WO2018133414A1 - 切包方法、请求处理方法、装置和计算机存储介质 - Google Patents

切包方法、请求处理方法、装置和计算机存储介质 Download PDF

Info

Publication number
WO2018133414A1
WO2018133414A1 PCT/CN2017/099344 CN2017099344W WO2018133414A1 WO 2018133414 A1 WO2018133414 A1 WO 2018133414A1 CN 2017099344 W CN2017099344 W CN 2017099344W WO 2018133414 A1 WO2018133414 A1 WO 2018133414A1
Authority
WO
WIPO (PCT)
Prior art keywords
sub
request
transmission request
source
packet
Prior art date
Application number
PCT/CN2017/099344
Other languages
English (en)
French (fr)
Inventor
王强
刘新良
李晓莎
刘万江
陈艳
Original Assignee
深圳市中兴微电子技术有限公司
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 深圳市中兴微电子技术有限公司 filed Critical 深圳市中兴微电子技术有限公司
Publication of WO2018133414A1 publication Critical patent/WO2018133414A1/zh

Links

Images

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L65/00Network arrangements, protocols or services for supporting real-time applications in data packet communication
    • H04L65/40Support for services or applications
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/01Protocols
    • H04L67/10Protocols in which an application is distributed across nodes in the network
    • H04L67/104Peer-to-peer [P2P] networks
    • H04L67/1074Peer-to-peer [P2P] networks for supporting data block transmission mechanisms
    • H04L67/1078Resource delivery mechanisms
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L67/00Network arrangements or protocols for supporting network services or applications
    • H04L67/50Network services
    • H04L67/60Scheduling or organising the servicing of application requests, e.g. requests for application data transmissions using the analysis and optimisation of the required network resources

