WO2011015130A1 - 设备接管方法和装置及双控系统 - Google Patents

设备接管方法和装置及双控系统 Download PDF

Info

Publication number
WO2011015130A1
WO2011015130A1 PCT/CN2010/075663 CN2010075663W WO2011015130A1 WO 2011015130 A1 WO2011015130 A1 WO 2011015130A1 CN 2010075663 W CN2010075663 W CN 2010075663W WO 2011015130 A1 WO2011015130 A1 WO 2011015130A1
Authority
WO
WIPO (PCT)
Prior art keywords
converter
control board
bridge
port
resources
Prior art date
Application number
PCT/CN2010/075663
Other languages
English (en)
French (fr)
Inventor
王学友
Original Assignee
成都市华为赛门铁克科技有限公司
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 成都市华为赛门铁克科技有限公司 filed Critical 成都市华为赛门铁克科技有限公司
Publication of WO2011015130A1 publication Critical patent/WO2011015130A1/zh
Priority to US13/368,106 priority Critical patent/US8285910B2/en

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
    • G06F13/40Bus structure
    • G06F13/4004Coupling between buses
    • G06F13/4022Coupling between buses using switching circuits, e.g. switching matrix, connection or expansion network
    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06FELECTRIC DIGITAL DATA PROCESSING
    • G06F2213/00Indexing scheme relating to interconnection of, or transfer of information or other signals between, memories, input/output devices or central processing units
    • G06F2213/0026PCI express

