[go: up one dir, main page]

CN105468562B - Chipset and server system - Google Patents

Chipset and server system Download PDF

Info

Publication number
CN105468562B
CN105468562B CN201510779281.XA CN201510779281A CN105468562B CN 105468562 B CN105468562 B CN 105468562B CN 201510779281 A CN201510779281 A CN 201510779281A CN 105468562 B CN105468562 B CN 105468562B
Authority
CN
China
Prior art keywords
management controller
server
node
embedded
server node
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
CN201510779281.XA
Other languages
Chinese (zh)
Other versions
CN105468562A (en
Inventor
秦双双
杨国俊
林皓琳
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.)
Shanghai Zhaoxin Semiconductor Co Ltd
Original Assignee
Shanghai Zhaoxin Integrated Circuit Co Ltd
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 Shanghai Zhaoxin Integrated Circuit Co Ltd filed Critical Shanghai Zhaoxin Integrated Circuit Co Ltd
Priority to CN201510779281.XA priority Critical patent/CN105468562B/en
Publication of CN105468562A publication Critical patent/CN105468562A/en
Application granted granted Critical
Publication of CN105468562B publication Critical patent/CN105468562B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR 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
    • G06F13/42Bus transfer protocol, e.g. handshake; Synchronisation
    • G06F13/4265Bus transfer protocol, e.g. handshake; Synchronisation on a point to point bus

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Debugging And Monitoring (AREA)

Abstract

应用于服务器系统的服务器节点的芯片组,具有南桥、北桥、且还具有一嵌入式管理控制器。该嵌入式管理控制器收集服务器节点的节点内部信息,供服务器系统管理使用。该嵌入式管理控制器耦接一底板管理控制器。该底板管理控制器设置在服务器节点外部,通过网络与远程终端通讯。

The chip set applied to the server node of the server system has a south bridge, a north bridge, and an embedded management controller. The embedded management controller collects node internal information of the server node for server system management. The embedded management controller is coupled to a baseboard management controller. The baseboard management controller is arranged outside the server node, and communicates with the remote terminal through the network.

Description

芯片组以及服务器系统Chipsets and Server Systems

技术领域technical field

本发明涉及服务器系统管理(server system management)技术领域,尤其涉及一种芯片组以及服务器系统。The present invention relates to the technical field of server system management (server system management), in particular to a chipset and a server system.

背景技术Background technique

在一服务器系统中,一项重要课题为监控网络以及每一服务器节点的物理状态。服务器系统上常常会应用智能平台管理界面(Intelligent Platform ManagementInterface,简称IPMI)协议,但其通讯协议非常复杂。低成本且低复杂度的服务器系统管理架构为本技术领域一项重要课题。In a server system, an important issue is to monitor the network and the physical status of each server node. The intelligent platform management interface (Intelligent Platform Management Interface, referred to as IPMI) protocol is often used on the server system, but its communication protocol is very complicated. A low-cost and low-complexity server system management architecture is an important topic in this technical field.

发明内容Contents of the invention

本发明提供一种芯片组以及使用该芯片组的服务器系统。The invention provides a chip set and a server system using the chip set.

本发明一种实施方式提供的一芯片组是应用于一服务器系统的一服务器节点,且具有一嵌入式管理控制器。该芯片组还包括南桥以及北桥。该嵌入式管理芯片组收集服务器节点的节点内部信息,供服务器系统管理使用。该嵌入式管理控制器耦接一底板管理控制器。该底板管理控制器设置在服务器节点外部,通过网络与远程终端通讯。A chipset provided by an embodiment of the present invention is applied to a server node of a server system and has an embedded management controller. The chipset also includes a southbridge as well as a northbridge. The embedded management chipset collects internal information of the server node for server system management. The embedded management controller is coupled to a baseboard management controller. The baseboard management controller is arranged outside the server node, and communicates with the remote terminal through the network.

本发明另一种实施方式是一服务器系统,包括复数个服务器节点、一底板管理控制器以及一开关模块。所述服务器节点各自包括一芯片组。各芯片组包括南桥与北桥。服务器节点上的芯片组是与一嵌入式管理控制器整合入一芯片。上述嵌入式管理控制器是用于收集对应的服务器节点的节点内部信息,供服务器系统管理使用。底板管理控制器设置于所述服务器节点外,经网络与一远程终端通讯。该开关模块选择性地将不同服务器节点的嵌入式管理控制器连结至该底板管理控制器。Another embodiment of the present invention is a server system, including a plurality of server nodes, a baseboard management controller and a switch module. The server nodes each include a chipset. Each chipset includes a south bridge and a north bridge. The chipset on the server node is integrated into a chip with an embedded management controller. The above-mentioned embedded management controller is used to collect node internal information of the corresponding server node for use in server system management. The baseboard management controller is arranged outside the server node, and communicates with a remote terminal through the network. The switch module selectively connects the embedded management controllers of different server nodes to the baseboard management controller.

在一种实施方式中,服务器系统包括一刀片(blade)。所述多个服务器节点、开关模块、以及底板管理控制器是设置在该刀片上,形成刀片服务器。In one embodiment, the server system includes a blade. The plurality of server nodes, switch modules, and baseboard management controllers are arranged on the blade to form a blade server.

在另一种实施方式中,该服务器系统包括一背板(backplane),用以插装复数个刀片(blades)。复数个服务器节点是划分为多个组别,分组设置在所述刀片上。该底板管理控制器是设置在该背板上,而非所述刀片上。该开关模块包括复数个开关子模块,分别设置在所述刀片上。各刀片上的嵌入式管理控制器是经由各刀片上的开关子模块选择性地耦接至该底板管理控制器。In another embodiment, the server system includes a backplane for inserting a plurality of blades. The plurality of server nodes are divided into a plurality of groups, and the groups are arranged on the blades. The baseboard management controller is arranged on the backboard instead of the blade. The switch module includes a plurality of switch sub-modules, which are respectively arranged on the blades. The embedded management controller on each blade is selectively coupled to the baseboard management controller via a switch sub-module on each blade.

