In order to work with Geneve options, we need to maintain a mapping
table between an option (defined by <class, type, length>) and
an NXM field that can be operated on for the purposes of matches,
actions, etc. This mapping must be explicitly specified by the
user.
Conceptually, this table could be communicated using either OpenFlow
or OVSDB. Using OVSDB requires less code and definition of extensions
than OpenFlow but introduces the possibility that mapping table
updates and flow modifications are desynchronized from each other.
This is dangerous because the mapping table signifcantly impacts the
way that flows using Geneve options are installed and processed by
OVS. Therefore, the mapping table is maintained using OpenFlow commands
instead, which opens the possibility of using synchronization between
table changes and flow modifications through barriers, bundles, etc.
There are two primary groups of OpenFlow messages that are introduced
as Nicira extensions: modification commands (add, delete, clear mappings)
and table status request/reply to dump the current table along with switch
information.
Note that mappings should not be changed while they are in active use by
a flow. The result of doing so is undefined.
This only adds the OpenFlow infrastructure but doesn't actually
do anything with the information yet after the messages have been
decoded.
Signed-off-by: Jesse Gross <jesse@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
rconn_get_connection_seqno() is documented to change only when an rconn
connects or disconnnects, but in fact it was also changing whenever an
rconn went into or out of the "idle" state (following sending an echo
request). This fixes the problem.
rconn_get_connection_seqno() didn't have any existing users, but an
upcoming commit adds one.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Russell Bryant <rbryant@redhat.com>
Acked-by: Justin Pettit <jpettit@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>
These functions had no callers, so remove them and the data maintained
just to implement them.
Found by inspection.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Thomas Graf <tgraf@noironetworks.com>
Use of OF 1.4 bundle messages by a controller should indicate that the
controller has decided to use the switch, hence make is_admitted_msg()
return 'true' for them.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Acked-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>
On Windows, when a peer terminates without calling a close
on socket fd, the server ends up printing "connection dropped"
warning messages. We probably don't want those warning messages
when the error is WSAECONNRESET.
(In OVS unit tests on Windows, anytime a client like ovs-ofctl
calls a ovs_fatal without clean close(fd) on the socket, the
server like ovs-vswitchd prints warnings that cause unit tests
to fail.)
Signed-off-by: Gurucharan Shetty <gshetty@nicira.com>
Acked-by: Ben Pfaff <blp@nicira.com>
An rconn has a human-readable name that typically designates both endpoints
of the connection. For a "reliable" rconn, that automatically reconnects,
the name remains constant regardless of whether the rconn is currently
connected. Until now, though, an "unreliable" rconn, that cannot
automatically reconnect, kept its name only until disconnection occurred.
This is OK for the uses currently in the OVS tree, which only use the name
of a rconn while it is connected, but an upcoming commit will add a final
log message following disconnection in some cases, and it makes the log
messages less useful if unreliable rconns just report "void" in that case.
This commit, therefore, modifies the rconn code so that unreliable rconns
preserve their names past disconnection, just like reliable ones.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Alex Wang <alexw@nicira.com>
rconn objects do not cache IP address and port information any longer.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Alex Wang <alexw@nicira.com>
This is only the communication part of the bundles functionality.
The actual message pre-validation and commits are not implemented.
We also enable OF1.4 for all the tests.
Signed-off-by: Alexandru Copot <alex.mihai.c@gmail.com>
Cc: Daniel Baluta <dbaluta@ixiacom.com>
[blp@nicira.com made ofputil_decode_bundle_add() more obviously correct]
Signed-off-by: Ben Pfaff <blp@nicira.com>
Rename 'l2' to 'frame' and add new ofpbuf_set_frame() and ofpbuf_l2().
ofpbuf_set_frame() alse resets all the layer offsets. ofpbuf_l2()
returns NULL if the packet has no Ethernet header, as indicated either
by unset l3 offset or NULL frame pointer. Callers of ofpbuf_l2() are
supposed to check the return value, unless they can otherwise be sure
that the packet has a valid Ethernet header.
The recent commit 437d0d22 made some assumptions that were not valid
regarding the use of the 'l2' pointer in rconn module and by
compose_rarp(). This is now fixed as follows: rconn now relies on the
fact that once OpenFlow messages are given to rconn for transport, the
frame pointer is no longer needed to refer to the OpenFlow header; and
compose_rarp() now sets the frame pointer and offsets as expected.
