Alow users to push down more labels per MPLS encap. Similar to LSR case,
move label array to the end of mpls_iptunnel_encap and allocate based on
the number of labels for the route.
For consistency with the LSR case, re-use the same maximum number of
labels.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Allow users to push down more labels per MPLS route. With the previous
patches, no memory allocations are based on MAX_NEW_LABELS; the limit
is only used to keep userspace in check.
At this point MAX_NEW_LABELS is only used for mpls_route_config (copying
route data from userspace) and processing nexthops looking for the max
number of labels across the route spec.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move labels to the end of mpls_nh as a 0-sized array and within mpls_route
move the via for a nexthop after the mpls_nh. The new layout becomes:
+----------------------+
| mpls_route |
+----------------------+
| mpls_nh 0 |
+----------------------+
| alignment padding | 4 bytes for odd number of labels; 0 for even
+----------------------+
| via[rt_max_alen] 0 |
+----------------------+
| alignment padding | via's aligned on sizeof(unsigned long)
+----------------------+
| ... |
+----------------------+
| mpls_nh n-1 |
+----------------------+
| via[rt_max_alen] n-1 |
+----------------------+
Memory allocated for nexthop + via is constant across all nexthops and
their via. It is based on the maximum number of labels across all nexthops
and the maximum via length. The size is saved in the mpls_route as
rt_nh_size. Accessing a nexthop becomes rt->rt_nh + index * rt->rt_nh_size.
The offset of the via address from a nexthop is saved as rt_via_offset
so that given an mpls_nh pointer the via for that hop is simply
nh + rt->rt_via_offset.
With prior code, memory allocated per mpls_route with 1 nexthop:
via is an ethernet address - 64 bytes
via is an ipv4 address - 64
via is an ipv6 address - 72
With this patch set, memory allocated per mpls_route with 1 nexthop and
1 or 2 labels:
via is an ethernet address - 56 bytes
via is an ipv4 address - 56
via is an ipv6 address - 64
The 8-byte reduction is due to the previous patch; the change introduced
by this patch has no impact on the size of allocations for 1 or 2 labels.
Performance impact of this change was examined using network namespaces
with veth pairs connecting namespaces. ns0 inserts the packet to the
label-switched path using an lwt route with encap mpls. ns1 adds 1 or 2
labels depending on test, ns2 (and ns3 for 2-label test) pops the label
and forwards. ns3 (or ns4) for a 2-label is the destination. Similar
series of namespaces used for 2-nexthop test.
Intent is to measure changes to latency (overhead in manipulating the
packet) in the forwarding path. Tests used netperf with UDP_RR.
IPv4: current patches
1 label, 1 nexthop 29908 30115
2 label, 1 nexthop 29071 29612
1 label, 2 nexthop 29582 29776
2 label, 2 nexthop 29086 29149
IPv6: current patches
1 label, 1 nexthop 24502 24960
2 label, 1 nexthop 24041 24407
1 label, 2 nexthop 23795 23899
2 label, 2 nexthop 23074 22959
In short, the change has no effect to a modest increase in performance.
This is expected since this patch does not really have an impact on routes
with 1 or 2 labels (the current limit) and 1 or 2 nexthops.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Number of nexthops and number of alive nexthops are tracked using an
unsigned int. A route should never have more than 255 nexthops so
convert both to u8. Update all references and intermediate variables
to consistently use u8 as well.
Shrinks the size of mpls_route from 32 bytes to 24 bytes with a 2-byte
hole before the nexthops.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The number of alive nexthops for a route (rt->rt_nhn_alive) and the
flags for a next hop (nh->nh_flags) are modified by netdev event
handlers. The event handlers run with rtnl_lock held so updates are
always done with the lock held. The packet path accesses the fields
under the rcu lock. Since those fields can change at any moment in
the packet path, both fields should be accessed using READ_ONCE. Updates
to both fields should use WRITE_ONCE.
Update mpls_select_multipath (packet path) and mpls_ifdown and mpls_ifup
(event handlers) accordingly.
