CN112130859A - An application running method, device, device and storage medium - Google Patents

Abstract

The invention discloses an application running method. The method uses a macro switch to decide to compile the platform source code of the target application including the assembly code corresponding to the target platform, so that the compiled execution program is executable under the target platform. program. The maintainability of the code is strong, which improves the convenience of code maintenance and reduces labor costs. The invention also discloses an application running method. In the process of running the target application on the target platform, for the target data packet to be sent, the packet header space is first expanded to obtain the reserved space, and the kernel module is in the reserved space. The padding switch label is sent to the switch chip, and the switch chip determines the network port used for transmitting the target data packet according to it. The target application can run smoothly on the application platform. The invention also discloses an application running device, equipment and storage medium, which have corresponding technical effects.

Description

An application running method, device, device and storage medium

technical field

The present invention relates to the technical field of computer applications, and in particular, to an application running method, apparatus, device and storage medium.

Background technique

With the rapid development of Internet technology and the rapid emergence of the cloud industry, the infrastructure network is gradually integrating into a network based on a general computing platform to support diverse network systems.

An application will often contain assembly code that depends on the hardware architecture of the operating environment. The implementation of these codes depends on the instruction set under a certain architecture, such as the x86 architecture. If the application is to be ported to run on platforms of other architectures, such as porting to On the platform of the Aarch64 architecture, because the instruction sets of different architectures are quite different, all the codes of the application need to be rewritten.

That is to say, if an application is to run on a platform of a certain architecture, it corresponds to a set of codes corresponding to the platform of this architecture. The code corresponding to the platform. In this way, during maintenance, it is necessary to separately maintain the code of the application corresponding to platforms with different architectures. An application not only contains code that depends on the hardware architecture of the operating environment, but also contains code that does not depend on the hardware architecture of the operating environment. If the code corresponding to platforms with different architectures is maintained separately, as long as there is a code update, multiple codes need to be changed at the same time, even if the content of the change is code that does not depend on the hardware architecture of the operating environment, each code needs to be updated. A separately maintained code with simultaneous changes. The maintainability is poor and the labor cost is high.

In addition, at present, most product architectures are based on the Linux kernel forwarding mode, which has certain problems. The first is the performance problem, which has performance bottlenecks and cannot meet the development needs of new services, and the second is the reliability problem, running in the kernel. Once the state process fails to execute, it will cause business interruption. Based on these two problems, an evolution trend recognized by the industry is to use the data plane provided by itself to send and receive data packets in user space, bypassing the processing of data packets by the Linux kernel protocol stack, so as to solve the performance and performance of traditional Linux kernel forwarding. reliability issues.

In view of the above problems, the network infrastructure platform NFP application emerges as the times require, and its data plane is an extensible framework based on VPP research and development. The VPP (Vector Packet Processing) platform is an extensible framework that provides out-of-the-box switch/router functionality. However, the existing NFP applications only support running on the hardware platform of the x86 architecture, and do not support running on the ARM hardware platform. As a result, most ARM hardware platforms still use the Linux kernel as the support for forwarding services, resulting in performance constraints on business development. .

SUMMARY OF THE INVENTION

The purpose of the present invention is to provide an application running method, device, device and storage medium, so as to improve the convenience of code maintenance, reduce labor costs, and enable the target application to run smoothly on the target platform.

In order to solve the above-mentioned technical problems, the present invention provides the following technical solutions:

An application running method, including:

When a preset version compilation trigger condition is reached, the target version of the target application to be compiled is determined according to the pre-imported environment variables and the macros required for compilation, where the target version corresponds to the target platform on which the target application is to be run version of;

Determine the storage directory of the dependency library of the binary file of the target platform and the generation directory of the binary file of the target platform;

Introducing the assembly code corresponding to the target platform of the target application;

A preset compiling tool is invoked to compile the platform source code including the assembly code, and a binary file of the target application is obtained, so as to run the target application on the target platform based on the binary file of the target application.

In a specific implementation manner of the present invention, the target version is an x86 version or an Aarch64 version.

In a specific embodiment of the present invention, the target application is a network infrastructure platform NFP application, the target version is an Aarch64 version, the target platform is an Aarch64 architecture platform, the NFP application is deployed in user space, and the Methods also include:

During the process of running the NFP application on the Aarch64 architecture platform, if the packet sending and receiving module of the data plane of the NFP application detects the target data packet to be sent, The header space is expanded to obtain reserved space;

Call the kernel function interface of the kernel space of the Aarch64 architecture platform to send the target data packet after the expansion processing, so that the target data packet filled with the switching label in the reserved space is sent to Marvell by the kernel module After switching chips, the Marvell switching chip determines the network port for transmitting the target data packet through the switching label.

In a specific embodiment of the present invention, after the packet header space of the target data packet is expanded by the packet sending and receiving module, the kernel function interface calling the kernel space of the Aarch64 architecture platform will Before the target data packet after the expansion processing is sent out, it also includes:

adding an extension mark to the target data packet;

After the kernel function interface that calls the kernel space of the Aarch64 architecture platform sends the expanded target data packet, it also includes:

If a data packet that fails to be sent is detected and the extended flag exists in the data packet, the data packet is recycled.

In a specific embodiment of the present invention, the target data packet is a data packet in the sending buffer of the user space.

An application running method, including:

In the process of running the target application on the target platform, if the target application detects the target data packet to be sent, the header space of the target data packet is expanded to obtain reserved space;

Call the kernel function interface of the target platform to send the target data packet after expansion processing, so that after the kernel module sends the target data packet filled with the switching label in the reserved space to the switching chip, the target data packet is sent out. The switch chip determines the network port for transmitting the target data packet through the switch label.

