[Bug c/60804] New: Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[Bug c/60804] New: Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[andi-gcc at firstfloor dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

            Bug ID: 60804
           Summary: Another CilkPlus ICE in gimplify_expr, at
                    gimplify.c:8335
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
          Assignee: unassigned at gcc dot gnu.org
          Reporter: andi-gcc at firstfloor dot org

Created attachment 32577
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32577&action=edit
test case from csmith

I hacked csmith to add some _Cilk_spawn keywords and let it run for a bit.

The only new ICE that turned up was this (in various incarnations)

Here's the unreduced file from csmith.

x.c:178:35: internal compiler error: in gimplify_expr, at gimplify.c:8335
     if ((safe_div_func_uint8_t_u_u(_Cilk_spawn func_4(((--(*l_9)) , g_10),
((((
void*)0 != g_13) , (&g_14[1][0] == ((safe_add_func_uint32_t_u_u(g_14[1][0],
(g_1
9 || (((*l_2079) = _Cilk_spawn
func_20(((safe_rshift_func_uint16_t_u_s((~((*l_39
) = ((safe_rshift_func_int16_t_s_u(((*l_37) =
(safe_add_func_int8_t_s_s((_Cilk_s
pawn func_28(g_31, g_32) <= g_19), 0x5FL))), 11)) , 0xC254L))), 5)) & g_19)))
!=
 (void*)0)))) , &g_33))) | l_2080), g_420[4][0][1], g_847), l_2080)))
                                   ^
0x79bad0 gimplify_expr(tree_node**, gimple_statement_base**,
gimple_statement_ba
se**, bool (*)(tree_node*), int)
        ../../gcc/gcc/gimplify.c:8335
0x799d38 gimplify_expr(tree_node**, gimple_statement_base**,
gimple_statement_ba
se**, bool (*)(tree_node*), int)
        ../../gcc/gcc/gimplify.c:7715
0x7a0ed3 gimplify_call_expr
        ../../gcc/gcc/gimplify.c:2354
0x79b1fc gimplify_expr(tree_node**, gimple_statement_base**,
gimple_statement_ba
se**, bool (*)(tree_node*), int)
        ../../gcc/gcc/gimplify.c:7402
0x79aacc gimplify_expr(tree_node**, gimple_statement_base**,
gimple_statement_ba
...

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2014-04-10
                 CC|                            |jakub at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Can't reproduce this, supposedly starting with r208382 when this is rejected as
invalid use of _Cilk_spawn.  So, do you have any testcases that ICE even with
current trunk?

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[andi-gcc at firstfloor dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

Andi Kleen <andi-gcc at firstfloor dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #32577|0                           |1
        is obsolete|                            |

--- Comment #2 from Andi Kleen <andi-gcc at firstfloor dot org> ---
Created attachment 32580
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32580&action=edit
this one still fails with current trunk

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Reduced testcase:
int bar (void);

int
foo ()
{
  if (_Cilk_spawn bar ())
    return 5;
  return 0;
}

Whether this is valid Cilk+ or not, no idea.

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
According to
http://software.intel.com/sites/products/documentation/doclib/stdxe/2013/composerxe/compiler/cpp-win/index.htm
as well as
http://www.cilkplus.org/sites/default/files/open_specifications/cilk_plus_language_specification_0_9.pdf
this isn't valid (but we should then diagnose it), but whether these two (and
which one of them) are the latest Cilk+ spec is unclear to me.

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[evstupac at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

Stupachenko Evgeny <evstupac at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |evstupac at gmail dot com

--- Comment #5 from Stupachenko Evgeny <evstupac at gmail dot com> ---
Yes it is not valid.
Please find latest spec at:
https://www.cilkplus.org/sites/default/files/open_specifications/Intel_Cilk_plus_lang_spec_1.2.htm

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[mpolacek at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

--- Comment #6 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
The following should disallow this in the C FE.

