CN102204183A - Message order-preserving processing method, order-preserving coprocessor and network equipment - Google Patents

Abstract

Embodiments of the present invention provide a packet order-preserving processing method, an order-preserving coprocessor, and a network device. The message order-preserving processing method includes: receiving a network message sent by a thread, wherein the network message includes a message identifier; determining an order-preserving flow corresponding to the message identifier, and the order-preserving flow and the network device's The packet processing sequence corresponding to a channel and included in the sequence-preserving flow is the sequence in which network packets enter the channel; according to the packet processing sequence, the sequence-preserving processing is performed on the network packets. In the embodiment of the present invention, in the case of asynchronous processing of messages by multiple threads, the order-preserving processing of messages can be performed.

Description

Packet order-preserving processing method, order-preserving coprocessor and network device

technical field

The embodiments of the present invention relate to the field of communications, and in particular, to a packet order-preserving processing method, an order-preserving coprocessor, and a network device.

Background technique

With the continuous development of technologies such as the Internet and mobile communication networks, the reliability of packet transmission also needs to be improved urgently, so as to minimize the possibility of packet disorder after network transmission.

The so-called packet out-of-order refers to the packet sequence sent by the sender, which is forwarded by several intermediate nodes, but does not reach the destination in the original order. Packet out-of-order will seriously affect the performance of some network applications. For example, transmission control protocol (Transmission Control Protocol, hereinafter referred to as: TCP) connection causes a large number of retransmissions due to out-of-order; real-time compression protocol (Compressed Real-Time Protocol, hereinafter referred to as: CRTP) business will frequently send FULL_HEADER due to out-of-order packets to synchronize the context, resulting in greatly reduced compression efficiency. Therefore, in order to improve network operation efficiency, existing communication standards require network devices, such as routers, to avoid packet disorder as much as possible.

However, most current network devices use multiple threads to process packets asynchronously and in parallel, and the speed at which different threads process packets is affected by many factors, which makes the processing of packet order preservation more complicated. Therefore, in the case of multi-threaded asynchronous processing of messages, how to process messages in order has become an urgent problem to be solved.

Contents of the invention

An embodiment of the present invention provides a message order-preserving processing method, an order-preserving coprocessor, and a network device, so as to perform order-preserving processing on messages in the case of multi-thread asynchronously processing messages.

An embodiment of the present invention provides a packet order-preserving processing method, including:

Receiving the network message sent by the thread, the network message includes a message identifier;

Determining an order-preserving flow corresponding to the message identifier, the order-preserving flow corresponding to a channel of the network device and the message processing sequence contained in the order-preserving flow is the order in which network messages enter the channel;

Perform sequence-preserving processing on the network packets according to the packet processing sequence.

An embodiment of the present invention provides an order-preserving coprocessor, including:

A receiving module, configured to receive a network message sent by a thread, where the network message includes a message identifier;

A determining module, configured to determine an order-preserving flow corresponding to the message identifier, the order-preserving flow corresponds to a channel of a network device and the message processing sequence contained in the order-preserving flow is that network messages enter the The sequence of the channels;

The sequence-preserving processing module is configured to perform sequence-preserving processing on the network packets according to the packet processing sequence.

An embodiment of the present invention provides a network device, including: a sequentially connected network processor unit, an order-preserving coprocessor, and a coprocessor, and the order-preserving coprocessor adopts the above-mentioned order-preserving coprocessor.

In the embodiment of the present invention, the corresponding order-preserving flow can be determined according to the message identifier contained in the network message sent by the thread, and the order-preserving process is performed on the network messages according to the order in which the network messages included in the order-preserving flow enter the channel . Therefore, even if the network packets are out of order due to the asynchronous parallel processing of each thread, the sending order of the network packets can be sorted to be consistent with the order in which the network packets enter the channel of the network device through the order-preserving flow, so that the network packets can They are sent sequentially in the order in which they entered the channel.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without any creative effort.

In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings that need to be used in the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description These are some embodiments of the present invention. For those skilled in the art, other drawings can also be obtained according to these drawings without any creative effort.

