In other words, murmurhash outperforms Jenkins for all input lengths other
than exactly 3 32-bit words (12 bytes). (It's understandable that Jenkins
would have a best case at 12 bytes, because Jenkins works in 12-byte
chunks.) Even in the case where Jenkins is faster, it's only by 5%. On
average within this data set, murmurhash is 15% faster, and for 4-word
input it is 30% faster.
We retain Jenkins for flow_hash_symmetric_l4() and flow_hash_fields(),
which are cases where the hash value is exposed externally.
This commit appears to improve "ovs-benchmark rate" results slightly by
a few hundred connections per second (under 1%), when used with an NVP
controller.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ben Pfaff [Fri, 14 Dec 2012 22:50:06 +0000 (14:50 -0800)]
hash: Change mhash_finish() data measurement from words to bytes.
murmurhash includes an xor with the number of bytes hashed in its finishing
step. Until now, we've only had murmurhash for full words, but an upcoming
commit adds murmurhash for bytes, so the finishing function will need to
take a count of bytes instead.
Ben Pfaff [Fri, 14 Dec 2012 21:43:54 +0000 (13:43 -0800)]
hash: Correct implementation of mhash_finish().
With rotates instead of shifts, the upper and lower 16 bits of the returned
hash are always the same.
I noticed this while working on replacing Jenkins hash by murmurhash,
because some of the database unit tests started failing, e.g. "row
hashing (scalars)".
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ben Pfaff [Wed, 16 Jan 2013 05:48:25 +0000 (21:48 -0800)]
tests: Fix dependencies on hash function in ofproto-dpif tests.
These tests relied on luck to ensure that OpenFlow ports received the
expected OpenFlow port numbers. With a different hash function, or (I
expect) on a big-endian architecture, the port numbers were assigned
differently and the tests failed.
This commit fixes the problem by requesting the specific expected port
numbers explicitly.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Jesse Gross [Mon, 21 Jan 2013 11:59:17 +0000 (03:59 -0800)]
datapath: Return correct error code when dumping flow actions.
Currently, if there isn't enough space to store the actions in a
flow during a dump we return -ENOMEM. However, the standard error
in this situation is -EMSGSIZE so this changes the behavior to match.
This issue was introduced in 354d4c98a8cdaae3525848f564e58a9016bcd3af
(datapath: Fix nelink attribute size for flow.).
Signed-off-by: Jesse Gross <jesse@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Mon, 21 Jan 2013 11:59:16 +0000 (03:59 -0800)]
datapath: Don't dump partial action lists in flows.
After commit 9b405f1aa8d175dc63ad3ffe5d0fe05d5ee09162 (datapath: More
flexible kernel/userspace tunneling attribute.), it was possible for a
flow dump to return a partial action list. It's better to return no
action list at all in this situation since then userspace will know
that it should request the full thing if it wants rather than have
incorrect results. Therefore, this prevents those partial lists in
situations where we have a very large number of actions.
Signed-off-by: Ben Pfaff <blp@nicira.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
In general, it should not be possible have a NULL return value from
skb_gso_segment() since we're not actually trying to verify the
header integrity. No other callers with similar needs have NULL
checks. The actual cause of the problem was LRO packets, which
OVS isn't equipped to handle. The commit 33e031e99cc630baf1b0cb9256710dee7d9ab66d (datapath: Move LRO check
from transmit to receive.) solves that problem by fixing the LRO
check. In order to avoid possibly masking any other problems, this
reverts the GSO check which should no longer be needed.
Jesse Gross [Mon, 21 Jan 2013 13:13:32 +0000 (05:13 -0800)]
datapath: Move LRO check from transmit to receive.
Commit 24b019f808211a95078efd916064af0975ca5733 (datapath: Disable
LRO from userspace instead of the kernel.) accidentally moved the
check for LRO packets from the receive path to transmit. Since
this check is supposed to protect OVS (and other parts of the system)
from packets that it cannot handle it is obviously not useful on
egress. Therefore, this commit moves it back to the receive side.
The primary problem that this caused is upcalls to userspace tried
to segment the packet even though no segmentation information is
available. This would later cause NULL pointer dereferences when
skb_gso_segment() did nothing.
