In the commit referred to below we eliminated sending of the 'gap'
indicator in regular ACK messages, reserving this to explicit NACK
ditto.
Unfortunately we missed to also eliminate building of the 'gap block'
area in ACK messages. This area is meant to report gaps in the
received packet sequence following the initial gap, so that lost
packets can be retransmitted earlier and received out-of-sequence
packets can be released earlier. However, the interpretation of those
blocks is dependent on a complete and correct sequence of gaps and
acks. Hence, when the initial gap indicator is missing a single gap
block will be interpreted as an acknowledgment of all preceding
packets. This may lead to packets being released prematurely from the
sender's transmit queue, with easily predicatble consequences.
We now fix this by not building any gap block area if there is no
initial gap to report.
Fixes: commit 02288248b0 ("tipc: eliminate gap indicator from ACK messages")
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We introduce a simple variable window congestion control for links.
The algorithm is inspired by the Reno algorithm, covering both 'slow
start', 'congestion avoidance', and 'fast recovery' modes.
- We introduce hard lower and upper window limits per link, still
different and configurable per bearer type.
- We introduce a 'slow start theshold' variable, initially set to
the maximum window size.
- We let a link start at the minimum congestion window, i.e. in slow
start mode, and then let is grow rapidly (+1 per rceived ACK) until
it reaches the slow start threshold and enters congestion avoidance
mode.
- In congestion avoidance mode we increment the congestion window for
each window-size number of acked packets, up to a possible maximum
equal to the configured maximum window.
- For each non-duplicate NACK received, we drop back to fast recovery
mode, by setting the both the slow start threshold to and the
congestion window to (current_congestion_window / 2).
- If the timeout handler finds that the transmit queue has not moved
since the previous timeout, it drops the link back to slow start
and forces a probe containing the last sent sequence number to the
sent to the peer, so that this can discover the stale situation.
This change does in reality have effect only on unicast ethernet
transport, as we have seen that there is no room whatsoever for
increasing the window max size for the UDP bearer.
For now, we also choose to keep the limits for the broadcast link
unchanged and equal.
This algorithm seems to give a 50-100% throughput improvement for
messages larger than MTU.
Suggested-by: Xin Long <lucien.xin@gmail.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we increase the link tranmsit window we often observe the following
scenario:
1) A STATE message bypasses a sequence of traffic packets and arrives
far ahead of those to the receiver. STATE messages contain a
'peers_nxt_snt' field to indicate which was the last packet sent
from the peer. This mechanism is intended as a last resort for the
receiver to detect missing packets, e.g., during very low traffic
when there is no packet flow to help early loss detection.
3) The receiving link compares the 'peer_nxt_snt' field to its own
'rcv_nxt', finds that there is a gap, and immediately sends a
NACK message back to the peer.
4) When this NACKs arrives at the sender, all the requested
retransmissions are performed, since it is a first-time request.
Just like in the scenario described in the previous commit this leads
to many redundant retransmissions, with decreased throughput as a
consequence.
We fix this by adding two more conditions before we send a NACK in
this sitution. First, the deferred queue must be empty, so we cannot
assume that the potential packet loss has already been detected by
other means. Second, we check the 'peers_snd_nxt' field only in probe/
probe_reply messages, thus turning this into a true mechanism of last
resort as it was really meant to be.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When we increase the link send window we sometimes observe the
following scenario:
1) A packet #N arrives out of order far ahead of a sequence of older
packets which are still under way. The packet is added to the
deferred queue.
2) The missing packets arrive in sequence, and for each 16th of them
an ACK is sent back to the receiver, as it should be.
3) When building those ACK messages, it is checked if there is a gap
between the link's 'rcv_nxt' and the first packet in the deferred
queue. This is always the case until packet number #N-1 arrives, and
a 'gap' indicator is added, effectively turning them into NACK
messages.
4) When those NACKs arrive at the sender, all the requested
retransmissions are done, since it is a first-time request.
This sometimes leads to a huge amount of redundant retransmissions,
causing a drop in max throughput. This problem gets worse when we
in a later commit introduce variable window congestion control,
since it drops the link back to 'fast recovery' much more often
than necessary.
We now fix this by not sending any 'gap' indicator in regular ACK
messages. We already have a mechanism for sending explicit NACKs
in place, and this is sufficient to keep up the packet flow.
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
ipv6_stub uses the ip6_dst_lookup function to allow other modules to
perform IPv6 lookups. However, this function skips the XFRM layer
entirely.
All users of ipv6_stub->ip6_dst_lookup use ip_route_output_flow (via the
ip_route_output_key and ip_route_output helpers) for their IPv4 lookups,
which calls xfrm_lookup_route(). This patch fixes this inconsistent
behavior by switching the stub to ip6_dst_lookup_flow, which also calls
xfrm_lookup_route().
This requires some changes in all the callers, as these two functions
take different arguments and have different return types.
