CN104951385A - Channel health state recording device of dynamic reconfigurable bus monitoring system - Google Patents

Abstract

A channel health state recording device in a dynamic reconfigurable bus monitoring system, characterized in that it includes a channel detection and identification module, a row and column mapping module, a three-dimensional channel table memory, a channel data organization module and a communication data buffer memory, using row nodes, The channel health status storage table in three dimensions of column nodes and channels realizes the storage and management of channel health status between all nodes on the dynamically reconfigurable bus.

Description

Channel health state recording device in dynamic reconfigurable bus monitoring system

technical field

The invention relates to a channel health state recording device in an embedded system bus monitoring system, in particular to a channel health state recording device in a dynamically reconfigurable bus.

Background technique

The dynamic reconfigurable bus (UM-BUS) is a high-speed serial bus based on the M-LVDS bus topology defined for high-reliability embedded applications. As shown in Figure 1, the bus topology is adopted. It supports direct interconnection of multiple nodes, and can use up to 32 channels to transmit data concurrently. If some channels fail, the bus controller can detect them in real time, dynamically allocate data to the remaining effective channels for transmission, realize dynamic reconstruction, and dynamically fault-tolerant communication failures.

The communication protocol model of the UM-BUS bus is shown in Figure 2, which is the processing layer, data link layer, and physical layer from top to bottom. The processing layer is responsible for the management of the entire bus, protocol encapsulation and conversion of the upper layer application interface. The data link layer is divided into two parts, the transmission sublayer and the MAC sublayer. The transmission sublayer groups and dynamically reconstructs data according to the existing effective lines. Information about valid lines is provided by the MAC sublayer through link detection. The physical layer is the bottom layer of the protocol. It provides transmission media and interconnection equipment for data communication, realizes the physical connection of the network, completes serial-to-parallel conversion, 8b/10b codec, clock synchronization and other functions, and provides reliable communication for the bus. Base.

The UM-BUS bus adopts the communication mode of master-slave command response, and exchanges information in the form of data packets. The nodes on the bus can be divided into master nodes, slave nodes and monitoring nodes according to different functions. A communication process can only be initiated by the master node and responded by the slave nodes. The monitoring nodes are used to monitor the communication process on the bus. As shown in Figure 3, during data communication, at the sending end, the processing layer obtains data from the upper-layer interface and stores it in the data buffer, and at the transmission sub-layer, dynamically and evenly distributes the data packets to the effective line information provided by the MAC sub-layer. On the channel, after the packet data is packaged at the physical layer, it is encoded into a bit stream by 8b/10b and sent to the link. At the receiving end, the physical layer performs clock synchronization, 8b/10b decoding, and serial-to-parallel conversion on the received data, unpacks the channel data, and then dynamically organizes the data at the transmission sublayer according to the effective line information provided by the MAC sublayer And stored in the data buffer, and finally handed over to the application layer by the processing layer for processing.

The UM-BUS bus adopts the "command-response-confirm" three-stage channel fault detection algorithm. After reset or when a fault occurs, the master node starts the process of checking the health status of the bus channel from the slave node. As shown in Figure 4, the channel detection process is divided into three stages in terms of time: ① detection command sending stage, the master node sends detection command packets from all bus channels to the detected slave nodes, and starts the channel detection process; ② detection status In the response phase, the detected slave node returns a detection response packet to the master node that initiated the detection from all the bus channels that received the detection command; Send a detection confirmation packet to the detected slave node. For the master node, the channel that can receive the detection response packet is a healthy channel; for the slave node, the channel that can receive the detection confirmation packet is a healthy channel. If the master node does not receive the detection response packet from the slave node within the specified time, or the slave node does not receive the detection confirmation packet from the master node, it is considered that the detection process has timed out and the channel is unavailable. After the channel detection is completed, the master and slave nodes participating in the detection process update their respective channel health status tables according to the reception of detection information packets.

