Management method of peripheral component interconnection interface error report

Abstract

The invention discloses a data management method, and particularly relates to a method for managing a peripheral component interconnection interface error report, which comprises the following steps: a processor runs a first program when executing a basic input / output system program, and the first program comprises the following steps: receiving an error event to trigger a system management interrupt; reading an error report closing instruction; judging whether the error report closing instruction corresponds to an error event or not; when it is judged that the error report closing instruction corresponds to the error event, closing the permission of triggering system management interruption by the error event; and sending the system event record to the substrate management controller. The baseboard management controller is suitable for running a second program, and the second program comprises the following steps: receiving a system event record; processing an error event set of the system event record according to a default algorithm to judge whether an error report closing instruction is generated or not; and sending and storing a generated error report closing instruction.

Description

Technical Field

[0001] The invention discloses a data management method, and more particularly, a data management method applicable to a server. Background Art

[0002] In a computer system, the processor can connect and communicate with peripheral components (such as display adapters, memory, and network cards) through the peripheral component interconnect express (PCI-e) interface. When an error occurs in a peripheral component (such as data error, cable damage, or unstable power supply), the detection mechanism of the peripheral component can transmit the error signal to the processor through the PCI-e interface, so that the system can perform self-error repair or notify the user to perform manual repair. The above errors can be divided into uncorrectable errors and correctable errors. Uncorrectable errors can be divided into fatal errors and non-fatal errors. Some uncorrectable errors can be repaired by system software, while some uncorrectable errors must be reset or replaced by the user. Correctable errors can usually be self-repaired by peripheral components, such as errors caused by short-term transmission errors, electromagnetic interference, or unstable power supply.

[0003] After the processor receives an error signal from a peripheral component, it can trigger a System Management Interrupt (SMI) through the Basic Input / Output system (BIOS) to enter System Management Mode (SMM), which allows the processor to suspend the processing program in progress and perform maintenance procedures on peripheral components. When the processor executes the System Management Mode, other devices and programs in the system cannot interfere with or modify the program being executed by the processor, thereby ensuring the safety and reliability of the overall system.

[0004] However, even though most correctable errors can usually be self-repaired by peripheral components, too frequent system management interrupts will cause the processor's processing program to be constantly interrupted, resulting in low execution efficiency and even system performance crashes. In particular, correctable errors will send a large number of error signals in a short period of time, causing the basic input and output system to frequently interrupt system management, which will cause the entire system to stop operating. Summary of the invention

[0005] In view of this, the applicant proposes a plurality of management methods for peripheral component interconnect interface error reports. In some embodiments, the management method for peripheral component interconnect interface error reports is applicable to a server, the server comprising a processor, a first memory coupled to the processor, a baseboard management controller, and a second memory coupled to the baseboard management controller, the processor being suitable for running a basic input and output system. The management method of this embodiment comprises: the processor runs a first program when executing the basic input and output system program, comprising: receiving an error event to trigger a system management interrupt; reading an error report closing instruction from the first memory; determining whether the error report closing instruction corresponds to the error event; when determining that the error report closing instruction corresponds to the error event, closing the permission of the error event to trigger the system management interrupt; and sending a system event record to the baseboard management controller, the system event record comprising an error event set, the error event set comprising the error event; and the baseboard management controller is suitable for running a second program, comprising: receiving the system event record; processing the error event set according to a default algorithm to determine whether the error report closing instruction is generated; and when the error report closing instruction is generated, sending and storing the error report closing instruction to the first memory.

[0006] In particular, the default algorithm includes: capturing a subset of error events corresponding to the same error type in the error event set; determining whether the number of the error event subset within a default period is greater than or equal to a counting threshold; and generating the error report closing instruction when the number of the error event subset within the default period is greater than or equal to the counting threshold.

[0007] In particular, the method further includes waiting to receive a next system event record when the number of the error event subset within the default period is less than the counting threshold.

[0008] Particularly, the baseboard management controller is further used to set the counting threshold. The management method further comprises the baseboard management controller setting and storing the counting threshold or the default period in the second memory after receiving a setting instruction for setting the counting threshold or the default period from an external device.

