KR930002137B1 - Method for processing e-bus protocol - Google Patents

Abstract

The method controls E-bus protocol to work efficiently and perceives the board for a host to control and the status of LT's. It is composed of three stages; the 1st stage for referring to the status of a request command signal (REG-F) from a slave terminal, the 2nd stage for performing abnormal group falling when the REG-F signal is set, and the 3rd stage for performing normal group falling when the REG-F signal is reset.

Description

How to handle E bus protocol

1 is a structural diagram of an E bus to which the present invention is applied.

2 is a state transition diagram of a slave node viewed from a host.

3 (a) and 3 (b) are flowcharts of a method for processing an E bus protocol of a slave according to the present invention.

4 is a schematic polling flowchart of an E bus protocol processing method according to the present invention.

Figure 5 is an embodiment according to the E bus protocol processing method according to the present invention.

* Explanation of symbols for main parts of the drawings

11: E bus 12: E bus host

13: E bus slave a, b, c: transmission loop

The present invention relates to an E-bus protocol processing method used to process D-channel information in an ISDN subscriber module.

In order to process D channel in ISDN subscriber module, PASB or BASB board and CPB board are needed, and bus and bus controller for connecting these boards is needed. At this time, E bus was proposed as a new bus structure to connect multiple PASB or BASB boards to one CPB board. The CPB must always know the status of the PASB or BASB boards it manages and the status of the LTs maintained within each board. PASB or BASB boards have more than one LT.

It is an object of the present invention to provide an E-bus protocol processing method for enabling an E-bus to operate efficiently and for a host to know a state of a board and an LT to be controlled.

In order to achieve the above object, the present invention provides a clock line for supplying a control clock to a master terminal and a slave terminal, a first data line consisting of an RS-485 loop for supplying data from the master terminal to the slave terminal, and E-bus protocol processing method consisting of a second data line consisting of an RS-485 loop for supplying data from the slave end to the master end to exchange D-channel signals at one master end and a plurality of slave ends The first step of searching for the status of the registration request signal of the slave end, the second step of performing and processing an abnormal group polling when the registration request signal is set, and the normal group polling when the registration request signal is reset It is characterized by being carried out by a third step of performing and processing.

Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

1 is a structural diagram of an E bus to which the present invention is applied, where 11 is an E bus, 12 is an E bus host, 13 is an E bus slave, and a, b, and c are transmission loops, respectively.

The E bus is a full duplex, implemented as an RS-485 loop, which controls two unidirectional data transmission lines (b, c), each transferring data from slave to master or from master to slave and from master to slave, respectively. It consists of a 2MHz clock transmission loop (a) for transmitting the clock.

The E bus connects up to 16 slave nodes with the CPB, and transfers data from the BASB or PASB to the CPB, or from the CPB to the BASB or PASB. The IOB must prevent data loss during transmission on the E bus and manage the slaves through the management of the E bus. Various methods can be used to prevent data loss during transmission, but polling method that minimizes the degradation of system performance, facilitates instantaneous recovery, and does not require additional hardware even under some malfunction of slave nodes. Considering the advantages of the polling (Polling) method.

For effective polling, the 16 buses are divided into three groups on the E bus. The first is an off-board group where no board is mounted, and the second is an on-borard state in which the board is currently mounted in a slot or an abnormal board (node that does not serve as a slave). The third group is the normal board (node) group in which the board is currently mounted in the slot and is operating normally.

The distinction between on board and off board is determined by an on-bordflag (OB-F) connected from the slave through the back panel. The master does not poll nodes that do not have the on-board flag set. Nodes that do not respond to polling a certain number of times in the on-board group are transferred to an abnormal group and no further polling is performed. In order for a fault board to be assigned polling, that is, to change from a faulty group to a normal group, the master sets a registration flag (REG-F) common to all slave boards.

2 is a state transition diagram of a slave node viewed from a host. When the board is inserted in the off state in which the board is hermetically inserted, the board transitions to an abnormal state (f). After the registration request flag is set in the abnormal state, the terminal transitions to the normal state if it responds to the polling request of the master (d). The slave node determines that it is an abnormal board when it is not assigned polling for 200msec. After detecting the registration request, the master gives three polling opportunities to the abnormal group and receives registration until the registration request flag is removed.

The protocol on the E bus basically consists of a pair of 'query-responses' that the slave responds to the IOB's request. In other words, if the slaves are all normal, the slave responds to the IOB once query and decides the response message based on its own status.

The commands used on the E bus are shown in Table 1 below.

