KR100913799B1 - CPU Degenerate System and Degenerate Method Using Service Processor - Google Patents

Abstract

The present invention provides an error CPU information (not directly recognized by a micro program) obtained by a service processor immediately before the end of a micro program in order to enable the degeneracy of a CPU with a preliminary example of unstable operation before starting the OS. The microprocessor acquires from the processor and requests the degeneracy of the CPU which has a foreseeable operation of the service processor based on the acquired error CPU information, and the service processor executes the degeneracy of the CPU having a foreseeable operation before OS startup. For the purpose of

Description

CPU SUPPRESSION SYSTEM AND CPU SUPPRESSION METHOD USING SERVICE PROCESSOR}

The present invention relates to a CPU degenerate system and a degenerate method using a service processor that detects a CPU having a preliminary unstable operation before operation and does not incorporate it into the system.

1 is a schematic diagram showing the configuration of a CPU degenerate system using a service processor for realizing a conventional CPU degenerate. The micro program 11 of Fig. 1 functions as a module for diagnosing the units constituting the system, in response to a system reset such as when the system is powered on or rebooted / reset. The micro program 11 shown in FIG. 1 is executed by the CPU on the operating system side, that is, the CPU (A) 10 of FIG. The CPU (A) 10 on the operating system side is depicted as one CPU in the drawing, but of course, the CPU on the operating system side is configured by a plurality of CPUs.

When the CPU control on the operating system side is transmitted to the Micro Program 11 shown in Fig. 1, since the operation of the CPU is checked by the service processor side shown by the arrow 1, the operating system side The CPU is in a state where it can at least perform basic operations using its own hardware resources. The service processor CPU (B) 20 performs CPU stop processing for the CPU on the operating system side determined to be NG (No Good) by the operation check of the CPU on the operating system side by the service processor CPU (B) 20. In order to execute the control, control is not transferred to the microprogram 11.

The micro program 11 shown in Fig. 1 is represented by arrows (2), (3), and (4) in Fig. 1, and includes a CPU (A) 10 and a memory (12) of the operating system. And initial setting and diagnosis of a system configuration unit called an I / O unit 13 to exchange information with the service processor side.

Fig. 2 is a schematic diagram showing a processing sequence for explaining the operation of a CPU degenerate system using a conventional service processor between a micro program, a system CPU, and a service processor. Since the numbers in the sequence chart shown in FIG. 2 correspond to the numbers in the configuration diagram of the CPU degenerate system using the conventional service processor shown in FIG. 1, the conventional service processor is used using FIG. 1 and FIG. The operation of the CPU degenerate system will be described. As shown in FIG. 2, first, system power-on is performed on the service processor side (A1). Next, the operation check (Fig. 1 (1)) of the system CPU 10 is performed (A2). Here, when the CPU 10 on the operating system side is capable of performing at least basic operations using its own hardware resources, the CPU control is transmitted to the microprogram 11 (A4), but the operation of the CPU on the operating system side is checked. If it is determined as NG (No Good) by the service processor, CPU stop processing is performed by the service processor, and control is not transmitted to the microprogram 11 (A3).

