poll-loop: Make poll_fd_wait_event() cross-platform.
This is helpful if we want to wait either on 'fd' for POSIX or
events for Windows.
For Windows, if both 'fd' and 'wevent' is specified, we associate
that event with the 'fd' using WSAEventSelect() in poll_block().
So any 'events' on that 'fd' will wake us up from WaitForMultipleObjects().
WSAEventSelect() does not understand POLLIN, POLLOUT etc. Instead the
macros it understands are FD_READ, FD_ACCEPT, FD_CONNECT, FD_CLOSE etc.
So we need to make that transition.
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Tue, 11 Feb 2014 16:24:16 +0000 (08:24 -0800)]
ofproto-dpif-xlate: Make flows that match ICMP fields revalidate correctly.
ICMPv4 and ICMPv6 have 8-bit "type" and "code" fields. struct flow
uses the low 8 bits of the 16-bit tp_src and tp_dst members to
represent these fields. The datapath interface, on the other hand,
represents them with just 8 bits each. This means that if the high 8
bits of the masks for these fields somehow become set (meaning to
match on the nonexistent "high bits" of these fields) during
translation, then they will get chopped off by a round trip through
the datapath, and revalidation will spot that as an inconsistency and
delete the flow. This commit avoids the problem by making sure that
only the low 8 bits of either field can be unwildcarded for ICMP.
This seems like the minimal fix for this problem, appropriate for
backporting to earlier branches. The root of the issue is that these high
bits can get set in the match at all. I have some leads on that, but they
require more invasive changes elsewhere.
Bug #23320. Reported-by: Krishna Miriyala <miriyalak@vmware.com> Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Andy Zhou <azhou@nicira.com>
Ben Pfaff [Tue, 4 Feb 2014 17:01:16 +0000 (09:01 -0800)]
pcap-file: Allow capturing TCP streams where the SYN is not seen.
Until now, the tcp_stream() code has ignored any TCP packets that don't
correspond to a stream that began with a TCP SYN. This commit changes the
code so that any TCP packet that contains a payload starts a new stream.
Signed-off-by: Ben Pfaff <blp@nicira.com> Reported-by: Vasu Dasari <vdasari@gmail.com>
Ben Pfaff [Tue, 4 Feb 2014 20:39:37 +0000 (12:39 -0800)]
Always insert MPLS labels after VLAN tags.
OpenFlow 1.1 and 1.2 always inserted MPLS labels after VLAN tags.
OpenFlow 1.3 and 1.4 insert MPLS labels before VLAN tags.
OpenFlow 1.3.4 and 1.5, both in preparation, recognize that the change in
1.3 was an error and revert it. This commit implements that reversion
in Open vSwitch.
EXT-457. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Simon Horman <horms@verge.net.au>
YAMAMOTO Takashi [Fri, 24 Jan 2014 01:31:13 +0000 (10:31 +0900)]
daemon.at: Fix stderr races
When spawning ovsdb-server in background, redirect stderr to /dev/null.
Otherwise, the banner output (eg. "ovsdb-server (Open vSwitch) 2.1.90")
can mess stderr of the following commands and make these tests fail.
Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Signed-off-by: Ben Pfaff <blp@nicira.com>
daemon-windows: Ability to handle windows service calls.
The following code does not add any users yet.
The visioned workflow that this piece of code should work with is:
* Create a windows service through a startup script with
a tool like 'sc'
ex: sc create ovsdb-server binpath=
"C:\openvswitch\usr\sbin\ovsdb-server.exe -vconsole:off
-vsyslog:off -vfile:info --remote=ptcp:6632:127.0.0.1 --log-file
--service-monitor --service"
* Start the service from the startup script.
ex: sc start ovsdb-server
* Terminate the service during shutdown process.
ex: sc stop ovsdb-server
* Abrupt termination will restart the service.