Definitions

  • the present invention relates to the field of electronic technologies, and in particular, to a device take-over method and device and a dual control system. Background technique
  • SW, Switch are non-transparent (NT, NON- Transparent) in the Peripheral Component Interconnect Express (PCIE) bus.
  • PCIE Peripheral Component Interconnect Express
  • the NT bridge is a non-transparent PCIE bridge.
  • the NT bridge of the converter prevents the host from enumerating the devices connected to the back of the NT bridge. It also allows windows to be opened to exchange data and isolate the devices and memory space behind the NT bridge.
  • the dual control of the device is mainly implemented by the PCIE NT bridge mirroring technology.
  • another control board B connected to the control board through the NT bridge can use mirroring technology to
  • the request of the device controlled by the control board A is indirectly processed, thereby implementing dual control of the device through the control boards A and B, thereby improving the anti-fault capability of the system.
  • the embodiment of the invention provides a device take-over method and device and a dual control system, which can achieve a dual control by directly taking over a faulty device, thereby improving the processing response speed of the device request.
  • a device takeover method includes:
  • the first control board performs topology discovery on the fast peripheral component interconnect PCIE bus; When the topology discovers the non-transparent NT bridge, the resource is reserved for the converter where the NT bridge is located and the device connected to the downlink port of the converter according to the set resource reservation policy, where the NT bridge is located.
  • the converter and the device connected to the downstream port of the converter are currently controlled by the second control board; when the second control board fails, the converter is taken over from the second control board and the downstream port of the converter a connected device, the reserved resource is allocated to the converter and a device connected to a downlink port of the converter, so that the converter and a device connected to a downlink port of the converter are first Work under the control of the control panel.
  • a device take-over device includes:
  • Topology discovery module for topology discovery of the PCIE bus
  • a resource reservation module configured to: when the topology discovery module topology discovers to the NT bridge, connect the converter where the NT bridge is located and the downlink port of the converter according to a set resource reservation policy The device reserves resources, and the converter where the NT bridge is located and the device connected to the downlink port of the converter are currently controlled by the second control board;
  • a takeover module configured to: when the second control board fails, take over the converter and the device connected to the downlink port of the converter from the second control board, and allocate resources reserved by the resource reservation module to The converter and the device coupled to the downstream port of the converter operate the converter and the device connected to the downstream port of the converter under control of the device takeover device.
  • a dual control system includes a first control board and a second control board
  • the first control board is configured to perform topology discovery on the PCIE bus; when the topology discovers the NT bridge, according to the set resource reservation policy, the converter where the NT bridge is located and the converter
  • the device connected to the downlink port reserves resources, and the converter where the NT bridge is located and the device connected to the downlink port of the converter are currently controlled by the second control board; when the second control board fails, the second control a board taking over the converter and a device connected to a downstream port of the converter, the reserved resource being allocated to the converter and a device connected to a downstream port of the converter, the converter being And the device connected to the downstream port of the converter operates under the control of the first control board.
  • the technical solution adopted in the embodiment of the present invention has the following advantages: since the first control board is previously a converter with an NT bridge currently controlled by the second control board and A device connected to the downstream port of the converter reserves resources such that when the switch is taken over from the second control board and the device connected to the downstream port of the converter, resources reserved in advance may be allocated to the above
  • the converter and the device connected to the downstream port of the converter enable the taken over device to have relatively sufficient operable resources, can work normally under the control of the first control board, and realize dual control by directly taking over the device, which can improve The response speed of the request processing.
  • FIG. 1 is a flowchart of a method for taking over a device according to Embodiment 1 of the present invention
  • FIG. 2 is a flowchart of a method for taking over a device according to Embodiment 2 of the present invention
  • FIG. 3 is a schematic diagram of a dual control architecture for resource reservation according to Embodiment 2 of the present invention
  • FIG. 3B is a schematic diagram of a dual control architecture for providing reserved resource allocation according to Embodiment 2 of the present invention
  • Embodiment 3 provides a flowchart of a method for taking over a device
  • FIG. 5-a is a schematic diagram of a dual control architecture in which a resource is not reserved according to Embodiment 3 of the present invention
  • FIG. 5-b is a schematic diagram of a dual control architecture for resource reservation according to Embodiment 3 of the present invention
  • the third embodiment of the present invention provides a schematic diagram of a dual-control architecture for resource allocation.
  • FIG. 6 is a schematic diagram of a device take-over device according to Embodiment 4 of the present invention;
  • FIG. 7 is a schematic diagram of a dual control system according to Embodiment 5 of the present invention. detailed description
  • the embodiment of the invention provides a method and a device for taking over a device and a dual control system.
  • the processing response speed of the device request can be improved.
  • the converter with NT bridge can set the failover (FV, Failover) function
  • the so-called failover refers to the function of port type conversion with NT bridge converter, that is, Under the control of the control board, the NT bridge of the converter can be converted to an up port, the up port of the converter can also be converted to an NT bridge, and so on.
  • Embodiment 1 mainly utilizes the failover function of the converter with the NT bridge to realize direct takeover of the device, which will be described in detail below through specific embodiments.
  • Embodiment 1 mainly utilizes the failover function of the converter with the NT bridge to realize direct takeover of the device, which will be described in detail below through specific embodiments.
  • a device takeover method the flow chart is shown in Figure 1, the method may include:
  • Step 110 The first control board performs topology discovery on the fast peripheral component interconnection bus.
  • topology discovery refers to discovering different buses and their connected devices and functions through the PCIE bus of the scanning system.
  • Step 120 When the topology discovers the NT bridge, reserve resources for the converter where the NT bridge is located and the device connected to the downlink port of the converter according to the set resource reservation policy, where the NT bridge is converted.
  • the device and the device connected to the downstream port of the converter are currently controlled by a second control board.
  • the first control board can perform topology discovery on all branches of the PCIE bus in a deep traversal manner when the operating system is started.
  • the first control board can also perform topology discovery on the branch where the NT bridge is located in the PCIE bus according to the deep traversal process during the stable operation of the operating system.
  • the converter that is currently controlled by the second control board and the downlink port of the converter may be configured according to the set resource reservation policy.
  • the connected device reserves bus resources and/or address space resources, and the like.
  • the resource reservation policy set by the foregoing may include: setting, in advance, the number of resources reserved for the converter where the NT bridge is located and the device connected to the downlink port of the converter in different scenarios. .
  • the number of reserved address space resources can be set according to the requirements of the network card device address space resource, and as much as possible for the network card device. Address space resources, and so on.
  • Step 130 When the second control board fails, the second control board takes over the converter and the device connected to the downlink port of the converter, and allocates the reserved resource to the converter. And a device connected to the downstream port of the converter, such that the converter and the device of the downstream port connected to the converter operate under the control of the first control board.
  • the first control board can set the NT bridge conversion of the converter to an up port, that is, by using the failed takeover function of the converter. Directly taking over the above converter and the device connected to the downstream port of the converter.
  • the first control board may allocate the reserved resources to the foregoing according to the bus resources and/or address space resources required by the converter and the device connected to the downlink port of the converter. And a converter and a device connected to the downstream port of the converter, enabling the converter previously controlled by the second control board and the device connected to the downstream port of the converter to continue normal under the control of the first control board jobs. For example, if the first control board reserves 3 bus resources and 6M address space resources in advance, the converter taken over from the second control board currently includes one up port and two down ports, and two down ports are connected.
  • the first control board can allocate three bus resources reserved in advance to one up port and two down ports of the converter, and the 6M address reserved in advance.
  • the spatial resources are allocated on average or non-average to the two endpoint devices connected to the down port of the above converter.
  • the device connected to the downstream port of the converter may include an endpoint device and/or a bridge device. It can be understood that the foregoing technical solutions in this embodiment may be specifically implemented by the first control board or a part of the functional modules of the first control board.
  • resources are reserved for the converter where the NT bridge controlled by the first control board is controlled by the second control board and the device connected to the downlink port of the converter; when the second control board is faulty, The first control board takes over the above-mentioned converter and the device of the downlink port connected to the converter from the second control board, and allocates reserved resources as an example for description.
  • the second control board may also be a converter in which the NT bridge controlled by the first control board is located and a device connected to the downstream port of the converter reserve resources; when the first control board fails, the second control board takes over the converter from the first control board and And the device connected to the downlink port of the converter, and allocates reserved resources for the second control board to take over the device from the first control board, and realize dual control of the device.
  • the first control board of the embodiment reserves resources in advance for the converter with the NT bridge controlled by the second control board and the device connected to the downlink port of the converter. After the second control board takes over the converter and the device connected to the downstream port of the converter, the reserved resources are allocated to the converter and the device connected to the downstream port of the converter, so that the taken device has Relatively enough operable resources can work normally under the control of the first control board, and dual control can be realized by directly taking over the device, which can improve the response speed of the request processing.
  • Embodiment 2 Embodiment 2
  • the present embodiment uses the control board A to reserve resources for the converter with the NT bridge currently controlled by the control board B and the device connected to the downstream port of the converter at the time of system startup, taking over the above conversion. After the device is connected to the downlink port of the converter, the reserved resources are allocated to the taken device as an example.
  • a device takeover method the flow chart is shown in Figure 2, the method may include:
  • Step 201 The control board A performs topology discovery on all branches of the PCIE bus.
  • Step 202 When the control board A topology discovers the NT bridge, the converter that is currently controlled by the control board B and the device connected to the downlink port of the converter are configured according to the set resource reservation policy. Reserve resources.
  • the system core of the control board A can perform topology discovery on all branches of the PCIE bus in a deep traversal manner.
  • control board A can determine the type of the device and the resource requirement according to the type identifier carried by each device under the PCIE bus, so that the bus resource and/or the address space resource can be allocated according to the resource requirements of the different types of devices.
  • the devices under the PCIE bus may include endpoint devices (for example, network devices, graphics cards, and the like) and bridge devices (for example, converters, etc.).
  • endpoint devices for example, network devices, graphics cards, and the like
  • bridge devices for example, converters, etc.
  • the resource reservation policy set by the foregoing may include a preset amount of resources reserved for the converter where the NT bridge is located in the different scenarios and the device connected to the downlink port of the converter, which may be different according to the application scenario. To set the amount of resources that need to be reserved.