The micro program 11 to which control is transmitted from the service processor 20 starts initial setting and diagnosis of a system configuration unit, and performs CPU diagnostic processing (FIG. 1 (2)) first (A5). During the initial setting and diagnosis of the system configuration unit, the service processor 20 always monitors the occurrence of the error of the system CPU 10 and at the same time immediately before the micro program transitions to the next control, the error occurrence situation is transmitted to the system side CPU. Notify (A6). When the Micro Program 11 is executing the initial setting / diagnosis (arrows (2), (3), (4) in FIG. 1) of the units constituting the system, and is delivered after OS startup. Until the occurrence of the CPU hardware error is recognized by the service processor CPU (B) 20 (arrow 5 in FIG. 1), and the number of occurrences of the error exceeds a preset threshold value, "a serious error will occur later." It was a standby state by the OS, considering a possible CPU with a predictable and unstable behavior. That is, conventionally, the CPU is not detached physically but the CPU is softly separated without the process being allocated to the target CPU. Then, the micro program 11 causes the micro-program 11 to be reset. Was running degenerate. As a result of the diagnosis by the microprogram 11, if it is OK, the program processing is continued in the normal CPU and the diagnosis processing is terminated (A8). On the other hand, when the system-side CPU itself indicates the NG situation as a result of the CPU diagnostic processing by the microprogram, the microprogram 11 itself degenerates the CPU (A7), and even if the CPU degenerates, the normal CPU remains. If the program processing can be continued, the processing continues and the diagnostic processing by the microprogram 11 is terminated (A8). That is, the process is not assigned to a CPU that has an example of unstable operation from the system, but it is already after the OS (system) is started. This was the timing of rebooting and the like. (See Patent Document 1)

[Patent Document 1] Japanese Unexamined Patent Application Publication No. 08-087341

As described above, since the CPU degenerate system using the conventional service processor recognizes a CPU having an unstable operation after OS startup, and executes CPU degeneracy at the next system reset, the CPU degenerate system has an unstable operation before OS startup. There was a problem that the degeneracy of the existing CPU could not be performed.

However, in the computer field, the quality fluctuation caused by the production process of the CPU is inevitable. Due to this change in quality, stable operation of the system is important by identifying a CPU with a preliminary example of unstable operation, recognizing what level of unstable operation is, and separating them from the constituent unit before operating the system.

Recognizing and degenerating CPUs with unpredictable CPUs that perform unstable behavior before operation, improves the robustness of the operating system, and trouble-shooting time when a failure occurs after operation It is important to reduce the cost of maintenance. However, when these are realized by a hardware function, the increase in the cost of the system and the increase in the size of the system due to the increase of the hardware to be mounted have been an obstacle to overcoming the problem.

In order to solve the above problems, the present invention provides a CPU degenerate system using a service processor that recognizes a CPU with a preliminary example of unstable operation by a micro program, determines a CPU to be degenerate, and is not assembled to an operating system before OS startup. It is an object to provide a degeneracy method.

The present invention obtains the error CPU information (not directly recognized by the micro program) acquired by the service processor from the service processor immediately before the micro program ends, and operates unstable to the service processor based on the acquired error CPU information. Request degeneracy of the CPU with a foreword, and the service processor is characterized in that degeneracy of the CPU with a foreplay of unstable operation before starting the OS.

According to the present invention, by degenerating a CPU with a preliminary example of unstable operation before OS startup, it is possible to improve the fastness in the operating system and to provide the cost of troubleshooting and maintenance work time when a failure occurs after the operation. We can save.

1 is a schematic diagram showing the configuration of a CPU degenerate system using a service processor for realizing a conventional CPU degenerate.

Fig. 2 is a schematic diagram showing a processing sequence for explaining the operation of a CPU degenerate system using a conventional service processor between a micro program, a system CPU, and a service processor.

Fig. 3 is a schematic diagram showing the configuration of a CPU degenerate system using a service processor for realizing CPU degeneracy of the present invention.

Fig. 4 is a schematic diagram showing a processing sequence for explaining the operation of the CPU degenerate system using the service processor of the present invention between a Micro program, a system CPU, and a service processor.

Fig. 3 is a schematic diagram showing the configuration of a CPU degenerate system using a service processor for realizing CPU degeneracy of the present invention. The micro program 11 of FIG. 3 functions as a module for diagnosing the units constituting the system, in response to a system reset such as when the system is powered on or rebooted or reset. The micro program 11 shown in FIG. 3 is executed by the CPU on the operating system side, that is, the CPU (A) 10 of FIG. The CPU (A) 10 on the operating system side is depicted as one CPU in the drawing, but of course, the CPU on the operating system side is configured by a plurality of CPUs.