Fig. 1 is a flow chart of an embodiment of the message order preserving processing method of the present invention;

Fig. 2 is a flow chart of another embodiment of the message order preserving processing method of the present invention;

Fig. 3 is a kind of schematic structural diagram of the corresponding relationship between the message identifier and the order-preserving flow generated in the embodiment shown in Fig. 2;

Fig. 4 is a flow chart of another embodiment of the message order preserving processing method of the present invention;

FIG. 5 is a schematic diagram of a processing procedure of the embodiment shown in FIG. 4 in a scenario;

FIG. 6 is a schematic structural diagram of an embodiment of the order-preserving coprocessor of the present invention;

FIG. 7 is a schematic structural diagram of another embodiment of the order-preserving coprocessor of the present invention;

FIG. 8 is a schematic structural diagram of another embodiment of the order-preserving coprocessor of the present invention;

FIG. 9 is a schematic structural diagram of an embodiment of a network device according to the present invention.

Detailed ways

In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments It is a part of embodiments of the present invention, but not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without creative efforts fall within the protection scope of the present invention.

Fig. 1 is a flow chart of an embodiment of the message order preserving processing method of the present invention, as shown in Fig. 1, the method of this embodiment may include:

Step 101, receiving a network message sent by a thread, where the network message includes a message identifier.

For example, an order co-processor (Order Co-Processor, hereinafter referred to as: OCP) can receive a network message sent by a thread, and the network message can include a message identifier, and this embodiment does not limit the message identifier. Specific forms. The OCP in this embodiment can be implemented using application specific integrated circuits (Application Specific Integrated Circuits, hereinafter referred to as: ASIC) technology, and can be packaged together with network processor units (Network Processer Units, hereinafter referred to as: NPU).

Specifically, each channel of a network device, such as a router, can receive network packets sent by other network devices, and then the network packets of each channel can be assigned to multiple threads for parallel processing, and each thread can only hold and process a network message. When network packets enter each channel, there is a sequence, that is, when they are transmitted in each channel, the sequence of network packets is correct, but when the packets of each channel are assigned to multiple threads for processing, due to the The processing of threads is affected by many factors, and the network packets sent from each thread are out of order. Therefore, the network packets received by OCP are out of order.

Step 102: Determine the sequence-preserving flow corresponding to the packet identifier, the sequence-preserving flow corresponds to a channel of the network device, and the packet processing sequence contained in the sequence-preserving flow is the order in which network packets enter the channel sequence.

Specifically, this embodiment may include multiple order-preserving streams, and each order-preserving stream corresponds to a channel of the network device. Therefore, the number of order-preserving streams in this embodiment may be equal to the number of channels of the network device. For each sequence-preserving flow, the packet processing sequence contained therein may be the sequence in which network packets enter the channel. For example, taking the order-preserving flow n corresponding to the Nth channel as an example, the order-preserving flow may include the message identifiers of all network packets entering the Nth channel, and the message identifiers may be based on the corresponding network Packets are arranged in the order in which they enter the Nth channel.

For example, OCP can monitor the packets received by each channel. For each channel, OCP can record the sequence in which network packets are received. According to the sequence, OCP can generate the sequence-preserving flow corresponding to the channel. It should be noted that this embodiment does not limit how the OCP learns the order in which network packets enter channels, and those skilled in the art may use any means to monitor the order in which network packets enter channels and notify the OCP. Therefore, when using the sequence-preserving streams to perform sequence-preserving processing on the network packets, the out-of-sequence network packets can be restored to the sequence in which the network packets entered the channel.

In actual implementation, OCP can extract the message ID from each network message. For each message ID, OCP can determine the sequence-preserving flow where the message ID is located, so as to determine the channel corresponding to the message ID. , that is, the sequence-preserving flow corresponding to the network message or the channel through which the network message enters can be determined.

It can be seen that, after step 102 is executed, the OCP can map the out-of-order network packets received from each thread to each order-preserving flow.

Step 103: Perform sequence-preserving processing on the network packets according to the packet processing sequence.

For the network packets corresponding to each channel, OCP can process the network packets according to the packet processing sequence contained in the sequence-preserving flow corresponding to the channel, that is, according to the network packets recorded in the sequence-preserving flow The order in which the packets enter the channel is to send the network packets sequentially, so that the sending order of the network packets that have been processed in order is consistent with the order in which the network packets enter the channel.

