[go: up one dir, main page]

CN118331905B - Extensible on-chip interconnection bus structure - Google Patents

Extensible on-chip interconnection bus structure Download PDF

Info

Publication number
CN118331905B
CN118331905B CN202410753080.1A CN202410753080A CN118331905B CN 118331905 B CN118331905 B CN 118331905B CN 202410753080 A CN202410753080 A CN 202410753080A CN 118331905 B CN118331905 B CN 118331905B
Authority
CN
China
Prior art keywords
host
slave
priority
arbitration
signal
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Active
Application number
CN202410753080.1A
Other languages
Chinese (zh)
Other versions
CN118331905A (en
Inventor
崔媛媛
赵晓冬
张海金
张洵颖
李万通
李臻
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northwestern Polytechnical University
Original Assignee
Northwestern Polytechnical University
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Northwestern Polytechnical University filed Critical Northwestern Polytechnical University
Priority to CN202410753080.1A priority Critical patent/CN118331905B/en
Publication of CN118331905A publication Critical patent/CN118331905A/en
Application granted granted Critical
Publication of CN118331905B publication Critical patent/CN118331905B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/14Handling requests for interconnection or transfer
    • G06F13/20Handling requests for interconnection or transfer for access to input/output bus
    • G06F13/28Handling requests for interconnection or transfer for access to input/output bus using burst mode transfer, e.g. direct memory access DMA, cycle steal
    • G06F13/30Handling requests for interconnection or transfer for access to input/output bus using burst mode transfer, e.g. direct memory access DMA, cycle steal with priority control
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/14Handling requests for interconnection or transfer
    • G06F13/36Handling requests for interconnection or transfer for access to common bus or bus system
    • G06F13/368Handling requests for interconnection or transfer for access to common bus or bus system with decentralised access control
    • G06F13/374Handling requests for interconnection or transfer for access to common bus or bus system with decentralised access control using a self-select method with individual priority code comparator
    • GPHYSICS
    • G06COMPUTING OR CALCULATING; COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F13/00Interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F13/38Information transfer, e.g. on bus
    • G06F13/40Bus structure
    • G06F13/4004Coupling between buses
    • G06F13/4027Coupling between buses using bus bridges
    • G06F13/4031Coupling between buses using bus bridges with arbitration
    • YGENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02DCLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D10/00Energy efficient computing, e.g. low power processors, power management or thermal management

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • General Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Hardware Design (AREA)
  • Bus Control (AREA)

Abstract

本发明涉及片上互联总线设计技术领域,具体提出一种可扩展的片上互联总线结构,其包括多个主机接口电路与多个从机仲裁选择逻辑电路,多个主机接口电路与多个从机仲裁选择逻辑电路连接构成硬件互联矩阵结构,其中,主机端口的管理由主机接口电路完成,从机端口的管理由从机仲裁选择逻辑电路完成;主机端口和从机端口扩展或者缩减,均通过添加或去除主机接口电路和从机仲裁选择逻辑电路完成;从机仲裁选择逻辑电路支持固定优先级比较算法、循环优先级算法两种仲裁逻辑,通过选择不同的仲裁模式,满足不同应用场景下多个总线主机访问同一个从机时的仲裁需求。

The invention relates to the technical field of on-chip interconnection bus design, and specifically proposes an expandable on-chip interconnection bus structure, which comprises a plurality of host interface circuits and a plurality of slave arbitration selection logic circuits, wherein the plurality of host interface circuits and the plurality of slave arbitration selection logic circuits are connected to form a hardware interconnection matrix structure, wherein the management of the host port is completed by the host interface circuit, and the management of the slave port is completed by the slave arbitration selection logic circuit; the expansion or reduction of the host port and the slave port is completed by adding or removing the host interface circuit and the slave arbitration selection logic circuit; the slave arbitration selection logic circuit supports two arbitration logics, namely a fixed priority comparison algorithm and a cyclic priority algorithm, and satisfies the arbitration requirements when multiple bus hosts access the same slave in different application scenarios by selecting different arbitration modes.

Description

一种可扩展的片上互联总线结构A Scalable On-Chip Interconnect Bus Architecture

技术领域Technical Field

本发明属于片上互联总线设计技术领域,特别涉及一种可扩展的片上互联总线结构。The invention belongs to the technical field of on-chip interconnection bus design, and in particular relates to an expandable on-chip interconnection bus structure.

背景技术Background technique

随着集成电路制造水平不断提高,系统芯片设计规模不断增大,系统芯片集成的功能模块种类和数量越来越多,所有功能模块依靠片上互联总线实现片上交互,片上互联总线设计决定着系统芯片运行效能。As the level of integrated circuit manufacturing continues to improve, the scale of system chip design continues to increase, the types and numbers of functional modules integrated in system chips are increasing, and all functional modules rely on on-chip interconnect buses to achieve on-chip interaction. The design of the on-chip interconnect bus determines the operating performance of the system chip.

目前大部分系统芯片为了保证运行效能,其片上互联总线设计结构定制化程度较高,灵活扩展能力不强。In order to ensure operational performance, most current system chips have a high degree of customization in their on-chip interconnect bus design structure and have limited flexible expansion capabilities.

发明内容Summary of the invention

为了解决上述背景技术中存在的技术问题,本发明提出了一种可扩展的片上互联总线结构(以下简称:可扩展总线),在保障良好的总线运行效能的前提下,实现了片上互联总线主机和从机的灵活扩展,显著提高片上互联总线结构的可扩展性与通用性。In order to solve the technical problems existing in the above-mentioned background technology, the present invention proposes an expandable on-chip interconnect bus structure (hereinafter referred to as: expandable bus), which realizes the flexible expansion of the on-chip interconnect bus master and slaves while ensuring good bus operation performance, and significantly improves the scalability and versatility of the on-chip interconnect bus structure.

具体的,本发明提供了一种可扩展的片上互联总线结构,包括多个主机接口电路与多个从机仲裁选择逻辑电路,多个所述主机接口电路与多个所述从机仲裁选择逻辑电路连接构成硬件互联矩阵结构,所述硬件互联矩阵结构支持将多个总线通过总线主机扩展接口与扩展接入接口进行连接,主机能够对多个总线上任意从机进行访问,共同构成一个片上互联总线;其中,主机端口的管理由所述主机接口电路完成,从机端口的管理由所述从机仲裁选择逻辑电路完成;Specifically, the present invention provides an expandable on-chip interconnect bus structure, including multiple host interface circuits and multiple slave arbitration selection logic circuits, multiple host interface circuits and multiple slave arbitration selection logic circuits are connected to form a hardware interconnect matrix structure, the hardware interconnect matrix structure supports connecting multiple buses through bus host extension interfaces and extension access interfaces, and the host can access any slave on the multiple buses, together forming an on-chip interconnect bus; wherein the management of the host port is completed by the host interface circuit, and the management of the slave port is completed by the slave arbitration selection logic circuit;

主机端口和从机端口扩展或者缩减,一方面通过将多个总线通过总线主机扩展接口与扩展接入接口进行连接构成一个片上互联总线完成;另一方面通过添加同一个总线或去除主机接口电路和从机仲裁选择逻辑电路完成;The expansion or reduction of the host port and the slave port is accomplished, on the one hand, by connecting multiple buses through the bus host expansion interface and the expansion access interface to form an on-chip interconnect bus; on the other hand, by adding the same bus or removing the host interface circuit and the slave arbitration selection logic circuit;

所述从机仲裁选择逻辑电路支持层次化仲裁逻辑,每一个从机接口通过层次化仲裁逻辑电路将连接到该总线上主机分为若干个组,同一组内主机采用一种优先级模式,不同组之间采用另一种优先级模式;所述优先级模式支持固定优先级比较算法或者循环优先级算法。层次化仲裁逻辑电路支持主机访问从机的精细化控制,可以更好的满足不同应用场景下多个总线主机访问同一个从机时的仲裁需求,提高访问效率。The slave arbitration selection logic circuit supports hierarchical arbitration logic. Each slave interface divides the hosts connected to the bus into several groups through the hierarchical arbitration logic circuit. The hosts in the same group adopt one priority mode, and different groups adopt another priority mode. The priority mode supports a fixed priority comparison algorithm or a circular priority algorithm. The hierarchical arbitration logic circuit supports refined control of the host access to the slave, which can better meet the arbitration requirements when multiple bus hosts access the same slave in different application scenarios, and improve access efficiency.

