introduce Stateful Flow Table
cover introduce Stateful Flow Table
The RFC introduces Stateful Flow Table (SFT) API and changes needed in both ethdev an RTE flow to support SFT functionality.
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.
Andrey Vesnovaty (3):
ethdev: add item/action for SFT
ethdev: support SFT APIs
sft: introduce API
lib/librte_ethdev/rte_ethdev.c | 7 + lib/librte_ethdev/rte_ethdev.h | 16 + lib/librte_ethdev/rte_ethdev_core.h | 1 + lib/librte_ethdev/rte_flow.h | 84 +++ lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 ++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 10 files changed, 1213 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.map
Signature · 1 line
2.26.2
RFC 1/3 ethdev: add item/action for SFT
Attach SFT flow context to packet with SFT action. Match on SFT flow context (attached to packet), with SFT item.
--- lib/librte_ethdev/rte_flow.h | 84 ++++++++++++++++++++++++++++++++++++ 1 file changed, 84 insertions(+)
lib/librte_ethdev/rte_flow.h +84 -0
@@ -537,6 +537,12 @@ enum rte_flow_item_type {*/ RTE_FLOW_ITEM_TYPE_ECPRI , +/** +* Matches SFT context (see fields of struct rte_flow_item_sft). +* +* See struct rte_flow_item_sft. +*/ + RTE_FLOW_ITEM_TYPE_SFT, };/**
@@ -1579,6 +1585,54 @@ static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { };#endif +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ITEM_TYPE_SFT +* +* Matches context of flow in SFT table. +* +* 5-tuple: src/dest IP + src/dest port + IP protocol. +* zone: application defined value cupled with 5-tuple to identify flow, +* example - VxLAN, VLAN. +* SFT: Statfull flow table +* SFT in scope of ethernet device (port) is HW offloaded lookup table +* where key is zone + 5-tuple & value is statefull flow context. +* Contents of the SFT maintained by SFT PMD (see SFT PMD API in rte_sft). +* +* The structure describes SFT flow context. +* All the fields of the structure, except @p fid, should be considered as +* user defined. +* The @p fid assigned by RTE SFT & used as unique flow identifier. +* SFT context attached to packet by action ``SFT`` (see RTE_FLOW_ACTION_SFT). +* +* SFT default context defined as context attached to packet when there is no +* entry for the flow in SFT. The @p state has application reserved value +* meaning that SFT context for the packet undefined since entry wasn't found +* in SFT. If state 'undefined' then @p zone should be valid othervice @p fid +* should be valid. +* +* Context considered virtual since the method of storing this info on packet +* is PMD/implementation specific & may involve mapping methods if there is +* 'not enough bits' to store entire contents of struct rte_flow_item_sft. +* +* Maximal value/size of each field depends on HW capabilities and considered +* as implementation specific. +*/ +struct rte_flow_item_sft { +union { +uint32_t fid;/**< SFT flow identifier. */ +uint32_t zone;/**< Zone assigned to flow. */ + }; +uint8_t state;/**< User defined flow state. */ +uint8_t fid_valid:1 ;/**< fid field validity bit. */ +uint8_t zone_valid:1 ;/**< zone fieald validity bit. */ +uint8_t state_valid:1 ;/**< state fieald validity bit. */ +uint8_t user_data_size;/**< user_data buffer size. */ +uint8_t *user_data;/**< Arbitrary user data. */ +}; +/** * Matching pattern item definition. *
@@ -2132,6 +2186,15 @@ enum rte_flow_action_type {* see enum RTE_ETH_EVENT_FLOW_AGED */ RTE_FLOW_ACTION_TYPE_AGE , + +/** +* RTE_FLOW_ACTION_TYPE_SFT +* +* Set SFT context and redirect to continue processing. +* +* See struct rte_flow_action_sft. +*/ + RTE_FLOW_ACTION_TYPE_SFT, };/**
@@ -2721,6 +2784,27 @@ rte_flow_dynf_metadata_set(struct rte_mbuf *m, uint32_t v) *RTE_FLOW_DYNF_METADATA (m) = v; } +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ACTION_TYPE_SFT +* +* Attaches an SFT context (see struct rte_flow_item_sft) to packet. +* +* Performs lookup by *zone* and 5-tuple in SFT; if entry found the related SFT +* context will be attached othervise default SFT context attached (see +* 'SFT default context' in struct rte_flow_item_sft description). +* Adding action of type ``SFT`` to the list of rule actions may impose +* limitations on other rule actions added to the list, depending on specific +* PMD implementation. +* +* For 5-tuple, zone & SFT definitions see `struct rte_flow_item_sft`. +*/ +struct rte_flow_action_sft { +uint32_t zone;/**< Zone for lookup in SFT */ +}; +/* * Definition of a single action. *
Signature · 1 line
2.26.2
Re: [dpdk-dev] [RFC 1/3] ethdev: add item/action for SFT
Hi Andrey,
PSB
Quoted 18 lines
-----Original Message-----
From: Andrey Vesnovaty <[email protected]>
Sent: Wednesday, September 9, 2020 11:30 PMAttach SFT flow context to packet with SFT action. Match on SFT flow context (attached to packet), with SFT item.
Signed-off-by Andrey Vesnovaty <[email protected]>--- lib/librte_ethdev/rte_flow.h | 84 ++++++++++++++++++++++++++++++++++++ 1 file changed, 84 insertions(+)lib/librte_ethdev/rte_flow.h +0
-0
lib/librte_ethdev/rte_flow.h enum rte_flow_item_type { · line 540
Quoted 4 lines
lib/librte_ethdev/rte_flow.h +1
-0@@ -537,6 +537,12 @@ enum rte_flow_item_type {*/ RTE_FLOW_ITEM_TYPE_ECPRI, +/**
You are missing the Meta, tag not relevant for RFC but please notice for the patch.
Quoted 53 lines
lib/librte_ethdev/rte_flow.h +26
-0+* Matches SFT context (see fields of struct rte_flow_item_sft). +* +* See struct rte_flow_item_sft. +*/ + RTE_FLOW_ITEM_TYPE_SFT, };/** @@ -1579,6 +1585,54 @@ static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { };#endif +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ITEM_TYPE_SFT +* +* Matches context of flow in SFT table. +* +* 5-tuple: src/dest IP + src/dest port + IP protocol. +* zone: application defined value cupled with 5-tuple to identify flow, +* example - VxLAN, VLAN. +* SFT: Statfull flow table +* SFT in scope of ethernet device (port) is HW offloaded lookup table +* where key is zone + 5-tuple & value is statefull flow context. +* Contents of the SFT maintained by SFT PMD (see SFT PMD API in rte_sft). +* +* The structure describes SFT flow context. +* All the fields of the structure, except @p fid, should be considered as +* user defined. +* The @p fid assigned by RTE SFT & used as unique flow identifier. +* SFT context attached to packet by action ``SFT`` (see
RTE_FLOW_ACTION_SFT).
+ *
+ * SFT default context defined as context attached to packet when there is no
+ * entry for the flow in SFT. The @p state has application reserved value
+ * meaning that SFT context for the packet undefined since entry wasn't found
+ * in SFT. If state 'undefined' then @p zone should be valid othervice @p fid
+ * should be valid.
+ *
+ * Context considered virtual since the method of storing this info on packet
+ * is PMD/implementation specific & may involve mapping methods if there is
+ * 'not enough bits' to store entire contents of structrte_flow_item_sft.
+ *
+ * Maximal value/size of each field depends on HW capabilities and considered
+ * as implementation specific.
+ */lib/librte_ethdev/rte_flow.h +3
-0+struct rte_flow_item_sft { +union { +uint32_t fid;/**< SFT flow identifier. */
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1635
Quoted 10 lines
lib/librte_ethdev/rte_flow.h +10
-0+uint32_t zone;/**< Zone assigned to flow. */ + }; +uint8_t state;/**< User defined flow state. */ + uint8_t fid_valid:1; /**< fid field validity bit. */ + uint8_t zone_valid:1; /**< zone fieald validity bit. */ + uint8_t state_valid:1; /**< state fieald validity bit. */ +uint8_t user_data_size;/**< user_data buffer size. */ +uint8_t *user_data;/**< Arbitrary user data. */ +}; +
This object is only used to match and not set so
why do we need the union? I understand that later when reporting to the SFT in the application layer sometimes you will get zone while other time you will get fid.
From rte flow you are matching on given object which is 32 bit.
What are the matchable fields? (fid / zone / user_data / fid_valid ... )
Do you think that some of the times the match will be on he fid other on the zone? If so they should not be union.
I think zone is the responsibility of the application to save and to match. So I don't see why it is needed here.
Quoted 47 lines
/**
* Matching pattern item definition.
*lib/librte_ethdev/rte_flow.h +30
-0@@ -2132,6 +2186,15 @@ enum rte_flow_action_type {* see enum RTE_ETH_EVENT_FLOW_AGED */ RTE_FLOW_ACTION_TYPE_AGE , + +/** +* RTE_FLOW_ACTION_TYPE_SFT +* +* Set SFT context and redirect to continue processing. +* +* See struct rte_flow_action_sft. +*/ + RTE_FLOW_ACTION_TYPE_SFT, };/** @@ -2721,6 +2784,27 @@ rte_flow_dynf_metadata_set(struct rte_mbuf *m, uint32_t v) *RTE_FLOW_DYNF_METADATA (m ) = v; } +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ACTION_TYPE_SFT +* +* Attaches an SFT context (see struct rte_flow_item_sft) to packet. +* +* Performs lookup by *zone* and 5-tuple in SFT; if entry found the related SFT +* context will be attached othervise default SFT context attached (see +* 'SFT default context' in struct rte_flow_item_sft description). +* Adding action of type ``SFT`` to the list of rule actions may impose +* limitations on other rule actions added to the list, depending on specific +* PMD implementation. +* +* For 5-tuple, zone & SFT definitions see `struct rte_flow_item_sft`. +*/ +struct rte_flow_action_sft { +uint32_t zone;/**< Zone for lookup in SFT */ +}; +/* * Definition of a single action.
Quoted 3 lines
*
--
2.26.2
Thanks,
Ori
Re: [dpdk-dev] [RFC 1/3] ethdev: add item/action for SFT
Quoted 11 lines · Ori Kam
Hi Andrey,
PSB
Quoted 7 lines
-----Original Message-----
From: Andrey Vesnovaty <[email protected]>
Sent: Wednesday, September 9, 2020 11:30 PMAttach SFT flow context to packet with SFT action. Match on SFT flow context (attached to packet), with SFT item.
Since it is the first patch which introduces SFT, it would be useful define abbreviation in the changeset description.
The description does not explain what is SFT flow context. It does not explain why we should attach it using action and why we should match on it using pattern items.
Please, help the reader to understand how it is supposed to be used in the future.
Quoted 30 lines
Quoted 11 lines
Signed-off-by Andrey Vesnovaty <[email protected]>--- lib/librte_ethdev/rte_flow.h | 84 ++++++++++++++++++++++++++++++++++++ 1 file changed, 84 insertions(+)lib/librte_ethdev/rte_flow.h +0
-0lib/librte_ethdev/rte_flow.h enum rte_flow_item_type { · line 540
Quoted 4 lines
lib/librte_ethdev/rte_flow.h +1
-0@@ -537,6 +537,12 @@ enum rte_flow_item_type {*/ RTE_FLOW_ITEM_TYPE_ECPRI, +/** You are missing the Meta, tag not relevant for RFC but please notice for the patch.
Quoted 13 lines
lib/librte_ethdev/rte_flow.h +5
-0+* Matches SFT context (see fields of struct rte_flow_item_sft). +* +* See struct rte_flow_item_sft. +*/ + RTE_FLOW_ITEM_TYPE_SFT, };/** @@ -1579,6 +1585,54 @@ static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { };#endif
Quoted 8 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1595
Quoted 8 lines
lib/librte_ethdev/rte_flow.h +8
-0+/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ITEM_TYPE_SFT +* +* Matches context of flow in SFT table. +*
It looks like the content of SFT context is defined below. If so, please, say so.
Quoted 2 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1597
Quoted 2 lines
lib/librte_ethdev/rte_flow.h +2
-0+* 5-tuple: src/dest IP + src/dest port + IP protocol. +* zone: application defined value cupled with 5-tuple to identify flow,
cupled -> coupled
Quoted 2 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1599
Quoted 2 lines
lib/librte_ethdev/rte_flow.h +2
-0+* example - VxLAN, VLAN. +* SFT: Statfull flow table
Statfull -> Stateful
Quoted 1 line
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1600
Quoted 1 line
lib/librte_ethdev/rte_flow.h +1
-0+* SFT in scope of ethernet device (port) is HW offloaded lookup table
ethernet -> Ethernet
Quoted 1 line
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1601
Quoted 1 line
lib/librte_ethdev/rte_flow.h +1
-0+* where key is zone + 5-tuple & value is statefull flow context.
statefull -> stateful
Quoted 4 lines
Quoted 4 lines
lib/librte_ethdev/rte_flow.h +4
-0+* Contents of the SFT maintained by SFT PMD (see SFT PMD API in rte_sft). +* +* The structure describes SFT flow context. +* All the fields of the structure, except @p fid, should be considered as
Quoted 8 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1613
Quoted 8 lines
lib/librte_ethdev/rte_flow.h +8
-0+* user defined. +* The @p fid assigned by RTE SFT & used as unique flow identifier. +* SFT context attached to packet by action ``SFT`` (see RTE_FLOW_ACTION_SFT). +* +* SFT default context defined as context attached to packet when there is no +* entry for the flow in SFT. The @p state has application reserved value +* meaning that SFT context for the packet undefined since entry wasn't found +* in SFT. If state 'undefined' then @p zone should be valid othervice @p fid
othervice -> otherwise
Quoted 6 lines
Quoted 6 lines
lib/librte_ethdev/rte_flow.h +6
-0+* should be valid. +* +* Context considered virtual since the method of storing this info on packet +* is PMD/implementation specific & may involve mapping methods if there is +* 'not enough bits' to store entire contents of struct rte_flow_item_sft. +*
Quoted 8 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1627
Quoted 8 lines
lib/librte_ethdev/rte_flow.h +8
-0+* Maximal value/size of each field depends on HW capabilities and considered +* as implementation specific. +*/ +struct rte_flow_item_sft { +union { +uint32_t fid;/**< SFT flow identifier. */ +uint32_t zone;/**< Zone assigned to flow. */ + };
Is RTE_STD_C11 missing?
Quoted 3 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1630
Quoted 3 lines
lib/librte_ethdev/rte_flow.h +3
-0+uint8_t state;/**< User defined flow state. */ + uint8_t fid_valid:1; /**< fid field validity bit. */ + uint8_t zone_valid:1; /**< zone fieald validity bit. */
fieald -> field
Quoted 1 line
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1631
Quoted 1 line
lib/librte_ethdev/rte_flow.h +1
-0+ uint8_t state_valid:1; /**< state fieald validity bit. */
fieald -> field
Quoted 5 lines
lib/librte_ethdev/rte_flow.h static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { · line 1635
Quoted 4 lines
lib/librte_ethdev/rte_flow.h +4
-0+uint8_t user_data_size;/**< user_data buffer size. */ +uint8_t *user_data;/**< Arbitrary user data. */ +}; +This object is only used to match and not set so
Quoted 8 lines
why do we need the union? I understand that later when reporting to the SFT in the application layer sometimes you will get zone while other time you will get fid.
From rte flow you are matching on given object which is 32 bit.
What are the matchable fields? (fid / zone /user_data/fid_valid... )
Do you think that some of the times the match will be on he fid other on the zone? If so they should not be union.
I think zone is the responsibility of the application to save and to match. So I don't see why it is needed here.
+1
Quoted 25 lines
Quoted 25 lines
lib/librte_ethdev/rte_flow.h +0
-0/** * Matching pattern item definition. *
lib/librte_ethdev/rte_flow.h +9
-0@@ -2132,6 +2186,15 @@ enum rte_flow_action_type {* see enum RTE_ETH_EVENT_FLOW_AGED */ RTE_FLOW_ACTION_TYPE_AGE , + +/** +* RTE_FLOW_ACTION_TYPE_SFT +* +* Set SFT context and redirect to continue processing. +* +* See struct rte_flow_action_sft. +*/ + RTE_FLOW_ACTION_TYPE_SFT, };/** @@ -2721,6 +2784,27 @@ rte_flow_dynf_metadata_set(struct rte_mbuf *m, uint32_t v) *RTE_FLOW_DYNF_METADATA (m ) = v; }
Quoted 10 lines
lib/librte_ethdev/rte_flow.h rte_flow_dynf_metadata_set(struct rte_mbuf *m, uint32_t v) · line 2796
Quoted 10 lines
lib/librte_ethdev/rte_flow.h +10
-0+/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ACTION_TYPE_SFT +* +* Attaches an SFT context (see struct rte_flow_item_sft) to packet. +* +* Performs lookup by *zone* and 5-tuple in SFT; if entry found the related SFT +* context will be attached othervise default SFT context attached (see
othervise -> otherwise
Quoted 20 lines
Quoted 13 lines
lib/librte_ethdev/rte_flow.h +11
-0+* 'SFT default context' in struct rte_flow_item_sft description). +* Adding action of type ``SFT`` to the list of rule actions may impose +* limitations on other rule actions added to the list, depending on specific +* PMD implementation. +* +* For 5-tuple, zone & SFT definitions see `struct rte_flow_item_sft`. +*/ +struct rte_flow_action_sft { +uint32_t zone;/**< Zone for lookup in SFT */ +}; +/* * Definition of a single action. Quoted 3 lines
*
--
2.26.2Thanks,
Ori
RFC 2/3 ethdev: support SFT APIs
In reply to Andrey Vesnovatyethdev updated to support SFT lookup offload
to ethernet device.
--- lib/librte_ethdev/rte_ethdev.c | 7 +++++++ lib/librte_ethdev/rte_ethdev.h | 16 ++++++++++++++++ lib/librte_ethdev/rte_ethdev_core.h | 1 + 3 files changed, 24 insertions(+)
lib/librte_ethdev/rte_ethdev.c +7 -0
@@ -752,6 +752,13 @@ rte_eth_dev_get_sec_ctx(uint16_t port_id)return rte_eth_devices[port_id].security_ctx; } +void * +rte_eth_dev_get_sft_ctx (uint16_t port_id) +{ +RTE_ETH_VALID_PORTID_OR_ERR_RET (port_id,NULL ); +return rte_eth_devices[port_id].sft_ctx; +} +uint16_t rte_eth_dev_count_avail (void ) {
lib/librte_ethdev/rte_ethdev.h +16 -0
@@ -1228,6 +1228,7 @@ struct rte_eth_conf {#define DEV_RX_OFFLOAD_SCTP_CKSUM 0x00020000#define DEV_RX_OFFLOAD_OUTER_UDP_CKSUM 0x00040000#define DEV_RX_OFFLOAD_RSS_HASH 0x00080000 +#define DEV_RX_OFFLOAD_SFT 0x00100000#define DEV_RX_OFFLOAD_CHECKSUM (DEV_RX_OFFLOAD_IPV4_CKSUM | \ DEV_RX_OFFLOAD_UDP_CKSUM | \
@@ -4388,6 +4389,21 @@ rte_eth_dev_pool_ops_supported(uint16_t port_id, const char *pool);void *rte_eth_dev_get_sec_ctx (uint16_t port_id); +/** +* @warning +* @b EXPERIMENTAL: this API may change, or be removed, without prior notice +* +* Get the SFT context for the Ethernet device. +* +* @param port_id +* Port identifier of the Ethernet device +* @return +* - NULL on error. +* - pointer to SFT context on success. +*/ +void * +rte_eth_dev_get_sft_ctx (uint16_t port_id); +/** * @warning * @b EXPERIMENTAL: this API may change, or be removed, without prior notice
lib/librte_ethdev/rte_ethdev_core.h +1 -0
@@ -806,6 +806,7 @@ struct rte_eth_dev {struct rte_eth_rxtx_callback *pre_tx_burst_cbs[RTE_MAX_QUEUES_PER_PORT ];enum rte_eth_dev_state state;/**< Flag indicating the port state */ void *security_ctx;/**< Context for security ops */ +void *sft_ctx;/**< Context for SFT ops */ uint64_t reserved_64s[4 ];/**< Reserved for future fields */ void *reserved_ptrs[4 ];/**< Reserved for future fields */
Signature · 1 line
2.26.2
RFC 3/3 sft: introduce API
In reply to Andrey VesnovatyDefines RTE SFT APIs for Statefull Flow Table library.
SFT General description:
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.
--- lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 +++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 6 files changed, 1105 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.map
lib/librte_sft/Makefile new file +28 -0
@@ -0,0 +1,28 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +include $(RTE_SDK)/mk/rte.vars.mk + +# library name +LIB = librte_sft.a + +# library version +LIBABIVER := 1 + +# build flags +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) +LDLIBS += -lrte_eal -lrte_mbuf + +# library source files +# all source are stored in SRCS-y +SRCS-$(CONFIG_RTE_LIBRTE_REGEXDEV) := rte_sft.c + +# export include files +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft.h +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft_driver.h + +# versioning export map +EXPORT_MAP := rte_sft_version.map + +include $(RTE_SDK)/mk/rte.lib.mk
lib/librte_sft/meson.build new file +7 -0
@@ -0,0 +1,7 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +sources = files('rte_sft.c') +headers = files('rte_sft.h', + 'rte_sft_driver.h') +deps += ['mbuf']
lib/librte_sft/rte_sft.c new file +9 -0
@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */
lib/librte_sft/rte_sft.h new file +845 -0
845 lines not on this page · raw message
lib/librte_sft/rte_sft_driver.h new file +195 -0
195 lines not on this page · raw message
lib/librte_sft/rte_sft_version.map new file +21 -0
@@ -0,0 +1,21 @@ +EXPERIMENTAL { + global: + + rte_sft_flow_get_status; + rte_sft_flow_set_ctx; + rte_sft_init; + rte_sft_fini; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + rte_sft_drain_mbuf; + rte_sft_flow_activate; + rte_sft_flow_create; + rte_sft_flow_lock; + rte_sft_flow_unlock; + rte_sft_flow_destroy; + rte_sft_flow_touch; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_get_client_obj; + rte_sft_flow_reset_client_obj; +};
Signature · 1 line
2.26.2
Re: [dpdk-dev] [RFC 3/3] sft: introduce API
Hi Andery,
PSB
Quoted 299 lines
-----Original Message-----
From: Andrey Vesnovaty <[email protected]>
Sent: Wednesday, September 9, 2020 11:30 PM
To: [email protected]
Subject: [RFC 3/3] sft: introduce APIDefines RTE SFT APIs for Statefull Flow Table library.
