From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1765210Ab3DDXTG (ORCPT ); Thu, 4 Apr 2013 19:19:06 -0400 Received: from haggis.pcug.org.au ([203.10.76.10]:55435 "EHLO members.tip.net.au" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1765142Ab3DDXTE (ORCPT ); Thu, 4 Apr 2013 19:19:04 -0400 Date: Fri, 5 Apr 2013 10:18:53 +1100 From: Stephen Rothwell To: Al Viro Cc: Andrew Morton , linux-next@vger.kernel.org, linux-kernel@vger.kernel.org, Nathan Zimmer Subject: Re: linux-next: manual merge of the akpm tree with the vfs tree Message-Id: <20130405101853.769d3b915cacd8a4863b7c19@canb.auug.org.au> In-Reply-To: <20130404081011.GO21522@ZenIV.linux.org.uk> References: <20130404172648.27834d12cbb68338f4b89482@canb.auug.org.au> <20130403235634.6bc72c39.akpm@linux-foundation.org> <20130404000253.2fc9c75d.akpm@linux-foundation.org> <20130404081011.GO21522@ZenIV.linux.org.uk> X-Mailer: Sylpheed 3.3.0 (GTK+ 2.24.10; i486-pc-linux-gnu) Mime-Version: 1.0 Content-Type: multipart/signed; protocol="application/pgp-signature"; micalg="PGP-SHA256"; boundary="Signature=_Fri__5_Apr_2013_10_18_53_+1100_q0CGuO8FIiErykVS" Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org --Signature=_Fri__5_Apr_2013_10_18_53_+1100_q0CGuO8FIiErykVS Content-Type: text/plain; charset=US-ASCII Content-Disposition: inline Content-Transfer-Encoding: quoted-printable Hi Al, On Thu, 4 Apr 2013 09:10:11 +0100 Al Viro wrote: > > On Thu, Apr 04, 2013 at 12:02:53AM -0700, Andrew Morton wrote: >=20 > > > Well perhaps the vfs tree should start paying some attention to the > > > rest of the world, particularly after -rc5. > >=20 > > I can't even find this "lift sb_start_write() out of ->write()". Not o= n fsdevel, > > not on lkml. What the heck is it and why was it so important? >=20 > Deadlocks around splice; see the threads re overlayfs/unionmount/aufs and > deadlocks in their copyup implementations. See also XFS freeze-related > deadlocks, etc. >=20 > The thing is, sb_start_write() is pretty high in locking hierarchy (outsi= de > ->i_mutex, etc.), but ->splice_write() and friends had it buried pretty > deep. With distinctly unpleasant results, including ->..._write() instan= ces > using generic ones (which took the lock) *and* doing some IO outside of t= hose > (ext4, for example; ocfs2 also looked fishy in that respect, IIRC). >=20 > The obvious solution is to lift taking that lock out of the methods, which > had been done. It had been discussed on fsdevel and sat in #experimental= for > several weeks; time for it to go into #for-next. It would have been useful to put something like that in the commit message = ... --=20 Cheers, Stephen Rothwell sfr@canb.auug.org.au --Signature=_Fri__5_Apr_2013_10_18_53_+1100_q0CGuO8FIiErykVS Content-Type: application/pgp-signature -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.12 (GNU/Linux) iQIcBAEBCAAGBQJRXgpdAAoJEECxmPOUX5FEfVkP/RUeL6JuiTZSoUXc02z0PcdN hS26kfzYp6Ziz0MUHx/rauzVqNm6+swCQqnwc8v4smzc9acBIMhIvpSl3UuAw4HO T673B7MYqLyJ0WwsX21ChZ9wETGY2BIXCE+mzVtAv+7LPRVX0ZRFBBgyCnsBgn6n v7oIIBH76eM2ZDY4f+aUNaxlVVdLlfLy3liG+vpal3pRa5WLfXxvyL3Ax0iYGRJk V3Td7/j175gWyOs93srej7ryNYVpk1e0BRB7wnXHEZu2HEAmW8IsDrUZYVJEQR6D 71p06t+Vl7KzG8+R6DjKMnRIOT/iLjv5OkrKkYopLoX0wzmRwcj5vwvNAcdOJAmc DdVcgsqmKIKIfV4gUhImBLhF/S+NnQv3m3DegRC1j4MpeKczZIRaeuhT73H1RjzA TcevApShCmp3YHEHSvjilBbPBPwjEnu4I2V4M5TzX/EmtTnJtEc5Unu0lJ5eu/+N 5LXUxWoU6EUg0+kCrfCNq3cFXLGy7jAXaoiB2z4zVA7LswvHTbqF5Vm8DNjJwuQ+ qF0dhdsORObCrFL7SBBxqRFp/CM0UhPv7I7rmcqReHxbFwSbuZ5fDZKWzvbzxRKM WKoXNxmnGh8LejrPrdExnm77dS61W9uWbv8Q9XZ9x8Z8PQoch5QihZBB1Dzva6nA YLWeGP/utc+7lrv+dZXV =Yn9n -----END PGP SIGNATURE----- --Signature=_Fri__5_Apr_2013_10_18_53_+1100_q0CGuO8FIiErykVS--