[Bug middle-end/59711] New: ICE in force_constant_size, at gimplify.c:619 (nested function and variably-modified type)

[Bug middle-end/59711] New: ICE in force_constant_size, at gimplify.c:619 (nested function and variably-modified type)

[fweimer at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

            Bug ID: 59711
           Summary: ICE in force_constant_size, at gimplify.c:619 (nested
                    function and variably-modified type)
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: middle-end
          Assignee: unassigned at gcc dot gnu.org
          Reporter: fweimer at redhat dot com

Created attachment 31764
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31764&action=edit
u.c

The attached test case ICEs on x86_64-redhat-linux-gnu with the following
backtrace:

#2  0x00000000007f5875 in force_constant_size (var=0x7ffff032a8e8) at
../../trunk/gcc/gimplify.c:619
#3  gimple_add_tmp_var (tmp=0x7ffff032a8e8) at ../../trunk/gcc/gimplify.c:638
#4  0x00000000007d58fb in create_tmp_var (type=type@entry=0x7ffff0328000,
prefix=<optimized out>)
    at ../../trunk/gcc/gimple-expr.c:509
#5  0x00000000007f5b15 in create_tmp_from_val (is_formal=true,
val=0x7ffff0324990) at ../../trunk/gcc/gimplify.c:457
#6  lookup_tmp_var (is_formal=true, val=0x7ffff0324990) at
../../trunk/gcc/gimplify.c:479
#7  internal_get_tmp_var (val=0x7ffff0324990, pre_p=0x7fffffffd968,
post_p=<optimized out>, is_formal=<optimized out>)
    at ../../trunk/gcc/gimplify.c:523
#8  0x00000000007f627e in gimplify_expr (expr_p=expr_p@entry=0x7ffff03249d8,
pre_p=pre_p@entry=0x7fffffffd968, 
    post_p=post_p@entry=0x7fffffffd518, gimple_test_f=<optimized out>,
fallback=fallback@entry=3)
    at ../../trunk/gcc/gimplify.c:8322
#9  0x00000000007fdea2 in gimplify_compound_lval
(expr_p=expr_p@entry=0x7ffff0305ac0, 
    pre_p=pre_p@entry=0x7fffffffd968, post_p=post_p@entry=0x7fffffffd518,
fallback=fallback@entry=1)
    at ../../trunk/gcc/gimplify.c:1976
#10 0x00000000007f73cd in gimplify_expr (expr_p=expr_p@entry=0x7ffff0305ac0,
pre_p=pre_p@entry=0x7fffffffd968, 
    post_p=post_p@entry=0x7fffffffd518, gimple_test_f=<optimized out>,
fallback=fallback@entry=1)
    at ../../trunk/gcc/gimplify.c:7383
#11 0x00000000008029bb in gimplify_modify_expr
(expr_p=expr_p@entry=0x7fffffffd6b0, pre_p=pre_p@entry=0x7fffffffd968, 
    post_p=post_p@entry=0x7fffffffd518, want_value=<optimized out>) at
../../trunk/gcc/gimplify.c:4493
#12 0x00000000007f771a in gimplify_expr (expr_p=0x7fffffffd6b0,
pre_p=pre_p@entry=0x7fffffffd968, 
    post_p=0x7fffffffd518, post_p@entry=0x0,
gimple_test_f=gimple_test_f@entry=0x7efef0 <is_gimple_stmt(tree)>, 
    fallback=fallback@entry=0) at ../../trunk/gcc/gimplify.c:7431
#13 0x00000000007fb0f7 in gimplify_stmt (stmt_p=stmt_p@entry=0x7fffffffd6b0,
seq_p=seq_p@entry=0x7fffffffd968)
    at ../../trunk/gcc/gimplify.c:5334
#14 0x00000000007f970c in gimplify_and_add (seq_p=0x7fffffffd968,