Fixes: 5f81bd2e5d ("ipv6: export a stub for IPv6 symbols used by vxlan")
Reported-by: Xiumei Mu <xmu@redhat.com>
Signed-off-by: Sabrina Dubroca <sd@queasysnail.net>
Signed-off-by: David S. Miller <davem@davemloft.net>
Scenario:
1. A client socket initiates a SYN message to a listening socket.
2. The send link is congested, the SYN message is put in the
send link and a wakeup message is put in wakeup queue.
3. The congestion situation is abated, the wakeup message is
pulled out of the wakeup queue. Function tipc_sk_push_backlog()
is called to send out delayed messages by Nagle. However,
the client socket is still in CONNECTING state. So, it sends
the SYN message in the socket write queue to the listening socket
again.
4. The listening socket receives the first SYN message and creates
first server socket. The client socket receives ACK- and establishes
a connection to the first server socket. The client socket closes
its connection with the first server socket.
5. The listening socket receives the second SYN message and creates
second server socket. The second server socket sends ACK- to the
client socket, but it has been closed. It results in connection
reset error when reading from the server socket in user space.
Solution: return from function tipc_sk_push_backlog() immediately
if there is pending SYN message in the socket write queue.
Fixes: c0bceb97db ("tipc: add smart nagle feature")
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In function __tipc_shutdown(), the timeout value passed to
tipc_wait_for_cond() is not jiffies.
This commit fixes it by converting that value from milliseconds
to jiffies.
Fixes: 365ad353c2 ("tipc: reduce risk of user starvation during link congestion")
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When tipc_sk_timeout() is executed but user space is grabbing
ownership, this function rearms itself and returns. However, the
socket reference counter is not reduced. This causes potential
unexpected behavior.
This commit fixes it by calling sock_put() before tipc_sk_timeout()
returns in the above-mentioned case.
Fixes: afe8792fec ("tipc: refactor function tipc_sk_timeout()")
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Acked-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
When initiating a connection message to a server side, the connection
message is cloned and added to the socket write queue. However, if the
cloning is failed, only the socket write queue is purged. It causes
memory leak because the original connection message is not freed.
This commit fixes it by purging the list of connection message when
it cannot be cloned.
Fixes: 6787927475 ("tipc: buffer overflow handling in listener socket")
Reported-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Acked-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 4f07b80c97 ("tipc: check msg->req data len in
tipc_nl_compat_bearer_disable") the same patch code was copied into
routines: tipc_nl_compat_bearer_disable(),
tipc_nl_compat_link_stat_dump() and tipc_nl_compat_link_reset_stats().
The two link routine occurrences should have been modified to check
the maximum link name length and not bearer name length.
Fixes: 4f07b80c97 ("tipc: check msg->reg data len in tipc_nl_compat_bearer_disable")
Signed-off-by: John Rutherford <john.rutherford@dektech.com.au>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
rhashtable_lookup_fast() internally calls rcu_read_lock() then,
calls rhashtable_lookup(). So if rcu_read_lock() is already held,
rhashtable_lookup() is enough.
Signed-off-by: Taehee Yoo <ap420073@gmail.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
It is observed that TIPC service binding order will not be kept in the
publication event report to user if the service is subscribed after the
bindings.
For example, services are bound by application in the following order:
Server: bound port A to {18888,66,66} scope 2
Server: bound port A to {18888,33,33} scope 2
Now, if a client subscribes to the service range (e.g. {18888, 0-100}),
it will get the 'TIPC_PUBLISHED' events in that binding order only when
the subscription is started before the bindings.
Otherwise, if started after the bindings, the events will arrive in the
opposite order:
Client: received event for published {18888,33,33}
Client: received event for published {18888,66,66}
For the latter case, it is clear that the bindings have existed in the
name table already, so when reported, the events' order will follow the
order of the rbtree binding nodes (- a node with lesser 'lower'/'upper'
range value will be first).
This is correct as we provide the tracking on a specific service status
(available or not), not the relationship between multiple services.
However, some users expect to see the same order of arriving events
irrespective of when the subscription is issued. This turns out to be
easy to fix. We now add functionality to ensure that publication events
always are issued in the same temporal order as the corresponding
bindings were performed.
v2: replace the unnecessary macro - 'publication_after()' with inline
function.
v3: reuse 'time_after32()' instead of reinventing the same exact code.
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
When setting up a cluster with non-replicast/replicast capability
supported. This capability will be disabled for broadcast send link
in order to be backwards compatible.
However, when these non-support nodes left and be removed out the cluster.
We don't update this capability on broadcast send link. Then, some of
features that based on this capability will also disabling as unexpected.
In this commit, we make sure the broadcast send link capabilities will
be re-calculated as soon as a node removed/rejoined a cluster.
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The tipc prefix for log messages generated by tipc was
removed in commit 07f6c4bc04 ("tipc: convert tipc reference
table to use generic rhashtable").
This is still a useful prefix so add it back.
