Return-Path: Received: from lo.gmane.org ([80.91.229.12]:39288 "EHLO lo.gmane.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752061Ab1C0AH0 (ORCPT ); Sat, 26 Mar 2011 20:07:26 -0400 Received: from list by lo.gmane.org with local (Exim 4.69) (envelope-from ) id 1Q3dVt-0005AQ-O3 for linux-nfs@vger.kernel.org; Sun, 27 Mar 2011 01:07:25 +0100 Received: from d67-193-214-242.home3.cgocable.net ([67.193.214.242]) by main.gmane.org with esmtp (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 27 Mar 2011 01:07:25 +0100 Received: from brian by d67-193-214-242.home3.cgocable.net with local (Gmexim 0.1 (Debian)) id 1AlnuQ-0007hv-00 for ; Sun, 27 Mar 2011 01:07:25 +0100 To: linux-nfs@vger.kernel.org From: "Brian J. Murrell" Subject: Re: different kernels mean NFS4/GSSAPI works or doesn't Date: Sat, 26 Mar 2011 20:07:12 -0400 Message-ID: <4D8E7FB0.9060109@interlinx.bc.ca> References: <4D89D36D.3090207@interlinx.bc.ca> <4D8A0C9F.8090300@interlinx.bc.ca> <1300900035.11677.12.camel@lade.trondhjem.org> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="------------enig51F3B30CECD4604F801EDDC6" Cc: Kevin Coffman In-Reply-To: <1300900035.11677.12.camel@lade.trondhjem.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: MIME-Version: 1.0 --------------enig51F3B30CECD4604F801EDDC6 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable On 11-03-23 01:07 PM, Trond Myklebust wrote: > Please don't remove people from the Cc list. Fair enough. Personally, I dislike getting two copies, but I can respect others' wishes. > You said this happens only with RPCSEC_GSS? Indeed. Hrm. Well, to be honest my two scenarios are NFS4 with RPCSEC_GSS and NFSv3. I've never really looked into NFS4 without RPCSEC_GSS. Is that even possible? What's the point? I guess there are protocol improvements in NFS4 aside from RPCSEC_GSS? > Have you tried bisecting, to see which changes caused it? I just did, using Ubuntu's "mainline" kernels, which AFAIU are upstream kernel source without any of their special sauce, but built with their =2Econfigs. It seems the problem was introduced in 2.6.32.12. I don't really know which commit between 2.6.32.11 and 2.6.32.12 though. In fact I don't even know where to find these so-called .11 and .12 4th order of releases to see what all went into it. I did find http://www.kernel.org/pub/linux/kernel/v2.6/ChangeLog-2.6.32.11 but a gitweb interface to the changes would be much more useful I think. Although the number of changes that show up there for matches of /nfs/ or /gss/ are quite small (only one for the latter). > You can > probably restrict yourself to bisecting changes to include/linux/sunrpc= , > net/sunrpc and net/sunrpc/auth_gss. If I could find a gitweb tree to grovel through. :-) b. --------------enig51F3B30CECD4604F801EDDC6 Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.10 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/ iEYEARECAAYFAk2Of7AACgkQl3EQlGLyuXA6KACgndbSOAncrfUCQI2BEWlgkZlc 0IoAoMpQWD3/fGsY8Ql1aN1a7DSnFCXK =juBK -----END PGP SIGNATURE----- --------------enig51F3B30CECD4604F801EDDC6--