280 results about "Linux kernel" patented technology

The invention belongs to the technical field of Linux kernel schedulers, and particularly relates to a Linux kernel scheduler parameter optimization system and method.The Linux kernel scheduler parameter optimization system comprises a data collection module, a large model training and optimization module, a kernel scheduler module, a code generation and verification module, a small model training module and a small model reasoning module; the large model is responsible for periodically providing a visual angle of global optimization, generating an initial weight and a scheduler prototype of the small model, extracting a rule from long-term accumulated data, and ensuring that an optimization strategy has a global visual field. Meanwhile, the small model serves as a local adapter to process the dynamic scheduling requirement of a kernel mode in real time. The layered optimization mechanism can balance short-term adaptability and long-term performance improvement.

The invention discloses a compatible operation method for an Android camera HAL in a container based on memory access virtualization, and the method comprises the steps: taking a DMA-Buf memory heap of a Linux kernel host system as a target memory heap, taking an ION memory heap of an Android container system as a source memory heap, building a process exclusive memory management context, an FD cache table and a synchronous fence pool, and carrying out the execution of the process exclusive memory management context, the FD cache table and the synchronous fence pool; kernel registration and node binding of virtual ION equipment, pre-allocation of a first memory pool and access hook registration of kernel layer equipment are completed, and when an ION file descriptor is obtained in an HAL process, legality of a container process is verified, and context binding is initialized to the file descriptor; intercepting a memory allocation request of the HAL process, analyzing and adapting parameters, and preferentially multiplexing first memory pool resources to obtain an ION handle; when a data sharing request is processed, matching an FD cache or generating a new FD through an ION handle; when the HAL process releases the memory, resources are recycled according to a memory source, and when the HAL process exits, the context is cached or destroyed, so that cross-system memory operation compatibility is realized.

The invention relates to the technical field of data processing, and discloses a system kernel fuzzy test seed generation method and system based on a large language model. The method comprises the following steps: performing feature extraction on Linux kernel system calling on a Host host machine to obtain a feature data set; performing CFG analysis on the feature data set to obtain a system call sequence diagram and extracting a state transition table; constructing a test scene for the state transition table, and performing Syzkaller verification to obtain an effective scene library; generating an initial seed case set for the effective scene library, and executing a test in the QEMU to obtain result data; and performing coverage analysis on the result data to obtain a coverage rate report, and optimizing to generate a target test case set. According to the invention, the test coverage rate and the defect discovery capability are improved.

Embodiments of the present application relate to the technical field of communications, and provide a packet distribution method, a network device and a storage medium. The method comprises: acquiring a protocol type corresponding to a network packet; if the protocol type is a first preset protocol type, acquiring a target port number corresponding to the network packet; if the target port number belongs to a preset port number group, transmitting the network packet to a corresponding user-mode protocol stack, so that the user-mode protocol stack processes the network packet; and if the target port number does not belong to the preset port number group, transmitting the network packet to a Linux kernel protocol stack, so that the Linux kernel protocol stack processes the network packet.

The invention discloses an eBPF-based kernel optimization method, system, equipment and medium, belongs to the technical field of Linux kernel network optimization, and aims to solve the technical problem of how to improve the transmission performance for a specific data stream and specify the data stream to carry out a special routing and transmission process so as to provide services for a special network adapter and improve the transmission performance of the specific data stream. According to the technical scheme, the method comprises the steps that data packet interception is carried out on a plurality of kernel eBPF loading points and program interception points selected by a user through an eBPF kernel hook program, routing query is carried out with a routing kernel module, and selection of an operation function and routing parameters is completed; the routing kernel module completes a routing decision based on various complex routing requirements, configuration conditions, data packet conditions and environment conditions, and returns a query result to the eBPF kernel hook program, and the query result is executed by the eBPF kernel hook program. The system comprises an eBPF user mode management program module, a user mode control tool, an eBPF kernel hook program module and a routing kernel module.

