[builtins] Align addresses to cache lines in __clear_cache for aarch64

This makes sure that the last cache line gets invalidated properly.

This matches the example code at
http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc.den0024a/BABJDBHI.html,
and also matches what libgcc does.

Differential Revision: https://reviews.llvm.org/D42196

llvm-svn: 323315