Bug #14772
Signed-off-by: Jesse Gross <jesse@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Mon, 21 Jan 2013 22:29:15 +0000 (14:29 -0800)]
datapath: Avoid null deref when GSO is for verifying header integrity only.
skb_gso_segment() has the following comment:
* It may return NULL if the skb requires no segmentation. This is
* only possible when GSO is used for verifying header integrity.
Somehow queue_gso_packets() has never hit this case before, but some
failures have suddenly been reported. This commit should fix the problem.
Additional commentary by Jesse: We shouldn't normally be hitting this case
because we're actually trying to do GSO, not header validation. However, I
guess the guest/backend must be generating a packet with an MSS, which
tricks us into thinking that it's GSO, but no GSO is actually requested.
In the case of the bridge, header validation does take place so the
situation is handled already. It seems not ideal that the network backend
doesn't sanitize these packets but it's probably good that we handle
it in any case.
Pravin B Shelar [Mon, 21 Jan 2013 07:14:36 +0000 (23:14 -0800)]
datapath: Fix Flow dump operation.
Following patch adds null check while inserting new netlink attribute.
This was introduced by commit 9b405f1aa8d175d (datapath: More
flexible kernel/userspace tunneling attribute.)
Signed-off-by: Pravin B Shelar <pshelar@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Bug #14767
Pravin B Shelar [Sat, 19 Jan 2013 02:10:59 +0000 (18:10 -0800)]
datapath: More flexible kernel/userspace tunneling attribute.
Following patch breaks down single ipv4_tunnel netlink attribute into
individual member attributes. It will help when we extend tunneling
parameters in future.
debian: Change openvswitch restart logic for kernel package.
Currently, every time we install the non-dkms openvswitch kernel package,
we restart openvswitch. This has no useful effect when an Open vSwitch kernel
module is already loaded, because "/etc/init.d/openvswitch restart" does not
load the new version of the kernel module.
With this patch, we start openvswitch if there is no kernel module
currently loaded (which is a case for a fresh install of openvswitch). This
will load the kernel module.
If someone wants to use the latest kernel module after an upgrade, they need
to run "/etc/init.d/openvswitch-switch force-reload-kmod" manually.
Justin Pettit [Thu, 17 Jan 2013 08:31:20 +0000 (00:31 -0800)]
ofproto: Add "actions=" to drop action in "ovs-appctl bridge/dump-flows".
The function flow_stats_ds() did its own formatting when there were no
actions, but this didn't include "action=". There's actually no reason
to special-case this, since ofpact_format() properly handles actions
with a zero length and prints it properly.
Justin Pettit [Wed, 16 Jan 2013 23:53:14 +0000 (15:53 -0800)]
dpif-linux: Fix segfault when a port already exists.
Commit 78a2d59c (dpif-linux.c: Let the kernel pick a port number if one
not requested.) changed the logic for port assignment, but didn't
properly handle some error conditions. An attempt to add a tunnel port
that already exists would lead to a segfault. This commit fixes the
logic to stop processing and return an error.
Ben Pfaff [Thu, 17 Jan 2013 00:03:03 +0000 (16:03 -0800)]
util: Introduce ovs_assert macro.
An occasionally significant problem with the standard "assert" macro is
that it writes the failure message to stderr. In our daemons, stderr is
generally redirected to /dev/null. It's more useful to write the failure
message to the log, which is what the new ovs_assert macro introduced in
this patch does.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
ovs-ctl.in: Restore ofport values across force-reload-kmod.
If we want to upgrade from pre-1.10 branch to 1.10 or later
and restore flows across force-reload-kmod, we need to maintain
the value of ofport.
To do this, we fetch the value of ofport before the upgrade and use
it to populate the column 'ofport_request' after ovsdb is upgraded.
We need to do this before vswitchd is restarted as vswitchd
needs the ofport_request column populated before it starts.
ovs-save: Add a helper command to maintain ofport value.
This patch adds a helper command: save-ofports. The command
will o/p a shell script that will set the 'ofport_request' column
of the interface table with the value of the corresponding ofport.
This command takes a list of bridges as input and will be used by
a future commit to restore the ofport value across force-reload-kmod.
