Re: [PATCH, libpthread] Correct logic for PTHREAD_KEY_INVALID slots.

[Neal H. Walfield]

I haven't looked at this change in detail, but it sounds sane and if
Samuel says its okay, that's good enough for me.

Neal