Received-SPF: pass (google.com: domain of linux-kernel+bounces-3936-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Date: Mon, 18 Dec 2023 16:32:27 +0100
From: Jan Kara <jack@suse.cz>
To: Baokun Li <libaokun1@huawei.com>
Cc: linux-ext4@vger.kernel.org, tytso@mit.edu, adilger.kernel@dilger.ca,
	jack@suse.cz, ritesh.list@gmail.com, linux-kernel@vger.kernel.org,
	yi.zhang@huawei.com, yangerkun@huawei.com, yukuai3@huawei.com,
	Wei Chen <harperchen1110@gmail.com>,
	xingwei lee <xrivendell7@gmail.com>, stable@vger.kernel.org
Subject: Re: [PATCH 1/4] ext4: fix double-free of blocks due to wrong extents
 moved_len
Message-ID: <20231218153227.x273h3c5t2dqgh76@quack3>
References: <20231218141814.1477338-1-libaokun1@huawei.com>
 <20231218141814.1477338-2-libaokun1@huawei.com>
Precedence: bulk
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20231218141814.1477338-2-libaokun1@huawei.com>

On Mon 18-12-23 22:18:11, Baokun Li wrote:
> In ext4_move_extents(), moved_len is only updated when all moves are
> successfully executed, and only discards orig_inode and donor_inode
> preallocations when moved_len is not zero. When the loop fails to exit
> after successfully moving some extents, moved_len is not updated and
> remains at 0, so it does not discard the preallocations.
> 
> If the moved extents overlap with the preallocated extents, the
> overlapped extents are freed twice in ext4_mb_release_inode_pa() and
> ext4_process_freed_data() (as described in commit 94d7c16cbbbd), and
> bb_free is incremented twice. Hence when trim is executed, a zero-division
> bug is triggered in mb_update_avg_fragment_size() because bb_free is not
> zero and bb_fragments is zero.
> 
> Therefore, update move_len after each extent move to avoid the issue.
> 
> Reported-by: Wei Chen <harperchen1110@gmail.com>
> Reported-by: xingwei lee <xrivendell7@gmail.com>
> Closes: https://lore.kernel.org/r/CAO4mrferzqBUnCag8R3m2zf897ts9UEuhjFQGPtODT92rYyR2Q@mail.gmail.com
> Fixes: fcf6b1b729bc ("ext4: refactor ext4_move_extents code base")
> CC: stable@vger.kernel.org # 3.18
> Signed-off-by: Baokun Li <libaokun1@huawei.com>
> ---
>  fs/ext4/move_extent.c | 3 +--
>  1 file changed, 1 insertion(+), 2 deletions(-)
> 
> diff --git a/fs/ext4/move_extent.c b/fs/ext4/move_extent.c
> index 3aa57376d9c2..4b9b503c6346 100644
> --- a/fs/ext4/move_extent.c
> +++ b/fs/ext4/move_extent.c
> @@ -672,7 +672,7 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp, __u64 orig_blk,
>  		 */
>  		ext4_double_up_write_data_sem(orig_inode, donor_inode);
>  		/* Swap original branches with new branches */
> -		move_extent_per_page(o_filp, donor_inode,
> +		*moved_len += move_extent_per_page(o_filp, donor_inode,
>  				     orig_page_index, donor_page_index,
>  				     offset_in_page, cur_len,
>  				     unwritten, &ret);

Although this is currently fine, I think ext4_move_extents() should be
careful and zero out *moved_len before it starts using it.

> @@ -682,7 +682,6 @@ ext4_move_extents(struct file *o_filp, struct file *d_filp, __u64 orig_blk,
>  		o_start += cur_len;
>  		d_start += cur_len;
>  	}
> -	*moved_len = o_start - orig_blk;
>  	if (*moved_len > len)
>  		*moved_len = len;

So I'm not sure the *moved_len > len condition can ever trigger but if it
does, we'd need to check it whenever we are returning moved_len. So either
I'd delete the condition or move it to the out label.

								Honza
-- 
Jan Kara <jack@suse.com>
SUSE Labs, CR