Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
Date:   Fri, 25 Feb 2022 21:23:37 -0500
From:   "Theodore Ts'o" <tytso@mit.edu>
To:     Ye Bin <yebin10@huawei.com>
Cc:     adilger.kernel@dilger.ca, linux-ext4@vger.kernel.org,
        jaegeuk@kernel.org, chao@kernel.org,
        linux-f2fs-devel@lists.sourceforge.net,
        linux-kernel@vger.kernel.org, jack@suse.cz, lczerner@redhat.com
Subject: Re: [PATCH -next v2] ext4:fix file system corrupted when rmdir non
 empty directory with IO error
Message-ID: <YhmPKVrVHhTeKOzl@mit.edu>
References: <20220211093527.3335518-1-yebin10@huawei.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20220211093527.3335518-1-yebin10@huawei.com>
Precedence: bulk

On Fri, Feb 11, 2022 at 05:35:27PM +0800, Ye Bin wrote:
> Now if read directory block failed, 'ext4_empty_dir' will return true, assume
> directory is empty. Obviously, it will lead to above issue.
> To solve this issue, if read directory block failed 'ext4_empty_dir' just assume
> directory isn't empty. To avoid making things worse when file system is already
> corrupted, 'ext4_empty_dir' also assume directory isn't empty.
> To distinguish the error type, return the exact error code to the caller.
>

Does the same issue exist for f2fs and ubifs?  We could solve the
specific bug much more simply by having ext4_empty_dir() return FALSE
if we aren't able to read the directory block.  Yes, it means that we
don't return as specific an error code in the case of an I/O error ---
although I believe we do syslog a warning --- but it makes for a much
simpler patch that doesn't requiring getting acked-by's from the
fscrypt, f2fs and ubifs folks.

							- Ted