In a specific embodiment of the present invention, after the expansion processing is performed on the header space of the target data packet, the invoking the kernel function interface of the target platform will perform the expansion processing on the target data packet. Before sending, also include:

adding an extension mark to the target data packet;

After the call to the kernel function interface of the target platform sends the expanded target data packet, the method further includes:

If a data packet that fails to be sent is detected and the extended flag exists in the data packet, the data packet is recycled.

In a specific embodiment of the present invention, the target application is deployed in the user space of the target platform, and the target data packet is a data packet in the sending buffer of the user space.

In a specific embodiment of the present invention, the target platform is an Aarch64 architecture platform, the target application is a network basic platform NFP application, and the switch chip is a Marvell switch chip.

In a specific embodiment of the present invention, it also includes:

When a preset version compilation trigger condition is reached, the target version to be compiled of the target application is determined according to the pre-imported environment variables and the macros required for compilation, and the target version is the same as the target application to be run. the version corresponding to the target platform;

Determine the storage directory of the dependency library of the binary file of the target platform and the generation directory of the binary file of the target platform;

Introducing the assembly code corresponding to the target platform of the target application;

A preset compiling tool is invoked to compile the platform source code including the assembly code, and a binary file of the target application is obtained, so as to run the target application on the target platform based on the binary file of the target application.

In a specific implementation manner of the present invention, the target version is an x86 version or an Aarch64 version.

An application running device, comprising:

The version determination unit is used to determine the target version to be compiled of the target application according to the pre-imported environment variables and the macros required for compilation when a preset version compilation trigger condition is reached, and the target version is the same as the target version to be executed. The version corresponding to the target platform of the target application;

a directory determination unit, configured to determine the storage directory of the dependency library of the binary file of the target platform and the generation directory of the binary file of the target platform;

a code introduction unit, used for introducing the assembly code corresponding to the target platform of the target application;

a code compiling unit, configured to invoke a preset compiling tool to compile the platform source code containing the assembly code, and obtain the binary file of the target application, so as to generate a binary file on the target platform based on the binary file of the target application Run the target application.

An application running device, comprising:

The expansion unit is used for running the target application process on the target platform, if the target application detects the target data packet to be sent, then the header space of the target data packet is expanded to obtain reserved space;

A sending unit, configured to call the kernel function interface of the target platform to send the expanded target data packet, so that the kernel module sends the target data packet filled with the switching label in the reserved space to the target data packet. After switching chips, the switching chip determines the network port for transmitting the target data packet through the switching label.

An application running device including:

memory for storing computer programs;

The processor is configured to implement the steps of the above application running method when executing the computer program.

A computer-readable storage medium stores a computer program on the computer-readable storage medium, and when the computer program is executed by a processor, implements the steps of the above application running method.

The technical solution provided by the embodiment of the present invention is applied, and the target version to be compiled of the target application is determined according to the pre-imported environment variables and the macros required for compilation, and the storage directory and target of the dependency library of the binary file of the target platform are further determined. The directory for generating the binary files of the platform, introducing the assembly code of the target application corresponding to the target platform, calling the preset compilation tool to compile the platform source code containing the assembly code, and obtaining the binary file of the target application, which can be based on the target application. Binaries that run the target application on the target platform. That is, when the application is to be migrated from the platform of the first architecture to the platform of the second architecture, the macro switch is used to determine the platform source of the target application that contains the assembly code corresponding to the platform of the second architecture. The code is compiled, so that the compiled executable program is an executable program under the platform of the second architecture. The code of the target application that depends on the hardware architecture of different operating environments can be maintained separately, and only one copy of the code of the target application that does not depend on the hardware architecture of the operating environment can be maintained. When updating, it only needs to be updated once. The maintainability is strong, which improves the convenience of code maintenance and reduces labor costs.

In addition, in the process of running the target application on the target platform, if the target application detects the target data packet to be sent, it first expands the header space of the target data packet to obtain the reserved space, and then calls the kernel function interface of the target platform. The target data packet after expansion processing is sent out, so that after the kernel module sends the target data packet filled with the switching label in the reserved space to the switching chip, the switching chip can determine the target data packet used to transmit the target data packet through the switching label. network port. The target application can run smoothly on the target platform. In addition, it better solves the problem that the current Netmap driver does not support the switch tag DSA TAG feature. Therefore, when using the native Netmap as the sending and receiving packet module, when a data packet is sent, after the data packet is sent to the kernel module through Netmap, the kernel module supplements the DSA. The TAG is marked in the Ethernet frame, but the data packet does not reserve space to fill in the DSA TAG, so the kernel module will crash when expanding the header of the data packet to fill the DSA TAG.

Description of drawings

In order to explain the embodiments of the present invention or the technical solutions in the prior art more clearly, the following briefly introduces the accompanying drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are only These are some embodiments of the present invention. For those of ordinary skill in the art, other drawings can also be obtained according to these drawings without creative efforts.

FIG. 1 is an implementation flowchart of an application running method in an embodiment of the present invention;

2 is an overall frame diagram of the operation of the network infrastructure platform NFP in an embodiment of the present invention;

3 is an implementation flowchart of another application running method in an embodiment of the present invention;

FIG. 4 is a schematic structural diagram of an application running device corresponding to the method embodiment shown in FIG. 1;

5 is a schematic structural diagram of an application running device corresponding to the method embodiment shown in FIG. 3;

FIG. 6 is a schematic structural diagram of an application running device in an embodiment of the present invention.

Detailed ways

In order to make those skilled in the art better understand the solution of the present invention, the present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. Obviously, the described embodiments are only some, but not all, embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present invention.

