JP5545771B2 - Diagnostic device, diagnostic method, and diagnostic program diagnostic method - Google Patents

Description

  The present invention relates to a diagnostic device, a diagnostic method, and a diagnostic program, and more particularly, to a diagnostic device, a diagnostic method, and a diagnostic program for diagnosing a memory unit of a computer system.

  In a computer system, for example, when an error in data read from a memory unit is detected, it is important to immediately replace the faulty part of the memory unit and continue operation. Generally, when an error is detected in read data from the memory unit, the read address of the memory unit and the error bit of the read data are obtained. However, depending on the faulty part, there may be a fault spreading beyond the error detection part. Therefore, it is difficult to accurately identify the faulty site only with the memory read address and the error bit of the read data.

  For this reason, in conventional computer systems, as a method of determining the suspected part when a failure occurs in the memory unit, is it possible to point out all the parts that are likely to fail, judging from the failure status detected by the patrol of the diagnostic circuit? Or, a general diagnostic program was executed to identify a suspected failure site.

  For example, Japanese Patent Application Laid-Open No. 5-210597, “Patrol circuit of storage device” of Patent Document 1 describes a diagnostic circuit of a computer system, and checks error information detected by patrol to check the possibility of a failure. All parts are extracted.

  Further, for example, Japanese Patent Application Laid-Open No. 9-218798, “Intermittent Fault Diagnosis Method and Electronic Device” describes a computer system diagnosis method, and uses a general-purpose diagnosis program to suspect a fault. Diagnosis is performed to identify the site.

Japanese Patent Laid-Open No. 5-210597 (page 2-3) JP-A-9-218798 (page 3-4)

  However, the conventional diagnosis method has the following problems.

  The first problem is that the failure information detected by the diagnostic circuit is often general-purpose and vague information, and thus the failure-probable site cannot be accurately identified.

  The second problem is that diagnosis is often performed by a general-purpose diagnosis program regardless of the circuit configuration of the computer system, so that it is not possible to make a detailed diagnosis specialized to the detected failure situation, and the failure suspected part This means that it cannot be accurately specified, and it takes time to complete the diagnosis execution by the diagnostic program.

(Object of the present invention)
The present invention has been made in view of such circumstances, and an object of the present invention is to provide a diagnostic device, a diagnostic method, and a diagnostic program for quickly and accurately identifying a suspected site of failure in a memory unit of a computer system. .

  In order to solve the above-described problems, the diagnostic apparatus, diagnostic method, and diagnostic program according to the present invention mainly adopt the following characteristic configuration.

  (1) A diagnostic device according to the present invention is a diagnostic device for identifying a suspected failure site where an error has occurred when an error occurs by accessing a memory unit of a computer system, and includes a memory read address and an error Whether the error bit of the memory read data is temporarily held and at least means for re-accessing and retrying the same memory read address at the time of error occurrence, and whether the same error bit occurs again as a result of the retry Based on whether or not it is a fixed failure, if it is determined as a fixed failure, refer to the memory read address and the error bit of the memory read data at the time of occurrence of the error held in the holding means By performing the isolation test centering on the site where the resulting error occurred, Characterized in that it comprises means for identifying a suspect site.

  (2) A diagnostic method according to the present invention is a diagnostic method for specifying a suspected failure site where an error has occurred when an error occurs when accessing a memory unit of a computer system, and includes a memory read address and an error Whether or not it is a fixed failure based on whether or not the same error bit occurs again as a result of retrying by holding again the error bit of the memory read data and accessing the same memory read address at the time of error again If the error is obtained by referring to the memory read address and the error bit of the memory read data when the error is temporarily held, It is characterized by identifying a suspected fault site by performing a carving test.

  (3) The diagnostic program according to the present invention is characterized in that at least the diagnostic method described in (2) is implemented as a program executable by a computer.

  According to the diagnostic apparatus, diagnostic method, and diagnostic program of the present invention, the following effects can be obtained.

  The first effect is that when an error is detected in the memory read operation, the diagnosis is immediately performed, and the failure suspected site is narrowed down by performing a separation test around the site where the error has occurred. It is possible to prevent maintenance errors due to oversight or analysis errors.

  The second effect is that a diagnosis specialized for the error location detected by the memory read operation is performed, so that the suspected failure site can be identified quickly and accurately.

  The third effect is that since the diagnosis is performed immediately after the error is detected in the memory read operation, the error is highly reproducible and an accurate diagnosis including a marginal failure can be performed.

