JP4603088B2 - How to build a network - Google Patents

Description

The present invention relates to a network construction method and is suitable for an interface circuit conforming to IEEE 1394, which is a serial interface standard.
In recent years, with the development of multimedia, an increase in data transfer amount and an increase in transfer speed are required. In particular, as an interface for connecting a peripheral device such as a digital video camera or a color page printer that handles a large amount of audio and image data and a personal computer, a serial interface adopting the IEEE 1394 standard has attracted attention.

  In a packet transfer system compliant with IEEE 1394 or the like, when a bus reset occurs, bus initialization (Bus Initialize) is first performed. In other words, each node constituting the topology is provided with an interface circuit, and the device is provided with a state machine. Then, bus initialization is performed in the bus reset state (bus reset start to bus reset wait) in the state machine. In the bus initialization, each node erases information on the topology.

  Next, when the bus initialization is completed, tree identification is performed based on the tree ID state in the state machine. In tree identification, each node connected individually is identified so that it can be treated as a tree-like connection. Identifying in a tree form means selecting an appropriate root node and identifying a route in which other nodes are connected to branches.

  More specifically, as shown in FIG. 7, the tree ID state conforming to the IEEE 1394 draft includes a T0 state (tree ID start) 50, a T1 state (child handshake) 51, and a T2 state (parent handshake) 52. And a T3 state 53.

  When the processing of the bus reset wait R1 in the bus initialization is completed, the state transits to the T0 state 50 in the tree identification. The state machine waits until there is one port that has not received a parent notify signal (hereinafter referred to as a parent notification signal) in the T0 state 50, and transitions to the T1 state 51. Note that the parent notification signal conveys the desire to become a parent. When the node is a leaf node (Leaf), the T0 state 50 passes immediately because there is only one port connected to another node from the beginning.

  The state machine transmits a signal “child notify” (hereinafter referred to as a child notification signal) to the transmission source of the parent notification signal received in the T1 state 51, and transits to the T2 state 52. The child notification signal conveys that the child is recognized as the child.

  Next, the state machine transmits a parent notification signal to one port that has not received the parent notification signal in the T2 state 52, and waits until a child notification signal is received. When this process is finished, the tree identification is finished. The T3 state 53 is used for arbitration at the time of tree identification.

  Therefore, the parent notification signal is transmitted in order from the leaf node. The branch node receives a parent notification signal from each node. When there is one port that has not received the parent notification signal, the branch node transmits a child notification signal to the transmission source of the received parent notification signal, and transmits a parent notification signal from the remaining ports. This process is repeated, and finally the node that receives the parent notification signal from all the ports and sends back the child notification signal becomes the root, and each node recognizes the parent-child relationship.

  Next, when the tree identification is completed, self identification is performed by a self ID state (self ID start S0 or the like) in the state machine. In self-identification, each node selects a physical ID used by an address and creates management information related to bus connection. After this self-identification is completed, normal packet transfer can be performed between the nodes.

By the way, in the packet transfer method stipulated in the IEEE 1394 standard or the like, a loop topology is not allowed.
That is, the nodes constituting the loop do not have one port that has not received the parent notification signal, and are stacked in the T0 state 50. In addition, since the node that does not constitute the loop does not receive the child notification signal from the node that constitutes the loop, the node is stuck in the T2 state 52. When a node that does not constitute a loop is stacked for a predetermined time in the T2 state 52, the bus initialization process is attempted again. However, as long as the topology does not change, these processes are only repeated.

  In this case, the user needs to confirm the connection state of the node (device such as a personal computer, digital video camera, printer, etc.) and the bus cable, and reconnect to a topology without a loop. However, there is a problem that the operation is complicated when a large number of nodes are connected or when the nodes are connected across a plurality of rooms.

  The present invention has been made to solve the above problems, and even if a loop exists in the constructed topology, the loop is automatically resolved without the user changing the topology. An object of the present invention is to provide a network construction method that can be constructed.

