CN114139164A - Variation method for kernel fuzzy test of trusted operating system - Google Patents
Variation method for kernel fuzzy test of trusted operating system Download PDFInfo
- Publication number
- CN114139164A CN114139164A CN202111472930.3A CN202111472930A CN114139164A CN 114139164 A CN114139164 A CN 114139164A CN 202111472930 A CN202111472930 A CN 202111472930A CN 114139164 A CN114139164 A CN 114139164A
- Authority
- CN
- China
- Prior art keywords
- operating system
- seeds
- trusted operating
- variation
- test
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/50—Monitoring users, programs or devices to maintain the integrity of platforms, e.g. of processors, firmware or operating systems
- G06F21/57—Certifying or maintaining trusted computer platforms, e.g. secure boots or power-downs, version controls, system software checks, secure updates or assessing vulnerabilities
- G06F21/577—Assessing vulnerabilities and evaluating computer system security
Landscapes
- Engineering & Computer Science (AREA)
- Computer Hardware Design (AREA)
- Computer Security & Cryptography (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Debugging And Monitoring (AREA)
Abstract
The invention relates to the field of vulnerability mining of an operating system, in particular to a seed mutation method suitable for a fuzzy test of a trusted operating system, which is designed aiming at the field of vulnerability mining of the trusted operating system and aiming at solving the problem that the quality of seeds generated by mutation in the process of the fuzzy test of the trusted operating system is poor. The process of the invention is shown in figure 1, and the adopted technical means are as follows: after the seeds of the test set are subjected to randomized variation, effective seeds with correct format and semantics are screened out by utilizing an effectiveness judgment function; and (3) jointly guiding the adjustment of the mutation algorithm by using two feedback information, namely the coverage rate of the effectiveness judgment function and the coverage rate of a core of a trusted operating system, and continuously updating the test set used by mutation by using seeds with positive feedback. Based on the method of the invention, the trusted operating system can rapidly generate high-quality seeds in the variation stage in the process of carrying out the fuzzing test, thereby improving the execution efficiency of the fuzzing test and the tested range of the kernel code of the trusted operating system.
Description
Technical Field
The invention relates to the field of vulnerability discovery of operating systems, in particular to a seed mutation method suitable for a fuzzy test of a trusted operating system aiming at the field of vulnerability discovery of the trusted operating system. Aiming at the problems of low efficiency and poor quality of generated seeds when the fuzzy test of the trusted operating system adopts random variation, on one hand, varied invalid seeds are filtered out, on the other hand, the random variation is guided by two feedback information of effectiveness judgment function coverage and trusted operating system kernel code coverage, and the effective seed generation capability of a variation algorithm is improved.
Background
The trusted operating system is designed aiming at sensitive data and key code leakage problems. To this end, an effective method is to abstract a physical CPU into two virtual CPUs by a virtualization method, and run a Normal World (Normal World) and a Secure World (Secure World) respectively, where the Normal World and the Secure World have their own independent apps and operating systems respectively, and the trusted operating system is an operating system running in the Secure World. The Normal operation of the trusted operating system can ensure that sensitive data is not transmitted back to Normal World, so the security of the trusted operating system is crucial.
In order to enhance the security of the trusted operating system, security testing and vulnerability mining need to be performed on the trusted operating system, and a fuzzy testing technology is a feasible security scheme. The fuzzy testing technology is a mature vulnerability mining technology of an operating system due to high automation degree, simple using method and wide application range, and can make up the defects of high difficulty, low efficiency and the like in manual vulnerability mining. When using the fuzz testing method, the tester needs to provide a small number of executable seeds to the fuzz tester as the initial test set of the operating system. The fuzzy tester can perform small-range and multi-method variation on the seeds in the test set based on a certain strategy to generate a large number of new seeds, and the seeds in the test set and the new seeds generated by variation are sent to an operating system for execution. The fuzzy tester can monitor whether the operating system is abnormal in real time to judge whether a bug exists in the operating system. Meanwhile, the seeds transmitted into the operating system record the coverage rate information of the kernel in the executed process and feed the coverage rate information back to the fuzz tester. The tester screens the mutated seeds according to the feedback information, adds the seeds contributing to the coverage of the kernel code into the test set for further mutation, and discards the non-contributing seeds.