SFT General description:
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.Signed-off-by Andrey Vesnovaty <[email protected]>--- lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 +++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 6 files changed, 1105 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.maplib/librte_sft/Makefile new file +28
-0@@ -0,0 +1,28 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +include $(RTE_SDK)/mk/rte.vars.mk + +# library name +LIB = librte_sft.a + +# library version +LIBABIVER := 1 + +# build flags +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) +LDLIBS += -lrte_eal -lrte_mbuf + +# library source files +# all source are stored in SRCS-y +SRCS-$(CONFIG_RTE_LIBRTE_REGEXDEV) := rte_sft.c + +# export include files +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft.h +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft_driver.h + +# versioning export map +EXPORT_MAP := rte_sft_version.map + +include $(RTE_SDK)/mk/rte.lib.mklib/librte_sft/meson.build new file +7
-0@@ -0,0 +1,7 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +sources = files('rte_sft.c') +headers = files('rte_sft.h', + 'rte_sft_driver.h') +deps += ['mbuf']lib/librte_sft/rte_sft.c new file +9
-0@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */ lib/librte_sft/rte_sft.h new file +119
-0@@ -0,0 +1,845 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_H_ +#define _RTE_SFT_H_ + +/** +* @file +* +* RTE SFT API +* +* Defines RTE SFT APIs for Statefull Flow Table library. +* +* SFT General description: +* SFT library provides a framework for applications that need to maintain +* context across different packets of the connection. +* Examples for such applications: +* - Next-generation firewalls +* - Intrusion detection/prevention systems (IDS/IPS): Suricata, Snort +* - SW/Virtual Switching: OVS +* The goals of the SFT library: +* - Accelerate flow recognition & its context retrieval for further lookaside +* processing. +* - Enable context-aware flow handling offload. +* +* Definitions and Abbreviations: +* - 5-tuple: defined by: +* -- Source IP address +* -- Source port +* -- Destination IP address +* -- Destination port +* -- IP protocol number +* - 7-tuple: 5-tuple zone and port (see struct rte_sft_7tuple) +* - 5/7-tuple: 5/7-tuple of the packet from connection initiator +* - revers 5/7-tuple: 5/7-tuple of the packet from connection initiate +* - application: SFT library API consumer +* - APP: see application +* - CID: client ID +* - CT: connection tracking +* - FID: Flow identifier +* - FIF: First In Flow +* - Flow: defined by 7-tuple and its reverse i.e. flow is bidirectional +* - SFT: Stateful Flow Table +* - user: see application +* - zone: additional user defined value used as differentiator for +* connections having same 5-tuple (for example different VxLan +* connections with same inner 5-tuple). +* +* SFT components: +* +* +-----------------------------------+ +* | RTE flow | +* | | +* | +-------------------------------+ | +----------------+ +* | | group X | | | RTE_SFT | +* | | | | | | +* | | +---------------------------+ | | | | +* | | | rule ... | | | | | +* | | | . | | | +-----------+----+ +* | | | . | | | | +* | | | . | | | entry +* | | +---------------------------+ | | create +* | | | rule | | | | +* | | | patterns ... +---------+ | +* | | | actions | | | | | +* | | | SFT (zone=Z) | | | | | +* | | | JUMP (group=Y) | | | lookup | +* | | +---------------------------+ | | zone=Z, | +* | | | rule ... | | | 5tuple | +* | | | . | | | | | +* | | | . | | | +--v-------------+ +* | | | . | | | | SFT | | +* | | | | | | | | | +* | | +---------------------------+ | | | +--v--+ | +* | | | | | | | | +* | +-------------------------------+ | | | PMD | | +* | | | | | | +* | | | +-----+ | +* | +-------------------------------+ | | | +* | | group Y | | | | +* | | | | | set flow CTX | +* | | +---------------------------+ | | | | +* | | | rule | | | +--------+-------+ +* | | | patterns | | | | +* | | | SFT (state=UNDEFINED) | | | | +* | | | actions RSS | | | | +* | | +---------------------------+ | | | +* | | | rule | | | | +* | | | patterns | | | | +* | | | SFT (state=INVALID) | <-------------+ +* | | | actions DROP | | | forward +* | | +---------------------------+ | | group=Y +* | | | rule | | | +* | | | patterns | | | +* | | | SFT (state=ACCEPTED) | | | +* | | | actions PORT | | | +* | | +---------------------------+ | | +* | | ... | | +* | | | | +* | +-------------------------------+ | +* | ... | +* | | +* +-----------------------------------+ +* +* SFT as datastructure: +* SFT can be treated as datastructure maintaining flow context across its +* lifetime. SFT flow entry represent bidirectional network flow and defined by +* 7-tuple & its reverse 7-tuple. +* Each entry in SFT has: +* - FID: 1:1 mapped & used as entry handle & encapsulating internal +* implementation of the entry. +* - State: user-defined value attached to each entry, the only library +* reserved value for state unset (the actual value defined by SFT +* configuration). The application should define flow state encodings and +* set it for flow via rte_sft_flow_set_ctx() than what actions should be +* applied on packets can be defined via related RTE flow rule matching SFT +* state (see rules in SFT components diagram above). +* - Timestamp: for the last seen in flow packet used for flow aging mechanism
+ * implementation.
+ * - Client Objects: user-defined flow contexts attached as opaques to flow.
+ * - Acceleration & offloading - utilize RTE flow capabilities, when supported
+ * (see action ``SFT``), for flow lookup acceleration and further
+ * context-aware flow handling offload.
+ * - CT state: optionally for TCP connections CT state can be maintained
+ * (see enumrte_sft_flow_ct_state).
+ * - Out of order TCP packets: optionally SFT can keep out of order TCP
+ * packets aside the flow context till the arrival of the missing in-order
+ * packet.
+ *
+ * RTE flow changes:
+ * The SFT flow state (or context) for RTE flow is defined by fields of
+ * structrte_flow_item_sft.
+ * To utilize SFT capabilities new item and action types introduced:
+ * - item SFT: matching on SFT flow state (seeRTE_FLOW_ITEM_TYPE_SFT).
+ * - action SFT: retrieve SFT flow context and attache it to the processed
+ * packet (seeRTE_FLOW_ACTION_TYPE_SFT).
+ *
+ * The contents of per port SFT serving RTE flow action ``SFT`` managed via
+ * SFT PMD APIs (see structrte_sft_ops).
+ * The SFT flow state/context retrieval performed by user-defined zone ``SFT``
+ * action argument and processed packet 5-tuple.
+ * If in scope of action ``SFT`` there is no context/state for the flow in SFT
+ * undefined sate attached to the packet meaning that the flow is not
+ * recognized by SFT, most probably FIF packet.
+ *
+ * Once the SFT state set for a packet it can match on item SFT
+ * (seeRTE_FLOW_ITEM_TYPE_SFT) and forwarding design can be done for the
+ * packet, for example:
+ * - if state value == x than queue for further processing by the application
+ * - if state value == y than forward it to eth port (full offload)
+ * - if state value == 'undefined' than queue for further processing by
+ * the application (handle FIF packets)
+ *
+ * Processing packets with SFT library:
+ *
+ * FIF packet:
+ * To recognize upcoming packets of the SFT flow every FIF packet should be
+ * forwarded to the application utilizing the SFT library. Non-FIF packets can
+ * be processed by the application or its processing can be fully offloaded.
+ * Processing of the packets in SFT library starts withrte_sft_process_mbuf
+ * orrte_sft_process_mbuf_with_zone. If mbuf recognized as FIF application
+ * should make a design to destroy flow or complete flow creation process in
+ * SFT usingrte_sft_flow_activate.
+ *
+ * Recognized SFT flow:
+ * Once structrte_sft_flow_statuswith valid fid field posesed by application
+ * it can:
+ * - mange client objects on it (seeclient_objfield in
+ * structrte_sft_flow_status) using rte_sft_flow_<OP>_client_obj APIs
+ * - analyze user-defined flow state and CT state (see state &ct_satefields
+ * in structrte_sft_flow_status).
+ * - set flow state to be attached to the upcoming packets by action ``SFT``
+ * via structrte_sft_flow_statusAPI.
+ * - decide to destroy flow viarte_sft_flow_destroyAPI.
+ *
+ * Flow aging:
+ *
+ * SFT library manages the aging for each flow. On flow creation, it's
+ * assigned an aging value, the maximal number of seconds passed since the
+ * last flow packet arrived, once exceeded flow considered aged.
+ * The application notified of aged flow asynchronously via event queues.
+ * The device and port IDs tuple to identify the event queue to enqueue
+ * flow aged events passed on flow creation as arguments
+ * (seerte_sft_flow_activate). It's the application responsibility to
+ * initialize event queues and assign them to each flow for EOF event
+ * notifications.
+ * Aged EOF event handling:
+ * - Should be considered as application responsibility.
+ * - The last stage should be the release of the flow resources via
+ *rte_sft_flow_destroyAPI.
+ * - All client objects should be removed from flow before the
+ *rte_sft_flow_destroyAPI call.
Quoted 8 lines
+ * See the description of
rete_sft_flow_destroyfor an example of aged flow
+ * handling.
+ *
+ * SFT API thread safety:
+ *
+ * SFT library APIs are thread-safe while handling of specific flow can be
+ * done in a single thread simultaneously. Exclusive access to specific SFT
+ * flow guaranteed by:
The line above contradict itself, if you are working with single thread you can't work simultaneously. Does the SFT allow the access to a single flow from two threads in the same time? or it is the responsibility Of the application to protect itself. I think it should be the application responsibility the SFT should protect itself only on SFT global functions. For example calling process_mbuf should be protected, so application can call the same function from different threads.
I think we can assume that all packets from a specific flow will arrive to the same queue and the same thread.
So I don't see the usage of the lock API.
Quoted 180 lines
+ * -
rte_sft_process_mbuf
+ * -rte_sft_process_mbuf_with_zone
+ * -rte_sft_flow_create
+ * -rte_sft_flow_lock
+ * When application is done with the flow handling for the current packet it
+ * should callrte_sft_flow_unlockAPI to maintain exclusive access to the
+ * flow with other threads.
+ *
+ * SFT Library initialization and cleanup:
+ *
+ * SFT library should be considered as a single instance, preconfigured and
+ * initialized viarte_sft_init()API.
+ * SFT library resource deallocation and cleanup should be done via
+ *rte_sft_init()API as a stage of the application termination procedure.
+ */lib/librte_sft/rte_sft.h new file +165
-0+ +#ifdef __cplusplus +extern "C" { +#endif + +#include <rte_common.h> +#include <rte_config.h> +#include <rte_errno.h> +#include <rte_mbuf.h> +#include <rte_ethdev.h> +#include <rte_flow.h> + +/** +* L3/L4 5-tuple - src/dest IP and port and IP protocol. +* +* Used for flow/connection identification. +*/ +struct rte_sft_5tuple { +union { +struct { +rte_be32_t src_addr;/**< IPv4 source address. */ +rte_be32_t dst_addr;/**< IPv4 destination address. */ + } ipv4; +struct { +uint8_t src_addr[16 ];/**< IPv6 source address. */ +uint8_t dst_addr[16 ];/**< IPv6 destination address. */ + } ipv6; + }; +uint16_t src_port;/**< Source port. */ +uint16_t dst_port;/**< Destination port. */ +uint8_t proto;/**< IP protocol. */ +uint8_t is_ipv6:1 ;/**< True for valid IPv6 fields. Otherwise IPv4. */ +}; + +/** +* Port flow identification. +* +* @p zone used for setups where 5-tuple is not enough to identify flow. +* For example different VLANs/VXLANs may have similar 5-tuples. +*/ +struct rte_sft_7tuple { +struct rte_sft_5tuple flow_5tuple;/**< L3/L4 5-tuple. */ +uint32_t zone;/**< Zone assigned to flow. */ +uint16_t port_id;/** <Port identifier of Ethernet device. */ +}; + +/** +* Flow connection tracking states +*/ +enum rte_sft_flow_ct_state { +RTE_SFT_FLOW_CT_STATE_NEW = (1 <<0 ), +RTE_SFT_FLOW_CT_STATE_EST = (1 <<1 ), +RTE_SFT_FLOW_CT_STATE_REL = (1 <<2 ), +RTE_SFT_FLOW_CT_STATE_RPL = (1 <<3 ), +RTE_SFT_FLOW_CT_STATE_INV = (1 <<4 ), +RTE_SFT_FLOW_CT_STATE_TRK = (1 <<5 ), +RTE_SFT_FLOW_CT_STATE_SNAT = (1 <<6 ), +RTE_SFT_FLOW_CT_STATE_DNAT = (1 <<7 ), +}; + +/** +* Structure describes SFT library configuration +*/ +struct rte_sft_conf { +uint32_t UDP_aging;/**< UDP proto default aging. */ +uint32_t TCP_aging;/**< TCP proto default aging. */ +uint32_t TCP_SYN_aging;/**< TCP SYN default aging. */ +uint32_t OTHER_aging;/**< All unlisted proto default aging. */ +uint32_t size;/**< Max entries in SFT. */ +uint8_t undefined_state;/**< Undefined state constant. */ +uint8_t reorder_enable:1 ; +/**< TCP packet reordering feature enabled bit. */ +uint8_t ct_enable:1 ;/**< Connection tracking feature enabled bit. */ +}; + +/** +* Structure describes the state of the flow in SFT. +*/ +struct rte_sft_flow_status { +uint32_t fid;/**< SFT flow id. */ +uint32_t zone;/**< Zone for lookup in SFT */ +uint8_t state;/**< Application defined bidirectional flow state. */ +uint8_t ct_state;/**< Connection tracking flow state. */ +uint32_t age;/**< Seconds passed since last flown packet. */ +uint32_t aging; +/**< Flow considered aged once this age (seconds) reached. */ +uint32_t nb_in_order_mbufs; +/**< Number of in-order mbufs available for drain */ +void **client_obj;/**< Array of clients attached to flow. */ +int nb_clients;/**< Number of clients attached to flow. */ +uint8_t defined:1 ;/**< Flow defined in SFT bit. */ +uint8_t activated:1 ;/**< Flow activation bit. */ +uint8_t fragmented:1 ;/**< Last flow mbuf was fragmented. */ +uint8_t out_of_order:1 ;/**< Last flow mbuf was out of order (TCP). */ +}; + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get SFT flow status. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump actual SFT flow status. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_get_status (const uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set user defined context. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Updates per ethernet dev SFT entries: +* - flow lookup acceleration +* - partial/full flow offloading managed by flow context +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param ctx +* User defined state to set. +* Update of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_set_ctx (uint32_t fid, +const struct rte_flow_item_sft *ctx, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Initialize SFT library instance. +* +* @param conf +* SFT library instance configuration. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_init (const struct rte_sft_conf *conf); +
lib/librte_sft/rte_sft.h line 391
Quoted 10 lines
lib/librte_sft/rte_sft.h new file +10
-0+/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Finalize SFT library instance. +* Cleanup & release allocated resources. +*/ +void +rte_sft_fini (void ); +
I think we should use stop. It is not commons in DPDK to have fini functions. Maybe we should also add start function, so the app can init and then start the SFT.
Quoted 4 lines
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice.
+ *
Quoted 8 lines
+ * Process mbuf received on RX queue.
+ *
+ * Fragmentation handling (SFT fragmentation feature configured):
+ * If *mbuf_in* of fragmented packet received it will be stored by SFT library.
+ *status->fragmentedbit will be set and *mbuf_out* will be set to NULL.
+ * On reception of all related fragments of IP packet it will be reassembled
+ * and further processed by this function on reception of last fragment.
+ *
Does this function allocate a new mbuf? Does it releases all old mbufs?
Quoted 25 lines
+ * Flow definition:
+ * SFT flow defined by one of its 7-tuples, since there is no zone value as
+ * argument flow should be defined by context attached to mbuf with action
+ * ``SFT`` (see RTE flowRTE_FLOW_ACTION_TYPE_SFT). Otherwise status-Quoted 1 line
defined
+ * field will be turned off & *mbuf_out* will be set to *mbuf_in*.
+ * In order to define flow for *mbuf_in* without attached sft context
+ *rte_sft_process_mbuf_with_zone()should be used with *zone* argument
+ * supplied by caller.
+ *
+ * Flow lookup:
+ * If SFT flow identifier can't be retrieved from SFT context attached to
+ * *mbuf_in* by action ``SFT`` - SFT lookup should be performmed by zone,
+ * retrieved from SFT context attached to *mbuf_in*, and 5-tuple, extracted
+ * form mbuf outer header contents.
+ *
+ * Flow defined but does not exists:
+ * If flow not found in SFT inactivated flow will be created in SFT.
+ *status->activatedfield will be turned off & *mbuf_out* be set to *mbuf_in*.
+ * In order to activate created flowrte_sft_flow_activate()should be used
+ * with reverse 7-tuple supplied by caller.
+ * This is first phase of flow creation in SFT for second phase & more detailed
+ * descriotion of flow creation seerte_sft_flow_activate.
+ *
Quoted 8 lines
+ * Out of order (SFT out of oreder feature configured):
+ * If flow defined & activated but *mbuf_in* is TCP out of order packet it will
+ * be stored by SFT library.status->out_of_orderbit will be set & *mbuf_out*
+ * will be set to NULL. On reception of the first missing in order packet
+ *status->nb_in_order_mbufswill be set to number of mbufs that available for
+ * processing withrte_sft_drain_mbuf().
+ *
It is possible that some packets will get trapped in the SFT do to this feature. if it supports ordering. For example the following case:
Packets arrive to the application. After draining the packets the Application changed the flow to full offload. This means that
all future packets will not arrive to the application.
But until the flow is offloaded some packets do arrive not in order. Then the flow is offloaded, this will result in the situation that no more packets will arrive to the application so some packets will get stack in the SFT.
I think we must have some force drain or, notify the SFT that no more packets should arrive to even if the packets are not in order it will release them.
Also the same with fragmented does this function allocate new mbufs? are you releasing the old ones?
Quoted 6 lines
+ * Flow defined & activated, mbuf not fragmented and 'in order':
+ * - Flow aging related data (see age field instruct rte_sft_flow_status)
+ * will be updated according to *mbuf_in* timestamp.
+ * - Flow connection tracking state (seect_statefield in
+ *struct rte_sft_flow_status) will be updated according to *mbuf_in* L4
+ * header contents.
Quoted 8 lines
+ * - *mbuf_out* will be set to last processed mbuf.
+ *
+ * @param[in]mbuf_in
+ * mbuf to process; mbuf pinter considered 'consumed' and should not be
used
+ * after successful call to this function.
+ * @param[out]mbuf_out
+ * last processed not fragmented and in order mbuf.
If the in mbuf is not fragmented and in order, this pointer will point to the in one?
Quoted 3 lines
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contents of
+ * *mbuf_in*.
Does the status bits for example fragmented is kept per connection or per flow? Since it is possible to get fragmented packets from both sides. The same goes for out of order packets
Quoted 85 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success:
+ * - *mbuf_out* contains valid mbuf pointer, locked SFT flow recognized by
+ *status->fid.
+ * - *mbuf_out* is NULL andstatus->fragmentedbit on in case of
+ * non last fragment *mbuf_in*.
+ * - *mbuf_out* is NULL andstatus->out_of_orderbit on in case of out of
+ * order *mbuf_in*, locked SFT flow recognized bystatus->fid.
+ * On failure a negative errno value andrte_errnois set.
+ */lib/librte_sft/rte_sft.h new file +20
-0+int +rte_sft_process_mbuf (struct rte_mbuf *mbuf_in, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue while zone value provided by caller. +* +* The behaviour of this function is similar to rte_sft_process_mbuf except +* the lookup in SFT procedure. The lookup in SFT always done by the *zone* +* arg and 5-tuple 5-tuple, extracted form mbuf outer header contents. +* +* @see rte_sft_process_mbuf +* +* @param[in] mbuf_in +* mbuf to process; mbuf pinter considered 'consumed' and should not be used
+ * after successful call to this function.
+ * @param[out]mbuf_out
+ * last processed not fragmented and in order mbuf.
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contents of
+ * *mbuf_in*.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return+ * 0 on success: + * - *mbuf_out* contains valid mbuf pointer. + * - *mbuf_out* is NULL and status->fragmented bit on in case of+ * non last fragment *mbuf_in*.
+ * - *mbuf_out* is NULL andstatus->out_of_orderbit on in case of out of
+ * order *mbuf_in*.
+ * On failure a negative errno value andrte_errnois set.
+ */lib/librte_sft/rte_sft.h new file +31
-0+int +rte_sft_process_mbuf_with_zone (struct rte_mbuf *mbuf_in, +uint32_t zone, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Drain next in order mbuf. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* This function behaves similar to rte_sft_process_mbuf() but acts on packets +* accumulated in SFT flow due to missing in order packet. Processing done on +* single mbuf at a time and `in order`. Other than above the behavior is +* same as of rte_sft_process_mbuf for flow defined & activated & mbuf isn't +* fragmented & 'in order'. This function should be called when +* rte_sft_process_mbuf or rte_sft_process_mbuf_with_zone sets +* status->nb_in_order_mbufs output param !=0 and until +* status->nb_in_order_mbufs == 0. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump SFT flow status once updated according to contents of +* *mbuf_in*. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this
lib/librte_sft/rte_sft.h line 543
Quoted 10 lines
lib/librte_sft/rte_sft.h new file +10
-0+* structure in case of error only. +* +* @return +* A valid mbuf in case of success, NULL otherwise and rte_errno is set. +*/ +struct rte_mbuf * +rte_sft_drain_mbuf (uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); +
Fid represent a connection, so which direction do we drain the packets? since we can have inordered packet in from both directions right?
Quoted 371 lines
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice.
+ *
+ * Activate flow in SFT.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * This function performs second phase of flow creation in SFT.
+ * The reasons for 2 phase flow creation procedure:
+ * 1. Missing reverse flow - flow context is shared for both flow directions+ * i.e. in order maintain bidirectional flow context in RTE SFT packets + * arriving from both dirrections should be identified as packets of the + * RTE SFT flow. Consequently before creation of the SFT flow caller should + * provide reverse flow direction 7-tuple. + * 2. The caller of rte_sft_process_mbuf/rte_sft_process_mbuf_with_zone should + * be notified that arrived mbuf is first in flow & decide weather to + * create new flow or it distroy before it was activated with + * rte_sft_flow_destroy.+ * This function completes creation of the bidirectional SFT flow & creates
+ * entry for 7-tuple on SFT PMD defined by the tuple port for both
+ * initiator/initiate 7-tuples.
+ * Flow aging, connection tracking state & out of order handling will be
+ * initialized according to the content of the *mbuf_in* passes to
+ * rte_sft_process_mbuf/_with_zone during the phase 1 of flow creation.
+ * Once this function returns upcoming calls rte_sft_process_mbuf/_with_zone
+ * with 7-tuple or its reverse will return handle to this flow.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * @param fid
+ * SFT flow ID.
+ * @paramreverse_tuple
+ * Expected response flow 7-tuple.
+ * @param ctx
+ * User defined state to set.
+ * Update of *fid* or *zone* fields in structrte_flow_item_sftunsupported.
+ * @paramct_enable
+ * Enables maintenance ofstatus->ct_stateconnection tracking value for the
+ * flow; otherwisestatus->ct_statewill be initialized with zeros.
+ * @paramevdev_id
+ * Event dev ID to enqueue end of flow event.
+ * @paramevport_id
+ * Event port ID to enqueue end of flow event.
