A method and device for realizing solid state hard disk data channel virtualization
Through the queue management module of the hardware acceleration control chip, the queue management module handles the submission and completion of queue elements, and directly conducts data interaction at the hardware layer, solving the problem of data transmission delay and performance degradation caused by solid-state hard disk virtualization in the prior art, and achieving low cost and high efficiency data channel virtualization.
Patent Information
- Application Number
- CN202410969357.4
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-19
- Publication Date
- 2024-12-03
- Estimated Expiration
- 2044-07-19
AI Technical Summary
The prior art realizes solid-state drive virtualization through software, resulting in increased data transmission delays and reduced performance, and the system structure is complex and costly.
The hardware acceleration control chip is adopted to process the submission queue elements and the completion queue elements through the queue management module, which realizes the update of the queue mapping relationship between the virtual hard disk and the physical hard disk, and directly interacts data at the hardware layer to avoid the delay in software virtualization.
Reduces data transmission delay, improves data transmission efficiency, reduces system complexity and production costs, and supports more virtualization needs of solid-state drives.
Smart Images

Figure CN118519586B_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of computer interface devices, and in particular to a method for implementing solid state hard disk data channel virtualization, and a device for implementing the method. Background Art
[0002] With the development of computer technology, more and more computer devices use computer interface expansion technology to expand computer interfaces. Among them, PCIE is a common computer interface expansion technology.
[0003] PCIE, also known as PCI-Express (peripheral component interconnect express), is a standard interface technology proposed by the PCI-SIG organization for connecting internal components of a computer. Its architecture consists of PCIE devices such as root components RC (root complex), switches, and endpoint devices (endpoint / ep). The switch is used to expand the PCIE interface, so that the host or root component can connect to more endpoint devices.
[0004] PCIE interface technology is widely used in data storage in data centers. For example, each virtual machine corresponds to a physical solid-state drive (NVMe SSD), which provides independent data storage functions for the corresponding virtual machine. Due to the development of storage technology, the storage capacity of a single solid-state drive has become very large, and the performance has reached the requirements of PCIE Gen5×4. However, due to the limited number of PCIE slots supported by the host, as the number of virtual machines continues to grow, the existing technology needs to virtualize the physical solid-state drive into multiple small virtual hard drives through virtualization technology and provide them to multiple virtual machines for use. The existing practice is usually to use software to achieve the virtualization of the solid-state drive. However, virtualization achieved by software will increase the delay of data transmission and will also lead to a decrease in the performance of the solid-state drive.
[0005] The invention patent application with announcement number CN115840620A discloses a data path construction method, which detects the read and write operation request initiated by the client virtual machine through the FPGA card, determines the diversion mark according to the corresponding instruction read and write operation or data read and write operation, and adds it to the submission queue element corresponding to the read and write operation request, and writes the submission queue element to the submission queue; the solid-state drive initiates a direct memory access operation for executing the read and write operation request according to the address information in the submission queue element; the FPGA card responds to the initiated direct read and write operation, and selectively diverts the direct memory access operation to the memory of the FPGA card or the host according to the diversion mark. When the direct memory access operation is diverted to the host, a data path between the solid-state drive and the host is constructed, which is used for the solid-state drive to read and write the memory of a given client virtual machine. In this way, the NVMe virtualized data path is realized, the latency is reduced, the bandwidth is increased, and the native NVMe driver can be used.
[0006] However, in the data path constructed by the above method, the direct memory access operation of the solid-state drive is diverted to the host through the FPGA card, which increases the data transmission delay of the data channel, and the maximum performance of the data channel is also limited by the interface performance of the FPGA card itself. In order to plug in more solid-state drives (usually up to 4), the physical size of the FPGA card must be full height and full length, and a host can generally support 24 solid-state drives, which requires 6 FPGA cards. The system structure is complex, and due to the high production cost of FPGA cards, the production cost of the entire system is too high. Summary of the invention
[0007] The first object of the present invention is to provide a method for realizing data channel virtualization of a solid state drive with low production cost and less data transmission delay.