在本发明前述实施例的芯片组以及使用如此芯片组的服务器系统中,由于各服务器节点的节点内部信息是经由芯片组所支持的芯片内通讯路径被搜集和传递至服务器节点上设置的嵌入式管理控制器eMC,故服务器节点以及其上的嵌入式管理控制器eMC之间的通讯无需依赖于复杂的智能平台管理接口(IPMI)协议。低成本且低复杂度的服务器系统管理架构因而建立。In the chip set of the foregoing embodiments of the present invention and the server system using such a chip set, since the node internal information of each server node is collected and delivered to the embedded The management controller eMC, so the communication between the server node and the embedded management controller eMC thereon does not need to rely on the complex intelligent platform management interface (IPMI) protocol. A low-cost and low-complexity server system management architecture is thus established.

下文特举实施例,并配合所附图示,详细说明本发明内容。Hereinafter, specific embodiments are cited, and the contents of the present invention are described in detail in conjunction with the accompanying drawings.

附图说明Description of drawings

图1根据本发明一种实施方式图解一种服务器系统100;FIG. 1 illustrates a server system 100 according to an embodiment of the present invention;

图2A根据本发明一种实施方式图解具有一刀片(blade)202的一种服务器系统;FIG. 2A illustrates a server system with a blade 202 according to an embodiment of the present invention;

图2B根据本发明另一种实施方式图解一服务器系统,具有一背板(backplane)204、以及插装于该背板204上的复数个刀片(blades)Blade1与Blade2;Fig. 2B illustrates a server system according to another embodiment of the present invention, has a backplane (backplane) 204, and a plurality of blades (blades) Blade1 and Blade2 inserted on the backplane 204;

图3为方块图,图解根据本发明一种实施方式所实现的嵌入式管理控制器eMC;3 is a block diagram illustrating an embedded management controller eMC implemented according to an embodiment of the present invention;

图4为流程图,图解嵌入式管理控制器eMC如何将节点内部信息经第一开关(例如I2C开关)传送至该底板管理控制器BMC;4 is a flowchart illustrating how the embedded management controller eMC transmits the internal information of the node to the baseboard management controller BMC through a first switch (such as an I2C switch);

图5图解非挥发性存储器312的空间配置,其中,该非挥发性存储器312是由一服务器节点以及该服务器节点上所安装的一嵌入式管理控制器eMC共享;FIG. 5 illustrates the spatial configuration of the non-volatile memory 312, wherein the non-volatile memory 312 is shared by a server node and an embedded management controller eMC installed on the server node;

图6图解非挥发性存储器312的空间配置,其中,该非挥发性存储器312是由复数个服务器节点以及所述服务器节点上安装的复数个嵌入式管理控制器eMC共享;Fig. 6 illustrates the spatial configuration of the non-volatile memory 312, wherein the non-volatile memory 312 is shared by a plurality of server nodes and a plurality of embedded management controllers eMC installed on the server nodes;

图7A图解一逻辑电路,用于设定一服务器系统的开机顺序,该服务器系统具有复数个服务器节点Node1…Node4,且不同服务器节点Node1…Node4上的芯片组以及嵌入式管理控制器eMC如图6所示共享同一闪存312;以及FIG. 7A illustrates a logic circuit for setting the boot sequence of a server system. The server system has a plurality of server nodes Node1...Node4, and chipsets and embedded management controllers eMC on different server nodes Node1...Node4 are shown in the figure. 6 share the same flash memory 312; and

图7B根据本发明一种实施方式图解服务器系统700上所安装的组件的开机顺序。FIG. 7B illustrates a boot sequence of components installed on server system 700 according to one embodiment of the present invention.

附图说明:Description of drawings:

100:服务器系统; 102:远程终端;100: server system; 102: remote terminal;

104:网络; 202:刀片(blade);104: network; 202: blade;

204:背板(backplane); 302:节点内部信息收集模块;204: backplane; 302: node internal information collection module;

304:第一通讯接口控制器; 306:第二通讯接口控制器;304: the first communication interface controller; 306: the second communication interface controller;

308:存储器接口控制器; 312:非挥发性存储器;308: memory interface controller; 312: non-volatile memory;

314:多工器; 402…414:步骤;314: multiplexer; 402...414: steps;

502:系统的基本输入输出系统(system BIOS);502: basic input and output system (system BIOS) of the system;

602:服务器节点基本输入输出系统的读取区;602: the reading area of the basic input and output system of the server node;

604:服务器节点Node1的写入区;604: the writing area of the server node Node1;

606:服务器节点Node2的写入区;606: the writing area of the server node Node2;

608:服务器节点Node3的写入区;608: the writing area of the server node Node3;

610:服务器节点Node4的写入区;610: the writing area of the server node Node4;

702…738:步骤;702...738: steps;

BMC:底板管理控制器; CI_1…CI_M:接口模块;BMC: baseboard management controller; CI_1…CI_M: interface module;

CPU:中央处理器; CS:控制信号;CPU: central processing unit; CS: control signal;

eMC:嵌入式管理控制器; Image1、Image2:固件映像;eMC: Embedded Management Controller; Image1, Image2: Firmware image;

MCU:微控制器; Mux:多工器;MCU: microcontroller; Mux: multiplexer;

NB:北桥; Node1…NodeM:服务器节点;NB: North Bridge; Node1...NodeM: server node;

SB:南桥; SB_Bus:南桥SB所支持的芯片内路径;SB: South Bridge; SB_Bus: In-chip path supported by South Bridge SB;

SW:开关模块; SW1、SW2:第一、第二开关。SW: switch module; SW1, SW2: first and second switches.

具体实施方式Detailed ways

以下叙述列举本发明的多种实施例。以下叙述介绍本发明的基本概念,且并非意图限制本发明内容。实际发明范围应依照申请专利范围而界定。The following description lists various embodiments of the present invention. The following description introduces the basic concept of the present invention and is not intended to limit the content of the present invention. The actual scope of the invention should be defined according to the scope of the patent application.

图1图解根据本发明一种实施方式所实现的一服务器系统100。服务器系统100包括复数个服务器节点Node1…NodeM、一底板管理控制器BMC以及一开关模块SW。FIG. 1 illustrates a server system 100 implemented according to an embodiment of the present invention. The server system 100 includes a plurality of server nodes Node1 . . . NodeM, a baseboard management controller BMC, and a switch module SW.