Signed-off-by: Matt Bennett <matt.bennett@alliedtelesis.co.nz>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 25b0b9c4e8 ("tipc: handle collisions of 32-bit node address
hash values"), the 32-bit node address only generated after one second
trial period expired. However the self's addr in struct tipc_monitor do
not update according to node address generated. This lead to it is
always zero as initial value. As result, sorting algorithm using this
value does not work as expected, neither neighbor monitoring framework.
In this commit, we add a fix to update self's addr when 32-bit node
address generated.
Fixes: 25b0b9c4e8 ("tipc: handle collisions of 32-bit node address hash values")
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Variable err is not uninitialized and hence can potentially contain
any garbage value. This may cause an error when logical or'ing the
return values from the calls to functions crypto_aead_setauthsize or
crypto_aead_setkey. Fix this by setting err to the return of
crypto_aead_setauthsize rather than or'ing in the return into the
uninitialized variable
Addresses-Coverity: ("Uninitialized scalar variable")
Fixes: fc1b6d6de2 ("tipc: introduce TIPC encryption & authentication")
Signed-off-by: Colin Ian King <colin.king@canonical.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit adds two netlink commands to TIPC in order for user to be
able to set or remove AEAD keys:
- TIPC_NL_KEY_SET
- TIPC_NL_KEY_FLUSH
When the 'KEY_SET' is given along with the key data, the key will be
initiated and attached to TIPC crypto. On the other hand, the
'KEY_FLUSH' command will remove all existing keys if any.
Acked-by: Ying Xue <ying.xue@windreiver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit offers an option to encrypt and authenticate all messaging,
including the neighbor discovery messages. The currently most advanced
algorithm supported is the AEAD AES-GCM (like IPSec or TLS). All
encryption/decryption is done at the bearer layer, just before leaving
or after entering TIPC.
Supported features:
- Encryption & authentication of all TIPC messages (header + data);
- Two symmetric-key modes: Cluster and Per-node;
- Automatic key switching;
- Key-expired revoking (sequence number wrapped);
- Lock-free encryption/decryption (RCU);
- Asynchronous crypto, Intel AES-NI supported;
- Multiple cipher transforms;
- Logs & statistics;
Two key modes:
- Cluster key mode: One single key is used for both TX & RX in all
nodes in the cluster.
- Per-node key mode: Each nodes in the cluster has one specific TX key.
For RX, a node requires its peers' TX key to be able to decrypt the
messages from those peers.
Key setting from user-space is performed via netlink by a user program
(e.g. the iproute2 'tipc' tool).
Internal key state machine:
Attach Align(RX)
+-+ +-+
| V | V
+---------+ Attach +---------+
| IDLE |---------------->| PENDING |(user = 0)
+---------+ +---------+
A A Switch| A
| | | |
| | Free(switch/revoked) | |
(Free)| +----------------------+ | |Timeout
| (TX) | | |(RX)
| | | |
| | v |
+---------+ Switch +---------+
| PASSIVE |<----------------| ACTIVE |
+---------+ (RX) +---------+
(user = 1) (user >= 1)
The number of TFMs is 10 by default and can be changed via the procfs
'net/tipc/max_tfms'. At this moment, as for simplicity, this file is
also used to print the crypto statistics at runtime:
echo 0xfff1 > /proc/sys/net/tipc/max_tfms
The patch defines a new TIPC version (v7) for the encryption message (-
backward compatibility as well). The message is basically encapsulated
as follows:
+----------------------------------------------------------+
| TIPCv7 encryption | Original TIPCv2 | Authentication |
| header | packet (encrypted) | Tag |
+----------------------------------------------------------+
The throughput is about ~40% for small messages (compared with non-
encryption) and ~9% for large messages. With the support from hardware
crypto i.e. the Intel AES-NI CPU instructions, the throughput increases
upto ~85% for small messages and ~55% for large messages.
By default, the new feature is inactive (i.e. no encryption) until user
sets a key for TIPC. There is however also a new option - "TIPC_CRYPTO"
in the kernel configuration to enable/disable the new code when needed.
MAINTAINERS | add two new files 'crypto.h' & 'crypto.c' in tipc
Acked-by: Ying Xue <ying.xue@windreiver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
When user sets RX key for a peer not existing on the own node, a new
node entry is needed to which the RX key will be attached. However,
since the peer node address (& capabilities) is unknown at that moment,
only the node-ID is provided, this commit allows the creation of a node
with only the data that we call as “preliminary”.
A preliminary node is not the object of the “tipc_node_find()” but the
“tipc_node_find_by_id()”. Once the first message i.e. LINK_CONFIG comes
from that peer, and is successfully decrypted by the own node, the
actual peer node data will be properly updated and the node will
function as usual.
In addition, the node timer always starts when a node object is created
so if a preliminary node is not used, it will be cleaned up.
The later encryption functions will also use the node timer and be able
to create a preliminary node automatically when needed.
Acked-by: Ying Xue <ying.xue@windreiver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
As a need to support the crypto asynchronous operations in the later
commits, apart from the current RCU mechanism for bearer pointer, we
add a 'refcnt' to the bearer object as well.
