https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79529
Bug ID: 79529
Summary: [7 Regression] ICE in is_maybe_undefined
(tree-ssa-loop-unswitch.c:162)
Product: gcc
Version: unknown
Status: UNCONFIRMED
Keywords: ice-on-valid-code
Severity: normal
Priority: P3
Component: tree-optimization
Assignee: unassigned at gcc dot gnu.org
Reporter: marxin at gcc dot gnu.org
CC: aldyh at gcc dot gnu.org, jamborm at gcc dot gnu.org
Target Milestone: ---
Building Firefox with -flto and -O3, one can get:
/home/marxin/Programming/gecko-dev/intl/icu/source/i18n/dtptngen.cpp:1379:1:
internal compiler error: Segmentation fault
DateTimePatternGenerator::getBestAppending(int32_t missingFields, int32_t
flags, UDateTimePatternMatchOptions options) {
^
0xcaea52 crash_signal
../../gcc/toplev.c:333
0x888c28 dominated_by_p(cdi_direction, basic_block_def const*, basic_block_def
const*)
../../gcc/dominance.c:1117
0xe465f0 is_maybe_undefined
../../gcc/tree-ssa-loop-unswitch.c:162
0xe46863 tree_may_unswitch_on
../../gcc/tree-ssa-loop-unswitch.c:222
0xe46beb tree_unswitch_single_loop
../../gcc/tree-ssa-loop-unswitch.c:322
0xe46379 tree_ssa_unswitch_loops()
../../gcc/tree-ssa-loop-unswitch.c:102
0xe48337 execute
../../gcc/tree-ssa-loop-unswitch.c:935
Problem is that:
if (dominated_by_p (CDI_DOMINATORS, loop->header, gimple_bb (def)))
where def is a gimple nop, which has gimple_bb (def) == NULL.
I guess following patch should fix that:
diff --git a/gcc/tree-ssa-loop-unswitch.c b/gcc/tree-ssa-loop-unswitch.c
index 4ef3a6bf80a..5716c99e58c 100644
--- a/gcc/tree-ssa-loop-unswitch.c
+++ b/gcc/tree-ssa-loop-unswitch.c
@@ -141,6 +141,9 @@ is_maybe_undefined (const tree name, gimple *stmt, struct
loop *loop)
gimple *def = SSA_NAME_DEF_STMT (t);
+ if (gimple_nop_p (def))
+ return true;
+
/* Check that all the PHI args are fully defined. */
if (gphi *phi = dyn_cast <gphi *> (def))
{