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Provide the ability to control on a per-route basis whether the TTL
value from an MPLS packet is propagated to an IPv4/IPv6 packet when
the last label is popped as per the theoretical model in RFC 3443
through a new route attribute, RTA_TTL_PROPAGATE which can be 0 to
mean disable propagation and 1 to mean enable propagation.
In order to provide the ability to change the behaviour for packets
arriving with IPv4/IPv6 Explicit Null labels and to provide an easy
way for a user to change the behaviour for all existing routes without
having to reprogram them, a global knob is provided. This is done
through the addition of a new per-namespace sysctl,
"net.mpls.ip_ttl_propagate", which defaults to enabled. If the
per-route attribute is set (either enabled or disabled) then it
overrides the global configuration.
Signed-off-by: Robert Shearman <rshearma@brocade.com>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Tested-by: David Ahern <dsa@cumulusnetworks.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add netconf support to MPLS. Allows userpsace to learn and be notified
of changes to 'input' enable setting per interface.
Acked-by: Nicolas Dichtel <nicolas.dichtel@6wind.com>
Signed-off-by: David Ahern <dsa@cumulusnetworks.com>
Acked-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Having MPLS packet stats is useful for observing network operation and
for diagnosing network problems. In the absence of anything better,
RFC2863 and RFC3813 are used for guidance for which stats to expose
and the semantics of them. In particular rx_noroutes maps to in
unknown protos in RFC2863. The stats are exposed to userspace via
AF_MPLS attributes embedded in the IFLA_STATS_AF_SPEC attribute of
RTM_GETSTATS messages.
All the introduced fields are 64-bit, even error ones, to ensure no
overflow with long uptimes. Per-CPU counters are used to avoid
cache-line contention on the commonly used fields. The other fields
have also been made per-CPU for code to avoid performance problems in
error conditions on the assumption that on some platforms the cost of
atomic operations could be more expensive than sending the packet
(which is what would be done in the success case). If that's not the
case, we could instead not use per-CPU counters for these fields.
Only unicast and non-fragment are exposed at the moment, but other
counters can be exposed in the future either by adding to the end of
struct mpls_link_stats or by additional netlink attributes in the
AF_MPLS IFLA_STATS_AF_SPEC nested attribute.
Signed-off-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This will be also used by openvswitch.
Signed-off-by: Jiri Benc <jbenc@redhat.com>
Acked-by: David Ahern <dsa@cumulusnetworks.com>
Acked-by: Pravin B Shelar <pshelar@ovn.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
Adds support for RTNH_F_DEAD and RTNH_F_LINKDOWN flags on mpls
routes due to link events. Also adds code to ignore dead
routes during route selection.
Unlike ip routes, mpls routes are not deleted when the route goes
dead. This is current mpls behaviour and this patch does not change
that. With this patch however, routes will be marked dead.
dead routes are not notified to userspace (this is consistent with ipv4
routes).