In this embodiment, the corresponding order-preserving flow can be determined according to the message identifier included in the network message sent by the thread, and the order-preserving process is performed on the network messages according to the order in which the network messages included in the order-preserving flow enter the channel. Therefore, even if the network packets are out of order due to the asynchronous parallel processing of each thread, the sending order of the network packets can be sorted to be consistent with the order in which the network packets enter the channel of the network device through the order-preserving flow, so that the network packets can They are sent sequentially in the order in which they entered the channel.

Fig. 2 is a flow chart of another embodiment of the message order preserving processing method of the present invention, as shown in Fig. 2, the method of this embodiment may include:

Step 201 : Generate an order-preserving flow corresponding to the channel according to the order in which network packets enter the channel, and generate a correspondence between the message identifiers of the network packets and the order-preserving flow.

OCP can generate an order-preserving flow corresponding to the channel according to the order in which the network messages enter the channel when network messages enter the channel, and generate a link between the message identifier of the network message and the order-preserving flow Corresponding relationship, for each sequence-preserving flow, the sequence in which network packets enter the channel is the packet processing sequence. This embodiment does not limit the technical means for the OCP to learn the order in which network packets enter the channel. For example, the sequence-preserving flow can adopt a linked list structure. FIG. 3 is a schematic structural diagram of the corresponding relationship between the message identifier and the sequence-preserving flow generated in the embodiment shown in FIG. 2. As shown in FIG. Sequence flow linked list (Link) 0～sequence-preserving flow link n correspond to n+1 channels of the network device respectively, Valid indicates that the corresponding sequence-preserving flow link is valid, head is the head of the table, tail is the tail of the table, and next is the next node Pointer, PKT_ID0~PKT_IDm are the identifiers of network packets received on each channel of the network device. For example, the corresponding relationship indicated by the arrow in Figure 3 is the corresponding relationship between the sequence-preserving flow link1 and PKT_ID, where the header of the sequence-preserving flow link1 is at PKT_ID1, and the next pointer at PKT_ID1 points to PKT_ID3, PKT_ID3 The next pointer at point points to PKT_ID5, and PKT_ID5 is also the end of the order-preserving stream link1, and PKT_ID1, PKT_ID3, and PKT_ID5 have a sequence when entering the channel corresponding to the order-preserving stream link1, and so on, PKT_ID0～PKT_IDm can correspond to each order-preserving stream link.

Step 202, receiving network packets sent by each thread, where the network packets include packet identifiers.

Step 203: Buffer the network packets sent by each thread, and generate an order-preserving request queue including an order-preserving processing order, the order-preserving processing order being the sequence in which each thread sends network packets.

Specifically, each thread may sequentially send an order-preserving request (hereinafter referred to as REQ) to the OCP, and the REQ includes the data content of the network message and the like. Then the OCP can cache the REQ sent by each thread, for example, cache the REQ in BUFFER. Moreover, the OCP can also generate an order-preserving request queue, which includes an order-preserving processing order, and the order-preserving processing order is the order in which each thread sends network packets.

For example, the order-preserving processing sequence may be reflected by a request descriptor (hereinafter referred to as: REQ-ID). Table 1 is a description of REQ-ID.

Table 1

NPU_ID THREAD_ID PKT_ID

Wherein, NPU_ID is the identifier of the parallel processor, and THREAD_ID is the thread identifier of each NPU. Each REQ_ID in the order-preserving processing queue can index one REQ, and the number of REQ_IDs is equal to the product of the number of co-processors (Co-process, COP for short) and the number of threads.

Table 2 is a description of REQ cached in BUFFER.

Table 2

OP ADDR CTRL DATA

Among them, OP represents the operation type of REQ, for example, the type of read operation is READ, the type of write operation is WRITE, and the type of read and write operation is WRITE&READ; ADDR represents the peripheral address accessed by REQ; CTRL is used to describe the data of REQ Control information such as length, order preservation or not; DATA indicates the request data carried by REQ and sent to COP.

There is a corresponding relationship between the REQ cached in the BUFFER, such as in Table 2, and the REQ_ID in the order-preserving request queue, such as shown in Table 1.

Step 204: Determine the sequence-preserving flow corresponding to the packet identifier of each network packet in sequence according to the sequence-preserving processing sequence.

