dsa.rst 27 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259260261262263264265266267268269270271272273274275276277278279280281282283284285286287288289290291292293294295296297298299300301302303304305306307308309310311312313314315316317318319320321322323324325326327328329330331332333334335336337338339340341342343344345346347348349350351352353354355356357358359360361362363364365366367368369370371372373374375376377378379380381382383384385386387388389390391392393394395396397398399400401402403404405406407408409410411412413414415416417418419420421422423424425426427428429430431432433434435436437438439440441442443444445446447448449450451452453454455456457458459460461462463464465466467468469470471472473474475476477478479480481482483484485486487488489490491492493494495496497498499500501502503504505506507508509510511512513514515516517518519520521522523524525526527528529530531532533534535536537538539540541542543544545546547548549550551552553554555556557558559560561562563564565566567568569570571572573574575576577578579580581582583584585586587
  1. ============
  2. Architecture
  3. ============
  4. This document describes the **Distributed Switch Architecture (DSA)** subsystem
  5. design principles, limitations, interactions with other subsystems, and how to
  6. develop drivers for this subsystem as well as a TODO for developers interested
  7. in joining the effort.
  8. Design principles
  9. =================
  10. The Distributed Switch Architecture is a subsystem which was primarily designed
  11. to support Marvell Ethernet switches (MV88E6xxx, a.k.a Linkstreet product line)
  12. using Linux, but has since evolved to support other vendors as well.
  13. The original philosophy behind this design was to be able to use unmodified
  14. Linux tools such as bridge, iproute2, ifconfig to work transparently whether
  15. they configured/queried a switch port network device or a regular network
  16. device.
  17. An Ethernet switch is typically comprised of multiple front-panel ports, and one
  18. or more CPU or management port. The DSA subsystem currently relies on the
  19. presence of a management port connected to an Ethernet controller capable of
  20. receiving Ethernet frames from the switch. This is a very common setup for all
  21. kinds of Ethernet switches found in Small Home and Office products: routers,
  22. gateways, or even top-of-the rack switches. This host Ethernet controller will
  23. be later referred to as "master" and "cpu" in DSA terminology and code.
  24. The D in DSA stands for Distributed, because the subsystem has been designed
  25. with the ability to configure and manage cascaded switches on top of each other
  26. using upstream and downstream Ethernet links between switches. These specific
  27. ports are referred to as "dsa" ports in DSA terminology and code. A collection
  28. of multiple switches connected to each other is called a "switch tree".
  29. For each front-panel port, DSA will create specialized network devices which are
  30. used as controlling and data-flowing endpoints for use by the Linux networking
  31. stack. These specialized network interfaces are referred to as "slave" network
  32. interfaces in DSA terminology and code.
  33. The ideal case for using DSA is when an Ethernet switch supports a "switch tag"
  34. which is a hardware feature making the switch insert a specific tag for each
  35. Ethernet frames it received to/from specific ports to help the management
  36. interface figure out:
  37. - what port is this frame coming from
  38. - what was the reason why this frame got forwarded
  39. - how to send CPU originated traffic to specific ports
  40. The subsystem does support switches not capable of inserting/stripping tags, but
  41. the features might be slightly limited in that case (traffic separation relies
  42. on Port-based VLAN IDs).
