[libc-commits] [libc] [libc][aarch64] Use inline_memcpy_aligned_access_64bit under -mstrict-align (PR #227428)

[via libc-commits]

PiJoules wrote:

> Also take a look at all the other `aarch64/inline_*.h` implementations and make sure they are good for the strict-align case. It could be a separate change for each function, though I won't object if multiple are folded together.
> 
> As well as examining the source, it would be wise to look at the disassembly of how these come out in our actual optimized strict-align build (Fuchsia phys, where boot shims, physload, and physboot should all wind up with their own copies of the same compilation of the libc string functions). Make sure they don't by eyeball look obviously worse than what we had before.

Noted. I shall file bugs separately.

https://github.com/llvm/llvm-project/pull/227428