During the primitive, SM-DATA and SM-POLL provide a format for recording the LT state of the slave board so that the master can know the LT state in all slave nodes in the normal state.

When the host queries, two decisions are needed.

The first is the question of what command to send. That is, whether to send MS-POLL or MS-DATA. The second problem is which node to send MS-POLL to.

There are three main types of IOB queries. The first is an attempt to change to a normal board by responding to an MS-POLL request of an IOB for an abnormal board. This is the case when the slave in normal state responds to the MS-POLL request of the IOB.

In the second case of the three types of requests, the MS-DATA is executed irregularly because the MS-DATA is output only when there is a packet to transmit. However, a check to see if there is a transport packet (MS-DATA) is searched each time just before sending the first or third type of request. MS-DATA can be output regardless of the current polling node value. If there are always packets to be output in the IOB, the second request, MS-DATA, will always be printed after the first and third types of requests. Before sending the output packet to the slave as MS-DATA, IOB checks the slave node to output and if the node is not a normal node, abandon the output. If time-out occurs, repeat the message up to 3 times and output the same message as MS-DATA.

However, the first and third types of requests are made periodically and this entire cycle is called the polling cycle.

The slave decides the type of response message based only on its own state (there is no data to send to the IOB). That is, it is irrelevant to the state of the host node and the type of command sent from the host to the slave. If there is data in the E bus bidirectional for a certain node at any one time, piggyback can be made by responding to the MS-DATA request of the host to the SM-DATA of the slave node. The SM-DATA delivered to the IOB within the time limit is regarded as a legitimate response to all types of queries and forwarded to the CPB.

3 (a) and 3 (b) are flowcharts illustrating a method of processing an E bus protocol of a slave according to the present invention, (a) is a flowchart of an initial state of an E bus slave, and (b) is of a normal state of an E bus slave. It is a flow chart.

The E bus slave initializes the software and hardware as shown in FIG. 3 (a) in the initial state, triggers the WDT (Watch dog Timer) signal, and sets the registration request flag (REG-F) to 1 (100, 110, 120, 130). ).

When the E bus slave is in a normal state, for the request of the host, the slave determines the type of message to respond to whether there is data to send or not. When MS-DATA is transmitted from the master, the transmission packet is stored in the transmission buffer and the WDT signal is triggered (150, 170, 180). In addition, when MS-POLL is transmitted from the master, the WDT signal is directly triggered without going through the step of storing the transmission packet (140, 180). After triggering the WDT signal, the registration request flag (REG-F) is set to "0", and it is checked whether a transmission packet exists in the transmission buffer (190, 200). If there is no transmission packet, it responds to the master with SM-POLL (220). If there are two transmission packets, it searches for two or more transmission packets (210). (240) If not more than two, respond with SM-DATA (F = 0) (230) update the transmit buffer (250) and return to the E bus slave normal state. That is, if there is no packet to send to IOB, it responds with SM-POLL. If there is a packet to send to IOB, it responds with SM-DATA.

When sending an SM-DATA, whether or not there is an extra packet to be sent in the current slave node is included in the SM-DATA, and if there is an F = 1, if not, it is reported as an IOB.

When the watchdog timer enters the fault in the E bus slave normal state, it triggers the WDT signal and sets the registration request flag (REG-F) to "0" and enters the E bus slave initial state (160, 260, 270).

Full polling consists of reading and status output of onboard flag values, polling for abnormal groups and polling for normal groups. The abnormal group polling is performed only when the registration request flag (REG-F) is set. This is performed to check the node number of the abnormal node to be converted from the abnormal state to the normal state. Normal node polling is performed to receive SM-DATA while preventing data collisions between slave nodes on the E bus.

Polling for anomalous groups is done once for one node and then for anomalous nodes if there is no response. If there is no response for all abnormal nodes, up to three abnormal group polls are performed. Abnormal group polling is performed until the registration request flag (REG-F) is cleared, polling is performed three times in total, or there is no abnormal node. If a response from the slave is sent to the IOB within a predetermined time for the MS-POLL request of the IOB, the IOB transfers the abnormal node to the normal group regardless of the response type.

Polling for a normal group may respond with an SM-POLL indicating three consecutive timeouts for a node or with a slave indicating that there are no packets to transmit, or SM-DATA (F = 0) indicating that there are no more packets to transmit. Repeat until. If three timeouts occur in succession, the error count is set once (if the request for MS-DATA occurs, the error count is increased even if the timeout occurs three times in succession). Transfer the slave from normal group to abnormal group. If a valid response (SM-POLL or SM-DATA) enters the IOB, the error count is set to zero.

