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 * make: Support for building DPDK with "make" has been deprecated and
15 support will be removed in the 20.11 release. From 20.11 onwards, DPDK
16 should be built using meson and ninja. For basic instructions see the
17 `Quick-start Guide <https://core.dpdk.org/doc/quick-start/>`_ on the
18 website or the `Getting Started Guide
19 <https://doc.dpdk.org/guides/linux_gsg/build_dpdk.html>`_ document.
21 * meson: The minimum supported version of meson for configuring and building
22 DPDK will be increased to v0.47.1 (from 0.41) from DPDK 19.05 onwards. For
23 those users with a version earlier than 0.47.1, an updated copy of meson
24 can be got using the ``pip``, or ``pip3``, tool for downloading python
27 * kvargs: The function ``rte_kvargs_process`` will get a new parameter
28 for returning key match count. It will ease handling of no-match case.
30 * eal: The function ``rte_eal_remote_launch`` will return new error codes
31 after read or write error on the pipe, instead of calling ``rte_panic``.
33 * eal: The ``rte_logs`` struct and global symbol will be made private to
34 remove it from the externally visible ABI and allow it to be updated in the
37 * rte_atomicNN_xxx: These APIs do not take memory order parameter. This does
38 not allow for writing optimized code for all the CPU architectures supported
39 in DPDK. DPDK will adopt C11 atomic operations semantics and provide wrappers
40 using C11 atomic built-ins. These wrappers must be used for patches that
41 need to be merged in 20.08 onwards. This change will not introduce any
42 performance degradation.
44 * rte_smp_*mb: These APIs provide full barrier functionality. However, many
45 use cases do not require full barriers. To support such use cases, DPDK will
46 adopt C11 barrier semantics and provide wrappers using C11 atomic built-ins.
47 These wrappers must be used for patches that need to be merged in 20.08
48 onwards. This change will not introduce any performance degradation.
50 * rte_cio_*mb: Since the IO barriers for ARMv8 platforms are relaxed from DSB
51 to DMB, rte_cio_*mb APIs provide the same functionality as rte_io_*mb
52 APIs (taking all platforms into consideration). rte_io_*mb APIs should be
53 used in the place of rte_cio_*mb APIs. The rte_cio_*mb APIs will be
54 deprecated in 20.11 release.
56 * igb_uio: In the view of reducing the kernel dependency from the main tree,
57 as a first step, the Technical Board decided to move ``igb_uio``
58 kernel module to the dpdk-kmods repository in the /linux/igb_uio/ directory
60 Minutes of Technical Board Meeting of `2019-11-06
61 <https://mails.dpdk.org/archives/dev/2019-November/151763.html>`_.
63 * lib: will fix extending some enum/define breaking the ABI. There are multiple
64 samples in DPDK that enum/define terminated with a ``.*MAX.*`` value which is
65 used by iterators, and arrays holding these values are sized with this
66 ``.*MAX.*`` value. So extending this enum/define increases the ``.*MAX.*``
67 value which increases the size of the array and depending on how/where the
68 array is used this may break the ABI.
69 ``RTE_ETH_FLOW_MAX`` is one sample of the mentioned case, adding a new flow
70 type will break the ABI because of ``flex_mask[RTE_ETH_FLOW_MAX]`` array
71 usage in following public struct hierarchy:
72 ``rte_eth_fdir_flex_conf -> rte_fdir_conf -> rte_eth_conf (in the middle)``.
73 Need to identify this kind of usages and fix in 20.11, otherwise this blocks
74 us extending existing enum/define.
75 One solution can be using a fixed size array instead of ``.*MAX.*`` value.
77 * pci: The ``rte_kernel_driver`` enum defined in rte_dev.h will be made private
78 to the PCI subsystem as it is used only by the PCI bus driver and PCI
80 The associated field ``kdrv`` in the ethdev ``rte_eth_dev_data`` structure
81 will be removed as it gave no useful abstracted information to the
82 applications and had no user (neither internal nor external).
84 * ethdev: Split the ``struct eth_dev_ops`` struct to hide it as much as possible
85 will be done in 20.11.
86 Currently the ``struct eth_dev_ops`` struct is accessible by the application
87 because some inline functions, like ``rte_eth_tx_descriptor_status()``,
88 access the struct directly.
89 The struct will be separate in two, the ops used by inline functions will be
90 moved next to Rx/Tx burst functions, rest of the ``struct eth_dev_ops`` struct
91 will be moved to header file for drivers to hide it from applications.
