linux_dsm_epyc7002/net/tipc/node.c

2743 lines
68 KiB
C
Raw Normal View History

/*
* net/tipc/node.c: TIPC node management routines
*
* Copyright (c) 2000-2006, 2012-2016, Ericsson AB
* Copyright (c) 2005-2006, 2010-2014, Wind River Systems
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions are met:
*
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. Neither the names of the copyright holders nor the names of its
* contributors may be used to endorse or promote products derived from
* this software without specific prior written permission.
*
* Alternatively, this software may be distributed under the terms of the
* GNU General Public License ("GPL") version 2 as published by the Free
* Software Foundation.
*
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS"
* AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE
* LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR
* CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF
* SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS
* INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN
* CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)
* ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE
* POSSIBILITY OF SUCH DAMAGE.
*/
#include "core.h"
#include "link.h"
#include "node.h"
#include "name_distr.h"
#include "socket.h"
#include "bcast.h"
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
#include "monitor.h"
#include "discover.h"
#include "netlink.h"
tipc: enable tracepoints in tipc As for the sake of debugging/tracing, the commit enables tracepoints in TIPC along with some general trace_events as shown below. It also defines some 'tipc_*_dump()' functions that allow to dump TIPC object data whenever needed, that is, for general debug purposes, ie. not just for the trace_events. The following trace_events are now available: - trace_tipc_skb_dump(): allows to trace and dump TIPC msg & skb data, e.g. message type, user, droppable, skb truesize, cloned skb, etc. - trace_tipc_list_dump(): allows to trace and dump any TIPC buffers or queues, e.g. TIPC link transmq, socket receive queue, etc. - trace_tipc_sk_dump(): allows to trace and dump TIPC socket data, e.g. sk state, sk type, connection type, rmem_alloc, socket queues, etc. - trace_tipc_link_dump(): allows to trace and dump TIPC link data, e.g. link state, silent_intv_cnt, gap, bc_gap, link queues, etc. - trace_tipc_node_dump(): allows to trace and dump TIPC node data, e.g. node state, active links, capabilities, link entries, etc. How to use: Put the trace functions at any places where we want to dump TIPC data or events. Note: a) The dump functions will generate raw data only, that is, to offload the trace event's processing, it can require a tool or script to parse the data but this should be simple. b) The trace_tipc_*_dump() should be reserved for a failure cases only (e.g. the retransmission failure case) or where we do not expect to happen too often, then we can consider enabling these events by default since they will almost not take any effects under normal conditions, but once the rare condition or failure occurs, we get the dumped data fully for post-analysis. For other trace purposes, we can reuse these trace classes as template but different events. c) A trace_event is only effective when we enable it. To enable the TIPC trace_events, echo 1 to 'enable' files in the events/tipc/ directory in the 'debugfs' file system. Normally, they are located at: /sys/kernel/debug/tracing/events/tipc/ For example: To enable the tipc_link_dump event: echo 1 > /sys/kernel/debug/tracing/events/tipc/tipc_link_dump/enable To enable all the TIPC trace_events: echo 1 > /sys/kernel/debug/tracing/events/tipc/enable To collect the trace data: cat trace or cat trace_pipe > /trace.out & To disable all the TIPC trace_events: echo 0 > /sys/kernel/debug/tracing/events/tipc/enable To clear the trace buffer: echo > trace d) Like the other trace_events, the feature like 'filter' or 'trigger' is also usable for the tipc trace_events. For more details, have a look at: Documentation/trace/ftrace.txt MAINTAINERS | add two new files 'trace.h' & 'trace.c' in tipc Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:56 +07:00
#include "trace.h"
#define INVALID_NODE_SIG 0x10000
#define NODE_CLEANUP_AFTER 300000
/* Flags used to take different actions according to flag type
* TIPC_NOTIFY_NODE_DOWN: notify node is down
* TIPC_NOTIFY_NODE_UP: notify node is up
* TIPC_DISTRIBUTE_NAME: publish or withdraw link state name type
*/
enum {
TIPC_NOTIFY_NODE_DOWN = (1 << 3),
TIPC_NOTIFY_NODE_UP = (1 << 4),
TIPC_NOTIFY_LINK_UP = (1 << 6),
TIPC_NOTIFY_LINK_DOWN = (1 << 7)
};
struct tipc_link_entry {
struct tipc_link *link;
spinlock_t lock; /* per link */
u32 mtu;
struct sk_buff_head inputq;
struct tipc_media_addr maddr;
};
struct tipc_bclink_entry {
struct tipc_link *link;
struct sk_buff_head inputq1;
struct sk_buff_head arrvq;
struct sk_buff_head inputq2;
struct sk_buff_head namedq;
};
/**
* struct tipc_node - TIPC node structure
* @addr: network address of node
* @ref: reference counter to node object
* @lock: rwlock governing access to structure
* @net: the applicable net namespace
* @hash: links to adjacent nodes in unsorted hash chain
* @inputq: pointer to input queue containing messages for msg event
* @namedq: pointer to name table input queue with name table messages
* @active_links: bearer ids of active links, used as index into links[] array
* @links: array containing references to all links to node
* @action_flags: bit mask of different types of node actions
* @state: connectivity state vs peer node
* @sync_point: sequence number where synch/failover is finished
* @list: links to adjacent nodes in sorted list of cluster's nodes
* @working_links: number of working links to node (both active and standby)
* @link_cnt: number of links to node
* @capabilities: bitmap, indicating peer node's functional capabilities
* @signature: node instance identifier
* @link_id: local and remote bearer ids of changing link, if any
* @publ_list: list of publications
* @rcu: rcu struct for tipc_node
* @delete_at: indicates the time for deleting a down node
*/
struct tipc_node {
u32 addr;
struct kref kref;
rwlock_t lock;
struct net *net;
struct hlist_node hash;
int active_links[2];
struct tipc_link_entry links[MAX_BEARERS];
struct tipc_bclink_entry bc_entry;
int action_flags;
struct list_head list;
int state;
bool failover_sent;
u16 sync_point;
int link_cnt;
u16 working_links;
u16 capabilities;
u32 signature;
u32 link_id;
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
u8 peer_id[16];
struct list_head publ_list;
struct list_head conn_sks;
unsigned long keepalive_intv;
struct timer_list timer;
struct rcu_head rcu;
unsigned long delete_at;
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
struct net *peer_net;
u32 peer_hash_mix;
};
/* Node FSM states and events:
*/
enum {
SELF_DOWN_PEER_DOWN = 0xdd,
SELF_UP_PEER_UP = 0xaa,
SELF_DOWN_PEER_LEAVING = 0xd1,
SELF_UP_PEER_COMING = 0xac,
SELF_COMING_PEER_UP = 0xca,
SELF_LEAVING_PEER_DOWN = 0x1d,
NODE_FAILINGOVER = 0xf0,
NODE_SYNCHING = 0xcc
};
enum {
SELF_ESTABL_CONTACT_EVT = 0xece,
SELF_LOST_CONTACT_EVT = 0x1ce,
PEER_ESTABL_CONTACT_EVT = 0x9ece,
PEER_LOST_CONTACT_EVT = 0x91ce,
NODE_FAILOVER_BEGIN_EVT = 0xfbe,
NODE_FAILOVER_END_EVT = 0xfee,
NODE_SYNCH_BEGIN_EVT = 0xcbe,
NODE_SYNCH_END_EVT = 0xcee
};
static void __tipc_node_link_down(struct tipc_node *n, int *bearer_id,
struct sk_buff_head *xmitq,
struct tipc_media_addr **maddr);
static void tipc_node_link_down(struct tipc_node *n, int bearer_id,
bool delete);
static void node_lost_contact(struct tipc_node *n, struct sk_buff_head *inputq);
static void tipc_node_delete(struct tipc_node *node);
static void tipc_node_timeout(struct timer_list *t);
static void tipc_node_fsm_evt(struct tipc_node *n, int evt);
static struct tipc_node *tipc_node_find(struct net *net, u32 addr);
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
static struct tipc_node *tipc_node_find_by_id(struct net *net, u8 *id);
static void tipc_node_put(struct tipc_node *node);
static bool node_is_up(struct tipc_node *n);
static void tipc_node_delete_from_list(struct tipc_node *node);
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
struct tipc_sock_conn {
u32 port;
u32 peer_port;
u32 peer_node;
struct list_head list;
};
static struct tipc_link *node_active_link(struct tipc_node *n, int sel)
{
int bearer_id = n->active_links[sel & 1];
if (unlikely(bearer_id == INVALID_BEARER_ID))
return NULL;
return n->links[bearer_id].link;
}
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
int tipc_node_get_mtu(struct net *net, u32 addr, u32 sel, bool connected)
{
struct tipc_node *n;
int bearer_id;
unsigned int mtu = MAX_MSG_SIZE;
n = tipc_node_find(net, addr);
if (unlikely(!n))
return mtu;
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
/* Allow MAX_MSG_SIZE when building connection oriented message
* if they are in the same core network
*/
if (n->peer_net && connected) {
tipc_node_put(n);
return mtu;
}
bearer_id = n->active_links[sel & 1];
if (likely(bearer_id != INVALID_BEARER_ID))
mtu = n->links[bearer_id].mtu;
tipc_node_put(n);
return mtu;
}
bool tipc_node_get_id(struct net *net, u32 addr, u8 *id)
{
u8 *own_id = tipc_own_id(net);
struct tipc_node *n;
if (!own_id)
return true;
if (addr == tipc_own_addr(net)) {
memcpy(id, own_id, TIPC_NODEID_LEN);
return true;
}
n = tipc_node_find(net, addr);
if (!n)
return false;
memcpy(id, &n->peer_id, TIPC_NODEID_LEN);
tipc_node_put(n);
return true;
}
u16 tipc_node_get_capabilities(struct net *net, u32 addr)
{
struct tipc_node *n;
u16 caps;
n = tipc_node_find(net, addr);
if (unlikely(!n))
return TIPC_NODE_CAPABILITIES;
caps = n->capabilities;
tipc_node_put(n);
return caps;
}
static void tipc_node_kref_release(struct kref *kref)
{
struct tipc_node *n = container_of(kref, struct tipc_node, kref);
kfree(n->bc_entry.link);
kfree_rcu(n, rcu);
}
static void tipc_node_put(struct tipc_node *node)
{
kref_put(&node->kref, tipc_node_kref_release);
}
static void tipc_node_get(struct tipc_node *node)
{
kref_get(&node->kref);
}
/*
* tipc_node_find - locate specified node object, if it exists
*/
static struct tipc_node *tipc_node_find(struct net *net, u32 addr)
{
struct tipc_net *tn = tipc_net(net);
struct tipc_node *node;
unsigned int thash = tipc_hashfn(addr);
rcu_read_lock();
hlist_for_each_entry_rcu(node, &tn->node_htable[thash], hash) {
if (node->addr != addr)
continue;
if (!kref_get_unless_zero(&node->kref))
node = NULL;
break;
}
rcu_read_unlock();
return node;
}
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
/* tipc_node_find_by_id - locate specified node object by its 128-bit id
* Note: this function is called only when a discovery request failed
* to find the node by its 32-bit id, and is not time critical
*/
static struct tipc_node *tipc_node_find_by_id(struct net *net, u8 *id)
{
struct tipc_net *tn = tipc_net(net);
struct tipc_node *n;
bool found = false;
rcu_read_lock();
list_for_each_entry_rcu(n, &tn->node_list, list) {
read_lock_bh(&n->lock);
if (!memcmp(id, n->peer_id, 16) &&
kref_get_unless_zero(&n->kref))
found = true;
read_unlock_bh(&n->lock);
if (found)
break;
}
rcu_read_unlock();
return found ? n : NULL;
}
static void tipc_node_read_lock(struct tipc_node *n)
{
read_lock_bh(&n->lock);
}
static void tipc_node_read_unlock(struct tipc_node *n)
{
read_unlock_bh(&n->lock);
}
static void tipc_node_write_lock(struct tipc_node *n)
{
write_lock_bh(&n->lock);
}
2017-01-24 19:00:43 +07:00
static void tipc_node_write_unlock_fast(struct tipc_node *n)
{
write_unlock_bh(&n->lock);
}
static void tipc_node_write_unlock(struct tipc_node *n)
{
struct net *net = n->net;
u32 addr = 0;
u32 flags = n->action_flags;
u32 link_id = 0;
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
u32 bearer_id;
struct list_head *publ_list;
if (likely(!flags)) {
write_unlock_bh(&n->lock);
return;
}
addr = n->addr;
link_id = n->link_id;
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
bearer_id = link_id & 0xffff;
publ_list = &n->publ_list;
n->action_flags &= ~(TIPC_NOTIFY_NODE_DOWN | TIPC_NOTIFY_NODE_UP |
TIPC_NOTIFY_LINK_DOWN | TIPC_NOTIFY_LINK_UP);
write_unlock_bh(&n->lock);
if (flags & TIPC_NOTIFY_NODE_DOWN)
tipc_publ_notify(net, publ_list, addr);
if (flags & TIPC_NOTIFY_NODE_UP)
tipc_named_node_up(net, addr);
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
if (flags & TIPC_NOTIFY_LINK_UP) {
tipc_mon_peer_up(net, addr, bearer_id);
tipc_nametbl_publish(net, TIPC_LINK_STATE, addr, addr,
TIPC_NODE_SCOPE, link_id, link_id);
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
}
if (flags & TIPC_NOTIFY_LINK_DOWN) {
tipc_mon_peer_down(net, addr, bearer_id);
tipc_nametbl_withdraw(net, TIPC_LINK_STATE, addr,
addr, link_id);
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
}
}
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
static void tipc_node_assign_peer_net(struct tipc_node *n, u32 hash_mixes)
{
int net_id = tipc_netid(n->net);
struct tipc_net *tn_peer;
struct net *tmp;
u32 hash_chk;
if (n->peer_net)
return;
for_each_net_rcu(tmp) {
tn_peer = tipc_net(tmp);
if (!tn_peer)
continue;
/* Integrity checking whether node exists in namespace or not */
if (tn_peer->net_id != net_id)
continue;
if (memcmp(n->peer_id, tn_peer->node_id, NODE_ID_LEN))
continue;
hash_chk = tipc_net_hash_mixes(tmp, tn_peer->random);
if (hash_mixes ^ hash_chk)
continue;
n->peer_net = tmp;
n->peer_hash_mix = hash_mixes;
break;
}
}
static struct tipc_node *tipc_node_create(struct net *net, u32 addr,
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
u8 *peer_id, u16 capabilities,
u32 signature, u32 hash_mixes)
{
struct tipc_net *tn = net_generic(net, tipc_net_id);
struct tipc_node *n, *temp_node;
struct tipc_link *l;
int bearer_id;
int i;
spin_lock_bh(&tn->node_list_lock);
n = tipc_node_find(net, addr);
if (n) {
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
if (n->peer_hash_mix ^ hash_mixes)
tipc_node_assign_peer_net(n, hash_mixes);
if (n->capabilities == capabilities)
goto exit;
/* Same node may come back with new capabilities */
tipc_node_write_lock(n);
n->capabilities = capabilities;
for (bearer_id = 0; bearer_id < MAX_BEARERS; bearer_id++) {
l = n->links[bearer_id].link;
if (l)
tipc_link_update_caps(l, capabilities);
}
tipc_node_write_unlock_fast(n);
/* Calculate cluster capabilities */
tn->capabilities = TIPC_NODE_CAPABILITIES;
list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
tn->capabilities &= temp_node->capabilities;
}
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
goto exit;
}
n = kzalloc(sizeof(*n), GFP_ATOMIC);
if (!n) {
pr_warn("Node creation failed, no memory\n");
goto exit;
}
n->addr = addr;
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
memcpy(&n->peer_id, peer_id, 16);
n->net = net;
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
n->peer_net = NULL;
n->peer_hash_mix = 0;
/* Assign kernel local namespace if exists */
tipc_node_assign_peer_net(n, hash_mixes);
n->capabilities = capabilities;
kref_init(&n->kref);
rwlock_init(&n->lock);
INIT_HLIST_NODE(&n->hash);
INIT_LIST_HEAD(&n->list);
INIT_LIST_HEAD(&n->publ_list);
INIT_LIST_HEAD(&n->conn_sks);
skb_queue_head_init(&n->bc_entry.namedq);
skb_queue_head_init(&n->bc_entry.inputq1);
__skb_queue_head_init(&n->bc_entry.arrvq);
skb_queue_head_init(&n->bc_entry.inputq2);
for (i = 0; i < MAX_BEARERS; i++)
spin_lock_init(&n->links[i].lock);
n->state = SELF_DOWN_PEER_LEAVING;
n->delete_at = jiffies + msecs_to_jiffies(NODE_CLEANUP_AFTER);
n->signature = INVALID_NODE_SIG;
n->active_links[0] = INVALID_BEARER_ID;
n->active_links[1] = INVALID_BEARER_ID;
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
if (!tipc_link_bc_create(net, tipc_own_addr(net),
addr, U16_MAX,
tipc_link_window(tipc_bc_sndlink(net)),
n->capabilities,
&n->bc_entry.inputq1,
&n->bc_entry.namedq,
tipc_bc_sndlink(net),
&n->bc_entry.link)) {
pr_warn("Broadcast rcv link creation failed, no memory\n");
kfree(n);
n = NULL;
goto exit;
}
tipc_node_get(n);
timer_setup(&n->timer, tipc_node_timeout, 0);
n->keepalive_intv = U32_MAX;
hlist_add_head_rcu(&n->hash, &tn->node_htable[tipc_hashfn(addr)]);
list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
if (n->addr < temp_node->addr)
break;
}
list_add_tail_rcu(&n->list, &temp_node->list);
/* Calculate cluster capabilities */
tn->capabilities = TIPC_NODE_CAPABILITIES;
list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
tn->capabilities &= temp_node->capabilities;
}
trace_tipc_node_create(n, true, " ");
exit:
spin_unlock_bh(&tn->node_list_lock);
return n;
}
static void tipc_node_calculate_timer(struct tipc_node *n, struct tipc_link *l)
{
unsigned long tol = tipc_link_tolerance(l);
unsigned long intv = ((tol / 4) > 500) ? 500 : tol / 4;
/* Link with lowest tolerance determines timer interval */
if (intv < n->keepalive_intv)
n->keepalive_intv = intv;
/* Ensure link's abort limit corresponds to current tolerance */
tipc_link_set_abort_limit(l, tol / n->keepalive_intv);
}
static void tipc_node_delete_from_list(struct tipc_node *node)
{
list_del_rcu(&node->list);
hlist_del_rcu(&node->hash);
tipc_node_put(node);
}
static void tipc_node_delete(struct tipc_node *node)
{
trace_tipc_node_delete(node, true, " ");
tipc_node_delete_from_list(node);
del_timer_sync(&node->timer);
tipc_node_put(node);
}
void tipc_node_stop(struct net *net)
{
struct tipc_net *tn = tipc_net(net);
struct tipc_node *node, *t_node;
spin_lock_bh(&tn->node_list_lock);
list_for_each_entry_safe(node, t_node, &tn->node_list, list)
tipc_node_delete(node);
spin_unlock_bh(&tn->node_list_lock);
}
void tipc_node_subscribe(struct net *net, struct list_head *subscr, u32 addr)
{
struct tipc_node *n;
if (in_own_node(net, addr))
return;
n = tipc_node_find(net, addr);
if (!n) {
pr_warn("Node subscribe rejected, unknown node 0x%x\n", addr);
return;
}
tipc_node_write_lock(n);
list_add_tail(subscr, &n->publ_list);
2017-01-24 19:00:43 +07:00
tipc_node_write_unlock_fast(n);
tipc_node_put(n);
}
void tipc_node_unsubscribe(struct net *net, struct list_head *subscr, u32 addr)
{
struct tipc_node *n;
if (in_own_node(net, addr))
return;
n = tipc_node_find(net, addr);
if (!n) {
pr_warn("Node unsubscribe rejected, unknown node 0x%x\n", addr);
return;
}
tipc_node_write_lock(n);
list_del_init(subscr);
2017-01-24 19:00:43 +07:00
tipc_node_write_unlock_fast(n);
tipc_node_put(n);
}
int tipc_node_add_conn(struct net *net, u32 dnode, u32 port, u32 peer_port)
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
{
struct tipc_node *node;
struct tipc_sock_conn *conn;
int err = 0;
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
if (in_own_node(net, dnode))
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
return 0;
node = tipc_node_find(net, dnode);
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
if (!node) {
pr_warn("Connecting sock to node 0x%x failed\n", dnode);
return -EHOSTUNREACH;
}
conn = kmalloc(sizeof(*conn), GFP_ATOMIC);
if (!conn) {
err = -EHOSTUNREACH;
goto exit;
}
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
conn->peer_node = dnode;
conn->port = port;
conn->peer_port = peer_port;
tipc_node_write_lock(node);
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
list_add_tail(&conn->list, &node->conn_sks);
tipc_node_write_unlock(node);
exit:
tipc_node_put(node);
return err;
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
}
void tipc_node_remove_conn(struct net *net, u32 dnode, u32 port)
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
{
struct tipc_node *node;
struct tipc_sock_conn *conn, *safe;
if (in_own_node(net, dnode))
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
return;
node = tipc_node_find(net, dnode);
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
if (!node)
return;
tipc_node_write_lock(node);
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
list_for_each_entry_safe(conn, safe, &node->conn_sks, list) {
if (port != conn->port)
continue;
list_del(&conn->list);
kfree(conn);
}
tipc_node_write_unlock(node);
tipc_node_put(node);
tipc: use message to abort connections when losing contact to node In the current implementation, each 'struct tipc_node' instance keeps a linked list of those ports/sockets that are connected to the node represented by that struct. The purpose of this is to let the node object know which sockets to alert when it loses contact with its peer node, i.e., which sockets need to have their connections aborted. This entails an unwanted direct reference from the node structure back to the port/socket structure, and a need to grab port_lock when we have to make an upcall to the port. We want to get rid of this unecessary BH entry point into the socket, and also eliminate its use of port_lock. In this commit, we instead let the node struct keep list of "connected socket" structs, which each represents a connected socket, but is allocated independently by the node at the moment of connection. If the node loses contact with its peer node, the list is traversed, and a "connection abort" message is created for each entry in the list. The message is sent to it respective connected socket using the ordinary data path, and the receiving socket aborts its connections upon reception of the message. This enables us to get rid of the direct reference from 'struct node' to ´struct port', and another unwanted BH access point to the latter. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Reviewed-by: Erik Hugne <erik.hugne@ericsson.com> Reviewed-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2014-08-23 05:09:08 +07:00