This command is only useful while trying to upgrade from a pre-1.10
branch to 1.10 or a later branch.
Ben Pfaff [Tue, 15 Jan 2013 21:23:53 +0000 (13:23 -0800)]
ovs-vsctl: Add --if-exists option to many database commands.
A few ovs-vsctl commands have accepted --if-exists options for some time,
to make it possible to execute them in cases where it doesn't really
matter if the records they touch exist. This commit adds this option to
other commands.
This is intended for initial use with "ovs-vsctl set interface <iface>
ofport_request=<number>" commands in ovs-ctl for upgrades from OVS 1.9
to later versions.
This makes a very small incompatible change to the ovs-vsctl "get"
command: previously with --if-exists, only keys in map columns were
allowed to be missing; now with --if-eixsts, also records are allowed
to be missing.
CC: Gurucharan Shetty <gshetty@nicira.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Pavithra Ramesh [Tue, 15 Jan 2013 20:23:46 +0000 (12:23 -0800)]
bridge: Remove restriction on socket name.
Following patch removes restriction on the listening socket name that gets
configured as bridge controller. Currently, we only connect to sockets in a
specific directory with the name of the bridge. This patch removes the
restriction on the bridge name, keeping the directory restriction.
Bug #14029. Signed-off-by: Pavithra Ramesh <paramesh@vmware.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Ed Maste [Wed, 16 Jan 2013 14:57:52 +0000 (09:57 -0500)]
netdev: Add NULL get_tunnel_config for BSD.
Commit f431bf7d78f3212d32bb3d122f783c5c796a1576 (netdev: Parse and make
available tunnel configuration.) added new method get_tunnel_config to
netdev_class but did not update netdev-bsd.
Signed-off-by: Ed Maste <emaste@freebsd.org> Signed-off-by: Ben Pfaff <blp@nicira.com>
James Page [Wed, 16 Jan 2013 11:12:09 +0000 (11:12 +0000)]
python: Workaround UNIX socket path length limits
From aa28e8bfb646a54ba91e3545f3c0b9db39eddb7f Mon Sep 17 00:00:00 2001
From: James Page <james.page@ubuntu.com>
Date: Wed, 16 Jan 2013 10:52:59 +0000
Subject: [PATCH] python: Workaround UNIX socket path length limits
To: dev@openvswitch.org
UNIX socket paths have a limit in terms of length. On Linux
this is 103 characters.
This is a problem in Debian/Ubuntu buildds which can generate
quite long paths.
This patch works around this issue by detecting when socket
connection/binding fails due to this issue and accessing the
socket path using a file descriptor path in /proc/self/fd.
The workaround is limited to Linux.
This is based on similar code in the C parts of OpenvSwitch.
Signed-off-by: James Page <james.page@ubuntu.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Jesse Gross [Tue, 15 Jan 2013 21:45:41 +0000 (13:45 -0800)]
ofproto-dpif: Maintain tun_id across action lists.
Previously, a tunnel ID received on input would be used by default
on output if the packet was later sent out a tunnel. This was not
actually intentional behavior - tunnel information is not supposed
to carry over. However, it is useful in the case where move actions
are used to load the original tunnel ID into registers for further
processing or additional lookups in resubmits since otherwise the
information is lost before handling actions.
When the initial components for flow based tunneling were introduced
this behavior for tunnel headers caused problems since it resulted
in packets being sent to the IP address they were originally received
on. All of this behavior was "fixed" in commit 47d4a9db26329f9d93eb945c1fcc0e248cf2656a (ofproto-dpif: Initialize
tunnel metadata in both 'flow' and 'base_flow'.), breaking the use
cases the require tun_id during action processing.
For the time being, at least, this restores the original behavior
for tun_id, while keeping the new behavior for the rest of the
tunnel headers. The latter are not exposed through OpenFlow and
therefore do not have similar complications. If we do expose these
headers then we might have to revist this.
Thanks to Ben Pfaff for identifying the commit that introduced the
problem.