According to one aspect of the present invention, to send diagnostic information to the network from the first port of the first node, said first node, said diagnostic information which the second node transmits to said network, said When received at a second port different from the first port, it is determined that there is a loop including the first node and the second node in the network, and the first port and the second port Is deactivated with respect to the network, and information on the deactivated port is stored in a register included in the first node, and the register holds data when the network is initialized. Ru is set to.

As described above in detail, according to the inventions described in claims 1 to 5 , even if a loop exists in the constructed topology, the loop is automatically canceled without the user changing the topology. A network construction method capable of constructing a topology can be provided.

FIG. 3 is a block circuit diagram of an interface circuit of the present embodiment. Explanatory drawing which shows the tree ID state of this Embodiment. The flowchart of the disconnection process performed in a Tx state. (A)-(g) Explanatory drawing for demonstrating the disconnection process of this Embodiment. (A)-(j) Explanatory drawing for demonstrating the disconnection process of another example. (A)-(d) Explanatory drawing for demonstrating the disconnection process of another example. Explanatory drawing which shows the tree ID state of a prior art.

  Hereinafter, an embodiment of the present invention will be described with reference to FIGS. In the present embodiment, components having the same configurations as those of the prior art are denoted by the same reference numerals, and a detailed description thereof is partially omitted.

  FIG. 4A shows a system configuration (topology) in which a plurality of nodes A1 to D1 are connected by IEEE 1394 bus cables (hereinafter simply referred to as bus cables) 1a to 1d. The nodes A1 to D1 are generic names of connection points such as a personal computer, a printer, a digital video camera, and a set top box.

  Node B1 is connected to node D1 via bus cable 1a. Node B1 is connected to node B1 via bus cable 1b, node A1 is connected to node A1 via bus cable 1c, and node B1 is connected to node A1 via bus cable 1d. In the topology constructed in this way, a loop is constituted by the nodes A1 to C1 and the bus cables 1b to 1d.

The node A1 includes the interface circuit 11 shown in FIG. The other nodes B1 to D1 also include the interface circuit 11 like the node A1.
The interface circuit 11 includes a physical layer processing circuit 12, a link layer processing circuit 13, a host computer (not shown) (hereinafter referred to as a host), and the like. The physical layer processing circuit 12 includes a port unit 14, an analysis circuit 15, a generation circuit 16, and a state machine 17. The port unit 14 includes a port 18, a decoder 19, an encoder 20, and a port control unit 21. In the interface circuit 11 (FIG. 1) of the present embodiment, only one port 18 is illustrated, but up to eight ports 18 are provided in one node, and the port unit 14 is parallel to each port 18. Is provided.

  Packets from other nodes are input to the state machine 17 via the port 18, the decoder 19 and the analysis circuit 15. Then, the state machine 17 outputs the packet to the link layer processing circuit 13. At this time, if there is another connected node, the state machine 17 transmits the packet to the node via the generation circuit 16, the encoder 20, and the port 18.

  The state machine 17 receives a packet from the link layer processing circuit 13. Then, the state machine 17 transmits the packet to another node via the generation circuit 16, the encoder 20 and the port 18.

  When determining that the link layer processing circuit 13 is addressed to itself based on the header information of the packet input from the state machine 17 (physical layer processing circuit 12), the link layer processing circuit 13 supplies the data in the packet to a host or the like. The link layer processing circuit 13 outputs the data received from the host to the state machine 17 (physical layer processing circuit 12).

  The state machine 17 includes the tree ID state 31 shown in FIG. 2, the bus reset state (bus reset start R0 to bus reset wait R1) and the self ID state (self ID start S0, etc.) described in the prior art. Yes.

  The tree ID state 31 includes a T0 state (tree ID start) 50a, a T1 state (child handshake) 51a, a T2 state (parent handshake) 52a, a T3 state 53, and a Tx state 32 as a diagnostic state. Prepare.

  In the tree ID state 31, when the state stacked in the T0 state 50a, the T1 state 51a, and the T2 state 52a continues for a predetermined time, the state machine 17 transitions from these states to the Tx state 32.

In the Tx state 32, the state machine 17 performs a disconnection process according to the flowchart shown in FIG.
In step S1, the state machine 17 determines whether or not its own node is a specific node, in this embodiment, an initiate node. The initiate node is a node that has started itself and has generated a bus reset.

