Some legitimate system activity can cause network devices to be destroyed
just before they are removed from the database, added to the database
just before they are created, or destroyed and then recreated under the
same name while other database activity is going on. Logging these
events as errors makes it sound like something unexpectedly bad is going
on, but in fact these events are the most common instances of these log
messages, so downgrade them to warnings.
Reported-by: Reid Price <reid@nicira.com>
Bug #2584.
We had retained but deprecated the use of the older 'managers' column in the
'Open vSwitch' table for compatibility with applications that might still use
it, but that created more problems than it solved. This commit removes the
'managers' column from the schema, and removes all references to it from the
code, init scripts, documentation, and tests.
This feature was included only to allow Citrix QA to run some tests that
interacted directly with the bridge. This feature hasn't been turned on
for some time, so it should not be necessary any longer.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Before this patch, special packets such as LACP and CFM messages
were only processed if they had NORMAL open flow actions. With
this patch these messages are always processed unless originated in
ofproto_send_packet().
Otherwise when we add support for saving and restoring configuration
of internal devices around kernel module unload and reload, there's
no easy way for the "restore" code to tell when all the interfaces
should be set up and ready for configuration.
Much of the LACP status information attached to an interface is
moved from an enum to a bit mask in this commit. The main reason
to do this is to allow a link to be concurrently expired and
defaulted. With this commit, if a link enters an expired state,
but has never had its partner information update, it will properly
set the defaulted flag in its LACP messages.
Hardware sets the default LACP partner information to 0 so this
commit follows. The collector delay is a more interesting
case. Hardware sets it to 32768 and Linux sets it to 0. The
collector delay relates to a part of the LACP protocol which we
don't implement so we follow Linux in this case.
It looks to me like, once a CFM object is configured on an interface, it
will never get deconfigured. This fixes the problem.
Found by inspection. I haven't tested the original code or the fix.
Until now, tunnel vports have had a specific MTU, in the same way that
ordinary network devices have an MTU, but treating them this way does not
always make sense. For example, consider a datapath that has three ports:
the local port, a GRE tunnel to another host, and a physical port. If
the physical port is configured with a jumbo MTU, it should be possible to
send jumbo packets across the tunnel: the tunnel can do fragmentation or
the physical port traversed by the tunnel might have a jumbo MTU.
However, until now, tunnels always had a 1500-byte MTU by default. It
could be adjusted using ODP_VPORT_MTU_SET, but nothing actually did this.
One alternative would be to make ovs-vswitchd able to set the vport's MTU.
This commit, however, takes a different approach, of dropping the concept
of MTU entirely for tunnel vports. This also solves the problem described
above, without making any additional work for anyone.
I tested that, without this change, I could not send 1600-byte "pings"
between two machines whose NICs had 2000-byte MTUs that were connected to
vswitches that were in turn connected over GRE tunnels with the default
1500-byte MTU. With this change, it worked OK, regardless of the MTU of
the network traversed by the GRE tunnel.
This patch also makes "patch" ports MTU-less.
It might make sense to remove vport_set_mtu() and the associated callback
now, since ordinary network devices are the only vports that support it
now.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Suggested-by: Jesse Gross <jesse@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Bug #3728.
This commit implements a new bonding mode "balance-tcp" which takes
into account L4 flow information when hashing. If LACP negotiation
is unsuccessful it automatically falls back to "balance-slb" bonding.
Bug #4213.
This commit implements LACP, a protocol which allows directly
connected switches to automatically negotiate which links may
participate in bonds. This commit disables LACP by default. Once
sufficiently tested, LACP will be enabled in "active" mode on
bonded ports, and "passive" mode on all others.
Bug #4213.
This commit caches each interface's carrier status when bonding is
enabled. This allows port_update_bond_compat() to be called less
often, and will prove useful in future patches.
Jesse suggested this naming scheme, so I'm adjusting existing names to
fit it.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
None of the remaining dpif implementations have more than one name per
dpif, so there's no need for this function anymore.
Suggested-by: Jesse Gross <jesse@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Following this commit, "struct odp_port" is only used in Linux-specific
parts of OVS userspace code. This allows the actual Linux datapath
interface to evolve more freely.
Reviewed by Justin Pettit.
This gives network device implementations the opportunity to fetch an
existing device's configuration and store it as their arguments, so that
netdev clients can find out how an existing device is configured.
So far netdev-vport is the only implementation that needs to use this.
The next commit will add use by clients.
Reviewed by Justin Pettit.
One of the goals for Open vSwitch is to decouple kernel and userspace
software, so that either one can be upgraded or rolled back independent of
the other. To do this in full generality, it must be possible to add new
features to the kernel vport layer without changing userspace software. In
turn, that means that the odp_port structure must become variable-length.
This does not, however, fit in well with the ODP_PORT_LIST ioctl in its
current form, because that would require userspace to know how much space
to allocate for each port in advance, or to allocate as much space as
could possibly be needed. Neither choice is very attractive.
This commit prepares for a different solution, by replacing ODP_PORT_LIST
by a new ioctl ODP_VPORT_DUMP that retrieves information about a single
vport from the datapath on each call. It is much cleaner to allocate the
maximum amount of space for a single vport than to do so for possibly a
large number of vports.
It would be faster to retrieve a number of vports in batch instead of just
one at a time, but that will naturally happen later when the kernel
datapath interface is changed to use Netlink, so this patch does not bother
with it.
The Netlink version won't need to take the starting port number from
userspace, since Netlink sockets can keep track of that state as part
of their "dump" feature.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
New columns in Interface table: admin_state, link_state, link_speed, duplex,
mtu.
New keys in status map in Interface table: driver_name, driver_version,
firmware_version.
Requested-by: Peter Balland <pballand@nicira.com>
Bug #4299.
Causes bond/show to display the miimon interval for bonded ports
when configured to use miimon instead of carrier.
Requested-by: Michael Mao <mmao@nicira.com>
Bug #4418.
This commit removes the tunnel_egress_iface column from the
interface table and moves it's data to the status column. In the
process it reverts the database to version 1.0.0.
Since GRE-over-IPsec is a proper tunnel type and no longer configured
through "other_config", we can remove this function that folded an
interface's "other_confg" into "options".
Until now, compose_actions() has allocated enough "struct dst"s on the
stack for a worst-case flow, one that floods packets with the maximum
number of ports and mirrors. When the code was written this was correct.
However, now the number of ports is no longer known at compile time. The
maximum number, 65535, would require (65536 * (32 + 1) * 4) == 8 MB of
stack space, which is a lot. So this commit fixes the problem a different
way, by allocating the "struct dst"s dynamically when necessary.
This is a bug fix, but not a very serious one, because it could only
become a buffer overflow with a large number of mirrors.
There's no reason to allocate the bals[] array on the stack here, since
this is not on any fast-path.
As an alternative, we could limit the number of interfaces on a single
bond to some reasonable maximum, such as 8 or 32, but this commit's change
is simpler.
Currently the type of the datapath action length is mixture of
size_t and unsigned int. However, size_t is really defined as an
unsigned long, which causes the build to fail on 64-bit platforms.
This consistently uses size_t.
In the medium term, we plan to migrate the datapath to use Netlink as its
communication channel. In the short term, we need to be able to have
actions with 64-bit arguments but "struct odp_action" only has room for
48 bits. So this patch shifts to variable-length arguments using Netlink
attributes, which starts in on the Netlink transition and makes 64-bit
arguments possible at the same time.
Signed-off-by: Ben Pfaff <blp@nicira.com>
Acked-by: Jesse Gross <jesse@nicira.com>
Before this patch one could specify a mac address as part of the
bond/migrate command. This will no longer make sense as bond
hashing becomes more complicated.