According to the Linux real-time enhancement method based on the domestic real-time kernel, by introducing a kernel adaptation layer and designing a packaging mechanism of the kernel adaptation layer, it is ensured that code space between the real-time kernel and the Linux kernel is effectively isolated, the code coupling degree is reduced, the kernel and the Linux kernel can harmoniously coexist in the same environment and do not interfere with each other, and the real-time performance of the Linux kernel is improved. The real-time task can be called by a non-real-time system and inherits the existing functions and services of the Linux, so that the ecological advantages of the Linux are brought into full play, effective management of hardware and memory resources is ensured through flexible and diversified resource allocation and sharing strategies, and the problems of resource contention and performance reduction are reduced.

The invention provides a dual-storage system starting method and device. An architecture in which a master control chip SOC works cooperatively with an external memory DDR, a nonvolatile memory SPI NOR and an SD card is adopted. After the SOC is powered on, a BootROM guide code is executed, a system is initialized, ID Blocks containing a DDR initialization program and an SPL secondary loading program are sequentially obtained and verified from multiple storage media, the DDR initialization program is loaded, DDR is initialized, and the SPL secondary loading program is loaded to the DDR to run. The method comprises the following steps of: firstly, loading an SD (Secure Digital) card, then loading U-Boot firmware, completing initialization of residual hardware and security mirror verification, then detecting and verifying a kernel mirror format and Header information in the SD card, then loading a Linux kernel, an equipment tree and a ramdisk, mounting rootfs to an RAM (Random Access Memory) space, and mounting the SD card as a user data storage space. According to the method, high reliability of system guiding is ensured through the SPI NOR, large-capacity storage and flexible upgrading are achieved through the SD card, and the comprehensive performance of the embedded system in the aspects of reliability, storage expansibility and cost effectiveness is remarkably improved.

The invention provides a time synchronization method and system for a wireless deterministic network, and relates to the technical field of communication. The method comprises the following steps: designing a handshake protocol based on a Linux kernel, and generating a message with an identifier; the message is identified and a timestamp is inserted when the message is sent to pass through the MAC layer, and the message is identified and the timestamp is captured when the message is received to pass through the MAC layer; performing synchronization error compensation calculation according to the acquired timestamp information, wherein synchronization error compensation comprises phase error compensation and frequency error compensation; designing a compensation filter, and inputting the calculated compensation data into the compensation filter for screening; and inputting the screened compensation data into a clock system for compensation to finish synchronization. According to the method, the influence of asymmetric transmission delay and wireless transmission abnormal compensation data on the time synchronization precision in a wireless time synchronization mechanism can be effectively inhibited, so that the time synchronization precision and reliability are improved.

The invention discloses a power distribution terminal upgrading method, system and device based on an SD card and a medium. The method specifically comprises the steps of generating an incremental package of a Linux kernel and / or a root file system, calculating a hash value file of the incremental package, and generating a block chain certificate with a timestamp; encrypting and packaging the incremental package, the hash value file and the block chain certificate into a container mirror image, performing digital signature to form an upgrade package, and writing the upgrade package into an SD card; after it is detected that the SD card is inserted, mounting a boot partition, analyzing metadata of an upgrade package, and setting an upgrade flag bit and an upgrade operation type of Boot; and performing digital signature verification, hash value file verification and block chain certificate verification on the upgrade package in the SD card, if the verification is passed, backing up user parameters to the encrypted storage area, and restarting to enter the Recovery OS to complete replacement upgrade of the Linux kernel and / or the root file system. According to the method, safe, efficient and flexible upgrading of the operating system of the power distribution automation terminal is achieved, the rollback function is achieved, and the upgrading risk and cost are effectively reduced.

The invention discloses a kernel probe generation method and device, computer equipment and a storage medium, and relates to the technical field of operating system kernel monitoring, and the kernel probe generation method comprises the steps of obtaining a target metadata file corresponding to a current operating system kernel in response to a kernel probe generation request; determining a target declarative strategy file; and determining a target program logic file corresponding to the target declarative strategy file, and generating a probe code according to the target program logic file and the target metadata file to obtain a kernel probe program. The probe program source code adaptive to the current kernel version can be dynamically generated in combination with the target metadata file and the target declarative strategy file, kernel probe development logic and underlying kernel knowledge can be decoupled through the source code generation function, and developers do not need to deeply understand kernel internal data structures and version differences; the development threshold is lower, a specific operating system or probe type is supported, and a universal Linux kernel and multiple probe types are also supported.

