Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Wed, 30 Jun 2021 11:59:23 +0200
From:   David Sterba <dsterba@suse.cz>
To:     Johannes Thumshirn <Johannes.Thumshirn@wdc.com>
Cc:     "lijian_8010a29@163.com" <lijian_8010a29@163.com>,
        "clm@fb.com" <clm@fb.com>,
        "josef@toxicpanda.com" <josef@toxicpanda.com>,
        "dsterba@suse.com" <dsterba@suse.com>,
        "linux-btrfs@vger.kernel.org" <linux-btrfs@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        lijian <lijian@yulong.com>
Subject: Re: [PATCH] fs: btrfs: extent_map: removed unneeded variable
Message-ID: <20210630095923.GH2610@twin.jikos.cz>
Reply-To: dsterba@suse.cz
Mail-Followup-To: dsterba@suse.cz,
        Johannes Thumshirn <Johannes.Thumshirn@wdc.com>,
        "lijian_8010a29@163.com" <lijian_8010a29@163.com>,
        "clm@fb.com" <clm@fb.com>,
        "josef@toxicpanda.com" <josef@toxicpanda.com>,
        "dsterba@suse.com" <dsterba@suse.com>,
        "linux-btrfs@vger.kernel.org" <linux-btrfs@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        lijian <lijian@yulong.com>
References: <20210629085025.98437-1-lijian_8010a29@163.com>
 <PH0PR04MB741666CF29DB96CF0DB0C26B9B029@PH0PR04MB7416.namprd04.prod.outlook.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <PH0PR04MB741666CF29DB96CF0DB0C26B9B029@PH0PR04MB7416.namprd04.prod.outlook.com>
User-Agent: Mutt/1.5.23.1-rc1 (2014-03-12)
Precedence: bulk

On Tue, Jun 29, 2021 at 09:04:40AM +0000, Johannes Thumshirn wrote:
> On 29/06/2021 10:51, lijian_8010a29@163.com wrote:
> > From: lijian <lijian@yulong.com>
> > 
> > removed unneeded variable 'ret'.
> 
> Wouldn't it make more sense to return an error (-ENOENT??) in case
> the em lookup fails and handle the error in btrfs_finish_ordered_io()
> as this is the only caller of unpin_extent_cache()?
> 
> I've actually tripped over this a couple of times already when 
> investigating extent map and ordered extent splitting problems
> on zoned filesystems:
> 
> 	em = lookup_extent_mapping(tree, start, len);
> 	WARN_ON(!em || em->start != start);
> 
> Maybe even turn this WARN_ON() into an ASSERT() when introducing proper
> error handling, as we shouldn't really get there unless we have a logical
> error.

If you have real workloads hitting the warning then it really should be
proper error handling, not even an assert.