[0009] In particular, the error event subset includes a plurality of error event subsets, and the error events included in each error event subset correspond to different error types. The default algorithm further includes sequentially capturing each error event subset to determine whether the number of each error event subset within the default period is greater than or equal to the counting threshold.

[0010] In particular, the default algorithm includes: extracting a subset of error events corresponding to the same error type in the error event set; determining whether the time interval between the last two data in the error event subset is less than or equal to a time threshold; and generating the error report closing instruction when the time interval between the last two data in the error event subset is less than or equal to the time threshold.

[0011] In particular, it further includes waiting to receive the next system event record when the time interval between the last two data in the error event subset is greater than the time threshold.

[0012] Particularly, the baseboard management controller is further used to set the time threshold. The management method further comprises the following steps: after the baseboard management controller receives a setting instruction for setting the time threshold from an external device, the baseboard management controller sets and stores the time threshold in the second memory.

[0013] In particular, the default algorithm includes: extracting a subset of error events corresponding to the same error type in the error event set; determining whether the average value of the time intervals between multiple data items in the error event subset is less than or equal to a time threshold; and generating the error report closing instruction when the average value of the time intervals between multiple data items in the error event subset is less than or equal to the time threshold.

[0014] In particular, when it is determined that the error report closing instruction does not correspond to the error event, the system event record is sent to the baseboard management controller and the system management interrupt is terminated.

[0015] Compared with the prior art, the management method of the peripheral component interconnect interface error report of the present invention has the advantage of adjusting the judgment conditions of the error report closing command through the baseboard management controller in that the parameters of the default algorithm, including the default period, count threshold and time threshold, can be modified in real time in response to the adjustment of the system status or management conditions without modifying the program code of the basic input and output system. For example, the baseboard management controller runs an application to present the system event record to the user. The user or the analysis program can evaluate the system status of the server based on the system event record and change the management plan of the server, connect to the management interface through the network interface of the baseboard management controller through a remote external device, or enter the management interface of the baseboard management controller through a transmission line, and then adjust the parameters of the default algorithm in real time, thereby affecting the SMI triggering authority of various types of error events. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Other features and effects of the present invention will be clearly presented in the embodiments with reference to the accompanying drawings, in which: Figure 1 A block diagram of a server according to some embodiments of the present invention; Figure 2A flow chart of a first procedure of a method for managing peripheral component interconnect interface error reporting according to some embodiments of the present invention; and Figure 3 A flow chart of a second procedure of a method for managing peripheral component interconnect interface error reporting according to some embodiments of the present invention. DETAILED DESCRIPTION

[0017] Before the present invention is described in detail, it should be noted that similar elements are denoted by the same reference numerals in the following description.

[0018] Figure 1 is a block diagram of a server 10 according to some embodiments of the present invention, please refer to Figure 1 In some embodiments, the server 10 includes a processor 101, a platform controller hub (Platform Controller Hub, PCH) 102, a baseboard management controller (BMC) 103, a first memory 104, and a second memory 105. The processor 101 is coupled to the first memory 104, the baseboard management controller 103, and the peripheral component 20 through the platform controller hub 102, and the baseboard management controller 103 is coupled to the second memory 105. The coupling is not limited to direct connection or indirect connection, and only allows data transmission between components.

[0019] The processor 101 may be, but is not limited to, a SoC chip, a central processing unit (CPU), a microcontroller unit (MCU), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA) or a logic circuit. The processor 101 is suitable for running a basic input / output system to initialize, detect or drive a hardware device. The basic input / output system receives an error signal sent by the peripheral component 20 to trigger a system management interrupt. In some embodiments, the basic input / output system sets permissions for triggering a system management interrupt for different peripheral components 20. For example, when the temperature of the display adapter is too high and an error signal is generated, the basic input / output system determines that the display adapter has the SMI trigger permission (e.g., logic value 1), and then triggers a system management interrupt; otherwise, the basic input / output system determines that the display adapter does not have the SMI trigger permission (e.g., logic value 0), and then does not trigger a system management interrupt. In other embodiments, the basic input / output system sets permissions for triggering a system management interrupt for different error signals sent by the peripheral component 20. For example, when the temperature of the display adapter is too high and an error signal is generated, the basic input and output system determines that the temperature error event of the display adapter has SMI trigger permission (for example, logical value 1), and thus triggers the system management interrupt; at the same time, the basic input and output system determines that the communication error event of the display adapter does not have SMI trigger permission (for example, logical value 0), and thus does not trigger the system management interrupt.

