JPH069033B2 - Abnormality diagnosis method for multi-CPU system - Google Patents

Description

Detailed Description of the Invention A. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a multi-CPU system in which a plurality of CPU nodes are connected by a bus, and more particularly to a method for diagnosing a data transmission / reception function abnormality by a bus between nodes.

B. SUMMARY OF THE INVENTION According to the present invention, the transmitting node monitors the data on the bus at the same time as transmitting and compares the transmitting data and the receiving data. When it is detected, the node is disconnected from the bus, self-diagnosis processing is executed, and if the self-diagnosis result is not defective, the node is reconnected to the bus. It is now possible to take appropriate measures depending on the cause of failure such as noise, bursty noise on the bus, or node or bus failure.

C. 2. Description of the Related Art In a multi-CPU system in which a plurality of CPU nodes are connected to, for example, an asynchronous serial bus, a receiving node performs a parity check or a framing check during data transmission via the bus, and an error is detected in this check. When this happens, a bus failure is dealt with, and the receiving node stops operating the bus and the system goes down.

D. Problems to be Solved by the Invention In the conventional system described above, even if a transient noise occurs on the bus, if the noise coincides with the transmission timing, the receiving node recognizes it as a bus failure and stops the operation of the bus. Becomes If the above state occurs while a large number of nodes are simultaneously receiving bus signals, all receiving nodes stop operating the bus, and transient noise causes the entire system to go down.

The present invention has been made in view of the above-mentioned conventional problems, and an object of the present invention is to provide a multi-CPU system capable of appropriately dealing with each cause such as transient noise, burst noise, and node or bus failure. It is to provide a failure diagnosis method.

E. In the present invention, therefore, in the present invention, the transmitting node monitors the bus signal at the same time as the transmission and compares the transmitted data with the received data. When the two data do not match, the data transmission is performed again and the data is retransmitted. When the data mismatch is detected, the node is disconnected from the bus and the self-diagnosis process is executed. If the result of the self-diagnosis is not bad, the node is reconnected to the bus.

F. Effect According to the abnormality diagnosing method of the present invention, when transient noise occurs on the bus and the transmission data and the reception data become inconsistent with each other, there is no influence of noise at the time of data retransmission, and normal transmission is possible. When a certain amount of bursty noise occurs on the bus, a mismatch between the transmitted and received data is detected even when the data is retransmitted. In this case, the self-diagnosis process is executed, and since there is no abnormality in the transmission node itself, the diagnosis result becomes normal and the bus is reconnected to the bus. The node stops the bus operation only when the self-diagnosis result is bad.

G. Embodiment FIG. 1 is a functional block diagram of one CPU node for implementing the method of the present invention.

Reference numeral 1 is a serial bus for interconnecting a plurality of CPU nodes having similar configurations, 2 is a transmission control CPU, 3 is a host CP
U and _RY are relays that connect the bus 1 to the reception monitoring circuit 4 and the transmission control circuit 5, and 6 is a priority determination circuit that determines whether or not the right to use the bus 1 is given to the CPU node. Is a relay circuit connecting the transmission control CPU 2 and the host CPU 3, and 8 is the received data R from the CPU 2 to the CPU 3.
_{The X} transfer register 9 is a register for transferring the transmission data T _X from the CPU 3 to the CPU 2.

The transmission control CPU 2 uses a UART (universal asynchronization).
ous receiver and transmitter) 21 and the receiving function unit 2
2, a mismatch detection function unit 23, and a transmission function unit 24 are included.

The host CPU 3 includes an application function unit 31, a reception function unit 32, a transmission function unit 33, an interrupt control unit 34,
It includes the self-diagnosis function unit 35.

When receiving data, the signal on bus 1 is relay R _Y
Is input to the reception monitoring circuit 4 via the serial output SD
I is processed by the UART 21 and becomes the reception data R _X.
The reception function unit 22 checks the parity of the reception data R _X , and transfers the data R _X to the reception function unit 32 of the host CPU 3. If a parity error is detected, a signal R _X ERR to that effect is transmitted to the interrupt control unit 34.