When the state machine 17 determines that its own node is an initiate node, the state machine 17 proceeds to step S2.
In step S2, the state machine 17 determines whether there is an unmarked (unidentified) port 18 of its own. The mark is made when a diagnostic packet SP described later is transmitted / received, and all ports are not marked at first. If the state machine 17 determines that there is an unidentified port 18 in itself, the state machine 17 proceeds to step S3.

  In step S3, the state machine 17 transmits the diagnostic packet SP from the unidentified port 18, marks the port 18, and proceeds to step S4. In the present embodiment, when there are a plurality of unidentified ports 18, one of them is arbitrarily selected, the diagnostic packet SP is transmitted, and it is marked.

On the other hand, when the state machine 17 determines that there is no unidentified port 18 in step S2, the process proceeds to step S5.
In step S5, the state machine 17 transmits the diagnostic packet SP from the identified port 18, marks the port 18, and proceeds to step S4.

On the other hand, if it is determined in step S1 that the own node is not an initiate node, the process proceeds to step S4.
In step S4, the state machine 17 determines whether or not the diagnostic packet SP has been received. If the state machine 17 determines that the diagnostic packet SP has not been received, the state machine 17 proceeds to step S6.

  In step S6, the state machine 17 determines whether or not the time during which the diagnostic packet SP has not been received has exceeded a predetermined time. If the state machine 17 determines that the predetermined time has not been exceeded, the process returns to step S4. If the state machine 17 determines that the predetermined time has been exceeded, the state machine 17 generates a bus reset and transitions to the bus reset state (bus reset start R0) described in the prior art. It should be noted that the predetermined time set in step S6 is set to a length that is reached in the event of an abnormality such as transmission / reception of the diagnostic packet SP being interrupted for some reason. That is, in this step S6, it is prevented from being stuck by the process of step S4 at the time of abnormality.

On the other hand, when the state machine 17 determines in step S4 that the diagnostic packet SP has been received, the process proceeds to step S7.
In step S7, the state machine 17 marks the port 18 that has received the diagnostic packet SP, and proceeds to step S8.

  In step S8, the state machine 17 determines whether or not the diagnostic packet SP is transmitted in step S3 or step S5. When the state machine 17 determines that the diagnostic packet SP is not transmitted, the state machine 17 returns to step S2.

On the other hand, if the state machine 17 determines in step S8 that the diagnostic packet SP has been transmitted, the process proceeds to step S9.
In step S9, the state machine 17 determines whether or not the port 18 marked in step S7 is the same as the port 18 marked in step S3. That is, it is determined whether or not the port 18 that has transmitted the diagnostic packet SP is the same as the port 18 that has received the diagnostic packet SP. If it is determined that the ports 18 marked in steps S7 and S3 are the same, that is, if it is determined that the ports 18 that have transmitted and received the diagnostic packet SP are the same, the process returns to step S2.

  On the other hand, if the state machine 17 determines in step S9 that the port 18 marked in steps S7 and S3 is different, that is, if it determines that the port 18 that transmitted the diagnostic packet SP and the received port 18 are different, both of them. It is determined that there is a loop including the bus cable connected to the port 18, and the process proceeds to step S10.

  In step S10, the state machine 17 outputs a disconnection signal Z for electrically disconnecting one of the ports 18 marked in steps S3 and S7 to the port control unit 21. The disconnection signal Z of the present embodiment is used to electrically disconnect the port 18 marked in step S7, that is, the port 18 that has received the diagnostic packet SP.

Next, when the process of step S10 is completed, a bus reset is generated, and the state transits to the bus reset state (bus reset start R0) described in the prior art.
The port control unit 21 electrically disconnects the bus cable connected to the port 18 based on the disconnection signal Z from the state machine 17 and sets the disconnection state. The port control unit 21 is composed of a register that is not initialized when the bus is initialized.

The operation of the interface circuit 11 configured as described above will be described.
In this embodiment, a case will be described in which the node D1 is an initiate node that generates a bus reset in the topology shown in FIG.

