Date: Tue, 30 Aug 2005 21:16:44 -0400 From: Alexander Kabaev <kabaev@gmail.com> To: Alfred Perlstein <alfred@freebsd.org> Cc: Alexander Kabaev <kan@FreeBSD.org>, cvs-src@FreeBSD.org, src-committers@FreeBSD.org, cvs-all@FreeBSD.org Subject: Re: cvs commit: src/sys/kern vfs_lookup.c Message-ID: <20050830211644.339c86db@kan.dnsalias.net> In-Reply-To: <20050830224704.GI77060@elvis.mu.org> References: <200508301612.j7UGC3CM085774@repoman.freebsd.org> <20050830224704.GI77060@elvis.mu.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--Signature_Tue__30_Aug_2005_21_16_44_-0400_jKUEwb3gwzt1VkZm
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: quoted-printable
On Tue, 30 Aug 2005 15:47:04 -0700
Alfred Perlstein <alfred@freebsd.org> wrote:
> Ok, maybe I'm missing something, but...
>=20
> while (dp->v_type =3D=3D VDIR && (mp =3D dp->v_mountedhere) &&
> (cnp->cn_flags & NOCROSSMOUNT) =3D=3D 0) {
> KASSERT(dp !=3D ndp->ni_dvp, ("XXX"));
> if (vfs_busy(mp, 0, 0, td))
> continue;
> here ->>> vput(dp);
> tvfslocked =3D VFS_LOCK_GIANT(mp);
> VFS_UNLOCK_GIANT(vfslocked);
> vfslocked =3D tvfslocked;
> VOP_UNLOCK(ndp->ni_dvp, 0, td);
> error =3D VFS_ROOT(mp, cnp->cn_lkflags, &tdp, td);
> VOP_LOCK(ndp->ni_dvp, cnp->cn_lkflags | LK_RETRY, td);
> vfs_unbusy(mp, td);
> if (error) {
> dpunlocked =3D 1;
> goto bad2;
> }
> ndp->ni_vp =3D dp =3D tdp;
> }
>=20
> Isn't dp already unlocked here? vput should be unlocking the vnode
> and we have the above KASSERT showing that they should be the same
> vnode...
dp !=3D ndp->ni_dvp at that point. If we have amd serving /home,
ndp->ni_dvp is '/' directory
dp is /home directory =20
Having / locked while waiting for VFS_ROOT on amd filesystem serves no
purpose.
> Also, I think lockparent doesn't need to be respected because
> we're trading parent vnodes. Basically, we're hitting a vnode
> that has another vnode stacked on it (root vnode), and switching to
> it. we're not actually decending.
>=20
> Is this OK?
>=20
> -Alfred
>=20
>=20
>=20
> * Alexander Kabaev <kan@FreeBSD.org> [050830 09:12] wrote:
> > kan 2005-08-30 16:12:03 UTC
> >=20
> > FreeBSD src repository
> >=20
> > Modified files: (Branch: RELENG_6)
> > sys/kern vfs_lookup.c=20
> > Log:
> > MFC r1.81:
> > Do not keep parent directory locked while calling VFS_ROOT to
> > traverse mount points in lookup(). The lock can be dropped safely
> > around VFS_ROOT because LOCKPARENT semantics with child and parent
> > vnodes coming from different FSes does not really have any
> > meaningful use and we do not really care about parent state after
> > we relock it. This prevents easily triggered deadlock on systems
> > using automounter daemon.
> > =20
> > Approved by: re (scottl)
> > =20
> > Revision Changes Path
> > 1.80.2.1 +2 -0 src/sys/kern/vfs_lookup.c
>=20
> --=20
> - Alfred Perlstein
--=20
Alexander Kabaev
--Signature_Tue__30_Aug_2005_21_16_44_-0400_jKUEwb3gwzt1VkZm
Content-Type: application/pgp-signature
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.2 (FreeBSD)
iD8DBQFDFQT/Q6z1jMm+XZYRAvsAAJ9yxJbuZrb5j+rrzrD1L4yhjPp5fQCgySmJ
K7ppNeA7oXphk3xywnxaNDQ=
=GE0Q
-----END PGP SIGNATURE-----
--Signature_Tue__30_Aug_2005_21_16_44_-0400_jKUEwb3gwzt1VkZm--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20050830211644.339c86db>