The health status management in the UM-BUS bus master and slave nodes is based on the node-to-node communication requirements, and the two-dimensional form shown in Table 1 can be used to provide the channel availability between the node and any node to the transmission sublayer. In Table 1, each row corresponds to a bus node, indicating the availability of the channel between it and the node, each column corresponds to a communication channel, and stores the health status X of the channel. X is 0, indicating that the channel is unavailable, and 1, indicating that the channel is available. .

Table 1 Node channel health status table

The UM-BUS bus monitoring node is connected to the bus according to the topology shown in Figure 1, monitors and records the communication, detection and other activities on the bus, and transmits the recorded bus activity information to the background processor for classification and analysis, so that Bus users can fully grasp the working conditions of each node device on the bus, which is convenient for debugging and problem analysis of the bus application system. The bus monitoring node only monitors and records the activities on the bus, but cannot send information to the bus. The core of the bus monitoring node is composed of a bus monitoring system. Its function is to monitor all channels on the bus, collect and record communication information such as commands, status, data and time of various communication processes, detection processes and other active processes on the bus. Work status data.

During the working process of the bus monitoring system, it is necessary to organize the communication data on the bus channel and extract the original communication data according to the monitoring results of the bus channel and the channel health status table between the master and slave nodes involved in the bus communication process. Packets, which record communication commands, status and data. To complete the bus channel data organization, the bus monitoring system must be able to obtain the channel health status table between the nodes of the two communication parties, but the bus monitoring system cannot obtain their respective channel health status from the master and slave nodes involved in the communication process during the monitoring process information. This requires the bus monitoring system to be able to monitor the bus channel detection process. According to the monitoring results, the channel health status storage table between the nodes on the bus is established by itself, and the bus channel health status is recorded for recovery of bus communication information during the monitoring process. Use when packing. At the same time, because the channel health status of the UM-BUS bus is based on node-to-node, this makes it impossible to use the two-dimensional channel health status table form in the master and slave communication nodes in the bus monitoring system, otherwise, the monitoring system cannot satisfy any two The organization needs to listen to the communication packets between communication nodes.

Contents of the invention

The purpose of the present invention is to design a device for recording channel health status between bus nodes in a dynamic reconfigurable bus monitoring system, which can record the channel health status between all nodes on the dynamic reconfigurable bus, and can It is used for data organization and recovery of communication packets during monitoring.

To achieve the above object, the technical solution adopted in the present invention is:

A channel health state recording device for a dynamic reconfigurable bus monitoring system, characterized in that: the channel health state recording device includes a channel detection and identification module, a row and column mapping module, a three-dimensional channel table memory, a channel data organization module and a communication data buffer memory.

Further, the channel detection identification module is used to monitor the channel detection process of the dynamic reconfigurable bus. When a channel detection process is monitored, the source node number and target node number of channel detection are generated according to the monitored detection information packet. Number, channel health status word and channel detection completion signal are sent to the row-column mapping module.

Further, the row-column mapping module is used to generate the access address and control signal of the three-dimensional channel table memory. When receiving the channel detection completion signal sent by the channel detection and identification module, the source node number sent by the channel detection and identification module is used as The row node address of the left port, the target node number is used as the column node address of the left port, and the channel health status word is written from the left port into the three-dimensional channel table memory; the target node number sent by the channel detection and identification module is used as the row node of the right port The address and the source node number are used as the column node address of the right port, and the channel health status word is simultaneously written into the three-dimensional channel table memory from the right port.

Further, the three-dimensional channel table memory is a dual-port memory, which is used to store the channel health status between all nodes on the dynamically reconfigurable bus, and the effective data width of the left and right ports and the dynamic reconfigurable The number of channels of the structural bus is the same. The effective address bits of the left and right ports are equally divided into two parts: row node address and column node address. Each data word represents the channel health status between the corresponding row node and column node. Each Bits correspond to a channel of the dynamically reconfigurable bus, forming a channel health state storage table in three dimensions of row nodes, column nodes and channels.

Further, the communication data buffer memory is used for storing detected bus communication information and bus detection information in the dynamically reconfigurable bus monitoring system.

