1 .. SPDX-License-Identifier: BSD-3-Clause
2 Copyright 2018 The DPDK contributors
4 ABI and API Deprecation
5 =======================
7 See the :doc:`guidelines document for details of the ABI policy </contributing/versioning>`.
8 API and ABI deprecation notices are to be posted here.
14 * meson: The minimum supported version of meson for configuring and building
15 DPDK will be increased to v0.47.1 (from 0.41) from DPDK 19.05 onwards. For
16 those users with a version earlier than 0.47.1, an updated copy of meson
17 can be got using the ``pip``, or ``pip3``, tool for downloading python
20 * kvargs: The function ``rte_kvargs_process`` will get a new parameter
21 for returning key match count. It will ease handling of no-match case.
23 * eal: The function ``rte_eal_remote_launch`` will return new error codes
24 after read or write error on the pipe, instead of calling ``rte_panic``.
26 * eal: both declaring and identifying devices will be streamlined in v18.11.
27 New functions will appear to query a specific port from buses, classes of
28 device and device drivers. Device declaration will be made coherent with the
29 new scheme of device identification.
30 As such, ``rte_devargs`` device representation will change.
32 - The enum ``rte_devtype`` was used to identify a bus and will disappear.
33 - Functions previously deprecated will change or disappear:
35 + ``rte_eal_devargs_type_count``
37 * eal: The ``rte_logs`` struct and global symbol will be made private to
38 remove it from the externally visible ABI and allow it to be updated in the
41 * vfio: removal of ``rte_vfio_dma_map`` and ``rte_vfio_dma_unmap`` APIs which
42 have been replaced with ``rte_dev_dma_map`` and ``rte_dev_dma_unmap``
43 functions. The due date for the removal targets DPDK 20.02.
45 * dpaa2: removal of ``rte_dpaa2_memsegs`` structure which has been replaced
46 by a pa-va search library. This structure was earlier being used for holding
47 memory segments used by dpaa2 driver for faster pa->va translation. This
48 structure would be made internal (or removed if all dependencies are cleared)
51 * net: The Ethernet address and header definitions will change
52 attributes. The Ethernet address struct will no longer be marked as
53 packed since the packed attribute is meaningless on a byte
54 array. The Ethernet header will be marked as aligned on a 2-byte
55 boundary and will no longer have the packed attribute. This allows
56 for efficient access on CPU architectures where unaligned access is
57 expensive. These changes should not impact normal usage because drivers
58 naturally align the Ethernet header on receive and all known
59 encapsulations preserve the alignment of the header.
61 * ethdev: The function ``rte_eth_dev_count`` will be removed in DPDK 20.02.
62 It is replaced by the function ``rte_eth_dev_count_avail``.
63 If the intent is to iterate over ports, ``RTE_ETH_FOREACH_*`` macros
64 are better port iterators.
66 * ethdev: the legacy filter API, including
67 ``rte_eth_dev_filter_supported()``, ``rte_eth_dev_filter_ctrl()`` as well
68 as filter types MACVLAN, ETHERTYPE, FLEXIBLE, SYN, NTUPLE, TUNNEL, FDIR,
69 HASH and L2_TUNNEL, is superseded by the generic flow API (rte_flow) in
70 PMDs that implement the latter.
71 Target release for removal of the legacy API will be defined once most
72 PMDs have switched to rte_flow.
74 * ethdev: Update API functions returning ``void`` to return ``int`` with
75 negative errno values to indicate various error conditions (e.g.
76 invalid port ID, unsupported operation, failed operation):
78 - ``rte_eth_dev_stop``
79 - ``rte_eth_dev_close``
81 * ethdev: New offload flags ``DEV_RX_OFFLOAD_RSS_HASH`` and
82 ``DEV_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
83 This will allow application to enable or disable PMDs from updating
84 ``rte_mbuf::hash::rss`` and ``rte_mbuf::hash::fdir`` respectively.
85 This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
86 thereby improve Rx performance if application wishes do so.
87 In 19.11 PMDs will still update the fields even when the offloads are not
90 * ethdev: New function ``rte_eth_dev_set_supported_ptypes`` will be added in
92 This will allow application to request PMD to set specific ptypes defined
93 through ``rte_eth_dev_set_supported_ptypes`` in ``rte_mbuf::packet_type``.
94 If application doesn't want any ptype information it can call
95 ``rte_eth_dev_set_supported_ptypes(ethdev_id, RTE_PTYPE_UNKNOWN)`` and PMD
96 will set ``rte_mbuf::packet_type`` to ``0``.
97 If application doesn't call ``rte_eth_dev_set_supported_ptypes`` PMD can
98 return ``rte_mbuf::packet_type`` with ``rte_eth_dev_get_supported_ptypes``.
99 If application is interested only in L2/L3 layer, it can inform the PMD
100 to update ``rte_mbuf::packet_type`` with L2/L3 ptype by calling
101 ``rte_eth_dev_set_supported_ptypes(ethdev_id, RTE_PTYPE_L2_MASK | RTE_PTYPE_L3_MASK)``.
102 This scheme will allow PMDs to avoid lookup to internal ptype table on Rx and
103 thereby improve Rx performance if application wishes do so.
105 * ethdev: New 32-bit fields may be added for maximum LRO session size, in
106 struct ``rte_eth_dev_info`` for the port capability and in struct
107 ``rte_eth_rxmode`` for the port configuration.
109 * cryptodev: support for using IV with all sizes is added, J0 still can
110 be used but only when IV length in following structs ``rte_crypto_auth_xform``,
111 ``rte_crypto_aead_xform`` is set to zero. When IV length is greater or equal
112 to one it means it represents IV, when is set to zero it means J0 is used
113 directly, in this case 16 bytes of J0 need to be passed.
115 * sched: To allow more traffic classes, flexible mapping of pipe queues to
116 traffic classes, and subport level configuration of pipes and queues
117 changes will be made to macros, data structures and API functions defined
118 in "rte_sched.h". These changes are aligned to improvements suggested in the
119 RFC https://mails.dpdk.org/archives/dev/2018-November/120035.html.
121 * metrics: The function ``rte_metrics_init`` will have a non-void return
122 in order to notify errors instead of calling ``rte_exit``.
124 * power: ``rte_power_set_env`` function will no longer return 0 on attempt
125 to set new power environment if power environment was already initialized.
126 In this case the function will return -1 unless the environment is unset first
127 (using ``rte_power_unset_env``). Other function usage scenarios will not change.