  • the address space resource may be allocated to the endpoint device according to the resource requirement of the endpoint device; when the topology finds a converter without the NT bridge, the Allocating a bus resource for the converter, and then continuing to topology the device under the bus; when the topology discovers the NT bridge, the bus resource is reserved for the converter where the NT bridge is located according to the set resource reservation policy (or pre- Retaining bus resources and address space resources), and for the downstream port of the converter The connected device reserves bus resources and/or address space resources, and so on, until the topology discovery of all branches of the PCIE bus is completed.
  • control board A when performing topology discovery of the PCIE bus, the control board A may topologically discover one or more NT bridges, and the converters of the one or more NT bridges and the converters
  • the device connected to the downstream port may be actually controlled by another one or more control boards, and the control board A may be a converter for each NT bridge discovered by the topology and a device connected to the downstream port of the converter.
  • the control board A only discovers one NT bridge and the converter where the NT bridge is located is currently controlled by the control board B as an example for specific description.
  • Step 203 When the control board B fails, the control board A takes over the converter and the device connected to the downlink port of the converter from the control board B, and allocates the reserved resources to the converter and the conversion. The device connected to the downstream port of the device.
  • the control board A when the control board B fails, the control board A can convert the up port connected to the converter and the control board B into an NT bridge, and convert the NT bridge connected to the converter and the control board A into The up port, in turn, takes over the converter and the device connected to the downstream port of the converter from the control board B.
  • the control board A can perform topology discovery on the converter and the device of the downlink port connected to the converter according to the deep traversal manner, according to the bus resource requirement of the converter and the downlink port connected to the converter.
  • the bus resource and/or address space resource requirement of the device allocate some or all of the bus resources reserved in advance to the uplink port and the downlink port of the above converter, and reserve some or all of the bus resources and/or address spaces reserved in advance.
  • the resource is allocated to the device of the downstream port connected to the converter such that the converter and the device of the downstream port of the converter can continue to operate normally under the control of the control board A.
  • control board A can use the above manner to perform device takeover and reserve resource allocation, and details are not described herein again.
  • the present embodiment is described by taking as an example a converter in which the control board A previously controls the NT bridge controlled by the control board B and a device reserved for the downlink port connected to the converter; when the control board B fails The control board A takes over the above-mentioned converter and the device connected to the downlink port of the converter from the control board B, and allocates reserved resources as an example for description.
  • the control board B can also be a control board in advance.
  • the device connected to the downlink port is allocated with the reserved resources.
  • the process of implementing the above-mentioned takeover of the control board B is the same as that of the control board A, and is not described here.
  • control board 310 is the above-mentioned control board A
  • control board 320 is the above-mentioned control board B as an example.
  • the control board 310 includes a converter 311, a converter 312, and a converter 313.
  • the control board 320 includes a converter 321, a converter 322, and a converter 323.
  • the down port 3111 of the converter 311 is connected to the converter 312, and the down port 3112 is connected to the converter 313.
  • the NT bridge port of the converter 312 is connected to a down port of the converter 322; the port 3231 of the converter 323 is connected to one down port of the converter 313.
  • the system core of the control board 310 topology discovers all the branches under the PCIE bus according to the deep traversal manner, and allocates or reserves address space resources and/or buses for the devices discovered by the topology. Resources.
  • port 3231 of converter 323 is currently an NT bridge, and port 3132 is currently an up port.
  • the control board 310 allocates the No. 0 bus resource to the up port of the converter 311, the No. 1 bus resource to the down port 3111, the No. 6 bus resource to the down port 3112, and the No. 2 bus resource and the down port to the up port of the converter 312.
  • the bus number and address resource of the PCIE are sequentially incremented, and the valid bus number range allowed by the port 3111 of the converter 311 is: 1 to 5, and the address space range is: 0x00000000 ⁇ 0x002ffff (3M);
  • the valid bus number range allowed by 3112 is: 6 ⁇ 13 and the address space range is: 0x00e00000 ⁇ 0x017ffff (10M).
  • the control board 310 takes over the converter 323 from the control board 320 and the endpoint device 343 and the endpoint device 344, that is, the port 3232 of the converter 323 is converted to NT. Bridge, set port 3231 conversion to up port.
  • the control board 310 can perform topology discovery on the device under the No. 8 bus in a deep traversal manner, and according to the resource requirements of the converter 323 and the endpoint device 343 and the endpoint device 344, the reserved buses 9, 10 and 11 will be reserved.
  • the resource is allocated to one up port and two down ports of the converter 323, and the reserved 6M (0x00e00000 ⁇ 0x013fffff) address space resources are allocated to the endpoint devices 343 and 344.
  • the endpoint device 343 can be allocated 2M (0x00e00000 ⁇ 0x00fffff). ), assign 4M (0x01000000 ⁇ 0x013fffff) to endpoint device 344
  • the converter 323 and the endpoint device 343 and the endpoint device 344 can be enabled on the control board 310.
  • Control continues to work normally, so that the control board 310 and the control board 320 implement dual control of the converter 323 and the endpoint device 343 and the endpoint device 344 by using the NT bridge, thereby improving the system's anti-fault capability.
  • control board 320 can also perform topology discovery on the PCIE bus in the above manner, and is the converter 312 where the NT bridge of the control board 310 is located, and the endpoint device 331 of the downlink port connected to the converter.
  • the endpoint device 332 reserves resources.
  • the control board 320 can also take over the converter 312 and the endpoint device 331 and the endpoint device 332 in the manner described above, by directly taking over the converter 312 and the endpoint device 331 and the endpoint device 332. Double control.
  • the control board A is reserved for the converter with the NT bridge currently controlled by the control board B and the device connected to the downlink port of the converter when the operating system is started.
  • the resource after failing to take over the above converter and the device connected to the downstream port of the converter from the control board B, allocate the reserved resources to the converter and the device connected to the downstream port of the converter, so that the failure is caused.
  • the taken over device has relatively sufficient operable resources, can work normally under the control of the control board A, and realizes dual control by directly taking over the device, which can improve The processing response speed of the device request.
  • the present embodiment reserves resources for the converter with the NT bridge currently controlled by the control board B and the device connected to the downstream port of the converter at any time after the system is stably operated.
  • an example is given by taking an example of allocating reserved resources for the taken-over device.
  • a device takeover method the flow chart can be as shown in FIG. 4, the method can include:
  • Step 401 When the operating system of the control board A is started, the control board A performs topology discovery on each branch of the PCIE bus, and allocates an address space for the endpoint device discovered by the topology and the converter without the NT bridge. Resource and / or bus resources.
  • the system core of the control board A can perform topology discovery on all branches of the PCIE bus in a deep traversal manner.
  • control board A can determine the type of the device and the resource requirement according to the type identifier carried by each device under the PCIE bus, so that the bus resource and/or the address space resource can be allocated according to the resource requirements of the different types of devices.
  • the topology discovery of the converter where the NT bridge is located and the device connected to the downstream port of the converter is skipped.
  • the address space resource may be allocated to the endpoint device according to the resource requirement of the endpoint device; when the topology finds a converter without the NT bridge, the Allocating a bus resource to the converter, and then continuing to topology the device under the bus; when the topology discovers the NT bridge, the location of the NT bridge can be recorded, and the topology is found to find the converter where the NT bridge is located and the conversion The device connected below the device, and does not allocate resources for it, and so on, until the topology discovery of all branches of the PCIE bus is completed.
  • Step 402 During the stable operation of the operating system, the control board A re-tops and discovers the branch where the NT bridge is located in the PCIE bus.
  • control board A can re-top the branch where the NT bridge is located in the PCIE bus at any time after the operating system is stably running.
  • the equipment of the branch reserves resources or reallocates resources.
  • Step 403 When the topology board finds the NT bridge, the control board A selects the converter where the NT bridge currently controlled by the control board B is located according to the set resource reservation policy, and connects to the downlink port of the converter.
  • the device reserves resources.
  • the address space resource may be re-allocated for the endpoint device according to the resource requirement of the endpoint device.
  • a bus resource can be re-allocated for the converter, and then the topology is found to find the device under the bus; when the topology is found to the NT bridge, the resource can be pre-configured according to the set resource.
  • the retention policy reserves bus resources and/or address space resources for the converter where the NT bridge is located and the device connected to the downstream port of the converter, and so on, until the above topology of the branch where the NT bridge is located is completed. Find.
  • the resource reservation policy set by the foregoing may include the preset number of resources reserved for the converter where the NT bridge is located in different scenarios and the device connected to the downlink port of the converter, which may be according to an application scenario. The difference is to set the amount of resources that need to be reserved.
  • control board A may topologically discover one or more NT bridges when the system is started, and the converter where the one or more NT bridges are located and the device connected to the downstream port of the converter are currently It may be actually controlled by one or more other control boards.
  • the control board A can re-top discover the branch where each NT bridge in the PCIE bus is located, and find it for the topology.
  • the converter in which each NT bridge is located and the device connected to the downstream port of the converter reserve resources to reallocate resources for other devices discovered by the topology.
  • only one NT bridge is found in the topology of the control board A when the system is started, and the converter where the NT bridge is located is currently controlled by the control board B as an example for specific description.
  • Step 404 When the control board B fails, the control board A fails to take over the converter and the device connected to the downlink port of the converter from the control board B, and allocates the reserved resources to the converter and the The device connected to the downstream port of the converter.
  • control board A when the control board B fails, the control board A can convert the up port connected to the converter and the control board B into an NT bridge, and convert the NT bridge connected to the converter and the control board A into The up port, in turn, fails from the control board B to take over the converter and the device connected to the downstream port of the converter.
  • control board A can follow the depth traversal to the above converter and Performing topology discovery on the device connected to the downstream port of the converter, according to the bus resource requirements of the converter and the bus resource and/or address space resource requirements of the device connected to the downstream port of the converter, All or part of the bus resources remaining are allocated to the uplink port and the downlink port of the above converter, and part or all of the bus resource and/or address space resources reserved in advance are allocated to the device connected to the downlink port of the converter, so that The converter and the device of the downstream port of the converter can continue to operate normally under the control of the control board A.
  • control board A topology discovers multiple NT bridges, and the converters where the plurality of NT bridges are currently controlled by other one or more control boards, the conversion for each NT bridge is performed. And the device connected to the downlink port of the converter, and the control board A can perform device takeover and reserved resource allocation in the foregoing manner, and details are not described herein again.
  • the present embodiment reserves resources for the converter where the control board A is in advance controlled by the control board B and the device connected to the downstream port of the converter; when the control board B fails, the control board A