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Ken Ajiro [Tue, 28 Jan 2014 01:20:43 +0000 (01:20 +0000)]
ovs-vsctl: Update will be discarded when multiple ovs-vsctl are executed
When two ovs-vsctl update map type column at same time, one ovs-vsctl's
update will be discarded although all ovs-vsctl succeeded. This patch
fixes this issue.
Signed-off-by: Ken Ajiro <ajiro@mxw.nes.nec.co.jp> Signed-off-by: Ben Pfaff <blp@nicira.com>
Simon Horman [Wed, 29 Jan 2014 06:13:19 +0000 (15:13 +0900)]
ofproto-dpif-xlate: Remove unused fitnessp pararameter from xlate_receive
Some functions pass a non-NULL value as this parameter
but none of those function uses the value xlate_receive()
returns there. So simply remove the parameter all together.
Also remove the now unused key_fitness field of struct flow_miss.
Signed-off-by: Simon Horman <horms@verge.net.au> Signed-off-by: Ben Pfaff <blp@nicira.com>
This function must cast a const value to a non const value.
By adding an uintptr_t cast the warning is suppressed.
To avoid the cast (proper solution) several function signatures
must be changed.
Signed-off-by: Daniele Di Proietto <daniele.di.proietto@gmail.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
This change, firstly, avoids declaring the formal parameter const,
since it is treated as non const. (to avoid -Wcast-qual)
Secondly, it cast the pointer from void* to u8*, since it is used
in arithmetic (to avoid -Wpointer-arith)
Signed-off-by: Daniele Di Proietto <daniele.di.proietto@gmail.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
Andy Zhou [Fri, 31 Jan 2014 23:47:58 +0000 (15:47 -0800)]
datapth: Suppress error messages on megaflow updates
With subfacets, we'd expect megaflow updates message to carry
the original micro flow. If not, EINVAL is returned and kernel
logs an error message. Now that the user space subfacet layer is
removed, it is expected that flow updates can arrive with a
micro flow other than the original. Change the return code to
EEXIST and remove the kernel error log message.
Reported-by: Ben Pfaff <blp@nicira.com> Signed-off-by: Andy Zhou <azhou@nicira.com>
Ben Pfaff [Tue, 10 Dec 2013 01:28:32 +0000 (17:28 -0800)]
socket-util: Remove unused functions.
A Windows porter mentioned to me that these functions caused special
trouble in the Windows port. However, they are no longer used, so we
might as well remove them.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Andy Zhou <azhou@nicira.com>
Pravin B Shelar [Tue, 28 Jan 2014 02:18:33 +0000 (18:18 -0800)]
datapath: Fix ovs_flow_free() ovs-lock assert.
ovs_flow_free() is not called under ovs-lock during packet
execute path (ovs_packet_cmd_execute()). Since packet execute
does not touch flow->mask, there is no need to take that
lock either. So move assert in case where flow->mask is checked.
Found by code inspection.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com> Acked-by: Andy Zhou <azhou@nicira.com> Acked-by: Jesse Gross <jesse@nicira.com>
cccl: Handle library paths for one shot compilation.
When one wishes to compile and link a program with an
external library in one shot, additional
option "-link" is expected after all the other options. For example,
Ben Pfaff [Fri, 31 Jan 2014 00:57:16 +0000 (16:57 -0800)]
bridge: Set ofport column in every database transaction.
Database transactions can occasionally fail due to concurrent changes in
the database. When that happens, the next transaction should repeat the
changes that ovs-vswitchd tried to make the first time (adjusted for the
changes to the database).
The code to report the OpenFlow port number in use didn't do that. It set
the ofport field once when it created the port and never set it again, even
if the transaction to set it failed. This commit fixes the problem.
Ethan Jackson [Tue, 28 Jan 2014 00:40:27 +0000 (16:40 -0800)]
ofproto-dpif-upcall: Hardcode max_idle to 1500ms.