So, a bearer can be hold via 'tipc_bearer_hold()' without being freed
even though the bearer or interface can be disabled in the meanwhile.
If that happens, the bearer will be released then when the crypto
operation is completed and 'tipc_bearer_put()' is called.
Acked-by: Ying Xue <ying.xue@windreiver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, we scan over all network namespaces at each received
discovery message in order to check if the sending peer might be
present in a host local namespaces.
This is unnecessary since we can assume that a peer will not change its
location during an established session.
We now improve the condition for this testing so that we don't perform
any redundant scans.
Fixes: f73b12812a ("tipc: improve throughput between nodes in netns")
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
When preparing tunnel packets for the link failover or synchronization,
as for the safe algorithm, we added a dummy packet on the pair link but
never sent it out. In the case of failover, the pair link will be reset
anyway. But for link synching, it will always result in retransmission
of the dummy packet after that.
We have also observed that such the retransmission at the early stage
when a new node comes in a large cluster will take some time and hard
to be done, leading to the repeated retransmit failures and the link is
reset.
Since in commit 4929a932be ("tipc: optimize link synching mechanism")
we have already built a dummy 'TUNNEL_PROTOCOL' message on the new link
for the synchronization, there's no need for the dummy on the pair one,
this commit will skip it when the new mechanism takes in place. In case
nothing exists in the pair link's transmq, the link synching will just
start and stop shortly on the peer side.
The patch is backward compatible.
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Tested-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
With huge cluster (e.g >200nodes), the amount of that flow:
gap -> retransmit packet -> acked will take time in case of STATE_MSG
dropped/delayed because a lot of traffic. This lead to 1.5 sec tolerance
value criteria made link easy failure around 2nd, 3rd of failed
retransmission attempts.
Instead of re-introduced criteria of 99 faled retransmissions to fix the
issue, we increase failure detection timer to ten times tolerance value.
Fixes: 77cf8edbc0 ("tipc: simplify stale link failure criteria")
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Acked-by: Jon
Signed-off-by: David S. Miller <davem@davemloft.net>
There are two improvements when re-calculate cluster capabilities:
- When deleting a specific down node, need to re-calculate.
- In tipc_node_cleanup(), do not need to re-calculate if node
is still existing in cluster.
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Acked-by: Jon
Signed-off-by: David S. Miller <davem@davemloft.net>
As mentioned in commit e95584a889 ("tipc: fix unlimited bundling of
small messages"), the current message bundling algorithm is inefficient
that can generate bundles of only one payload message, that causes
unnecessary overheads for both the sender and receiver.
This commit re-designs the 'tipc_msg_make_bundle()' function (now named
as 'tipc_msg_try_bundle()'), so that when a message comes at the first
place, we will just check & keep a reference to it if the message is
suitable for bundling. The message buffer will be put into the link
backlog queue and processed as normal. Later on, when another one comes
we will make a bundle with the first message if possible and so on...
This way, a bundle if really needed will always consist of at least two
payload messages. Otherwise, we let the first buffer go its way without
any need of bundling, so reduce the overheads to zero.
Moreover, since now we have both the messages in hand, we can even
optimize the 'tipc_msg_bundle()' function, make bundle of a very large
(size ~ MSS) and small messages which is not with the current algorithm
e.g. [1400-byte message] + [10-byte message] (MTU = 1500).
Acked-by: Ying Xue <ying.xue@windreiver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The only slightly tricky merge conflict was the netdevsim because the
mutex locking fix overlapped a lot of driver reload reorganization.
The rest were (relatively) trivial in nature.
Signed-off-by: David S. Miller <davem@davemloft.net>
We introduce a feature that works like a combination of TCP_NAGLE and
TCP_CORK, but without some of the weaknesses of those. In particular,
we will not observe long delivery delays because of delayed acks, since
the algorithm itself decides if and when acks are to be sent from the
receiving peer.
- The nagle property as such is determined by manipulating a new
'maxnagle' field in struct tipc_sock. If certain conditions are met,
'maxnagle' will define max size of the messages which can be bundled.
If it is set to zero no messages are ever bundled, implying that the
nagle property is disabled.
- A socket with the nagle property enabled enters nagle mode when more
than 4 messages have been sent out without receiving any data message
from the peer.
- A socket leaves nagle mode whenever it receives a data message from
the peer.
In nagle mode, messages smaller than 'maxnagle' are accumulated in the
socket write queue. The last buffer in the queue is marked with a new
'ack_required' bit, which forces the receiving peer to send a CONN_ACK
message back to the sender upon reception.
The accumulated contents of the write queue is transmitted when one of
the following events or conditions occur.
- A CONN_ACK message is received from the peer.
- A data message is received from the peer.
- A SOCK_WAKEUP pseudo message is received from the link level.
- The write queue contains more than 64 1k blocks of data.
- The connection is being shut down.
- There is no CONN_ACK message to expect. I.e., there is currently
no outstanding message where the 'ack_required' bit was set. As a
consequence, the first message added after we enter nagle mode
is always sent directly with this bit set.