  • the control board B takes over the above-mentioned converter and the device connected to the downlink port of the converter, and allocates reserved resources as an example for description.
  • the control board B can also be an NT bridge controlled by the control board A in advance.
  • the converter in which it is located and the device connected to the downstream port of the converter reserve resources; when the control board A fails, the control board B takes over the above-mentioned converter and the device of the downstream port connected to the converter from the control board A
  • the process of the device B is the same as that of the control board A, and is not described here.
  • control board 510 is the above-described control board A
  • control board 520 is the above-described control board B
  • the control board 510 includes a converter 511, a converter 512, and a converter 513
  • the control board 520 includes a converter 521, a converter 522, and a converter 523.
  • the down port 5111 of the converter 511 is connected to the converter 512, and the down port 5112 is connected to the converter 513.
  • the NT bridge port of the converter 512 is connected to a down port of the converter 522;
  • the port 5231 of the converter 523 is connected to one down port of the converter 513.
  • the system core of the control board 510 can discover all the branches under the PCIE bus in a deep traversal manner, and the endpoint devices discovered by the topology and the converter without the NT bridge. Allocate address space resources and/or bus resources.
  • port 5231 of converter 523 is currently an NT bridge, and port 5232 is currently an up port.
  • the control board 510 allocates the No. 0 bus resource to the up port of the converter 511, the down port 5111 allocates the No. 1 bus resource, and the down port 5112 allocates the No.
  • the range of the bus number allowed by the port 5111 of the converter 511 is: 1 to 5, and the allowable address space range is: 0x00000000 to 0x002fffff (3M); the port 5112 allows control.
  • the bus number range is: 6 ⁇ 10, the allowable address space range is: 0x00e00000 ⁇ 0x011 fffff ( 4M )##
  • control board 510 re-tops the branch where the NT bridge is located, starting from the top of the branch where the NT bridge is located (the 0 bus).
  • the topology found that resources were reserved and reallocated for devices connected to the following branches of bus 0.
  • the control board 510 can re-assign the number 31 bus to the down port 5112 of the converter 511; re-up to the up port of the converter 513 and 3
  • One down port allocates buses 32, 33, 37, 38; reserved buses 34, 35, and 36 for converter 523, and 6M (0x02000000 ⁇ 0x025fffff) address space resources for endpoint devices 543 and 544; 533 reallocates 2M (0x02600000 ⁇ 0x027fffff) address space resources, and reallocates 2M (0x02800000 ⁇ 0x029fffff) address space resources for endpoint device 534.
  • the port number of the converter 5111 allows the control of the bus number range to be: 1 to 5, and the allowed address space range is: 0x00000000 ⁇ 0x002fffff (3M); Port 5112 allows control of the bus number range: 31 ⁇ 38, the allowable address space range is: 0x02000000 - 0x029fffff (10M).
  • control board 520 fails or otherwise causes the control board 510 to take over the converter 523 and the endpoint device 543 and the endpoint device 544 from the control board 520, and the port 5232 of the converter 523 is converted to NT. Bridge, set port 5231 conversion to the up port.
  • the control board 510 can perform topology discovery on the device under the No. 33 bus in a deep traversal manner, and reserve the 34, 35 and 36 bus resources according to the resource requirements of the converter 523 and the endpoint device 543 and the endpoint device 544.
  • the 1 up port and the 2 down ports assigned to the converter 523 allocate the reserved 6M (0x02000000 ⁇ 0x025fffff) address space resources to the endpoint devices 543 and 544.
  • the endpoint device 543 can be allocated 3M (0x02000000 ⁇ 0x022ffffff). ), assign 3M (0x02300000 ⁇ 0x025fffff) to the endpoint device 544.
  • the control board 510 can control its normal operation, and the control board 510 and the control board 520 utilize the NT bridge.
  • the dual control of the converter 523 and the endpoint device 543 and the endpoint device 544 is implemented, thereby improving the system's anti-fault capability.
  • control board 520 can also perform topology discovery on the PCIE bus in the above manner, and is the converter 512 where the NT bridge of the control board 510 is located, and the endpoint device 531 of the downlink port connected to the converter.
  • the endpoint device 532 reserves resources.
  • the control board 520 can also take over the converter 512 and the endpoint device 531 and the endpoint device 532 in the manner described above, by directly taking over the converter 512 and the endpoint device 531 and the endpoint device 532. Double control.
  • control board A in this embodiment is connected to the converter with the NT bridge currently controlled by the control board B and the downlink port of the converter at any time when the operating system is stably operated.
  • the device reserves resources, and after the controller B fails to take over the converter and the device connected to the downlink port of the converter, allocates the reserved resources to the converter and the device connected to the downlink port of the converter.
  • the device that is failed to take over has relatively sufficient operable resources, can work normally under the control of the control board A, and realizes dual control by directly taking over the device, which can relatively improve the processing response speed of the device.
  • an apparatus take-over device is further provided in the embodiment of the present invention.
  • a device take-over device 600 may include: a topology discovery module 610, a resource reservation module 620, and a takeover module 630.
  • the topology discovery module 610 is configured to perform topology discovery on the PCIE bus;
  • the resource reservation module 620 is configured to: when the topology discovery module 610 topology discovers to the NT bridge, the converter in which the NT bridge is located and the device connected to the downlink port of the converter according to the set resource reservation policy Reserving resources, the converter where the NT bridge is located and the device connected to the downlink port of the converter are currently controlled by the second control board;
  • the resource reservation policy set by the foregoing may be: a preset amount of reserved resources for a converter where the NT bridge is located and a device connected to the downlink port of the converter in different scenarios. For example, if the downstream port of the converter with the NT bridge is usually connected to the network card device, the number of reserved address space resources can be set according to the requirements of the network card device address space resource, and enough address is reserved for the network card device. Space resources, and so on.
  • the takeover module 630 is configured to: when the second control board fails, fail to take over the converter and the device connected to the downlink port of the converter from the second control board, and allocate resources reserved by the resource reservation module 620 to The converter and the device connected to the downstream port of the converter operate the converter and the downstream port device connected to the converter under the control of the device take-up device 600.
  • the topology discovery module 610 may be specifically configured to perform topology discovery on all branches of the PCIE bus when the system is started.
  • the topology discovery module 610 can be specifically used to perform topology discovery on the branch where the NT bridge is located in the PCIE bus during system operation.
  • the resource reservation module 620 may be specifically configured to: when the topology discovery module 610 topology discovers the NT bridge, according to the set resource reservation policy, the NT that is currently controlled by the second control board.
  • the converter in which the bridge resides reserves bus resources and reserves bus resources and/or address space resources for devices connected to the downstream ports of the converter.
  • the takeover module 630 can include:
  • a pipe module configured to take over the converter and the device connected to the downstream port of the converter from the second control board when the second control board fails.
  • a resource allocation sub-module configured to allocate, by the resource reservation module 620, the bus resource reserved for the converter to the uplink port and the downlink port of the converter, and the resource reservation module 620 is connected to the downlink port of the converter.
  • the bus resources and/or address space resources reserved by the device are allocated to devices connected to the downstream port of the above converter.
  • connection pipe module when the second control board fails, can set the NT bridge conversion of the converter to an up port, thereby taking over the converter and the device connected to the downstream port of the converter, wherein
  • the device connected to the downstream port of the above converter may include an endpoint device and/or a bridge device.
  • the resource allocation sub-module may perform topology discovery on the converter and the device connected to the downlink port of the converter in a deep traversal manner, according to the converter and the device connected to the downlink port of the converter.
  • the bus resource and/or address space resource and the like allocate the resources reserved by the resource reservation module 620 to the above-mentioned converter and the device connected to the downstream port of the converter.
  • the converter that fails the pipe module to take over from the second control board currently includes one up port and two down ports, 2
  • the downlink port is connected to two endpoint devices that respectively require a 2M address space, and the resource allocation sub-module can allocate three bus resources reserved in advance by the resource reservation module 620 to one up port and two of the converters respectively.
  • the down port allocates 6M address space resources reserved in advance by the resource reservation module 620 to the two endpoint devices connected to the down port of the converter.
  • the device take-over device 600 of the embodiment may be disposed in the control board A or other device according to the second or third embodiment, and the functions of the respective functional modules may be specifically implemented according to the method in the second or third embodiment.
  • the device take-over device is a converter with an NT bridge controlled by the second control board in advance.
  • a device connected to a downlink port of the converter reserves resources, and after failing to take over the converter and the device connected to the downlink port of the converter from the second control board, allocating resources reserved in advance to the converter And a device connected to the downstream port of the converter, such that the device that failed to take over has relatively sufficient operational resources, capable of Under the control of the device, the dual control is realized by directly taking over the device, which can relatively improve the processing response speed requested by the device.
  • Embodiment 5 In order to facilitate the better implementation of the technical solution of the above embodiments of the present invention, a dual control system is also provided in the embodiment of the present invention.
  • a dual control system may include: a first control board 710 and a second control board 720.
  • the first control board 710 is configured to perform topology discovery on the PCIE bus; and when the topology is discovered to the NT bridge, according to the set resource reservation policy, the converter where the NT bridge is located and the conversion
  • the device connected to the downlink port of the device reserves resources, and the converter where the NT bridge is located and the device connected to the downlink port of the converter are currently controlled by the second control board; when the second control board 720 fails, the second The second control board 720 takes over the converter and the device connected to the downlink port of the converter, and allocates the reserved resources to the converter and the device connected to the downlink port of the converter, so that the converter and the converter A device connected to the downstream port of the converter operates under the control of the first control board 710.
  • the second control board 720 can also perform topology discovery on the PCIE bus.
  • the NT bridge currently controlled by the first control board 710 is located according to the set resource reservation policy.
  • the converter and the device connected to the downstream port of the converter reserve resources; when the first control board 710 fails, the first control board 710 takes over the converter and the downstream port of the converter And the device allocates the reserved resource to the converter and the device connected to the downlink port of the converter, so that the converter and the device connected to the downlink port of the converter are controlled by the second control board 720. Work under.
  • the first control board 710 of the dual control system of this embodiment may include the device 600 that is taken over by the device in the fourth embodiment.
  • the functions of the entities of the dual control system may be specifically according to the method in the second or third embodiment.
  • the first control board is controlled by the second control board in advance.
  • the NT bridge converter and the device connected to the downstream port of the converter reserve resources, and the resources reserved in advance after taking over the converter and the device connected to the downstream port of the converter from the second control board
  • the device is allocated to the above converter and the device connected to the downstream port of the converter, so that the taken over device has relatively sufficient operable resources, can work normally under the control of the first control board, and realizes dual control by directly taking over the device. , can relatively improve the processing response speed of the device request.
  • the storage medium may be a magnetic disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Mathematical Physics (AREA)
  • Computer Hardware Design (AREA)
  • General Physics & Mathematics (AREA)
  • Small-Scale Networks (AREA)
  • Bus Control (AREA)

