1Linux Base Driver for the Intel(R) Ethernet Controller XL710 Family
2===================================================================
3
4Intel i40e Linux driver.
5Copyright(c) 2013 Intel Corporation.
6
7Contents
8========
9
10- Identifying Your Adapter
11- Additional Configurations
12- Performance Tuning
13- Known Issues
14- Support
15
16
17Identifying Your Adapter
18========================
19
20The driver in this release is compatible with the Intel Ethernet
21Controller XL710 Family.
22
23For more information on how to identify your adapter, go to the Adapter &
24Driver ID Guide at:
25
26    http://support.intel.com/support/network/sb/CS-012904.htm
27
28
29Enabling the driver
30===================
31
32The driver is enabled via the standard kernel configuration system,
33using the make command:
34
35     make config/oldconfig/menuconfig/etc.
36
37The driver is located in the menu structure at:
38
39	-> Device Drivers
40	  -> Network device support (NETDEVICES [=y])
41	    -> Ethernet driver support
42	      -> Intel devices
43	        -> Intel(R) Ethernet Controller XL710 Family
44
45Additional Configurations
46=========================
47
48  Generic Receive Offload (GRO)
49  -----------------------------
50  The driver supports the in-kernel software implementation of GRO.  GRO has
51  shown that by coalescing Rx traffic into larger chunks of data, CPU
52  utilization can be significantly reduced when under large Rx load.  GRO is
53  an evolution of the previously-used LRO interface.  GRO is able to coalesce
54  other protocols besides TCP.  It's also safe to use with configurations that
55  are problematic for LRO, namely bridging and iSCSI.
56
57  Ethtool
58  -------
59  The driver utilizes the ethtool interface for driver configuration and
60  diagnostics, as well as displaying statistical information. The latest
61  ethtool version is required for this functionality.
62
63  The latest release of ethtool can be found from
64  https://www.kernel.org/pub/software/network/ethtool
65
66
67  Flow Director n-ntuple traffic filters (FDir)
68  ---------------------------------------------
69  The driver utilizes the ethtool interface for configuring ntuple filters,
70  via "ethtool -N <device> <filter>".
71
72  The sctp4, ip4, udp4, and tcp4 flow types are supported with the standard
73  fields including src-ip, dst-ip, src-port and dst-port. The driver only
74  supports fully enabling or fully masking the fields, so use of the mask
75  fields for partial matches is not supported.
76
77  Additionally, the driver supports using the action to specify filters for a
78  Virtual Function. You can specify the action as a 64bit value, where the
79  lower 32 bits represents the queue number, while the next 8 bits represent
80  which VF. Note that 0 is the PF, so the VF identifier is offset by 1. For
81  example:
82
83    ... action 0x800000002 ...
84
85  Would indicate to direct traffic for Virtual Function 7 (8 minus 1) on queue
86  2 of that VF.
87
88  The driver also supports using the user-defined field to specify 2 bytes of
89  arbitrary data to match within the packet payload in addition to the regular
90  fields. The data is specified in the lower 32bits of the user-def field in
91  the following way:
92
93  +----------------------------+---------------------------+
94  | 31    28    24    20    16 | 15    12     8     4     0|
95  +----------------------------+---------------------------+
96  | offset into packet payload |  2 bytes of flexible data |
97  +----------------------------+---------------------------+
98
99  As an example,
100
101    ... user-def 0x4FFFF ....
102
103  means to match the value 0xFFFF 4 bytes into the packet payload. Note that
104  the offset is based on the beginning of the payload, and not the beginning
105  of the packet. Thus
106
107    flow-type tcp4 ... user-def 0x8BEAF ....
108
109  would match TCP/IPv4 packets which have the value 0xBEAF 8bytes into the
110  TCP/IPv4 payload.
111
112  For ICMP, the hardware parses the ICMP header as 4 bytes of header and 4
113  bytes of payload, so if you want to match an ICMP frames payload you may need
114  to add 4 to the offset in order to match the data.
115
116  Furthermore, the offset can only be up to a value of 64, as the hardware
117  will only read up to 64 bytes of data from the payload. It must also be even
118  as the flexible data is 2 bytes long and must be aligned to byte 0 of the
119  packet payload.
120
121  When programming filters, the hardware is limited to using a single input
122  set for each flow type. This means that it is an error to program two
123  different filters with the same type that don't match on the same fields.
124  Thus the second of the following two commands will fail:
125
126    ethtool -N <device> flow-type tcp4 src-ip 192.168.0.7 action 5
127    ethtool -N <device> flow-type tcp4 dst-ip 192.168.15.18 action 1
128
129  This is because the first filter will be accepted and reprogram the input
130  set for TCPv4 filters, but the second filter will be unable to reprogram the
131  input set until all the conflicting TCPv4 filters are first removed.
132
133  Note that the user-defined flexible offset is also considered part of the
134  input set and cannot be programmed separately for multiple filters of the
135  same type. However, the flexible data is not part of the input set and
136  multiple filters may use the same offset but match against different data.
137
138  Data Center Bridging (DCB)
139  --------------------------
140  DCB configuration is not currently supported.
141
142  FCoE
143  ----
144  The driver supports Fiber Channel over Ethernet (FCoE) and Data Center
145  Bridging (DCB) functionality. Configuring DCB and FCoE is outside the scope
146  of this driver doc. Refer to http://www.open-fcoe.org/ for FCoE project
147  information and http://www.open-lldp.org/ or email list
148  e1000-eedc@lists.sourceforge.net for DCB information.
149
150  MAC and VLAN anti-spoofing feature
151  ----------------------------------
152  When a malicious driver attempts to send a spoofed packet, it is dropped by
153  the hardware and not transmitted.  An interrupt is sent to the PF driver
154  notifying it of the spoof attempt.
155
156  When a spoofed packet is detected the PF driver will send the following
157  message to the system log (displayed by  the "dmesg" command):
158
159  Spoof event(s) detected on VF (n)
160
161  Where n=the VF that attempted to do the spoofing.
162
163
164Performance Tuning
165==================
166
167An excellent article on performance tuning can be found at:
168
169http://www.redhat.com/promo/summit/2008/downloads/pdf/Thursday/Mark_Wagner.pdf
170
171
172Known Issues
173============
174
175
176Support
177=======
178
179For general information, go to the Intel support website at:
180
181    http://support.intel.com
182
183or the Intel Wired Networking project hosted by Sourceforge at:
184
185    http://e1000.sourceforge.net
186
187If an issue is identified with the released source code on the supported
188kernel with a supported adapter, email the specific information related
189to the issue to e1000-devel@lists.sourceforge.net and copy
190netdev@vger.kernel.org.
191