}
static void tipc_node_clear_links(struct tipc_node *node)
{
int i;
for (i = 0; i < MAX_BEARERS; i++) {
struct tipc_link_entry *le = &node->links[i];
if (le->link) {
kfree(le->link);
le->link = NULL;
node->link_cnt--;
}
}
}
/* tipc_node_cleanup - delete nodes that does not
* have active links for NODE_CLEANUP_AFTER time
*/
tipc: fix lockdep warning during node delete We see the following lockdep warning: [ 2284.078521] ====================================================== [ 2284.078604] WARNING: possible circular locking dependency detected [ 2284.078604] 4.19.0+ #42 Tainted: G E [ 2284.078604] ------------------------------------------------------ [ 2284.078604] rmmod/254 is trying to acquire lock: [ 2284.078604] 00000000acd94e28 ((&n->timer)#2){+.-.}, at: del_timer_sync+0x5/0xa0 [ 2284.078604] [ 2284.078604] but task is already holding lock: [ 2284.078604] 00000000f997afc0 (&(&tn->node_list_lock)->rlock){+.-.}, at: tipc_node_stop+0xac/0x190 [tipc] [ 2284.078604] [ 2284.078604] which lock already depends on the new lock. [ 2284.078604] [ 2284.078604] [ 2284.078604] the existing dependency chain (in reverse order) is: [ 2284.078604] [ 2284.078604] -> #1 (&(&tn->node_list_lock)->rlock){+.-.}: [ 2284.078604] tipc_node_timeout+0x20a/0x330 [tipc] [ 2284.078604] call_timer_fn+0xa1/0x280 [ 2284.078604] run_timer_softirq+0x1f2/0x4d0 [ 2284.078604] __do_softirq+0xfc/0x413 [ 2284.078604] irq_exit+0xb5/0xc0 [ 2284.078604] smp_apic_timer_interrupt+0xac/0x210 [ 2284.078604] apic_timer_interrupt+0xf/0x20 [ 2284.078604] default_idle+0x1c/0x140 [ 2284.078604] do_idle+0x1bc/0x280 [ 2284.078604] cpu_startup_entry+0x19/0x20 [ 2284.078604] start_secondary+0x187/0x1c0 [ 2284.078604] secondary_startup_64+0xa4/0xb0 [ 2284.078604] [ 2284.078604] -> #0 ((&n->timer)#2){+.-.}: [ 2284.078604] del_timer_sync+0x34/0xa0 [ 2284.078604] tipc_node_delete+0x1a/0x40 [tipc] [ 2284.078604] tipc_node_stop+0xcb/0x190 [tipc] [ 2284.078604] tipc_net_stop+0x154/0x170 [tipc] [ 2284.078604] tipc_exit_net+0x16/0x30 [tipc] [ 2284.078604] ops_exit_list.isra.8+0x36/0x70 [ 2284.078604] unregister_pernet_operations+0x87/0xd0 [ 2284.078604] unregister_pernet_subsys+0x1d/0x30 [ 2284.078604] tipc_exit+0x11/0x6f2 [tipc] [ 2284.078604] __x64_sys_delete_module+0x1df/0x240 [ 2284.078604] do_syscall_64+0x66/0x460 [ 2284.078604] entry_SYSCALL_64_after_hwframe+0x49/0xbe [ 2284.078604] [ 2284.078604] other info that might help us debug this: [ 2284.078604] [ 2284.078604] Possible unsafe locking scenario: [ 2284.078604] [ 2284.078604] CPU0 CPU1 [ 2284.078604] ---- ---- [ 2284.078604] lock(&(&tn->node_list_lock)->rlock); [ 2284.078604] lock((&n->timer)#2); [ 2284.078604] lock(&(&tn->node_list_lock)->rlock); [ 2284.078604] lock((&n->timer)#2); [ 2284.078604] [ 2284.078604] *** DEADLOCK *** [ 2284.078604] [ 2284.078604] 3 locks held by rmmod/254: [ 2284.078604] #0: 000000003368be9b (pernet_ops_rwsem){+.+.}, at: unregister_pernet_subsys+0x15/0x30 [ 2284.078604] #1: 0000000046ed9c86 (rtnl_mutex){+.+.}, at: tipc_net_stop+0x144/0x170 [tipc] [ 2284.078604] #2: 00000000f997afc0 (&(&tn->node_list_lock)->rlock){+.-.}, at: tipc_node_stop+0xac/0x19 [...} The reason is that the node timer handler sometimes needs to delete a node which has been disconnected for too long. To do this, it grabs the lock 'node_list_lock', which may at the same time be held by the generic node cleanup function, tipc_node_stop(), during module removal. Since the latter is calling del_timer_sync() inside the same lock, we have a potential deadlock. We fix this letting the timer cleanup function use spin_trylock() instead of just spin_lock(), and when it fails to grab the lock it just returns so that the timer handler can terminate its execution. This is safe to do, since tipc_node_stop() anyway is about to delete both the timer and the node instance. Fixes: 6a939f365bdb ("tipc: Auto removal of peer down node instance") 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>
2018-11-27 00:26:14 +07:00
static bool tipc_node_cleanup(struct tipc_node *peer)
{
struct tipc_node *temp_node;
struct tipc_net *tn = tipc_net(peer->net);
bool deleted = false;
tipc: fix lockdep warning during node delete We see the following lockdep warning: [ 2284.078521] ====================================================== [ 2284.078604] WARNING: possible circular locking dependency detected [ 2284.078604] 4.19.0+ #42 Tainted: G E [ 2284.078604] ------------------------------------------------------ [ 2284.078604] rmmod/254 is trying to acquire lock: [ 2284.078604] 00000000acd94e28 ((&n->timer)#2){+.-.}, at: del_timer_sync+0x5/0xa0 [ 2284.078604] [ 2284.078604] but task is already holding lock: [ 2284.078604] 00000000f997afc0 (&(&tn->node_list_lock)->rlock){+.-.}, at: tipc_node_stop+0xac/0x190 [tipc] [ 2284.078604] [ 2284.078604] which lock already depends on the new lock. [ 2284.078604] [ 2284.078604] [ 2284.078604] the existing dependency chain (in reverse order) is: [ 2284.078604] [ 2284.078604] -> #1 (&(&tn->node_list_lock)->rlock){+.-.}: [ 2284.078604] tipc_node_timeout+0x20a/0x330 [tipc] [ 2284.078604] call_timer_fn+0xa1/0x280 [ 2284.078604] run_timer_softirq+0x1f2/0x4d0 [ 2284.078604] __do_softirq+0xfc/0x413 [ 2284.078604] irq_exit+0xb5/0xc0 [ 2284.078604] smp_apic_timer_interrupt+0xac/0x210 [ 2284.078604] apic_timer_interrupt+0xf/0x20 [ 2284.078604] default_idle+0x1c/0x140 [ 2284.078604] do_idle+0x1bc/0x280 [ 2284.078604] cpu_startup_entry+0x19/0x20 [ 2284.078604] start_secondary+0x187/0x1c0 [ 2284.078604] secondary_startup_64+0xa4/0xb0 [ 2284.078604] [ 2284.078604] -> #0 ((&n->timer)#2){+.-.}: [ 2284.078604] del_timer_sync+0x34/0xa0 [ 2284.078604] tipc_node_delete+0x1a/0x40 [tipc] [ 2284.078604] tipc_node_stop+0xcb/0x190 [tipc] [ 2284.078604] tipc_net_stop+0x154/0x170 [tipc] [ 2284.078604] tipc_exit_net+0x16/0x30 [tipc] [ 2284.078604] ops_exit_list.isra.8+0x36/0x70 [ 2284.078604] unregister_pernet_operations+0x87/0xd0 [ 2284.078604] unregister_pernet_subsys+0x1d/0x30 [ 2284.078604] tipc_exit+0x11/0x6f2 [tipc] [ 2284.078604] __x64_sys_delete_module+0x1df/0x240 [ 2284.078604] do_syscall_64+0x66/0x460 [ 2284.078604] entry_SYSCALL_64_after_hwframe+0x49/0xbe [ 2284.078604] [ 2284.078604] other info that might help us debug this: [ 2284.078604] [ 2284.078604] Possible unsafe locking scenario: [ 2284.078604] [ 2284.078604] CPU0 CPU1 [ 2284.078604] ---- ---- [ 2284.078604] lock(&(&tn->node_list_lock)->rlock); [ 2284.078604] lock((&n->timer)#2); [ 2284.078604] lock(&(&tn->node_list_lock)->rlock); [ 2284.078604] lock((&n->timer)#2); [ 2284.078604] [ 2284.078604] *** DEADLOCK *** [ 2284.078604] [ 2284.078604] 3 locks held by rmmod/254: [ 2284.078604] #0: 000000003368be9b (pernet_ops_rwsem){+.+.}, at: unregister_pernet_subsys+0x15/0x30 [ 2284.078604] #1: 0000000046ed9c86 (rtnl_mutex){+.+.}, at: tipc_net_stop+0x144/0x170 [tipc] [ 2284.078604] #2: 00000000f997afc0 (&(&tn->node_list_lock)->rlock){+.-.}, at: tipc_node_stop+0xac/0x19 [...} The reason is that the node timer handler sometimes needs to delete a node which has been disconnected for too long. To do this, it grabs the lock 'node_list_lock', which may at the same time be held by the generic node cleanup function, tipc_node_stop(), during module removal. Since the latter is calling del_timer_sync() inside the same lock, we have a potential deadlock. We fix this letting the timer cleanup function use spin_trylock() instead of just spin_lock(), and when it fails to grab the lock it just returns so that the timer handler can terminate its execution. This is safe to do, since tipc_node_stop() anyway is about to delete both the timer and the node instance. Fixes: 6a939f365bdb ("tipc: Auto removal of peer down node instance") 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>
2018-11-27 00:26:14 +07:00
/* If lock held by tipc_node_stop() the node will be deleted anyway */
if (!spin_trylock_bh(&tn->node_list_lock))
return false;
tipc_node_write_lock(peer);
if (!node_is_up(peer) && time_after(jiffies, peer->delete_at)) {
tipc_node_clear_links(peer);
tipc_node_delete_from_list(peer);
deleted = true;
}
tipc_node_write_unlock(peer);
if (!deleted) {
spin_unlock_bh(&tn->node_list_lock);
return deleted;
}
/* Calculate cluster capabilities */
tn->capabilities = TIPC_NODE_CAPABILITIES;
list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
tn->capabilities &= temp_node->capabilities;
}
spin_unlock_bh(&tn->node_list_lock);
return deleted;
}
/* tipc_node_timeout - handle expiration of node timer
*/
static void tipc_node_timeout(struct timer_list *t)
{
struct tipc_node *n = from_timer(n, t, timer);
struct tipc_link_entry *le;
struct sk_buff_head xmitq;
int remains = n->link_cnt;
int bearer_id;
int rc = 0;
trace_tipc_node_timeout(n, false, " ");
if (!node_is_up(n) && tipc_node_cleanup(n)) {
/*Removing the reference of Timer*/
tipc_node_put(n);
return;
}
__skb_queue_head_init(&xmitq);
tipc: fix node keep alive interval calculation When setting LINK tolerance, node timer interval will be calculated base on the LINK with lowest tolerance. But when calculated, the old node timer interval only updated if current setting value (tolerance/4) less than old ones regardless of number of links as well as links' lowest tolerance value. This caused to two cases missing if tolerance changed as following: Case 1: 1.1/ There is one link (L1) available in the system 1.2/ Set L1's tolerance from 1500ms => lower (i.e 500ms) 1.3/ Then, fallback to default (1500ms) or higher (i.e 2000ms) Expected: node timer interval is 1500/4=375ms after 1.3 Result: node timer interval will not being updated after changing tolerance at 1.3 since its value 1500/4=375ms is not less than 500/4=125ms at 1.2. Case 2: 2.1/ There are two links (L1, L2) available in the system 2.2/ L1 and L2 tolerance value are 2000ms as initial 2.3/ Set L2's tolerance from 2000ms => lower 1500ms 2.4/ Disable link L2 (bring down its bearer) Expected: node timer interval is 2000ms/4=500ms after 2.4 Result: node timer interval will not being updated after disabling L2 since its value 2000ms/4=500ms is still not less than 1500/4=375ms at 2.3 although L2 is already not available in the system. To fix this, we start the node interval calculation by initializing it to a value larger than any conceivable calculated value. This way, the link with the lowest tolerance will always determine the calculated value. 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>
2018-12-06 09:00:09 +07:00
/* Initial node interval to value larger (10 seconds), then it will be
* recalculated with link lowest tolerance
*/
tipc_node_read_lock(n);
n->keepalive_intv = 10000;
tipc_node_read_unlock(n);
for (bearer_id = 0; remains && (bearer_id < MAX_BEARERS); bearer_id++) {
tipc_node_read_lock(n);
le = &n->links[bearer_id];
if (le->link) {
spin_lock_bh(&le->lock);
/* Link tolerance may change asynchronously: */
tipc_node_calculate_timer(n, le->link);
rc = tipc_link_timeout(le->link, &xmitq);
spin_unlock_bh(&le->lock);
remains--;
}
tipc_node_read_unlock(n);
tipc_bearer_xmit(n->net, bearer_id, &xmitq, &le->maddr);
if (rc & TIPC_LINK_DOWN_EVT)
tipc_node_link_down(n, bearer_id, false);
}
mod_timer(&n->timer, jiffies + msecs_to_jiffies(n->keepalive_intv));
}
/**
* __tipc_node_link_up - handle addition of link
* Node lock must be held by caller
* Link becomes active (alone or shared) or standby, depending on its priority.
*/
static void __tipc_node_link_up(struct tipc_node *n, int bearer_id,
struct sk_buff_head *xmitq)
{
int *slot0 = &n->active_links[0];
int *slot1 = &n->active_links[1];
struct tipc_link *ol = node_active_link(n, 0);
struct tipc_link *nl = n->links[bearer_id].link;
if (!nl || tipc_link_is_up(nl))
tipc: delay ESTABLISH state event when link is established Link establishing, just like link teardown, is a non-atomic action, in the sense that discovering that conditions are right to establish a link, and the actual adding of the link to one of the node's send slots is done in two different lock contexts. The link FSM is designed to help bridging the gap between the two contexts in a safe manner. We have now discovered a weakness in the implementaton of this FSM. Because we directly let the link go from state LINK_ESTABLISHING to state LINK_ESTABLISHED already in the first lock context, we are unable to distinguish between a fully established link, i.e., a link that has been added to its slot, and a link that has not yet reached the second lock context. It may hence happen that a manual intervention, e.g., when disabling an interface, causes the function tipc_node_link_down() to try removing the link from the node slots, decrementing its active link counter etc, although the link was never added there in the first place. We solve this by delaying the actual state change until we reach the second lock context, inside the function tipc_node_link_up(). This makes it possible for potentail callers of __tipc_node_link_down() to know if they should proceed or not, and the problem is solved. Unforunately, the situation described above also has a second problem. Since there by necessity is a tipc_node_link_up() call pending once the node lock has been released, we must defuse that call by setting the link back from LINK_ESTABLISHING to LINK_RESET state. This forces us to make a slight modification to the link FSM, which will now look as follows. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | +----------------+ | | | RESET_EVT| |RESET_EVT | | | | | | | | | | |ESTABLISH_EVT | | | | +-------------+ | | | | | | RESET_EVT | | | | | | | | | | | V V V | | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-16 01:52:44 +07:00
return;
tipc_link_fsm_evt(nl, LINK_ESTABLISH_EVT);
if (!tipc_link_is_up(nl))
return;
n->working_links++;
n->action_flags |= TIPC_NOTIFY_LINK_UP;
n->link_id = tipc_link_id(nl);
/* Leave room for tunnel header when returning 'mtu' to users: */
n->links[bearer_id].mtu = tipc_link_mtu(nl) - INT_H_SIZE;
tipc_bearer_add_dest(n->net, bearer_id, n->addr);
tipc_bcast_inc_bearer_dst_cnt(n->net, bearer_id);
pr_debug("Established link <%s> on network plane %c\n",
tipc_link_name(nl), tipc_link_plane(nl));
trace_tipc_node_link_up(n, true, " ");
/* Ensure that a STATE message goes first */
tipc_link_build_state_msg(nl, xmitq);
/* First link? => give it both slots */
if (!ol) {
*slot0 = bearer_id;
*slot1 = bearer_id;
tipc_node_fsm_evt(n, SELF_ESTABL_CONTACT_EVT);
n->action_flags |= TIPC_NOTIFY_NODE_UP;
tipc_link_set_active(nl, true);
tipc_bcast_add_peer(n->net, nl, xmitq);
return;
}
/* Second link => redistribute slots */
if (tipc_link_prio(nl) > tipc_link_prio(ol)) {
pr_debug("Old link <%s> becomes standby\n", tipc_link_name(ol));
*slot0 = bearer_id;
*slot1 = bearer_id;
tipc_link_set_active(nl, true);
tipc_link_set_active(ol, false);
} else if (tipc_link_prio(nl) == tipc_link_prio(ol)) {
tipc_link_set_active(nl, true);
*slot1 = bearer_id;
} else {
pr_debug("New link <%s> is standby\n", tipc_link_name(nl));
}
/* Prepare synchronization with first link */
tipc_link_tnl_prepare(ol, nl, SYNCH_MSG, xmitq);
}
/**
* tipc_node_link_up - handle addition of link
*
* Link becomes active (alone or shared) or standby, depending on its priority.
*/
static void tipc_node_link_up(struct tipc_node *n, int bearer_id,
struct sk_buff_head *xmitq)
{
struct tipc_media_addr *maddr;
tipc_node_write_lock(n);
__tipc_node_link_up(n, bearer_id, xmitq);
maddr = &n->links[bearer_id].maddr;
tipc_bearer_xmit(n->net, bearer_id, xmitq, maddr);
tipc_node_write_unlock(n);
}
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
/**
* tipc_node_link_failover() - start failover in case "half-failover"
*
* This function is only called in a very special situation where link
* failover can be already started on peer node but not on this node.
* This can happen when e.g.
* 1. Both links <1A-2A>, <1B-2B> down
* 2. Link endpoint 2A up, but 1A still down (e.g. due to network
* disturbance, wrong session, etc.)
* 3. Link <1B-2B> up
* 4. Link endpoint 2A down (e.g. due to link tolerance timeout)
tipc: fix issues with early FAILOVER_MSG from peer It appears that a FAILOVER_MSG can come from peer even when the failure link is resetting (i.e. just after the 'node_write_unlock()'...). This means the failover procedure on the node has not been started yet. The situation is as follows: node1 node2 linkb linka linka linkb | | | | | | x failure | | | RESETTING | | | | | | x failure RESET | | RESETTING FAILINGOVER | | | (FAILOVER_MSG) | | |<-------------------------------------------------| | *FAILINGOVER | | | | | (dummy FAILOVER_MSG) | | |------------------------------------------------->| | RESET | | FAILOVER_END | FAILINGOVER RESET | . . . . . . . . . . . . Once this happens, the link failover procedure will be triggered wrongly on the receiving node since the node isn't in FAILINGOVER state but then another link failover will be carried out. The consequences are: 1) A peer might get stuck in FAILINGOVER state because the 'sync_point' was set, reset and set incorrectly, the criteria to end the failover would not be met, it could keep waiting for a message that has already received. 2) The early FAILOVER_MSG(s) could be queued in the link failover deferdq but would be purged or not pulled out because the 'drop_point' was not set correctly. 3) The early FAILOVER_MSG(s) could be dropped too. 4) The dummy FAILOVER_MSG could make the peer leaving FAILINGOVER state shortly, but later on it would be restarted. The same situation can also happen when the link is in PEER_RESET state and a FAILOVER_MSG arrives. The commit resolves the issues by forcing the link down immediately, so the failover procedure will be started normally (which is the same as when receiving a FAILOVER_MSG and the link is in up state). Also, the function "tipc_node_link_failover()" is toughen to avoid such a situation from happening. Acked-by: Jon Maloy <jon.maloy@ericsson.se> Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-06-17 11:56:12 +07:00
* 5. Node 2 starts failover onto link <1B-2B>
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
*
tipc: fix issues with early FAILOVER_MSG from peer It appears that a FAILOVER_MSG can come from peer even when the failure link is resetting (i.e. just after the 'node_write_unlock()'...). This means the failover procedure on the node has not been started yet. The situation is as follows: node1 node2 linkb linka linka linkb | | | | | | x failure | | | RESETTING | | | | | | x failure RESET | | RESETTING FAILINGOVER | | | (FAILOVER_MSG) | | |<-------------------------------------------------| | *FAILINGOVER | | | | | (dummy FAILOVER_MSG) | | |------------------------------------------------->| | RESET | | FAILOVER_END | FAILINGOVER RESET | . . . . . . . . . . . . Once this happens, the link failover procedure will be triggered wrongly on the receiving node since the node isn't in FAILINGOVER state but then another link failover will be carried out. The consequences are: 1) A peer might get stuck in FAILINGOVER state because the 'sync_point' was set, reset and set incorrectly, the criteria to end the failover would not be met, it could keep waiting for a message that has already received. 2) The early FAILOVER_MSG(s) could be queued in the link failover deferdq but would be purged or not pulled out because the 'drop_point' was not set correctly. 3) The early FAILOVER_MSG(s) could be dropped too. 4) The dummy FAILOVER_MSG could make the peer leaving FAILINGOVER state shortly, but later on it would be restarted. The same situation can also happen when the link is in PEER_RESET state and a FAILOVER_MSG arrives. The commit resolves the issues by forcing the link down immediately, so the failover procedure will be started normally (which is the same as when receiving a FAILOVER_MSG and the link is in up state). Also, the function "tipc_node_link_failover()" is toughen to avoid such a situation from happening. Acked-by: Jon Maloy <jon.maloy@ericsson.se> Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-06-17 11:56:12 +07:00
* ==> Node 1 does never start link/node failover!