Description

设备接管方法和装置及双控系统 本申请要求于 2009 年 8 月 7 日提交中国专利局, 申请号为: 200910161060.0, 发明名称为 "设备接管方法和装置及双控系统" 的中国专 利申请的优先权, 其全部内容通过引用结合在本申请中。 技术领域
本发明涉及电子技术领域, 具体涉及设备接管方法和装置及双控系统。 背景技术
目前, 随着总线技术和双控技术的发展, 在快速外设组件互联( PCIE, Peripheral Component Interconnect Express ) 总线中, 许多的转换器 ( SW, Switch ) 都带有非透明 (NT, NON- Transparent )桥, 现有技术中也常常利 用 NT桥来实现对设备的双控。
NT桥是一种非透明的 PCIE连接桥, 转换器的 NT桥可以防止主机列举 连接到 NT桥后面的设备, 同时允许打开窗口以交换数据, 并隔离连接到 NT 桥后面的设备及内存空间。
现有双控系统中主要通过 PCIE NT桥镜像技术实现对设备的双控,例如 当一个控制板 A出现故障时, 与该控制板通过 NT桥连接的另一个控制板 B 可以采用镜像技术对此前由控制板 A控制的设备的请求进行间接处理,从而 通过控制板 A和 B实现对设备的双控, 提高系统的抗故障能力。
在实现本发明的过程中, 发明人发现, 通过镜像技术间接对设备的请求 进行处理, 其处理响应速度通常相对较慢。 发明内容
本发明实施例提供一种设备接管方法和装置及双控系统, 通过直接接管 故障设备实现双控, 能够提高对设备请求的处理响应速度。
为解决上述技术问题, 本发明实施例中提供以下技术方案:
一种设备接管方法, 包括:
第一控制板对快速外设组件互联 PCIE总线进行拓朴发现; 在拓朴发现到非透明 NT桥时,按照设定的资源预留策略为所述 NT桥所在 的转换器以及和所述转换器的下行端口连接的设备预留资源, 所述 NT桥所在 的转换器以及和所述转换器的下行端口连接的设备当前由第二控制板控制; 在第二控制板出现故障时, 从第二控制板接管所述转换器以及和所述转 换器的下行端口连接的设备, 将所述预留的资源分配给所述转换器以及和所 述转换器的下行端口连接的设备, 使所述转换器以及和所述转换器的下行端 口连接的设备在第一控制板的控制下工作。
一种设备接管装置, 包括:
拓朴发现模块, 用于对 PCIE总线进行拓朴发现;
资源预留模块, 用于在所述拓朴发现模块拓朴发现到 NT桥时, 按照设定 的资源预留策略为所述 NT桥所在的转换器以及和所述转换器的下行端口连接 的设备预留资源, 所述 NT桥所在的转换器以及和所述转换器的下行端口连接 的设备当前由第二控制板控制;
接管模块, 用于在第二控制板出现故障时, 从第二控制板接管所述转换 器以及和所述转换器的下行端口连接的设备, 将所述资源预留模块预留的资 源分配给所述转换器以及和所述转换器的下行端口连接的设备, 使所述转换 器以及和所述转换器的下行端口连接的设备在所述设备接管装置的控制下工 作。
一种双控系统, 包括第一控制板和第二控制板,
其中, 第一控制板, 用于对 PCIE总线进行拓朴发现; 在拓朴发现到 NT桥 时, 按照设定的资源预留策略为所述 NT桥所在的转换器以及和所述转换器的 下行端口连接的设备预留资源, 所述 NT桥所在的转换器以及和所述转换器的 下行端口连接的设备当前由第二控制板控制; 在第二控制板出现故障时, 从 第二控制板接管所述转换器以及和所述转换器的下行端口连接的设备, 将所 述预留的资源分配给所述转换器以及和所述转换器的下行端口连接的设备, 使所述转换器以及和所述转换器的下行端口连接的设备在第一控制板的控制 下工作。
由上述技术方案可以看出, 本发明实施例中釆用的技术方案具有如下优 点: 由于第一控制板事先为当前由第二控制板控制的带有 NT桥的转换器以及 和所述转换器的下行端口连接的设备预留了资源, 使得在从第二控制板接管 上述转换器以及与该转换器的下行端口连接的设备时, 可以将事先预留的资 源分配给上述转换器以及与该转换器下行端口连接的设备, 使被接管的设备 具有相对足够的可操作资源, 能够在第一控制板的控制下正常工作, 通过直 接接管设备的方式实现双控, 可以提高请求处理的响应速度。
附图说明
为了更清楚地说明本发明实施例和现有技术中的技术方案, 下面将对实 施例和现有技术描述中所需要使用的附图作简单地介绍, 显而易见地, 下面 描述中的附图仅仅是本发明的一些实施例, 对于本领域普通技术人员来讲, 在不付出创造性劳动性的前提下, 还可以根据这些附图获得其他的附图。
图 1是本发明实施例一提供一种设备接管方法流程图;
图 2是本发明实施例二提供一种设备接管方法流程图;
图 3-a是本发明实施例二提供一种资源预留的双控架构示意图; 图 3-b是本发明实施例二提供一种预留资源分配的双控架构示意图; 图 4是本发明实施例三提供一种设备接管方法流程图;
图 5-a是本发明实施例三提供一种资源未预留的双控架构示意图; 图 5-b是本发明实施例三提供一种资源预留的双控架构示意图; 图 5-c是本发明实施例三提供一种预留资源分配的双控架构示意图; 图 6是本发明实施例四提供一种设备接管装置示意图;
图 7是本发明实施例五提供一种双控系统示意图。 具体实施方式
本发明实施例提供一种设备接管的方法和装置及双控系统, 通过直接接 管设备的方式实现双控, 可以提高对设备请求的处理响应速度。
下面将结合本发明实施例中的附图, 对本发明实施例中的技术方案进行 清楚、 完整地描述, 显然, 所描述的实施例仅仅是本发明一部分实施例, 而 不是全部的实施例。 基于本发明中的实施例, 本领域普通技术人员在没有做 出创造性劳动前提下所获得的所有其他实施例, 都属于本发明保护的范围。 首先说明的是, 随着转换器的发展, 带有 NT桥的转换器可以设置失败 接管(FV, Failover )功能, 所谓失败接管是指带有 NT桥转换器具有端口类 型转换的功能, 即在控制板的控制下, 转换器的 NT桥可以转换为 up端口, 转换器的 up端口也可以转换为 NT桥, 以此类推。
本发明实施例主要是利用带有 NT桥的转换器的失败接管功能, 实现对 设备的直接接管, 下面通过具体实施例进行详细的说明。 实施例一、
一种设备接管方法, 流程图如图 1所示, 该方法可以包括:
步骤 110、 第一控制板对快速外设组件互联总线进行拓朴发现。
其中, 所谓拓朴发现, 是指通过扫描系统的 PCIE总线, 发现不同总线及 其连接的设备和功能。
步驟 120、 在拓朴发现到 NT桥时, 按照设定的资源预留策略为该 NT桥所 在的转换器以及和所述转换器的下行端口连接的设备预留资源, 上述 NT桥所 在的转换器以及和所述转换器的下行端口连接的设备当前由第二控制板控 制。
在一种应用场景下, 第一控制板可以在操作系统启动时, 按照深度遍历 的方式对 PCIE总线的所有支路进行拓朴发现。
在另一种应用场景下, 第一控制板也可以在操作系统稳定运行过程中, 按照深度遍历的方式重新对 PCIE总线中 NT桥所在的支路进行拓朴发现。
在上述两种应用场景下, 当拓朴发现到 NT桥时, 可以按照设定的资源预 留策略为当前由第二控制板控制的 NT桥所在的转换器以及和所述转换器的下 行端口连接的设备预留总线资源和 /或地址空间资源等。
在一种应用场景下, 上述设定的资源预留策略可以包括: 事先设定的在 不同场景下为 NT桥所在的转换器以及和所述转换器的下行端口连接的设备预 留资源的数量。 举例来说, 若某个带有 NT桥的转换器下行端口通常连接网卡 设备, 则可以根据网卡设备地址空间资源的需求, 设定预留地址空间资源的 数量, 尽量为该网卡设备预留足够的地址空间资源, 以此类推。
步骤 130、 在第二控制板出现故障时, 从第二控制板接管上述转换器以及 和所述转换器的下行端口连接的设备, 将上述预留的资源分配给上述转换器 以及和所述转换器的下行端口连接的设备, 使上述转换器以及和所述转换器 连接的下行端口的设备在第一控制板的控制下工作。
在一种应用场景下, 在当前控制上述转换器的第二控制板出现故障时, 第一控制板可以将上述转换器的 NT桥转换设置为 up端口, 即利用上述转换器 的失败接管功能, 直接接管上述转换器以及和所述转换器的下行端口连接的 设备。
在进行设备接管的同时, 第一控制板可以根据上述转换器以及和所述转 换器的下行端口连接的设备所需的总线资源和 /或地址空间资源等, 将上述预 留的资源分配给上述转换器以及和所述转换器的下行端口连接的设备, 使之 前由第二控制板控制下的转换器以及和所述转换器的下行端口连接的设备能 够在第一控制板的控制下继续正常工作。 举例来说, 若第一控制板事先预留 了 3条总线资源和 6M地址空间资源, 从第二控制板接管的转换器当前包括 1个 up端口和 2个 down端口, 2个 down端口连接了两个分别需要 2M地址空间的端点 设备, 则第一控制板可以将事先预留的 3条总线资源分别分配给上述转换器的 1个 up端口和 2个 down端口, 将事先预留的 6M地址空间资源平均或非平均的分 配给连接到上述转换器的 down端口的两个端点设备。
其中,连接到上述转换器下行端口的设备可以包括端点设备和 /或桥设备。 可以理解的是, 本实施例上述的技术方案可以由第一控制板或第一控制 板的部分功能模块来具体实施。
可以理解, 本实施例是以第一控制板事先为第二控制板控制的 NT桥所在 的转换器以及和所述转换器的下行端口连接的设备预留资源; 当第二控制板 故障时, 第一控制板从第二控制板接管上述转换器以及和所述转换器连接的 下行端口的设备, 并为其分配预留的资源为例进行说明的, 当然, 第二控制 板也可以事先为第一控制板控制的 NT桥所在的转换器以及和所述转换器的下 行端口连接的设备预留资源; 当第一控制板故障时, 第二控制板从第一控制 板接管上述转换器以及和所述转换器的下行端口连接的设备, 并为其分配预 留的资源, 实现第二控制板从第一控制板接管设备, 实现对设备的双控。
由上述技术方案可以看出, 本实施例第一控制板事先为第二控制板控制 的带有 NT桥的转换器以及和所述转换器的下行端口连接的设备预留资源, 在 从第二控制板接管上述转换器以及与该转换器的下行端口连接的设备后, 将 事先预留的资源分配给上述转换器以及与该转换器下行端口连接的设备, 使 被接管的设备具有相对足够的可操作的资源, 能够在第一控制板的控制下正 常工作, 通过直接接管设备的方式实现双控, 可以提高请求处理的响应速度。 实施例二、
为便于理解, 本实施以控制板 A在系统启动时, 为当前由控制板 B控制的 带有 NT桥的转换器以及和所述转换器的下行端口连接的设备预留资源, 在接 管上述转换器以及和所述转换器的下行端口连接的设备后, 为被接管的设备 分配预留的资源为例, 进行具体说明。
一种设备接管方法, 流程图如图 2所示, 该方法可以包括:
步骤 201、 控制板 A对 PCIE总线的全部支路进行拓朴发现。
步骤 202、 当控制板 A拓朴发现到 NT桥时, 按照设定的资源预留策略为当 前由控制板 B控制的上述 NT桥所在的转换器以及和所述转换器的下行端口连 接的设备预留资源。
在一种应用场景下, 在控制板 A的操作系统启动时, 控制板 A的系统内核 可以按照深度遍历的方式, 依次对 PCIE总线的所有支路进行拓朴发现。
具体的,控制板 A可以根据 PCIE总线下的各个设备自身携带的类型标识判 断出该设备的类型以及资源需求, 从而可以根据不同类型设备的资源需求为 其分配总线资源和 /或地址空间资源。
其中, PCIE总线下的设备可以包括端点设备(例如, 网卡、 显卡等功能 设备) 和桥设备(例如, 转换器等)。
上述设定的资源预留策略可以包括事先设定的在不同场景下为 NT桥所在 的转换器以及和所述转换器的下行端口连接的设备预留资源的数量, 具体可 以根据应用场景的不同来设定需要预留的资源数量。
举例来说, 当控制板 A拓朴发现到端点设备时, 可以根据该端点设备的资 源需求为该端点设备分配地址空间资源; 在拓朴发现到不带有 NT桥的转换器 时, 可以先为该转换器分配一条总线资源, 然后继续拓朴该总线下的设备; 在拓朴发现到 NT桥时,按照设定的资源预留策略为 NT桥所在的转换器预留总 线资源 (或预留总线资源和地址空间资源), 以及为和所述转换器的下行端口 连接的设备预留总线资源和 /或地址空间资源, 以此类推, 直至完成 PCIE总线 全部支路的拓朴发现。
需要说明的是,在进行 PCIE总线的拓朴发现时,控制板 A可能拓朴发现到 一个或多个 NT桥, 且上述 1个或多个 NT桥所在的转换器以及和所述转换器的 下行端口连接的设备当前可能由其它的一个或多个控制板实际控制,控制板 A 可以为拓朴发现到的每个 NT桥所在的转换器以及和所述转换器的下行端口连 接的设备预留资源。本实施例以控制板 A只拓朴发现到 1个 NT桥且该 NT桥所在 的转换器当前由控制板 B控制为例进行具体说明。
步驟 203、 在控制板 B出现故障时, 控制板 A从控制板 B接管上述转换器以 及和所述转换器的下行端口连接的设备, 将预留的资源分配给上述转换器以 及和所述转换器的下行端口连接的设备。