The OCP can sequentially process corresponding REQs according to the sequence-preserving processing sequence in the sequence-preserving request queue, that is, process network packets sequentially. When specifically processing network packets, the OCP may determine the sequence-preserving flow corresponding to each network packet according to, for example, the correspondence between PKT_ID and sequence-preserving flow in Table 1.

Step 205: Send the network packets corresponding to each channel to the data bus sequentially according to the packet processing sequence of each sequence-preserving flow.

For the network packets corresponding to each sequence-preserving flow, each network packet needs to be sent to the data bus BUS sequentially according to the processing sequence of the packets in the sequence-preserving flow.

Specifically, for each sequence-preserving flow, if the PKT_ID of the network packet is at the head of the link list of the sequence-preserving flow, the OCP can send the network packet to the bus BUS, and then the OCP can send the PKT_ID of the sequence-preserving flow The next PKT_ID is set as the head of the linked list, so that the next sequence-preserving processing is for the next network message. If the PKT_ID of the network packet is not at the head of the sequence-preserving flow list, OCP can process the next REQ cached in BUFFER according to the REQ_ID in the sequence-preserving request queue, while the current REQ needs to wait for the next one polling.

In this embodiment, when network messages enter each channel, the OCP can record the order in which the network messages enter for each channel, and the OCP can cache the network messages sent by each thread, and according to the network messages sent by each thread The packet identifier contained in the , determines the corresponding sequence-preserving flow, and performs sequence-preserving processing on the network packets according to the order in which the network packets included in the sequence-preserving flow enter the channel. Therefore, even if the network packets are out of order due to the asynchronous parallel processing of each thread, the sending order of the network packets can be sorted to be consistent with the order in which the network packets enter the channel of the network device through the order-preserving flow, so that the network packets can They are sent sequentially in the order in which they entered the channel.

For the above-mentioned embodiment shown in FIG. 1 or FIG. 2 , it can not only perform order-preserving processing on network packets sent by threads on the NPU side, but also perform order-preserving processing on network packets sent by threads on the COP side.

In an application scenario, the network message sent by the thread on the NPU side requires the thread on the COP side to send a corresponding response message. Therefore, the OCP can perform order-preserving processing on the network packets sent by the thread on the NPU side, and when the thread on the COP side sends a corresponding response message, the OCP can perform order-preserving processing on the response message sent by the thread on the COP side again.

Fig. 4 is a flow chart of another embodiment of the message order-preserving processing method of the present invention, and Fig. 5 is a schematic diagram of the processing process of the embodiment shown in Fig. 4 in a scenario, as shown in Fig. 4 and Fig. 5 , without loss of generality Specifically, this embodiment only shows three threads, namely thread A, thread B and thread C, and the PKT_ID of the network message currently held by thread A is 0, and the PKT_ID of the network message currently held by thread B is 0. is 1, the PKT_ID of the network packet currently held by thread C is 2, and these three network packets belong to the same channel. The corresponding message processing order in the sequence-preserving flow corresponding to this channel is 0, 1, 2, that is, the order in which network messages enter the channel is 0, 1, 2, and the order in which asynchronous threads send network messages to OCP The order is thread B, thread A, thread C. According to the scenario, this embodiment may include:

Step 401, thread B, thread A and thread C send REQ to OCP in sequence.

Thread B, thread A, and thread C send REQs to the OCP in sequence, and each REQ can use the description shown in Table 2.

Step 402, the REQ BUFFER in the OCP caches the REQ, and generates a REQ_ID queue (hereinafter referred to as: REQ_ID Queue).

In this embodiment, the order-preserving processing order in the REQ_ID queue is 1, 0, 2.

Step 403, the scheduling unit in the OCP (hereinafter referred to as: Schedule) schedules a REQ_ID from the REQ_ID Queue according to the order-preserving processing order.

Step 404, Schedule judges whether the REQ_ID is at the head of the linked list of the sequence-preserving flow, if so, executes step 405, otherwise executes step 403.

Step 405, Schedule sends REQ authorization to REQ BUFFER, and sets the next REQ_ID in the sequence-preserving stream as the head of the linked list.

Step 406, the REQ BUFFER sends the REQ corresponding to the REQ_ID to the REQ First In First Out (hereinafter referred to as: FIFO) queue on the BUS.