dead routes:
-----------
$ip -f mpls route show
100
nexthop as to 200 via inet 10.1.1.2 dev swp1
nexthop as to 700 via inet 10.1.1.6 dev swp2
$ip link set dev swp1 down
$ip link show dev swp1
4: swp1: <BROADCAST,MULTICAST> mtu 1500 qdisc pfifo_fast state DOWN mode
DEFAULT group default qlen 1000
link/ether 00:02:00:00:00:01 brd ff:ff:ff:ff:ff:ff
$ip -f mpls route show
100
nexthop as to 200 via inet 10.1.1.2 dev swp1 dead linkdown
nexthop as to 700 via inet 10.1.1.6 dev swp2
linkdown routes:
----------------
$ip -f mpls route show
100
nexthop as to 200 via inet 10.1.1.2 dev swp1
nexthop as to 700 via inet 10.1.1.6 dev swp2
$ip link show dev swp1
4: swp1: <BROADCAST,MULTICAST,UP,LOWER_UP> mtu 1500 qdisc pfifo_fast
state UP mode DEFAULT group default qlen 1000
link/ether 00:02:00:00:00:01 brd ff:ff:ff:ff:ff:ff
/* carrier goes down */
$ip link show dev swp1
4: swp1: <NO-CARRIER,BROADCAST,MULTICAST,UP> mtu 1500 qdisc pfifo_fast
state DOWN mode DEFAULT group default qlen 1000
link/ether 00:02:00:00:00:01 brd ff:ff:ff:ff:ff:ff
$ip -f mpls route show
100
nexthop as to 200 via inet 10.1.1.2 dev swp1 linkdown
nexthop as to 700 via inet 10.1.1.6 dev swp2
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Acked-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Nexthops for MPLS routes have a via address field sized for the
largest via address that is expected, which is 32 bytes. This means
that in the most common case of having ipv4 via addresses, 28 bytes of
memory more than required are used per nexthop. In the other common
case of an ipv6 nexthop then 16 bytes more than required are
used. With large numbers of MPLS routes this extra memory usage could
start to become significant.
To avoid allocating memory for a maximum length via address when not
all of it is required and to allow for ease of iterating over
nexthops, then the via addresses are changed to be stored in the same
memory block as the route and nexthops, but in an array after the end
of the array of nexthops. New accessors are provided to retrieve a
pointer to the via address.
To allow for O(1) access without having to store a pointer or offset
per nh, the via address for each nexthop is sized according to the
maximum via address for any nexthop in the route, which is stored in a
new route field, rt_max_alen, but this is in an existing hole in
struct mpls_route so it doesn't increase the size of the
structure. Each via address is ensured to be aligned to VIA_ALEN_ALIGN
to account for architectures that don't allow unaligned accesses.
Signed-off-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This patch adds support for MPLS multipath routes.
Includes following changes to support multipath:
- splits struct mpls_route into 'struct mpls_route + struct mpls_nh'
- 'struct mpls_nh' represents a mpls nexthop label forwarding entry
- moves mpls route and nexthop structures into internal.h
- A mpls_route can point to multiple mpls_nh structs
- the nexthops are maintained as a array (similar to ipv4 fib)
- In the process of restructuring, this patch also consistently changes
all labels to u8
- Adds support to parse/fill RTA_MULTIPATH netlink attribute for
multipath routes similar to ipv4/v6 fib
- In this patch, the multipath route nexthop selection algorithm
simply returns the first nexthop. It is replaced by a
hash based algorithm from Robert Shearman in the next patch
- mpls_route_update cleanup: remove 'dev' handling in mpls_route_update.
mpls_route_update though implemented to update based on dev, it was
never used that way. And the dev handling gets tricky with multiple
nexthops. Cannot match against any single nexthops dev. So, this patch
removes the unused 'dev' handling in mpls_route_update.
- dead route/path handling will be implemented in a subsequent patch
Example:
$ip -f mpls route add 100 nexthop as 200 via inet 10.1.1.2 dev swp1 \
nexthop as 700 via inet 10.1.1.6 dev swp2 \
nexthop as 800 via inet 40.1.1.2 dev swp3
$ip -f mpls route show
100
nexthop as to 200 via inet 10.1.1.2 dev swp1
nexthop as to 700 via inet 10.1.1.6 dev swp2
nexthop as to 800 via inet 40.1.1.2 dev swp3
Signed-off-by: Roopa Prabhu <roopa@cumulusnetworks.com>
Acked-by: Robert Shearman <rshearma@brocade.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
The mpls device is used in an RCU read context without a lock being
held. As the memory is freed without waiting for the RCU grace period
to elapse, the freed memory could still be in use.
Address this by using kfree_rcu to free the memory for the mpls device
after the RCU grace period has elapsed.
Fixes: 03c57747a7 ("mpls: Per-device MPLS state")
Signed-off-by: Robert Shearman <rshearma@brocade.com>
Acked-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Move to include/uapi/linux/mpls.h to be externally visibile.