Further, the channel data organization module is used to monitor the dynamic reconfigurable bus, and send the source node number and target node number of the monitored communication packet to the row and column mapping module whenever a bus communication process is monitored , the row and column mapping module uses the source node number and the target node number as the row node address and column node address, reads the channel health status word from any port of the three-dimensional channel table memory, and sends it to the channel data organization module; the channel data organization module is based on The channel health state word reads the communication data from the available bus channel, composes the communication information packet, adds information including the time code, and stores it in the communication data buffer memory.

Further, the channel detection and identification module is also used to compose a detection information packet together with information including time code and store the detection information in the communication data buffer memory when a channel detection process is monitored.

The channel health state recording device in the dynamic reconfigurable bus monitoring system realized by the present invention solves the problem that the dynamic reconfigurable bus monitoring system collects, records and uses the channel health status between two bus nodes, and provides a monitoring system for the monitoring system. Real-time and complete monitoring records of bus communication activities provide technical support.

Description of drawings

Fig. 1 is a topological structure diagram of the UM-BUS bus;

Fig. 2 is a hierarchical model diagram of the UM-BUS bus protocol;

Fig. 3 is a schematic diagram of UM-BUS bus data transmission process and data path;

Fig. 4 is a schematic diagram of a UM-BUS bus channel fault detection method;

5 is a structural diagram of a channel health state recording device in a dynamic reconfigurable bus monitoring system;

Fig. 6 is a schematic diagram of a three-dimensional storage space for channel health status.

Detailed ways

As shown in Figure 1, the dynamic reconfigurable bus (UM-BUS) adopts a multi-channel intelligent dynamic redundant bus topology based on M-LVDS, and supports direct interconnection of up to 30 communication nodes, and the node numbers are 1 to 30 , numbers 0 and 31 are reserved; use up to 32 channels to transmit data concurrently, and each channel is numbered from 0 to 31; when two nodes communicate, if a channel fails, it can be automatically shielded through channel dynamic redundancy and fault reconstruction technology Faulty channels, data exchange is only performed on healthy channels.

UM-BUS adopts the master-slave response mode, and exchanges information between nodes in the form of data packets. Nodes on the bus can be divided into master nodes, slave nodes and monitoring nodes according to different functions. A communication process can only be started by the master node by sending a command packet to the slave node, and the slave node returns a response packet to end. Monitoring nodes are used to monitor the communication process on the bus.

As shown in Figure 3, during data communication, at the sending end, the processing layer obtains data from the upper-layer interface and stores it in the data buffer, and at the transmission sub-layer, dynamically and evenly distributes the data packets to the effective line information provided by the MAC sub-layer. On the channel, the packet data is packaged at the physical layer and then encoded into a bit stream by 8b/10b and sent to the link. At the receiving end, the physical layer performs clock synchronization, 8b/10b decoding, and serial-to-parallel conversion on the received data to unpack the channel data, and then at the transmission sublayer, according to the channel health status information provided by the MAC sublayer, all valid channels The data on the network is dynamically organized to form a communication data packet (command packet or response packet), which is stored in the data buffer and finally handed over to the application layer by the processing layer for processing.

After reset, all master nodes on the UM-BUS will take turns to use the "command-response-confirm" three-stage channel fault detection algorithm to check the channel health status of all other nodes on the bus one by one. As shown in Figure 4, the channel detection process is divided into three stages in terms of time: ① detection command sending stage, the master node sends detection command packets from all bus channels to the detected slave nodes, and starts the channel detection process; ② detection status In the response phase, the detected slave node returns a detection response packet to the master node that initiated the detection from all the bus channels that received the detection command; Send a detection confirmation packet to the detected slave node. For the master node, the channel that can receive the detection response packet is a healthy channel; for the slave node, the channel that can receive the detection confirmation packet is a healthy channel. If the master node does not receive the detection response packet from the slave node within the specified time, or the slave node does not receive the detection confirmation packet from the master node, it is considered that the detection process has timed out, and the channel is a faulty channel and cannot be used. After the channel detection is completed, the master and slave nodes participating in the detection process update their respective channel health status tables according to the reception of detection information packets.