It is a block block diagram which shows an example of the block configuration of the diagnostic apparatus by this invention. It is a block block diagram which shows the example different from FIG. 1 of the block configuration of the computer system containing the diagnostic apparatus by this invention.

  Preferred embodiments of a diagnostic apparatus, a diagnostic method, and a diagnostic program according to the present invention will be described below with reference to the accompanying drawings. In the following description, the diagnostic apparatus and diagnostic method according to the present invention will be described. However, the diagnostic method may be implemented as a diagnostic program that can be executed by a computer, or the diagnostic program may be read by a computer. Needless to say, it may be recorded on a possible recording medium.

(Features of the present invention)
Prior to the description of the embodiments of the present invention, an outline of the features of the present invention will be described first. The present invention is characterized in that when a fault is detected in a memory device of a computer system, a suspected part can be identified at high speed and accurately.

  More specifically, at least a data error detection circuit, an error bit holding circuit, an error bit comparison circuit, a diagnostic read command generation circuit, a read address holding circuit, a diagnostic read address generation circuit, and a diagnostic control circuit are used for diagnosis of the memory device. If an error occurs in the read data of the memory device, a retry operation is performed to access the same memory read address again. If the same error bit is detected again, it is determined as a fixed fault and an error occurs. By carrying out the diagnostic read-out operation as a separate test centering on the selected site, it is possible to quickly and accurately specify the faulty site.

(Configuration example of embodiment)
Next, embodiments of the diagnostic apparatus and diagnostic method according to the present invention will be described. FIG. 1 is a block diagram showing an example of a block configuration of a diagnostic apparatus according to the present invention. The diagnostic apparatus is equipped in the computer system in order to diagnose a memory unit constituting the computer system. . Note that the computer system 100 shown in FIG. 1 shows only parts related to the diagnostic apparatus 10 according to the present invention, and omits parts that are not related.

  As shown in FIG. 1, a computer system 100 includes at least a diagnostic device 10 according to the present invention, a memory read control unit 2 and a memory unit 3 to be diagnosed.

  The diagnostic device 10 is a part that detects an error in data read from the memory unit 3 and performs diagnosis, and the memory read control unit 2 is a part that performs control to read data from the memory unit 3.

  Further, the memory read control unit 2 performs not only a read operation from the memory unit 3 during a normal operation but also a retry operation and a fault site isolation test for diagnosis of the memory unit 3 according to an instruction from the diagnostic device 10. This is a part that also performs read control of data from the memory unit 3, and includes a memory read command generation circuit 21 and a memory read address generation circuit 22.

  Here, the memory read command generation circuit 21 receives a memory read command 211 for a DIMM (Dual Inline Memory Module) 31 that is a first memory module of the memory unit 3 and a memory read command 212 for a DIMM 32 that is a second memory module. This is a part that is generated and output. The memory read address generation circuit 22 is a part that generates and outputs a memory read address 221 for the DIMM 31 that is the first memory module and the DIMM 32 that is the second memory module.

  The memory unit 3 includes at least a DIMM 31 that is a first memory module and a DIMM 32 that is a second memory module. The DIMM 31 is configured by a RAM (Random Access Memory) 311, and the DIMM 32 is configured by a RAM 321. When the memory read command 211 from the memory read command generation circuit 21 is valid, the DIMM 31 reads the data at the address specified by the memory read address 221 from the memory read address generation circuit 22 and converts it into the memory read data 33. Output. On the other hand, when the memory read command 212 from the memory read command generation circuit 21 is valid, the DIMM 32 reads the data at the address specified by the memory read address 221 from the memory read address generation circuit 22 to read the memory read data. To 33.

  In addition to the diagnostic readout control unit 1 that forms the main part of the diagnostic device 10, the diagnostic device 10 notifies the user of an input unit 20 for inputting an instruction related to diagnosis execution from the user, and the diagnosis execution result. The output part 30 for doing is comprised at least.

  The diagnostic read control unit 1 includes a data error detection circuit 11, an error bit holding circuit 12, an error bit comparison circuit 13, a diagnostic read command generation circuit 14, a read address holding circuit 15, a diagnostic read address generation circuit 16, and a diagnostic control circuit 17. It is comprised including at least.

  The data error detection circuit 11 pulls in the memory read data 33 read from the DIMM 31 and DIMM 32 of the memory unit 3 to detect an error in the memory read data, outputs the error report 112 to the diagnosis control circuit 17 and detects it. If the error is a correctable failure, an error bit 111 indicating the bit in which the error has occurred is output to the error bit holding circuit 12 and the error bit comparison circuit 13.

