From mboxrd@z Thu Jan 1 00:00:00 1970 From: Boaz Harrosh Subject: Re: [pnfs] [GIT BISECT] first bad commit: 1f36f774 Switch !O_CREAT case to use of do_last() Date: Thu, 25 Mar 2010 15:11:52 +0200 Message-ID: <4BAB6118.4080201@panasas.com> References: <4BAA4CAC.6060104@panasas.com> <4BAA5035.1060906@panasas.com> <4BAA52A8.2080304@panasas.com> <20100324180622.GS30031@ZenIV.linux.org.uk> <4BAA5955.4000001@gmail.com> <20100324185604.GT30031@ZenIV.linux.org.uk> <4BAB2F5A.30409@panasas.com> <20100325101231.GU30031@ZenIV.linux.org.uk> <4BAB3967.6000702@panasas.com> <20100325104924.GV30031@ZenIV.linux.org.uk> <20100325115514.GY30031@ZenIV.linux.org.uk> <4BAB5E69.4090407@panasas.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 7bit Cc: pNFS Mailing List , "J. Bruce Fields" , Trond Myklebust , linux-kernel , Doug Nazar , linux-fsdevel To: Al Viro Return-path: In-Reply-To: <4BAB5E69.4090407@panasas.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: linux-fsdevel.vger.kernel.org On 03/25/2010 03:00 PM, Boaz Harrosh wrote: > On 03/25/2010 01:55 PM, Al Viro wrote: >> On Thu, Mar 25, 2010 at 10:49:24AM +0000, Al Viro wrote: >> >>> Does that happen with nfsv3 or is that v4-only? I'm going to set up v4 >>> server and client and see what happens, but that information could be >>> useful... >> >> *grumble* >> I'd managed to forget that rpc.imapd in its infinite wisdom relies on >> the FPOS I normally disable. Oh, well... rebuilding testbox kernel >> with dnotify enabled ;-/ > > I just wanted to make sure, on top of 2.6.34-rc2 i did: > git revert 781b16775ba0bb55fac0e1757bf0bd87c8879632 > Revert "Fix a dumb typo - use of & instead of &&" > and: > git revert 1f36f774b22a0ceb7dd33eca626746c81a97b6a5 > Revert "Switch !O_CREAT case to use of do_last()" > > The revert is now clean and the problem does not show. > Are you able to reproduce this? Anything I can do to help? > I'm not sure I mentioned this. If you are able to reproduce the error, then after the error in "git status" the tree becomes unusable. Not even from the local_mount. I have not found a method to fix the tree. So what I do is: - Before I start playing I save "cp some_tree/.git/index index.back" - then after the error and tree damage On local mount I: "rm some_tree/.git/index" "cp index.back cp some_tree/.git/index" "cd some_tree; git status" this will fix it so I can try a new iteration > Thanks > Boaz