The great core of the fuzzy test technology is to design a high-quality seed variation algorithm, and the variation method directly influences the number of generated effective seeds and the diversity of the seeds. The variation algorithm adopted by the fuzzy test is a random algorithm, particularly, a special fuzzy tester designed for a trusted operating system is not provided at the present stage, while the seeds in the fuzzy test process of the traditional operating system are in a binary format, and the variation process is usually to perform modification operations such as bit inversion, bit grouping calculation, bit deletion and the like on random positions in seed files of the seeds. However, when the conventional fuzz testing method is directly applied to a trusted operating system, the testing efficiency is very low. Therefore, the invention is improved as follows:
(1) and (4) screening effective seeds. The purely random approach necessarily generates a large number of invalid seeds that are misformatted and cannot be executed, and the invalid seeds cause the test program to exit early and return to an error state. Therefore, a seed validity judgment is added between the fuzz tester and the trusted operating system to screen valid seeds, and only seeds that are deemed to be correct are executed.
(2) The test set is updated based on the plurality of feedback information. The traditional method only refers to the kernel code coverage rate condition for evaluating the mutated seeds, and the seeds which can increase the coverage rate are added into a test set. That is, only when a new effective seed is generated by mutation, the new seed can be added to the test set for further mutation, which is difficult to achieve, and the single judgment criterion greatly reduces the speed of generating the effective seed. In order to solve the problem, the invention provides a multi-feedback mutation method, which selects the seeds for mutation by taking the coverage rate of an effectiveness judgment function as the standard for selecting the seeds besides the traditional kernel code coverage rate, and retains the seeds when the coverage rate of the effectiveness judgment function or the increase value of the kernel coverage rate of a trusted operating system exceeds a threshold value.
Disclosure of Invention
The invention designs a multi-feedback variation method for the kernel fuzzy test of the trusted operating system in order to improve the security of the trusted operating system and enhance the capability of excavating the vulnerability of the trusted kernel, the method has higher adaptability to the trusted operating system, particularly alleviates the defect caused by strong randomness in the binary file variation process, improves the effective seed generation efficiency in the variation process, and is beneficial to testing each branch of the trusted operating system more quickly.
The mutation method of the invention can be divided into three basic links: randomly mutating seeds, screening the mutated effective seeds, and updating a test set used by mutation based on feedback information. The invention innovates two links of effective seed screening and test set updating in the method for testing the variation of the credible operating system in a fuzzy way. On one hand, the effectiveness of the seeds is judged after random variation, and the trusted operating system only needs to execute the effective seeds which are randomly varied; on the other hand, whether the mutated seeds need to be added into the test set is evaluated according to two indexes, which are respectively: and the coverage rate of the validity judgment function and the coverage rate of the kernel of the trusted operating system.
The invention adds an effectiveness judgment function between a credible operating system and a fuzzy tester, and is used for screening out the seeds with correct format and semantics generated after random variation, the coverage rate of the effectiveness judgment function can reflect whether the seeds can trigger a new jump branch of the effectiveness judgment function, if the new branch is triggered, the random variation is proved to be helpful for generating effective seeds, and the random variation of the varied seeds can generate the effective seeds with a larger probability. The coverage rate of the kernel of the trusted operating system reflects whether the seed can trigger a new code path in the trusted operating system, and the continuous random variation of the seed which can trigger the new path is beneficial to testing more kernel code branches of the trusted operating system.
Therefore, the tester selects the thresholds α and β (0< α, β < ═ 1) for percentage increase, respectively, and when the seed after random variation increases the coverage of the validity judgment function beyond the threshold α or the coverage of the trusted operating system beyond the threshold β, the random variation is considered to be significant, and the seed after this random variation is added to the test set. Based on the principle, the seeds in the test set used for random variation can be eliminated, and the speed of generating effective seeds by variation in the fuzzy test process is continuously accelerated, so that the fuzzy test speed is improved, and the testing range of the fuzzy test is expanded. The steps of the above-described method are shown in fig. 1, and the specific steps include: the method comprises the steps of randomly mutating seeds, screening effective seeds generated after random mutation based on an effectiveness judgment function, counting feedback information in a seed execution process, and randomly updating a test set used by the mutation based on the feedback information by a fuzzy tester.