Before this patch, OVS tried to guess an optimal max idle time for
datapath flows based on the number of datapath flows relative to the
limit. This caused instability because the limit was based on the
dump duration which was affected by the max idle time. This patch
chooses instead to hardcode the max idle time to 1.5s except in
extreme case where the datapath flow limit is exceeded. 1.5s was
chosen to ensure pings occurring at once per second stay cached in the
datapath.
Signed-off-by: Ethan Jackson <ethan@nicira.com> Acked-by: Joe Stringer <joestringer@nicira.com>
getopt_long: Copy over the implementation from netbsd.
Windows does not have a getopt_long function. This commit
copies over the getopt_long implementation from netbsd with
some minor modifications and is used only on Windows platform.
Modifications on top of the version in NetBSD repo.
* Remove header files not available in Visual Studio.
* Remove some unwanted #defines.
* Add Open vSwitch specific header files like config.h, vlog.h, util.h
* Add the following #define's
define __UNCONST(a) ((void *)(unsigned long)(const void *)(a))
define _DIAGASSERT(q) ovs_assert(q)
define warnx VLOG_WARN
* Add extern declaration in getopt.h for optarg, optind.
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
syslog: Provide stub functions for openlog and syslog.
One option to implement openlog and syslog functionality in Windows
is to use windows event logger. But it looks like it involves changing
registry settings and in general looks complicated.
For the time being, do nothing for syslog. All the information needed for
debugging will be present through the 'file' option anyways.
We can start OVS daemons on Windows with "-vfile:info -vsyslog:off".
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Andy Zhou [Wed, 18 Dec 2013 21:55:25 +0000 (13:55 -0800)]
netdev-dummy: Add support for active stream
The dummy ports thus far only support passive connections. It can
listen for multiple incoming connection requests but not make active
connections. This patch adds support of active stream, so that a
dummy port can be configured with either passive or active connections.
The net result is that dummy ports can now connect to each other,
without being patch ports. This feature will be useful in adding test
cases of future commits.
commit c58cc9a460fd158e5250e59902e96ac677dc115f (datapath: Allow user space to
announce ability to accept unaligned Netlink messages) introduced
OVS_DP_ATTR_USER_FEATURES netlink attribute in datapath responses,
but the attribute size was not taken into account in ovs_dp_cmd_msg_size().
Signed-off-by: Daniele Di Proietto <daniele.di.proietto@gmail.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
Luigi Rizzo [Thu, 23 Jan 2014 16:24:03 +0000 (17:24 +0100)]
lib/pcap-file: add 'ovs_' prefix to pcap functions
This is done to avoid collisions and confusions with libpcap symbols,
like pcap_read()
Signed-off-by: Luigi Rizzo <rizzo@iet.unipi.it> Signed-off-by: Daniele Di Proietto <daniele.di.proietto@gmail.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Ethan Jackson [Wed, 22 Jan 2014 01:55:41 +0000 (17:55 -0800)]
Makefile: Use AM_V_GEN instead of echo.
The AM_V_GEN macro fits more cleanly with the automake silent rules
option. When enabled it will print "GEN <filename>" instead of simply
echoing the command as before.
Signed-off-by: Ethan Jackson <ethan@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Ansis Atteka [Tue, 21 Jan 2014 01:16:39 +0000 (17:16 -0800)]
ipsec: install iptables rules that set IPsec bit in skb mark
Without these two iptables rules (one for UDP encapsulated IPsec and
another for direct IPsec), ovs-vswitchd would incorrectly conclude
that GRE packet belonged to a plain GRE tunnel instead of IPsec GRE
tunnel.
Joe Stringer [Wed, 22 Jan 2014 06:50:49 +0000 (06:50 +0000)]
upcall: Cache the number of flows from the datapath.
Fetching the number of flows in the datapath has been causing
unnecessary contention on the kernel ovs_lock in recent TCP CRR tests.
This patch caches this number for up to 100ms in the userspace to reduce
such kernel calls.