  The error bit holding circuit 12 temporarily holds the error bit 111 output from the data error detection circuit 11 and outputs the held error bit as an error bit 121 to the error bit comparison circuit 13. The error bit comparison circuit 13 compares the error bit 111 output from the data error detection circuit 11 with the error bit 121 from the error bit holding circuit 12 and outputs the comparison result 131 to the diagnostic read command generation circuit 14. To do.

  Based on the comparison result 131 from the error bit comparison circuit 13, the diagnostic read command generation circuit 14 focuses on the memory read command for diagnosis of the memory unit 3 (ie, a temporary error) centering on the portion at the time of error occurrence. Memory read command for retry to determine whether or not, and a memory read command for isolation test to isolate the faulty part when it is determined that the fault is a fixed fault, and read the memory of the memory read control unit 2 The command generation circuit 21 is instructed.

  The read address holding circuit 15 temporarily holds the memory read address 221 output from the memory read address generation circuit 22 of the memory read control unit 2, and uses the held memory read address as the memory read address 151 to the diagnostic read address generation circuit 16. Output. Further, the diagnostic read address generation circuit 16 is based on the memory read address 151 output from the read address holding circuit 15 and focuses on the memory read address for diagnosis of the memory unit 3 (ie, the portion at the time of error occurrence). A memory read control unit that generates a memory read address for retry for determining whether or not it is a temporary error and a memory read address for isolation test for isolating a faulty part when it is determined as a fixed fault 2 is output to the second memory read address generation circuit 22.

  The diagnostic read command generation circuit 14 not only generates the same memory read command as the read operation when an error occurs based on the comparison result 131, but also a different system centering on the command in which the error has occurred. It is also possible to generate a memory read command or the like for diagnosis. Further, the diagnostic read address generation circuit 16 is not limited to the address in the DIMM in which the error has occurred, but is different from the different address in the DIMM, the different RAM in the DIMM, or the DIMM around the address in which the error has occurred. It is also possible to generate a RAM address or the like in a different DIMM for diagnosis.

  The diagnosis control circuit 17 controls diagnosis of the memory unit 3 based on the error report 112 output from the data error detection circuit 11.

  In the diagnostic apparatus 10 having such a configuration, a diagnostic read command and a diagnostic read address generated based on a memory read address and an error bit at the time of error detection are used. Therefore, the memory reading operation for diagnosis of the memory unit 3 can be performed in detail, so that the occurrence site of the failure in the memory unit 3 can be narrowed down quickly and accurately.

(Description of operation of embodiment)
Next, an example of the operation of the diagnostic apparatus 10 shown in FIG. 1 will be described in detail.

  In the computer system 100 shown in FIG. 1, as described above, when the memory read command 211 from the memory read command generation circuit 21 of the memory read control unit 2 is valid during normal operation, the memory read control unit 2 Data of the address of the RAM 311 of the DIMM 31 designated by the memory read address 221 from the memory read address generation circuit 22 is read and output to the memory read data 33. The memory read data 33 read from the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10. Here, the memory read address 221 at this time is referred to as “address X”.

  At this time, when an error is detected in the data error detection circuit 11, the error report 112 reports to the diagnosis control circuit 17 that the error has been detected. The diagnostic control circuit 17 holds the error bit 111 at this time in the error bit holding circuit 12 and reads the memory from the same memory read address of the DIMM 31 again as a retry operation for determining whether or not there is a temporary error. The diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 are instructed to perform the operation. The diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 refer to the memory read address 151 held in the read address holding circuit 15 and generate the same memory read command and memory read address as when an error is detected. The memory read control unit 2 is instructed to perform the memory read operation from the “address X” where the error is detected (that is, retry operation) again.

  The memory read command generation circuit 21 and the memory read address generation circuit 22 of the memory read control unit 2 that have received the instructions from the diagnosis read command generation circuit 14 and the diagnosis read address generation circuit 16 of the diagnostic apparatus 10 The same memory read command 211 and the same memory read address 221 are set, and data is read again from the “address X” of the DIMM 31 of the memory unit 3 and output to the memory read data 33. The memory read data 33 read again from the “address X” of the DIMM 31 of the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10.

  If no error is detected in the data error detection circuit 11, the diagnosis control circuit 17 notifies the diagnosis control circuit 17 that the error is a temporary error, and the diagnosis control circuit 17 that receives the notification notifies the error bit holding circuit. The error bit 111 temporarily held in 12 is deleted, and the diagnosis operation is terminated.