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
*
* @n: tipc node structure
* @l: link peer endpoint failingover (- can be NULL)
* @tnl: tunnel link
* @xmitq: queue for messages to be xmited on tnl link later
*/
static void tipc_node_link_failover(struct tipc_node *n, struct tipc_link *l,
struct tipc_link *tnl,
struct sk_buff_head *xmitq)
{
/* Avoid to be "self-failover" that can never end */
if (!tipc_link_is_up(tnl))
return;
tipc: fix issues with early FAILOVER_MSG from peer It appears that a FAILOVER_MSG can come from peer even when the failure link is resetting (i.e. just after the 'node_write_unlock()'...). This means the failover procedure on the node has not been started yet. The situation is as follows: node1 node2 linkb linka linka linkb | | | | | | x failure | | | RESETTING | | | | | | x failure RESET | | RESETTING FAILINGOVER | | | (FAILOVER_MSG) | | |<-------------------------------------------------| | *FAILINGOVER | | | | | (dummy FAILOVER_MSG) | | |------------------------------------------------->| | RESET | | FAILOVER_END | FAILINGOVER RESET | . . . . . . . . . . . . Once this happens, the link failover procedure will be triggered wrongly on the receiving node since the node isn't in FAILINGOVER state but then another link failover will be carried out. The consequences are: 1) A peer might get stuck in FAILINGOVER state because the 'sync_point' was set, reset and set incorrectly, the criteria to end the failover would not be met, it could keep waiting for a message that has already received. 2) The early FAILOVER_MSG(s) could be queued in the link failover deferdq but would be purged or not pulled out because the 'drop_point' was not set correctly. 3) The early FAILOVER_MSG(s) could be dropped too. 4) The dummy FAILOVER_MSG could make the peer leaving FAILINGOVER state shortly, but later on it would be restarted. The same situation can also happen when the link is in PEER_RESET state and a FAILOVER_MSG arrives. The commit resolves the issues by forcing the link down immediately, so the failover procedure will be started normally (which is the same as when receiving a FAILOVER_MSG and the link is in up state). Also, the function "tipc_node_link_failover()" is toughen to avoid such a situation from happening. Acked-by: Jon Maloy <jon.maloy@ericsson.se> Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-06-17 11:56:12 +07:00
/* Don't rush, failure link may be in the process of resetting */
if (l && !tipc_link_is_reset(l))
return;
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT);
tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT);
n->sync_point = tipc_link_rcv_nxt(tnl) + (U16_MAX / 2 - 1);
tipc_link_failover_prepare(l, tnl, xmitq);
if (l)
tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT);
tipc_node_fsm_evt(n, NODE_FAILOVER_BEGIN_EVT);
}
/**
* __tipc_node_link_down - handle loss of link
*/
static void __tipc_node_link_down(struct tipc_node *n, int *bearer_id,
struct sk_buff_head *xmitq,
struct tipc_media_addr **maddr)
{
struct tipc_link_entry *le = &n->links[*bearer_id];
int *slot0 = &n->active_links[0];
int *slot1 = &n->active_links[1];
int i, highest = 0, prio;
struct tipc_link *l, *_l, *tnl;
l = n->links[*bearer_id].link;
tipc: merge link->exec_mode and link->state into one FSM Until now, we have been handling link failover and synchronization by using an additional link state variable, "exec_mode". This variable is not independent of the link FSM state, something causing a risk of inconsistencies, apart from the fact that it clutters the code. The conditions are now in place to define a new link FSM that covers all existing use cases, including failover and synchronization, and eliminate the "exec_mode" field altogether. The FSM must also support non-atomic resetting of links, which will be introduced later. The new link FSM is shown below, with 7 states and 8 events. Only events leading to state change are shown as edges. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | | | | | | +--------------+ | | | RESET_EVT| |RESET_EVT |ESTABLISH_EVT | | | | | | | | | | | | V V | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ These changes are fully backwards compatible. Tested-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>
2015-07-31 05:24:21 +07:00
if (!l || tipc_link_is_reset(l))
return;
n->working_links--;
n->action_flags |= TIPC_NOTIFY_LINK_DOWN;
n->link_id = tipc_link_id(l);
tipc_bearer_remove_dest(n->net, *bearer_id, n->addr);
pr_debug("Lost link <%s> on network plane %c\n",
tipc_link_name(l), tipc_link_plane(l));
/* Select new active link if any available */
*slot0 = INVALID_BEARER_ID;
*slot1 = INVALID_BEARER_ID;
for (i = 0; i < MAX_BEARERS; i++) {
_l = n->links[i].link;
if (!_l || !tipc_link_is_up(_l))
continue;
if (_l == l)
continue;
prio = tipc_link_prio(_l);
if (prio < highest)
continue;
if (prio > highest) {
highest = prio;
*slot0 = i;
*slot1 = i;
continue;
}
*slot1 = i;
}
if (!node_is_up(n)) {
if (tipc_link_peer_is_down(l))
tipc_node_fsm_evt(n, PEER_LOST_CONTACT_EVT);
tipc_node_fsm_evt(n, SELF_LOST_CONTACT_EVT);
tipc: add trace_events for tipc link The commit adds the new trace_events for TIPC link object: trace_tipc_link_timeout() trace_tipc_link_fsm() trace_tipc_link_reset() trace_tipc_link_too_silent() trace_tipc_link_retrans() trace_tipc_link_bc_ack() trace_tipc_link_conges() And the traces for PROTOCOL messages at building and receiving: trace_tipc_proto_build() trace_tipc_proto_rcv() Note: a) The 'tipc_link_too_silent' event will only happen when the 'silent_intv_cnt' is about to reach the 'abort_limit' value (and the event is enabled). The benefit for this kind of event is that we can get an early indication about TIPC link loss issue due to timeout, then can do some necessary actions for troubleshooting. For example: To trigger the 'tipc_proto_rcv' when the 'too_silent' event occurs: echo 'enable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_too_silent/trigger And disable it when TIPC link is reset: echo 'disable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_reset/trigger b) The 'tipc_link_retrans' or 'tipc_link_bc_ack' event is useful to trace TIPC retransmission issues. In addition, the commit adds the 'trace_tipc_list/link_dump()' at the 'retransmission failure' case. Then, if the issue occurs, the link 'transmq' along with the link data can be dumped for post-analysis. These dump events should be enabled by default since it will only take effect when the failure happens. The same approach is also applied for the faulty case that the validation of protocol message is failed. Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:57 +07:00
trace_tipc_link_reset(l, TIPC_DUMP_ALL, "link down!");
tipc_link_fsm_evt(l, LINK_RESET_EVT);
tipc_link_reset(l);
tipc_link_build_reset_msg(l, xmitq);
*maddr = &n->links[*bearer_id].maddr;
node_lost_contact(n, &le->inputq);
tipc_bcast_dec_bearer_dst_cnt(n->net, *bearer_id);
return;
}
tipc_bcast_dec_bearer_dst_cnt(n->net, *bearer_id);
/* There is still a working link => initiate failover */
*bearer_id = n->active_links[0];
tnl = n->links[*bearer_id].link;
tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT);
tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT);
n->sync_point = tipc_link_rcv_nxt(tnl) + (U16_MAX / 2 - 1);
tipc_link_tnl_prepare(l, tnl, FAILOVER_MSG, xmitq);
tipc: add trace_events for tipc link The commit adds the new trace_events for TIPC link object: trace_tipc_link_timeout() trace_tipc_link_fsm() trace_tipc_link_reset() trace_tipc_link_too_silent() trace_tipc_link_retrans() trace_tipc_link_bc_ack() trace_tipc_link_conges() And the traces for PROTOCOL messages at building and receiving: trace_tipc_proto_build() trace_tipc_proto_rcv() Note: a) The 'tipc_link_too_silent' event will only happen when the 'silent_intv_cnt' is about to reach the 'abort_limit' value (and the event is enabled). The benefit for this kind of event is that we can get an early indication about TIPC link loss issue due to timeout, then can do some necessary actions for troubleshooting. For example: To trigger the 'tipc_proto_rcv' when the 'too_silent' event occurs: echo 'enable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_too_silent/trigger And disable it when TIPC link is reset: echo 'disable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_reset/trigger b) The 'tipc_link_retrans' or 'tipc_link_bc_ack' event is useful to trace TIPC retransmission issues. In addition, the commit adds the 'trace_tipc_list/link_dump()' at the 'retransmission failure' case. Then, if the issue occurs, the link 'transmq' along with the link data can be dumped for post-analysis. These dump events should be enabled by default since it will only take effect when the failure happens. The same approach is also applied for the faulty case that the validation of protocol message is failed. Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:57 +07:00
trace_tipc_link_reset(l, TIPC_DUMP_ALL, "link down -> failover!");
tipc_link_reset(l);
tipc_link_fsm_evt(l, LINK_RESET_EVT);
tipc: merge link->exec_mode and link->state into one FSM Until now, we have been handling link failover and synchronization by using an additional link state variable, "exec_mode". This variable is not independent of the link FSM state, something causing a risk of inconsistencies, apart from the fact that it clutters the code. The conditions are now in place to define a new link FSM that covers all existing use cases, including failover and synchronization, and eliminate the "exec_mode" field altogether. The FSM must also support non-atomic resetting of links, which will be introduced later. The new link FSM is shown below, with 7 states and 8 events. Only events leading to state change are shown as edges. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | | | | | | +--------------+ | | | RESET_EVT| |RESET_EVT |ESTABLISH_EVT | | | | | | | | | | | | V V | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ These changes are fully backwards compatible. Tested-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>
2015-07-31 05:24:21 +07:00
tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT);
tipc_node_fsm_evt(n, NODE_FAILOVER_BEGIN_EVT);
*maddr = &n->links[*bearer_id].maddr;
}
static void tipc_node_link_down(struct tipc_node *n, int bearer_id, bool delete)
{
struct tipc_link_entry *le = &n->links[bearer_id];
struct tipc_media_addr *maddr = NULL;
tipc: delay ESTABLISH state event when link is established Link establishing, just like link teardown, is a non-atomic action, in the sense that discovering that conditions are right to establish a link, and the actual adding of the link to one of the node's send slots is done in two different lock contexts. The link FSM is designed to help bridging the gap between the two contexts in a safe manner. We have now discovered a weakness in the implementaton of this FSM. Because we directly let the link go from state LINK_ESTABLISHING to state LINK_ESTABLISHED already in the first lock context, we are unable to distinguish between a fully established link, i.e., a link that has been added to its slot, and a link that has not yet reached the second lock context. It may hence happen that a manual intervention, e.g., when disabling an interface, causes the function tipc_node_link_down() to try removing the link from the node slots, decrementing its active link counter etc, although the link was never added there in the first place. We solve this by delaying the actual state change until we reach the second lock context, inside the function tipc_node_link_up(). This makes it possible for potentail callers of __tipc_node_link_down() to know if they should proceed or not, and the problem is solved. Unforunately, the situation described above also has a second problem. Since there by necessity is a tipc_node_link_up() call pending once the node lock has been released, we must defuse that call by setting the link back from LINK_ESTABLISHING to LINK_RESET state. This forces us to make a slight modification to the link FSM, which will now look as follows. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | +----------------+ | | | RESET_EVT| |RESET_EVT | | | | | | | | | | |ESTABLISH_EVT | | | | +-------------+ | | | | | | RESET_EVT | | | | | | | | | | | V V V | | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-16 01:52:44 +07:00
struct tipc_link *l = le->link;
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
int old_bearer_id = bearer_id;
struct sk_buff_head xmitq;
tipc: delay ESTABLISH state event when link is established Link establishing, just like link teardown, is a non-atomic action, in the sense that discovering that conditions are right to establish a link, and the actual adding of the link to one of the node's send slots is done in two different lock contexts. The link FSM is designed to help bridging the gap between the two contexts in a safe manner. We have now discovered a weakness in the implementaton of this FSM. Because we directly let the link go from state LINK_ESTABLISHING to state LINK_ESTABLISHED already in the first lock context, we are unable to distinguish between a fully established link, i.e., a link that has been added to its slot, and a link that has not yet reached the second lock context. It may hence happen that a manual intervention, e.g., when disabling an interface, causes the function tipc_node_link_down() to try removing the link from the node slots, decrementing its active link counter etc, although the link was never added there in the first place. We solve this by delaying the actual state change until we reach the second lock context, inside the function tipc_node_link_up(). This makes it possible for potentail callers of __tipc_node_link_down() to know if they should proceed or not, and the problem is solved. Unforunately, the situation described above also has a second problem. Since there by necessity is a tipc_node_link_up() call pending once the node lock has been released, we must defuse that call by setting the link back from LINK_ESTABLISHING to LINK_RESET state. This forces us to make a slight modification to the link FSM, which will now look as follows. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | +----------------+ | | | RESET_EVT| |RESET_EVT | | | | | | | | | | |ESTABLISH_EVT | | | | +-------------+ | | | | | | RESET_EVT | | | | | | | | | | | V V V | | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-16 01:52:44 +07:00
if (!l)
return;
__skb_queue_head_init(&xmitq);
tipc_node_write_lock(n);
tipc: delay ESTABLISH state event when link is established Link establishing, just like link teardown, is a non-atomic action, in the sense that discovering that conditions are right to establish a link, and the actual adding of the link to one of the node's send slots is done in two different lock contexts. The link FSM is designed to help bridging the gap between the two contexts in a safe manner. We have now discovered a weakness in the implementaton of this FSM. Because we directly let the link go from state LINK_ESTABLISHING to state LINK_ESTABLISHED already in the first lock context, we are unable to distinguish between a fully established link, i.e., a link that has been added to its slot, and a link that has not yet reached the second lock context. It may hence happen that a manual intervention, e.g., when disabling an interface, causes the function tipc_node_link_down() to try removing the link from the node slots, decrementing its active link counter etc, although the link was never added there in the first place. We solve this by delaying the actual state change until we reach the second lock context, inside the function tipc_node_link_up(). This makes it possible for potentail callers of __tipc_node_link_down() to know if they should proceed or not, and the problem is solved. Unforunately, the situation described above also has a second problem. Since there by necessity is a tipc_node_link_up() call pending once the node lock has been released, we must defuse that call by setting the link back from LINK_ESTABLISHING to LINK_RESET state. This forces us to make a slight modification to the link FSM, which will now look as follows. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | +----------------+ | | | RESET_EVT| |RESET_EVT | | | | | | | | | | |ESTABLISH_EVT | | | | +-------------+ | | | | | | RESET_EVT | | | | | | | | | | | V V V | | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-16 01:52:44 +07:00
if (!tipc_link_is_establishing(l)) {
__tipc_node_link_down(n, &bearer_id, &xmitq, &maddr);
} else {
/* Defuse pending tipc_node_link_up() */
tipc: fix link session and re-establish issues When a link endpoint is re-created (e.g. after a node reboot or interface reset), the link session number is varied by random, the peer endpoint will be synced with this new session number before the link is re-established. However, there is a shortcoming in this mechanism that can lead to the link never re-established or faced with a failure then. It happens when the peer endpoint is ready in ESTABLISHING state, the 'peer_session' as well as the 'in_session' flag have been set, but suddenly this link endpoint leaves. When it comes back with a random session number, there are two situations possible: 1/ If the random session number is larger than (or equal to) the previous one, the peer endpoint will be updated with this new session upon receipt of a RESET_MSG from this endpoint, and the link can be re- established as normal. Otherwise, all the RESET_MSGs from this endpoint will be rejected by the peer. In turn, when this link endpoint receives one ACTIVATE_MSG from the peer, it will move to ESTABLISHED and start to send STATE_MSGs, but again these messages will be dropped by the peer due to wrong session. The peer link endpoint can still become ESTABLISHED after receiving a traffic message from this endpoint (e.g. a BCAST_PROTOCOL or NAME_DISTRIBUTOR), but since all the STATE_MSGs are invalid, the link will be forced down sooner or later! Even in case the random session number is larger than the previous one, it can be that the ACTIVATE_MSG from the peer arrives first, and this link endpoint moves quickly to ESTABLISHED without sending out any RESET_MSG yet. Consequently, the peer link will not be updated with the new session number, and the same link failure scenario as above will happen. 2/ Another situation can be that, the peer link endpoint was reset due to any reasons in the meantime, its link state was set to RESET from ESTABLISHING but still in session, i.e. the 'in_session' flag is not reset... Now, if the random session number from this endpoint is less than the previous one, all the RESET_MSGs from this endpoint will be rejected by the peer. In the other direction, when this link endpoint receives a RESET_MSG from the peer, it moves to ESTABLISHING and starts to send ACTIVATE_MSGs, but all these messages will be rejected by the peer too. As a result, the link cannot be re-established but gets stuck with this link endpoint in state ESTABLISHING and the peer in RESET! Solution: =========== This link endpoint should not go directly to ESTABLISHED when getting ACTIVATE_MSG from the peer which may belong to the old session if the link was re-created. To ensure the session to be correct before the link is re-established, the peer endpoint in ESTABLISHING state will send back the last session number in ACTIVATE_MSG for a verification at this endpoint. Then, if needed, a new and more appropriate session number will be regenerated to force a re-synch first. In addition, when a link in ESTABLISHING state is reset, its state will move to RESET according to the link FSM, along with resetting the 'in_session' flag (and the other data) as a normal link reset, it will also be deleted if requested. The solution is backward compatible. Acked-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-11 13:29:43 +07:00
tipc_link_reset(l);
tipc: delay ESTABLISH state event when link is established Link establishing, just like link teardown, is a non-atomic action, in the sense that discovering that conditions are right to establish a link, and the actual adding of the link to one of the node's send slots is done in two different lock contexts. The link FSM is designed to help bridging the gap between the two contexts in a safe manner. We have now discovered a weakness in the implementaton of this FSM. Because we directly let the link go from state LINK_ESTABLISHING to state LINK_ESTABLISHED already in the first lock context, we are unable to distinguish between a fully established link, i.e., a link that has been added to its slot, and a link that has not yet reached the second lock context. It may hence happen that a manual intervention, e.g., when disabling an interface, causes the function tipc_node_link_down() to try removing the link from the node slots, decrementing its active link counter etc, although the link was never added there in the first place. We solve this by delaying the actual state change until we reach the second lock context, inside the function tipc_node_link_up(). This makes it possible for potentail callers of __tipc_node_link_down() to know if they should proceed or not, and the problem is solved. Unforunately, the situation described above also has a second problem. Since there by necessity is a tipc_node_link_up() call pending once the node lock has been released, we must defuse that call by setting the link back from LINK_ESTABLISHING to LINK_RESET state. This forces us to make a slight modification to the link FSM, which will now look as follows. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | +----------------+ | | | RESET_EVT| |RESET_EVT | | | | | | | | | | |ESTABLISH_EVT | | | | +-------------+ | | | | | | RESET_EVT | | | | | | | | | | | V V V | | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-16 01:52:44 +07:00
tipc_link_fsm_evt(l, LINK_RESET_EVT);
}
tipc: fix link session and re-establish issues When a link endpoint is re-created (e.g. after a node reboot or interface reset), the link session number is varied by random, the peer endpoint will be synced with this new session number before the link is re-established. However, there is a shortcoming in this mechanism that can lead to the link never re-established or faced with a failure then. It happens when the peer endpoint is ready in ESTABLISHING state, the 'peer_session' as well as the 'in_session' flag have been set, but suddenly this link endpoint leaves. When it comes back with a random session number, there are two situations possible: 1/ If the random session number is larger than (or equal to) the previous one, the peer endpoint will be updated with this new session upon receipt of a RESET_MSG from this endpoint, and the link can be re- established as normal. Otherwise, all the RESET_MSGs from this endpoint will be rejected by the peer. In turn, when this link endpoint receives one ACTIVATE_MSG from the peer, it will move to ESTABLISHED and start to send STATE_MSGs, but again these messages will be dropped by the peer due to wrong session. The peer link endpoint can still become ESTABLISHED after receiving a traffic message from this endpoint (e.g. a BCAST_PROTOCOL or NAME_DISTRIBUTOR), but since all the STATE_MSGs are invalid, the link will be forced down sooner or later! Even in case the random session number is larger than the previous one, it can be that the ACTIVATE_MSG from the peer arrives first, and this link endpoint moves quickly to ESTABLISHED without sending out any RESET_MSG yet. Consequently, the peer link will not be updated with the new session number, and the same link failure scenario as above will happen. 2/ Another situation can be that, the peer link endpoint was reset due to any reasons in the meantime, its link state was set to RESET from ESTABLISHING but still in session, i.e. the 'in_session' flag is not reset... Now, if the random session number from this endpoint is less than the previous one, all the RESET_MSGs from this endpoint will be rejected by the peer. In the other direction, when this link endpoint receives a RESET_MSG from the peer, it moves to ESTABLISHING and starts to send ACTIVATE_MSGs, but all these messages will be rejected by the peer too. As a result, the link cannot be re-established but gets stuck with this link endpoint in state ESTABLISHING and the peer in RESET! Solution: =========== This link endpoint should not go directly to ESTABLISHED when getting ACTIVATE_MSG from the peer which may belong to the old session if the link was re-created. To ensure the session to be correct before the link is re-established, the peer endpoint in ESTABLISHING state will send back the last session number in ACTIVATE_MSG for a verification at this endpoint. Then, if needed, a new and more appropriate session number will be regenerated to force a re-synch first. In addition, when a link in ESTABLISHING state is reset, its state will move to RESET according to the link FSM, along with resetting the 'in_session' flag (and the other data) as a normal link reset, it will also be deleted if requested. The solution is backward compatible. Acked-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-02-11 13:29:43 +07:00
if (delete) {
kfree(l);
le->link = NULL;
n->link_cnt--;
}
trace_tipc_node_link_down(n, true, "node link down or deleted!");
tipc_node_write_unlock(n);
tipc: add neighbor monitoring framework TIPC based clusters are by default set up with full-mesh link connectivity between all nodes. Those links are expected to provide a short failure detection time, by default set to 1500 ms. Because of this, the background load for neighbor monitoring in an N-node cluster increases with a factor N on each node, while the overall monitoring traffic through the network infrastructure increases at a ~(N * (N - 1)) rate. Experience has shown that such clusters don't scale well beyond ~100 nodes unless we significantly increase failure discovery tolerance. This commit introduces a framework and an algorithm that drastically reduces this background load, while basically maintaining the original failure detection times across the whole cluster. Using this algorithm, background load will now grow at a rate of ~(2 * sqrt(N)) per node, and at ~(2 * N * sqrt(N)) in traffic overhead. As an example, each node will now have to actively monitor 38 neighbors in a 400-node cluster, instead of as before 399. This "Overlapping Ring Supervision Algorithm" is completely distributed and employs no centralized or coordinated state. It goes as follows: - Each node makes up a linearly ascending, circular list of all its N known neighbors, based on their TIPC node identity. This algorithm must be the same on all nodes. - The node then selects the next M = sqrt(N) - 1 nodes downstream from itself in the list, and chooses to actively monitor those. This is called its "local monitoring domain". - It creates a domain record describing the monitoring domain, and piggy-backs this in the data area of all neighbor monitoring messages (LINK_PROTOCOL/STATE) leaving that node. This means that all nodes in the cluster eventually (default within 400 ms) will learn about its monitoring domain. - Whenever a node discovers a change in its local domain, e.g., a node has been added or has gone down, it creates and sends out a new version of its node record to inform all neighbors about the change. - A node receiving a domain record from anybody outside its local domain matches this against its own list (which may not look the same), and chooses to not actively monitor those members of the received domain record that are also present in its own list. Instead, it relies on indications from the direct monitoring nodes if an indirectly monitored node has gone up or down. If a node is indicated lost, the receiving node temporarily activates its own direct monitoring towards that node in order to confirm, or not, that it is actually gone. - Since each node is actively monitoring sqrt(N) downstream neighbors, each node is also actively monitored by the same number of upstream neighbors. This means that all non-direct monitoring nodes normally will receive sqrt(N) indications that a node is gone. - A major drawback with ring monitoring is how it handles failures that cause massive network partitionings. If both a lost node and all its direct monitoring neighbors are inside the lost partition, the nodes in the remaining partition will never receive indications about the loss. To overcome this, each node also chooses to actively monitor some nodes outside its local domain. Those nodes are called remote domain "heads", and are selected in such a way that no node in the cluster will be more than two direct monitoring hops away. Because of this, each node, apart from monitoring the member of its local domain, will also typically monitor sqrt(N) remote head nodes. - As an optimization, local list status, domain status and domain records are marked with a generation number. This saves senders from unnecessarily conveying unaltered domain records, and receivers from performing unneeded re-adaptations of their node monitoring list, such as re-assigning domain heads. - As a measure of caution we have added the possibility to disable the new algorithm through configuration. We do this by keeping a threshold value for the cluster size; a cluster that grows beyond this value will switch from full-mesh to ring monitoring, and vice versa when it shrinks below the value. This means that if the threshold is set to a value larger than any anticipated cluster size (default size is 32) the new algorithm is effectively disabled. A patch set for altering the threshold value and for listing the table contents will follow shortly. - This change is fully backwards compatible. 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>
2016-06-14 07:46:22 +07:00
if (delete)
tipc_mon_remove_peer(n->net, n->addr, old_bearer_id);
if (!skb_queue_empty(&xmitq))
tipc_bearer_xmit(n->net, bearer_id, &xmitq, maddr);
tipc_sk_rcv(n->net, &le->inputq);
}
static bool node_is_up(struct tipc_node *n)
{
return n->active_links[0] != INVALID_BEARER_ID;
}
bool tipc_node_is_up(struct net *net, u32 addr)
{
struct tipc_node *n;
bool retval = false;
if (in_own_node(net, addr))
return true;
n = tipc_node_find(net, addr);
if (!n)
return false;
retval = node_is_up(n);
tipc_node_put(n);
return retval;
}
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
static u32 tipc_node_suggest_addr(struct net *net, u32 addr)
{
struct tipc_node *n;
addr ^= tipc_net(net)->random;
while ((n = tipc_node_find(net, addr))) {
tipc_node_put(n);
addr++;
}
return addr;
}
/* tipc_node_try_addr(): Check if addr can be used by peer, suggest other if not
* Returns suggested address if any, otherwise 0
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
*/
u32 tipc_node_try_addr(struct net *net, u8 *id, u32 addr)
{
struct tipc_net *tn = tipc_net(net);
struct tipc_node *n;
/* Suggest new address if some other peer is using this one */
n = tipc_node_find(net, addr);
if (n) {
if (!memcmp(n->peer_id, id, NODE_ID_LEN))
addr = 0;
tipc_node_put(n);
if (!addr)
return 0;
return tipc_node_suggest_addr(net, addr);
}
/* Suggest previously used address if peer is known */
n = tipc_node_find_by_id(net, id);
if (n) {
addr = n->addr;
tipc_node_put(n);
return addr;
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
}
/* Even this node may be in conflict */
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
if (tn->trial_addr == addr)
return tipc_node_suggest_addr(net, addr);
return 0;
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
}
void tipc_node_check_dest(struct net *net, u32 addr,
u8 *peer_id, struct tipc_bearer *b,
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
u16 capabilities, u32 signature, u32 hash_mixes,
struct tipc_media_addr *maddr,
bool *respond, bool *dupl_addr)
{
struct tipc_node *n;
struct tipc_link *l;
struct tipc_link_entry *le;
bool addr_match = false;
bool sign_match = false;
bool link_up = false;
bool accept_addr = false;
bool reset = true;
char *if_name;
unsigned long intv;
u16 session;
*dupl_addr = false;
*respond = false;
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
n = tipc_node_create(net, addr, peer_id, capabilities, signature,
hash_mixes);
if (!n)
return;
tipc_node_write_lock(n);
le = &n->links[b->identity];
/* Prepare to validate requesting node's signature and media address */
l = le->link;
link_up = l && tipc_link_is_up(l);
addr_match = l && !memcmp(&le->maddr, maddr, sizeof(*maddr));
sign_match = (signature == n->signature);
/* These three flags give us eight permutations: */
if (sign_match && addr_match && link_up) {
/* All is fine. Do nothing. */
reset = false;
/* Peer node is not a container/local namespace */
if (!n->peer_hash_mix)
n->peer_hash_mix = hash_mixes;
} else if (sign_match && addr_match && !link_up) {
/* Respond. The link will come up in due time */
*respond = true;
} else if (sign_match && !addr_match && link_up) {
/* Peer has changed i/f address without rebooting.
* If so, the link will reset soon, and the next
* discovery will be accepted. So we can ignore it.
* It may also be an cloned or malicious peer having
* chosen the same node address and signature as an
* existing one.
* Ignore requests until the link goes down, if ever.
*/
*dupl_addr = true;
} else if (sign_match && !addr_match && !link_up) {
/* Peer link has changed i/f address without rebooting.
* It may also be a cloned or malicious peer; we can't
* distinguish between the two.
* The signature is correct, so we must accept.
*/
accept_addr = true;
*respond = true;
} else if (!sign_match && addr_match && link_up) {
/* Peer node rebooted. Two possibilities:
* - Delayed re-discovery; this link endpoint has already
* reset and re-established contact with the peer, before
* receiving a discovery message from that node.
* (The peer happened to receive one from this node first).
* - The peer came back so fast that our side has not
* discovered it yet. Probing from this side will soon
* reset the link, since there can be no working link
* endpoint at the peer end, and the link will re-establish.
* Accept the signature, since it comes from a known peer.
*/
n->signature = signature;
} else if (!sign_match && addr_match && !link_up) {
/* The peer node has rebooted.
* Accept signature, since it is a known peer.
*/
n->signature = signature;
*respond = true;
} else if (!sign_match && !addr_match && link_up) {
/* Peer rebooted with new address, or a new/duplicate peer.
* Ignore until the link goes down, if ever.
*/
*dupl_addr = true;
} else if (!sign_match && !addr_match && !link_up) {
/* Peer rebooted with new address, or it is a new peer.
* Accept signature and address.