服务器节点Node1…NodeM可为相似架构。每个服务器节点包括一中央处理单元CPU、还包括北桥NB以及南桥SB组成的一芯片组(chipset)。北桥NB与南桥SB可为两芯片、也可制作在单一芯片上。嵌入式管理控制器eMC对应设置在各服务器节点上。在某些实施例中,嵌入式管理控制器eMC整合至含南桥SB的芯片中。在图1所示的实施方式中,嵌入式管理控制器eMC整合在芯片组内。嵌入式管理控制器eMC各自收集所对应的服务器节点的节点内部信息,供服务器系统管理(server system management)使用。上述节点内部信息可包括中央处理器CPU、北桥NB或南桥SB的信息。此外,设置在各服务器节点上的传感器(如,风扇传感器)的感测数据也可由对应的嵌入式管理控制器eMC收集,作为节点内部信息。由于各服务器节点的节点内部信息是经由芯片组所支持的芯片内通讯路径传递,故服务器节点以及其上所安装的嵌入式管理控制器eMC之间的通讯无需依赖于前述的智能平台管理接口(IPMI)协议。低成本且低复杂度的服务器系统管理架构因而建立。此外,除了节点内部信息的收集(如,收集且报导(collecting and reporting)节点内部状态),嵌入式管理控制器eMC还可支持采用局域网络(LAN)或串行局域网络(SOL,serial over LAN简写)的远程文本终端(remote text console);远程重开机(reboot)/上电(power-on)/断电(power-off)操作、远程决策的基本输入输出系统设定(remote audit BIOS setting)、事件日志(SEL)或现场可更换单元(FRU)、基本输入输出系统开机顺序调整(BIOS boot ordermodification)、或以上操作的任意组合。The server nodes Node1 . . . NodeM may be of similar architecture. Each server node includes a central processing unit CPU, and a chip set (chipset) composed of a north bridge NB and a south bridge SB. The north bridge NB and the south bridge SB can be two chips, or can be fabricated on a single chip. The embedded management controller eMC is correspondingly set on each server node. In some embodiments, the embedded management controller eMC is integrated into a chip including the south bridge SB. In the embodiment shown in FIG. 1, the embedded management controller eMC is integrated in the chipset. The embedded management controllers eMC respectively collect node internal information of the corresponding server nodes for use by server system management. The above-mentioned internal information of the node may include the information of the central processing unit CPU, the north bridge NB or the south bridge SB. In addition, the sensing data of sensors (eg, fan sensors) arranged on each server node may also be collected by the corresponding embedded management controller eMC as internal information of the node. Since the internal information of each server node is transmitted through the on-chip communication path supported by the chipset, the communication between the server node and the embedded management controller eMC installed on it does not need to rely on the aforementioned intelligent platform management interface ( IPMI) protocol. A low-cost and low-complexity server system management architecture is thus established. In addition, in addition to the collection of internal information of nodes (for example, collecting and reporting (collecting and reporting) the internal status of nodes), the embedded management controller eMC can also support the use of local area network (LAN) or serial local area network (SOL, serial over LAN) abbreviation) remote text console (remote text console); remote reboot/power-on/power-off operations, remote decision-making basic input and output system settings (remote audit BIOS setting ), event log (SEL) or field replaceable unit (FRU), BIOS boot order modification, or any combination of the above operations.

如图所示,底板管理控制器BMC设置在服务器节点Node1…NodeM外部,经由网络104与远程终端102通讯。该开关模块SW将不同服务器节点Node1…NodeM上的嵌入式管理控制器(全标号为eMC)选择性地耦接至该底板管理控制器BMC。As shown in the figure, the baseboard management controller BMC is arranged outside the server nodes Node1 . . . NodeM, and communicates with the remote terminal 102 via the network 104 . The switch module SW selectively couples embedded management controllers (full symbol eMC) on different server nodes Node1 . . . NodeM to the baseboard management controller BMC.

在图1所示实施方式中,开关模块SW对应一第一通讯接口(如,内部整合电路接口,简称I2C)提供一第一开关SW1、且对应一第二通讯接口(如,通用异步收发传输器,简称UART)提供一第二开关SW2。经由第一开关SW1,不同服务器节点Node1…NodeM的节点内部信息得以传递至该底板管理控制器BMC,且节点外部信息(如,远程终端102发出的节点控制指令以及其它信息)得以自底板管理控制器BMC传送至服务器节点Node1至NodeM。如图所示,第一开关SW1可具有复数个接口模块CI_1…CI_M(如,I2C模块)以及一多工器Mux。接口模块CI_1…CI_M分别耦接不同服务器节点Node_1…Node_M上的嵌入式管理控制器eMC。多工器Mux用于选择性地将所述接口模块CI_1…CI_M之一耦接至该底板管理控制器BMC。特别说明的,嵌入式管理控制器eMC是以一第一串行总线(如,I2C总线)耦接该第一开关SW1以及该底板管理控制器BMC,第一串行总线的通讯协议远比IPMI协议简单。整个服务器系统的连接以及通讯因而简化;此外,以上通讯可由底板管理控制器BMC发起、或是由嵌入式管理控制器eMC发起。稍后将参考图4详细讨论之。至于该底板管理控制器BMC为了服务器系统管理所传送出的某些特定节点外部要求(request),则是由第二开关SW2决定如何由该底板管理控制器BMC传递至所述节点Node1…NodeM。举例来说,SOL操作为一种特定节点外部要求,用以使能一服务器节点的串行端口的输入与输出经由互联网络协议(IP)重新导向。如此的特定节点外部要求的通讯是由该底板管理控制器BMC发起,但不会由嵌入式管理控制器eMC发起。第二开关SW2用于建立该底板管理控制器BMC与不同服务器节点Node1…NodeM上的嵌入式管理控制器eMC之间的连接,以经由第二通讯接口传输上述节点外部要求。特别说明的,嵌入式管理控制器eMC是经由一第二串行总线(如UART总线)耦接第二开关SW2以及底板管理控制器BMC。第二串行总线的通讯协议也远简单于IPMI协议。In the embodiment shown in FIG. 1 , the switch module SW corresponds to a first communication interface (such as an internal integrated circuit interface, referred to as I2C) to provide a first switch SW1 and corresponds to a second communication interface (such as Universal Asynchronous Transceiver Transmission device, referred to as UART) provides a second switch SW2. Through the first switch SW1, the node internal information of different server nodes Node1...NodeM can be transmitted to the baseboard management controller BMC, and the node external information (such as node control commands and other information sent by the remote terminal 102) can be controlled from the baseboard management The server BMC is sent to the server nodes Node1 to NodeM. As shown in the figure, the first switch SW1 may have a plurality of interface modules CI_1 . . . CI_M (eg, I2C module) and a multiplexer Mux. The interface modules CI_1...CI_M are respectively coupled to the embedded management controllers eMC on different server nodes Node_1...Node_M. The multiplexer Mux is used for selectively coupling one of the interface modules CI_1 . . . CI_M to the baseboard management controller BMC. Specifically, the embedded management controller eMC is coupled to the first switch SW1 and the baseboard management controller BMC with a first serial bus (eg, I2C bus). The protocol is simple. The connection and communication of the whole server system are thus simplified; in addition, the above communication can be initiated by the baseboard management controller BMC or the embedded management controller eMC. This will be discussed in detail later with reference to FIG. 4 . As for some specific node external requests sent by the BMC for server system management, the second switch SW2 determines how the BMC sends them to the nodes Node1 . . . NodeM. For example, SOL operation is a node-specific external requirement to enable serial port input and output of a server node to be redirected via Internet Protocol (IP). Such communications required externally to a particular node are initiated by the baseboard management controller BMC, but not by the embedded management controller eMC. The second switch SW2 is used to establish a connection between the baseboard management controller BMC and the embedded management controllers eMC on different server nodes Node1 . Specifically, the embedded management controller eMC is coupled to the second switch SW2 and the baseboard management controller BMC via a second serial bus (such as a UART bus). The communication protocol of the second serial bus is also far simpler than the IPMI protocol.