  43. Note that DSA does not currently create network interfaces for the "cpu" and
  44. "dsa" ports because:
  45. - the "cpu" port is the Ethernet switch facing side of the management
  46. controller, and as such, would create a duplication of feature, since you
  47. would get two interfaces for the same conduit: master netdev, and "cpu" netdev
  48. - the "dsa" port(s) are just conduits between two or more switches, and as such
  49. cannot really be used as proper network interfaces either, only the
  50. downstream, or the top-most upstream interface makes sense with that model
  51. Switch tagging protocols
  52. ------------------------
  53. DSA currently supports 5 different tagging protocols, and a tag-less mode as
  54. well. The different protocols are implemented in:
  55. - ``net/dsa/tag_trailer.c``: Marvell's 4 trailer tag mode (legacy)
  56. - ``net/dsa/tag_dsa.c``: Marvell's original DSA tag
  57. - ``net/dsa/tag_edsa.c``: Marvell's enhanced DSA tag
  58. - ``net/dsa/tag_brcm.c``: Broadcom's 4 bytes tag
  59. - ``net/dsa/tag_qca.c``: Qualcomm's 2 bytes tag
  60. The exact format of the tag protocol is vendor specific, but in general, they
  61. all contain something which:
  62. - identifies which port the Ethernet frame came from/should be sent to
  63. - provides a reason why this frame was forwarded to the management interface
  64. Master network devices
  65. ----------------------
  66. Master network devices are regular, unmodified Linux network device drivers for
  67. the CPU/management Ethernet interface. Such a driver might occasionally need to
  68. know whether DSA is enabled (e.g.: to enable/disable specific offload features),
  69. but the DSA subsystem has been proven to work with industry standard drivers:
  70. ``e1000e,`` ``mv643xx_eth`` etc. without having to introduce modifications to these
  71. drivers. Such network devices are also often referred to as conduit network
  72. devices since they act as a pipe between the host processor and the hardware
  73. Ethernet switch.
  74. Networking stack hooks
  75. ----------------------
  76. When a master netdev is used with DSA, a small hook is placed in the
  77. networking stack is in order to have the DSA subsystem process the Ethernet
  78. switch specific tagging protocol. DSA accomplishes this by registering a
  79. specific (and fake) Ethernet type (later becoming ``skb->protocol``) with the
  80. networking stack, this is also known as a ``ptype`` or ``packet_type``. A typical
  81. Ethernet Frame receive sequence looks like this:
  82. Master network device (e.g.: e1000e):
  83. 1. Receive interrupt fires:
  84. - receive function is invoked
  85. - basic packet processing is done: getting length, status etc.
  86. - packet is prepared to be processed by the Ethernet layer by calling
  87. ``eth_type_trans``
  88. 2. net/ethernet/eth.c::
  89. eth_type_trans(skb, dev)
  90. if (dev->dsa_ptr != NULL)
  91. -> skb->protocol = ETH_P_XDSA
  92. 3. drivers/net/ethernet/\*::
  93. netif_receive_skb(skb)
  94. -> iterate over registered packet_type
  95. -> invoke handler for ETH_P_XDSA, calls dsa_switch_rcv()
  96. 4. net/dsa/dsa.c::
  97. -> dsa_switch_rcv()
  98. -> invoke switch tag specific protocol handler in 'net/dsa/tag_*.c'
  99. 5. net/dsa/tag_*.c:
  100. - inspect and strip switch tag protocol to determine originating port
  101. - locate per-port network device
  102. - invoke ``eth_type_trans()`` with the DSA slave network device
  103. - invoked ``netif_receive_skb()``
  104. Past this point, the DSA slave network devices get delivered regular Ethernet
  105. frames that can be processed by the networking stack.
  106. Slave network devices
  107. ---------------------
  108. Slave network devices created by DSA are stacked on top of their master network
  109. device, each of these network interfaces will be responsible for being a
  110. controlling and data-flowing end-point for each front-panel port of the switch.
  111. These interfaces are specialized in order to:
  112. - insert/remove the switch tag protocol (if it exists) when sending traffic
  113. to/from specific switch ports
  114. - query the switch for ethtool operations: statistics, link state,
  115. Wake-on-LAN, register dumps...
  116. - external/internal PHY management: link, auto-negotiation etc.
  117. These slave network devices have custom net_device_ops and ethtool_ops function
  118. pointers which allow DSA to introduce a level of layering between the networking
  119. stack/ethtool, and the switch driver implementation.