  Now, a bus reset is generated starting from the node D1. Then, in each node A1 to D1, bus initialization (Bus Initialize) is first performed in the bus reset state (bus reset start R0 to bus reset wait R1). Thereby, the information regarding the topology of each of the nodes A1 to D1 is deleted.

  Next, in each of the nodes A1 to D1, when the bus initialization is completed, tree identification (Tree Identify) is performed by the tree ID state 31. That is, when the processing of the bus reset wait R1 in the bus initialization is completed, the process transits to the T0 state 50a in the tree identification and the processing is performed.

Then, as shown in FIG. 4B, the node D1, which is a leaf node, immediately transitions to the T2 state 52a, and the parent notification signal PN is transmitted to the node B1.
In the nodes A1 to C1, the number of ports that have not received the parent notification signal PN is not one, so that the stack is made in the T0 state 50a. Further, since the node D1 does not receive the child notification signal from the node B1, it is stacked in the T2 state 52a.

Then, in each of the nodes A1 to D1, the state stacked in the T0 state 50a or the T2 state 52a continues for a predetermined time, transitions to the Tx state 32, and the processing is performed.
In the node D1, it is determined in the process of step S1 that it is an initiate node, and then in the process of step S2, it is determined that there is an unmarked port 18 (unidentified).

Then, as shown in FIG. 4C, the diagnostic packet SP is output from the unidentified port 18 to the node B1 in the process of step S3, and the port 18 is marked.
On the other hand, it is determined that the nodes A1 to C1 are not initiate nodes in the process of step S1.

The node B1 determines that the diagnostic packet SP is received from the node D1 in the process of step S4, and the port 18 is marked in the process of step S7.
Then, the node B1 determines that its own node B1 has not transmitted the diagnostic packet SP in the process of step S8, and determines that there is an unidentified port 18 in the process of step S2.

  Then, from the node B1, as shown in FIG. 4D, the diagnostic packet SP is output from the unidentified port 18 to the node C1 in the process of step S3, and the port 18 is marked. At this time, the unidentified port 18 includes one connected to the node A1 and one connected to the node C1, but in this embodiment, the port 18 on the node C1 side is arbitrarily selected. To do.

Then, in the node C1, the same processing as that of the node B1 is performed in the processing of step S4, step S7, step S8, and step S2.
Then, as shown in FIG. 4E, the node C1 outputs the diagnostic packet SP from the unidentified port 18 to the node A1 in the process of step S3, and the port 18 is marked.

Then, in the node A1, the same processing as that of the node C1 is performed in the processing of step S4, step S7, step S8, and step S2.
Then, as shown in FIG. 4F, the node A1 outputs the diagnostic packet SP from the unidentified port 18 to the node B1 in the process of step S3, and the port 18 is marked.

Then, in the node B1, it is determined that the diagnostic packet SP is received from the node A1 in the process of step S4, and the port 18 is marked in the process of step S7.
The node B1 then determines that its own node B1 has transmitted the diagnostic packet SP in the process of step S8. Then, it is determined in step S9 that the port 18 that transmitted the diagnostic packet SP is different from the received port 18, and it is determined that both the ports 18 constitute a loop.

  Then, the disconnection signal Z for electrically disconnecting the bus cable 1d connected to the port 18 that has received the diagnostic packet SP in the process of step S10 is sent from the state machine 17 in the node B1 to the port controller 21. Is output.

  Then, as indicated by a broken line in FIG. 4G, the bus cable 1d connected to the port 18 that has received the diagnostic packet SP by the port control unit 21 is electrically disconnected. This eliminates the topology loop.

Then, a bus reset is generated by the node B1, and the bus initialization is performed again at each of the nodes A1 to D1.
Next, when the bus initialization is completed, tree identification is normally performed by the tree ID state 31. Thereby, the root node is determined, and the parent-child relationship is recognized by each of the nodes A1 to D1.

  Next, when the tree identification is completed, self identification is performed by the self ID state. Thereby, management information relating to bus connection is created. Therefore, normal packet transfer is enabled between the nodes A1 to D1 thereafter.