Bug #14625
Reported-by: Michael Hu <humichael@vmware.com> Signed-off-by: Jesse Gross <jesse@nicira.com>
Krishna Kondaka [Fri, 11 Jan 2013 05:20:22 +0000 (21:20 -0800)]
alloc_ofp_port does not allocate the port number correctly
alloc_ofp_port() does not allocate the port number correctly if the port
number passed initially is already in use. The following if block
if (ofp_port >= ofproto->max_ports
|| bitmap_is_set(ofproto->ofp_port_ids, ofp_port)) {
is entered when either of the two conditions is true but the while block
after this is not entered if the second condition above is true and the
first condition is false.
This results in an existing port_number to be re-assigned!
which caused problems because worker_request() is not reentrant. In a
little more detail, the second worker_request() shoves its RPC protocol
data into the middle of the first. This means that, first, you get
some binary crud in the log (the header for the second RPC). And,
second, text from the first RPC log message gets treated by the worker
as the subsequent RPC's header. That, in turn, typically causes the
worker to try to xmalloc() a huge number of bytes (0x20000000 or more,
since "space" has ASCII value 0x20), which causes the worker to die
with "virtual memory exhausted". The main process then dies because
the worker's death closes the socket it uses to communicate with it
("connection reset").
Bug #14616. Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Justin Pettit [Sat, 5 Jan 2013 02:34:26 +0000 (18:34 -0800)]
dpif-linux: Give each port its own userspace-kernel channel.
Userspace-kernel communication is a possible bottleneck when OVS is
receiving a large number of flow set up requests. To help prevent a bad
actor from consuming too much of this resource, we introduced channels
to segegrate traffic. Previously, we created 17 channels and
round-robin assigned ports to one of 16 channels (the 17th was reserved
for use by the system). This meant if there were more than 16 ports,
sharing of channels would occur.
This commit creates a new channel for each port, so that there is no
more sharing and better isolation. The special system port uses the
"ovs-system"'s channel (port 0), since it is not heavily loaded.
This also fixes an issue introduced in commit acf60855 (ofproto-dpif:
Use a single underlying datapath across multiple bridges.) where ports
that were added at run-time were given the special system channel.
Ben Pfaff [Fri, 21 Dec 2012 22:59:11 +0000 (14:59 -0800)]
classifier: Fix theoretical leak in classifier_destroy().
The open-coded version of destroy_table() in classifier_destroy() didn't
free the table's minimatch. Use destroy_table() to do it properly.
This is only a theoretical leak because all the existing callers actually
remove all the rules from their classifiers before they destroy them
(outside of the tests/ directory, which I didn't examine) and so they don't
ever have anything left to remove in classifier_destroy().
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Justin Pettit [Wed, 9 Jan 2013 00:39:15 +0000 (16:39 -0800)]
ofproto: Fix remaining incorrect users of OVSP_LOCAL.
In the switch to a single datapath, the simple mapping between
OFPP_LOCAL and OVSP_LOCAL was broken. Unfortunately, a couple of
translations were missed, so this commit fixes them.
Kyle Mestery [Wed, 9 Jan 2013 19:47:41 +0000 (14:47 -0500)]
datapath: Fix a kernel crash when adding a 2nd VXLAN port with existing UDP port
When creating a second VXLAN tunnel port with the same UDP
port as an existing VXLAN tunnel port, make sure to set
*vxport before returning. Fixes a kernel crash.
Currently we use "*" or ANY to mark a field in flow syntax
as a wildcard. With ANY being a valid openflow port now,
there is a conflict for in_port field. So at the least, we
need to remove ANY from being considered as a wildcard for
in_port. But this may cause general confusion and it may be
a better idea to remove 'ANY' as a wildcard for all fields.
Ben Pfaff [Wed, 26 Dec 2012 17:42:33 +0000 (09:42 -0800)]
connmgr: Fix memory leak destroying ofmonitor.
This should be pretty minor because it only leaks memory if the match does
not fit in the inline area of a miniflow and minimask. I imagine that the
most common match used for an ofmonitor is "match everything", which
wouldn't leak.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Zoltan Kiss [Mon, 7 Jan 2013 22:47:51 +0000 (22:47 +0000)]
bond: Change rebalance ratio to be more fair
The old algorithm tries to converge to 0, despite it would mean a very
unbalanced situation. We're aiming for an ideal ratio of 1, meaning both
the 'from' and 'to' slave have the same load. Therefore, we only move an
entry if it decreases the load on 'from', and brings us closer to equal
traffic load.