  120. Upon frame transmission from these slave network devices, DSA will look up which
  121. switch tagging protocol is currently registered with these network devices, and
  122. invoke a specific transmit routine which takes care of adding the relevant
  123. switch tag in the Ethernet frames.
  124. These frames are then queued for transmission using the master network device
  125. ``ndo_start_xmit()`` function, since they contain the appropriate switch tag, the
  126. Ethernet switch will be able to process these incoming frames from the
  127. management interface and delivers these frames to the physical switch port.
  128. Graphical representation
  129. ------------------------
  130. Summarized, this is basically how DSA looks like from a network device
  131. perspective::
  132. |---------------------------
  133. | CPU network device (eth0)|
  134. ----------------------------
  135. | <tag added by switch |
  136. | |
  137. | |
  138. | tag added by CPU> |
  139. |--------------------------------------------|
  140. | Switch driver |
  141. |--------------------------------------------|
  142. || || ||
  143. |-------| |-------| |-------|
  144. | sw0p0 | | sw0p1 | | sw0p2 |
  145. |-------| |-------| |-------|
  146. Slave MDIO bus
  147. --------------
  148. In order to be able to read to/from a switch PHY built into it, DSA creates a
  149. slave MDIO bus which allows a specific switch driver to divert and intercept
  150. MDIO reads/writes towards specific PHY addresses. In most MDIO-connected
  151. switches, these functions would utilize direct or indirect PHY addressing mode
  152. to return standard MII registers from the switch builtin PHYs, allowing the PHY
  153. library and/or to return link status, link partner pages, auto-negotiation
  154. results etc..
  155. For Ethernet switches which have both external and internal MDIO busses, the
  156. slave MII bus can be utilized to mux/demux MDIO reads and writes towards either
  157. internal or external MDIO devices this switch might be connected to: internal
  158. PHYs, external PHYs, or even external switches.
  159. Data structures
  160. ---------------
  161. DSA data structures are defined in ``include/net/dsa.h`` as well as
  162. ``net/dsa/dsa_priv.h``:
  163. - ``dsa_chip_data``: platform data configuration for a given switch device,
  164. this structure describes a switch device's parent device, its address, as
  165. well as various properties of its ports: names/labels, and finally a routing
  166. table indication (when cascading switches)
  167. - ``dsa_platform_data``: platform device configuration data which can reference
  168. a collection of dsa_chip_data structure if multiples switches are cascaded,
  169. the master network device this switch tree is attached to needs to be
  170. referenced
  171. - ``dsa_switch_tree``: structure assigned to the master network device under
  172. ``dsa_ptr``, this structure references a dsa_platform_data structure as well as
  173. the tagging protocol supported by the switch tree, and which receive/transmit
  174. function hooks should be invoked, information about the directly attached
  175. switch is also provided: CPU port. Finally, a collection of dsa_switch are
  176. referenced to address individual switches in the tree.
  177. - ``dsa_switch``: structure describing a switch device in the tree, referencing
  178. a ``dsa_switch_tree`` as a backpointer, slave network devices, master network
  179. device, and a reference to the backing``dsa_switch_ops``
  180. - ``dsa_switch_ops``: structure referencing function pointers, see below for a
  181. full description.
  182. Design limitations
  183. ==================
  184. Limits on the number of devices and ports
  185. -----------------------------------------
  186. DSA currently limits the number of maximum switches within a tree to 4
  187. (``DSA_MAX_SWITCHES``), and the number of ports per switch to 12 (``DSA_MAX_PORTS``).
  188. These limits could be extended to support larger configurations would this need
  189. arise.