Referring to FIG. 1, it is an implementation flowchart of an application running method provided by an embodiment of the present invention, and the method may include the following steps:

S110: When a preset version compilation trigger condition is reached, determine a target version to be compiled of the target application according to a pre-imported environment variable and a macro required for compilation.

The target version is the version corresponding to the target platform of the target application to be run.

In the embodiment of the present invention, a version compilation trigger condition may be preset. For example, when a compilation instruction is received, it is considered that the preset version compilation trigger condition is reached, or when an update of the assembly code is detected, it is considered that the predetermined version compilation trigger condition is reached. Preset version build trigger conditions.

When the preset version compilation trigger condition is reached, the pre-imported environment variables and the macros required for compilation can be obtained first. In practical applications, in the compilation preparation stage, the environment build script of the version can be executed according to the version to be compiled of the target application, and the environment variables and macros required for compiling the running program of this version are imported. For example, if the version to be compiled of the target application is the x86 version, execute the environment build script of the x86 version, import the environment variables required to compile the x86 version running program and the macros required for compilation; if the version to be compiled of the target application is For the Aarch64 version, execute the environment build script of the Aarch64 version, import the environment variables required to compile the Aarch64 version running program and the macros required for compilation; if the version to be compiled of the target application is another version, execute the environment build of the other version script, import the environment variables required to compile the other version of the running program and the macros required for compilation.

According to the pre-imported environment variables and the macros required for compilation, the target version of the target application to be compiled can be determined, and the target version is the version corresponding to the target platform of the target application to be run. It can be the x86 version, it can also be the Aarch64 version, of course, it can also be other versions.

x86 generally refers to a series of CPU instruction set architectures based on Intel 8086 and backward compatible, usually the 32-bit x86 architecture is called i386, x86, and the 64-bit x86 architecture is called x86-64.

Aarch64 is the 64-bit execution state of the ARMv8 instruction set architecture. ARM (Advanced RISC Machine, advanced reduced instruction set machine) is a 32-bit (or 64-bit) meta-reduced instruction set (RISC) processor architecture, which is widely used in many embedded system designs. There are also many achievements in other fields. Due to the characteristics of energy saving, the ARM processor is very suitable for the field of mobile communication, which is in line with its main design goals of low cost, high performance and low power consumption.

S120: Determine the storage directory of the dependency library of the binary file of the target platform and the generation directory of the binary file of the target platform.

After the target version of the target application to be compiled is determined, the storage directory of the dependent library of the binary file of the target platform and the generation directory of the binary file of the target platform can be further determined. The code corresponding to the platforms of different architectures is compiled, and the dependent libraries, environment variables and macros are different. Therefore, the storage directories of the dependent libraries of the binary files corresponding to the target platforms of different architectures and the generation directory of the binary files are different.

For example, if the target version to be compiled of the determined target application is the x86 version, it can be further determined that the storage directory of the dependent library of the binary file of the target platform of the x86 architecture is lib_x86, and the generation directory of the binary file of the target platform of the x86 architecture is bin_x86; if the target version to be compiled of the determined target application is the Aarch64 version, it can be further determined that the storage directory of the dependent library of the binary file of the target platform of the Aarch64 architecture is lib_arm64, and the generation directory of the binary file of the target platform of the Aarch64 architecture is bin_arm64.

S130: The assembly code corresponding to the target platform of the target application is introduced.

The assembly code corresponding to the target platform of the target application can be pre-written by a technician. After determining the target version of the target application to be compiled, the storage directory of the dependency library of the target platform binary file and the generation directory of the target platform binary file, the assembly code of the target application corresponding to the target platform can be imported.

S140: Invoke a preset compiling tool to compile the platform source code including the assembly code, obtain a binary file of the target application, and run the target application on the target platform based on the binary file of the target application.

In practical applications, the compilation tool to be used may be preset, for example, the compilation tool is set as a GCC (GNU Compiler Collection, GNU Compiler Suite) compilation tool. Calling the preset compilation tool to compile the platform source code including the assembly code, the binary file of the target application can be obtained, and the binary file is the executable program of the target application on the target platform. Based on the binary file of the target application, the target application can be run on the target platform.

The obtained binary file of the target application may be stored in the generation directory of the binary file of the previously determined target platform, and information such as compilation time may be marked in the file name of the binary file to facilitate later search.

The dependency library of the binary file of the target application is stored in the storage directory of the dependency library of the binary file of the target platform determined before, so that the binary file can be called when running the binary file.

Using the method provided by the embodiment of the present invention, according to the pre-imported environment variables and the macros required for compilation, determine the target version of the target application to be compiled, and further determine the storage directory of the dependency library of the binary file of the target platform and the target platform The generation directory of the binary file of the target application, import the assembly code corresponding to the target platform of the target application, call the preset compilation tool to compile the platform source code containing the assembly code, and obtain the binary file of the target application, which can be based on the target application. file to run the target application on the target platform. That is, when the application is to be migrated from the platform of the first architecture to the platform of the second architecture, the macro switch is used to determine the platform source of the target application that contains the assembly code corresponding to the platform of the second architecture. The code is compiled, so that the compiled executable program is an executable program under the platform of the second architecture. The code of the target application that depends on the hardware architecture of different operating environments can be maintained separately, and only one copy of the code of the target application that does not depend on the hardware architecture of the operating environment can be maintained. When updating, it only needs to be updated once. The maintainability is strong, which improves the convenience of code maintenance and reduces labor costs.

The technical solution provided by the embodiment of the present invention enhances the scalability of the application. If the application needs to be transplanted to run on the target platform of the new architecture, the developer only needs to write the environment construction script of the target platform, and the rewrite depends on the target platform. The assembly code of the new architecture can be compiled or not controlled by the macro switch, so that a code can be compiled to run programs corresponding to multiple platforms.

In one embodiment of the present invention, the target application is a network infrastructure platform NFP application, the target version is an Aarch64 version, the target platform is an Aarch64 architecture platform, and the NFP application is deployed in the user space, and the method may further include the following steps:

Step 1: During the process of running the NFP application on the Aarch64 architecture platform, if the transceiver packet module of the data plane of the NFP application detects the target data packet to be sent, the packet header space of the target data packet is expanded by the transceiver packet module to obtain Reserved space;

Step 2: Call the kernel function interface of the kernel space of the Aarch64 architecture platform to send the expanded target data packet, so that after the kernel module sends the target data packet filled with the switching label in the reserved space to the Marvell switch chip, Marvell The switch chip determines the network port for transmitting the target data packet by exchanging the label.

For the convenience of description, the above two steps are combined for description.

In the embodiment of the present invention, the target application may be a network basic platform NFP application, and NFP is a network function basic platform. The advantages of this platform are high performance and mature technology, its modularity and flexibility, and its rich function set. , which can be adapted to different product lines for building a basic network platform.

The target version can be an Aarch64 version, and the target platform is an Aarch64 architecture platform, such as a hardware platform based on a Marvell home switch, which belongs to the Aarch64 architecture.

At present, most product architectures are based on the Linux kernel forwarding mode. This mode has certain problems. The first is the performance problem, which has performance bottlenecks and cannot meet the development needs of new services. The second is the reliability problem. Once the process fails to execute, it will cause business interruption. Based on these two problems, an evolution trend recognized by the industry is to use the data plane provided by itself to send and receive data packets in user space, bypassing the processing of data packets by the Linux kernel protocol stack, so as to solve the performance and performance of traditional Linux kernel forwarding. reliability issues.

In view of the above problems, the network infrastructure platform NFP application emerges as the times require, and its data plane is an extensible framework based on VPP research and development. The VPP (Vector Packet Processing) platform is an extensible framework that provides out-of-the-box switch/router functionality. However, the existing NFP applications only support running on the hardware platform of the x86 architecture, and do not support running on the ARM hardware platform. As a result, most ARM hardware platforms still use the Linux kernel as the support for forwarding services, resulting in performance constraints on business development. .

The implementation of the present invention deploys the network basic platform NFP application in the user space and runs on the platform of the Aarch64 architecture. That is, by using the instruction set under the Aarch64 architecture to rewrite the architecture-dependent code of the NFP application of the network infrastructure platform, and then compiling the corresponding code, the network infrastructure platform NFP can run on the platform of the Aarch64 architecture.

In the process of running the NFP application on the Aarch64 architecture platform, the sending and receiving module of the data plane of the NFP application can perform real-time detection. Specifically, it can detect the sending buffer of the user space. If there are data packets in the sending buffer of the user space , you can select a data packet as the target data packet to be sent. Specifically, the module for sending and receiving packets can be a Netmap module. Netmap is a framework for processing high-speed I/O. It is implemented as a single kernel module and can be used in FreeBSD, Linux and Windows. Netmap supports access to network cards, virtual interfaces and Netmap pipes, which can easily reach the wire-speed rate of 10G network cards.

After detecting the target data packet to be sent, the packet header space of the target data packet can be expanded and processed by the sending and receiving packet module to obtain the reserved space. Specifically, a space of 4 or 8 bytes may be reserved in the header of the target data packet according to a specific network protocol.

After expanding the header space of the target data packet, the target platform, that is, the kernel function interface of the kernel space of the Aarch64 architecture platform, can be called to send the expanded target data packet. In this way, the kernel module can fill the DSA TAG in the reserved space of the target packet. Then, the target data packet filled with the switching label is sent to the Marvell switching chip. The Marvell switching chip identifies the switching label and can determine the network port used to transmit the target data packet, that is, determine which network port to transmit the target data packet through. go out. The full name of DSA TAG is DSA Switch Tag. DSA (Distributed Switch Architecture) is a term for Marvell home switches. The DSA technology enables devices to form a large-scale switching network in a cascaded manner, which speeds up packet forwarding and processing to the greatest extent.

The hardware framework of Marvell's home switch is characterized in that the data packet does not flow directly to the network port after flowing out of the Ethernet controller, but flows to the Marvell switch chip. In the Marvell switch chip, the data flow is determined by parsing the switch label DSA TAG of the data packet. The specific direction, that is, which network port to send from.

The embodiments of the present invention better solve the following technical problems: at present, the Netmap driver does not support the switch tag DSA TAG feature. Therefore, when the native Netmap is used as the module for sending and receiving packets, when data packets are sent, the data packets are sent to the kernel module through Netmap. After that, the kernel module adds the DSA TAG tag in the Ethernet frame, but the data packet does not reserve space to fill in the DSA TAG, so the kernel module will crash when expanding the header of the data packet to fill the DSA TAG.

As shown in Figure 2, NFP applications are deployed in user space, with data plane and multiple control planes. Data packets are sent from data plane. Netmap module detects the send buffer of user space to determine whether there are data packets to be sent. If so, then The data packets to be sent can be expanded and placed in the Netmap queue, and then the kernel function interface of the kernel space can be called to send the data packets. After the data packets are sent to the kernel space through the Netmap module, the kernel module will reserve the header of the data packet. The space is filled with switching labels, which are put into the network card queue based on the host protocol stack Host stack, flow out through the Ethernet controller, and flow to the Marvell switching chip. The Marvell switching chip identifies the switching label and can determine the specific direction of the data flow, that is, which network port, such as through Port1, Port2, etc.

In an embodiment of the present invention, after the first step and before the second step, the method may further include the following steps:

Add an extension tag to the target packet;

After the second step, the method may further include the following steps:

If a data packet that fails to be sent is detected and the data packet has an extension mark, the data packet is recycled.

In the embodiment of the present invention, after the packet header space of the target data packet is expanded by the sending and receiving module, the expansion mark may be added to the target data packet, and specifically, the addition may be performed by the file name of the target data packet. In this way, after calling the kernel function interface of the kernel space of the target platform to send the expanded target data packet, if the data packet that fails to be sent is detected and it is determined that the data packet has an expansion mark, the data packet can be sent out. for recycling.

Because there is reserved space in the header of the data packet for expansion processing, such a data packet will not be correctly parsed when it is transmitted to other modules. The packet can be recycled, such as deleted, to free up packet space.

Referring to FIG. 3, it is an implementation flowchart of an application running method provided by an embodiment of the present invention, and the method may include the following steps:

S310 : During the process of running the target application on the target platform, if the target application detects a target data packet to be sent, it expands the header space of the target data packet to obtain a reserved space.

In this embodiment of the present invention, the target application may be deployed in the user space of the target platform, and during the process of running the target application on the target platform, the sending and receiving module of the data plane of the target application may perform real-time detection. The sending buffer is used for detection. If there is a data packet in the sending buffer of the user space, a data packet can be selected as the target data packet to be sent. Specifically, the module for sending and receiving packets can be a Netmap module. Netmap is a framework for processing high-speed I/O. It is implemented as a single kernel module and can be used in FreeBSD, Linux and Windows. Netmap supports access to network cards, virtual interfaces and Netmap pipes, which can easily reach the wire-speed rate of 10G network cards.

After detecting the target data packet to be sent, the packet header space of the target data packet can be expanded and processed by the sending and receiving packet module to obtain the reserved space. Specifically, a space of 4 or 8 bytes may be reserved in the header of the target data packet according to a specific network protocol.

In this embodiment of the present invention, the target platform may be an Aarch64 architecture platform, and the target application may be a network basic platform NFP application.

S320 : calling the kernel function interface of the target platform to send the expanded target data packet, so that after the kernel module sends the target data packet filled with the switch label in the reserved space to the switch chip, the switch chip determines by exchanging the label The network port used to transmit target data packets.

After expanding the header space of the target data packet, the target platform, that is, the kernel function interface of the kernel space of the Aarch64 architecture platform, can be called to send the expanded target data packet. In this way, the kernel module can fill the DSA TAG in the reserved space of the target packet. Then, the target data packet filled with the switching label is sent to the switching chip, such as the Marvell switching chip. The Marvell switching chip identifies the switching label and can determine the network port used to transmit the target data packet, that is, determine which network port to use to transmit the target data packet. The destination packet is transmitted. The full name of DSA TAG is DSA Switch Tag. DSA (Distributed Switch Architecture) is a term for Marvell home switches. The DSA technology enables devices to form a large-scale switching network in a cascaded manner, which speeds up packet forwarding and processing to the greatest extent.

The hardware framework of Marvell's home switch is characterized in that the data packet does not flow directly to the network port after flowing out of the Ethernet controller, but flows to the Marvell switch chip. In the Marvell switch chip, the data flow is determined by parsing the switch label DSA TAG of the data packet. The specific direction, that is, which network port to send from.

Applying the method provided by the embodiment of the present invention, in the process of running the target application on the target platform, if the target application detects the target data packet to be sent, the header space of the target data packet is expanded and processed to obtain reserved space, Then call the kernel function interface of the target platform to send the expanded target data packet. In this way, after the kernel module sends the target data packet filled with the switching label in the reserved space to the switching chip, the switching chip can exchange the label. Determines the network port used to transmit the destination packet. The target application can run smoothly on the target platform. In addition, it better solves the problem that the current Netmap driver does not support the switch tag DSA TAG feature. Therefore, when using the native Netmap as the sending and receiving packet module, when a data packet is sent, after the data packet is sent to the kernel module through Netmap, the kernel module supplements the DSA. The TAG is marked in the Ethernet frame, but the data packet does not reserve space to fill in the DSA TAG, so the kernel module will crash when expanding the header of the data packet to fill the DSA TAG.

In an embodiment of the present invention, after step S310 and before step S320, the method may further include the following steps:

Add an extension tag to the target packet;

After step S320, the method may further include the following steps:

If a data packet that fails to be sent is detected and the data packet has an extension mark, the data packet is recycled.

In the embodiment of the present invention, after the packet header space of the target data packet is expanded by the sending and receiving module, the expansion mark may be added to the target data packet, and specifically, the addition may be performed by the file name of the target data packet. In this way, after calling the kernel function interface of the kernel space of the target platform to send the expanded target data packet, if the data packet that fails to be sent is detected and it is determined that the data packet has an expansion mark, the data packet can be sent out. for recycling.

Because there is reserved space in the header of the data packet for expansion processing, such a data packet will not be correctly parsed when it is transmitted to other modules. The packet can be recycled, such as deleted, to free up packet space.

In one embodiment of the present invention, the method may further include the following steps:

Step 1: When the preset version compilation trigger condition is reached, the target version of the target application to be compiled is determined according to the pre-imported environment variables and the macros required for compilation, and the target version is the target version corresponding to the target platform of the target application to be run. Version;

Step 2: Determine the storage directory of the dependency library of the binary file of the target platform and the generation directory of the binary file of the target platform;

Step 3: Introduce the assembly code corresponding to the target platform of the target application;

Step 4: Invoke a preset compiling tool to compile the platform source code including the assembly code, obtain the binary file of the target application, and run the target application on the target platform based on the binary file of the target application.

For the convenience of description, the above four steps are combined for description.

In the embodiment of the present invention, a version compilation trigger condition may be preset. For example, when a compilation instruction is received, it is considered that the preset version compilation trigger condition is reached, or when an update of the assembly code is detected, it is considered that the predetermined version compilation trigger condition is reached. Preset version build trigger conditions.

When the preset version compilation trigger condition is reached, the pre-imported environment variables and the macros required for compilation can be obtained first. In practical applications, in the compilation preparation stage, the environment build script of the version can be executed according to the version to be compiled of the target application, and the environment variables and macros required for compiling the running program of this version are imported. For example, if the version to be compiled of the target application is the x86 version, execute the environment build script of the x86 version, import the environment variables required to compile the x86 version running program and the macros required for compilation; if the version to be compiled of the target application is For the Aarch64 version, execute the environment build script of the Aarch64 version, import the environment variables required to compile the Aarch64 version running program and the macros required for compilation; if the version to be compiled of the target application is another version, execute the environment build of the other version script, import the environment variables required to compile the other version of the running program and the macros required for compilation.

According to the pre-imported environment variables and the macros required for compilation, the target version of the target application to be compiled can be determined, and the target version is the version corresponding to the target platform of the target application to be run. It can be an x86 version, an Aarch64 version, or, of course, other versions.

After the target version of the target application to be compiled is determined, the storage directory of the dependent library of the binary file of the target platform and the generation directory of the binary file of the target platform can be further determined. The code corresponding to the platforms of different architectures is compiled, and the dependent libraries, environment variables and macros are different. Therefore, the storage directories of the dependent libraries of the binary files corresponding to the target platforms of different architectures and the generation directory of the binary files are different.

For example, if the target version to be compiled of the determined target application is the x86 version, it can be further determined that the storage directory of the dependent library of the binary file of the target platform of the x86 architecture is lib_x86, and the generation directory of the binary file of the target platform of the x86 architecture is bin_x86; if the target version to be compiled of the determined target application is the Aarch64 version, it can be further determined that the storage directory of the dependent library of the binary file of the target platform of the Aarch64 architecture is lib_arm64, and the generation directory of the binary file of the target platform of the Aarch64 architecture is bin_arm64.

The assembly code corresponding to the target platform of the target application can be pre-written by a technician. After determining the target version of the target application to be compiled, the storage directory of the dependency library of the target platform binary file and the generation directory of the target platform binary file, the assembly code of the target application corresponding to the target platform can be imported.

In practical applications, the compilation tool to be used may be preset, for example, the compilation tool is set as a GCC (GNU Compiler Collection, GNU Compiler Suite) compilation tool. Calling the preset compilation tool to compile the platform source code including the assembly code, the binary file of the target application can be obtained, and the binary file is the executable program of the target application on the target platform. Based on the binary file of the target application, the target application can be run on the target platform.

The obtained binary file of the target application may be stored in the generation directory of the binary file of the previously determined target platform, and information such as compilation time may be marked in the file name of the binary file to facilitate later search.

The dependency library of the binary file of the target application is stored in the storage directory of the dependency library of the binary file of the target platform determined before, so that the binary file can be called when running the binary file.

According to the pre-imported environment variables and the macros required for compilation, determine the target version of the target application to be compiled, and further determine the storage directory of the dependent library of the target platform's binary files and the generation directory of the target platform's binary files, and introduce the target application. The assembly code corresponding to the target platform is called, the preset compilation tool is called to compile the platform source code containing the assembly code, and the binary file of the target application is obtained, so that the target application can be run on the target platform based on the binary file of the target application. That is, when the application is to be migrated from the platform of the first architecture to the platform of the second architecture, the macro switch is used to determine the platform source of the target application that contains the assembly code corresponding to the platform of the second architecture. The code is compiled, so that the compiled executable program is an executable program under the platform of the second architecture. The code of the target application that depends on the hardware architecture of different operating environments can be maintained separately, and only one copy of the code of the target application that does not depend on the hardware architecture of the operating environment can be maintained. When updating, it only needs to be updated once. The maintainability is strong, which improves the convenience of code maintenance and reduces labor costs.

Corresponding to the method embodiment shown in FIG. 1 above, an embodiment of the present invention further provides an application running apparatus, and an application running apparatus described below and an application running method described above may refer to each other correspondingly.

Referring to Figure 4, the device includes the following units:

The version determination unit 410 is configured to determine the target version to be compiled of the target application according to the pre-imported environment variables and the macros required for compilation when the preset version compilation trigger condition is reached, and the target version is the target version of the target application to be run. The version corresponding to the target platform;

A directory determination unit 420, configured to determine the storage directory of the dependent library of the binary file of the target platform and the generation directory of the binary file of the target platform;

a code importing unit 430 for importing the assembly code corresponding to the target platform of the target application;

The code compiling unit 440 is configured to invoke a preset compiling tool to compile the platform source code including the assembly code, obtain the binary file of the target application, and run the target application on the target platform based on the binary file of the target application.

Using the device provided by the embodiment of the present invention, according to the pre-imported environment variables and the macros required for compilation, the target version to be compiled of the target application is determined, and the storage directory of the dependency library of the binary file of the target platform and the target platform are further determined. The generation directory of the binary file of the target application, import the assembly code corresponding to the target platform of the target application, call the preset compilation tool to compile the platform source code containing the assembly code, and obtain the binary file of the target application, which can be based on the target application. file to run the target application on the target platform. That is, when the application is to be migrated from the platform of the first architecture to the platform of the second architecture, the macro switch is used to determine the platform source of the target application that contains the assembly code corresponding to the platform of the second architecture. The code is compiled, so that the compiled executable program is an executable program under the platform of the second architecture. The code of the target application that depends on the hardware architecture of different operating environments can be maintained separately, and only one copy of the code of the target application that does not depend on the hardware architecture of the operating environment can be maintained. When updating, it only needs to be updated once. The maintainability is strong, which improves the convenience of code maintenance and reduces labor costs.

In a specific implementation manner of the present invention, the target version is an x86 version or an Aarch64 version.

In a specific embodiment of the present invention, the target application is a network basic platform NFP application, the target version is an Aarch64 version, the target platform is an Aarch64 architecture platform, the NFP application is deployed in the user space, and the device further includes:

The expansion unit is used to run the NFP application process on the Aarch64 architecture platform. If the transceiver packet module of the data plane of the NFP application detects the target data packet to be sent, the packet header space of the target data packet is expanded through the transceiver packet module. , get reserved space;

The sending unit is used to call the kernel function interface of the kernel space of the Aarch64 architecture platform to send the expanded target data packet, so that after the kernel module sends the target data packet filled with the switching label in the reserved space to the Marvell switch chip , the Marvell switching chip determines the network port used to transmit the target data packet by exchanging labels.

In a specific embodiment of the present invention, it also includes:

Marking unit: After the packet header space of the target data packet is expanded and processed by the transceiver module, and the kernel function interface of the kernel space of the Aarch64 architecture platform is called to send the expanded target data packet, the target data packet is sent out. add extension tag;

The recycling unit is used to send the expanded target data packet after calling the kernel function interface of the kernel space of the Aarch64 architecture platform. Packets are recycled.

In a specific embodiment of the present invention, the target data packet is a data packet in the sending buffer of the user space.

Corresponding to the method embodiment shown in FIG. 3 above, an embodiment of the present invention further provides an application running device, and an application running device described below and an application running method described above may refer to each other correspondingly.

Referring to Figure 5, the device includes the following units:

The expansion unit 510 is used to run the target application process on the target platform, if the target application detects the target data packet to be sent, then the header space of the target data packet is expanded to obtain reserved space;

The sending unit 520 is used to call the kernel function interface of the target platform to send out the target data packet after the expansion processing, so that after the kernel module sends the target data packet filled with the switching label in the reserved space to the switching chip, the switching chip passes through. Switch labels to determine the network port used to transmit the target packet.

By applying the device provided by the embodiment of the present invention, in the process of running the target application on the target platform, if the target application detects the target data packet to be sent, the header space of the target data packet is expanded and processed to obtain the reserved space, Then call the kernel function interface of the target platform to send the expanded target data packet. In this way, after the kernel module sends the target data packet filled with the switching label in the reserved space to the switching chip, the switching chip can exchange the label. Determines the network port used to transmit the destination packet. The target application can run smoothly on the target platform. In addition, it better solves the problem that the current Netmap driver does not support the switch tag DSA TAG feature. Therefore, when using the native Netmap as the sending and receiving packet module, when a data packet is sent, after the data packet is sent to the kernel module through Netmap, the kernel module supplements the DSA. The TAG is marked in the Ethernet frame, but the data packet does not reserve space to fill in the DSA TAG, so the kernel module will crash when expanding the header of the data packet to fill the DSA TAG.

In a specific embodiment of the present invention, it also includes:

The marking unit is used to add an expansion mark to the target data packet after the expansion processing is performed on the header space of the target data packet and before the target data packet after the expansion processing is sent out by calling the kernel function interface of the target platform;

The recycling unit is used to recycle the data packet if it is detected that the data packet that fails to be sent is detected and the data packet has an expansion mark after calling the kernel function interface of the target platform to send the expanded target data packet. deal with.

In a specific embodiment of the present invention, the target application is deployed in the user space of the target platform, and the target data packet is a data packet in the sending buffer of the user space.

In a specific embodiment of the present invention, the target platform is an Aarch64 architecture platform, the target application is a network basic platform NFP application, and the switching chip is a Marvell switching chip.

In a specific embodiment of the present invention, it also includes:

The version determination unit is used to determine the target version to be compiled of the target application according to the pre-imported environment variables and the macros required for compilation when the preset version compilation trigger condition is reached, and the target version is the target of the target application to be run. The version corresponding to the platform;

The directory determination unit is used to determine the storage directory of the dependency library of the binary file of the target platform and the generation directory of the binary file of the target platform;

The code introduction unit is used to introduce the assembly code corresponding to the target platform of the target application;

The code compiling unit is used for invoking a preset compiling tool to compile the platform source code including the assembly code, obtaining the binary file of the target application, and running the target application on the target platform based on the binary file of the target application.

In a specific implementation manner of the present invention, the target version is an x86 version or an Aarch64 version.

Corresponding to the above method embodiments, an embodiment of the present invention further provides an application running device, as shown in FIG. 6 , including:

memory 610 for storing computer programs;

The processor 620 is configured to implement the steps of the above application running method when executing the computer program.

Corresponding to the above method embodiments, embodiments of the present invention further provide a computer-readable storage medium, where a computer program is stored on the computer-readable storage medium, and when the computer program is executed by a processor, the steps of the above application running method are implemented.

The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments, and the same or similar parts of the various embodiments may be referred to each other.

Professionals may further realize that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two, in order to clearly illustrate the possibilities of hardware and software. Interchangeability, the above description has generally described the components and steps of each example in terms of functionality. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Skilled artisans may implement the described functionality using different methods for each particular application, but such implementations should not be considered beyond the scope of the present invention.

The steps of a method or algorithm described in conjunction with the embodiments disclosed herein may be directly implemented in hardware, a software module executed by a processor, or a combination of the two. A software module can be placed in random access memory (RAM), internal memory, read only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, removable disk, CD-ROM, or any other in the technical field. in any other known form of storage medium.

The principles and implementations of the present invention are described herein by using specific examples, and the descriptions of the above embodiments are only used to help understand the technical solutions and core ideas of the present invention. It should be pointed out that for those skilled in the art, without departing from the principle of the present invention, several improvements and modifications can also be made to the present invention, and these improvements and modifications also fall within the protection scope of the claims of the present invention.

Claims (15)

1. An application running method, comprising:

when a preset version compiling triggering condition is reached, determining a target version to be compiled of a target application according to a pre-imported environment variable and a macro required by compiling, wherein the target version is a version corresponding to a target platform of the target application to be run;

determining a storage directory of a dependency library of the binary file of the target platform and a generation directory of the binary file of the target platform;

introducing assembly code of the target application corresponding to the target platform;

and calling a preset compiling tool to compile the platform source code containing the assembly code to obtain the binary file of the target application, and running the target application on the target platform based on the binary file of the target application.

2. The method of claim 1, wherein the target version is version x86 or version Aarch 64.

3. The method of claim 1 or 2, wherein the target application is a network infrastructure platform (NFP) application, wherein the target version is a version Aarch64, wherein the target platform is a version Aarch64 architecture platform, wherein the NFP application is deployed in user space, and wherein the method further comprises:

in the process of running the NFP application on the Aarch64 architecture platform, if a packet receiving and transmitting module of a data plane of the NFP application detects a target data packet to be sent, performing expansion processing on a packet header space of the target data packet through the packet receiving and transmitting module to obtain a reserved space;

and calling a kernel function interface of a kernel space of the Aarch64 architecture platform to send the target data packet after expansion processing, so that after the kernel module sends the target data packet with the reserved space filled with the exchange tag to a Marvell exchange chip, the Marvell exchange chip determines a network port for transmitting the target data packet through the exchange tag.

4. The method according to claim 3, wherein after the performing the expansion processing on the header space of the target data packet by the packet transceiving module and before the invoking the kernel function interface of the kernel space of the Aarch64 architecture platform sends out the target data packet after the performing the expansion processing, further comprising:

adding an expansion mark to the target data packet;

after the calling the kernel function interface of the kernel space of the Aarch64 architecture platform sends out the target data packet after the expansion processing, the method further includes:

and if the data packet which fails to be sent is detected and the data packet has the expansion mark, carrying out recovery processing on the data packet.

5. The method of claim 3, wherein the destination packet is a packet in a transmit buffer of the user space.

6. An application running method, comprising:

in the process of running a target application on a target platform, if the target application detects a target data packet to be sent, expanding the packet header space of the target data packet to obtain a reserved space;

and calling a kernel function interface of the target platform to send out the target data packet after the expansion processing, so that after the kernel module sends the target data packet filled with the exchange label in the reserved space to an exchange chip, the exchange chip determines a network port for transmitting the target data packet through the exchange label.

7. The method of claim 6, further comprising, after the expanding the header space of the target packet and before the invoking the kernel function interface of the target platform sends out the expanded target packet, the method further comprising:

adding an expansion mark to the target data packet;

after the calling the kernel function interface of the target platform sends out the target data packet after the expansion processing, the method further comprises the following steps:

and if the data packet which fails to be sent is detected and the data packet has the expansion mark, carrying out recovery processing on the data packet.

8. The method of claim 6, wherein the target application is deployed in a user space of the target platform, and wherein the target packet is a packet in a transmit buffer of the user space.

9. The method according to any one of claims 6 to 8, wherein the target platform is an Aarch64 architecture platform, the target application is a network base platform (NFP) application, and the switch chip is a Marvell switch chip.

10. The method of claim 6, further comprising:

when a preset version compiling triggering condition is reached, determining a target version to be compiled of the target application according to a pre-imported environment variable and a macro required by compiling, wherein the target version is a version corresponding to a target platform of the target application to be run;

determining a storage directory of a dependency library of the binary file of the target platform and a generation directory of the binary file of the target platform;

introducing assembly code of the target application corresponding to the target platform;

and calling a preset compiling tool to compile the platform source code containing the assembly code to obtain the binary file of the target application, and running the target application on the target platform based on the binary file of the target application.

11. The method of claim 10, wherein the target version is version x86 or version Aarch 64.

12. An application execution apparatus, comprising:

the version determining unit is used for determining a target version to be compiled of a target application according to a pre-imported environment variable and a macro required by compiling when a preset version compiling triggering condition is reached, wherein the target version is a version corresponding to a target platform to run the target application;

the directory determining unit is used for determining a storage directory of a dependency library of the binary file of the target platform and a generation directory of the binary file of the target platform;

the code introducing unit is used for introducing assembly codes of the target application, which correspond to the target platform;

and the code compiling unit is used for calling a preset compiling tool to compile the platform source code containing the assembly code to obtain the binary file of the target application, and running the target application on the target platform based on the binary file of the target application.

13. An application execution apparatus, comprising:

the system comprises an expansion unit, a receiving unit and a sending unit, wherein the expansion unit is used for expanding the packet header space of a target data packet to obtain a reserved space if the target application detects the target data packet to be sent in the process of running the target application on a target platform;

and the sending unit is used for calling a kernel function interface of the target platform to send out the target data packet after the expansion processing, so that after the kernel module sends the target data packet filled with the exchange tag in the reserved space to an exchange chip, the exchange chip determines a network port for transmitting the target data packet through the exchange tag.

14. An application execution apparatus, comprising:

a memory for storing a computer program;

a processor for implementing the steps of the application execution method according to any one of claims 1 to 5 and/or the application execution method according to any one of claims 6 to 11 when executing the computer program.

15. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the application execution method according to any one of claims 1 to 5 and/or the application execution method according to any one of claims 6 to 11.

Images