In addition to storing network frames, ofpbufs are also used for
handling OpenFlow messages and action lists. lib/ofpbuf.h now has a
comment documenting the current usage conventions and invariants.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
This patch shrinks the struct ofpbuf from 104 to 48 bytes on 64-bit
systems, or from 52 to 36 bytes on 32-bit systems (counting in the
'l7' removal from an earlier patch). This may help contribute to
cache efficiency, and will speed up initializing, copying and
manipulating ofpbufs. This is potentially important for the DPDK
datapath, but the rest of the code base may also see a little benefit.
Changes are:
- Remove 'l7' pointer (previous patch).
- Use offsets instead of layer pointers for l2_5, l3, and l4 using
'l2' as basis. Usually 'data' is the same as 'l2', but this is not
always the case (e.g., when parsing or constructing a packet), so it
can not be easily used as the offset basis. Also, packet parsing is
faster if we do not need to maintain the offsets each time we pull
data from the ofpbuf.
- Use uint32_t for 'allocated' and 'size', as 2^32 is enough even for
largest possible messages/packets.
- Use packed enum for 'source'.
- Rearrange to avoid unnecessary padding.
- Remove 'private_p', which was used only in two cases, both of which
had the invariant ('l2' == 'data'), so we can temporarily use 'l2'
as a private pointer.
Signed-off-by: Jarno Rajahalme <jrajahalme@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
rconn_recv() calls vconn_recv(), which will report a connection error and
cause rconn_recv() to disconnect the rconn. Most rconn users regularly
call rconn_recv(), so that disconnection happens promptly. However,
the lswitch code only calls rconn_recv() after the connection negotiates an
OpenFlow version, so a connection that failed before negotiation would
never be detected as failed. This commit fixes the problem by making
rconn_run() also detect and handle connection errors.
The lswitch code is only used by the test-controller program, so this is
not an important bug fix.
Reported-by: Vasu Dasari <vdasari@gmail.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
These functions don't have any ultimate users. The in-band control code
used to use them, but not anymore, so we might as well delete them all.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
This allows other libraries to use util.h that has already
defined NOT_REACHED.
Signed-off-by: Harold Lim <haroldl@vmware.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
This should make sending OFPT_FLOW_REMOVED and NXST_FLOW_MONITOR safe from
miss handler threads.
Bug #20271.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
Added infrastructure to support Openflow OFPT_TABLE_MOD message. This patch
does not include the flexible table miss handling code that is necessary to
support the semantics specified in OFPT_TABLE_MOD messages.
Current flow miss behavior continues to conform to Openflow 1.0. Future
commits to add more flexible table miss support are needed to fully support
OPFT_TABLE_MOD for Openflow-1.1+.
Signed-off-by: Andy Zhou <azhou@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Otherwise, if a monitor connection happens to be talking to a (misguided?)
peer that sends it messages, such as replies to what the peer perceives as
echo requests meant for it, then the peer will eventually hang trying to
send data because the monitor connection never sinks it.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
rconn_add_monitor() tries to check the version of the controller
connection being monitored, so that it can decide what OpenFlow version to
tell the monitor connection to negotiate. But at any given time an rconn
may not have a controller connection (e.g. during backoff), so rc->vconn
may be null and thus vconn_get_version(rc->vconn) dereferences a null
pointer.
Fixing the problem in a local way would require the rconn to remember the
previous version negotiated, and that fails if the rconn hasn't yet
connected or if the next connection negotiates a new version.
This commit instead adds the ability to a vconn to accept any OpenFlow
message version and modifies "ovs-ofctl snoop" to use that feature, thus
removing the need to negotiate the "correct" version on snoops.
Bug #14265.
Reported-by: Pratap Reddy <preddy@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
This is a straight search-and-replace, except that I also removed #include
<assert.h> from each file where there were no assert calls left.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Ethan Jackson <ethan@nicira.com>
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>
rconn_create() was substituting OFPUTIL_DEFAULT_VERSIONS if an
allowed_versions of 0 was passed in. At the same time,
connmgr_set_controllers() compared the adjusted value of allowed_versions
against the original value, saw that they were different, and concluded
that it should kill off and recreate the rconn with the "corrected"
allowed_versions.
This commit fixes the problem by no longer adjusting allowed_versions.
There is no need, because it is only used in contexts where the original
version is OK.
This problem was introduced by commit 90ef0206ea (connmgr:
Reinitialise controllers if protocols changes).
Bug #14126.
CC: Simon Horman <horms@verge.net.au>
Reported-by: Natasha Gude <natasha@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Initial OpenFlow 1.3 support with new include/openflow/openflow-1.3.h.
Most of the messages that differ from 1.2 are implemented. OFPT_SET_ASYNC
is implemented via NX_SET_ASYNC_CONFIG, other new message types are yet to
be implemented. Stats replies that add duration fields are implemented at
encode/decode level only. Test cases for implemented features are included.