  On the other hand, when the data error detection circuit 11 detects the error again, the error report 112 reports to the diagnosis control circuit 17 that the error has been detected again. The diagnostic control circuit 17 uses the error bit comparison circuit 13 to compare the error bit 111 at this time with the error bit 121 previously held in the error bit holding circuit 12.

  If the error bit 111 and the error bit 121 do not match, it is determined that the failure is a one-time temporary failure, the suspected failure portion is determined to be a marginal failure of the DIMM 31, and the diagnosis operation is terminated.

  On the other hand, if the error bit 111 and the error bit 121 match, the diagnosis control circuit 17 determines that a fixed failure has occurred, and then an error has occurred in order to isolate the fixed failure portion. The diagnostic read command generation circuit 14 and the diagnostic read that the memory read operation is performed for an arbitrary address different from the “address X” in the same DIMM 31 in which the error has occurred are transferred to the separation test operation centered on the part. An instruction is given to the address generation circuit 16.

  The diagnostic read address generation circuit 16 refers to the memory read address 151 held in the read address holding circuit 15, calculates an arbitrary address different from the “address X” of the DIMM 31, for example, “address Y”, and performs diagnostic read The command generation circuit 14 generates the same memory read command when an error occurs, and performs a memory read operation from “address Y” that is different from “address X” in the same DIMM 31 that detected the error. To the memory read control unit 2.

  The memory read command generation circuit 21 and the memory read address generation circuit 22 of the memory read control unit 2 that have received the instructions from the diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 of the diagnostic device 10 detect when an error is detected. The same memory read command 211 and memory read address 221 of “address Y” within the same DIMM 31 and different from “address X” are generated, and data is generated from “address Y” of DIMM 31 of the memory unit 3. Is output to the memory read data 33. Memory read data 33 read from the “address Y” of the DIMM 31 of the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10.

  When the data error detection circuit 11 does not detect an error, the diagnosis control circuit 17 that has received the notification to that effect determines that the suspected failure part is a part unique to the “address X” of the DIMM 31. End the diagnostic operation.

  On the other hand, when an error is detected in the data error detection circuit 11, an error is detected by the error report 112 even in “address Y” that is different from “address X” in the same DIMM 31 that previously detected the error. This is reported to the diagnosis control circuit 17. The diagnostic control circuit 17 uses the error bit comparison circuit 13 to compare the error bit 111 at this time with the error bit 121 previously held in the error bit holding circuit 12.

  If the error bit 111 and the error bit 121 do not match, it is determined that the suspected fault is a fixed fault in the address portion common to the “address X” and “address Y” in the DIMM 31. End the diagnostic operation.

  On the other hand, when the error bit 111 and the error bit 121 match, the diagnosis control circuit 17 determines that the failure is a fixed failure related to the bit position of the memory read data, and then uses the memory read data portion for the separation. As a test, the diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 are instructed to perform a memory read operation on an arbitrary address in the DIMM 32 different from the DIMM 31 in which an error has occurred, for example, “address Z”. .

  The diagnostic read address generation circuit 16 refers to the memory read address 151 held in the read address holding circuit 15 and calculates an arbitrary address in the DIMM 32 different from the DIMM 31 in which an error has occurred, for example, “address Z”. The diagnostic read command generation circuit 14 generates the same memory read command when an error occurs, and performs the memory read operation from “address Z” in the DIMM 32 different from the DIMM 31 that detected the error. The controller 2 is instructed.

  The memory read command generation circuit 21 of the memory read control unit 2 that has received instructions from the diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 of the diagnostic device 10 uses the same system of memory read commands as when an error was detected. However, the memory read command 212 for the DIMM 32 whose output destination is different from the DIMM 31 when the error is detected is generated. The memory read address generation circuit 22 generates a memory read address 221 of “address Z” in an arbitrary address in the DIMM 32 different from the DIMM 31 at the time when the error is detected.

  Therefore, based on the memory read command 212 and the memory read address 221 generated by the memory read command generation circuit 21 and the memory read address generation circuit 22, data is read from the “address Z” of the DIMM 32 of the memory unit 3 to read the memory read data. To 33. The memory read data 33 read from the “address Z” of the DIMM 32 of the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10.

  If no error is detected in the data error detection circuit 11, the diagnosis control circuit 17 that has received a notification to that effect determines that the suspected fault is a part unique to the DIMM 31 and terminates the diagnostic operation. To do.