[0020] The processor 101 is connected to and controls peripheral chips or devices through the platform path controller 102. The platform path controller 102 may be provided with a control chip of one or more hardware communication interfaces, and its communication interface may be but not limited to a Universal Serial Bus (USB), a Serial Advanced Technology Attachment (SATA), a High Definition Multimedia Interface (HDMI), an Ethernet or a peripheral component interconnection interface. In addition, in the server 10, the processor 101 communicates with the baseboard management controller 103 through the platform path controller 102, so as to allow the baseboard management controller 103 to parse the data generated by the processor 101 and notify the user, or transmit the control signal to the processor 101 for control, repair or maintenance.

[0021] The memory may be a flash memory or a read-only memory (ROM), such as but not limited to an erasable programmable read-only memory (EPROM), a flash read-only memory (Flash ROM), an electrically erasable programmable read-only memory (EEPROM), a field replaceable unit (FRU), or other storage technologies, such as a CD-ROM, a DVD, a disk, a tape, a USB flash drive, a hard disk, or a memory card. In some embodiments, the first memory 104 is a BIOS memory, which is used to store the program code of the basic input and output system, and the processor 101 runs the first program of the peripheral component interconnect interface error report management method when executing the above program code. In other embodiments, there are multiple first memories 104, one of which is a BIOS memory, and another first memory 104 is used to store the program code of the first program of the peripheral component interconnect interface error report management method. In some embodiments, the second memory 105 is used to store program codes of a second program of the peripheral component interconnect interface error report management method. When the baseboard management controller 103 executes the program codes, the second program of the peripheral component interconnect interface error report management method is executed.

[0022] Figure 2 is a flowchart of the first procedure of the peripheral component interconnect interface error report management method according to some embodiments of the present invention, please refer to Figure 2. The processor 101 is suitable for running the first program when executing the basic input and output system. After the peripheral component 20 transmits the error signal to the processor 101, the basic input and output system determines whether to trigger the system management interrupt according to the SMI trigger permission of the error signal. When the error signal does not have the SMI trigger permission, the basic input and output system does not continue to execute the first program; when the error signal has the SMI trigger permission, the basic input and output system executes the system management interrupt (step S101) to enter the system management mode. Here, different types of error signals, such as different devices or different causes of errors, can have different SMI trigger permissions. The basic input and output system can set the SMI trigger permission for a specific error event, so when the SMI trigger permission of a specific error event is turned off, it does not affect other error events triggering the system management interrupt. In some embodiments, the error event may refer to a correctable error. By turning off the SMI trigger permission of the correctable error, it is avoided that a large number of error signals of correctable errors occur in a short period of time, causing the basic input and output system to frequently interrupt the system management.

[0023] When the error event triggers the system management interrupt, the basic input and output system reads the error report closing instruction from the memory (step S102), and determines whether the error report closing instruction corresponds to the error event that currently triggers the system management interrupt (step S103). In detail, the error report closing instruction can be generated by the baseboard management controller 103 and stored in the first memory 104. The storage can mean that the baseboard management controller 103 directly stores the error report closing instruction in the first memory 104, or the baseboard management controller 103 sends the error report closing instruction to the processor 101 and requests the processor 101 to store the error report closing instruction in the first memory 104. The first memory 104 can be used to store a database of multiple error report closing instructions containing different error events, and the basic input and output system reads one or more error report closing instructions from the database.

[0024] In step S103, the basic input and output system determines whether the error report closing instruction read corresponds to the same type of error event that currently triggers the system management interrupt. In detail, multiple error events constitute an error event set, and the error event set is divided into a plurality of error event subsets, each of which corresponds to a different error event type. For example, the error event set can include 10 error events, and the number of error event types can be 8, including a first error type, a second error type, ..., and an eighth error type. Then, the 10 error events will be classified into one of the eight error event types. For example, the first error type includes 3 error events that occur at different times, and the second to eighth error types each include 1 error event. The error report closing instruction can be generated by the server 10 in the past for any of the first to eighth error types, and is only used to set the SMI trigger permission to close the error event of that type. When the error event currently triggering the system management interrupt is of the same type as an error event that occurred in the past, and the baseboard management controller 103 generates an error report shutdown instruction for the error event that occurred in the past, the correspondence may refer to the relative relationship between the error event currently triggering the system management interrupt and the aforementioned error report shutdown instruction.