在一种实施方式中,底板管理控制器BMC可与复数个服务器节点设置在一刀片(blade)上。图2A根据本发明一种实施方式所实现的一服务器系统,包括一刀片202。该刀片202上设置复数个服务器节点Node1…Node4、一开关模块(包括第一以及第二开关SW1以及SW2)以及一底板管理控制器BMC。远程终端102是经由底板管理控制器BMC上的管理网络接口卡(management network interface card)与不同服务器节点Node1…Node4上的嵌入式管理控制器eMC通讯,以收集节点内部信息、并控制节点内部状态。在一种实施方式中,第一开关SW1是基于I2C协议实现,且第二开关SW2是基于UART协议实现。底板管理控制器BMC经I2C总线自所述嵌入式管理控制器eMC取得节点内部信息,并将取得的节点内部信息依照IPMI协议经网络104传送至远程终端102。SOL以及某些其它特定操作是立基于该底板管理控制器BMC以及服务器节点Node1…Node4之间的UART总线。特别说明的,刀片202虽然仅绘制4个服务器节点,但并不意图限定刀片上的服务器节点数量。刀片上可设置其它数量的服务器节点。In one embodiment, the baseboard management controller BMC and a plurality of server nodes can be set on a blade. FIG. 2A is a server system implemented according to an embodiment of the present invention, including a blade 202 . The blade 202 is provided with a plurality of server nodes Node1 . . . Node4 , a switch module (including first and second switches SW1 and SW2 ), and a baseboard management controller BMC. The remote terminal 102 communicates with the embedded management controllers eMC on different server nodes Node1...Node4 via the management network interface card on the baseboard management controller BMC to collect internal information of the nodes and control the internal status of the nodes . In one embodiment, the first switch SW1 is implemented based on the I2C protocol, and the second switch SW2 is implemented based on the UART protocol. The baseboard management controller BMC obtains the internal information of the node from the embedded management controller eMC via the I2C bus, and transmits the obtained internal information of the node to the remote terminal 102 via the network 104 according to the IPMI protocol. SOL and some other specific operations are based on the UART bus between the Baseboard Management Controller BMC and server nodes Node1...Node4. In particular, although the blade 202 only draws 4 server nodes, it is not intended to limit the number of server nodes on the blade. Other numbers of server nodes can be set on the blade.

如图1所示,服务器系统100的服务器节点Node1…NodeM可划分为复数组,不同组的服务器节点设置在不同刀片(blade)上。图2B图解一种服务器系统,具有一背板(backplane)204以及插装于背板204上的复数个刀片Blade1以及Blade2。如图所示,刀片Blade1上供应一组服务器节点Node1以及Node2,且刀片Blade2上供应一组服务器节点Node3以及Node4。底板管理控制器BMC是设置在背板204上,而非设置在刀片Blade1或刀片Blade2上。开关模块包括复数个开关子模块SW_s1以及SW_s2,分别设置在刀片Blade1以及刀片Blade2上,使各刀片上的嵌入式管理控制器eMC得以耦接至该底板管理控制器BMC。如图所示,各开关子模块SW_s1以及SW_s2包括一第一开关以及一第二开关。如前述讨论,第一开关可依循I2C协定且第二开关可依循UART协定。特别说明的,假设各个刀片具有同样数量服务器节点(例如每个刀片具有4个服务器节点),图2B背板204上的底板管理控制器BMC需比图2A刀片202上的底板管理控制器BMC强大。但相比的,图2B服务器系统较图2A服务器系统低成本,因为图2A服务器系统的各刀片上设置了专属的底板管理控制器BMC。As shown in FIG. 1 , the server nodes Node1 . . . NodeM of the server system 100 can be divided into multiple groups, and server nodes of different groups are arranged on different blades. FIG. 2B illustrates a server system with a backplane 204 and a plurality of blades Blade1 and Blade2 inserted on the backplane 204 . As shown in the figure, a group of server nodes Node1 and Node2 are provided on the blade Blade1, and a group of server nodes Node3 and Node4 are provided on the blade Blade2. The baseboard management controller BMC is set on the backplane 204 instead of the blade Blade1 or the blade Blade2. The switch module includes a plurality of switch sub-modules SW_s1 and SW_s2, respectively disposed on the blade Blade1 and the blade Blade2, so that the embedded management controller eMC on each blade can be coupled to the baseboard management controller BMC. As shown in the figure, each switch sub-module SW_s1 and SW_s2 includes a first switch and a second switch. As previously discussed, the first switch can follow the I2C protocol and the second switch can follow the UART protocol. In particular, assuming that each blade has the same number of server nodes (for example, each blade has 4 server nodes), the baseboard management controller BMC on the backplane 204 of FIG. 2B needs to be stronger than the baseboard management controller BMC on the blade 202 of FIG. 2A . But in comparison, the server system in FIG. 2B is cheaper than the server system in FIG. 2A because each blade of the server system in FIG. 2A is provided with a dedicated baseboard management controller BMC.