diff --git a/gcc/c/c-parser.c b/gcc/c/c-parser.c
index 5653e49..cbbb272 100644
--- a/gcc/c/c-parser.c
+++ b/gcc/c/c-parser.c
@@ -5153,6 +5153,9 @@ c_parser_if_statement (c_parser *parser)
   /* If the if statement contains array notations, then we expand them.  */
   if (flag_cilkplus && contains_array_notation_expr (if_stmt))
     if_stmt = fix_conditional_array_notations (if_stmt);
+  /* But Cilk_spawn in controlling expression is invalid.  */
+  if (flag_cilkplus && contains_cilk_spawn_stmt (cond))
+    error_at (loc, "%<_Cilk_spawn%> in if statement is not permitted");
   add_stmt (if_stmt);
 }

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

--- Comment #7 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
But then it is not valid in many other places (for/while/do condition, etc.).
Perhaps if _Cilk_spawn is allowed in the grammer only in 3 specific cases it
might be better to disallow it everywhere and check for it only in the 3 listed
cases (whole statement, var initializer and rhs of assignment (but for the last
one it has to be only toplevel assignment I guess, a = b = _Cilk_spawn foo ();
is presumably also invalid, ditto a + (b = _Cilk_spawn foo ()); etc.

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[andi-gcc at firstfloor dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

--- Comment #8 from Andi Kleen <andi-gcc at firstfloor dot org> ---
I went through my collection of gimplify:8335 from the generator.

Not all of them are special statements. So some more general check would be
needed.

Some examples:

     (*l_11) = (g_9 = _Cilk_spawn func_2(l_4));


(*g_682) = _Cilk_spawn func_2(g_5, (((((g_5 |
((safe_add_func_int32_t_s_s((((*g_682) = _Cilk_spawn func_8(g_5, _Cilk_spawn
func_12(g_5), (l_2841 = l_2841))) != (g_3060 = l_3057)),
(safe_mod_func_uint16_t_u_u(g_868.f0, g_2089.f3)))) && 1UL)) , l_3063) & g_88)
| (*l_3057)) , l_3064));

l_1679 ^= _Cilk_spawn func_10((((safe_lshift_func_int8_t_s_u(((((*l_1669) ^=
(0x4BAFC8D6799626DALL || (((~((*l_1668) = (safe_rshift_func_uint16_t_u_u(((g_3
== (safe_mul_func_int16_t_s_s((safe_rshift_func_uint16_t_u_s(g_3, 2)),
((safe_mul_func_uint8_t_u_u(_Cilk_spawn func_24(&g_3, ((*l_32) &=
(safe_rshift_func_int8_t_s_s(g_7, g_6[8][1][3]))), ((((_Cilk_spawn
func_34(_Cilk_spawn func_37(((-1L) > (*l_2)), ((*l_925) = _Cilk_spawn
func_43(g_6[7][3][2])), &g_198[1][1][7], l_2, (*l_2)), l_5[2][0][1]) && 0x43L)
, g_317) ^ l_1665[0]) < g_604), l_2, l_5[4][8][2]), 1UL)) < 0xF63AL)))) <=
l_1667), (*l_2))))) | 2L) , 1L))) > (-1L)) | l_1671[0][4]), 2)) , 1UL) >
0x8FL), l_5[4][3][1], (*l_2));

l_3901 |= (safe_sub_func_int16_t_s_s(((((safe_div_func_uint64_t_u_u(_Cilk_spawn
func_6(l_12[0][3],
(safe_add_func_uint64_t_u_u(((safe_sub_func_int32_t_s_s(_Cilk_spawn
func_17(g_20[0], (safe_lshift_func_int8_t_s_u(((_Cilk_spawn
func_23(((safe_add_func_uint32_t_u_u(((l_12[0][3] | l_12[8][2]) ,
(((safe_mod_func_int32_t_s_s(_Cilk_spawn func_30(g_20[0], (0x6E31BB00L >= (l_34
& 0L)), (((safe_div_func_uint64_t_u_u(g_20[0], l_12[5][2])) , l_12[0][3]) ,
l_12[0][3])), l_2523)) , l_12[0][3]) , 0xFC5A8F50L)), l_2523)) < l_12[0][3]),
(*g_1492)) > (-1L)) ^ (*g_1492)), l_2690))), l_2690)) | 0xBE8DAE3C6AAFCCBCLL),
l_2523)), l_2523, l_12[8][3], l_2523), l_3900)) != (*g_1492)) && l_12[8][4]) ,
(**g_3267)), l_12[0][3]));