The invention provides a DPU equipment initialization method, and relates to the field of computer equipment. According to the method, a BAR space mapping process, a PCI device initialization process, a PF data channel initialization process and a VP data channel initialization process are executed respectively; the method comprises the following steps: firstly, carrying out IO remap mapping operation on a BAR register, and mapping a physical address of the BAR register to a virtual address space of a Linux kernel to obtain a BAR space; then PCI device initialization is carried out through the BAR space; secondly, initializing a PCI driver and a kernel timer, and on the basis, maintaining a work queue task and periodically executing port state polling; and finally, distributing an interrupt vector to the VP data channel, and managing the on-off of the VP data channel through the interrupt vector. The invention provides a complete and systematized initialization processing scheme of the DPU high-speed daughter card based on the domestic chip, which is beneficial to improving the development efficiency of domestic equipment.

According to the kernel indirect control flow transfer dynamic verification method based on the eBPF, an eBPF attachment point of an indirect call instruction is positioned, a verification program is dynamically attached, a source address and a target address are captured to generate a compound key, and the legality of a control flow is verified in a pre-constructed BPF hash table; and meanwhile, three indirect addressing modes of a differential strategy adaptive register, an immediate operand and a memory are adopted. Based on an eBPF technology, an EV program can be dynamically loaded into a running kernel as required during running, and the characteristic of dynamic expansion of a modern kernel can be perfectly adapted; all the capabilities are constructed on a universal software mechanism of a Linux kernel and do not depend on any hardware extension; while powerful security defense is provided, extremely low overhead during operation is realized. Benefited from an innovative hybrid verification strategy and an efficient CFG storage scheme, control flow hijacking attacks in the real world can be effectively defended with extremely low performance loss.

The invention discloses a file hiding method and system based on an ext4 file system, and the method comprises the steps: obtaining a target file and metadata information of a directory where the target file is located according to a target file path needing to be hidden, and distributing a new mapping item in an internal mapping table of the ext4 file system for recording; creating a new hard link in a directory where the target file is located, and associating the new hard link with an inode node for storing metadata information of the target file; reopening the ext4 file system, and deleting the directory entry of the target file, so that the target file cannot be found in the ext4 file system; and re-mounting the ext4 file system, refreshing the Linux kernel cache, and recording the new hard link, the original name of the target file, the inode node number of the target file and the file type into a new mapping item. According to the invention, data security can be ensured, and efficient hidden storage of files is realized.

The invention discloses a method for detecting linux kernel memory leak. The method comprises the following steps: acquiring a use index of a system kernel memory through a monitoring program; determining whether memory leak occurs or not according to the use index; and in response to determining that the memory leak occurs, starting an eBPF detection program to monitor memory allocation and release activities in the kernel, and analyzing the memory allocation and release activities to generate a memory leak report to locate the memory leak. According to the method, the system kernel memory is monitored by depending on the system information, the use index of the system kernel memory is acquired through the monitoring program, and whether memory leakage occurs or not is judged according to the use index, so that the use condition of the kernel memory can be monitored in real time with very small system overhead, and the system reliability is improved. And detecting whether the kernel memory leak occurs or not.

The invention discloses a container system process life cycle management method and system, and relates to the technical field of computers, and the method comprises the following steps: constructing a container cluster system covering a single machine and a cluster mode; a configuration management mapping scheme from container process management to host process management is designed, and process configuration in a container is converted into process configuration on a host by using a namespace and a resource management subsystem of a Linux kernel. According to the container system process life cycle management method and system, the resource utilization rate and the system performance are improved through fine resource allocation and NUMA optimization, under the NUMA architecture, by identifying and configuring NUMA nodes, it is ensured that the process preferentially uses a memory on a specific NUMA node, memory access delay is reduced, the data processing efficiency is improved, and the service life of the container system process is prolonged. CPU and memory resource allocation is dynamically adjusted according to the priority and the dependency relationship of the processes, and it is ensured that the key processes obtain enough resources to maintain the optimal performance.