  190. Lack of CPU/DSA network devices
  191. -------------------------------
  192. DSA does not currently create slave network devices for the CPU or DSA ports, as
  193. described before. This might be an issue in the following cases:
  194. - inability to fetch switch CPU port statistics counters using ethtool, which
  195. can make it harder to debug MDIO switch connected using xMII interfaces
  196. - inability to configure the CPU port link parameters based on the Ethernet
  197. controller capabilities attached to it: http://patchwork.ozlabs.org/patch/509806/
  198. - inability to configure specific VLAN IDs / trunking VLANs between switches
  199. when using a cascaded setup
  200. Common pitfalls using DSA setups
  201. --------------------------------
  202. Once a master network device is configured to use DSA (dev->dsa_ptr becomes
  203. non-NULL), and the switch behind it expects a tagging protocol, this network
  204. interface can only exclusively be used as a conduit interface. Sending packets
  205. directly through this interface (e.g.: opening a socket using this interface)
  206. will not make us go through the switch tagging protocol transmit function, so
  207. the Ethernet switch on the other end, expecting a tag will typically drop this
  208. frame.
  209. Slave network devices check that the master network device is UP before allowing
  210. you to administratively bring UP these slave network devices. A common
  211. configuration mistake is forgetting to bring UP the master network device first.
  212. Interactions with other subsystems
  213. ==================================
  214. DSA currently leverages the following subsystems:
  215. - MDIO/PHY library: ``drivers/net/phy/phy.c``, ``mdio_bus.c``
  216. - Switchdev:``net/switchdev/*``
  217. - Device Tree for various of_* functions
  218. MDIO/PHY library
  219. ----------------
  220. Slave network devices exposed by DSA may or may not be interfacing with PHY
  221. devices (``struct phy_device`` as defined in ``include/linux/phy.h)``, but the DSA
  222. subsystem deals with all possible combinations:
  223. - internal PHY devices, built into the Ethernet switch hardware
  224. - external PHY devices, connected via an internal or external MDIO bus
  225. - internal PHY devices, connected via an internal MDIO bus
  226. - special, non-autonegotiated or non MDIO-managed PHY devices: SFPs, MoCA; a.k.a
  227. fixed PHYs
  228. The PHY configuration is done by the ``dsa_slave_phy_setup()`` function and the
  229. logic basically looks like this:
  230. - if Device Tree is used, the PHY device is looked up using the standard
  231. "phy-handle" property, if found, this PHY device is created and registered
  232. using ``of_phy_connect()``
  233. - if Device Tree is used, and the PHY device is "fixed", that is, conforms to
  234. the definition of a non-MDIO managed PHY as defined in
  235. ``Documentation/devicetree/bindings/net/fixed-link.txt``, the PHY is registered
  236. and connected transparently using the special fixed MDIO bus driver
  237. - finally, if the PHY is built into the switch, as is very common with
  238. standalone switch packages, the PHY is probed using the slave MII bus created
  239. by DSA
  240. SWITCHDEV
  241. ---------
  242. DSA directly utilizes SWITCHDEV when interfacing with the bridge layer, and
  243. more specifically with its VLAN filtering portion when configuring VLANs on top
  244. of per-port slave network devices. Since DSA primarily deals with
  245. MDIO-connected switches, although not exclusively, SWITCHDEV's
  246. prepare/abort/commit phases are often simplified into a prepare phase which
  247. checks whether the operation is supported by the DSA switch driver, and a commit
  248. phase which applies the changes.
  249. As of today, the only SWITCHDEV objects supported by DSA are the FDB and VLAN
  250. objects.
  251. Device Tree
  252. -----------
  253. DSA features a standardized binding which is documented in
  254. ``Documentation/devicetree/bindings/net/dsa/dsa.txt``. PHY/MDIO library helper
  255. functions such as ``of_get_phy_mode()``, ``of_phy_connect()`` are also used to query
  256. per-port PHY specific details: interface connection, MDIO bus location etc..
  257. Driver development
  258. ==================
  259. DSA switch drivers need to implement a dsa_switch_ops structure which will
  260. contain the various members described below.
  261. ``register_switch_driver()`` registers this dsa_switch_ops in its internal list
  262. of drivers to probe for. ``unregister_switch_driver()`` does the exact opposite.