[0025] In some embodiments, when the basic input and output system determines that the error report shutdown command does not correspond to the error event that currently triggers the system management interrupt (step S103, the judgment result is "no"), the system event record is sent to the baseboard management controller 103 (step S105), and then the system management interrupt is terminated (step S106). The system event log (System EventLog, SEL) is used to monitor and record the operating status of the server 10 system. The system event log may include warnings, errors, and abnormal conditions that occur in the peripheral component 20, and record them as error events of the peripheral component 20. The system event log includes an error event set. In this embodiment, the error event set may include error event records that have occurred from the past to the present, including the error event that triggered the system management interrupt at that time.

[0026] For example, when the temperature of the display adapter is too high and an error signal triggers the current system management interrupt (step S101), the basic input and output system reads the database of the first memory 104 and determines that there is an error report shutdown command. The error report shutdown command is generated by the display adapter due to multiple signal transmission errors in the past. Because the error report shutdown command does not correspond to the error event that triggers the system management interrupt this time (step S103, the judgment result is "No"), the basic input and output system directly sends the system event record containing the error event that triggers the system management interrupt this time to the baseboard management controller 103 (step S105). In some embodiments, when the basic input and output system determines that there is no error report shutdown command in the first memory 104, it is determined that the error report shutdown command does not correspond to the error event that currently triggers the system management interrupt (step S103, the judgment result is "No").

[0027] When the basic input and output system determines that the error report closing instruction corresponds to the error event that currently triggers the system management interrupt (step S103, the judgment result is "yes"), the permission for the error event to trigger the system management interrupt is closed (step S104). That is, when the peripheral component 20 has an error event of the same type as this time again in the future, its error signal will no longer trigger the basic input and output system to execute the system management interrupt (step S101). In some embodiments, after the basic input and output system determines to close the permission for the error event to trigger the system management interrupt (step S104), the system event record is sent to the baseboard management controller 103 (step S105), and the system management interrupt is ended (step S106). In this embodiment, the error event set of the system event record can include the error event that triggered the system management interrupt this time.

[0028] For example, when the display adapter cannot be driven and triggers the current system management interrupt (step S101), the basic input and output system reads the database of the first memory 104 and determines that there is an error report shutdown instruction. The error report shutdown instruction is generated by the display adapter because it cannot be driven many times in the past. Because the error report shutdown instruction corresponds to the error event that triggers the system management interrupt this time (step S103, the judgment result is "yes"), the basic input and output system closes the permission to trigger the system management interrupt corresponding to the error event that the display adapter cannot be driven (step S104), and sends the system event record containing the error event that triggers the system management interrupt this time to the baseboard management controller 103 (step S105). In this way, when the error event that the display adapter cannot be driven occurs again in the future, the basic input and output system will no longer be triggered to execute the system management interrupt (step S101); however, if other display adapters installed in the server 10 also have an error event that cannot be driven, before the SMI triggering permission of the error event is closed, it is still allowed to trigger the basic input and output system to execute the system management interrupt (step S101).

[0029] Figure 3 is a flow chart of the second procedure of the method for managing peripheral component interconnect interface error reporting according to some embodiments of the present invention, please refer to Figure 3 . The baseboard management controller 103 is suitable for running a second program. In the first program, the basic input and output system run by the processor 101 sends the system event record to the baseboard management controller 103 (step S105). In the second program, when the baseboard management controller 103 receives the system event record (step S201), it processes the error event set according to the default algorithm. The default algorithm may include the calculation logic and its parameters programmed by the user, or the calculation logic and its parameters set through statistics or machine learning. The default algorithm can be used to analyze or count the error event set contained in the system event record. For example, sort or filter error events; analyze the error type (for example, based on the error code), occurrence time, number of occurrences, and frequency of occurrence of the error event; analyze the peripheral components 20 that generate the error event. The baseboard management controller 103 analyzes the error event set according to the default algorithm to determine whether to generate an error report shutdown instruction.

