This commit relaxes the whitelist format for punix path of
service controller. Instead of only allowing
punix:<ovs_rundir>/<bridge_name>.controller, the new format
allows any suffix, like punix:<ovs_rundir>/<bridge_name>.*.
(except one containing '/').
Signed-off-by: Alex Wang <ee07b291@gmail.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Define struct eth_addr and use it instead of a uint8_t array for all
ethernet addresses in OVS userspace. The struct is always the right
size, and it can be assigned without an explicit memcpy, which makes
code more readable.
"struct eth_addr" is a good type name for this as many utility
functions are already named accordingly.
struct eth_addr can be accessed as bytes as well as ovs_be16's, which
makes the struct 16-bit aligned. All use seems to be 16-bit aligned,
so some algorithms on the ethernet addresses can be made a bit more
efficient making use of this fact.
As the struct fits into a register (in 64-bit systems) we pass it by
value when possible.
This patch also changes the few uses of Linux specific ETH_ALEN to
OVS's own ETH_ADDR_LEN, and removes the OFP_ETH_ALEN, as it is no
longer needed.
This work stemmed from a desire to make all struct flow members
assignable for unrelated exploration purposes. However, I think this
might be a nice code readability improvement by itself.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
GRE64 was introduced to extend gre key from 32-bit to 64-bit using
gre-key and sequence number field. But GRE64 is not standard
protocol. There are not many users of this protocol. Therefore we
have decided to remove it.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Whenever system interfaces are removed, added or change state, reconfigure
bridge. This allows late interfaces to be added to the datapath when they are
added to the system after ovs-vswitchd is started.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@redhat.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Until now, mirroring has been implemented by accumulating, across the whole
translation process, a set of mirrors that should receive a mirrored
packet. After translation was complete, mirroring restored the original
version of the packet and sent that version to the mirrors.
That implementation was ugly for multiple reasons. First, it means that
we have to keep a copy of the original packet (or its headers, actually),
which is expensive. Second, it doesn't really make sense to mirror a
version of a packet that is different from the one originally output.
Third, it interacted with recirculation; mirroring needed to happen only
after recirculation was complete, but this was never properly implemented,
so that (I think) mirroring never happened for packets that were
recirculated.
This commit changes how mirroring works. Now, a packet is mirrored at the
point in translation when it becomes eligible for it: for mirrors based on
ingress port, this is at ingress; for mirrors based on egress port, this
is at egress. (Duplicates are dropped.) Mirroring happens on the version
of the packet as it exists when it becomes eligible. Finally, since
mirroring happens immediately, it interacts better with recirculation
(it still isn't perfect, since duplicate mirroring will occur if a packet
is eligible for mirroring both before and after recirculation; this is
not difficult to fix and an upcoming commit later in this series will do so).
Finally, this commit removes more code from xlate_actions() than it adds,
which in my opinion makes it easier to understand.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
Some commands of ovs-appctl were lazily registered when first
bridg or bfd was added. Before that, calling these commands raised a
error("xxx is not a valid command"). The problem commangs included
"bfd/...", "upcall/...","dpif/...","fdb/..." and so on.
Fix this by moving the register into the "bridge_init" and
"bridge_init_ofproto". All commands are registered at the moment
ovs-vswitchd starts.
Signed-off-by: Huanle Han <hanxueluo@gmail.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
OpenFlow 1.4 introduces the ability to turn on flow table eviction with an
OFPT_TABLE_MOD message specifying OFPTC_EVICTION. It also adds related
machinery to other messages that mention OFPTC_* fields. This commit adds
support for the new feature, implementing it as a second, parallel way to
enable flow table eviction. It takes more work than it seems like it
should because there is so much weirdness with the treatment of OFPTC_*
flags over the evolution of OpenFlow; please refer to the explanation in
DESIGN.md for more information.
This commit also adds related support to ovs-ofctl, plus tests.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Co-authored-by: Saloni Jain <saloni.jain@tcs.com>
Signed-off-by: Saloni Jain <saloni.jain@tcs.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
The "importance" field is considered before flow timeout because I figure
that if you set the importance, you think it's important.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Co-authored-by: Saloni Jain <saloni.jain@tcs.com>
Signed-off-by: Saloni Jain <saloni.jain@tcs.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
Add support for MLDv1 and MLDv2. The behavior is not that different from
IGMP. Packets to all-hosts address and queries are always flooded,
reports go to routers, routers are added when a query is observed, and
all MLD packets go through slow path.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@redhat.com>
Cc: Flavio Leitner <fbl@redhat.com>
Cc: Ben Pfaff <blp@nicira.com>
[blp@nicira.com moved an assignment out of an 'if' statement]
Signed-off-by: Ben Pfaff <blp@nicira.com>
The current support for Geneve in OVS is exactly equivalent to VXLAN:
it is possible to set and match on the VNI but not on any options
contained in the header. This patch enables the use of options.