  On the other hand, when the data error detection circuit 11 detects an error, the error report 112 reports to the diagnostic control circuit 17 that the error has been detected in the DIMM 32 different from the DIMM 31 that has detected the error. . The diagnostic control circuit 17 uses the error bit comparison circuit 13 to compare the error bit 111 at this time with the error bit 121 previously held in the error bit holding circuit 12.

  When the error bit 111 and the error bit 121 do not match, the suspected failure part is a fixed failure of a part common to “address X”, “address Y” of the DIMM 31 and “address Z” of the DIMM 32. It judges that it is a thing, and a diagnostic operation is complete | finished.

  On the other hand, if the error bit 111 and the error bit 121 match, the diagnosis control circuit 17 determines that the suspected fault is a fixed fault related to the bit position of the memory read data common to the DIMMs 31 and 32. The diagnostic operation is terminated.

  In the above description, if the error bit 111 and the error bit 121 do not match, the suspected faulty part is immediately identified. However, the memory read operation for diagnosis for further isolating the faulty part To repeat the diagnostic isolation test until errors are no longer detected at different addresses or until the same error bit is detected, allowing more accurate isolation of the suspected failure site. May be.

  In the diagnostic read command generation circuit 14 of the diagnostic device 10, when generating a memory read command for diagnosis, a memory read command having the same command system as that at the time of error detection is generated. A memory read command 211 at the time of error detection from the circuit 21 is further provided, and a memory read command having a command system different from that at the time of error detection is also provided in order to determine whether there is a failure in the memory read command system circuit unit. You may make it add and produce | generate for a carving test.

(Explanation of effect of embodiment)
As described in detail above, in the present embodiment, the following effects can be obtained.

  The first effect is that when an error is detected in the memory read operation, the diagnosis is immediately performed, and the failure suspected site is narrowed down by performing a separation test around the site where the error has occurred. It is possible to prevent maintenance errors due to oversight or analysis errors.

  The second effect is that a diagnosis specialized for the error location detected by the memory read operation is performed, so that the suspected failure site can be identified quickly and accurately.

  The third effect is that since the diagnosis is performed immediately after the error is detected in the memory read operation, the error is highly reproducible and an accurate diagnosis including a marginal failure can be performed.

(Configuration example of other embodiment)
In the computer system 100 shown in FIG. 1 described above, the DIMM 31 and the DIMM 32 in the memory unit 3 to be diagnosed are each configured by the RAM 311 and the RAM 321 and a single RAM. However, the present invention is not limited to this, and the DIMM in the memory unit to be diagnosed may be composed of a plurality of RAMs. In the case of such a configuration, the failure suspected part is further finely divided with respect to the plurality of RAMs constituting the DIMM in the memory unit 3 in the diagnostic device 10, and among the failures of the DIMM, the plurality of RAMs in the DIMM Among them, it is possible to specify which RAM has a failure.

  FIG. 2 is a block configuration diagram showing an example different from FIG. 1 of the block configuration of the computer system including the diagnostic apparatus according to the present invention. A computer system 100A shown in FIG. 2 includes a diagnostic device 10 according to the present invention, a memory read control unit 2 and a memory unit 3A to be diagnosed, at least, as in the case of FIG. Here, the diagnosis device 10 and the memory read control unit 2 are exactly the same as the block configuration shown in FIG. 1, and further description thereof is omitted here.

  On the other hand, the memory unit 3A is different from the case of FIG. 1 in that the memory module is only a single DIMM 31A. The DIMM 31A is composed of two RAMs, a first RAM 311 and a second RAM 312. Has been. When the first RAM 311 and the second RAM 312 are assigned different read addresses, and the memory read command 211 generated by the memory read command generation circuit 21 of the memory read control unit 2 is valid, Of the first RAM 311 and the second RAM 312, for the RAM designated by the memory read address 221 from the memory read address generation circuit 22, the address data in the RAM is read and output to the memory read data 33. To do.

  The memory unit 3A shown in FIG. 2 is different from the memory unit 3 shown in FIG. 1 as described above. Since the memory module is only a single DIMM 31A as described above, the memory read command generation circuit 21 of the memory read control unit 2 is used. As for the output from, the memory read command 212 provided for enabling the DIMM 32 of the second memory module in the case of FIG. 1 is not necessary and is not shown in FIG.

(Description of operation of other embodiment)
Next, an example of the operation of the diagnostic apparatus 10 shown in FIG. 2 will be described in detail.