Next, characteristic operation and effects of the interface circuit 11 configured as described above and its system will be described below.
(1) When a loop exists in the topology, the port 18 constituting the loop is automatically determined, the port 18 is automatically disconnected, and the loop is automatically canceled. As a result, the topology is reliably constructed, and packet transfer between the nodes can be automatically performed. That is, even if a topology having a loop as shown in FIG. 4 (a) is constructed, the user is troublesome to confirm the connection state of the nodes A1 to D1 and the bus cables 1a to 1d and reconnect to the topology having no loop. No need to perform tedious work. In other words, the user does not have to construct the topology carefully so as not to loop.

  (2) Since the Tx state 32 is added to the tree ID state 31 as compared with the conventional IEEE 1394 draft, the tree ID state 31 can be easily created without significantly changing the conventional product.

(3) Since the port unit 14 includes only the port control unit 21 compared to the conventional product, it can be easily created without significantly changing the conventional product.
(4) Since only the node D1, which is the initiate node, outputs the diagnostic packet SP first, the diagnostic packet SP is not output simultaneously from a plurality of nodes. Thereby, malfunction is prevented.

The above embodiment may be modified as follows.
In the above embodiment, the topology is composed of four nodes A1 to D1 and four bus cables 1a to 1d and has one loop (see FIG. 4A). However, there are many nodes and bus cables. However, it may be implemented in a complex topology having a plurality of loops.

  For example, you may implement in the topology shown to Fig.5 (a). Node B2 is connected to node A2 via bus cable 1e, and nodes C2 and D2 are connected to node B2 via bus cables 1f and 1g. Node D2 is connected to node C2 via bus cable 1h. Node E2 is connected to node D2 via bus cable 1i, and node E2 is connected via bus cable 1j. Further, the node E2 is connected to the node E2 via the bus cable 1k. In the topology constructed in this way, the first loop is configured by the nodes B2 to D2 and the bus cables 1f to 1h, and the second loop is configured by the nodes D2 and E2 and the bus cables 1i and 1j. Each node A <b> 2 to F <b> 2 includes an interface circuit 11.

In this example, a case where the node C2 is an initiate node will be described.
Now, a bus reset is generated starting from the node C2. Then, in each of the nodes A2 to F2, bus initialization is performed, and then tree identification is performed. Then, the nodes B2 to E2 are stacked in the T0 state 50a, and the nodes A2 and F2 are stacked in the T2 state 52a. Then, each of the nodes A2 to F2 transitions to the Tx state 32 and is processed.

  First, as shown in FIG. 5A, the diagnostic packet SP is transmitted from the unidentified port 18 of the node C2, which is the initiate node, to the node B2, and the port 18 is marked (step S3). At this time, it is assumed that the port 18 on the node B2 side is arbitrarily selected as the port 18 to which the diagnostic packet SP is transmitted.

  Then, as shown in FIG. 5B, the diagnostic packet SP is transmitted from the unidentified port 18 of the node B2 to the node A2, and the port 18 is marked (step S3). At this time, it is assumed that the port 18 on the node A2 side is arbitrarily selected as the port 18 to which the diagnostic packet SP is transmitted.

  Then, as shown in FIG. 5C, since the node A2 does not have the unidentified port 18, the diagnostic packet SP is transmitted from the identified port 18 to the node B2, and the port 18 is marked ( Step S5).

Then, in the node B2, since it is determined that the port 18 that transmits / receives the diagnostic packet SP is the same, it is determined that the port 18 does not constitute a loop (step S9).
Then, as shown in FIG. 5D, the diagnostic packet SP is transmitted from the unidentified port 18 of the node B2 to the node D2, and the port 18 is marked (step S3).

  Then, as shown in FIG. 5E, the diagnostic packet SP is transmitted from the unidentified port 18 connected to the bus cable 1i of the node D2 to the node E2, and the port 18 is marked (step S3). ). At this time, it is assumed that the port 18 to which the diagnostic packet SP is transmitted is arbitrarily selected from the port 18 connected to the bus cable 1i.