Signed-off-by: Joe Stringer <joestringer@nicira.com> Co-authored-by: Jarno Rajahalme <jrajahalme@nicira.com> Signed-off--by: Jarno Rajahalme <jrajahalme@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Alex Wang [Tue, 21 Jan 2014 22:23:27 +0000 (14:23 -0800)]
bfd: Add bfd_src_ip and bfd_dst_ip.
This commit adds two new options, bfd_src_ip and bfd_dst_ip
respectively, which allows user to configure the source and
destination IP address of bfd control packet. If the user
specified address cannot be parsed, the default address
will be used.
Signed-off-by: Alex Wang <alexw@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ben Pfaff [Fri, 6 Dec 2013 00:59:13 +0000 (16:59 -0800)]
vlog: Avoid deadlock in vlog_init__() corner case.
Anything inside vlog_init__() that tried to log a message was going to
deadlock, since it would hit pthread_once() recursively and wait for the
previous call to complete. Unfortunately, there was a VLOG_ERR call inside
vlog_init__(), only called in the corner case where the system's clock was
wrong.
This fixes the problem by rearranging code so that the logging isn't
inside the "once-only" region.
Ben Pfaff [Fri, 6 Dec 2013 05:29:14 +0000 (21:29 -0800)]
ofp-util: Ignore out_group in flow_mods for commands other than delete.
OpenFlow 1.1 through 1.4 say that the out_group in a flow_mod, for commands
other than OFPFC_DELETE or OFPFC_DELETE_STRICT, should be ignored. OVS
actually reported an error. This fixes the problem.
Reported-by: chen zhang <3zhangchen9211@gmail.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Linda Sun [Fri, 10 Jan 2014 00:26:12 +0000 (16:26 -0800)]
poll-loop: Port to Windows.
Use WaitForMultipleObjects for polling on windows. This works on all kinds
of objects, e.g. sockets, files, especially ioctl calls to the kernel.
poll_fd_wait_event() is used if events need to be passed to pollfds. latch
is signaled with event, to be waited/polled by WaitForMultipleObjects() as
well. Changed array of fds to hmap to check for duplicate fds.
Signed-off-by: Linda Sun <lsun@vmware.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Andy Zhou [Fri, 10 Jan 2014 23:57:04 +0000 (15:57 -0800)]
datapath: Fix kernel panic on ovs_flow_free
Both mega flow mask's reference counter and per flow table mask list
should only be accessed when holding ovs_mutex() lock. However
this is not true with ovs_flow_table_flush(). The patch fixes this bug.
Reported-by: Joe Stringer <joestringer@nicira.com> Signed-off-by: Andy Zhou <azhou@nicira.com>
Ben Pfaff [Fri, 17 Jan 2014 01:21:55 +0000 (17:21 -0800)]
netdev-linux: Fix build break on RHEL 6.1.
Commit 73c85181d (netdev-linux: Read packet auxdata to obtain vlan_tid)
added #include <linux/if_packet.h> to this file, to get the definition
of PACKET_AUXDATA and some other definitions, but on RHEL 6.1 this
provoked compiler errors:
In file included from /usr/include/linux/rtnetlink.h:5,
from lib/netdev-linux.c:34:
/usr/include/linux/netlink.h:34: error: expected specifier-qualifier-list
before 'sa_family_t'
Since the old #includes worked everywhere, and this file already defined
its own versions of most of the new macros that it needed, this commit just
reverts the old #includes and adds the one macro definition it didn't
already have.
(RHEL 6.1 isn't necessarily the only platform where this is a problem, but
it's the first one for which we noticed the problem.)
This switches the definition of sockaddr_ll used from the Linux one, which
uses __be16 for sll_protocol, to the glibc one, which uses plain "unsigned
short int". This makes sparse complain (rightly), so this commit also
adds a sparse-specific header that uses ovs_be16 to prevent the warning.