+ * @param[out] status
+ * Structure to dump SFT flow status once activated.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */lib/librte_sft/rte_sft.h new file +226
-0+int +rte_sft_flow_activate (uint32_t fid, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +uint8_t dev_id, +uint8_t port_id, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Artificially create SFT flow. +* +* Function to create SFT flow before reception of the first flow packet. +* +* @param tuple +* Expected initiator flow 7-tuple. +* @param reverse_tuple +* Expected initiate flow 7-tuple. +* @param ctx +* User defined state to set. +* Setting of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param[out] ct_enable +* Enables maintenance of status->ct_state connection tracking value for the +* flow; otherwise status->ct_state will be initialized with zeros. +* @param[out] status +* Structure to dump SFT flow status once created. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* - on success: 0, locked SFT flow recognized by status->fid. +* - on error: a negative errno value otherwise and rte_errno is set. +*/ + +int +rte_sft_flow_create (const struct rte_sft_7tuple *tuple, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Lock exclusively SFT flow. +* +* Explicit flow locking; used for handling aged flows. +* +* @param fid +* SFT flow ID. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_lock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Release exclusively locked SFT flow. +* +* When rte_sft_process_mbuf/_with_zone and rte_sft_flow_create +* return *status* containing fid with defined bit on the flow considered +* exclusively locked and should be unlocked with this function. +* +* @param fid +* SFT flow ID. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_unlock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Removes flow from SFT. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* - Flow should be locked by caller in order to remove it. +* - Flow should have no client objects attached. +* +* Should be applied on aged flows, when flow aged event received. +* +* @code{.c} +* while (1) { +* rte_event_dequeue_burst(...); +* FOR_EACH_EV(ev) { +* uint32_t fid = ev.u64; +* rte_sft_flow_lock(fid); +* FOR_EACH_CLIENT(fid, client_id) { +* rte_sft_flow_reset_client_obj(fid, client_obj); +* // detached client object handling +* } +* rte_sft_flow_destroy(fid, &error); +* } +* } +* @endcode +* +* @param fid +* SFT flow ID to destroy. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_destroy (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Reset flow age to zero. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Simulates last flow packet with timestamp set to just now. +* +* @param fid +* SFT flow ID. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_touch (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set flow aging to specific value. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param aging +* New flow aging value. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_set_aging (uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to set object for. +* @param client_obj +* Pointer to opaque client object structure. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_set_client_obj (uint32_t fid, +uint8_t client_id, +void *client_obj, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to get object for. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid client object opaque pointer in case of success, NULL otherwise +* and rte_errno is set. +*/ +void * +rte_sft_flow_get_client_obj (const uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Remove client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Detaches client object from SFT flow and returns the ownership for the +* client object to the caller by returning client object pointer value. +* The pointer returned by this function won't be accessed any more, the caller
+ * may release all client obj related resources & the memory allocated for
+ * this client object.
+ *
+ * @param fid
+ * SFT flow ID.
+ * @paramclient_id
+ * Client ID to remove object for.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid client object opaque pointer in case of success, NULL otherwise
+ * andrte_errnois set.
+ */lib/librte_sft/rte_sft.h new file +10
-0+void * +rte_sft_flow_reset_client_obj (uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_H_ */ lib/librte_sft/rte_sft_driver.h new file +60
-0@@ -0,0 +1,195 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_DRIVER_H_ +#define _RTE_SFT_DRIVER_H_ + +/** + * @file + * + * RTE SFT Ethernet device PMD API + * + * APIs that are used by the SFT library to offload SFT operationons + * to Ethernet device. + */ + +#include "rte_sft.h" + +#ifdef __cplusplus +extern "C" { +#endif + +/** + * Opaque type returned after successfully creating an entry in SFT. + * + * This handle can be used to manage and query the related entry (e.g. to + * destroy it or update age). + */ +struct rte_sft_entry; + +/** + * Create SFT entry in eth_dev SFT. + * + * @param dev + * Pointer to Ethernet device structure. + * @param tuple + * L3/L4 5-tuple - src/dest IP and port and IP protocol. + * @param nat_tuple + * L3/L4 5-tuple to replace in packet original 5-tuple in order to implement + * NAT offloading; if NULL NAT offloading won't be configured for the flow. + * @param aging + * Flow aging timeout in seconds. + * @param ctx + * Initial values in SFT flow context + * (see RTE flow struct rte_flow_item_sft). + * ctx->zone should be valid. + * @param fid + * SFT flow ID for the entry to create on *device*. + * If there is an entry for the *fid* in PMD it will be updated with the + * values of *ctx*. + * @param[out] queue_index + * if PMD can figure out the queue where the flow packets will + * arrive in RX data path it will set the value of queue_index; otherwise + * all bits will be turned on. + * @param[out] error + * Perform verbose error reporting if not NULL. PMDs initialize this + * structure in case of error only. + * + * @return + * A valid handle in case of success, NULL otherwise and rte_errno is set.
lib/librte_sft/rte_sft_driver.h line 70
Quoted 10 lines
lib/librte_sft/rte_sft_driver.h new file +10
-0+*/ +typedef struct rte_sft_entry *(*sft_entry_create_t ) (struct rte_eth_dev *dev, +const struct rte_sft_5tuple *tuple, +const struct rte_sft_5tuple *nat_tuple, +const uint32_t aging, +const struct rte_flow_item_sft *ctx, +const uint32_t fid, +uint16_t *queue_index, +struct rte_sft_error *error); +
I think for easier reading, the API should change to have 6 tuple (5 + zone) the ctx should be removed and replaced with the state.
Then add new API to modify the ctx
typedef int (*sft_modify_state)(struct rte_eth_dev *dev, uint8 state);
The main issue we my suggestion is that it will force the PMD to store the information to recreate the rule, data that is already saved by the SFT.
Also I don't see why we need queue index, since the RSS and queue will be configured by the RTE flow in a different group.
Quoted 29 lines
lib/librte_sft/rte_sft_driver.h new file +29
-0+/** +* Destroy SFT entry in eth_dev SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param entry +* Handle to the SFT entry to destroy. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_destroy_t )(struct rte_eth_dev *dev, +struct rte_sft_entry *entry, +struct rte_sft_error *error); + +/** +* Decodes SFT flow context if attached to mbuf by action ``SFT``. +* @see RTE flow RTE_FLOW_ACTION_TYPE_SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param mbuf +* mbuf of the packet to decode attached state from. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only.
lib/librte_sft/rte_sft_driver.h line 109
Quoted 10 lines
lib/librte_sft/rte_sft_driver.h new file +10
-0+* +* @return +* A valid SFT flow context in case of success, NULL otherwise and rte_errno +* is set. +*/ +typedef struct rte_flow_item_sft *(*sft_entry_mbuf_decode_ctx_t )( +struct rte_eth_dev *dev, +const struct rte_mbuf *mbuf, +struct rte_sft_error *error); +
What about returning int as error code, and return the rte_flow_item_sft as out parameter?
This will remove the allocation and free.
Quoted 115 lines
lib/librte_sft/rte_sft_driver.h new file +86
-0+/** +* Get aged-out SFT entries. +* +* Report entry as aged-out if timeout passed without any matching +* on the SFT entry. +* +* @param[in] dev +* Pointer to Ethernet device structure. +* @param[in, out] fid_aged +* The address of an array of aged-out SFT flow IDs. +* @param[in] nb_aged +* The length of *fid_aged* array pointers. +* @param[out] error +* Perform verbose error reporting if not NULL. Initialized in case of +* error only. +* +* @return +* if nb_aged is 0, return the amount of all aged flows. +* if nb_aged is not 0 , return the amount of aged flows reported +* in the *fid_aged* array, otherwise negative errno value. +*/ +typedef int (*sft_entry_get_aged_entries_t )(struct rte_eth_dev *dev, +uint32_t *fid_aged, +int nb_aged, +struct rte_sft_error *error); + +/** +* Simulate SFT entry match in terms of entry aging. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_touch_t )(struct rte_eth_dev *dev, +uint32_t fid, +struct rte_sft_error *error); + +/** +* Set SFT entry aging to specific value. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param aging +* New entry aging value. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_set_aging_t )(struct rte_eth_dev *dev, +uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** SFT operations function pointer table */ +struct rte_sft_ops { +sft_entry_create_t entry_create; +/**< Create SFT entry in eth_dev SFT. */ +sft_entry_destroy_t entry_destroy; +/**< Destroy SFT entry in eth_dev SFT. */ +sft_entry_mbuf_decode_ctx_t mbuf_decode_ctx; +/**< Decodes SFT flow context if attached to mbuf by action ``SFT``. */ +sft_entry_get_aged_entries_t get_aged_entries; +/**< Get aged-out SFT entries. */ +sft_entry_touch_t entry_touch; +/**< Simulate SFT entry match in terms of entry aging. */ +sft_entry_set_aging_t set_aging; +/**< Set SFT entry aging to specific value. */ +}; + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_DRIVER_H_ */ lib/librte_sft/rte_sft_version.map new file +21
-1@@ -0,0 +1,21 @@ +EXPERIMENTAL { + global: + + rte_sft_flow_get_status; + rte_sft_flow_set_ctx; + rte_sft_init; + rte_sft_fini; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + rte_sft_drain_mbuf; + rte_sft_flow_activate; + rte_sft_flow_create; + rte_sft_flow_lock; + rte_sft_flow_unlock; + rte_sft_flow_destroy; + rte_sft_flow_touch; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_get_client_obj; + rte_sft_flow_reset_client_obj; +}; --
Quoted 1 line
2.26.2
Best,
Ori
Re: [dpdk-dev] [RFC 3/3] sft: introduce API
Hi Andrey,
looks very interesting, but a bit hard to review. I hope I'll do deeper review on the next version. Right not just few cosmetic things to make the next version a bit clearer.
Do you plan to create/publish an example appliction which uses the API and demonstrates how to do it?
Plesee, see below.
Thanks,
Andrew.
Quoted 122 lines · Ori Kam
Hi Andery,
PSBQuoted 119 lines
-----Original Message-----
From: Andrey Vesnovaty <[email protected]>
Sent: Wednesday, September 9, 2020 11:30 PM
To: [email protected]
Subject: [RFC 3/3] sft: introduce APIDefines RTE SFT APIs for Statefull Flow Table library.
SFT General description:
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.Signed-off-by Andrey Vesnovaty <[email protected]>--- lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 +++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 6 files changed, 1105 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.maplib/librte_sft/Makefile new file +28
-0@@ -0,0 +1,28 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +include $(RTE_SDK)/mk/rte.vars.mk + +# library name +LIB = librte_sft.a + +# library version +LIBABIVER := 1 + +# build flags +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) +LDLIBS += -lrte_eal -lrte_mbuf + +# library source files +# all source are stored in SRCS-y +SRCS-$(CONFIG_RTE_LIBRTE_REGEXDEV) := rte_sft.c + +# export include files +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft.h +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft_driver.h + +# versioning export map +EXPORT_MAP := rte_sft_version.map + +include $(RTE_SDK)/mk/rte.lib.mklib/librte_sft/meson.build new file +7
-0@@ -0,0 +1,7 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +sources = files('rte_sft.c') +headers = files('rte_sft.h', + 'rte_sft_driver.h') +deps += ['mbuf']lib/librte_sft/rte_sft.c new file +9
-0@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */ lib/librte_sft/rte_sft.h new file +15
-0@@ -0,0 +1,845 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_H_ +#define _RTE_SFT_H_ + +/** +* @file +* +* RTE SFT API +* +* Defines RTE SFT APIs for Statefull Flow Table library. +* +* SFT General description:
Quoted 8 lines
lib/librte_sft/rte_sft.h line 23
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* SFT library provides a framework for applications that need to maintain +* context across different packets of the connection. +* Examples for such applications: +* - Next-generation firewalls +* - Intrusion detection/prevention systems (IDS/IPS): Suricata, Snort +* - SW/Virtual Switching: OVS +* The goals of the SFT library: +* - Accelerate flow recognition & its context retrieval for further lookaside
lookaside -> look-aside
Quoted 11 lines
lib/librte_sft/rte_sft.h line 34
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+* processing. +* - Enable context-aware flow handling offload. +* +* Definitions and Abbreviations: +* - 5-tuple: defined by: +* -- Source IP address +* -- Source port +* -- Destination IP address +* -- Destination port +* -- IP protocol number +* - 7-tuple: 5-tuple zone and port (see struct rte_sft_7tuple)
I guess comma is missing after "5-tuple", since I read it as "5-tuple zone"???
Quoted 5 lines
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+* - 5/7-tuple: 5/7-tuple of the packet from connection initiator +* - revers 5/7-tuple: 5/7-tuple of the packet from connection initiate +* - application: SFT library API consumer +* - APP: see application +* - CID: client ID
Quoted 8 lines
lib/librte_sft/rte_sft.h line 47
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* - CT: connection tracking +* - FID: Flow identifier +* - FIF: First In Flow +* - Flow: defined by 7-tuple and its reverse i.e. flow is bidirectional +* - SFT: Stateful Flow Table +* - user: see application +* - zone: additional user defined value used as differentiator for +* connections having same 5-tuple (for example different VxLan
VxLan -> VXLAN (see devtools/words-case.txt)
Quoted 53 lines
Quoted 53 lines
lib/librte_sft/rte_sft.h new file +53
-0+* connections with same inner 5-tuple). +* +* SFT components: +* +* +-----------------------------------+ +* | RTE flow | +* | | +* | +-------------------------------+ | +----------------+ +* | | group X | | | RTE_SFT | +* | | | | | | +* | | +---------------------------+ | | | | +* | | | rule ... | | | | | +* | | | . | | | +-----------+----+ +* | | | . | | | | +* | | | . | | | entry +* | | +---------------------------+ | | create +* | | | rule | | | | +* | | | patterns ... +---------+ | +* | | | actions | | | | | +* | | | SFT (zone=Z) | | | | | +* | | | JUMP (group=Y) | | | lookup | +* | | +---------------------------+ | | zone=Z, | +* | | | rule ... | | | 5tuple | +* | | | . | | | | | +* | | | . | | | +--v-------------+ +* | | | . | | | | SFT | | +* | | | | | | | | | +* | | +---------------------------+ | | | +--v--+ | +* | | | | | | | | +* | +-------------------------------+ | | | PMD | | +* | | | | | | +* | | | +-----+ | +* | +-------------------------------+ | | | +* | | group Y | | | | +* | | | | | set flow CTX | +* | | +---------------------------+ | | | | +* | | | rule | | | +--------+-------+ +* | | | patterns | | | | +* | | | SFT (state=UNDEFINED) | | | | +* | | | actions RSS | | | | +* | | +---------------------------+ | | | +* | | | rule | | | | +* | | | patterns | | | | +* | | | SFT (state=INVALID) | <-------------+ +* | | | actions DROP | | | forward +* | | +---------------------------+ | | group=Y +* | | | rule | | | +* | | | patterns | | | +* | | | SFT (state=ACCEPTED) | | | +* | | | actions PORT | | | +* | | +---------------------------+ | | +* | | ... | | +* | | | |
Quoted 8 lines
lib/librte_sft/rte_sft.h line 108
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* | +-------------------------------+ | +* | ... | +* | | +* +-----------------------------------+ +* +* SFT as datastructure: +* SFT can be treated as datastructure maintaining flow context across its +* lifetime. SFT flow entry represent bidirectional network flow and defined by
represent -> represents
Quoted 51 lines
Quoted 51 lines
lib/librte_sft/rte_sft.h new file +51
-0+* 7-tuple & its reverse 7-tuple. +* Each entry in SFT has: +* - FID: 1:1 mapped & used as entry handle & encapsulating internal +* implementation of the entry. +* - State: user-defined value attached to each entry, the only library +* reserved value for state unset (the actual value defined by SFT +* configuration). The application should define flow state encodings and +* set it for flow via rte_sft_flow_set_ctx() than what actions should be +* applied on packets can be defined via related RTE flow rule matching SFT +* state (see rules in SFT components diagram above). +* - Timestamp: for the last seen in flow packet used for flow aging mechanism +* implementation. +* - Client Objects: user-defined flow contexts attached as opaques to flow. +* - Acceleration & offloading - utilize RTE flow capabilities, when supported +* (see action ``SFT``), for flow lookup acceleration and further +* context-aware flow handling offload. +* - CT state: optionally for TCP connections CT state can be maintained +* (see enum rte_sft_flow_ct_state). +* - Out of order TCP packets: optionally SFT can keep out of order TCP +* packets aside the flow context till the arrival of the missing in-order +* packet. +* +* RTE flow changes: +* The SFT flow state (or context) for RTE flow is defined by fields of +* struct rte_flow_item_sft. +* To utilize SFT capabilities new item and action types introduced: +* - item SFT: matching on SFT flow state (see RTE_FLOW_ITEM_TYPE_SFT). +* - action SFT: retrieve SFT flow context and attache it to the processed +* packet (see RTE_FLOW_ACTION_TYPE_SFT). +* +* The contents of per port SFT serving RTE flow action ``SFT`` managed via +* SFT PMD APIs (see struct rte_sft_ops). +* The SFT flow state/context retrieval performed by user-defined zone ``SFT`` +* action argument and processed packet 5-tuple. +* If in scope of action ``SFT`` there is no context/state for the flow in SFT +* undefined sate attached to the packet meaning that the flow is not +* recognized by SFT, most probably FIF packet. +* +* Once the SFT state set for a packet it can match on item SFT +* (see RTE_FLOW_ITEM_TYPE_SFT) and forwarding design can be done for the +* packet, for example: +* - if state value == x than queue for further processing by the application +* - if state value == y than forward it to eth port (full offload) +* - if state value == 'undefined' than queue for further processing by +* the application (handle FIF packets) +* +* Processing packets with SFT library: +* +* FIF packet: +* To recognize upcoming packets of the SFT flow every FIF packet should be +* forwarded to the application utilizing the SFT library. Non-FIF packets can
Quoted 8 lines
lib/librte_sft/rte_sft.h line 167
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* be processed by the application or its processing can be fully offloaded. +* Processing of the packets in SFT library starts with rte_sft_process_mbuf +* or rte_sft_process_mbuf_with_zone. If mbuf recognized as FIF application +* should make a design to destroy flow or complete flow creation process in +* SFT using rte_sft_flow_activate. +* +* Recognized SFT flow: +* Once struct rte_sft_flow_status with valid fid field posesed by application
posesed -> possessed
Quoted 19 lines
Quoted 19 lines
lib/librte_sft/rte_sft.h new file +19
-0+* it can: +* - mange client objects on it (see client_obj field in +* struct rte_sft_flow_status) using rte_sft_flow_<OP>_client_obj APIs +* - analyze user-defined flow state and CT state (see state & ct_sate fields +* in struct rte_sft_flow_status). +* - set flow state to be attached to the upcoming packets by action ``SFT`` +* via struct rte_sft_flow_status API. +* - decide to destroy flow via rte_sft_flow_destroy API. +* +* Flow aging: +* +* SFT library manages the aging for each flow. On flow creation, it's +* assigned an aging value, the maximal number of seconds passed since the +* last flow packet arrived, once exceeded flow considered aged. +* The application notified of aged flow asynchronously via event queues. +* The device and port IDs tuple to identify the event queue to enqueue +* flow aged events passed on flow creation as arguments +* (see rte_sft_flow_activate). It's the application responsibility to +* initialize event queues and assign them to each flow for EOF event
Quoted 8 lines
lib/librte_sft/rte_sft.h line 194
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* notifications. +* Aged EOF event handling: +* - Should be considered as application responsibility. +* - The last stage should be the release of the flow resources via +* rte_sft_flow_destroy API. +* - All client objects should be removed from flow before the +* rte_sft_flow_destroy API call. +* See the description of rete_sft_flow_destroy for an example of aged flow
rete_sft_flow_destroy -> rte_sft_flow_destroy
Quoted 52 lines
lib/librte_sft/rte_sft.h line 201
Quoted 7 lines
lib/librte_sft/rte_sft.h new file +7
-0+* handling. +* +* SFT API thread safety: +* +* SFT library APIs are thread-safe while handling of specific flow can be +* done in a single thread simultaneously. Exclusive access to specific SFT +* flow guaranteed by: The line above contradict itself, if you are working with single thread you can't work simultaneously. Does the SFT allow the access to a single flow from two threads in the same time? or it is the responsibility Of the application to protect itself. I think it should be the application responsibility the SFT should protect itself only on SFT global functions. For example calling
process_mbufshould be protected, so application can call the same function from different threads.
I think we can assume that all packets from a specific flow will arrive to the same queue and the same thread.So I don't see the usage of the lock API.
Quoted 35 lines
lib/librte_sft/rte_sft.h new file +35
-0+* - rte_sft_process_mbuf +* - rte_sft_process_mbuf_with_zone +* - rte_sft_flow_create +* - rte_sft_flow_lock +* When application is done with the flow handling for the current packet it +* should call rte_sft_flow_unlock API to maintain exclusive access to the +* flow with other threads. +* +* SFT Library initialization and cleanup: +* +* SFT library should be considered as a single instance, preconfigured and +* initialized via rte_sft_init() API. +* SFT library resource deallocation and cleanup should be done via +* rte_sft_init() API as a stage of the application termination procedure. +*/ + +#ifdef __cplusplus +extern "C" { +#endif + +#include <rte_common.h> +#include <rte_config.h> +#include <rte_errno.h> +#include <rte_mbuf.h> +#include <rte_ethdev.h> +#include <rte_flow.h> + +/** + * L3/L4 5-tuple - src/dest IP and port and IP protocol. + * + * Used for flow/connection identification. + */ +struct rte_sft_5tuple { + union { + struct {
Quoted 8 lines
lib/librte_sft/rte_sft.h line 244
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+rte_be32_t src_addr;/**< IPv4 source address. */ +rte_be32_t dst_addr;/**< IPv4 destination address. */ + } ipv4; +struct { +uint8_t src_addr[16 ];/**< IPv6 source address. */ +uint8_t dst_addr[16 ];/**< IPv6 destination address. */ + } ipv6; + };
RTE_STD_C11 missing?
Quoted 2 lines
lib/librte_sft/rte_sft.h line 246
Quoted 2 lines
lib/librte_sft/rte_sft.h new file +2
-0+uint16_t src_port;/**< Source port. */ +uint16_t dst_port;/**< Destination port. */
If it is really host-endian, please, highlight it in above descriptions. Also it would be interesting to understand why.
Quoted 30 lines
Quoted 30 lines
lib/librte_sft/rte_sft.h new file +30
-0+uint8_t proto;/**< IP protocol. */ + uint8_t is_ipv6: 1; /**< True for valid IPv6 fields. Otherwise IPv4. */ +}; + +/** +* Port flow identification. +* +* @p zone used for setups where 5-tuple is not enough to identify flow. +* For example different VLANs/VXLANs may have similar 5-tuples. +*/ +struct rte_sft_7tuple { +struct rte_sft_5tuple flow_5tuple;/**< L3/L4 5-tuple. */ +uint32_t zone;/**< Zone assigned to flow. */ +uint16_t port_id;/** <Port identifier of Ethernet device. */ +}; + +/** +* Flow connection tracking states +*/ +enum rte_sft_flow_ct_state { +RTE_SFT_FLOW_CT_STATE_NEW = (1 <<0 ), +RTE_SFT_FLOW_CT_STATE_EST = (1 <<1 ), +RTE_SFT_FLOW_CT_STATE_REL = (1 <<2 ), +RTE_SFT_FLOW_CT_STATE_RPL = (1 <<3 ), +RTE_SFT_FLOW_CT_STATE_INV = (1 <<4 ), +RTE_SFT_FLOW_CT_STATE_TRK = (1 <<5 ), +RTE_SFT_FLOW_CT_STATE_SNAT = (1 <<6 ), +RTE_SFT_FLOW_CT_STATE_DNAT = (1 <<7 ), +}; +
Quoted 8 lines
lib/librte_sft/rte_sft.h line 284
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+/** +* Structure describes SFT library configuration +*/ +struct rte_sft_conf { +uint32_t UDP_aging;/**< UDP proto default aging. */ +uint32_t TCP_aging;/**< TCP proto default aging. */ +uint32_t TCP_SYN_aging;/**< TCP SYN default aging. */ +uint32_t OTHER_aging;/**< All unlisted proto default aging. */
May I suggest to stick to lowercase fields, please.
