[Bug tree-optimization/110269] New: 13/14 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[Bug tree-optimization/110269] New: 13/14 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[theodort at inf dot ethz.ch]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

            Bug ID: 110269
           Summary: 13/14 Regression] Missed Dead Code Elimination when
                    using __builtin_unreachable since
                    r13-4607-g2dc5d6b1e7e
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: tree-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: theodort at inf dot ethz.ch
  Target Milestone: ---

https://godbolt.org/z/5qqPMsxEh

Given the following code:

void foo(void);
static int a = 1, c;
static int *b = &a;
static int **d = &b;
static int ***e = &d;
void __assert_fail() __attribute__((__noreturn__));
static int f() {
    if (a) return a;
    for (; c;) *e = 0;
    if (b) __assert_fail();
    return 6;
}
int main() {
    if (f()) {
        *d = 0;
        if (b == 0)
            ;
        else {
            __builtin_unreachable();
            __assert_fail();
        }
    }
    if (b == 0)
        ;
    else
        foo();
    ;
}

gcc-trunk -O3 generates:

main:
        subq    $8, %rsp
        movl    a(%rip), %eax
        testl   %eax, %eax
        jne     .L2
        cmpq    $0, b(%rip)
        jne     .L6
.L2:
        movq    d(%rip), %rax
        movq    $0, (%rax)
        cmpq    $0, b(%rip)
        je      .L3
        call    foo
.L3:
        xorl    %eax, %eax
        addq    $8, %rsp
        ret
.L6:
        xorl    %eax, %eax
        call    __assert_fail

gcc-12.3.0 -O3 generates:

main:
        movl    a(%rip), %edx
        testl   %edx, %edx
        jne     .L2
        cmpq    $0, b(%rip)
        jne     .L7
.L2:
        movq    d(%rip), %rax
        movq    $0, (%rax)
        xorl    %eax, %eax
        ret
.L7:
        pushq   %rax
        xorl    %eax, %eax
        call    __assert_fail

Bisects to r13-4607-g2dc5d6b1e7e

[Bug tree-optimization/110269] [13/14 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Target Milestone|---                         |13.2
     Ever confirmed|0                           |1
           Keywords|                            |missed-optimization
   Last reconfirmed|                            |2023-06-15

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
GCC 12.3 ccp2:
```
Folding statement: _20 = _15;
Queued stmt for removal.  Folds to: _15
Folding statement: _1 = _20;
Queued stmt for removal.  Folds to: _15
Folding statement: if (_1 != 0)
Matching expression match.pd:2119, gimple-match.cc:819
Matching expression match.pd:2122, gimple-match.cc:892
Matching expression match.pd:2129, gimple-match.cc:952
which is likely CONSTANT
Matching expression match.pd:2119, generic-match.cc:693
Matching expression match.pd:2122, generic-match.cc:753
Matching expression match.pd:2129, generic-match.cc:776
Folding predicate _15 != 0 to 1
Folded into: if (1 != 0)
```

GCC 13+ ccp2:
```
Folding statement: _20 = _15;
Queued stmt for removal.  Folds to: _15
Folding statement: _1 = _20;
Queued stmt for removal.  Folds to: _15
Folding statement: if (_1 != 0)
Matching expression match.pd:2404, gimple-match.cc:852
Matching expression match.pd:2407, gimple-match.cc:925
Matching expression match.pd:2414, gimple-match.cc:985
which is likely CONSTANT
Matching expression match.pd:2404, gimple-match.cc:852
Matching expression match.pd:2407, gimple-match.cc:925
Matching expression match.pd:2414, gimple-match.cc:985
Folded into: if (_15 != 0)
```


oh it was the ccp_fold change that misses it ...
But why I have not looked.

[Bug tree-optimization/110269] [13/14 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Maybe gimple_fold_stmt_to_constant_1 should handle GIMPLE_COND ...

[Bug tree-optimization/110269] [13/14 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

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

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

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
It does handle GIMPLE_COND.  What the change did was avoid using GENERIC
fold_const and instead resort to match.pd only.  I can't see how it
arrived at folding it before.  We (now) have

Visiting PHI node: _15 = PHI <a.3_11(2), 6(5)>
        Argument #0 (2 -> 7 executable)
        a.3_11  Value: CONSTANT a.3_11
        Argument #1 (5 -> 7 executable)
        6       Value: CONSTANT 6

    PHI node value: VARYING

Lattice value changed to VARYING.  Adding SSA edges to worklist.

so there's no nonzero bits info.  The input to CCP2 is almost the
same:

 int main ()
 {
-  int D.2014;
+  int D.2786;
   int _1;
   int * * d.0_2;
   int * b.1_3;
@@ -46,7 +46,7 @@
   __assert_fail ();

   <bb 7> [local count: 1073458354]:
-  # RANGE ~[0, 0]
+  # RANGE [irange] int [-INF, -1][1, +INF]
   # _15 = PHI <a.3_11(2), 6(5)>
   _20 = _15;
   _1 = _20;

so it looks like range representation changes make a difference here?

Ah, so match.pd lacks a simple (ne @0 integer_zerop) check using
tree_single_nonzero_warnv_p.  We have

 (simplify
  (cmp (convert? addr@0) integer_zerop)
  (if (tree_single_nonzero_warnv_p (@0, NULL))
   { constant_boolean_node (cmp == NE_EXPR, type); }))

[Bug tree-optimization/110269] [13/14 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

--- Comment #4 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:827e208fa64771f15fc8e53970a2297e637277b5

commit r14-1880-g827e208fa64771f15fc8e53970a2297e637277b5
Author: Richard Biener <rguenther@suse.de>
Date:   Fri Jun 16 08:55:12 2023 +0200

    tree-optimization/110269 - restore missed condition folding

    The following makes sure we optimize x != 0 using range info
    via tree_expr_nonzero_p via match.pd.

            PR tree-optimization/110269
            * fold-const.cc (fold_binary_loc): Merge x != 0 folding
            with tree_expr_nonzero_p ...
            * match.pd (cmp (convert? addr@0) integer_zerop): With this
            pattern.

            * gcc.dg/tree-ssa/pr110269.c: New testcase.

[Bug tree-optimization/110269] [13 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269
Bug 110269 depends on bug 110278, which changed state.

Bug 110278 Summary: [14 Regression] FAIL: gcc.dg/tree-ssa/pr103257-1.c scan-tree-dump-times optimized "link_error" 0 from r14-1880-g827e208fa64771
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110278

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

[Bug tree-optimization/110269] [13 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.2                        |13.3

--- Comment #5 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 13.2 is being released, retargeting bugs to GCC 13.3.

[Bug tree-optimization/110269] [13 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

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

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

[Bug tree-optimization/110269] [13 Regression] Missed Dead Code Elimination when using __builtin_unreachable since r13-4607-g2dc5d6b1e7e

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110269

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.3                        |13.4

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 13.3 is being released, retargeting bugs to GCC 13.4.