1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2018 The DPDK contributors
4 ABI and API Deprecation
5 =======================
7 See the guidelines document for details of the :doc:`ABI policy
8 <../contributing/abi_policy>`. API and ABI deprecation notices are to be posted
14 * kvargs: The function ``rte_kvargs_process`` will get a new parameter
15 for returning key match count. It will ease handling of no-match case.
17 * eal: The function ``rte_eal_remote_launch`` will return new error codes
18 after read or write error on the pipe, instead of calling ``rte_panic``.
20 * rte_atomicNN_xxx: These APIs do not take memory order parameter. This does
21 not allow for writing optimized code for all the CPU architectures supported
22 in DPDK. DPDK will adopt C11 atomic operations semantics and provide wrappers
23 using C11 atomic built-ins. These wrappers must be used for patches that
24 need to be merged in 20.08 onwards. This change will not introduce any
25 performance degradation.
27 * rte_smp_*mb: These APIs provide full barrier functionality. However, many
28 use cases do not require full barriers. To support such use cases, DPDK will
29 adopt C11 barrier semantics and provide wrappers using C11 atomic built-ins.
30 These wrappers must be used for patches that need to be merged in 20.08
31 onwards. This change will not introduce any performance degradation.
33 * lib: will fix extending some enum/define breaking the ABI. There are multiple
34 samples in DPDK that enum/define terminated with a ``.*MAX.*`` value which is
35 used by iterators, and arrays holding these values are sized with this
36 ``.*MAX.*`` value. So extending this enum/define increases the ``.*MAX.*``
37 value which increases the size of the array and depending on how/where the
38 array is used this may break the ABI.
39 ``RTE_ETH_FLOW_MAX`` is one sample of the mentioned case, adding a new flow
40 type will break the ABI because of ``flex_mask[RTE_ETH_FLOW_MAX]`` array
41 usage in following public struct hierarchy:
42 ``rte_eth_fdir_flex_conf -> rte_fdir_conf -> rte_eth_conf (in the middle)``.
43 Need to identify this kind of usages and fix in 20.11, otherwise this blocks
44 us extending existing enum/define.
45 One solution can be using a fixed size array instead of ``.*MAX.*`` value.
47 * ethdev: The flow director API, including ``rte_eth_conf.fdir_conf`` field,
48 and the related structures (``rte_fdir_*`` and ``rte_eth_fdir_*``),
49 will be removed in DPDK 20.11.
51 * ethdev: New offload flags ``DEV_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
52 This will allow application to enable or disable PMDs from updating
53 ``rte_mbuf::hash::fdir``.
54 This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
55 thereby improve Rx performance if application wishes do so.
56 In 19.11 PMDs will still update the field even when the offload is not
59 * ethdev: ``uint32_t max_rx_pkt_len`` field of ``struct rte_eth_rxmode``, will be
60 replaced by a new ``uint32_t mtu`` field of ``struct rte_eth_conf`` in v21.11.
61 The new ``mtu`` field will be used to configure the initial device MTU via
62 ``rte_eth_dev_configure()`` API.
63 Later MTU can be changed by ``rte_eth_dev_set_mtu()`` API as done now.
64 The existing ``(struct rte_eth_dev)->data->mtu`` variable will be used to store
65 the configured ``mtu`` value,
66 and this new ``(struct rte_eth_dev)->data->dev_conf.mtu`` variable will
67 be used to store the user configuration request.
68 Unlike ``max_rx_pkt_len``, which was valid only when ``JUMBO_FRAME`` enabled,
69 ``mtu`` field will be always valid.
70 When ``mtu`` config is not provided by the application, default ``RTE_ETHER_MTU``
72 ``(struct rte_eth_dev)->data->mtu`` should be updated after MTU set successfully,
73 either by ``rte_eth_dev_configure()`` or ``rte_eth_dev_set_mtu()``.
75 An application may need to configure device for a specific Rx packet size, like for
76 cases ``DEV_RX_OFFLOAD_SCATTER`` is not supported and device received packet size
77 can't be bigger than Rx buffer size.
78 To cover these cases an application needs to know the device packet overhead to be
79 able to calculate the ``mtu`` corresponding to a Rx buffer size, for this
80 ``(struct rte_eth_dev_info).max_rx_pktlen`` will be kept,
81 the device packet overhead can be calculated as:
82 ``(struct rte_eth_dev_info).max_rx_pktlen - (struct rte_eth_dev_info).max_mtu``
84 * ethdev: ``rx_descriptor_done`` dev_ops and ``rte_eth_rx_descriptor_done``
85 will be removed in 21.11.
86 Existing ``rte_eth_rx_descriptor_status`` and ``rte_eth_tx_descriptor_status``
87 APIs can be used as replacement.
89 * ethdev: The port mirroring API can be replaced with a more fine grain flow API.
90 The structs ``rte_eth_mirror_conf``, ``rte_eth_vlan_mirror`` and the functions
91 ``rte_eth_mirror_rule_set``, ``rte_eth_mirror_rule_reset`` will be marked
92 as deprecated in DPDK 20.11, along with the associated macros ``ETH_MIRROR_*``.
93 This API will be fully removed in DPDK 21.11.
95 * ethdev: Attribute ``shared`` of the ``struct rte_flow_action_count``
96 is deprecated and will be removed in DPDK 21.11. Shared counters should
97 be managed using shared actions API (``rte_flow_shared_action_create`` etc).
99 * ethdev: The flow API matching pattern structures, ``struct rte_flow_item_*``,
100 should start with relevant protocol header.
101 Some matching pattern structures implements this by duplicating protocol header
102 fields in the struct. To clarify the intention and to be sure protocol header
103 is intact, will replace those fields with relevant protocol header struct.
104 In v21.02 both individual protocol header fields and the protocol header struct
105 will be added as union, target is switch usage to the protocol header by time.
106 In v21.11 LTS, protocol header fields will be cleaned and only protocol header
109 * ethdev: Queue specific stats fields will be removed from ``struct rte_eth_stats``.
110 Mentioned fields are: ``q_ipackets``, ``q_opackets``, ``q_ibytes``, ``q_obytes``,
112 Instead queue stats will be received via xstats API. Current method support
113 will be limited to maximum 256 queues.
114 Also compile time flag ``RTE_ETHDEV_QUEUE_STAT_CNTRS`` will be removed.
116 * ethdev: The offload flag ``PKT_RX_EIP_CKSUM_BAD`` will be removed and
117 replaced by the new flag ``PKT_RX_OUTER_IP_CKSUM_BAD``. The new name is more
118 consistent with existing outer header checksum status flag naming, which
119 should help in reducing confusion about its usage.
121 * sched: To allow more traffic classes, flexible mapping of pipe queues to
122 traffic classes, and subport level configuration of pipes and queues
123 changes will be made to macros, data structures and API functions defined
124 in "rte_sched.h". These changes are aligned to improvements suggested in the
125 RFC https://mails.dpdk.org/archives/dev/2018-November/120035.html.
127 * metrics: The function ``rte_metrics_init`` will have a non-void return
128 in order to notify errors instead of calling ``rte_exit``.
130 * cmdline: ``cmdline`` structure will be made opaque to hide platform-specific
131 content. On Linux and FreeBSD, supported prior to DPDK 20.11,
132 original structure will be kept until DPDK 21.11.