图3为方块图,根据本发明一种实施方式图解一嵌入式管理控制器eMC。嵌入式管理控制器eMC包括一微控制器MCU、一节点内部信息收集模块302、两个通讯接口控制器(包括一第一通讯接口控制器304以及一第二通讯接口控制器306)以及一存储器接口控制器308。节点内部信息收集模块302是由微控制器MCU操作,遵循芯片组所支持的芯片内部通讯协议,以收集服务器节点上组件(如,中央处理单元CPU、南桥SB、北桥NB以及风扇等)的节点内部信息。第一通讯接口控制器304是由微控制器MCU控制,经第一开关SW1传送节点内部信息至底板管理控制器BMC,且经第一开关SW1自底板管理控制器BMC接收节点外部信息。特别说明的,第一通讯接口控制器304是经由第一串行总线(如,I2C总线)耦接第一开关SW1以及底板管理控制器BMC。第一串行总线的通讯协议远较IPMI协议简易。第二通讯接口控制器306耦接该第二开关SW2,且是由微控制器MCU控制,以接收底板管理控制器BMC为了服务器系统管理所发来的节点外部要求。特别说明的,第二通讯接口控制器306是经由一第二串行总线(如UART总线)耦接该第二开关SW2以及该底板管理控制器BMC。第二串行总线的通讯协议远较IPMI协议简易。存储器接口控制器308是由微控制器MCU控制,以存取一非挥发性存储器312。非挥发性存储器312是由服务器节点以及服务器节点上的嵌入式管理控制器eMC共享。非挥发性存储器312可为闪存,且该存储器接口控制器308可依循一串行外设接口(SPI)协议。多工器314用于设定该非挥发性存储器312的存取路径。经由切换控制信号CS,非挥发性存储器312可由该嵌入式管理控制器eMC经该存储器接口控制器308存取、或是由南桥SB经由南桥SB所支持的芯片内路径SB_Bus存取。FIG. 3 is a block diagram illustrating an embedded management controller eMC according to one embodiment of the present invention. The embedded management controller eMC includes a microcontroller MCU, a node internal information collection module 302, two communication interface controllers (including a first communication interface controller 304 and a second communication interface controller 306) and a memory interface controller 308 . The node internal information collection module 302 is operated by the microcontroller MCU, and follows the chip internal communication protocol supported by the chipset to collect information on components (such as the central processing unit CPU, South Bridge SB, North Bridge NB and fans, etc.) on the server node. Node internal information. The first communication interface controller 304 is controlled by the microcontroller MCU, transmits node internal information to the baseboard management controller BMC through the first switch SW1, and receives node external information from the baseboard management controller BMC through the first switch SW1. Specifically, the first communication interface controller 304 is coupled to the first switch SW1 and the baseboard management controller BMC via a first serial bus (eg, I2C bus). The communication protocol of the first serial bus is much simpler than the IPMI protocol. The second communication interface controller 306 is coupled to the second switch SW2 and is controlled by the microcontroller MCU to receive external node requirements sent by the baseboard management controller BMC for server system management. Specifically, the second communication interface controller 306 is coupled to the second switch SW2 and the baseboard management controller BMC via a second serial bus (such as a UART bus). The communication protocol of the second serial bus is much simpler than the IPMI protocol. The memory interface controller 308 is controlled by the microcontroller MCU to access a non-volatile memory 312 . The non-volatile memory 312 is shared by the server node and the embedded management controller eMC on the server node. The non-volatile memory 312 can be flash memory, and the memory interface controller 308 can follow a serial peripheral interface (SPI) protocol. The multiplexer 314 is used to set the access path of the non-volatile memory 312 . Through the switching control signal CS, the non-volatile memory 312 can be accessed by the embedded management controller eMC via the memory interface controller 308 , or accessed by the south bridge SB via the on-chip path SB_Bus supported by the south bridge SB.

本段落讨论如何将节点内部信息经嵌入式管理控制器eMC传送至底板管理控制器BMC。由于多个服务器节点的嵌入式管理控制器eMC经由开关模块耦接至同一个底板管理控制器BMC,当某个服务器节点的嵌入式管理控制器eMC需要向底板管理控制器BMC发起传送节点内部信息的传输时,需要使用以下中断机制:在传递节点内部信息至底板管理控制器BMC之前,微控制器MCU确立(assert)一中断状态“INT”,使底板管理控制器BMC得以将该中断状态“INT”确立的服务器节点与其它服务器节点区别。倘若微控制器MCU接收到底板管理控制器BMC输出给中断状态“INT”确立的服务器节点的批准指令“Grant”,微控制器MCU控制该第一通讯接口控制器304开始传送节点内部信息至底板管理控制器BMC。一种实施方式是以8051架构的控制器实现该嵌入式管理控制器eMC。该微控制器MCU可以是8051微控制器。This paragraph discusses how to transmit the internal information of the node to the baseboard management controller BMC through the embedded management controller eMC. Since the embedded management controllers eMC of multiple server nodes are coupled to the same baseboard management controller BMC via the switch module, when the embedded management controller eMC of a certain server node needs to initiate the transmission of node internal information to the baseboard management controller BMC During transmission, the following interrupt mechanism needs to be used: before transmitting the internal information of the node to the baseboard management controller BMC, the microcontroller MCU establishes (assert) an interrupt status "INT", so that the baseboard management controller BMC can use the interrupt status "INT" INT" to distinguish the server node from other server nodes. If the microcontroller MCU receives the approval command “Grant” output by the baseboard management controller BMC to the server node whose interrupt state “INT” is established, the microcontroller MCU controls the first communication interface controller 304 to start transmitting the internal information of the node to the baseboard Management controller BMC. One embodiment implements the embedded management controller eMC with a controller of 8051 architecture. The microcontroller MCU may be an 8051 microcontroller.