The embodiment of the invention provides a message distribution method, network equipment and a storage medium, and relates to the technical field of communication. The method comprises the following steps: acquiring a protocol type corresponding to a network message; if the protocol type is a first preset protocol type, obtaining a target port number corresponding to the network message; if the target port number belongs to a preset port number group, transmitting the network message to a corresponding user mode protocol stack, so that the user mode protocol stack processes the network message; and if the target port number does not belong to the preset port number group, transmitting the network message to a Linux kernel protocol stack, so that the Linux kernel protocol stack processes the network message. The embodiment of the invention aims to solve the problem of resource conflict when a user mode protocol stack based on the AFXDP and a Linux kernel protocol stack share a network interface.

The invention discloses a unified multi-version kernel service network traffic acquisition method and system based on a high-performance non-intrusive technology, and belongs to the technical field of computers. The method comprises the following steps: collecting traffic according to a kernel type of an operating system; and analyzing the collected flow data. Through the scheme of the invention, efficient collection of network traffic in various kernel versions and operation environments can be realized, and technical obstacles caused by kernel version differences do not need to be worried about. No matter a high-performance server, a device with a low-version Linux kernel or a container or a cloud environment with limited authority, seamless flow capture and analysis can be realized by using the system provided by the invention. Meanwhile, the protocol analysis and flow topology function helps the user to quickly understand the network condition and output a unified flow index, thereby greatly improving the efficiency of network operation, maintenance and analysis.

The invention discloses a method for realizing quick start of a Linux system, electronic equipment and a storage medium, the method comprises the following steps: reading parameters of a UEFI (Unified Extensible Firmware Interface) start item of a BIOS (Basic Input / Output System) of the system, and entering a quick start mode according to the parameters of the UEFI start item; in the quick start mode, loading a Linux kernel through a specified path, and transmitting UEFI (Unified Extensible Firmware Interface) start parameters of the kernel to the Linux kernel; loading the optimized drive program according to blacklist information set by the optimization strategy; optimizing the system service, and loading the optimized system service; and displaying a Linux system login interface. According to the embodiment of the invention, the dependence on additional storage space is reduced, so that the requirement on the storage capacity of the system is reduced, and the starting process is simplified and the complexity of configuration and management is reduced by optimizing the starting of the driving program and adjusting the starting sequence of the system service.

The invention discloses a method for ensuring hard real-time in a Linux user mode based on a multi-core CPU, and belongs to the technical field of embedded systems. According to the method, a Linux kernel is configured to be in a completely preemptable mode, user threads with the number equal to that of CPU kernels are created and bound to different kernels, a real-time scheduling kernel is selected, the thread scheduling strategy of the real-time scheduling kernel is set to be SCHEDFIFO, the priority of the real-time scheduling kernel is equal to that of the lowest-priority soft interrupt, the thread stack space is locked, the real-time kernel is isolated through kernel parameters isolcplus and nohzfull, and the cost of tick is avoided. Finally, the real-time core is exclusively used for executing the real-time task. According to the scheme, the problem that the real-time performance of the Linux user state task is poor due to the fact that a kernel cannot be preempted, scheduling delay, memory page changing, system interference and the like is effectively solved, and extremely low delay and high certainty of user state task response are ensured through multi-level collaborative optimization.

The invention relates to the technical field of embedded operating systems, in particular to a multi-core processor AMP architecture real-time operating system starting method, which comprises the following steps: reserving a real-time operating system memory in U-Boot and Linux equipment trees by defining nodes, and setting specific attributes to prevent resource conflicts; a real-time operating system core is forbidden in a Linux device tree, and a node is defined to reserve an inter-core communication shared memory. After the related mirror image is solidified by utilizing a burning tool, the U-Boot firstly loads the mirror image of the real-time operating system to a reserved address, starts a corresponding core by virtue of a CPUON function of a power management interface specification, then loads a Linux kernel and a file system to start the Linux, and the dual systems communicate through a shared memory segment; the starting speed of a real-time operating system is increased, the high-real-time scene requirement is met, and the starting delay risk is reduced.