[0030] Figure 3Steps S202 to S204 are the default algorithms of an embodiment disclosed by the present invention. In this embodiment, the baseboard management controller 103 extracts a subset of error events corresponding to the same error type from the error event set of the system event record (step S202). For example, the error event set includes five error events, namely, error event 1 of display adapter overtemperature warning, error event 2 of display adapter overtemperature warning, error event 1 of fan unstable system power supply warning, error event 3 of display adapter overtemperature warning, and error event 1 of sound card driver abnormality. Among them, error events 1 to 3 of display adapter overtemperature warning correspond to the same error type (display adapter overtemperature), and the three error events of the same error type constitute the error event subset. In some embodiments, the error events extracted by the baseboard management controller 103 are all correctable error events, so as to judge a large number of correctable error events in a short period of time to avoid affecting the system operation. The baseboard management controller 103 analyzes the occurrence time and quantity of each error event in the error event subset to determine whether the quantity of the error event subset within the default period is greater than or equal to the counting threshold (step S203). The default period may be a period calculated from the recording time of the last error event in the error event subset, and the counting threshold may be the upper limit of the number of error events allowed by the administrator. When the baseboard management controller 103 determines that the number of the error event subset within the default period is greater than or equal to the counting threshold (step S203, the judgment result is "yes"), it further determines whether the interval between the last two data in the error event subset is less than or equal to the time threshold (step S204). The time threshold may be the minimum time interval between two error events allowed by the administrator. The default period, counting threshold and time threshold are parameters of the default algorithm, which can be set by the user or generated by a statistical algorithm or a machine learning algorithm. When the baseboard management controller 103 determines that the interval between the last two data in the error event subset is less than or equal to the time threshold (step S204, the judgment result is "yes"), an error report closing instruction is generated (step S205). In other embodiments, step S204 may also be to determine whether the average value of the interval between multiple data in the error event subset is less than or equal to the time threshold. When the baseboard management controller 103 determines that the average value of the time intervals of multiple data in the error event subset is less than or equal to the time threshold, an error report closing instruction is generated (step S205). The aforementioned multiple data time can refer to all the data in the error event subset, and can also refer to data that meets specific conditions. For example, the last N (N is a non-zero positive integer value) data in the error event subset, or the data in the error event subset within the most recent time range (for example, the data within the last 5 seconds). It should be understood that in different embodiments, some of the aforementioned steps can be omitted or replaced.For example, step S203 or step S204 is omitted, or the order of step S203 and step S204 is swapped.

[0031] In some embodiments, when the baseboard management controller 103 determines that the number of the error event subset within the default period is less than the counting threshold (step S203, the judgment result is "No"), it returns to step S201 and waits for receiving the next system event record. In some embodiments, when the baseboard management controller 103 determines whether the time interval between the last two data in the error event subset is greater than the time threshold (step S204, the judgment result is "No"), it returns to step S201 and waits for receiving the next system event record. Therefore, in this embodiment, the second program triggered by this system event record does not generate an error report closing instruction, so the SMI trigger permission corresponding to the same type of error event in the first program is not closed.

[0032] After the baseboard management controller 103 generates the error report closing instruction (step S205), the error report closing instruction is sent and stored in the first memory 104. For example, please refer to Figure 1 , the baseboard management controller 103 controls the platform path controller 102 to store the error report closing instruction in the first memory 104; or, the baseboard management controller 103 controls the platform path controller 102 to send the error report closing instruction to the processor 101, and the processor 101 stores the error report closing instruction in the first memory 104. In some embodiments, the baseboard management controller 103 directly sends the error report closing instruction to the basic input and output system running on the processor 101, so that the basic input and output system directly sets the SMI trigger permission of the corresponding type of error event according to the error report closing instruction.

[0033] In some embodiments, the baseboard management controller 103 is coupled to an external device (not shown), for example, a remote external device is connected through its network interface, or an external device is connected through a transmission line. The external device can provide the user with the ability to control the baseboard management controller 103 through the network to set the parameters of the default algorithm. For example, the calculation process of step S202 to step S204 is adjusted, or parameters such as a default period, a counting threshold, or a time threshold are set. The baseboard management controller 103 can receive the parameters set by the external device and store them in the second memory 105. In some embodiments, the baseboard management controller 103 receives the program code or parameters set by the external device while running the second program to adjust the judgment conditions for generating the error report shutdown command in real time.

[0034] The benefit of adjusting the judgment condition of the error report shutdown command through the baseboard management controller 103 is that the parameters of the default algorithm, including the default period, count threshold and time threshold, can be modified in real time in response to the adjustment of the system status or management conditions without modifying the program code of the basic input and output system. For example, the baseboard management controller 103 runs an application to present the system event record to the user. The user or the analysis program can evaluate the system status of the server 10 based on the system event record and change the management plan of the server 10, connect to the management interface through a remote external device and the network interface of the baseboard management controller 103, or enter the management interface of the baseboard management controller 103 through a transmission line, and then adjust the parameters of the default algorithm in real time, thereby affecting the SMI triggering authority of various types of error events.

[0035] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any person skilled in the art can easily think of changes or substitutions within the technical scope disclosed by the present invention, which should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A peripheral component interconnect interface error report management method, applicable to a server, the server comprising a processor, a first memory coupled to the processor, a baseboard management controller, a second memory coupled to the baseboard management controller, the processor is suitable for running a basic input and output system, characterized in that: This management approach includes: The processor runs a first program when executing the basic input and output system program, including: Receiving an error event and triggering a system management interrupt; Reading an error report closing instruction from the first memory; Determining whether the error report closing instruction corresponds to the error event; When it is determined that the error report closing instruction corresponds to the error event, closing the permission of the error event to trigger the system management interrupt; as well as Sending a system event record to the baseboard management controller, the system event record comprising an error event set, the error event set comprising the error event; as well as The baseboard management controller is suitable for running a second program, including: Receive the system event log; Processing the error event set according to a default algorithm to determine whether to generate the error report closing instruction; and When the error report closing instruction is generated, the error report closing instruction is sent and stored in the first memory.

2. The management method according to claim 1, characterized in that: The default algorithm contains: Retrieving a sub-set of error events corresponding to the same error type in the error event set; Determining whether the number of the error event subset within a default period is greater than or equal to a counting threshold; as well as When the number of the error event subset within the default period is greater than or equal to the counting threshold, the error report closing instruction is generated.

3. The management method according to claim 2, characterized in that: The method further includes waiting to receive a next system event record when the number of the error event subset within the default period is less than the counting threshold.

4. The management method according to claim 2, characterized in that: The baseboard management controller is further used to set the counting threshold. The management method further includes setting and storing the counting threshold or the default period in the second memory after the baseboard management controller receives a setting instruction for setting the counting threshold or the default period from an external device.

5. The management method according to claim 2, characterized in that: The error event subset includes a plurality of error event subsets, and the error events included in each error event subset correspond to different error types. The default algorithm further includes sequentially capturing each error event subset to determine whether the number of each error event subset within the default period is greater than or equal to the counting threshold.

6. The management method according to claim 1, characterized in that: The default algorithm contains: Retrieving a sub-set of error events corresponding to the same error type in the error event set; Determine whether the time interval between the last two data in the error event subset is less than or equal to a time threshold; as well as When the time interval between the last two data in the error event subset is less than or equal to the time threshold, the error report closing instruction is generated.

7. The management method according to claim 6, characterized in that: The method further includes waiting to receive a next system event record when the time interval between the last two data records in the error event subset is greater than the time threshold.

8. The management method according to claim 6, characterized in that: The baseboard management controller is further used to set the time threshold. The management method further includes setting and storing the time threshold in the second memory after the baseboard management controller receives a setting instruction for setting the time threshold from an external device.

9. The management method according to claim 1, characterized in that: The default algorithm contains: Retrieving a sub-set of error events corresponding to the same error type in the error event set; Determine whether an average value of time intervals between multiple data in the error event subset is less than or equal to a time threshold; as well as When the average value of the time intervals between multiple data in the error event subset is less than or equal to the time threshold, the error report closing instruction is generated.

10. The management method according to claim 1, characterized in that: The method further includes sending the system event record to the baseboard management controller and terminating the system management interruption when it is determined that the error report closing instruction does not correspond to the error event.

Images