A method and apparatus for locating memory CA signal errors
By performing parity checks on memory through bit-by-bit flipping test packets, the problem of quickly locating memory CA signal errors was solved, enabling automatic identification of faulty CA pins and improving server maintenance efficiency.
Patent Information
- Application Number
- CN202211633516.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-12-19
- Publication Date
- 2026-11-10
- Estimated Expiration
- 2042-12-19
AI Technical Summary
Existing technologies struggle to quickly locate errors in memory CA signals, especially in automatically identifying specific CA pin faults, leading to low repair efficiency.
Multiple bit-by-bit flip test packets are used to perform parity checks on the memory. The faulty CA pin is located by using the parity check results. The fault is then precisely located by sending a special test packet from the CPU in BIOS SMM mode.
This system enables automatic identification of specific CA signal fault pins during system operation, improving fault location efficiency, reducing server maintenance time, and enhancing user service quality.
Smart Images

Figure CN115981912B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of server technology, and more specifically to a method and apparatus for locating memory CA signal errors. Background Technology
[0002] With the widespread application of cloud computing and big data, higher requirements have been placed on the reliability, stability and maintainability of server systems. Among these, maintainability is crucial for extending the lifespan of servers.
[0003] Memory is the component that most frequently causes server crashes. When memory problems occur, quickly locating the fault can shorten server downtime for maintenance, reduce user losses, and improve the quality of user service.
[0004] The memory module and CPU are connected via the motherboard. The signals from the memory module to the CPU include data signals (DQ / DQS), control signals (Command), and address signals (Address). These signals can all fail during actual use. When the CA (Command / Address) signal malfunctions, such as a CA parity error, the system retransmits the data via retry. However, if a persistent error occurs in the link, retransmission cannot resolve the problem, and the system cannot pinpoint which CA pin is faulty. Typically, a restart and accessing the BIOS serial port logs are required for further analysis of the CA problem, which is inconvenient for troubleshooting and locating memory issues, resulting in low repair efficiency. Summary of the Invention
[0005] To address the deficiencies in existing technologies, the first aspect of this invention provides a method for locating errors in memory CA signals. When a CA signal fault occurs during system operation, the method can automatically locate which specific CA signal is faulty.
[0006] To achieve the above objectives, the technical solution adopted by the present invention is as follows:
[0007] A method for locating memory CA signal errors, the method comprising the following steps:
[0008] The memory is parity checked using multiple bit-by-bit flip test packets. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of CA pins and parity pins.
[0009] In some embodiments, the step of performing parity checks on the memory using multiple bit-by-bit flip test packets, and locating the faulty CA pin based on the parity check results, includes:
[0010] Send the initial test packet to memory, and determine whether a parity error has occurred based on the data returned by memory.
[0011] When a parity error occurs when sending the initial test packet, a test packet based on a single bit flip of the initial test packet is sent to memory for polling and detection to determine whether a parity error has occurred.
[0012] The faulty CA pin was located based on the test package that did not show a parity error during polling.
[0013] In some embodiments, the initial test packet is sent to memory, and based on the data fed back from the memory, it is determined whether a parity error has occurred.
[0014] The first initial test packet and the second initial test packet are sent to memory in sequence, wherein the second initial test packet is the inverted bit of the first initial test packet;
[0015] When the first initial test packet and the second initial test packet are sent, if a parityerror occurs under either test packet, it is determined that a parity error has occurred.
[0016] Since no parity error occurred after sending the first and second initial test packets, it was determined that no parity error occurred.
[0017] In some embodiments, it also includes:
[0018] If no parity error occurs after sending the first and second initial test packets, the third initial test packet is sent to memory. The third initial test packet is the inverted n bits of the first or second initial test packet, where n is an odd number.
[0019] A parity error occurred after sending the third initial test packet, indicating a faulty parity pin.
[0020] If no parity error occurs after sending the third initial test packet, it is determined that no parity error has occurred.
