From mboxrd@z Thu Jan 1 00:00:00 1970 From: Dekel Peled Subject: Re: [PATCH 1/3] ethdev: add actions to modify TCP header fields Date: Sun, 31 Mar 2019 13:09:46 +0000 Message-ID: References: <1553177917-43297-1-git-send-email-dekelp@mellanox.com> <1553177917-43297-2-git-send-email-dekelp@mellanox.com> <20190329135850.GH4889@6wind.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable Cc: "wenzhuo.lu@intel.com" , "jingjing.wu@intel.com" , "bernard.iremonger@intel.com" , Yongseok Koh , Shahaf Shuler , "dev@dpdk.org" , Ori Kam To: Adrien Mazarguil Return-path: Received: from EUR03-VE1-obe.outbound.protection.outlook.com (mail-eopbgr50062.outbound.protection.outlook.com [40.107.5.62]) by dpdk.org (Postfix) with ESMTP id 6BFFD11A4 for ; Sun, 31 Mar 2019 15:09:48 +0200 (CEST) In-Reply-To: <20190329135850.GH4889@6wind.com> Content-Language: en-US List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" Thanks, PSB. > -----Original Message----- > From: Adrien Mazarguil > Sent: Friday, March 29, 2019 4:59 PM > To: Dekel Peled > Cc: wenzhuo.lu@intel.com; jingjing.wu@intel.com; > bernard.iremonger@intel.com; Yongseok Koh ; > Shahaf Shuler ; dev@dpdk.org; Ori Kam > > Subject: Re: [PATCH 1/3] ethdev: add actions to modify TCP header fields >=20 > Hi Derek, It's Dekel, not Derek. >=20 > It's been a while since I last reviewed something, sorry it took so long. > Some comments below. >=20 > On Thu, Mar 21, 2019 at 04:18:35PM +0200, Dekel Peled wrote: > > Add actions: > > - INC_TCP_SEQ - Increase sequence number in the outermost TCP header. > > - DEC_TCP_SEQ - Decrease sequence number in the outermost TCP > header. > > - INC_TCP_ACK - Increase acknowledgment number in the outermost TCP > > header. > > - DEC_TCP_ACK - Decrease acknowledgment number in the outermost TCP > > header. >=20 > Out of curiosity, are these upcoming/existing OpenFlow actions? If so you > should consider prefixing them with "OF_", otherwise leave them as is. Not related to OpenFlow. >=20 > > Original work by Xiaoyu Min. > > > > Signed-off-by: Dekel Peled > > --- > > doc/guides/prog_guide/rte_flow.rst | 72 > ++++++++++++++++++++++++++++++++++++++ > > lib/librte_ethdev/rte_flow.c | 8 +++++ > > lib/librte_ethdev/rte_flow.h | 60 > +++++++++++++++++++++++++++++++ > > 3 files changed, 140 insertions(+) > > > > diff --git a/doc/guides/prog_guide/rte_flow.rst > > b/doc/guides/prog_guide/rte_flow.rst > > index 0203f4f..bdb817a 100644 > > --- a/doc/guides/prog_guide/rte_flow.rst > > +++ b/doc/guides/prog_guide/rte_flow.rst > > @@ -2345,6 +2345,78 @@ Otherwise, RTE_FLOW_ERROR_TYPE_ACTION > error will be returned. > > | ``mac_addr`` | MAC address | > > +--------------+---------------+ > > > > +Action: ``INC_TCP_SEQ`` > > +^^^^^^^^^^^^^^^^^^^^^^^ > > + > > +Increase sequence number in the outermost TCP header. > > + > > +It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP flow pattern > item. > > +Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned. >=20 > I know this is already present for various SET* actions, but please do no= t tie > actions to the presence of specific pattern items at the API level. You c= an > describe that the lack of a TCP header in matched *traffic* results in > unspecified behavior though. I accept, will change it in all relevant locations. >=20 > Then PMD documentation can describe that the lack of a TCP pattern item > with this action results in RTE_FLOW_ERROR_TYPE_ACTION as a PMD-specific > constraint. >=20 > > + > > +.. _table_rte_flow_action_inc_tcp_seq: > > + > > +.. table:: INC_TCP_SEQ > > + > > + +-----------+--------------------------------------------+ > > + | Field | Value | > > + > +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D+=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D+ > > + | ``value`` | Value to increase TCP sequence number by | > > + +-----------+--------------------------------------------+ >=20 > Nit: unnecessary empty space after "by". Will remove. >=20 > > + > > +Action: ``DEC_TCP_SEQ`` > > +^^^^^^^^^^^^^^^^^^^^^^^ > > + > > +Decrease sequence number in the outermost TCP header. > > + > > +It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP flow pattern > item. > > +Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned. >=20 > Same comment as above. >=20 > > + > > +.. _table_rte_flow_action_dec_tcp_seq: > > + > > +.. table:: DEC_TCP_SEQ > > + > > + +-----------+--------------------------------------------+ > > + | Field | Value | > > + > +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D+=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D+ > > + | ``value`` | Value to decrease TCP sequence number by | > > + +-----------+--------------------------------------------+ > > + > > +Action: ``INC_TCP_ACK`` > > +^^^^^^^^^^^^^^^^^^^^^^^ > > + > > +Increase acknowledgment number in the outermost TCP header. > > + > > +It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP flow pattern > item. > > +Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned. >=20 > Ditto. >=20 > > + > > +.. _table_rte_flow_action_inc_tcp_ack: > > + > > +.. table:: INC_TCP_ACK > > + > > + +-----------+--------------------------------------------------+ > > + | Field | Value | > > + > +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D+=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D > =3D=3D=3D=3D=3D+ > > + | ``value`` | Value to increase TCP acknowledgment number by | > > + +-----------+--------------------------------------------------+ > > + > > +Action: ``DEC_TCP_ACK`` > > +^^^^^^^^^^^^^^^^^^^^^^^ > > + > > +Decrease acknowledgment number in the outermost TCP header. > > + > > +It must be used with a valid RTE_FLOW_ITEM_TYPE_TCP flow pattern > item. > > +Otherwise, RTE_FLOW_ERROR_TYPE_ACTION error will be returned. >=20 > Ditto. >=20 > > + > > +.. _table_rte_flow_action_dec_tcp_ack: > > + > > +.. table:: DEC_TCP_ACK > > + > > + +-----------+--------------------------------------------------+ > > + | Field | Value | > > + > +=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D+=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D= =3D=3D=3D=3D=3D=3D=3D=3D > =3D=3D=3D=3D=3D+ > > + | ``value`` | Value to decrease TCP acknowledgment number by | > > + +-----------+--------------------------------------------------+ > > + > > Negative types > > ~~~~~~~~~~~~~~ > > > > diff --git a/lib/librte_ethdev/rte_flow.c > > b/lib/librte_ethdev/rte_flow.c index 3277be1..589d0b9 100644 > > --- a/lib/librte_ethdev/rte_flow.c > > +++ b/lib/librte_ethdev/rte_flow.c > > @@ -143,6 +143,14 @@ struct rte_flow_desc_data { > > MK_FLOW_ACTION(SET_TTL, sizeof(struct rte_flow_action_set_ttl)), > > MK_FLOW_ACTION(SET_MAC_SRC, sizeof(struct > rte_flow_action_set_mac)), > > MK_FLOW_ACTION(SET_MAC_DST, sizeof(struct > rte_flow_action_set_mac)), > > + MK_FLOW_ACTION(INC_TCP_SEQ, > > + sizeof(struct rte_flow_action_modify_tcp_seq)), > > + MK_FLOW_ACTION(DEC_TCP_SEQ, > > + sizeof(struct rte_flow_action_modify_tcp_seq)), > > + MK_FLOW_ACTION(INC_TCP_ACK, > > + sizeof(struct rte_flow_action_modify_tcp_ack)), > > + MK_FLOW_ACTION(DEC_TCP_ACK, > > + sizeof(struct rte_flow_action_modify_tcp_ack)), > > }; > > > > static int > > diff --git a/lib/librte_ethdev/rte_flow.h > > b/lib/librte_ethdev/rte_flow.h index c0fe879..74cd03e 