WO2016131299A1

WO2016131299A1 - Packet reception method and apparatus for processor network interface inside device

Info

Publication number: WO2016131299A1
Application number: PCT/CN2015/092731
Authority: WO
Inventors: 何清
Original assignee: 中兴通讯股份有限公司
Priority date: 2015-07-13
Filing date: 2015-10-23
Publication date: 2016-08-25
Also published as: CN106357554A

Abstract

Disclosed are a packet reception method and apparatus for a processor network interface inside a device. The apparatus comprises: a packet reception hardware module configured to deliver a received data packet message to a corresponding VCPU queue according to a pre-configuration; a packet reception threading module configured to receive the data packet message in the bound VCPU queue, determine the priority of the data packet message according to the Ethernet type of the received data packet message, and deliver the received data packet message to a corresponding common queue according to a pre-configured correspondence between the priority of the data packet message and the priority of the common queue; and a service module configured to read the data packet message in the common queue according to the priority order of each common queue.

Description

Device internal processor network port receiving method and device

Technical field

This application relates to, but is not limited to, communication technology.

Background technique

The common way for the internal network port of the device to send and receive packets is to create a socket on the network port through the receiving thread, and then guard the socket to receive the packet. The received packets of each type are not distinguished and stored in a cache. The queue is for upper-layer services to read; this collection mode brings the following problems:

1. The message must be processed by the kernel protocol stack and share CPU resources, resulting in lower performance of the packet receiving;

2. The packet type is not differentiated. As a result, when a large number of packets are flooded into the network port, packets are not discarded. The loss of some keep-alive packets will cause the service protocol to be broken.

3. There is no rate limit for packets. As a result, a large number of packets may be occupied by the CPU. As a result, other applications may not be scheduled, resulting in abnormalities.

Summary of the invention

The following is an overview of the topics detailed in this document. This Summary is not intended to limit the scope of the claims.

This paper provides a method and device for collecting the internal processor network port of the device to solve the problem of bottleneck performance bottleneck in the related art.

A device internal processor network port receiving device, the device comprising:

The receiving hardware module is configured to: deliver the received data packet message to the corresponding virtual central processing unit VCPU queue according to the pre-configuration;

The receiving thread module is configured to: receive the data packet in the VCPU queue bound to the packet, and determine the priority of the Ethernet packet according to the Ethernet type of the received packet, according to the pre-configured Correspondence between the priority of the packet and the priority of the public queue a public queue, which delivers the received packet message to the determined corresponding public queue;

The service module is configured to: read the packet information in the public queue according to the priority order of each public queue.

Optionally, the foregoing apparatus further includes:

The packet management module is configured to: pre-configure a correspondence between the data packet received by the receiving hardware module and the VCPU queue.

Optionally, in the foregoing apparatus, the receiving management module is further configured to: establish a specified number of public queues in the shared memory, and pre-configure a correspondence between a priority of the data packet and a priority of the public queue. relationship.

Optionally, in the foregoing apparatus, the receiving management module is further configured to: obtain, from each service module, a flow limit value of each type of Ethernet packet message;

The receiving thread module is further configured to: stop sending data receiving packets to the corresponding public queue when one or more types of Ethernet packet data flows reach or exceed the flow limit value.

Optionally, in the foregoing apparatus, the receiving thread module is configured to: the receiving thread module, when the packet packet flow reaches or exceeds the flow limit value for the first time, the receiving flag is set to be rejected. Receiving the packet; when the number of times the packet packet reaches or exceeds the stream limit value reaches the set number of times, the packet count is cleared, and the packet receipt flag is reset to allow the packet to be received.

A device internal device network port receiving method, the method comprising:

Delivering the received packet message to the corresponding virtual CPU VCPU queue according to the pre-configuration;

Determining the priority of the packet according to the Ethernet type of the packet in the VCPU queue, according to the correspondence between the priority of the pre-configured packet and the priority of the public queue. Queueing, and delivering the packet message to the determined corresponding public queue;

Packets in the received public queue are read in the order of priority of each of the pre-configured public queues.

Optionally, the foregoing method further includes:

Pre-configure the correspondence between received packet messages and VCPU queues.

Optionally, the foregoing method further includes:

Establish a specified number of public queues in the shared memory, and pre-configure the correspondence between the priority of the packet and the priority of the public queue.

