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;
Message-ID: <5ab59623-4317-5aff-5173-d7285b5b224c@squashfs.org.uk>
Date:   Fri, 31 Mar 2023 02:10:15 +0100
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:102.0) Gecko/20100101
 Thunderbird/102.9.1
Subject: Re: [PATCH 1/2] block: Rework bio_for_each_segment_all()
From:   Phillip Lougher <phillip@squashfs.org.uk>
To:     Kent Overstreet <kent.overstreet@linux.dev>
Cc:     linux-kernel@vger.kernel.org, linux-block@vger.kernel.org,
        willy@infradead.org, axboe@kernel.dk,
        Ming Lei <ming.lei@redhat.com>
References: <20230327174402.1655365-1-kent.overstreet@linux.dev>
 <20230327174402.1655365-2-kent.overstreet@linux.dev>
 <52a5bd5c-d3a1-71d7-e1e5-7965501818bd@squashfs.org.uk>
 <ZCXNDQ6Eslhj+9g5@moria.home.lan>
 <4753802a-685f-ab56-fed2-22d6eb4cfccd@squashfs.org.uk>
In-Reply-To: <4753802a-685f-ab56-fed2-22d6eb4cfccd@squashfs.org.uk>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 30/03/2023 19:59, Phillip Lougher wrote:
> On 30/03/2023 18:55, Kent Overstreet wrote:
>> On Wed, Mar 29, 2023 at 05:50:27PM +0100, Phillip Lougher wrote:
>>> There is a problem with the above code, on testing I get the following
>>> results:
>>>
>>> Index 78018785329, offset 49, bvec.bv_len 1024: In same bio, metadata 
>>> length
>>> 32780
>>
>> Could you share how you're doing your testing? I set up a basic test
>> (created images using every compression type, tested reading them) and
>> so far I'm not able to reproduce this.
> 
> I use a very large Squashfs file that triggers the edge case.
> 
> That is too big to post, and so I'll see if I can artifically generate
> a small Squashfs filesystem that triggers the edge case.
> 
> Phillip
> 
> 

Hi,

This is a Google drive link to a file that triggers the issue.

https://drive.google.com/file/d/1-3-a1BKq62hZGQ6ynioreMSWFMuCV9B4/view?usp=sharing

To reproduce the issue

% mount -t squashfs <the file> /mnt -o errors=panic

then either one of the following will produce a panic

% ls /mnt
% find /mnt -type f | xargs wc > /dev/null


Phillip