  263. Unless requested differently by setting the priv_size member accordingly, DSA
  264. does not allocate any driver private context space.
  265. Switch configuration
  266. --------------------
  267. - ``tag_protocol``: this is to indicate what kind of tagging protocol is supported,
  268. should be a valid value from the ``dsa_tag_protocol`` enum
  269. - ``probe``: probe routine which will be invoked by the DSA platform device upon
  270. registration to test for the presence/absence of a switch device. For MDIO
  271. devices, it is recommended to issue a read towards internal registers using
  272. the switch pseudo-PHY and return whether this is a supported device. For other
  273. buses, return a non-NULL string
  274. - ``setup``: setup function for the switch, this function is responsible for setting
  275. up the ``dsa_switch_ops`` private structure with all it needs: register maps,
  276. interrupts, mutexes, locks etc.. This function is also expected to properly
  277. configure the switch to separate all network interfaces from each other, that
  278. is, they should be isolated by the switch hardware itself, typically by creating
  279. a Port-based VLAN ID for each port and allowing only the CPU port and the
  280. specific port to be in the forwarding vector. Ports that are unused by the
  281. platform should be disabled. Past this function, the switch is expected to be
  282. fully configured and ready to serve any kind of request. It is recommended
  283. to issue a software reset of the switch during this setup function in order to
  284. avoid relying on what a previous software agent such as a bootloader/firmware
  285. may have previously configured.
  286. PHY devices and link management
  287. -------------------------------
  288. - ``get_phy_flags``: Some switches are interfaced to various kinds of Ethernet PHYs,
  289. if the PHY library PHY driver needs to know about information it cannot obtain
  290. on its own (e.g.: coming from switch memory mapped registers), this function
  291. should return a 32-bits bitmask of "flags", that is private between the switch
  292. driver and the Ethernet PHY driver in ``drivers/net/phy/\*``.
  293. - ``phy_read``: Function invoked by the DSA slave MDIO bus when attempting to read
  294. the switch port MDIO registers. If unavailable, return 0xffff for each read.
  295. For builtin switch Ethernet PHYs, this function should allow reading the link
  296. status, auto-negotiation results, link partner pages etc..
  297. - ``phy_write``: Function invoked by the DSA slave MDIO bus when attempting to write
  298. to the switch port MDIO registers. If unavailable return a negative error
  299. code.
  300. - ``adjust_link``: Function invoked by the PHY library when a slave network device
  301. is attached to a PHY device. This function is responsible for appropriately
  302. configuring the switch port link parameters: speed, duplex, pause based on
  303. what the ``phy_device`` is providing.
  304. - ``fixed_link_update``: Function invoked by the PHY library, and specifically by
  305. the fixed PHY driver asking the switch driver for link parameters that could
  306. not be auto-negotiated, or obtained by reading the PHY registers through MDIO.
  307. This is particularly useful for specific kinds of hardware such as QSGMII,
  308. MoCA or other kinds of non-MDIO managed PHYs where out of band link
  309. information is obtained
  310. Ethtool operations
  311. ------------------
  312. - ``get_strings``: ethtool function used to query the driver's strings, will
  313. typically return statistics strings, private flags strings etc.