Optionally, the foregoing method further includes:

Obtaining a flow limit value of each type of Ethernet packet received by each service;

When one or more Ethertype packet flow reaches or exceeds the flow limit value, the delivery of the received packet message to the corresponding public queue is stopped.

Optionally, in the foregoing method, when one or more types of Ethernet packet flow meet or exceed the flow limit value, the process of stopping delivery of the data collection message to the corresponding public queue includes:

When the packet packet stream reaches or exceeds the stream limit value for the first time, the packet receiving flag is set to refuse to receive the packet; when the number of times the packet packet stream reaches or exceeds the stream limit value reaches a set number of times, The receipt count is cleared and the receipt flag is reset to allow for collection.

A computer readable storage medium storing computer executable instructions for performing the method of any of the above.

The technical solution of the present application directly receives the packet by binding the hardware thread, so that the performance of the device collection is greatly improved, and the performance bottleneck of the previous packet collection is solved. At the same time, the technical solution of the present application adds a priority queue according to the Ethernet type of the packet, and the priority queue is preferentially processed by the priority queue. In the optional solution, the rate limit is configured to process packets according to the packet type. The rate limit mode is adopted to ensure that security is improved when a large number of attack packets are flooded into the device.

Other aspects will be apparent upon reading and understanding the drawings and detailed description.

BRIEF abstract

1 is a schematic structural diagram of a device internal network port receiving device of a device according to Embodiment 1 of the present invention;

FIG. 2 is a flowchart of receiving a packet of an internal processor network port of a device according to an embodiment of the present invention;

FIG. 3 is a flowchart of a method for collecting a packet of a device internal processor network port according to Embodiment 2 of the present invention.

Embodiments of the invention

Embodiments of the present invention will be described in detail below with reference to the accompanying drawings. It should be noted that, in the case of no conflict, the features in the embodiments and the embodiments of the present application may be combined with each other arbitrarily.

Example 1

As shown in FIG. 1 , this embodiment provides a device internal device network port receiving device, which includes the following modules:

The receiving hardware module 11 is configured to: deliver the received data packet message to the corresponding VCPU queue according to the pre-configuration;

The packet receiving hardware module 11 is a processor network port on the main control board, and the network port receives the service data packet message sent from each line card. The processor chip in the receiving device is multi-core and supports hardware receiving.

The receiving thread module 12 is configured to: receive the data packet in the VCPU queue bound to the packet, and determine the priority of the packet according to the Ethernet type of the received packet, according to the pre-configured data. Corresponding relationship between the priority of the packet and the priority of the public queue, and delivering the received packet to the corresponding public queue;

The service module 13 is configured to: read the packet information in the public queue according to the priority order of each public queue.

In addition, the foregoing apparatus may further include a receiving management module 14, which sets the processor to directly receive the packet from the network port and arbitrarily binds the core, that is, pre-configures the data packet received by the receiving hardware module 11 in an initial stage. Correspondence between packets and VCPU queues.

In addition to the foregoing operations, the collection management module 14 is further configured to: interact with each business task, obtain a public queue corresponding to the Ethernet type of the data packet by interacting with the business task, that is, establish a specified number of public in the shared memory. Queue, and pre-configure the correspondence between the priority of packet packets and the priority of the public queue. Among them, the public type can be determined according to the ether type of the packet message The number of queues, for example, the number of public queues is the same as the number of Ethernet types of packet packets, where a type of packet message corresponds to a public queue.

In addition to considering the priority of packet types of different types of Ethernet packets, it is also possible to detect whether the current packet receiving flow rate meets the speed limit requirement by combining the traffic rate limit of different types of Ethernet packet packets, thereby correspondingly processing. That is, each service module 13 sets a flow limit value of the packet type of the packet according to the Ethernet type of the packet read by its own service, for example, the number of public queues that can be established by the packet management module 14 and Its corresponding priority, set the flow limit value corresponding to its message type. At this time, the packet management module 14 is further configured to: obtain, from each service module, a flow limit value of each type of Ethernet packet. The receiving thread module 12 is further configured to stop delivering the packet message to the corresponding public queue when the one or more Ether type packet flow reaches or exceeds the flow limit value.

The receiving thread module 12 stops sending the data packet message to the public queue in two steps. The first step is that when the data packet flow reaches or exceeds the corresponding flow limit value for the first time, the receiving thread module 12 can receive the packet. The packet flag is set to reject the packet (deny); in the second step, when the number of times the packet packet flow reaches or exceeds the corresponding packet stream limit value reaches a set number of times, the packetizing thread module 12 can count the packet collection. Cleared, the receipt flag is reset to allow for a permit.

