A single event upset fault injection device and an impact detection method
Through the single-particle flip fault injection device and detection method in the aviation field, the fault injection device of DSP and CPLD architectures is used to simulate and verify the impact of single-particle flip on the system, solving the problems of logical errors and system failure, and improving the software Reliability and security.
Patent Information
- Application Number
- CN202211644842.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2022-12-19
- Publication Date
- 2024-11-26
- Estimated Expiration
- 2042-12-19
AI Technical Summary
Large-scale integrated circuits in the aviation field are sensitive to single-particle flips, resulting in logic errors and system failures. The lack of effective simulation and verification methods makes it difficult to verify the impact of single-particle flips on the system at the software level.
Design a single-particle flip fault injection device, adopts DSP and CPLD architecture, communicates with the upper computer through the USB interface, supports manual or automatic fault injection, and uses ARINC429, RS422 and discrete interface to inject simulated faults into the embedded software platform, and passes detection Software verification failure impact.
It can simulate a single-particle flip scenario at the software level, verify its impact on the system, improve the reliability and security of the software, discover potential hidden dangers, and not destroy the original software structure.
Smart Images

Figure CN115981279B_ABST
Abstract
Description
Technical Field
[0001] The invention belongs to the field of aviation electromechanical control and relates to a single particle upset fault injection device and an impact detection method. Background Art
[0002] The aviation industry uses more and more large-scale integrated circuits. These circuits are sensitive to radiation. Once a single-particle upset occurs in a critical system, it may cause a catastrophic accident. A single-particle upset is a phenomenon in which a single high-energy particle in the universe enters the sensitive area of a semiconductor device, causing the device's logic state to flip. There are a large number of high-energy particles in the space environment where aircraft and spacecraft are located. Computer-type CMOS electronic components are irradiated by particles from the earth's radiation belt, solar rays, cosmic rays, etc., causing the potential state to jump, such as logic "0" jumping to logic "1", logic "1" jumping to logic "0". The occurrence of logical errors is likely to cause software failure, and then lead to system failure. Therefore, on SRAM, DRAM, D flip-flops, latches, registers or integrated circuits that use the above functions, it is necessary to consider the impact of single-particle upsets (SEU). Usually, the strategy to eliminate the impact of single-particle upsets on the system needs to be considered from the system and hardware levels, but it is difficult to simulate the scenario of single-particle upsets at the system and hardware levels. At the software level, it is usually transmitted to the software in the form of security requirements, but there is no way to verify the impact of single-particle upsets. Therefore, it is very necessary to provide a detection method that can simulate the occurrence of single-particle upsets and verify the impact of single-particle upsets on the system. It can verify the impact of single-particle upsets on software and systems, help software discover hidden dangers in advance, and improve software reliability and safety, which has great application value in the aviation field. Summary of the invention
[0003] Purpose of the present invention:
[0004] The present invention provides a single particle upset fault injection device and an impact detection method. The single particle upset fault injection device adopts DSP and CPLD architecture, adopts USB interface to communicate with host computer, and can inject simulated single particle upset fault into embedded software platform through RS422 interface, ARINC429 interface and discrete interface through integration of host computer instruction and discrete signal of injection device, and can support manual or automatic fault injection mode, and receive and record feedback information after fault injection. The detection method runs on embedded software platform, and can simulate the scenario of single particle upset from software level through tested software and single particle upset fault injection device, and will not destroy the structure of tested software. After fault injection is completed, it can return to the original running state of tested software fault injection. A single particle upset fault injection device and an impact detection method verify the impact of single particle upset effect on software and system before the product is put into actual use, discover software hidden dangers, and improve software reliability and security.
[0005] The technical solution of the present invention:
[0006] A single particle upset fault injection device comprises a USB interface, an RS422 interface, an ARINC429 interface and a discrete interface; the device is implemented by using a DSP and CPLD architecture, and the device performs data interaction with a host computer through USB communication; the start / stop control, communication mode and manual / automatic mode of an embedded software platform are selected through a high-open discrete input group; the ARINC429 interface / RS422 interface / discrete interface is used to send single particle upset fault execution information to the embedded software platform and receive execution completion information fed back by the embedded software platform; and the discrete interface execution information instruction is completed through a low-open discrete input group.
[0007] Furthermore, the interface connected to the embedded software platform at least includes an ARINC429 interface, an RS422 interface and a discrete interface.
[0008] The USB interface includes 1 channel.
[0009] The low-open discrete input group includes 16 channels, and the high-open discrete input group includes 8 channels.
[0010] The high-open discrete input group at least includes mode selection, communication selection and startup control functions.
[0011] The RS422 interface fault injection and ARINC429 interface fault injection are sent from the host computer, and the discrete interface fault injection is sent from the low-open discrete input group.