Signed-off-by: Zoltan Kiss <zoltan.kiss@citrix.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Saul St. John [Sat, 5 Jan 2013 02:11:40 +0000 (20:11 -0600)]
manpages: Fix build
This patch fixes regressions in the dist and distclean targets, and in
the Debian build process, that were introduced by commit d2cb6c956b
(manpages: Put version number instead of date at bottom of page.).
Signed-off-by: Saul St. John <sstjohn@cs.wisc.edu> Signed-off-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Thu, 3 Jan 2013 01:10:43 +0000 (17:10 -0800)]
vconn: Fix parameters for vconn_open(), vconn_open_block(), pvconn_open().
The customary parameter order in Open vSwitch is to put input parameters
before output parameters, but vconn_open() and pvconn_open() had the 'dscp'
input parameter at the end, which bugged me a bit. Also,
vconn_open_block() didn't take a 'dscp' parameter at all even though it's
otherwise a wrapper around vconn_open(). This commit fixes all that up.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ben Pfaff [Wed, 2 Jan 2013 18:19:14 +0000 (10:19 -0800)]
manpages: Put version number instead of date at bottom of page.
We're really good about keeping manpages up to date, but terrible at
updating the dates at the bottom of the manpages. So, instead of using
manually updated dates, this commit switches to using automatically updated
version numbers.
We can only use automatically updated version numbers for manpages that
we preprocess, that is, the manpages whose sources end with ".in". This
excludes a couple of manpages that don't actually get installed with OVS,
such as the manpages for ovs-ctl and ovsdb-idlc. This commit doesn't
change those manpages. It does change the ovs-bugtool manpage to one that
is preprocessed so that we can use the version there.
Signed-off-by: Ben Pfaff <blp@nicira.com> Acked-by: Ethan Jackson <ethan@nicira.com>
Ethan Jackson [Fri, 30 Nov 2012 19:36:42 +0000 (11:36 -0800)]
netdev-vport: Manage ethernet addresses in userspace.
Letting netdev-vport manage ethernet addresses itself instead of
relying on the datapath has several advantages. It simplifies the
code, is significantly more efficient, and will work when there is
no longer a one to one mapping from netdev-vports to datapath
vports.
Theoretically, its possible for netdev_get_status() to be called
on a netdev-vport which hasn't had its configuration set yet. In
this case, netdev-vport would dereference a null pointer.
Reported-by: Jesse Gross <jesse@nicira.com> Signed-off-by: Ethan Jackson <ethan@nicira.com>
Ethan Jackson [Wed, 19 Dec 2012 04:54:28 +0000 (20:54 -0800)]
ofproto-dpif: New function ofproto_receive().
Before translating a datapath flow key into actions, ofproto-dpif
must parse it, tweak it, and figure out what ofproto_dpif it
belongs to. This patch brings all this logic into one place where
it will be easier to extend in the future.
Ben Pfaff [Thu, 3 Jan 2013 17:02:52 +0000 (09:02 -0800)]
ofp-actions: Fix the check for instruction ordering and duplication.
Open vSwitch enforces that, when instructions appear as Nicira extensions
in OpenFlow 1.0 action lists, the instructions appear in the order that
an OpenFlow 1.1+ switch would execute them and that no duplicates appear.
But the check wasn't general enough, because it had been implemented when
only one instruction was implemented and never later generalized. This
commit fixes the problem.
One of the tests was actually testing for the wrong behavior that the
check implemented, so this commit corrects that test. It also updates
another test with updated log messages.
Reported-by: Jing Ai <ai_jing2000@hotmail.com> Tested-by: Jing Ai <ai_jing2000@hotmail.com> Signed-off-by: Ben Pfaff <blp@nicira.com>
Anoob Soman [Wed, 2 Jan 2013 17:23:07 +0000 (17:23 +0000)]
lacp: Fix dumping of the aggregation key.
While dumping lacp information using ovs-appctl, the "aggregation
key" field displays the port_id even though the aggregation key is
set using "other-config:lacp-aggregation-key".