Specifically, Schedule can schedule REQ_ID "1" from the REQ_ID Queue, and then Schedule can compare the scheduled REQ_ID "1" with the REQ_ID contained in the head of the linked list of the sequence-preserving flow, thereby judging the scheduled REQ_ID "1" Whether it is at the head of the linked list of the order-preserving flow. In this embodiment, the current linked list head of the sequence-preserving flow is "0", therefore, the REQ corresponding to the REQ_ID "1" scheduled by the Schedule cannot be sent to the BUS FIFO queue. At this point, Schedule can schedule REQ_ID "0" from the REQ_ID Queue, and then Schedule can compare the scheduled REQ_ID "0" with the REQ_ID contained in the sequence-preserving link list header to determine whether the scheduled REQ_ID "1" is The head of the linked list in the order-preserving flow. It can be seen from the judgment that REQ_ID "0" is at the head of the linked list of the sequence-preserving flow. Therefore, Schedule can send REQ authorization to REQ BUFFER, so that REQ BUFFER sends REQ corresponding to REQ_ID "0" to BUS FIFO, and Schedule can send The next REQ_ID "1" in the sequence-preserving stream is set as the head of the linked list. Next, the Schedule can schedule REQ_ID "2" from the REQ_ID Queue, and then the Schedule can compare the scheduled REQ_ID "2" with the REQ_ID contained in the current link list header of the sequence-preserving flow, so as to determine whether the scheduled REQ_ID "2" is The head of the linked list in the order-preserving flow. Since the REQ_ID contained in the current linked list header of the sequence-preserving flow is "1", the REQ corresponding to the REQ_ID "2' scheduled by Schedule cannot be sent to the BUS FIFO queue. At this time, Schedule can schedule REQ_ID "1" from REQ_IDQueue ", and then, Schedule can compare the scheduled REQ_ID "1" with the REQ_ID contained in the current link list header of the sequence-preserving flow, so as to judge whether the scheduled REQ_ID "1" is in the link list header of the sequence-preserving flow. Through the judgment, it can be seen that, REQ_ID "1" is at the head of the current linked list of the sequence-preserving flow. Therefore, Schedule can send REQ authorization to REQ BUFFER, so that REQ BUFFER sends the REQ corresponding to REQ_ID "1" to the BUS FIFO. Moreover, Schedule can send the sequence-preserving flow The next REQ_ID "2" in the list is set as the head of the linked list. Schedule can compare the scheduled REQ_ID "2" with the REQ_ID contained in the current linked list header of the sequence-preserving stream, so as to judge whether the scheduled REQ_ID "2" is in the sequence-preserving The head of the linked list of the stream. It can be seen from the judgment that REQ_ID "2" is at the current linked list head of the sequence-preserving stream. Therefore, Schedule can send REQ authorization to REQ BUFFER, so that REQ BUFFER sends the REQ corresponding to REQ_ID "2" to the BUS FIFO .

For more network packets, the above-mentioned processing procedure can be adopted, and details will not be repeated here.

Through the above process, the REQ sequence in the REQ FIFO queue is REQA, REQB, REQC.

Step 407, REQA, REQB, and REQC access the COP in sequence.

For the processing process of the COP feedback response message, it can process the out-of-order response message (hereinafter referred to as: RSP) in Figure 5, that is, RSPC, RSPA, and RSPB are sequence-preserved as RSPA by adopting steps similar to the above steps 401-407 , RSPB, RSPC, which are sent to the corresponding thread. Wherein, the description structures of RSP_ID and RSP are correspondingly similar to the description structures of REQ_ID and REQ, and will not be repeated here.

In this embodiment, on the basis of the above embodiment shown in FIG. 1 or FIG. 2 , further, order-preserving processing may be performed on the response message sent by the COP.

FIG. 6 is a schematic structural diagram of an embodiment of the order-preserving coprocessor of the present invention. As shown in FIG. It is used to receive a network message sent by a thread, the network message includes a message identifier; the determination module 12 is used to determine an order-preserving flow corresponding to the message identifier, and the order-preserving flow is connected to a channel of the network device Correspondingly, the packet processing sequence included in the sequence-preserving flow is the sequence in which network packets enter the channel; the sequence-preserving processing module 13 is used to perform sequence-preserving on the network packets according to the packet processing sequence deal with.