Remaining FIXME:s should not cause runtime aborts. Make check comes out
clean.
Signed-off-by: Jarno Rajahalme <jarno.rajahalme@nsn.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Override the allowed versions of the snoop vconn so that
only the version of the controller connection is allowed.
This is because the snoop will see the same messages as the
controller.
Without this change the snoop will try to negotiate a connection
using the default allowed versions, or in other words only
allow OpenFlow 1.0. This breaks snoops for controller connections
using other OpenFlow versions, that is OpenFlow 1.2.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Ben Pfaff <blp@nicira.com>
Add allowed OpenFlow versions to struct rconn to allow
reconnect to use these parameters rather than hard-coded defaults.
This is in preparation for allowing configuration of the
allowed OpenFlow versions.
Signed-off-by: Simon Horman <horms@verge.net.au>
Signed-off-by: Ben Pfaff <blp@nicira.com>
This is in preparation for allowing the range of allowed OpenFlow versions
to be configured.
As part of this change pvconn_open() is now paramatised over the allowed
versions. this is to avoid avoids needing to provide version information
as a parameter to pvconn_accept(). This will in turn avoid the need to
pass version information to connmgr_run().
Signed-off-by: Simon Horman <horms@verge.net.au>
[blp@nicira.com simplified slightly and generalize log messages]
Signed-off-by: Ben Pfaff <blp@nicira.com>
Until now, the rconn module has used messages received from the
controller as the sole means to determine that the connection is up.
This can interact badly with the OVS connection manager in ofproto,
which stops reading and processing messages from the receive queue
when there is a backlog in the send queue for a given connection
(because reading and processes messages is the main cause of messages
getting pushed onto the send queue). So, if a send queue backlog
lasts more than twice the inactivity probe interval, then the
connection drops, whether the controller is sending messages or not.
Dumping a large flow table can trigger this behavior if the controller
becomes temporarily busy or if the network between OVS and a
controller is slow. The problem can easily repeat itself, since upon
reconnection the controller will generally dump the flow table.
This commit fixes the problem by expanding the definition of
"activity" to include successfully sending an OpenFlow message that
was previously queued.
Bug #12789.
Reported-by: Natasha Gude <natasha@nicira.com>
Signed-off-by: Ben Pfaff <blp@nicira.com>
OpenFlow headers are not as uniform as they could be, with size, alignment,
and numbering changes from one version to another and across varieties
(e.g. ordinary messages vs. "stats" messages). Until now the Open vSwitch
internal APIs haven't done a good job of abstracting those differences in
header formats. This commit changes that; from this commit forward very
little code actually needs to understand the header format or numbering.
Instead, it can just encode or decode, or pull or put, the header using
a more abstract API using the ofpraw_, ofptype_, and other APIs in the
new ofp-msgs module.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Tested-by: Simon Horman <horms@verge.net.au>
Reviewed-by: Simon Horman <horms@verge.net.au>
This patch reapplies the changes that were reverted with the commit 59efa47
(Revert DSCP update changes.). It also addresses the problem introduced by
the original commits, cd8fca2 ((jsonrpc: Correctly setting the dscp value
before reconnect.) and b2e18d (No need to restart DB / OVS on changing
dscp value.), that caused numerous unit test failures on some systems (as
diagnosed by valgrind).
With this change there is no need to restart the DB or OVS on configuring a
different value for the manager or controller connection respectively. On
detecting a change in the dscp value on the socket, the previous socket is
closed and a new socket is created and connection is established with the new
configured dscp value.
Signed-off-by: Mehak Mahajan <mmahajan@nicira.com>
This reverts commit cd8fca2ba0 (jsonrpc:
Correctly setting the dscp value before reconnect.) and commit
b2e18db292 (No need to restart DB / OVS on
changing dscp value.), which on some systems causes numerous unit test
failures that valgrind diagnoses as:
Conditional jump or move depends on uninitialised value(s)
at 0x805F63F: jsonrpc_session_set_dscp (jsonrpc.c:1061)
by 0x804F45D: ovsdb_jsonrpc_server_set_remotes (jsonrpc-server.c:417)
by 0x804B775: reconfigure_from_db (ovsdb-server.c:656)
by 0x804C231: main (ovsdb-server.c:159)
Signed-off-by: Ben Pfaff <blp@nicira.com>
With this change there is no need to restart the DB or OVS on configuring a
different value for the manager or controller connection respectively. On
detecting a change in the dscp value on the socket, the previous socket is
closed and a new socket is created and connection is established with the new
configured dscp value.
Signed-off-by: Mehak Mahajan <mmahajan@nicira.com>
