1. 31 Jul, 2015 13 commits
    • David S. Miller's avatar
      Merge branch 'tipc-next' · 1df33a11
      David S. Miller authored
      Jon Maloy says:
      
      ====================
      tipc: separate link aggregation from link layer
      
      We continue the work on separating the roles of the link aggregation and
      link layers, as well as making code cleanups in general.
      
      This second commit batch focuses on moving the orchestration of link
      failover and synchronization to the node level, as well as preparing the
      node lock structure for further future impovements. We also make some
      changes to message delivery between link and socket layer, in order to
      make this mechanism safer and less obscure.
      ====================
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      1df33a11
    • Jon Paul Maloy's avatar
      tipc: clean up link creation · 440d8963
      Jon Paul Maloy authored
      We simplify the link creation function tipc_link_create() and the way
      the link struct it is connected to the node struct. In particular, we
      remove the duplicate initialization of some fields which are anyway set
      in tipc_link_reset().
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      440d8963
    • Jon Paul Maloy's avatar
      tipc: use temporary, non-protected skb queue for bundle reception · 9073fb8b
      Jon Paul Maloy authored
      Currently, when we extract small messages from a message bundle, or
      when many messages have accumulated in the link arrival queue, those
      messages are added one by one to the lock protected link input queue.
      This may increase contention with the reader of that queue, in
      the function tipc_sk_rcv().
      
      This commit introduces a temporary, unprotected input queue in
      tipc_link_rcv() for such cases. Only when the arrival queue has been
      emptied, and the function is ready to return, does it splice the whole
      temporary queue into the real input queue.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      9073fb8b
    • Jon Paul Maloy's avatar
      tipc: remove implicit message delivery in node_unlock() · 23d8335d
      Jon Paul Maloy authored
      After the most recent changes, all access calls to a link which
      may entail addition of messages to the link's input queue are
      postpended by an explicit call to tipc_sk_rcv(), using a reference
      to the correct queue.
      
      This means that the potentially hazardous implicit delivery, using
      tipc_node_unlock() in combination with a binary flag and a cached
      queue pointer, now has become redundant.
      
      This commit removes this implicit delivery mechanism both for regular
      data messages and for binding table update messages.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      23d8335d
    • Jon Paul Maloy's avatar
      tipc: make resetting of links non-atomic · 598411d7
      Jon Paul Maloy authored
      In order to facilitate future improvements to the locking structure, we
      want to make resetting and establishing of links non-atomic. I.e., the
      functions tipc_node_link_up() and tipc_node_link_down() should be called
      from outside the node lock context, and grab/release the node lock
      themselves. This requires that we can freeze the link state from the
      moment it is set to RESETTING or PEER_RESET in one lock context until
      it is set to RESET or ESTABLISHING in a later context. The recently
      introduced link FSM makes this possible, so we are now ready to introduce
      the above change.
      
      This commit implements this.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      598411d7
    • Jon Paul Maloy's avatar
      tipc: move received discovery data evaluation inside node.c · cf148816
      Jon Paul Maloy authored
      The node lock is currently grabbed and and released in the function
      tipc_disc_rcv() in the file discover.c. As a preparation for the next
      commits, we need to move this node lock handling, along with the code
      area it is covering, to node.c.
      
      This commit introduces this change.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      cf148816
    • Jon Paul Maloy's avatar
      tipc: merge link->exec_mode and link->state into one FSM · 662921cd
      Jon Paul Maloy authored
      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: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      662921cd
    • Jon Paul Maloy's avatar
      tipc: move protocol message sending away from link FSM · 5045f7b9
      Jon Paul Maloy authored
      The implementation of the link FSM currently takes decisions about and
      sends out link protocol messages. This is unnecessary, since such
      actions are not the result of any link state change, and are even
      decided based on non-FSM state information ("silent_intv_cnt").
      
      We now move the sending of unicast link protocol messages to the
      function tipc_link_timeout(), and the initial broadcast synchronization
      message to tipc_node_link_up(). The latter is done because a link
      instance should not need to know whether it is the first or second
      link to a destination. Such information is now restricted to and
      handled by the link aggregation layer in node.c
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      5045f7b9
    • Jon Paul Maloy's avatar
      tipc: move link synch and failover to link aggregation level · 6e498158
      Jon Paul Maloy authored
      Link failover and synchronization have until now been handled by the
      links themselves, forcing them to have knowledge about and to access
      parallel links in order to make the two algorithms work correctly.
      
      In this commit, we move the control part of this functionality to the
      link aggregation level in node.c, which is the right location for this.
      As a result, the two algorithms become easier to follow, and the link
      implementation becomes simpler.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      6e498158
    • Jon Paul Maloy's avatar
      tipc: extend node FSM · 66996b6c
      Jon Paul Maloy authored
      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: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      66996b6c
    • Jon Paul Maloy's avatar
      tipc: reverse call order for link_reset()->node_link_down() · 655fb243
      Jon Paul Maloy authored
      In many cases the call order when a link is reset goes as follows:
      tipc_node_xx()->tipc_link_reset()->tipc_node_link_down()
      
      This is not the right order if we want the node to be in control,
      so in this commit we change the order to:
      tipc_node_xx()->tipc_node_link_down()->tipc_link_reset()
      
      The fact that tipc_link_reset() now is called from only one
      location with a well-defined state will also facilitate later
      simplifications of tipc_link_reset() and the link FSM.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      655fb243
    • Jon Paul Maloy's avatar
      tipc: move all link_reset() calls to link aggregation level · 6144a996
      Jon Paul Maloy authored
      In line with our effort to let the node level have full control over
      its links, we want to move all link reset calls from link.c to node.c.
      Some of the calls can be moved by simply moving the calling function,
      when this is the right thing to do. For the remaining calls we use
      the now established technique of returning a TIPC_LINK_DOWN_EVT
      flag from tipc_link_rcv(), whereafter we perform the reset call when
      the call returns.
      
      This change serves as a preparation for the coming commits.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      6144a996
    • Jon Paul Maloy's avatar
      tipc: eliminate function tipc_link_activate() · cbeb83ca
      Jon Paul Maloy authored
      The function tipc_link_activate() is redundant, since it mostly performs
      settings that have already been done in a preceding tipc_link_reset().
      
      There are three exceptions to this:
      - The actual state change to TIPC_LINK_WORKING. This should anyway be done
        in the FSM, and not in a separate function.
      - Registration of the link with the bearer. This should be done by the
        node, since we don't want the link to have any knowledge about its
        specific bearer.
      - Call to tipc_node_link_up() for user access registration. With the new
        role distribution between link aggregation and link level this becomes
        the wrong call order; tipc_node_link_up() should instead be called
        directly as a result of a TIPC_LINK_UP event, hence by the node itself.
      
      This commit implements those changes.
      Tested-by: default avatarYing Xue <ying.xue@windriver.com>
      Signed-off-by: default avatarJon Maloy <jon.maloy@ericsson.com>
      Signed-off-by: default avatarDavid S. Miller <davem@davemloft.net>
      cbeb83ca
  2. 30 Jul, 2015 27 commits