[AArch64] Suboptimal expansion of `__builtin_bswap16`

https://godbolt.org/z/jcxb4n97o

```c
uint16_t rev16(uint16_t x) { return __builtin_bswap16(x); }
```

expands to
```asm
rev16:
        rev     w8, w0
        lsr     w0, w8, #16
        ret
```

instead of
```asm
rev16:
        rev16   w0, w0
        ret
```

`__builtin_bswap32` and `__builtin_bswap64` don't have the same issue