1. Screening for effective seeds
According to the method, the interface between the fuzzy tester and the kernel of the trusted operating system is expanded, the validity judgment function is added at the interface, the trusted operating system screens the seeds before the trusted operating system executes the seeds generated by random variation, and finally the trusted operating system only needs to execute the seeds which are determined to be valid. The process of screening the seeds avoids the waste of time and resources caused by invalid seeds obtained by executing random variation by a trusted operating system, improves the efficiency of the fuzzy test, and plays an important role in accelerating the variation method for generating the valid seeds in subsequent design by using an effectiveness judgment function used in the process.
The design of the validity judgment function should take two aspects into consideration: the format is correct and the semantics are correct. Based on the two aspects, the validity judgment function needs to sequentially implement the following functions:
(1) and (6) resolving the seeds. The binary seed file obtained by random variation is analyzed into a system call description structure defined in the memory, namely a system call serial number, the number of system call parameters, the type of each parameter and a specific parameter value, and a system call can be uniquely represented by using the description structure. Firstly, the length of each system call can be judged according to the system call sequence number, so that the file is cut into a plurality of system call sequences, and the sequence of the system calls in the binary file is the actual execution sequence called when the seeds are executed; next, the types of the parameters are sequentially read out, and the offset of the data representing the parameter values in the binary file is determined according to the representation range of the parameter types, so that each parameter value is correctly read. The format of the binary seed file is shown in fig. 2, and the system call sequence form shown in fig. 3 can be obtained after parsing.
(2) And judging the correctness of the format. The format of each system call is defined in the official document of the trusted operating system, and the description data of the parsed system call and the official document are consistent in basic data type and value size range.
(3) And judging that the semantics is correct. Semantic correctness is primarily associated with a parameter, which requires that the parameter should not only fit within the range of values for the parameter type, but also fit within the semantics of the parameter type. For example: when detecting the read-write buffer type parameter, it should be determined whether the buffer is already allocated, when detecting the buffer for allocation in the parameter, it should be determined whether the trusted operating system allows allocating the buffer of the size, and when detecting the pointer parameter to be analyzed, it is necessary to determine whether the pointer points to a legal area.
2. Statistics of feedback information
Due to the uneven quality of the seeds generated after random variation of the fuzzy tester, the fuzzy tester needs to decide which seeds to use to update the test set based on some feedback information, thereby realizing dynamic tuning of the variation method. Therefore, the trusted operating system needs to collect certain information in real time and feed back the information to the mutation module of the fuzz tester during the process of executing the seeds.
The invention selects two feedback information: the coverage rate of the trusted operating system code and the coverage rate of the validity judgment function. The fuzzy tester carries out random variation by using the test set adjusted based on the two feedback information, so that effective seeds for testing can be generated more quickly, and the fuzzy testing process is accelerated by the quickly increased scale of the test case set, so that the tested code range is widened.
The coverage of the validity judgment function is related to the speed of generating valid seeds. The jumping among the basic blocks corresponds to different judgment jumping branches, so that when the coverage rate of the randomly mutated seeds can be increased, the judgment condition which is not passed before is met, and the seeds are easier to generate effective seeds compared with the seeds before random mutation.
Because the bitmap data structure storage has the advantage of saving space, aiming at the characteristic that the effectiveness judgment function and the jump branches of the kernel code are numerous, the coverage rate of the effectiveness judgment function and the coverage rate of the kernel of the trusted operating system are respectively stored in the global bitmap data structures of the effectiveness judgment function and the kernel of the trusted operating system. The code whose coverage is to be counted can be divided into several basic blocks, which refer to code blocks having only unique entries and exits. When in storage, each basic block is used as a bit of the bitmap, and one bit type data is uniquely corresponding between two bits. If the data is 0, it means that there is no skip between two basic blocks, and if it is 1, it means that there is a skip. The procedure for counting the coverage is as follows:
(1) and (5) instrumentation the target code needing the statistical coverage rate by using an LLVM tool. Firstly, the LLVM compiles a code to LLVM IR language, and jump keywords such as ret, jmp and the like in the LLVM are used as basic block dividing boundaries; next, the basic block is uniquely represented using a randomly generated integer data.