When CPU control on the operating system side is transmitted to the Micro Program 11 shown in Fig. 3, since the operation of the CPU is checked by the service processor side shown by the arrow 1, the operating system side The CPU is in a state where it can at least perform basic operations using its own hardware resources. The service processor CPU (B) 20 performs CPU stop processing for the CPU on the operating system side determined to be NG (No Good) by the operation check of the CPU on the operating system side by the service processor CPU (B) 20. In order to execute the control, control is not transferred to the microprogram 11.

The micro program 11 shown in FIG. 3 is represented by arrows 2, 3, and 4 of FIG. 1, and includes a CPU (A) 10 and a memory (12) of the operating system. And initial setting and diagnosis of a system configuration unit called an I / O unit 13 to exchange information with the service processor side.

Fig. 4 is a schematic diagram showing a processing sequence for explaining the operation of the CPU degenerate system using the service processor of the present invention between a micro program, a system CPU, and a service processor. Since the numbers in the sequence chart shown in FIG. 4 correspond to the numbers in the configuration diagram of the CPU degenerate system using the service processor of the present invention shown in FIG. 3, the service processor of the present invention will be described with reference to FIGS. The operation of the used CPU degenerate system will be described. As shown in FIG. 4, first, system power-on is performed on the service processor side (B1). Subsequently, an operation check (Fig. 3 (1)) of the system CPU 10 is performed (B2). Here, when the CPU 10 on the operating system side is in a state capable of at least basic operation using its own hardware resources, CPU control is transferred to the microprogram 11 (B4), but the operation of the CPU on the operating system side. When it is judged as NG (No Good) by the check, CPU stop processing is performed, and control is not transmitted to the microprogram 11 (B3).

The microprogram 11 to which control is transferred from the service processor 20 starts initial setting and diagnosis of the system configuration unit, and first performs CPU diagnostic processing (Fig. 3 (2)) (B5). During the CPU diagnostic process, the service processor 20 always monitors the occurrence of the error of the system CPU 10 and notifies the system side CPU of the error occurrence only when the microprogram moves to the next control (B6). That is, the occurrence of CPU hardware error during the initial setting / diagnosis (arrows 2, 3, and 4 of FIG. 3) of the unit constituting the system is executed by the micro program 11. If the service processor CPU (B) 20 recognizes (the arrow 5 in Fig. 3), and the number of occurrences of the error exceeds a preset threshold value, "preparation that may cause a serious error later and perform unstable operation. CPU ”is requested to the service processor 20 to degenerate the CPU (B13). If the result of the diagnosis by the microprogram 11 is OK, the program processing is continued in the normal CPU (B9). On the other hand, when the system-side CPU itself indicates the situation of NG (NoGood) as a result of the CPU diagnostic processing by the microprogram, the service processor 20 is requested to degenerate the CPU (B7). The service processor 20 continues the process if it is possible to continue the program processing in the normal CPU left by executing the CPU degeneration requested by the microprogram 11 (B8) (B9).

By the way, the Micro Program detects the occurrence of the CPU hardware error during the initial setting / diagnosis of the units constituting the system ((2), (3), (4) of FIG. 3). 20 always monitors and recognizes the occurrence (Fig. 3 (5)) (B6), and accumulates the error CPU information (described later) on the system side until immediately before the micro program ends. Let Info "21 be created ([Error Info] of FIG. 3) (FIG. 3 (6)) (B10). Therefore, "Error Info" 21 constitutes an error CPU information accumulating unit.

The microprogram 11 requests notification of the CPU in which an error has occurred to the service processor 20 immediately before the microprogram ends (Fig. 3 (7)) (B11), and the service processor 20 receives an error. The generated CPU is notified to the microprogram 11 based on "Error Info" 21 (Fig. 3 (8)) (B12).

