On Tue, 2006-10-24 at 16:50 -0400, Chris Mason wrote:
quoted text > On Tue, Oct 24, 2006 at 01:28:31PM -0700, Badari Pulavarty wrote:
> > On Tue, 2006-10-24 at 15:34 -0400, Chris Mason wrote:
> > > Hello,
> > >=20
> > > Here is a new cut of my O_DIRECT locking rework. It has a much lower
> > > cpu cost than the last set, and simple benchmarks no longer show a
> > > regression here in system time. But, the complexity for inserting
> > > placeholder pages has gone up.
> > >=20
> > > I've also changed the way I test for a place holder page (from
> > > mm/filemap.c):
> > >=20
> > > static struct address_space placeholder_address_space;
> > > #define PagePlaceHolder(page) ((page)->mapping =3D=3D &placeholder_ad=
dress_space)
quoted text > > >=20
> > > This is more stable than the last one but I'm just starting to run ra=
ce
quoted text > > > and load testing on it.
> > >=20
> > > -chris
> >=20
> > Gave it a spin with simple fsx tests and ran into ..
>=20
> Thanks, I missed an i_mutex change. Let me audit mutexes coming in and
> out again and resend.
>=20
> I think the new rules should be:
>=20
> reads: don't need i_mutex at all
> writes: don't need i_mutex for io inside i_size.
>=20
> Oh, and don't mess with i_mutex under DIO_OWN_LOCKING.
Can we get rid of DIO_OWN_LOCKING altogether.
It's horribly confusing and overload with regards to the
create flag in get_blocks.
quoted text > -chris
> -
> To unsubscribe from this list: send the line "unsubscribe linux-fsdevel" =
in
quoted text --=20
Russell Cattelan <cattelan@thebarn.com>