(2) Coverage was counted using LLVM IR Builder tool. In this step, an insert _ before interface provided in the tool is used to add a bitmap update instruction at the front end of the basic block, and the instruction will set the data between the last basic block before the bitmap jumps to the basic block and the basic block to 1, so as to realize the statistics of the coverage rate.
3. Updating test sets
The test set used by random variation is continuously updated based on two feedback information, namely the coverage rate of the trusted operating system code and the coverage rate of the effectiveness judgment function, and more high-quality seeds are contained. High quality seeds are divided into two categories: the first type is a seed that can test to a new branch of the trusted operating system, and the second type is an invalid seed that is more formally and semantically close to a valid seed. The principle of test set update is as follows:
(1) selecting an increased threshold value alpha for the coverage rate of the effectiveness judging function, and selecting an increased threshold value beta for the coverage rate of the trusted operating system, wherein 0< alpha, beta < ═ 1;
(2) and for all the randomly mutated seeds, if the coverage rate increase value of the validity judgment function exceeds alpha, adding the seeds into the test set, otherwise, discarding the seeds and setting all the positions in the bitmap corresponding to the validity judgment function to be 0.
(3) And for the seeds judged to be effective, if the kernel coverage value of the trusted operating system can be improved to exceed beta when the trusted operating system executes the seeds, adding the seeds into the test set, otherwise, discarding the seeds.
Drawings
FIG. 1: variant flow chart of kernel fuzzy test of trusted operating system
FIG. 2: binary format seed schematic
FIG. 3: system call sequence diagram
Detailed Description
The hardware environment of the invention is mainly a PC host. The CPU of the PC host selects the CPU of an ARM architecture as the best, an x86 architecture can be adopted, and the memory requires 4GB RAM and above 64-bit operating systems.
The software implementation of the invention takes Ubuntu16.04 and above as a platform, C language development is used, and a development tool is VS Code 2017. By means of the AFL tool and the LLVM tool, the compilation tool selects gcc, and if the CPU adopts an x86 architecture, a cross compilation tool arm-linux-gcc needs to be selected. The trusted operating system selects OP-TEE, version v3.3.0.
The experimental data are the initial seeds required during the fuzz testing, these seeds being derived from the test suite Xtest officially maintained by OP-TEE.
The invention mainly innovates two stages of effective seed screening and test set updating in a variation method of the fuzzy test process of the trusted operating system.
The two stages are specifically realized as follows:
1. screening for effective seeds
(1) Description of algorithms
Inputting an algorithm: seed, Seed _ Len, syscalls _ template.
And (3) outputting an algorithm: seed'.
Description of the drawings: seed is the initial Seed provided by the tester, Seed _ Len is the length of the initial Seed, syscalls _ template is the officially defined system call template, and Seed' is the valid Seed returned.
The algorithm comprises the following steps:
step1, analyzing Seed file Seed, defining a structure array for storing system call information in a memory, and performing a step of analyzing the Seed file Seed
The calling information obtained by analysis is stored in an array;
step2, judging whether the analyzed system call meets the format correctness or not based on the definition of each system call in the official template and the system call serial number range and the parameter type serial number range set by the trusted operating system;
step3, aiming at different types of parameters, detecting whether the analyzed system call meets semantic correctness;
seed' returned by screening.
Its pseudo code is as follows:
2. updating test sets
(1) Description of algorithms
Inputting an algorithm: and S.
And (3) outputting an algorithm: s'.
Description of the drawings: s is an old test set, and S' is an updated new test set.