[0008] A second object of the present invention is to provide a device for implementing the above-mentioned method for virtualizing a solid state drive data channel.
[0009] To achieve the first purpose of the present invention, the method for realizing solid-state hard disk data channel virtualization provided by the present invention is applied to a hardware acceleration control chip, and the hardware acceleration control chip has a queue management module, and the method includes: a submission queue element processing module of the queue management module obtains the submission queue element currently to be processed from the host, and modifies the submission queue element recorded in the queue management module according to the mapping relationship between the submission queue of the virtual hard disk and the submission queue of the physical hard disk; the solid-state hard disk reads the submission queue element recorded in the queue management module, and obtains the information of the physical area page list corresponding to the host by parsing the submission queue element, and exchanges data with the host according to the information of the submission queue element and the physical area page list; the solid-state hard disk sends a completion queue element to the queue management module, and the queue management module updates the completion queue element according to the queue mapping relationship between the virtual hard disk and the physical hard disk and sends it to the host.
[0010] As can be seen from the above scheme, the present invention realizes the management of submission queue elements and completion queue elements by setting a hardware acceleration control chip, and can realize the update of submission queue elements and completion queue elements according to the queue mapping relationship between the virtual hard disk and the physical hard disk. Since the present invention does not realize the virtualization of the solid-state hard disk by software, but by a physical hardware acceleration control chip, the solid-state hard disk directly interacts with the host, the delay of data transmission is very small, and the efficiency of data transmission can be improved.
[0011] In addition, since the hardware acceleration control chip is not directly connected to the SSD through hardware connection, the hardware acceleration control chip and the SSD communicate through the root component or PCIE switch, and the data transmission between the SSD and the host will not be limited by bandwidth due to the hardware acceleration control chip, the board size of the hardware acceleration control chip can be made very small, and it can simultaneously support the virtualization requirements of all SSDs under the same root component. In addition, since multiple SSDs can be plugged in without setting up multiple FPGAs, the production cost of the system is low.
[0012] A preferred solution is that before the submission queue element processing module obtains the current submission queue element to be processed from the host: the host updates the submission queue doorbell status of the virtual hard disk and triggers the virtual hard disk queue scheduler, and the virtual hard disk queue scheduler sends the scheduling result to the submission queue element processing module.
[0013] It can be seen that the host first triggers the virtual hard disk queue scheduler by updating the doorbell state of the submission queue of the virtual hard disk, so that the host can actively obtain the data of the submission queue and completion queue of each virtual hard disk.
[0014] A further solution is that after the queue management module modifies the submission queue element recorded in the queue management module, the submission queue element processing module updates the submission queue doorbell status recorded in the doorbell register of the solid state drive in a point-to-point manner.
[0015] It can be seen that after the submission queue element is recorded in the queue management module, the submission queue doorbell status recorded in the doorbell register of the solid-state drive can be updated in time, thereby triggering the update of the submission queue element of the solid-state drive.
[0016] A further solution is that after the queue management module updates the completion queue element, the queue management module also updates the completion queue doorbell status of the solid state drive in a point-to-point manner.
[0017] It can be seen that after the queue management module updates the completion queue element, it triggers the update of the completion queue element of the solid-state drive by updating the completion queue doorbell state of the solid-state drive.
[0018] A further solution is that after the queue management module updates the completion queue doorbell status of the solid state drive, the host updates the completion queue doorbell status recorded by the queue management module after receiving the interrupt request information or polling the completion queue, if it determines that the current IO operation is completed.
[0019] It can be seen that through the above operations, the IO operation initiated by the host can be completed. The entire operation process does not involve hardware such as FPGA. The host and the solid-state drive can directly transmit data, making the data transmission delay time very short, and the bandwidth of data communication is not limited by hardware performance.