The microprogram 11 judges the contents of the CPU in which the error has occurred, determines the CPU to make a degeneracy request to the service processor 20, and requests the CPU to degenerate the service processor 20 (Fig. 3 ( 9)] (B13). In the present invention, CPU resources are required as a resource for initial setting / diagnostic processing of the microprogram. Therefore, degeneracy of the CPU having a preliminary operation of unstable operation is performed immediately before the end of the microprogram.

In response to the degeneracy request from the microprogram 11, the service processor 20 executes degeneracy of the CPU (B14).

If no CPU hardware error occurs during the operation of the microprogram (during initial setting / diagnosis of the system configuration unit), the initial setting / diagnosis processing by the microprogram is terminated as usual.

As described above, according to the present invention, by acquiring the error CPU information that the microprogram cannot directly recognize from the service processor side, the degeneracy of the CPU having a preliminary operation that makes the operation unstable before the OS startup becomes feasible, and thus the solidity in the operating system. It is possible to improve the cost and reduce the time for maintenance work and maintenance when a failure occurs after operation.

When the error CPU information on the system side is described, the error CPU information belongs to a Synchronous Error, and an Instruction Access Error (IAE), cache, which is shown in a memory UE (Uncorrectable Error) or the like that is being fetched from an instruction. DAE (Data Access Error, etc.) that is shown to the UE when Load / Store is being accessed to the UE. A high error level, such as Instruction Urgent Error (I_UGE) shown to a UE in a CCR (Condition Codes Register, etc.), and is also harmful to a RE (Restrainable Error) belonging to an Asynchronous Error, that is, a program currently running. Errors that do not affect, for example, low error levels such as CE (correctable error) corrected by hard, are included in the present invention. Regardless of the height of the bell, if it is determined that the CPU has a foreseeable behavior, a request for degeneracy of the CPU is sent to the service processor to improve the fastness of the operating system and to troubleshoot when a failure occurs after operation. The cost of working time and maintenance can be reduced.

In a computer system in which a hardware function that detects a CPU that has an unstable operation and recognizes or degenerates its level cannot be completely implemented for reasons such as cost, these functions are performed by a micro program. By implementing the above, an inexpensive and high quality computer system having a function of detecting a CPU having a preliminary example of unstable operation and recognizing the level thereof can be constructed.

Claims (6)

A service processor that performs CPU check of an operating system and transmits CPU control to a microprogram when it is determined that the CPU has a basic operating capability by its own hardware resource; and a CPU of the operating system upon system reset. A microprogram for diagnosing a system configuration unit, wherein the microprogram obtains error CPU information from the service processor until immediately before the end of diagnosis by the microprogram, and is unstable to the service processor based on the error CPU information. Requesting degeneracy of a CPU with a working example CPU degeneration system using a service processor, characterized in that. 2. The service processor of claim 1, wherein the service processor includes an error CPU information storage unit that monitors an error occurrence state of a CPU of the operating system and stores the error occurrence state until immediately before the end of the diagnostic processing by the microprogram. CPU degenerate system using a service processor. The service processor of claim 2, wherein the service processor notifies the microprogram of the error CPU information stored in the error CPU information storage unit when receiving the request for the transmission of the error CPU information from the microprogram. CPU degenerate system. The microprocessor of claim 3, wherein the microprogram determines a CPU to degenerate based on the error CPU information notified from the service processor, and requests the service processor to degenerate the CPU. A CPU degenerate system using a service processor, characterized in that the CPU degenerate. The service processor recognizing the occurrence of a CPU hardware error while the microprogram is performing initial setup or diagnostics of the units making up the system; Creating error CPU information on the system side, Requesting notification of the CPU information in which the error occurred to the service processor immediately before the end of the micro program, wherein the service processor notifies the microprogram of the CPU information in which the error is received by the request; Determining, by the micro-program, the content of the notified error CPU information, and requesting degeneracy to the service processor by deciding a CPU to degenerate the service processor; Receiving a degeneracy request from the microprogram, and the service processor executing degeneracy of the CPU CPU degeneracy method using a service processor comprising a. delete

Images