*/
n->signature = signature;
accept_addr = true;
*respond = true;
}
if (!accept_addr)
goto exit;
/* Now create new link if not already existing */
if (!l) {
tipc: remove restrictions on node address values Nominally, TIPC organizes network nodes into a three-level network hierarchy consisting of the levels 'zone', 'cluster' and 'node'. This hierarchy is reflected in the node address format, - it is sub-divided into an 8-bit zone id, and 12 bit cluster id, and a 12-bit node id. However, the 'zone' and 'cluster' levels have in reality never been fully implemented,and never will be. The result of this has been that the first 20 bits the node identity structure have been wasted, and the usable node identity range within a cluster has been limited to 12 bits. This is starting to become a problem. In the following commits, we will need to be able to connect between nodes which are using the whole 32-bit value space of the node address. We therefore remove the restrictions on which values can be assigned to node identity, -it is from now on only a 32-bit integer with no assumed internal structure. Isolation between clusters is now achieved only by setting different values for the 'network id' field used during neighbor discovery, in practice leading to the latter becoming the new cluster identity. The rules for accepting discovery requests/responses from neighboring nodes now become: - If the user is using legacy address format on both peers, reception of discovery messages is subject to the legacy lookup domain check in addition to the cluster id check. - Otherwise, the discovery request/response is always accepted, provided both peers have the same network id. This secures backwards compatibility for users who have been using zone or cluster identities as cluster separators, instead of the intended 'network id'. 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>
2018-03-23 02:42:47 +07:00
if (n->link_cnt == 2)
goto exit;
tipc: remove restrictions on node address values Nominally, TIPC organizes network nodes into a three-level network hierarchy consisting of the levels 'zone', 'cluster' and 'node'. This hierarchy is reflected in the node address format, - it is sub-divided into an 8-bit zone id, and 12 bit cluster id, and a 12-bit node id. However, the 'zone' and 'cluster' levels have in reality never been fully implemented,and never will be. The result of this has been that the first 20 bits the node identity structure have been wasted, and the usable node identity range within a cluster has been limited to 12 bits. This is starting to become a problem. In the following commits, we will need to be able to connect between nodes which are using the whole 32-bit value space of the node address. We therefore remove the restrictions on which values can be assigned to node identity, -it is from now on only a 32-bit integer with no assumed internal structure. Isolation between clusters is now achieved only by setting different values for the 'network id' field used during neighbor discovery, in practice leading to the latter becoming the new cluster identity. The rules for accepting discovery requests/responses from neighboring nodes now become: - If the user is using legacy address format on both peers, reception of discovery messages is subject to the legacy lookup domain check in addition to the cluster id check. - Otherwise, the discovery request/response is always accepted, provided both peers have the same network id. This secures backwards compatibility for users who have been using zone or cluster identities as cluster separators, instead of the intended 'network id'. 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>
2018-03-23 02:42:47 +07:00
if_name = strchr(b->name, ':') + 1;
get_random_bytes(&session, sizeof(u16));
if (!tipc_link_create(net, if_name, b->identity, b->tolerance,
b->net_plane, b->mtu, b->priority,
b->window, session,
tipc: handle collisions of 32-bit node address hash values When a 32-bit node address is generated from a 128-bit identifier, there is a risk of collisions which must be discovered and handled. We do this as follows: - We don't apply the generated address immediately to the node, but do instead initiate a 1 sec trial period to allow other cluster members to discover and handle such collisions. - During the trial period the node periodically sends out a new type of message, DSC_TRIAL_MSG, using broadcast or emulated broadcast, to all the other nodes in the cluster. - When a node is receiving such a message, it must check that the presented 32-bit identifier either is unused, or was used by the very same peer in a previous session. In both cases it accepts the request by not responding to it. - If it finds that the same node has been up before using a different address, it responds with a DSC_TRIAL_FAIL_MSG containing that address. - If it finds that the address has already been taken by some other node, it generates a new, unused address and returns it to the requester. - During the trial period the requesting node must always be prepared to accept a failure message, i.e., a message where a peer suggests a different (or equal) address to the one tried. In those cases it must apply the suggested value as trial address and restart the trial period. This algorithm ensures that in the vast majority of cases a node will have the same address before and after a reboot. If a legacy user configures the address explicitly, there will be no trial period and messages, so this protocol addition is completely backwards compatible. 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>
2018-03-23 02:42:51 +07:00
tipc_own_addr(net), addr, peer_id,
n->capabilities,
tipc_bc_sndlink(n->net), n->bc_entry.link,
&le->inputq,
&n->bc_entry.namedq, &l)) {
*respond = false;
goto exit;
}
tipc: add trace_events for tipc link The commit adds the new trace_events for TIPC link object: trace_tipc_link_timeout() trace_tipc_link_fsm() trace_tipc_link_reset() trace_tipc_link_too_silent() trace_tipc_link_retrans() trace_tipc_link_bc_ack() trace_tipc_link_conges() And the traces for PROTOCOL messages at building and receiving: trace_tipc_proto_build() trace_tipc_proto_rcv() Note: a) The 'tipc_link_too_silent' event will only happen when the 'silent_intv_cnt' is about to reach the 'abort_limit' value (and the event is enabled). The benefit for this kind of event is that we can get an early indication about TIPC link loss issue due to timeout, then can do some necessary actions for troubleshooting. For example: To trigger the 'tipc_proto_rcv' when the 'too_silent' event occurs: echo 'enable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_too_silent/trigger And disable it when TIPC link is reset: echo 'disable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_reset/trigger b) The 'tipc_link_retrans' or 'tipc_link_bc_ack' event is useful to trace TIPC retransmission issues. In addition, the commit adds the 'trace_tipc_list/link_dump()' at the 'retransmission failure' case. Then, if the issue occurs, the link 'transmq' along with the link data can be dumped for post-analysis. These dump events should be enabled by default since it will only take effect when the failure happens. The same approach is also applied for the faulty case that the validation of protocol message is failed. Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:57 +07:00
trace_tipc_link_reset(l, TIPC_DUMP_ALL, "link created!");
tipc_link_reset(l);
tipc_link_fsm_evt(l, LINK_RESET_EVT);
tipc: eliminate risk of premature link setup during failover When a link goes down, and there is still a working link towards its destination node, a failover is initiated, and the failed link is not allowed to re-establish until that procedure is finished. To ensure this, the concerned link endpoints are set to state LINK_FAILINGOVER, and the node endpoints to NODE_FAILINGOVER during the failover period. However, if the link reset is due to a disabled bearer, the corres- ponding link endpoint is deleted, and only the node endpoint knows about the ongoing failover. Now, if the disabled bearer is re-enabled during the failover period, the discovery mechanism may create a new link endpoint that is ready to be established, despite that this is not permitted. This situation may cause both the ongoing failover and any subsequent link synchronization to fail. In this commit, we ensure that a newly created link goes directly to state LINK_FAILINGOVER if the corresponding node state is NODE_FAILINGOVER. This eliminates the problem described above. Furthermore, we tighten the criteria for which packets are allowed to end a failover state in the function tipc_node_check_state(). By checking that the receiving link is up and running, instead of just checking that it is not in failover mode, we eliminate the risk that protocol packets from the re-created link may cause the failover to be prematurely terminated. Reviewed-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>
2015-08-20 13:12:54 +07:00
if (n->state == NODE_FAILINGOVER)
tipc_link_fsm_evt(l, LINK_FAILOVER_BEGIN_EVT);
le->link = l;
n->link_cnt++;
tipc_node_calculate_timer(n, l);
if (n->link_cnt == 1) {
intv = jiffies + msecs_to_jiffies(n->keepalive_intv);
if (!mod_timer(&n->timer, intv))
tipc_node_get(n);
}
}
memcpy(&le->maddr, maddr, sizeof(*maddr));
exit:
tipc_node_write_unlock(n);
if (reset && l && !tipc_link_is_reset(l))
tipc_node_link_down(n, b->identity, false);
tipc_node_put(n);
}
void tipc_node_delete_links(struct net *net, int bearer_id)
{
struct tipc_net *tn = net_generic(net, tipc_net_id);
struct tipc_node *n;
rcu_read_lock();
list_for_each_entry_rcu(n, &tn->node_list, list) {
tipc_node_link_down(n, bearer_id, true);
}
rcu_read_unlock();
}
static void tipc_node_reset_links(struct tipc_node *n)
{
int i;
pr_warn("Resetting all links to %x\n", n->addr);
trace_tipc_node_reset_links(n, true, " ");
for (i = 0; i < MAX_BEARERS; i++) {
tipc_node_link_down(n, i, false);
}
}
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
/* tipc_node_fsm_evt - node finite state machine
* Determines when contact is allowed with peer node
*/
static void tipc_node_fsm_evt(struct tipc_node *n, int evt)
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
{
int state = n->state;
switch (state) {
case SELF_DOWN_PEER_DOWN:
switch (evt) {
case SELF_ESTABL_CONTACT_EVT:
state = SELF_UP_PEER_COMING;
break;
case PEER_ESTABL_CONTACT_EVT:
state = SELF_COMING_PEER_UP;
break;
case SELF_LOST_CONTACT_EVT:
case PEER_LOST_CONTACT_EVT:
break;
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_END_EVT:
case NODE_SYNCH_BEGIN_EVT:
case NODE_FAILOVER_BEGIN_EVT:
case NODE_FAILOVER_END_EVT:
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
default:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
goto illegal_evt;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
break;
case SELF_UP_PEER_UP:
switch (evt) {
case SELF_LOST_CONTACT_EVT:
state = SELF_DOWN_PEER_LEAVING;
break;
case PEER_LOST_CONTACT_EVT:
state = SELF_LEAVING_PEER_DOWN;
break;
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_BEGIN_EVT:
state = NODE_SYNCHING;
break;
case NODE_FAILOVER_BEGIN_EVT:
state = NODE_FAILINGOVER;
break;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
case SELF_ESTABL_CONTACT_EVT:
case PEER_ESTABL_CONTACT_EVT:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_END_EVT:
case NODE_FAILOVER_END_EVT:
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
break;
default:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
goto illegal_evt;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
break;
case SELF_DOWN_PEER_LEAVING:
switch (evt) {
case PEER_LOST_CONTACT_EVT:
state = SELF_DOWN_PEER_DOWN;
break;
case SELF_ESTABL_CONTACT_EVT:
case PEER_ESTABL_CONTACT_EVT:
case SELF_LOST_CONTACT_EVT:
break;
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_END_EVT:
case NODE_SYNCH_BEGIN_EVT:
case NODE_FAILOVER_BEGIN_EVT:
case NODE_FAILOVER_END_EVT:
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
default:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
goto illegal_evt;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
break;
case SELF_UP_PEER_COMING:
switch (evt) {
case PEER_ESTABL_CONTACT_EVT:
state = SELF_UP_PEER_UP;
break;
case SELF_LOST_CONTACT_EVT:
tipc: correct error in node fsm commit 88e8ac7000dc ("tipc: reduce transmission rate of reset messages when link is down") revealed a flaw in the node FSM, as defined in the log of commit 66996b6c47ed ("tipc: extend node FSM"). We see the following scenario: 1: Node B receives a RESET message from node A before its link endpoint is fully up, i.e., the node FSM is in state SELF_UP_PEER_COMING. This event will not change the node FSM state, but the (distinct) link FSM will move to state RESETTING. 2: As an effect of the previous event, the local endpoint on B will declare node A lost, and post the event SELF_DOWN to the its node FSM. This moves the FSM state to SELF_DOWN_PEER_LEAVING, meaning that no messages will be accepted from A until it receives another RESET message that confirms that A's endpoint has been reset. This is wasteful, since we know this as a fact already from the first received RESET, but worse is that the link instance's FSM has not wasted this information, but instead moved on to state ESTABLISHING, meaning that it repeatedly sends out ACTIVATE messages to the reset peer A. 3: Node A will receive one of the ACTIVATE messages, move its link FSM to state ESTABLISHED, and start repeatedly sending out STATE messages to node B. 4: Node B will consistently drop these messages, since it can only accept accept a RESET according to its node FSM. 5: After four lost STATE messages node A will reset its link and start repeatedly sending out RESET messages to B. 6: Because of the reduced send rate for RESET messages, it is very likely that A will receive an ACTIVATE (which is sent out at a much higher frequency) before it gets the chance to send a RESET, and A may hence quickly move back to state ESTABLISHED and continue sending out STATE messages, which will again be dropped by B. 7: GOTO 5. 8: After having repeated the cycle 5-7 a number of times, node A will by chance get in between with sending a RESET, and the situation is resolved. Unfortunately, we have seen that it may take a substantial amount of time before this vicious loop is broken, sometimes in the order of minutes. We correct this by making a small correction to the node FSM: When a node in state SELF_UP_PEER_COMING receives a SELF_DOWN event, it now moves directly back to state SELF_DOWN_PEER_DOWN, instead of as now SELF_DOWN_PEER_LEAVING. This is logically consistent, since we don't need to wait for RESET confirmation from of an endpoint that we alread know has been reset. It also means that node B in the scenario above will not be dropping incoming STATE messages, and the link can come up immediately. Finally, a symmetry comparison reveals that the FSM has a similar error when receiving the event PEER_DOWN in state PEER_UP_SELF_COMING. Instead of moving to PERR_DOWN_SELF_LEAVING, it should move directly to SELF_DOWN_PEER_DOWN. Although we have never seen any negative effect of this logical error, we choose fix this one, too. The node FSM looks as follows after those changes: +----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |-----------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_ |FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +-------->| SELF_UP_ |<-------+ | | | | +-----------------| PEER_UP |----------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING| |PEER_COMING| |SELF_COMING| |SELF_LEAVING| +------------+ +-----------+ +-----------+ +------------+ | | A A | | | | | | | | | SELF_ | |SELF_ |PEER_ |PEER_ | | DOWN_EVT| |UP_EVT |UP_EVT |DOWN_EVT | | | | | | | | | | | | | | | +--------------+ | | |PEER_DOWN_EVT +--->| SELF_DOWN_ |<---+ SELF_DOWN_EVT| +------------------->| PEER_DOWN |<--------------------+ +--------------+ 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>
2016-06-08 23:00:04 +07:00
state = SELF_DOWN_PEER_DOWN;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
break;
case SELF_ESTABL_CONTACT_EVT:
case PEER_LOST_CONTACT_EVT:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_END_EVT:
case NODE_FAILOVER_BEGIN_EVT:
tipc: delay ESTABLISH state event when link is established Link establishing, just like link teardown, is a non-atomic action, in the sense that discovering that conditions are right to establish a link, and the actual adding of the link to one of the node's send slots is done in two different lock contexts. The link FSM is designed to help bridging the gap between the two contexts in a safe manner. We have now discovered a weakness in the implementaton of this FSM. Because we directly let the link go from state LINK_ESTABLISHING to state LINK_ESTABLISHED already in the first lock context, we are unable to distinguish between a fully established link, i.e., a link that has been added to its slot, and a link that has not yet reached the second lock context. It may hence happen that a manual intervention, e.g., when disabling an interface, causes the function tipc_node_link_down() to try removing the link from the node slots, decrementing its active link counter etc, although the link was never added there in the first place. We solve this by delaying the actual state change until we reach the second lock context, inside the function tipc_node_link_up(). This makes it possible for potentail callers of __tipc_node_link_down() to know if they should proceed or not, and the problem is solved. Unforunately, the situation described above also has a second problem. Since there by necessity is a tipc_node_link_up() call pending once the node lock has been released, we must defuse that call by setting the link back from LINK_ESTABLISHING to LINK_RESET state. This forces us to make a slight modification to the link FSM, which will now look as follows. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | +----------------+ | | | RESET_EVT| |RESET_EVT | | | | | | | | | | |ESTABLISH_EVT | | | | +-------------+ | | | | | | RESET_EVT | | | | | | | | | | | V V V | | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-16 01:52:44 +07:00
break;
case NODE_SYNCH_BEGIN_EVT:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_FAILOVER_END_EVT:
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
default:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
goto illegal_evt;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
break;
case SELF_COMING_PEER_UP:
switch (evt) {
case SELF_ESTABL_CONTACT_EVT:
state = SELF_UP_PEER_UP;
break;
case PEER_LOST_CONTACT_EVT:
tipc: correct error in node fsm commit 88e8ac7000dc ("tipc: reduce transmission rate of reset messages when link is down") revealed a flaw in the node FSM, as defined in the log of commit 66996b6c47ed ("tipc: extend node FSM"). We see the following scenario: 1: Node B receives a RESET message from node A before its link endpoint is fully up, i.e., the node FSM is in state SELF_UP_PEER_COMING. This event will not change the node FSM state, but the (distinct) link FSM will move to state RESETTING. 2: As an effect of the previous event, the local endpoint on B will declare node A lost, and post the event SELF_DOWN to the its node FSM. This moves the FSM state to SELF_DOWN_PEER_LEAVING, meaning that no messages will be accepted from A until it receives another RESET message that confirms that A's endpoint has been reset. This is wasteful, since we know this as a fact already from the first received RESET, but worse is that the link instance's FSM has not wasted this information, but instead moved on to state ESTABLISHING, meaning that it repeatedly sends out ACTIVATE messages to the reset peer A. 3: Node A will receive one of the ACTIVATE messages, move its link FSM to state ESTABLISHED, and start repeatedly sending out STATE messages to node B. 4: Node B will consistently drop these messages, since it can only accept accept a RESET according to its node FSM. 5: After four lost STATE messages node A will reset its link and start repeatedly sending out RESET messages to B. 6: Because of the reduced send rate for RESET messages, it is very likely that A will receive an ACTIVATE (which is sent out at a much higher frequency) before it gets the chance to send a RESET, and A may hence quickly move back to state ESTABLISHED and continue sending out STATE messages, which will again be dropped by B. 7: GOTO 5. 8: After having repeated the cycle 5-7 a number of times, node A will by chance get in between with sending a RESET, and the situation is resolved. Unfortunately, we have seen that it may take a substantial amount of time before this vicious loop is broken, sometimes in the order of minutes. We correct this by making a small correction to the node FSM: When a node in state SELF_UP_PEER_COMING receives a SELF_DOWN event, it now moves directly back to state SELF_DOWN_PEER_DOWN, instead of as now SELF_DOWN_PEER_LEAVING. This is logically consistent, since we don't need to wait for RESET confirmation from of an endpoint that we alread know has been reset. It also means that node B in the scenario above will not be dropping incoming STATE messages, and the link can come up immediately. Finally, a symmetry comparison reveals that the FSM has a similar error when receiving the event PEER_DOWN in state PEER_UP_SELF_COMING. Instead of moving to PERR_DOWN_SELF_LEAVING, it should move directly to SELF_DOWN_PEER_DOWN. Although we have never seen any negative effect of this logical error, we choose fix this one, too. The node FSM looks as follows after those changes: +----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |-----------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_ |FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +-------->| SELF_UP_ |<-------+ | | | | +-----------------| PEER_UP |----------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING| |PEER_COMING| |SELF_COMING| |SELF_LEAVING| +------------+ +-----------+ +-----------+ +------------+ | | A A | | | | | | | | | SELF_ | |SELF_ |PEER_ |PEER_ | | DOWN_EVT| |UP_EVT |UP_EVT |DOWN_EVT | | | | | | | | | | | | | | | +--------------+ | | |PEER_DOWN_EVT +--->| SELF_DOWN_ |<---+ SELF_DOWN_EVT| +------------------->| PEER_DOWN |<--------------------+ +--------------+ 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>
2016-06-08 23:00:04 +07:00
state = SELF_DOWN_PEER_DOWN;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
break;
case SELF_LOST_CONTACT_EVT:
case PEER_ESTABL_CONTACT_EVT:
break;
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_END_EVT:
case NODE_SYNCH_BEGIN_EVT:
case NODE_FAILOVER_BEGIN_EVT:
case NODE_FAILOVER_END_EVT:
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
default:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
goto illegal_evt;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
break;
case SELF_LEAVING_PEER_DOWN:
switch (evt) {
case SELF_LOST_CONTACT_EVT:
state = SELF_DOWN_PEER_DOWN;
break;
case SELF_ESTABL_CONTACT_EVT:
case PEER_ESTABL_CONTACT_EVT:
case PEER_LOST_CONTACT_EVT:
break;
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
case NODE_SYNCH_END_EVT:
case NODE_SYNCH_BEGIN_EVT:
case NODE_FAILOVER_BEGIN_EVT:
case NODE_FAILOVER_END_EVT:
default:
goto illegal_evt;
}
break;
case NODE_FAILINGOVER:
switch (evt) {
case SELF_LOST_CONTACT_EVT:
state = SELF_DOWN_PEER_LEAVING;
break;
case PEER_LOST_CONTACT_EVT:
state = SELF_LEAVING_PEER_DOWN;
break;
case NODE_FAILOVER_END_EVT:
state = SELF_UP_PEER_UP;
break;
case NODE_FAILOVER_BEGIN_EVT:
case SELF_ESTABL_CONTACT_EVT:
case PEER_ESTABL_CONTACT_EVT:
break;
case NODE_SYNCH_BEGIN_EVT:
case NODE_SYNCH_END_EVT:
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
default:
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
goto illegal_evt;
}
break;
case NODE_SYNCHING:
switch (evt) {
case SELF_LOST_CONTACT_EVT:
state = SELF_DOWN_PEER_LEAVING;
break;
case PEER_LOST_CONTACT_EVT:
state = SELF_LEAVING_PEER_DOWN;
break;
case NODE_SYNCH_END_EVT:
state = SELF_UP_PEER_UP;
break;
case NODE_FAILOVER_BEGIN_EVT:
state = NODE_FAILINGOVER;
break;
case NODE_SYNCH_BEGIN_EVT:
case SELF_ESTABL_CONTACT_EVT:
case PEER_ESTABL_CONTACT_EVT:
break;
case NODE_FAILOVER_END_EVT:
default:
goto illegal_evt;
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
break;
default:
pr_err("Unknown node fsm state %x\n", state);
break;
}
trace_tipc_node_fsm(n->peer_id, n->state, state, evt);
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
n->state = state;
tipc: extend node FSM In the next commit, we will move link synch/failover orchestration to the link aggregation level. In order to do this, we first need to extend the node FSM with two more states, NODE_SYNCHING and NODE_FAILINGOVER, plus four new events to enter and leave those states. This commit introduces this change, without yet making use of it. The node FSM now looks as follows: +-----------------------------------------+ | PEER_DOWN_EVT| | | +------------------------+----------------+ | |SELF_DOWN_EVT | | | | | | | | +-----------+ +-----------+ | | |NODE_ | |NODE_ | | | +----------|FAILINGOVER|<---------|SYNCHING |------------+ | | |SELF_ +-----------+ FAILOVER_+-----------+ PEER_ | | | |DOWN_EVT | A BEGIN_EVT A | DOWN_EVT| | | | | | | | | | | | | | | | | | | | |FAILOVER_|FAILOVER_ |SYNCH_ |SYNCH_ | | | | |END_EVT |BEGIN_EVT |BEGIN_EVT|END_EVT | | | | | | | | | | | | | | | | | | | | | +--------------+ | | | | | +------->| SELF_UP_ |<-------+ | | | | +----------------| PEER_UP |------------------+ | | | | |SELF_DOWN_EVT +--------------+ PEER_DOWN_EVT| | | | | | A A | | | | | | | | | | | | | | PEER_UP_EVT| |SELF_UP_EVT | | | | | | | | | | | V V V | | V V V +------------+ +-----------+ +-----------+ +------------+ |SELF_DOWN_ | |SELF_UP_ | |PEER_UP_ | |PEER_DOWN | |PEER_LEAVING|<------|PEER_COMING| |SELF_COMING|------>|SELF_LEAVING| +------------+ SELF_ +-----------+ +-----------+ PEER_ +------------+ | DOWN_EVT A A DOWN_EVT | | | | | | | | | | SELF_UP_EVT| |PEER_UP_EVT | | | | | | | | | |PEER_DOWN_EVT +--------------+ SELF_DOWN_EVT| +------------------->| SELF_DOWN_ |<--------------------+ | PEER_DOWN | +--------------+ Tested-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>
2015-07-31 05:24:18 +07:00
return;
illegal_evt:
pr_err("Illegal node fsm evt %x in state %x\n", evt, state);
trace_tipc_node_fsm(n->peer_id, n->state, state, evt);
tipc: introduce node contact FSM The logics for determining when a node is permitted to establish and maintain contact with its peer node becomes non-trivial in the presence of multiple parallel links that may come and go independently. A known failure scenario is that one endpoint registers both its links to the peer lost, cleans up it binding table, and prepares for a table update once contact is re-establihed, while the other endpoint may see its links reset and re-established one by one, hence seeing no need to re-synchronize the binding table. To avoid this, a node must not allow re-establishing contact until it has confirmation that even the peer has lost both links. Currently, the mechanism for handling this consists of setting and resetting two state flags from different locations in the code. This solution is hard to understand and maintain. A closer analysis even reveals that it is not completely safe. In this commit we do instead introduce an FSM that keeps track of the conditions for when the node can establish and maintain links. It has six states and four events, and is strictly based on explicit knowledge about the own node's and the peer node's contact states. Only events leading to state change are shown as edges in the figure below. +--------------+ | SELF_UP/ | +---------------->| PEER_COMING |-----------------+ SELF_ | +--------------+ |PEER_ ESTBL_ | | |ESTBL_ CONTACT| SELF_LOST_CONTACT | |CONTACT | v | | +--------------+ | | PEER_ | SELF_DOWN/ | SELF_ | | LOST_ +--| PEER_LEAVING |<--+ LOST_ v +-------------+ CONTACT | +--------------+ | CONTACT +-----------+ | SELF_DOWN/ |<----------+ +----------| SELF_UP/ | | PEER_DOWN |<----------+ +----------| PEER_UP | +-------------+ SELF_ | +--------------+ | PEER_ +-----------+ | LOST_ +--| SELF_LEAVING/|<--+ LOST_ A | CONTACT | PEER_DOWN | CONTACT | | +--------------+ | | A | PEER_ | PEER_LOST_CONTACT | |SELF_ ESTBL_ | | |ESTBL_ CONTACT| +--------------+ |CONTACT +---------------->| PEER_UP/ |-----------------+ | SELF_COMING | +--------------+ Reviewed-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>
2015-07-17 03:54:30 +07:00
}
static void node_lost_contact(struct tipc_node *n,
struct sk_buff_head *inputq)
{
struct tipc_sock_conn *conn, *safe;
struct tipc_link *l;
struct list_head *conns = &n->conn_sks;
struct sk_buff *skb;
uint i;
pr_debug("Lost contact with %x\n", n->addr);
n->delete_at = jiffies + msecs_to_jiffies(NODE_CLEANUP_AFTER);
trace_tipc_node_lost_contact(n, true, " ");
/* Clean up broadcast state */
tipc_bcast_remove_peer(n->net, n->bc_entry.link);
/* Abort any ongoing link failover */
for (i = 0; i < MAX_BEARERS; i++) {
l = n->links[i].link;
if (l)
tipc_link_fsm_evt(l, LINK_FAILOVER_END_EVT);
}
/* Notify publications from this node */
n->action_flags |= TIPC_NOTIFY_NODE_DOWN;
n->peer_net = NULL;
n->peer_hash_mix = 0;
/* Notify sockets connected to node */
list_for_each_entry_safe(conn, safe, conns, list) {
skb = tipc_msg_create(TIPC_CRITICAL_IMPORTANCE, TIPC_CONN_MSG,
SHORT_H_SIZE, 0, tipc_own_addr(n->net),
conn->peer_node, conn->port,
conn->peer_port, TIPC_ERR_NO_NODE);
if (likely(skb))
skb_queue_tail(inputq, skb);
list_del(&conn->list);
kfree(conn);
}
}
/**
* tipc_node_get_linkname - get the name of a link
*
* @bearer_id: id of the bearer
* @node: peer node address
* @linkname: link name output buffer
*
* Returns 0 on success
*/
int tipc_node_get_linkname(struct net *net, u32 bearer_id, u32 addr,
char *linkname, size_t len)
{
struct tipc_link *link;
int err = -EINVAL;
struct tipc_node *node = tipc_node_find(net, addr);
if (!node)
return err;
if (bearer_id >= MAX_BEARERS)
goto exit;
tipc_node_read_lock(node);
link = node->links[bearer_id].link;
if (link) {
strncpy(linkname, tipc_link_name(link), len);
err = 0;
}
tipc_node_read_unlock(node);
exit:
tipc_node_put(node);
return err;
}
/* Caller should hold node lock for the passed node */
static int __tipc_nl_add_node(struct tipc_nl_msg *msg, struct tipc_node *node)
{
void *hdr;
struct nlattr *attrs;
hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
NLM_F_MULTI, TIPC_NL_NODE_GET);
if (!hdr)
return -EMSGSIZE;
attrs = nla_nest_start_noflag(msg->skb, TIPC_NLA_NODE);
if (!attrs)
goto msg_full;
if (nla_put_u32(msg->skb, TIPC_NLA_NODE_ADDR, node->addr))
goto attr_msg_full;
if (node_is_up(node))
if (nla_put_flag(msg->skb, TIPC_NLA_NODE_UP))
goto attr_msg_full;
nla_nest_end(msg->skb, attrs);
genlmsg_end(msg->skb, hdr);
return 0;
attr_msg_full:
nla_nest_cancel(msg->skb, attrs);
msg_full:
genlmsg_cancel(msg->skb, hdr);
return -EMSGSIZE;
}
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
static void tipc_lxc_xmit(struct net *peer_net, struct sk_buff_head *list)
{
struct tipc_msg *hdr = buf_msg(skb_peek(list));
struct sk_buff_head inputq;
switch (msg_user(hdr)) {
case TIPC_LOW_IMPORTANCE:
case TIPC_MEDIUM_IMPORTANCE:
case TIPC_HIGH_IMPORTANCE:
case TIPC_CRITICAL_IMPORTANCE:
if (msg_connected(hdr) || msg_named(hdr)) {
tipc_loopback_trace(peer_net, list);
spin_lock_init(&list->lock);
tipc_sk_rcv(peer_net, list);
return;
}
if (msg_mcast(hdr)) {
tipc_loopback_trace(peer_net, list);
skb_queue_head_init(&inputq);
tipc_sk_mcast_rcv(peer_net, list, &inputq);
__skb_queue_purge(list);
skb_queue_purge(&inputq);
return;
}
return;
case MSG_FRAGMENTER:
if (tipc_msg_assemble(list)) {
tipc_loopback_trace(peer_net, list);
skb_queue_head_init(&inputq);
tipc_sk_mcast_rcv(peer_net, list, &inputq);
__skb_queue_purge(list);
skb_queue_purge(&inputq);
}
return;
case GROUP_PROTOCOL:
case CONN_MANAGER:
tipc_loopback_trace(peer_net, list);
spin_lock_init(&list->lock);
tipc_sk_rcv(peer_net, list);
return;
case LINK_PROTOCOL:
case NAME_DISTRIBUTOR:
case TUNNEL_PROTOCOL:
case BCAST_PROTOCOL:
return;
default:
return;
};
}
/**
* tipc_node_xmit() is the general link level function for message sending
* @net: the applicable net namespace
* @list: chain of buffers containing message
* @dnode: address of destination node
* @selector: a number used for deterministic link selection
tipc: reduce risk of user starvation during link congestion The socket code currently handles link congestion by either blocking and trying to send again when the congestion has abated, or just returning to the user with -EAGAIN and let him re-try later. This mechanism is prone to starvation, because the wakeup algorithm is non-atomic. During the time the link issues a wakeup signal, until the socket wakes up and re-attempts sending, other senders may have come in between and occupied the free buffer space in the link. This in turn may lead to a socket having to make many send attempts before it is successful. In extremely loaded systems we have observed latency times of several seconds before a low-priority socket is able to send out a message. In this commit, we simplify this mechanism and reduce the risk of the described scenario happening. When a message is attempted sent via a congested link, we now let it be added to the link's backlog queue anyway, thus permitting an oversubscription of one message per source socket. We still create a wakeup item and return an error code, hence instructing the sender to block or stop sending. Only when enough space has been freed up in the link's backlog queue do we issue a wakeup event that allows the sender to continue with the next message, if any. The fact that a socket now can consider a message sent even when the link returns a congestion code means that the sending socket code can be simplified. Also, since this is a good opportunity to get rid of the obsolete 'mtu change' condition in the three socket send functions, we now choose to refactor those functions completely. Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.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>
2017-01-03 22:55:11 +07:00
* Consumes the buffer chain.
