Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-5.2 required=3.0 tests=BAYES_00, HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, USER_AGENT_SANE_1 autolearn=no autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id E67C6C2D0E4 for ; Fri, 27 Nov 2020 09:48:46 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 927CE221F7 for ; Fri, 27 Nov 2020 09:48:46 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728741AbgK0Jsq (ORCPT ); Fri, 27 Nov 2020 04:48:46 -0500 Received: from verein.lst.de ([213.95.11.211]:37093 "EHLO verein.lst.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1727303AbgK0Jsq (ORCPT ); Fri, 27 Nov 2020 04:48:46 -0500 Received: by verein.lst.de (Postfix, from userid 2407) id 9707D68B05; Fri, 27 Nov 2020 10:48:42 +0100 (CET) Date: Fri, 27 Nov 2020 10:48:42 +0100 From: Christoph Hellwig To: Jan Kara Cc: Christoph Hellwig , Jens Axboe , Tejun Heo , Josef Bacik , Coly Li , Mike Snitzer , Greg Kroah-Hartman , Johannes Thumshirn , dm-devel@redhat.com, Jan Kara , linux-block@vger.kernel.org, linux-bcache@vger.kernel.org, linux-mtd@lists.infradead.org, linux-fsdevel@vger.kernel.org, linux-mm@kvack.org, Chao Yu Subject: Re: [PATCH 37/44] block: switch partition lookup to use struct block_device Message-ID: <20201127094842.GA15984@lst.de> References: <20201126130422.92945-1-hch@lst.de> <20201126130422.92945-38-hch@lst.de> <20201126182219.GC422@quack2.suse.cz> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20201126182219.GC422@quack2.suse.cz> User-Agent: Mutt/1.5.17 (2007-11-01) Precedence: bulk List-ID: X-Mailing-List: linux-bcache@vger.kernel.org On Thu, Nov 26, 2020 at 07:22:19PM +0100, Jan Kara wrote: > On Thu 26-11-20 14:04:15, Christoph Hellwig wrote: > > struct hd_struct *disk_get_part(struct gendisk *disk, int partno) > > { > > - struct hd_struct *part; > > + struct block_device *part; > > > > rcu_read_lock(); > > part = __disk_get_part(disk, partno); > > - if (part) > > - get_device(part_to_dev(part)); > > - rcu_read_unlock(); > > + if (!part) { > > + rcu_read_unlock(); > > + return NULL; > > + } > > > > - return part; > > + get_device(part_to_dev(part->bd_part)); > > + rcu_read_unlock(); > > + return part->bd_part; > > } > > This is not directly related to this particular patch but I'm wondering: > What prevents say del_gendisk() from racing with disk_get_part(), so that > delete_partition() is called just after we fetched 'part' pointer and the > last 'part' kobject ref is dropped before disk_get_part() calls > get_device()? I don't see anything preventing that and so we'd hand out > 'part' that is soon to be freed (after RCU grace period expires). At this point the hd_struct is already allocated together with the block_device, and thus only freed after the last block_device reference goes away plus the inode freeing RCU grace period. So the device model ref to part is indeed gone, but that simply does not matter any more.