The working principle of the above device will be described in detail below with reference to the embodiments of the device internal processor network port receiving device provided in the accompanying drawings.

As shown in FIG. 1 , the main control board unit includes a multi-core CPU unit module, and the CPU needs to support multiple cores and can receive packets directly by hardware. The main control board unit includes: receiving hardware. The module 11, the collection threading module 12, the collection management module 14, and the business module 13.

The working process between each module in the device internal processor network port receiving device shown in FIG. 1 is as follows:

First, the collection management module 14 sets the collection hardware module 11 in the initialization phase to perform initialization of the underlying hardware transceiver package; and through the setting of the collection hardware module 11, the hardware of the device system is delivered to a VCPU after being packaged. Queue VC0; at this time, the receiving thread module 12 is tied The packet is received from the queue VC0 in the corresponding VCPU; in addition, the packet management module 14 obtains the Ethernet type of the packet message of each service module 13 through interaction with each service module in the initial stage. The priority of the packet type and the shared memory are determined by the Ethernet type of the packet, and a specified number of public queues are established in the shared memory, and the correspondence between the priority of each packet and the priority of the public queue is configured. The receiving thread module 12 can then deliver the data packet received from the VC0 to the corresponding public queue according to the Ethernet type and priority of the data packet; the receiving hardware module 11 obtains in the initialization phase. The public queue with the priority is set, and then the received packet message is sent to the service module 13. In addition to the above, the service module 13 also informs the packet management module 14 of the stream limit value of the packet packet through the shared memory. First, the packet management module 14 in the initialization phase acquires the flow rate limit of each packet packet through the database. The value is saved. The second is the change of the packet flow during the operation. At this time, the database notifies the packet management module 14 to modify the flow rate limit value of the saved packet packet, and obtains the flow rate limit value of the packet packet. The receiving management module 14 can periodically detect the current actual message flow rate by using a timer of 100 milliseconds. If the previously specified speed limit value is exceeded, the receiving flag is set to deny; the timer expires 10 times, that is, 1 second. The receipt count is cleared and the receipt flag is reset to permit.

A more detailed interaction process between the packet threading module and the service module before and after receiving the first packet message is described below with reference to FIG.

Step 201, the receiving management module first initializes the underlying chip so that it can receive the packet directly into the VCPU through the hardware;

Step 202: The service module first initializes the receiving parameter, and reports the Ethernet type and the speed limit value to the receiving management module.

Step 203: The collection management module acquires an Ethernet type and a speed limit value.

Step 205: The collection management module sets the number of the public queues according to the number of the acquired Ethernet types, and the index of each public queue corresponds to the corresponding priority.

Step 205, opening a shared memory to establish a public queue;

Step 206: When the receiving threading module receives the first data packet, the service module is notified by the message to receive the packet.

Step 207: At this time, the service module reads the packet message from the public queue.

Example 2

As shown in FIG. 3, this embodiment provides a method for receiving a network port of a processor internal device, and the method mainly includes the following operations:

Step 301: Deliver the received data packet message to the corresponding VCPU queue according to the pre-configuration;

Step 302: Determine the priority of the packet according to the Ethernet type of the packet in the VCPU queue, and set the data according to the correspondence between the priority of the pre-configured packet and the priority of the public queue. The packet message is delivered to the corresponding public queue;

Step 303: Read the packet information in the public queue according to the priority order of each public queue.

On the basis of the foregoing method, the correspondence between the received packet message and the VCPU queue may also be pre-configured. In addition, a specified number of public queues can be established in the shared memory, and the correspondence between the priority of the packet and the priority of the public queue is pre-configured.

In addition to considering the priority of each packet, you can also consider the flow limit value of packets corresponding to different types of Ethernet packets, that is, obtain each type of Ethernet packet read by each service. The stream limit value of the text, such that when one or more Ethertype packet flow reaches or exceeds the flow limit value, the delivery of the packet message to the corresponding public queue is stopped. The process of stopping the delivery of the packet message to the public queue is further divided into the following operations:

When the packet packet stream reaches or exceeds the corresponding stream limit value for the first time, the packet receiving flag is set to deny; when the packet packet stream reaches or exceeds the corresponding packet stream limit value for a set number of times, The receipt count is cleared and the receipt flag is reset to permit.

