[PATCH v2] Remove upper limit on tunable MALLOC_MMAP_THRESHOLD

[DJ Delorie]

Patrick McGehearty <patrick.mcgehearty@oracle.com> writes:
> DJ, does your final comments mean this patch is
> OK to commit?

Sorry, yes.  Please commit.

> I should say that I intend to also change HEAP_MAX to be
> a tunable to handle the issues we discussed as a separate
> patch. It does not look to be time consuming. I just
> can't predict when I'll have it ready because it's behind
> a couple of other things in my task list and I can't
> predict what new work tasks might cause further delays.

Please pay attention to the performance metrics on that one; you're
replacing a compiled-in constant (with associated masks and shifts,
which may not be obviously dependent) with a runtime computation, and we
do those computations a lot.  Otherwise, good luck ;-)