Quoted 29 lines
Quoted 29 lines
lib/librte_sft/rte_sft.h new file +29
-0+uint32_t size;/**< Max entries in SFT. */ +uint8_t undefined_state;/**< Undefined state constant. */ + uint8_t reorder_enable: 1; +/**< TCP packet reordering feature enabled bit. */ + uint8_t ct_enable: 1; /**< Connection tracking feature enabled bit. */ +}; + +/** +* Structure describes the state of the flow in SFT. +*/ +struct rte_sft_flow_status { +uint32_t fid;/**< SFT flow id. */ +uint32_t zone;/**< Zone for lookup in SFT */ +uint8_t state;/**< Application defined bidirectional flow state. */ +uint8_t ct_state;/**< Connection tracking flow state. */ +uint32_t age;/**< Seconds passed since last flown packet. */ +uint32_t aging; +/**< Flow considered aged once this age (seconds) reached. */ +uint32_t nb_in_order_mbufs; +/**< Number of in-order mbufs available for drain */ +void **client_obj;/**< Array of clients attached to flow. */ +int nb_clients;/**< Number of clients attached to flow. */ +uint8_t defined:1 ;/**< Flow defined in SFT bit. */ +uint8_t activated:1 ;/**< Flow activation bit. */ +uint8_t fragmented:1 ;/**< Last flow mbuf was fragmented. */ +uint8_t out_of_order:1 ;/**< Last flow mbuf was out of order (TCP). */ +}; + +/**
Quoted 8 lines
lib/librte_sft/rte_sft.h line 321
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get SFT flow status. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Flow should be locked by caller (see rte_sft_flow_lock). +*
Dup lines above
Quoted 11 lines
lib/librte_sft/rte_sft.h line 332
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump actual SFT flow status. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/
__rte_experimental
Quoted 5 lines
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+int +rte_sft_flow_get_status (const uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); +
Quoted 8 lines
lib/librte_sft/rte_sft.h line 345
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set user defined context. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Updates per ethernet dev SFT entries:
ethernet -> Ethernet
dev -> device
Quoted 7 lines
Quoted 7 lines
lib/librte_sft/rte_sft.h new file +7
-0+* - flow lookup acceleration +* - partial/full flow offloading managed by flow context +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param ctx
Quoted 8 lines
lib/librte_sft/rte_sft.h line 360
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* User defined state to set. +* Update of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set.
Remove space before comma
Quoted 1 line
lib/librte_sft/rte_sft.h line 361
Quoted 1 line
lib/librte_sft/rte_sft.h new file +1
-0+*/
__rte_experimental
Quoted 9 lines
Quoted 9 lines
lib/librte_sft/rte_sft.h new file +9
-0+int +rte_sft_flow_set_ctx (uint32_t fid, +const struct rte_flow_item_sft *ctx, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +*
Quoted 8 lines
lib/librte_sft/rte_sft.h line 378
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* Initialize SFT library instance. +* +* @param conf +* SFT library instance configuration. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. +*/
__rte_experimental
Quoted 29 lines
lib/librte_sft/rte_sft.h line 391
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +13
-0+int +rte_sft_init (const struct rte_sft_conf *conf); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Finalize SFT library instance. +* Cleanup & release allocated resources. +*/ +void +rte_sft_fini (void ); +I think we should use stop. It is not commons in DPDK to have fini functions. Maybe we should also add start function, so the app can init and then start the SFT.
lib/librte_sft/rte_sft.h line 403
Quoted 12 lines
lib/librte_sft/rte_sft.h new file +12
-0+/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue. +* +* Fragmentation handling (SFT fragmentation feature configured): +* If *mbuf_in* of fragmented packet received it will be stored by SFT library. +* status->fragmented bit will be set and *mbuf_out* will be set to NULL. +* On reception of all related fragments of IP packet it will be reassembled +* and further processed by this function on reception of last fragment. +*
Quoted 15 lines
Does this function allocate a new mbuf? Does it releases all old mbufs?
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +4
-0+* Flow definition: +* SFT flow defined by one of its 7-tuples, since there is no zone value as +* argument flow should be defined by context attached to mbuf with action +* ``SFT`` (see RTE flow RTE_FLOW_ACTION_TYPE_SFT). Otherwise status- Quoted 1 line
defined
+ * field will be turned off & *mbuf_out* will be set to *mbuf_in*.
+ * In order to define flow for *mbuf_in* without attached sft context
+ *rte_sft_process_mbuf_with_zone()should be used with *zone* argument
+ * supplied by caller.
+ *
+ * Flow lookup:
+ * If SFT flow identifier can't be retrieved from SFT context attached to
+ * *mbuf_in* by action ``SFT`` - SFT lookup should be performmed by zone,
performmed -> performed
Quoted 10 lines
Quoted 10 lines
+ * retrieved from SFT context attached to *mbuf_in*, and 5-tuple, extracted
+ * form mbuf outer header contents.
+ *
+ * Flow defined but does not exists:
+ * If flow not found in SFT inactivated flow will be created in SFT.
+ *status->activatedfield will be turned off & *mbuf_out* be set to *mbuf_in*.
+ * In order to activate created flowrte_sft_flow_activate()should be used
+ * with reverse 7-tuple supplied by caller.
+ * This is first phase of flow creation in SFT for second phase & more detailed
+ * descriotion of flow creation seerte_sft_flow_activate.
descriotion -> description
Quoted 2 lines
Quoted 2 lines
+ *
+ * Out of order (SFT out of oreder feature configured):
oreder -> order
Quoted 39 lines
Quoted 7 lines
+ * If flow defined & activated but *mbuf_in* is TCP out of order packet it will
+ * be stored by SFT library.status->out_of_orderbit will be set & *mbuf_out*
+ * will be set to NULL. On reception of the first missing in order packet
+ *status->nb_in_order_mbufswill be set to number of mbufs that available for
+ * processing withrte_sft_drain_mbuf().
+ *It is possible that some packets will get trapped in the SFT do to this feature. if it supports ordering. For example the following case:
Packets arrive to the application. After draining the packets the Application changed the flow to full offload. This means that
all future packets will not arrive to the application.
But until the flow is offloaded some packets do arrive not in order. Then the flow is offloaded, this will result in the situation that no more packets will arrive to the application so some packets will get stack in the SFT.
I think we must have some force drain or, notify the SFT that no more packets should arrive to even if the packets are not in order it will release them.Also the same with fragmented does this function allocate new mbufs? are you releasing the old ones?
Quoted 6 lines
+ * Flow defined & activated, mbuf not fragmented and 'in order':
+ * - Flow aging related data (see age field instruct rte_sft_flow_status)
+ * will be updated according to *mbuf_in* timestamp.
+ * - Flow connection tracking state (seect_statefield in
+ *struct rte_sft_flow_status) will be updated according to *mbuf_in* L4
+ * header contents.Quoted 8 lines
+ * - *mbuf_out* will be set to last processed mbuf.
+ *
+ * @param[in]mbuf_in
+ * mbuf to process; mbuf pinter considered 'consumed' and should not be
used
+ * after successful call to this function.
+ * @param[out]mbuf_out
+ * last processed not fragmented and in order mbuf.If the in mbuf is not fragmented and in order, this pointer will point to the in one?
Quoted 23 lines
Quoted 3 lines
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contents of
+ * *mbuf_in*.Does the status bits for example fragmented is kept per connection or per flow? Since it is possible to get fragmented packets from both sides. The same goes for out of order packets
Quoted 14 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success:
+ * - *mbuf_out* contains valid mbuf pointer, locked SFT flow recognized by
+ *status->fid.
+ * - *mbuf_out* is NULL andstatus->fragmentedbit on in case of
+ * non last fragment *mbuf_in*.
+ * - *mbuf_out* is NULL andstatus->out_of_orderbit on in case of out of
+ * order *mbuf_in*, locked SFT flow recognized bystatus->fid.
+ * On failure a negative errno value andrte_errnois set.
+ */
__rte_experimental
Quoted 12 lines
Quoted 12 lines
lib/librte_sft/rte_sft.h new file +12
-0+int +rte_sft_process_mbuf (struct rte_mbuf *mbuf_in, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue while zone value provided by caller. +*
Quoted 8 lines
Quoted 8 lines
+ * The behaviour of this function is similar to
rte_sft_process_mbufexcept
+ * the lookup in SFT procedure. The lookup in SFT always done by the *zone*
+ * arg and 5-tuple 5-tuple, extracted form mbuf outer header contents.
+ *
+ * @seerte_sft_process_mbuf
+ *
+ * @param[in]mbuf_in
+ * mbuf to process; mbuf pinter considered 'consumed' and should not be used
pinter -> pointer
Quoted 11 lines
Quoted 11 lines
+ * after successful call to this function.
+ * @param[out]mbuf_out
+ * last processed not fragmented and in order mbuf.
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contents of
+ * *mbuf_in*.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
Quoted 8 lines
Quoted 8 lines
+ * 0 on success: + * - *mbuf_out* contains valid mbuf pointer. + * - *mbuf_out* is NULL and status->fragmented bit on in case of+ * non last fragment *mbuf_in*.
+ * - *mbuf_out* is NULL andstatus->out_of_orderbit on in case of out of
+ * order *mbuf_in*.
+ * On failure a negative errno value andrte_errnois set.
+ */
__rte_experimental
Quoted 28 lines
Quoted 28 lines
lib/librte_sft/rte_sft.h new file +28
-0+int +rte_sft_process_mbuf_with_zone (struct rte_mbuf *mbuf_in, +uint32_t zone, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Drain next in order mbuf. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* This function behaves similar to rte_sft_process_mbuf() but acts on packets +* accumulated in SFT flow due to missing in order packet. Processing done on +* single mbuf at a time and `in order`. Other than above the behavior is +* same as of rte_sft_process_mbuf for flow defined & activated & mbuf isn't +* fragmented & 'in order'. This function should be called when +* rte_sft_process_mbuf or rte_sft_process_mbuf_with_zone sets +* status->nb_in_order_mbufs output param !=0 and until +* status->nb_in_order_mbufs == 0. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump SFT flow status once updated according to contents of
Quoted 8 lines
Quoted 8 lines
+ * *mbuf_in*.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid mbuf in case of success, NULL otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 62 lines
lib/librte_sft/rte_sft.h line 543
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+struct rte_mbuf * +rte_sft_drain_mbuf (uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); +Fid represent a connection, so which direction do we drain the packets? since we can have inordered packet in from both directions right?
Quoted 53 lines
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice.
+ *
+ * Activate flow in SFT.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * This function performs second phase of flow creation in SFT.
+ * The reasons for 2 phase flow creation procedure:
+ * 1. Missing reverse flow - flow context is shared for both flow directions+ * i.e. in order maintain bidirectional flow context in RTE SFT packets + * arriving from both dirrections should be identified as packets of the + * RTE SFT flow. Consequently before creation of the SFT flow caller should + * provide reverse flow direction 7-tuple. + * 2. The caller of rte_sft_process_mbuf/rte_sft_process_mbuf_with_zone should + * be notified that arrived mbuf is first in flow & decide weather to + * create new flow or it distroy before it was activated with + * rte_sft_flow_destroy.+ * This function completes creation of the bidirectional SFT flow & creates
+ * entry for 7-tuple on SFT PMD defined by the tuple port for both
+ * initiator/initiate 7-tuples.
+ * Flow aging, connection tracking state & out of order handling will be
+ * initialized according to the content of the *mbuf_in* passes to
+ * rte_sft_process_mbuf/_with_zone during the phase 1 of flow creation.
+ * Once this function returns upcoming calls rte_sft_process_mbuf/_with_zone
+ * with 7-tuple or its reverse will return handle to this flow.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * @param fid
+ * SFT flow ID.
+ * @paramreverse_tuple
+ * Expected response flow 7-tuple.
+ * @param ctx
+ * User defined state to set.
+ * Update of *fid* or *zone* fields in structrte_flow_item_sftunsupported.
+ * @paramct_enable
+ * Enables maintenance ofstatus->ct_stateconnection tracking value for the
+ * flow; otherwisestatus->ct_statewill be initialized with zeros.
+ * @paramevdev_id
+ * Event dev ID to enqueue end of flow event.
+ * @paramevport_id
+ * Event port ID to enqueue end of flow event.
+ * @param[out] status
+ * Structure to dump SFT flow status once activated.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 31 lines
Quoted 31 lines
lib/librte_sft/rte_sft.h new file +31
-0+int +rte_sft_flow_activate (uint32_t fid, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +uint8_t dev_id, +uint8_t port_id, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Artificially create SFT flow. +* +* Function to create SFT flow before reception of the first flow packet. +* +* @param tuple +* Expected initiator flow 7-tuple. +* @param reverse_tuple +* Expected initiate flow 7-tuple. +* @param ctx +* User defined state to set. +* Setting of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param[out] ct_enable +* Enables maintenance of status->ct_state connection tracking value for the +* flow; otherwise status->ct_state will be initialized with zeros. +* @param[out] status +* Structure to dump SFT flow status once created. +* @param[out] error
Quoted 8 lines
Quoted 8 lines
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * - on success: 0, locked SFT flow recognized bystatus->fid.
+ * - on error: a negative errno value otherwise andrte_errnois set.
+ */
+
No extra empty line and __rte_experimental
Quoted 14 lines
Quoted 14 lines
lib/librte_sft/rte_sft.h new file +14
-0+int +rte_sft_flow_create (const struct rte_sft_7tuple *tuple, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Lock exclusively SFT flow. +*
Quoted 8 lines
Quoted 8 lines
+ * Explicit flow locking; used for handling aged flows.
+ *
+ * @param fid
+ * SFT flow ID.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 11 lines
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+int +rte_sft_flow_lock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Release exclusively locked SFT flow. +* +* When rte_sft_process_mbuf/_with_zone and rte_sft_flow_create +* return *status* containing fid with defined bit on the flow considered
Quoted 8 lines
Quoted 8 lines
+ * exclusively locked and should be unlocked with this function.
+ *
+ * @param fid
+ * SFT flow ID.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 31 lines
Quoted 31 lines
lib/librte_sft/rte_sft.h new file +31
-0+int +rte_sft_flow_unlock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Removes flow from SFT. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* - Flow should be locked by caller in order to remove it. +* - Flow should have no client objects attached. +* +* Should be applied on aged flows, when flow aged event received. +* +* @code{.c} +* while (1) { +* rte_event_dequeue_burst(...); +* FOR_EACH_EV(ev) { +* uint32_t fid = ev.u64; +* rte_sft_flow_lock(fid); +* FOR_EACH_CLIENT(fid, client_id) { +* rte_sft_flow_reset_client_obj(fid, client_obj); +* // detached client object handling +* } +* rte_sft_flow_destroy(fid, &error); +* } +* } +* @endcode +* +* @param fid
Quoted 8 lines
Quoted 8 lines
+ * SFT flow ID to destroy.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 13 lines
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +13
-0+int +rte_sft_flow_destroy (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Reset flow age to zero. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Simulates last flow packet with timestamp set to just now. +* +* @param fid
Quoted 8 lines
Quoted 8 lines
+ * SFT flow ID.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 13 lines
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +13
-0+int +rte_sft_flow_touch (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set flow aging to specific value. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param aging
Quoted 8 lines
Quoted 8 lines
+ * New flow aging value.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 17 lines
Quoted 17 lines
lib/librte_sft/rte_sft.h new file +17
-0+int +rte_sft_flow_set_aging (uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to set object for. +* @param client_obj
Quoted 8 lines
Quoted 8 lines
+ * Pointer to opaque client object structure.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimental
Quoted 17 lines
Quoted 17 lines
lib/librte_sft/rte_sft.h new file +17
-0+int +rte_sft_flow_set_client_obj (uint32_t fid, +uint8_t client_id, +void *client_obj, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to get object for.
Quoted 8 lines
Quoted 8 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid client object opaque pointer in case of success, NULL otherwise
+ * andrte_errnois set.
+ */
__rte_experimental
Quoted 10 lines
lib/librte_sft/rte_sft.h line 815
Quoted 10 lines
lib/librte_sft/rte_sft.h new file +10
-0+void * +rte_sft_flow_get_client_obj (const uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Remove client object for given client ID.
Function name uses "reset", but description says "remove". May be synchronize it?
Quoted 12 lines
Quoted 12 lines
+ * Flow should be locked by caller (see
rte_sft_flow_lock).
+ *
+ * Detaches client object from SFT flow and returns the ownership for the
+ * client object to the caller by returning client object pointer value.
+ * The pointer returned by this function won't be accessed any more, the caller
+ * may release all client obj related resources & the memory allocated for
+ * this client object.
+ *
+ * @param fid
+ * SFT flow ID.
+ * @paramclient_id
+ * Client ID to remove object for.
Quoted 8 lines
Quoted 8 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid client object opaque pointer in case of success, NULL otherwise
+ * andrte_errnois set.
+ */
__rte_experimental
Quoted 247 lines
Quoted 76 lines
lib/librte_sft/rte_sft.h new file +10
-0+void * +rte_sft_flow_reset_client_obj (uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_H_ */ lib/librte_sft/rte_sft_driver.h new file +60
-0@@ -0,0 +1,195 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_DRIVER_H_ +#define _RTE_SFT_DRIVER_H_ + +/** + * @file + * + * RTE SFT Ethernet device PMD API + * + * APIs that are used by the SFT library to offload SFT operationons + * to Ethernet device. + */ + +#include "rte_sft.h" + +#ifdef __cplusplus +extern "C" { +#endif + +/** + * Opaque type returned after successfully creating an entry in SFT. + * + * This handle can be used to manage and query the related entry (e.g. to + * destroy it or update age). + */ +struct rte_sft_entry; + +/** + * Create SFT entry in eth_dev SFT. + * + * @param dev + * Pointer to Ethernet device structure. + * @param tuple + * L3/L4 5-tuple - src/dest IP and port and IP protocol. + * @param nat_tuple + * L3/L4 5-tuple to replace in packet original 5-tuple in order to implement + * NAT offloading; if NULL NAT offloading won't be configured for the flow. + * @param aging + * Flow aging timeout in seconds. + * @param ctx + * Initial values in SFT flow context + * (see RTE flow struct rte_flow_item_sft). + * ctx->zone should be valid. + * @param fid + * SFT flow ID for the entry to create on *device*. + * If there is an entry for the *fid* in PMD it will be updated with the + * values of *ctx*. + * @param[out] queue_index + * if PMD can figure out the queue where the flow packets will + * arrive in RX data path it will set the value of queue_index; otherwise + * all bits will be turned on. + * @param[out] error + * Perform verbose error reporting if not NULL. PMDs initialize this + * structure in case of error only. + * + * @return + * A valid handle in case of success, NULL otherwise and rte_errno is set.lib/librte_sft/rte_sft_driver.h line 70
Quoted 10 lines
lib/librte_sft/rte_sft_driver.h new file +10
-0+*/ +typedef struct rte_sft_entry *(*sft_entry_create_t ) (struct rte_eth_dev *dev, +const struct rte_sft_5tuple *tuple, +const struct rte_sft_5tuple *nat_tuple, +const uint32_t aging, +const struct rte_flow_item_sft *ctx, +const uint32_t fid, +uint16_t *queue_index, +struct rte_sft_error *error); +I think for easier reading, the API should change to have 6 tuple (5 + zone) the ctx should be removed and replaced with the state.
Then add new API to modify the ctx
typedef int (*sft_modify_state)(structrte_eth_dev*dev, uint8 state);
The main issue we my suggestion is that it will force the PMD to store the information to recreate the rule, data that is already saved by the SFT.Also I don't see why we need queue index, since the RSS and queue will be configured by the RTE flow in a different group.
Quoted 29 lines
lib/librte_sft/rte_sft_driver.h new file +29
-0+/** +* Destroy SFT entry in eth_dev SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param entry +* Handle to the SFT entry to destroy. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_destroy_t )(struct rte_eth_dev *dev, +struct rte_sft_entry *entry, +struct rte_sft_error *error); + +/** +* Decodes SFT flow context if attached to mbuf by action ``SFT``. +* @see RTE flow RTE_FLOW_ACTION_TYPE_SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param mbuf +* mbuf of the packet to decode attached state from. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. lib/librte_sft/rte_sft_driver.h line 109
Quoted 10 lines
lib/librte_sft/rte_sft_driver.h new file +10
-0+* +* @return +* A valid SFT flow context in case of success, NULL otherwise and rte_errno +* is set. +*/ +typedef struct rte_flow_item_sft *(*sft_entry_mbuf_decode_ctx_t )( +struct rte_eth_dev *dev, +const struct rte_mbuf *mbuf, +struct rte_sft_error *error); +What about returning int as error code, and return the
rte_flow_item_sftas out parameter?
This will remove the allocation and free.Quoted 105 lines
lib/librte_sft/rte_sft_driver.h new file +86
-0+/** +* Get aged-out SFT entries. +* +* Report entry as aged-out if timeout passed without any matching +* on the SFT entry. +* +* @param[in] dev +* Pointer to Ethernet device structure. +* @param[in, out] fid_aged +* The address of an array of aged-out SFT flow IDs. +* @param[in] nb_aged +* The length of *fid_aged* array pointers. +* @param[out] error +* Perform verbose error reporting if not NULL. Initialized in case of +* error only. +* +* @return +* if nb_aged is 0, return the amount of all aged flows. +* if nb_aged is not 0 , return the amount of aged flows reported +* in the *fid_aged* array, otherwise negative errno value. +*/ +typedef int (*sft_entry_get_aged_entries_t )(struct rte_eth_dev *dev, +uint32_t *fid_aged, +int nb_aged, +struct rte_sft_error *error); + +/** +* Simulate SFT entry match in terms of entry aging. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_touch_t )(struct rte_eth_dev *dev, +uint32_t fid, +struct rte_sft_error *error); + +/** +* Set SFT entry aging to specific value. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param aging +* New entry aging value. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_set_aging_t )(struct rte_eth_dev *dev, +uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** SFT operations function pointer table */ +struct rte_sft_ops { +sft_entry_create_t entry_create; +/**< Create SFT entry in eth_dev SFT. */ +sft_entry_destroy_t entry_destroy; +/**< Destroy SFT entry in eth_dev SFT. */ +sft_entry_mbuf_decode_ctx_t mbuf_decode_ctx; +/**< Decodes SFT flow context if attached to mbuf by action ``SFT``. */ +sft_entry_get_aged_entries_t get_aged_entries; +/**< Get aged-out SFT entries. */ +sft_entry_touch_t entry_touch; +/**< Simulate SFT entry match in terms of entry aging. */ +sft_entry_set_aging_t set_aging; +/**< Set SFT entry aging to specific value. */ +}; + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_DRIVER_H_ */ lib/librte_sft/rte_sft_version.map new file +12
-0@@ -0,0 +1,21 @@ +EXPERIMENTAL { + global: + + rte_sft_flow_get_status; + rte_sft_flow_set_ctx; + rte_sft_init; + rte_sft_fini; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + rte_sft_drain_mbuf; + rte_sft_flow_activate; + rte_sft_flow_create;
Quoted 8 lines
lib/librte_sft/rte_sft_version.map line 20
Quoted 8 lines
lib/librte_sft/rte_sft_version.map new file +8
-0+ rte_sft_flow_lock; + rte_sft_flow_unlock; + rte_sft_flow_destroy; + rte_sft_flow_touch; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_get_client_obj; + rte_sft_flow_reset_client_obj;
If I'm not mistaken, it should be alphabetially sorted.
Quoted 7 lines
Quoted 3 lines
lib/librte_sft/rte_sft_version.map new file +1
-1+}; --2.26.2
Best,
Ori
Re: [dpdk-dev] [RFC 3/3] sft: introduce API
Hi Andrew, Ray and everybody reviewing SFT RFC patch series.