  Also in the computer system 100A shown in FIG. 2, as described above, when the memory read command 211 from the memory read command generation circuit 21 of the memory read control unit 2 is valid during normal operation, the memory read control unit 2 For example, the data of the address of the first RAM 311 of the DIMM 31 A designated by the memory read address 221 from the memory read address generation circuit 22 is read and output to the memory read data 33. The memory read data 33 read from the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10. Here, the memory read address 221 at this time is referred to as “address X”. That is, “address X” is an address in the first RAM 311.

  At this time, when an error is detected in the data error detection circuit 11, the error report 112 reports to the diagnosis control circuit 17 that the error has been detected. The diagnosis control circuit 17 holds the error bit 111 at this time in the error bit holding circuit 12, and again, as a retry operation for determining whether or not there is a temporary error, the same memory in the first RAM 311 of the DIMM 31A again. The diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 are instructed to perform the memory read operation from the read address. The diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 refer to the memory read address 151 held in the read address holding circuit 15 and generate the same memory read command and memory read address as when an error is detected. The memory read control unit 2 is instructed to perform the memory read operation from the “address X” where the error is detected (that is, retry operation) again.

  The memory read command generation circuit 21 and the memory read address generation circuit 22 of the memory read control unit 2 that have received the instructions from the diagnosis read command generation circuit 14 and the diagnosis read address generation circuit 16 of the diagnostic apparatus 10 The same memory read command 211 and the same memory read address 221 are set, and the data is read again from the “address X” of the first RAM 311 of the DIMM 31 A of the memory unit 3 and output to the memory read data 33. The memory read data 33 read again from the “address X” of the DIMM 31 </ b> A of the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10.

  If no error is detected in the data error detection circuit 11, the diagnosis control circuit 17 notifies the diagnosis control circuit 17 that the error is a temporary error, and the diagnosis control circuit 17 that receives the notification notifies the error bit holding circuit. The error bit 111 temporarily held in 12 is deleted, and the diagnosis operation is terminated.

  On the other hand, when the data error detection circuit 11 detects the error again, the error report 112 reports to the diagnosis control circuit 17 that the error has been detected again. The diagnostic control circuit 17 uses the error bit comparison circuit 13 to compare the error bit 111 at this time with the error bit 121 previously held in the error bit holding circuit 12.

  If the error bit 111 and the error bit 121 do not match, it is determined that the failure is a one-time temporary failure, and the suspected failure portion is determined to be a marginal failure in the first RAM 311 of the DIMM 31A. Exit.

  On the other hand, if the error bit 111 and the error bit 121 match, the diagnosis control circuit 17 determines that a fixed failure has occurred, and then an error has occurred in order to isolate the fixed failure portion. Shift to the segmentation test operation centered on the part, and generate a diagnostic read command to perform a memory read operation for an arbitrary address different from “address X” in the first RAM 311 of the same DIMM 31A where the error has occurred The circuit 14 and the diagnostic read address generation circuit 16 are instructed.

  The diagnostic read address generation circuit 16 refers to the memory read address 151 held in the read address holding circuit 15, and any address different from “address X” in the first RAM 311 of the DIMM 31 </ b> A, for example, “address Y”. The diagnostic read command generation circuit 14 generates the same memory read command when an error occurs, and is different from the “address X” in the first RAM 311 of the same DIMM 31A that detected the error. The memory read control unit 2 is instructed to perform a memory read operation from “address Y”.

  The memory read command generation circuit 21 and the memory read address generation circuit 22 of the memory read control unit 2 that have received the instructions from the diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 of the diagnostic device 10 detect when an error is detected. The memory read command 211 and the memory read address 221 of the “address Y” different from the “address X” in the first RAM 311 of the same DIMM 31A are generated, and the DIMM 31A of the DIMM 31A of the memory unit 3 Data is read from “address Y” in the RAM 311 of 1 and output to the memory read data 33. Memory read data 33 read from “address Y” in the first RAM 311 of the DIMM 31 </ b> A of the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10.

  When the data error detection circuit 11 does not detect an error, the diagnosis control circuit 17 that has received a notification to that effect indicates that the suspected failure part is a part unique to “address X” in the first RAM 311 of the DIMM 31A. It judges that it is a thing, and a diagnostic operation is complete | finished.

  On the other hand, when an error is detected in the data error detection circuit 11, an “address Y” that is different from “address X” in the first RAM 311 of the same DIMM 31 </ b> A that previously detected the error is detected based on the error report 112. The fact that an error has been detected is also reported to the diagnostic control circuit 17. The diagnostic control circuit 17 uses the error bit comparison circuit 13 to compare the error bit 111 at this time with the error bit 121 previously held in the error bit holding circuit 12.