4 is a schematic polling flowchart of a method for processing an E bus protocol according to the present invention. Referring to FIG.

ANNi is the case where MS-POLL is sent to abnormal node i, NMNi is the case where MS-POLL is sent to healthy node i, and '_' is SM-POLL or SM-DATA (F = 0) is entered, " "→" indicates when timeout occurs. "→" indicates when SM-POLL, SM-DATA (F = 0 or 1) is input or a timeout occurs.

If the registration request flag REG-F coming from the back panel is checked and set, polling for an abnormal node is performed (300), and if it is reset, polling for a normal node is performed (310). Polling of the abnormal node is performed until three polls have been made for all abnormal nodes or until the registration request flag (REG-F) is reset (320). Polling for the normal node is performed up to three times when a timeout occurs (330).

5 is an exemplary embodiment according to the E bus protocol processing method according to the present invention. "_ _" Indicates timeout, "-→" indicates SM-POLL, "_" indicates SM-DATA (F = 0), and "→" indicates SM-DATA (F = 1).

Assume that the initial normal node is 3, 9 and the abnormal node is 1, 4, 6, 8. A registration request flag (REG-F) is set to poll for an abnormal node. Node 1 switches to the normal node by responding to the polling request with the SM-POLL (500). Nodes 4, 6, and 8 failed to register due to timeout (510). In polling for node 2, node 4 responds with SM-DATA (F = 0) and returns to the normal node (520). Node 6 receives SM-DATA (F = 1) for the third poll. It is cleared and polling for node 8 is not performed. Polling of the unhealthy node is complete and now it goes to polling the healthy node. Polling for node 1 responds with SM-DATA (F = 0) and reports it to CPB (540), and node 3 receives SM-POLL in response to polling and goes to the next normal node (550). . Node 4 has three timeouts, incrementing the error count by one, and moving on to the next node. If this timeout occurs three times in succession, the node is changed to an abnormal node (560). Node 6 has a first timeout but responds to the second poll with SM-POLL and proceeds to the next node (570). Node 9 can respond to the polling request with four consecutive SM-DATA (F = 1) and then reply with SM-DATA (F = 0) to send five consecutive data to the host (600).

The present invention constructed and operated as described above provides a subscriber system access technology for accessing ISDN functions to a domestic electronic switchboard TDX-1 (registered trademark) series, thereby providing a multi-drop that requires real-time processing of the proposed E bus protocol. It can be applied to a similar communication structure using a polling scheme as a master and a slave in a message transfer unit requiring inter-processor communication through a communication medium of a bus structure having a structure of.

Claims (8)

A clock line (a) for supplying control clocks to the master and slave stages, a first data line (b) consisting of an RS-485 loop for supplying data from the master stage to the slave stage, and the slave stage The second data line (c) consisting of an RS-485 loop for supplying data to the master terminal is provided with a protocol processing method of an E bus for exchanging D channel signals at one master terminal and a plurality of slave terminals. In; A first step of retrieving the status of the registration request signal REG-F of the slave end; a second step of performing and processing an abnormal group polling when the registration request signal REG-F is set; and the registration request signal And (REG-F) is reset by the third step of performing and processing normal group polling. The method of claim 1, wherein the second step includes: a first process of polling one abnormal node once, and searching for the presence or absence of a response to the polling, and transitioning to a normal group if a valid response within a predetermined time is reached; If it is not a valid response, a second process proceeding to the next step, and a third process of polling the next abnormal node and repeating the second process at most twice. 2. The method of claim 1, wherein the second step repeats the abnormal group polling until the registration request signal (REG-F) is cleared. The method of claim 1, wherein the second step repeats the abnormal group polling until there is no abnormal node. The method of claim 1, wherein the third step includes: a first process of polling one normal node, a second process of searching for the presence or absence of a response to the polling; A third process of repeating the first and second processes until the timeout count is 3, a fourth process of increasing the error count by 1 when the timeout count is 3, and an abnormality when the error count is 3 E bus protocol processing characterized in that performed by a fifth process of transition to a group. The method of claim 1, wherein the third step is performed until the slave station responds with a control signal SM-POLL indicating that there is no packet to transmit. The method of claim 1, wherein the third step is repeated until the slave station responds with a control signal (SM-DATA, F = 0) indicating that there are no more packets to transmit. . The E bus of claim 2, wherein the legitimate response is a control signal SM-POLL indicating that the slave terminal has no packet to transmit and a control signal SM-DATA transmitting the packet to the slave terminal. Protocol handling method.