The goal for Geneve support is not to add support for any particular option
but to allow end users or controllers to specify what they would like to
match. That is, the full range of Geneve's capabilities should be exposed
without modifying the code (the one exception being options that require
per-packet computation in the fast path).
The main issue with supporting Geneve options is how to integrate the
fields into the existing OpenFlow pipeline. All existing operations
are referred to by their NXM/OXM field name - matches, action generation,
arithmetic operations (i.e. tranfer to a register). However, the Geneve
option space is exactly the same as the OXM space, so a direct mapping
is not feasible. Instead, we create a pool of 64 NXMs that are then
dynamically mapped on Geneve option TLVs using OpenFlow. Once mapped,
these fields become first-class citizens in the OpenFlow pipeline.
An example of how to use Geneve options:
ovs-ofctl add-geneve-map br0 {class=0xffff,type=0,len=4}->tun_metadata0
ovs-ofctl add-flow br0 in_port=LOCAL,actions=set_field:0xffffffff->tun_metadata0,1
This will add a 4 bytes option (filled will all 1's) to all packets
coming from the LOCAL port and then send then out to port 1.
A limitation of this patch is that although the option table is specified
for a particular switch over OpenFlow, it is currently global to all
switches. This will be addressed in a future patch.
Based on work originally done by Madhu Challa. Ben Pfaff also significantly
improved the comments.
Signed-off-by: Madhu Challa <challa@noironetworks.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Support IGMPv3 messages with multiple records. Make sure all IGMPv3
messages go through slow path, since they may carry multiple multicast
addresses, unlike IGMPv2.
Tests done:
* multiple addresses in IGMPv3 report are inserted in mdb;
* address is removed from IGMPv3 if record is INCLUDE_MODE;
* reports sent on a burst with same flow all go to userspace;
* IGMPv3 reports go to mrouters, i.e., ports that have issued a query.
Signed-off-by: Thadeu Lima de Souza Cascardo <cascardo@redhat.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Until now there have been two variants for --enable-dummy:
* --enable-dummy: This adds support for "dummy" dpif and netdev.
* --enable-dummy=override: In addition, this replaces *every* existing
dpif and netdev by the dummy type.
The latter is useful for testing but it defeats the possibility of using
the userspace native tunneling implementation (because all the tunnel
netdevs get replaced by dummy netdevs). Thus, this commit adds a third
variant:
* --enable-dummy=system: This replaces the "system" dpif and netdev
by dummies but leaves the others untouched.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Alex Wang <alexw@nicira.com>
This patch adds support for a new port type to the userspace
datapath called dpdkvhostuser.
A new dpdkvhostuser port will create a unix domain socket which
when provided to QEMU is used to facilitate communication between
the virtio-net device on the VM and the OVS port on the host.
vhost-cuse ('dpdkvhost') ports are still available as 'dpdkvhostcuse'
ports and will be enabled if vhost-cuse support is detected in the
DPDK build specified during compilation of the switch. Otherwise,
vhost-user ports are enabled.
Signed-off-by: Ciara Loftus <ciara.loftus@intel.com>
Acked-by: Flavio Leitner <fbl@redhat.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Commit 7a6cf343a410d77e05ebd7bf5b5ade52803879ae raised the MAXFD
limit from 7500 to 65535.
Signed-off-by: Mijo Safradin <mijo@linux.vnet.ibm.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
This commit changes the semantics of 'netdev_set_multiq()' to allow OVS
DPDK to run on device with limited multi queue support.
* If a netdev doesn't have the requested number of rxqs it can simply
inform the datapath without failing.
* If a netdev doesn't have the requested number of txqs it should try
to create as many as possible and use locking.
Signed-off-by: Daniele Di Proietto <diproiettod@vmware.com>
Signed-off-by: Ethan Jackson <ethan@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
The Stateless TCP Tunnel (STT) protocol encapsulates traffic in
IPv4/TCP packets.
STT uses TCP segmentation offload available in most of NIC. On
packet xmit STT driver appends STT header along with TCP header
to the packet. For GSO packet GSO parameters are set according
to tunnel configuration and packet is handed over to networking
stack. This allows use of segmentation offload available in NICs
The protocol is documented at
http://www.ietf.org/archive/id/draft-davie-stt-06.txt
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
I guess that this was missed when the corresponding feature was removed.
Reported-by: David Evans <davidjoshuaevans@gmail.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Russell Bryant <rbryant@redhat.com>
If the manager resides on the same host as ovs, the manager target will
be the loopback address. Then, if in-band is enabled on a bridge, the
in-band module will constantly checks the connection to the manager to
make sure the manager is reachable. However, the connection checking
implementation cannot identify the route for the loopback address and
will keep issuing the following warning:
|in_band|WARN|cannot find route for controller (127.0.0.1): No such
device or address.
