Align initial-exec TLS segments to the p_vaddr % align.

This is continuation of D21163/r359634, which handled the alignment
for global mode.

Non-x86 arches are not handled, maintainers are welcomed.

Tested by: emaste
Sponsored by: The FreeBSD Foundation
MFC after: 2 weeks
Differential revision: https://reviews.freebsd.org/D24366