[0021] In some embodiments, when all test packets in the polling detection show a parity error, the problem is identified as a paritypin failure.
[0022] A second aspect of this invention provides a device for locating errors in memory CA signals. When a CA signal fault occurs during system operation, the device can automatically pinpoint which specific CA signal is faulty.
[0023] To achieve the above objectives, the technical solution adopted by the present invention is as follows:
[0024] An apparatus for locating memory CA signal errors, comprising:
[0025] The positioning module is used to drive the CPU to send multiple bit-by-bit flip test packets to perform parity checks on the memory. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of CA pins and the number of parity pins.
[0026] In some embodiments, the positioning module drives the CPU to send multiple bit-by-bit flip test packets to perform parity checks on the memory, and locates the faulty CA pin based on the parity check results, including:
[0027] The CPU is driven to send an initial test packet to the memory, and based on the data fed back from the memory, it is determined whether a parityerror has occurred.
[0028] When a parity error occurs when sending the initial test packet, the CPU is driven to send a test packet based on a single bit flip of the initial test packet to the memory for polling and detection to determine whether a parity error has occurred.
[0029] The faulty CA pin was located based on the test package that did not show a parity error during polling.
[0030] In some embodiments, the positioning module drives the CPU to send an initial test packet to memory, and determines whether a parity error has occurred based on the data fed back from memory, including:
[0031] The CPU is driven to send the first initial test packet and the second initial test packet to the memory in sequence, wherein the second initial test packet is the inverted bit of the first initial test packet;
[0032] When the first initial test packet and the second initial test packet are sent, if a parityerror occurs under either test packet, it is determined that a parity error has occurred.
[0033] Since no parity error occurred after sending the first and second initial test packets, it was determined that no parity error occurred.
[0034] In some embodiments, the positioning module is further used for:
[0035] If no parity error occurs after sending the first and second initial test packets, the CPU is driven to continue sending a third initial test packet to memory. The third initial test packet is the inverted n bits of the first or second initial test packet, where n is an odd number.
[0036] A parity error occurred after sending the third initial test packet, indicating a faulty parity pin.
[0037] If no parity error occurs after sending the third initial test packet, it is determined that no parity error has occurred.
[0038] In some embodiments, the positioning module is further used for:
[0039] When all test packages in the polling detection show a parity error, the problem is identified as a faulty parity pin.
[0040] Compared with the prior art, the advantages of the present invention are as follows:
[0041] The method for locating memory CA signal errors in this invention utilizes multiple bit-by-bit toggling test packets to perform parity checks on the memory. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of CA pins and the number of parity pins. When a memory CA signal error occurs, the system generates an SMI interrupt and enters BIOS SMM mode. The CPU sends special test packets to the memory. Because the data received by the memory is toggled when a pin error occurs, the specific bit toggled can be located by using the bit-by-bit toggling test packets, thus pinpointing the specific pin that is experiencing the error. Attached Figure Description
[0042] Figure 1 This is a flowchart of a method for locating memory CA signal errors in an embodiment of the present invention;
[0043] Figure 2 yes Figure 1 Flowchart of step S1;
[0044] Figure 3 This is an embodiment of the present invention. Detailed Implementation
[0045] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.
[0046] This invention discloses a method for locating memory CA signal errors, characterized in that the method includes the following steps:
[0047] The memory is parity checked using multiple bit-by-bit flip test packets. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of CA pins and parity pins.
[0048] It is worth noting that when a CA signal error occurs in the memory, the system generates an SMI interrupt and enters the BIOS SMM mode. The CPU sends a special test packet to the memory. Because when a pin error occurs, the data received by the memory is flipped, the test packet that flips a single bit can pinpoint which bit was flipped, thus pinpointing which pin has an error.
[0049] See Figure 1 As shown, in a specific implementation, the above steps include:
[0050] S1. Send the initial test packet to memory, and determine whether a parity error has occurred based on the data returned by memory.