  314. - ``get_ethtool_stats``: ethtool function used to query per-port statistics and
  315. return their values. DSA overlays slave network devices general statistics:
  316. RX/TX counters from the network device, with switch driver specific statistics
  317. per port
  318. - ``get_sset_count``: ethtool function used to query the number of statistics items
  319. - ``get_wol``: ethtool function used to obtain Wake-on-LAN settings per-port, this
  320. function may, for certain implementations also query the master network device
  321. Wake-on-LAN settings if this interface needs to participate in Wake-on-LAN
  322. - ``set_wol``: ethtool function used to configure Wake-on-LAN settings per-port,
  323. direct counterpart to set_wol with similar restrictions
  324. - ``set_eee``: ethtool function which is used to configure a switch port EEE (Green
  325. Ethernet) settings, can optionally invoke the PHY library to enable EEE at the
  326. PHY level if relevant. This function should enable EEE at the switch port MAC
  327. controller and data-processing logic
  328. - ``get_eee``: ethtool function which is used to query a switch port EEE settings,
  329. this function should return the EEE state of the switch port MAC controller
  330. and data-processing logic as well as query the PHY for its currently configured
  331. EEE settings
  332. - ``get_eeprom_len``: ethtool function returning for a given switch the EEPROM
  333. length/size in bytes
  334. - ``get_eeprom``: ethtool function returning for a given switch the EEPROM contents
  335. - ``set_eeprom``: ethtool function writing specified data to a given switch EEPROM
  336. - ``get_regs_len``: ethtool function returning the register length for a given
  337. switch
  338. - ``get_regs``: ethtool function returning the Ethernet switch internal register
  339. contents. This function might require user-land code in ethtool to
  340. pretty-print register values and registers
  341. Power management
  342. ----------------
  343. - ``suspend``: function invoked by the DSA platform device when the system goes to
  344. suspend, should quiesce all Ethernet switch activities, but keep ports
  345. participating in Wake-on-LAN active as well as additional wake-up logic if
  346. supported
  347. - ``resume``: function invoked by the DSA platform device when the system resumes,
  348. should resume all Ethernet switch activities and re-configure the switch to be
  349. in a fully active state
  350. - ``port_enable``: function invoked by the DSA slave network device ndo_open
  351. function when a port is administratively brought up, this function should be
  352. fully enabling a given switch port. DSA takes care of marking the port with
  353. ``BR_STATE_BLOCKING`` if the port is a bridge member, or ``BR_STATE_FORWARDING`` if it
  354. was not, and propagating these changes down to the hardware
  355. - ``port_disable``: function invoked by the DSA slave network device ndo_close
  356. function when a port is administratively brought down, this function should be
  357. fully disabling a given switch port. DSA takes care of marking the port with
  358. ``BR_STATE_DISABLED`` and propagating changes to the hardware if this port is
  359. disabled while being a bridge member
  360. Bridge layer
  361. ------------
  362. - ``port_bridge_join``: bridge layer function invoked when a given switch port is
  363. added to a bridge, this function should be doing the necessary at the switch
  364. level to permit the joining port from being added to the relevant logical
  365. domain for it to ingress/egress traffic with other members of the bridge.
  366. - ``port_bridge_leave``: bridge layer function invoked when a given switch port is
  367. removed from a bridge, this function should be doing the necessary at the
  368. switch level to deny the leaving port from ingress/egress traffic from the
  369. remaining bridge members. When the port leaves the bridge, it should be aged
  370. out at the switch hardware for the switch to (re) learn MAC addresses behind
  371. this port.
  372. - ``port_stp_state_set``: bridge layer function invoked when a given switch port STP
  373. state is computed by the bridge layer and should be propagated to switch
  374. hardware to forward/block/learn traffic. The switch driver is responsible for
  375. computing a STP state change based on current and asked parameters and perform
  376. the relevant ageing based on the intersection results
  377. Bridge VLAN filtering
  378. ---------------------
  379. - ``port_vlan_filtering``: bridge layer function invoked when the bridge gets
  380. configured for turning on or off VLAN filtering. If nothing specific needs to
  381. be done at the hardware level, this callback does not need to be implemented.
  382. When VLAN filtering is turned on, the hardware must be programmed with
  383. rejecting 802.1Q frames which have VLAN IDs outside of the programmed allowed
  384. VLAN ID map/rules. If there is no PVID programmed into the switch port,
  385. untagged frames must be rejected as well. When turned off the switch must
  386. accept any 802.1Q frames irrespective of their VLAN ID, and untagged frames are
  387. allowed.
