Date: Fri, 13 Mar 2009 23:22:29 +0200 From: Kostik Belousov <kostikbel@gmail.com> To: John Baldwin <jhb@freebsd.org> Cc: svn-src-head@freebsd.org, svn-src-all@freebsd.org, src-committers@freebsd.org Subject: Re: svn commit: r186276 - head/sys/kern Message-ID: <20090313212229.GW41617@deviant.kiev.zoral.com.ua> In-Reply-To: <49BAA2C6.2000807@FreeBSD.org> References: <200812181158.mBIBwC50039690@svn.freebsd.org> <49BAA2C6.2000807@FreeBSD.org>
next in thread | previous in thread | raw e-mail | index | archive | help
--2UfnThGAjlIhdWh/ Content-Type: text/plain; charset=us-ascii Content-Disposition: inline Content-Transfer-Encoding: quoted-printable On Fri, Mar 13, 2009 at 02:15:34PM -0400, John Baldwin wrote: > Konstantin Belousov wrote: > >Author: kib > >Date: Thu Dec 18 11:58:12 2008 > >New Revision: 186276 > >URL: http://svn.freebsd.org/changeset/base/186276 > > > >Log: > > Do not return success and doomed vnode from lookup. LK_UPGRADE allows > > the vnode to be reclaimed. > > =20 > > Tested by: pho > > MFC after: 1 month >=20 > Would EBADF be more appropriate? That is what other places that check=20 > VI_DOOMED return for this type of check (e.g. in cache_lookup()). I do not think so. When we do namei lookup, there is actually no file descriptor to be invalid. The fact the we lost the race with forced unmount actually means that there is no more node with supplied name. --2UfnThGAjlIhdWh/ Content-Type: application/pgp-signature Content-Disposition: inline -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (FreeBSD) iEYEARECAAYFAkm6zpQACgkQC3+MBN1Mb4hx2ACgj4wTci0VmNugCYOeh1gBMCza M3IAoNDpUMKztX5lFK8cZm2g9Rv8MxnA =qyS1 -----END PGP SIGNATURE----- --2UfnThGAjlIhdWh/--
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20090313212229.GW41617>