Signed-off-by: Anoob Soman <anoob.soman@citrix.com> Signed-off-by: Ethan Jackson <ethan@nicira.com>
Jesse Gross [Mon, 31 Dec 2012 20:38:26 +0000 (12:38 -0800)]
ofproto-dpif: Initialize tunnel metadata in both 'flow' and 'base_flow'.
The datapath doesn't carry over tunnel metadata between input and
output because such an operation doesn't generally make sense. In
order to keep the flow information in sync, userspace needs to do
this as well. However, userspace was clearing out only the base
flow, which means that it later generated an action with the original
tunnel metadata. If a packet is both being sent and received on a
tunnel then this will cause the output tunnel's configuration to be
overriden with incorrect information. This was recently exposed
when the ability to interpret tunnel metadata was added to userspace.
Reported-by: Anand Krishnamurthy <krishnamurt4@wisc.edu> Reported-by: Saul St. John <sstjohn@cs.wisc.edu> Signed-off-by: Jesse Gross <jesse@nicira.com>
Jesse Gross [Mon, 31 Dec 2012 20:28:41 +0000 (12:28 -0800)]
datapath: Initialize tunnel_key pad member.
When a packet is received on a tunnel the pad member is currently
left uninitialized. This didn't previously cause problems because
userspace didn't interprete the IPV4_TUNNEL attribute and blindly
copied back the uninitialized data. However, now that userspace
knows how to serialize this attribute it was zeroing it out, which
prevented flows that had been previously installed from being
deleted. In addition to zeroing out the padding on packet reception,
it also does the same thing on flow setup since we should be ignoring
the value.
Reported-by: Anand Krishnamurthy <krishnamurt4@wisc.edu> Reported-by: Saul St. John <sstjohn@cs.wisc.edu> Signed-off-by: Jesse Gross <jesse@nicira.com> Acked-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Sun, 30 Dec 2012 17:56:46 +0000 (09:56 -0800)]
debian: Fix typo in comment in openvswitch-controller.default.
Reported-by: Saul St. John Reported-by: Logan Rosen <logatronico@gmail.com>
Reported-at: https://bugs.launchpad.net/ubuntu/+source/openvswitch/+bug/1094556 Signed-off-by: Ben Pfaff <blp@nicira.com>
Jarno Rajahalme [Fri, 28 Dec 2012 16:28:49 +0000 (18:28 +0200)]
Add Openflow 1.2 role request/reply processing, update OF 1.2 tests. Add struct ofputil_role_request and encode/decode functions for role request/reply.
Signed-off-by: Jarno Rajahalme <jarno.rajahalme@nsn.com>
[blp@nicira.com made ofp-print print any error, renamed a function,
added a comment] Signed-off-by: Ben Pfaff <blp@nicira.com>
SUGYO Kazushi [Tue, 18 Dec 2012 12:51:11 +0000 (21:51 +0900)]
ofp-util: features reply of openflow13 does not have ports
Signed-off-by: SUGYO Kazushi <sugyo.org@gmail.com>
[blp@nicira.com moved the test from ofproto.c to ofp-util.c] Signed-off-by: Ben Pfaff <blp@nicira.com>
debian: Integrate interfaces config with OVS startup script.
Currently, there is an option to use the /etc/network/interfaces
file to configure Open vSwitch bridges and ports. To create and
destroy bridges, one has to manually run the following commands.
Ben Pfaff [Wed, 26 Dec 2012 23:00:06 +0000 (15:00 -0800)]
ovs-vswitch: Make OpenFlow "dp_desc" configurable.
Signed-off-by: Felician Nemeth <nemethf@tmit.bme.hu>
[blp@nicira.com refactored the code that this was based on and revised the
patch accordingly] Signed-off-by: Ben Pfaff <blp@nicira.com>
Ben Pfaff [Wed, 26 Dec 2012 22:56:01 +0000 (14:56 -0800)]
ofproto: Use NULL "desc" values to indicate defaults.
This saves a small amount of memory. It should also save a small amount of
time for reconfiguration starting with an upcoming commit where it becomes
possible to change the dp_desc away from the default (because it will not
be necessary to compare the current value against the default).
projects / ovs.git / log