  388. - ``port_vlan_prepare``: bridge layer function invoked when the bridge prepares the
  389. configuration of a VLAN on the given port. If the operation is not supported
  390. by the hardware, this function should return ``-EOPNOTSUPP`` to inform the bridge
  391. code to fallback to a software implementation. No hardware setup must be done
  392. in this function. See port_vlan_add for this and details.
  393. - ``port_vlan_add``: bridge layer function invoked when a VLAN is configured
  394. (tagged or untagged) for the given switch port
  395. - ``port_vlan_del``: bridge layer function invoked when a VLAN is removed from the
  396. given switch port
  397. - ``port_vlan_dump``: bridge layer function invoked with a switchdev callback
  398. function that the driver has to call for each VLAN the given port is a member
  399. of. A switchdev object is used to carry the VID and bridge flags.
  400. - ``port_fdb_add``: bridge layer function invoked when the bridge wants to install a
  401. Forwarding Database entry, the switch hardware should be programmed with the
  402. specified address in the specified VLAN Id in the forwarding database
  403. associated with this VLAN ID. If the operation is not supported, this
  404. function should return ``-EOPNOTSUPP`` to inform the bridge code to fallback to
  405. a software implementation.
  406. .. note:: VLAN ID 0 corresponds to the port private database, which, in the context
  407. of DSA, would be its port-based VLAN, used by the associated bridge device.
  408. - ``port_fdb_del``: bridge layer function invoked when the bridge wants to remove a
  409. Forwarding Database entry, the switch hardware should be programmed to delete
  410. the specified MAC address from the specified VLAN ID if it was mapped into
  411. this port forwarding database
  412. - ``port_fdb_dump``: bridge layer function invoked with a switchdev callback
  413. function that the driver has to call for each MAC address known to be behind
  414. the given port. A switchdev object is used to carry the VID and FDB info.
  415. - ``port_mdb_prepare``: bridge layer function invoked when the bridge prepares the
  416. installation of a multicast database entry. If the operation is not supported,
  417. this function should return ``-EOPNOTSUPP`` to inform the bridge code to fallback
  418. to a software implementation. No hardware setup must be done in this function.
  419. See ``port_fdb_add`` for this and details.
  420. - ``port_mdb_add``: bridge layer function invoked when the bridge wants to install
  421. a multicast database entry, the switch hardware should be programmed with the
  422. specified address in the specified VLAN ID in the forwarding database
  423. associated with this VLAN ID.
  424. .. note:: VLAN ID 0 corresponds to the port private database, which, in the context
  425. of DSA, would be its port-based VLAN, used by the associated bridge device.
  426. - ``port_mdb_del``: bridge layer function invoked when the bridge wants to remove a
  427. multicast database entry, the switch hardware should be programmed to delete
  428. the specified MAC address from the specified VLAN ID if it was mapped into
  429. this port forwarding database.
  430. - ``port_mdb_dump``: bridge layer function invoked with a switchdev callback
  431. function that the driver has to call for each MAC address known to be behind
  432. the given port. A switchdev object is used to carry the VID and MDB info.
  433. TODO
  434. ====
  435. Making SWITCHDEV and DSA converge towards an unified codebase
  436. -------------------------------------------------------------
  437. SWITCHDEV properly takes care of abstracting the networking stack with offload
  438. capable hardware, but does not enforce a strict switch device driver model. On
  439. the other DSA enforces a fairly strict device driver model, and deals with most
  440. of the switch specific. At some point we should envision a merger between these
  441. two subsystems and get the best of both worlds.
  442. Other hanging fruits
  443. --------------------
  444. - making the number of ports fully dynamic and not dependent on ``DSA_MAX_PORTS``
  445. - allowing more than one CPU/management interface:
  446. http://comments.gmane.org/gmane.linux.network/365657
  447. - porting more drivers from other vendors:
  448. http://comments.gmane.org/gmane.linux.network/365510