https://gcc.gnu.org/bugzilla/show_bug.cgi?id=68786
Richard Biener <rguenth at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Keywords| |wrong-code
Status|UNCONFIRMED |NEW
Last reconfirmed| |2015-12-08
CC| |jakub at gcc dot gnu.org,
| |rguenth at gcc dot gnu.org
Ever confirmed|0 |1
--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
The zero comes from bump_vector_ptr I think, a missed optimization. So the
issue is that we don't properly use an unaligned type, likely when we expand:
static void
expand_mask_store_optab_fn (internal_fn, gcall *stmt, convert_optab optab)
{
struct expand_operand ops[3];
tree type, lhs, rhs, maskt;
rtx mem, reg, mask;
maskt = gimple_call_arg (stmt, 2);
rhs = gimple_call_arg (stmt, 3);
type = TREE_TYPE (rhs);
lhs = fold_build2 (MEM_REF, type, gimple_call_arg (stmt, 0),
gimple_call_arg (stmt, 1));
it just uses 'type' from rhs but that is just the value being stored. type
needs to be adjusted with build_aligned_type () for the alignment of the
access.