图4为流程图,图解嵌入式管理控制器eMC如何经由第一开关(如,I2C开关)传送节点内部信息至底板管理控制器BMC。步骤402,嵌入式管理控制器eMC的微控制器MCU填写该第一通讯接口控制器304的一从属数据输入/输出缓冲缓存器(slave data IO bufferregister),以确立一中断状态“INT”。步骤404,微控制器MCU使用通用输入输出(GPIO)管脚对第一开关SW1显示中断状态“INT”的确立。步骤406,底板管理控制器BMC经由第一开关SW1观察到中断状态“INT”的确立。步骤408,底板管理控制器BMC读取各嵌入式管理控制器eMC的第一通讯接口控制器304的从属数据输入/输出缓冲缓存器中的一从属中断状态,辨识出中断状态“INT”确立的服务器节点。步骤410,底板管理控制器BMC输出批准指令“Grant”给中断状态“INT”确立的服务器节点的第一通讯接口控制器304,且第一开关SW1切换成在底板管理控制器BMC与中断状态“INT”确立的服务器节点之间建立连接。一种实施方式经由启动所述接口模块CI_1…CI_M中对应者作连接建立。步骤412,中断状态“INT”确立的服务器节点的第一通讯接口控制器304输出一中断信号至自身的微控制器MCU。对应已在步骤402经从属数据输入/输出缓冲缓存器确立的中断状态“INT”,微控制器MCU将允许数据经由第一开关SW1传送至该底板管理控制器BMC。步骤414,微控制器MCU填写第一通讯接口控制器304内的主相关缓存器(master-related registers),以启始数据传输。至于底板管理控制器BMC如何将节点外部信息经由第一开关(如,I2C开关)传送至嵌入式管理控制器eMC的技术则更为简易。底板管理控制器BMC直接启动目标服务器节点所对应的接口模块(CI_1…CI_M之一),接着开始传输节点外部信息至目标服务器节点上的嵌入式管理控制器eMC。FIG. 4 is a flowchart illustrating how the embedded management controller eMC transmits node internal information to the baseboard management controller BMC via a first switch (eg, an I2C switch). Step 402 , the microcontroller MCU of the embedded management controller eMC fills in a slave data I/O buffer register of the first communication interface controller 304 to establish an interrupt state “INT”. In step 404, the microcontroller MCU uses a GPIO pin to indicate the assertion of the interrupt state "INT" to the first switch SW1. Step 406, the Baseboard Management Controller BMC observes the assertion of the interrupt state "INT" via the first switch SW1. Step 408, the baseboard management controller BMC reads a slave interrupt status in the slave data input/output buffer register of the first communication interface controller 304 of each embedded management controller eMC, and recognizes that the interrupt status "INT" is established server node. In step 410, the baseboard management controller BMC outputs the approval command "Grant" to the first communication interface controller 304 of the server node established in the interrupt state "INT", and the first switch SW1 is switched to be between the baseboard management controller BMC and the interrupt state "INT". INT" to establish a connection between the server nodes established. One embodiment is to establish a connection by initiating a corresponding one of the interface modules CI_1 . . . CI_M. Step 412 , the first communication interface controller 304 of the server node whose interrupt status “INT” is established outputs an interrupt signal to its own microcontroller MCU. Corresponding to the interrupt state "INT" asserted by the slave data I/O buffer in step 402, the microcontroller MCU will allow data to be transmitted to the baseboard management controller BMC via the first switch SW1. Step 414 , the microcontroller MCU fills in master-related registers in the first communication interface controller 304 to start data transmission. As for how the baseboard management controller BMC transmits the node external information to the embedded management controller eMC through the first switch (eg, I2C switch), it is simpler. The baseboard management controller BMC directly starts the interface module (one of CI_1 . . . CI_M) corresponding to the target server node, and then starts to transmit the external information of the node to the embedded management controller eMC on the target server node.

图5图解非挥发性存储器312的一种空间配置,其中,非挥发性存储器312是由一服务器节点以及该服务器节点上设置的嵌入式管理控制器eMC共享。除了嵌入式管理控制器固件映像(eMC firmware image)Image1以及Image2(Image2例如为Image1的备份)外,非挥发性存储器312还配置空间502给服务器节点的系统固件。系统固件例如可为系统的基本输入输出系统(system BIOS)。FIG. 5 illustrates a space configuration of the non-volatile memory 312, wherein the non-volatile memory 312 is shared by a server node and an embedded management controller (eMC) disposed on the server node. In addition to embedded management controller firmware images (eMC firmware images) Image1 and Image2 (Image2 is, for example, a backup of Image1), the non-volatile memory 312 also configures a space 502 for the system firmware of the server node. The system firmware can be, for example, a basic input output system (system BIOS) of the system.

图6图解非挥发性存储器312的另一种空间配置,其中,非挥发性存储器312是由复数个服务器节点以及所述服务器节点上设置的嵌入式管理控制器eMC共享。以下以四个服务器节点Node1…Node4为例。除了嵌入式管理控制器固件映像(eMC firmware image)Image1以及Image2外,非挥发性存储器312还配置空间602,作为服务器节点基本输入输出系统的读取区(例如,储存服务器节点Node1…Node4的基本输入输出系统程序代码)、以及作为不同服务器节点Node1…Node4的写入区。例如,对应服务器节点Node1的写入区604、对应服务器节点Node2的写入区606、对应服务器节点Node3的写入区608以及对应服务器节点Node4的写入区610。FIG. 6 illustrates another space configuration of the non-volatile memory 312, wherein the non-volatile memory 312 is shared by a plurality of server nodes and embedded management controllers eMC disposed on the server nodes. The following takes four server nodes Node1...Node4 as an example. In addition to the embedded management controller firmware image (eMC firmware image) Image1 and Image2, the non-volatile memory 312 also configures a space 602, which is used as a reading area of the server node basic input and output system (for example, storing the basic input and output system program code), and as the writing area of different server nodes Node1...Node4. For example, the writing area 604 corresponding to the server node Node1, the writing area 606 corresponding to the server node Node2, the writing area 608 corresponding to the server node Node3, and the writing area 610 corresponding to the server node Node4.