Definitions

  • the present invention relates to a baseband chip architecture technology in the field of wireless communication networks, and in particular, to a packet cutting method, a request processing method and a device.
  • LTE Long Term Evolution
  • UMTS Universal Mobile Telecommunications System
  • GSM Global System for Mobile Communication
  • the chip embeds a baseband-specific accelerator for a variety of system processing and numerous CPU/digital signal processors.
  • the dedicated accelerator and the embedded software on the central processing unit work together to complete all the baseband functions of the multiple hybrid systems.
  • the whole chip is divided into clock reset management, P subsystem and U subsystem according to functions.
  • the P subsystem includes an interconnection matrix and an accelerator closely associated with the digital signal processor.
  • the U subsystem includes a CCN closely associated with the central processing unit. (Cache Coherent Network) and accelerator, its architecture is shown in Figure 1.
  • CCN is a ring topology, the external interconnection relationship is shown in Figure 2, four sets of coherent bus interfaces are respectively connected to a cluster of central processors, each cluster contains four processors, AXI_SLV is the memory of CCN chip, low-speed peripherals Interface; AXI_MST is the accelerator of the chip, DMA (Direct Memory Access), high-speed / medium-speed peripherals, etc., and accesses the CCN through the RNI (IO coherent Request Node) interface.
  • AXI_SLV is the memory of CCN chip, low-speed peripherals Interface
  • AXI_MST is the accelerator of the chip
  • DMA Direct Memory Access
  • high-speed / medium-speed peripherals etc.
  • RNI IO coherent Request Node
  • RNI will only send a read or write request of the same AxID (Advanced aXtensible Interface, AXI standard agreement in axid and arid) after the response of the previous request is returned;
  • AxID Advanced aXtensible Interface
  • the RNI automatically splits the request into multiple 64-byte boundary-aligned requests with the same AxID and executes them one by one.
  • the AXID value of the request sent by the CCN chip accelerator, DMA, high-speed/medium-speed peripheral, etc. is mostly fixed or changed within a small range, and the data carried by one request is mostly in the case. It will be larger than 64 bytes and smaller than 64 bytes but spans 64 byte boundaries. In this scenario, the CCN will align the request by 64-byte boundary alignment and then process it in order. Therefore, the processing efficiency is low and the bandwidth utilization rate is utilized. small.
  • an embodiment of the present invention provides a packet cutting method, a data request method, a device, and a computer storage medium, which can improve processing efficiency and improve bandwidth utilization.
  • an embodiment of the present invention provides a packet cutting method, where the method includes:
  • the source transmission request is split into multiple sub-transmission requests
  • a response to the source transmission request is returned to the upstream device based on an order relationship between the plurality of sub-transmission requests.
  • an embodiment of the present invention further provides a packet cutting device, where the packet cutting device includes:
  • the first determining module is configured to obtain a source transmission request sent by the upstream device, and determine whether the source transmission request needs to be packet-cutted;
  • the packet cutting module is configured to split the source transmission request into multiple sub-transmission requests when the first judging module determines that the source transmission request needs to be packet-cut processing;
  • a second determining module configured to determine whether it is necessary to perform ID conversion on the multiple sub-transmission requests
  • the ID conversion module is configured to perform ID conversion on the plurality of sub-transmission requests if the second judging module determines that ID conversion is required;
  • a sequence ordering module configured to record an order relationship between the plurality of sub-transmission requests; acquire a response of the downstream device for the plurality of sub-transmission requests; and to the upstream device based on an order relationship between the plurality of sub-transmission requests Returns a response to the source transfer request.
  • an embodiment of the present invention provides a data processing method, including:
  • the source transmission request is split into multiple sub-transmission requests
  • the triggering target device responds to the sub-transmission request in parallel when the identifiers of the sub-transmission requests are different.
  • an embodiment of the present invention further provides a computer storage medium, where the computer storage medium stores computer executable instructions, where the computer executable instructions are used to execute the foregoing packet cutting method or the foregoing request processing method.
  • acquiring a source transmission request sent by an upstream device determining Whether the source transmission request needs to be packet-cutted; when it is determined that the source transmission request needs to be packet-cut, the source transmission request is split into multiple sub-transmission requests;
  • the sub-transmission request performs ID conversion. If ID conversion is required, ID conversion is performed on the plurality of sub-transmission requests, and an order relationship between the plurality of sub-transmission requests is recorded; and the downstream device is requested to obtain the plurality of sub-transmission requests. Responding to returning a response to the source transmission request to the upstream device based on an order relationship between the plurality of sub-transmission requests.
  • ID conversion is performed on the multiple sub-transmission requests, which can improve processing efficiency and improve bandwidth utilization.
  • 1 is a block diagram of a system structure of a soft baseband chip
  • FIG. 2 is a schematic diagram of a connection relationship between an accelerator, a peripheral, and the like and a CCN;
  • FIG. 3 is a schematic flowchart diagram of a packet cutting method according to an embodiment of the present invention.
  • FIG. 4 is a schematic diagram of a connection relationship between a packet cutting device, an upstream device, and a downstream device according to an embodiment of the present invention
  • FIG. 5 is a block diagram showing an implementation principle of a packet cutting device according to an embodiment of the present invention.
  • FIG. 6 is a schematic structural diagram of a packet cutting device according to an embodiment of the present invention.
  • FIG. 7 is a schematic flowchart of processing after a packet cutting device receives a source transmission request sent by an upstream device according to an embodiment of the present invention.
  • FIG. 3 is a schematic flowchart of a packet cutting method according to an embodiment of the present invention, which is applied to a packet In the device, as shown in FIG. 3, the packet cutting method includes the following steps:
  • Step 301 Acquire a source transmission request sent by the upstream device.
  • the source transmission request includes but is not limited to:
  • the packet cutting device is located between the upstream device and the downstream device, and when the upstream device sends a source transmission request to the downstream device, the packet cutting device acquires the source transmission request, and The source transmission request performs an analysis and determination to determine whether the source transmission request needs to be packetized.
  • the downstream device refers to a CCN device.
  • the upstream device refers to an AXI_MST connected to the CCN device
  • the downstream device may include a network accelerator, a DMA, a high speed/intermediate speed peripheral, and the like.
  • FIG. 4 a schematic diagram of a connection relationship between the packet cutting device and an upstream device and a downstream device is shown in FIG. 4.
  • an input end of the packet cutting device is connected to an AXI_MST, and an output end of the packet cutting device is The RNI interface in the CCN device is connected.
  • Step 302 Determine whether it is necessary to perform packet cutting processing on the source transmission request.
  • the determining whether the source transmission request needs to be packet-cutted includes:
  • the static bypass is in an active state, it is determined that the source transmission request does not need to be packetized, and the source transmission request is directly sent to the downstream device.
  • Step 303 When it is determined that the source transmission request needs to be packet-cut, the source transmission request is split into multiple sub-transmission requests.
  • the splitting the source transmission request into multiple sub-transmission requests includes:
  • the source transmission request is packet-cutted in combination with the first address, size, and length of the source transmission request, and split into multiple sub-transmission requests.
  • the source transmission request is packet-cut, and the IDs of the split multiple sub-transmission requests are the same as the ID of the source transmission request.
  • the source transmission request is split into three sub-transmission requests, and then the IDs of the three sub-transmission requests are all A.
  • the first source address, the size, and the length of the source transmission request are combined, and the source transmission request is packet-cut and split into multiple sub-transmission requests, including:
  • each sub-transport before the last sub-transmission is divided into 64 bytes, and the first sub-transport is started from the first address to the first 64-byte aligned address.
  • the i-th 64-byte aligned address is the i+1th sub-transport before the i+1th byte-aligned address, where i is a positive integer greater than or equal to 1, and so on, until the last sub-transmission
  • the starting address is the i+1th 64-byte aligned address
  • the last sub-transport is the i+1th 64-byte aligned address to the end of the transmission.
  • the first sub-transport is performed from the first address to the first 64-byte aligned address
  • the first 64-byte aligned address is The 2 byte alignment address precedes the second subtransmission
  • the 2nd 64 byte alignment address to the 3rd byte alignment address is the third subtransmission
  • the fourth subtransmission is the 3rd 64 byte alignment address to The location at which the transfer ends.
  • the first source address, the size, and the length of the request are transmitted, and the source transmission request is packetized and split into multiple sub-transmission requests, including:
  • the first subtransmission is from the first address to the first 64-byte aligned address, from the first 64 words.
  • the section aligns the address to the end of the transfer for the second subtransmission.
  • Step 304 Determine whether it is necessary to perform ID conversion on the plurality of sub-transmission requests. If ID conversion is required, perform ID conversion on the plurality of sub-transmission requests, and record an order relationship between the plurality of sub-transmission requests.
  • the purpose of performing ID conversion on the plurality of sub-transmission requests is to make the IDs of the respective sub-transmission requests corresponding to the same source transmission request different, so that the RNIs in the CCN device can make the respective ones belonging to the same source transmission request Different sub-transmission requests can be sent to the downstream processing module of the CCN device in parallel, thereby improving bandwidth utilization and helping to improve the processing efficiency of the source transmission request.
  • the determining whether it is necessary to perform ID conversion on the multiple sub-transmission requests includes:
  • the determining whether the ID conversion is needed according to the dynamic bypass address segment and the enable bit in the configuration register includes:
  • the performing ID conversion on the multiple sub-transmission requests includes:
  • the smallest idle ID value is selected from the resource IDs as the ID of the subtransmission request.
  • the smallest idle ID value is selected as the ID of the subtransmission request in order to facilitate more orderly determination of the ID value of each subtransmission request.
  • any idle ID value may be selected from the resource IDs as the ID of the sub-transmission request.
  • the method further includes:
  • the existing processing manner includes: transmitting the multiple sub-transmission requests to the downstream device in sequence, and delivering one at a time, and receiving, by the downstream device, one of the multiple sub-transmission requests. After the requested feedback, another sub-transmission request of the same ID in the plurality of sub-transmission requests is delivered.
  • Step 305 Acquire a response of the downstream device for the multiple sub-transmission requests.
  • the method before the obtaining the response of the downstream device to the multiple sub-transmission requests, the method further includes:
  • Step 306 Return a response to the source transmission request to the upstream device based on an order relationship between the multiple sub-transmission requests.
  • the returning, according to the sequence relationship between the multiple sub-transmission requests, a response to the source transmission request to the upstream device includes:
  • the source transmission request is a read request
  • the downstream device when receiving a response for each sub-transmission request fed back by the downstream device, returning to the upstream device according to the order relationship between the respective sub-transmission requests of the same source transmission request The read response of each subtransmission request of the source transfer request;
  • the source transmission request is a write request
  • the responses to all sub-transmission requests for the same source transmission request are merged, and after the merge, the The upstream device returns a write response to the same source transfer request.
  • the packet-cutting method provided by the embodiment of the present invention performs ID conversion on the multiple sub-transmission requests when ID conversion is required, and the corresponding IDs of the sub-transmission requests corresponding to the same source transmission request are different after the conversion, in the CCN device.
  • the RNI can deliver each sub-transmission request corresponding to the same source transmission request with different IDs at one time, thereby improving processing efficiency and increasing bandwidth. Utilization rate.
  • this embodiment provides a block diagram of an implementation principle of the packet cutting device.
  • the packet cutting device is mainly divided into a packet cutting process, an ID conversion process, and a guard sequence. Handling three major parts, the functions of each part are as follows:
  • the packet processing part performs packet cutting processing on the source transmission request in combination with the received first address, size, and length of the AXI bus, and splits into multiple sub-transmissions.
  • ID conversion processing part compares the packetized sub-transmission request with the dynamic bypass configuration window and the enable bit to determine whether ID conversion is required, and if the match is valid, crosses the ID conversion module and sets the sixth bit of the AxID The position is 1, and the lower five bits remain unchanged, indicating that the transmission has not been converted to the AxID value and is directly output to the order-preserving module; the transmission of the dynamic bypass configuration window and the enable bit fails, and the sixth bit position of the AxID is zero. After regenerating the lower five digits, output to the order-preserving module.
  • the order-preserving processing part performs the order-preserving processing on the received transmission. Since the AxID of each sub-transmission is different after the same source transmission and cutting, the downstream equipment processes the sub-transmission in parallel, so the response order of each sub-transmission may be returned. Out of order. The response from the downstream device is staged and returned to the upstream device in the same order as sent to the downstream device.
  • the bandwidth is 5GB/s and 20GN/s, respectively, and the difference is up to 4 times.
  • the embodiment of the present invention provides a packet cutting device.
  • the structure of the packet cutting device is as shown in FIG. 6.
  • the packet cutting device includes: a first determining module 61, a packet cutting module 62, and a second determining module 63.
  • the first determining module 61 is configured to acquire a source transmission request sent by the upstream device, and determine whether the packet transmission process needs to be performed on the source transmission request.
  • the packet cutting module 62 is configured to, when the first determining module 61 determines that the source transmission request needs to be packet-cut, process the source transmission request into multiple sub-transmission requests;
  • the second determining module 63 is configured to determine whether it is necessary to perform ID conversion on the multiple sub-transmission requests.
  • the ID conversion module 64 is configured to perform ID conversion on the plurality of sub-transmission requests if the second judging module 63 determines that ID conversion is required;
  • the order-preserving module 65 is configured to record an order relationship between the plurality of sub-transmission requests; acquire a response of the downstream device to the plurality of sub-transmission requests; and based on an order relationship between the plurality of sub-transmission requests, The upstream device returns a response to the source transmission request.
  • the first determining module 61 is further configured to:
  • the static bypass is in an active state, it is determined that the source transmission request does not need to be packetized, and the source transmission request is directly sent to the downstream device.
  • the packet cutting module 62 is further configured to:
  • the source transmission request is packet-cutted in combination with the first address, size, and length of the source transmission request, and split into multiple sub-transmission requests.
  • the packet cutting module 62 is configured to:
  • each sub-transport before the last sub-transmission is divided into 64 bytes, starting from the first address to the first 64-byte aligned address.
  • the i-th 64-byte aligned address is the i+1th sub-transmission before the i+1th byte-aligned address, where i is a positive integer greater than or equal to 1, and so on.
  • the start address of the last subtransmission is the i+1th 64 byte aligned address
  • the last subtransmission is the i+1th 64 byte aligned address to the end of the transmission;
  • the first subtransmission is from the first address to the first 64-byte aligned address, from the first 64 words.
  • the section aligns the address to the end of the transfer for the second subtransmission.
  • the second determining module 63 is further configured to:
  • the second determining module 63 is configured to:
  • the ID conversion module 64 is configured to:
  • the smallest idle ID value is selected from the resource IDs as the ID of the subtransmission request.
  • the order-preserving module 65 is further configured to:
  • the source transmission request is a read request
  • the downstream device when receiving a response for each sub-transmission request fed back by the downstream device, returning to the upstream device according to the order relationship between the respective sub-transmission requests of the same source transmission request The read response of each subtransmission request of the source transfer request;
  • the source transmission request is a write request
  • the responses to all sub-transmission requests for the same source transmission request are merged, and after the merge, the The upstream device returns a write response to the same source transfer request.
  • the first determining module 61, the packet cutting module 62, the second determining module 63, the ID converting module 64, and the sorting module 65 in the packet cutting device may be used in the packet cutting device in practical applications. Realized by a central processing unit (CPU), a digital signal processor (DSP), or a field-programmable gate array (FPGA).
  • CPU central processing unit
  • DSP digital signal processor
  • FPGA field-programmable gate array
  • the packet cutting device can be disposed between the upstream device and the downstream device.
  • the downstream device refers to a CCN device.
  • the upstream device refers to the AXI_MST connected to the CCN device, including a network accelerator, a DMA, a high speed/intermediate speed peripheral, and the like.
  • the input end of the packet cutting device is connected to the AXI_MST, and the output end of the packet cutting device is connected to the RNI interface in the CCN device.
  • the packet cutting device of the embodiment of the invention can improve processing efficiency and improve bandwidth utilization.
  • this embodiment provides a processing flow after the packet-cutting device receives the source transmission request sent by the upstream device, as shown in FIG. 7, the specific processing steps are as follows:
  • Step 701 The packet cutting device determines the state of the static bypass, if the static bypass is in the active state, the source transmission request is sent to the RNI in the downstream device; if the static bypass is in the invalid state, step 702 is performed;
  • the first determining module in the packet cutting device determines whether the received source transmission request needs to be packetized by determining the state of the static bypass. If the static bypass is in the active state, the source transmission request from the upstream device will directly communicate with the downstream device across the subsequent processing flow of the packet cutting device.
  • Step 702 The packet cutting device performs packet cutting on the received source transmission request, and then performs step 703;
  • the source transmission request is packetized by the packet cutting module in the packet cutting device.
  • the first sub-address is used from the first address to the first 64-byte aligned address.
  • the transfer, the subsequent 64-byte aligned address to the next 64-byte aligned address position is split by 64 bytes, and so on, and the last sub-transfer is a 64-byte aligned address to the end of the transfer.
  • Step 703 The packet cutting device determines whether the ID conversion is required according to the dynamic bypass, if necessary, step 704 is performed; if not, step 705 is performed;
  • the second determining module in the packet cutting device determines whether the ID conversion is needed according to the dynamic bypass address segment and the enable bit in the configuration register.
  • all packet devices support up to 32 configurable bypass address segments, and the address segment granularity is 2 megabytes; each address segment supports reading and writing dynamic bypass enable bits to read and write channels for ID conversion.
  • Control wherein the control information of the 32 address segments is composed of dynbp_addr_begin, dynbp_addr_end, and dynbp_addr_bypass, that is, the start segment address, the end segment address, and the enable signal.
  • dynbp_addr_begin and dynbp_addr_end are 14-bit signals, corresponding to the upper 14 bits of the 35-bit AXI address; dynbp_addr_bypass is a 2-bit signal, and the high and low bits respectively control the dynamic bypass enable of the address segment write and read channels.
  • the source transmission request in the corresponding address segment will not perform ID conversion.
  • the packet cutting device can be flexibly applied to different scenarios, and the two can coexist. For example, for data space, packet cutting and ID conversion processing are usually required to improve system performance. For some configuration spaces, the order of configuration is strictly required. At this time, the corresponding address segment is configured to the dynamic bypass window to make it strict. Execute in the order of the original request.
  • Step 704 The packet cutting device performs ID conversion, and then step 705 is performed;
  • ID conversion is performed by an ID conversion module in the packet cutting device.
  • all packet devices support up to 32 read requests and 32 write requests.
  • the bypass judgment if the subtransmission needs to perform ID conversion, the smallest idle ID value will be selected from the resource ID as the AxID of the subtransmission, and then the The sub-transmission is sent to the order-preserving module in the packet-cutting device.
  • Step 705 The packet cutting device performs the order-preserving process, and then the sub-transmission after the order-preserving process is requested. Request to send to the RNI in the downstream device.
  • the order-preserving process is performed by the order-preserving module in the packet cutting device.
  • the response and data requirements for the return are processed in accordance with the source AxID information before the ID conversion.
  • the outstanding includes a read request and a write request.
  • each read request sent to the downstream device will occupy a different AxID value.
  • the downstream device receives the read data and temporarily stores it, receives the complete read data and returns the data to the upstream device. Release the AxID. Since the downstream device receives sub-transmissions of different AxIDs, the read data returned to the packet-cutting device will be out of order.
  • the sequence-preserving module in the packet-cutting device performs the order-preserving processing on the out-of-order read data, and returns the response to the upstream device in the same order sent to the downstream device, and the order is saved.
  • the method is mainly implemented by transmitting the AxID group chain of the corresponding sub-transport to the same source.
  • the AxID of a source transmission is 1, the AxIDs when the packets are divided into 3 sub-transmissions and sent to the downstream device sequentially are 4, 5, and 6, respectively.
  • the three AxID values will be combined into one chain, id4->id5->id6, and there is no other AxID in front of id4, indicating that the read data received by id4 can be returned to the upstream device according to the AxID transmitted by the source, etc.
  • id4 After all the data is returned to the upstream device, the linked list is updated to id5->id6, and so on, the data of id5 and id6 is returned to the upstream device according to the AxID transmitted by the source.
  • the write channel ID conversion and the order-preserving processing are basically the same as the read channel. The difference is that the write response of several sub-transmissions transmitted by the same source will be combined into a response and then returned to the upstream. device.
  • the processing steps are similar to source transfer requests greater than 64 bytes, with the difference being: step 602 packet shredding processing and step 605 order processing.
  • the step 602 packet processing includes: for a source transmission that is less than 64 bytes but spans a 64-byte boundary, the first sub-transport after the packet is changed from the first address to the first 64-byte aligned address.
  • the first subtransmission, the second subtransmission is a 64-byte aligned address to the end of the transmission, a total of only two sub-transmissions.
  • the AxID of the source transmission is 7, and it is only split into 2 sub-transmissions.
  • the two AxID values will be combined into one chain, id8->id9, id8
  • There is no other AxID in the front indicating that the read data received by id8 can be returned to the upstream device according to the AxID transmitted by the source.
  • the linked list is updated to id9, and the data of id9 is returned according to the AxID transmitted by the source. Give the upstream device.
  • the packet cutting device receives the response for each sub-transmission request fed back by the downstream device, based on the order relationship between the plurality of sub-transmission requests recorded during the sequence-preserving process in step 705,
  • the upstream device returns a response to the source transmission request.
  • the source transmission request is a read request
  • the source transmission request when receiving a response for each sub-transport request fed back by the downstream device, returning to the upstream device according to an order relationship between the sub-transmission requests of the same source transmission request a read response for each of the sub-transmission requests of the same source transmission request
  • the source transmission request is a write request
  • the source transmission request when receiving a response for each sub-transmission request fed back by the downstream device, transmitting a request for the same source
  • the responses of all subtransmission requests are merged and a write response to the same source transmission request is returned to the upstream device upon merging.
  • the packet cutting device feeds back the response of the downstream device to the upstream device, only some modules in the packet cutting device pass through, and no processing is performed.
  • the first determining module in step 701 and the step 702 in step 702 The packet cutting module and the second judging module in step 703 do not need to perform any processing.
  • the packet cutting process provided by the embodiment of the present invention can improve processing efficiency and improve bandwidth utilization.
  • An embodiment of the present invention provides a request processing method, where the method includes:
  • the source transmission request is split into multiple sub-transmission requests
  • the triggering target device responds to the sub-transmission request in parallel when the identifiers of the sub-transmission requests are different.
  • the method can be applied to a request splitting device, where the request splitting device can correspond to the front
  • the unpacking device may be an intermediate device that connects an upstream device and a downstream device, the upstream device is a device that sends a source transmission request, and the downstream device may be a device that responds to the source transmission request.
  • the specific upstream device may be AXI_MST; the downstream device may be a CCN device.
  • the split request may be a split component applied to the downstream device, or may be a split component applied to the upstream device.
  • different IDs are set for each sub-transmission request, so that when the downstream device performs data transmission, it does not share an ID because multiple sub-transmission requests, and only the transmission rate caused by serial transmission is possible. Slow question.
  • the IDs of the plurality of sub-transmission requests corresponding to one source transmission request are different, and the sub-transmission requests can be responded in parallel, which obviously greatly improves the transmission rate.
  • the method further includes:
  • Each of the sub-transmission requests is assigned a different identification number, including:
  • each of the sub-transmission requests is assigned a different identification number.
  • the upstream device explicitly needs to receive the requested data in order, and does not run the parallel response. If the parallel response may cause the upstream device to fail to receive the packets in order.
  • the sub-transmission request is assigned a different identification number, and only the parallel response is allowed.
  • the same identification number is assigned to each of the sub-transmission allocation requests and the response order of the sub-transmission request is recorded. In this way, the transmitting device that responds to the sub-transmission request returns data to the device that sent the source transmission request according to the response sequence.
  • the source transmission request is split into three sub-transmission requests when the parallel response is not allowed. Then, the IDs of the three sub-transmission requests may all be A.
  • it is required to record the correspondence between the ID of the subtransmission request and the source transmission request ID.
  • the ID of the source transmission request is recorded as A
  • the source transmission request is split into three sub-transmission requests, and then the ID of the three sub-transmission requests is A1.
  • A1, A2, and A3 are different from each other.
  • it is required to record the correspondence between the ID of the subtransmission request and the source transmission request ID.
  • whether the parallel response is allowed may be determined by the enabled state of the second identity.
  • the second identifier here may correspond to the identifier corresponding to the foregoing dynamic bypass address segment; the enabled state may correspond to an enabled state of the dynamic bypass address segment in the foregoing embodiment.
  • the determining whether the source transmission request needs to be split is performed, including:
  • the preset threshold here can be the set value.
  • the data transmission request request packet is larger than the preset byte, and the transmission request is split according to the preset byte to form a plurality of sub-transmission requests each of which is not greater than the preset byte.
  • the preset threshold in this embodiment may be 64 bytes or a predetermined number of bytes such as 32 bytes or 128 bytes.
  • the determining whether the source transmission request needs to be split is performed, including:
  • the target data spans at least two storage units, it is determined that the source transmission request needs to be split. That is, the target data of the source transmission request is stored in a plurality of storage units, and then splitting is required.
  • the storage capacity corresponding to the storage unit is equal to the preset threshold in the foregoing embodiment.
  • the source transmission request is split into multiple sub-transmission requests, including at least one of the following:
  • the request is required according to the source transmission request Separating the source data request by the length L of the target data and the transmission data bit width W, wherein the number of the sub-transport requests is equal to a minimum integer not less than the L/W;
  • the source transmission request is split into m sub-transmission requests.
  • the transmission data bit width is not greater than the bit width of the bus between the transmitting end of the transmission request and the receiving end of the data.
  • the method further includes:
  • Determining whether the source transmission request needs to be split including:
  • the method further includes:
  • the source transmission request is transparently transmitted.
  • the first identifier here may be the aforementioned static bypass identifier.
  • the maximum length of the data packets corresponding to different service types is different. Therefore, the state of the first identifier can be configured according to the length of the data packet corresponding to the service type. If the first identifier is directly set to the second state, the source transmission request is transparently transmitted, and the operation of determining whether the source transmission request needs to be split and splitting is performed, thereby improving transmission efficiency without splitting. .
  • the first identifier may be a 1-bit state value. If the value of the bit is "0", corresponding to the first state, when the value of the bit is "1", the corresponding state is the second state. If the value of the bit is "1”, corresponding to the first state, when the value of the bit is "0", the corresponding state is the second state.
  • An embodiment of the present invention provides a computer storage medium having stored therein computer executable instructions for performing a packet cutting method or one or more implementations provided by one or more embodiments.
  • the request processing method provided by the example.
  • the computer storage medium may be various types of storage media such as a mobile hard disk, an optical disk, a magnetic tape, and the like, and may be a non-transitory storage medium.
  • the embodiment of the present invention further provides a device, including: a memory and a processor; the processor is connected to the memory, and can implement the cutting provided by one or more of the foregoing technical solutions by executing computer executable instructions such as a computer program. Package method and/or request processing method.
  • the disclosed method and smart device may be implemented in other manners.
  • the device embodiments described above are merely illustrative.
  • the division of the unit is only a logical function division.
  • there may be another division manner such as: multiple units or components may be combined, or Can be integrated into another system, or some features can be ignored or not executed.
  • the coupling, or direct coupling, or communication connection of the components shown or discussed may be indirect coupling or communication connection through some interfaces, devices or units, and may be electrical, mechanical or other forms. of.
  • the units described above as separate components may or may not be physically separated, and the components displayed as the unit may or may not be physical units, that is, may be located in one place or distributed to multiple network units; Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of the embodiment.
  • each functional unit in each embodiment of the present invention may be integrated into one second processing unit, or each unit may be separately used as one unit, or two or more units may be integrated into one unit;
  • the above integrated unit can be implemented in the form of hardware or in the form of hardware plus software functional units.
  • the splitting of the source transmission request and the allocation of the different identification numbers of the sub-transmission request, so that the transmitting device responding to the transmission request may be different based on different
  • the sub-transmission of the identification number requests a parallel response of the data transmission, thereby increasing the rate of data transmission, thereby having a beneficial industrial effect, being simple to implement, and being industrially reproducible.