The invention discloses a kernel-level resource creation management and control method and system based on an eBPF. The method comprises the steps that when a system kernel receives a creation request of a virtual machine and / or a container, an eBPF verification module reads an authorization strategy corresponding to the creation request from an eBPF Map; judging whether the creation request of the virtual machine and / or the container conforms to the authorization strategy or not; if the authorization strategy is met, the execution of the virtual machine and / or container creation request is allowed, and the creation of the virtual machine and / or container is completed; if the authorization policy is not met, stopping the execution of the virtual machine and / or container creation request; and the eBPF verification module runs in a system kernel and executes verification operation of the authorization strategy in real time. According to the method, the eBPF strategy engine is deployed in the Linux kernel, the creation requests of the container and the virtual machine are intercepted and verified in real time, and dynamic blocking and automatic destroying of unauthorized resources are achieved.

The invention discloses a system call hijacking method and device compatible with multiple versions of Linux kernels, a first hijacking channel and a second hijacking channel are deployed in parallel in a target version of Linux kernel in advance, the first hijacking channel is adapted to a Linux kernel version executing system call through a system call table, and the second hijacking channel is adapted to a second hijacking channel which is adapted to a Linux kernel version executing system call through a system call table. The second hijacking channel is adapted to a Linux kernel version which executes system calling through a system calling distribution function, and the method comprises the following steps: in response to a system calling request, hijacking the system calling request to a pre-defined target address through a target hijacking channel which actually takes effect in a Linux kernel of a target version, determining a corresponding target function based on the target address; sorting system call parameters carried in the system call request into a data format required by a preset hook entry function through a target function, and sending the data subjected to format sorting to the hook entry function; and determining a security processing strategy for the system call request through a hook entry function on the basis of the formatted data.

The invention provides a body area network networking method fusing multiple physical networks. The method comprises the following steps: 1, performing network initialization: controlling discovery, configuration and management of network equipment through an intelligent terminal SDN (Software Defined Network); 2, sending data: coordinating communication of different types of equipment by using an IP-based network protocol stack; and step 3, data receiving: de-encapsulating the received IP data packet through a network stack of the Linux kernel and transmitting the de-encapsulated IP data packet to target equipment. According to the invention, a unified IP network stack is introduced on body area network controller software. The existing network capability of Linux is utilized, and the method is compatible with the existing traditional network equipment at low cost.

The invention discloses a Linux-based virtual input method and terminal, and the method comprises the steps: creating a character device file of a virtual device corresponding to an HMI device by using a Linux kernel, reading an input event coordinate of a physical touch node of the HMI device, obtaining a backlight state of the HMI device, and if the backlight state is a normally-on state, executing the operation of the HMI device; if the backlight state is an off state, it is indicated that the current input event is the real operation of the user on the HMI device, the coordinates of the input event are written into the character device file of the virtual device, and if the backlight state is an off state, it is indicated that the current input event is only the user wants to light the screen, and a system interface is called to light the backlight. According to the method, the input event coordinate is not written into the character device file of the virtual device, and the button operation corresponding to the coordinate is ignored, so that the input event is selectively filtered by using the character device file of the virtual device according to specific conditions, and misoperation of the HMI device is effectively avoided.

The invention provides a monitoring method and system for a high-frequency transaction scene based on an eBPF. The monitoring method comprises the steps that a pre-compiled eBPF byte code program is loaded into a Linux kernel; an eBPF probe is deployed at an event trigger point of each layer of the network protocol stack; establishing a data storage and processing mechanism in a kernel space by using an eBPF Map, acquiring real-time data through a probe, and performing preliminary analysis; further processing is carried out based on the analysis result, and if abnormity is detected, an alarm is given and information is recorded. The technical problem that microsecond-level continuous monitoring and accurate acquisition of abnormal events cannot be realized in high-frequency transactions is solved.