To fix this, this commit makes ovs not consider manager with loopback
for in-band control at all, since the manager is always reachable
on the same host.
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Found by LLVM scan-build.
Reported-by: Kevin Lo <kevlo@FreeBSD.org>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Dennis Flynn <drflynn@avaya.com>
When "-V" was supplied to ovs-vswitchd, it was returning a max supported
OpenFlow version of 1.0, which is no longer true. There are other
methods to determine the supported OpenFlow versions in ovs-vswitchd, so
remove it from the "-V" output.
Reported-by: Russell Bryant <rbryant@redhat.com>
Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Russell Bryant <rbryant@redhat.com>
These commands can be used to get packets and cycles counters on a pmd
thread basis. They're useful to get a clearer picture about the
performance of the userspace datapath.
They export these pieces of information:
- A (per-thread) view of the caches hit rate. Hits in the exact match
cache are reported separately from hits in the masked classifier
- A rough cycles count. This will allow to estimate the load of OVS and
the polling overhead.
Signed-off-by: Daniele Di Proietto <diproiettod@vmware.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
When ovsdb happens to get blocked for few seconds, the 'stats_timer'
will not be updated due to incompletion of previous transaction.
When the current time passes the 'stats_timer', the call to
poll_timer_wait_until(stats_timer) will keep waking up ovs-vswitchd,
causing 100% cpu utilization.
This commit fixes this issue by making ovs-vswitchd wait on the
previous idl transaction when it is incomplete rather than the
'stats_timer'.
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
During upgrade of ovs-vswitchd, we do not want to recreate the already
configured kernel interfaces. Especially when IP address is assigned to
the internal port, the recreation will cause the lost of connection.
Therefore, ovs-vswitchd should read current ovsdb content first and then
reuse the existing kernel interfaces that are configured in ovsdb. In
terms of the code language, ovs-vswitchd should only execute bridge_run()
after it finishes reading the ovsdb content.
However, this expected behavior is broken by the recent commit d18e52e
(ovsdb-idl: Tolerate missing tables and columns.) which causes the
execution of bridge_run() before getting the hint of configured interfaces
from ovsdb.
To fix the issue, this commit makes sure that the execution of bridge_run()
happens only after retrieving the ovsdb contents.
VMware-BZ: #1424342
Signed-off-by: Alex Wang <alexw@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
The description was at best incomplete and difficult to understand.
Reported-by: Brian Field <Brian_Field@cable.comcast.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Justin Pettit <jpettit@nicira.com>
The kernel module can already support outer UDP checksums for
Geneve and VXLAN using the standard checksum flag in tunnel
metadata. This makes userspace aware of the capability so that
users can enable it on tunnel ports.
There is a complication in that there is no way for userspace to
probe or detect if the kernel does not support this capability
in order to warn the user. In this case, connectivity will appear
to function normally but packets will not be checksum protected.
This is mainly an issue for VXLAN which has existed in the kernel
for a some time without checksum support - while there are also
a few kernel versions that support Geneve only without checksums,
they are much less common.
There isn't a particularly good solution to the compatibility
issue without introducing a larger capabilities structure. However,
UDP checksums are likely to be used only rarely at this point in
time and the VXLAN spec (where the main problem lies) recommends
against them. Therefore, this is considered to be an advanced user
feature and we settle for just documenting the issue.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Pritesh Kothari <pritesh.kothari@cisco.com>
At startup enumerate datapath and port types and add this information to
the datapath_types and port_types columns in the ovsdb.
This allows an ovsdb client to query the datapath in order to determine
if certain datapath and port types exist. For example, by querying the
port_types column, an ovsdb client will be able to determine if this
instance of ovs-vswitchd was compiled with DPDK support.
Signed-off-by: Mark D. Gray <mark.d.gray@intel.com>
Signed-off-by: Billy O'Mahony <billy.o.mahony@intel.com>
[blp@nicira.com made several changes]
Signed-off-by: Ben Pfaff <blp@nicira.com>
This patch adds support for SFQ, CoDel and FQ_CoDel classless qdiscs to Open vSwitch. It also removes the requirement for a QoS to have at least one Queue (as this makes no sense when using classless qdiscs). I have also not implemented class_{get,set,delete,get_stats,dump_stats} because they are meant for qdiscs with classes.
Signed-off-by: Jonathan Vestin <jonavest@kau.se>
[blp@nicira.com mostly applied stylistic changes]
Signed-off-by: Ben Pfaff <blp@nicira.com>
This patch adds support for a new port type to userspace datapath
called dpdkvhost. This allows KVM (QEMU) to offload the servicing
of virtio-net devices to its associated dpdkvhost port. Instructions
for use are in INSTALL.DPDK.