The OCP of this embodiment can be used to execute the method of the method embodiment shown in FIG. 1 , and its implementation principle and technical effect are similar, and will not be repeated here.

FIG. 7 is a schematic structural diagram of another embodiment of the order-preserving coprocessor of the present invention. As shown in FIG. 7 , the OCP of this embodiment may further include: a generation processing module 14 and a cache memory on the basis of the OCP shown in FIG. 6 The processing module 15, wherein, the generation processing module 14 is used to generate the sequence-preserving flow corresponding to the channel according to the order in which the network messages enter the channel, and generate the message identifier of the network message and the order-preserving flow Correspondence between streams; the determination module 12 is specifically configured to determine the sequence-preserving stream corresponding to the network identifier according to the correspondence generated by the generation processing module; a cache processing module 15, the cache processing module 15 is used for caching The network messages sent by each thread, and generate an order-preserving request queue that includes an order-preserving processing order, and the order-preserving processing order is the order in which each thread sends network messages; the determination module 12 can also be specifically used to The sequence processing sequence is determined sequentially to determine the sequence-preserving flow corresponding to the packet identifier of each network packet. The order-preserving processing module 13 includes a judging unit 131 and a processing unit 132, wherein the judging unit 131 is used to judge whether the message identifier of the network message is at the head of the linked list of the order-preserving flow; the processing unit 132 is used for if the The message identifier of the network message is at the head of the linked list of the order-preserving flow, then the network message is sent, and the next message identifier in the order-preserving flow is set as the head of the linked list; if the network message If the packet identifier is not at the head of the linked list of the sequence-preserving flow, then perform sequence-preserving processing on the next network packet according to the sequence-preserving processing sequence.

The OCP of this embodiment can be used to execute the method of the method embodiment shown in FIG. 2 , and its implementation principle and technical effect are similar, and will not be repeated here.

FIG. 8 is a schematic structural diagram of another embodiment of the order-preserving coprocessor of the present invention. As shown in FIG. 8, the OCP of this embodiment is based on the OCP shown in FIG. 6. Further, the receiving module 11 may include: first The receiving unit 111 and the second receiving unit 112, wherein the first receiving unit 111 is used to receive the network message sent by the thread on the network processor unit side; the second receiving unit 112 is used to receive the network packet sent by the thread on the coprocessor side. message.

The OCP of this embodiment can be used to execute the method of the method embodiment shown in FIG. 4 , and its implementation principle and technical effect are similar, and will not be repeated here.

It should be noted that the first receiving unit 111 and the second receiving unit 112 in this embodiment can also choose one of them, which can execute the method of the embodiment shown in FIG. 1 or FIG. 2 .

In the embodiments shown in FIGS. 6-8 , the order-preserving coprocessors can all be implemented using ASIC technology, and those skilled in the art can design specific circuit diagrams by themselves according to the concepts of the embodiments of the present invention, which is not limited in this embodiment.

FIG. 9 is a schematic structural diagram of an embodiment of the network device of the present invention. As shown in FIG. 9, the network device of this embodiment may include a network processor unit 1, an order-preserving coprocessor 2, and a coprocessor 3 connected in sequence, wherein The order-preserving coprocessor 2 can adopt the structure shown in any one of the embodiments in FIGS. 6 to 8, which can correspondingly execute the method of the method embodiment shown in FIG. I won't repeat them here.

Those of ordinary skill in the art can understand that all or part of the steps for realizing the above-mentioned method embodiments can be completed by hardware related to program instructions, and the aforementioned program can be stored in a computer-readable storage medium. When the program is executed, the It includes the steps of the above method embodiments; and the aforementioned storage medium includes: ROM, RAM, magnetic disk or optical disk and other various media that can store program codes.

Finally, it should be noted that: the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: it can still be Modifications are made to the technical solutions described in the foregoing embodiments, or equivalent replacements are made to some of the technical features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the various embodiments of the present invention.