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Multimedia (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明公开了一种切包方法、请求处理方法及装置,所述方法包括:获取上游设备发送的源传输请求;判断是否需要对所述源传输请求进行切包处理;当判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆成多个子传输请求;判断是否需要对所述多个子传输请求进行标识号(ID)转换,若需要进行ID转换,则对所述多个子传输请求进行ID转换,并记录所述多个子传输请求之间的顺序关系;获取下游设备针对所述多个子传输请求的响应;基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。本发明实施例还提供一种计算机存储介质。

Description

切包方法、请求处理方法、装置和计算机存储介质
本申请基于申请号为201710051983.5、申请日为2017年01月20日的中国专利申请提出,并要求该中国专利申请的优先权,该中国专利申请的全部内容在此引入本申请作为参考。
技术领域
本发明涉及无线通讯网络领域中基带芯片架构技术,尤其涉及一种切包方法、请求处理方法及装置。
背景技术
目前各大运营商都存在多种制式并存的问题,并将长期存在,运营商普遍采用基站机柜插不同制式的基带板卡的方式来实现一体化基站,造成基站机房建设支出巨大,运维成本非常高。运营商希望通讯设备商提供的基站能够支持2G/3G/4G等网络的平滑升级,机柜小型化,绿色低功耗,以减少设备投资,降低运维成本。
为应对这一需求,各主流通讯设备商启动了支持LTE(3GPP Long Term Evolution,长期演进技术)、UMTS(Universal Mobile Telecommunications System,环球移动通信系统)、GSM(Global System for Mobile Communication,全球移动通信系统)等多种制式共存的单芯片软基带芯片研发。芯片内嵌适合多种制式处理的基带专用加速器和众多中央处理器/数字信号处理器。专用加速器和中央处理器上的内嵌软件协同分工,共同完成多种混合制式上下行的所有基带功能。整个芯片根据功能分为时钟复位管理、P子系统和U子系统,其中P子系统包括与数字信号处理器有紧密联系的互联矩阵及加速器,U子系统包括与中央处理器有紧密联系的CCN(Cache Coherent Network,一致性高速缓存网络)及加速器,其架构如图1所示。
其中,CCN为环状拓扑结构,对外互联关系如图2所示,四组一致性总线接口分别连接一簇中央处理器,每簇包含四个处理器,AXI_SLV为CCN芯片的存储器、低速外设接口;AXI_MST为本芯片的加速器、DMA(Direct Memory Access,直接存储器存取设备)、高速/中速外设等,通过RNI(IO coherent Request Node,输入输出一致性请求节点)接口接入CCN。根据CCN的资料描述,由于RNI只有在前一个请求的响应返回后才会发下一个相同AxID(Advanced eXtensible Interface,AXI标准协议中的awid和arid的总称)的读或写请求;且当收到的请求大于64字节或者小于64字节但跨64字节边界时,RNI会自动将请求拆分为多个AxID相同的64字节边界对齐的请求,并按顺序逐一执行。
实际应用中,CCN芯片的加速器、DMA、高速/中速外设等端口发出的请求中的AxID值多数情况下是固定的或者在很小的范围内变化,而且一个请求携带的数据多数情况下会大于64字节和小于64字节但是跨64字节边界,此种场景下CCN将对该请求按64字节边界对齐进行切包,然后按顺序处理,因此,处理效率低,带宽利用率小。
发明内容
为解决上述技术问题,本发明实施例提供了一种切包方法、数据请求方法、装置及计算机存储介质,能提高处理效率,提升带宽利用率。
第一方面,本发明实施例提供了一种切包方法,所述方法包括:
获取上游设备发送的源传输请求;
判断是否需要对所述源传输请求进行切包处理;
当判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆分成多个子传输请求;
判断是否需要对所述多个子传输请求进行标识号(ID,Identity)转换,若需要进行ID转换,则对所述多个子传输请求进行ID转换,并记录所述多个子传输请求之间的顺序关系;
获取下游设备针对所述多个子传输请求的响应;
基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
第二方面,本发明实施例还提供了一种切包装置,所述切包装置包括:
第一判断模块,配置为获取上游设备发送的源传输请求,判断是否需要对所述源传输请求进行切包处理;
切包模块,配置为当所述第一判断模块判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆分成多个子传输请求;
第二判断模块,配置为判断是否需要对所述多个子传输请求进行ID转换;
ID转换模块,配置为若所述第二判断模块判定需要进行ID转换,则对所述多个子传输请求进行ID转换;
保序模块,配置为记录所述多个子传输请求之间的顺序关系;获取下游设备针对所述多个子传输请求的响应;基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
第三方面,本发明实施例提供一种数据处理方法,包括:
获取源传输请求;
确定是否需要对所述源传输请求进行拆分处理;
当判断出需要对所述源传输请求进行拆分处理时,将所述源传输请求拆分成多个子传输请求;
为每一个所述子传输请求分配不同的标识号;
其中,当所述子传输请求的标识号不同时,用于触发目标设备并行响应所述子传输请求。
第四方面,本发明实施例还提供一种计算机存储介质,所述计算机存储介质中存储有计算机可执行指令,所述计算机可执行指令用于执行前述的切包方法或前述的请求处理方法。
本发明实施例的技术方案中,获取上游设备发送的源传输请求;判断 是否需要对所述源传输请求进行切包处理;当判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆分成多个子传输请求;判断是否需要对所述多个子传输请求进行ID转换,若需要进行ID转换,则对所述多个子传输请求进行ID转换,并记录所述多个子传输请求之间的顺序关系;获取下游设备针对所述多个子传输请求的响应;基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。通过本发明实施例所述技术方案,在需要进行ID转换时,对所述多个子传输请求进行ID转换,能够提高处理效率,提升带宽利用率。
附图说明
图1为软基带芯片的系统结构框图;
图2为加速器、外设等与CCN的一种连接关系示意图;
图3为本发明实施例提供的一种切包方法的流程示意图;
图4为本发明实施例提供的切包装置与上游设备、下游设备连接关系示意图;
图5为本发明实施例的切包装置的一种实现原理框图;
图6为本发明实施例的切包装置的一种结构组成示意图;
图7为本发明实施例的切包装置接收到上游设备发送的源传输请求之后的处理流程示意图。
具体实施方式
为了能够更加详尽地了解本发明实施例的特点与技术内容,下面结合附图对本发明实施例的实现进行详细阐述,所附附图仅供参考说明之用,并非用来限定本发明实施例。
实施例一
图3为本发明实施例提供的一种切包方法的流程示意图,应用于切包 装置中,如图3所示,所述切包方法包括以下步骤:
步骤301:获取上游设备发送的源传输请求。
这里,所述源传输请求包括但不限于:
读请求、写请求。
本实施例中,所述切包装置位于上游设备与下游设备之间,当所述上游设备向所述下游设备发送源传输请求时,所述切包装置获取所述源传输请求,并对所述源传输请求进行分析判断,以便于确定是否需要对所述源传输请求进行切包处理。
这里,所述下游设备是指CCN设备。
这里,所述上游设备是指与CCN设备连接的AXI_MST,该下游设备可包括网络加速器、DMA、高速/中速外设等。
本实施例中,所述切包装置与上游设备、下游设备连接关系示意图如图4所示,可选地,所述切包装置的输入端与AXI_MST连接,所述切包装置的输出端与所述CCN设备中的RNI接口连接。
步骤302:判断是否需要对所述源传输请求进行切包处理。
作为一种可选实施方式,所述判断是否需要对所述源传输请求进行切包处理,包括:
检测静态bypass的状态;
若所述静态bypass处于无效状态,则判定需要对所述源传输请求进行切包处理;
若所述静态bypass处于有效状态,则判定不需要对所述源传输请求进行切包处理,且直接将所述源传输请求发送至下游设备。
也就是说,若所述静态bypass处于有效状态,则无需要对所述源传输请求进行切包处理,直接将所述源传输请求发送至下游设备。
步骤303:当判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆成多个子传输请求。作为一种可选实施方式,所述将所述源传输请求拆成多个子传输请求,包括:
结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求。
其中,对源传输请求进行切包处理,拆分成的多个子传输请求的ID与源传输请求的ID相同。
举例来说,源传输请求的ID记为A,那么,该源传输请求拆分为3个子传输请求,那么,所述3个子传输请求的ID均为A。
在一具体实施方式中,所述结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求,包括:
如果源传输请求的数据包大于64字节,则对最后一个子传输之前的各个子传输按照64字节进行划分,从首地址开始到第1个64字节对齐地址之前为第一个子传输,第i个64字节对齐地址到第i+1个字节对齐地址之前为第i+1个子传输,其中,i为大于或等于1的正整数,以此类推,直至最后一个子传输的起始地址为第i+1个64字节对齐地址,且所述最后一个子传输为第i+1个64字节对齐地址到传输结束的位置。
也就是说,对于大于64字节的源传输请求,至少能拆成两个子传输请求。
举例来说,若一源传输请求可拆分成4个子传输请求,则从首地址开始到第1个64字节对齐地址之前为第一个子传输,第1个64字节对齐地址到第2个字节对齐地址之前为第二个子传输,第2个64字节对齐地址到第3个字节对齐地址之前为第三个子传输,第四个子传输为第3个64字节对齐地址到传输结束的位置。
在另一具体实施方式中,所述结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求,包括:
如果源传输请求的数据包小于64字节但是跨64字节边界,则切包后,从首地址开始到第一个64字节对齐地址之前为第一个子传输,从第一个64字节对齐地址到传输结束的位置为第二个子传输。
也就是说,对于小于64字节但是跨64字节边界的源传输请求,只能拆成两个子传输请求。
步骤304:判断是否需要对所述多个子传输请求进行ID转换,若需要进行ID转换,则对所述多个子传输请求进行ID转换,并记录所述多个子传输请求之间的顺序关系。
这里,对所述多个子传输请求进行ID转换的目的是,使同一源传输请求对应的各个子传输请求的ID不同,如此,能够使得CCN设备中的RNI将所述属于同一源传输请求的各个不同的子传输请求能够并行下发至CCN设备的下游处理模块,进而提升带宽利用率,也有助于提高对所述源传输请求的处理效率。