* Returns 0 if success, otherwise: -ELINKCONG,-EHOSTUNREACH,-EMSGSIZE,-ENOBUF
*/
int tipc_node_xmit(struct net *net, struct sk_buff_head *list,
u32 dnode, int selector)
{
struct tipc_link_entry *le = NULL;
struct tipc_node *n;
struct sk_buff_head xmitq;
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
bool node_up = false;
int bearer_id;
int rc;
if (in_own_node(net, dnode)) {
tipc_loopback_trace(net, list);
tipc: clean up skb list lock handling on send path 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 d12cffe9329f ("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>
2019-08-15 21:42:50 +07:00
spin_lock_init(&list->lock);
tipc_sk_rcv(net, list);
return 0;
}
n = tipc_node_find(net, dnode);
if (unlikely(!n)) {
tipc: clean up skb list lock handling on send path 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 d12cffe9329f ("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>
2019-08-15 21:42:50 +07:00
__skb_queue_purge(list);
return -EHOSTUNREACH;
}
tipc_node_read_lock(n);
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
node_up = node_is_up(n);
if (node_up && n->peer_net && check_net(n->peer_net)) {
/* xmit inner linux container */
tipc_lxc_xmit(n->peer_net, list);
if (likely(skb_queue_empty(list))) {
tipc_node_read_unlock(n);
tipc_node_put(n);
return 0;
}
}
bearer_id = n->active_links[selector & 1];
if (unlikely(bearer_id == INVALID_BEARER_ID)) {
tipc_node_read_unlock(n);
tipc_node_put(n);
tipc: clean up skb list lock handling on send path 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 d12cffe9329f ("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>
2019-08-15 21:42:50 +07:00
__skb_queue_purge(list);
return -EHOSTUNREACH;
}
__skb_queue_head_init(&xmitq);
le = &n->links[bearer_id];
spin_lock_bh(&le->lock);
rc = tipc_link_xmit(le->link, list, &xmitq);
spin_unlock_bh(&le->lock);
tipc_node_read_unlock(n);
tipc: reduce risk of user starvation during link congestion The socket code currently handles link congestion by either blocking and trying to send again when the congestion has abated, or just returning to the user with -EAGAIN and let him re-try later. This mechanism is prone to starvation, because the wakeup algorithm is non-atomic. During the time the link issues a wakeup signal, until the socket wakes up and re-attempts sending, other senders may have come in between and occupied the free buffer space in the link. This in turn may lead to a socket having to make many send attempts before it is successful. In extremely loaded systems we have observed latency times of several seconds before a low-priority socket is able to send out a message. In this commit, we simplify this mechanism and reduce the risk of the described scenario happening. When a message is attempted sent via a congested link, we now let it be added to the link's backlog queue anyway, thus permitting an oversubscription of one message per source socket. We still create a wakeup item and return an error code, hence instructing the sender to block or stop sending. Only when enough space has been freed up in the link's backlog queue do we issue a wakeup event that allows the sender to continue with the next message, if any. The fact that a socket now can consider a message sent even when the link returns a congestion code means that the sending socket code can be simplified. Also, since this is a good opportunity to get rid of the obsolete 'mtu change' condition in the three socket send functions, we now choose to refactor those functions completely. Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.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>
2017-01-03 22:55:11 +07:00
if (unlikely(rc == -ENOBUFS))
tipc_node_link_down(n, bearer_id, false);
tipc: reduce risk of user starvation during link congestion The socket code currently handles link congestion by either blocking and trying to send again when the congestion has abated, or just returning to the user with -EAGAIN and let him re-try later. This mechanism is prone to starvation, because the wakeup algorithm is non-atomic. During the time the link issues a wakeup signal, until the socket wakes up and re-attempts sending, other senders may have come in between and occupied the free buffer space in the link. This in turn may lead to a socket having to make many send attempts before it is successful. In extremely loaded systems we have observed latency times of several seconds before a low-priority socket is able to send out a message. In this commit, we simplify this mechanism and reduce the risk of the described scenario happening. When a message is attempted sent via a congested link, we now let it be added to the link's backlog queue anyway, thus permitting an oversubscription of one message per source socket. We still create a wakeup item and return an error code, hence instructing the sender to block or stop sending. Only when enough space has been freed up in the link's backlog queue do we issue a wakeup event that allows the sender to continue with the next message, if any. The fact that a socket now can consider a message sent even when the link returns a congestion code means that the sending socket code can be simplified. Also, since this is a good opportunity to get rid of the obsolete 'mtu change' condition in the three socket send functions, we now choose to refactor those functions completely. Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.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>
2017-01-03 22:55:11 +07:00
else
tipc_bearer_xmit(net, bearer_id, &xmitq, &le->maddr);
tipc_node_put(n);
return rc;
}
/* tipc_node_xmit_skb(): send single buffer to destination
* Buffers sent via this functon are generally TIPC_SYSTEM_IMPORTANCE
* messages, which will not be rejected
* The only exception is datagram messages rerouted after secondary
* lookup, which are rare and safe to dispose of anyway.
*/
int tipc_node_xmit_skb(struct net *net, struct sk_buff *skb, u32 dnode,
u32 selector)
{
struct sk_buff_head head;
tipc: clean up skb list lock handling on send path 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 d12cffe9329f ("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>
2019-08-15 21:42:50 +07:00
__skb_queue_head_init(&head);
__skb_queue_tail(&head, skb);
tipc: reduce risk of user starvation during link congestion The socket code currently handles link congestion by either blocking and trying to send again when the congestion has abated, or just returning to the user with -EAGAIN and let him re-try later. This mechanism is prone to starvation, because the wakeup algorithm is non-atomic. During the time the link issues a wakeup signal, until the socket wakes up and re-attempts sending, other senders may have come in between and occupied the free buffer space in the link. This in turn may lead to a socket having to make many send attempts before it is successful. In extremely loaded systems we have observed latency times of several seconds before a low-priority socket is able to send out a message. In this commit, we simplify this mechanism and reduce the risk of the described scenario happening. When a message is attempted sent via a congested link, we now let it be added to the link's backlog queue anyway, thus permitting an oversubscription of one message per source socket. We still create a wakeup item and return an error code, hence instructing the sender to block or stop sending. Only when enough space has been freed up in the link's backlog queue do we issue a wakeup event that allows the sender to continue with the next message, if any. The fact that a socket now can consider a message sent even when the link returns a congestion code means that the sending socket code can be simplified. Also, since this is a good opportunity to get rid of the obsolete 'mtu change' condition in the three socket send functions, we now choose to refactor those functions completely. Signed-off-by: Parthasarathy Bhuvaragan <parthasarathy.bhuvaragan@ericsson.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>
2017-01-03 22:55:11 +07:00
tipc_node_xmit(net, &head, dnode, selector);
return 0;
}
/* tipc_node_distr_xmit(): send single buffer msgs to individual destinations
* Note: this is only for SYSTEM_IMPORTANCE messages, which cannot be rejected
*/
int tipc_node_distr_xmit(struct net *net, struct sk_buff_head *xmitq)
{
struct sk_buff *skb;
u32 selector, dnode;
while ((skb = __skb_dequeue(xmitq))) {
selector = msg_origport(buf_msg(skb));
dnode = msg_destnode(buf_msg(skb));
tipc_node_xmit_skb(net, skb, dnode, selector);
}
return 0;
}
void tipc_node_broadcast(struct net *net, struct sk_buff *skb)
{
struct sk_buff *txskb;
struct tipc_node *n;
u32 dst;
rcu_read_lock();
list_for_each_entry_rcu(n, tipc_nodes(net), list) {
dst = n->addr;
if (in_own_node(net, dst))
continue;
if (!node_is_up(n))
continue;
txskb = pskb_copy(skb, GFP_ATOMIC);
if (!txskb)
break;
msg_set_destnode(buf_msg(txskb), dst);
tipc_node_xmit_skb(net, txskb, dst, 0);
}
rcu_read_unlock();
kfree_skb(skb);
}
static void tipc_node_mcast_rcv(struct tipc_node *n)
{
struct tipc_bclink_entry *be = &n->bc_entry;
/* 'arrvq' is under inputq2's lock protection */
spin_lock_bh(&be->inputq2.lock);
spin_lock_bh(&be->inputq1.lock);
skb_queue_splice_tail_init(&be->inputq1, &be->arrvq);
spin_unlock_bh(&be->inputq1.lock);
spin_unlock_bh(&be->inputq2.lock);
tipc_sk_mcast_rcv(n->net, &be->arrvq, &be->inputq2);
}
static void tipc_node_bc_sync_rcv(struct tipc_node *n, struct tipc_msg *hdr,
int bearer_id, struct sk_buff_head *xmitq)
{
struct tipc_link *ucl;
int rc;
rc = tipc_bcast_sync_rcv(n->net, n->bc_entry.link, hdr);
if (rc & TIPC_LINK_DOWN_EVT) {
tipc_node_reset_links(n);
return;
}
if (!(rc & TIPC_LINK_SND_STATE))
return;
/* If probe message, a STATE response will be sent anyway */
if (msg_probe(hdr))
return;
/* Produce a STATE message carrying broadcast NACK */
tipc_node_read_lock(n);
ucl = n->links[bearer_id].link;
if (ucl)
tipc_link_build_state_msg(ucl, xmitq);
tipc_node_read_unlock(n);
}
/**
* tipc_node_bc_rcv - process TIPC broadcast packet arriving from off-node
* @net: the applicable net namespace
* @skb: TIPC packet
* @bearer_id: id of bearer message arrived on
*
* Invoked with no locks held.
*/
static void tipc_node_bc_rcv(struct net *net, struct sk_buff *skb, int bearer_id)
{
int rc;
struct sk_buff_head xmitq;
struct tipc_bclink_entry *be;
struct tipc_link_entry *le;
struct tipc_msg *hdr = buf_msg(skb);
int usr = msg_user(hdr);
u32 dnode = msg_destnode(hdr);
struct tipc_node *n;
__skb_queue_head_init(&xmitq);
/* If NACK for other node, let rcv link for that node peek into it */
if ((usr == BCAST_PROTOCOL) && (dnode != tipc_own_addr(net)))
n = tipc_node_find(net, dnode);
else
n = tipc_node_find(net, msg_prevnode(hdr));
if (!n) {
kfree_skb(skb);
return;
}
be = &n->bc_entry;
le = &n->links[bearer_id];
rc = tipc_bcast_rcv(net, be->link, skb);
/* Broadcast ACKs are sent on a unicast link */
if (rc & TIPC_LINK_SND_STATE) {
tipc_node_read_lock(n);
tipc_link_build_state_msg(le->link, &xmitq);
tipc_node_read_unlock(n);
}
if (!skb_queue_empty(&xmitq))
tipc_bearer_xmit(net, bearer_id, &xmitq, &le->maddr);
if (!skb_queue_empty(&be->inputq1))
tipc_node_mcast_rcv(n);
/* Handle NAME_DISTRIBUTOR messages sent from 1.7 nodes */
if (!skb_queue_empty(&n->bc_entry.namedq))
tipc_named_rcv(net, &n->bc_entry.namedq);
/* If reassembly or retransmission failure => reset all links to peer */
if (rc & TIPC_LINK_DOWN_EVT)
tipc_node_reset_links(n);
tipc_node_put(n);
}
/**
* tipc_node_check_state - check and if necessary update node state
* @skb: TIPC packet
* @bearer_id: identity of bearer delivering the packet
* Returns true if state and msg are ok, otherwise false
*/
static bool tipc_node_check_state(struct tipc_node *n, struct sk_buff *skb,
tipc: merge link->exec_mode and link->state into one FSM Until now, we have been handling link failover and synchronization by using an additional link state variable, "exec_mode". This variable is not independent of the link FSM state, something causing a risk of inconsistencies, apart from the fact that it clutters the code. The conditions are now in place to define a new link FSM that covers all existing use cases, including failover and synchronization, and eliminate the "exec_mode" field altogether. The FSM must also support non-atomic resetting of links, which will be introduced later. The new link FSM is shown below, with 7 states and 8 events. Only events leading to state change are shown as edges. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | | | | | | +--------------+ | | | RESET_EVT| |RESET_EVT |ESTABLISH_EVT | | | | | | | | | | | | V V | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ These changes are fully backwards compatible. Tested-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>
2015-07-31 05:24:21 +07:00
int bearer_id, struct sk_buff_head *xmitq)
{
struct tipc_msg *hdr = buf_msg(skb);
int usr = msg_user(hdr);
int mtyp = msg_type(hdr);
u16 oseqno = msg_seqno(hdr);
u16 exp_pkts = msg_msgcnt(hdr);
u16 rcv_nxt, syncpt, dlv_nxt, inputq_len;
int state = n->state;
tipc: fix stale link problem during synchronization Recent changes to the link synchronization means that we can now just drop packets arriving on the synchronizing link before the synch point is reached. This has lead to significant simplifications to the implementation, but also turns out to have a flip side that we need to consider. Under unlucky circumstances, the two endpoints may end up repeatedly dropping each other's packets, while immediately asking for retransmission of the same packets, just to drop them once more. This pattern will eventually be broken when the synch point is reached on the other link, but before that, the endpoints may have arrived at the retransmission limit (stale counter) that indicates that the link should be broken. We see this happen at rare occasions. The fix for this is to not ask for retransmissions when a link is in state LINK_SYNCHING. The fact that the link has reached this state means that it has already received the first SYNCH packet, and that it knows the synch point. Hence, it doesn't need any more packets until the other link has reached the synch point, whereafter it can go ahead and ask for the missing packets. However, because of the reduced traffic on the synching link that follows this change, it may now take longer to discover that the synch point has been reached. We compensate for this by letting all packets, on any of the links, trig a check for synchronization termination. This is possible because the packets themselves don't contain any information that is needed for discovering this condition. Reviewed-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>
2015-08-20 13:12:56 +07:00
struct tipc_link *l, *tnl, *pl = NULL;
struct tipc_media_addr *maddr;
int pb_id;
if (trace_tipc_node_check_state_enabled()) {
trace_tipc_skb_dump(skb, false, "skb for node state check");
trace_tipc_node_check_state(n, true, " ");
}
l = n->links[bearer_id].link;
if (!l)
return false;
rcv_nxt = tipc_link_rcv_nxt(l);
if (likely((state == SELF_UP_PEER_UP) && (usr != TUNNEL_PROTOCOL)))
return true;
/* Find parallel link, if any */
for (pb_id = 0; pb_id < MAX_BEARERS; pb_id++) {
if ((pb_id != bearer_id) && n->links[pb_id].link) {
pl = n->links[pb_id].link;
break;
}
}
tipc: add trace_events for tipc link The commit adds the new trace_events for TIPC link object: trace_tipc_link_timeout() trace_tipc_link_fsm() trace_tipc_link_reset() trace_tipc_link_too_silent() trace_tipc_link_retrans() trace_tipc_link_bc_ack() trace_tipc_link_conges() And the traces for PROTOCOL messages at building and receiving: trace_tipc_proto_build() trace_tipc_proto_rcv() Note: a) The 'tipc_link_too_silent' event will only happen when the 'silent_intv_cnt' is about to reach the 'abort_limit' value (and the event is enabled). The benefit for this kind of event is that we can get an early indication about TIPC link loss issue due to timeout, then can do some necessary actions for troubleshooting. For example: To trigger the 'tipc_proto_rcv' when the 'too_silent' event occurs: echo 'enable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_too_silent/trigger And disable it when TIPC link is reset: echo 'disable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_reset/trigger b) The 'tipc_link_retrans' or 'tipc_link_bc_ack' event is useful to trace TIPC retransmission issues. In addition, the commit adds the 'trace_tipc_list/link_dump()' at the 'retransmission failure' case. Then, if the issue occurs, the link 'transmq' along with the link data can be dumped for post-analysis. These dump events should be enabled by default since it will only take effect when the failure happens. The same approach is also applied for the faulty case that the validation of protocol message is failed. Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:57 +07:00
if (!tipc_link_validate_msg(l, hdr)) {
trace_tipc_skb_dump(skb, false, "PROTO invalid (2)!");
trace_tipc_link_dump(l, TIPC_DUMP_NONE, "PROTO invalid (2)!");
return false;
tipc: add trace_events for tipc link The commit adds the new trace_events for TIPC link object: trace_tipc_link_timeout() trace_tipc_link_fsm() trace_tipc_link_reset() trace_tipc_link_too_silent() trace_tipc_link_retrans() trace_tipc_link_bc_ack() trace_tipc_link_conges() And the traces for PROTOCOL messages at building and receiving: trace_tipc_proto_build() trace_tipc_proto_rcv() Note: a) The 'tipc_link_too_silent' event will only happen when the 'silent_intv_cnt' is about to reach the 'abort_limit' value (and the event is enabled). The benefit for this kind of event is that we can get an early indication about TIPC link loss issue due to timeout, then can do some necessary actions for troubleshooting. For example: To trigger the 'tipc_proto_rcv' when the 'too_silent' event occurs: echo 'enable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_too_silent/trigger And disable it when TIPC link is reset: echo 'disable_event:tipc:tipc_proto_rcv' > \ events/tipc/tipc_link_reset/trigger b) The 'tipc_link_retrans' or 'tipc_link_bc_ack' event is useful to trace TIPC retransmission issues. In addition, the commit adds the 'trace_tipc_list/link_dump()' at the 'retransmission failure' case. Then, if the issue occurs, the link 'transmq' along with the link data can be dumped for post-analysis. These dump events should be enabled by default since it will only take effect when the failure happens. The same approach is also applied for the faulty case that the validation of protocol message is failed. Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:57 +07:00
}
/* Check and update node accesibility if applicable */
if (state == SELF_UP_PEER_COMING) {
if (!tipc_link_is_up(l))
return true;
if (!msg_peer_link_is_up(hdr))
return true;
tipc_node_fsm_evt(n, PEER_ESTABL_CONTACT_EVT);
}
if (state == SELF_DOWN_PEER_LEAVING) {
if (msg_peer_node_is_up(hdr))
return false;
tipc_node_fsm_evt(n, PEER_LOST_CONTACT_EVT);
2015-11-20 02:30:41 +07:00
return true;
}
if (state == SELF_LEAVING_PEER_DOWN)
return false;
/* Ignore duplicate packets */
tipc: eliminate risk of stalled link synchronization In commit 6e498158a827 ("tipc: move link synch and failover to link aggregation level") we introduced a new mechanism for performing link failover and synchronization. We have now detected a bug in this mechanism. During link synchronization we use the arrival of any packet on the tunnel link to trig a check for whether it has reached the synchronization point or not. This has turned out to be too permissive, since it may cause an arriving non-last SYNCH packet to end the synch state, just to see the next SYNCH packet initiate a new synch state with a new, higher synch point. This is not fatal, but should be avoided, because it may significantly extend the synchronization period, while at the same time we are not allowed to send NACKs if packets are lost. In the worst case, a low-traffic user may see its traffic stall until a LINK_PROTOCOL state message trigs the link to leave synchronization state. At the same time, LINK_PROTOCOL packets which happen to have a (non- valid) sequence number lower than the tunnel link's rcv_nxt value will be consistently dropped, and will never be able to resolve the situation described above. We fix this by exempting LINK_PROTOCOL packets from the sequence number check, as they should be. We also reduce (but don't completely eliminate) the risk of entering multiple synchronization states by only allowing the (logically) first SYNCH packet to initiate a synchronization state. This works independently of actual packet arrival order. Fixes: commit 6e498158a827 ("tipc: move link synch and failover to link aggregation level") Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-13 23:41:51 +07:00
if ((usr != LINK_PROTOCOL) && less(oseqno, rcv_nxt))
return true;
/* Initiate or update failover mode if applicable */
if ((usr == TUNNEL_PROTOCOL) && (mtyp == FAILOVER_MSG)) {
syncpt = oseqno + exp_pkts - 1;
tipc: fix issues with early FAILOVER_MSG from peer It appears that a FAILOVER_MSG can come from peer even when the failure link is resetting (i.e. just after the 'node_write_unlock()'...). This means the failover procedure on the node has not been started yet. The situation is as follows: node1 node2 linkb linka linka linkb | | | | | | x failure | | | RESETTING | | | | | | x failure RESET | | RESETTING FAILINGOVER | | | (FAILOVER_MSG) | | |<-------------------------------------------------| | *FAILINGOVER | | | | | (dummy FAILOVER_MSG) | | |------------------------------------------------->| | RESET | | FAILOVER_END | FAILINGOVER RESET | . . . . . . . . . . . . Once this happens, the link failover procedure will be triggered wrongly on the receiving node since the node isn't in FAILINGOVER state but then another link failover will be carried out. The consequences are: 1) A peer might get stuck in FAILINGOVER state because the 'sync_point' was set, reset and set incorrectly, the criteria to end the failover would not be met, it could keep waiting for a message that has already received. 2) The early FAILOVER_MSG(s) could be queued in the link failover deferdq but would be purged or not pulled out because the 'drop_point' was not set correctly. 3) The early FAILOVER_MSG(s) could be dropped too. 4) The dummy FAILOVER_MSG could make the peer leaving FAILINGOVER state shortly, but later on it would be restarted. The same situation can also happen when the link is in PEER_RESET state and a FAILOVER_MSG arrives. The commit resolves the issues by forcing the link down immediately, so the failover procedure will be started normally (which is the same as when receiving a FAILOVER_MSG and the link is in up state). Also, the function "tipc_node_link_failover()" is toughen to avoid such a situation from happening. Acked-by: Jon Maloy <jon.maloy@ericsson.se> Signed-off-by: Tuong Lien <tuong.t.lien@dektech.com.au> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-06-17 11:56:12 +07:00
if (pl && !tipc_link_is_reset(pl)) {
__tipc_node_link_down(n, &pb_id, xmitq, &maddr);
trace_tipc_node_link_down(n, true,
"node link down <- failover!");
tipc_skb_queue_splice_tail_init(tipc_link_inputq(pl),
tipc_link_inputq(l));
}
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
/* If parallel link was already down, and this happened before
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
* the tunnel link came up, node failover was never started.