Anything you comments to this patchset coupled with other open issues made me & Ori rethink the suggested API.
Ori is about to publish V2 for SFT RFC very soon addressing al the mentioned above.
Thanks,
Andrey
Quoted 1,428 lines
-----Original Message-----
From: Andrew Rybchenko <[email protected]>
Sent: Friday, September 18, 2020 10:44 AM
To: Ori Kam <[email protected]>; Andrey Vesnovaty <[email protected]>; [email protected]
Cc: [email protected]; Slava Ovsiienko <[email protected]>; [email protected]; Oz Shlomo <[email protected]>; Eli Britstein <[email protected]>; Alex Rosenbaum <[email protected]>; Roni Bar Yanai <[email protected]>; Ray Kinsella <[email protected]>; Neil Horman <[email protected]>; Ferruh Yigit <[email protected]>
Subject: Re: [RFC 3/3] sft: introduce APIHi Andrey,
looks very interesting, but a bit hard to review.
I hope I'll do deeper review on the next version.
Right not just few cosmetic things to make the
next version a bit clearer.Do you plan to create/publish an example appliction which uses the API and demonstrates how to do it?
Plesee, see below.
Thanks,
Andrew.Quoted 122 lines · Ori Kam
Hi Andery,
PSBQuoted 119 lines
-----Original Message-----
From: Andrey Vesnovaty <[email protected]>
Sent: Wednesday, September 9, 2020 11:30 PM
To: [email protected]
Subject: [RFC 3/3] sft: introduce APIDefines RTE SFT APIs for Statefull Flow Table library.
SFT General description:
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.Signed-off-by Andrey Vesnovaty <[email protected]>--- lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 +++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 6 files changed, 1105 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.maplib/librte_sft/Makefile new file +28
-0@@ -0,0 +1,28 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +include $(RTE_SDK)/mk/rte.vars.mk + +# library name +LIB = librte_sft.a + +# library version +LIBABIVER := 1 + +# build flags +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) +LDLIBS += -lrte_eal -lrte_mbuf + +# library source files +# all source are stored in SRCS-y +SRCS-$(CONFIG_RTE_LIBRTE_REGEXDEV) := rte_sft.c + +# export include files +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft.h +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft_driver.h + +# versioning export map +EXPORT_MAP := rte_sft_version.map + +include $(RTE_SDK)/mk/rte.lib.mklib/librte_sft/meson.build new file +7
-0@@ -0,0 +1,7 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +sources = files('rte_sft.c') +headers = files('rte_sft.h', + 'rte_sft_driver.h') +deps += ['mbuf']lib/librte_sft/rte_sft.c new file +9
-0@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */ lib/librte_sft/rte_sft.h new file +15
-0@@ -0,0 +1,845 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_H_ +#define _RTE_SFT_H_ + +/** +* @file +* +* RTE SFT API +* +* Defines RTE SFT APIs for Statefull Flow Table library. +* +* SFT General description: Quoted 8 lines
lib/librte_sft/rte_sft.h line 23
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* SFT library provides a framework for applications that need to maintain +* context across different packets of the connection. +* Examples for such applications: +* - Next-generation firewalls +* - Intrusion detection/prevention systems (IDS/IPS): Suricata, Snort +* - SW/Virtual Switching: OVS +* The goals of the SFT library: +* - Accelerate flow recognition & its context retrieval for further lookaside lookaside -> look-aside
Quoted 11 lines
lib/librte_sft/rte_sft.h line 34
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+* processing. +* - Enable context-aware flow handling offload. +* +* Definitions and Abbreviations: +* - 5-tuple: defined by: +* -- Source IP address +* -- Source port +* -- Destination IP address +* -- Destination port +* -- IP protocol number +* - 7-tuple: 5-tuple zone and port (see struct rte_sft_7tuple) I guess comma is missing after "5-tuple", since I read it as "5-tuple zone"???
Quoted 5 lines
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+* - 5/7-tuple: 5/7-tuple of the packet from connection initiator +* - revers 5/7-tuple: 5/7-tuple of the packet from connection initiate +* - application: SFT library API consumer +* - APP: see application +* - CID: client ID Quoted 8 lines
lib/librte_sft/rte_sft.h line 47
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* - CT: connection tracking +* - FID: Flow identifier +* - FIF: First In Flow +* - Flow: defined by 7-tuple and its reverse i.e. flow is bidirectional +* - SFT: Stateful Flow Table +* - user: see application +* - zone: additional user defined value used as differentiator for +* connections having same 5-tuple (for example different VxLan VxLan -> VXLAN (see
devtools/words-case.txt)Quoted 53 lines
Quoted 53 lines
lib/librte_sft/rte_sft.h new file +53
-0+* connections with same inner 5-tuple). +* +* SFT components: +* +* +-----------------------------------+ +* | RTE flow | +* | | +* | +-------------------------------+ | +----------------+ +* | | group X | | | RTE_SFT | +* | | | | | | +* | | +---------------------------+ | | | | +* | | | rule ... | | | | | +* | | | . | | | +-----------+----+ +* | | | . | | | | +* | | | . | | | entry +* | | +---------------------------+ | | create +* | | | rule | | | | +* | | | patterns ... +---------+ | +* | | | actions | | | | | +* | | | SFT (zone=Z) | | | | | +* | | | JUMP (group=Y) | | | lookup | +* | | +---------------------------+ | | zone=Z, | +* | | | rule ... | | | 5tuple | +* | | | . | | | | | +* | | | . | | | +--v-------------+ +* | | | . | | | | SFT | | +* | | | | | | | | | +* | | +---------------------------+ | | | +--v--+ | +* | | | | | | | | +* | +-------------------------------+ | | | PMD | | +* | | | | | | +* | | | +-----+ | +* | +-------------------------------+ | | | +* | | group Y | | | | +* | | | | | set flow CTX | +* | | +---------------------------+ | | | | +* | | | rule | | | +--------+-------+ +* | | | patterns | | | | +* | | | SFT (state=UNDEFINED) | | | | +* | | | actions RSS | | | | +* | | +---------------------------+ | | | +* | | | rule | | | | +* | | | patterns | | | | +* | | | SFT (state=INVALID) | <-------------+ +* | | | actions DROP | | | forward +* | | +---------------------------+ | | group=Y +* | | | rule | | | +* | | | patterns | | | +* | | | SFT (state=ACCEPTED) | | | +* | | | actions PORT | | | +* | | +---------------------------+ | | +* | | ... | | +* | | | | Quoted 8 lines
lib/librte_sft/rte_sft.h
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* | +-------------------------------+ | +* | ... | +* | | +* +-----------------------------------+ +* +* SFT as datastructure: +* SFT can be treated as datastructure maintaining flow context across its +* lifetime. SFT flow entry represent bidirectional network flow and defined by
represent -> represents
Quoted 11 lines
lib/librte_sft/rte_sft.h
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+* 7-tuple & its reverse 7-tuple. +* Each entry in SFT has: +* - FID: 1:1 mapped & used as entry handle & encapsulating internal +* implementation of the entry. +* - State: user-defined value attached to each entry, the only library +* reserved value for state unset (the actual value defined by SFT +* configuration). The application should define flow state encodings and +* set it for flow via rte_sft_flow_set_ctx() than what actions should be +* applied on packets can be defined via related RTE flow rule matching SFT +* state (see rules in SFT components diagram above). +* - Timestamp: for the last seen in flow packet used for flow aging mechanism
Quoted 14 lines
Quoted 14 lines
lib/librte_sft/rte_sft.h new file +14
-0+* implementation. +* - Client Objects: user-defined flow contexts attached as opaques to flow. +* - Acceleration & offloading - utilize RTE flow capabilities, when supported +* (see action ``SFT``), for flow lookup acceleration and further +* context-aware flow handling offload. +* - CT state: optionally for TCP connections CT state can be maintained +* (see enum rte_sft_flow_ct_state). +* - Out of order TCP packets: optionally SFT can keep out of order TCP +* packets aside the flow context till the arrival of the missing in-order +* packet. +* +* RTE flow changes: +* The SFT flow state (or context) for RTE flow is defined by fields of +* struct rte_flow_item_sft. Quoted 8 lines
lib/librte_sft/rte_sft.h
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* To utilize SFT capabilities new item and action types introduced: +* - item SFT: matching on SFT flow state (see RTE_FLOW_ITEM_TYPE_SFT). +* - action SFT: retrieve SFT flow context and attache it to the processed +* packet (see RTE_FLOW_ACTION_TYPE_SFT). +* +* The contents of per port SFT serving RTE flow action ``SFT`` managed via +* SFT PMD APIs (see struct rte_sft_ops). +* The SFT flow state/context retrieval performed by user-defined zone ``SFT``
Quoted 7 lines
lib/librte_sft/rte_sft.h
Quoted 7 lines
lib/librte_sft/rte_sft.h new file +7
-0+* action argument and processed packet 5-tuple. +* If in scope of action ``SFT`` there is no context/state for the flow in SFT +* undefined sate attached to the packet meaning that the flow is not +* recognized by SFT, most probably FIF packet. +* +* Once the SFT state set for a packet it can match on item SFT +* (see RTE_FLOW_ITEM_TYPE_SFT) and forwarding design can be done for the
Quoted 6 lines
Quoted 6 lines
lib/librte_sft/rte_sft.h new file +6
-0+* packet, for example: +* - if state value == x than queue for further processing by the application +* - if state value == y than forward it to eth port (full offload) +* - if state value == 'undefined' than queue for further processing by +* the application (handle FIF packets) +* Quoted 8 lines
lib/librte_sft/rte_sft.h
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* Processing packets with SFT library: +* +* FIF packet: +* To recognize upcoming packets of the SFT flow every FIF packet should be +* forwarded to the application utilizing the SFT library. Non-FIF packets can +* be processed by the application or its processing can be fully offloaded. +* Processing of the packets in SFT library starts with rte_sft_process_mbuf +* or rte_sft_process_mbuf_with_zone. If mbuf recognized as FIF application
Quoted 1 line
lib/librte_sft/rte_sft.h line 163
Quoted 1 line
lib/librte_sft/rte_sft.h new file +1
-0+* should make a design to destroy flow or complete flow creation process in
Quoted 4 lines
lib/librte_sft/rte_sft.h line 167
Quoted 4 lines
lib/librte_sft/rte_sft.h new file +4
-0+* SFT using rte_sft_flow_activate. +* +* Recognized SFT flow: +* Once struct rte_sft_flow_status with valid fid field posesed by application posesed -> possessed
Quoted 19 lines
Quoted 19 lines
lib/librte_sft/rte_sft.h new file +19
-0+* it can: +* - mange client objects on it (see client_obj field in +* struct rte_sft_flow_status) using rte_sft_flow_<OP>_client_obj APIs +* - analyze user-defined flow state and CT state (see state & ct_sate fields +* in struct rte_sft_flow_status). +* - set flow state to be attached to the upcoming packets by action ``SFT`` +* via struct rte_sft_flow_status API. +* - decide to destroy flow via rte_sft_flow_destroy API. +* +* Flow aging: +* +* SFT library manages the aging for each flow. On flow creation, it's +* assigned an aging value, the maximal number of seconds passed since the +* last flow packet arrived, once exceeded flow considered aged. +* The application notified of aged flow asynchronously via event queues. +* The device and port IDs tuple to identify the event queue to enqueue +* flow aged events passed on flow creation as arguments +* (see rte_sft_flow_activate). It's the application responsibility to +* initialize event queues and assign them to each flow for EOF event Quoted 8 lines
lib/librte_sft/rte_sft.h line 194
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* notifications. +* Aged EOF event handling: +* - Should be considered as application responsibility. +* - The last stage should be the release of the flow resources via +* rte_sft_flow_destroy API. +* - All client objects should be removed from flow before the +* rte_sft_flow_destroy API call. +* See the description of rete_sft_flow_destroy for an example of aged flow
rete_sft_flow_destroy->rte_sft_flow_destroyQuoted 9 lines
lib/librte_sft/rte_sft.h line 201
Quoted 7 lines
lib/librte_sft/rte_sft.h new file +7
-0+* handling. +* +* SFT API thread safety: +* +* SFT library APIs are thread-safe while handling of specific flow can be +* done in a single thread simultaneously. Exclusive access to specific SFT +* flow guaranteed by: The line above contradict itself, if you are working with single thread you can't
work simultaneously.
Quoted 1 line
Does the SFT allow the access to a single flow from two threads in the same
time? or it is the responsibility
Quoted 1 line
Of the application to protect itself. I think it should be the application
responsibility the SFT should protect
Quoted 1 line
itself only on SFT global functions. For example calling
process_mbufshould beprotected, so application can
Quoted 2 lines
call the same function from different threads.
I think we can assume that all packets from a specific flow will arrive to thesame queue and the same thread.
Quoted 38 lines
So I don't see the usage of the lock API.
Quoted 35 lines
lib/librte_sft/rte_sft.h new file +35
-0+* - rte_sft_process_mbuf +* - rte_sft_process_mbuf_with_zone +* - rte_sft_flow_create +* - rte_sft_flow_lock +* When application is done with the flow handling for the current packet it +* should call rte_sft_flow_unlock API to maintain exclusive access to the +* flow with other threads. +* +* SFT Library initialization and cleanup: +* +* SFT library should be considered as a single instance, preconfigured and +* initialized via rte_sft_init() API. +* SFT library resource deallocation and cleanup should be done via +* rte_sft_init() API as a stage of the application termination procedure. +*/ + +#ifdef __cplusplus +extern "C" { +#endif + +#include <rte_common.h> +#include <rte_config.h> +#include <rte_errno.h> +#include <rte_mbuf.h> +#include <rte_ethdev.h> +#include <rte_flow.h> + +/** + * L3/L4 5-tuple - src/dest IP and port and IP protocol. + * + * Used for flow/connection identification. + */ +struct rte_sft_5tuple { + union { + struct {Quoted 8 lines
lib/librte_sft/rte_sft.h line 244
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+rte_be32_t src_addr;/**< IPv4 source address. */ +rte_be32_t dst_addr;/**< IPv4 destination address. */ + } ipv4; +struct { +uint8_t src_addr[16 ];/**< IPv6 source address. */ +uint8_t dst_addr[16 ];/**< IPv6 destination address. */ + } ipv6; + };
RTE_STD_C11missing?Quoted 2 lines
lib/librte_sft/rte_sft.h line 246
Quoted 2 lines
lib/librte_sft/rte_sft.h new file +2
-0+uint16_t src_port;/**< Source port. */ +uint16_t dst_port;/**< Destination port. */ If it is really host-endian, please, highlight it in above descriptions. Also it would be interesting to understand why.
Quoted 30 lines
Quoted 30 lines
lib/librte_sft/rte_sft.h new file +30
-0+uint8_t proto;/**< IP protocol. */ + uint8_t is_ipv6: 1; /**< True for valid IPv6 fields. Otherwise IPv4. */ +}; + +/** +* Port flow identification. +* +* @p zone used for setups where 5-tuple is not enough to identify flow. +* For example different VLANs/VXLANs may have similar 5-tuples. +*/ +struct rte_sft_7tuple { +struct rte_sft_5tuple flow_5tuple;/**< L3/L4 5-tuple. */ +uint32_t zone;/**< Zone assigned to flow. */ +uint16_t port_id;/** <Port identifier of Ethernet device. */ +}; + +/** +* Flow connection tracking states +*/ +enum rte_sft_flow_ct_state { +RTE_SFT_FLOW_CT_STATE_NEW = (1 <<0 ), +RTE_SFT_FLOW_CT_STATE_EST = (1 <<1 ), +RTE_SFT_FLOW_CT_STATE_REL = (1 <<2 ), +RTE_SFT_FLOW_CT_STATE_RPL = (1 <<3 ), +RTE_SFT_FLOW_CT_STATE_INV = (1 <<4 ), +RTE_SFT_FLOW_CT_STATE_TRK = (1 <<5 ), +RTE_SFT_FLOW_CT_STATE_SNAT = (1 <<6 ), +RTE_SFT_FLOW_CT_STATE_DNAT = (1 <<7 ), +}; +Quoted 8 lines
lib/librte_sft/rte_sft.h line 284
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+/** +* Structure describes SFT library configuration +*/ +struct rte_sft_conf { +uint32_t UDP_aging;/**< UDP proto default aging. */ +uint32_t TCP_aging;/**< TCP proto default aging. */ +uint32_t TCP_SYN_aging;/**< TCP SYN default aging. */ +uint32_t OTHER_aging;/**< All unlisted proto default aging. */ May I suggest to stick to lowercase fields, please.
Quoted 29 lines
Quoted 29 lines
lib/librte_sft/rte_sft.h new file +29
-0+uint32_t size;/**< Max entries in SFT. */ +uint8_t undefined_state;/**< Undefined state constant. */ + uint8_t reorder_enable: 1; +/**< TCP packet reordering feature enabled bit. */ + uint8_t ct_enable: 1; /**< Connection tracking feature enabled bit. */ +}; + +/** +* Structure describes the state of the flow in SFT. +*/ +struct rte_sft_flow_status { +uint32_t fid;/**< SFT flow id. */ +uint32_t zone;/**< Zone for lookup in SFT */ +uint8_t state;/**< Application defined bidirectional flow state. */ +uint8_t ct_state;/**< Connection tracking flow state. */ +uint32_t age;/**< Seconds passed since last flown packet. */ +uint32_t aging; +/**< Flow considered aged once this age (seconds) reached. */ +uint32_t nb_in_order_mbufs; +/**< Number of in-order mbufs available for drain */ +void **client_obj;/**< Array of clients attached to flow. */ +int nb_clients;/**< Number of clients attached to flow. */ +uint8_t defined:1 ;/**< Flow defined in SFT bit. */ +uint8_t activated:1 ;/**< Flow activation bit. */ +uint8_t fragmented:1 ;/**< Last flow mbuf was fragmented. */ +uint8_t out_of_order:1 ;/**< Last flow mbuf was out of order (TCP). */ +}; + +/** Quoted 8 lines
lib/librte_sft/rte_sft.h line 321
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get SFT flow status. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Flow should be locked by caller (see rte_sft_flow_lock). +* Dup lines above
Quoted 11 lines
lib/librte_sft/rte_sft.h line 332
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump actual SFT flow status. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/
__rte_experimentalQuoted 5 lines
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+int +rte_sft_flow_get_status (const uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); +Quoted 8 lines
lib/librte_sft/rte_sft.h line 345
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set user defined context. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Updates per ethernet dev SFT entries: ethernet -> Ethernet
dev -> deviceQuoted 9 lines
lib/librte_sft/rte_sft.h
Quoted 9 lines
lib/librte_sft/rte_sft.h new file +9
-0+* - flow lookup acceleration +* - partial/full flow offloading managed by flow context +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param ctx +* User defined state to set. +* Update of *fid* or *zone* fields in struct rte_flow_item_sft unsupported.
Quoted 6 lines
lib/librte_sft/rte_sft.h line 360
Quoted 6 lines
lib/librte_sft/rte_sft.h new file +6
-0+* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. Remove space before comma
Quoted 1 line
lib/librte_sft/rte_sft.h line 361
Quoted 1 line
lib/librte_sft/rte_sft.h new file +1
-0+*/
__rte_experimentalQuoted 9 lines
Quoted 9 lines
lib/librte_sft/rte_sft.h new file +9
-0+int +rte_sft_flow_set_ctx (uint32_t fid, +const struct rte_flow_item_sft *ctx, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* Quoted 8 lines
lib/librte_sft/rte_sft.h line 378
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+* Initialize SFT library instance. +* +* @param conf +* SFT library instance configuration. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. +*/
__rte_experimentalQuoted 8 lines
Quoted 8 lines
lib/librte_sft/rte_sft.h new file +8
-0+int +rte_sft_init (const struct rte_sft_conf *conf); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Finalize SFT library instance. Quoted 8 lines
lib/librte_sft/rte_sft.h line 391
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+* Cleanup & release allocated resources. +*/ +void +rte_sft_fini (void ); +I think we should use stop. It is not commons in DPDK to have fini functions. Maybe we should also add start function, so the app can init and then start the
SFT.
Quoted 9 lines
lib/librte_sft/rte_sft.h
Quoted 9 lines
lib/librte_sft/rte_sft.h new file +8
-0+/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue. +* +* Fragmentation handling (SFT fragmentation feature configured): +* If *mbuf_in* of fragmented packet received it will be stored by SFT library.
Quoted 4 lines
lib/librte_sft/rte_sft.h line 403
Quoted 4 lines
lib/librte_sft/rte_sft.h new file +4
-0+* status->fragmented bit will be set and *mbuf_out* will be set to NULL. +* On reception of all related fragments of IP packet it will be reassembled +* and further processed by this function on reception of last fragment. +* Quoted 15 lines
Does this function allocate a new mbuf? Does it releases all old mbufs?
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +4
-0+* Flow definition: +* SFT flow defined by one of its 7-tuples, since there is no zone value as +* argument flow should be defined by context attached to mbuf with action +* ``SFT`` (see RTE flow RTE_FLOW_ACTION_TYPE_SFT). Otherwise status- Quoted 1 line
defined
+ * field will be turned off & *mbuf_out* will be set to *mbuf_in*.
+ * In order to define flow for *mbuf_in* without attached sft context
+ *rte_sft_process_mbuf_with_zone()should be used with *zone* argument
+ * supplied by caller.
+ *
+ * Flow lookup:
+ * If SFT flow identifier can't be retrieved from SFT context attached to
+ * *mbuf_in* by action ``SFT`` - SFT lookup should be performmed by zone,performmed -> performed
Quoted 6 lines
Quoted 6 lines
+ * retrieved from SFT context attached to *mbuf_in*, and 5-tuple, extracted
+ * form mbuf outer header contents.
+ *
+ * Flow defined but does not exists:
+ * If flow not found in SFT inactivated flow will be created in SFT.
+ *status->activatedfield will be turned off & *mbuf_out* be set to*mbuf_in*.
Quoted 3 lines
Quoted 3 lines
+ * In order to activate created flow
rte_sft_flow_activate()should be used
+ * with reverse 7-tuple supplied by caller.
+ * This is first phase of flow creation in SFT for second phase & moredetailed
Quoted 1 line
Quoted 1 line
+ * descriotion of flow creation see
rte_sft_flow_activate.descriotion -> description
Quoted 2 lines
Quoted 2 lines
+ *
+ * Out of order (SFT out of oreder feature configured):oreder -> order
Quoted 1 line
Quoted 1 line
+ * If flow defined & activated but *mbuf_in* is TCP out of order packet it
will
Quoted 1 line
Quoted 1 line
+ * be stored by SFT library.
status->out_of_orderbit will be set &*mbuf_out*
Quoted 5 lines
Quoted 5 lines
+ * will be set to NULL. On reception of the first missing in order packet
+ *status->nb_in_order_mbufswill be set to number of mbufs that available for
+ * processing withrte_sft_drain_mbuf().
+ *Quoted 11 lines
It is possible that some packets will get trapped in the SFT do to this feature. if it supports ordering. For example the following case:
Packets arrive to the application. After draining the packets the Application changed the flow to full offload. This means that
all future packets will not arrive to the application.
But until the flow is offloaded some packets do arrive not in order. Then the flow is offloaded, this will result in the situation that no more packets will arrive to the application so some packets will get stack in the SFT.
I think we must have some force drain or, notify the SFT that no more packets should arrive to even if the packets are not in order it will releasethem.
