In message <3DB5A2E6.6000305@redhat.com>, > : Ulrich Drepper writes:
> -----BEGIN PGP SIGNED MESSAGE-----
> Hash: SHA1
>
> Dave McCracken wrote:
>
> > 3) The current large page implementation is only for applications
> > that want anonymous *non-pageable* shared memory. Shared page
> > tables reduce resource usage for any shared area that's mapped
> > at a common address and is large enough to span entire pte pages.
>
>
> Does this happen automatically (i.e., without modifying th emmap call)?
>
> In any case, a system using prelinking will likely have all users of a
> DSO mapping the DSO at the same address. Will a system benefit in this
> case? If not directly, perhaps with some help from ld.so since we do
> know when we expect the same is used everywhere.
One important thing to watch out for is to make sure that the PLT
and GOT fixups (where typically pages are mprotected to RW, modified,
then back to RO) are not in the range of pages that are shared.
And, it helps if everything that is shared read-only is 4 MB aligned.
If ld.so could do that under linux, we'd have the biggest win.
gerrit
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Wed Oct 23 2002 - 22:01:00 EST