From mboxrd@z Thu Jan 1 00:00:00 1970 From: Daniel Phillips Date: Wed, 14 Jun 2006 12:08:06 -0700 Subject: [Ocfs2-devel] ext2 attributes for OCFS2 In-Reply-To: <20060614185431.GE3082@ca-server1.us.oracle.com> References: <20060610123253.GA23924@MAIL.13thfloor.at> <20060611174510.GW3082@ca-server1.us.oracle.com> <20060612214116.GA13582@MAIL.13thfloor.at> <20060613235625.GD3082@ca-server1.us.oracle.com> <20060614104400.GA19984@MAIL.13thfloor.at> <20060614185431.GE3082@ca-server1.us.oracle.com> Message-ID: <44905E96.9040002@google.com> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ocfs2-devel@oss.oracle.com Mark Fasheh wrote: > On Wed, Jun 14, 2006 at 12:44:00PM +0200, Herbert Poetzl wrote: >>>Yeah, it should mostly work just fine, but I'm betting there are some >>>paths which will need a manual re-check as things may have changed out >>>from under the VFS while we weren't holding a cluster lock. It should >>>be easy enough to check things like the immutable flag in those places >>>though. >> >>might be, I already wondered why things like >>mark_inode_dirty() do not work with ocfs2, >>IMHO the vfs/ocfs2 synchronization is broken >>(to some degree) but maybe that is intentional > > Yeah, it's difficult to get the VFS to "cluster lock" without affecting the > runtime of other file systems, so we have to work within that constraint :/ Please don't be shy about asking. Can you describe the mark_inode_dirty issue? Regards, Daniel