xref: /dpdk/doc/guides/nics/fail_safe.rst (revision db27370b57202632ad8830352c1c0ee2dde4542f)
1..  SPDX-License-Identifier: BSD-3-Clause
2    Copyright 2017 6WIND S.A.
3
4Fail-safe poll mode driver library
5==================================
6
7The Fail-safe poll mode driver library (**librte_net_failsafe**) implements a
8virtual device that allows using device supporting hotplug, without modifying
9other components relying on such device (application, other PMDs).
10In this context, hotplug support is meant as plugging or removing a device
11from its bus suddenly.
12
13Additionally to the Seamless Hotplug feature, the Fail-safe PMD offers the
14ability to redirect operations to a secondary device when the primary has been
15removed from the system.
16
17
18Features
19--------
20
21The Fail-safe PMD only supports a limited set of features. If you plan to use a
22device underneath the Fail-safe PMD with a specific feature, this feature must
23also be supported by the Fail-safe PMD.
24
25A notable exception is the device removal feature. The fail-safe PMD is not
26meant to be removed itself, unlike its sub-devices which should support it.
27If a sub-device supports hotplugging, the fail-safe PMD will enable its use
28automatically by detecting capable devices and registering the relevant handler.
29
30Check the feature matrix for the complete set of supported features.
31
32
33Using the Fail-safe PMD from the EAL command line
34-------------------------------------------------
35
36The Fail-safe PMD can be used like most other DPDK virtual devices, by passing a
37``--vdev`` parameter to the EAL when starting the application. The device name
38must start with the *net_failsafe* prefix, followed by numbers or letters. This
39name must be unique for each device. Each fail-safe instance must have at least one
40sub-device, and at most two.
41
42A sub-device can be any DPDK device, including possibly another fail-safe device.
43
44Fail-safe command line parameters
45~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
46
47- **dev(<iface>)** parameter
48
49  This parameter allows the user to define a sub-device. The ``<iface>`` part of
50  this parameter must be a valid device definition. It follows the same format
51  provided to any ``-a`` or ``--vdev`` options.
52
53  Enclosing the device definition within parentheses here allows using
54  additional sub-device parameters if need be. They will be passed on to the
55  sub-device.
56
57.. note::
58
59   In case where the sub-device is also used as an allowed device, using ``-a``
60   on the EAL command line, the fail-safe PMD will use the device with the
61   options provided to the EAL instead of its own parameters.
62
63   When trying to use a PCI device automatically probed by the command line,
64   the name for the fail-safe sub-device must be the full PCI id:
65   Domain:Bus:Device.Function, *i.e.* ``00:00:00.0`` instead of ``00:00.0``,
66   as the second form is historically accepted by the DPDK.
67
68- **exec(<shell command>)** parameter
69
70  This parameter allows the user to provide a command to the fail-safe PMD to
71  execute and define a sub-device.
72  It is done within a regular shell context.
73  The first line of its output is read by the fail-safe PMD and otherwise
74  interpreted as if passed to a **dev** parameter.
75  Any other line is discarded.
76  If the command fails or output an incorrect string, the sub-device is not
77  initialized.
78  All commas within the ``shell command`` are replaced by spaces before
79  executing the command. This helps using scripts to specify devices.
80
81- **fd(<file descriptor number>)** parameter
82
83  This parameter reads a device definition from an arbitrary file descriptor
84  number in ``<iface>`` format as described above.
85
86  The file descriptor is read in non-blocking mode and is never closed in
87  order to take only the last line into account (unlike ``exec()``) at every
88  probe attempt.
89
90- **mac** parameter [MAC address]
91
92  This parameter allows the user to set a default MAC address to the fail-safe
93  and all of its sub-devices.
94  If no default mac address is provided, the fail-safe PMD will read the MAC
95  address of the first of its sub-device to be successfully probed and use it as
96  its default MAC address, trying to set it to all of its other sub-devices.
97  If no sub-device was successfully probed at initialization, then a random MAC
98  address is generated, that will be subsequently applied to all sub-devices once
99  they are probed.
100
101- **hotplug_poll** parameter [UINT64] (default **2000**)
102
103  This parameter allows the user to configure the amount of time in milliseconds
104  between two sub-device upkeep round.
105
106Usage example
107~~~~~~~~~~~~~
108
109This section shows some example of using **testpmd** with a fail-safe PMD.
110
111#. To build a PMD and configure DPDK, refer to the document
112   :ref:`compiling and testing a PMD for a NIC <pmd_build_and_test>`.
113
114#. Start testpmd. The sub-device ``84:00.0`` should be blocked from normal EAL
115   operations to avoid probing it twice, as the PCI bus is in blocklist mode.
116
117   .. code-block:: console
118
119      ./<build_dir>/app/dpdk-testpmd -c 0xff -n 4 \
120         --vdev 'net_failsafe0,mac=de:ad:be:ef:01:02,dev(84:00.0),dev(net_ring0)' \
121         -b 84:00.0 -b 00:04.0 -- -i
122
123   If the sub-device ``84:00.0`` is not blocked, it will be probed by the
124   EAL first. When the fail-safe then tries to initialize it the probe operation
125   fails.
126
127   Note that PCI blocklist mode is the default PCI operating mode.
128
129#. Alternatively, it can be used alongside any other device in allow mode.
130
131   .. code-block:: console
132
133      ./<build_dir>/app/dpdk-testpmd -c 0xff -n 4 \
134         --vdev 'net_failsafe0,mac=de:ad:be:ef:01:02,dev(84:00.0),dev(net_ring0)' \
135         -a 81:00.0 -- -i
136
137#. Start testpmd using a flexible device definition
138
139   .. code-block:: console
140
141      ./<build_dir>/app/dpdk-testpmd -c 0xff -n 4 -a ff:ff.f \
142         --vdev='net_failsafe0,exec(echo 84:00.0)' -- -i
143
144#. Start testpmd, automatically probing the device 84:00.0 and using it with
145   the fail-safe.
146
147   .. code-block:: console
148
149      ./<build_dir>/app/dpdk-testpmd -c 0xff -n 4 \
150         --vdev 'net_failsafe0,dev(0000:84:00.0),dev(net_ring0)' -- -i
151
152
153Using the Fail-safe PMD from an application
154-------------------------------------------
155
156This driver strives to be as seamless as possible to existing applications, in
157order to propose the hotplug functionality in the easiest way possible.
158
159Care must be taken, however, to respect the **ether** API concerning device
160access, and in particular, using the ``RTE_ETH_FOREACH_DEV`` macro to iterate
161over ethernet devices, instead of directly accessing them or by writing one's
162own device iterator.
163
164   .. code-block:: C
165
166      unsigned int i;
167
168      /* VALID iteration over eth-dev. */
169      RTE_ETH_FOREACH_DEV(i) {
170              [...]
171      }
172
173      /* INVALID iteration over eth-dev. */
174      for (i = 0; i < RTE_MAX_ETHPORTS; i++) {
175              [...]
176      }
177
178Plug-in feature
179---------------
180
181A sub-device can be defined without existing on the system when the fail-safe
182PMD is initialized. Upon probing this device, the fail-safe PMD will detect its
183absence and postpone its use. It will then register for a periodic check on any
184missing sub-device.
185
186During this time, the fail-safe PMD can be used normally, configured and told to
187emit and receive packets. It will store any applied configuration but will fail
188to emit anything, returning ``0`` from its TX function. Any unsent packet must
189be freed.
190
191Upon the probing of its missing sub-device, the current stored configuration
192will be applied. After this configuration pass, the new sub-device will be
193synchronized with other sub-devices, i.e. be started if the fail-safe PMD has
194been started by the user before.
195
196Plug-out feature
197----------------
198
199A sub-device supporting the device removal event can be removed from its bus at
200any time. The fail-safe PMD will register a callback for such event and react
201accordingly. It will try to safely stop, close and uninit the sub-device having
202emitted this event, allowing it to free its eventual resources.
203
204Fail-safe glossary
205------------------
206
207Fallback device
208    Also called **Secondary device**.
209
210    The fail-safe will fail-over onto this device when the preferred device is
211    absent.
212
213Preferred device
214    Also called **Primary device**.
215
216    The first declared sub-device in the fail-safe parameters.
217    When this device is plugged, it is always used as emitting device.
218    It is the main sub-device and is used as target for configuration
219    operations if there is any ambiguity.
220
221Upkeep round
222    Periodical event during which sub-devices are serviced. Each devices having a state
223    different to that of the fail-safe device itself, is synchronized with it
224    (brought down or up accordingly). Additionally, any sub-device marked for
225    removal is cleaned-up.
226
227Slave
228    In the context of the fail-safe PMD, synonymous to sub-device.
229
230Sub-device
231    A device being utilized by the fail-safe PMD.
232    This is another PMD running underneath the fail-safe PMD.
233    Any sub-device can disappear at any time. The fail-safe will ensure
234    that the device removal happens gracefully.
235