JP2000148709A - Multi-cpu constitution system and system reconstituting method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the system reliability by allowing a CPU having detected a fault to obtain and refer to system data regarding the CPU which has got out of order, newly load the functional program that the faulty CPU is taking partial charge of, and take partial charge of the faulty CPU function in addition to the original functions. SOLUTION: The CPU having detected one CPU having got out of order obtains the program head address where a common program is stored and the program transfer size according to an obtained CP#. Transfer from a system data memory 4 to local memories 8-1 to 8-5 of respective CPUs is performed from the obtained common program head address by the program transfer size. The function set in the CP# of the CPU having got out of order and the head address and program transfer size of the function program are obtained. Transfer from the system data memory 4 to the local memories 8-1 to 8-5 of the CPUs is performed from the obtained head address of the function program by the program transfer size.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a function-distributed multi-CPU configuration system and a system reconfiguration method capable of maintaining the operation of the system even when a CPU failure occurs.

[0002]

2. Description of the Related Art In a multi-CPU configuration system having a mechanism for arbitrating a conflict that occurs when a plurality of CPUs randomly issue a request for a right to use one bus in parallel, a function distribution / load distribution system and a redundant configuration are applied. In general, a method of configuring a system having high fault-tolerant performance is performed.

FIG. 6 is a diagram showing an example of a load distribution method in a conventional multi-CPU system disclosed in Japanese Patent Application Laid-Open No. Hei 10-11412.

[0004]

As described above, in a multi-CPU configuration system, a method of configuring a system with high fault tolerance by applying a function distribution / load distribution method and a redundant configuration is generally used. In some cases, a redundant configuration and load distribution cannot be adopted due to the limitation of the amount of applied hardware. Multi-CP without such redundant configuration and load sharing
The U configuration system has a problem in that if a certain CPU fails during operation, the entire system becomes inoperable.

[0005] An object of the present invention is to solve the problem in the case where a redundant configuration and load distribution are not adopted due to the limitation of the amount of applied hardware.
An object of the present invention is to provide a multi-CPU configuration system and a system reconfiguration method that can maintain the operation of the system even when a U failure occurs.

[0006]

SUMMARY OF THE INVENTION A multi-CPU according to the present invention
When a failure occurs in one CPU and the system becomes inoperable, one of the Cs in the other normal CPUs is configured.
A means for the PU to detect the failure, and the CPU that has detected the failure acquires and refers to system data relating to the CPU in which the failure has occurred, and newly loads a function program that is originally assigned to the failed CPU, and In addition to the above, the system can be reconfigured by having means for sharing the function of the failed CPU and operating the function merge.

Further, the present invention provides a function-distributed multi-CP
In a system reconfiguration method in a U configuration system,
If one CPU fails and becomes inoperable,
Any one of the other normal CPUs detects the failure, acquires and refers to system data relating to the failed CPU, newly loads a function program originally assigned to the failed CPU, It is characterized in that the function merge operation is performed by sharing the faulty CPU function in addition to the original function.

[0008]

Next, an embodiment of the present invention will be described with reference to the drawings.

FIG. 1 is a system configuration diagram showing an embodiment of a function-distributed multi-CPU configuration system without a redundant configuration and load distribution according to the present invention.

FIG. 2 is a table for transmitting a function division to each CPU in the multi-CPU configuration system of the present invention, and stores a function to be executed and a corresponding program for each number of the mounted CPU. Stores the address and program size. FIG. 3 is an example of storing a program for each function. FIG. 4 is a diagram showing an example of a CP # management table, and FIG. 5 is a flowchart showing an example of a program load process according to the present invention.

A system reconfiguration method according to the present invention can
In the example in which five are mounted, first, when a system reset is executed in the multi-CPU configuration system shown in FIG. 1, each CPU is reset. Thereafter, the program loading process shown in FIG. 5 is executed using the CP # management table shown in FIG.

The program load processing will be described with reference to FIG. 5. First, the LO # is stored in the CP # management table shown in FIG.
CK is acquired (601). If the content of LOCK is "OF
If "F", the next process is performed, and if "ON", LOCK is acquired again (602). Next, "ON" is set to LOCK (603). Next, the CP shown in FIG.
# Get the CP # from the management table (604), and
It is determined whether or not is less than 5 (605). If it is less than 5, the acquired CP # is incremented by 1, and the CP # management table is reset (606). If it is 5 or more, the following processing is performed. Finally, LOCK is set to “OFF” (60
7).

