This MR enforces alignment of TLS slots as described in PE file.
I'm not 100% happy with the deallocation scheme, as it could be slow. If someone has a better idea, be welcome! Didn't open the option to introduce helpers in heap.c (but maybe that's the way to go).
-- v2: ntdll: Enforce the alignment of TLS directory entries. kernel32/tests: Add a test about TLS slot alignment.