Helmut Schaa [Fri, 6 Dec 2013 15:18:42 +0000 (16:18 +0100)]
vswitchd: Inherit parents mac address for fake bridges
When adding a physical port to the main bridge the mac address
of the bridge is updated. We can do the same for fake bridges by
copying the mac address of the parent bridge.
There exists only one fake bridge per vlan, hence it is safe
to copy the mac address of the parent bridge.
Signed-off-by: Helmut Schaa <helmut.schaa@googlemail.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Simon Horman [Wed, 15 Jan 2014 08:17:02 +0000 (17:17 +0900)]
netdev-linux: Read packet auxdata to obtain vlan_tid
If VLAN acceleration is used when the kernel receives a packet
then the outer-most VLAN tag will not be present in the packet
when it is received by netdev-linux. Rather, it will be present
in auxdata.
This patch uses recvmsg() instead of recv() to read auxdata for
each packet and if the vlan_tid is set then it is added to the packet.
Adding the vlan_tid makes use of headroom available
in the buffer parameter of rx_recv.
Signed-off-by: Simon Horman <horms@verge.net.au> Co-authored-by: Ben Pfaff <blp@nicira.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Simon Horman [Wed, 15 Jan 2014 08:17:00 +0000 (17:17 +0900)]
netdev_class: Pass a struct ofpbuf * to rx_recv()
Update the netdev_class so that struct ofpbuf * is passed to rx_recv()
to provide both the data and size of the data to read a packet into.
On success, update struct ofpbuf size inside netdev_class rx_recv
implementation and return 0. This moves logic from the caller.
On error a positive error code is returned, whereas previously
a negative error code was returned. This is a more common convention.
This patch should not have any behavioural changes.
This patch is in preparation for the netdev-linux variant of rx_recv()
making use of headroom in the struct ofpbuf * parameter to push a VLAN tag
obtained from auxdata.
Signed-off-by: Simon Horman <horms@verge.net.au> Co-authored-by: Ben Pfaff <blp@nicira.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Thomas Graf [Tue, 14 Jan 2014 09:27:02 +0000 (01:27 -0800)]
datapath: Pad OVS_PACKET_ATTR_PACKET if linear copy was performed
While the zerocopy method is correctly omitted if user space
does not support unaligned Netlink messages. The attribute is
still not padded correctly as skb_zerocopy() will not ensure
padding and the attribute size is no longer pre calculated
though nla_reserve() which ensured padding previously.
This patch applies appropriate padding if a linear data copy
was performed in skb_zerocopy().
Signed-off-by: Thomas Graf <tgraf@suug.ch> Acked-by: Zoltan Kiss <zoltan.kiss@citrix.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
YAMAMOTO Takashi [Wed, 15 Jan 2014 18:06:40 +0000 (10:06 -0800)]
ofproto-dpif-xlate: Avoid recursive acquisition of xlate_rwlock.
Currently xlate_rwlock is recursively acquired.
(xlate_send_packet -> ofproto_dpif_execute_actions -> xlate_actions)
Due to writer-preference in rwlock implementations, this causes
deadlock if another thread tries to acquire the lock exclusively
behind us.
This change avoids the problem by making xlate_send_packet drop
the lock before calling ofproto_dpif_execute_actions. This is the
simplest fix but opens a race window against port reconfigurations.
Given the way xlate_send_packet is currently used, the race does not
seem a big problem. An alternative would be passing down the
"xlate_rwlock is held" info to ofproto_dpif_execute_actions.
Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Signed-off-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Mon, 13 Jan 2014 19:21:12 +0000 (11:21 -0800)]
classifier: Use fat_rwlock instead of ovs_rwlock.
Jarno Rajahalme reported up to 40% performance gain on netperf TCP_CRR with
an earlier version of this patch in combination with a kernel NUMA patch,
together with a reduction in variance:
http://openvswitch.org/pipermail/dev/2014-January/035867.html
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ben Pfaff [Mon, 13 Jan 2014 19:17:55 +0000 (11:17 -0800)]
fat-rwlock: New big but fast synchronization primitive.