Claims (12)

1. a packet order preserving processing method is characterized in that, comprising:

The network message that receiving thread sends comprises message identification in the described network message;

Determine the order-preserving stream corresponding with described message identification, described order-preserving stream is the sequencing that network message enters described passage with the message processing sequence that comprises during the corresponding and described order-preserving of a passage of the network equipment is flowed;

According to described message processing sequence, described network message is carried out order-preserving handle.

2. method according to claim 1 is characterized in that, before the network message that described receiving thread sends, also comprises:

According to the sequencing of network message admission passage, generate the order-preserving stream corresponding, and generate the message identification of described network message and the corresponding relation between the described order-preserving stream with this passage;

The described definite order-preserving corresponding with described message identification stream comprises:

According to described corresponding relation, determine the order-preserving stream of described network identity correspondence.

3. method according to claim 1 and 2 is characterized in that, determines also to comprise before the order-preserving stream corresponding with described message identification:

The network message that each thread of buffer memory sends, and generate the order-preserving request queue that comprises the order-preserving processing sequence, described order-preserving processing sequence sends the sequencing of network message for each thread;

The described definite order-preserving corresponding with described message identification stream comprises:

According to described order-preserving processing sequence, determine the order-preserving stream of the message identification correspondence of each network message successively.

4. method according to claim 3 is characterized in that, and is described according to the message processing sequence in the described order-preserving stream, described network message carried out order-preserving handle, and comprising:

If the message identification of described network message is in the linked list head of described order-preserving stream, then sends described network message, and the next message identification in the described order-preserving stream is made as linked list head;

If the message identification of described network message is not in the linked list head of described order-preserving stream, then next network message is carried out order-preserving and handle according to described order-preserving processing sequence.

5. method according to claim 1 and 2 is characterized in that, the network message that described receiving thread sends comprises:

Receive the network message of the thread transmission of network processor unit side; And/or,

Receive the network message of the thread transmission of coprocessor side.

6. an order-preserving coprocessor is characterized in that, comprising:

Receiver module is used for the network message that receiving thread sends, and comprises message identification in the described network message;

Determination module is used for determining the order-preserving stream corresponding with described message identification, and described order-preserving stream is the sequencing that network message enters described passage with the message processing sequence that comprises during the corresponding and described order-preserving of a passage of the network equipment is flowed;

The order-preserving processing module is used for according to described message processing sequence, described network message is carried out order-preserving handle.

7. order-preserving coprocessor according to claim 6 is characterized in that, also comprises:

Generate processing module, be used for sequencing, generate the order-preserving stream corresponding, and generate the message identification of described network message and the corresponding relation between the described order-preserving stream with this passage according to described network message admission passage;

Described determination module specifically is used for the described corresponding relation according to described generation processing module generation, determines the order-preserving stream of described network identity correspondence.

8. according to claim 6 or 7 described order-preserving coprocessors, it is characterized in that, also comprise:

Buffer process module is used for the network message that each thread of buffer memory sends, and generates the order-preserving request queue that comprises the order-preserving processing sequence, and described order-preserving processing sequence sends the sequencing of network message for each thread;

Described determination module specifically is used for according to described order-preserving processing sequence, determines the order-preserving stream of the message identification correspondence of each network message successively.

9. according to claim 6 or 7 described order-preserving coprocessors, it is characterized in that described order-preserving processing module comprises:

Judging unit is used to judge whether the message identification of described network message is in the linked list head of described order-preserving stream;

Processing unit is used for then sending described network message if the message identification of described network message is in the linked list head of described order-preserving stream, and the next message identification in the described order-preserving stream is made as linked list head; If the message identification of described network message is not in the linked list head of described order-preserving stream, then next network message is carried out order-preserving and handle according to described order-preserving processing sequence.

10. according to claim 6 or 7 described order-preserving coprocessors, it is characterized in that described receiver module comprises following at least one unit:

First receiving element is used to receive the network message that the thread of network processor unit side sends;

Second receiving element is used to receive the network message that the thread of coprocessor side sends.

11., it is characterized in that described order-preserving coprocessor adopts application-specific integrated circuit (ASIC) to realize according to claim 6 or 7 described order-preserving coprocessors.

12. a network equipment is characterized in that, comprising: the network processor unit of Lian Jieing, order-preserving coprocessor and coprocessor successively, described order-preserving coprocessor adopts the described order-preserving coprocessor of the arbitrary claim of claim 6～11.

Images