Next, it is determined whether or not a CP # of less than 5 has been acquired (608). If the CP # cannot be acquired, it is regarded as abnormal, an abnormal process is performed (609), and an infinite loop is performed (610). It will be in a disconnected state. Here, the CPU 2-1 sets CP # 0, and the CPU 2-2 sets C
The following processing will be described on the assumption that the CPU 2-3 is set to CP # 2, the CPU 2-4 is set to CP # 3, and the CPU 2-5 is set to CP # 4.

Next, based on the acquired CP #, each CPU refers to the common program data in the CPU function table shown in FIG. 2 and acquires the program start address and the program transfer size where the common program is stored. (611). From the acquired common program start address, the program transfer size is transferred from the system data memory 4 to the local memories 8-1 to 8-5 of each CPU (612).

Next, referring to the program data for each CP # in the CPU function table shown in FIG. 2, the function set for each CP #, the start address of the set function program and the program transfer size are set. Is acquired (613). From the start address of each acquired function program to the size of the program transfer size, the system data memory 4
Then, the data is transferred to the local memories 8-1 to 8-5 of each CPU (614).

A case where a failure occurs in a certain CPU will be described. If a certain CPU fails, the other CP
Any one of U detects a failure. The CPU that detects the failure performs the following processing based on the CP # of the failed CPU.

The CPU that detects the failure of a certain CPU refers to the common program data in the CPU function table shown in FIG. 2 based on the acquired CP #, and determines the program start address where the common program is stored and the program transfer address. The size is obtained (611). A local memory 8-1 of each CPU from the system data memory 4 by the size of the program transfer size from the acquired common program start address.
To 8-5 (612).

Next, referring to the program data for each CP # in the CPU function table shown in FIG. 2, the function set in the CP # of the failed CPU and the head address of the set function program And the program transfer size is acquired (613). From the start address of the acquired functional program by the program transfer size, the local memory 8-1 of the own CPU is stored in the system data memory 4.
To 8-5 (614). After that, the function assigned to the own CPU and the function assigned to the failed CPU function together.

According to the present invention, as described in the above embodiment, an active CPU or a spare CPU in alternative operation
However, the present invention relates to a system reconfiguration technique for maintaining a system operation in response to an emergency trouble by loading a program corresponding to a failed CPU sharing function into its own CPU and performing a combined operation of the functions.

At the time of the above-mentioned CPU failure, by applying the program load processing of the present invention in the system reconfiguration processing, the active CPU can combine the functions of the failed CPU and continue the system operation.

In this system reconfiguration method, in principle, even if a maximum of half of the system-mounted CPUs fail, the remaining half of the CPUs operate as function-merged CPUs.
System reconfiguration comparable to N-redundant configuration is possible.

In the above-described embodiment, a system reconfiguration method in a function-distributed multi-CPU configuration system without a redundant configuration and load distribution has been described. However, the present invention provides an N + 1 spare CPU configuration without load distribution. This reconfiguration method can be applied to the system. 1
Even in the case where another CPU failure occurs while the spare CPU is being replaced by another CPU due to a CPU failure, a system reconfiguration comparable to the N + 2 redundant configuration can be performed by combining the functions of the spare CPU. In this case, the same applies to the above-described program transfer except that the CPU that detects the failure of the other CPU is the spare CPU.

When this method is applied to an N + 1 spare CPU configuration system, the processing capacity of the spare CPU is
If the processing capacity of U is 3 times, it can withstand 3 CPU failures, and if it is 4 times, 4 CPUs due to 4 CPU failures
The system can be reconfigured as a function merge alternative. Spare CPU
And a highly reliable redundant configuration that evolves into a system reconfiguration that depends on the processing capability.

Further, the present invention provides an N +
This reconfiguration method can also be applied to one spare CPU configuration system. In this case, the processing capacity of the spare CPU is
If the processing capacity of the PU is 3 times, it can withstand 3 CPU failures, and if it is 4 times, 4 CP due to 4 CPU failures
It is possible to reconfigure the system as an alternative to merging U functions.