[0012] A method for detecting the impact of a single-event upset effect is performed using a single-event upset fault injection device, comprising the following steps:
[0013] Step 1: Embed the single-particle upset detection software into the embedded software platform, assign the single-particle upset detection software to a FLASH sector different from the software under test in the embedded software platform, set the startup function address of the single-particle upset detection software, and divide the RAM address dedicated to the detection software;
[0014] Step 2: Add a single event upset fault injection function to the software under test;
[0015] Step 3: In the software under test, the single-particle upset fault injection function is called periodically, and the bit information to be flipped, mode information, software under test running information and stack information are passed to the single-particle upset detection software;
[0016] Step 4: The detection software flips the corresponding bit according to the instructions of the single-event upset fault injection device ARINC429 interface, RS422 interface and discrete interface and the information transmitted by the software under test;
[0017] Step 5: The detection software feeds back the fault injection execution information to the single event upset fault injection device through the ARINC429 interface, the RS422 interface and the discrete interface;
[0018] Step 6: The single event upset fault injection device feeds back the fault injection execution information to the host computer through the USB interface;
[0019] Step 7: The detection software returns the original running address of the software under test by obtaining the stack information of the software under test;
[0020] Step 8: If the software under test runs normally, return to step 3. If the software under test fails, the host computer records the fault information.
[0021] further,
[0022] The FLASH sectors and RAM addresses in step 1 can be configured;
[0023] The single event upset fault injection function in step 2 is provided by the single event upset impact detection software;
[0024] In step 3, the single-particle upset fault injection function is called periodically, which means that the jump time is at least configurable; wherein the single-particle upset fault injection function has the function of jumping from the detection software to the software under test;
[0025] In step 4, the instructions of the single event upset fault injection device refer to at least discrete instructions or communication instructions, wherein ARINC429 interface and RS422 interface instructions are communication instructions;
[0026] In step 5, the fault injection execution information is fed back, and at least the feedback information is transmitted via discrete or communication instructions;
[0027] In step 7, the original running address of the tested software is returned, and after the detection software is running, the original running state of the tested software can be restored;
[0028] The recording of fault information in step 8 refers to the transmission of the fault information from the single event upset fault injection device to the host computer for recording.
[0029] The beneficial technical effects of the present invention are as follows:
[0030] The present invention provides a single particle upset fault injection device and an impact detection method, which can simulate the injection of single particle upset faults into an embedded software platform, and supports multiple communication modes and fault injection modes. It has strong versatility, can verify the impact of single particle upsets on software and systems, and help software discover problems caused by single particle upsets in advance without destroying the structure of the original software. It can also automatically perform tests and fault records, thereby improving the reliability and safety of the software, and has great application value in the aviation field. BRIEF DESCRIPTION OF THE DRAWINGS
[0031] Figure 1 A schematic diagram of a single event upset fault injection device and a method for affecting detection thereof;
[0032] Figure 2 It is a schematic diagram of a detection method of a single event upset fault injection device and a method for affecting detection;
[0033] Figure 3 The invention discloses a flow chart of a detection method of a single-particle upset fault injection device and a method for affecting detection. Specific implementation methods
[0034] The present invention will be further described in detail below in conjunction with the accompanying drawings.
[0035] like Figure 1 As shown, a single particle upset fault injection device includes a USB interface, an RS422 interface, an ARINC429 interface, and a discrete interface; the device is implemented using a DSP and CPLD architecture, and the device interacts with a host computer through USB communication; the start / stop control, communication mode, and manual / automatic mode of the embedded software platform are selected through a high-open discrete input group; the ARINC429 interface / RS422 interface / discrete interface is used to send single particle upset fault execution information to the embedded software platform and receive execution completion information fed back by the embedded software platform; the discrete interface execution information instruction is completed through a low-open discrete input group.
[0036] Among them, the interface connected to the embedded software platform at least includes an ARINC429 interface, an RS422 interface and a discrete interface; the USB interface includes 1 channel; the low-open discrete input group includes 16 channels, and the high-open discrete input group includes 8 channels; the high-open discrete input group at least includes mode selection, communication selection and startup control functions; RS422 interface fault injection and ARINC429 interface fault injection are issued by the upper computer, and discrete interface fault injection is issued by the low-open discrete input group.