Signed-off-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
An MPLS network is a single trust domain where the edges must be in
control of what labels make their way into the core. The simplest way
of ensuring this is for the edge device to always impose the labels,
and not allow forward labeled traffic from untrusted neighbours. This
is achieved by allowing a per-device configuration of whether MPLS
traffic input from that interface should be processed or not.
To be secure by default, the default state is changed to MPLS being
disabled on all interfaces unless explicitly enabled and no global
option is provided to change the default. Whilst this differs from
other protocols (e.g. IPv6), network operators are used to explicitly
enabling MPLS forwarding on interfaces, and with the number of links
to the MPLS core typically fairly low this doesn't present too much of
a burden on operators.
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Robert Shearman <rshearma@brocade.com>
Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Add per-device MPLS state to supported interfaces. Use the presence of
this state in mpls_route_add to determine that this is a supported
interface.
Use the presence of mpls_dev to drop packets that arrived on an
unsupported interface - previously they were allowed through.
Cc: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: Robert Shearman <rshearma@brocade.com>
Reviewed-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Reading and writing addresses in network byte order in netlink is
traditional and I see no reason to change that. MPLS is interesting
as effectively it has variabely length addresses (the MPLS label
stack). To represent these variable length addresses in netlink
I use a valid MPLS label stack (complete with stop bit).
This achieves two things: a well defined existing format is used,
and the data can be interpreted without looking at it's length.
Not needed to look at the length to decode the variable length
network representation allows existing userspace functions
such as inet_ntop to be used without needed to change their
prototype.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This change adds a new Kconfig option MPLS_ROUTING.
The core of this change is the code to look at an mpls packet received
from another machine. Look that packet up in a routing table and
forward the packet on.
Support of MPLS over ATM is not considered or attempted here. This
implemntation follows RFC3032 and implements the MPLS shim header that
can pass over essentially any network.
What RFC3021 refers to as the as the Incoming Label Map (ILM) I call
net->mpls.platform_label[]. What RFC3031 refers to as the Next Label
Hop Forwarding Entry (NHLFE) I call mpls_route. Though calling it the
label fordwarding information base (lfib) might also be valid.
Further the implemntation forwards packets as described in RFC3032.
There is no need and given the original motivation for MPLS a strong
discincentive to have a flexible label forwarding path. In essence
the logic is the topmost label is read, looked up, removed, and
replaced by 0 or more new lables and the sent out the specified
interface to it's next hop.
Quite a few optional features are not implemented here. Among them
are generation of ICMP errors when the TTL is exceeded or the packet
is larger than the next hop MTU (those conditions are detected and the
packets are dropped instead of generating an icmp error). The traffic
class field is always set to 0. The implementation focuses on IP over
MPLS and does not handle egress of other kinds of protocols.
Instead of implementing coordination with the neighbour table and
sorting out how to input next hops in a different address family (for
which there is value). I was lazy and implemented a next hop mac
address instead. The code is simpler and there are flavor of MPLS
such as MPLS-TP where neither an IPv4 nor an IPv6 next hop is
appropriate so a next hop by mac address would need to be implemented
at some point.
Two new definitions AF_MPLS and PF_MPLS are exposed to userspace.
Decoding the mpls header must be done by first byeswapping a 32bit bit
endian word into the local cpu endian and then bit shifting to extract
the pieces. There is no C bit-field that can represent a wire format
mpls header on a little endian machine as the low bits of the 20bit
label wind up in the wrong half of third byte. Therefore internally
everything is deal with in cpu native byte order except when writing
to and reading from a packet.
For management simplicity if a label is configured to forward out
an interface that is down the packet is dropped early. Similarly
if an network interface is removed rt_dev is updated to NULL
(so no reference is preserved) and any packets for that label
are dropped. Keeping the label entries in the kernel allows
the kernel label table to function as the definitive source
of which labels are allocated and which are not.
Signed-off-by: "Eric W. Biederman" <ebiederm@xmission.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
David Ahern