[Bug other/88733] New: [9 Regression] ICE in wide_int_to_tree_1, at tree.c:1561
doko at debian dot org
gcc-bugzilla@gcc.gnu.org
Mon Jan 7 08:58:00 GMT 2019
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88733
Bug ID: 88733
Summary: [9 Regression] ICE in wide_int_to_tree_1, at
tree.c:1561
Product: gcc
Version: 9.0
Status: UNCONFIRMED
Severity: normal
Priority: P3
Component: other
Assignee: unassigned at gcc dot gnu.org
Reporter: doko at debian dot org
Target Milestone: ---
seen with trunk 20190103, extracted from the rna-star package. Both -fopenmp
-flto are needed to see the ICE.
$ cat STAR.ii
class a {
public:
int b;
int c;
};
class e {
public:
long *d;
};
class h {
public:
e *f;
};
long g;
void k() {
a i;
h *j[i.b];
for (;;)
#pragma omp parallel num_threads(i.c)
g = j[0]->f->d[0];
}
$ g++ -std=c++11 -Wno-unused-result -fopenmp -flto -fstack-protector-strong -c
STAR.ii
during IPA pass: *free_lang_data
STAR.ii:21:1: internal compiler error: in wide_int_to_tree_1, at tree.c:1561
21 | }
| ^
0x5ce5ac wide_int_to_tree_1
../../src/gcc/tree.c:1561
0xdbed1b wide_int_to_tree(tree_node*, poly_int<1u,
generic_wide_int<wide_int_ref_storage<false, true> > > const&)
../../src/gcc/tree.c:1686
0xdbed1b build_int_cst(tree_node*, poly_int<1u, long>)
../../src/gcc/tree.c:1355
0x9107d3 round_up_loc(unsigned int, tree_node*, unsigned int)
../../src/gcc/fold-const.c:14374
0xb714f4 finalize_type_size
../../src/gcc/stor-layout.c:1931
0xb72e9b layout_type(tree_node*)
../../src/gcc/stor-layout.c:2596
0xdc02e1 build_array_type_1
../../src/gcc/tree.c:8302
0xdc55f3 fld_process_array_type
../../src/gcc/tree.c:5249
0xdc712a fld_incomplete_type_of
../../src/gcc/tree.c:5303
0xdc7068 fld_incomplete_type_of
../../src/gcc/tree.c:5285
0xdc9334 fld_simplified_type
../../src/gcc/tree.c:5374
0xdc9334 free_lang_data_in_decl
../../src/gcc/tree.c:5746
0xdc9334 free_lang_data_in_cgraph
../../src/gcc/tree.c:6175
0xdc9334 free_lang_data
../../src/gcc/tree.c:6216
0xdc9334 execute
../../src/gcc/tree.c:6288
Please submit a full bug report,
with preprocessed source if appropriate.
More information about the Gcc-bugs
mailing list