Quoted 2 lines
Also the same with fragmented does this function allocate new mbufs? are
you releasing the
Quoted 17 lines
old ones?
Quoted 6 lines
+ * Flow defined & activated, mbuf not fragmented and 'in order':
+ * - Flow aging related data (see age field instruct rte_sft_flow_status)
+ * will be updated according to *mbuf_in* timestamp.
+ * - Flow connection tracking state (seect_statefield in
+ *struct rte_sft_flow_status) will be updated according to *mbuf_in* L4
+ * header contents.Quoted 8 lines
+ * - *mbuf_out* will be set to last processed mbuf.
+ *
+ * @param[in]mbuf_in
+ * mbuf to process; mbuf pinter considered 'consumed' and should not be
used
+ * after successful call to this function.
+ * @param[out]mbuf_out
+ * last processed not fragmented and in order mbuf.Quoted 1 line
If the in mbuf is not fragmented and in order, this pointer will point to the in
one?
Quoted 3 lines
Quoted 3 lines
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contentsof
Quoted 3 lines
Quoted 1 line
+ * *mbuf_in*.
Does the status bits for example fragmented is kept per connection or per
flow?
Quoted 18 lines
Since it is possible to get fragmented packets from both sides. The same goes for out of order packets
Quoted 14 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success:
+ * - *mbuf_out* contains valid mbuf pointer, locked SFT flow recognized by
+ *status->fid.
+ * - *mbuf_out* is NULL andstatus->fragmentedbit on in case of
+ * non last fragment *mbuf_in*.
+ * - *mbuf_out* is NULL andstatus->out_of_orderbit on in case of out of
+ * order *mbuf_in*, locked SFT flow recognized bystatus->fid.
+ * On failure a negative errno value andrte_errnois set.
+ */
__rte_experimentalQuoted 12 lines
Quoted 12 lines
lib/librte_sft/rte_sft.h new file +12
-0+int +rte_sft_process_mbuf (struct rte_mbuf *mbuf_in, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue while zone value provided by caller. +* Quoted 8 lines
Quoted 8 lines
+ * The behaviour of this function is similar to
rte_sft_process_mbufexcept
+ * the lookup in SFT procedure. The lookup in SFT always done by the *zone*
+ * arg and 5-tuple 5-tuple, extracted form mbuf outer header contents.
+ *
+ * @seerte_sft_process_mbuf
+ *
+ * @param[in]mbuf_in
+ * mbuf to process; mbuf pinter considered 'consumed' and should not beused
pinter -> pointer
Quoted 5 lines
Quoted 5 lines
+ * after successful call to this function.
+ * @param[out]mbuf_out
+ * last processed not fragmented and in order mbuf.
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contentsof
Quoted 6 lines
Quoted 6 lines
+ * *mbuf_in*.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @returnQuoted 8 lines
Quoted 8 lines
+ * 0 on success: + * - *mbuf_out* contains valid mbuf pointer. + * - *mbuf_out* is NULL and status->fragmented bit on in case of+ * non last fragment *mbuf_in*.
+ * - *mbuf_out* is NULL andstatus->out_of_orderbit on in case of out of
+ * order *mbuf_in*.
+ * On failure a negative errno value andrte_errnois set.
+ */
__rte_experimentalQuoted 7 lines
Quoted 7 lines
lib/librte_sft/rte_sft.h new file +7
-0+int +rte_sft_process_mbuf_with_zone (struct rte_mbuf *mbuf_in, +uint32_t zone, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); +Quoted 8 lines
Quoted 8 lines
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice.
+ *
+ * Drain next in order mbuf.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * This function behaves similar torte_sft_process_mbuf()but acts onpackets
Quoted 1 line
Quoted 1 line
+ * accumulated in SFT flow due to missing in order packet. Processing done
on
Quoted 2 lines
Quoted 2 lines
+ * single mbuf at a time and
in order. Other than above the behavior is
+ * same as ofrte_sft_process_mbuffor flow defined & activated & mbufisn't
Quoted 10 lines
Quoted 10 lines
+ * fragmented & 'in order'. This function should be called when
+ *rte_sft_process_mbuforrte_sft_process_mbuf_with_zonesets
+ *status->nb_in_order_mbufsoutput param !=0 and until
+ *status->nb_in_order_mbufs== 0.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * @param fid
+ * SFT flow ID.
+ * @param[out] status
+ * Structure to dump SFT flow status once updated according to contentsof
Quoted 8 lines
Quoted 8 lines
+ * *mbuf_in*.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid mbuf in case of success, NULL otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 22 lines
lib/librte_sft/rte_sft.h line 543
Quoted 5 lines
lib/librte_sft/rte_sft.h new file +5
-0+struct rte_mbuf * +rte_sft_drain_mbuf (uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); +Fid represent a connection, so which direction do we drain the packets? since we can have inordered packet in from both directions right?
Quoted 13 lines
+/**
+ * @warning
+ * @b EXPERIMENTAL: this API may change without prior notice.
+ *
+ * Activate flow in SFT.
+ * Flow should be locked by caller (seerte_sft_flow_lock).
+ *
+ * This function performs second phase of flow creation in SFT.
+ * The reasons for 2 phase flow creation procedure:
+ * 1. Missing reverse flow - flow context is shared for both flow directions+ * i.e. in order maintain bidirectional flow context in RTE SFT packets + * arriving from both dirrections should be identified as packets of the + * RTE SFT flow. Consequently before creation of the SFT flow callershould
Quoted 16 lines
Quoted 16 lines
+ * provide reverse flow direction 7-tuple.
+ * 2. The caller of rte_sft_process_mbuf/rte_sft_process_mbuf_with_zone should+ * be notified that arrived mbuf is first in flow & decide weather to + * create new flow or it distroy before it was activated with + * rte_sft_flow_destroy.+ * This function completes creation of the bidirectional SFT flow & creates
+ * entry for 7-tuple on SFT PMD defined by the tuple port for both
+ * initiator/initiate 7-tuples.
+ * Flow aging, connection tracking state & out of order handling will be
+ * initialized according to the content of the *mbuf_in* passes to
+ * rte_sft_process_mbuf/_with_zone during the phase 1 of flow creation.
+ * Once this function returns upcoming calls rte_sft_process_mbuf/_with_zone
+ * with 7-tuple or its reverse will return handle to this flow.
+ * Flow should be locked by caller (seerte_sft_flow_lock).Quoted 8 lines
Quoted 8 lines
+ *
+ * @param fid
+ * SFT flow ID.
+ * @paramreverse_tuple
+ * Expected response flow 7-tuple.
+ * @param ctx
+ * User defined state to set.
+ * Update of *fid* or *zone* fields in structrte_flow_item_sftunsupported.
Quoted 2 lines
Quoted 2 lines
+ * @param
ct_enable
+ * Enables maintenance ofstatus->ct_stateconnection tracking value forthe
Quoted 6 lines
Quoted 6 lines
+ * flow; otherwise
status->ct_statewill be initialized with zeros.
+ * @paramevdev_id
+ * Event dev ID to enqueue end of flow event.
+ * @paramevport_id
+ * Event port ID to enqueue end of flow event.
+ * @param[out] statusQuoted 8 lines
Quoted 8 lines
+ * Structure to dump SFT flow status once activated.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 17 lines
Quoted 17 lines
lib/librte_sft/rte_sft.h new file +17
-0+int +rte_sft_flow_activate (uint32_t fid, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +uint8_t dev_id, +uint8_t port_id, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Artificially create SFT flow. +* +* Function to create SFT flow before reception of the first flow packet. Quoted 8 lines
Quoted 8 lines
+ *
+ * @param tuple
+ * Expected initiator flow 7-tuple.
+ * @paramreverse_tuple
+ * Expected initiate flow 7-tuple.
+ * @param ctx
+ * User defined state to set.
+ * Setting of *fid* or *zone* fields in structrte_flow_item_sftunsupported.
Quoted 2 lines
Quoted 2 lines
+ * @param[out]
ct_enable
+ * Enables maintenance ofstatus->ct_stateconnection tracking value forthe
Quoted 4 lines
Quoted 4 lines
+ * flow; otherwise
status->ct_statewill be initialized with zeros.
+ * @param[out] status
+ * Structure to dump SFT flow status once created.
+ * @param[out] errorQuoted 8 lines
Quoted 8 lines
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * - on success: 0, locked SFT flow recognized bystatus->fid.
+ * - on error: a negative errno value otherwise andrte_errnois set.
+ */
+No extra empty line and
__rte_experimentalQuoted 14 lines
Quoted 14 lines
lib/librte_sft/rte_sft.h new file +14
-0+int +rte_sft_flow_create (const struct rte_sft_7tuple *tuple, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Lock exclusively SFT flow. +* Quoted 8 lines
Quoted 8 lines
+ * Explicit flow locking; used for handling aged flows.
+ *
+ * @param fid
+ * SFT flow ID.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 11 lines
Quoted 11 lines
lib/librte_sft/rte_sft.h new file +11
-0+int +rte_sft_flow_lock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Release exclusively locked SFT flow. +* +* When rte_sft_process_mbuf/_with_zone and rte_sft_flow_create +* return *status* containing fid with defined bit on the flow considered Quoted 8 lines
Quoted 8 lines
+ * exclusively locked and should be unlocked with this function.
+ *
+ * @param fid
+ * SFT flow ID.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 31 lines
Quoted 31 lines
lib/librte_sft/rte_sft.h new file +31
-0+int +rte_sft_flow_unlock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Removes flow from SFT. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* - Flow should be locked by caller in order to remove it. +* - Flow should have no client objects attached. +* +* Should be applied on aged flows, when flow aged event received. +* +* @code{.c} +* while (1) { +* rte_event_dequeue_burst(...); +* FOR_EACH_EV(ev) { +* uint32_t fid = ev.u64; +* rte_sft_flow_lock(fid); +* FOR_EACH_CLIENT(fid, client_id) { +* rte_sft_flow_reset_client_obj(fid, client_obj); +* // detached client object handling +* } +* rte_sft_flow_destroy(fid, &error); +* } +* } +* @endcode +* +* @param fid Quoted 8 lines
Quoted 8 lines
+ * SFT flow ID to destroy.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 13 lines
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +13
-0+int +rte_sft_flow_destroy (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Reset flow age to zero. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Simulates last flow packet with timestamp set to just now. +* +* @param fid Quoted 8 lines
Quoted 8 lines
+ * SFT flow ID.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 13 lines
Quoted 13 lines
lib/librte_sft/rte_sft.h new file +13
-0+int +rte_sft_flow_touch (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set flow aging to specific value. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param aging Quoted 8 lines
Quoted 8 lines
+ * New flow aging value.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 17 lines
Quoted 17 lines
lib/librte_sft/rte_sft.h new file +17
-0+int +rte_sft_flow_set_aging (uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to set object for. +* @param client_obj Quoted 8 lines
Quoted 8 lines
+ * Pointer to opaque client object structure.
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * 0 on success, a negative errno value otherwise andrte_errnois set.
+ */
__rte_experimentalQuoted 17 lines
Quoted 17 lines
lib/librte_sft/rte_sft.h new file +17
-0+int +rte_sft_flow_set_client_obj (uint32_t fid, +uint8_t client_id, +void *client_obj, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to get object for. Quoted 8 lines
Quoted 8 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid client object opaque pointer in case of success, NULL otherwise
+ * andrte_errnois set.
+ */
__rte_experimentalQuoted 10 lines
lib/librte_sft/rte_sft.h line 815
Quoted 10 lines
lib/librte_sft/rte_sft.h new file +10
-0+void * +rte_sft_flow_get_client_obj (const uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Remove client object for given client ID. Function name uses "reset", but description says "remove". May be synchronize it?
Quoted 5 lines
Quoted 5 lines
+ * Flow should be locked by caller (see
rte_sft_flow_lock).
+ *
+ * Detaches client object from SFT flow and returns the ownership for the
+ * client object to the caller by returning client object pointer value.
+ * The pointer returned by this function won't be accessed any more, thecaller
Quoted 7 lines
Quoted 7 lines
+ * may release all client obj related resources & the memory allocated for
+ * this client object.
+ *
+ * @param fid
+ * SFT flow ID.
+ * @paramclient_id
+ * Client ID to remove object for.Quoted 8 lines
Quoted 8 lines
+ * @param[out] error
+ * Perform verbose error reporting if not NULL. PMDs initialize this
+ * structure in case of error only.
+ *
+ * @return
+ * A valid client object opaque pointer in case of success, NULL otherwise
+ * andrte_errnois set.
+ */
__rte_experimentalQuoted 48 lines
Quoted 48 lines
lib/librte_sft/rte_sft.h new file +10
-0+void * +rte_sft_flow_reset_client_obj (uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_H_ */ lib/librte_sft/rte_sft_driver.h new file +32
-0@@ -0,0 +1,195 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_DRIVER_H_ +#define _RTE_SFT_DRIVER_H_ + +/** + * @file + * + * RTE SFT Ethernet device PMD API + * + * APIs that are used by the SFT library to offload SFT operationons + * to Ethernet device. + */ + +#include "rte_sft.h" + +#ifdef __cplusplus +extern "C" { +#endif + +/** + * Opaque type returned after successfully creating an entry in SFT. + * + * This handle can be used to manage and query the related entry (e.g. to + * destroy it or update age). + */ +struct rte_sft_entry; + +/** + * Create SFT entry in eth_dev SFT.Quoted 8 lines
lib/librte_sft/rte_sft_driver.h
Quoted 8 lines
lib/librte_sft/rte_sft_driver.h new file +8
-0+* +* @param dev +* Pointer to Ethernet device structure. +* @param tuple +* L3/L4 5-tuple - src/dest IP and port and IP protocol. +* @param nat_tuple +* L3/L4 5-tuple to replace in packet original 5-tuple in order to implement +* NAT offloading; if NULL NAT offloading won't be configured for the flow.
Quoted 14 lines
Quoted 14 lines
lib/librte_sft/rte_sft_driver.h new file +14
-0+* @param aging +* Flow aging timeout in seconds. +* @param ctx +* Initial values in SFT flow context +* (see RTE flow struct rte_flow_item_sft). +* ctx->zone should be valid. +* @param fid +* SFT flow ID for the entry to create on *device*. +* If there is an entry for the *fid* in PMD it will be updated with the +* values of *ctx*. +* @param[out] queue_index +* if PMD can figure out the queue where the flow packets will +* arrive in RX data path it will set the value of queue_index; otherwise +* all bits will be turned on. Quoted 8 lines
lib/librte_sft/rte_sft_driver.h
Quoted 8 lines
lib/librte_sft/rte_sft_driver.h new file +8
-0+* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid handle in case of success, NULL otherwise and rte_errno is set. +*/ +typedef struct rte_sft_entry *(*sft_entry_create_t) (struct rte_eth_dev*dev,
Quoted 12 lines
lib/librte_sft/rte_sft_driver.h line 70
Quoted 8 lines
lib/librte_sft/rte_sft_driver.h new file +8
-0+const struct rte_sft_5tuple *tuple, + const struct rte_sft_5tuple *nat_tuple, + const uint32_t aging, + const struct rte_flow_item_sft *ctx, + const uint32_t fid, + uint16_t *queue_index, + struct rte_sft_error *error); +I think for easier reading, the API should change to have 6 tuple (5 + zone) the ctx should be removed and replaced with the state.
Quoted 3 lines
Then add new API to modify the ctx
typedef int (*sft_modify_state)(structrte_eth_dev*dev, uint8 state); The main issue we my suggestion is that it will force the PMD to store theinformation to recreate
Quoted 3 lines
the rule, data that is already saved by the SFT.
Also I don't see why we need queue index, since the RSS and queue will be
configured by the RTE flow
Quoted 26 lines
in a different group.
Quoted 24 lines
lib/librte_sft/rte_sft_driver.h new file +24
-0+/** +* Destroy SFT entry in eth_dev SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param entry +* Handle to the SFT entry to destroy. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_destroy_t )(struct rte_eth_dev *dev, +struct rte_sft_entry *entry, +struct rte_sft_error *error); + +/** +* Decodes SFT flow context if attached to mbuf by action ``SFT``. +* @see RTE flow RTE_FLOW_ACTION_TYPE_SFT. +* +* @param dev +* Pointer to Ethernet device structure. Quoted 8 lines
lib/librte_sft/rte_sft_driver.h
Quoted 8 lines
lib/librte_sft/rte_sft_driver.h new file +8
-0+* @param mbuf +* mbuf of the packet to decode attached state from. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid SFT flow context in case of success, NULL otherwise and
rte_errnoQuoted 117 lines
lib/librte_sft/rte_sft_driver.h line 109
Quoted 7 lines
lib/librte_sft/rte_sft_driver.h new file +7
-0+* is set. +*/ +typedef struct rte_flow_item_sft *(*sft_entry_mbuf_decode_ctx_t )( +struct rte_eth_dev *dev, +const struct rte_mbuf *mbuf, +struct rte_sft_error *error); +What about returning int as error code, and return the
rte_flow_item_sftas out parameter?
This will remove the allocation and free.Quoted 105 lines
lib/librte_sft/rte_sft_driver.h new file +86
-0+/** +* Get aged-out SFT entries. +* +* Report entry as aged-out if timeout passed without any matching +* on the SFT entry. +* +* @param[in] dev +* Pointer to Ethernet device structure. +* @param[in, out] fid_aged +* The address of an array of aged-out SFT flow IDs. +* @param[in] nb_aged +* The length of *fid_aged* array pointers. +* @param[out] error +* Perform verbose error reporting if not NULL. Initialized in case of +* error only. +* +* @return +* if nb_aged is 0, return the amount of all aged flows. +* if nb_aged is not 0 , return the amount of aged flows reported +* in the *fid_aged* array, otherwise negative errno value. +*/ +typedef int (*sft_entry_get_aged_entries_t )(struct rte_eth_dev *dev, +uint32_t *fid_aged, +int nb_aged, +struct rte_sft_error *error); + +/** +* Simulate SFT entry match in terms of entry aging. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_touch_t )(struct rte_eth_dev *dev, +uint32_t fid, +struct rte_sft_error *error); + +/** +* Set SFT entry aging to specific value. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param aging +* New entry aging value. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_set_aging_t )(struct rte_eth_dev *dev, +uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** SFT operations function pointer table */ +struct rte_sft_ops { +sft_entry_create_t entry_create; +/**< Create SFT entry in eth_dev SFT. */ +sft_entry_destroy_t entry_destroy; +/**< Destroy SFT entry in eth_dev SFT. */ +sft_entry_mbuf_decode_ctx_t mbuf_decode_ctx; +/**< Decodes SFT flow context if attached to mbuf by action ``SFT``. */ +sft_entry_get_aged_entries_t get_aged_entries; +/**< Get aged-out SFT entries. */ +sft_entry_touch_t entry_touch; +/**< Simulate SFT entry match in terms of entry aging. */ +sft_entry_set_aging_t set_aging; +/**< Set SFT entry aging to specific value. */ +}; + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_DRIVER_H_ */ lib/librte_sft/rte_sft_version.map new file +12
-0@@ -0,0 +1,21 @@ +EXPERIMENTAL { + global: + + rte_sft_flow_get_status; + rte_sft_flow_set_ctx; + rte_sft_init; + rte_sft_fini; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + rte_sft_drain_mbuf; + rte_sft_flow_activate; + rte_sft_flow_create;Quoted 8 lines
lib/librte_sft/rte_sft_version.map line 20
Quoted 8 lines
lib/librte_sft/rte_sft_version.map new file +8
-0+ rte_sft_flow_lock; + rte_sft_flow_unlock; + rte_sft_flow_destroy; + rte_sft_flow_touch; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_get_client_obj; + rte_sft_flow_reset_client_obj;If I'm not mistaken, it should be alphabetially sorted.
Quoted 7 lines
Quoted 3 lines
lib/librte_sft/rte_sft_version.map new file +1
-1+}; --2.26.2
Best,
Ori
Re: [dpdk-dev] [RFC 3/3] sft: introduce API
In reply to Andrey VesnovatyQuoted 1,162 lines · Andrey Vesnovaty
Defines RTE SFT APIs for Statefull Flow Table library.