The UM-BUS bus monitoring node is connected to the bus, monitors and records the communication, detection and other activities on the bus, and transmits the recorded bus activity information to the background processor for classification and analysis. The bus monitoring node only monitors and records the activities on the bus, but cannot send information to the bus. The bus monitoring node is usually composed of a bus monitoring system, a communication data buffer memory, a transmission control module and a background processing computer. The bus monitoring system is responsible for monitoring all channels on the bus, collecting communication information and working status data such as commands, status, data and time of various communication processes, detection processes and other active processes on the bus, and recording these information into the communication data buffer in memory. The transmission control module transmits the record data in the communication data buffer memory to the background processing computer under the control of the background processing computer. The background processing computer completes the classified analysis and storage of the monitoring data for users to use.

The bus monitoring system is the core component of the bus monitoring node, and usually adopts the organizational structure shown in Figure 3, but only realizes the one-way data path from the physical channel to the processing layer. The channel data flow is received by the physical layer, after clock synchronization, 8b/10b decoding, and serial-to-parallel conversion, the channel data is unpacked, and then in the transmission sublayer, according to the channel health status information provided by the MAC sublayer, the The data is dynamically organized to form a communication data packet (command packet or response packet), and stored in the data buffer of the data buffer sublayer, and finally the processing layer writes the communication data packet together with other information such as time code into the communication data buffer memory , waiting to be sent to the background processing computer.

Based on the operating principle of the above-mentioned UM-BUS bus and bus monitoring nodes, a specific implementation of the channel health state recording device of the present invention is as follows:

A channel health state recording device as shown in Figure 5 is set in the bus monitoring system, including a channel detection and identification module, a row and column mapping module, a three-dimensional channel table memory, a channel data organization module and a communication data buffer memory. The channel detection and identification module monitors the channel detection process on the bus at the MAC sublayer, and generates source node numbers, target node numbers and channel health status words for channel detection. The row-column mapping module reads and writes the three-dimensional channel table memory according to the request of the channel detection and identification module and the channel data organization module. The channel data organization module is the specific implementation of the transmission sublayer, data buffer sublayer and processing layer, responsible for the dynamic organization and transmission control of bus communication data packets. The communication data buffer memory is used to store the monitored bus communication information and bus detection information in the dynamic reconfigurable bus monitoring system. When the background processor needs it, it will take the monitored bus activity information from the communication data buffer memory.

The three-dimensional channel table memory is a dual-port memory, which is used to store the channel health status between all nodes on the dynamically reconfigurable bus. As shown in Figure 5, the effective data width of the left and right ports of the three-dimensional channel table memory is the same as the number of channels of the dynamic reconfigurable bus, and the effective address bits of the left and right ports are equally divided into row node address and column node address. Each data word represents the channel health status between the corresponding row node and column node, and each bit corresponds to a channel of the dynamic reconfigurable bus, which constitutes the channel health status in three dimensions of row node, column node and channel Store the table.

The channel health status recording device adopts the following methods and steps to obtain and record the channel health status between all nodes on the bus for use when the transmission sublayer dynamically organizes communication data packets:

(1) The channel detection and identification module is in the MAC sublayer, and monitors the channel data packets output by each channel MAC module (MAC1~MAC32) in real time, and when a detection information packet (detection command packet, detection response packet or detection confirmation packet) is found , enter the corresponding stage of the channel detection process, start the timer to start the tracking of the bus channel detection process, and extract the source node number and target node number of the channel detection from the detection information packet. When the timing of the bus channel detection result confirmation phase ends, the channel detection and identification module determines the health status of each channel according to whether the detection confirmation packet is monitored, and the channel health status word that constitutes this channel detection, the channels that can monitor the detection confirmation packet are A healthy channel is available, otherwise a faulty channel. After completing the tracking of the bus channel detection process, the channel detection and identification module sends the source node number, target node number and channel health status word of this detection to the row and column mapping module, and generates a channel detection completion signal to inform the row and column mapping module of a channel detection After the process is completed, the row-column mapping module is requested to write the channel health status word into the three-dimensional channel table memory. When the channel detection and identification module completes the channel detection, it also needs to combine the detection information (including source node number, target node number, channel health status word, three detection information packet monitoring conditions, etc.) together with time code and other information to form a detection information packet , write to the communication data buffer memory.