在一种应用场景下, 当控制板 B出现故障时, 控制板 A可以将上述转换器 与控制板 B连接的 up端口转换为 NT桥, 将上述转换器与控制板 A连接的 NT桥 转换为 up端口 ,进而从控制板 B接管上述转换器以及和所述转换器的下行端口 连接的设备。 同时, 控制板 A可以按照深度遍历的方式对上述转换器以及和所 迷转换器连接的下行端口的设备进行拓朴发现, 根据上述转换器的总线资源 需求以及和所述转换器连接的下行端口的设备的总线资源和 /或地址空间资源 需求, 将事先预留的部分或全部总线资源分配给上述转换器的上行端口和下 行端口, 将事先预留的部分或全部总线资源和 /或地址空间资源分配给和所述 转换器连接的下行端口的设备, 使得该转换器以及该转换器的下行端口的设 备可以在控制板 A的控制下继续正常工作。
可以看出, 由于对被控制板 A接管的上述转换器以及和所述转换器的下行 端口连接的设备分配到了足够的可操作资源, 因而可以在控制板 A的控制下正 常的工作, 成功实现控制板 B和控制板 A对上述转换器以及和所述转换器的下 行端口连接的设备的双控。
可以理解的是, 若控制板 A拓朴发现到多个 NT桥, 且上述多个 NT桥所在 的转换器当前由其它 1个或多个控制板控制的情况,对于每个 NT桥所在的转换 器以及和所述转换器的下行端口连接的设备,控制板 A均可釆用上述方式进行 设备接管和预留资源的分配, 此处不再赘述。 可以理解,本实施例是以控制板 A事先为控制板 B控制的 NT桥所在的转换 器以及和所述转换器连接的下行端口的设备预留资源为例进行说明; 当控制 板 B故障时, 控制板 A从控制板 B接管上述转换器以及和所述转换器的下行端 口连接的设备, 并为其分配预留的资源为例进行说明的, 当然, 控制板 B也可 以事先为控制板 A控制的 NT桥所在的转换器以及和所迷转换器的下行端口连 接的设备预留资源; 当控制板 A故障时, 控制板 B从控制板 A接管上述转换器 以及和所述转换器的下行端口连接的设备, 并为其分配预留的资源, 控制板 B 的实现上述接管的过程可以与控制板 A的实现过程相同, 此处不再赘述。
为更好的理解本实施例上述的技术方案, 下面通过具体的实例对本实例 技术方案做进一步详细的描迷。
请参见图 3-a和 3-b, 以控制板 310为上述的控制板 A, 控制板 320为上述的 控制板 B为例。 如图所示, 控制板 310包括: 转换器 311、 转换器 312和转换器 313。 控制板 320包括: 转换器 321、 转换器 322和转换器 323。
其中, 转换器 311的 down端口 3111和转换器 312连接, down端口 3112和转 换器 313连接。 转换器 312的 NT桥端口和转换器 322的一个 down端口连接; 转 换器 323的端口 3231和转换器 313的 1个 down端口连接。
在控制板 310的操作系统启动时, 控制板 310的系统内核按照深度遍历的 方式拓朴发现 PCIE总线下的全部支路, 为拓朴发现到的设备分配或预留地址 空间资源和 /或总线资源。
如图 3-a所示, 转换器 323的端口 3231当前为 NT桥, 端口 3132当前为 up端 口。 控制板 310为转换器 311的 up端口分配 0号总线资源、 down端口 3111分配 1 号总线资源、 down端口 3112分配了 6号总线资源; 为转换器 312的 up端口分配 2 号总线资源、 down端口分配 3、 4、 5号总线资源; 为端点设备 331分配 1M ( OxOOOOOOOO-OxOOOfffff ) 的地址空间资源、 为端点设备 332分配 2M ( 0x00100000〜0x002fffff )的地址空间资源; 为转换器 313的 up端口和 3个 down 端口分配 7、 8、 12、 13号总线; 为转换器 323预留 9、 10、 11号总线, 为端点 设备 343和 344预留 6M ( 0x00e00000~0x013ffff ) 的地址空间资源; 为端点设 备 333分配 2M ( 0x01400000〜0x015fffff ) 的地址空间资源、 为端点设备 334分 配 2M ( 0x01600000^0x017fffff ) 的地址空间资源。 上述拓朴发现完成后, PCIE的总线号和地址资源按照顺序递增, 转换器 311的端口 3111允许的有效总线号范围是: 1 ~ 5号, 地址空间范围是: 0x00000000^ 0x002ffff ( 3M ); 端口 3112允许的有效总线号范围是: 6 ~ 13号, 地址空间范围是: 0x00e00000〜 0x017fffff ( 10M )。
如图 3-b所示, 当控制板 320出现故障或其它原因时, 控制板 310从控制板 320接管转换器 323以及端点设备 343和端点设备 344, 即将转换器 323的端口 3232转换设置为 NT桥, 将端口 3231转换设置为 up端口。 同时, 控制板 310可以 按照深度遍历的方式对 8号总线下的设备进行拓朴发现, 根据转换器 323以及 端点设备 343和端点设备 344的资源需求, 将预留的 9、 10、 11号总线资源分配 给转换器 323 的 1 个 up端 口 和 2个 down端 口 , 将预 留 的 6M ( 0x00e00000~0x013fffff)地址空间资源分配给端点设备 343和 344, 例如可以 为端点设备 343分配 2M ( 0x00e00000〜0x00ffffff )、 为端点设备 344分配 4M ( 0x01000000^0x013fffff )„
此时, 对于控制板 310来说, 由于为转换器 323以及端点设备 343和端点设 备 344具有相对足够的可操作资源, 因此使得转换器 323以及端点设备 343和端 点设备 344能够在控制板 310的控制下继续正常工作, 从而控制板 310和控制板 320利用 NT桥实现了对转换器 323以及端点设备 343和端点设备 344的双控, 进 而提高了系统抗故障能力。
可以理解的是,控制板 320也可以按照上述方式为对 PCIE总线进行拓朴发 现, 并为控制板 310的 NT桥所在的转换器 312以及和所述转换器连接的下行端 口的端点设备 331和端点设备 332预留资源, 当控制板 310故障时, 控制板 320 也可以按照上述方式接管转换器 312以及端点设备 331和端点设备 332, 通过直 接接管转换器 312以及端点设备 331和端点设备 332实现双控。
由上述技术方案可以看出, 本实施例中控制板 A在操作系统启动时, 为当 前由控制板 B控制的带有 NT桥的转换器以及和所述转换器的下行端口连接的 设备预留资源,在从控制板 B失败接管上述转换器以及与该转换器的下行端口 连接的设备后, 将事先预留的资源分配给上述转换器以及与该转换器下行端 口连接的设备, 使得被失败接管的设备具有相对足够的可操作的资源, 能够 在控制板 A的控制下正常工作, 通过直接接管设备的方式实现双控, 可以提高 设备请求的处理响应速度。 实施例三、
为便于理解, 本实施以控制板 A在系统稳定运行后的任意时刻, 为当前由 控制板 B控制的带有 NT桥的转换器以及和所述转换器的下行端口连接的设备 预留资源, 在接管上述转换器以及和所述转换器的下行端口连接的设备后, 为被接管的设备分配预留的资源为例, 进行具体说明。
一种设备接管方法 , 流程图可如图 4所示, 该方法可以包括:
步骤 401、 在控制板 A的操作系统启动时, 控制板 A对其 PCIE总线的 各条支路进行拓朴发现, 为拓朴发现到的端点设备和不带有 NT桥的转换器 分配地址空间资源和 /或总线资源。
在一种应用场景下, 在控制板 A的操作系统启动时, 控制板 A的系统内核 可以按照深度遍历的方式依次对 PCIE总线的所有支路进行拓朴发现。
具体的,控制板 A可以根据 PCIE总线下的各个设备自身携带的类型标识判 断出该设备的类型以及资源需求, 从而可以根据不同类型设备的资源需求为 其分配总线资源和 /或地址空间资源。
其中, 由于 NT桥的属性特点, 在拓朴发现 PCIE总线时, 控制板 A若拓朴 发现到 NT桥, 并不能对该 NT桥所在的转换器以及和所述转换器的下行端口连 接的设备进行进一步的拓朴发现, 故而直接跳过对 NT桥所在的转换器以及和 所述转换器的下行端口连接的设备进行拓朴发现。
举例来说, 当控制板 A拓朴发现到端点设备时, 可以根据该端点设备的资 源需求为该端点设备分配地址空间资源; 在拓朴发现到不带有 NT桥的转换器 时, 可以先为该转换器分配一条总线资源, 然后继续拓朴该总线下的设备; 在拓朴发现到 NT桥时, 可以记录该 NT桥的位置, 并停止拓朴发现该 NT桥所 在转换器以及该转换器下面连接的设备, 且暂不为其分配资源, 以此类推, 直至完成 PCIE总线全部支路的拓朴发现。
步骤 402、在操作系统稳定运行中, 控制板 A重新拓朴发现 PCIE总线中 NT 桥所在的支路。
在一种应用场景下, 控制板 A可以在操作系统稳定运行后的任意时刻, 按 照深度遍历的方式对 PCIE总线中 NT桥所在的支路重新进行拓朴发现, 为该条 支路的设备预留资源或重新分配资源。
步骤 403、 控制板 A在拓朴发现到带有 NT桥时, 按照设定的资源预留策略 为当前由控制板 B控制的上述 NT桥所在的转换器以及和所述转换器的下行端 口连接的设备预留资源。
具体的, 在控制板 A重新拓朴发现 NT桥所在的支路过程中, 在拓朴发现 到端点设备时, 可以根据该端点设备的资源需求重新为该端点设备分配地址 空间资源, 在拓朴发现到不带有 NT桥的转换器时, 可以重新为该转换器分配 一条总线资源, 然后继续拓朴发现该总线下的设备; 在拓朴发现到 NT桥时, 可以按照设定的资源预留策略为 NT桥所在的转换器以及和所述转换器的下行 端口连接的设备预留总线资源和 /或地址空间资源等, 以此类推, 直至完成上 述对 NT桥所在的支路的拓朴发现。
其中, 上述设定的资源预留策略可以包括事先设定的在不同场景下为 NT 桥所在的转换器以及和所述转换器的下行端口连接的设备预留资源的数量, 具体可以根据应用场景的不同来设定需要预留的资源数量。
可以理解的是, 控制板 A在系统启动时, 可能拓朴发现到一个或多个 NT 桥, 且上述一个或多个 NT桥所在的转换器以及和所述转换器的下行端口连接 的设备当前可能由其它的 1个或多个控制板实际控制, 在系统稳定运行后的任 意时刻, 控制板 A可以重新拓朴发现 PCIE总线中每个 NT桥所在的支路, 并为 拓朴发现到的每个 NT桥所在的转换器以及和所述转换器的下行端口连接的设 备预留资源, 为拓朴发现到的其它设备重新分配资源。 本实施例以控制板 A在 系统启动时只拓朴发现到 1个 NT桥, 且该 NT桥所在的转换器当前由控制板 B 控制为例, 进行具体说明。
步骤 404、 在控制板 B出现故障时, 控制板 A从控制板 B失败接管上述转换 器以及和所述转换器的下行端口连接的设备, 将预留的资源分配给上述转换 器以及和所述转换器的下行端口连接的设备。
在一种应用场景下, 当控制板 B出现故障时, 控制板 A可以将上述转换器 与控制板 B连接的 up端口转换为 NT桥, 将上述转换器与控制板 A连接的 NT桥 转换为 up端口,进而从控制板 B失败接管上述转换器以及和所述转换器的下行 端口连接的设备。 同时, 控制板 A可以按照深度遍历的方式对上述转换器以及 和所述转换器的下行端口连接的设备进行拓朴发现, 根据上述转换器的总线 资源需求以及和所述转换器的下行端口连接的设备的总线资源和 /或地址空间 资源需求, 将事先预留的部分或全部总线资源分配给上述转换器的上行端口 和下行端口, 将事先预留的部分或全部总线资源和 /或地址空间资源分配给和 所述转换器的下行端口连接的设备, 使得该转换器以及该转换器的下行端口 的设备可以在控制板 A的控制下继续正常工作。
可以看出, 由于被控制板 A接管的上述转换器以及和所述转换器的下行端 口连接的设备分配到了足够的可操作资源, 因而可以在控制板 A的控制下正常 的工作, 成功实现控制板 B和控制板 A对上述转换器以及和所述转换器的下行 端口连接的设备的双控。
可以理解的是, 若控制板 A拓朴发现到多个 NT桥, 且上述多个 NT桥所在 的转换器当前由其它 1个或多个控制板控制的情况,对于每个 NT桥所在的转换 器以及和所述转换器的下行端口连接的设备,控制板 A均可采用上述方式进行 设备接管和预留资源的分配, 此处不再赘述。
可以理解,本实施是以控制板 A事先为控制板 B控制的 NT桥所在的转换器 以及和所述转换器的下行端口连接的设备预留资源; 当控制板 B故障时, 控制 板 A从控制板 B接管上述转换器以及和所述转换器的下行端口连接的设备, 并 为其分配预留的资源为例进行说明的, 当然, 控制板 B也可以事先为控制板 A 控制的 NT桥所在的转换器以及和所述转换器的下行端口连接的设备预留资 源; 当控制板 A故障时, 控制板 B从控制板 A接管上述转换器以及和所述转换 器连接的下行端口的设备, 并为其分配预留的资源, 控制板 B实现设备接管的 过程可以与控制板 A的实现过程相同, 此处不再赘述。
为更好的理解本实施例的技术方案, 下面通过具体的实例对本实施例的 技术方案做进一步详细的描述。
请参见图 5-a、 5-b和 5-c, 以控制板 510为上述的控制板 A, 控制板 520为上 述的控制板 B为例。 如图所示, 控制板 510包括: 转换器 511、 转换器 512和转 换器 513。 控制板 520包括: 转换器 521、 转换器 522和转换器 523。