优选的,所述主机接口电路用于向目标主机发送主机访问控制信号、写数据,同时接收来自从机的响应信号、读数据,包括读数据选择电路和主机接口控制状态机;Preferably, the host interface circuit is used to send a host access control signal and write data to the target host, and simultaneously receive a response signal and read data from the slave, and includes a read data selection circuit and a host interface control state machine;

所述主机接口电路的输入信号包括时钟与复位信号、主机数据信号、主机控制信号、从机读数据信号、从机响应信号,输出信号包括主机数据信号、主机控制信号、目标从机读数据信号、从机响应信号、主机状态指示信号;The input signals of the host interface circuit include clock and reset signals, host data signals, host control signals, slave read data signals, and slave response signals, and the output signals include host data signals, host control signals, target slave read data signals, slave response signals, and host status indication signals;

所述主机接口电路根据从机选择信号,将所述目标从机读数据信号输出至主机,将所述主机数据信号与所述主机控制信号输出至所述从机仲裁选择逻辑电路中;所述主机接口控制状态机将所述从机响应信号与所述主机控制信号作为输入,控制所述主机接口电路状态,输出所述从机响应信号、所述主机状态指示信号、当前从机选择信号。The host interface circuit outputs the target slave read data signal to the host according to the slave selection signal, and outputs the host data signal and the host control signal to the slave arbitration selection logic circuit; the host interface control state machine takes the slave response signal and the host control signal as input, controls the state of the host interface circuit, and outputs the slave response signal, the host state indication signal, and the current slave selection signal.

优选的,所述读数据选择电路的输入信号为总线上所有从机读数据信号及从机选择信号,输出信号为目标从机读数据信号;Preferably, the input signal of the read data selection circuit is the read data signal of all slaves on the bus and the slave selection signal, and the output signal is the read data signal of the target slave;

所述读数据选择电路根据所述从机选择信号,从输入的所有从机读数据信号选择目标从机读数据信号;将其输出至对应主机中;当前主机无访问请求时,所述读数据选择电路的输出保持恒定状态。The read data selection circuit selects a target slave read data signal from all input slave read data signals according to the slave selection signal, and outputs it to the corresponding host; when the current host has no access request, the output of the read data selection circuit remains constant.

优选的,所述主机接口控制状态机的输入信号为总线的时钟与复位信号、从机响应信号及主机控制信号,输出信号为主机状态指示信号、从机响应信号及从机选择信号;Preferably, the input signals of the host interface control state machine are the clock and reset signals of the bus, the slave response signal and the host control signal, and the output signals are the host state indication signal, the slave response signal and the slave selection signal;

主机数据信号、主机控制信号从主机输出,经过主机接口电路后,输出至各个从机仲裁选择逻辑电路中;当前主机无访问请求时,主机控制信号与主机数据信号输出保持恒定状态。The host data signal and host control signal are output from the host, and after passing through the host interface circuit, are output to each slave arbitration selection logic circuit; when the current host has no access request, the host control signal and host data signal output remain in a constant state.

优选的,所述从机仲裁选择逻辑电路用于接收来自总线上各主机的数据与控制信号,基于优先级模式与配置给出同时访问的多个主机的访问优先级,根据同时访问的各主机优先级选择当前有效的主机访问控制与写数据输出至从机,同将从机的响应信号与读数据输出至对应的主机端口,包括主机选择电路、层次化仲裁逻辑电路、从机仲裁选择逻辑状态机及从机寄存器;Preferably, the slave arbitration selection logic circuit is used to receive data and control signals from each host on the bus, give access priorities of multiple hosts accessed simultaneously based on the priority mode and configuration, select the currently valid host access control and write data to output to the slave according to the priority of each host accessed simultaneously, and output the response signal and read data of the slave to the corresponding host port, including a host selection circuit, a hierarchical arbitration logic circuit, a slave arbitration selection logic state machine and a slave register;

所述从机仲裁选择逻辑电路的输入信号包括总线上所有主机数据与控制信号、从机配置信号、从机读数据信号与响应信号、从机配置时钟与复位信号、从机仲裁选择逻辑电路工作时钟与复位信号,输出信号包括当前主机数据与控制信号、从机访问状态信号、从机读数据信号与响应信号;The input signals of the slave arbitration selection logic circuit include all host data and control signals on the bus, slave configuration signals, slave read data signals and response signals, slave configuration clock and reset signals, slave arbitration selection logic circuit working clock and reset signals, and the output signals include current host data and control signals, slave access status signals, slave read data signals and response signals;

所述从机仲裁选择逻辑电路依据从机状态与主机控制信号,基于优先级模式与配置信号,在总线上全部主机端口中选择有效的主机端口,并将该主机数据与控制信号输出至从机。The slave arbitration selection logic circuit selects a valid host port from all host ports on the bus according to the slave state and the host control signal, based on the priority mode and the configuration signal, and outputs the host data and control signal to the slave.

优选的,主机选择电路的输入信号包括总线上所有主机数据与控制信号、主机选择信号,输出信号为当前主机数据与控制信号;Preferably, the input signal of the host selection circuit includes all host data and control signals on the bus, and the host selection signal, and the output signal is the current host data and control signal;

所述主机选择信号由从机仲裁选择逻辑状态机生成,所述主机选择电路依据主机选择信号,确定当前有效的主机端口并给予其对当前主机的控制权限,输出当前主机数据与控制信号;从机无主机端口访问时,当前主机数据与控制信号输出保持恒定状态。The host selection signal is generated by the slave arbitration selection logic state machine. The host selection circuit determines the current valid host port and grants it control authority over the current host based on the host selection signal, and outputs the current host data and control signals; when the slave has no host port access, the current host data and control signal outputs remain in a constant state.

优选的,所述从机寄存器的输入信号包括从机配置信号、从机配置时钟与复位信号;所述从机配置信号包括从机仲裁模式、仲裁优先级配置;Preferably, the input signal of the slave register includes a slave configuration signal, a slave configuration clock and a reset signal; the slave configuration signal includes a slave arbitration mode and an arbitration priority configuration;

所述从机寄存器保存从机仲裁模式与仲裁优先级的配置,并将优先级模式与配置信号输出至所述层次化仲裁逻辑电路。The slave register stores the configuration of the slave arbitration mode and arbitration priority, and outputs the priority mode and configuration signals to the hierarchical arbitration logic circuit.

优选的,所述从机仲裁选择逻辑状态机的输入信号为从机仲裁选择逻辑电路工作时钟与复位信号、层次化仲裁逻辑电路的仲裁结果,输出为从机访问状态信号;Preferably, the input signal of the slave arbitration selection logic state machine is the slave arbitration selection logic circuit working clock and reset signal, and the arbitration result of the hierarchical arbitration logic circuit, and the output is the slave access state signal;

所述从机仲裁选择逻辑状态机根据所述仲裁结果、当前从机被控状态,产生最新的主机选择信号、从机访问状态信号,最新的主机选择信号输出至所述层次化仲裁逻辑电路及所述主机选择电路。The slave arbitration selection logic state machine generates the latest host selection signal and slave access status signal according to the arbitration result and the current slave controlled state, and the latest host selection signal is output to the hierarchical arbitration logic circuit and the host selection circuit.

优选的,所述层次化仲裁逻辑电路用于实现可扩展总线的层次化仲裁逻辑,其输入信号为主机选择信号、优先级模式与配置信号,输出信号为各个主机访问优先级仲裁结果;Preferably, the hierarchical arbitration logic circuit is used to implement the hierarchical arbitration logic of the scalable bus, and its input signal is the host selection signal, the priority mode and the configuration signal, and the output signal is the access priority arbitration result of each host;

所述层次化仲裁逻辑电路的每一个从机接口都能够通过层次化仲裁逻辑结构对可扩展总线上所有主机的优先级进行分层配置,对于任意一个从机接口,同一层次下的主机基于一种优先级仲裁逻辑完成仲裁,不同层次之间优先级仲裁逻辑不同,优先级仲裁逻辑算法包括固定优先级比较算法、循环优先级算法两种优先级仲裁算法,不同层次选择不同的仲裁逻辑算法作为该层次的仲裁逻辑;Each slave interface of the hierarchical arbitration logic circuit can hierarchically configure the priorities of all hosts on the expandable bus through the hierarchical arbitration logic structure. For any slave interface, the hosts at the same level complete arbitration based on a priority arbitration logic. The priority arbitration logics at different levels are different. The priority arbitration logic algorithms include two priority arbitration algorithms: a fixed priority comparison algorithm and a circular priority algorithm. Different levels select different arbitration logic algorithms as the arbitration logic of the level.

