TREE_TYPE (@2)) <= TYPE_PRECISION (type))
   (lshift (convert @2) @3)))

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

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

https://gcc.gnu.org/g:78fa5112b4c2dcd94b78ee79baddebbf14d6ad98

commit r12-3104-g78fa5112b4c2dcd94b78ee79baddebbf14d6ad98
Author: Roger Sayle <roger@nextmovesoftware.com>
Date:   Tue Aug 24 02:59:02 2021 +0100

    [Committed] PR middle-end/102029: Stricter typing in LSHIFT_EXPR sign
folding.

    My sincere apologies to everyone (again).  As diagnosed by
    Jakub Jelinek, my recent patch to fold the signedness of LSHIFT_EXPR
    needs to be careful not to attempt transforming a left shift in an
    integer type into an invalid left shift of a pointer type.

    2021-08-24  Roger Sayle  <roger@nextmovesoftware.com>
                Jakub Jelinek  <jakub@redhat.com>

    gcc/ChangeLog
            PR middle-end/102029
            * match.pd (shift transformations): Add an additional check for
            !POINTER_TYPE_P in the recently added left shift transformation.

    gcc/testsuite/ChangeLog
            PR middle-end/102029
            * gcc.dg/fold-convlshift-3.c: New test case.

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

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

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

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
(In reply to Roger Sayle from comment #3)
> Here's the fix that I'm currently testing...
> 
>  (simplify
>   (convert (lshift:s@0 (convert:s@1 @2) INTEGER_CST@3))
> - (if (tree_nop_conversion_p (type, TREE_TYPE (@0))
> + (if (INTEGRAL_TYPE_P (type)
> +      && !POINTER_TYPE_P (type)

either of both is redundant, I think keeping INTEGRAL_TYPE_P is better.

> +      && tree_nop_conversion_p (type, TREE_TYPE (@0))
>        && INTEGRAL_TYPE_P (TREE_TYPE (@2))
>        && TYPE_PRECISION (TREE_TYPE (@2)) <= TYPE_PRECISION (type))
>    (lshift (convert @2) @3)))

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[dje at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

David Edelsohn <dje at gcc dot gnu.org> changed:

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

--- Comment #6 from David Edelsohn <dje at gcc dot gnu.org> ---
This also had broken 32 bit powerpc, so I want to track any future updates.

[Bug middle-end/102029] [12 Regression] ice: error: type mismatch in ‘lshift_expr’

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=102029

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Roger Sayle <sayle@gcc.gnu.org>:

https://gcc.gnu.org/g:a2d9b558299df91e9c34a583eb0f0b14d1cacce9

commit r12-3163-ga2d9b558299df91e9c34a583eb0f0b14d1cacce9
Author: Roger Sayle <roger@nextmovesoftware.com>
Date:   Thu Aug 26 18:51:46 2021 +0100

    [Committed] Tidy up !POINTER_TYPE_P test in match.pd LSHIFT_EXPR folding

    As suggested by Richard Biener in the comments of PR middle-end/102029,
    the new test "INTEGRAL_TYPE_P (type) && !POINTER_TYPE_P (type) ..." is
    redundant, and just "INTEGRAL_TYPE_P (type)" is the preferred form.

    2021-08-26  Roger Sayle  <roger@nextmovesoftware.com>
                Richard Biener  <rguenther@suse.de>

    gcc/ChangeLog
            * match.pd (shift transformations): Remove a redundant
            !POINTER_TYPE_P check.