其中, 转换器 511的 down端口 5111和转换器 512连接, down端口 5112和转 换器 513连接。 转换器 512的 NT桥端口和转换器 522的一个 down端口连接; 转 换器 523的端口 5231和转换器 513的 1个 down端口连接。
在控制板 510的操作系统启动时, 控制板 510的系统内核可以按照深度遍 历的方式拓朴发现 PCIE总线下的全部支路, 为拓朴发现到的端点设备和不带 有 NT桥的转换器分配地址空间资源和 /或总线资源。
如图 5-a所示, 转换器 523的端口 5231当前为 NT桥, 端口 5232当前为 up端 口。 在操作系统启动时, 控制板 510为转换器 511的 up端口分配 0号总线资源、 down端口 5111分配 1号总线资源、 down端口 5112分配了 6号总线资源; 为转换 器 512的 up端口和 down端口分配 2、 3、 4、 5号总线资源; 为端点设备 531分配 1M ( 0x00000000~0x000ffff ) 的地址空间资源、 为端点设备 532分配 2M ( 0x00100000~0x002fffff ) 的地址空间资源; 为转换器 513的 up端口和 down端 口分配 7、 8、 9、 10号总线, 为端点设备 533分配 2M ( 0x00e00000~0x00ffffff ) 的地址空间资源, 为端点设备 534分配 2M ( 0x01000000^0x011 fffff ) 的地址空 间资源; 转换器 523以及端点设备 543和 544暂不分配资源。
上述系统启动时进行的拓朴发现完成后, 转换器 511的端口 5111允许控制 的总线号范围是: 1 ~ 5号, 允许的地址空间范围是: 0x00000000〜 0x002fffff ( 3M ); 端口 5112允许控制的总线号范围是: 6 ~ 10号, 允许地址空间范围是: 0x00e00000〜 0x011 fffff ( 4M )„
如图 5-b所示,在操作系统稳定运行后的任意时刻,控制板 510重新拓朴发 现 NT桥所在的支路, 从 NT桥所在的支路的顶端 ( 0号总线) 开始向下重新拓 朴发现, 为连接到 0号总线以下支路的设备预留和重新分配资源。
举例来说, 若此前的地址空间资源分配到了 OxOlffffff, 总线号分配到了 30号, 则控制板 510可以重新为转换器 511的 down端口 5112分配 31号总线; 重 新为转换器 513的 up端口和 3个 down端口分配 32、 33、 37、 38号总线; 为转换 器 523预留 34 、 35和 36号总线, 为端点设备 543和 544预留 6M ( 0x02000000〜0x025fffff ) 的地址空间资源; 为端点设备 533重新分配 2M ( 0x02600000~0x027fffff ) 的地址空间资源, 为端点设备 534重新分配 2M ( 0x02800000~0x029fffff ) 的地址空间资源。
上述重新进行的拓朴发现完成后, 转换器 511的端口 5111允许控制的总线 号范围是: 1 ~ 5, 允许的地址空间范围是: 0x00000000〜 0x002fffff ( 3M ); 端 口 5112允许控制的总线号范围是: 31 ~ 38,允许地址空间范围是: 0x02000000- 0x029fffff ( 10M )。
如图 5-c所示, 当控制板 520出现故障或其它原因时, 控制板 510从控制板 520接管转换器 523以及端点设备 543和端点设备 544, 将转换器 523的端口 5232 转换设置为 NT桥, 将端口 5231转换设置为 up端口。
控制板 510可以按照深度遍历的方式, 对 33号总线下的设备进行拓朴发 现,根据转换器 523以及端点设备 543和端点设备 544的资源需求,将预留的 34、 35和 36号总线资源分配给转换器 523的 1个 up端口和 2个 down端口, 将预留的 6M ( 0x02000000〜0x025fffff )地址空间资源分配给端点设备 543和 544, 例如, 可以为端点设备 543分配 3M ( 0x02000000〜0x022fffff )、为端点设备 544分配 3M ( 0x02300000~0x025fffff )。
此时, 对于控制板 510来说, 转换器 523以及端点设备 543和端点设备 544 具有相对足够的可操作资源, 控制板 510能够控制其正常的工作, 控制板 510 和控制板 520利用 NT桥, 实现了对转换器 523以及端点设备 543和端点设备 544 的双控, 进而相对提高了系统抗故障能力。
可以理解的是,控制板 520也可以按照上述方式为对 PCIE总线进行拓朴发 现, 并为控制板 510的 NT桥所在的转换器 512以及和所述转换器连接的下行端 口的端点设备 531和端点设备 532预留资源, 当控制板 510故障时, 控制板 520 也可以按照上述方式接管转换器 512以及端点设备 531和端点设备 532 , 通过直 接接管转换器 512以及端点设备 531和端点设备 532实现双控。
由上述技术方案可以看出,本实施例中控制板 A在操作系统稳定运行的任 意时刻, 为当前由控制板 B控制的带有 NT桥的转换器以及和所述转换器的下 行端口连接的设备预留资源,在从控制板 B失败接管上述转换器以及与该转换 器的下行端口连接的设备后, 将事先预留的资源分配给上述转换器以及与该 转换器下行端口连接的设备, 使得被失败接管的设备具有相对足够的可操作 的资源, 能够在控制板 A的控制下正常工作, 通过直接接管设备的方式实现双 控, 可以相对提高设备情求的处理响应速度。
为便于更好的实施本发明上述实施例的技术方案, 本发明实施例中还提 供一种设备接管装置。 实施例四
请参见图 6, 本发明实施例四的一种设备接管装置 600可以包括: 拓朴发 现模块 610、 资源预留模块 620和接管模块 630。
其中, 拓朴发现模块 610, 用于对 PCIE总线进行拓朴发现;
资源预留模块 620, 用于在拓朴发现模块 610拓朴发现到 NT桥时, 按照设 定的资源预留策略为该 NT桥所在的转换器以及和所述转换器的下行端口连接 的设备预留资源, 上述 NT桥所在的转换器以及和所述转换器的下行端口连接 的设备当前由第二控制板控制;
在具体应用时, 上述设定的资源预留策略可以是: 事先设定的在不同场 景下为 NT桥所在的转换器以及和所述转换器的下行端口连接的设备预留资源 的多少。 举例来说, 若带有 NT桥的转换器下行端口通常连接网卡设备, 则可 以根据网卡设备地址空间资源的需求, 设定预留地址空间资源的数量, 尽量 为该网卡设备预留足够的地址空间资源, 以此类推。
接管模块 630, 用于在第二控制板出现故障时, 从第二控制板失败接管上 述转换器以及和所述转换器的下行端口连接的设备, 将资源预留模块 620预留 的资源分配给上述转换器以及和所述转换器的下行端口连接的设备, 使上述 转换器以及和所述转换器连接的下行端口的设备在设备接管装置 600控制下 工作。
在一种应用场景下, 拓朴发现模块 610具体可以用于, 在系统启动时, 对 PCIE总线的全部支路进行拓朴发现。
在一种应用场景下,拓朴发现模块 610具体可以用于,在系统运行过程中, 对 PCIE总线中 NT桥所在的支路进行拓朴发现。
在一种应用场景下,资源预留模块 620具体可以用于,在拓朴发现模块 610 拓朴发现到 NT桥时, 按照设定的资源预留策略为当前由第二控制板控制的上 述 NT桥所在的转换器预留总线资源以及为和所述转换器的下行端口连接的设 备预留总线资源和 /或地址空间资源。
在一种应用场景下, 接管模块 630可以包括:
接管子模块, 用于在第二控制板出现故障时, 从第二控制板接管上述转 换器以及和所述转换器的下行端口连接的设备。 资源分配子模块, 用于将资源预留模块 620为上述转换器预留的总线资源 分配给上述转换器的上行端口和下行端口, 将资源预留模块 620为连接到上述 转换器的下行端口的设备预留的总线资源和 /或地址空间资源分配给连接到上 述转换器的下行端口的设备。
在具体应用中, 在第二控制板出现故障时, 接管子模块可以将上述转换 器的 NT桥转换设置为 up端口, 进而接管上述转换器以及和所述转换器的下行 端口连接的设备, 其中, 连接到上述转换器下行端口的设备可以包括端点设 备和 /或桥设备。
资源分配子模块可以按照深度遍历的方式对上述转换器以及和所述转换 器的下行端口连接的设备进行拓朴发现, 根据上述转换器以及和所述转换器 的下行端口连接的设备所需的总线资源和 /或地址空间资源等, 将资源预留模 块 620预留的资源分配给上述转换器以及和所述转换器的下行端口连接的设 备。
举例来说, 若资源预留模块 620事先预留了 3条总线资源和 6M地址空间资 源, 失败接管子模块从第二控制板接管的转换器当前包括 1个 up端口和 2个 down端口, 2个 down端口连接了两个分别需要 2M地址空间的端点设备, 则资 源分配子模块可以将资源预留模块 620事先预留的 3条总线资源分别分配给上 述转换器的 1个 up端口和 2个 down端口, 将资源预留模块 620事先预留的 6M地 址空间资源平均或非平均的分配给连接到上述转换器的 down端口的两个端点 设备。
可以理解是的, 本实施例的设备接管装置 600可以设置于如实施例二或三 的控制板 A或其它设备中,其各个功能模块的功能可以根据实施例二或三中的 方法具体实现, 其具体实现过程可参照上述实施例中的相关描述, 此处不再 由上述技术方案可以看出, 本实施例中, 设备接管装置事先为第二控制 板控制的带有 NT桥的转换器以及和所述转换器的下行端口连接的设备预留资 源, 在从第二控制板失败接管上述转换器以及与该转换器的下行端口连接的 设备后, 将事先预留的资源分配给上述转换器以及与该转换器下行端口连接 的设备, 使得被失败接管的设备具有相对足够的可操作的资源, 能够在上述 装置的控制下正常工作, 通过直接接管设备的方式实现双控, 可以相对提高 设备请求的处理响应速度。
为便于更好的实施本发明上述实施例的技术方案, 本发明实施例中还提 供一种双控系统。 实施例五
请参见图 7, 本发明实施例五的一种双控系统可以包括: 第一控制板 710 和第二控制板 720。
其中, 第一控制板 710, 用于对 PCIE总线进行拓朴发现; 用于在拓朴发现 到 NT桥时,按照设定的资源预留策略为该 NT桥所在的转换器以及和所述转换 器的下行端口连接的设备预留资源, 上述 NT桥所在的转换器以及和所述转换 器的下行端口连接的设备当前由第二控制板控制; 在第二控制板 720出现故障 时, 从第二控制板 720接管上述转换器以及和所述转换器的下行端口连接的设 备, 将上述预留的资源分配给上述转换器以及和所迷转换器的下行端口连接 的设备, 使上述转换器以及和所述转换器的下行端口连接的设备在第一控制 板 710的控制下工作。
此外, 第二控制板 720也可以对 PCIE总线进行拓朴发现; 用于在拓朴发现 到 NT桥时, 按照设定的资源预留策略为当前由第一控制板 710控制的上述 NT 桥所在的转换器以及和所述转换器的下行端口连接的设备预留资源; 在第一 控制板 710出现故障时, 从第一控制板 710接管上述转换器以及和所述转换器 的下行端口连接的设备, 将上述预留的资源分配给该转换器以及和所述转换 器的下行端口连接的设备, 使上述转换器以及和所述转换器的下行端口连接 的设备在第二控制板 720的控制下工作。
可以理解是的, 本实施例双控系统的第一控制板 710可以包括如实施例四 中的设备接管的装置 600 , 双控系统的各个实体的功能可以根据实施例二或三 中的方法具体实现, 其具体实现过程可参照上述实施例中的相关描述, 此处 不再赘述。
在上述实施例中, 对各个实施例的描述都各有侧重, 某个实施例中没 有详述的部分, 可以参见其他实施例的相关描述。
综上所述, 本发明实施例中, 第一控制板事先为第二控制板控制的带有 NT桥的转换器以及和所述转换器的下行端口连接的设备预留资源, 在从第二 控制板接管上述转换器以及与该转换器的下行端口连接的设备后, 将事先预 留的资源分配给上述转换器以及与该转换器下行端口连接的设备, 使被接管 设备具有相对足够的可操作的资源, 能够在第一控制板的控制下正常工作, 通过直接接管设备的方式实现双控, 可以相对提高设备请求的处理响应速度。
本领域普通技术人员可以理解实现上述实施例方法中的全部或部分流 程, 是可以通过计算机程序来指令相关的硬件来完成, 所述的程序可存储于 计算机可读取存储介质中, 该程序在执行时, 可包括如上述各方法的实施例 的流程。其中,所述的存储介质可为磁碟、光盘、只读存储记忆体(Read-Only Memory, ROM ) 或随机存储记忆体 ( Random Access Memory, RAM ) 等。
以上对本发明实施例所提供的一种设备接管方法和装置及双控系统进行 了详细介绍, 本文中应用了具体个例对本发明的原理及实施方式进行了阐述, 以上实施例的说明只是用于帮助理解本发明的方法及其核心思想; 同时, 对 于本领域的一般技术人员, 依据本发明的思想, 在具体实施方式及应用范围 上均会有改变之处, 综上所述, 本说明书内容不应理解为对本发明的限制。