[0020] A further solution is that the method also includes a virtual device initialization operation: after the host is started, all interface devices are enumerated, and a physical driver is loaded into the solid-state hard disk, a submission queue and a completion queue are created in the host's memory, and when creating the submission queue and the completion queue, the base address of the submission queue and the completion queue is configured as the base address register space of the computer interface device virtualized by the hardware acceleration control chip; the virtual driver is loaded into the solid-state hard disk, and all base address register space read and write access operations except the IO queue are transmitted from the computer interface to the controller of the hardware acceleration control chip for processing, and the controller returns the completion TLP information according to the base address register information, and transmits it to the host through the computer interface.
[0021] It can be seen that before the host initiates an IO operation, the virtual device is first initialized, that is, a virtual driver is pre-loaded to the solid-state drive, so that the solid-state drive can virtualize multiple virtual hard disks.
[0022] A further solution is that when the host enumerates all interface devices, the hardware acceleration control chip transmits the request information for configuring TLP information to the processor, and the processor encapsulates the saved configuration space information and forms TLP information and returns it to the computer interface, and then returns it to the host through the computer interface.
[0023] To achieve the above-mentioned second purpose, the device for realizing solid-state hard disk data channel virtualization provided by the present invention includes a host, a solid-state hard disk and a hardware acceleration control chip, and the hardware acceleration control chip has a queue management module; the queue management module has a submission queue element processing module, which is used to obtain the submission queue element currently to be processed from the host, and modify the submission queue element recorded in the queue management module according to the mapping relationship between the submission queue of the virtual hard disk and the submission queue of the physical hard disk; the solid-state hard disk is used to read the submission queue element recorded in the queue management module, and obtain the information of the physical area page list corresponding to the host by parsing the submission queue element, and exchange data with the host according to the information of the submission queue element and the physical area page list; the solid-state hard disk sends the completion queue element to the queue management module, and the queue management module updates the completion queue element according to the queue mapping relationship between the virtual hard disk and the physical hard disk and sends it to the host.
[0024] As can be seen from the above scheme, the hardware acceleration control chip provided by the present invention can manage the submission queue elements and the completion queue elements by the queue management module, and can update the submission queue elements and the completion queue elements according to the queue mapping relationship between the virtual hard disk and the physical hard disk. Since the present invention does not realize the virtualization of the solid-state hard disk by software, but by a physical hardware acceleration control chip, the solid-state hard disk directly interacts with the host for data, which can effectively solve the problem of data transmission delay, thereby improving the efficiency of data transmission.
[0025] A preferred solution is that the hardware acceleration control chip also includes a computer interface and an embedded controller, the computer interface is connected to the host, and the computer interface also has a controller for data exchange.
[0026] It can be seen from the above scheme that the hardware acceleration control chip can communicate with the host through the computer interface. The computer interface can be connected to the host on the one hand and to the embedded controller on the other hand. The queue management module can communicate with the host through the computer interface to meet the needs of data transmission.
[0027] A further solution is that the host includes two processor cores, and the computer interfaces are connected to the two processor cores.
[0028] It can be seen that a hardware acceleration control chip can exchange data with two processor cores, which can meet the usage requirements of a host with two processor cores, and the two processor cores can share one hardware acceleration control chip, which can make the most of the resources of the hardware acceleration control chip and reduce the cost of the entire system. BRIEF DESCRIPTION OF THE DRAWINGS
[0029] Figure 1 It is a structural schematic diagram of an embodiment of a device for realizing solid state hard disk data channel virtualization of the present invention.
[0030] Figure 2 It is a schematic diagram of virtual device management queue processing of an embodiment of the apparatus for implementing solid state hard disk data channel virtualization of the present invention.
[0031] Figure 3 It is a schematic diagram of virtual device IO queue processing in an embodiment of the apparatus for implementing solid state hard disk data channel virtualization of the present invention.
[0032] Figure 4 This is the first part of the flow chart of the method embodiment for realizing the virtualization of the solid state drive data channel of the present invention.
