https://gcc.gnu.org/bugzilla/show_bug.cgi?id=87802
--- Comment #8 from Richard Biener <rguenth at gcc dot gnu.org> ---
Ah, ok.
/home/seurer/gcc/gcc-test/gcc/testsuite/g++.dg/vect/slp-pr87105.cc:98:1:
missed: not vectorized: bad data alignment in basic block.
looks like with the flags we do not support unaligned accesses. Removing
the -mno-allow-movmisalign flag fixes this.
Does the following work? If so please go ahead and install it.
Index: gcc/testsuite/g++.dg/vect/slp-pr87105.cc
===================================================================
--- gcc/testsuite/g++.dg/vect/slp-pr87105.cc (revision 265619)
+++ gcc/testsuite/g++.dg/vect/slp-pr87105.cc (working copy)
@@ -99,7 +99,7 @@ void quadBoundingBoxA(const Point bez[3]
// We should have if-converted everything down to straight-line code
// { dg-final { scan-tree-dump-times "<bb \[0-9\]+>" 1 "slp2" } }
-// { dg-final { scan-tree-dump-times "basic block part vectorized" 1 "slp2" }
}
+// { dg-final { scan-tree-dump-times "basic block part vectorized" 1 "slp2" {
xfail { { ! vect_element_align } && { ! vect_hw_misalign } } } } }
// It's a bit awkward to detect that all stores were vectorized but the
// following more or less does the trick
-// { dg-final { scan-tree-dump "vect_iftmp\[^\r\m\]* = MIN" "slp2" } }
+// { dg-final { scan-tree-dump "vect_iftmp\[^\r\m\]* = MIN" "slp2" { xfail { {
! vect_element_align } && { ! vect_hw_misalign } } } } }
