Re: [PATCH] smb/server: fix SMB2_MIN_SUPPORTED_HEADER_SIZE value

[ChenXiaoSong <chenxiaosong.chenxiaosong@linux.dev>]

I will analyze the code in more detail. Thanks for your review.

Thanks,
ChenXiaoSong.

On 12/19/25 09:42, Namjae Jeon wrote:
> On Fri, Dec 19, 2025 at 10:31 AM ChenXiaoSong
> <chenxiaosong.chenxiaosong@linux.dev> wrote:
>>
>> Hi Namjae,
>>
>> `SMB1_MIN_SUPPORTED_HEADER_SIZE` and `SMB2_MIN_SUPPORTED_HEADER_SIZE`
>> are only used in `ksmbd_conn_handler_loop()` to check the PDU size, and
>> seems not to cause slab-out-of-bounds issues.
> Okay, I explained it but you didn't listen... So I can not ACK this patch.
>>
>> Thanks,
>> ChenXiaoSong.
>>
>> On 12/19/25 09:16, Namjae Jeon wrote:
>>> On Fri, Dec 19, 2025 at 10:00 AM ChenXiaoSong
>>> <chenxiaosong.chenxiaosong@linux.dev> wrote:
>>>>
>>>> Hi Namjae,
>>>>
>>>> We should rename them to `SMB1_MIN_SUPPORTED_PDU_SIZE` and
>>>> `SMB2_MIN_SUPPORTED_PDU_SIZE`.
>>>>
>>>> But we "should not" add "+4" to them.
>>> Not adding the +4 will trigger a slab-out-of-bounds issue.
>>> You should check ksmbd_smb2_check_message() and
>>> ksmbd_negotiate_smb_dialect() as well as ksmbd_conn_handler_loop().
>>>>
>>>> The `ksmbd_conn_handler_loop()` function is as follows:
>>>>
>>>> ksmbd_conn_handler_loop()
>>>> {
>>>>      ...
>>>>      pdu_size = get_rfc1002_len(hdr_buf);
>>>>      ...
>>>>      if (pdu_size < SMB1_MIN_SUPPORTED_HEADER_SIZE)
>>>>      ...
>>>>      if (pdu_size < SMB2_MIN_SUPPORTED_HEADER_SIZE)
>>>>      ...
>>>> }
>>>>
>>>> Thanks,
>>>> ChenXiaoSong.
>>>>