The algorithm comprises the following steps:
step1, adding the seeds in the S into a queue, and then sequentially taking out the seeds from the head of the queue for mutation;
step2, selecting a threshold value alpha for increasing the coverage rate of the effectiveness judging function and a threshold value beta for increasing the coverage rate of the trusted operating system;
step3, screening seeds by using an effectiveness judgment function, meanwhile, counting the coverage rate change of the effectiveness judgment function in the execution process, and if the coverage rate increase value exceeds a threshold value alpha, adding the coverage rate increase value into a test set; otherwise, discarding;
step4, executing the screened seed by the trusted operating system, and adding the seed to the test set if the coverage rate increase value of the trusted operating system exceeds the threshold value beta; otherwise, discarding;
step5. return the new test set S'.
Its pseudo code is as follows:
Claims (4)
1. a variation method of kernel fuzzy test of a trusted operating system is characterized by comprising the following implementation steps:
(1) screening the randomly mutated seeds, wherein the trusted operating system only executes the seeds with correct format and semantic meaning;
(2) updating two feedback information of the coverage rate of the effectiveness judgment function and the coverage rate of the credible operating system in real time by the credible operating system in the process of the fuzzy test;
(3) based on the two feedback information, valuable seeds are screened and added to the test set for further variation to adjust the variation process.
2. The kernel fuzz test variation method of the trusted operating system according to claim 1, wherein the method adds a validity judgment function of the seed by expanding an interface between the fuzz tester and the trusted operating system, so as to screen out valid seeds after random variation, and the validity judgment needs to pass through the following three steps:
(1) defining a data structure for representing system call in a memory, and analyzing the binary seed file after random variation into the structure;
(2) judging whether the analyzed data in the system call description structure is consistent with the official document of the trusted operating system in the basic data type and the numerical value range or not so as to judge whether the format is correct or not;
(3) and judging whether each parameter accords with the semantics of the parameter type or not, and judging whether the seed is legal in semantics or not.
3. The kernel fuzz testing variation method of the trusted operating system according to claim 1, wherein the method adds a process of counting the coverage of the validity determination function at the stage of the coverage statistics of the variation process, and the specific statistical method is as follows:
(1) setting a global bitmap data structure, wherein each bit in the bitmap is initialized to be 0 and used for storing the coverage rate of the validity judgment function;
(2) performing basic block division and pile inserting on the effectiveness judging function, and taking each basic block as a vertex of a bitmap;
(3) adding a bitmap updating instruction at the front part of each basic block for recording the basic block before jumping and the current basic block;
(4) and in the process of running the validity judging function each time, updating the basic block jumping condition in the bitmap in real time, and when jumping between two basic blocks exists, setting the data corresponding to the two basic blocks as 1.
4. The kernel fuzzy test variation method of the trusted operating system according to claim 1, wherein at the stage of updating the test set during the variation process, the method selects valuable seeds according to two feedback information of validity determination function coverage and trusted operating system kernel coverage, and updates the varied test set using the valuable seeds, and the specific update strategy is as follows:
(1) selecting a threshold alpha of percentage increase for the coverage rate of the effectiveness judging function, and selecting a threshold beta of percentage increase for the coverage rate of the trusted operating system, wherein 0< alpha, beta < 1;
(2) for all the varied seeds, if the coverage rate increase value of the validity judgment function exceeds a threshold value alpha, adding the varied seeds into a test set, otherwise, discarding the seeds and setting all the positions in the bitmap corresponding to the validity judgment function to be 0;
