https://gcc.gnu.org/bugzilla/show_bug.cgi?id=118150

--- Comment #3 from Spencer Abson <sabson at gcc dot gnu.org> ---
>It looks like we also miss AND+PTEST because the existing combine matches the 
>>(and (<logical> ((a) (b)) (c)) canonicalisation of predicated ior and xor, 
>>where c is all-true.

Sorry - I mean that c is an all-true here if expanded via the generic optab
(without it's own predicate operand).

Reply via email to