作为一种可选实施方式,所述判断是否需要对所述多个子传输请求进行ID转换,包括:
根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换。
作为一种可选实施方式,所述根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换,包括:
若使能位处于无效状态,判定不需要进行ID转换;
若使能位处于有效状态,判断子传输请求对应的地址是否在动态bypass地址段中,如果不在,判定需要对所述子传输请求进行ID转换;如果在,判定不需要对所述子传输请求进行ID转换。
作为一种可选实施方式,所述对所述多个子传输请求进行ID转换,包括:
从资源ID中选择最小的空闲ID值作为所述子传输请求的ID。
这里,选择最小的空闲ID值作为所述子传输请求的ID,是为了便于更有序地确定各个子传输请求的ID值。
当然,也可以从资源ID中选择任意一空闲ID值作为所述子传输请求的ID。
例如,假设资源ID包括0~31共32个ID值,且当前处于空闲状态的资源ID有5、13、17、19、21、30六个可用资源ID,那么,可以将资源ID=5作为子传输请求A1的ID值,并在将资源ID=5作为所述子传输请求 A1的ID值后,将资源ID=5标记为当前已被占用。这样,后续的子传输请求A2则从当前处于空闲状态的资源ID即13、17、19、21、30五个可用资源ID值中,选择一个空闲ID值作为所述子传输请求A2的ID值。
可选地,所述方法还包括:
若不需要进行ID转换,则按照现有处理方式进行处理。
其中,所述现有处理方式包括:将所述多个子传输请求按照顺序依次下发至下游设备,且每次下发一个,并在接收到下游设备对所述多个子传输请求中的一个传输请求的反馈后,才下发下所述多个子传输请求中的另一个相同ID的子传输请求。
步骤305:获取下游设备针对所述多个子传输请求的响应。
可选地,在所述获取下游设备针对所述多个子传输请求的响应之前,所述方法还包括:
将所述多个子传输请求下发至下游设备,以由所述下游设备对所述多个子传输请求作出响应。
步骤306:基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
作为一种可选实施方式,所述基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应,包括:
若所述源传输请求为读请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,按照同一源传输请求的各个子传输请求之间的顺序关系向上游设备返回针对所述同一源传输请求的各个子传输请求的读响应;
若所述源传输请求为写请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,对针对同一源传输请求的所有子传输请求的响应进行合并,并在合并后向所述上游设备返回针对所述同一源传输请求的写响应。
本发明实施例提供的切包方法,在需要进行ID转换时,对所述多个子传输请求进行ID转换,由于同一源传输请求对应的各个子传输请求在转换之后对应的ID不同,CCN设备中的RNI能够一次性下发ID不同的所述同一源传输请求对应的各个子传输请求,从而能够提高处理效率,提升带宽 利用率。
实施例二
基于上述实施例一所述技术方案,本实施例给出了切包装置的一种实现原理框图,如图5所示,在切包装置中主要分为切包处理、ID转换处理、保序处理三大部分,各部分的功能如下:
切包处理部分:结合收到的AXI总线中请求的首地址、大小、长度三者信息对源传输请求进行切包处理,拆分成多个子传输。
ID转换处理部分:将切包后的子传输请求与动态bypass配置窗口和使能位进行比较,判断是否需要进行ID转换,如果匹配有效,则跨过ID转换模块,并将AxID的第六比特位置为1,低五位保持不变,表示该传输没有进行AxID值的转换,直接输出到保序模块;与动态bypass配置窗口和使能位匹配失败的传输,将AxID的第六比特位置零,重新生成低五位后,输出到保序模块。
保序处理部分:对收到的传输进行保序处理,由于同一个源传输切包后,各个子传输的AxID不同,下游设备对子传输是并行处理的,因此各个子传输的响应返回顺序可能乱序。将收到来自下游设备的响应进行暂存,并按照与发送给下游设备的相同顺序将响应返回给上游设备。
有益效果:理性情况下,切包装置的实际流量如下:
1)写数据,在切包装置的Master口方向,每两个源传输之间的写数据只有一个时钟周期间隔;
2)读数据,在切包装置的Slave口方向,来自上游设备的AxID相同的情况下,每两个子传输之间的读数据只有一个时钟周期的间隔;来自上游设备的AxID不同的情况下,读数据达到满带宽。
经实测发现,在RNI工作频率1.4GHz,理想激励情况下,即AXI_MST最大限度的发送相同和不同AxID的激励给RNI处理,其带宽分别为5GB/s和20GN/s,相差达4倍之多。
实施例三
本实施例提供了一种切包装置,所述切包装置的组成结构示意图如图6所示,所述切包装置包括:第一判断模块61、切包模块62、第二判断模块63、ID转换模块64、保序模块65;其中,
所述第一判断模块61,配置为获取上游设备发送的源传输请求,判断是否需要对所述源传输请求进行切包处理;
所述切包模块62,配置为当所述第一判断模块61判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆成多个子传输请求;
第二判断模块63,配置为判断是否需要对所述多个子传输请求进行ID转换;
所述ID转换模块64,配置为若所述第二判断模块63判定需要进行ID转换,则对所述多个子传输请求进行ID转换;
所述保序模块65,配置为记录所述多个子传输请求之间的顺序关系;获取下游设备针对所述多个子传输请求的响应;基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
在一可选实施方式中,所述第一判断模块61,还配置为:
检测静态bypass的状态;
若所述静态bypass处于无效状态,则判定需要对所述源传输请求进行切包处理;
若所述静态bypass处于有效状态,则判定不需要对所述源传输请求进行切包处理,且直接将所述源传输请求发送至下游设备。
在一可选实施方式中,所述切包模块62,还配置为:
结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求。
在一具体实施方式中,所述切包模块62,配置为:
如果源传输请求的数据包大于64字节,则对最后一个子传输之前的各个子传输按照64字节进行划分,从首地址开始到第1个64字节对齐地址 之前为第一个子传输,第i个64字节对齐地址到第i+1个字节对齐地址之前为第i+1个子传输,其中,i为大于或等于1的正整数,以此类推,直至最后一个子传输的起始地址为第i+1个64字节对齐地址,且所述最后一个子传输为第i+1个64字节对齐地址到传输结束的位置;
如果源传输请求的数据包小于64字节但是跨64字节边界,则切包后,从首地址开始到第一个64字节对齐地址之前为第一个子传输,从第一个64字节对齐地址到传输结束的位置为第二个子传输。
在一可选实施方式中,所述第二判断模块63,还配置为:
根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换。
在一具体实施方式中,所述第二判断模块63,配置为:
若使能位处于无效状态,判定不需要进行ID转换;
若使能位处于有效状态,判断子传输请求对应的地址是否在动态bypass地址段中,如果不在,判定需要对所述子传输请求进行ID转换;如果在,判定不需要对所述子传输请求进行ID转换。
在一可选实施方式中,所述ID转换模块64,配置为:
从资源ID中选择最小的空闲ID值作为所述子传输请求的ID。
在一可选实施方式中,所述保序模块65,还配置为:
若所述源传输请求为读请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,按照同一源传输请求的各个子传输请求之间的顺序关系向上游设备返回针对所述同一源传输请求的各个子传输请求的读响应;
若所述源传输请求为写请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,对针对同一源传输请求的所有子传输请求的响应进行合并,并在合并后向所述上游设备返回针对所述同一源传输请求的写响应。
本领域技术人员应当理解,本发明实施例的切包装置中各处理模块的功能,可参照前述切包方法的相关描述而理解,本发明实施例的切包装置中各处理模块,可通过实现本发明实施例所述的功能的模拟电路而实现,也可以通过执行本发明实施例所述的功能的软件在智能终端上的运行而实 现。
本实施例中,所述切包装置中的第一判断模块61、切包模块62、第二判断模块63、ID转换模块64、保序模块65,在实际应用中可由所述切包装置中的中央处理器(CPU,Central Processing Unit)、数字信号处理器(DSP,Digital Signal Processor)或可编程门阵列(FPGA,Field-Programmable Gate Array)等实现。
其中,所述切包装置可设置于上游设备和下游设备之间。这里,所述下游设备是指CCN设备。所述上游设备是指与CCN设备连接的AXI_MST,包括网络加速器、DMA、高速/中速外设,等。
可选地,所述切包装置的输入端与AXI_MST连接,所述切包装置的输出端与所述CCN设备中的RNI接口连接。
本发明实施例的切包装置,能够提高处理效率,提升带宽利用率。
实施例四
对于大于64字节的源传输请求,本实施例提供了切包装置接收到上游设备发送的源传输请求之后的一种处理流程,如图7所示,具体处理步骤如下:
步骤701:切包装置判断静态bypass的状态,如果静态bypass处于有效状态,将源传输请求发送至下游设备中的RNI;如果静态bypass处于无效状态,执行步骤702;
可选地,切包装置中的第一判断模块通过判断静态bypass的状态来判断是否需要对收到的源传输请求进行切包。如果静态bypass处于有效状态,来自上游设备的源传输请求将跨过切包装置的后续处理流程,直接与下游设备连通。
步骤702:切包装置对收到的源传输请求进行切包,然后执行步骤703;
可选地,由切包装置中的切包模块对源传输请求进行切包。
例如,对于地址递增类型的源传输请求,如果整个源传输请求的数据包大于64字节,从首地址开始到第一个64字节对齐地址之前为第一个子 传输,后续第一个64字节对齐地址到下一个64字节对齐地址位置按照64字节进行切分,以此类推,最后一个子传输为64字节对齐地址到传输结束的位置。
步骤703:切包装置根据动态bypass判断是否需要进行ID转换,如果需要,执行步骤704;如果不需要,执行步骤705;
可选地,源传输请求经过切包处理后,由切包装置中的第二判断模块根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换。
举例来说,一切包装置支持最多32个可配置bypass地址段,地址段颗粒度为2兆字节;每个地址段支持读、写动态bypass使能位对读写通道进行是否ID转换的分别控制,其中,32个地址段的控制信息都由dynbp_addr_begin、dynbp_addr_end和dynbp_addr_bypass组成,即起始段地址,结束段地址和使能信号组成。其中,dynbp_addr_begin和dynbp_addr_end为14位信号,对应35位AXI地址的高14位;dynbp_addr_bypass为2位信号,高低位分别控制该地址段写和读通道的动态bypass使能。
在动态bypass使能位处于使能的情况下,对应地址段的内的源传输请求将不进行ID转换,通过该配置,可以让切包装置灵活应用于不同的场景,且两者可以并存。例如,对于数据空间通常需要切包和ID转换处理以提高系统性能,而对于一些配置空间,对配置的先后顺序有严格要求,此时把相应的地址段配置到动态bypass窗口,以使其严格按照原始请求的顺序执行。
步骤704:切包装置进行ID转换,然后执行步骤705;
可选地,由切包装置中的ID转换模块进行ID转换。
例如,一切包装置支持最多32个读请求和32个写请求,经过bypass判断后,若子传输需要进行ID转换,将从资源ID中选择最小的空闲ID值作为该子传输的AxID,然后将该子传输发给切包装置中的保序模块。
步骤705:切包装置进行保序处理,然后将经保序处理之后的子传输请 求发送至下游设备中的RNI。
可选地,由切包装置中的保序模块进行保序处理。
这里,为保证在所支持的outstanding内AxID值不同,对返回的响应和数据要求根据ID转换前的源AxID信息进行保序处理。这里,所述outstanding包括读请求和写请求。
以读传输为例,每一笔发到下游设备中的读请求都将占用一个不同的AxID值,下游设备收到读数据先暂存,收齐全部读数据并把数据返回给上游设备,再释放该AxID。由于下游设备收到的是不同AxID的子传输,因此返回给切包装置的读数据将会是乱序的。对于同一个源传输切包后的子传输,切包装置中的保序模块将对乱序返回的读数据进行保序处理,按照发给下游设备的相同顺序将响应返回给上游设备,保序的方式主要通过对同一源传输对应的子传输的AxID组链来实现。
例如,某源传输的AxID为1,切分成3个子传输后顺序发到下游设备时的AxID分别为4,5,6。在ID转换内部会将这三个AxID值组成一条链,id4->id5->id6,id4前面没有其他的AxID,说明收到id4的读数据可以按源传输的AxID返回给上游设备,等id4的数据全部返回给上游设备后,更新该链表为id5->id6,以此类推将id5和id6的数据按源传输的AxID返回给上游设备。
对于写传输,写通道ID转换和保序处理方式与读通道基本相同,区别在于对同一个源传输的几个子传输写响应,会进行“或”的逻辑操作合并成一笔响应后再返回给上游设备。
对于小于64字节但是跨64字节边界的源传输请求,处理步骤与大于64字节的源传输请求类似,区别在于:步骤602切包处理和步骤605保序处理。可选地,步骤602切包处理包括:对于小于64字节但是跨64字节边界的源传输,切包后第一个子传输为从首地址开始到第一个64字节对齐地址之前为第一个子传输,第二个子传输为64字节对齐地址到传输结束的位置,一共只有两个子传输。同理,对于小于64字节但是跨64字节边界的源传输,步骤605保序处理的时候,源传输的AxID为7,只会切分成2个子传输。在ID转换时,会将这二个AxID值组成一条链,id8->id9,id8 前面没有其他的AxID,说明收到id8的读数据可以按源传输的AxID返回给上游设备,等id8的数据全部返回给上游设备后,更新该链表为id9,id9的数据按源传输的AxID返回给上游设备。
这里,需要说明的是,当切包装置接收到下游设备反馈的针对各个子传输请求的响应时,基于步骤705中进行保序处理时记录的所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
可选地,若所述源传输请求为读请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,按照同一源传输请求的各个子传输请求之间的顺序关系向上游设备返回针对所述同一源传输请求的各个子传输请求的读响应;若所述源传输请求为写请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,对针对同一源传输请求的所有子传输请求的响应进行合并,并在合并后向所述上游设备返回针对所述同一源传输请求的写响应。
需要说明的是,当切包装置向上游设备反馈下游设备的响应时,对于切包装置中的一些模块只经过,不进行任何处理,例如,步骤701中的第一判断模块、步骤702中的切包模块、步骤703中的第二判断模块不需要进行任何处理。
本发明实施例提供的切包处理流程,能够提高处理效率,提升带宽利用率。
本发明实施例提供一种请求处理方法,所述方法包括:
获取源传输请求;
确定是否需要对所述源传输请求进行拆分处理;
当判断出需要对所述源传输请求进行拆分处理时,将所述源传输请求拆分成多个子传输请求;
为每一个所述子传输请求分配不同的标识号;
其中,当所述子传输请求的标识号不同时,用于触发目标设备并行响应所述子传输请求。
该方法可应用于请求拆分装置中,这里的请求拆分装置可以对应于前 述的拆包装置,可为连接上游设备和下游设备的中间设备,该上游设备为发送源传输请求的设备,下游设备可为响应所述源传输请求的设备。具体的上游设备可为AXI_MST;所述下游设备可为CCN设备。
在一些实施例中所述拆分请求可为应用于所述下游设备的一个拆分组件,还可以是应用于所述上游设备中的一个拆分组件。
在本实施例中为每一个子传输请求都设置了不同的ID,这样的话,下游设备在进行数据传输时,就不会因为多个子传输请求共用一个ID,仅能够串行传输导致的传输速率慢的问题。在本实施例中,一个源传输请求对应的多个子传输请求的ID不同,这些子传输请求可以并行响应,这样显然大大的提升了传输速率。
可选地,所述方法还包括:
确定所述子传输请求是否允许并行响应;
所述为每一个所述子传输请求分配不同的标识号,包括:
当所述传输请求允许并行响应时,为每一个所述子传输请求分配不同的标识号。
在有些情况下,上游设备明确需要按顺序接收请求的数据,则不运行并行响应,若并行响应可能会导致上游设备无法按顺序接收到数据包。
故在本实施例中为子传输请求分配不同的标识号,仅允许并行响应的情况下才可以。当所述传输请求不允许并行响应时,为每一个所述子传输分配请求分配相同的标识号并记录所述子传输请求的响应顺序。这样后续,响应所述子传输请求的发送设备,就会按照所述响应顺序向发送所述源传输请求的设备返回数据。
举例来说,源传输请求的ID记为A,那么,不允许并行响应时,该源传输请求拆分为3个子传输请求,那么,所述3个子传输请求的ID可均为A。当然,在一些实施例中需要记录子传输请求的ID与源传输请求ID的对应关系。
举例来说,源传输请求的ID记为A,那么,允许并行响应时,该源传输请求拆分为3个子传输请求,那么,所述3个子传输请求的ID为A1、 A2及A3。A1、A2及A3互不相同。当然,在一些实施例中需要记录子传输请求的ID与源传输请求ID的对应关系。
在一些实施例中,是否允许并行响应可以由第二标识的使能状态来确定。这里的第二标识可对应于前述的动态bypass地址段对应的标识;所述使能状态可以对应于前述实施例中动态bypass地址段的使能状态。
在一些实施例中,所述确定是否需要对所述源传输请求进行拆分处理,包括:
确定所述源传输请求所需请求的目标数据的长度是否大于预设阈值;
当所述目标数据的长度大于所述预设阈值时,确定需要对所述源传输请求进行拆分处理。
这里的预设阈值可为设定的值。这里的源传输请求需求请求的数据包大于预设字节,进行按照预设字节进行传输请求的拆分,以形成每一个均不大于所述预设字节的多个子传输请求。在本实施例所述预设阈值可为64字节或32个字节或128字节等预定的字节数。
可选地,所述确定是否需要对所述源传输请求进行拆分处理,包括:
确定所述源传输请求所需请求的目标数据是否跨了至少两个存储单元;
当所述目标数据跨了至少两个存储单元时,确定需要对所述源传输请求进行拆分。即所述源传输请求的目标数据存储在多个存储单元中,则需要进行拆分。
一般情况下,所述存储单元对应的存储容量等于前述实施例中的预设阈值。
可选地,所述当判断出需要对所述源传输请求进行拆分处理时,将所述源传输请求拆分成多个子传输请求,包括以下至少之一:
根据所述源传输请求所需请求的目标数据的长度L和预设阈值Th,进行所述源传输请求的拆分,其中,所述子传输请求的个数等于不小于所述L/Th的最小整数;
当传输数据位宽W小于所述Th时,根据所述源传输请求所需请求的 目标数据的长度L和传输数据位宽W,进行所述源数据请求的拆分,其中,所述子传输请求的个数等于不小于所述L/W的最小整数;
当所述目标数据在存储区域的首地址和尾地址跨了m个存储单元时,将所述源传输请求拆分成m个子传输请求。
在本实施例中所述传输数据位宽不大于传输请求的发送端和数据的接收端之间的总线的位宽。
可选地,所述方法还包括:
根据当前传输的业务类型对应的数据包长度,配置第一标识的状态;
所述确定是否需要对所述源传输请求进行拆分处理,包括:
当所述第一标识的状态为第一状态时,确定时序需要对所述源传输请求进行拆分处理。
可选地,所述方法还包括:
当所述第一标识的状态为第二状态时,透传所述源传输请求。
这里的第一标识可为前述的静态bypass标识。
不同的业务类型对应的数据包的最大长度不同,故可以根据业务类型对应的数据包长度,来配置第一标识的状态。若第一标识直接设置为第二状态时,则透传所述源传输请求,则会进行是否需要进行源传输请求的拆分的确定及拆分等操作,提升无需拆分情况下的传输效率。
所述第一标识可为1比特的状态值,若该比特的取值为“0”时,对应于第一状态,则该比特的取值为“1”时,对应的为第二状态。若该比特的取值为“1”时,对应于第一状态,则该比特的取值为“0”时,对应的为第二状态。
本发明实施例提供一种计算机存储介质,所述计算机存储介质中存储有计算机可执行指令,所述计算机可执行指令用于执行一个或多个实施例提供的切包方法或一个或多个实施例提供的请求处理方法。
所述计算机存储介质可为移动硬盘、光盘、磁带等各种类型的存储介质,可选为非瞬间存储介质。
本发明实施例还提供一种设备,该设备包括:存储器及处理器;该处理器与存储器连接,能够通过计算机程序等计算机可执行指令的执行,实现前述的一个或多个技术方案提供的切包方法和/或请求处理方法。
本发明实施例所记载的技术方案之间,在不冲突的情况下,可以任意组合。
在本发明所提供的几个实施例中,应该理解到,所揭露的方法和智能设备,可以通过其它的方式实现。以上所描述的设备实施例仅仅是示意性的,例如,所述单元的划分,仅仅为一种逻辑功能划分,实际实现时可以有另外的划分方式,如:多个单元或组件可以结合,或可以集成到另一个系统,或一些特征可以忽略,或不执行。另外,所显示或讨论的各组成部分相互之间的耦合、或直接耦合、或通信连接可以是通过一些接口,设备或单元的间接耦合或通信连接,可以是电性的、机械的或其它形式的。
上述作为分离部件说明的单元可以是、或也可以不是物理上分开的,作为单元显示的部件可以是、或也可以不是物理单元,即可以位于一个地方,也可以分布到多个网络单元上;可以根据实际的需要选择其中的部分或全部单元来实现本实施例方案的目的。
另外,在本发明各实施例中的各功能单元可以全部集成在一个第二处理单元中,也可以是各单元分别单独作为一个单元,也可以两个或两个以上单元集成在一个单元中;上述集成的单元既可以采用硬件的形式实现,也可以采用硬件加软件功能单元的形式实现。
以上所述,仅为本发明的具体实施方式,但本发明的保护范围并不局限于此,凡按照本发明原理所作的修改,都应当理解为落入本发明的保护范围。
工业实用性
本发明实施例中提供的技术方案中,通过源传输请求的拆分及子传输请求的不同标识号的分配,这样响应传输请求的发送设备,可以基于不同 标识号的子传输请求进行数据传输的并行响应,从而提升数据传输的速率,从而具有有益的工业效果,且实现简便,在工业上的可复制性强。

Claims (25)

  1. 一种切包方法,所述方法包括:
    获取上游设备发送的源传输请求;
    判断是否需要对所述源传输请求进行切包处理;
    当判断出需要对所述源传输请求进行切包处理时,将所述源传输请求拆分成多个子传输请求;
    判断是否需要对所述多个子传输请求进行标识号ID转换,若需要进行ID转换,则对所述多个子传输请求进行ID转换,并记录所述多个子传输请求之间的顺序关系;
    获取下游设备针对所述多个子传输请求的响应;
    基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
  2. 根据权利要求1所述的切包方法,其中,所述判断是否需要对所述源传输请求进行切包处理,包括:
    检测静态bypass的状态;
    若所述静态bypass处于无效状态,则判定需要对所述源传输请求进行切包处理;
    若所述静态bypass处于有效状态,则判定不需要对所述源传输请求进行切包处理,且直接将所述源传输请求发送至下游设备。
  3. 根据权利要求1所述的切包方法,其中,所述将所述源传输请求拆分成多个子传输请求,包括:
    结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求。
  4. 根据权利要求3所述的切包方法,其中,所述结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求,包括:
    如果源传输请求的数据包大于64字节,则对最后一个子传输之前的各个子传输按照64字节进行划分,从首地址开始到第1个64字节对齐地址之前为第一个子传输,第i个64字节对齐地址到第i+1个字节对齐地址之前为第i+1个子传输,其中,i为大于或等于1的正整数,以此类推,直至最后一个子传输的起始地址为第i+1个64字节对齐地址,且所述最后一个子传输为第i+1个64字节对齐地址到传输结束的位置;
    如果源传输请求的数据包小于64字节但是跨64字节边界,则切包后,从首地址开始到第一个64字节对齐地址之前为第一个子传输,从第一个64字节对齐地址到传输结束的位置为第二个子传输。
  5. 根据权利要求1所述的切包方法,其中,所述判断是否需要对所述多个子传输请求进行ID转换,包括:
    根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换。
  6. 根据权利要求5所述的切包方法,其中,所述根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换,包括:
    若使能位处于无效状态,判定不需要进行ID转换;
    若使能位处于有效状态,判断子传输请求对应的地址是否在第二标识的地址段中,如果不在,判定需要对所述子传输请求进行ID转换;如果在,判定不需要对所述子传输请求进行ID转换。
  7. 根据权利要求6所述的切包方法,其中,所述对所述多个子传输请求进行ID转换,包括:
    从资源ID中选择最小的空闲ID值作为所述子传输请求的ID。
  8. 根据权利要求1所述的切包方法,其中,所述基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应,包括:
    若所述源传输请求为读请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,按照同一源传输请求的各个子传输请求之间的顺序关系向上游设备返回针对所述同一源传输请求的各个子传输请求的读响应;
    若所述源传输请求为写请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,对针对同一源传输请求的所有子传输请求的响应进行合并,并在合并后向所述上游设备返回针对所述同一源传输请求的写响应。
  9. 一种切包装置,所述切包装置包括:
    第一判断模块,配置为获取上游设备发送的源传输请求,判断是否需要对所述源传输请求进行切包处理;
    切包模块,配置为当所述第一判断模块判断出需要对所述源传输请求进行拆分处理时,将所述源传输请求拆分成多个子传输请求;
    第二判断模块,配置为判断是否需要对所述多个子传输请求进行ID转换;
    ID转换模块,配置为若所述第二判断模块判定需要进行ID转换,则对所述多个子传输请求进行ID转换;
    保序模块,配置为记录所述多个子传输请求之间的顺序关系;获取下游设备针对所述多个子传输请求的响应;基于所述多个子传输请求之间的顺序关系,向所述上游设备返回针对所述源传输请求的响应。
  10. 根据权利要求9所述的切包装置,其中,所述第一判断模块,还配置为:检测第一预设标识的状态;
    若所述第一预设标识处于无效状态,则判定需要对所述源传输请求进行切包处理;
    若所述第一预设标识处于有效状态,则判定不需要对所述源传输请求进行切包处理,且直接将所述源传输请求发送至下游设备。
  11. 根据权利要求9所述的切包装置,其中,所述切包模块,还配置为:
    结合源传输请求的首地址、大小、长度,对源传输请求进行切包处理,并拆分成多个子传输请求。
  12. 根据权利要求11所述的切包装置,其中,所述切包模块,配置为:
    如果源传输请求的数据包大于64字节,则对最后一个子传输之前的各个子传输按照64字节进行划分,从首地址开始到第1个64字节对齐地址 之前为第一个子传输,第i个64字节对齐地址到第i+1个字节对齐地址之前为第i+1个子传输,其中,i为大于或等于1的正整数,以此类推,直至最后一个子传输的起始地址为第i+1个64字节对齐地址,且所述最后一个子传输为第i+1个64字节对齐地址到传输结束的位置;
    如果源传输请求的数据包小于64字节但是跨64字节边界,则切包后,从首地址开始到第一个64字节对齐地址之前为第一个子传输,从第一个64字节对齐地址到传输结束的位置为第二个子传输。
  13. 根据权利要求9所述的切包装置,其中,所述第二判断模块,还配置为:
    根据配置寄存器中的动态bypass地址段和使能位来判断是否需要进行ID转换。
  14. 根据权利要求13所述的切包装置,其中,所述第二判断模块,具体用于:
    若使能位处于无效状态,判定不需要进行ID转换;
    若使能位处于有效状态,判断子传输请求对应的地址是否在动态bypass地址段中,如果不在,判定需要对所述子传输请求进行ID转换;如果在,判定不需要对所述子传输请求进行ID转换。
  15. 根据权利要求14所述的切包装置,其中,所述ID转换模块,还配置为:
    从资源ID中选择最小的空闲ID值作为所述子传输请求的ID。
  16. 根据权利要求9所述的切包装置,其中,所述保序模块,还配置为:
    若所述源传输请求为读请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,按照同一源传输请求的各个子传输请求之间的顺序关系向上游设备返回针对所述同一源传输请求的各个子传输请求的读响应;
    若所述源传输请求为写请求,则当接收到下游设备反馈的针对各个子传输请求的响应时,对针对同一源传输请求的所有子传输请求的响应进行合并,并在合并后向所述上游设备返回针对所述同一源传输请求的写响应。
  17. 一种请求处理方法,所述方法包括:
    获取源传输请求;
    确定是否需要对所述源传输请求进行拆分处理;
    当判断出需要对所述源传输请求进行拆分处理时,将所述源传输请求拆分成多个子传输请求;
    为每一个所述子传输请求分配不同的标识号;
    其中,当所述子传输请求的标识号不同时,用于触发目标设备并行响应所述子传输请求。
  18. 根据权利要求17所述的方法,还包括:
    确定所述子传输请求是否允许并行响应;
    所述为每一个所述子传输请求分配不同的标识号,包括:
    当所述传输请求允许并行响应时,为每一个所述子传输请求分配不同的标识号。
  19. 根据权利要求17所述的方法,其中,
    所述方法还包括:
    当所述传输请求不允许并行响应时,为每一个所述子传输分配请求分配相同的标识号并记录所述子传输请求的响应顺序,其中,所述响应顺序用于指示所述子传输请求被响应的顺序关系。
  20. 根据权利要求17至19任一项所述的方法,其中,
    所述确定是否需要对所述源传输请求进行拆分处理,包括:
    确定所述源传输请求所需请求的目标数据的长度是否大于预设阈值;
    当所述目标数据的长度大于所述预设阈值时,确定需要对所述源传输请求进行拆分处理。
  21. 根据权利要求17至19任一项所述的方法,其中,
    所述确定是否需要对所述源传输请求进行拆分处理,包括:
    确定所述源传输请求所需请求的目标数据是否跨了至少两个存储单 元;
    当所述目标数据跨了至少两个存储单元时,确定需要对所述源传输请求进行拆分。
  22. 根据权利要求17至19任一项所述的方法,其中,
    所述当判断出需要对所述源传输请求进行拆分处理时,将所述源传输请求拆分成多个子传输请求,包括以下至少之一:
    根据所述源传输请求所需请求的目标数据的长度L和预设阈值Th,进行所述源传输请求的拆分,其中,所述子传输请求的个数等于不小于所述L/Th的最小整数;
    当传输数据位宽W小于所述Th时,根据所述源传输请求所需请求的目标数据的长度L和传输数据位宽W,进行所述源数据请求的拆分,其中,所述子传输请求的个数等于不小于所述L/W的最小整数;
    当所述目标数据在存储区域的首地址和尾地址跨了m个存储单元时,将所述源传输请求拆分成m个子传输请求。
  23. 根据权利要求17至19任一项所述的方法,其中,
    所述方法还包括:
    根据当前传输的业务类型对应的数据包长度,配置第一标识的状态;
    所述确定是否需要对所述源传输请求进行拆分处理,包括:
    当所述第一标识的状态为第一状态时,确定时序需要对所述源传输请求进行拆分处理。
  24. 根据权利要求23所述的方法,其中,
    所述方法还包括:
    当所述第一标识的状态为第二状态时,透传所述源传输请求。
  25. 一种计算机存储介质,所述计算机存储介质中存储有计算机可执行指令,所述计算机可执行指令用于执行权利要求1至8任一项提供的切包方法或17至24任一项提供的请求处理方法。
PCT/CN2017/099344 2017-01-20 2017-08-28 切包方法、请求处理方法、装置和计算机存储介质 WO2018133414A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201710051983.5A CN108337286A (zh) 2017-01-20 2017-01-20 一种切包方法及装置
CN201710051983.5 2017-01-20

Publications (1)

Publication Number Publication Date
WO2018133414A1 true WO2018133414A1 (zh) 2018-07-26

Family

ID=62909011

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/099344 WO2018133414A1 (zh) 2017-01-20 2017-08-28 切包方法、请求处理方法、装置和计算机存储介质

Country Status (2)

Country Link
CN (1) CN108337286A (zh)
WO (1) WO2018133414A1 (zh)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111988235A (zh) * 2020-08-13 2020-11-24 暨南大学 一种基于多http/3连接的并行传输机制
CN113051076A (zh) * 2021-03-23 2021-06-29 湖南快乐阳光互动娱乐传媒有限公司 一种请求数据处理方法和装置
CN113691630A (zh) * 2021-08-26 2021-11-23 北京字节跳动网络技术有限公司 Dns请求处理方法、装置、系统、电子设备和存储介质
CN115374032A (zh) * 2022-10-24 2022-11-22 北京智芯微电子科技有限公司 存储器中非对齐数据传输方法、写入方法和装置
CN118113230A (zh) * 2024-04-23 2024-05-31 沐曦集成电路(上海)有限公司 一种基于地址重映射的读存储器数据的方法

Families Citing this family (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112328520B (zh) * 2020-09-30 2022-02-11 郑州信大捷安信息技术股份有限公司 一种pcie设备、基于pcie设备的数据传输方法和系统

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101382955A (zh) * 2008-09-28 2009-03-11 中国科学院计算技术研究所 一种机群文件系统中文件读取的方法及系统
CN102014111A (zh) * 2009-09-04 2011-04-13 无锡江南计算技术研究所 数据传输方法、消息引擎、通信节点及网络系统
WO2013107393A1 (zh) * 2012-01-18 2013-07-25 华为技术有限公司 一种消息式内存访问装置及其访问方法
CN103532875A (zh) * 2013-10-12 2014-01-22 江苏华丽网络工程有限公司 一种用于pcie应用层接口的重排序技术

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101321293B (zh) * 2007-06-06 2010-08-18 中兴通讯股份有限公司 一种实现多路节目复用的装置和方法

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101382955A (zh) * 2008-09-28 2009-03-11 中国科学院计算技术研究所 一种机群文件系统中文件读取的方法及系统
CN102014111A (zh) * 2009-09-04 2011-04-13 无锡江南计算技术研究所 数据传输方法、消息引擎、通信节点及网络系统
WO2013107393A1 (zh) * 2012-01-18 2013-07-25 华为技术有限公司 一种消息式内存访问装置及其访问方法
CN103532875A (zh) * 2013-10-12 2014-01-22 江苏华丽网络工程有限公司 一种用于pcie应用层接口的重排序技术

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111988235A (zh) * 2020-08-13 2020-11-24 暨南大学 一种基于多http/3连接的并行传输机制
CN111988235B (zh) * 2020-08-13 2023-05-09 暨南大学 一种基于多http/3连接的并行传输方法
CN113051076A (zh) * 2021-03-23 2021-06-29 湖南快乐阳光互动娱乐传媒有限公司 一种请求数据处理方法和装置
CN113691630A (zh) * 2021-08-26 2021-11-23 北京字节跳动网络技术有限公司 Dns请求处理方法、装置、系统、电子设备和存储介质
CN115374032A (zh) * 2022-10-24 2022-11-22 北京智芯微电子科技有限公司 存储器中非对齐数据传输方法、写入方法和装置
CN115374032B (zh) * 2022-10-24 2023-02-17 北京智芯微电子科技有限公司 存储器中非对齐数据传输方法、写入方法和装置
CN118113230A (zh) * 2024-04-23 2024-05-31 沐曦集成电路(上海)有限公司 一种基于地址重映射的读存储器数据的方法
CN118113230B (zh) * 2024-04-23 2024-07-12 沐曦集成电路(上海)有限公司 一种基于地址重映射的读存储器数据的方法

Also Published As

Publication number Publication date
CN108337286A (zh) 2018-07-27

Similar Documents

Publication Publication Date Title
WO2018133414A1 (zh) 切包方法、请求处理方法、装置和计算机存储介质
US20200364080A1 (en) Interrupt processing method and apparatus and server
US11755513B2 (en) Data processing and writing method based on virtual machine memory identification field and devise
CN106919452B (zh) 多核异构系统及其硬件资源的管理方法
CN108848006B (zh) 一种端口状态的监测方法与装置
JP7375209B2 (ja) マルチリンク通信方法、装置、およびシステム
US7814178B2 (en) Method and apparatus for data configuration in communication device
WO2017128905A1 (zh) 一种单主机多环most网络通信系统
CN113396565A (zh) 一种控制方法及装置
CN113316218B (zh) 接入方法及装置
WO2020232895A1 (zh) 动态分库路由方法、装置、服务器及存储介质
WO2023072065A1 (zh) 数据处理方法、装置、电子设备和存储介质
CN111581136B (zh) 一种dma控制器及其实现方法
CN113297124B (zh) 设备枚举方法、设备、系统及存储介质
US11758391B2 (en) User equipment (UE) identification in a wireless communication network
IL272881B2 (en) Method and device for wireless communication
WO2023030178A1 (zh) 一种基于用户态协议栈的通信方法及相应装置
CN113691466A (zh) 一种数据的传输方法、智能网卡、计算设备及存储介质
CN114579505A (zh) 芯片以及核间通信方法
CN111314272B (zh) 一种任务处理方法及装置
CN115756296A (zh) 缓存管理方法和装置、控制程序及控制器
CN112527722A (zh) 异构多核嵌入式系统及该系统的数据交互方法
CN108738035B (zh) 一种多制式基带芯片的数据处理方法及装置、处理设备
WO2024055660A1 (zh) 一种数据运算方法、装置及设备
CN104301000A (zh) 利用样点级加速器进行数据处理的方法和样点级加速器

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: 17893264

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17893264

Country of ref document: EP

Kind code of ref document: A1