[0033] Figure 5 This is the second part of the flow chart of the method embodiment for realizing the virtualization of the solid state drive data channel of the present invention.
[0034] The present invention is further described below in conjunction with the accompanying drawings and embodiments. DETAILED DESCRIPTION
[0035] The method for realizing solid-state hard disk data channel virtualization of the present invention is used to virtualize the solid-state hard disk, so that a physical solid-state hard disk is virtualized into multiple virtual solid-state hard disks, and a corresponding virtual hard disk storage space is provided to each virtual machine to meet the usage requirements of the growing number of virtual machines.
[0036] Embodiment of a device for implementing solid state drive data channel virtualization:
[0037] See also Figure 1 In this embodiment, without changing the existing host structure, a hardware acceleration control chip 30 is set, so that each solid-state hard disk can still remain in the original slot, and the hardware acceleration control chip 30 and each solid-state hard disk are connected to the root component RC (root complex) of the same group.
[0038] Figure 1What is shown is a typical peer-to-peer virtualization system that meets NVMe (Non Volatile Memory Host Controller Interface Specification). Among them, the host is provided with two independent processor cores, namely CPU10 and CPU11, and the two processor cores CPU10 and CPU11 are respectively connected to the PCIE bus. In this embodiment, each processor core CPU10 and CPU11 are connected to two solid-state hard disks, for example, the processor core CUP10 is connected to solid-state hard disks 21 and 22, and the processor core CUP11 is connected to solid-state hard disks 23 and 24. Of course, in actual applications, each processor core can also be connected to more solid-state hard disks.
[0039] The hardware acceleration control chip 30 includes a computer interface 31, a queue management module 32 and an embedded controller 33, wherein the computer interface 31 is preferably a PCIE interface, and the computer interface 31 is used to connect to the host and the controller 33, and the computer interface 31 needs to support two-way communication, one of which is used to communicate with the controller 33 to realize the configuration of virtual devices and the management of queues, and the other needs to exchange data with the queue management module 32 to transmit queue data between the queue management module 32 and the host, and also transmit queue data between the queue management module 32 and each solid-state hard disk.
[0040] In addition, the computer interface 31 and controller 33 of the hardware acceleration control chip 30 also need to complete the PCIE enumeration of virtual devices, base address register space and management queue processing, and present virtual devices to the host, such as presenting multiple virtual downstream ports, downstream devices, etc. In addition, the computer interface 31 and controller 33 of the hardware acceleration control chip 30 also virtualize a PCIE device, mapping the IO queue space of each solid-state hard disk to the base address register space, so that each solid-state hard disk can access the IO queue space located on the hardware acceleration control chip 30 through a PCIE peer-to-peer (Peer to Peer) method instead of accessing the host's memory, thereby realizing the virtualization of each solid-state hard disk.
[0041] See also Figure 2 The management queue of the virtual device needs to be applied to the virtual NVMe driver and the physical NVMe driver, and needs to be implemented in combination with the hardware acceleration control chip 30 and multiple solid-state drives. Figure 2 The solid state drive 21 is taken as an example for description, and the working principles of other solid state drives are the same.
[0042] Before the host initiates an IO operation, it is first necessary to initialize the virtual device, including creating an IO queue. First, when the host starts, all PCIE devices need to be enumerated, and the host's BIOS (Basic Input Output System) needs to initiate a read and write operation of the configuration space to the hardware acceleration control chip 30. The computer interface 31 of the hardware acceleration control chip 30 transmits the request information of the configuration TLP (Transaction Layer Package) to the embedded controller 33, and the controller 33 processes the operation of configuring the TLP information. After the controller 33 encapsulates the saved configuration space information into TLP information, it sends the encapsulated TLP information to the computer interface 31, and the computer interface 31 returns it to the host's BIOS. By performing the above operations multiple times, the host will complete the enumeration of virtual devices, including completing the enumeration of virtual NVMe devices and virtual PCIE devices, and the virtual PCIE device will provide a cache of the IO queue of the physical hard disk.
