--C1iGAkRnbeBonpVg
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Jul 05, 2007 at 03:14:02PM +0200, Joerg Sonnenberger wrote:
> On Thu, Jul 05, 2007 at 10:06:07PM +0900, YAMAMOTO Takashi wrote:
> > > On Thu, Jul 05, 2007 at 09:46:56PM +0900, YAMAMOTO Takashi wrote:
> > > > before it goes in the tree, i'd like to see an actual user of the s=
yscall.
> > >=20
> > > I have an alternative malloc implementation using it.
> > >=20
> > > Joerg
> >=20
> > do you have any numbers with and without mremap?
>=20
> Something like
> 	ptr =3D malloc(1 << 29);
> 	realloc(ptr, 1 << 30);
> finishes in an instant with mremap and forces paging on my machine
> otherwise. That alone should be an pretty obvious difference.

True, but what app actually does that?

What apps would make use of this, and what typical realloc sizes do they=20
play with?

Take care,

Bill

--C1iGAkRnbeBonpVg
Content-Type: application/pgp-signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.3 (NetBSD)

iD8DBQFGjVbzWz+3JHUci9cRAlgAAJ9I48WLG4pW2/gRgPJnOC8yUCwcLwCfZnav
vNFEONo8Rq4CrCDDdv1FZcI=
=+TCr
-----END PGP SIGNATURE-----

--C1iGAkRnbeBonpVg--