(2) When the row and column mapping module receives the channel detection completion signal sent by the channel detection and identification module, the source node number sent by the channel detection and identification module is used as the row node address of the left port, and the target node number is used as the column node address of the left port , write the channel health status word from the left port into the three-dimensional channel table memory; at the same time, use the target node number sent by the channel detection and identification module as the row node address of the right port, and the source node number as the column node address of the right port, write the channel health The status word is also written into the three-dimensional channel table memory from the right port.

(3) The channel data organization module monitors the dynamic reconfigurable bus by monitoring the channel data packets output by each channel MAC module (MAC1-MAC32) in real time. Whenever the channel MAC module receives a data packet, if it is a part of the communication information packet, the channel data organization module thinks that a bus communication activity has been monitored, and extracts the source of this bus communication activity from the channel data packet according to the bus communication protocol. The node number and target node number are sent to the row-column mapping module. The row and column mapping module uses the source node number and target node number as the row node address and column node address of the left port, reads the channel health status word from the left port of the three-dimensional channel table memory, and sends it to the channel data organization module. Afterwards, the channel data organization module reads the communication data from the available bus channel according to the channel health status word, forms a communication information packet, adds time code and other information, and stores it in the communication data buffer memory.

Table 2 The storage form of the health state of the memory bus channel in the three-dimensional channel table

In this embodiment, the UM-BUS bus supports 30 communication nodes and 32 bus channels, including two reserved nodes 0 and 31, a total of 32 nodes, and both the node number and the channel number need to be represented by 5-digit binary numbers. In order to store the channel health states between all nodes, the three-dimensional channel table memory uses a 32-bit dual-port memory with 1024 words. The storage form of the health state of the bus channel in the three-dimensional channel table memory is shown in Table 2. The row node dimension and the column node dimension each occupy 5 bits, constituting the 10-bit access address of the three-dimensional channel table memory. Each 32-bit word of the three-dimensional channel table memory corresponds to a physical channel bit by bit, and stores the health status X of the channel. When X is 0, it indicates that the channel is faulty and unavailable, and when it is 1, it indicates that the channel is in a healthy and available state.

When the channel data organization module reads the channel health status word, the source node number is used as the row node dimension address, and the target node number is used as the column node dimension address to read the channel health status when the source node sends communication data packets to the target node Character. At this time, each bit in the three-dimensional channel table memory constitutes a three-dimensional access space as shown in FIG. 6 , which can give the health status of any channel between two communication nodes. In Fig. 6, M (2, 30, 6) = 1 represents that the health status of the sixth channel is healthy and available when node 2 communicates with node 30 in the three-dimensional channel table memory, and M (3, 4, 4) = 1 represents the node When communicating from node 3 to node 4, the health status of the fourth channel is healthy and available.

According to the UM-BUS bus protocol, the bus communication process is started by the master node sending a communication command packet, and the slave node sends a communication response packet to end. At any time, only one communication data packet can be transmitted on the bus. Therefore, during the monitoring process, the channel data organization module only reads the channel health status word from the source node to the target node from the three-dimensional channel table memory, and accesses it from one port of the three-dimensional channel table memory. However, during the communication process, both master and slave nodes will serve as source nodes to send communication data packets to each other, therefore, the three-dimensional channel table memory must be able to provide channel health status from any node as source node to other nodes.

However, according to the UM-BUS bus protocol, the UM-BUS bus detection process can only be initiated by the master node, that is, the source node in the channel detection process can only be the master node that initiates the detection, and the target node is the detected node. In this way, it is impossible for the slave node to appear as the source node in the detection process. Without special treatment, the channel health status word when the slave node is used as the source node cannot be established in the three-dimensional channel table memory, resulting in the inability to monitor the communication response packet sent by the slave node. . Consider the following two issues: 1) There is no communication process from node to slave node and from node to self in the UM-BUS bus; 2) During the bus channel detection process, the bidirectional channels of the master and slave nodes are detected, so The health status of the channel from the master node to the slave node is consistent with the health status of the channel from the slave node to the master node. Therefore, the following processing is performed on the storage of the channel health state of the three-dimensional channel table memory:

(1) When monitoring to the end of a detection process, the row-column mapping module maps the channel from the left and right ports to The health status word is symmetrically written to two corresponding locations in the three-dimensional channel table memory.