图7A图解服务器系统700上控制组件开机顺序的一逻辑电路。服务器系统700包括多个服务器节点Node1…Node4以及不同服务器节点Node1…Node4的芯片组。服务器节点与各服务器节点的芯片组上所安装的嵌入式管理控制器eMC如图6所示方式共享同一闪存(312)。当底板管理控制器BMC确立一上电信号BMC_RSMRST#,服务器节点Node1的嵌入式管理控制器eMC因此而启动、并确立信号704。因此,逻辑‘与(AND)’门706的输出被拉起,服务器节点Node2的嵌入式管理控制器eMC因此而启动、且确立信号708。因此,逻辑‘与’门710的输出被拉起,服务器节点Node3的嵌入式管理控制器eMC因此而启动、且确立信号712。因此,逻辑‘与’门714的输出被拉起,服务器节点Node4的嵌入式管理控制器eMC因此而启动。服务器系统700内的服务器节点Node1…Node4的嵌入式管理控制器eMC因而循序分开启动,分时存取该闪存(312)读取其中的嵌入式管理控制器固件码。服务器系统700的服务器节点Node1…Node4上的嵌入式管理控制器eMC启动后将分别启动所对应的服务器节点的芯片组。服务器节点Node1…Node4上的芯片组是一一启动,分时存取该闪存(312)获取其中服务器节点固件码(如,基本输入输出系统码(BIOS code))。一种实施方式中,服务器系统700的服务器节点Node1…Node4上的嵌入式管理控制器eMC的启动顺序与服务器节点Node1…Node4的启动顺序相同。图7B根据本发明一种实施方式图解服务器系统700上组件的启动顺序。步骤722,底板管理控制器BMC为服务器系统700确立上电信号BMC_RSMRST#。步骤724…730,服务器系统700的服务器节点Node1…Node4的嵌入式管理控制器eMC依序启动。步骤732…738,服务器系统700的服务器节点Node1…Node4的芯片组依序启动。FIG. 7A illustrates a logic circuit for controlling the boot sequence of components on the server system 700 . The server system 700 includes a plurality of server nodes Node1...Node4 and chipsets of different server nodes Node1...Node4. The server node and the embedded management controller eMC installed on the chip set of each server node share the same flash memory ( 312 ) as shown in FIG. 6 . When the baseboard management controller BMC asserts a power-on signal BMC_RSMRST#, the embedded management controller eMC of the server node Node1 is thus activated and asserts the signal 704 . Therefore, the output of the logic 'AND' gate 706 is pulled up, the embedded management controller eMC of the server node Node2 is thus activated, and the signal 708 is asserted. Therefore, the output of the logic 'AND' gate 710 is pulled up, and the embedded management controller eMC of the server node Node3 is thus activated and the signal 712 is asserted. Therefore, the output of the logic 'AND' gate 714 is pulled up, and the embedded management controller eMC of the server node Node4 is started accordingly. The embedded management controllers eMC of the server nodes Node1 . . . Node4 in the server system 700 are thus started separately sequentially, accessing the flash memory ( 312 ) time-sharingly to read the embedded management controller firmware code therein. After the embedded management controllers eMC on the server nodes Node1 . . . Node4 of the server system 700 start, they will respectively start the chip sets of the corresponding server nodes. The chipsets on the server nodes Node1...Node4 are started one by one, and access the flash memory (312) in time-sharing to obtain the server node firmware code (eg, BIOS code). In one embodiment, the startup sequence of the embedded management controllers eMC on the server nodes Node1...Node4 of the server system 700 is the same as the startup sequence of the server nodes Node1...Node4. FIG. 7B illustrates the startup sequence of components on server system 700 according to one embodiment of the invention. Step 722 , the baseboard management controller BMC asserts a power-on signal BMC_RSMRST# for the server system 700 . In steps 724...730, the embedded management controllers eMC of the server nodes Node1...Node4 of the server system 700 start up sequentially. In steps 732...738, the chip sets of the server nodes Node1...Node4 of the server system 700 are started up sequentially.

虽然本发明已以较佳实施例揭露如上,然其并非用以限定本发明,任何本领域技术人员,在不脱离本发明的精神和范围内,当可做些许更动与润饰,因此本发明的保护范围当视权利要求书所界定的为准。Although the present invention has been disclosed above with preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art may make some changes and modifications without departing from the spirit and scope of the present invention. Therefore, the present invention The scope of protection should depend on what is defined in the claims.

Claims (6)