This new feature gives a 50-100% improvement of throughput for small
(i.e., less than MTU size) messages, while it might add up to one RTT
to latency time when the socket is in nagle mode.
Acked-by: Ying Xue <ying.xue@windreiver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Currently, TIPC transports intra-node user data messages directly
socket to socket, hence shortcutting all the lower layers of the
communication stack. This gives TIPC very good intra node performance,
both regarding throughput and latency.
We now introduce a similar mechanism for TIPC data traffic across
network namespaces located in the same kernel. On the send path, the
call chain is as always accompanied by the sending node's network name
space pointer. However, once we have reliably established that the
receiving node is represented by a namespace on the same host, we just
replace the namespace pointer with the receiving node/namespace's
ditto, and follow the regular socket receive patch though the receiving
node. This technique gives us a throughput similar to the node internal
throughput, several times larger than if we let the traffic go though
the full network stacks. As a comparison, max throughput for 64k
messages is four times larger than TCP throughput for the same type of
traffic.
To meet any security concerns, the following should be noted.
- All nodes joining a cluster are supposed to have been be certified
and authenticated by mechanisms outside TIPC. This is no different for
nodes/namespaces on the same host; they have to auto discover each
other using the attached interfaces, and establish links which are
supervised via the regular link monitoring mechanism. Hence, a kernel
local node has no other way to join a cluster than any other node, and
have to obey to policies set in the IP or device layers of the stack.
- Only when a sender has established with 100% certainty that the peer
node is located in a kernel local namespace does it choose to let user
data messages, and only those, take the crossover path to the receiving
node/namespace.
- If the receiving node/namespace is removed, its namespace pointer
is invalidated at all peer nodes, and their neighbor link monitoring
will eventually note that this node is gone.
- To ensure the "100% certainty" criteria, and prevent any possible
spoofing, received discovery messages must contain a proof that the
sender knows a common secret. We use the hash mix of the sending
node/namespace for this purpose, since it can be accessed directly by
all other namespaces in the kernel. Upon reception of a discovery
message, the receiver checks this proof against all the local
namespaces'hash_mix:es. If it finds a match, that, along with a
matching node id and cluster id, this is deemed sufficient proof that
the peer node in question is in a local namespace, and a wormhole can
be opened.
- We should also consider that TIPC is intended to be a cluster local
IPC mechanism (just like e.g. UNIX sockets) rather than a network
protocol, and hence we think it can justified to allow it to shortcut the
lower protocol layers.
Regarding traceability, we should notice that since commit 6c9081a391
("tipc: add loopback device tracking") it is possible to follow the node
internal packet flow by just activating tcpdump on the loopback
interface. This will be true even for this mechanism; by activating
tcpdump on the involved nodes' loopback interfaces their inter-name
space messaging can easily be tracked.
v2:
- update 'net' pointer when node left/rejoined
v3:
- grab read/write lock when using node ref obj
v4:
- clone traffics between netns to loopback
Suggested-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Hoang Le <hoang.h.le@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
Many poll() handlers are lockless. Using skb_queue_empty_lockless()
instead of skb_queue_empty() is more appropriate.
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
sk->sk_backlog.len can be written by BH handlers, and read
from process contexts in a lockless way.
Note the write side should also use WRITE_ONCE() or a variant.
We need some agreement about the best way to do this.
syzbot reported :
BUG: KCSAN: data-race in tcp_add_backlog / tcp_grow_window.isra.0
write to 0xffff88812665f32c of 4 bytes by interrupt on cpu 1:
sk_add_backlog include/net/sock.h:934 [inline]
tcp_add_backlog+0x4a0/0xcc0 net/ipv4/tcp_ipv4.c:1737
tcp_v4_rcv+0x1aba/0x1bf0 net/ipv4/tcp_ipv4.c:1925
ip_protocol_deliver_rcu+0x51/0x470 net/ipv4/ip_input.c:204
ip_local_deliver_finish+0x110/0x140 net/ipv4/ip_input.c:231
NF_HOOK include/linux/netfilter.h:305 [inline]
NF_HOOK include/linux/netfilter.h:299 [inline]
ip_local_deliver+0x133/0x210 net/ipv4/ip_input.c:252
dst_input include/net/dst.h:442 [inline]