[0037] A method for detecting the impact of a single event upset effect, the method comprising the following steps:
[0038] Step 1: Select the communication method with the embedded software platform, including ARINC429 interface, RS422 interface and discrete interface;
[0039] Step 2: Select the communication method with the host computer, including the USB interface;
[0040] Step 3: Configure high open discrete input signals, including 8 channels;
[0041] Step 4: When the discrete interface is selected, configure the low-open discrete input signal, including 16 channels;
[0042] Step 5: RS422 interface fault injection and ARINC429 interface fault injection are sent by the host computer, and discrete interface fault injection is sent by the low-open discrete input group;
[0043] Step 6: In the address allocation file, set the fixed address of the single-particle upset detection software and the fixed address of the software under test to different FLASH sectors, set the startup function address of the single-particle upset detection software, and set the RAM address dedicated to the detection software;
[0044] Step 7: Find a suitable location in the software under test and insert the single event upset fault injection interface function provided by the detection software;
[0045] Step 8: In the software under test, the single-particle upset fault injection interface function is called regularly, and the bit information to be flipped, the running information of the software under test, and the stack information are passed to the RAM address dedicated to the detection software through the function parameters.
[0046] Step 9: The detection software reads the discrete instructions or communication instructions of the single-particle upset fault injection device, reads the upset information and manual / automatic information transmitted by the software under test, and after synthesizing the instructions, performs the flipping of the corresponding bit;
[0047] Step 10: After the single event upset is completed, the detection software feeds back the fault injection execution information to the single event upset fault injection device through the ARINC429 interface, RS422 interface and discrete interface;
[0048] Step 11: The single event upset fault injection device feeds back fault injection execution information to the host computer through the USB interface;
[0049] Step 12: Return to the original running state of the tested software after fault injection by obtaining the running information and stack information of the tested software, and continue running the tested software;
[0050] Step 13: If the software under test runs normally, return to step 8. If the software under test fails, the host computer records the fault information.
[0051] Example 1
[0052] This section is an embodiment of the present invention, which is used to explain and illustrate the technical solution of the present invention.
[0053] The method for detecting the impact of single-particle upset effect is to perform the following steps on the embedded software platform DSP28335 and the host computer LabVIEW:
[0054] Step 1: Connect the fault injection device and the embedded software platform, select the communication mode, and select discrete instructions through discrete signals;
[0055] Step 2: Connect the fault injection device and the host computer via USB;
[0056] Step 3: In the address allocation file FLASH.CMD, set the fixed address of the single-particle upset detection software to FALSHA, the fixed address of the software under test to FLASHB, and set the startup function address of the single-particle upset detection software to 0x33FFF6, and set the RAM address dedicated to the detection software;
[0057] Step 4: In the 10ms periodic task in the software under test, call the single event upset fault injection interface function provided by the detection software;
[0058] Step 5: Configure the high open discrete input signal and select the mode as automatic;
[0059] Step 6: Configure the low-open discrete input signal and select the single event upset start address information;
[0060] Step 7: Configure the low-open discrete input signal, turn on the start-stop control discrete interface, and perform single event upset injection;
[0061] Step 8: After the 10ms period is up, the single-particle upset fault injection interface function is called, and the bit address information, automatic test mode, operation information and stack information to be flipped are written through the function parameters. The interface function writes the above information into a dedicated RAM area.
[0062] Step 9: The detection software read mode is the automatic mode, only the discrete instruction start address information of the single-event upset fault injection device is read, and the next bit to be flipped in the RAM area of the software under test is automatically flipped;
[0063] Step 10: After the single event upset is completed, the detection software upsets the discrete output interface of the execution completion information and transmits the execution completion information to the fault injection device;
[0064] Step 11: by obtaining the running information and stack information of the tested software, the running address of the tested software before the fault injection is returned through the jump function, and the tested software continues to execute;
[0065] Step 12: Monitor the software under test. If it runs normally, return to step 3 in 10ms. When the limit time is exceeded and the fault injection device still cannot receive the execution completion information through the discrete interface, the software under test is recorded as failed.
[0066] Step 13: The single event upset fault injection device feeds back fault injection execution information to the host computer through the USB interface;
[0067] Step 14: The host computer records the feedback information of the embedded software platform and records whether the software runs normally or fails.
[0068] Example 2
[0069] This section is an embodiment of the present invention, which is used to explain and illustrate the technical solution of the present invention.
[0070] The method for detecting the impact of single-particle upset effect is to perform the following steps on the embedded software platform DSP28335, the host computer is LabVIEW, and the communication mode is 422 communication:
[0071] Step 1: Connect the fault injection device and the embedded software platform, select the communication mode, and select the 422 interface through the discrete signal;
[0072] Step 2: Connect the fault injection device and the host computer via USB;
[0073] Step 3: In the address allocation file FLASH.CMD, set the fixed address of the single-particle upset detection software to FALSHA, the fixed address of the software under test to FLASHB, and set the startup function address of the single-particle upset detection software to 0x33FFF6, and set the RAM address dedicated to the detection software;
[0074] Step 4: In the 10ms periodic task in the software under test, call the single event upset fault injection interface function provided by the detection software;
[0075] Step 5: Configure the high open discrete input signal and select the mode as manual;
[0076] Step 6: Send the single event upset address information to the fault injection device regularly through the host computer;
[0077] Step 7: Configure the low-open discrete input signal, turn on the start-stop control discrete interface, and perform single event upset injection;
[0078] Step 8: After the 100ms period is up, the single-particle upset fault injection interface function is called, and the bit address information, automatic test mode, operation information and stack information to be flipped are written through the function parameters. The interface function writes the above information into a dedicated RAM area.