* Ensure that a FAILOVER_MSG is sent to get peer out of
* NODE_FAILINGOVER state, also this node must accept
* TUNNEL_MSGs from peer.
*/
tipc: fix missing Name entries due to half-failover TIPC link can temporarily fall into "half-establish" that only one of the link endpoints is ESTABLISHED and starts to send traffic, PROTOCOL messages, whereas the other link endpoint is not up (e.g. immediately when the endpoint receives ACTIVATE_MSG, the network interface goes down...). This is a normal situation and will be settled because the link endpoint will be eventually brought down after the link tolerance time. However, the situation will become worse when the second link is established before the first link endpoint goes down, For example: 1. Both links <1A-2A>, <1B-2B> down 2. Link endpoint 2A up, but 1A still down (e.g. due to network disturbance, wrong session, etc.) 3. Link <1B-2B> up 4. Link endpoint 2A down (e.g. due to link tolerance timeout) 5. Node B starts failover onto link <1B-2B> ==> Node A does never start link failover. When the "half-failover" situation happens, two consequences have been observed: a) Peer link/node gets stuck in FAILINGOVER state; b) Traffic or user messages that peer node is trying to failover onto the second link can be partially or completely dropped by this node. The consequence a) was actually solved by commit c140eb166d68 ("tipc: fix failover problem"), but that commit didn't cover the b). It's due to the fact that the tunnel link endpoint has never been prepared for a failover, so the 'l->drop_point' (and the other data...) is not set correctly. When a TUNNEL_MSG from peer node arrives on the link, depending on the inner message's seqno and the current 'l->drop_point' value, the message can be dropped (- treated as a duplicate message) or processed. At this early stage, the traffic messages from peer are likely to be NAME_DISTRIBUTORs, this means some name table entries will be missed on the node forever! The commit resolves the issue by starting the FAILOVER process on this node as well. Another benefit from this solution is that we ensure the link will not be re-established until the failover ends. 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>
2019-05-02 17:23:23 +07:00
if (n->state != NODE_FAILINGOVER)
tipc_node_link_failover(n, pl, l, xmitq);
/* If pkts arrive out of order, use lowest calculated syncpt */
if (less(syncpt, n->sync_point))
n->sync_point = syncpt;
}
/* Open parallel link when tunnel link reaches synch point */
tipc: eliminate risk of premature link setup during failover When a link goes down, and there is still a working link towards its destination node, a failover is initiated, and the failed link is not allowed to re-establish until that procedure is finished. To ensure this, the concerned link endpoints are set to state LINK_FAILINGOVER, and the node endpoints to NODE_FAILINGOVER during the failover period. However, if the link reset is due to a disabled bearer, the corres- ponding link endpoint is deleted, and only the node endpoint knows about the ongoing failover. Now, if the disabled bearer is re-enabled during the failover period, the discovery mechanism may create a new link endpoint that is ready to be established, despite that this is not permitted. This situation may cause both the ongoing failover and any subsequent link synchronization to fail. In this commit, we ensure that a newly created link goes directly to state LINK_FAILINGOVER if the corresponding node state is NODE_FAILINGOVER. This eliminates the problem described above. Furthermore, we tighten the criteria for which packets are allowed to end a failover state in the function tipc_node_check_state(). By checking that the receiving link is up and running, instead of just checking that it is not in failover mode, we eliminate the risk that protocol packets from the re-created link may cause the failover to be prematurely terminated. Reviewed-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>
2015-08-20 13:12:54 +07:00
if ((n->state == NODE_FAILINGOVER) && tipc_link_is_up(l)) {
tipc: merge link->exec_mode and link->state into one FSM Until now, we have been handling link failover and synchronization by using an additional link state variable, "exec_mode". This variable is not independent of the link FSM state, something causing a risk of inconsistencies, apart from the fact that it clutters the code. The conditions are now in place to define a new link FSM that covers all existing use cases, including failover and synchronization, and eliminate the "exec_mode" field altogether. The FSM must also support non-atomic resetting of links, which will be introduced later. The new link FSM is shown below, with 7 states and 8 events. Only events leading to state change are shown as edges. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | | | | | | +--------------+ | | | RESET_EVT| |RESET_EVT |ESTABLISH_EVT | | | | | | | | | | | | V V | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ These changes are fully backwards compatible. Tested-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>
2015-07-31 05:24:21 +07:00
if (!more(rcv_nxt, n->sync_point))
return true;
tipc_node_fsm_evt(n, NODE_FAILOVER_END_EVT);
if (pl)
tipc: merge link->exec_mode and link->state into one FSM Until now, we have been handling link failover and synchronization by using an additional link state variable, "exec_mode". This variable is not independent of the link FSM state, something causing a risk of inconsistencies, apart from the fact that it clutters the code. The conditions are now in place to define a new link FSM that covers all existing use cases, including failover and synchronization, and eliminate the "exec_mode" field altogether. The FSM must also support non-atomic resetting of links, which will be introduced later. The new link FSM is shown below, with 7 states and 8 events. Only events leading to state change are shown as edges. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | | | | | | +--------------+ | | | RESET_EVT| |RESET_EVT |ESTABLISH_EVT | | | | | | | | | | | | V V | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ These changes are fully backwards compatible. Tested-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>
2015-07-31 05:24:21 +07:00
tipc_link_fsm_evt(pl, LINK_FAILOVER_END_EVT);
return true;
}
/* No synching needed if only one link */
if (!pl || !tipc_link_is_up(pl))
return true;
tipc: eliminate risk of stalled link synchronization In commit 6e498158a827 ("tipc: move link synch and failover to link aggregation level") we introduced a new mechanism for performing link failover and synchronization. We have now detected a bug in this mechanism. During link synchronization we use the arrival of any packet on the tunnel link to trig a check for whether it has reached the synchronization point or not. This has turned out to be too permissive, since it may cause an arriving non-last SYNCH packet to end the synch state, just to see the next SYNCH packet initiate a new synch state with a new, higher synch point. This is not fatal, but should be avoided, because it may significantly extend the synchronization period, while at the same time we are not allowed to send NACKs if packets are lost. In the worst case, a low-traffic user may see its traffic stall until a LINK_PROTOCOL state message trigs the link to leave synchronization state. At the same time, LINK_PROTOCOL packets which happen to have a (non- valid) sequence number lower than the tunnel link's rcv_nxt value will be consistently dropped, and will never be able to resolve the situation described above. We fix this by exempting LINK_PROTOCOL packets from the sequence number check, as they should be. We also reduce (but don't completely eliminate) the risk of entering multiple synchronization states by only allowing the (logically) first SYNCH packet to initiate a synchronization state. This works independently of actual packet arrival order. Fixes: commit 6e498158a827 ("tipc: move link synch and failover to link aggregation level") Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Acked-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2015-10-13 23:41:51 +07:00
/* Initiate synch mode if applicable */
if ((usr == TUNNEL_PROTOCOL) && (mtyp == SYNCH_MSG) && (oseqno == 1)) {
if (n->capabilities & TIPC_TUNNEL_ENHANCED)
syncpt = msg_syncpt(hdr);
else
syncpt = msg_seqno(msg_inner_hdr(hdr)) + exp_pkts - 1;
tipc: remove premature ESTABLISH FSM event at link synchronization When a link between two nodes come up, both endpoints will initially send out a STATE message to the peer, to increase the probability that the peer endpoint also is up when the first traffic message arrives. Thereafter, if the establishing link is the second link between two nodes, this first "traffic" message is a TUNNEL_PROTOCOL/SYNCH message, helping the peer to perform initial synchronization between the two links. However, the initial STATE message may be lost, in which case the SYNCH message will be the first one arriving at the peer. This should also work, as the SYNCH message itself will be used to take up the link endpoint before initializing synchronization. Unfortunately the code for this case is broken. Currently, the link is brought up through a tipc_link_fsm_evt(ESTABLISHED) when a SYNCH arrives, whereupon __tipc_node_link_up() is called to distribute the link slots and take the link into traffic. But, __tipc_node_link_up() is itself starting with a test for whether the link is up, and if true, returns without action. Clearly, the tipc_link_fsm_evt(ESTABLISHED) call is unnecessary, since tipc_node_link_up() is itself issuing such an event, but also harmful, since it inhibits tipc_node_link_up() to perform the test of its tasks, and the link endpoint in question hence is never taken into traffic. This problem has been exposed when we set up dual links between pre- and post-4.4 kernels, because the former ones don't send out the initial STATE message described above. We fix this by removing the unnecessary event call. Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2017-08-09 03:23:56 +07:00
if (!tipc_link_is_up(l))
__tipc_node_link_up(n, bearer_id, xmitq);
if (n->state == SELF_UP_PEER_UP) {
n->sync_point = syncpt;
tipc: merge link->exec_mode and link->state into one FSM Until now, we have been handling link failover and synchronization by using an additional link state variable, "exec_mode". This variable is not independent of the link FSM state, something causing a risk of inconsistencies, apart from the fact that it clutters the code. The conditions are now in place to define a new link FSM that covers all existing use cases, including failover and synchronization, and eliminate the "exec_mode" field altogether. The FSM must also support non-atomic resetting of links, which will be introduced later. The new link FSM is shown below, with 7 states and 8 events. Only events leading to state change are shown as edges. +------------------------------------+ |RESET_EVT | | | | +--------------+ | +-----------------| SYNCHING |-----------------+ | |FAILURE_EVT +--------------+ PEER_RESET_EVT| | | A | | | | | | | | | | | | | | |SYNCH_ |SYNCH_ | | | |BEGIN_EVT |END_EVT | | | | | | | V | V V | +-------------+ +--------------+ +------------+ | | RESETTING |<---------| ESTABLISHED |--------->| PEER_RESET | | +-------------+ FAILURE_ +--------------+ PEER_ +------------+ | | EVT | A RESET_EVT | | | | | | | | | | | | | +--------------+ | | | RESET_EVT| |RESET_EVT |ESTABLISH_EVT | | | | | | | | | | | | V V | | | +-------------+ +--------------+ RESET_EVT| +--->| RESET |--------->| ESTABLISHING |<----------------+ +-------------+ PEER_ +--------------+ | A RESET_EVT | | | | | | | |FAILOVER_ |FAILOVER_ |FAILOVER_ |BEGIN_EVT |END_EVT |BEGIN_EVT | | | V | | +-------------+ | | FAILINGOVER |<----------------+ +-------------+ These changes are fully backwards compatible. Tested-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>
2015-07-31 05:24:21 +07:00
tipc_link_fsm_evt(l, LINK_SYNCH_BEGIN_EVT);
tipc_node_fsm_evt(n, NODE_SYNCH_BEGIN_EVT);
}
}
/* Open tunnel link when parallel link reaches synch point */
2015-11-20 02:30:41 +07:00
if (n->state == NODE_SYNCHING) {
tipc: fix stale link problem during synchronization Recent changes to the link synchronization means that we can now just drop packets arriving on the synchronizing link before the synch point is reached. This has lead to significant simplifications to the implementation, but also turns out to have a flip side that we need to consider. Under unlucky circumstances, the two endpoints may end up repeatedly dropping each other's packets, while immediately asking for retransmission of the same packets, just to drop them once more. This pattern will eventually be broken when the synch point is reached on the other link, but before that, the endpoints may have arrived at the retransmission limit (stale counter) that indicates that the link should be broken. We see this happen at rare occasions. The fix for this is to not ask for retransmissions when a link is in state LINK_SYNCHING. The fact that the link has reached this state means that it has already received the first SYNCH packet, and that it knows the synch point. Hence, it doesn't need any more packets until the other link has reached the synch point, whereafter it can go ahead and ask for the missing packets. However, because of the reduced traffic on the synching link that follows this change, it may now take longer to discover that the synch point has been reached. We compensate for this by letting all packets, on any of the links, trig a check for synchronization termination. This is possible because the packets themselves don't contain any information that is needed for discovering this condition. Reviewed-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>
2015-08-20 13:12:56 +07:00
if (tipc_link_is_synching(l)) {
tnl = l;
} else {
tnl = pl;
pl = l;
}
inputq_len = skb_queue_len(tipc_link_inputq(pl));
dlv_nxt = tipc_link_rcv_nxt(pl) - inputq_len;
if (more(dlv_nxt, n->sync_point)) {
tipc: fix stale link problem during synchronization Recent changes to the link synchronization means that we can now just drop packets arriving on the synchronizing link before the synch point is reached. This has lead to significant simplifications to the implementation, but also turns out to have a flip side that we need to consider. Under unlucky circumstances, the two endpoints may end up repeatedly dropping each other's packets, while immediately asking for retransmission of the same packets, just to drop them once more. This pattern will eventually be broken when the synch point is reached on the other link, but before that, the endpoints may have arrived at the retransmission limit (stale counter) that indicates that the link should be broken. We see this happen at rare occasions. The fix for this is to not ask for retransmissions when a link is in state LINK_SYNCHING. The fact that the link has reached this state means that it has already received the first SYNCH packet, and that it knows the synch point. Hence, it doesn't need any more packets until the other link has reached the synch point, whereafter it can go ahead and ask for the missing packets. However, because of the reduced traffic on the synching link that follows this change, it may now take longer to discover that the synch point has been reached. We compensate for this by letting all packets, on any of the links, trig a check for synchronization termination. This is possible because the packets themselves don't contain any information that is needed for discovering this condition. Reviewed-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>
2015-08-20 13:12:56 +07:00
tipc_link_fsm_evt(tnl, LINK_SYNCH_END_EVT);
tipc_node_fsm_evt(n, NODE_SYNCH_END_EVT);
return true;
}
tipc: fix stale link problem during synchronization Recent changes to the link synchronization means that we can now just drop packets arriving on the synchronizing link before the synch point is reached. This has lead to significant simplifications to the implementation, but also turns out to have a flip side that we need to consider. Under unlucky circumstances, the two endpoints may end up repeatedly dropping each other's packets, while immediately asking for retransmission of the same packets, just to drop them once more. This pattern will eventually be broken when the synch point is reached on the other link, but before that, the endpoints may have arrived at the retransmission limit (stale counter) that indicates that the link should be broken. We see this happen at rare occasions. The fix for this is to not ask for retransmissions when a link is in state LINK_SYNCHING. The fact that the link has reached this state means that it has already received the first SYNCH packet, and that it knows the synch point. Hence, it doesn't need any more packets until the other link has reached the synch point, whereafter it can go ahead and ask for the missing packets. However, because of the reduced traffic on the synching link that follows this change, it may now take longer to discover that the synch point has been reached. We compensate for this by letting all packets, on any of the links, trig a check for synchronization termination. This is possible because the packets themselves don't contain any information that is needed for discovering this condition. Reviewed-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>
2015-08-20 13:12:56 +07:00
if (l == pl)
return true;
if ((usr == TUNNEL_PROTOCOL) && (mtyp == SYNCH_MSG))
return true;
if (usr == LINK_PROTOCOL)
return true;
return false;
}
return true;
}
/**
* tipc_rcv - process TIPC packets/messages arriving from off-node
* @net: the applicable net namespace
* @skb: TIPC packet
* @bearer: pointer to bearer message arrived on
*
* Invoked with no locks held. Bearer pointer must point to a valid bearer
* structure (i.e. cannot be NULL), but bearer can be inactive.
*/
void tipc_rcv(struct net *net, struct sk_buff *skb, struct tipc_bearer *b)
{
struct sk_buff_head xmitq;
struct tipc_node *n;
struct tipc_msg *hdr;
int bearer_id = b->identity;
struct tipc_link_entry *le;
tipc: only process unicast on intended node We have observed complete lock up of broadcast-link transmission due to unacknowledged packets never being removed from the 'transmq' queue. This is traced to nodes having their ack field set beyond the sequence number of packets that have actually been transmitted to them. Consider an example where node 1 has sent 10 packets to node 2 on a link and node 3 has sent 20 packets to node 2 on another link. We see examples of an ack from node 2 destined for node 3 being treated as an ack from node 2 at node 1. This leads to the ack on the node 1 to node 2 link being increased to 20 even though we have only sent 10 packets. When node 1 does get around to sending further packets, none of the packets with sequence numbers less than 21 are actually removed from the transmq. To resolve this we reinstate some code lost in commit d999297c3dbb ("tipc: reduce locking scope during packet reception") which ensures that only messages destined for the receiving node are processed by that node. This prevents the sequence numbers from getting out of sync and resolves the packet leakage, thereby resolving the broadcast-link transmission lock-ups we observed. While we are aware that this change only patches over a root problem that we still haven't identified, this is a sanity test that it is always legitimate to do. It will remain in the code even after we identify and fix the real problem. Reviewed-by: Chris Packham <chris.packham@alliedtelesis.co.nz> Reviewed-by: John Thompson <john.thompson@alliedtelesis.co.nz> Signed-off-by: Hamish Martin <hamish.martin@alliedtelesis.co.nz> Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-04-29 21:40:24 +07:00
u32 self = tipc_own_addr(net);
int usr, rc = 0;
u16 bc_ack;
__skb_queue_head_init(&xmitq);
/* Ensure message is well-formed before touching the header */
TIPC_SKB_CB(skb)->validated = false;
if (unlikely(!tipc_msg_validate(&skb)))
goto discard;
hdr = buf_msg(skb);
usr = msg_user(hdr);
bc_ack = msg_bcast_ack(hdr);
/* Handle arrival of discovery or broadcast packet */
if (unlikely(msg_non_seq(hdr))) {
if (unlikely(usr == LINK_CONFIG))
return tipc_disc_rcv(net, skb, b);
else
return tipc_node_bc_rcv(net, skb, bearer_id);
}
tipc: only process unicast on intended node We have observed complete lock up of broadcast-link transmission due to unacknowledged packets never being removed from the 'transmq' queue. This is traced to nodes having their ack field set beyond the sequence number of packets that have actually been transmitted to them. Consider an example where node 1 has sent 10 packets to node 2 on a link and node 3 has sent 20 packets to node 2 on another link. We see examples of an ack from node 2 destined for node 3 being treated as an ack from node 2 at node 1. This leads to the ack on the node 1 to node 2 link being increased to 20 even though we have only sent 10 packets. When node 1 does get around to sending further packets, none of the packets with sequence numbers less than 21 are actually removed from the transmq. To resolve this we reinstate some code lost in commit d999297c3dbb ("tipc: reduce locking scope during packet reception") which ensures that only messages destined for the receiving node are processed by that node. This prevents the sequence numbers from getting out of sync and resolves the packet leakage, thereby resolving the broadcast-link transmission lock-ups we observed. While we are aware that this change only patches over a root problem that we still haven't identified, this is a sanity test that it is always legitimate to do. It will remain in the code even after we identify and fix the real problem. Reviewed-by: Chris Packham <chris.packham@alliedtelesis.co.nz> Reviewed-by: John Thompson <john.thompson@alliedtelesis.co.nz> Signed-off-by: Hamish Martin <hamish.martin@alliedtelesis.co.nz> Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-04-29 21:40:24 +07:00
/* Discard unicast link messages destined for another node */
if (unlikely(!msg_short(hdr) && (msg_destnode(hdr) != self)))
goto discard;
/* Locate neighboring node that sent packet */
n = tipc_node_find(net, msg_prevnode(hdr));
if (unlikely(!n))
goto discard;
le = &n->links[bearer_id];
/* Ensure broadcast reception is in synch with peer's send state */
if (unlikely(usr == LINK_PROTOCOL))
tipc_node_bc_sync_rcv(n, hdr, bearer_id, &xmitq);
else if (unlikely(tipc_link_acked(n->bc_entry.link) != bc_ack))
tipc: fix broadcast link synchronization problem In commit 2d18ac4ba745 ("tipc: extend broadcast link initialization criteria") we tried to fix a problem with the initial synchronization of broadcast link acknowledge values. Unfortunately that solution is not sufficient to solve the issue. We have seen it happen that LINK_PROTOCOL/STATE packets with a valid non-zero unicast acknowledge number may bypass BCAST_PROTOCOL initialization, NAME_DISTRIBUTOR and other STATE packets with invalid broadcast acknowledge numbers, leading to premature opening of the broadcast link. When the bypassed packets finally arrive, they are inadvertently accepted, and the already correctly initialized acknowledge number in the broadcast receive link is overwritten by the invalid (zero) value of the said packets. After this the broadcast link goes stale. We now fix this by marking the packets where we know the acknowledge value is or may be invalid, and then ignoring the acks from those. To this purpose, we claim an unused bit in the header to indicate that the value is invalid. We set the bit to 1 in the initial BCAST_PROTOCOL synchronization packet and all initial ("bulk") NAME_DISTRIBUTOR packets, plus those LINK_PROTOCOL packets sent out before the broadcast links are fully synchronized. This minor protocol update is fully backwards compatible. Reported-by: John Thompson <thompa.atl@gmail.com> Tested-by: John Thompson <thompa.atl@gmail.com> Signed-off-by: Jon Maloy <jon.maloy@ericsson.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2016-10-28 05:51:55 +07:00
tipc_bcast_ack_rcv(net, n->bc_entry.link, hdr);
/* Receive packet directly if conditions permit */
tipc_node_read_lock(n);
if (likely((n->state == SELF_UP_PEER_UP) && (usr != TUNNEL_PROTOCOL))) {
spin_lock_bh(&le->lock);
if (le->link) {
rc = tipc_link_rcv(le->link, skb, &xmitq);
skb = NULL;
}
spin_unlock_bh(&le->lock);
}
tipc_node_read_unlock(n);
/* Check/update node state before receiving */
if (unlikely(skb)) {
if (unlikely(skb_linearize(skb)))
goto discard;
tipc_node_write_lock(n);
if (tipc_node_check_state(n, skb, bearer_id, &xmitq)) {
if (le->link) {
rc = tipc_link_rcv(le->link, skb, &xmitq);
skb = NULL;
}
}
tipc_node_write_unlock(n);
}
if (unlikely(rc & TIPC_LINK_UP_EVT))
tipc_node_link_up(n, bearer_id, &xmitq);
if (unlikely(rc & TIPC_LINK_DOWN_EVT))
tipc_node_link_down(n, bearer_id, false);
if (unlikely(!skb_queue_empty(&n->bc_entry.namedq)))
tipc_named_rcv(net, &n->bc_entry.namedq);
if (unlikely(!skb_queue_empty(&n->bc_entry.inputq1)))
tipc_node_mcast_rcv(n);
if (!skb_queue_empty(&le->inputq))
tipc_sk_rcv(net, &le->inputq);
if (!skb_queue_empty(&xmitq))
tipc_bearer_xmit(net, bearer_id, &xmitq, &le->maddr);
tipc_node_put(n);
discard:
kfree_skb(skb);
}
void tipc_node_apply_property(struct net *net, struct tipc_bearer *b,
int prop)
{
struct tipc_net *tn = tipc_net(net);
int bearer_id = b->identity;
struct sk_buff_head xmitq;
struct tipc_link_entry *e;
struct tipc_node *n;
__skb_queue_head_init(&xmitq);
rcu_read_lock();
list_for_each_entry_rcu(n, &tn->node_list, list) {
tipc_node_write_lock(n);
e = &n->links[bearer_id];
if (e->link) {
if (prop == TIPC_NLA_PROP_TOL)
tipc_link_set_tolerance(e->link, b->tolerance,
&xmitq);
else if (prop == TIPC_NLA_PROP_MTU)
tipc_link_set_mtu(e->link, b->mtu);
}
tipc_node_write_unlock(n);
tipc_bearer_xmit(net, bearer_id, &xmitq, &e->maddr);
}
rcu_read_unlock();
}
int tipc_nl_peer_rm(struct sk_buff *skb, struct genl_info *info)
{
struct net *net = sock_net(skb->sk);
struct tipc_net *tn = net_generic(net, tipc_net_id);
struct nlattr *attrs[TIPC_NLA_NET_MAX + 1];
struct tipc_node *peer, *temp_node;
u32 addr;
int err;
/* We identify the peer by its net */
if (!info->attrs[TIPC_NLA_NET])
return -EINVAL;
netlink: make validation more configurable for future strictness We currently have two levels of strict validation: 1) liberal (default) - undefined (type >= max) & NLA_UNSPEC attributes accepted - attribute length >= expected accepted - garbage at end of message accepted 2) strict (opt-in) - NLA_UNSPEC attributes accepted - attribute length >= expected accepted Split out parsing strictness into four different options: * TRAILING - check that there's no trailing data after parsing attributes (in message or nested) * MAXTYPE - reject attrs > max known type * UNSPEC - reject attributes with NLA_UNSPEC policy entries * STRICT_ATTRS - strictly validate attribute size The default for future things should be *everything*. The current *_strict() is a combination of TRAILING and MAXTYPE, and is renamed to _deprecated_strict(). The current regular parsing has none of this, and is renamed to *_parse_deprecated(). Additionally it allows us to selectively set one of the new flags even on old policies. Notably, the UNSPEC flag could be useful in this case, since it can be arranged (by filling in the policy) to not be an incompatible userspace ABI change, but would then going forward prevent forgetting attribute entries. Similar can apply to the POLICY flag. We end up with the following renames: * nla_parse -> nla_parse_deprecated * nla_parse_strict -> nla_parse_deprecated_strict * nlmsg_parse -> nlmsg_parse_deprecated * nlmsg_parse_strict -> nlmsg_parse_deprecated_strict * nla_parse_nested -> nla_parse_nested_deprecated * nla_validate_nested -> nla_validate_nested_deprecated Using spatch, of course: @@ expression TB, MAX, HEAD, LEN, POL, EXT; @@ -nla_parse(TB, MAX, HEAD, LEN, POL, EXT) +nla_parse_deprecated(TB, MAX, HEAD, LEN, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse_strict(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated_strict(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression TB, MAX, NLA, POL, EXT; @@ -nla_parse_nested(TB, MAX, NLA, POL, EXT) +nla_parse_nested_deprecated(TB, MAX, NLA, POL, EXT) @@ expression START, MAX, POL, EXT; @@ -nla_validate_nested(START, MAX, POL, EXT) +nla_validate_nested_deprecated(START, MAX, POL, EXT) @@ expression NLH, HDRLEN, MAX, POL, EXT; @@ -nlmsg_validate(NLH, HDRLEN, MAX, POL, EXT) +nlmsg_validate_deprecated(NLH, HDRLEN, MAX, POL, EXT) For this patch, don't actually add the strict, non-renamed versions yet so that it breaks compile if I get it wrong. Also, while at it, make nla_validate and nla_parse go down to a common __nla_validate_parse() function to avoid code duplication. Ultimately, this allows us to have very strict validation for every new caller of nla_parse()/nlmsg_parse() etc as re-introduced in the next patch, while existing things will continue to work as is. In effect then, this adds fully strict validation for any new command. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-04-26 19:07:28 +07:00
err = nla_parse_nested_deprecated(attrs, TIPC_NLA_NET_MAX,
info->attrs[TIPC_NLA_NET],
tipc_nl_net_policy, info->extack);
if (err)
return err;
if (!attrs[TIPC_NLA_NET_ADDR])
return -EINVAL;
addr = nla_get_u32(attrs[TIPC_NLA_NET_ADDR]);
if (in_own_node(net, addr))
return -ENOTSUPP;
spin_lock_bh(&tn->node_list_lock);
peer = tipc_node_find(net, addr);
if (!peer) {
spin_unlock_bh(&tn->node_list_lock);
return -ENXIO;
}
tipc_node_write_lock(peer);
if (peer->state != SELF_DOWN_PEER_DOWN &&
peer->state != SELF_DOWN_PEER_LEAVING) {
tipc_node_write_unlock(peer);
err = -EBUSY;
goto err_out;
}
tipc_node_clear_links(peer);
tipc_node_write_unlock(peer);
tipc_node_delete(peer);
/* Calculate cluster capabilities */
tn->capabilities = TIPC_NODE_CAPABILITIES;
list_for_each_entry_rcu(temp_node, &tn->node_list, list) {
tn->capabilities &= temp_node->capabilities;
}
err = 0;
err_out:
tipc_node_put(peer);
spin_unlock_bh(&tn->node_list_lock);
return err;
}
int tipc_nl_node_dump(struct sk_buff *skb, struct netlink_callback *cb)
{
int err;
struct net *net = sock_net(skb->sk);
struct tipc_net *tn = net_generic(net, tipc_net_id);
int done = cb->args[0];
int last_addr = cb->args[1];
struct tipc_node *node;
struct tipc_nl_msg msg;
if (done)
return 0;
msg.skb = skb;
msg.portid = NETLINK_CB(cb->skb).portid;
msg.seq = cb->nlh->nlmsg_seq;
rcu_read_lock();
if (last_addr) {
node = tipc_node_find(net, last_addr);
if (!node) {
rcu_read_unlock();
/* We never set seq or call nl_dump_check_consistent()
* this means that setting prev_seq here will cause the
* consistence check to fail in the netlink callback
* handler. Resulting in the NLMSG_DONE message having
* the NLM_F_DUMP_INTR flag set if the node state
* changed while we released the lock.
