From mboxrd@z Thu Jan 1 00:00:00 1970 From: Goldwyn Rodrigues Date: Sun, 03 Nov 2013 21:46:01 -0600 Subject: [Ocfs2-devel] [PATCH 3/6] Shift allocation ocfs2_live_connection to user_connect() In-Reply-To: <20131103231457.GI29346@wotan.suse.de> References: <20131018144534.GA4583@shrek.lan> <20131103231457.GI29346@wotan.suse.de> Message-ID: <52771879.5060808@suse.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: ocfs2-devel@oss.oracle.com On 11/03/2013 05:14 PM, Mark Fasheh wrote: > On Fri, Oct 18, 2013 at 09:45:39AM -0500, Goldwyn Rodrigues wrote: >> We perform this because the DLM recovery callbacks will require >> the ocfs2_live_connection structure to record the node information >> when dlm_new_lockspace() is updated. > > Ok but what I see below is that you took the alloc out of > ocfs2_live_connecion_new() and just do it above that call in > user_cluster_connect(), then call ocfs2_live_connection_new(). > > Aside from that though this doesn't seem to add new functionality, which is > fine if the later patches use what you've done here but I don't where this > change has made any impact on patches 4-6. So my guess is that there's one > of two things going on here: > > 1) This change was just left in by accident in which case you can just > remove this patch from your series. > > 2) I made an error reading the later patches and this _is_ used :) > If that's the case do me a favor and point out what I missed :) When dlm_new_lockspace is called, it calls recover_done() to inform our node number. This is recorded in the live_connection data structure and needs ocfs2_live_connection to be available before the call to dlm_lockspce_new(). I will put this information in the next series. > Also, if this needs to stay I would rename ocfs2_live_connecion_new() since > _new() usually makes me think "allocates" and we're not doing that any more. > Maybe something like ocfs2_live_connection_attach()? Yes, we should change the name of the function. -- Goldwyn