ip_rcv_finish+0x121/0x160 net/ipv4/ip_input.c:413
NF_HOOK include/linux/netfilter.h:305 [inline]
NF_HOOK include/linux/netfilter.h:299 [inline]
ip_rcv+0x18f/0x1a0 net/ipv4/ip_input.c:523
__netif_receive_skb_one_core+0xa7/0xe0 net/core/dev.c:5004
__netif_receive_skb+0x37/0xf0 net/core/dev.c:5118
netif_receive_skb_internal+0x59/0x190 net/core/dev.c:5208
napi_skb_finish net/core/dev.c:5671 [inline]
napi_gro_receive+0x28f/0x330 net/core/dev.c:5704
receive_buf+0x284/0x30b0 drivers/net/virtio_net.c:1061
virtnet_receive drivers/net/virtio_net.c:1323 [inline]
virtnet_poll+0x436/0x7d0 drivers/net/virtio_net.c:1428
napi_poll net/core/dev.c:6352 [inline]
net_rx_action+0x3ae/0xa50 net/core/dev.c:6418
read to 0xffff88812665f32c of 4 bytes by task 7292 on cpu 0:
tcp_space include/net/tcp.h:1373 [inline]
tcp_grow_window.isra.0+0x6b/0x480 net/ipv4/tcp_input.c:413
tcp_event_data_recv+0x68f/0x990 net/ipv4/tcp_input.c:717
tcp_rcv_established+0xbfe/0xf50 net/ipv4/tcp_input.c:5618
tcp_v4_do_rcv+0x381/0x4e0 net/ipv4/tcp_ipv4.c:1542
sk_backlog_rcv include/net/sock.h:945 [inline]
__release_sock+0x135/0x1e0 net/core/sock.c:2427
release_sock+0x61/0x160 net/core/sock.c:2943
tcp_recvmsg+0x63b/0x1a30 net/ipv4/tcp.c:2181
inet_recvmsg+0xbb/0x250 net/ipv4/af_inet.c:838
sock_recvmsg_nosec net/socket.c:871 [inline]
sock_recvmsg net/socket.c:889 [inline]
sock_recvmsg+0x92/0xb0 net/socket.c:885
sock_read_iter+0x15f/0x1e0 net/socket.c:967
call_read_iter include/linux/fs.h:1864 [inline]
new_sync_read+0x389/0x4f0 fs/read_write.c:414
__vfs_read+0xb1/0xc0 fs/read_write.c:427
vfs_read fs/read_write.c:461 [inline]
vfs_read+0x143/0x2c0 fs/read_write.c:446
Reported by Kernel Concurrency Sanitizer on:
CPU: 0 PID: 7292 Comm: syz-fuzzer Not tainted 5.3.0+ #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
sk_add_backlog() callers usually read sk->sk_rcvbuf without
owning the socket lock. This means sk_rcvbuf value can
be changed by other cpus, and KCSAN complains.
Add READ_ONCE() annotations to document the lockless nature
of these reads.
Note that writes over sk_rcvbuf should also use WRITE_ONCE(),
but this will be done in separate patches to ease stable
backports (if we decide this is relevant for stable trees).
BUG: KCSAN: data-race in tcp_add_backlog / tcp_recvmsg
write to 0xffff88812ab369f8 of 8 bytes by interrupt on cpu 1:
__sk_add_backlog include/net/sock.h:902 [inline]
sk_add_backlog include/net/sock.h:933 [inline]
tcp_add_backlog+0x45a/0xcc0 net/ipv4/tcp_ipv4.c:1737
tcp_v4_rcv+0x1aba/0x1bf0 net/ipv4/tcp_ipv4.c:1925
ip_protocol_deliver_rcu+0x51/0x470 net/ipv4/ip_input.c:204
ip_local_deliver_finish+0x110/0x140 net/ipv4/ip_input.c:231
NF_HOOK include/linux/netfilter.h:305 [inline]
NF_HOOK include/linux/netfilter.h:299 [inline]
ip_local_deliver+0x133/0x210 net/ipv4/ip_input.c:252
dst_input include/net/dst.h:442 [inline]
ip_rcv_finish+0x121/0x160 net/ipv4/ip_input.c:413
NF_HOOK include/linux/netfilter.h:305 [inline]
NF_HOOK include/linux/netfilter.h:299 [inline]
ip_rcv+0x18f/0x1a0 net/ipv4/ip_input.c:523
__netif_receive_skb_one_core+0xa7/0xe0 net/core/dev.c:5004
__netif_receive_skb+0x37/0xf0 net/core/dev.c:5118
netif_receive_skb_internal+0x59/0x190 net/core/dev.c:5208
napi_skb_finish net/core/dev.c:5671 [inline]
napi_gro_receive+0x28f/0x330 net/core/dev.c:5704
receive_buf+0x284/0x30b0 drivers/net/virtio_net.c:1061
virtnet_receive drivers/net/virtio_net.c:1323 [inline]
virtnet_poll+0x436/0x7d0 drivers/net/virtio_net.c:1428
napi_poll net/core/dev.c:6352 [inline]
net_rx_action+0x3ae/0xa50 net/core/dev.c:6418
read to 0xffff88812ab369f8 of 8 bytes by task 7271 on cpu 0:
tcp_recvmsg+0x470/0x1a30 net/ipv4/tcp.c:2047
inet_recvmsg+0xbb/0x250 net/ipv4/af_inet.c:838
sock_recvmsg_nosec net/socket.c:871 [inline]
sock_recvmsg net/socket.c:889 [inline]
sock_recvmsg+0x92/0xb0 net/socket.c:885
sock_read_iter+0x15f/0x1e0 net/socket.c:967
call_read_iter include/linux/fs.h:1864 [inline]
new_sync_read+0x389/0x4f0 fs/read_write.c:414
__vfs_read+0xb1/0xc0 fs/read_write.c:427
vfs_read fs/read_write.c:461 [inline]
vfs_read+0x143/0x2c0 fs/read_write.c:446
ksys_read+0xd5/0x1b0 fs/read_write.c:587
__do_sys_read fs/read_write.c:597 [inline]
__se_sys_read fs/read_write.c:595 [inline]
__x64_sys_read+0x4c/0x60 fs/read_write.c:595
do_syscall_64+0xcf/0x2f0 arch/x86/entry/common.c:296
entry_SYSCALL_64_after_hwframe+0x44/0xa9
Reported by Kernel Concurrency Sanitizer on:
CPU: 0 PID: 7271 Comm: syz-fuzzer Not tainted 5.3.0+ #0
Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: syzbot <syzkaller@googlegroups.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
__tipc_nl_compat_dumpit() calls tipc_nl_publ_dump() which expects
the attrs to be available by genl_dumpit_info(cb)->attrs. Add info
struct and attr parsing in compat dumpit function.