[0051] It is worth noting that there are two main situations where a pin error occurs: one is floating and the other is grounded. When a pin is floating or grounded, the transmitted value remains unchanged, with 1 for a floating pin and 0 for a grounded pin. To address these two situations, this embodiment of the invention uses two initial test packages to determine whether a parity error occurs.
[0052] See Figure 2 As shown, step S1 specifically includes:
[0053] S11. Send the first initial test packet and the second initial test packet to memory in sequence, wherein the second initial test packet is the inverted bit of the first initial test packet.
[0054] It is understood that the bit length of each test packet mentioned above corresponds to the number of pins in the CApin and paritypin. The Command signal is 7 bits, the Address signal is also 7 bits, and the parity pin is 1 bit. Therefore, the test packet in this embodiment of the invention is 15 bits.
[0055] In specific implementation, the first initial test packet in this embodiment is "000000000000000", and the second initial test packet is "111111111111110". The last bit is 0, which is a parity check bit; 0 indicates an even check. It can be understood that only the sum of the first 14 bits of the first and second initial test packets needs to be even; that is, the test packets are not limited to the example given above. Of course, odd checks can also be performed by adjusting the test packets, and this invention does not impose any restrictions on this.
[0056] S12. After sending the first initial test packet and the second initial test packet, if a parity error occurs in either test packet, it is determined that a parity error has occurred.
[0057] S13. If no parity error occurs after sending the first initial test packet and the second initial test packet, it is determined that no parity error has occurred.
[0058] Understandably, the first initial test packet is mainly to detect whether there is a floating situation. If there is a floating situation, the value transmitted by the pin will still be 1, not 0 as in the test packet, which means that the pin cannot transmit signals normally.
[0059] If a parity error occurs after sending the first initial test packet, it can be considered that an error has occurred on a certain pin. At this time, there is no need to send a second initial test packet, and proceed directly to the subsequent step S2.
[0060] If no parity error occurs after sending the first initial test packet, a second initial test packet needs to be sent. It is understood that the second initial test packet is mainly to detect whether a grounding situation has occurred. If a grounding situation occurs, the value transmitted by the pin will still be 0, not 1 as in the test packet, which means that the pin cannot transmit signals normally.
[0061] If no parity error occurs after sending the first and second initial test packets, it is determined that no parity error has occurred, and the process can proceed to step S2.
[0062] Furthermore, in some preferred embodiments, the method further includes the following after step S13:
[0063] If no parity error occurs after sending the first and second initial test packets, the third initial test packet is sent to memory. The third initial test packet is the inverted n bits of the first or second initial test packet, where n is an odd number.
[0064] If a parity error occurs after sending the third initial test packet, it is determined that the parity pin is faulty; if no parity error occurs after sending the third initial test packet, it is determined that no parity error has occurred.
[0065] It is understandable that after step S13, it can be assumed that the pins corresponding to the control signal Command and the address signal Address are both functioning correctly under the current parity pin. However, the parity pin itself may still be faulty, hence the above steps are needed to test it. The third initial test package set above can determine whether the parity pin is faulty, allowing for processing before step S2. Alternatively, step S2 can be used to determine whether the parity pin is faulty.
[0066] S2. When a parity error occurs when sending the initial test packet, send a test packet based on a single bit flip of the initial test packet to memory for polling and detection to determine whether a parity error has occurred.
[0067] S3. Locate the faulty CApin based on the test packages that do not show a parity error during polling.
[0068] Understandably, step S1 mainly determines whether a parity error exists, but it cannot pinpoint which pin is causing the problem.
[0069] For this reason, see Figure 3 As shown, it illustrates the specific behavior of a test packet based on single-bit flipping of the initial test packet. From Figure 3 As can be seen, it uses a bit-by-bit flipping method to generate a series of test packages.