1. a kind of server system, which is characterized in that including:
A plurality of server nodes include respectively a chipset with north bridge and south bridge, wherein corresponding to each server node also An embedded management controller is integrated respectively in the chipset of corresponding server node, and the embedded management controls The intra-node information that device collects corresponding server node is used for server system management;
One Baseboard Management Controller is made in the outside of the server node, and is communicated via network and a remote terminal;With And
Embedded management controller on the different server node is selectively coupled to the substrate tubes by one switch module Manage controller;
Wherein, the embedded management controller transmission node internal information of one of described server node to the baseboard management controls Before device, an interrupt status, the server node for enabling the Baseboard Management Controller to establish the interrupt status are also established It is distinguished with other server nodes;
The embedded management controller further includes:
One memory interface controller is accessed in a non-volatility memorizer, which is also shared on storage institute The system firmware of the server node of category;
The embedded management controller of a plurality of server nodes shares the non-volatility memorizer, and the non-volatile holographic storage Device is allocated for storing firmware code for embedded management controller;
The non-volatility memorizer is further configured the write area for providing different server node;
The embedded management controller sequential start of server node starts it in the embedded management controller of server node Afterwards, the chipset sequential start of server node.
2. server system according to claim 1, which is characterized in that further include:
The server node, the switch module and the Baseboard Management Controller is arranged in one blade.
3. server system according to claim 1, which is characterized in that further include:
One backboard and a plurality of blades for being plugged in the backboard;
Wherein:
The server node is divided into plural groups, and the server node of different groups is arranged on different blades;And
The Baseboard Management Controller is arranged on the backboard, rather than is arranged on the blade;
The switch module includes a plurality of switch submodules, is separately positioned on the blade, to couple the embedded of each blade Management Controller is to the Baseboard Management Controller.
4. server system according to claim 1, which is characterized in that each embedded management controller includes:
One microcontroller;And
One first communication interface controller, transmission node internal information to the baseboard management controls under microcontroller control Device and from the Baseboard Management Controller receiving node external information,
Wherein, which couples the Baseboard Management Controller through one first universal serial bus.
5. server system according to claim 1, it is characterised in that:
When receiving a collection of quasi instruction that the Baseboard Management Controller sends out the server node that the interrupt status is established, Embedded management controller in the server node that the interrupt status is established starts transmission node internal information to the bottom Board management controller.
6. server system according to claim 4, which is characterized in that the embedded management controller respectively also wraps It includes:
One second communication interface controller is controlled by the microcontroller of corresponding embedded management controller, is somebody's turn to do with receiving The node external demand that Baseboard Management Controller is sent for the server system management,
Wherein, which couples the Baseboard Management Controller via one second universal serial bus.
CN201510779281.XA 2015-11-13 2015-11-13 Chipset and server system Active CN105468562B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201510779281.XA CN105468562B (en) 2015-11-13 2015-11-13 Chipset and server system

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201510779281.XA CN105468562B (en) 2015-11-13 2015-11-13 Chipset and server system

Publications (2)

Publication Number Publication Date
CN105468562A CN105468562A (en) 2016-04-06
CN105468562B true CN105468562B (en) 2018-09-14

Family

ID=55606282

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201510779281.XA Active CN105468562B (en) 2015-11-13 2015-11-13 Chipset and server system

Country Status (1)

Country Link
CN (1) CN105468562B (en)

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108337175A (en) * 2018-01-31 2018-07-27 郑州云海信息技术有限公司 A kind of multipath server and its node communication method

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101488118A (en) * 2008-01-18 2009-07-22 株式会社日立制作所 Management method of combined computer system and combined computer system
CN102236629A (en) * 2010-04-21 2011-11-09 鸿富锦精密工业(深圳)有限公司 System and method for identifying serial interface signal
CN104049692A (en) * 2014-06-27 2014-09-17 浪潮(北京)电子信息产业有限公司 Blade server
CN104283709A (en) * 2013-07-10 2015-01-14 英业达科技有限公司 Server system and data transmission method thereof
CN104571294A (en) * 2013-10-22 2015-04-29 英业达科技有限公司 Server system

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US9367419B2 (en) * 2013-01-08 2016-06-14 American Megatrends, Inc. Implementation on baseboard management controller of single out-of-band communication access to multiple managed computer nodes

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101488118A (en) * 2008-01-18 2009-07-22 株式会社日立制作所 Management method of combined computer system and combined computer system
CN102236629A (en) * 2010-04-21 2011-11-09 鸿富锦精密工业(深圳)有限公司 System and method for identifying serial interface signal
CN104283709A (en) * 2013-07-10 2015-01-14 英业达科技有限公司 Server system and data transmission method thereof
CN104571294A (en) * 2013-10-22 2015-04-29 英业达科技有限公司 Server system
CN104049692A (en) * 2014-06-27 2014-09-17 浪潮(北京)电子信息产业有限公司 Blade server

Also Published As

Publication number Publication date
CN105468562A (en) 2016-04-06

Similar Documents

Publication Publication Date Title
US10126954B1 (en) Chipset and server system using the same
JP5272265B2 (en) PCI device sharing method
US10333865B2 (en) Transformation of peripheral component interconnect express compliant virtual devices in a network environment
TW201937381A (en) System and method for supporting multi-mode and/or multi-speed non-volatile memory (NVM) express (NVMe) over fabrics (NVME-oF) devices and storage device
US7631050B2 (en) Method for confirming identity of a master node selected to control I/O fabric configuration in a multi-host environment
US20190286590A1 (en) Cpld cache application in a multi-master topology system
US11983136B2 (en) PCIe device and operating method thereof
US11995019B2 (en) PCIe device with changeable function types and operating method thereof
TWI739127B (en) Method, system, and server for providing the system data
US11928070B2 (en) PCIe device
US20130262700A1 (en) Information processing system and virtual address setting method
JP2009043256A (en) Accessing method and arrangement of memory unit
US12242625B2 (en) PCIe function and operating method thereof
CN108023762B (en) Server data port learning for data exchanger
TWI447670B (en) High speed baseboard management controller and transmission method thereof
US8996734B2 (en) I/O virtualization and switching system
US20150207731A1 (en) System and method of forwarding ipmi message packets based on logical unit number (lun)
US20120324078A1 (en) Apparatus and method for sharing i/o device
CN105468562B (en) Chipset and server system
JP6760579B2 (en) Network line card (LC) integration into host operating system (OS)
CN116627472A (en) Firmware program upgrading method and server of high-speed peripheral component equipment
US12169465B2 (en) Peripheral component interconnect express device and operating method thereof
JP2001027920A (en) Bus bridge circuit and its power management method
CN120256373A (en) Device control apparatus, device control method, storage medium, and electronic device

Legal Events

Date Code Title Description
C06 Publication
PB01 Publication
C10 Entry into substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant
CP01 Change in the name or title of a patent holder
CP01 Change in the name or title of a patent holder

Address after: Room 301, 2537 Jinke Road, Zhangjiang High Tech Park, Pudong New Area, Shanghai 201203

Patentee after: Shanghai Zhaoxin Semiconductor Co.,Ltd.

Address before: Room 301, 2537 Jinke Road, Zhangjiang High Tech Park, Pudong New Area, Shanghai 201203

Patentee before: VIA ALLIANCE SEMICONDUCTOR Co.,Ltd.