93 * ethdev: the legacy filter API, including
94 ``rte_eth_dev_filter_supported()``, ``rte_eth_dev_filter_ctrl()`` as well
95 as filter types MACVLAN, ETHERTYPE, FLEXIBLE, SYN, NTUPLE, TUNNEL, FDIR,
96 HASH and L2_TUNNEL, is superseded by the generic flow API (rte_flow) in
97 PMDs that implement the latter.
98 Target release for removal of the legacy API will be defined once most
99 PMDs have switched to rte_flow.
101 * ethdev: Update API functions returning ``void`` to return ``int`` with
102 negative errno values to indicate various error conditions (e.g.
103 invalid port ID, unsupported operation, failed operation):
105 - ``rte_eth_dev_stop``
106 - ``rte_eth_dev_close``
108 * ethdev: New offload flags ``DEV_RX_OFFLOAD_FLOW_MARK`` will be added in 19.11.
109 This will allow application to enable or disable PMDs from updating
110 ``rte_mbuf::hash::fdir``.
111 This scheme will allow PMDs to avoid writes to ``rte_mbuf`` fields on Rx and
112 thereby improve Rx performance if application wishes do so.
113 In 19.11 PMDs will still update the field even when the offload is not
116 * ethdev: ``rx_descriptor_done`` dev_ops and ``rte_eth_rx_descriptor_done``
117 will be deprecated in 20.11 and will be removed in 21.11.
118 Existing ``rte_eth_rx_descriptor_status`` and ``rte_eth_tx_descriptor_status``
119 APIs can be used as replacement.
121 * ethdev: Some internal APIs for driver usage are exported in the .map file.
122 Now DPDK has ``__rte_internal`` marker so we can mark internal APIs and move
123 them to the INTERNAL block in .map. Although these APIs are internal it will
124 break the ABI checks, that is why change is planned for 20.11.
125 The list of internal APIs are mainly ones listed in ``rte_ethdev_driver.h``.
127 * traffic manager: All traffic manager API's in ``rte_tm.h`` were mistakenly made
128 ABI stable in the v19.11 release. The TM maintainer and other contributors have
129 agreed to keep the TM APIs as experimental in expectation of additional spec
130 improvements. Therefore, all APIs in ``rte_tm.h`` will be marked back as
131 experimental in v20.11 DPDK release. For more details, please see `the thread
132 <https://mails.dpdk.org/archives/dev/2020-April/164970.html>`_.
134 * pmd_dpaa: The API ``rte_pmd_dpaa_set_tx_loopback`` will have extended
135 ``port_id`` definition from ``uint8_t`` to ``uint16_t``.
137 * cryptodev: support for using IV with all sizes is added, J0 still can
138 be used but only when IV length in following structs ``rte_crypto_auth_xform``,
139 ``rte_crypto_aead_xform`` is set to zero. When IV length is greater or equal
140 to one it means it represents IV, when is set to zero it means J0 is used
141 directly, in this case 16 bytes of J0 need to be passed.
143 * rawdev: The rawdev APIs which take a device-specific structure as
144 parameter directly, or indirectly via a "private" pointer inside another
145 structure, will be modified to take an additional parameter of the
146 structure size. The affected APIs will include ``rte_rawdev_info_get``,
147 ``rte_rawdev_configure``, ``rte_rawdev_queue_conf_get`` and
148 ``rte_rawdev_queue_setup``.
150 * sched: To allow more traffic classes, flexible mapping of pipe queues to
151 traffic classes, and subport level configuration of pipes and queues
152 changes will be made to macros, data structures and API functions defined
153 in "rte_sched.h". These changes are aligned to improvements suggested in the
154 RFC https://mails.dpdk.org/archives/dev/2018-November/120035.html.
156 * mbuf: ``refcnt_atomic`` member in structures ``rte_mbuf`` and
157 ``rte_mbuf_ext_shared_info`` is of type ``rte_atomic16_t``. Due to adoption
158 of C11 atomic builtins it will be of type ``uint16_t``. ``refcnt_atomic``
159 will be removed in 20.11. It will be replaced with ``refcnt`` of type
162 * metrics: The function ``rte_metrics_init`` will have a non-void return
163 in order to notify errors instead of calling ``rte_exit``.
165 * power: ``rte_power_set_env`` function will no longer return 0 on attempt
166 to set new power environment if power environment was already initialized.
167 In this case the function will return -1 unless the environment is unset first
168 (using ``rte_power_unset_env``). Other function usage scenarios will not change.
170 * python: Since the beginning of 2020, Python 2 has officially reached
171 end-of-support: https://www.python.org/doc/sunset-python-2/.
172 Python 2 support will be completely removed in 20.11.
173 In 20.08, explicit deprecation warnings will be displayed when running
174 scripts with Python 2.