(3) and for the seeds judged to be effective, if the kernel coverage value of the trusted operating system can be improved to exceed the threshold value beta when the trusted operating system executes the seeds, adding the seeds into the test set, otherwise, discarding the seeds.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111472930.3A CN114139164A (en) | 2021-12-01 | 2021-12-01 | Variation method for kernel fuzzy test of trusted operating system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202111472930.3A CN114139164A (en) | 2021-12-01 | 2021-12-01 | Variation method for kernel fuzzy test of trusted operating system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114139164A true CN114139164A (en) | 2022-03-04 |
Family
ID=80388009
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202111472930.3A Pending CN114139164A (en) | 2021-12-01 | 2021-12-01 | Variation method for kernel fuzzy test of trusted operating system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114139164A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114840437A (en) * | 2022-05-24 | 2022-08-02 | 中南大学 | Operating system kernel fuzzy test seed evaluation distribution method |
| CN115774677A (en) * | 2022-12-20 | 2023-03-10 | 上海安般信息科技有限公司 | Fuzzy test method and device based on multi-parameter input |
| CN117370994A (en) * | 2023-12-08 | 2024-01-09 | 浙江君同智能科技有限责任公司 | Large language model vulnerability detection method and device based on fuzzy test |
-
2021
- 2021-12-01 CN CN202111472930.3A patent/CN114139164A/en active Pending
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN114840437A (en) * | 2022-05-24 | 2022-08-02 | 中南大学 | Operating system kernel fuzzy test seed evaluation distribution method |
| CN114840437B (en) * | 2022-05-24 | 2023-04-07 | 中南大学 | Operating system kernel fuzzy test seed evaluation distribution method |
| CN115774677A (en) * | 2022-12-20 | 2023-03-10 | 上海安般信息科技有限公司 | Fuzzy test method and device based on multi-parameter input |
| CN115774677B (en) * | 2022-12-20 | 2024-02-23 | 上海安般信息科技有限公司 | Fuzzy test method and device based on multi-parameter input |
| CN117370994A (en) * | 2023-12-08 | 2024-01-09 | 浙江君同智能科技有限责任公司 | Large language model vulnerability detection method and device based on fuzzy test |
| CN117370994B (en) * | 2023-12-08 | 2024-02-27 | 浙江君同智能科技有限责任公司 | Large language model vulnerability detection method and device based on fuzzy test |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114139164A (en) | Variation method for kernel fuzzy test of trusted operating system | |
| You et al. | Profuzzer: On-the-fly input type probing for better zero-day vulnerability discovery | |
| Petsios et al. | Slowfuzz: Automated domain-independent detection of algorithmic complexity vulnerabilities | |
| CN109032942B (en) | Fuzzy test frame based on AFL | |
| Chen et al. | Learning to prioritize test programs for compiler testing | |
| US5933640A (en) | Method for analyzing and presenting test execution flows of programs | |
| JP4918544B2 (en) | Obfuscation evaluation method, obfuscation evaluation apparatus, obfuscation evaluation program, storage medium, and integrated circuit | |
| CN111859388B (en) | Multi-level mixed vulnerability automatic mining method | |
| CN111382067A (en) | Method and system for generating high-quality seeds in fuzzy test | |
| CN111428247B (en) | Method for improving computer leak library | |
| CN116305162A (en) | Concurrent program vulnerability detection method based on fuzzy test and static analysis | |
| CN115757149A (en) | Automatic testing method, system, electronic equipment and storage medium | |
| Sun et al. | AFLTurbo: Speed up path discovery for greybox fuzzing | |
| CN114756471A (en) | Vulnerability type oriented fuzzy test method and system based on byte sensitive energy distribution | |
| CN113935041A (en) | Vulnerability detection system and method for real-time operating system equipment | |
| CN109766690A (en) | A kind of ROP and mutation attacks dynamic testing method based on the detection of more policy instructions | |
| Sagdeo et al. | Precis: Inferring invariants using program path guided clustering | |
| CN109002723B (en) | Sectional type symbol execution method | |
| CN113419960B (en) | Seed generation method and system for kernel fuzzy test of trusted operating system | |
| US11822463B2 (en) | Computer-implemented method and device for selecting a fuzzing method for testing a program code | |
| US20200012581A1 (en) | Method for Semantic Preserving Transform Mutation Discovery and Vetting | |
| CN116166535A (en) | Method for automatically mining non-volatile memory heap loopholes based on fuzzy test | |
| Borrello et al. | Predictive context-sensitive fuzzing | |
| CN115048298A (en) | Seed scheduling weight value distribution method for data packet splicing | |
| WO2021104027A1 (en) | Code performance testing method, apparatus and device, and storage medium |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |