When adding TC rules we save the prio so can reuse same prio
for same mask since different mask will have to use different prio.
The multi mask per prio probe broke this by using a prio but
get_prio_for_tc_flower() didn't know about it.
Also multi mask per prio support changes the hash calculation.
It's best the probe will add and del the ingress qdisc to have a clean start
after it.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Acked-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Add a new class op for netdevs to get the block_id if one exists. The
block_id is used in offload ops to group multiple qdiscs together.
Stub calls are made to the new class op (implementation to follow in
further patches). The default block_id of 0 (no block) will be used in
these cases.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Blocks, in tc classifiers, allow the grouping of multiple qdiscs with an
associated block id. Whenever a filter is added to/removed from this
block, the filter is added to/removed from all associated qdiscs.
Extend TC offload functions to take a block id as a parameter. If the id
is zero then the dqisc is not considered part of a block.
Signed-off-by: John Hurley <john.hurley@netronome.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Reviewed-by: Dirk van der Merwe <dirk.vandermerwe@netronome.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Previously, any rule that is offloaded via a netdev, not necessarily
to the HW, would be reported as "offloaded". This patch fixes this
misalignment, and introduces the 'dp' state, as follows:
rule is in HW via TC offload -> offloaded=yes dp:tc
rule is in not HW over TC DP -> offloaded=no dp:tc
rule is in not HW over OVS DP -> offloaded=no dp:ovs
To achieve this, the flows's 'offloaded' flag was encapsulated in a new
attrs struct, which contains the offloaded state of the flow and the
DP layer the flow is handled in, and instead of setting the flow's
'offloaded' state based solely on the type of dump it was acquired
via, for netdev flows it now sends the new attrs struct to be
collected along with the rest of the flow via the netdev, allowing
it to be set per flow.
For TC offloads, the offloaded state is set based on the 'in_hw' and
'not_in_hw' flags received from the TC as part of the flower. If no
such flag was received, due to lack of kernel support, it defaults
to true.
Signed-off-by: Gavi Teitz <gavi@mellanox.com>
Acked-by: Roi Dayan <roid@mellanox.com>
[simon: resolved conflict in lib/dpctl.man]
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Presumably this was meant to be all-one-bits but it wasn't. It also didn't
have the right endianness for an ovs_be16, so "sparse" complained.
CC: Paul Blakey <paulb@mellanox.com>
CC: Simon Horman <simon.horman@netronome.com>
Fixes: d00eeded6a9b ("netdev-tc-offloads: Probe for allowing multiple masks on single priority")
Signed-off-by: Ben Pfaff <blp@ovn.org>
Acked-by: Justin Pettit <jpettit@ovn.org>
OVS datapath flows aren't overlapping, so having their tc flower
counterparts be prioritized makes no sense, we did so because of a
tc flower restriction.
Kernel tc flower added support for multiple masks on a single flower
instance (there's an instance per priority) to remove this restriction.
Probe for this once on first added port, and if available, use a
single priority per ethertype when inserting tc flower flows.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
When OVS DP passes the actions to TC library, we save all the
actions in data structure tc_flower and each action type has its
own field in tc_flower. So when TC library passes the actions to
kernel, actually the actions order is lost.
We add an actions array in tc_flower to keep the actions order
in this patch.
Signed-off-by: Chris Mi <chrism@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Fragment mask (any and later) always exists so we need to test
for FLOW_NW_FRAG_LATER only if the state is FLOW_NW_FRAG_ANY.
Before this fix we could pass frag no and first at the same time to TC
which is also not tested there for bad frag state.
This fix make sure we only pass frag first/later if is frag.
Fixes: 83e866067ea6 ("netdev-tc-offloads: Add support for IP fragmentation")
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Add support for frag no, first and later.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Shahar Klein <shahark@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
This function uses local array set_buff[] to store Netlink attributes.
It declares set_buff as an array of character pointers, which is a strange
type for a buffer of non-character-pointer objects. In OVS it is
customary to use an ofpbuf with a stub of uint64_t objecs (to ensure
proper alignment, otherwise uint8_t would be more usual). This commit
changes to that more usual form.
Signed-off-by: Ben Pfaff <blp@ovn.org>
OVS action set always has a single nested OVS_KEY_ATTR_*
so there is no need to iterate it.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Travis complains about variable length array usage, use
a fixed size instead.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Currently, when an offloaded DP flow is deleted, the related stats
are unconditionally cleared. As a result the counters for the
originating open flow are corrupted.
This change addresses the issue updating the DP stats with the current
values provided by the flower APIs before deleting the tc filter, as
currently done by others DP providers.
Tested vs different OVS H/W offload devices, verifying that the open flow
stats are correct after the expiration of the related, H/W offloaded DP
flow.
Fixes: 30b6b047260b ("netdev-tc-offloads: Implement netdev flow del using tc interface")
CC: Paul Blakey <paulb@mellanox.com>
Reported-by: Kumar Sanghvi <kumaras@chelsio.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Implement support for offloading ovs action set using
tc header rewrite action.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
This commit fixes offloading when tunnel id is 0 by checking
for valid tunnel dst ipv4 or ipv6, like done in the normal flow.
Also need to reset the tunnel mask outside the condition
because in the special case, when adding a port with key=flow,
the match for tunnel_set gets a tunnel key id 0 with a mask.
If we don't reset the mask we'll fail later when we check
for unknown attributes that the mask was not reset.
Fixes: 8f283af89298 ("netdev-tc-offloads: Implement netdev flow put using tc interface")
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
There is an identical log msg from multiple api calls.
Add a prefix to identify the source function of the log msg.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Joe Stringer <joe@ovn.org>
Add support to offload rules matching on tcp_flags.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Joe Stringer <joe@ovn.org>
Add matching (put/dump) on ip layer ttl.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Simon Horman <simon.horman@netronome.com>
Signed-off-by: Joe Stringer <joe@ovn.org>
After splitting the unions of tcp/udp the sctp was forgotten
when parsing flower back to match.
Fixes: 2b1d9fa90909 ("tc: Split IPs and transport layer ports unions in flower struct")
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Ben Pfaff <blp@ovn.org>
Acked-by: Simon Horman <simon.horman@netronome.com>
There is no need to parse ip related fields if eth type is not ip.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Ben Pfaff <blp@ovn.org>
Split dst/src_port and ipv4/ipv6 union so we can
distingush them easily for later features.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Ben Pfaff <blp@ovn.org>
It's basically what is being passed today and passing a specific
type adds a compiler type check.
Signed-off-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Implement SCTP source and destination ports support for flower.
Signed-off-by: Vlad Buslov <vladbu@mellanox.com>
Reviewed-by: Paul Blakey <paulb@mellanox.com>
Acked-by: Roi Dayan <roid@mellanox.com>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Fix minor style variations and unnecessary includes.
Signed-off-by: Joe Stringer <joe@ovn.org>
Tested-by: Greg Rose <gvrose8192@gmail.com>
Acked-by: Greg Rose <gvrose8192@gmail.com>
Search the requested ufid for a offloaded flow, and if found,
dump and parse it back to required format.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Reviewed-by: Simon Horman <simon.horman@netronome.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Currently only tunnel offload is supported.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Flower classifer requires a different priority per mask,
so we hash the mask and generate a new priority for
each new mask used.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Flows offloaded to tc are identified by priority
and handle pair while OVS flows are identified by ufid.
Added a hash map to convert between the two for later
retrieval and deleting of offloaded flows.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
Add a new API interface for offloading dpif flows to netdev.
The API consist on the following:
flow_put - offload a new flow
flow_get - query an offloaded flow
flow_del - delete an offloaded flow
flow_flush - flush all offloaded flows
flow_dump_* - dump all offloaded flows
In upcoming commits we will introduce an implementation of this
API for netdev-linux.
Signed-off-by: Paul Blakey <paulb@mellanox.com>
Reviewed-by: Roi Dayan <roid@mellanox.com>
Acked-by: Flavio Leitner <fbl@sysclose.org>
Signed-off-by: Simon Horman <simon.horman@netronome.com>