  Then, as shown in FIG. 5 (f), the diagnostic packet SP is transmitted from the unidentified port 18 connected to the bus cable 1j of the node E2 to the node D2, and the port 18 is marked (step S3). ). At this time, the port 18 to which the diagnostic packet SP is transmitted is arbitrarily selected from the port 18 connected to the bus cable 1j.

Then, in the node D2, it is determined that the ports 18 that transmit / receive the diagnostic packet SP are different, and it is determined that both the ports 18 constitute a loop (step S9).
Then, a disconnection signal Z for electrically disconnecting the bus cable 1j from the state machine 17 of the node D2 is output to the port control unit 21 (step S10), and as shown by a broken line in FIG. The bus cable 1j is electrically disconnected by the control unit 21. This eliminates the second loop.

  Then, a second bus reset is generated starting from the node D2. At this time, since the first loop is not eliminated, the nodes B2 to D2 are stacked in the T0 state 50a, and the nodes A2, E2, and F2 are stacked in the T2 state 52a. Then, each of the nodes A2 to F2 transitions to the Tx state 32 and is processed.

  Then, as shown in FIG. 5G, the diagnostic packet SP is transmitted from the unidentified port 18 of the node D2, which is the initial node at this time, to the node B2, and the port 18 is marked (step S3). ). At this time, it is assumed that the port 18 on the node B2 side is arbitrarily selected as the port 18 to which the diagnostic packet SP is transmitted.

  Then, as shown in FIG. 5 (h), the diagnostic packet SP is transmitted from the unidentified port 18 of the node B2 to the node C2, and the port 18 is marked (step S3). At this time, it is assumed that the port 18 on the node C2 side is arbitrarily selected as the port 18 to which the diagnostic packet SP is transmitted.

  Then, as shown in FIG. 5 (i), the diagnostic packet SP is transmitted from the unidentified port 18 of the node C2 to the node D2, and the port 18 is marked (step S3).

Then, in the node D2, it is determined that the ports 18 that transmit / receive the diagnostic packet SP are different, and it is determined that both the ports 18 constitute a loop (step S9).
Then, a disconnection signal Z for electrically disconnecting the bus cable 1h from the state machine 17 at the node D2 is output to the port control unit 21 (step S10), and as shown by a broken line in FIG. The bus cable 1h is electrically disconnected by the control unit 21. This eliminates the first loop.

  Then, the third bus reset is generated starting from the node D2. At this time, since the first and second loops are eliminated, bus initialization, tree identification, and self-identification are normally performed thereafter, and normal packet transfer is enabled between the nodes A2 to F2. As described above, even in a complicated topology having a plurality of loops, the ports 18 constituting each loop are automatically disconnected, and the same effect as in the above embodiment can be obtained.

  In the case where there are a plurality of unidentified ports 18 in step S3 of the Tx state 32, the state machine 17 of the above embodiment arbitrarily selects one of them and transmits the diagnostic packet SP. Priorities may be set in advance and selected in that order. In this way, when the loop is resolved, the port side with the higher priority remains and the port side with the lower priority is disconnected.

  In the above-described embodiment, the description has been made on the construction of the topology, but the same operation is performed when the topology is changed after the loop is eliminated. That is, when any abnormality occurs in the port 18 or the bus cable on the side that has not been disconnected, the state machine 17 detects a change in topology and starts a bus reset. The cable may be again electrically connected, and the port 18 on the side that is not disconnected may be disconnected. The abnormality includes, for example, when the bus cable is physically disconnected, or when an error of a predetermined amount or more is detected during packet transfer between the nodes. In this way, even if any abnormality occurs in the port 18 or bus cable on the side that has not been disconnected after the loop is eliminated, the connection is automatically switched to the port 18 and bus cable on the disconnected side.

  -When constructing a topology, create a loop intentionally in advance, and after the loop is resolved as described above, if any abnormality occurs on the side that was not disconnected, the connected side is electrically connected again. It is good also as a state and it is good also considering the side which was not made into the disconnection state as a disconnection state. For example, the topology shown in FIG.

