libthr: drop use of try_gc in the end of thr_join

[aokblast]

Summary

thr_try_gc unlock and lock again to make sure the lock order. However,
there is no reason for such roundtrip if the caller holds the correct
oder. As a result, we create a new helper function to add the thread
to gc list without unlock + lock and wire the join_common to this new helper.

[kib]

The change now takes the thread list write lock unconditionally. Before the change, it only happen if the thread is not referenced. Did you checked that this is innocent?

lib/libthr/thread/thr_join.c

96

This helper as is does not make much sense after your change. Perhaps another helper that does everything up to, but not including, the thr_try_gc() call would be useful, but not this single-use function that can be inlined in the single use place.

118

There is no need in tmp, you can assign directly. It is arguably not needed in the join_common_joined().