Reported-by: syzbot+8d37c50ffb0f52941a5e@syzkaller.appspotmail.com
Fixes: 057af70713 ("net: tipc: have genetlink code to parse the attrs during dumpit")
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Jakub Kicinski <jakub.kicinski@netronome.com>
As this is the last user of genl_family_attrbuf, convert to allocate
attrs locally and do it in a similar way this is done in compat_doit().
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Benefit from the fact that the generic netlink code can parse the attrs
for dumpit op and avoid need to parse it in the op callback.
Signed-off-by: Jiri Pirko <jiri@mellanox.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
We have identified a problem with the "oversubscription" policy in the
link transmission code.
When small messages are transmitted, and the sending link has reached
the transmit window limit, those messages will be bundled and put into
the link backlog queue. However, bundles of data messages are counted
at the 'CRITICAL' level, so that the counter for that level, instead of
the counter for the real, bundled message's level is the one being
increased.
Subsequent, to-be-bundled data messages at non-CRITICAL levels continue
to be tested against the unchanged counter for their own level, while
contributing to an unrestrained increase at the CRITICAL backlog level.
This leaves a gap in congestion control algorithm for small messages
that can result in starvation for other users or a "real" CRITICAL
user. Even that eventually can lead to buffer exhaustion & link reset.
We fix this by keeping a 'target_bskb' buffer pointer at each levels,
then when bundling, we only bundle messages at the same importance
level only. This way, we know exactly how many slots a certain level
have occupied in the queue, so can manage level congestion accurately.
By bundling messages at the same level, we even have more benefits. Let
consider this:
- One socket sends 64-byte messages at the 'CRITICAL' level;
- Another sends 4096-byte messages at the 'LOW' level;
When a 64-byte message comes and is bundled the first time, we put the
overhead of message bundle to it (+ 40-byte header, data copy, etc.)
for later use, but the next message can be a 4096-byte one that cannot
be bundled to the previous one. This means the last bundle carries only
one payload message which is totally inefficient, as for the receiver
also! Later on, another 64-byte message comes, now we make a new bundle
and the same story repeats...
With the new bundling algorithm, this will not happen, the 64-byte
messages will be bundled together even when the 4096-byte message(s)
comes in between. However, if the 4096-byte messages are sent at the
same level i.e. 'CRITICAL', the bundling algorithm will again cause the
same overhead.
Also, the same will happen even with only one socket sending small
messages at a rate close to the link transmit's one, so that, when one
message is bundled, it's transmitted shortly. Then, another message
comes, a new bundle is created and so on...
We will solve this issue radically by another patch.
Fixes: 365ad353c2 ("tipc: reduce risk of user starvation during link congestion")
Reported-by: Hoang Le <hoang.h.le@dektech.com.au>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
The policy for handling the skb list locks on the send and receive paths
is simple.
- On the send path we never need to grab the lock on the 'xmitq' list
when the destination is an exernal node.
- On the receive path we always need to grab the lock on the 'inputq'
list, irrespective of source node.
However, when transmitting node local messages those will eventually
end up on the receive path of a local socket, meaning that the argument
'xmitq' in tipc_node_xmit() will become the 'ínputq' argument in the
function tipc_sk_rcv(). This has been handled by always initializing
the spinlock of the 'xmitq' list at message creation, just in case it
may end up on the receive path later, and despite knowing that the lock
in most cases never will be used.
This approach is inaccurate and confusing, and has also concealed the
fact that the stated 'no lock grabbing' policy for the send path is
violated in some cases.
We now clean up this by never initializing the lock at message creation,
instead doing this at the moment we find that the message actually will
enter the receive path. At the same time we fix the four locations
where we incorrectly access the spinlock on the send/error path.
This patch also reverts commit d12cffe932 ("tipc: ensure head->lock
is initialised") which has now become redundant.