  Node B3 is connected to node A3 via bus cable 1p, and node B3 is connected to node C3 via bus cable 1q. The node D3 is connected to the node B3 through three bus cables 1r, 1s, and 1t, respectively. The node D3 is connected to the node D3 via the bus cable 1u, and the node F3 is connected to the node D3 via the bus cable 1v. Node E3 is connected to node F3 via bus cable 1w. Here, a high priority is set for both ports 18 using the bus cable 1r as a main path.

  Then, by turning on the power of each of the nodes A3 to G3, when a bus reset is generated starting from the node C3, the diagnostic packet SP is transmitted from C3 to B3, B3, as shown in FIG. 6B, for example. To D3, D3 to G3, G3 to D3, and D3 to B3. At this time, the diagnostic packet SP transmitted from the node B3 to the node D3 is transmitted via the bus cable 1r because it passes through the port 18 having a high priority, and the diagnostic packet SP transmitted from the node D3 to the node B3 is, for example, It is transmitted via the bus cable 1t.

Then, in the node B3, it is determined that the port 18 that transmits / receives the diagnostic packet SP forms a loop, and the received bus cable 1t on the port 18 side is electrically disconnected.
Then, a second bus reset is generated starting from the node B3. Then, for example, as shown in FIG. 6C, the diagnostic packet SP is transmitted in the order of B3 to D3, D3 to F3, F3 to E3, E3 to F3, F3 to D3, and D3 to B3. At this time, the diagnostic packet SP transmitted from the node B3 to the node D3 is transmitted via the bus cable 1r because it passes through the port 18 having a high priority, and the diagnostic packet SP transmitted from the node D3 to the node B3 is the bus. It is transmitted via the cable 1s.

  Then, in the node B3, it is determined that the port 18 that has transmitted / received the diagnostic packet SP forms a loop, and the received bus cable 1s on the port 18 side is electrically disconnected. This eliminates the loop. In addition, at this time, the bus cable 1r, which is the main path between the ports 18 with high priority, remains connected.

Then, bus initialization, tree identification, and self-identification are normally performed, and normal packet transfer is enabled between the nodes A3 to G3.
Thereafter, as shown in FIG. 6D, for example, when the bus cable 1r as the main path is physically disconnected, the bus machine 1s is again connected by the state machine 17 of the node B3. As a result, the packet transfer between the node B3 and the node D3 is automatically enabled again. Thus, for example, if a loop is intentionally created in advance between the node B3 and the node D3 that are difficult to maintain, even if an abnormality occurs between them, the packet transfer between them is easily resumed in a short time.

  In the above embodiment, the Tx state 32 is provided in the tree ID state 31 of the state machine 17 and the port control unit 21 is controlled by the processing in the Tx state 32. However, the processing similar to that in the Tx state 32 (shown in FIG. 3) The same control may be performed in terms of software using a recording medium (software) on which program data capable of executing the flowchart is recorded.

  In the above embodiment, the initial diagnostic packet SP is transmitted with the initiate node as a specific node, but other nodes may be set as specific nodes. Alternatively, the first diagnostic packet SP may be transmitted to an arbitrary node.

The various embodiments described above are summarized as follows.
(1) A topology construction method in which a plurality of nodes are connected by a bus cable, wherein the loop location is automatically determined, the location is automatically disconnected, and the loop is eliminated. How to build a topology.

  (2) In the topology construction method described in (1) above, in the tree ID state conforming to a draft of a predetermined standard, when a loop exists in the topology, a diagnostic state that transitions after stacking for a predetermined time is provided, and the diagnostic state A topology construction method characterized in that the loop location is determined.

  (3) In the topology construction method according to the above (1) or (2), when the loop exists, the specific or arbitrary node transmits a diagnostic packet or a diagnostic signal and marks the transmitted port. The node that received the diagnostic packet or diagnostic signal marks the received port, and if there is an unmarked port, transmits the diagnostic packet or diagnostic signal from the port and marks the transmitted port, When the node that has transmitted the diagnostic packet or diagnostic signal receives the diagnostic packet or diagnostic signal from an unmarked port, the node determines that the transmitted port and the received port constitute the loop, and the transmission port or the reception port A method for constructing a topology, characterized in that a disconnection state is established.