(2) When resetting, write all bits of the three-dimensional channel table memory into 0. In this way, after the bus detection process is all over, the channel health status word between the master node and other nodes and the slave node to the master node will be written into the actual channel health status word during the detection process, and each slave node to other slave nodes And the channel health status word from each node to itself will remain all 0.

The channel health state recording device in the dynamic reconfigurable bus monitoring system realized by the present invention solves the problem of obtaining, recording and using the channel health status between nodes of the bus in the dynamic reconfigurable bus monitoring system, and realizes the dynamic Proper monitoring of reconfigurable bus communication activity.

Within the scope of not departing from the spirit of the present invention, the present invention can have various modifications, such as: the number of nodes supported by the bus, the number of channels, the numbering method of nodes and channels, etc., which can be changed in different implementations. These modifications are also included within the scope of protection claimed by the present invention.

Claims (7)

1. A channel health state recording device of a dynamic reconfigurable bus monitoring system, characterized in that: the channel health state recording device includes a channel detection and identification module, a row and column mapping module, a three-dimensional channel table memory, a channel data organization module and a communication Data buffer memory. 2. The channel health state recording device according to claim 1, characterized in that: the channel detection identification module is used to monitor the channel detection process of the dynamic reconfigurable bus, when a channel detection process is monitored, according to The monitored detection information packet generates the source node number, target node number, channel health status word and channel detection completion signal of the channel detection and sends them to the row-column mapping module. 3. The channel health state recording device according to claim 1, characterized in that: the row and column mapping module is used to generate the access address and control signal of the three-dimensional channel table memory, when receiving the channel detection and identification module sent by the channel detection and identification module When the signal is completed, the source node number sent by the channel detection and identification module is used as the row node address of the left port, the target node number is used as the column node address of the left port, and the channel health status word is written into the three-dimensional channel table memory from the left port; The target node number sent by the channel detection and identification module is used as the row node address of the right port, the source node number is used as the column node address of the right port, and the channel health status word is simultaneously written into the three-dimensional channel table memory from the right port. 4. The channel health state recording device according to claim 1, characterized in that: the three-dimensional channel table memory is a dual-port memory for storing the channel health states between all nodes on the dynamically reconfigurable bus , the effective data width of the left and right ports is the same as the number of channels of the dynamic reconfigurable bus, the effective address bits of the left and right ports are equally divided into two parts: the row node address and the column node address, and each data word represents Corresponding to the channel health status between row nodes and column nodes, each bit corresponds to a channel of the dynamic reconfigurable bus, forming a channel health status storage table in three dimensions of row nodes, column nodes and channels. 5 . The channel health state recording device according to claim 1 , wherein the communication data buffer memory is used for storing monitored bus communication information and bus detection information in the dynamically reconfigurable bus monitoring system. 5 . 6. The channel health state recording device according to claim 1, characterized in that: the channel data organization module is used to monitor the dynamic reconfigurable bus, and whenever a bus communication process is monitored, the monitored The source node number and target node number of the communication information packet are sent to the row-column mapping module, and the row-column mapping module uses the source node number and the target node number as the row node address and the column node address, and reads the channel from any port of the three-dimensional channel table memory The health status word is sent to the channel data organization module; the channel data organization module reads the communication data from the available bus channel according to the channel health status word, forms a communication information packet, and stores it in the communication after adding the information including the time code. in the data buffer memory. 7. The channel health state recording device according to claim 2, characterized in that: the channel detection identification module is also used to compose the detection information together with the information including the time code when a channel detection process is monitored The detection packet is stored in the communication data buffer memory.