100644 > > --- a/lib/librte_ethdev/rte_flow.h > > +++ b/lib/librte_ethdev/rte_flow.h > > @@ -1651,6 +1651,46 @@ enum rte_flow_action_type { > > * See struct rte_flow_action_set_mac. > > */ > > RTE_FLOW_ACTION_TYPE_SET_MAC_DST, > > + > > + /** > > + * Increase sequence number in the outermost TCP header. > > + * > > + * If flow pattern does not define a valid > RTE_FLOW_ITEM_TYPE_TCP, > > + * the PMD should return a RTE_FLOW_ERROR_TYPE_ACTION error. >=20 > Ditto. >=20 > > + * > > + * See struct rte_flow_action_modify_tcp_seq > > + */ > > + RTE_FLOW_ACTION_TYPE_INC_TCP_SEQ, > > + > > + /** > > + * Decrease sequence number in the outermost TCP header. > > + * > > + * If flow pattern does not define a valid > RTE_FLOW_ITEM_TYPE_TCP, > > + * the PMD should return a RTE_FLOW_ERROR_TYPE_ACTION error. >=20 > Ditto. >=20 > > + * > > + * See struct rte_flow_action_modify_tcp_seq > > + */ > > + RTE_FLOW_ACTION_TYPE_DEC_TCP_SEQ, > > + > > + /** > > + * Increase acknowledgment number in the outermost TCP header. > > + * > > + * If flow pattern does not define a valid > RTE_FLOW_ITEM_TYPE_TCP, > > + * the PMD should return a RTE_FLOW_ERROR_TYPE_ACTION error. > > + * >=20 > Ditto. >=20 > > + * See struct rte_flow_action_modify_tcp_ack > > + */ > > + RTE_FLOW_ACTION_TYPE_INC_TCP_ACK, > > + > > + /** > > + * Decrease acknowledgment number in the outermost TCP header. > > + * > > + * If flow pattern does not define a valid > RTE_FLOW_ITEM_TYPE_TCP, > > + * the PMD should return a RTE_FLOW_ERROR_TYPE_ACTION error. > > + * >=20 > Ditto. >=20 > > + * See struct rte_flow_action_modify_tcp_ack > > + */ > > + RTE_FLOW_ACTION_TYPE_DEC_TCP_ACK, > > }; > > > > /** > > @@ -2122,6 +2162,26 @@ struct rte_flow_action_set_mac { > > uint8_t mac_addr[ETHER_ADDR_LEN]; > > }; > > > > +/** >=20 > Experimental tag is missing. Will add it here and elsewhere. >=20 > > + * RTE_FLOW_ACTION_TYPE_INC_TCP_SEQ > > + * RTE_FLOW_ACTION_TYPE_DEC_TCP_SEQ > > + * > > + * Increase/Decrease outermost TCP's sequence number >=20 > Suggestion: >=20 > Increase/decrease outermost TCP sequence number. Accepted, will change here and elsewhere. >=20 > > + */ > > +struct rte_flow_action_modify_tcp_seq { > > + rte_be32_t value; >=20 > Field documentation is mandatory, e.g.: >=20 > rte_be32_t value; /**< Value to add/subtract. */ Will add it. >=20 > Beside, I'm not sure this value should be big endian since it's not store= d as is > in the TCP header; it's used by the host system to compute a new value. It isn't used by the host, but sent to NIC as part of action. >=20 > Also what about having another field to specify what needs to be done wit= h > this value (e.g. add/sub/set - in which case big endian makes sense) to > reduce the number of new actions? Something like: >=20 > struct rte_flow_action_mod_tcp_seq { > enum rte_flow_action_mod_tcp_seq_op { > RTE_FLOW_ACTION_MOD_TCP_SEQ_OP_ADD, > RTE_FLOW_ACTION_MOD_TCP_SEQ_OP_SUB, > RTE_FLOW_ACTION_MOD_TCP_SEQ_OP_SET, > } op; /**< Operation to perform. */ > rte_be32_t value; /**< Value to use with operation. */ }; >=20 > > +}; > > + > > +/** >=20 > Experimental tag also missing. >=20 > > + * RTE_FLOW_ACTION_TYPE_INC_TCP_ACK > > + * RTE_FLOW_ACTION_TYPE_DEC_TCP_ACK > > + * > > + * Increase/Decrease TCP's acknowledgment number. >=20 > Suggestion: >=20 > Increase/decrease outermost TCP acknowledgment number. Accepted, will change. >=20 > > + */ > > +struct rte_flow_action_modify_tcp_ack { > > + rte_be32_t value; >=20 > Field documentation also missing. >=20 > > +}; > > + > > /* > > * Definition of a single action. > > * > > -- > > 1.8.3.1 > > >=20 > -- > Adrien Mazarguil > 6WIND