(*g_20) = _Cilk_spawn func_2(g_7, l_8, ((*l_3296) = _Cilk_spawn func_9(g_7,
((**g_2525) = (((g_15 = g_7) | (((((safe_rshift_func_int16_t_s_u(_Cilk_spawn
func_18(g_20), 7)) >=
(safe_mod_func_int64_t_s_s((safe_mod_func_uint64_t_u_u(l_3288, ((*l_3296) &=
((l_3289 ||
(safe_mul_func_int16_t_s_s(((((safe_unary_minus_func_int8_t_s(((((**g_2282) ^=
(&g_528[0][0][1] != l_3293)) , l_3294[3][1][1]) & (***l_3293)))) , (***g_243))
> (*g_2850)) > (*g_242)), l_3295))) > (***l_3293))))), (***l_3293)))) , l_3297)
, 1L) > 0x65L)) ^ l_3298)), (*g_242), (***l_3293))), l_3324);

(**g_388) = _Cilk_spawn func_8((_Cilk_spawn func_10((g_12 , g_13)) | ((0x439DL
|| 0L) ^ (((*l_1983) = g_5[g_4]) , (((~((safe_mul_func_uint16_t_u_u(g_5[g_4],
g_1613.f6)) , (((((g_161.f1 , &g_291) == l_1986) , (*l_1983)) != (*l_1983)) ||
(*g_379)))) ^ l_1987) | l_1987)))));

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[ak at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

ak at gcc dot gnu.org changed:

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

--- Comment #9 from ak at gcc dot gnu.org ---
This still ICEs with gcc version 5.0.0 20140926 (experimental) (GCC)

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[ak at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

--- Comment #10 from ak at gcc dot gnu.org ---
Reduced test case. It's probably invalid cilk, but gcc shouldn't ICE:

fn1() {
  if (_Cilk_spawn func_2())
    ;
}

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[ak at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

--- Comment #11 from ak at gcc dot gnu.org ---
Author: ak
Date: Tue Nov 11 05:10:58 2014
New Revision: 217336

URL: https://gcc.gnu.org/viewcvs?rev=217336&root=gcc&view=rev
Log:
Error out for Cilk_spawn or array expression in forbidden places

_Cilk_spawn or Cilk array expressions are only allowed on their own,
but not in for(), if(), switch, do, while, goto, etc.
The C parser didn't always check for that, which lead to ICEs earlier
for invalid code.

Add a generic helper that checks this and call it where needed
in the C frontend.

I chose to allow spawn/array for for init and increment expressions.
While the Cilk spec could be interpreted to forbid it there too
there didn't seem any reason to not allow it.

One dark corner is spawn, array in statement expressions not at
the end. Right now that's forbidden too.

gcc/c-family/:

2014-11-10  Andi Kleen  <ak@linux.intel.com>

    PR c/60804
    * c-common.h (check_no_cilk): Declare.
    * cilk.c (get_error_location): New function.
    (check_no_cilk): Dito.

gcc/c/:

2014-11-10  Andi Kleen  <ak@linux.intel.com>

    PR c/60804
    * c-parser.c (c_parser_statement_after_labels): Call
    check_no_cilk.
    (c_parser_if_statement): Dito.
    (c_parser_switch_statement): Dito.
    (c_parser_while_statement): Dito.
    (c_parser_do_statement): Dito.
    (c_parser_for_statement): Dito.
    * c-typeck.c (c_finish_loop): Dito.

Modified:
    trunk/gcc/c-family/ChangeLog
    trunk/gcc/c-family/c-common.h
    trunk/gcc/c-family/cilk.c
    trunk/gcc/c/ChangeLog
    trunk/gcc/c/c-parser.c
    trunk/gcc/c/c-typeck.c

[Bug c/60804] Another CilkPlus ICE in gimplify_expr, at gimplify.c:8335

[andi-gcc at firstfloor dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=60804

Andi Kleen <andi-gcc at firstfloor dot org> changed:

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

--- Comment #12 from Andi Kleen <andi-gcc at firstfloor dot org> ---
Should be all fixed now in mainline.