[0043] After the host is started, the physical NVMe driver is loaded to the solid-state drive 21. Preferably, the loaded physical NVMe driver is pre-stored on the host. The physical NVMe driver processes the management queue part of the solid-state drive 21 in the same way as the traditional driver, and the submission queue SQ and the completion queue CQ are both created on the host's memory. Among them, the submission queue SQ includes one or more submission queue elements SQE, and the completion queue CQ includes one or more submission queue elements CQE. In addition, when creating an IO queue, the base address of the submission queue and the completion queue is configured to the base address corresponding to the base address register space of the PCIE device virtualized by the hardware acceleration control chip 30.
[0044] After the SSD 21 is initialized, a virtual NVMe driver is loaded to the SSD, preferably, the loaded virtual NVMe driver is also pre-stored on the host, and the virtual NVMe driver is a standard OS driver. In the process of loading the virtual NVMe driver, all read and write accesses of the base address register space except the IO queue are transmitted to the controller 33 by the computer interface 31 for processing, and the controller 33 returns the TLP information according to the information of the cached base address register, and transmits it to the host through the computer interface 31. In addition, the management queue is all created in the host's memory, and the direct memory access (Direct Memory Access, DMA) operations of the submission queue, the completion queue, and the data are all controlled by the software of the controller 33. After the base address register access and direct memory access interaction between the host and the controller 33, the initialization of the virtual NVMe device is completed.
[0045] At this point, the virtual device is initialized and the host can initiate IO operations. Figures 3 to 5 This section describes the process of the host performing IO operations.
[0046] Method embodiment for implementing solid state drive data channel virtualization:
[0047] See also Figure 3 The IO operation initiated by the host involves the processing of the doorbell register, submission queue element (SQE), completion queue element (CQE), physical region page (PRP) and data, and an application interrupt is required to realize the data response. In this embodiment, the hardware acceleration control chip 30 is used to perform the processing of the doorbell, submission queue element, completion queue element and interrupt, while the processing of the physical region page and data is completed by the direct memory access of the solid state drive 21.
[0048] The queue management module 32 is provided with a virtual doorbell register 41, a virtual hard disk queue scheduler 42, a submission queue element processing module 43 and a completion queue element processing module 44, and stores a cache of submission queue elements. Among them, the virtual doorbell register 41 can store the doorbell status of the submission queue and the doorbell status of the completion queue.
[0049] Combination Figure 4 and Figure 5 When the host initiates an IO operation, step S11 is first executed. The host updates the submission queue doorbell status of the virtual hard disk through the virtual NVMe driver, for example, updates the submission queue doorbell status data stored in the virtual doorbell register 41, thereby triggering the virtual disk queue scheduler 42 to work.
[0050] Then, step S12 is executed, the virtual hard disk queue scheduler 42 sends the scheduling result to the submission queue element processing module 43, and then the submission queue element processing module 43 executes step S13 to read the current submission queue element SQE from the host. Figure 3 It can be seen that the host's memory stores the submission queue element SQE. In step S13, the submission queue element processing module 43 reads the submission queue element currently to be processed stored in the host from the host's memory through the computer interface 31.
[0051] Then, step S14 is executed, and the submission queue element processing module 43 modifies the submission queue element recorded by the queue management module 32 according to the mapping relationship between the IO queue of the virtual hard disk and the IO queue of the physical hard disk, and writes it into the cache of the submission queue element. Next, step S15 is executed, and the submission queue element processing module 43 updates the doorbell state of the submission queue of the solid state drive 21 in a point-to-point manner of PCIE, that is, the submission queue doorbell register of the solid state drive 21 is modified, so as to notify the solid state drive 21 that the current submission queue element has changed.
[0052] After the solid state drive 21 obtains the update of the submission queue doorbell status, it executes step S16 to read the cache of the submission queue element in the queue management module 32 in a PCIE point-to-point manner, and then executes step S17. The solid state drive 21 parses the submission queue element and obtains the internal address of the physical area page list (PRP List) corresponding to the host, and reads the corresponding physical area page list information from the host's memory.