t=0x7ffff0305aa0) at ../../trunk/gcc/gimplify.c:384
#15 gimplify_return_expr (pre_p=0x7fffffffd968, stmt=0x7ffff0314a60) at
../../trunk/gcc/gimplify.c:1234
#16 gimplify_expr (expr_p=0x7ffff0312bc8, pre_p=pre_p@entry=0x7fffffffd968,
post_p=0x7fffffffd698, post_p@entry=0x0, 
    gimple_test_f=gimple_test_f@entry=0x7efef0 <is_gimple_stmt(tree)>,
fallback=fallback@entry=0)
    at ../../trunk/gcc/gimplify.c:7671
#17 0x00000000007fb0f7 in gimplify_stmt (stmt_p=<optimized out>,
seq_p=seq_p@entry=0x7fffffffd968)
    at ../../trunk/gcc/gimplify.c:5334
#18 0x00000000007f7ee4 in gimplify_statement_list (pre_p=0x7fffffffd968,
expr_p=0x7ffff0324a10)
    at ../../trunk/gcc/gimplify.c:1405
#19 gimplify_expr (expr_p=0x7ffff0324a10, pre_p=pre_p@entry=0x7fffffffd968,
post_p=0x7fffffffd818, post_p@entry=0x0, 
    gimple_test_f=gimple_test_f@entry=0x7efef0 <is_gimple_stmt(tree)>,
fallback=fallback@entry=0)
    at ../../trunk/gcc/gimplify.c:7839
#20 0x00000000007fb0f7 in gimplify_stmt (stmt_p=stmt_p@entry=0x7ffff0324a10,
seq_p=seq_p@entry=0x7fffffffd968)
    at ../../trunk/gcc/gimplify.c:5334
#21 0x00000000007fc01e in gimplify_bind_expr
(expr_p=expr_p@entry=0x7ffff030a698, pre_p=pre_p@entry=0x7fffffffdb58)
    at ../../trunk/gcc/gimplify.c:1072
#22 0x00000000007f7856 in gimplify_expr (expr_p=0x7ffff030a698,
pre_p=pre_p@entry=0x7fffffffdb58, 
    post_p=0x7fffffffda28, post_p@entry=0x0,
gimple_test_f=gimple_test_f@entry=0x7efef0 <is_gimple_stmt(tree)>, 
    fallback=fallback@entry=0) at ../../trunk/gcc/gimplify.c:7621
#23 0x00000000007fb0f7 in gimplify_stmt (stmt_p=stmt_p@entry=0x7ffff030a698,
seq_p=seq_p@entry=0x7fffffffdb58)
    at ../../trunk/gcc/gimplify.c:5334
#24 0x00000000007fc9d0 in gimplify_body (fndecl=fndecl@entry=0x7ffff030a600,
do_parms=do_parms@entry=true)
    at ../../trunk/gcc/gimplify.c:8531
#25 0x00000000007fd017 in gimplify_function_tree
(fndecl=fndecl@entry=0x7ffff030a600)
    at ../../trunk/gcc/gimplify.c:8684
#26 0x000000000068d638 in analyze_function (node=node@entry=0x7ffff032b000) at
../../trunk/gcc/cgraphunit.c:649
#27 0x000000000068e9f5 in analyze_functions () at
../../trunk/gcc/cgraphunit.c:1017
#28 0x0000000000690036 in finalize_compilation_unit () at
../../trunk/gcc/cgraphunit.c:2310
#29 0x0000000000552704 in c_write_global_declarations () at
../../trunk/gcc/c/c-decl.c:10400
#30 0x00000000009a661d in compile_file () at ../../trunk/gcc/toplev.c:561
#31 0x00000000009a84b2 in do_compile () at ../../trunk/gcc/toplev.c:1887
#32 toplev_main (argc=21, argv=0x7fffffffde48) at ../../trunk/gcc/toplev.c:1963
#33 0x00007ffff6933b45 in __libc_start_main (main=0x535ca0 <main(int, char**)>,
argc=21, ubp_av=0x7fffffffde48, 
    init=<optimized out>, fini=<optimized out>, rtld_fini=<optimized out>,
stack_end=0x7fffffffde38)
    at libc-start.c:274
#34 0x0000000000535ce5 in _start ()

The crash happens at:

617      max_size = max_int_size_in_bytes (TREE_TYPE (var));
618    
619      gcc_assert (max_size >= 0);

var looks like this:

 <var_decl 0x7ffff032a8e8 D.1765
    type <record_type 0x7ffff0328000 S sizes-gimplified type_0 type_1 BLK
        size <var_decl 0x7ffff032a7b8 D.1763 type <integer_type 0x7ffff01fe150
bitsizetype>
            used unsigned ignored TI file /tmp/u.c line 3 col 10
            size <integer_cst 0x7ffff02000c0 constant 128>
            unit size <integer_cst 0x7ffff02000e0 constant 16>
            align 128 context <function_decl 0x7ffff030a600 f> chain <var_decl
0x7ffff032a720 D.1762>>
        unit size <var_decl 0x7ffff032a390 D.1756 type <integer_type
0x7ffff01fe0a8 sizetype>
            used unsigned ignored DI file /tmp/u.c line 3 col 10
            size <integer_cst 0x7ffff0200040 constant 64>
            unit size <integer_cst 0x7ffff0200060 constant 8>
            align 64 context <function_decl 0x7ffff030a600 f> chain <var_decl
0x7ffff032a2f8 D.1755>>
        align 32 symtab 0 alias set -1 canonical type 0x7ffff0328000
        fields <field_decl 0x7ffff032a000 a type <array_type 0x7ffff0328150>
            decl_0 BLK file /tmp/u.c line 4 col 9 size <var_decl 0x7ffff032a2f8
D.1755> unit size <var_decl 0x7ffff032a390 D.1756>
            align 32 offset_align 128
            offset <integer_cst 0x7ffff0200080 constant 0>
            bit offset <integer_cst 0x7ffff0200100 constant 0> context
<record_type 0x7ffff0328000 S>> context <function_decl 0x7ffff030a600 f>
        chain <type_decl 0x7ffff0329000 D.1744>>
    used ignored BLK file /tmp/u.c line 13 col 12 size <var_decl 0x7ffff032a7b8
D.1763> unit size <var_decl 0x7ffff032a390 D.1756>
    align 32>

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 (nested function and variably-modified type)

[fweimer at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

--- Comment #1 from Florian Weimer <fweimer at redhat dot com> ---
Created attachment 31765
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31765&action=edit
funcpointer.c

Test case without nested function.

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-01-07
                 CC|                            |mpolacek at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #2 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Confirmed.  Even GCC 3.4 ICEs on it.

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[ebotcazou at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

Eric Botcazou <ebotcazou at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #31765|0                           |1
        is obsolete|                            |
                 CC|                            |ebotcazou at gcc dot gnu.org

--- Comment #3 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
Created attachment 31768
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31768&action=edit
More reduced testcase

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

--- Comment #4 from Richard Biener <rguenth at gcc dot gnu.org> ---
int f(int n)
{
  struct S {
    int a[n];
  };

  struct S g(void) 
  {
    struct S s;
    s.a[0] = 0;
    return s;
  }
  return g ().a[0];
}

I believe this isn't fixable with the GIMPLE IL as-is - without lowering
to the fact that g () returns by storing to a return slot.  Not sure why
3.4 ICEd (2.95.2 also ICEs btw ...).

Well, maybe we can fix it if we allow creating of a variable-size temporary
during gimplification.  That is, the gimple code needs to look like

  tem = WITH_SIZE_EXPR <g (), sizeof (S)>;
  return tem.a[0];

which of course then has the issue that we can't have WITH_SIZE_EXPR around
calls.  Thus really the only way is to lower the return slot fact early.

  g (&tem);
  return tem.a[0];

Eventually not that way but using a special slot on the call stmt similar
to the chain / fn pointer.  Enlarges all call stmts, of course.

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[ebotcazou at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

--- Comment #5 from Eric Botcazou <ebotcazou at gcc dot gnu.org> ---
> I believe this isn't fixable with the GIMPLE IL as-is - without lowering
> to the fact that g () returns by storing to a return slot.  Not sure why
> 3.4 ICEd (2.95.2 also ICEs btw ...).

Because historically you cannot create temporaries with variable size in GCC.