SFT General description:
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.Signed-off-by Andrey Vesnovaty <[email protected]>--- lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 +++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 6 files changed, 1105 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.maplib/librte_sft/Makefile new file +28
-0@@ -0,0 +1,28 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +include $(RTE_SDK)/mk/rte.vars.mk + +# library name +LIB = librte_sft.a + +# library version +LIBABIVER := 1 + +# build flags +CFLAGS += -O3 +CFLAGS += $(WERROR_FLAGS) +LDLIBS += -lrte_eal -lrte_mbuf + +# library source files +# all source are stored in SRCS-y +SRCS-$(CONFIG_RTE_LIBRTE_REGEXDEV) := rte_sft.c + +# export include files +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft.h +SYMLINK-$(CONFIG_RTE_LIBRTE_REGEXDEV)-include += rte_sft_driver.h + +# versioning export map +EXPORT_MAP := rte_sft_version.map + +include $(RTE_SDK)/mk/rte.lib.mklib/librte_sft/meson.build new file +7
-0@@ -0,0 +1,7 @@ +# SPDX-License-Identifier: BSD-3-Clause +# Copyright 2020 Mellanox Technologies, Ltd + +sources = files('rte_sft.c') +headers = files('rte_sft.h', + 'rte_sft_driver.h') +deps += ['mbuf']lib/librte_sft/rte_sft.c new file +9
-0@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */ lib/librte_sft/rte_sft.h new file +845
-0@@ -0,0 +1,845 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_H_ +#define _RTE_SFT_H_ + +/** +* @file +* +* RTE SFT API +* +* Defines RTE SFT APIs for Statefull Flow Table library. +* +* SFT General description: +* SFT library provides a framework for applications that need to maintain +* context across different packets of the connection. +* Examples for such applications: +* - Next-generation firewalls +* - Intrusion detection/prevention systems (IDS/IPS): Suricata, Snort +* - SW/Virtual Switching: OVS +* The goals of the SFT library: +* - Accelerate flow recognition & its context retrieval for further lookaside +* processing. +* - Enable context-aware flow handling offload. +* +* Definitions and Abbreviations: +* - 5-tuple: defined by: +* -- Source IP address +* -- Source port +* -- Destination IP address +* -- Destination port +* -- IP protocol number +* - 7-tuple: 5-tuple zone and port (see struct rte_sft_7tuple) +* - 5/7-tuple: 5/7-tuple of the packet from connection initiator +* - revers 5/7-tuple: 5/7-tuple of the packet from connection initiate +* - application: SFT library API consumer +* - APP: see application +* - CID: client ID +* - CT: connection tracking +* - FID: Flow identifier +* - FIF: First In Flow +* - Flow: defined by 7-tuple and its reverse i.e. flow is bidirectional +* - SFT: Stateful Flow Table +* - user: see application +* - zone: additional user defined value used as differentiator for +* connections having same 5-tuple (for example different VxLan +* connections with same inner 5-tuple). +* +* SFT components: +* +* +-----------------------------------+ +* | RTE flow | +* | | +* | +-------------------------------+ | +----------------+ +* | | group X | | | RTE_SFT | +* | | | | | | +* | | +---------------------------+ | | | | +* | | | rule ... | | | | | +* | | | . | | | +-----------+----+ +* | | | . | | | | +* | | | . | | | entry +* | | +---------------------------+ | | create +* | | | rule | | | | +* | | | patterns ... +---------+ | +* | | | actions | | | | | +* | | | SFT (zone=Z) | | | | | +* | | | JUMP (group=Y) | | | lookup | +* | | +---------------------------+ | | zone=Z, | +* | | | rule ... | | | 5tuple | +* | | | . | | | | | +* | | | . | | | +--v-------------+ +* | | | . | | | | SFT | | +* | | | | | | | | | +* | | +---------------------------+ | | | +--v--+ | +* | | | | | | | | +* | +-------------------------------+ | | | PMD | | +* | | | | | | +* | | | +-----+ | +* | +-------------------------------+ | | | +* | | group Y | | | | +* | | | | | set flow CTX | +* | | +---------------------------+ | | | | +* | | | rule | | | +--------+-------+ +* | | | patterns | | | | +* | | | SFT (state=UNDEFINED) | | | | +* | | | actions RSS | | | | +* | | +---------------------------+ | | | +* | | | rule | | | | +* | | | patterns | | | | +* | | | SFT (state=INVALID) | <-------------+ +* | | | actions DROP | | | forward +* | | +---------------------------+ | | group=Y +* | | | rule | | | +* | | | patterns | | | +* | | | SFT (state=ACCEPTED) | | | +* | | | actions PORT | | | +* | | +---------------------------+ | | +* | | ... | | +* | | | | +* | +-------------------------------+ | +* | ... | +* | | +* +-----------------------------------+ +* +* SFT as datastructure: +* SFT can be treated as datastructure maintaining flow context across its +* lifetime. SFT flow entry represent bidirectional network flow and defined by +* 7-tuple & its reverse 7-tuple. +* Each entry in SFT has: +* - FID: 1:1 mapped & used as entry handle & encapsulating internal +* implementation of the entry. +* - State: user-defined value attached to each entry, the only library +* reserved value for state unset (the actual value defined by SFT +* configuration). The application should define flow state encodings and +* set it for flow via rte_sft_flow_set_ctx() than what actions should be +* applied on packets can be defined via related RTE flow rule matching SFT +* state (see rules in SFT components diagram above). +* - Timestamp: for the last seen in flow packet used for flow aging mechanism +* implementation. +* - Client Objects: user-defined flow contexts attached as opaques to flow. +* - Acceleration & offloading - utilize RTE flow capabilities, when supported +* (see action ``SFT``), for flow lookup acceleration and further +* context-aware flow handling offload. +* - CT state: optionally for TCP connections CT state can be maintained +* (see enum rte_sft_flow_ct_state). +* - Out of order TCP packets: optionally SFT can keep out of order TCP +* packets aside the flow context till the arrival of the missing in-order +* packet. +* +* RTE flow changes: +* The SFT flow state (or context) for RTE flow is defined by fields of +* struct rte_flow_item_sft. +* To utilize SFT capabilities new item and action types introduced: +* - item SFT: matching on SFT flow state (see RTE_FLOW_ITEM_TYPE_SFT). +* - action SFT: retrieve SFT flow context and attache it to the processed +* packet (see RTE_FLOW_ACTION_TYPE_SFT). +* +* The contents of per port SFT serving RTE flow action ``SFT`` managed via +* SFT PMD APIs (see struct rte_sft_ops). +* The SFT flow state/context retrieval performed by user-defined zone ``SFT`` +* action argument and processed packet 5-tuple. +* If in scope of action ``SFT`` there is no context/state for the flow in SFT +* undefined sate attached to the packet meaning that the flow is not +* recognized by SFT, most probably FIF packet. +* +* Once the SFT state set for a packet it can match on item SFT +* (see RTE_FLOW_ITEM_TYPE_SFT) and forwarding design can be done for the +* packet, for example: +* - if state value == x than queue for further processing by the application +* - if state value == y than forward it to eth port (full offload) +* - if state value == 'undefined' than queue for further processing by +* the application (handle FIF packets) +* +* Processing packets with SFT library: +* +* FIF packet: +* To recognize upcoming packets of the SFT flow every FIF packet should be +* forwarded to the application utilizing the SFT library. Non-FIF packets can +* be processed by the application or its processing can be fully offloaded. +* Processing of the packets in SFT library starts with rte_sft_process_mbuf +* or rte_sft_process_mbuf_with_zone. If mbuf recognized as FIF application +* should make a design to destroy flow or complete flow creation process in +* SFT using rte_sft_flow_activate. +* +* Recognized SFT flow: +* Once struct rte_sft_flow_status with valid fid field posesed by application +* it can: +* - mange client objects on it (see client_obj field in +* struct rte_sft_flow_status) using rte_sft_flow_<OP>_client_obj APIs +* - analyze user-defined flow state and CT state (see state & ct_sate fields +* in struct rte_sft_flow_status). +* - set flow state to be attached to the upcoming packets by action ``SFT`` +* via struct rte_sft_flow_status API. +* - decide to destroy flow via rte_sft_flow_destroy API. +* +* Flow aging: +* +* SFT library manages the aging for each flow. On flow creation, it's +* assigned an aging value, the maximal number of seconds passed since the +* last flow packet arrived, once exceeded flow considered aged. +* The application notified of aged flow asynchronously via event queues. +* The device and port IDs tuple to identify the event queue to enqueue +* flow aged events passed on flow creation as arguments +* (see rte_sft_flow_activate). It's the application responsibility to +* initialize event queues and assign them to each flow for EOF event +* notifications. +* Aged EOF event handling: +* - Should be considered as application responsibility. +* - The last stage should be the release of the flow resources via +* rte_sft_flow_destroy API. +* - All client objects should be removed from flow before the +* rte_sft_flow_destroy API call. +* See the description of rete_sft_flow_destroy for an example of aged flow +* handling. +* +* SFT API thread safety: +* +* SFT library APIs are thread-safe while handling of specific flow can be +* done in a single thread simultaneously. Exclusive access to specific SFT +* flow guaranteed by: +* - rte_sft_process_mbuf +* - rte_sft_process_mbuf_with_zone +* - rte_sft_flow_create +* - rte_sft_flow_lock +* When application is done with the flow handling for the current packet it +* should call rte_sft_flow_unlock API to maintain exclusive access to the +* flow with other threads. +* +* SFT Library initialization and cleanup: +* +* SFT library should be considered as a single instance, preconfigured and +* initialized via rte_sft_init() API. +* SFT library resource deallocation and cleanup should be done via +* rte_sft_init() API as a stage of the application termination procedure. +*/ + +#ifdef __cplusplus +extern "C" { +#endif + +#include <rte_common.h> +#include <rte_config.h> +#include <rte_errno.h> +#include <rte_mbuf.h> +#include <rte_ethdev.h> +#include <rte_flow.h> + +/** +* L3/L4 5-tuple - src/dest IP and port and IP protocol. +* +* Used for flow/connection identification. +*/ +struct rte_sft_5tuple { +union { +struct { +rte_be32_t src_addr;/**< IPv4 source address. */ +rte_be32_t dst_addr;/**< IPv4 destination address. */ + } ipv4; +struct { +uint8_t src_addr[16 ];/**< IPv6 source address. */ +uint8_t dst_addr[16 ];/**< IPv6 destination address. */ + } ipv6; + }; +uint16_t src_port;/**< Source port. */ +uint16_t dst_port;/**< Destination port. */ +uint8_t proto;/**< IP protocol. */ +uint8_t is_ipv6:1 ;/**< True for valid IPv6 fields. Otherwise IPv4. */ +}; + +/** +* Port flow identification. +* +* @p zone used for setups where 5-tuple is not enough to identify flow. +* For example different VLANs/VXLANs may have similar 5-tuples. +*/ +struct rte_sft_7tuple { +struct rte_sft_5tuple flow_5tuple;/**< L3/L4 5-tuple. */ +uint32_t zone;/**< Zone assigned to flow. */ +uint16_t port_id;/** <Port identifier of Ethernet device. */ +}; + +/** +* Flow connection tracking states +*/ +enum rte_sft_flow_ct_state { +RTE_SFT_FLOW_CT_STATE_NEW = (1 <<0 ), +RTE_SFT_FLOW_CT_STATE_EST = (1 <<1 ), +RTE_SFT_FLOW_CT_STATE_REL = (1 <<2 ), +RTE_SFT_FLOW_CT_STATE_RPL = (1 <<3 ), +RTE_SFT_FLOW_CT_STATE_INV = (1 <<4 ), +RTE_SFT_FLOW_CT_STATE_TRK = (1 <<5 ), +RTE_SFT_FLOW_CT_STATE_SNAT = (1 <<6 ), +RTE_SFT_FLOW_CT_STATE_DNAT = (1 <<7 ), +}; + +/** +* Structure describes SFT library configuration +*/ +struct rte_sft_conf { +uint32_t UDP_aging;/**< UDP proto default aging. */ +uint32_t TCP_aging;/**< TCP proto default aging. */ +uint32_t TCP_SYN_aging;/**< TCP SYN default aging. */ +uint32_t OTHER_aging;/**< All unlisted proto default aging. */ +uint32_t size;/**< Max entries in SFT. */ +uint8_t undefined_state;/**< Undefined state constant. */ +uint8_t reorder_enable:1 ; +/**< TCP packet reordering feature enabled bit. */ +uint8_t ct_enable:1 ;/**< Connection tracking feature enabled bit. */ +}; + +/** +* Structure describes the state of the flow in SFT. +*/ +struct rte_sft_flow_status { +uint32_t fid;/**< SFT flow id. */ +uint32_t zone;/**< Zone for lookup in SFT */ +uint8_t state;/**< Application defined bidirectional flow state. */ +uint8_t ct_state;/**< Connection tracking flow state. */ +uint32_t age;/**< Seconds passed since last flown packet. */ +uint32_t aging; +/**< Flow considered aged once this age (seconds) reached. */ +uint32_t nb_in_order_mbufs; +/**< Number of in-order mbufs available for drain */ +void **client_obj;/**< Array of clients attached to flow. */ +int nb_clients;/**< Number of clients attached to flow. */ +uint8_t defined:1 ;/**< Flow defined in SFT bit. */ +uint8_t activated:1 ;/**< Flow activation bit. */ +uint8_t fragmented:1 ;/**< Last flow mbuf was fragmented. */ +uint8_t out_of_order:1 ;/**< Last flow mbuf was out of order (TCP). */ +}; + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get SFT flow status. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump actual SFT flow status. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_get_status (const uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set user defined context. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Updates per ethernet dev SFT entries: +* - flow lookup acceleration +* - partial/full flow offloading managed by flow context +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param ctx +* User defined state to set. +* Update of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_set_ctx (uint32_t fid, +const struct rte_flow_item_sft *ctx, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Initialize SFT library instance. +* +* @param conf +* SFT library instance configuration. +* +* @return +* 0 on success , a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_init (const struct rte_sft_conf *conf); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Finalize SFT library instance. +* Cleanup & release allocated resources. +*/ +void +rte_sft_fini (void ); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue. +* +* Fragmentation handling (SFT fragmentation feature configured): +* If *mbuf_in* of fragmented packet received it will be stored by SFT library. +* status->fragmented bit will be set and *mbuf_out* will be set to NULL. +* On reception of all related fragments of IP packet it will be reassembled +* and further processed by this function on reception of last fragment. +* +* Flow definition: +* SFT flow defined by one of its 7-tuples, since there is no zone value as +* argument flow should be defined by context attached to mbuf with action +* ``SFT`` (see RTE flow RTE_FLOW_ACTION_TYPE_SFT). Otherwise status->defined +* field will be turned off & *mbuf_out* will be set to *mbuf_in*. +* In order to define flow for *mbuf_in* without attached sft context +* rte_sft_process_mbuf_with_zone() should be used with *zone* argument +* supplied by caller. +* +* Flow lookup: +* If SFT flow identifier can't be retrieved from SFT context attached to +* *mbuf_in* by action ``SFT`` - SFT lookup should be performmed by zone, +* retrieved from SFT context attached to *mbuf_in*, and 5-tuple, extracted +* form mbuf outer header contents. +* +* Flow defined but does not exists: +* If flow not found in SFT inactivated flow will be created in SFT. +* status->activated field will be turned off & *mbuf_out* be set to *mbuf_in*. +* In order to activate created flow rte_sft_flow_activate() should be used +* with reverse 7-tuple supplied by caller. +* This is first phase of flow creation in SFT for second phase & more detailed +* descriotion of flow creation see rte_sft_flow_activate. +* +* Out of order (SFT out of oreder feature configured): +* If flow defined & activated but *mbuf_in* is TCP out of order packet it will +* be stored by SFT library. status->out_of_order bit will be set & *mbuf_out* +* will be set to NULL. On reception of the first missing in order packet +* status->nb_in_order_mbufs will be set to number of mbufs that available for +* processing with rte_sft_drain_mbuf(). +* +* Flow defined & activated, mbuf not fragmented and 'in order': +* - Flow aging related data (see age field in `struct rte_sft_flow_status`) +* will be updated according to *mbuf_in* timestamp. +* - Flow connection tracking state (see ct_state field in +* `struct rte_sft_flow_status`) will be updated according to *mbuf_in* L4 +* header contents. +* - *mbuf_out* will be set to last processed mbuf. +* +* @param[in] mbuf_in +* mbuf to process; mbuf pinter considered 'consumed' and should not be used +* after successful call to this function. +* @param[out] mbuf_out +* last processed not fragmented and in order mbuf. +* @param[out] status +* Structure to dump SFT flow status once updated according to contents of +* *mbuf_in*. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success: +* - *mbuf_out* contains valid mbuf pointer, locked SFT flow recognized by +* status->fid. +* - *mbuf_out* is NULL and status->fragmented bit on in case of +* non last fragment *mbuf_in*. +* - *mbuf_out* is NULL and status->out_of_order bit on in case of out of +* order *mbuf_in*, locked SFT flow recognized by status->fid. +* On failure a negative errno value and rte_errno is set. +*/ +int +rte_sft_process_mbuf (struct rte_mbuf *mbuf_in, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Process mbuf received on RX queue while zone value provided by caller. +* +* The behaviour of this function is similar to rte_sft_process_mbuf except +* the lookup in SFT procedure. The lookup in SFT always done by the *zone* +* arg and 5-tuple 5-tuple, extracted form mbuf outer header contents. +* +* @see rte_sft_process_mbuf +* +* @param[in] mbuf_in +* mbuf to process; mbuf pinter considered 'consumed' and should not be used +* after successful call to this function. +* @param[out] mbuf_out +* last processed not fragmented and in order mbuf. +* @param[out] status +* Structure to dump SFT flow status once updated according to contents of +* *mbuf_in*. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success: +* - *mbuf_out* contains valid mbuf pointer. +* - *mbuf_out* is NULL and status->fragmented bit on in case of +* non last fragment *mbuf_in*. +* - *mbuf_out* is NULL and status->out_of_order bit on in case of out of +* order *mbuf_in*. +* On failure a negative errno value and rte_errno is set. +*/ +int +rte_sft_process_mbuf_with_zone (struct rte_mbuf *mbuf_in, +uint32_t zone, +struct rte_mbuf **mbuf_out, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Drain next in order mbuf. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* This function behaves similar to rte_sft_process_mbuf() but acts on packets +* accumulated in SFT flow due to missing in order packet. Processing done on +* single mbuf at a time and `in order`. Other than above the behavior is +* same as of rte_sft_process_mbuf for flow defined & activated & mbuf isn't +* fragmented & 'in order'. This function should be called when +* rte_sft_process_mbuf or rte_sft_process_mbuf_with_zone sets +* status->nb_in_order_mbufs output param !=0 and until +* status->nb_in_order_mbufs == 0. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param[out] status +* Structure to dump SFT flow status once updated according to contents of +* *mbuf_in*. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid mbuf in case of success, NULL otherwise and rte_errno is set. +*/ +struct rte_mbuf * +rte_sft_drain_mbuf (uint32_t fid, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Activate flow in SFT. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* This function performs second phase of flow creation in SFT. +* The reasons for 2 phase flow creation procedure: +* 1. Missing reverse flow - flow context is shared for both flow directions +* i.e. in order maintain bidirectional flow context in RTE SFT packets +* arriving from both dirrections should be identified as packets of the +* RTE SFT flow. Consequently before creation of the SFT flow caller should +* provide reverse flow direction 7-tuple. +* 2. The caller of rte_sft_process_mbuf/rte_sft_process_mbuf_with_zone should +* be notified that arrived mbuf is first in flow & decide weather to +* create new flow or it distroy before it was activated with +* rte_sft_flow_destroy. +* This function completes creation of the bidirectional SFT flow & creates +* entry for 7-tuple on SFT PMD defined by the tuple port for both +* initiator/initiate 7-tuples. +* Flow aging, connection tracking state & out of order handling will be +* initialized according to the content of the *mbuf_in* passes to +* rte_sft_process_mbuf/_with_zone during the phase 1 of flow creation. +* Once this function returns upcoming calls rte_sft_process_mbuf/_with_zone +* with 7-tuple or its reverse will return handle to this flow. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param reverse_tuple +* Expected response flow 7-tuple. +* @param ctx +* User defined state to set. +* Update of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param ct_enable +* Enables maintenance of status->ct_state connection tracking value for the +* flow; otherwise status->ct_state will be initialized with zeros. +* @param evdev_id +* Event dev ID to enqueue end of flow event. +* @param evport_id +* Event port ID to enqueue end of flow event. +* @param[out] status +* Structure to dump SFT flow status once activated. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_activate (uint32_t fid, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +uint8_t dev_id, +uint8_t port_id, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Artificially create SFT flow. +* +* Function to create SFT flow before reception of the first flow packet. +* +* @param tuple +* Expected initiator flow 7-tuple. +* @param reverse_tuple +* Expected initiate flow 7-tuple. +* @param ctx +* User defined state to set. +* Setting of *fid* or *zone* fields in struct rte_flow_item_sft unsupported. +* @param[out] ct_enable +* Enables maintenance of status->ct_state connection tracking value for the +* flow; otherwise status->ct_state will be initialized with zeros. +* @param[out] status +* Structure to dump SFT flow status once created. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* - on success: 0, locked SFT flow recognized by status->fid. +* - on error: a negative errno value otherwise and rte_errno is set. +*/ + +int +rte_sft_flow_create (const struct rte_sft_7tuple *tuple, +const struct rte_sft_7tuple *reverse_tuple, +const struct rte_flow_item_sft *ctx, +uint8_t ct_enable, +struct rte_sft_flow_status *status, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Lock exclusively SFT flow. +* +* Explicit flow locking; used for handling aged flows. +* +* @param fid +* SFT flow ID. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_lock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Release exclusively locked SFT flow. +* +* When rte_sft_process_mbuf/_with_zone and rte_sft_flow_create +* return *status* containing fid with defined bit on the flow considered +* exclusively locked and should be unlocked with this function. +* +* @param fid +* SFT flow ID. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_unlock (uint32_t fid); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Removes flow from SFT. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* - Flow should be locked by caller in order to remove it. +* - Flow should have no client objects attached. +* +* Should be applied on aged flows, when flow aged event received. +* +* @code{.c} +* while (1) { +* rte_event_dequeue_burst(...); +* FOR_EACH_EV(ev) { +* uint32_t fid = ev.u64; +* rte_sft_flow_lock(fid); +* FOR_EACH_CLIENT(fid, client_id) { +* rte_sft_flow_reset_client_obj(fid, client_obj); +* // detached client object handling +* } +* rte_sft_flow_destroy(fid, &error); +* } +* } +* @endcode +* +* @param fid +* SFT flow ID to destroy. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_destroy (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Reset flow age to zero. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Simulates last flow packet with timestamp set to just now. +* +* @param fid +* SFT flow ID. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_touch (uint32_t fid,struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set flow aging to specific value. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param aging +* New flow aging value. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_set_aging (uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Set client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to set object for. +* @param client_obj +* Pointer to opaque client object structure. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +int +rte_sft_flow_set_client_obj (uint32_t fid, +uint8_t client_id, +void *client_obj, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Get client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to get object for. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid client object opaque pointer in case of success, NULL otherwise +* and rte_errno is set. +*/ +void * +rte_sft_flow_get_client_obj (const uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +/** +* @warning +* @b EXPERIMENTAL: this API may change without prior notice. +* +* Remove client object for given client ID. +* Flow should be locked by caller (see rte_sft_flow_lock). +* +* Detaches client object from SFT flow and returns the ownership for the +* client object to the caller by returning client object pointer value. +* The pointer returned by this function won't be accessed any more, the caller +* may release all client obj related resources & the memory allocated for +* this client object. +* +* @param fid +* SFT flow ID. +* @param client_id +* Client ID to remove object for. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid client object opaque pointer in case of success, NULL otherwise +* and rte_errno is set. +*/ +void * +rte_sft_flow_reset_client_obj (uint32_t fid, +uint8_t client_id, +struct rte_sft_error *error); + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_H_ */ lib/librte_sft/rte_sft_driver.h new file +195
-0@@ -0,0 +1,195 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + +#ifndef _RTE_SFT_DRIVER_H_ +#define _RTE_SFT_DRIVER_H_ + +/** +* @file +* +* RTE SFT Ethernet device PMD API +* +* APIs that are used by the SFT library to offload SFT operationons +* to Ethernet device. +*/ + +#include "rte_sft.h" + +#ifdef __cplusplus +extern "C" { +#endif + +/** +* Opaque type returned after successfully creating an entry in SFT. +* +* This handle can be used to manage and query the related entry (e.g. to +* destroy it or update age). +*/ +struct rte_sft_entry ; + +/** +* Create SFT entry in eth_dev SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param tuple +* L3/L4 5-tuple - src/dest IP and port and IP protocol. +* @param nat_tuple +* L3/L4 5-tuple to replace in packet original 5-tuple in order to implement +* NAT offloading; if NULL NAT offloading won't be configured for the flow. +* @param aging +* Flow aging timeout in seconds. +* @param ctx +* Initial values in SFT flow context +* (see RTE flow struct rte_flow_item_sft). +* ctx->zone should be valid. +* @param fid +* SFT flow ID for the entry to create on *device*. +* If there is an entry for the *fid* in PMD it will be updated with the +* values of *ctx*. +* @param[out] queue_index +* if PMD can figure out the queue where the flow packets will +* arrive in RX data path it will set the value of queue_index; otherwise +* all bits will be turned on. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid handle in case of success, NULL otherwise and rte_errno is set. +*/ +typedef struct rte_sft_entry *(*sft_entry_create_t ) (struct rte_eth_dev *dev, +const struct rte_sft_5tuple *tuple, +const struct rte_sft_5tuple *nat_tuple, +const uint32_t aging, +const struct rte_flow_item_sft *ctx, +const uint32_t fid, +uint16_t *queue_index, +struct rte_sft_error *error); + +/** +* Destroy SFT entry in eth_dev SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param entry +* Handle to the SFT entry to destroy. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_destroy_t )(struct rte_eth_dev *dev, +struct rte_sft_entry *entry, +struct rte_sft_error *error); + +/** +* Decodes SFT flow context if attached to mbuf by action ``SFT``. +* @see RTE flow RTE_FLOW_ACTION_TYPE_SFT. +* +* @param dev +* Pointer to Ethernet device structure. +* @param mbuf +* mbuf of the packet to decode attached state from. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* A valid SFT flow context in case of success, NULL otherwise and rte_errno +* is set. +*/ +typedef struct rte_flow_item_sft *(*sft_entry_mbuf_decode_ctx_t )( +struct rte_eth_dev *dev, +const struct rte_mbuf *mbuf, +struct rte_sft_error *error); + +/** +* Get aged-out SFT entries. +* +* Report entry as aged-out if timeout passed without any matching +* on the SFT entry. +* +* @param[in] dev +* Pointer to Ethernet device structure. +* @param[in, out] fid_aged +* The address of an array of aged-out SFT flow IDs. +* @param[in] nb_aged +* The length of *fid_aged* array pointers. +* @param[out] error +* Perform verbose error reporting if not NULL. Initialized in case of +* error only. +* +* @return +* if nb_aged is 0, return the amount of all aged flows. +* if nb_aged is not 0 , return the amount of aged flows reported +* in the *fid_aged* array, otherwise negative errno value. +*/ +typedef int (*sft_entry_get_aged_entries_t )(struct rte_eth_dev *dev, +uint32_t *fid_aged, +int nb_aged, +struct rte_sft_error *error); + +/** +* Simulate SFT entry match in terms of entry aging. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_touch_t )(struct rte_eth_dev *dev, +uint32_t fid, +struct rte_sft_error *error); + +/** +* Set SFT entry aging to specific value. +* +* @param dev +* Pointer to Ethernet device structure. +* @param fid +* SFT flow ID paired with dev to retrieve related SFT entry. +* @param aging +* New entry aging value. +* @param[out] error +* Perform verbose error reporting if not NULL. PMDs initialize this +* structure in case of error only. +* +* @return +* 0 on success, a negative errno value otherwise and rte_errno is set. +*/ +typedef int (*sft_entry_set_aging_t )(struct rte_eth_dev *dev, +uint32_t fid, +uint32_t aging, +struct rte_sft_error *error); + +/** SFT operations function pointer table */ +struct rte_sft_ops { +sft_entry_create_t entry_create; +/**< Create SFT entry in eth_dev SFT. */ +sft_entry_destroy_t entry_destroy; +/**< Destroy SFT entry in eth_dev SFT. */ +sft_entry_mbuf_decode_ctx_t mbuf_decode_ctx; +/**< Decodes SFT flow context if attached to mbuf by action ``SFT``. */ +sft_entry_get_aged_entries_t get_aged_entries; +/**< Get aged-out SFT entries. */ +sft_entry_touch_t entry_touch; +/**< Simulate SFT entry match in terms of entry aging. */ +sft_entry_set_aging_t set_aging; +/**< Set SFT entry aging to specific value. */ +}; + +#ifdef __cplusplus +} +#endif + +#endif /* _RTE_SFT_DRIVER_H_ */ lib/librte_sft/rte_sft_version.map new file +11
-0@@ -0,0 +1,21 @@ +EXPERIMENTAL { + global: + + rte_sft_flow_get_status; + rte_sft_flow_set_ctx; + rte_sft_init; + rte_sft_fini; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + rte_sft_drain_mbuf; + rte_sft_flow_activate;
lib/librte_sft/rte_sft_version.map line 21
Quoted 10 lines
lib/librte_sft/rte_sft_version.map new file +10
-0+ rte_sft_flow_create; + rte_sft_flow_lock; + rte_sft_flow_unlock; + rte_sft_flow_destroy; + rte_sft_flow_touch; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_get_client_obj; + rte_sft_flow_reset_client_obj; +};
Missing the __rte_experimental attribute in front of all these in the header.
