Linux kernel mirror (for testing)
git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git
kernel
os
linux
Documentation: net: add flow control guide and document ethtool API
Introduce a new document, flow_control.rst, to provide a comprehensive
guide on Ethernet Flow Control in Linux. The guide explains how flow
control works, how autonegotiation resolves pause capabilities, and how
to configure it using ethtool and Netlink.
In parallel, document the pause and pause-stat attributes in the
ethtool.yaml netlink spec. This enables the ynl tool to generate
kernel-doc comments for the corresponding enums in the UAPI header,
making the C interface self-documenting.
Finally, replace the legacy flow control section in phy.rst with a
reference to the new document and add pointers in the relevant C source
files.
Signed-off-by: Oleksij Rempel <o.rempel@pengutronix.de>
Link: https://patch.msgid.link/20250924120241.724850-1-o.rempel@pengutronix.de
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
authored by
Oleksij Rempel
and committed by
Paolo Abeni
7bd80ed8
c5cb31c9
+453
-15
+27
Documentation/netlink/specs/ethtool.yaml
+27
Documentation/netlink/specs/ethtool.yaml
···
864
864
865
865
-
866
866
name: pause-stat
867
+
doc: Statistics counters for link-wide PAUSE frames (IEEE 802.3 Annex 31B).
867
868
attr-cnt-name: __ethtool-a-pause-stat-cnt
869
+
enum-name: ethtool-a-pause-stat
868
870
attributes:
869
871
-
870
872
name: unspec
···
877
875
type: pad
878
876
-
879
877
name: tx-frames
878
+
doc: Number of PAUSE frames transmitted.
880
879
type: u64
881
880
-
882
881
name: rx-frames
882
+
doc: Number of PAUSE frames received.
883
883
type: u64
884
884
-
885
885
name: pause
886
+
doc: Parameters for link-wide PAUSE (IEEE 802.3 Annex 31B).
886
887
attr-cnt-name: __ethtool-a-pause-cnt
888
+
enum-name: ethtool-a-pause
887
889
attributes:
888
890
-
889
891
name: unspec
···
899
893
nested-attributes: header
900
894
-
901
895
name: autoneg
896
+
doc: |
897
+
Acts as a mode selector for the driver.
898
+
On GET: indicates the driver's behavior. If true, the driver will
899
+
respect the negotiated outcome; if false, the driver will use a
900
+
forced configuration.
901
+
On SET: if true, the driver configures the PHY's advertisement based
902
+
on the rx and tx attributes. If false, the driver forces the MAC
903
+
into the state defined by the rx and tx attributes.
902
904
type: u8
903
905
-
904
906
name: rx
907
+
doc: |
908
+
Enable receiving PAUSE frames (pausing local TX).
909
+
On GET: reflects the currently preferred configuration state.
905
910
type: u8
906
911
-
907
912
name: tx
913
+
doc: |
914
+
Enable transmitting PAUSE frames (pausing peer TX).
915
+
On GET: reflects the currently preferred configuration state.
908
916
type: u8
909
917
-
910
918
name: stats
919
+
doc: |
920
+
Contains the pause statistics counters. The source of these
921
+
statistics is determined by stats-src.
911
922
type: nest
912
923
nested-attributes: pause-stat
913
924
-
914
925
name: stats-src
926
+
doc: |
927
+
Selects the source of the MAC statistics, values from
928
+
enum ethtool_mac_stats_src. This allows requesting statistics
929
+
from the individual components of the MAC Merge layer.
915
930
type: u32
916
931
-
917
932
name: eee
+373
Documentation/networking/flow_control.rst
+373
Documentation/networking/flow_control.rst
···
1
+
.. SPDX-License-Identifier: GPL-2.0
2
+
3
+
.. _ethernet-flow-control:
4
+
5
+
=====================
6
+
Ethernet Flow Control
7
+
=====================
8
+
9
+
This document is a practical guide to Ethernet Flow Control in Linux, covering
10
+
what it is, how it works, and how to configure it.