At the time of data transmission, the transmission data T _X is first generated from the transmission function unit 33 of the host CPU 3, becomes the serial signal SDO via the transmission function unit 24 of the CPU 2, and the UART 21, and is transmitted from the transmission control circuit 5 via the relay R _X to the bus 1. Is output to. At this time, at the same time, the signal on the bus 1 (the signal sent by itself) is monitored by the reception monitoring circuit 4, and the UART2
1, Notify the reception function unit 22. The mismatch detection circuit 23
Compare with the transmission data T _X. If normal, the data T _X and R _X match.

As shown in the timing chart of FIG. 2 (A), when transient noise occurs on the bus 1 during transmission of the data T _X , the mismatch detection circuit 23 detects the mismatch between the data T _X and R _X. To be done. In this case, the mismatch detection circuit 23 issues a retry command to the transmission function unit 24 and retransmits the same data T _X. If the cause is a transient noise as shown in FIG. 2 (A), the data T _X and R _X match at the time of retransmission.

FIG. 2 (B) shows the operation when bursty noise lasting for a relatively long time occurs in burst 1. In this case, the data T _X and R _X do not match even at the time of the above-mentioned retransmission. At this time, the mismatch detection function unit 23 gives the retry mismatch signal T _X ERR to the interrupt control unit 34.

In response to this, the self-diagnosis function unit 35 operates in the host CPU 3. First, the relay R _Y is turned off to disconnect the node from burst 1. In this state, the self-diagnosis function unit 35 generates transmission data T _X for self-diagnosis, which is transmitted to the transmission function unit 33. The transmission function unit 33 and the CPU 2 perform the same operation as above to transmit the transmission data T _X to the transmission control circuit 5
Output in series. This output does not appear in burst 1,
It is directly input to the reception monitoring circuit 4 in the same node.
Similar to the transmission described above, the mismatch detection function unit 23 of the CPU 2 causes the transmission data T _X and the reception data R _X at this time to be transmitted.
Are compared, and the comparison result and the reception data R _X are transmitted to the host CPU 3.

Even if the re-data match by the comparison of the mismatch detection function unit 23, the self-diagnosis function unit 35 checks the received data R _X received by the CPU 2 from the side and checks whether it is the same as the data T _X generated by itself. .

If the self-diagnosis described above determines that the node is normal, the internal function of the node is normal, and the cause of the abnormality is treated as noise on the burst side, and the relay R _Y is turned on to reconnect to the bus 1. . This operation is shown in FIG. 2 (B).

FIG. 2 (C) shows the operation when it is determined to be abnormal by self-diagnosis. In this case, it is recognized that there is an abnormality in its own node, and the relay R _Y remains off.

H. Effect of the Invention As described in detail above, the multi-CP according to the present invention
According to the above diagnostic method of the U system, when data transmission is not performed correctly and whether it is not performed correctly, it is appropriate depending on whether the cause is transient noise or burst noise, or abnormal in the transmitting node. Can be dealt with and the system will not be down unnecessarily, improving the operating rate and reliability of the system.

[Brief description of drawings]

FIG. 1 is a functional block diagram of one CPU node for implementing the method of the present invention, and FIG. 2 is a timing chart showing the operation of the present invention. 1 ... Serial bus, 2 ... Transmission control CPU, 3 ...
Host CPU.

Claims (1)

[Claims] 1. In a multi-CPU system in which a plurality of CPU nodes are connected by a bus, a transmitting node that has transmitted data to the bus simultaneously monitors the data on the bus at the same time as transmitting and compares the transmitted data with the received data. If the data does not match, the data is retransmitted.If data mismatch is detected during retransmission, the node is disconnected from the bus and self-diagnosis processing is executed. A method for diagnosing an abnormality in a multi-CPU system, which comprises reconnecting to