  If the error bit 111 and the error bit 121 do not match, the suspected fault is a fixed fault in the address part common to the “address X” and “address Y” in the first RAM 311 of the DIMM 31A. Judgment is made and the diagnosis operation is terminated.

  On the other hand, when the error bit 111 and the error bit 121 match, the diagnosis control circuit 17 determines that the failure is a fixed failure related to the bit position of the memory read data, and then uses the memory read data portion for the separation. As a test, the diagnostic read command generation circuit 14 and the diagnostic read address indicate that the memory read operation is performed for an arbitrary address in the second RAM 312 different from the first RAM 311 of the DIMM 31A in which an error has occurred, for example, “address Z”. The generation circuit 16 is instructed.

  The diagnostic read address generation circuit 16 refers to the memory read address 151 held in the read address holding circuit 15, and any address in the second RAM 312 that is different from the first RAM 311 of the DIMM 31 </ b> A in which an error has occurred, for example, The “address Z” is calculated, and the diagnostic read command generation circuit 14 generates the same memory read command at the time of occurrence of the error, and in the second RAM 312 different from the first RAM 311 of the DIMM 31A in which the error has occurred. The memory read control unit 2 is instructed to perform a memory read operation from “address Z”.

  The memory read command generation circuit 21 of the memory read control unit 2 that has received instructions from the diagnostic read command generation circuit 14 and the diagnostic read address generation circuit 16 of the diagnostic device 10 uses the same system of memory read commands as when an error was detected. 211 is generated. Further, the memory read address generation circuit 22 generates a memory read address 221 of “address Z” of an arbitrary address in the second RAM 312 different from the first RAM 311 of the DIMM 31A when an error is detected. .

  Therefore, based on the memory read command 211 and the memory read address 221 generated by the memory read command generation circuit 21 and the memory read address generation circuit 22, the data from the “address Z” in the second RAM 312 of the DIMM 31 </ b> A in the memory unit 3 is stored. Is output to the memory read data 33. Memory read data 33 read from “address Z” in the second RAM 312 of the DIMM 31 </ b> A of the memory unit 3 is input to the data error detection circuit 11 of the diagnostic device 10.

  If the data error detection circuit 11 does not detect an error, the diagnosis control circuit 17 that has received the notification to that effect determines that the suspected failure site is a site unique to the first RAM 311 of the DIMM 31A. The diagnostic operation is terminated.

  On the other hand, when the data error detection circuit 11 detects an error, it is diagnosed by the error report 112 that the error has been detected also in the second RAM 312 different from the first RAM 311 of the DIMM 31A that detected the error first. Report to the control circuit 17. The diagnostic control circuit 17 uses the error bit comparison circuit 13 to compare the error bit 111 at this time with the error bit 121 previously held in the error bit holding circuit 12.

  If the error bit 111 and the error bit 121 do not match, the suspected failure part is assigned to “address X” and “address Y” of the first RAM 311 of the DIMM 31A and “address Z” of the second RAM 312. The diagnosis operation is terminated by determining that the common part is a fixed failure.

  On the other hand, if the error bit 111 and the error bit 121 match, the diagnosis control circuit 17 indicates that the suspected failure part is fixed regarding the bit position of the memory read data common to the first RAM 311 and the second RAM 312 of the DIMM 31A. The diagnosis operation is terminated by determining that it is a failure.

  In the above description, if the error bit 111 and the error bit 121 do not match, the suspected fault site is immediately identified. However, as in the above-described embodiment of FIG. Add a memory read operation for diagnosis to isolate the faulty part and repeat the diagnosis isolation test until no error is detected at different addresses or until the same error bit is detected, You may make it perform more exact isolation | separation of a failure suspected site | part.

  Further, when the diagnostic read command generation circuit 14 of the diagnostic apparatus 10 generates a memory read command for diagnosis, a memory read command having the same command system as that at the time of error detection is generated. As in the case of the embodiment, the memory read command generation circuit 21 further includes a holding circuit that temporarily holds the memory read command 211 when an error is detected, and an error is detected in order to determine whether there is a failure in the memory read command circuit unit. A memory read command having a command system different from that at the time of detection may be additionally generated for the separation test.

  As described above, even in the case where the configuration of the memory unit is different from the case of the above-described embodiment of FIG. 1, as in the case of the embodiment of FIG. can do.