所述层次化仲裁逻辑电路的工作流程是所有主机选择信号、优先级模式与配置、优先级分组配置信号作为优先级分组配置电路的输入,当主机选择信号有效时其对应的主机ID也作为优先级分配电路的输入;优先级分组配置电路根据主机ID与优先级分组配置信号,对所有主机选择信号进行分组,将组内优先级模式与配置信号、同一组有效主机选择信号输入到层次化仲裁逻辑电路,每组包括i个主机,共有j组,同组主机选择信号经过层次化仲裁逻辑电路,输出组中每个主机的组内优先级;优先级分组配置电路将组内所有主机选择信号输入到或电路,当组内任意一个主机选择信号有效时,该组主机选择信号即为有效,所有组的组主机选择信号输入到层次化仲裁逻辑电路,该电路输出所有组的优先级;组优先级信号与该组中所有主机的组内优先级结果输入到优先级组合电路,优先级组合电路将执行主机组内优先级与对应组优先级的乘运算,输出各个主机的访问优先级,主机的访问优先级作为多个主机同时访问该从机时的仲裁依据。The workflow of the hierarchical arbitration logic circuit is that all host selection signals, priority modes and configurations, and priority group configuration signals are used as inputs of the priority group configuration circuit. When the host selection signal is valid, its corresponding host ID is also used as input of the priority allocation circuit; the priority group configuration circuit groups all host selection signals according to the host ID and the priority group configuration signal, and inputs the priority mode and configuration signal within the group and the valid host selection signals of the same group into the hierarchical arbitration logic circuit. Each group includes i hosts, and there are j groups in total. The host selection signals of the same group pass through the hierarchical arbitration logic circuit, and the group priority of each host in the group is output; the priority group configuration circuit inputs all host selection signals in the group into the OR circuit. When any host selection signal in the group is valid, the host selection signal of the group is valid. The group host selection signals of all groups are input into the hierarchical arbitration logic circuit, and the circuit outputs the priorities of all groups; the group priority signal and the group priority results of all hosts in the group are input into the priority combination circuit. The priority combination circuit will perform the multiplication operation of the host group priority and the corresponding group priority, and output the access priority of each host. The access priority of the host is used as the arbitration basis when multiple hosts access the slave at the same time.

优选的,当所述层次化仲裁逻辑电路采用固定优先级比较算法时,所述层次化仲裁逻辑电路将各主机的输入优先级配置作为各主机优先级仲裁结果直接输出;若两个主机请求访问从机端口,则具有较高优先级的主机将获得对从机端口的控制权限;当主机端口向从机端口发出请求时,从机端口将检查新的请求主机端口的优先级是否高于当前控制从机端口的主机端口的优先级;从机端口在每个时钟边沿执行仲裁检查,如果新请求主机的优先级高于当前控制从机端口的主机优先级,则新请求主机将在下一个时钟边沿获得对从机端口的控制权限;若当前控制从机端口的主机正在执行固定长度的突发传输或锁定的传输,则无论新请求主机的优先级是否高于当前正在执行访问的主机,均须等待突发传输或锁定传输结束后,才能获得对从机端口的控制权限;Preferably, when the hierarchical arbitration logic circuit adopts a fixed priority comparison algorithm, the hierarchical arbitration logic circuit directly outputs the input priority configuration of each host as the arbitration result of each host priority; if two hosts request to access the slave port, the host with a higher priority will obtain the control authority over the slave port; when the host port sends a request to the slave port, the slave port will check whether the priority of the new requesting host port is higher than the priority of the host port currently controlling the slave port; the slave port performs an arbitration check on each clock edge, and if the priority of the new requesting host is higher than the priority of the host currently controlling the slave port, the new requesting host will obtain the control authority over the slave port at the next clock edge; if the host currently controlling the slave port is performing a fixed-length burst transmission or a locked transmission, then regardless of whether the priority of the new requesting host is higher than that of the host currently performing access, it must wait until the burst transmission or the locked transmission ends before obtaining the control authority over the slave port;

当所述层次化仲裁逻辑电路采用循环优先级算法时,每个主机端口根据物理主机端口号分配相对优先级,并将所述相对优先级与在总线上执行传输的最后一个主机端口进行比较,最高优先级请求主机获得下一个传输边界从机的控制权限;当主机获得目标从机的控制权限后,主机能够向该端口执行数据传输,直到另一个主机向同一从机端口发出请求为止;下一个主机被允许在下一个传输边界访问从机端口,若当前主线没有挂起的访问请求,则在下一个时钟周期中访问。When the hierarchical arbitration logic circuit adopts a round-robin priority algorithm, each host port is assigned a relative priority according to the physical host port number, and the relative priority is compared with the last host port that performs transmission on the bus. The highest priority request host obtains control authority for the next transmission boundary slave; when the host obtains control authority for the target slave, the host can perform data transmission to the port until another host sends a request to the same slave port; the next host is allowed to access the slave port at the next transmission boundary, and if there is no pending access request on the current master line, it is accessed in the next clock cycle.

本发明的有益效果在于:The beneficial effects of the present invention are:

本发明提出的可扩展的片上互联总线结构,可以根据实际需要,灵活选择集成的主机和从机数量。该可扩展总线的主要优点如下:The scalable on-chip interconnect bus structure proposed in the present invention can flexibly select the number of integrated masters and slaves according to actual needs. The main advantages of the scalable bus are as follows:

(1)本发明中可扩展的片上互联总线支持将多个总线通过总线主机扩展接口与扩展接入接口进行连接,主机可以对多个总线上任意从机进行访问,共同构成一个片上互联总线,实现片上互联总线的扩展,提高了总线扩展的灵活性;(1) The expandable on-chip interconnect bus of the present invention supports connecting multiple buses through a bus host expansion interface and an expansion access interface. The host can access any slave on the multiple buses to form an on-chip interconnect bus, thereby realizing the expansion of the on-chip interconnect bus and improving the flexibility of bus expansion.

(2)本发明中的扩展总线采用硬件互连矩阵的方式连接总线上所有的主机与从机端口,支持灵活扩展,支持所有总线主机同时访问不同的总线从机,有效提高总线访问效率。同时为了缓解可扩展总线中硬件互联矩阵结构带来的功耗升高问题,对总线中所有不活跃的主机端口保持恒定输出以降低电路功耗;(2) The expansion bus in the present invention uses a hardware interconnection matrix to connect all host and slave ports on the bus, supports flexible expansion, supports all bus hosts to access different bus slaves at the same time, and effectively improves the bus access efficiency. At the same time, in order to alleviate the problem of increased power consumption caused by the hardware interconnection matrix structure in the expandable bus, all inactive host ports in the bus maintain a constant output to reduce circuit power consumption;

(3)本发明中的可扩展总线为每一个从机仲裁选择设计了层次化仲裁逻辑电路。从机可以根据主机类型,对连接到该从机的主机进行分类,将同一类型主机分为一组并采用一种优先级仲裁模式,包含不同类型主机的主机组采用一种优先级仲裁模式,该分层层次化仲裁逻辑电路设计实现了主机访问从机仲裁的精细化控制,可以更好的满足不同应用场景下多个总线主机访问同一个从机时的仲裁需求,提高访问效率。下面通过附图和实施例,对本发明的技术方案做进一步的详细描述。(3) The scalable bus in the present invention designs a hierarchical arbitration logic circuit for each slave arbitration selection. The slave can classify the hosts connected to the slave according to the host type, classify the hosts of the same type into a group and adopt a priority arbitration mode, and the host group containing different types of hosts adopts a priority arbitration mode. The hierarchical arbitration logic circuit design realizes the refined control of the host access to the slave arbitration, which can better meet the arbitration requirements when multiple bus hosts access the same slave in different application scenarios, and improve the access efficiency. The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments.

附图说明BRIEF DESCRIPTION OF THE DRAWINGS

附图用来提供对本发明的进一步理解,并且构成说明书的一部分,与本发明的实施例一起用于解释本发明,并不构成对本发明的限制。在附图中:The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

图1是本发明提供的可扩展的片上互联总线结构。FIG. 1 is a scalable on-chip interconnect bus structure provided by the present invention.

图2是本发明提供的可扩展的片上总线互联总线级联结构示例。FIG. 2 is an example of an expandable on-chip bus interconnection bus cascade structure provided by the present invention.

图3是本发明提供的主机端口电路结构。FIG. 3 is a circuit structure of a host port provided by the present invention.

图4是本发明提供的从机仲裁选择逻辑电路结构。FIG. 4 is a slave arbitration selection logic circuit structure provided by the present invention.

图5是本发明提供的层次化仲裁逻辑电路结构。FIG. 5 is a hierarchical arbitration logic circuit structure provided by the present invention.

图6是本发明提供的主机数量为8个的层次化仲裁逻辑电路结构。FIG6 is a hierarchical arbitration logic circuit structure for eight hosts provided by the present invention.

具体实施方式Detailed ways

以下结合附图对本发明的优选实施例进行说明,应当理解,此处所描述的优选实施例仅用于说明和解释本发明,并不用于限定本发明。The preferred embodiments of the present invention are described below in conjunction with the accompanying drawings. It should be understood that the preferred embodiments described herein are only used to illustrate and explain the present invention, and are not used to limit the present invention.

