DPDK-dev Archive on lore.kernel.org
 help / color / Atom feed
From: "Yigit, Ferruh" <ferruh.yigit@linux.intel.com>
To: Akhil Goyal <akhil.goyal@nxp.com>,
	Anoob Joseph <anoobj@marvell.com>,
	Adrien Mazarguil <adrien.mazarguil@6wind.com>,
	Declan Doherty <declan.doherty@intel.com>,
	Pablo de Lara <pablo.de.lara.guarch@intel.com>,
	Thomas Monjalon <thomas@monjalon.net>
Cc: Jerin Jacob Kollanukkaran <jerinj@marvell.com>,
	Narayana Prasad Raju Athreya <pathreya@marvell.com>,
	Ankur Dwivedi <adwivedi@marvell.com>,
	Shahaf Shuler <shahafs@mellanox.com>,
	Hemant Agrawal <hemant.agrawal@nxp.com>,
	Matan Azrad <matan@mellanox.com>,
	Yongseok Koh <yskoh@mellanox.com>,
	Wenzhuo Lu <wenzhuo.lu@intel.com>,
	Konstantin Ananyev <konstantin.ananyev@intel.com>,
	Radu Nicolau <radu.nicolau@intel.com>,
	"dev@dpdk.org" <dev@dpdk.org>
Subject: Re: [dpdk-dev] [RFC] ethdev: allow multiple security sessions to use one rte flow
Date: Tue, 8 Oct 2019 14:00:26 +0100
Message-ID: <b3d4b14c-42ab-f37d-1af6-94cf3f649a0f@linux.intel.com> (raw)
In-Reply-To: <VE1PR04MB6639DC2D5395F505C1A4DF92E6A80@VE1PR04MB6639.eurprd04.prod.outlook.com>

On 8/19/2019 8:09 AM, Akhil Goyal wrote:
> Hi Anoob,
>> Hi Akhil,
>>>>>>>> The rte_security API which enables inline protocol/crypto
>>>>>>>> feature mandates that for every security session an rte_flow is
>>> created.
>>>>>>>> This would internally translate to a rule in the hardware
>>>>>>>> which would do packet
>>>>>>> classification.
>>>>>>>> In rte_securty, one SA would be one security session. And if
>>>>>>>> an rte_flow need to be created for every session, the number
>>>>>>>> of SAs supported by an inline implementation would be limited
>>>>>>>> by the number of rte_flows the PMD would be able to support.
>>>>>>>> If the fields SPI & IP addresses are allowed to be a range,
>>>>>>>> then this limitation can be overcome. Multiple flows will be
>>>>>>>> able to use one rule for SECURITY processing. In this case,
>>>>>>>> the security session provided as
>>>>>>> conf would be NULL.
>>>>> SPI values are normally used to uniquely identify the SA that need
>>>>> to be applied on a particular flow.
>>>>> I believe SPI value should not be a range for applying a particular
>>>>> SA or session.
>>>>> Plain packet IP addresses can be a range. That is not an issue.
>>>>> Multiple plain packet flows can use the same session/SA.
>>>>> Why do you feel that security session provided should be NULL to
>>>>> support multiple flows.
>>>>> How will the keys and other SA related info will be passed to the
>>> driver/HW.
>>>> [Anoob] The SA configuration would be done via rte_security session.
>>>> The proposal here only changes the 1:1 dependency of rte_flow and
>>>> rte_security session.
>>> I don't see this dependency for rte_flow and security session.
>>> Multiple flows can be configured to use the same security session.
>>>> The h/w could use SPI field in the received packet to identify SA(ie,
>>>> rte_security session). If the h/w allows to index into a table which
>>>> holds SA information, then per SPI rte_flow is not required. This is
>>>> in fact our case. And for PMDs which doesn't do it this way,
>>>> rte_flow_validate() would fail and then per SPI rte_flow would require to
>>> be created.
>>> I am not able to understand the issue here. Flow are validated based on
>>> some pattern, You can identify the flow based on some parameter(currently
>>> it is spi in case of inline crypto and also your case).
>>> You can perform some action based on the security session that you have
>>> created before validating the flow And that session creation is nowhere
>>> linked to the type of flow. You can use the same session for as many flows
>>> you want.
>>>> In the present model, a security session is created, and then rte_flow
>>>> will connect ESP packets with one SPI to one security session.
>>>> Instead, when we create the security session, h/w can populate entries
>>>> in a DB that would be accessed during data path handling. And the
>>>> rte_flow could say, all SPI in some range gets inline processed with the
>>> security session identified with its SPI.
>>>> Our PMD supports limited number of flow entries but our h/w can do SA
>>>> lookup without flow entries(using SPI instead). So the current
>>>> approach of one flow per session is creating an artificial limit to the number
>>> of SAs that can be supported.
>>> Ok now I got it. You want to configure a single flow with multiple sessions in
>>> it.
>>> But defining a range in SPI and tunnel IP addresses does not make sense. In
>>> real world applications, Sessions can be created and destroyed at any time
>>> with varied values of SPI and tunnel IPs. How can One put a range to that.
>>> I would rather say, you actually do not need the rte_flows to be configured
>>> for Inline protocol processing. You have configured all the session info in the
>>> hw while Creating the session and your H/W will be able to identify on the
>>> basis of SPI value which It has stored in the DB and do all the processing.
>> [Anoob] Yes. That is the model being followed right now. Concern is, whether
>> this would be deviating from the spec. In other words, we could have devices
>> which would need rte_flow for every rte_security session (ixgbe needs for inline
>> crypto), and then we could have devices which doesn't need per session
>> rte_flow (which is our case). What do you think is the right approach for
>> supporting both kinds of devices?
> Inline proto case is not using rte_flow at the moment.
> And as far as I understand, you also do not need rte_flow to be configured.
> Inline crypto cases are mainly for Intel and Mellanox cases which only supported
> Inline crypto. For Protocol offload cases, I don't feel we need rte_flow as all information
> related to ipsec is already there when we call the session create. Rte_flows are used
> For segregation of ethernet traffic for classification which can be configured for various factors
> as well.
>>> What are the changes that you need in the ipsec-secgw for inline proto to
>>> work, there is No flow processing currently in the inline proto case. Will it not
>>> work as is for you?
>> [Anoob] In ipsec-secgw, a default flow would be created per security enabled
>> port with 'conf=NULL' & SPI = 'ANY'. Flow validate would be done to make sure
>> the underlying PMD supports it. For PMDs which doesn't support this model, per
>> SA flow would be created.
> Why do you need that flow as well. You have all the information in the session already.
> You can process the packets based on that information. Isn't it?
> Current implementation in application is good enough in my opinion.
>>> Atleast for NXP devices we are able to work as is without any issue.
>> [Anoob] Just curious, would having such a dependency on rte_flow be an issue
>> for NXP devices?
> As of now I do not have any comment on this. We are not using rte_flow in our work as of now.
> It is kind of POC for us, we may not upstream it.
> This will depend on the changes that will be done.

