+.IP
+The following actions are Nicira vendor extensions that, as of this writing, are
+only known to be implemented by Open vSwitch:
+.
+.RS
+.
+.IP \fBresubmit\fB:\fIport\fR
+Re-searches the OpenFlow flow table with the \fBin_port\fR field
+replaced by \fIport\fR and executes the actions found, if any, in
+addition to any other actions in this flow entry. Recursive
+\fBresubmit\fR actions are ignored.
+.
+.IP \fBset_tunnel\fB:\fIid\fR
+.IQ \fBset_tunnel64\fB:\fIid\fR
+If outputting to a port that encapsulates the packet in a tunnel and
+supports an identifier (such as GRE), sets the identifier to \fBid\fR.
+If the \fBset_tunnel\fR form is used and \fIid\fR fits in 32 bits,
+then this uses an action extension that is supported by Open vSwitch
+1.0 and later. Otherwise, if \fIid\fR is a 64-bit value, it requires
+Open vSwitch 1.1 or later.
+.
+.IP \fBdrop_spoofed_arp\fR
+Stops processing further actions, if the packet being processed is an
+Ethernet+IPv4 ARP packet for which the source Ethernet address inside
+the ARP packet differs from the source Ethernet address in the
+Ethernet header.
+.
+This is useful because OpenFlow does not provide a way to match on the
+Ethernet addresses inside ARP packets, so there is no other way to
+drop spoofed ARPs other than sending every ARP packet to a controller.
+.
+.IP \fBset_queue\fB:\fIqueue\fR
+Sets the queue that should be used to \fIqueue\fR when packets are
+output. The number of supported queues depends on the switch; some
+OpenFlow implementations do not support queuing at all.
+.
+.IP \fBpop_queue\fR
+Restores the queue to the value it was before any \fBset_queue\fR
+actions were applied.
+.
+.IP \fBnote:\fR[\fIhh\fR]...
+Does nothing at all. Any number of bytes represented as hex digits
+\fIhh\fR may be included. Pairs of hex digits may be separated by
+periods for readability.
+.
+.IP "\fBmove:\fIsrc\fB[\fIstart\fB..\fIend\fB]->\fIdst\fB[\fIstart\fB..\fIend\fB]\fR"
+Copies the named bits from field \fIsrc\fR to field \fIdst\fR.
+\fIsrc\fR and \fIdst\fR must be NXM field names as defined in
+\fBnicira\-ext.h\fR, e.g. \fBNXM_OF_UDP_SRC\fR or \fBNXM_NX_REG0\fR.
+Each \fIstart\fR and \fIend\fR pair, which are inclusive, must specify
+the same number of bits and must fit within its respective field.
+Shorthands for \fB[\fIstart\fB..\fIend\fB]\fR exist: use
+\fB[\fIbit\fB]\fR to specify a single bit or \fB[]\fR to specify an
+entire field.
+.IP
+Examples: \fBmove:NXM_NX_REG0[0..5]\->NXM_NX_REG1[26..31]\fR copies the
+six bits numbered 0 through 5, inclusive, in register 0 into bits 26
+through 31, inclusive;
+\fBmove:NXM_NX_REG0[0..15]->NXM_OF_VLAN_TCI[]\fR copies the least
+significant 16 bits of register 0 into the VLAN TCI field.
+.
+.IP "\fBload:\fIvalue\fB\->\fIdst\fB[\fIstart\fB..\fIend\fB]"
+Writes \fIvalue\fR to bits \fIstart\fR through \fIend\fR, inclusive,
+in field \fBdst\fR.
+.IP
+Example: \fBload:55\->NXM_NX_REG2[0..5]\fR loads value 55 (bit pattern
+\fB110111\fR) into bits 0 through 5, inclusive, in register 2.
+.
+.IP "\fBmultipath(\fIfields\fB, \fIbasis\fB, \fIalgorithm\fB, \fIn_links\fB, \fIarg\fB, \fIdst\fB[\fIstart\fB..\fIend\fB])\fR"
+Hashes \fIfields\fR using \fIbasis\fR as a universal hash parameter,
+then the applies multipath link selection \fIalgorithm\fR (with
+parameter \fIarg\fR) to choose one of \fIn_links\fR output links
+numbered 0 through \fIn_links\fR minus 1, and stores the link into
+\fIdst\fB[\fIstart\fB..\fIend\fB]\fR, which must be an NXM register as
+described above.
+.IP
+Currently, \fIfields\fR must be either \fBeth_src\fR or
+\fBsymmetric_l4\fR and \fIalgorithm\fR must be one of \fBmodulo_n\fR,
+\fBhash_threshold\fR, \fBhrw\fR, and \fBiter_hash\fR. Only
+the \fBiter_hash\fR algorithm uses \fIarg\fR.
+.IP
+Refer to \fBnicira\-ext.h\fR for more details.
+.RE
+.