This has been tested on Intel multi-core platforms and with clients
that have virtio-net interfaces.
Signed-off-by: Ciara Loftus <ciara.loftus@intel.com>
Signed-off-by: Kevin Traynor <kevin.traynor@intel.com>
Signed-off-by: Maryam Tahhan <maryam.tahhan@intel.com>
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Otherwise the policing limits could make no sense if large rates were
specified.
Reported-by: Zhangguanghui <zhang.guanghui@h3c.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ansis Atteka <aatteka@nicira.com>
The coding style guidelines include the following:
- Pick a unique name prefix (ending with an underscore) for each
module, and apply that prefix to all of that module's externally
visible names. Names of macro parameters, struct and union members,
and parameters in function prototypes are not considered externally
visible for this purpose.
This patch adds the new prefix to the externally visible names. This
makes it a bit more obvious what code is coming from common command
line handling code.
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
We've been warning about the change since 2.1, which was released a year
ago.
Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
The IPv6 examples in ovs-vswitchd.conf.db to configure the OpenFlow port
used the OVSDB port number, which is confusing.
Signed-off-by: Justin Pettit <jpettit@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
Until now, if both STP and RSTP were enabled, ovs-vswitchd would actually
enable only the one it first noticed to be enabled, and actually turn off
the setting for the other one in the database (!). This doesn't match
ovs-vswitchd behavior for other contradictory configurations, so this
commit changes its behavior so that, if both are enabled, RSTP takes
precedence.
Reported-by: Daniele Venturino <daniele.venturino@m3s.it>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ansis Atteka <aatteka@nicira.com>
An ISID is 24 bits, so it fits in a uint32_t. A VLAN is 12 bits, so it
fits in a uint16_t. Use these types consistently, instead of int64_t.
This removes a check in aa_mapping_unregister() that seems a little
mysterious to me: it previously checked for ISID and VLAN values >= 0. I
don't see a way that they could be < 0 in this situation though.
Signed-off-by: Ben Pfaff <blp@nicira.com>
This is the final commit in the series of commits that deliver initial support
for Auto-Attach. Specifically this commit delivers auto-attach support to the
OVS bridge layer as well as the new auto-attach commands. The OVSDB schema is
modified to define the new auto-attach entries. The man pages, unit tests, and
news and license notice files are also updated. A unit test is provided to
validate the construction of auto-attach packets.
Signed-off-by: Ludovic Beliveau <ludovic.beliveau@windriver.com>
Signed-off-by: Dennis Flynn <drflynn@avaya.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
ofpbuf was complicated due to its wide usage across all
layers of OVS, Now we have introduced independent dp_packet
which can be used for datapath packet, we can simplify ofpbuf.
Following patch removes DPDK mbuf and access API of ofpbuf
members.
Signed-off-by: Pravin B Shelar <pshelar@nicira.com>
Acked-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
VXLAN is now RFC, so we better to change the pointer to it.
Signed-off-by: Kentaro Ebisawa <ebiken.g@gmail.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
A fair number of columns had been overlooked. This documents them.
The patch is smaller than it appears because this rearranges the STP and
RSTP documentation to group configuration, status, and statistics together
in the documentation for clarity.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Gurucharan Shetty <gshetty@nicira.com>
The lower layers count errors but until now nothing actually reported them.
Found by inspection.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Gurucharan Shetty <gshetty@nicira.com>
Update the logic used in constify() to add const to a 'char **' while
still excluding all other cases of more than one level of indirection.
This results in adding const to a parameter of a generated setter
function where we're generally passing in array of constant strings.
As a result, this patch includes the other necessary fixes to the code
base to reflect the const addition.
Signed-off-by: Russell Bryant <rbryant@redhat.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Supports a new "exts" field in the tunnel configuration which takes a
comma separated list of enabled extensions.
The only extension supported so far is GBP but this can be used to
enable RCO and possibly others as soon as the OVS datapath supports
them.
Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Acked-by: Ben Pfaff <blp@nicira.com>
The RFC4541 section 2.1.1 item 1 allows the snooping switch
to provide an administrative control to allow Report messages
to be flooded to ports not connected to multicast routers.
Signed-off-by: Flavio Leitner <fbl@redhat.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Also moves definitions for struct vconn and pvconn to the public
header. The provider interface is kept private.
Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Acked-by: Ben Pfaff <blp@nicira.com>
A new function vlog_insert_module() is introduced to avoid using
list_insert() from the vlog.h header.
Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Acked-by: Ben Pfaff <blp@nicira.com>
struct list is a common name and can't be used in public headers.
Signed-off-by: Thomas Graf <tgraf@noironetworks.com>
Acked-by: Ben Pfaff <blp@nicira.com>