针对目前大部分系统芯片为了保证运行效能,其片上互联总线设计结构定制化程度较高,灵活扩展能力不强的问题,本发明提出了一种可扩展的片上互联总线结构,可扩展的片上互联总线使用硬件互联矩阵的方式连接主机与从机,支持将多个总线通过总线主机扩展接口与扩展接入接口进行连接,主机可以对多个总线上任意从机进行访问,共同构成一个片上互联总线,实现片上互联总线的扩展。如图1所示为可扩展总线结构,如图 2所示为级联扩展后的可扩展总线,图1与图2中M0、M1、Mn代表n+1个主机接口,S0、S1、Sm代表m+1个从机接口。可扩展的片上互联总线使用硬件互连矩阵的方式连接总线主机与从机,主机端口的管理由主机接口电路完成,从机端口的管理由从机仲裁选择逻辑电路完成。多个主机接口电路单元与多个从机仲裁选择逻辑电路单元连接构成硬件互联矩阵结构。In order to ensure the operation efficiency of most system chips at present, the on-chip interconnection bus design structure has a high degree of customization and weak flexible expansion capability. The present invention proposes an expandable on-chip interconnection bus structure. The expandable on-chip interconnection bus uses a hardware interconnection matrix to connect the host and the slave, supports connecting multiple buses through a bus host expansion interface and an expansion access interface, and the host can access any slave on multiple buses, which together constitute an on-chip interconnection bus to achieve the expansion of the on-chip interconnection bus. As shown in Figure 1, the expandable bus structure is shown, and as shown in Figure 2, the expandable bus after cascade expansion, in Figures 1 and 2, M0, M1, and Mn represent n+1 host interfaces, and S0, S1, and Sm represent m+1 slave interfaces. The expandable on-chip interconnection bus uses a hardware interconnection matrix to connect the bus host and the slave, and the management of the host port is completed by the host interface circuit, and the management of the slave port is completed by the slave arbitration selection logic circuit. Multiple host interface circuit units are connected with multiple slave arbitration selection logic circuit units to form a hardware interconnection matrix structure.

可扩展总线为每一个从机仲裁选择设计了层次化仲裁逻辑电路,每一个从机接口都可以通过层次化仲裁逻辑电路对连接到该总线上主机的优先级配置进行分层,即将主机分为若干个组,同一组内主机采用一种优先级模式,不同组之间采用一种优先级模式。优先级模式支持固定优先级比较算法或者循环优先级算法。对于主机访问某从机来说,该主机的组内优先级与所属组优先级相乘得到该主机访问某从机时的最终访问优先级。对于从机来说,可以根据主机类型,对连接到该从机的主机进行分类,将同一类型主机分为一组并采用一种优先级仲裁模式,例如相同类型的主机访问特性相同,可以配置为循环仲裁优先级,有利于各主机均衡获取访问权限;不同类型的主机组采用一种优先级仲裁模式。例如各个主机组之间采用固定优先级模式,各主机组主机特性不同,固定优先级可以更好的确认某些类型的主机被优先响应。层次化仲裁逻辑电路支持主机访问从机的精细化控制,可以更好的满足不同应用场景下多个总线主机访问同一个从机时的仲裁需求,提高访问效率。The scalable bus has designed a hierarchical arbitration logic circuit for each slave arbitration selection. Each slave interface can hierarchically configure the priority configuration of the host connected to the bus through the hierarchical arbitration logic circuit, that is, the host is divided into several groups, and the hosts in the same group adopt a priority mode, and different groups adopt a priority mode. The priority mode supports a fixed priority comparison algorithm or a circular priority algorithm. For a host accessing a slave, the host's intra-group priority is multiplied by the group priority to which it belongs to obtain the final access priority when the host accesses a slave. For the slave, the hosts connected to the slave can be classified according to the host type, and the hosts of the same type can be divided into a group and adopt a priority arbitration mode. For example, the access characteristics of hosts of the same type are the same, and the circular arbitration priority can be configured, which is conducive to the balanced access rights of each host; different types of host groups adopt a priority arbitration mode. For example, a fixed priority mode is adopted between each host group, and the host characteristics of each host group are different. The fixed priority can better confirm that certain types of hosts are responded to first. The hierarchical arbitration logic circuit supports the refined control of the host accessing the slave, which can better meet the arbitration requirements when multiple bus hosts access the same slave in different application scenarios, and improve the access efficiency.

可扩展总线上所有访问均由主机发起。当主机访问从机时,如果访问的目标从机可用(空闲),则主机可以立刻(零等待状态)访问目标从机,单时钟周期可通过可扩展总线访问到从机,可扩展总线将从机的所有响应信息返回给请求访问的主机;如果访问的目标从机端口忙,主机访问被插入等待状态,直到目标从机端口能够响应主机请求。服务请求的延迟取决于每个主机的优先级和响应从机的访问时间。在主机完成对目标从机访问之后,主机仍然保持控制该从机端口的权限,直到一个空闲周期之后或该主机对另一个从机发起访问。当另一个优先级较高的主机端口向从机端口发出请求时,如果当前主机在进行固定长度的突发传输,当前主机继续保留对从机端口的控制,直到该突发传输完成;否则,当前主机端口失去对从机端口的控制权限。All accesses on the scalable bus are initiated by the host. When the host accesses the slave, if the target slave to be accessed is available (idle), the host can access the target slave immediately (zero wait state), and access the slave through the scalable bus in a single clock cycle. The scalable bus returns all response information of the slave to the host requesting access; if the target slave port to be accessed is busy, the host access is inserted into the wait state until the target slave port can respond to the host request. The delay of the service request depends on the priority of each host and the access time of the responding slave. After the host completes the access to the target slave, the host still retains the right to control the slave port until an idle cycle or the host initiates access to another slave. When another host port with a higher priority sends a request to the slave port, if the current host is performing a fixed-length burst transmission, the current host continues to retain control of the slave port until the burst transmission is completed; otherwise, the current host port loses control of the slave port.

具体的,上述可扩展的片上互联总线结构包括:Specifically, the above-mentioned scalable on-chip interconnect bus structure includes:

1)主机接口电路,其电路结构如图3所示;1) Host interface circuit, the circuit structure of which is shown in Figure 3;

主机接口电路主要用于向目标主机发送主机访问控制信号、写数据,同时接收来自从机的响应信号、读数据,主要包括读数据选择电路、主机接口控制状态机两个部分。The host interface circuit is mainly used to send host access control signals and write data to the target host, and receive response signals and read data from the slave. It mainly includes two parts: read data selection circuit and host interface control state machine.

主机接口电路的输入信号包括时钟与复位信号、主机数据信号、主机控制信号、从机读数据信号、从机响应信号,输出信号主要包括主机数据信号、主机控制信号、目标从机读数据信号、从机响应信号、主机状态指示信号。The input signals of the host interface circuit include clock and reset signals, host data signals, host control signals, slave read data signals, and slave response signals. The output signals mainly include host data signals, host control signals, target slave read data signals, slave response signals, and host status indication signals.

主机接口电路根据从机选择信号,将目标从机读数据信号输出至主机,将主机数据与控制信号输出至可扩展总线的从机仲裁选择逻辑电路中。主机接口控制状态机将从机响应信号与主机控制信号作为输入,控制主机接口电路状态,输出从机响应信号、主机状态指示信号、当前从机选择信号。The host interface circuit outputs the target slave read data signal to the host according to the slave selection signal, and outputs the host data and control signal to the slave arbitration selection logic circuit of the expandable bus. The host interface control state machine takes the slave response signal and the host control signal as input, controls the state of the host interface circuit, and outputs the slave response signal, the host state indication signal, and the current slave selection signal.

(101)读数据选择电路(101) Read data selection circuit

读数据选择电路的输入信号为总线上所有从机读数据信号及从机选择信号,输出信号为目标从机读数据信号;读数据选择电路根据从机选择信号,从输入的所有从机的读数据信号选择目标从机读数据信号,将其输出至对应主机中。当前主机无访问请求时,读数据选择电路的输出保持恒定状态,以降低电路功耗。The input signal of the read data selection circuit is the read data signal of all slaves on the bus and the slave selection signal, and the output signal is the read data signal of the target slave; the read data selection circuit selects the read data signal of the target slave from the read data signals of all input slaves according to the slave selection signal, and outputs it to the corresponding host. When the current host has no access request, the output of the read data selection circuit remains constant to reduce the power consumption of the circuit.

(102)主机接口控制状态机(102) Host interface control state machine

主机接口控制状态机是主机接口电路控制器,其输入信号为可扩展总线的时钟与复位信号、从机响应信号、主机控制信号,输出信号为主机状态指示、从机响应信号、从机选择信号。同时主机数据信号、主机控制信号从主机输出,经过主机接口电路后,输出至各个从机仲裁选择逻辑电路中。当前主机无访问请求时,主机控制信号与主机数据信号输出保持恒定状态,以降低电路功耗。The host interface control state machine is a host interface circuit controller. Its input signals are the clock and reset signals of the expandable bus, the slave response signal, and the host control signal. Its output signals are the host status indication, the slave response signal, and the slave selection signal. At the same time, the host data signal and the host control signal are output from the host, and after passing through the host interface circuit, they are output to each slave arbitration selection logic circuit. When the current host has no access request, the host control signal and the host data signal output remain in a constant state to reduce circuit power consumption.