11
+
12
+
What is Flow Control?
13
+
=====================
14
+
15
+
Flow control is a mechanism to prevent a fast sender from overwhelming a
16
+
slow receiver with data, which would cause buffer overruns and dropped packets.
17
+
The receiver can signal the sender to temporarily stop transmitting, giving it
18
+
time to process its backlog.
19
+
20
+
Standards references
21
+
====================
22
+
23
+
Ethernet flow control mechanisms are specified across consolidated IEEE base
24
+
standards; some originated as amendments:
25
+
26
+
- Collision-based flow control is part of CSMA/CD in **IEEE 802.3**
27
+
(half-duplex).
28
+
- Link-wide PAUSE is defined in **IEEE 802.3 Annex 31B**
29
+
(originally **802.3x**).
30
+
- Priority-based Flow Control (PFC) is defined in **IEEE 802.1Q Clause 36**
31
+
(originally **802.1Qbb**).
32
+
33
+
In the remainder of this document, the consolidated clause numbers are used.
34
+
35
+
How It Works: The Mechanisms
36
+
============================
37
+
38
+
The method used for flow control depends on the link's duplex mode.
39
+
40
+
.. note::
41
+
The user-visible ``ethtool`` pause API described in this document controls
42
+
**link-wide PAUSE** (IEEE 802.3 Annex 31B) only. It does not control the
43
+
collision-based behavior that exists on half-duplex links.
44
+
45
+
1. Half-Duplex: Collision-Based Flow Control
46
+
--------------------------------------------
47
+
On half-duplex links, a device cannot send and receive simultaneously, so PAUSE
48
+
frames are not used. Flow control is achieved by leveraging the CSMA/CD
49
+
(Carrier Sense Multiple Access with Collision Detection) protocol itself.
50
+
51
+
* **How it works**: To inhibit incoming data, a receiving device can force a
52
+
collision on the line. When the sending station detects this collision, it
53
+
terminates its transmission, sends a "jam" signal, and then executes the
54
+
"Collision backoff and retransmission" procedure as defined in IEEE 802.3,
55
+
Section 4.2.3.2.5. This algorithm makes the sender wait for a random
56
+
period before attempting to retransmit. By repeatedly forcing collisions,
57
+
the receiver can effectively throttle the sender's transmission rate.
58
+
59
+
.. note::
60
+
While this mechanism is part of the IEEE standard, there is currently no
61
+
generic kernel API to configure or control it. Drivers should not enable
62
+
this feature until a standardized interface is available.
63
+
64
+
.. warning::
65
+
On shared-medium networks (e.g. 10BASE2, or twisted-pair networks using a
66
+
hub rather than a switch) forcing collisions inhibits traffic **across the
67
+
entire shared segment**, not just a single point-to-point link. Enabling
68
+
such behavior is generally undesirable.
69
+
70
+
2. Full-Duplex: Link-wide PAUSE (IEEE 802.3 Annex 31B)
71
+
------------------------------------------------------
72
+
On full-duplex links, devices can send and receive at the same time. Flow
73
+
control is achieved by sending a special **PAUSE frame**, defined by IEEE
74
+
802.3 Annex 31B. This mechanism pauses all traffic on the link and is therefore
75
+
called *link-wide PAUSE*.
76
+
77
+
* **What it is**: A standard Ethernet frame with a globally reserved
78
+
destination MAC address (``01-80-C2-00-00-01``). This address is in a range
79
+
that standard IEEE 802.1D-compliant bridges do not forward. However, some
80
+
unmanaged or misconfigured bridges have been reported to forward these
81
+
frames, which can disrupt flow control across a network.
82
+
83
+
* **How it works**: The frame contains a MAC Control opcode for PAUSE
84
+
(``0x0001``) and a ``pause_time`` value, telling the sender how long to
85
+
wait before sending more data frames. This time is specified in units of
86
+
"pause quantum", where one quantum is the time it takes to transmit 512 bits.
87
+
For example, one pause quantum is 51.2 microseconds on a 10 Mbit/s link,
88
+
and 512 nanoseconds on a 1 Gbit/s link. A ``pause_time`` of zero indicates
89
+
that the transmitter can resume transmission, even if a previous non-zero
90
+
pause time has not yet elapsed.
91
+
92
+
* **Who uses it**: Any full-duplex link, from 10 Mbit/s to multi-gigabit speeds.
93
+
94
+
3. Full-Duplex: Priority-based Flow Control (PFC) (IEEE 802.1Q Clause 36)
95
+
-------------------------------------------------------------------------
96
+
Priority-based Flow Control is an enhancement to the standard PAUSE mechanism
97
+
that allows flow control to be applied independently to different classes of
98
+
traffic, identified by their priority level.
99
+
100
+
* **What it is**: PFC allows a receiver to pause traffic for one or more of the
101
+
8 standard priority levels without stopping traffic for other priorities.
102
+
This is critical in data center environments for protocols that cannot
103
+
tolerate packet loss due to congestion (e.g., Fibre Channel over Ethernet
104
+
or RoCE).
105
+
106
+
* **How it works**: PFC uses a specific PAUSE frame format. It shares the same
107
+
globally reserved destination MAC address (``01-80-C2-00-00-01``) as legacy
108
+
PAUSE frames but uses a unique opcode (``0x0101``). The frame payload
109
+
contains two key fields:
110
+
111
+
- **``priority_enable_vector``**: An 8-bit mask where each bit corresponds to
112
+
one of the 8 priorities. If a bit is set to 1, it means the pause time
113
+
for that priority is active.
114
+
- **``time_vector``**: A list of eight 2-octet fields, one for each priority.
115
+
Each field specifies the ``pause_time`` for its corresponding priority,
116
+
measured in units of ``pause_quanta`` (the time to transmit 512 bits).
117
+
118
+
.. note::
119
+
When PFC is enabled for at least one priority on a port, the standard
120
+
**link-wide PAUSE** (IEEE 802.3 Annex 31B) must be disabled for that port.
121
+
The two mechanisms are mutually exclusive (IEEE 802.1Q Clause 36).
122
+
123
+
Configuring Flow Control
124
+
========================
125
+
126
+
Link-wide PAUSE and Priority-based Flow Control are configured with different
127
+
tools.
128
+
129
+
Configuring Link-wide PAUSE with ``ethtool`` (IEEE 802.3 Annex 31B)
130
+
-------------------------------------------------------------------
131
+
Use ``ethtool -a <interface>`` to view and ``ethtool -A <interface>`` to change
132
+
the link-wide PAUSE settings.
133
+
134
+
.. code-block:: bash
135
+
136
+
# View current link-wide PAUSE settings
137
+
ethtool -a eth0
138
+
139
+
# Enable RX and TX pause, with autonegotiation
140
+
ethtool -A eth0 autoneg on rx on tx on
141
+
142
+
**Key Configuration Concepts**:
143
+
144
+
* **Pause Autoneg vs Generic Autoneg**: ``ethtool -A ... autoneg {on,off}``
145
+
controls **Pause Autoneg** (Annex 31B) only. It is independent from the
146
+
**Generic link autonegotiation** configured with ``ethtool -s``. A device can
147
+
have Generic autoneg **on** while Pause Autoneg is **off**, and vice versa.
148
+
149
+
* **If Pause Autoneg is off** (``-A ... autoneg off``): the device will **not**
150
+
advertise pause in the PHY. The MAC PAUSE state is **forced** according to
151
+
``rx``/``tx`` and does not depend on partner capabilities or resolution.
152
+
Ensure the peer is configured complementarily for PAUSE to be effective.
153
+
154
+
* **If generic autoneg is off** but **Pause Autoneg is on**, the pause policy
155
+
is **remembered** by the kernel and applied later when Generic autoneg is
156
+
enabled again.
157
+
158
+
* **Autonegotiation Mode**: The PHY will *advertise* the ``rx`` and ``tx``
159
+
capabilities. The final active state is determined by what both sides of the
160
+
link agree on. See the "PHY (Physical Layer Transceiver)" section below,
161
+
especially the *Resolution* subsection, for details of the negotiation rules.
162
+
163
+
* **Forced Mode**: This mode is necessary when autonegotiation is not used or
164
+
not possible. This includes links where one or both partners have
165
+
autonegotiation disabled, or in setups without a PHY (e.g., direct
166
+
MAC-to-MAC connections). The driver bypasses PHY advertisement and
167
+
directly forces the MAC into the specified ``rx``/``tx`` state. The
168
+
configuration on both sides of the link must be complementary. For
169
+
example, if one side is set to ``tx on`` ``rx off``, the link partner must be
170
+
set to ``tx off`` ``rx on`` for flow control to function correctly.
171
+
172
+
Configuring PFC with ``dcb`` (IEEE 802.1Q Clause 36)
173
+
----------------------------------------------------
174
+
PFC is part of the Data Center Bridging (DCB) subsystem and is managed with the
175
+
``dcb`` tool (iproute2). Some deployments use ``dcbtool`` (lldpad) instead; this
176
+
document shows ``dcb(8)`` examples.
177
+
178
+
**Viewing PFC Settings**:
179
+
180
+
.. code-block:: text
181
+
182
+
$ dcb pfc show dev eth0
183
+
pfc-cap 8 macsec-bypass off delay 4096
184
+
prio-pfc 0:off 1:off 2:off 3:off 4:off 5:off 6:on 7:on
185
+
186
+
This shows the PFC state (on/off) for each priority (0-7).
187
+
188
+
**Changing PFC Settings**:
189
+
190
+
.. code-block:: bash
191
+
192
+
# Enable PFC on priorities 6 and 7, leaving others as they are
193
+
$ dcb pfc set dev eth0 prio-pfc 6:on 7:on
194
+
195
+
# Disable PFC for all priorities except 6 and 7
196
+
$ dcb pfc set dev eth0 prio-pfc all:off 6:on 7:on
197
+
198
+
Monitoring Flow Control
199
+
=======================
200
+
201
+
The standard way to check if flow control is actively being used is to view the
202
+
pause-related statistics.
203
+
204
+
**Monitoring Link-wide PAUSE**:
205
+
Use ``ethtool --include-statistics -a <interface>``.
206
+
207
+
.. code-block:: text
208
+
209
+
$ ethtool --include-statistics -a eth0
210
+
Pause parameters for eth0:
211
+
...
212
+
Statistics:
213
+
tx_pause_frames: 0
214
+
rx_pause_frames: 0
215
+
216
+
**Monitoring PFC**:
217
+
PFC statistics (sent and received frames per priority) are available
218
+
through the ``dcb`` tool.
219
+
220
+
.. code-block:: text
221
+
222
+
$ dcb pfc show dev eth0 requests indications
223
+
requests 0:0 1:0 2:0 3:1024 4:2048 5:0 6:0 7:0
224
+
indications 0:0 1:0 2:0 3:512 4:4096 5:0 6:0 7:0
225
+
226
+
The ``requests`` counters track transmitted PFC frames (TX), and the
227
+
``indications`` counters track received PFC frames (RX).
228
+
229
+
Link-wide PAUSE Autonegotiation Details
230
+
=======================================
231
+
232
+
The autonegotiation process for link-wide PAUSE is managed by the PHY and
233
+
involves advertising capabilities and resolving the outcome.
234
+
235
+
* Terminology (link-wide PAUSE):
236
+
237
+
- **Symmetric pause**: both directions are paused when requested (TX+RX
238
+
enabled).
239
+
- **Asymmetric pause**: only one direction is paused (e.g., RX-only or
240
+
TX-only).
241
+
242
+
In IEEE 802.3 advertisement/resolution, symmetric/asymmetric are encoded
243
+
using two bits (Pause/Asym) and resolved per the standard truth tables
244
+
below.
245
+
246
+
* **Advertisement**: The PHY advertises the MAC's flow control capabilities.
247
+
This is done using two bits in the advertisement register: "Symmetric
248
+
Pause" (Pause) and "Asymmetric Pause" (Asym). These bits should be
249
+
interpreted as a combined value, not as independent flags. The kernel
250
+
converts the user's ``rx`` and ``tx`` settings into this two-bit value as
251
+
follows:
252
+
253
+
.. code-block:: text
254
+
255
+
tx rx | Pause Asym
256
+
-------+-------------
257
+
0 0 | 0 0
258
+
0 1 | 1 1
259
+
1 0 | 0 1
260
+
1 1 | 1 0
261
+
262
+
* **Resolution**: After negotiation, the PHY reports the link partner's
263
+
advertised Pause and Asym bits. The final flow control mode is determined
264
+
by the combination of the local and partner advertisements, according to
265
+
the IEEE 802.3 standard:
266
+
267
+
.. code-block:: text
268
+
269
+
Local Device | Link Partner | Result
270
+
Pause Asym | Pause Asym |
271
+
-------------------+--------------------+---------
272
+
0 X | 0 X | Disabled
273
+
0 1 | 1 0 | Disabled
274
+
0 1 | 1 1 | TX only
275
+
1 0 | 0 X | Disabled
276
+
1 X | 1 X | TX + RX
277
+
1 1 | 0 1 | RX only
278
+
279
+
It is important to note that the advertised bits reflect the *current
280
+
configuration* of the MAC, which may not represent its full hardware
281
+
capabilities.
282
+
283
+
Kernel Policy: "Set and Trust"
284
+
==============================
285
+
286
+
The ethtool pause API is defined as a **wish policy** for
287
+
IEEE 802.3 link-wide PAUSE only. A user request is always accepted
288
+
as the preferred configuration, but it may not be possible to apply
289
+
it in all link states.
290
+
291
+
Key constraints:
292
+
293
+
- Link-wide PAUSE is not valid on half-duplex links.
294
+
- Link-wide PAUSE cannot be used together with Priority-based Flow Control
295
+
(PFC, IEEE 802.1Q Clause 36).
296
+
- If autonegotiation is active and the link is currently down, the future
297
+
mode is not yet known.
298
+
299
+
Because of these constraints, the kernel stores the requested setting
300
+
and applies it only when the link is in a compatible state.
301
+
302
+
Implications for userspace:
303
+
304
+
1. Set once (the "wish"): the requested Rx/Tx PAUSE policy is
305
+
remembered even if it cannot be applied immediately.
306
+
2. Applied conditionally: when the link comes up, the kernel enables
307
+
PAUSE only if the active mode allows it.
308
+
309
+
Component Roles in Flow Control
310
+
===============================
311
+
312
+
The configuration of flow control involves several components, each with a
313
+
distinct role.
314
+
315
+
The MAC (Media Access Controller)
316
+
---------------------------------
317
+
The MAC is the hardware component that actually sends and receives PAUSE
318
+
frames. Its capabilities define the upper limit of what the driver can support.
319
+
For link-wide PAUSE, MACs can vary in their support for symmetric (both
320
+
directions) or asymmetric (independent TX/RX) flow control.
321
+
322
+
For PFC, the MAC must be capable of generating and interpreting the
323
+
priority-based PAUSE frames and managing separate pause states for each
324
+
traffic class.
325
+
326
+
Many MACs also implement automatic PAUSE frame transmission based on the fill
327
+
level of their internal RX FIFO. This is typically configured with two
328
+
thresholds:
329
+
330
+
* **FLOW_ON (High Water Mark)**: When the RX FIFO usage reaches this
331
+
threshold, the MAC automatically transmits a PAUSE frame to stop the sender.
332
+
333
+
* **FLOW_OFF (Low Water Mark)**: When the RX FIFO usage drops below this
334
+
threshold, the MAC transmits a PAUSE frame with a quantum of zero to tell
335
+
the sender it can resume transmission.
336
+
337
+
The PHY (Physical Layer Transceiver)
338
+
------------------------------------
339
+
The PHY's role is distinct for each flow control mechanism:
340
+
341
+
* **Link-wide PAUSE**: During the autonegotiation process, the PHY is
342
+
responsible for advertising the device's flow control capabilities. See the
343
+
"Link-wide PAUSE Autonegotiation Details" section for more information.
344
+
345
+
* **Half-Duplex Collision-Based Flow Control**: The PHY is fundamental to the
346
+
CSMA/CD process. It performs carrier sensing (checking if the line is idle)
347
+
and collision detection, which is the mechanism leveraged to throttle the
348
+
sender.
349
+
350
+
* **Priority-based Flow Control (PFC)**: The PHY is not directly involved in
351
+
negotiating PFC capabilities. Its role is to establish the physical link.
352
+
PFC negotiation happens at a higher layer via the Data Center Bridging
353
+
Capability Exchange Protocol (DCBX).
354
+
355
+
User Space Interface
356
+
====================
357
+
The primary user space tools are ``ethtool`` for link-wide PAUSE and ``dcb`` for
358
+
PFC. They communicate with the kernel to configure the network device driver
359
+
and underlying hardware.
360
+
361
+
**Link-wide PAUSE Netlink Interface (``ethtool``)**
362
+
363
+
See the ethtool Netlink spec (``Documentation/netlink/specs/ethtool.yaml``)
364
+
for the authoritative definition of the Pause control and Pause statistics
365
+
attributes. The generated UAPI is in
366
+
``include/uapi/linux/ethtool_netlink_generated.h``.
367
+
368
+
**PFC Netlink Interface (``dcb``)**
369
+
370
+
The authoritative definitions for DCB/PFC netlink attributes and commands are in
371
+
``include/uapi/linux/dcbnl.h``. See also the ``dcb(8)`` manual page and the DCB
372
+
subsystem documentation for userspace configuration details.
373
+
+1
Documentation/networking/index.rst
+1
Documentation/networking/index.rst
+2
-10
Documentation/networking/phy.rst
+2
-10
Documentation/networking/phy.rst
···
343
343
Pause frames / flow control
344
344
===========================
345
345
346
-
The PHY does not participate directly in flow control/pause frames except by
347
-
making sure that the SUPPORTED_Pause and SUPPORTED_AsymPause bits are set in
348
-
MII_ADVERTISE to indicate towards the link partner that the Ethernet MAC
349
-
controller supports such a thing. Since flow control/pause frames generation
350
-
involves the Ethernet MAC driver, it is recommended that this driver takes care
351
-
of properly indicating advertisement and support for such features by setting
352
-
the SUPPORTED_Pause and SUPPORTED_AsymPause bits accordingly. This can be done
353
-
either before or after phy_connect() and/or as a result of implementing the
354
-
ethtool::set_pauseparam feature.
355
-
346
+
For detailed link-wide PAUSE and PFC behavior and configuration, see
347
+
flow_control.rst.
356
348
357
349
Keeping Close Tabs on the PAL
358
350
=============================
+42
-3
include/linux/ethtool.h
+42
-3
include/linux/ethtool.h
···
953
953
* @get_pause_stats: Report pause frame statistics. Drivers must not zero
954
954
* statistics which they don't report. The stats structure is initialized
955
955
* to ETHTOOL_STAT_NOT_SET indicating driver does not report statistics.
956
-
* @get_pauseparam: Report pause parameters
957
-
* @set_pauseparam: Set pause parameters. Returns a negative error code
958
-
* or zero.
956
+
*
957
+
* @get_pauseparam: Report the configured policy for link-wide PAUSE
958
+
* (IEEE 802.3 Annex 31B). Drivers must fill struct ethtool_pauseparam
959
+
* such that:
960
+
* @autoneg:
961
+
* This refers to **Pause Autoneg** (IEEE 802.3 Annex 31B) only
962
+
* and is independent of generic link autonegotiation configured
963
+
* via ethtool -s.
964
+
* true -> the device follows the negotiated result of pause
965
+
* autonegotiation (Pause/Asym);
966
+
* false -> the device uses a forced MAC state independent of
967
+
* negotiation.
968
+
* @rx_pause/@tx_pause:
969
+
* represent the desired policy (preferred configuration).
970
+
* In autoneg mode they describe what is to be advertised;
971
+
* in forced mode they describe the MAC state to apply.
972
+
*
973
+
* Drivers (and/or frameworks) should persist this policy across link
974
+
* changes and reapply appropriate MAC programming when link parameters
975
+
* change.
976
+
*
977
+
* @set_pauseparam: Apply a policy for link-wide PAUSE (IEEE 802.3 Annex 31B).
978
+
* If @autoneg is true:
979
+
* Arrange for pause advertisement (Pause/Asym) based on
980
+
* @rx_pause/@tx_pause and program the MAC to follow the
981
+
* negotiated result (which may be symmetric, asymmetric, or off
982
+
* depending on the link partner).
983
+
* If @autoneg is false:
984
+
* Do not rely on autonegotiation; force the MAC RX/TX pause
985
+
* state directly per @rx_pause/@tx_pause.
986
+
*
987
+
* Implementations that integrate with PHYLIB/PHYLINK should cooperate
988
+
* with those frameworks for advertisement and resolution; MAC drivers are
989
+
* still responsible for applying the required MAC state.
990
+
*
991
+
* Return: 0 on success or a negative errno. Return -EOPNOTSUPP if
992
+
* link-wide PAUSE is unsupported. If only symmetric pause is supported,
993
+
* reject unsupported asymmetric requests with -EINVAL (or document any
994
+
* coercion policy).
995
+
*
996
+
* See also: Documentation/networking/flow_control.rst
997
+
*
959
998
* @self_test: Run specified self-tests
960
999
* @get_strings: Return a set of strings that describe the requested objects
961
1000
* @set_phys_id: Identify the physical devices, e.g. by flashing an LED
+2
-2
include/uapi/linux/ethtool_netlink_generated.h
+2
-2
include/uapi/linux/ethtool_netlink_generated.h
···
375
375
ETHTOOL_A_COALESCE_MAX = (__ETHTOOL_A_COALESCE_CNT - 1)
376
376
};
377
377
378
-
enum {
378
+
enum ethtool_a_pause_stat {
379
379
ETHTOOL_A_PAUSE_STAT_UNSPEC,
380
380
ETHTOOL_A_PAUSE_STAT_PAD,
381
381
ETHTOOL_A_PAUSE_STAT_TX_FRAMES,
···
385
385
ETHTOOL_A_PAUSE_STAT_MAX = (__ETHTOOL_A_PAUSE_STAT_CNT - 1)
386
386
};
387
387
388
-
enum {
388
+
enum ethtool_a_pause {
389
389
ETHTOOL_A_PAUSE_UNSPEC,
390
390
ETHTOOL_A_PAUSE_HEADER,
391
391
ETHTOOL_A_PAUSE_AUTONEG,
+2
net/dcb/dcbnl.c
+2
net/dcb/dcbnl.c
···
27
27
*
28
28
* Priority-based Flow Control (PFC) - provides a flow control mechanism which
29
29
* can work independently for each 802.1p priority.
30
+
* See Documentation/networking/flow_control.rst for a high level description
31
+
* of the user space interface for Priority-based Flow Control (PFC).
30
32
*
31
33
* Congestion Notification - provides a mechanism for end-to-end congestion
32
34
* control for protocols which do not have built-in congestion management.