From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1753751AbeENNeI (ORCPT ); Mon, 14 May 2018 09:34:08 -0400 Received: from bombadil.infradead.org ([198.137.202.133]:57128 "EHLO bombadil.infradead.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753369AbeENNeH (ORCPT ); Mon, 14 May 2018 09:34:07 -0400 Date: Mon, 14 May 2018 06:34:03 -0700 From: Christoph Hellwig To: "Paul E. McKenney" Cc: Christoph Hellwig , Johannes Thumshirn , Keith Busch , Sagi Grimberg , Linux Kernel Mailinglist , Linux NVMe Mailinglist , Hannes Reinecke , Christoph Hellwig Subject: Re: [PATCH] nvme: fix lockdep warning in nvme_mpath_clear_current_path Message-ID: <20180514133403.GA14243@infradead.org> References: <20180514121312.13624-1-jthumshirn@suse.de> <20180514124230.GA654@infradead.org> <20180514133105.GU26088@linux.vnet.ibm.com> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20180514133105.GU26088@linux.vnet.ibm.com> User-Agent: Mutt/1.9.2 (2017-12-15) X-SRS-Rewrite: SMTP reverse-path rewritten from by bombadil.infradead.org. See http://www.infradead.org/rpr.html Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org On Mon, May 14, 2018 at 06:31:05AM -0700, Paul E. McKenney wrote: > > > + if (head && > > > + ns == rcu_dereference_protected(head->current_path, > > > + lockdep_is_held(&ns->ctrl->subsys->lock))) > > > rcu_assign_pointer(head->current_path, NULL); > > > } > > > struct nvme_ns *nvme_find_path(struct nvme_ns_head *head); > > > > We don't really dereference it at all in fact, but just check the > > pointers for equality. I wonder if there is a better way to do this, > > as my ANA patches add a caller without the lock (and withou SRU > > protection either now that I think of it) - for a pure pointer compare > > we really should not need any sort of protection. > > If you are just looking at the value of an RCU-protected pointer, then > using rcu_access_pointer() will cause RCU to just read out the value > and otherwise keeps its mouth shut. That is exactly the function I was looking for. And given that srcu and rcu use the same annotations I should have through of being able to use it of course. As you see above we only use the return value to do a comparison, so we should be perfectly fine. Johannes, can you respin the patch to use rcu_access_pointer?