2)从机仲裁选择逻辑电路,其电路结构如图4所示;2) Slave arbitration selection logic circuit, whose circuit structure is shown in Figure 4;

从机仲裁选择逻辑电路主要用于接收来自总线上各主机的数据与控制信号,基于优先级模式与配置给出同时访问的多个主机的访问优先级,根据同时访问的各主机优先级选择当前有效的主机访问控制与写数据输出至从机。同时将从机的响应信号与读数据输出至对应的主机端口,主要包括主机选择电路、层次化仲裁逻辑电路、从机仲裁选择逻辑状态机、从机寄存器四个部分。The slave arbitration selection logic circuit is mainly used to receive data and control signals from each host on the bus, give access priorities to multiple hosts accessing simultaneously based on the priority mode and configuration, and select the currently valid host access control and write data to output to the slave according to the priority of each host accessing simultaneously. At the same time, the response signal and read data of the slave are output to the corresponding host port, which mainly includes four parts: host selection circuit, hierarchical arbitration logic circuit, slave arbitration selection logic state machine, and slave register.

从机仲裁选择逻辑电路的输入信号主要包括可扩展总线上所有的主机端口的数据与控制信号、从机配置信号、从机读数据与响应信号、从机配置时钟与复位信号、从机仲裁选择逻辑电路工作时钟与复位信号,输出信号主要包括当前主机数据与控制信号、从机访问状态信号、从机读数据与响应信号。The input signals of the slave arbitration selection logic circuit mainly include the data and control signals of all host ports on the expandable bus, slave configuration signals, slave read data and response signals, slave configuration clock and reset signals, slave arbitration selection logic circuit working clock and reset signals; the output signals mainly include the current host data and control signals, slave access status signals, slave read data and response signals.

从机仲裁选择逻辑电路依据从机状态与主机访问控制信号,基于优先级模式与配置信号,在可扩展总线上全部主机端口中选择有效的主机端口,并将该主机数据与控制信号输出至从机。The slave arbitration selection logic circuit selects a valid host port from all host ports on the expandable bus according to the slave state and the host access control signal, based on the priority mode and the configuration signal, and outputs the host data and control signal to the slave.

(201)主机选择电路(201) Host selection circuit

主机选择电路的输入信号为主机数据与控制信号、主机选择信号。主机数据与控制信号包括可扩展总线上所有主机端口的数据与控制信号。主机选择信号由从机仲裁选择逻辑状态机生成。主机选择电路依据主机选择信号,确定当前有效的主机端口并给予其对当前主机的控制权限,输出当前主机数据与控制信号。从机无主机端口访问时,当前主机数据与控制信号输出保持恒定状态,以降低电路功耗。The input signals of the host selection circuit are the host data and control signal and the host selection signal. The host data and control signal include the data and control signals of all host ports on the expandable bus. The host selection signal is generated by the slave arbitration selection logic state machine. The host selection circuit determines the current valid host port based on the host selection signal and gives it control authority over the current host, and outputs the current host data and control signal. When the slave has no host port access, the current host data and control signal output remains constant to reduce circuit power consumption.

(202)从机寄存器(202) Slave register

从机寄存器的输入信号为从机配置信号、从机配置时钟与复位信号。从机配置信号包括仲裁模式、仲裁优先级配置。从机寄存器保存从机仲裁模式与仲裁优先级的配置,并将其输出至层次化仲裁逻辑电路。The input signals of the slave register are the slave configuration signal, the slave configuration clock and the reset signal. The slave configuration signal includes the arbitration mode and arbitration priority configuration. The slave register saves the configuration of the slave arbitration mode and arbitration priority and outputs it to the hierarchical arbitration logic circuit.

(203)从机仲裁选择逻辑状态机(203) Slave arbitration selection logic state machine

从机仲裁选择逻辑状态机的输入信号为从机仲裁选择逻辑电路工作时钟与复位信号、层次化仲裁逻辑电路的仲裁结果,输出为从机访问状态信号。从机仲裁选择逻辑状态机会根据仲裁结果、当前从机被控状态,产生最新的主机选择信号、从机访问状态信号,最新的主机选择信号将输出至层次化仲裁逻辑电路、主机选择电路。The input signals of the slave arbitration selection logic state machine are the slave arbitration selection logic circuit working clock and reset signal, and the arbitration result of the hierarchical arbitration logic circuit, and the output is the slave access status signal. The slave arbitration selection logic state machine will generate the latest host selection signal and slave access status signal according to the arbitration result and the current slave controlled state, and the latest host selection signal will be output to the hierarchical arbitration logic circuit and the host selection circuit.

(204)层次化仲裁逻辑电路(204) Hierarchical arbitration logic circuit

层次化仲裁逻辑电路能够实现可扩展总线的层次化仲裁结构。层次化仲裁逻辑电路设计如图5所示,其中M0_ID[7:0]、M1_ID[7:0]、Mn_ID[7:0]分别代表主机0、主机1、主机n的端口编号,INT(0)代表输入0值。每一个从机接口都可以通过层次化仲裁逻辑结构对可扩展总线上所有主机的优先级进行分层设置,不同从机接口的分层方法可以不同。对于任意一个从机接口,同一层次下的主机基于一种优先级仲裁逻辑完成仲裁,不同层次之间优先级仲裁逻辑可以不同。优先级仲裁逻辑算法包括固定优先级比较算法、循环优先级算法两种优先级仲裁算法,不同层次可以灵活选择不同的仲裁逻辑算法作为该层次的仲裁逻辑,提高不同类型的主机访问从机的仲裁灵活性。The hierarchical arbitration logic circuit can realize the hierarchical arbitration structure of the scalable bus. The hierarchical arbitration logic circuit design is shown in Figure 5, where M0_ID[7:0], M1_ID[7:0], and Mn_ID[7:0] represent the port numbers of host 0, host 1, and host n, respectively, and INT(0) represents the input value of 0. Each slave interface can hierarchically set the priority of all hosts on the scalable bus through the hierarchical arbitration logic structure, and the hierarchical methods of different slave interfaces can be different. For any slave interface, the hosts at the same level complete arbitration based on a priority arbitration logic, and the priority arbitration logics between different levels can be different. The priority arbitration logic algorithm includes two priority arbitration algorithms: a fixed priority comparison algorithm and a circular priority algorithm. Different levels can flexibly select different arbitration logic algorithms as the arbitration logic of the level, thereby improving the arbitration flexibility of different types of hosts accessing slaves.

层次化仲裁逻辑电路的工作逻辑是,所有主机选择信号、优先级模式与配置、优先级分组配置信号作为优先级分组配置电路的输入,当主机选择信号有效时其对应的主机ID也作为优先级分配电路的输入。优先级分组配置电路根据主机ID与优先级分组配置信号,对所有主机选择信号进行分组,将组内优先级模式与配置信号、同一组有效主机选择信号输入到层次化仲裁逻辑电路,每组包括i个主机,共有j组。同组主机选择信号经过层次化仲裁逻辑电路,输出组中每个主机的组内优先级;优先级分组配置电路将组内所有主机选择信号输入到或电路,当组内任意一个主机选择信号有效时,该组主机选择信号即为有效,所有组的组主机选择信号输入到层次化仲裁逻辑电路,该电路输出所有组的优先级;组优先级信号与该组中所有主机的组内优先级结果输入到优先级组合电路,优先级组合电路将执行主机组内优先级与对应组优先级的乘运算,输出各个主机的访问优先级。主机的访问优先级作为多个主机同时访问该从机时的仲裁依据。The working logic of the hierarchical arbitration logic circuit is that all host selection signals, priority modes and configurations, and priority group configuration signals are used as inputs of the priority group configuration circuit. When the host selection signal is valid, its corresponding host ID is also used as input of the priority allocation circuit. The priority group configuration circuit groups all host selection signals according to the host ID and the priority group configuration signal, and inputs the priority mode and configuration signals within the group and the valid host selection signals of the same group into the hierarchical arbitration logic circuit. Each group includes i hosts, and there are j groups in total. The host selection signals of the same group pass through the hierarchical arbitration logic circuit and output the group priority of each host in the group; the priority group configuration circuit inputs all host selection signals in the group into the OR circuit. When any host selection signal in the group is valid, the host selection signal of the group is valid. The group host selection signals of all groups are input into the hierarchical arbitration logic circuit, and the circuit outputs the priority of all groups; the group priority signal and the group priority of all hosts in the group are input into the priority combination circuit. The priority combination circuit will perform the multiplication operation of the host group priority and the corresponding group priority, and output the access priority of each host. The master's access priority serves as the arbitration basis when multiple masters access the slave at the same time.

(205)层次化仲裁逻辑电路(205) Hierarchical Arbitration Logic Circuit

层次化仲裁逻辑电路实现可扩展总线的两种仲裁方法,包括固定优先级比较算法、循环优先级算法。层次化仲裁逻辑电路的输入信号为主机选择信号、优先级模式与配置信号。输出信号为各个主机访问优先级仲裁结果。层次化仲裁逻辑电路通过优先级模式信号,选择执行固定优先级比较算法仲裁逻辑或者循环优先级算法仲裁逻辑。The hierarchical arbitration logic circuit implements two arbitration methods for the scalable bus, including a fixed priority comparison algorithm and a round-robin priority algorithm. The input signals of the hierarchical arbitration logic circuit are the host selection signal, the priority mode and the configuration signal. The output signal is the arbitration result of each host access priority. The hierarchical arbitration logic circuit selects to execute the fixed priority comparison algorithm arbitration logic or the round-robin priority algorithm arbitration logic through the priority mode signal.

