Re: [PATCH ipsec-next v9] xfrm: Add Direction to the SA in or out

[Nicolas Dichtel <nicolas.dichtel@6wind.com>]

Le 11/04/2024 à 09:14, Steffen Klassert a écrit :
> On Wed, Apr 10, 2024 at 08:32:20AM +0200, Nicolas Dichtel wrote:
>> Le 09/04/2024 à 19:56, Antony Antony a écrit :
>>> This patch introduces the 'dir' attribute, 'in' or 'out', to the
>>> xfrm_state, SA, enhancing usability by delineating the scope of values
>>> based on direction. An input SA will now exclusively encompass values
>>> pertinent to input, effectively segregating them from output-related
>>> values. This change aims to streamline the configuration process and
>>> improve the overall clarity of SA attributes.
>>>
>>> This feature sets the groundwork for future patches, including
>>> the upcoming IP-TFS patch.
>>>
>>> Signed-off-by: Antony Antony <antony.antony@secunet.com>
>>> ---
>>> v8->v9:
>>>  - add validation XFRM_STATE_ICMP not allowed on OUT SA.
>>>
>>> v7->v8:
>>>  - add extra validation check on replay window and seq
>>>  - XFRM_MSG_UPDSA old and new SA should match "dir"
>>>
>>> v6->v7:
>>>  - add replay-window check non-esn 0 and ESN 1.
>>>  - remove :XFRMA_SA_DIR only allowed with HW OFFLOAD
>> Why? I still think that having an 'input' SA used in the output path is wrong
>> and confusing.
> 
> I don't think this can happen. This patch does not change the
> state lookups, so we should match the correct state as it was
> before that patch.
This is the point. The user can set whatever direction in the SA, there is no
check. He can set the dir to 'output' even if the SA is used in input.

> 
> On the long run, we should make the direction a lookup key.
> That should have happened with the initial implemenatation
> already, now ~25 years later we would have to maintain the
> old input/output combined SADB and two new ones where input
> and output states are separated. 
>