JPH0588991A - Service processor - Google Patents

Abstract

PURPOSE:To prevent the generation of a memory fault (a 2-bit error) which can not be corrected due to the generation of an 1-bit fixed error in a memory device. CONSTITUTION:When an 1-bit error is generated in a main storage device, a fault information collecting means 102 in a service processor (SP) 1 collects fault information including an error address through a diagnostic path 100 and stores the collected information in a disk device 12. An 1-bit error fixed/ intermittent state judging means 103 judges whether the 1-bit error is a fixed error or an intermittent error. At the time of judging it as a fixed error, a page separation request sending means 101 requests the separation of a page including the error address to an operating system. An automatic informing means 104 reads out the fault information relating to the fixed error from the disk device 12 only at the time of judging the error as the fixed error and informs the fault information to a maintenance center through an automatic dialing means 105 and a line communication means 106.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a service processor, and more particularly, to a service capable of preventing an uncorrectable memory failure due to a fixed error when a 1-bit fixed error occurs in a memory device of a computer system. Regarding the processor.

[0002]

2. Description of the Related Art When a 1-bit error occurs in a memory device, a conventional service processor reports it to a maintenance center via a line regardless of whether it is a fixed error or an intermittent error.

The maintenance personnel of the maintenance center maintain the memory device only when a 1-bit fixed error is reported. That is, an intermittent error of 1 bit is caused by an ECC (Error Correcting Code).
Automatic correction is performed by the correction code such as e). On the other hand, a 1-bit fixed error due to a memory cell failure or the like becomes a 2-bit error due to collision with a 1-bit error due to an intermittent error in the same word. This is because a memory failure that cannot be corrected may occur.

[0004]

As described above, conventionally, when a 1-bit fixing error occurs, maintenance personnel maintain the memory device. Therefore, until the maintenance is completed, the 1-bit fixing error is fixed. The error will be neglected,
There is a problem that there is a high risk that an uncorrectable memory failure will occur.

Further, in the conventional service processor, when a 1-bit error occurs in a memory device, regardless of whether it is a fixed error or an intermittent error, all the 1-bit errors that occur are maintained via a line. Since it is reported to the center, the amount of communication with the maintenance center increases,
There are problems such as an increase in line usage charges, an increase in the amount of work for processing unnecessary reports (reports of intermittent error occurrences), and easy oversight of necessary reports (reports of fixed error occurrences). It was

An object of the present invention is to prevent an uncorrectable memory failure from occurring when a 1-bit fixed error occurs, and reduce the amount of communication with the maintenance center. The purpose is to provide a service processor that can be operated.

[0007]

SUMMARY OF THE INVENTION In order to prevent an uncorrectable memory failure from occurring due to a 1-bit fixed error that has occurred, the present invention provides: (A) When a 1-bit error occurs in a memory device, Failure information collecting means for collecting failure information including addresses, judging means for judging whether the 1-bit error is a fixed error or an intermittent error, and the judging means judging the 1-bit error as a fixed error. Obtaining a page address of a page including the error address based on the error address included in the failure information collected by the failure information collecting means,
A disconnection request sending means for requesting the operating system to disconnect the page corresponding to the page address is provided.

Further, according to the present invention, in order to reduce the amount of communication with the maintenance center, in addition to the configuration of (A), (B) the 1-bit error is determined to be a fixed error by the determination means, so that maintenance is performed. The center is connected to a line, and reporting means is provided for reporting failure information about the 1-bit error to the maintenance center.

[0009]

[Action]

In the configuration of (A), when the determination unit determines that the 1-bit error that has occurred in the memory device is a fixed error, the disconnection request transmission unit is based on the failure information collected by the failure information collection unit. The page address of the page including the fixed error is obtained, and the operating system is requested to disconnect the page.

In the configuration of (B), when the determining unit determines that the 1-bit error that has occurred in the memory device is a fixed error, the reporting unit connects the maintenance center and the line to detect the fixed error. Report the failure information to the maintenance center.

[0011]

Embodiments of the present invention will now be described in detail with reference to the drawings.

FIG. 1 is a block diagram showing an example of a computer system to which the present invention is applied. A service processor (SVP) 1, an arithmetic processing unit (EPU) 2, a main memory unit (MMU) 3, and an input unit. An output unit (IOP) 4, a system control unit (SCU) 5, a keyboard 11-1,
It is composed of a CRT 11-2 and a disk device 12.

Service processor 1 and arithmetic processing unit 2
The main storage device 3 and the input / output device 4 are connected via the system control device 5.

The service processor 1 is a processor controlled by a program, and is connected to the main storage device 3 via a diagnostic path 100. In addition to being able to access the main storage device 3 via this diagnostic path 100, the service processor 1 can also communicate with the arithmetic processing device 2. The service processor 1 also has a keyboard 11-1, a CRT 11-2 for conversation with the operator, and a disk device 12 for storing failure information.

The main memory 3 has a function of reporting an error to the service processor 1 via the diagnostic path 100 when a 1-bit error occurs. Further, as shown in FIG. 2, a communication area 22 for communication between the operating system and the service processor 1 is provided in the low address part of the OS (operating system) area 21 provided in the main storage device 3. There is.

FIG. 3 is a block diagram of an embodiment of the present invention, showing a configuration example of the service processor 1.

The service processor 1 collects failure information (including error content and error address) via the diagnostic path 100, and a main memory device 3.
1-bit error fixed intermittence judgment means 103 for judging whether the 1-bit error generated in 1 is a fixed error or an intermittent error
If the 1-bit error fixed intermittence determination means 103 determines that the error is a fixed error, the page address including the address where the fixed error occurs is stored in the communication area 22 shown in FIG. 2 and the SVP interrupt is sent to the arithmetic processing unit 2. It includes a page cut-off request sending means 101 for making a page. In FIG. 3, the same symbols as those in FIG. 1 represent the same parts.