当层次化仲裁逻辑电路采用固定优先级比较算法,层次化仲裁逻辑电路将各主机的输入优先级配置作为各主机优先级仲裁结果直接输出。在固定优先级模式下操作时,每个主机被分配唯一的固定优先级。如果两个主机请求访问从机端口,则具有较高优先级的主机将获得对从机端口的控制。当主机端口向从机端口发出请求时,从机端口将检查新的请求主机端口的优先级是否高于当前控制从机端口的主机端口的优先级。从机端口在每个时钟边沿执行仲裁检查。如果新请求主机的优先级高于当前控制从机端口的主机优先级,则新请求主机将在下一个时钟边沿获得对从机端口的控制权限。该仲裁方法的例外情况是,当前控制从机端口的主机正在执行固定长度的突发传输或锁定的传输。在这种情况下,即使无论新请求主机的优先级高于当前正在执行访问的主机,也必须等待突发传输或锁定传输结束后,才获得对从机端口的控制权限。When the hierarchical arbitration logic circuit adopts a fixed priority comparison algorithm, the hierarchical arbitration logic circuit directly outputs the input priority configuration of each host as the arbitration result of each host priority. When operating in fixed priority mode, each host is assigned a unique fixed priority. If two hosts request access to a slave port, the host with a higher priority will obtain control of the slave port. When a host port sends a request to a slave port, the slave port will check whether the priority of the new requesting host port is higher than the priority of the host port currently controlling the slave port. The slave port performs an arbitration check on each clock edge. If the priority of the new requesting host is higher than the priority of the host currently controlling the slave port, the new requesting host will obtain control of the slave port on the next clock edge. The exception to this arbitration method is that the host currently controlling the slave port is performing a fixed-length burst transfer or a locked transfer. In this case, even if the priority of the new requesting host is higher than that of the host currently performing access, it must wait until the burst transfer or locked transfer ends before obtaining control of the slave port.

当层次化仲裁逻辑电路采用循环优先级算法,每个主机端口根据物理主机端口号分配相对优先级。将此相对优先级与在总线上执行传输的最后一个主机端口进行比较。最高优先级请求主机获得下一个传输边界从机的控制权限。当主机获得目标从机的控制权限后,主机可以向该端口执行数据传输,直到另一个主机向同一从机端口发出请求为止。下一个主机被允许在下一个传输边界访问从机端口,如果当前主线没有挂起的访问请求,则在下一个时钟周期中访问。When the hierarchical arbitration logic circuit adopts a round-robin priority algorithm, each host port is assigned a relative priority based on the physical host port number. This relative priority is compared with the last host port that performed a transfer on the bus. The highest priority requesting host obtains control authority for the next transfer boundary slave. When the host obtains control authority for the target slave, the host can perform data transfers to that port until another host issues a request to the same slave port. The next host is allowed to access the slave port at the next transfer boundary, and in the next clock cycle if there is no pending access request for the current master line.

图6给出了一种主机数量为8个的层次化仲裁逻辑电路示例,其中current_m[0:7]代表8个主机的选择信号,current_m[0]代表主机0选择信号,current_m[7]代表主机7选择信号。当仲裁模式选择固定优先级比较算法模式时,层次化仲裁逻辑电路直接将固定优先级模式下仲裁优先级配置输出,作为各个主机唯一的固定仲裁优先级;当仲裁模式选择循环优先级算法模式时,每一个主机端口的仲裁优先级为相对优先级,某一个从机端口下主机的优先级将随着当前正在访问从机的主机端口的变化而改变。假设可扩展总线主机端口0、1、2、3、4、5和7被使用,如果此时控制某一个从机端口的主机端口是主机端口3,而主机端口0、2、4、5和7同时发出请求,则它们将按照4、5、7、0和2的顺序依次获得对目标从机的访问权限。FIG6 shows an example of a hierarchical arbitration logic circuit with 8 hosts, where current_m[0:7] represents the selection signals of the 8 hosts, current_m[0] represents the selection signal of host 0, and current_m[7] represents the selection signal of host 7. When the arbitration mode selects the fixed priority comparison algorithm mode, the hierarchical arbitration logic circuit directly outputs the arbitration priority configuration in the fixed priority mode as the unique fixed arbitration priority of each host; when the arbitration mode selects the round-robin priority algorithm mode, the arbitration priority of each host port is a relative priority, and the priority of the host under a certain slave port will change with the change of the host port currently accessing the slave. Assuming that the expandable bus host ports 0, 1, 2, 3, 4, 5, and 7 are used, if the host port controlling a certain slave port is host port 3 at this time, and host ports 0, 2, 4, 5, and 7 send requests at the same time, they will obtain access rights to the target slave in the order of 4, 5, 7, 0, and 2.

显然,本领域的技术人员可以对本技术方案进行各种改动和变型而不脱离本技术方案的精神和范围。这样,倘若本技术方案的这些修改和变型属于本技术方案权利要求及其等同技术的范围之内,则本技术方案也意图包含这些改动和变型在内。Obviously, those skilled in the art can make various changes and modifications to the technical solution without departing from the spirit and scope of the technical solution. Thus, if these modifications and variations of the technical solution fall within the scope of the claims of the technical solution and their equivalents, the technical solution is also intended to include these modifications and variations.

Claims (10)