[0070] Taking the first test packet "0111111111111111" as an example, compared with the second initial test packet "111111111111110", the first bit is flipped. It can be understood that after the flip, the sum is odd, so the final parity check bit is "1". The purpose of the first test packet is to detect whether the first bit is faulty. If the first bit is faulty, the problematic pin is located. It should be noted that because the first and second initial test packets can be reasonably configured, the specific form of subsequent test packets based on single-bit flips of the initial test packets will also change accordingly. This invention does not impose any restrictions on this.
[0071] Specifically, the first test packet "0111111111111111" is sent to determine if a parity error occurs. If a parity error occurs (because step S1 has already determined that a parity error exists, subsequent checks should normally show a parity error), it indicates that the first CApin can be controlled normally and that the pin is not faulty. If no parity error occurs, it means that the value transmitted through this pin is still 1, not 0 as in the test packet, indicating that the pin cannot transmit signals normally, thus locating the problem with this pin. Similarly, polling continues up to 111111111111101. If there is a problem, it indicates that the parity pin is faulty. This method can thus locate the faulty pin.
[0072] In summary, the method for locating memory CA signal errors in this invention utilizes multiple bit-by-bit toggling test packets to perform parity checks on the memory. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of the CA pin and the parity check pin. When a memory CA signal error occurs, the system generates an SMI interrupt and enters BIOS SMM mode. The CPU sends a special test packet to the memory. Because the data received by the memory is toggled when a pin error occurs, the specific bit toggled can be located by using the bit-by-bit toggling test packet, thus pinpointing the specific pin that is experiencing the error.
[0073] This invention also discloses a device for locating errors in memory CA signals, which includes a location module.
[0074] The positioning module is used to drive the CPU to send multiple bit-by-bit flip test packets to perform parity checks on the memory. Based on the parity check results, the faulty CApin is located. The number of bits in each test packet corresponds to the number of pins of the CApin and the parity check pin.
[0075] In some embodiments, the positioning module drives the CPU to send multiple bit-by-bit flip test packets to perform parity checks on the memory, and locates the faulty CA pin based on the parity check results, including:
[0076] The CPU is driven to send an initial test packet to the memory, and based on the data fed back from the memory, it is determined whether a parityerror has occurred.
[0077] When a parity error occurs when sending the initial test packet, the CPU is driven to send a test packet based on a single bit flip of the initial test packet to the memory for polling and detection to determine whether a parity error has occurred.
[0078] The faulty CA pin was located based on the test package that did not show a parity error during polling.
[0079] In some embodiments, the positioning module drives the CPU to send an initial test packet to memory, and determines whether a parity error has occurred based on the data fed back from memory, including:
[0080] The CPU is driven to send the first initial test packet and the second initial test packet to the memory in sequence, wherein the second initial test packet is the inverted bit of the first initial test packet;
[0081] When the first initial test packet and the second initial test packet are sent, if a parityerror occurs under either test packet, it is determined that a parity error has occurred.
[0082] Since no parity error occurred after sending the first and second initial test packets, it was determined that no parity error occurred.
[0083] In some embodiments, the positioning module is further used for:
[0084] If no parity error occurs after sending the first and second initial test packets, the CPU is driven to continue sending a third initial test packet to memory. The third initial test packet is the inverted n bits of the first or second initial test packet, where n is an odd number.
[0085] A parity error occurred after sending the third initial test packet, indicating a faulty parity pin.
[0086] If no parity error occurs after sending the third initial test packet, it is determined that no parity error has occurred.
[0087] In some embodiments, the positioning module is further used for:
[0088] When all test packages in the polling detection show a parity error, the problem is identified as a faulty parity pin.
[0089] In summary, the device for locating memory CA signal errors in this invention includes a location module. This module drives the CPU to send multiple bit-by-bit toggling test packets to perform parity checks on the memory. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of the CA pin and the parity pin. When a memory CA signal error occurs, the system generates an SMI interrupt and enters BIOS SMM mode. The CPU sends special test packets to the memory. Because the data received by the memory is toggled when a pin error occurs, the specific bit toggled can be located by using the bit-by-bit toggling test packets, thus pinpointing the specific pin that is experiencing the error.