*/
cb->prev_seq = 1;
return -EPIPE;
}
tipc_node_put(node);
}
list_for_each_entry_rcu(node, &tn->node_list, list) {
if (last_addr) {
if (node->addr == last_addr)
last_addr = 0;
else
continue;
}
tipc_node_read_lock(node);
err = __tipc_nl_add_node(&msg, node);
if (err) {
last_addr = node->addr;
tipc_node_read_unlock(node);
goto out;
}
tipc_node_read_unlock(node);
}
done = 1;
out:
cb->args[0] = done;
cb->args[1] = last_addr;
rcu_read_unlock();
return skb->len;
}
/* tipc_node_find_by_name - locate owner node of link by link's name
* @net: the applicable net namespace
* @name: pointer to link name string
* @bearer_id: pointer to index in 'node->links' array where the link was found.
*
* Returns pointer to node owning the link, or 0 if no matching link is found.
*/
static struct tipc_node *tipc_node_find_by_name(struct net *net,
const char *link_name,
unsigned int *bearer_id)
{
struct tipc_net *tn = net_generic(net, tipc_net_id);
struct tipc_link *l;
struct tipc_node *n;
struct tipc_node *found_node = NULL;
int i;
*bearer_id = 0;
rcu_read_lock();
list_for_each_entry_rcu(n, &tn->node_list, list) {
tipc_node_read_lock(n);
for (i = 0; i < MAX_BEARERS; i++) {
l = n->links[i].link;
if (l && !strcmp(tipc_link_name(l), link_name)) {
*bearer_id = i;
found_node = n;
break;
}
}
tipc_node_read_unlock(n);
if (found_node)
break;
}
rcu_read_unlock();
return found_node;
}
int tipc_nl_node_set_link(struct sk_buff *skb, struct genl_info *info)
{
int err;
int res = 0;
int bearer_id;
char *name;
struct tipc_link *link;
struct tipc_node *node;
struct sk_buff_head xmitq;
struct nlattr *attrs[TIPC_NLA_LINK_MAX + 1];
struct net *net = sock_net(skb->sk);
__skb_queue_head_init(&xmitq);
if (!info->attrs[TIPC_NLA_LINK])
return -EINVAL;
netlink: make validation more configurable for future strictness We currently have two levels of strict validation: 1) liberal (default) - undefined (type >= max) & NLA_UNSPEC attributes accepted - attribute length >= expected accepted - garbage at end of message accepted 2) strict (opt-in) - NLA_UNSPEC attributes accepted - attribute length >= expected accepted Split out parsing strictness into four different options: * TRAILING - check that there's no trailing data after parsing attributes (in message or nested) * MAXTYPE - reject attrs > max known type * UNSPEC - reject attributes with NLA_UNSPEC policy entries * STRICT_ATTRS - strictly validate attribute size The default for future things should be *everything*. The current *_strict() is a combination of TRAILING and MAXTYPE, and is renamed to _deprecated_strict(). The current regular parsing has none of this, and is renamed to *_parse_deprecated(). Additionally it allows us to selectively set one of the new flags even on old policies. Notably, the UNSPEC flag could be useful in this case, since it can be arranged (by filling in the policy) to not be an incompatible userspace ABI change, but would then going forward prevent forgetting attribute entries. Similar can apply to the POLICY flag. We end up with the following renames: * nla_parse -> nla_parse_deprecated * nla_parse_strict -> nla_parse_deprecated_strict * nlmsg_parse -> nlmsg_parse_deprecated * nlmsg_parse_strict -> nlmsg_parse_deprecated_strict * nla_parse_nested -> nla_parse_nested_deprecated * nla_validate_nested -> nla_validate_nested_deprecated Using spatch, of course: @@ expression TB, MAX, HEAD, LEN, POL, EXT; @@ -nla_parse(TB, MAX, HEAD, LEN, POL, EXT) +nla_parse_deprecated(TB, MAX, HEAD, LEN, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse_strict(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated_strict(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression TB, MAX, NLA, POL, EXT; @@ -nla_parse_nested(TB, MAX, NLA, POL, EXT) +nla_parse_nested_deprecated(TB, MAX, NLA, POL, EXT) @@ expression START, MAX, POL, EXT; @@ -nla_validate_nested(START, MAX, POL, EXT) +nla_validate_nested_deprecated(START, MAX, POL, EXT) @@ expression NLH, HDRLEN, MAX, POL, EXT; @@ -nlmsg_validate(NLH, HDRLEN, MAX, POL, EXT) +nlmsg_validate_deprecated(NLH, HDRLEN, MAX, POL, EXT) For this patch, don't actually add the strict, non-renamed versions yet so that it breaks compile if I get it wrong. Also, while at it, make nla_validate and nla_parse go down to a common __nla_validate_parse() function to avoid code duplication. Ultimately, this allows us to have very strict validation for every new caller of nla_parse()/nlmsg_parse() etc as re-introduced in the next patch, while existing things will continue to work as is. In effect then, this adds fully strict validation for any new command. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-04-26 19:07:28 +07:00
err = nla_parse_nested_deprecated(attrs, TIPC_NLA_LINK_MAX,
info->attrs[TIPC_NLA_LINK],
tipc_nl_link_policy, info->extack);
if (err)
return err;
if (!attrs[TIPC_NLA_LINK_NAME])
return -EINVAL;
name = nla_data(attrs[TIPC_NLA_LINK_NAME]);
if (strcmp(name, tipc_bclink_name) == 0)
return tipc_nl_bc_link_set(net, attrs);
node = tipc_node_find_by_name(net, name, &bearer_id);
if (!node)
return -EINVAL;
tipc_node_read_lock(node);
link = node->links[bearer_id].link;
if (!link) {
res = -EINVAL;
goto out;
}
if (attrs[TIPC_NLA_LINK_PROP]) {
struct nlattr *props[TIPC_NLA_PROP_MAX + 1];
err = tipc_nl_parse_link_prop(attrs[TIPC_NLA_LINK_PROP],
props);
if (err) {
res = err;
goto out;
}
if (props[TIPC_NLA_PROP_TOL]) {
u32 tol;
tol = nla_get_u32(props[TIPC_NLA_PROP_TOL]);
tipc_link_set_tolerance(link, tol, &xmitq);
}
if (props[TIPC_NLA_PROP_PRIO]) {
u32 prio;
prio = nla_get_u32(props[TIPC_NLA_PROP_PRIO]);
tipc_link_set_prio(link, prio, &xmitq);
}
if (props[TIPC_NLA_PROP_WIN]) {
u32 win;
win = nla_get_u32(props[TIPC_NLA_PROP_WIN]);
tipc_link_set_queue_limits(link, win);
}
}
out:
tipc_node_read_unlock(node);
tipc_bearer_xmit(net, bearer_id, &xmitq, &node->links[bearer_id].maddr);
return res;
}
int tipc_nl_node_get_link(struct sk_buff *skb, struct genl_info *info)
{
struct net *net = genl_info_net(info);
tipc: eliminate KMSAN uninit-value in strcmp complaint When we get link properties through netlink interface with tipc_nl_node_get_link(), we don't validate TIPC_NLA_LINK_NAME attribute at all, instead we directly use it. As a consequence, KMSAN detected the TIPC_NLA_LINK_NAME attribute was an uninitialized value, and then posted the following complaint: ================================================================== BUG: KMSAN: uninit-value in strcmp+0xf7/0x160 lib/string.c:329 CPU: 1 PID: 4527 Comm: syz-executor655 Not tainted 4.16.0+ #87 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:17 [inline] dump_stack+0x185/0x1d0 lib/dump_stack.c:53 kmsan_report+0x142/0x240 mm/kmsan/kmsan.c:1067 __msan_warning_32+0x6c/0xb0 mm/kmsan/kmsan_instr.c:683 strcmp+0xf7/0x160 lib/string.c:329 tipc_nl_node_get_link+0x220/0x6f0 net/tipc/node.c:1881 genl_family_rcv_msg net/netlink/genetlink.c:599 [inline] genl_rcv_msg+0x1686/0x1810 net/netlink/genetlink.c:624 netlink_rcv_skb+0x378/0x600 net/netlink/af_netlink.c:2447 genl_rcv+0x63/0x80 net/netlink/genetlink.c:635 netlink_unicast_kernel net/netlink/af_netlink.c:1311 [inline] netlink_unicast+0x166b/0x1740 net/netlink/af_netlink.c:1337 netlink_sendmsg+0x1048/0x1310 net/netlink/af_netlink.c:1900 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 RIP: 0033:0x445589 RSP: 002b:00007fb7ee66cdb8 EFLAGS: 00000246 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 00000000006dac24 RCX: 0000000000445589 RDX: 0000000000000000 RSI: 0000000020023000 RDI: 0000000000000003 RBP: 00000000006dac20 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 R13: 00007fffa2bf3f3f R14: 00007fb7ee66d9c0 R15: 0000000000000001 Uninit was created at: kmsan_save_stack_with_flags mm/kmsan/kmsan.c:278 [inline] kmsan_internal_poison_shadow+0xb8/0x1b0 mm/kmsan/kmsan.c:188 kmsan_kmalloc+0x94/0x100 mm/kmsan/kmsan.c:314 kmsan_slab_alloc+0x11/0x20 mm/kmsan/kmsan.c:321 slab_post_alloc_hook mm/slab.h:445 [inline] slab_alloc_node mm/slub.c:2737 [inline] __kmalloc_node_track_caller+0xaed/0x11c0 mm/slub.c:4369 __kmalloc_reserve net/core/skbuff.c:138 [inline] __alloc_skb+0x2cf/0x9f0 net/core/skbuff.c:206 alloc_skb include/linux/skbuff.h:984 [inline] netlink_alloc_large_skb net/netlink/af_netlink.c:1183 [inline] netlink_sendmsg+0x9a6/0x1310 net/netlink/af_netlink.c:1875 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 ================================================================== To quiet the complaint, TIPC_NLA_LINK_NAME attribute has been validated in tipc_nl_node_get_link() before it's used. Reported-by: syzbot+df0257c92ffd4fcc58cd@syzkaller.appspotmail.com Signed-off-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-08 20:44:06 +07:00
struct nlattr *attrs[TIPC_NLA_LINK_MAX + 1];
struct tipc_nl_msg msg;
char *name;
int err;
msg.portid = info->snd_portid;
msg.seq = info->snd_seq;
tipc: eliminate KMSAN uninit-value in strcmp complaint When we get link properties through netlink interface with tipc_nl_node_get_link(), we don't validate TIPC_NLA_LINK_NAME attribute at all, instead we directly use it. As a consequence, KMSAN detected the TIPC_NLA_LINK_NAME attribute was an uninitialized value, and then posted the following complaint: ================================================================== BUG: KMSAN: uninit-value in strcmp+0xf7/0x160 lib/string.c:329 CPU: 1 PID: 4527 Comm: syz-executor655 Not tainted 4.16.0+ #87 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:17 [inline] dump_stack+0x185/0x1d0 lib/dump_stack.c:53 kmsan_report+0x142/0x240 mm/kmsan/kmsan.c:1067 __msan_warning_32+0x6c/0xb0 mm/kmsan/kmsan_instr.c:683 strcmp+0xf7/0x160 lib/string.c:329 tipc_nl_node_get_link+0x220/0x6f0 net/tipc/node.c:1881 genl_family_rcv_msg net/netlink/genetlink.c:599 [inline] genl_rcv_msg+0x1686/0x1810 net/netlink/genetlink.c:624 netlink_rcv_skb+0x378/0x600 net/netlink/af_netlink.c:2447 genl_rcv+0x63/0x80 net/netlink/genetlink.c:635 netlink_unicast_kernel net/netlink/af_netlink.c:1311 [inline] netlink_unicast+0x166b/0x1740 net/netlink/af_netlink.c:1337 netlink_sendmsg+0x1048/0x1310 net/netlink/af_netlink.c:1900 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 RIP: 0033:0x445589 RSP: 002b:00007fb7ee66cdb8 EFLAGS: 00000246 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 00000000006dac24 RCX: 0000000000445589 RDX: 0000000000000000 RSI: 0000000020023000 RDI: 0000000000000003 RBP: 00000000006dac20 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 R13: 00007fffa2bf3f3f R14: 00007fb7ee66d9c0 R15: 0000000000000001 Uninit was created at: kmsan_save_stack_with_flags mm/kmsan/kmsan.c:278 [inline] kmsan_internal_poison_shadow+0xb8/0x1b0 mm/kmsan/kmsan.c:188 kmsan_kmalloc+0x94/0x100 mm/kmsan/kmsan.c:314 kmsan_slab_alloc+0x11/0x20 mm/kmsan/kmsan.c:321 slab_post_alloc_hook mm/slab.h:445 [inline] slab_alloc_node mm/slub.c:2737 [inline] __kmalloc_node_track_caller+0xaed/0x11c0 mm/slub.c:4369 __kmalloc_reserve net/core/skbuff.c:138 [inline] __alloc_skb+0x2cf/0x9f0 net/core/skbuff.c:206 alloc_skb include/linux/skbuff.h:984 [inline] netlink_alloc_large_skb net/netlink/af_netlink.c:1183 [inline] netlink_sendmsg+0x9a6/0x1310 net/netlink/af_netlink.c:1875 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 ================================================================== To quiet the complaint, TIPC_NLA_LINK_NAME attribute has been validated in tipc_nl_node_get_link() before it's used. Reported-by: syzbot+df0257c92ffd4fcc58cd@syzkaller.appspotmail.com Signed-off-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-08 20:44:06 +07:00
if (!info->attrs[TIPC_NLA_LINK])
return -EINVAL;
tipc: eliminate KMSAN uninit-value in strcmp complaint When we get link properties through netlink interface with tipc_nl_node_get_link(), we don't validate TIPC_NLA_LINK_NAME attribute at all, instead we directly use it. As a consequence, KMSAN detected the TIPC_NLA_LINK_NAME attribute was an uninitialized value, and then posted the following complaint: ================================================================== BUG: KMSAN: uninit-value in strcmp+0xf7/0x160 lib/string.c:329 CPU: 1 PID: 4527 Comm: syz-executor655 Not tainted 4.16.0+ #87 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:17 [inline] dump_stack+0x185/0x1d0 lib/dump_stack.c:53 kmsan_report+0x142/0x240 mm/kmsan/kmsan.c:1067 __msan_warning_32+0x6c/0xb0 mm/kmsan/kmsan_instr.c:683 strcmp+0xf7/0x160 lib/string.c:329 tipc_nl_node_get_link+0x220/0x6f0 net/tipc/node.c:1881 genl_family_rcv_msg net/netlink/genetlink.c:599 [inline] genl_rcv_msg+0x1686/0x1810 net/netlink/genetlink.c:624 netlink_rcv_skb+0x378/0x600 net/netlink/af_netlink.c:2447 genl_rcv+0x63/0x80 net/netlink/genetlink.c:635 netlink_unicast_kernel net/netlink/af_netlink.c:1311 [inline] netlink_unicast+0x166b/0x1740 net/netlink/af_netlink.c:1337 netlink_sendmsg+0x1048/0x1310 net/netlink/af_netlink.c:1900 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 RIP: 0033:0x445589 RSP: 002b:00007fb7ee66cdb8 EFLAGS: 00000246 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 00000000006dac24 RCX: 0000000000445589 RDX: 0000000000000000 RSI: 0000000020023000 RDI: 0000000000000003 RBP: 00000000006dac20 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 R13: 00007fffa2bf3f3f R14: 00007fb7ee66d9c0 R15: 0000000000000001 Uninit was created at: kmsan_save_stack_with_flags mm/kmsan/kmsan.c:278 [inline] kmsan_internal_poison_shadow+0xb8/0x1b0 mm/kmsan/kmsan.c:188 kmsan_kmalloc+0x94/0x100 mm/kmsan/kmsan.c:314 kmsan_slab_alloc+0x11/0x20 mm/kmsan/kmsan.c:321 slab_post_alloc_hook mm/slab.h:445 [inline] slab_alloc_node mm/slub.c:2737 [inline] __kmalloc_node_track_caller+0xaed/0x11c0 mm/slub.c:4369 __kmalloc_reserve net/core/skbuff.c:138 [inline] __alloc_skb+0x2cf/0x9f0 net/core/skbuff.c:206 alloc_skb include/linux/skbuff.h:984 [inline] netlink_alloc_large_skb net/netlink/af_netlink.c:1183 [inline] netlink_sendmsg+0x9a6/0x1310 net/netlink/af_netlink.c:1875 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 ================================================================== To quiet the complaint, TIPC_NLA_LINK_NAME attribute has been validated in tipc_nl_node_get_link() before it's used. Reported-by: syzbot+df0257c92ffd4fcc58cd@syzkaller.appspotmail.com Signed-off-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-08 20:44:06 +07:00
netlink: make validation more configurable for future strictness We currently have two levels of strict validation: 1) liberal (default) - undefined (type >= max) & NLA_UNSPEC attributes accepted - attribute length >= expected accepted - garbage at end of message accepted 2) strict (opt-in) - NLA_UNSPEC attributes accepted - attribute length >= expected accepted Split out parsing strictness into four different options: * TRAILING - check that there's no trailing data after parsing attributes (in message or nested) * MAXTYPE - reject attrs > max known type * UNSPEC - reject attributes with NLA_UNSPEC policy entries * STRICT_ATTRS - strictly validate attribute size The default for future things should be *everything*. The current *_strict() is a combination of TRAILING and MAXTYPE, and is renamed to _deprecated_strict(). The current regular parsing has none of this, and is renamed to *_parse_deprecated(). Additionally it allows us to selectively set one of the new flags even on old policies. Notably, the UNSPEC flag could be useful in this case, since it can be arranged (by filling in the policy) to not be an incompatible userspace ABI change, but would then going forward prevent forgetting attribute entries. Similar can apply to the POLICY flag. We end up with the following renames: * nla_parse -> nla_parse_deprecated * nla_parse_strict -> nla_parse_deprecated_strict * nlmsg_parse -> nlmsg_parse_deprecated * nlmsg_parse_strict -> nlmsg_parse_deprecated_strict * nla_parse_nested -> nla_parse_nested_deprecated * nla_validate_nested -> nla_validate_nested_deprecated Using spatch, of course: @@ expression TB, MAX, HEAD, LEN, POL, EXT; @@ -nla_parse(TB, MAX, HEAD, LEN, POL, EXT) +nla_parse_deprecated(TB, MAX, HEAD, LEN, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse_strict(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated_strict(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression TB, MAX, NLA, POL, EXT; @@ -nla_parse_nested(TB, MAX, NLA, POL, EXT) +nla_parse_nested_deprecated(TB, MAX, NLA, POL, EXT) @@ expression START, MAX, POL, EXT; @@ -nla_validate_nested(START, MAX, POL, EXT) +nla_validate_nested_deprecated(START, MAX, POL, EXT) @@ expression NLH, HDRLEN, MAX, POL, EXT; @@ -nlmsg_validate(NLH, HDRLEN, MAX, POL, EXT) +nlmsg_validate_deprecated(NLH, HDRLEN, MAX, POL, EXT) For this patch, don't actually add the strict, non-renamed versions yet so that it breaks compile if I get it wrong. Also, while at it, make nla_validate and nla_parse go down to a common __nla_validate_parse() function to avoid code duplication. Ultimately, this allows us to have very strict validation for every new caller of nla_parse()/nlmsg_parse() etc as re-introduced in the next patch, while existing things will continue to work as is. In effect then, this adds fully strict validation for any new command. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-04-26 19:07:28 +07:00
err = nla_parse_nested_deprecated(attrs, TIPC_NLA_LINK_MAX,
info->attrs[TIPC_NLA_LINK],
tipc_nl_link_policy, info->extack);
tipc: eliminate KMSAN uninit-value in strcmp complaint When we get link properties through netlink interface with tipc_nl_node_get_link(), we don't validate TIPC_NLA_LINK_NAME attribute at all, instead we directly use it. As a consequence, KMSAN detected the TIPC_NLA_LINK_NAME attribute was an uninitialized value, and then posted the following complaint: ================================================================== BUG: KMSAN: uninit-value in strcmp+0xf7/0x160 lib/string.c:329 CPU: 1 PID: 4527 Comm: syz-executor655 Not tainted 4.16.0+ #87 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 Call Trace: __dump_stack lib/dump_stack.c:17 [inline] dump_stack+0x185/0x1d0 lib/dump_stack.c:53 kmsan_report+0x142/0x240 mm/kmsan/kmsan.c:1067 __msan_warning_32+0x6c/0xb0 mm/kmsan/kmsan_instr.c:683 strcmp+0xf7/0x160 lib/string.c:329 tipc_nl_node_get_link+0x220/0x6f0 net/tipc/node.c:1881 genl_family_rcv_msg net/netlink/genetlink.c:599 [inline] genl_rcv_msg+0x1686/0x1810 net/netlink/genetlink.c:624 netlink_rcv_skb+0x378/0x600 net/netlink/af_netlink.c:2447 genl_rcv+0x63/0x80 net/netlink/genetlink.c:635 netlink_unicast_kernel net/netlink/af_netlink.c:1311 [inline] netlink_unicast+0x166b/0x1740 net/netlink/af_netlink.c:1337 netlink_sendmsg+0x1048/0x1310 net/netlink/af_netlink.c:1900 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 RIP: 0033:0x445589 RSP: 002b:00007fb7ee66cdb8 EFLAGS: 00000246 ORIG_RAX: 000000000000002e RAX: ffffffffffffffda RBX: 00000000006dac24 RCX: 0000000000445589 RDX: 0000000000000000 RSI: 0000000020023000 RDI: 0000000000000003 RBP: 00000000006dac20 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 0000000000000246 R12: 0000000000000000 R13: 00007fffa2bf3f3f R14: 00007fb7ee66d9c0 R15: 0000000000000001 Uninit was created at: kmsan_save_stack_with_flags mm/kmsan/kmsan.c:278 [inline] kmsan_internal_poison_shadow+0xb8/0x1b0 mm/kmsan/kmsan.c:188 kmsan_kmalloc+0x94/0x100 mm/kmsan/kmsan.c:314 kmsan_slab_alloc+0x11/0x20 mm/kmsan/kmsan.c:321 slab_post_alloc_hook mm/slab.h:445 [inline] slab_alloc_node mm/slub.c:2737 [inline] __kmalloc_node_track_caller+0xaed/0x11c0 mm/slub.c:4369 __kmalloc_reserve net/core/skbuff.c:138 [inline] __alloc_skb+0x2cf/0x9f0 net/core/skbuff.c:206 alloc_skb include/linux/skbuff.h:984 [inline] netlink_alloc_large_skb net/netlink/af_netlink.c:1183 [inline] netlink_sendmsg+0x9a6/0x1310 net/netlink/af_netlink.c:1875 sock_sendmsg_nosec net/socket.c:630 [inline] sock_sendmsg net/socket.c:640 [inline] ___sys_sendmsg+0xec0/0x1310 net/socket.c:2046 __sys_sendmsg net/socket.c:2080 [inline] SYSC_sendmsg+0x2a3/0x3d0 net/socket.c:2091 SyS_sendmsg+0x54/0x80 net/socket.c:2087 do_syscall_64+0x309/0x430 arch/x86/entry/common.c:287 entry_SYSCALL_64_after_hwframe+0x3d/0xa2 ================================================================== To quiet the complaint, TIPC_NLA_LINK_NAME attribute has been validated in tipc_nl_node_get_link() before it's used. Reported-by: syzbot+df0257c92ffd4fcc58cd@syzkaller.appspotmail.com Signed-off-by: Ying Xue <ying.xue@windriver.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2018-05-08 20:44:06 +07:00
if (err)
return err;
if (!attrs[TIPC_NLA_LINK_NAME])
return -EINVAL;
name = nla_data(attrs[TIPC_NLA_LINK_NAME]);
msg.skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
if (!msg.skb)
return -ENOMEM;
if (strcmp(name, tipc_bclink_name) == 0) {
err = tipc_nl_add_bc_link(net, &msg);
if (err)
goto err_free;
} else {
int bearer_id;
struct tipc_node *node;
struct tipc_link *link;
node = tipc_node_find_by_name(net, name, &bearer_id);
if (!node) {
err = -EINVAL;
goto err_free;
}