Next, the operation of this embodiment will be described.

When a 1-bit error occurs, the main memory 3 reports an error to the service processor 1 via the diagnostic path 100.

Upon receiving the error report, the failure information collecting means 102 in the service processor 1 collects the failure information including the error content and the error address from the main storage device 3 via the diagnostic path 100 and stores it in the disk device 12. ..

When the failure information is stored in the disk device 12, the 1-bit error fixed intermittent judgment means 103 obtains an error address from the failure information collected this time by the failure information collection means 102, and the same address as the error address is stored. Whether or not the 1-bit error occurred in the main storage device 3 is a fixed error by determining whether or not the error address is stored in the disk device 12 other than the error address stored this time.
Determine if it is an intermittent error. That is, if it is stored, it is determined as a fixed error, and if it is not stored, it is determined as an intermittent error.

When the 1-bit error fixed intermittency judging means 103 judges that the error is a fixed error, the page cut-off request sending means 101 detects the error from the error address included in the failure information collected this time by the failure information collecting means 102. Generate a page address for the page containing the address. Then, it is determined whether the same page address as the generated page address is already stored in the communication area 22 of FIG. 2, and only when it is determined that the same page address is not stored, the generated page address is stored in the communication area 22. After that, an SVP interrupt is issued to the arithmetic processing unit 2.

When the SVP interrupt is issued, the operating system operating on the arithmetic processing unit 2 reads the page address stored in the communication area 22, and thereafter prohibits new allocation of the page. Therefore, when the job using the page ends, the page is no longer used, and the probability of an uncorrectable memory failure can be sufficiently reduced.

FIG. 4 is a block diagram of another embodiment of the present invention, showing a configuration example of the service processor 1.

In addition to the configuration shown in FIG. 3, the service processor 1 automatically notifies the failure information when the 1-bit error fixed intermittent judgment means 103 judges that a 1-bit fixed error has occurred. And an automatic dial means 1 for transmitting a dial number registered in advance
05 and line communication means 106 for performing line control.

When a 1-bit error occurs in the main memory device 3, the failure information collecting means 102 stores the failure information in the disk device 12, and the 1-bit error fixed intermittent judgment means 103 causes the 1-bit error to occur as described above. Is determined as a fixed error or an intermittent error. Further, when the 1-bit error fixed intermittency judging means 103 judges that the error is a fixed error, the page cut-off requesting means 101 stores the page address of the page including the error address in the disk device 12 and the arithmetic processing device 2 S
Make a VP interrupt.

The automatic reporting means 104 carries out the following processing only when the 1-bit error fixed intermittency judging means 103 judges that the 1-bit error is a fixed error.

First, the automatic dial means 105 is called. When the automatic dial means 105 is called, the dial number of the connection destination (maintenance center) registered in advance is sent to the line communication means 106, and the line communication means 106 sends the added dial number to the line. Connect the line to the maintenance center.

When the maintenance center and the service processor 1 are connected by a line, the automatic notification means 104 reads out the failure information stored by the failure information collection means 102 this time from the disk device 12, and the read failure information is sent to the line communication means 106. Add to.

The line communication means 106 is an automatic notification means 104.
When the failure information is added from the, it is transmitted to the maintenance center.

[0031]

As described above, according to the present invention, when the 1-bit error generated in the memory device is a fixed error, the page including the fixed error is requested to the operating system and the page is disconnected. Since the page in which the 1-bit fixed error has occurred continues to be used for a long period of time thereafter, an uncorrectable memory failure can be prevented from occurring.

Further, according to the present invention, the failure information is sent to the maintenance center only when a 1-bit fixed error occurs in the memory device. It is possible to reduce the amount of communication. Therefore, according to the present invention, it is possible to reduce the line usage fee, the number of lines in the maintenance center, the labor for processing unnecessary reports (reports of intermittent error occurrences), and overlook necessary reports (reports of fixed error occurrences). It has the effect of reducing the risk.

[Brief description of drawings]

FIG. 1 is a block diagram showing an example of a computer system to which the present invention is applied.

FIG. 2 is a diagram showing a configuration example of an operating system area.

FIG. 3 is a block diagram of an embodiment of the present invention.

FIG. 4 is a block diagram of another embodiment of the present invention.

[Explanation of symbols]

 1 ... Service processor (SVP) 2 ... Arithmetic processing unit (EPU) 3 ... Main memory unit (MMU) 4 ... Input / output unit (IOP) 5 ... System control unit (SCU) 11-1 ... Keyboard 11-2 ... CRT 12 ... Disk device 100 ... Diagnostic path 101 ... Page disconnect request sending means 102 ... Failure information collecting means 103 ... 1-bit error fixed intermittent judgment means 104 ... Automatic reporting means 105 ... Automatic dialing means 106 ... Line communication means

Claims (3)

[Claims] 1. A failure information collecting means for collecting failure information including an error address when a 1-bit error occurs in the memory device, and a judging means for judging whether the 1-bit error is a fixed error or an intermittent error. When the determining unit determines that the 1-bit error is a fixed error, the page address of the page including the error address is obtained based on the error address included in the failure information collected by the failure information collecting unit, A service processor, comprising: disconnection request sending means for requesting an operating system to disconnect a page corresponding to a page address. 2. A notification means for connecting the maintenance center to a line when the determination means determines that the 1-bit error is a fixed error and for reporting failure information about the 1-bit error to the maintenance center. The service processor according to claim 1, wherein: 3. The service processor according to claim 1, wherein the determination unit determines whether the error is a fixed error or an intermittent error based on whether an error has occurred in the same address in the past.