[0025]

As described above, according to the present invention, when a CPU failure occurs in a redundant configuration and a function-distributed multi-CPU configuration without load distribution, the active CPU is replaced with the failed CPU in the system reconfiguration processing. Since the functions can be combined and the system operation can be continued, the system reliability can be dramatically improved. in this case,
Of the CPUs mounted and operated, up to half of the CPU failures can be tolerated.

Also, N + 1 spare CPUs without load distribution
In the function-distributed multi-CPU configuration, 1 CP
While the spare CPU is being replaced by U failure, another C
Even in the case of occurrence of a PU failure, the function combination can be expanded depending on the processing capacity of the spare CPU, so that it is possible to withstand failures of a plurality of CPUs.

When this method is applied to an N + 1 spare CPU configuration for load distribution, the processing capacity of the spare CPU is
If the processing capacity of the PU is 3 times, it can withstand 3 CPU failures, and if it is 4 times, 4 CP due to 4 CPU failures
It is possible to reconfigure the system as an alternative to merging U functions. Reserve CP
A highly reliable redundant configuration that evolves into a system reconfiguration depending on the processing capacity of U can be obtained.

Further, upgrading of the program version,
Even when the partial programs are exchanged, quick operation can be resumed very easily.

[Brief description of the drawings]

FIG. 1 is a system configuration diagram showing an embodiment of a multi-CPU configuration system of the present invention.

FIG. 2 is a diagram illustrating an example of a CPU function table.

FIG. 3 is a diagram showing an example of storing a program for each function in a system data area.

FIG. 4 is a diagram illustrating an example of a CP # management table.

FIG. 5 is a flowchart illustrating an example of a program load process.

FIG. 6 is a diagram showing a state of a conventional initial process in a multi-CPU configuration system.

[Explanation of symbols]

 1 Multi-CPU configuration system 2-1 to 2-5 CPU 3-1 and 3-2 Common memory 4 System data memory 5 Common timer 6 Backup memory 7 Input / output device 8-1 to 8-5 Local memory

Claims (7)

[Claims] When one CPU fails and becomes inoperable, any one of the other normal CPUs is disabled.
Means for detecting the failure, and the CPU which has detected the failure acquires and refers to system data related to the CPU in which the failure has occurred, and
U is newly loaded with a function program to be shared by the U, and further includes means for sharing a faulty CPU function in addition to the original function to perform the function merge operation, so that the system can be reconfigured. CPU configuration system. 2. When a failure occurs in one CPU and the operation becomes inoperable, the spare CPU detects the failure and the spare CPU acquires and refers to system data relating to the failed CPU. Then, a function program that is originally assigned to the failed CPU is newly loaded, and in addition to the original function, the failed CPU function is further assigned to perform a function merging operation to reconfigure the system. Multi-CPU configuration system. 3. Even if another spare CPU failure occurs during the replacement operation of the spare CPU due to a CPU failure, a plurality of failed CPU functions are also shared depending on the processing capacity of the spare CPU. 3. The multi-CPU configuration system according to claim 2, wherein: 4. A system reconfiguration method in a function-distributed multi-CPU configuration system, wherein when a failure occurs in one CPU and operation becomes impossible,
Any one of the other normal CPUs detects the failure, acquires and refers to system data relating to the failed CPU, newly loads a function program originally assigned to the failed CPU, A system reconfiguration method characterized in that in addition to an original function, a function of a malfunctioning CPU is further shared and a function merge operation is performed. 5. A system reconfiguration method in a function-distributed multi-CPU configuration system, wherein when one CPU fails and becomes inoperable,
The spare CPU detects the failure, and the failed CP
The system data related to U is acquired and referred to, a function program originally assigned by the failed CPU is newly loaded, and in addition to the original function, the failed CPU function is also assigned to perform the function merge operation. System reconfiguration method. 6. In the case where the spare CPU is replaced by a CPU failure and another CPU failure occurs during the replacement operation, a plurality of failed CPU functions are also shared depending on the processing capacity of the spare CPU to operate the function combination. The system reconfiguration method according to claim 5, wherein 7. The system reconfiguration method according to claim 4, wherein said function-distributed multi-CPU configuration system is a function-distributed multi-CPU configuration system with load distribution.