tipc_node_read_lock(node);
link = node->links[bearer_id].link;
if (!link) {
tipc_node_read_unlock(node);
err = -EINVAL;
goto err_free;
}
err = __tipc_nl_add_link(net, &msg, link, 0);
tipc_node_read_unlock(node);
if (err)
goto err_free;
}
return genlmsg_reply(msg.skb, info);
err_free:
nlmsg_free(msg.skb);
return err;
}
int tipc_nl_node_reset_link_stats(struct sk_buff *skb, struct genl_info *info)
{
int err;
char *link_name;
unsigned int bearer_id;
struct tipc_link *link;
struct tipc_node *node;
struct nlattr *attrs[TIPC_NLA_LINK_MAX + 1];
struct net *net = sock_net(skb->sk);
struct tipc_link_entry *le;
if (!info->attrs[TIPC_NLA_LINK])
return -EINVAL;
netlink: make validation more configurable for future strictness We currently have two levels of strict validation: 1) liberal (default) - undefined (type >= max) & NLA_UNSPEC attributes accepted - attribute length >= expected accepted - garbage at end of message accepted 2) strict (opt-in) - NLA_UNSPEC attributes accepted - attribute length >= expected accepted Split out parsing strictness into four different options: * TRAILING - check that there's no trailing data after parsing attributes (in message or nested) * MAXTYPE - reject attrs > max known type * UNSPEC - reject attributes with NLA_UNSPEC policy entries * STRICT_ATTRS - strictly validate attribute size The default for future things should be *everything*. The current *_strict() is a combination of TRAILING and MAXTYPE, and is renamed to _deprecated_strict(). The current regular parsing has none of this, and is renamed to *_parse_deprecated(). Additionally it allows us to selectively set one of the new flags even on old policies. Notably, the UNSPEC flag could be useful in this case, since it can be arranged (by filling in the policy) to not be an incompatible userspace ABI change, but would then going forward prevent forgetting attribute entries. Similar can apply to the POLICY flag. We end up with the following renames: * nla_parse -> nla_parse_deprecated * nla_parse_strict -> nla_parse_deprecated_strict * nlmsg_parse -> nlmsg_parse_deprecated * nlmsg_parse_strict -> nlmsg_parse_deprecated_strict * nla_parse_nested -> nla_parse_nested_deprecated * nla_validate_nested -> nla_validate_nested_deprecated Using spatch, of course: @@ expression TB, MAX, HEAD, LEN, POL, EXT; @@ -nla_parse(TB, MAX, HEAD, LEN, POL, EXT) +nla_parse_deprecated(TB, MAX, HEAD, LEN, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse_strict(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated_strict(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression TB, MAX, NLA, POL, EXT; @@ -nla_parse_nested(TB, MAX, NLA, POL, EXT) +nla_parse_nested_deprecated(TB, MAX, NLA, POL, EXT) @@ expression START, MAX, POL, EXT; @@ -nla_validate_nested(START, MAX, POL, EXT) +nla_validate_nested_deprecated(START, MAX, POL, EXT) @@ expression NLH, HDRLEN, MAX, POL, EXT; @@ -nlmsg_validate(NLH, HDRLEN, MAX, POL, EXT) +nlmsg_validate_deprecated(NLH, HDRLEN, MAX, POL, EXT) For this patch, don't actually add the strict, non-renamed versions yet so that it breaks compile if I get it wrong. Also, while at it, make nla_validate and nla_parse go down to a common __nla_validate_parse() function to avoid code duplication. Ultimately, this allows us to have very strict validation for every new caller of nla_parse()/nlmsg_parse() etc as re-introduced in the next patch, while existing things will continue to work as is. In effect then, this adds fully strict validation for any new command. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-04-26 19:07:28 +07:00
err = nla_parse_nested_deprecated(attrs, TIPC_NLA_LINK_MAX,
info->attrs[TIPC_NLA_LINK],
tipc_nl_link_policy, info->extack);
if (err)
return err;
if (!attrs[TIPC_NLA_LINK_NAME])
return -EINVAL;
link_name = nla_data(attrs[TIPC_NLA_LINK_NAME]);
if (strcmp(link_name, tipc_bclink_name) == 0) {
err = tipc_bclink_reset_stats(net);
if (err)
return err;
return 0;
}
node = tipc_node_find_by_name(net, link_name, &bearer_id);
if (!node)
return -EINVAL;
le = &node->links[bearer_id];
tipc_node_read_lock(node);
spin_lock_bh(&le->lock);
link = node->links[bearer_id].link;
if (!link) {
spin_unlock_bh(&le->lock);
tipc_node_read_unlock(node);
return -EINVAL;
}
tipc_link_reset_stats(link);
spin_unlock_bh(&le->lock);
tipc_node_read_unlock(node);
return 0;
}
/* Caller should hold node lock */
static int __tipc_nl_add_node_links(struct net *net, struct tipc_nl_msg *msg,
struct tipc_node *node, u32 *prev_link)
{
u32 i;
int err;
for (i = *prev_link; i < MAX_BEARERS; i++) {
*prev_link = i;
if (!node->links[i].link)
continue;
err = __tipc_nl_add_link(net, msg,
node->links[i].link, NLM_F_MULTI);
if (err)
return err;
}
*prev_link = 0;
return 0;
}
int tipc_nl_node_dump_link(struct sk_buff *skb, struct netlink_callback *cb)
{
struct net *net = sock_net(skb->sk);
struct tipc_net *tn = net_generic(net, tipc_net_id);
struct tipc_node *node;
struct tipc_nl_msg msg;
u32 prev_node = cb->args[0];
u32 prev_link = cb->args[1];
int done = cb->args[2];
int err;
if (done)
return 0;
msg.skb = skb;
msg.portid = NETLINK_CB(cb->skb).portid;
msg.seq = cb->nlh->nlmsg_seq;
rcu_read_lock();
if (prev_node) {
node = tipc_node_find(net, prev_node);
if (!node) {
/* We never set seq or call nl_dump_check_consistent()
* this means that setting prev_seq here will cause the
* consistence check to fail in the netlink callback
* handler. Resulting in the last NLMSG_DONE message
* having the NLM_F_DUMP_INTR flag set.
*/
cb->prev_seq = 1;
goto out;
}
tipc_node_put(node);
list_for_each_entry_continue_rcu(node, &tn->node_list,
list) {
tipc_node_read_lock(node);
err = __tipc_nl_add_node_links(net, &msg, node,
&prev_link);
tipc_node_read_unlock(node);
if (err)
goto out;
prev_node = node->addr;
}
} else {
err = tipc_nl_add_bc_link(net, &msg);
if (err)
goto out;
list_for_each_entry_rcu(node, &tn->node_list, list) {
tipc_node_read_lock(node);
err = __tipc_nl_add_node_links(net, &msg, node,
&prev_link);
tipc_node_read_unlock(node);
if (err)
goto out;
prev_node = node->addr;
}
}
done = 1;
out:
rcu_read_unlock();
cb->args[0] = prev_node;
cb->args[1] = prev_link;
cb->args[2] = done;
return skb->len;
}
int tipc_nl_node_set_monitor(struct sk_buff *skb, struct genl_info *info)
{
struct nlattr *attrs[TIPC_NLA_MON_MAX + 1];
struct net *net = sock_net(skb->sk);
int err;
if (!info->attrs[TIPC_NLA_MON])
return -EINVAL;
netlink: make validation more configurable for future strictness We currently have two levels of strict validation: 1) liberal (default) - undefined (type >= max) & NLA_UNSPEC attributes accepted - attribute length >= expected accepted - garbage at end of message accepted 2) strict (opt-in) - NLA_UNSPEC attributes accepted - attribute length >= expected accepted Split out parsing strictness into four different options: * TRAILING - check that there's no trailing data after parsing attributes (in message or nested) * MAXTYPE - reject attrs > max known type * UNSPEC - reject attributes with NLA_UNSPEC policy entries * STRICT_ATTRS - strictly validate attribute size The default for future things should be *everything*. The current *_strict() is a combination of TRAILING and MAXTYPE, and is renamed to _deprecated_strict(). The current regular parsing has none of this, and is renamed to *_parse_deprecated(). Additionally it allows us to selectively set one of the new flags even on old policies. Notably, the UNSPEC flag could be useful in this case, since it can be arranged (by filling in the policy) to not be an incompatible userspace ABI change, but would then going forward prevent forgetting attribute entries. Similar can apply to the POLICY flag. We end up with the following renames: * nla_parse -> nla_parse_deprecated * nla_parse_strict -> nla_parse_deprecated_strict * nlmsg_parse -> nlmsg_parse_deprecated * nlmsg_parse_strict -> nlmsg_parse_deprecated_strict * nla_parse_nested -> nla_parse_nested_deprecated * nla_validate_nested -> nla_validate_nested_deprecated Using spatch, of course: @@ expression TB, MAX, HEAD, LEN, POL, EXT; @@ -nla_parse(TB, MAX, HEAD, LEN, POL, EXT) +nla_parse_deprecated(TB, MAX, HEAD, LEN, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse_strict(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated_strict(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression TB, MAX, NLA, POL, EXT; @@ -nla_parse_nested(TB, MAX, NLA, POL, EXT) +nla_parse_nested_deprecated(TB, MAX, NLA, POL, EXT) @@ expression START, MAX, POL, EXT; @@ -nla_validate_nested(START, MAX, POL, EXT) +nla_validate_nested_deprecated(START, MAX, POL, EXT) @@ expression NLH, HDRLEN, MAX, POL, EXT; @@ -nlmsg_validate(NLH, HDRLEN, MAX, POL, EXT) +nlmsg_validate_deprecated(NLH, HDRLEN, MAX, POL, EXT) For this patch, don't actually add the strict, non-renamed versions yet so that it breaks compile if I get it wrong. Also, while at it, make nla_validate and nla_parse go down to a common __nla_validate_parse() function to avoid code duplication. Ultimately, this allows us to have very strict validation for every new caller of nla_parse()/nlmsg_parse() etc as re-introduced in the next patch, while existing things will continue to work as is. In effect then, this adds fully strict validation for any new command. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-04-26 19:07:28 +07:00
err = nla_parse_nested_deprecated(attrs, TIPC_NLA_MON_MAX,
info->attrs[TIPC_NLA_MON],
tipc_nl_monitor_policy,
info->extack);
if (err)
return err;
if (attrs[TIPC_NLA_MON_ACTIVATION_THRESHOLD]) {
u32 val;
val = nla_get_u32(attrs[TIPC_NLA_MON_ACTIVATION_THRESHOLD]);
err = tipc_nl_monitor_set_threshold(net, val);
if (err)
return err;
}
return 0;
}
static int __tipc_nl_add_monitor_prop(struct net *net, struct tipc_nl_msg *msg)
{
struct nlattr *attrs;
void *hdr;
u32 val;
hdr = genlmsg_put(msg->skb, msg->portid, msg->seq, &tipc_genl_family,
0, TIPC_NL_MON_GET);
if (!hdr)
return -EMSGSIZE;
attrs = nla_nest_start_noflag(msg->skb, TIPC_NLA_MON);
if (!attrs)
goto msg_full;
val = tipc_nl_monitor_get_threshold(net);
if (nla_put_u32(msg->skb, TIPC_NLA_MON_ACTIVATION_THRESHOLD, val))
goto attr_msg_full;
nla_nest_end(msg->skb, attrs);
genlmsg_end(msg->skb, hdr);
return 0;
attr_msg_full:
nla_nest_cancel(msg->skb, attrs);
msg_full:
genlmsg_cancel(msg->skb, hdr);
return -EMSGSIZE;
}
int tipc_nl_node_get_monitor(struct sk_buff *skb, struct genl_info *info)
{
struct net *net = sock_net(skb->sk);
struct tipc_nl_msg msg;
int err;
msg.skb = nlmsg_new(NLMSG_GOODSIZE, GFP_KERNEL);
if (!msg.skb)
return -ENOMEM;
msg.portid = info->snd_portid;
msg.seq = info->snd_seq;
err = __tipc_nl_add_monitor_prop(net, &msg);
if (err) {
nlmsg_free(msg.skb);
return err;
}
return genlmsg_reply(msg.skb, info);
}
int tipc_nl_node_dump_monitor(struct sk_buff *skb, struct netlink_callback *cb)
{
struct net *net = sock_net(skb->sk);
u32 prev_bearer = cb->args[0];
struct tipc_nl_msg msg;
int bearer_id;
int err;
if (prev_bearer == MAX_BEARERS)
return 0;
msg.skb = skb;
msg.portid = NETLINK_CB(cb->skb).portid;
msg.seq = cb->nlh->nlmsg_seq;
rtnl_lock();
for (bearer_id = prev_bearer; bearer_id < MAX_BEARERS; bearer_id++) {
err = __tipc_nl_add_monitor(net, &msg, bearer_id);
if (err)
break;
}
rtnl_unlock();
cb->args[0] = bearer_id;
return skb->len;
}
int tipc_nl_node_dump_monitor_peer(struct sk_buff *skb,
struct netlink_callback *cb)
{
struct net *net = sock_net(skb->sk);
u32 prev_node = cb->args[1];
u32 bearer_id = cb->args[2];
int done = cb->args[0];
struct tipc_nl_msg msg;
int err;
if (!prev_node) {
struct nlattr **attrs = genl_dumpit_info(cb)->attrs;
struct nlattr *mon[TIPC_NLA_MON_MAX + 1];
if (!attrs[TIPC_NLA_MON])
return -EINVAL;
netlink: make validation more configurable for future strictness We currently have two levels of strict validation: 1) liberal (default) - undefined (type >= max) & NLA_UNSPEC attributes accepted - attribute length >= expected accepted - garbage at end of message accepted 2) strict (opt-in) - NLA_UNSPEC attributes accepted - attribute length >= expected accepted Split out parsing strictness into four different options: * TRAILING - check that there's no trailing data after parsing attributes (in message or nested) * MAXTYPE - reject attrs > max known type * UNSPEC - reject attributes with NLA_UNSPEC policy entries * STRICT_ATTRS - strictly validate attribute size The default for future things should be *everything*. The current *_strict() is a combination of TRAILING and MAXTYPE, and is renamed to _deprecated_strict(). The current regular parsing has none of this, and is renamed to *_parse_deprecated(). Additionally it allows us to selectively set one of the new flags even on old policies. Notably, the UNSPEC flag could be useful in this case, since it can be arranged (by filling in the policy) to not be an incompatible userspace ABI change, but would then going forward prevent forgetting attribute entries. Similar can apply to the POLICY flag. We end up with the following renames: * nla_parse -> nla_parse_deprecated * nla_parse_strict -> nla_parse_deprecated_strict * nlmsg_parse -> nlmsg_parse_deprecated * nlmsg_parse_strict -> nlmsg_parse_deprecated_strict * nla_parse_nested -> nla_parse_nested_deprecated * nla_validate_nested -> nla_validate_nested_deprecated Using spatch, of course: @@ expression TB, MAX, HEAD, LEN, POL, EXT; @@ -nla_parse(TB, MAX, HEAD, LEN, POL, EXT) +nla_parse_deprecated(TB, MAX, HEAD, LEN, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression NLH, HDRLEN, TB, MAX, POL, EXT; @@ -nlmsg_parse_strict(NLH, HDRLEN, TB, MAX, POL, EXT) +nlmsg_parse_deprecated_strict(NLH, HDRLEN, TB, MAX, POL, EXT) @@ expression TB, MAX, NLA, POL, EXT; @@ -nla_parse_nested(TB, MAX, NLA, POL, EXT) +nla_parse_nested_deprecated(TB, MAX, NLA, POL, EXT) @@ expression START, MAX, POL, EXT; @@ -nla_validate_nested(START, MAX, POL, EXT) +nla_validate_nested_deprecated(START, MAX, POL, EXT) @@ expression NLH, HDRLEN, MAX, POL, EXT; @@ -nlmsg_validate(NLH, HDRLEN, MAX, POL, EXT) +nlmsg_validate_deprecated(NLH, HDRLEN, MAX, POL, EXT) For this patch, don't actually add the strict, non-renamed versions yet so that it breaks compile if I get it wrong. Also, while at it, make nla_validate and nla_parse go down to a common __nla_validate_parse() function to avoid code duplication. Ultimately, this allows us to have very strict validation for every new caller of nla_parse()/nlmsg_parse() etc as re-introduced in the next patch, while existing things will continue to work as is. In effect then, this adds fully strict validation for any new command. Signed-off-by: Johannes Berg <johannes.berg@intel.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2019-04-26 19:07:28 +07:00
err = nla_parse_nested_deprecated(mon, TIPC_NLA_MON_MAX,
attrs[TIPC_NLA_MON],
tipc_nl_monitor_policy,
NULL);
if (err)
return err;
if (!mon[TIPC_NLA_MON_REF])
return -EINVAL;
bearer_id = nla_get_u32(mon[TIPC_NLA_MON_REF]);
if (bearer_id >= MAX_BEARERS)
return -EINVAL;
}
if (done)
return 0;
msg.skb = skb;
msg.portid = NETLINK_CB(cb->skb).portid;
msg.seq = cb->nlh->nlmsg_seq;
rtnl_lock();
err = tipc_nl_add_monitor_peer(net, &msg, bearer_id, &prev_node);
if (!err)
done = 1;
rtnl_unlock();
cb->args[0] = done;
cb->args[1] = prev_node;
cb->args[2] = bearer_id;
return skb->len;
}
tipc: enable tracepoints in tipc As for the sake of debugging/tracing, the commit enables tracepoints in TIPC along with some general trace_events as shown below. It also defines some 'tipc_*_dump()' functions that allow to dump TIPC object data whenever needed, that is, for general debug purposes, ie. not just for the trace_events. The following trace_events are now available: - trace_tipc_skb_dump(): allows to trace and dump TIPC msg & skb data, e.g. message type, user, droppable, skb truesize, cloned skb, etc. - trace_tipc_list_dump(): allows to trace and dump any TIPC buffers or queues, e.g. TIPC link transmq, socket receive queue, etc. - trace_tipc_sk_dump(): allows to trace and dump TIPC socket data, e.g. sk state, sk type, connection type, rmem_alloc, socket queues, etc. - trace_tipc_link_dump(): allows to trace and dump TIPC link data, e.g. link state, silent_intv_cnt, gap, bc_gap, link queues, etc. - trace_tipc_node_dump(): allows to trace and dump TIPC node data, e.g. node state, active links, capabilities, link entries, etc. How to use: Put the trace functions at any places where we want to dump TIPC data or events. Note: a) The dump functions will generate raw data only, that is, to offload the trace event's processing, it can require a tool or script to parse the data but this should be simple. b) The trace_tipc_*_dump() should be reserved for a failure cases only (e.g. the retransmission failure case) or where we do not expect to happen too often, then we can consider enabling these events by default since they will almost not take any effects under normal conditions, but once the rare condition or failure occurs, we get the dumped data fully for post-analysis. For other trace purposes, we can reuse these trace classes as template but different events. c) A trace_event is only effective when we enable it. To enable the TIPC trace_events, echo 1 to 'enable' files in the events/tipc/ directory in the 'debugfs' file system. Normally, they are located at: /sys/kernel/debug/tracing/events/tipc/ For example: To enable the tipc_link_dump event: echo 1 > /sys/kernel/debug/tracing/events/tipc/tipc_link_dump/enable To enable all the TIPC trace_events: echo 1 > /sys/kernel/debug/tracing/events/tipc/enable To collect the trace data: cat trace or cat trace_pipe > /trace.out & To disable all the TIPC trace_events: echo 0 > /sys/kernel/debug/tracing/events/tipc/enable To clear the trace buffer: echo > trace d) Like the other trace_events, the feature like 'filter' or 'trigger' is also usable for the tipc trace_events. For more details, have a look at: Documentation/trace/ftrace.txt MAINTAINERS | add two new files 'trace.h' & 'trace.c' in tipc Acked-by: Ying Xue <ying.xue@windriver.com> Tested-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>
2018-12-19 09:17:56 +07:00
u32 tipc_node_get_addr(struct tipc_node *node)
{
return (node) ? node->addr : 0;
}
/**
* tipc_node_dump - dump TIPC node data
* @n: tipc node to be dumped
* @more: dump more?
* - false: dump only tipc node data
* - true: dump node link data as well
* @buf: returned buffer of dump data in format
*/
int tipc_node_dump(struct tipc_node *n, bool more, char *buf)
{
int i = 0;
size_t sz = (more) ? NODE_LMAX : NODE_LMIN;
if (!n) {
i += scnprintf(buf, sz, "node data: (null)\n");
return i;
}
i += scnprintf(buf, sz, "node data: %x", n->addr);
i += scnprintf(buf + i, sz - i, " %x", n->state);
i += scnprintf(buf + i, sz - i, " %d", n->active_links[0]);
i += scnprintf(buf + i, sz - i, " %d", n->active_links[1]);
i += scnprintf(buf + i, sz - i, " %x", n->action_flags);
i += scnprintf(buf + i, sz - i, " %u", n->failover_sent);
i += scnprintf(buf + i, sz - i, " %u", n->sync_point);
i += scnprintf(buf + i, sz - i, " %d", n->link_cnt);
i += scnprintf(buf + i, sz - i, " %u", n->working_links);
i += scnprintf(buf + i, sz - i, " %x", n->capabilities);
i += scnprintf(buf + i, sz - i, " %lu\n", n->keepalive_intv);
if (!more)
return i;
i += scnprintf(buf + i, sz - i, "link_entry[0]:\n");
i += scnprintf(buf + i, sz - i, " mtu: %u\n", n->links[0].mtu);
i += scnprintf(buf + i, sz - i, " media: ");
i += tipc_media_addr_printf(buf + i, sz - i, &n->links[0].maddr);
i += scnprintf(buf + i, sz - i, "\n");
i += tipc_link_dump(n->links[0].link, TIPC_DUMP_NONE, buf + i);
i += scnprintf(buf + i, sz - i, " inputq: ");
i += tipc_list_dump(&n->links[0].inputq, false, buf + i);
i += scnprintf(buf + i, sz - i, "link_entry[1]:\n");
i += scnprintf(buf + i, sz - i, " mtu: %u\n", n->links[1].mtu);
i += scnprintf(buf + i, sz - i, " media: ");
i += tipc_media_addr_printf(buf + i, sz - i, &n->links[1].maddr);
i += scnprintf(buf + i, sz - i, "\n");
i += tipc_link_dump(n->links[1].link, TIPC_DUMP_NONE, buf + i);
i += scnprintf(buf + i, sz - i, " inputq: ");
i += tipc_list_dump(&n->links[1].inputq, false, buf + i);
i += scnprintf(buf + i, sz - i, "bclink:\n ");
i += tipc_link_dump(n->bc_entry.link, TIPC_DUMP_NONE, buf + i);
return i;
}
tipc: improve throughput between nodes in netns 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 6c9081a3915d ("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>
2019-10-29 07:51:21 +07:00
void tipc_node_pre_cleanup_net(struct net *exit_net)
{
struct tipc_node *n;
struct tipc_net *tn;
struct net *tmp;
rcu_read_lock();
for_each_net_rcu(tmp) {
if (tmp == exit_net)
continue;
tn = tipc_net(tmp);
if (!tn)
continue;
spin_lock_bh(&tn->node_list_lock);
list_for_each_entry_rcu(n, &tn->node_list, list) {
if (!n->peer_net)
continue;
if (n->peer_net != exit_net)
continue;
tipc_node_write_lock(n);
n->peer_net = NULL;
n->peer_hash_mix = 0;
tipc_node_write_unlock_fast(n);
break;
}
spin_unlock_bh(&tn->node_list_lock);
}
rcu_read_unlock();
}