Is there any follow up to the RFC? Is it still valid?

>>>>>>>> Application should do an rte_flow_validate() to make sure the
>>>>>>>> flow is supported on the PMD.
>>>>>>>> Signed-off-by: Anoob Joseph <anoobj@marvell.com>
>>>>>>>> ---
>>>>>>>>  lib/librte_ethdev/rte_flow.h | 6 ++++++
>>>>>>>>  1 file changed, 6 insertions(+)
>>>>>>>> diff --git a/lib/librte_ethdev/rte_flow.h
>>>>>>>> b/lib/librte_ethdev/rte_flow.h index f3a8fb1..4977d3c 100644
>>>>>>>> --- a/lib/librte_ethdev/rte_flow.h
>>>>>>>> +++ b/lib/librte_ethdev/rte_flow.h
>>>>>>>> @@ -1879,6 +1879,12 @@ struct rte_flow_action_meter {
>>>>>>>>   * direction.
>>>>>>>>   *
>>>>>>>>   * Multiple flows can be configured to use the same security
>>> session.
>>>>>>>> + *
>>>>>>>> + * The NULL value is allowed for security session. If
>>>>>>>> + security session is NULL,
>>>>>>>> + * then SPI field in ESP flow item and IP addresses in flow
>>>>>>>> + items 'IPv4' and
>>>>>>>> + * 'IPv6' will be allowed to be a range. The rule thus
>>>>>>>> + created can enable
>>>>>>>> + * SECURITY processing on multiple flows.
>>> What you intent here is " The rule thus created can enable multiple security
>>> sessions on a single rte flow"
>>> Regards,
>>> Akhil

  reply index

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-07-24 14:17 Anoob Joseph
2019-08-02  5:35 ` Anoob Joseph
2019-08-14  9:22   ` Anoob Joseph
2019-08-14 11:07     ` Akhil Goyal
2019-08-15  6:49       ` Anoob Joseph
2019-08-15  9:48         ` Ananyev, Konstantin
2019-08-16  3:24           ` Anoob Joseph
2019-08-16  8:32         ` Akhil Goyal
2019-08-16 10:12           ` Anoob Joseph
2019-08-19  7:09             ` Akhil Goyal
2019-10-08 13:00               ` Yigit, Ferruh [this message]
2019-10-09 10:55                 ` [dpdk-dev] [EXT] " Anoob Joseph
2019-12-03  5:32                   ` Anoob Joseph

Reply instructions:

You may reply publically to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=b3d4b14c-42ab-f37d-1af6-94cf3f649a0f@linux.intel.com \
    --to=ferruh.yigit@linux.intel.com \
    --cc=adrien.mazarguil@6wind.com \
    --cc=adwivedi@marvell.com \
    --cc=akhil.goyal@nxp.com \
    --cc=anoobj@marvell.com \
    --cc=declan.doherty@intel.com \
    --cc=dev@dpdk.org \
    --cc=hemant.agrawal@nxp.com \
    --cc=jerinj@marvell.com \
    --cc=konstantin.ananyev@intel.com \
    --cc=matan@mellanox.com \
    --cc=pablo.de.lara.guarch@intel.com \
    --cc=pathreya@marvell.com \
    --cc=radu.nicolau@intel.com \
    --cc=shahafs@mellanox.com \
    --cc=thomas@monjalon.net \
    --cc=wenzhuo.lu@intel.com \
    --cc=yskoh@mellanox.com \


* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link

DPDK-dev Archive on lore.kernel.org

Archives are clonable:
	git clone --mirror https://lore.kernel.org/dpdk-dev/0 dpdk-dev/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 dpdk-dev dpdk-dev/ https://lore.kernel.org/dpdk-dev \
	public-inbox-index dpdk-dev

Example config snippet for mirrors

Newsgroup available over NNTP:

AGPL code for this site: git clone https://public-inbox.org/public-inbox.git