The invention discloses a kernel-level heterogeneous collaborative AI reasoning method and device, equipment and a storage medium, and relates to the field of AI reasoning acceleration. When the KDS receives a reasoning request submitted by a user space, generating a cross-device plan list according to a global system resource state and issuing the cross-device plan list to the HCEE; the HCEE drives the corresponding target execution equipment according to the task execution sequence, reads the data information from the UCMP, or restores intermediate / result data reasoned and output by the target execution equipment; after reasoning of the plan list is completed, the kernel space triggers an event notification mechanism, and a user process of the user space is awakened; and the user process directly reads the reasoning result from the UCMP. According to the scheme, deep collaboration of task scheduling, memory management and equipment execution is realized on a Linux kernel level, the problems of high overhead, resource islands, memory redundant copy, extensive scheduling granularity and the like of a user mode reasoning framework system are solved, and end-to-end low-delay, high-throughput and high-energy-efficiency AI reasoning acceleration is realized.

The invention provides an improved scheme of a fuzzy testing system of a Linux operating system. According to the scheme, fuzzy testing is guided by integrating device attributes, driving attributes and topological relations among devices. Specifically, a new syzlang description is generated by deeply analyzing a Linux kernel source code, so that a test case can introduce modification operation of equipment and drive attributes. Meanwhile, according to the scheme, the depth and the breadth of the fuzzy test are improved based on the corresponding relation between the equipment attribute file and the equipment file, so that the behavior of the kernel driver is more effectively explored. In addition, according to the scheme, concurrent testing is guided through the topological relation between the devices, testing cases related to device attributes and drive attributes are increased, the testing efficiency is further improved, the coverage rate of fuzzy testing is increased, and finally the testing process can reveal vulnerabilities caused by memory errors of a kernel of the Linux operating system in different states.

The invention provides a Linux kernel file access management and control method and device based on black and white lists, and the method comprises the steps: automatically recognizing a sensitive file, and adding a grading mark for the sensitive file according to a management and control strategy mode; the Linux kernel intercepts and checks whether a program accessing the sensitive file allocates a static permission mark or not; judging whether access is allowed or not according to the static permission mark of the program and the hierarchical mark of the sensitive file; extracting program information according to the management and control strategy mode, and matching black and white list features; judging whether access is allowed or not according to a matching result and a grading mark of the sensitive file; and switching a management and control strategy mode according to requirements in system operation, and modifying the grading marks of the sensitive files. According to the method, a blacklist and whitelist mechanism is realized in a Linux kernel, and operations such as access, modification and deletion of a program to a sensitive file are monitored and controlled in real time in combination with a configuration strategy, so that the security of a system is effectively improved, and risks caused by malicious programs or improper operations are prevented.

The invention discloses a distributed training method and device for communication acceleration, equipment and a storage medium, and relates to the technical field of distributed training, the method comprises the following steps: dividing gradient data of a current node into a plurality of gradient blocks, and storing each gradient block and corresponding metadata in a target cache region; receiving a gradient data packet sent by an upstream node; performing aggregation operation or overwriting operation on the gradient data packet and a corresponding gradient block in the target cache region according to the operation type of the gradient data packet, and writing an operation result back to a load of the gradient data packet; and modifying the destination IP address of the gradient data packet into the downstream node according to the metadata in the gradient data packet, and redirecting the gradient data packet to the downstream node. According to the method, the distributed training scheme can be realized through the eBPF program mounted on the Linux kernel, acceleration of gradient aggregation can be realized without depending on additional network equipment, the communication efficiency of distributed training can be improved, and the communication cost can be reduced.

The invention discloses a high-precision flow control protocol simulation system and method for a large-scale network, and relates to the field of network operation simulation services. In the system, a user flow control protocol integration module automatically converts a congestion control protocol algorithm or a flow control protocol provided by a user into a BPF program; and the congestion control protocol simulation module compiles the BPF program into a byte code which can be operated by the Linux kernel, and mounts the byte code to the hook function. The traffic generator generates traffic by using a mounted congestion control algorithm or a traffic control protocol; the data collecting and processing module collects and processes the flow of the flow control protocol simulation experiment by using a database; and the experimental network topology management tool generates a network topology and performs a simulation experiment based on the flow generated by the flow generator. According to the method and the device, the simulation overhead can be reduced, the complex condition in an actual network can be completely and truly reflected, and the applicability and authenticity of a simulation result cannot be limited.