Claims

权利要求
1、 一种设备接管方法, 其特征在于, 包括:
第一控制板对快速外设组件互联 PCIE总线进行拓朴发现;
在拓朴发现到非透明 NT桥时,按照设定的资源预留策略为所述 NT桥所在 的转换器以及和所述转换器的下行端口连接的设备预留资源, 所述 NT桥所在 的转换器以及和所述转换器的下行端口连接的设备当前由第二控制板控制; 在第二控制板出现故障时, 从第二控制板接管所述转换器以及和所述转 换器连接的下行端口的设备, 将所述预留的资源分配给所述转换器以及和所 述转换器的下行端口连接的设备, 使所述转换器以及和所述转换器的下行端 口连接的设备在第一控制板的控制下工作。
2、根据权利要求 1所述方法,其特征在于,所述对快速外设组件互联 PCIE 总线进行拓朴发现, 包括:
在第一控制板的操作系统启动时,对 PCIE总线的全部支路进行拓朴发现; 或
在第一控制板的操作系统运行过程中, 对 PCIE总线中 NT桥所在的支路进 行拓朴发现。
3、 根据权利要求 1或 2所述方法, 其特征在于, 所述按照设定的资源预留 策略为所述 NT桥所在的转换器以及连接到该转换器的下行端口的设备预留资 源, 包括:
按照设定的资源预留策略, 为当前由第二控制板控制的所述 NT桥所在的 转换器预留总线资源, 以及为连接到所述转换器的下行端口的设备预留总线 资源和 /或地址空间资源。
4、 根据权利要求 1或 2所述方法, 其特征在于, 所述将预留的资源分配给 所述转换器以及和所述转换器的下行端口连接的设备, 包括:
将为所述转换器预留的总线资源分配给所述转换器的上行端口和下行端 口, 将为和所述转换器连接的下行端口的设备预留的总线资源和 /或地址空间 资源分配给和所述转换器的下行端口连接的设备。
5、 一种设备接管装置, 其特征在于, 包括: 拓朴发现模块, 用于对 PCIE总线进行拓朴发现;
资源预留模块, 用于在所述拓朴发现模块拓朴发现到 NT桥时, 按照设定 的资源预留策略为所述 NT桥所在的转换器以及和所述转换器连接的下行端口 的设备预留资源, 所述 NT桥所在的转换器以及和所述转换器连接的下行端口 的设备当前由第二控制板控制;
接管模块, 用于在第二控制板出现故障时, 从第二控制板接管所述转换 器以及和所述转换器的下行端口连接的设备, 将所述资源预留模块预留的资 源分配给所述转换器以及和所述转换器的下行端口连接的设备, 使所述转换 器以及和所述转换器的下行端口连接的设备在所述设备接管装置的控制下工 作。
6、 根据权利要求 5所述装置, 其特征在于, 所述拓朴发现模块用于在系 统启动时, 对 PCIE总线的全部支路进行拓朴发现 。
7、 根据权利要求 5所述装置, 其特征在于, 所述拓朴发现模块还用于在 系统运行过程中, 对 PCIE总线中 NT桥所在的支路进行拓朴发现。
8、 根据权利要求 5-7中任一项所述装置, 其特征在于,
所述资源预留模块用于在所述拓朴发现模块拓朴发现到 NT桥时, 按照设 定的资源预留策略为所述 NT桥所在的转换器预留总线资源以及为和所述转换 器的下行端口连接的设备预留总线资源和 /或地址空间资源, 其中, 和上述转 换器连接的下行端口的设备包括端点设备和 /或桥设备。
9、 根据权利要求 8所述装置, 其特征在于, 所述接管模块包括: 接管子模块, 用于在第二控制板出现故障时, 从第二控制板接管所述转 换器以及和所述转换器的下行端口连接的设备;
资源分配子模块, 用于将所述资源预留模块为所述转换器预留的总线资 源分配给所述转换器的上行端口和下行端口, 将所述资源预留模块为和所述 转换器的下行端口连接的设备预留的总线资源和 /或地址空间资源分配给和所 迷转换器的下行端口连接的设备。
10、 一种双控系统, 其特征在于, 包括第一控制板和第二控制板, 其中, 第一控制板, 用于对 PCIE总线进行拓朴发现; 在拓朴发现到 NT桥 时, 按照设定的资源预留策略为所述 NT桥所在的转换器以及和所述转换器的 下行端口连接的设备预留资源, 所述 NT桥所在的转换器以及和所述转换器的 下行端口连接的设备当前由第二控制板控制; 在第二控制板出现故障时, 从 第二控制板接管所述转换器以及和所述转换器的下行端口连接的设备, 将所 述预留的资源分配给所述转换器以及和所述转换器的下行端口连接的设备, 使所述转换器以及和所述转换器连接的下行端口的设备在第一控制板的控制 下工作。
11、 根据权利要求 10所述的系统, 其特征在于, 所述将预留的资源分 配给所述转换器以及和所述转换器的下行端口连接的设备, 包括:
按照设定的资源预留策略, 为当前由第二控制板控制的 NT桥所在的转换 器预留总线资源, 以及为和所迷第二控制板控制的 NT桥所在的转换器连接的 下行端口的设备预留总线资源和 /或地址空间资源;
PCT/CN2010/075663 2009-08-07 2010-08-03 设备接管方法和装置及双控系统 WO2011015130A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US13/368,106 US8285910B2 (en) 2009-08-07 2012-02-07 Method and system for taking over devices

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN200910161060.0 2009-08-07
CN200910161060.0A CN101631083B (zh) 2009-08-07 2009-08-07 设备接管方法和装置及双控系统

Related Child Applications (1)

Application Number Title Priority Date Filing Date
US13/368,106 Continuation US8285910B2 (en) 2009-08-07 2012-02-07 Method and system for taking over devices

Publications (1)

Publication Number Publication Date
WO2011015130A1 true WO2011015130A1 (zh) 2011-02-10

Family

ID=41576034

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2010/075663 WO2011015130A1 (zh) 2009-08-07 2010-08-03 设备接管方法和装置及双控系统

Country Status (3)

Country Link
US (1) US8285910B2 (zh)
CN (1) CN101631083B (zh)
WO (1) WO2011015130A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8285910B2 (en) 2009-08-07 2012-10-09 Chengdu Huawei Symantec Technologies Co., Ltd. Method and system for taking over devices

Families Citing this family (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103384204B (zh) * 2011-12-31 2016-03-30 华为数字技术(成都)有限公司 串行并行转换电路故障的处理方法及装置
ES2687609T3 (es) * 2013-05-02 2018-10-26 Huawei Technologies Co., Ltd. Sistema informático, método para acceder a un terminal de interconexión de componentes periféricos exprés y equipo
CN104572508A (zh) * 2015-01-22 2015-04-29 山东超越数控电子有限公司 一种基于PCIE非透明桥的cache镜像方法
US10191882B2 (en) * 2015-06-29 2019-01-29 Futurewei Technologies, Inc. Method and system for aggregation-friendly address assignment to PCIe devices
TWI597953B (zh) * 2015-11-25 2017-09-01 財團法人工業技術研究院 具故障轉移能力的快速周邊元件互連網路系統與操作方法
US20230171099A1 (en) * 2021-11-27 2023-06-01 Oracle International Corporation Methods, systems, and computer readable media for sharing key identification and public certificate data for access token verification
CN115904520B (zh) * 2023-03-13 2023-06-06 珠海星云智联科技有限公司 基于pcie拓扑状态变更的配置保存方法及相关设备
CN117971504A (zh) * 2024-03-29 2024-05-03 井芯微电子技术(天津)有限公司 一种故障倒换Bus资源分配方法、系统及装置

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999059066A1 (en) * 1998-05-14 1999-11-18 Motorola, Inc. Controlling a bus with multiple system hosts
CN1373427A (zh) * 2001-03-01 2002-10-09 深圳市中兴通讯股份有限公司 一种实现双系统槽的装置和方法
TW200513844A (en) * 2003-10-09 2005-04-16 Mitac Int Corp Fault-tolerance RAID controller
CN101237315A (zh) * 2008-02-28 2008-08-06 浪潮电子信息产业股份有限公司 一种用于双控高可用系统的同步检测和故障隔离方法
CN101631083A (zh) * 2009-08-07 2010-01-20 成都市华为赛门铁克科技有限公司 设备接管方法和装置及双控系统

Family Cites Families (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP4810349B2 (ja) * 2006-08-11 2011-11-09 日本電気株式会社 I/o装置及び方法
TW201015336A (en) * 2008-10-03 2010-04-16 Accusys Technology Ltd Shared-storage bus switch

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO1999059066A1 (en) * 1998-05-14 1999-11-18 Motorola, Inc. Controlling a bus with multiple system hosts
CN1373427A (zh) * 2001-03-01 2002-10-09 深圳市中兴通讯股份有限公司 一种实现双系统槽的装置和方法
TW200513844A (en) * 2003-10-09 2005-04-16 Mitac Int Corp Fault-tolerance RAID controller
CN101237315A (zh) * 2008-02-28 2008-08-06 浪潮电子信息产业股份有限公司 一种用于双控高可用系统的同步检测和故障隔离方法
CN101631083A (zh) * 2009-08-07 2010-01-20 成都市华为赛门铁克科技有限公司 设备接管方法和装置及双控系统

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8285910B2 (en) 2009-08-07 2012-10-09 Chengdu Huawei Symantec Technologies Co., Ltd. Method and system for taking over devices

Also Published As

Publication number Publication date
CN101631083B (zh) 2012-04-04
CN101631083A (zh) 2010-01-20
US8285910B2 (en) 2012-10-09
US20120137042A1 (en) 2012-05-31

Similar Documents

Publication Publication Date Title
WO2011015130A1 (zh) 设备接管方法和装置及双控系统
JP5844373B2 (ja) ミドルウェアマシン環境におけるランナウェイサブネットマネージャインスタンスからの保護を促進するためのシステムおよび方法
US9350655B2 (en) Vertical converged framework
US9575919B2 (en) Information system capable of expanding drive and bus number allocation method of the information system
WO2015021629A1 (zh) 一种资源发放方法
CN110661894A (zh) 用户ip地址的分配方法、网络设备和存储介质
WO2015100973A1 (zh) 锁管理方法及系统、锁管理系统的配置方法及装置
US11140091B2 (en) Openflow protocol-based resource control method and system, and apparatus
WO2015158250A1 (zh) 一种资源分配方法、报文通信方法及装置
WO2014201623A1 (zh) 用于数据传输的方法、装置和系统以及物理网卡
EP3820083B1 (en) Physical function multiplexing method and apparatus and computer storage medium
CN104753707A (zh) 一种系统维护方法及网络交换设备
CN103001953A (zh) 虚拟机网络资源分配方法及装置
WO2009059476A1 (fr) Procédé pour distribuer le numéro de dispositif dans un système d'empilage
US20130246729A1 (en) Method for Managing a Memory of a Computer System, Memory Management Unit and Computer System
WO2015067051A1 (zh) 测试代理方法及其装置
US8873427B2 (en) Method for establishing topology structure of switching equipment, switching equipment and stacking system
CN110753093A (zh) 云计算系统中设备的管理方法和装置
CN111600795B (zh) 虚拟边缘设备的建立方法、控制器和服务器
CN111211927B (zh) 一种资源同步方法及装置
WO2009132555A1 (zh) 实现视讯会议的方法及设备
EP2939382B1 (en) Distributed data processing system
WO2023138390A1 (zh) 一种时隙分配的方法、网络设备和系统
CN101778034A (zh) 一种多数字信号处理器的加载方法和加载控制装置
JP2008097273A (ja) ネットワークインタフェース装置、ネットワークインタフェース制御方法、情報処理装置、データ転送方法

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 10806034

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205N DATED 16/04/2012)

122 Ep: pct application non-entry in european phase

Ref document number: 10806034

Country of ref document: EP

Kind code of ref document: A1