  The configuration of the preferred embodiment of the present invention has been described above. However, it should be noted that such embodiments are merely examples of the present invention and do not limit the present invention in any way. Those skilled in the art will readily understand that various modifications and changes can be made according to a specific application without departing from the gist of the present invention.

DESCRIPTION OF SYMBOLS 1 Diagnostic memory read control part 2 Memory read control part 3 Memory part 3A Memory part 10 Diagnostic apparatus 11 Data error detection circuit 12 Error bit holding circuit 13 Error bit comparison circuit 14 Diagnostic read command generation circuit 15 Read address holding circuit 16 Diagnostic read address Generation circuit 17 Diagnostic control circuit 20 Input unit 30 Output unit 21 Memory read command generation circuit 22 Memory read address generation circuit 31 DIMM (first memory module)
31A DIMM
32 DIMM (second memory module)
33 Memory read data 100 Computer system 100A Computer system 111 Error bit 112 Error report 121 Error bit 131 Comparison result 151 Memory read address 211 Memory read command 212 Memory read command 221 Memory read address 311 RAM
312 RAM
321 RAM

Claims (7)

  A diagnostic device that identifies a suspected failure site when an error occurs when accessing a memory unit of a computer system, and temporarily stores an error bit of a memory read address and a memory read data when the error occurs A holding means for holding, and a means for re-accessing and retrying the same memory read address at the time of error occurrence, and whether or not a fixed failure is determined based on whether or not the same error bit occurs again as a retry result If it is determined that the failure is a fixed failure, the error is obtained by referring to the memory read address and the error bit of the memory read data when the error is held in the holding means. A means to identify the suspected failure site by performing a carving test Diagnostic apparatus characterized by there.   As the means for identifying the suspected failure part by performing a separation test centering on the part where the error has occurred when determined as a fixed failure, different addresses in the same memory module of the memory unit at the time of the error, or When a plurality of RAMs (Random Access Memory) exist in the same memory module of the memory unit at the time of error occurrence Arbitrary address in RAM different from the RAM at the time of error occurrence or the memory unit at the time of error occurrence When there is a plurality of memory modules, diagnostic access is made to any one or more addresses in the memory module different from the memory module at the time of error occurrence, and the error bit at the time of error occurrence Based on whether or not the same error bit occurs The diagnostic device according to claim 1.   As a means for identifying a suspected faulty part by further comprising a holding means for temporarily holding a memory read command at the time of an error, and performing a separation test centering on the part where the error has occurred when it is determined as a fixed fault The memory at the time of error occurrence using either or both of the memory read command that is the same as the memory read command at the time of error occurrence or the memory read command having a command system different from the memory read command at the time of error occurrence 3. The diagnosis suspected part is accessed, and the suspected failure part is specified based on whether or not an error bit identical to the error bit at the time of occurrence of the error occurs. Diagnostic equipment.   A diagnostic method for identifying a suspected failure location when an error occurs by accessing a memory unit of a computer system, and temporarily stores an error bit of a memory read address and memory read data at the time of the error occurrence When it is determined that it is a fixed failure based on whether the same error bit occurs again as a result of retrying after holding the same memory read address at the time of error and retrying The failure suspected part is obtained by performing a separation test centering on the part where the error is obtained by referring to the memory read address and the error bit of the memory read data when the error is temporarily held. A diagnostic method characterized by specifying.   When identifying a suspected failure location by performing a separation test centering on the location where the error occurred when it is determined as a fixed failure, a different address in the same memory module of the memory unit at the time of the error or an error When a plurality of RAMs (Random Access Memory) exist in the same memory module of the memory unit at the time of occurrence, an arbitrary address in the RAM different from the RAM at the time of error occurrence or the memory unit at the time of error occurrence When there are multiple memory modules Diagnostic access is made to any one or more addresses in the memory module that are different from the memory module at the time of error occurrence, and the same as the error bit at the time of error occurrence A failure suspicious part is specified based on whether or not an error bit is generated. 4. The diagnostic method according to 4.   The memory read command at the time of error occurrence is further temporarily stored, and when it is determined that it is a fixed failure, when the failure is identified by performing the isolation test centered on the location where the error occurred, Diagnose the memory unit when an error occurs using either or both of the memory read command that is the same as the memory read command or a memory read command that has a different command system from the memory read command when an error occurs 6. The diagnosis method according to claim 4 or 5, wherein a fault suspected part is specified based on whether or not an error bit identical to the error bit at the time of occurrence of the error is generated.   7. A diagnostic program, wherein the diagnostic method according to claim 4 is implemented as a program executable by a computer.

Images