CC: Eric Dumazet <edumazet@google.com>
Reported-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Reviewed-by: Xin Long <lucien.xin@gmail.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit eliminates the use of the link 'stale_limit' & 'prev_from'
(besides the already removed - 'stale_cnt') variables in the detection
of repeated retransmit failures as there is no proper way to initialize
them to avoid a false detection, i.e. it is not really a retransmission
failure but due to a garbage values in the variables.
Instead, a jiffies variable will be added to individual skbs (like the
way we restrict the skb retransmissions) in order to mark the first skb
retransmit time. Later on, at the next retransmissions, the timestamp
will be checked to see if the skb in the link transmq is "too stale",
that is, the link tolerance time has passed, so that a link reset will
be ordered. Note, just checking on the first skb in the queue is fine
enough since it must be the oldest one.
A counter is also added to keep track the actual skb retransmissions'
number for later checking when the failure happens.
The downside of this approach is that the skb->cb[] buffer is about to
be exhausted, however it is always able to allocate another memory area
and keep a reference to it when needed.
Fixes: 77cf8edbc0 ("tipc: simplify stale link failure criteria")
Reported-by: Hoang Le <hoang.h.le@dektech.com.au>
Acked-by: Ying Xue <ying.xue@windriver.com>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au>
Signed-off-by: David S. Miller <davem@davemloft.net>
We set the field 'addr_trial_end' to 'jiffies', instead of the current
value 0, at the moment the node address is initialized. This guarantees
we don't inadvertently enter an address trial period when the node
address is explicitly set by the user.
Signed-off-by: Chris Packham <chris.packham@alliedtelesis.co.nz>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Since node internal messages are passed directly to the socket, it is not
possible to observe those messages via tcpdump or wireshark.
We now remedy this by making it possible to clone such messages and send
the clones to the loopback interface. The clones are dropped at reception
and have no functional role except making the traffic visible.
The feature is enabled if network taps are active for the loopback device.
pcap filtering restrictions require the messages to be presented to the
receiving side of the loopback device.
v3 - Function dev_nit_active used to check for network taps.
- Procedure netif_rx_ni used to send cloned messages to loopback device.
Signed-off-by: John Rutherford <john.rutherford@dektech.com.au>
Acked-by: Jon Maloy <jon.maloy@ericsson.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
In commit 365ad353c2 ("tipc: reduce risk of user starvation during
link congestion") we allowed senders to add exactly one list of extra
buffers to the link backlog queues during link congestion (aka
"oversubscription"). However, the criteria for when to stop adding
wakeup messages to the input queue when the overload abates is
inaccurate, and may cause starvation problems during very high load.
Currently, we stop adding wakeup messages after 10 total failed attempts
where we find that there is no space left in the backlog queue for a
certain importance level. The counter for this is accumulated across all
levels, which may lead the algorithm to leave the loop prematurely,
although there may still be plenty of space available at some levels.
The result is sometimes that messages near the wakeup queue tail are not
added to the input queue as they should be.
We now introduce a more exact algorithm, where we keep adding wakeup
messages to a level as long as the backlog queue has free slots for
the corresponding level, and stop at the moment there are no more such
slots or when there are no more wakeup messages to dequeue.
Fixes: 365ad35 ("tipc: reduce risk of user starvation during link congestion")
Reported-by: Tung Nguyen <tung.q.nguyen@dektech.com.au>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: Jon Maloy <jon.maloy@ericsson.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Commit 2753ca5d90 ("tipc: fix uninit-value in tipc_nl_compat_doit")
broke older tipc tools that use compat interface (e.g. tipc-config from
tipcutils package):
% tipc-config -p
operation not supported
The commit started to reject TIPC netlink compat messages that do not
have attributes. It is too restrictive because some of such messages are
valid (they don't need any arguments):
% grep 'tx none' include/uapi/linux/tipc_config.h
#define TIPC_CMD_NOOP 0x0000 /* tx none, rx none */
#define TIPC_CMD_GET_MEDIA_NAMES 0x0002 /* tx none, rx media_name(s) */
#define TIPC_CMD_GET_BEARER_NAMES 0x0003 /* tx none, rx bearer_name(s) */
#define TIPC_CMD_SHOW_PORTS 0x0006 /* tx none, rx ultra_string */
#define TIPC_CMD_GET_REMOTE_MNG 0x4003 /* tx none, rx unsigned */
#define TIPC_CMD_GET_MAX_PORTS 0x4004 /* tx none, rx unsigned */
#define TIPC_CMD_GET_NETID 0x400B /* tx none, rx unsigned */
#define TIPC_CMD_NOT_NET_ADMIN 0xC001 /* tx none, rx none */
This patch relaxes the original fix and rejects messages without
arguments only if such arguments are expected by a command (reg_type is
non zero).
Fixes: 2753ca5d90 ("tipc: fix uninit-value in tipc_nl_compat_doit")
Cc: stable@vger.kernel.org
Signed-off-by: Taras Kondratiuk <takondra@cisco.com>
Acked-by: Ying Xue <ying.xue@windriver.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
