From mboxrd@z Thu Jan 1 00:00:00 1970 From: Milosz Tanski Subject: Re: FSCache: Cookie create and destroy semantics on file open Date: Wed, 7 Aug 2013 08:08:28 -0400 Message-ID: <4307898982697141463@unknownmsgid> References: <26386.1375855465@warthog.procyon.org.uk> Mime-Version: 1.0 (1.0) Content-Type: text/plain; charset=ISO-8859-1 Cc: "linux-cachefs@redhat.com" , "linux-fsdevel@vger.kernel.org" To: David Howells Return-path: Received: from mail-wi0-f180.google.com ([209.85.212.180]:49130 "EHLO mail-wi0-f180.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S932544Ab3HGMIf (ORCPT ); Wed, 7 Aug 2013 08:08:35 -0400 Received: by mail-wi0-f180.google.com with SMTP id f14so1595365wiw.7 for ; Wed, 07 Aug 2013 05:08:33 -0700 (PDT) In-Reply-To: <26386.1375855465@warthog.procyon.org.uk> Sender: linux-fsdevel-owner@vger.kernel.org List-ID: Can you share those patches? I'd like to know what the approach you've taken since I'm trying to figure out how to solve it for Ceph. - M On Aug 7, 2013, at 2:04 AM, David Howells wrote: > Milosz Tanski wrote: > >> I have a question for you about how added / removing the cookie for a >> NFS inode works. Right the cache entry gets added when the file is >> open. It gets revoked when the file re-opened in write mode. What >> happens in this case. >> >> Open for read (create cookie) -> Open for write (destroy cookie) -> >> Open for read (create cookie again) >> >> In this case, none of the 3 files were closed. > > No. The way NFS is doing this is wrong. I have been given some patches that > should fix this, I think. > > David