+ Ferruh & Andrew.
Adding more people that may find this discussion relevant. Any feedback highly appreciated.
Thanks,
Andrey
Quoted 46 lines
-----Original Message-----
From: dev <[email protected]> On Behalf Of Andrey Vesnovaty
Sent: Wednesday, September 9, 2020 11:30 PM
To: [email protected]
Cc: [email protected]; Ori Kam <[email protected]>; Slava Ovsiienko <[email protected]>; [email protected]; Oz Shlomo <[email protected]>; Eli Britstein <[email protected]>; Alex Rosenbaum <[email protected]>; Roni Bar Yanai <[email protected]>
Subject: [dpdk-dev] [RFC 0/3] introduce Stateful Flow TableThe RFC introduces Stateful Flow Table (SFT) API and changes needed in both ethdev an RTE flow to support SFT functionality.
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.Andrey Vesnovaty (3):
ethdev: add item/action for SFT
ethdev: support SFT APIs
sft: introduce APIlib/librte_ethdev/rte_ethdev.c | 7 + lib/librte_ethdev/rte_ethdev.h | 16 + lib/librte_ethdev/rte_ethdev_core.h | 1 + lib/librte_ethdev/rte_flow.h | 84 +++ lib/librte_sft/Makefile | 28 + lib/librte_sft/meson.build | 7 + lib/librte_sft/rte_sft.c | 9 + lib/librte_sft/rte_sft.h | 845 ++++++++++++++++++++++++++++ lib/librte_sft/rte_sft_driver.h | 195 +++++++ lib/librte_sft/rte_sft_version.map | 21 + 10 files changed, 1213 insertions(+) create mode 100644 lib/librte_sft/Makefile create mode 100644 lib/librte_sft/meson.build create mode 100644 lib/librte_sft/rte_sft.c create mode 100644 lib/librte_sft/rte_sft.h create mode 100644 lib/librte_sft/rte_sft_driver.h create mode 100644 lib/librte_sft/rte_sft_version.map--
2.26.2
v2 cover introduce stateful flow table
In reply to Andrey VesnovatyThe RFC introduces Stateful Flow Table (SFT) API and changes needed in both ethdev an RTE flow to support SFT functionality.
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.
Andrey Vesnovaty (1):
ethdev: add item/action for SFT
Ori Kam (1):
ethdev: introduce sft lib
lib/librte_ethdev/meson.build | 3 + lib/librte_ethdev/rte_ethdev_version.map | 19 + lib/librte_ethdev/rte_flow.h | 75 ++ lib/librte_ethdev/rte_sft.c | 9 + lib/librte_ethdev/rte_sft.h | 878 +++++++++++++++++++++++ lib/librte_ethdev/rte_sft_driver.h | 201 ++++++ 6 files changed, 1185 insertions(+) create mode 100644 lib/librte_ethdev/rte_sft.c create mode 100644 lib/librte_ethdev/rte_sft.h create mode 100644 lib/librte_ethdev/rte_sft_driver.h
Signature · 1 line
2.25.1
PATCH v2 1/2 ethdev: add item/action for SFT
From: Andrey Vesnovaty <[email protected]>
Attach SFT flow context to packet with SFT action. Match on SFT flow context (attached to packet), with SFT item.
--- lib/librte_ethdev/rte_flow.h | 75 ++++++++++++++++++++++++++++++++++++ 1 file changed, 75 insertions(+)
lib/librte_ethdev/rte_flow.h +75 -0
@@ -537,6 +537,14 @@ enum rte_flow_item_type {*/ RTE_FLOW_ITEM_TYPE_ECPRI , +/** +* [META] +* +* Matches SFT context. +* +* See struct rte_flow_item_sft. +*/ + RTE_FLOW_ITEM_TYPE_SFT, };/**
@@ -1579,6 +1587,48 @@ static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { };#endif +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ITEM_TYPE_SFT +* +* Matches context of flow that wasw created by SFT. +* +* The structure describes SFT flow context. +* All the fields of the structure, except @p fid, should be considered as +* user defined. +* The @p fid assigned by RTE SFT & used as unique flow identifier. +* SFT context attached to packet by action ``SFT`` (see RTE_FLOW_ACTION_SFT), +* which sends the packet to the SFT module. +* +* SFT default context defined as context attached to packet when there is no +* entry for the flow in SFT. The @p state has application reserved value +* meaning that SFT context for the packet undefined since entry wasn't found +* in SFT. If state 'undefined' then @p zone should be valid othervice @p fid +* should be valid. +* +* Context considered virtual since the method of storing this info on packet +* is PMD/implementation specific & may involve mapping methods if there is +* 'not enough bits' to store entire contents of struct rte_flow_item_sft. +* +* Maximal value/size of each field depends on HW capabilities and considered +* as implementation specific. +*/ +RTE_STD_C11 +struct rte_flow_item_sft { +union { +uint32_t fid;/**< SFT flow identifier. */ +uint32_t zone;/**< Zone assigned to flow. */ + }; + uint32_t fid_valid:1; /**< The fid member is valid. */ + uint32_t zone_valid:1; /**< The zone member is valid. */ + uint32_t reserved:30; /**< Reserved. */ +uint8_t state;/**< User defined flow state. */ +uint8_t user_data_size;/**< user_data buffer size. */ +uint8_t *user_data;/**< Arbitrary user data. */ +}; +/** * Matching pattern item definition. *
@@ -2132,6 +2182,15 @@ enum rte_flow_action_type {* see enum RTE_ETH_EVENT_FLOW_AGED */ RTE_FLOW_ACTION_TYPE_AGE , + +/** +* RTE_FLOW_ACTION_TYPE_SFT +* +* Direct the packet to SFT module. +* +* See struct rte_flow_action_sft. +*/ + RTE_FLOW_ACTION_TYPE_SFT, };/**
@@ -2721,6 +2780,22 @@ rte_flow_dynf_metadata_set(struct rte_mbuf *m, uint32_t v) *RTE_FLOW_DYNF_METADATA (m) = v; } +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ACTION_TYPE_SFT +* +* Performs lookup by *zone* and 5-tuple in SFT. +* If entry is found the related SFT context will be attached otherwise, +* default SFT context will be attached. +* +* This action may result in termination of actions that following this action. +*/ +struct rte_flow_action_sft { +uint32_t zone;/**< Zone for lookup in SFT */ +}; +/* * Definition of a single action. *
Signature · 1 line
2.25.1
PATCH v2 2/2 ethdev: introduce sft lib
In reply to Ori KamDefines RTE SFT (Stateful Flow Table) APIs for Stateful Flow Table library.
Currently, DPDK enables only stateless offloading, using the rte_flow. stateless means that each packet is handled without any knowledge of privious or future packets.
As we look at the industry, there is much demand to save a context across packets that belong to a connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.
The solution suggested is to create a lib that will enable saving states between different packets that belong to the same connection.
The solution will also enable better HW abstraction than the one we get from using the rte_flow. The reason for this is that saving states is not atomic action like we have in rte_flow and also can't be done fully in HW (The first packets must be seen by the application). Saying the above this lib is based on interacting with the rte_flow but it doesn't replace it or encapsulate it.
Key design points.
- The SFT should offload as much as possible to HW.
- The SFT is designed to work alongside the rte_flow.
- The SFT has its own ops that the PMD needs to implement.
- The SFT works on 5 tuple + zone (a user-defined value)
Basic usage flow:
1. Application insert a flow that matches all eth traffic and have sft action along with jump action. (in the future this jump can be avoided and in doing so saving some jumps,
but for the most generic and complete solution we think that allow the application full control of the packet process using rte_flow is better.)
2. Application insert a flow in the target group that matches the packet state. Based on this state the application performs the needed actions. This flow can also be merged with other matching criteria. The application will also add a flow in the target group that will upload to the application any packet with a miss state.
3. First eth packet arrives and is routed to the SFT HW component. since this is the first packet the SFT will have a miss and will mark the packet with miss state and forward it to the target group.
4. The application will pull the packet from the queue and will send it to be processed by the sft lib.
5. The SFT will extract HW packet state and if valid the zone or the flow-id, and report it back to the application.
6. Application will see that this is a new connection, so it will issue SFT command to create a new connection with a selected state. The SFT will create a HW flow, that matches the 5 tuple + zone and sets the state of the packet. The state can be any u8 value, it is the responsibility of the application to match on the value.
7. On the next packet arriving to the HW it will jump to the SFT and in the SFT HW there will be a match which will result in setting the packet state and ID according to the application request.
8. In case of later miss (at some other group) or application logic that the packet should be routed back to the application. The application will call the SFT lib with the new mbuf, which will result in the flow-id returend to the application along with the context attached to this connection.
--- lib/librte_ethdev/meson.build | 3 + lib/librte_ethdev/rte_ethdev_version.map | 19 + lib/librte_ethdev/rte_sft.c | 9 + lib/librte_ethdev/rte_sft.h | 878 +++++++++++++++++++++++ lib/librte_ethdev/rte_sft_driver.h | 201 ++++++ 5 files changed, 1110 insertions(+) create mode 100644 lib/librte_ethdev/rte_sft.c create mode 100644 lib/librte_ethdev/rte_sft.h create mode 100644 lib/librte_ethdev/rte_sft_driver.h
lib/librte_ethdev/meson.build +3 -0
@@ -9,6 +9,7 @@ sources = files('ethdev_private.c', 'rte_ethdev.c', 'rte_flow.c', 'rte_mtr.c', + 'rte_sft.c', 'rte_tm.c') headers = files('rte_ethdev.h',
@@ -24,6 +25,8 @@ headers = files('rte_ethdev.h', 'rte_flow_driver.h', 'rte_mtr.h', 'rte_mtr_driver.h', + 'rte_sft.h', + 'rte_sft_driver.h', 'rte_tm.h', 'rte_tm_driver.h')
lib/librte_ethdev/rte_ethdev_version.map +19 -0
@@ -232,6 +232,25 @@ EXPERIMENTAL { rte_eth_fec_get_capability; rte_eth_fec_get; rte_eth_fec_set; + rte_sft_drain_mbuf; + rte_sft_fini; + rte_sft_flow_activate; + rte_sft_flow_create; + rte_sft_flow_destroy; + rte_sft_flow_get_client_obj; + rte_sft_flow_get_status; + rte_sft_flow_query; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_set_data; + rte_sft_flow_set_offload; + rte_sft_flow_set_state; + rte_sft_flow_touch; + rte_sft_init; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + + }; INTERNAL {
lib/librte_ethdev/rte_sft.c new file +9 -0
@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */
lib/librte_ethdev/rte_sft.h new file +878 -0
878 lines not on this page · raw message
lib/librte_ethdev/rte_sft_driver.h new file +201 -0
201 lines not on this page · raw message
Signature · 1 line
2.25.1
v3 cover introduce stateful flow table
In reply to Andrey VesnovatyThe RFC introduces Stateful Flow Table (SFT) API and changes needed in both ethdev an RTE flow to support SFT functionality.
SFT library provides a framework for applications that need to maintain context across different packets of the connection.
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.
Change log:
v3:
- add change log.
- change to RFC
v2:
- Add queue approach in the SFT.
- Move to ethdev.
- update based on ML commentes.
*** BLURB HERE ***
Andrey Vesnovaty (1):
ethdev: add item/action for SFT
Ori Kam (1):
ethdev: introduce sft lib
lib/librte_ethdev/meson.build | 3 + lib/librte_ethdev/rte_ethdev_version.map | 19 + lib/librte_ethdev/rte_flow.h | 75 ++ lib/librte_ethdev/rte_sft.c | 9 + lib/librte_ethdev/rte_sft.h | 877 +++++++++++++++++++++++ lib/librte_ethdev/rte_sft_driver.h | 201 ++++++ 6 files changed, 1184 insertions(+) create mode 100644 lib/librte_ethdev/rte_sft.c create mode 100644 lib/librte_ethdev/rte_sft.h create mode 100644 lib/librte_ethdev/rte_sft_driver.h
Signature · 1 line
2.25.1
RFC v3 1/2 ethdev: add item/action for SFT
From: Andrey Vesnovaty <[email protected]>
Attach SFT flow context to packet with SFT action. Match on SFT flow context (attached to packet), with SFT item.
--- lib/librte_ethdev/rte_flow.h | 75 ++++++++++++++++++++++++++++++++++++ 1 file changed, 75 insertions(+)
lib/librte_ethdev/rte_flow.h +75 -0
@@ -537,6 +537,14 @@ enum rte_flow_item_type {*/ RTE_FLOW_ITEM_TYPE_ECPRI , +/** +* [META] +* +* Matches SFT context. +* +* See struct rte_flow_item_sft. +*/ + RTE_FLOW_ITEM_TYPE_SFT, };/**
@@ -1579,6 +1587,48 @@ static const struct rte_flow_item_ecpri rte_flow_item_ecpri_mask = { };#endif +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ITEM_TYPE_SFT +* +* Matches context of flow that wasw created by SFT. +* +* The structure describes SFT flow context. +* All the fields of the structure, except @p fid, should be considered as +* user defined. +* The @p fid assigned by RTE SFT & used as unique flow identifier. +* SFT context attached to packet by action ``SFT`` (see RTE_FLOW_ACTION_SFT), +* which sends the packet to the SFT module. +* +* SFT default context defined as context attached to packet when there is no +* entry for the flow in SFT. The @p state has application reserved value +* meaning that SFT context for the packet undefined since entry wasn't found +* in SFT. If state 'undefined' then @p zone should be valid othervice @p fid +* should be valid. +* +* Context considered virtual since the method of storing this info on packet +* is PMD/implementation specific & may involve mapping methods if there is +* 'not enough bits' to store entire contents of struct rte_flow_item_sft. +* +* Maximal value/size of each field depends on HW capabilities and considered +* as implementation specific. +*/ +RTE_STD_C11 +struct rte_flow_item_sft { +union { +uint32_t fid;/**< SFT flow identifier. */ +uint32_t zone;/**< Zone assigned to flow. */ + }; + uint32_t fid_valid:1; /**< The fid member is valid. */ + uint32_t zone_valid:1; /**< The zone member is valid. */ + uint32_t reserved:30; /**< Reserved. */ +uint8_t state;/**< User defined flow state. */ +uint8_t user_data_size;/**< user_data buffer size. */ +uint8_t *user_data;/**< Arbitrary user data. */ +}; +/** * Matching pattern item definition. *
@@ -2132,6 +2182,15 @@ enum rte_flow_action_type {* see enum RTE_ETH_EVENT_FLOW_AGED */ RTE_FLOW_ACTION_TYPE_AGE , + +/** +* RTE_FLOW_ACTION_TYPE_SFT +* +* Direct the packet to SFT module. +* +* See struct rte_flow_action_sft. +*/ + RTE_FLOW_ACTION_TYPE_SFT, };/**
@@ -2721,6 +2780,22 @@ rte_flow_dynf_metadata_set(struct rte_mbuf *m, uint32_t v) *RTE_FLOW_DYNF_METADATA (m) = v; } +/** +* @warning +* @b EXPERIMENTAL: this structure may change without prior notice +* +* RTE_FLOW_ACTION_TYPE_SFT +* +* Performs lookup by *zone* and 5-tuple in SFT. +* If entry is found the related SFT context will be attached otherwise, +* default SFT context will be attached. +* +* This action may result in termination of actions that following this action. +*/ +struct rte_flow_action_sft { +uint32_t zone;/**< Zone for lookup in SFT */ +}; +/* * Definition of a single action. *
Signature · 1 line
2.25.1
RFC v3 2/2 ethdev: introduce sft lib
In reply to Ori KamDefines RTE SFT (Stateful Flow Table) APIs for Stateful Flow Table library.
Currently, DPDK enables only stateless offloading, using the rte_flow. stateless means that each packet is handled without any knowledge of previous or future packets.
As we look at the industry, there is much demand to save a context across packets that belong to a connection.
Examples for such applications:
- Next-generation firewalls
- Intrusion detection/prevention systems (IDS/IPS): Suricata, snort
- SW/Virtual Switching: OVS
The goals of the SFT library:
- Accelerate flow recognition & its context retrieval for further lookaside processing.
- Enable context-aware flow handling offload.
The solution suggested is to create a lib that will enable saving states between different packets that belong to the same connection.
The solution will also enable better HW abstraction than the one we get from using the rte_flow. The reason for this is that saving states is not atomic action like we have in rte_flow and also can't be done fully in HW (The first packets must be seen by the application). Saying the above this lib is based on interacting with the rte_flow but it doesn't replace it or encapsulate it.
Key design points.
- The SFT should offload as much as possible to HW.
- The SFT is designed to work alongside the rte_flow.
- The SFT has its own ops that the PMD needs to implement.
- The SFT works on 5 tuple + zone (a user-defined value)
Basic usage flow:
1. Application insert a flow that matches all eth traffic and have sft action along with jump action. (in the future this jump can be avoided and in doing so saving some jumps,
but for the most generic and complete solution we think that allow the application full control of the packet process using rte_flow is better.)
2. Application insert a flow in the target group that matches the packet state. Based on this state the application performs the needed actions. This flow can also be merged with other matching criteria. The application will also add a flow in the target group that will upload to the application any packet with a miss state.
3. First eth packet arrives and is routed to the SFT HW component. since this is the first packet the SFT will have a miss and will mark the packet with miss state and forward it to the target group.
4. The application will pull the packet from the queue and will send it to be processed by the sft lib.
5. The SFT will extract HW packet state and if valid the zone or the flow-id, and report it back to the application.
6. Application will see that this is a new connection, so it will issue SFT command to create a new connection with a selected state. The SFT will create a HW flow, that matches the 5 tuple + zone and sets the state of the packet. The state can be any u8 value, it is the responsibility of the application to match on the value.
7. On the next packet arriving to the HW it will jump to the SFT and in the SFT HW there will be a match which will result in setting the packet state and ID according to the application request.
8. In case of later miss (at some other group) or application logic that the packet should be routed back to the application. The application will call the SFT lib with the new mbuf, which will result in the flow-id returned to the application along with the context attached to this connection.
--- lib/librte_ethdev/meson.build | 3 + lib/librte_ethdev/rte_ethdev_version.map | 19 + lib/librte_ethdev/rte_sft.c | 9 + lib/librte_ethdev/rte_sft.h | 877 +++++++++++++++++++++++ lib/librte_ethdev/rte_sft_driver.h | 201 ++++++ 5 files changed, 1109 insertions(+) create mode 100644 lib/librte_ethdev/rte_sft.c create mode 100644 lib/librte_ethdev/rte_sft.h create mode 100644 lib/librte_ethdev/rte_sft_driver.h
lib/librte_ethdev/meson.build +3 -0
@@ -9,6 +9,7 @@ sources = files('ethdev_private.c', 'rte_ethdev.c', 'rte_flow.c', 'rte_mtr.c', + 'rte_sft.c', 'rte_tm.c') headers = files('rte_ethdev.h',
@@ -24,6 +25,8 @@ headers = files('rte_ethdev.h', 'rte_flow_driver.h', 'rte_mtr.h', 'rte_mtr_driver.h', + 'rte_sft.h', + 'rte_sft_driver.h', 'rte_tm.h', 'rte_tm_driver.h')
lib/librte_ethdev/rte_ethdev_version.map +19 -0
@@ -232,6 +232,25 @@ EXPERIMENTAL { rte_eth_fec_get_capability; rte_eth_fec_get; rte_eth_fec_set; + rte_sft_drain_mbuf; + rte_sft_fini; + rte_sft_flow_activate; + rte_sft_flow_create; + rte_sft_flow_destroy; + rte_sft_flow_get_client_obj; + rte_sft_flow_get_status; + rte_sft_flow_query; + rte_sft_flow_set_aging; + rte_sft_flow_set_client_obj; + rte_sft_flow_set_data; + rte_sft_flow_set_offload; + rte_sft_flow_set_state; + rte_sft_flow_touch; + rte_sft_init; + rte_sft_process_mbuf; + rte_sft_process_mbuf_with_zone; + + }; INTERNAL {
lib/librte_ethdev/rte_sft.c new file +9 -0
@@ -0,0 +1,9 @@ +/* SPDX-License-Identifier: BSD-3-Clause +* Copyright 2020 Mellanox Technologies, Ltd +*/ + + +#include "rte_sft.h" +#include "rte_sft_driver.h" + +/* Placeholder for RTE SFT library APIs implementation */
lib/librte_ethdev/rte_sft.h new file +877 -0
877 lines not on this page · raw message
lib/librte_ethdev/rte_sft_driver.h new file +201 -0
201 lines not on this page · raw message
Signature · 1 line
2.25.1