1.一种可扩展的片上互联总线结构,其特征在于,包括多个主机接口电路与多个从机仲裁选择逻辑电路,多个所述主机接口电路与多个所述从机仲裁选择逻辑电路连接构成硬件互联矩阵结构,所述硬件互联矩阵结构支持将多个总线通过总线主机扩展接口与扩展接入接口进行连接,主机能够对多个总线上任意从机进行访问,共同构成一个片上互联总线;其中,主机端口的管理由所述主机接口电路完成,从机端口的管理由所述从机仲裁选择逻辑电路完成;1. An expandable on-chip interconnect bus structure, characterized in that it comprises a plurality of host interface circuits and a plurality of slave arbitration selection logic circuits, wherein the plurality of host interface circuits and the plurality of slave arbitration selection logic circuits are connected to form a hardware interconnect matrix structure, wherein the hardware interconnect matrix structure supports connecting a plurality of buses through a bus host extension interface and an extension access interface, and the host can access any slave on the plurality of buses, which together form an on-chip interconnect bus; wherein the host port is managed by the host interface circuit, and the slave port is managed by the slave arbitration selection logic circuit; 主机端口和从机端口扩展或者缩减,一方面通过将多个总线通过总线主机扩展接口与扩展接入接口进行连接构成一个片上互联总线完成;另一方面通过添加同一个总线或去除主机接口电路和从机仲裁选择逻辑电路完成;The expansion or reduction of the host port and the slave port is accomplished, on the one hand, by connecting multiple buses through the bus host expansion interface and the expansion access interface to form an on-chip interconnect bus; on the other hand, by adding the same bus or removing the host interface circuit and the slave arbitration selection logic circuit; 所述从机仲裁选择逻辑电路支持层次化仲裁逻辑,每一个从机接口通过层次化仲裁逻辑电路将连接到该总线上主机分为若干个组,同一组内主机采用一种优先级模式,不同组之间采用另一种优先级模式;所述优先级模式支持固定优先级比较算法或者循环优先级算法。The slave arbitration selection logic circuit supports hierarchical arbitration logic. Each slave interface divides the hosts connected to the bus into several groups through the hierarchical arbitration logic circuit. The hosts in the same group adopt one priority mode, and different groups adopt another priority mode. The priority mode supports a fixed priority comparison algorithm or a circular priority algorithm. 2.如权利要求1所述的可扩展的片上互联总线结构,其特征在于,所述主机接口电路用于向目标主机发送主机访问控制信号、写数据,同时接收来自从机的响应信号、读数据,包括读数据选择电路和主机接口控制状态机;2. The scalable on-chip interconnect bus structure according to claim 1, wherein the host interface circuit is used to send a host access control signal and write data to a target host, and simultaneously receive a response signal and read data from a slave, and comprises a read data selection circuit and a host interface control state machine; 所述主机接口电路的输入信号包括时钟与复位信号、主机数据信号、主机控制信号、从机读数据信号、从机响应信号,输出信号包括主机数据信号、主机控制信号、目标从机读数据信号、从机响应信号、主机状态指示信号;The input signals of the host interface circuit include clock and reset signals, host data signals, host control signals, slave read data signals, and slave response signals, and the output signals include host data signals, host control signals, target slave read data signals, slave response signals, and host status indication signals; 所述主机接口电路根据从机选择信号,将所述目标从机读数据信号输出至主机,将所述主机数据信号与所述主机控制信号输出至所述从机仲裁选择逻辑电路中;所述主机接口控制状态机将所述从机响应信号与所述主机控制信号作为输入,控制所述主机接口电路状态,输出所述从机响应信号、所述主机状态指示信号、当前从机选择信号。The host interface circuit outputs the target slave read data signal to the host according to the slave selection signal, and outputs the host data signal and the host control signal to the slave arbitration selection logic circuit; the host interface control state machine takes the slave response signal and the host control signal as input, controls the state of the host interface circuit, and outputs the slave response signal, the host state indication signal, and the current slave selection signal. 3.如权利要求2所述的可扩展的片上互联总线结构,其特征在于,所述读数据选择电路的输入信号为总线上所有从机读数据信号及从机选择信号,输出信号为目标从机读数据信号;3. The scalable on-chip interconnect bus structure according to claim 2, wherein the input signal of the read data selection circuit is the read data signal of all slaves on the bus and the slave selection signal, and the output signal is the read data signal of the target slave; 所述读数据选择电路根据所述从机选择信号,从输入的所有从机读数据信号选择目标从机读数据信号;将其输出至对应主机中;当前主机无访问请求时,所述读数据选择电路的输出保持恒定状态。The read data selection circuit selects a target slave read data signal from all input slave read data signals according to the slave selection signal, and outputs it to the corresponding host; when the current host has no access request, the output of the read data selection circuit remains constant. 4.如权利要求2所述的可扩展的片上互联总线结构,其特征在于,所述主机接口控制状态机的输入信号为总线的时钟与复位信号、从机响应信号及主机控制信号,输出信号为主机状态指示信号、从机响应信号及从机选择信号;4. The expandable on-chip interconnect bus structure according to claim 2, wherein the input signals of the host interface control state machine are the clock and reset signals of the bus, the slave response signal and the host control signal, and the output signals are the host status indication signal, the slave response signal and the slave selection signal; 主机数据信号、主机控制信号从主机输出,经过主机接口电路后,输出至各个从机仲裁选择逻辑电路中;当前主机无访问请求时,主机控制信号与主机数据信号输出保持恒定状态。The host data signal and host control signal are output from the host, and after passing through the host interface circuit, are output to each slave arbitration selection logic circuit; when the current host has no access request, the host control signal and host data signal output remain in a constant state. 5.如权利要求1所述的可扩展的片上互联总线结构,其特征在于,所述从机仲裁选择逻辑电路用于接收来自总线上各主机的数据与控制信号,基于优先级模式与配置给出同时访问的多个主机的访问优先级,根据同时访问的各主机优先级选择当前有效的主机访问控制与写数据输出至从机,同将从机的响应信号与读数据输出至对应的主机端口,包括主机选择电路、层次化仲裁逻辑电路、从机仲裁选择逻辑状态机及从机寄存器;5. The scalable on-chip interconnect bus structure as claimed in claim 1, characterized in that the slave arbitration selection logic circuit is used to receive data and control signals from each host on the bus, give access priorities of multiple hosts accessed simultaneously based on priority mode and configuration, select the currently valid host access control and write data to output to the slave according to the priority of each host accessed simultaneously, and output the response signal and read data of the slave to the corresponding host port, including a host selection circuit, a hierarchical arbitration logic circuit, a slave arbitration selection logic state machine and a slave register; 所述从机仲裁选择逻辑电路的输入信号包括总线上所有主机数据与控制信号、从机配置信号、从机读数据信号与响应信号、从机配置时钟与复位信号、从机仲裁选择逻辑电路工作时钟与复位信号,输出信号包括当前主机数据与控制信号、从机访问状态信号、从机读数据信号与响应信号;The input signals of the slave arbitration selection logic circuit include all host data and control signals on the bus, slave configuration signals, slave read data signals and response signals, slave configuration clock and reset signals, slave arbitration selection logic circuit working clock and reset signals, and the output signals include current host data and control signals, slave access status signals, slave read data signals and response signals; 所述从机仲裁选择逻辑电路依据从机状态与主机控制信号,基于优先级模式与配置信号,在总线上全部主机端口中选择有效的主机端口,并将该主机数据与控制信号输出至从机。The slave arbitration selection logic circuit selects a valid host port from all host ports on the bus according to the slave state and the host control signal, based on the priority mode and the configuration signal, and outputs the host data and control signal to the slave. 6.如权利要求5所述的可扩展的片上互联总线结构,其特征在于,主机选择电路的输入信号包括总线上所有主机数据与控制信号、主机选择信号,输出信号为当前主机数据与控制信号;6. The expandable on-chip interconnect bus structure as claimed in claim 5, characterized in that the input signal of the host selection circuit includes all host data and control signals on the bus and the host selection signal, and the output signal is the current host data and control signal; 所述主机选择信号由从机仲裁选择逻辑状态机生成,所述主机选择电路依据主机选择信号,确定当前有效的主机端口并给予其对当前主机的控制权限,输出当前主机数据与控制信号;从机无主机端口访问时,当前主机数据与控制信号输出保持恒定状态。The host selection signal is generated by the slave arbitration selection logic state machine. The host selection circuit determines the current valid host port and grants it control authority over the current host based on the host selection signal, and outputs the current host data and control signals; when the slave has no host port access, the current host data and control signal outputs remain in a constant state. 7.如权利要求5所述的可扩展的片上互联总线结构,其特征在于,所述从机寄存器的输入信号包括从机配置信号、从机配置时钟与复位信号;所述从机配置信号包括从机仲裁模式、仲裁优先级配置;7. The scalable on-chip interconnect bus structure according to claim 5, characterized in that the input signal of the slave register includes a slave configuration signal, a slave configuration clock and a reset signal; the slave configuration signal includes a slave arbitration mode and an arbitration priority configuration; 所述从机寄存器保存从机仲裁模式与仲裁优先级的配置,并将优先级模式与配置信号输出至所述层次化仲裁逻辑电路。The slave register stores the configuration of the slave arbitration mode and arbitration priority, and outputs the priority mode and configuration signals to the hierarchical arbitration logic circuit. 8.如权利要求5所述的可扩展的片上互联总线结构,其特征在于,所述从机仲裁选择逻辑状态机的输入信号为从机仲裁选择逻辑电路工作时钟与复位信号、层次化仲裁逻辑电路的仲裁结果,输出为从机访问状态信号;8. The scalable on-chip interconnect bus structure according to claim 5, characterized in that the input signal of the slave arbitration selection logic state machine is the slave arbitration selection logic circuit working clock and reset signal, and the arbitration result of the hierarchical arbitration logic circuit, and the output is the slave access state signal; 所述从机仲裁选择逻辑状态机根据所述仲裁结果、当前从机被控状态,产生最新的主机选择信号、从机访问状态信号,最新的主机选择信号输出至所述层次化仲裁逻辑电路及所述主机选择电路。The slave arbitration selection logic state machine generates the latest host selection signal and slave access status signal according to the arbitration result and the current slave controlled state, and the latest host selection signal is output to the hierarchical arbitration logic circuit and the host selection circuit. 9.如权利要求5所述的可扩展的片上互联总线结构,其特征在于,所述层次化仲裁逻辑电路用于实现可扩展总线的层次化仲裁逻辑,其输入信号为主机选择信号、优先级模式与配置信号,输出信号为各个主机访问优先级仲裁结果;9. The scalable on-chip interconnect bus structure as claimed in claim 5, characterized in that the hierarchical arbitration logic circuit is used to implement the hierarchical arbitration logic of the scalable bus, and its input signal is the host selection signal, the priority mode and the configuration signal, and the output signal is the arbitration result of each host access priority; 所述层次化仲裁逻辑电路的每一个从机接口都能够通过层次化仲裁逻辑结构对可扩展总线上所有主机的优先级进行分层配置,对于任意一个从机接口,同一层次下的主机基于一种优先级仲裁逻辑完成仲裁,不同层次之间优先级仲裁逻辑不同,优先级仲裁逻辑算法包括固定优先级比较算法、循环优先级算法两种优先级仲裁算法,不同层次选择不同的仲裁逻辑算法作为该层次的仲裁逻辑;Each slave interface of the hierarchical arbitration logic circuit can hierarchically configure the priorities of all hosts on the expandable bus through the hierarchical arbitration logic structure. For any slave interface, the hosts at the same level complete arbitration based on a priority arbitration logic. The priority arbitration logics at different levels are different. The priority arbitration logic algorithms include two priority arbitration algorithms: a fixed priority comparison algorithm and a circular priority algorithm. Different levels select different arbitration logic algorithms as the arbitration logic of the level. 所述层次化仲裁逻辑电路的工作流程是所有主机选择信号、优先级模式与配置、优先级分组配置信号作为优先级分组配置电路的输入,当主机选择信号有效时其对应的主机ID也作为优先级分配电路的输入;优先级分组配置电路根据主机ID与优先级分组配置信号,对所有主机选择信号进行分组,将组内优先级模式与配置信号、同一组有效主机选择信号输入到层次化仲裁逻辑电路,每组包括i个主机,共有j组,同组主机选择信号经过层次化仲裁逻辑电路,输出组中每个主机的组内优先级;优先级分组配置电路将组内所有主机选择信号输入到或电路,当组内任意一个主机选择信号有效时,该组主机选择信号即为有效,所有组的组主机选择信号输入到层次化仲裁逻辑电路,该电路输出所有组的优先级;组优先级信号与该组中所有主机的组内优先级结果输入到优先级组合电路,优先级组合电路将执行主机组内优先级与对应组优先级的乘运算,输出各个主机的访问优先级,主机的访问优先级作为多个主机同时访问该从机时的仲裁依据。The workflow of the hierarchical arbitration logic circuit is that all host selection signals, priority modes and configurations, and priority group configuration signals are used as inputs of the priority group configuration circuit. When the host selection signal is valid, its corresponding host ID is also used as input of the priority allocation circuit; the priority group configuration circuit groups all host selection signals according to the host ID and the priority group configuration signal, and inputs the priority mode and configuration signal within the group and the valid host selection signals of the same group into the hierarchical arbitration logic circuit. Each group includes i hosts, and there are j groups in total. The host selection signals of the same group pass through the hierarchical arbitration logic circuit, and the group priority of each host in the group is output; the priority group configuration circuit inputs all host selection signals in the group into the OR circuit. When any host selection signal in the group is valid, the host selection signal of the group is valid. The group host selection signals of all groups are input into the hierarchical arbitration logic circuit, and the circuit outputs the priorities of all groups; the group priority signal and the group priority results of all hosts in the group are input into the priority combination circuit. The priority combination circuit will perform the multiplication operation of the host group priority and the corresponding group priority, and output the access priority of each host. The access priority of the host is used as the arbitration basis when multiple hosts access the slave at the same time. 10.如权利要求9所述的可扩展的片上互联总线结构,其特征在于,当所述层次化仲裁逻辑电路采用固定优先级比较算法时,所述层次化仲裁逻辑电路将各主机的输入优先级配置作为各主机优先级仲裁结果直接输出;若两个主机请求访问从机端口,则具有较高优先级的主机将获得对从机端口的控制权限;当主机端口向从机端口发出请求时,从机端口将检查新的请求主机端口的优先级是否高于当前控制从机端口的主机端口的优先级;从机端口在每个时钟边沿执行仲裁检查,如果新请求主机的优先级高于当前控制从机端口的主机优先级,则新请求主机将在下一个时钟边沿获得对从机端口的控制权限;若当前控制从机端口的主机正在执行固定长度的突发传输或锁定的传输,则无论新请求主机的优先级是否高于当前正在执行访问的主机,均须等待突发传输或锁定传输结束后,才能获得对从机端口的控制权限;10. The scalable on-chip interconnect bus structure as claimed in claim 9, characterized in that when the hierarchical arbitration logic circuit adopts a fixed priority comparison algorithm, the hierarchical arbitration logic circuit directly outputs the input priority configuration of each host as the arbitration result of each host priority; if two hosts request to access the slave port, the host with a higher priority will obtain the control authority over the slave port; when the host port sends a request to the slave port, the slave port will check whether the priority of the new requesting host port is higher than the priority of the host port currently controlling the slave port; the slave port performs arbitration check at each clock edge, and if the priority of the new requesting host is higher than the priority of the host currently controlling the slave port, the new requesting host will obtain the control authority over the slave port at the next clock edge; if the host currently controlling the slave port is performing a fixed-length burst transmission or a locked transmission, then regardless of whether the priority of the new requesting host is higher than that of the host currently performing access, it must wait until the burst transmission or the locked transmission ends before obtaining the control authority over the slave port; 当所述层次化仲裁逻辑电路采用循环优先级算法时,每个主机端口根据物理主机端口号分配相对优先级,并将所述相对优先级与在总线上执行传输的最后一个主机端口进行比较,最高优先级请求主机获得下一个传输边界从机的控制权限;当主机获得目标从机的控制权限后,主机能够向该端口执行数据传输,直到另一个主机向同一从机端口发出请求为止;下一个主机被允许在下一个传输边界访问从机端口,若当前主线没有挂起的访问请求,则在下一个时钟周期中访问。When the hierarchical arbitration logic circuit adopts a round-robin priority algorithm, each host port is assigned a relative priority according to the physical host port number, and the relative priority is compared with the last host port that performs transmission on the bus. The highest priority request host obtains control authority for the next transmission boundary slave; when the host obtains control authority for the target slave, the host can perform data transmission to the port until another host sends a request to the same slave port; the next host is allowed to access the slave port at the next transmission boundary, and if there is no pending access request on the current master line, it is accessed in the next clock cycle.
CN202410753080.1A 2024-06-12 2024-06-12 Extensible on-chip interconnection bus structure Active CN118331905B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202410753080.1A CN118331905B (en) 2024-06-12 2024-06-12 Extensible on-chip interconnection bus structure