One of ordinary skill in the art will appreciate that all or a portion of the steps of the above-described embodiments can be implemented using a computer program flow, which can be stored in a computer readable storage medium, such as on a corresponding hardware platform (eg, The system, device, device, device, etc. are executed, and when executed, include one or a combination of the steps of the method embodiments.

Alternatively, all or part of the steps of the above embodiments may also be implemented using an integrated circuit. The steps may be separately fabricated into individual integrated circuit modules, or a plurality of modules or steps may be fabricated into a single integrated circuit module.

The devices/function modules/functional units in the above embodiments may be implemented by a general-purpose computing device, which may be centralized on a single computing device or distributed over a network of multiple computing devices.

When the device/function module/functional unit in the above embodiment is implemented in the form of a software function module and sold or used as a stand-alone product, it can be stored in a computer readable storage medium. The above mentioned computer readable storage medium may be a read only memory, a magnetic disk or an optical disk or the like.

Industrial applicability

The embodiment of the invention directly receives the packet by binding the hardware thread, so that the performance of the device collection is greatly improved, and the performance bottleneck of the previous packet collection is solved. At the same time, on the basis of ensuring the performance improvement, a priority queue is set according to the Ethernet type of the packet, and the important packet is preferentially processed through the priority queue. The speed limit mode is adopted to improve the security when a large number of attack packets flood into the device.

Claims

A device internal processor network port receiving device, the device comprising:

The receiving hardware module is configured to: deliver the received data packet message to the corresponding virtual central processing unit VCPU queue according to the pre-configuration;

The receiving thread module is configured to: receive the data packet in the VCPU queue bound to the packet, and determine the priority of the data packet according to the Ethernet type of the received data packet, according to the pre-configured data packet. Corresponding relationship between the priority of the packet and the priority of the public queue, and the received packet packet is delivered to the corresponding public queue;

The service module is configured to: read the packet information in the public queue according to the priority order of each public queue.
The apparatus of claim 1 further comprising:

The packet management module is configured to: pre-configure a correspondence between the data packet received by the receiving hardware module and the VCPU queue.
The device according to claim 2, wherein

The receiving management module is further configured to: establish a specified number of public queues in the shared memory, and pre-configure a correspondence between a priority of the data packet and a priority of the public queue.
The device according to claim 2 or 3, wherein

The receiving management module is further configured to: obtain, from each service module, a flow limit value of each type of Ethernet packet data;

The receiving thread module is further configured to: stop sending the data packet message to the corresponding public queue when the one or more Ethernet packet data flows reach or exceed the flow limit value.
The apparatus of claim 4 wherein said receiving thread module is configured to:

When the packet packet stream reaches or exceeds the stream limit value for the first time, the packet receiving flag is set to refuse to receive the packet; when the number of times the packet packet stream reaches or exceeds the stream limit value reaches a set number of times, The receiving count is cleared, and the receiving flag is reset to allow the receiving.
A device internal device network port receiving method, the method comprising:

Deliver the received packet message to the corresponding virtual CPU VCPU according to the pre-configuration In the queue;

Determine the priority of the packet according to the Ethernet type of the packet in the VCPU queue, and set the data according to the correspondence between the priority of the pre-configured packet and the priority of the public queue. The packet message is delivered to the corresponding public queue;

Packets in the public queue are read in the order of priority of each public queue.
The method of claim 6 further comprising:

Pre-configure the correspondence between received packet messages and VCPU queues.
The method of claim 7 further comprising:

Establish a specified number of public queues in the shared memory, and pre-configure the correspondence between the priority of the packet and the priority of the public queue.
The method of claim 7 or 8, further comprising:

Obtaining a flow limit value of each type of Ethernet packet message read by each service;

When one or more Ethertype packet flow reaches or exceeds the flow limit value, the delivery of the packet message to the corresponding public queue is stopped.
The method of claim 9, wherein the stopping the delivery of the packet message to the corresponding public queue when the one or more Ether type packet flow reaches or exceeds the flow limit value comprises:

When the packet packet stream reaches or exceeds the stream limit value for the first time, the packet receiving flag is set to refuse to receive the packet; when the number of times the packet packet stream reaches or exceeds the stream limit value reaches a set number of times, The receipt count is cleared and the receipt flag is reset to allow for collection.
A computer readable storage medium storing computer executable instructions for performing the method of any of claims 6-10.