[0079] Step 9: The detection software read mode is in manual mode, the communication instruction of the single-particle upset fault injection device is read, and the bit that needs to be flipped in the RAM area of the software under test is flipped according to the 422 communication instruction;
[0080] Step 10: After the single event upset is completed, the detection software transmits the execution completion information to the fault injection device through the 422 communication instruction;
[0081] Step 11: by obtaining the running information and stack information of the tested software, the running address of the tested software before the fault injection is returned through the jump function, and the tested software continues to execute;
[0082] Step 12: Monitor the software under test. If it runs normally, return to step 3 after 100ms. When the limit time is exceeded and the fault injection device still cannot receive the execution completion information, the software under test is recorded as failed.
[0083] Step 13: The single event upset fault injection device feeds back fault injection execution information to the host computer through the USB interface;
[0084] Step 14: The host computer records the feedback information of the embedded software platform and records whether the software runs normally or fails.
Claims
1. A single event upset fault injection device, Features: It includes a USB interface, an RS422 interface, an ARINC429 interface, and a discrete interface; the device is implemented using a DSP and CPLD architecture, and the device interacts with a host computer through USB communication; the start / stop control, communication mode, and manual / automatic mode of the embedded software platform are selected through a high-open discrete input group; the ARINC429 interface / RS422 interface / discrete interface is used to send single-particle upset fault execution information to the embedded software platform and receive execution completion information fed back by the embedded software platform; Complete the discrete interface execution information instruction through the low open discrete input group; The method for performing single-event upset effect impact detection using the single-event upset fault injection device comprises the following steps: Step 1: Embed the single-particle upset detection software into the embedded software platform, assign the single-particle upset detection software to a FLASH sector different from the software under test in the embedded software platform, set the startup function address of the single-particle upset detection software, and divide the RAM address dedicated to the detection software; Step 2: Add a single event upset fault injection function to the software under test; Step 3: In the software under test, the single-particle upset fault injection function is called periodically, and the bit information to be flipped, mode information, software under test running information and stack information are passed to the single-particle upset detection software; Step 4: The detection software flips the corresponding bit according to the instructions of the single-event upset fault injection device ARINC429 interface, RS422 interface and discrete interface and the information transmitted by the software under test; Step 5: The detection software feeds back the fault injection execution information to the single event upset fault injection device through the ARINC429 interface, the RS422 interface and the discrete interface; Step 6: The single event upset fault injection device feeds back the fault injection execution information to the host computer through the USB interface; Step 7: The detection software returns the original running address of the software under test by obtaining the stack information of the software under test; Step 8: If the software under test runs normally, return to step 3. If the software under test fails, the host computer records the fault information.
2. A single event upset fault injection device according to claim 1, Features: The interface connected with the embedded software platform at least includes an ARINC429 interface, an RS422 interface and a discrete interface.
3. A single event upset fault injection device according to claim 1, Features: The USB interface includes 1 channel.
4. A single event upset fault injection device according to claim 1, Features: The low-open discrete input group includes 16 channels, and the high-open discrete input group includes 8 channels.
5. The single event upset fault injection device according to claim 1, Features: The high-open discrete input group at least includes mode selection, communication selection and startup control functions.
6. A single event upset fault injection device according to claim 1, Features: The RS422 interface fault injection and ARINC429 interface fault injection are sent from the host computer, and the discrete interface fault injection is sent from the low-open discrete input group.
7. A single event upset fault injection device according to claim 1, Features: The FLASH sectors and RAM addresses in step 1 can be configured; The single event upset fault injection function in step 2 is provided by the single event upset impact detection software; In step 3, the single-particle upset fault injection function is called periodically, which means that the jump time is at least configurable; wherein the single-particle upset fault injection function has the function of jumping from the detection software to the software under test; In step 4, the instructions of the single event upset fault injection device refer to at least discrete instructions or communication instructions, wherein ARINC429 interface and RS422 interface instructions are communication instructions; In step 5, the fault injection execution information is fed back, and at least the feedback information is transmitted via discrete or communication instructions; In step 7, the original running address of the tested software is returned, and after the detection software is running, the original running state of the tested software can be restored; The recording of fault information in step 8 refers to the transmission of the fault information from the single event upset fault injection device to the host computer for recording.