  (4) In the topology construction method according to any one of the above (1) to (3), a location where the loop is formed by electrically disconnecting a bus cable connected to a port is defined as a disconnected state. A topology construction method characterized by:

(5) The topology construction method according to any one of (3) to (4) above, wherein the specific or arbitrary node is an initiate node.
(6) In the topology construction method according to any one of (3) to (5), a node that has received the diagnostic packet or diagnostic signal has a preset priority when there are a plurality of unmarked ports. A topology construction method, wherein the diagnostic packet or diagnostic signal is transmitted from a port having a higher rank.

  (7) In the topology construction method according to any one of (1) to (6), when an abnormality occurs in a location that is not in the disconnected state, the disconnected portion is electrically connected again. Then, the topology construction method is characterized in that a portion that is not in the disconnected state is in a disconnected state.

(8) The topology construction method according to (7), wherein a loop is created in advance when the topology is constructed.
(9) The topology construction method according to any one of (1) to (8), wherein the disconnected node causes a bus reset next.

  (10) In the topology construction method described in the above (1), (3), (4) to (7), (9), each process when the loop exists is performed by software control. A feature topology construction method.

  (11) An interface circuit provided in a node constituting a topology connected by a bus cable, when a tree ID state is changed according to a draft of a predetermined standard, and is in a stacked state for a predetermined time in a predetermined state A state machine that automatically executes a diagnostic state that automatically determines a portion that constitutes a loop, and a port controller that electrically disconnects the portion that constitutes the loop based on the determination of the diagnostic state. An interface circuit characterized by that.

  (12) In the interface circuit according to (11) above, when the diagnostic state is a specific node, the diagnostic state transmits a diagnostic packet or a diagnostic signal and marks the port where the diagnostic state is transmitted. Regardless of whether it is a node or not, when the diagnostic packet or diagnostic signal is received, the received port is marked, and if there is an unmarked port, the diagnostic packet or diagnostic signal is transmitted from the port and the transmission is performed. When the diagnostic packet or diagnostic signal is received from an unmarked port after the diagnostic packet or diagnostic signal is transmitted after the diagnostic packet or diagnostic signal is transmitted, it is determined that the transmitted port and the received port constitute the loop. An interface circuit characterized by

(13) The interface circuit according to (12), wherein the specific node is an initiate node.
(14) In the interface circuit according to (12) or (13) above, when the diagnosis state receives a diagnosis packet or a diagnosis signal and there are a plurality of unmarked ports, a preset priority is set. An interface circuit which transmits the diagnostic packet or diagnostic signal from a high port.

  (15) In the interface circuit according to any one of (11) to (14), when an abnormality occurs in a location that is not in the disconnected state, the disconnected portion is electrically connected again. An interface circuit characterized in that a portion that is not in the disconnected state is in a disconnected state.

  (16) The interface circuit according to any one of (11) to (15), wherein a bus reset is generated after the disconnection state.

11 Interface Circuit 17 State Machine 18 Port 21 Port Control Unit 31 Tree ID State 32 Tx State (Diagnostic State)
1a-1k, 1q-1w Bus cable A1-D1, A2-F2, A3-G3 Node

Claims (5)

A network construction method including a plurality of nodes,
Sending diagnostic information from the first port of the first node to the network;
Wherein said first node, said diagnostic information which the second node transmits to the network, when received by the first port a second port different from, said in the network the first node It is determined that there is a loop including the second node,
Either one of the first port and the second port is made inactive with respect to the network, and information on the port made inactive is stored in a register included in the first node,
The network construction method , wherein the register is set to hold data when the network is initialized . 2. The network according to claim 1, wherein when the disconnection in the loop is detected when the first port is inactive to the network, the first port is made active for the network. How to build. 3. The network construction method according to claim 1, wherein the inactivity is an electrical disconnection state with respect to the network. The network construction method according to claim 2, wherein the activity is an electrical connection state with respect to the network.   The protocol used for constructing the network updates identifiers assigned to the plurality of nodes when detecting connection and disconnection to the ports of the plurality of nodes.
The network construction method according to any one of claims 1 to 4, wherein the network is constructed.

Images