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202410753080.1A CN118331905B (en) 2024-06-12 2024-06-12 Extensible on-chip interconnection bus structure

Publications (2)

Publication Number Publication Date
CN118331905A CN118331905A (en) 2024-07-12
CN118331905B true CN118331905B (en) 2024-08-06

Family

ID=91782827

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202410753080.1A Active CN118331905B (en) 2024-06-12 2024-06-12 Extensible on-chip interconnection bus structure

Country Status (1)

Country Link
CN (1) CN118331905B (en)

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107707491A (en) * 2017-09-28 2018-02-16 中国人民解放军国防科技大学 Device and method for realizing multilevel on-chip interconnection
CN112106035A (en) * 2018-04-03 2020-12-18 赛灵思公司 System-on-chip interface architecture

Family Cites Families (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8307147B2 (en) * 2005-09-09 2012-11-06 Freescale Semiconductor, Inc. Interconnect and a method for designing an interconnect
KR101620460B1 (en) * 2010-05-04 2016-05-13 삼성전자주식회사 Interconnect, bus system having interconnect and operating method of bus system
US8667197B2 (en) * 2010-09-08 2014-03-04 Intel Corporation Providing a fine-grained arbitration system
US20230297421A1 (en) * 2022-03-18 2023-09-21 Intel Corporation Hard partitioning via intra-soc composition
CN117155802A (en) * 2023-08-30 2023-12-01 山东云海国创云计算装备产业创新中心有限公司 Out-of-order transmission simulation method, device and system and electronic equipment
CN117806999B (en) * 2024-01-11 2025-05-06 电子科技大学 Bit width and channel adjustable on-chip bus
CN117951061A (en) * 2024-01-31 2024-04-30 重庆邮电大学 Arbitration method for multichannel data bus
CN118114615B (en) * 2024-04-30 2024-06-28 西北工业大学 Control and calculation separated system chip structure

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107707491A (en) * 2017-09-28 2018-02-16 中国人民解放军国防科技大学 Device and method for realizing multilevel on-chip interconnection
CN112106035A (en) * 2018-04-03 2020-12-18 赛灵思公司 System-on-chip interface architecture

Also Published As

Publication number Publication date
CN118331905A (en) 2024-07-12

Similar Documents

Publication Publication Date Title
CN100481052C (en) Internal bus system
CN100499556C (en) High-speed asynchronous interlinkage communication network of heterogeneous multi-nucleus processor
US8725919B1 (en) Device configuration for multiprocessor systems
JPH0713945A (en) Bus structure of multiprocessor system with separated arithmetic processing part and control/storage part
CN112597075A (en) Cache allocation method for router, network on chip and electronic equipment
US8213461B2 (en) Method of designating slots in a transmission frame for controlling transmission of data over an interconnect coupling a plurality of master units with a plurality of slave units
US10496565B2 (en) Micro-architectural techniques to minimize companion die firmware loading times in a server platform
CN117093438A (en) test system
CN115840592A (en) Flash access method, controller, system and readable storage medium
JP2004078929A (en) System which optimizes performance in processor level, method, and apparatus
CN100481048C (en) Bus arbitration method
CN116737604A (en) Memory control chips, server memory modules and servers
CN113688084B (en) A circuit, method and storage medium for realizing automatic allocation of PCIE resources
CN118331905B (en) Extensible on-chip interconnection bus structure
CN114661654A (en) Access processing device and method, processing device, electronic device, and storage medium
CN118114615B (en) Control and calculation separated system chip structure
CN112202600B (en) A multi-node single host and multi-host communication automatic switching device and method
CN117806999B (en) Bit width and channel adjustable on-chip bus
CN117609114B (en) Multi-core processor data sharing control method, device, storage module and chip
CN118349501A (en) An efficient data interaction mechanism for heterogeneous multi-computing core processors
CN117707426A (en) A chip, IO resource sharing method, system on a chip and electronic equipment
CN118430603A (en) Decoding method, first bare chip and second bare chip
Yin et al. A reconfigurable rack-scale interconnect architecture based on PCIe fabric
JP2780662B2 (en) Multiprocessor system
CN118210750B (en) A multi-bus interconnection architecture based on communication performance optimization

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant