[patch v3] malloc: avoid need for tcache == NULL checks

[DJ Delorie]

Wilco Dijkstra <Wilco.Dijkstra@arm.com> writes:
> Previously we agreed not to initialize tcache on the free path, and instead let
> the first malloc initialize tcache.

That breaks the testsuite, unless we also add tcache_init to all the
memalign calls too.

> If we want to add initialization, it has to be a no-inline tailcall
> and should be placed after the tcache accesses as Florian points out.

Done.

> All this complexity is also why we should think of a way to completely get rid
> of this initialization. It continues to create performance and correctness issues
> only because we want to save ~512 bytes per thread...

760, but who's counting?

And that's per thread, not per thread that does allocation.  That could
add up to a lot of cost to initialize a lot of unused data for each
short-lived computation-only thread, depending on the application.