From: dumas@centre-cired.fr (Patrice DUMAS - DOCT) Subject: Re: Locking patches (generic & nfs) Date: Tue, 23 Jul 2002 18:32:50 +0200 Sender: nfs-admin@lists.sourceforge.net Message-ID: <20020723183250.A30579@hermes.centre-cired.fr> References: <20020719101950.A15819@suse.de> <20020723170615.D1869@suse.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Return-path: Received: from boukha.centre-cired.fr ([193.51.120.234]) by usw-sf-list1.sourceforge.net with esmtp (Exim 3.31-VA-mm2 #1 (Debian)) id 17X2of-000567-00 for ; Tue, 23 Jul 2002 09:47:22 -0700 Received: from hermes.centre-cired.fr ([193.51.120.92]) by boukha.centre-cired.fr (8.9.3+Sun/jtpda-5.3.3) with ESMTP id SAA14346 for ; Tue, 23 Jul 2002 18:44:50 +0100 (WEST) Received: (from dumas@localhost) by hermes.centre-cired.fr (8.11.6/8.11.6) id g6NGWq816979 for nfs@lists.sourceforge.net; Tue, 23 Jul 2002 18:32:52 +0200 To: nfs@lists.sourceforge.net In-Reply-To: <20020723170615.D1869@suse.de>; from okir@suse.de on Tue, Jul 23, 2002 at 05:06:15PM +0200 Errors-To: nfs-admin@lists.sourceforge.net List-Help: List-Post: List-Subscribe: , List-Id: Discussion of NFS under Linux development, interoperability, and testing. List-Unsubscribe: , List-Archive: Hi, > > Patrice Dumas recently did some work on implementing this both for > > NLMv1,2,3 and NLM4, so I was planning on integrating his changes into > > 2.4.20. > > As you can see from the patch, it's not really much you need to add. > The functionality is all there, one only needs to decode the GRANTED_RES > call rather than dropping it. This is basically all what is needed, however there are some more issues raised when the server decode the GRANTED_RES calls I tried to answer in other patches: - if 2 clients use the same cookie, the blocks will be mixed up - if the linux client GRANTED_RES is lost, it will get a GRANTED_MSG twice, the second GRANTED_RES will have a status set to nlm_lck_denied because in nlmclnt_grant the lock request won't be found. This in turn will cause the server to unlock the lock, while the client thinks it is granted and potential file corruption. I made a fix on the client side. It doesn't solve the problem on the server side, but it seems to me that the server side behaviour is the right behaviour (except in very rare cases where there is a race condition I will try to solve in a next patch but it is very rare and not urgent in my opinion). Pat ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ NFS maillist - NFS@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/nfs