[0090] The above description is merely a specific embodiment of this application, enabling those skilled in the art to understand or implement this application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of this application. Therefore, this application is not to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features claimed herein.
Claims
1. A method for locating memory CA signal errors, characterized in that, The method includes the following steps: The memory is parity checked using multiple bit-by-bit flip test packets. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of CA pins and parity pins. The method of performing parity checks on the memory using multiple bit-by-bit flip test packets, and locating the faulty CA pin based on the parity check results, includes: Send the initial test packet to memory, and determine whether a parity error has occurred based on the data returned by memory. When a parity error occurs when sending the initial test packet, a test packet based on a single bit flip of the initial test packet is sent to memory for polling and detection to determine whether a parity error has occurred. The faulty CA pin was located based on the test package that did not show a parity error during polling detection; The process of sending the initial test packet to memory and determining whether a parity error has occurred based on the data returned from memory includes: The first initial test packet and the second initial test packet are sent to the memory in sequence. The second initial test packet is the inverted bit of the first initial test packet. The first initial test packet is used to detect whether there is a floating connection, and the second initial test packet is used to detect whether there is a ground connection. When the first initial test packet and the second initial test packet are sent, a parity error is determined to have occurred if a parity error occurs in either test packet. Since no parity error occurred after sending the first and second initial test packets, it was determined that no parity error occurred. Furthermore, when all test packages in the polling detection show a parity error, the problem is identified as a fault in the parity pin.
2. The method for locating memory CA signal errors according to claim 1, characterized in that, Also includes: If no parity error occurs after sending the first and second initial test packets, the third initial test packet is sent to memory. The third initial test packet is the inverted n bits of the first or second initial test packet, where n is an odd number. A parity error occurred after sending the third initial test packet, indicating a faulty parity pin. If no parity error occurs after sending the third initial test packet, it is determined that no parity error has occurred.
3. A device for locating errors in the memory CA signal, characterized in that, include: The positioning module is used to drive the CPU to send multiple bit-by-bit flip test packets to perform parity checks on the memory. Based on the parity check results, the faulty CA pin is located. The number of bits in each test packet corresponds to the number of pins of the CA pin and the parity check pin. The positioning module drives the CPU to send multiple bit-by-bit flip test packets to perform parity checks on the memory. Based on the parity check results, the faulty CA pin is located, including: The CPU is driven to send an initial test packet to the memory, and based on the data fed back from the memory, it is determined whether a parityerror has occurred. When a parity error occurs when sending the initial test packet, the CPU is driven to send a test packet based on a single bit flip of the initial test packet to the memory for polling and detection to determine whether a parity error has occurred. The faulty CA pin was located based on the test package that did not show a parity error during polling detection; The positioning module drives the CPU to send an initial test packet to the memory, and determines whether a parity error has occurred based on the data fed back from the memory, including: The CPU is driven to send a first initial test packet and a second initial test packet to the memory in sequence. The second initial test packet is the inverted bit of the first initial test packet. The first initial test packet is used to detect whether there is a floating connection, and the second initial test packet is used to detect whether there is a ground connection. When the first initial test packet and the second initial test packet are sent, a parity error is determined to have occurred if a parity error occurs in either test packet. Since no parity error occurred after sending the first and second initial test packets, it was determined that no parity error occurred. Furthermore, the positioning module is also used for: When all test packages in the polling detection show a parity error, the problem is identified as a faulty parity pin.
4. The device for locating memory CA signal errors according to claim 3, characterized in that, The positioning module is also used for: If no parity error occurs after sending the first and second initial test packets, the CPU is driven to continue sending a third initial test packet to memory. The third initial test packet is the inverted n bits of the first or second initial test packet, where n is an odd number. A parity error occurred after sending the third initial test packet, indicating a faulty parity pin. If no parity error occurs after sending the third initial test packet, it is determined that no parity error has occurred.
Citation Information
Patent Citations
Semiconductor system and method for testing semiconductor device
CN105938727A