This implements a reader-writer lock that uses a lot of memory (128 to 192
bytes per thread that takes the lock) but avoids cache line bouncing when
taking the read side. Thus, a fat_rwlock is a good choice for rwlocks
taken frequently by readers.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ben Pfaff [Tue, 14 Jan 2014 22:35:48 +0000 (14:35 -0800)]
ovs-thread: Add new support for thread-specific data.
A couple of times I've wanted to create a dynamic data structure that has
thread-specific data, but I've not been able to do that because
PTHREAD_KEYS_MAX is so low (POSIX says at least 128, glibc is only a little
bigger at 1024). This commit introduces a new form of thread-specific data
that supports a large number of items.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Windows path uses backward slashes. Also, the executable name
has a .exe extension in it. While creating log files, we use
the program name to create log file names. It feels a little odd
to have log file names like ovsdb-server.exe.log etc. Using
_splitpath_s() is a way to have same log file names on both
windows and linux platforms.
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Flavio Leitner [Tue, 14 Jan 2014 02:22:07 +0000 (00:22 -0200)]
rhel: Enable DHCP support for internal ports.
The current initscripts ifup-ovs brings up internal ports as
an ordinary ethernet device, so BOOTPROTO=dhcp|none does not
consider any OVS/bridge detail.
Since DHCP requires a port in the bridge to reach the server,
bring up the required port before in the same way it does for
OVS bridge.
Ben Pfaff [Fri, 10 Jan 2014 23:17:43 +0000 (15:17 -0800)]
ofproto-dpif: Un-wildcard nw_frag only for protocols that have fragments.
The revalidator code in ofproto-dpif-upcall.c, in revalidate_ukey(),
deletes any datapath flow for which the kernel reports wildcarded bits
that userspace requires to be matched. Until now, a couple of pieces of
code in ofproto-dpif always marked nw_frag (which tracks whether a packet
is an IPv4 or IPV6 fragment) as exact-match. For non-IP protocols, this
wasn't meaningful, so adding such a flow to the datapath and then receiving
it back caused nw_frag to become wildcarded, so revalidate_ukey() always
deleted them.
This fixes the problem by only un-wildcarding nw_frag for protocols where
it is defined (IPv4 and IPv6).
Noticed while observing CFM traffic (which isn't IP based) over a tunnel.
Reported-by: Guolin Yang <gyang@vmware.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Fri, 10 Jan 2014 23:14:27 +0000 (15:14 -0800)]
tunnel: Un-wildcard only flags that really exist in tnl_xlate_init().
The revalidator code in ofproto-dpif-upcall.c, in revalidate_ukey(),
deletes any datapath flow for which the kernel reports wildcarded bits
that userspace requires to be matched. Until now, tnl_xlate_init() marked
every bit in the tunnel flags as required to be matched. Since most of
those bits don't actually have defined flags, adding such a flow to the
datapath and then receiving it back caused those bits to become wildcarded,
which meant that revalidate_ukey() always deleted them.
This fixes the problem by only un-wildcarding defined flags.
Reported-by: Guolin Yang <gyang@vmware.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
handle_upcalls() always installed a flow for every packet, as long as
the datapath didn't already have too many flows, but there are cases where
we don't want to do this:
- If we get multiple packets in a single microflow all in one batch
(perhaps due to GSO breaking up a large TCP packet for sending to
userspace, or for another reason), then we only need to install the
datapath flow once.
- For a slow-pathed flow received via a slow-path action in the kernel,
we know that the kernel flow is already there (because otherwise it
would have been received as "no match" instead of an action), so
there is no benefit to reinstalling it.
Noticed because a CFM slow-pathed flow was getting reinstalled every time
a CFM packet was received.
Reported-by: Guolin Yang <gyang@nicira.com> Signed-off-by: Ben Pfaff <blp@nicira.com>