[Bug tree-optimization/83867] [8 Regression] ICE: Segmentation fault in nested_in_vect_loop_p
rguenth at gcc dot gnu.org
gcc-bugzilla@gcc.gnu.org
Tue Jan 16 09:45:00 GMT 2018
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83867
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
CC| |rsandifo at gcc dot gnu.org
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
The issue is that vectorized stmts replacing scalar stmts via
vect_finish_replace_stmt leave the scalar stmt def dangling in no BB. Here
9525 /* Handle inner-loop stmts whose DEF is used in the loop-nest that
9526 is being vectorized, but outside the immediately enclosing loop.
*/
9527 if (vec_stmt
9528 && STMT_VINFO_LOOP_VINFO (stmt_info)
9529 && nested_in_vect_loop_p (LOOP_VINFO_LOOP (
9530 STMT_VINFO_LOOP_VINFO (stmt_info)),
stmt)
passes the dangling 'stmt' to nested_in_vect_loop_p.
Replacing the scalar def in this way is somewhat hacky given the caller still
has its hand on 'stmt' :/
We could fix this particular fallout but why do we need to re-use the original
SSA def? This is from vectorize_fold_left_reduction.
Anyway, I see it is convenient but I do expect more fallout from this.
Index: gcc/tree-vect-stmts.c
===================================================================
--- gcc/tree-vect-stmts.c (revision 256722)
+++ gcc/tree-vect-stmts.c (working copy)
@@ -9426,6 +9426,11 @@ vect_transform_stmt (gimple *stmt, gimpl
gcc_assert (slp_node || !PURE_SLP_STMT (stmt_info));
gimple *old_vec_stmt = STMT_VINFO_VEC_STMT (stmt_info);
+ bool nested_p = (STMT_VINFO_LOOP_VINFO (stmt_info)
+ && nested_in_vect_loop_p
+ (LOOP_VINFO_LOOP (STMT_VINFO_LOOP_VINFO (stmt_info)),
+ stmt));
+
switch (STMT_VINFO_TYPE (stmt_info))
{
case type_demotion_vec_info_type:
@@ -9525,9 +9530,7 @@ vect_transform_stmt (gimple *stmt, gimpl
/* Handle inner-loop stmts whose DEF is used in the loop-nest that
is being vectorized, but outside the immediately enclosing loop. */
if (vec_stmt
- && STMT_VINFO_LOOP_VINFO (stmt_info)
- && nested_in_vect_loop_p (LOOP_VINFO_LOOP (
- STMT_VINFO_LOOP_VINFO (stmt_info)), stmt)
+ && nested_p
&& STMT_VINFO_TYPE (stmt_info) != reduc_vec_info_type
&& (STMT_VINFO_RELEVANT (stmt_info) == vect_used_in_outer
|| STMT_VINFO_RELEVANT (stmt_info) ==
More information about the Gcc-bugs
mailing list