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;
Subject: Re: [PATCH -next] block: Fix the partition start may overflow in
 add_partition()
To:     Yu Kuai <yukuai1@huaweicloud.com>,
        Zhong Jinghua <zhongjinghua@huaweicloud.com>, axboe@kernel.dk
Cc:     linux-block@vger.kernel.org, linux-kernel@vger.kernel.org,
        zhongjinghua@huawei.com, yi.zhang@huawei.com,
        chengzhihao1@huawei.com, yangerkun@huawei.com,
        "yukuai (C)" <yukuai3@huawei.com>
References: <20230522070615.1485014-1-zhongjinghua@huaweicloud.com>
 <655707b8-f741-3ed0-b1ea-974b3ea496fb@huaweicloud.com>
From:   Yu Kuai <yukuai1@huaweicloud.com>
Message-ID: <03df2341-5730-fb78-0d11-9c1e102f8b63@huaweicloud.com>
Date:   Thu, 25 May 2023 10:27:26 +0800
User-Agent: Mozilla/5.0 (Windows NT 10.0; WOW64; rv:60.0) Gecko/20100101
 Thunderbird/60.8.0
MIME-Version: 1.0
In-Reply-To: <655707b8-f741-3ed0-b1ea-974b3ea496fb@huaweicloud.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 8bit
Precedence: bulk

Hi,

在 2023/05/25 10:14, Yu Kuai 写道:
> Hi,
> 
> 在 2023/05/22 15:06, Zhong Jinghua 写道:
>> From: Zhong Jinghua <zhongjinghua@huawei.com>
>>
>> In the block_ioctl, we can pass in the unsigned number 0x8000000000000000
>> as an input parameter, like below:
>>
>> block_ioctl
>>    blkdev_ioctl
>>      blkpg_ioctl
>>        blkpg_do_ioctl
>>          copy_from_user
>>          bdev_add_partition
>>            add_partition
>>              p->start_sect = start; // start = 0x8000000000000000
> 
> start_sect is 0x8..... >> SECTOR_SHIFT.
> 
>>
>> Then, there was an warning when submit bio:
>>
>> WARNING: CPU: 0 PID: 382 at fs/iomap/apply.c:54
>> Call trace:
>>   iomap_apply+0x644/0x6e0
>>   __iomap_dio_rw+0x5cc/0xa24
>>   iomap_dio_rw+0x4c/0xcc
>>   ext4_dio_read_iter
>>   ext4_file_read_iter
>>   ext4_file_read_iter+0x318/0x39c
>>   call_read_iter
>>   lo_rw_aio.isra.0+0x748/0x75c
>>   do_req_filebacked+0x2d4/0x370
>>   loop_handle_cmd
>>   loop_queue_work+0x94/0x23c
>>   kthread_worker_fn+0x160/0x6bc
>>   loop_kthread_worker_fn+0x3c/0x50
>>   kthread+0x20c/0x25c
>>   ret_from_fork+0x10/0x18
>>
>> Stack:
>>
>> submit_bio_noacct
>>    submit_bio_checks
>>      blk_partition_remap
>>        bio->bi_iter.bi_sector += p->start_sect
>>        // bio->bi_iter.bi_sector = 0xffc0000000000000 + 65408
>> ..
>> loop_queue_work
>>   loop_handle_cmd
>>    do_req_filebacked
>>     pos = ((loff_t) blk_rq_pos(rq) << 9) + lo->lo_offset // pos < 0
>>     lo_rw_aio
>>       call_read_iter
>>        ext4_dio_read_iter
>>         __iomap_dio_rw
>>          iomap_apply
>>           ext4_iomap_begin
>>             map.m_lblk = offset >> blkbits
>>               ext4_set_iomap
>>               iomap->offset = (u64) map->m_lblk << blkbits
>>               // iomap->offset = 64512
>>           WARN_ON(iomap.offset > pos) // iomap.offset = 64512 and pos < 0

This is wrong, and please update above stack, it seems they're not based
on latest kernel.

Thanks,
Kuai
>>
>> This is unreasonable for start + length > disk->part0.nr_sects. There is
>> already a similar check in blk_add_partition().
>> Fix it by adding a check in bdev_add_partition().
> 
> The checking that you add is blkpg_do_ioctl().
> 
>>
>> Reported-by: Zhihao Cheng <chengzhihao1@huawei.com>
>> Signed-off-by: Zhong Jinghua <zhongjinghua@huawei.com>
>> ---
>>   block/ioctl.c | 7 +++++++
>>   1 file changed, 7 insertions(+)
>>
>> diff --git a/block/ioctl.c b/block/ioctl.c
>> index 9c5f637ff153..3223ea862523 100644
>> --- a/block/ioctl.c
>> +++ b/block/ioctl.c
>> @@ -33,9 +33,16 @@ static int blkpg_do_ioctl(struct block_device *bdev,
>>       if (op == BLKPG_DEL_PARTITION)
>>           return bdev_del_partition(disk, p.pno);
>> +    if (p.start < 0 || p.length <= 0 || p.start + p.length < 0)
>> +        return -EINVAL;
>> +
>>       start = p.start >> SECTOR_SHIFT;
>>       length = p.length >> SECTOR_SHIFT;
>> +    /* length may be equal to 0 after right shift */
>> +    if (!length || start + length > get_capacity(bdev->bd_disk))
>> +        return -EINVAL;
>> +
>>       switch (op) {
>>       case BLKPG_ADD_PARTITION:
>>           /* check if partition is aligned to blocksize */
>>
> 
> The change itself looks good to me, feel free to add:
> 
> Reviewed-by: Yu Kuai <yukuai3@huawei.com>
> 
> .
>