> Well, maybe we can fix it if we allow creating of a variable-size temporary
> during gimplification.  That is, the gimple code needs to look like
> 
>   tem = WITH_SIZE_EXPR <g (), sizeof (S)>;
>   return tem.a[0];
> 
> which of course then has the issue that we can't have WITH_SIZE_EXPR around
> calls.  Thus really the only way is to lower the return slot fact early.
> 
>   g (&tem);
>   return tem.a[0];

Of course this situation occurs very frequently in Ada and we explicitly create
the temporary in gigi.

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[fweimer at redhat dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

--- Comment #6 from Florian Weimer <fweimer at redhat dot com> ---
Created attachment 31772
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31772&action=edit
pr59711.adb

Here's the same thing in Ada (where it actually works—it's actually fairly
idiomatic code, I believe).  The generated GIMPLE looks like this (with GCC
4.8.2 from Fedora):

PR59711 (const natural___XDLU_0__2147483647 n)
{
  integer n.3;
  sizetype D.2600;
  bitsizetype D.2601;
  bitsizetype D.2602;
  struct pr59711__s * R.4;
  character D.2604;
  void * saved_stack.5;
  natural___XDLU_0__2147483647 pr59711__s__T1b___U;
  typedef pr59711__s__T1b___XDL_1 pr59711__s__T1b___XDL_1;
  typedef signed long struct signed long;
  typedef character pr59711__s__T2b[1:D.2600];
  typedef struct pr59711__s pr59711__s;
  struct pr59711__s R.0 [value-expr: *R.4];

  saved_stack.5 = .builtin_stack_save ();
  try
    {
      pr59711__s__T1b___U = n;
      n.3 = (integer) n;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      D.2601 = (bitsizetype) D.2600;
      D.2602 = D.2601 * 8;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      D.2601 = (bitsizetype) D.2600;
      D.2602 = D.2601 * 8;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      D.2601 = (bitsizetype) D.2600;
      D.2602 = D.2601 * 8;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      D.2601 = (bitsizetype) D.2600;
      D.2602 = D.2601 * 8;
      n.3 = (integer) n;
      D.2600 = (sizetype) n.3;
      R.4 = .builtin_alloca_with_align (D.2600, 8);
      *R.4 = pr59711.f (); [return slot optimization]
      D.2604 = R.4->a[1]{lb: 1 sz: 1};
      return D.2604;
    }
  finally
    {
      .builtin_stack_restore (saved_stack.5);
    }
}
>From gcc-bugs-return-439625-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Wed Jan 08 12:18:42 2014
Return-Path: <gcc-bugs-return-439625-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 27207 invoked by alias); 8 Jan 2014 12:18:42 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 27154 invoked by uid 48); 8 Jan 2014 12:18:39 -0000
From: "vries at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug tree-optimization/59715] [4.7/4.8/4.9 Regression] wrong code at -Os and above on x86_64-linux-gnu
Date: Wed, 08 Jan 2014 12:18:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: tree-optimization
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords: wrong-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: vries at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P2
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: 4.7.4
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: cc
Message-ID: <bug-59715-4-T3K0woO2bw@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-59715-4@http.gcc.gnu.org/bugzilla/>
References: <bug-59715-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-01/txt/msg00767.txt.bz2
Content-length: 472

http://gcc.gnu.org/bugzilla/show_bug.cgi?idY715

vries at gcc dot gnu.org changed:

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

--- Comment #6 from vries at gcc dot gnu.org ---
The redundant store elimination in tree-ssa-dom.c:optimize_stmt (introduced in
PR45705) removes the non-redundant store b = 0.

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
      *R.4 = pr59711.f (); [return slot optimization]

ah, so we don't require a WITH_SIZE_EXPR on a RSO call.  Obviously.  So it
should be fixable either by the gimplifier or by the frontends (as Ada
does it).

[Bug middle-end/59711] ICE in force_constant_size, at gimplify.c:619 with variably-modified return type

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=59711

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |6.0
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=70093
         Resolution|---                         |FIXED
           Keywords|                            |ice-on-valid-code
             Status|NEW                         |RESOLVED

--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Fixed in GCC 6, by the patch which fixed PR 70093.