> It would need strict_low_part unless we're dealing with subwords. I > think the patch should maybe check for that in the !W_R_O case.
The code already does the check, it simply won't mess with strict_low_part. > Does WORD_REGISTER_OPERATIONS really buy much on targets that use it? Yes, it makes it possible for the combiner to eliminate 2 redundant extensions on Alpha for the reduced testcase provided by Uros (that's why there is the mode mismatch). And it's even more helpful if the loads are not explicitly extended, like on SPARC (in combination with LOAD_EXTEND_OP of course). -- Eric Botcazou
