[AArch64] Add support for efficient bitcast in vector truncate store.

Following the changes in D145301, we now also support the efficient bitcast
when storing the bool vector. Previously, this was expanded.

Differential Revision: https://reviews.llvm.org/D148316