[0053] Next, the solid state drive 21 executes step S18. According to the information of the submission queue element and the corresponding physical area page list information, the solid state drive 21 exchanges data with the host's memory, that is, executes corresponding instructions according to the information of the submission queue element, such as reading data from the host's memory, or writing data to the host's memory.
[0054] After the solid state drive 21 completes the data exchange with the host, the solid state drive 21 executes step S19, and sends the completion queue element to the queue processing module 32 in a PCIE point-to-point manner. At this time, the queue management module 32 executes step S20, and the completion queue element processing module 44 modifies the recorded completion queue element and writes it to the host's memory according to the mapping relationship between the IO queue of the virtual hard disk and the IO queue of the physical hard disk. Preferably, the completion queue element processing module 44 needs to send an interrupt to the host before sending the completion queue element.
[0055] Finally, step S21 is executed, and the completion queue element processing module 44 updates the doorbell status of the completion queue element of the solid state drive 21 in a point-to-point manner of PCIE, for example, updating the doorbell status of the completion queue element recorded by the doorbell register of the solid state drive 21. At this time, step S22 is executed, and after the virtual NVMe driver on the host receives an interrupt request or polls the completion queue, if it is confirmed that the current IO operation has been completed, the doorbell register of the virtual hard disk is updated, that is, the doorbell status of the completion queue element of the corresponding virtual hard disk is updated. At this point, the IO operation initiated by the host is completed.
[0056] It can be seen that the main steps of this embodiment are that the queue management module 32 of the hardware acceleration control chip 30 receives the doorbell signal, reads the submission queue element in the submission queue of the queue management module 32, stores the submission queue element in the cache, and the cache is mapped to the base address register space of the virtual device, and then updates the submission queue doorbell status of the solid state drive 21.
[0057] After the solid-state hard disk receives the doorbell status update information of the submission queue, it reads the submission queue elements stored in the acceleration control chip 30 in a point-to-point manner and directly accesses the host's memory. That is, the solid-state hard disk does not access the acceleration control chip 30 during data processing, so that the solid-state hard disk 21 can directly access the host's memory, avoiding the problem of data transmission delay due to hardware limitations.
[0058] After the solid-state hard disk has completed processing the data, the completion queue elements of the acceleration control chip 30 are accessed in a point-to-point manner. The queue management module 32 can receive the completion queue elements of the solid-state hard disk 21 and convert them into completion queue elements of the virtual device, write these completion queue elements into the completion queue of the host, and update the completion queue doorbell status information of the solid-state hard disk 21 in a point-to-point manner.
[0059] In this way, the method of the present invention can realize the virtualization of the solid state drive, and the solid state drive can directly exchange data with the host's memory, which can avoid data transmission delay. In addition, since the data exchange between the solid state drive and the host does not need to go through the acceleration control chip, the acceleration control chip does not need to be connected to the solid state drive, and the data transmission bandwidth between the solid state drive and the host will not be limited by the hardware performance of the acceleration control chip. The acceleration control chip has very low bandwidth requirements for data transmission, and the size of the acceleration control chip can be made very small. One acceleration control chip can simultaneously support the virtualization requirements of all solid state drives under the same root component, and the production cost is low.
[0060] Finally, it should be emphasized that the above are only preferred embodiments of the present invention and are not intended to limit the present invention. For those skilled in the art, the present invention may have various changes and modifications. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.
Claims
1. A method for implementing solid state drive data channel virtualization, characterized in that: The method is applied to a hardware acceleration control chip, wherein the hardware acceleration control chip has a queue management module, and the method comprises: The submission queue element processing module of the queue management module obtains the submission queue element to be processed currently from the host, and modifies the submission queue element recorded in the queue management module according to the mapping relationship between the submission queue of the virtual hard disk and the submission queue of the physical hard disk; The solid state drive reads the submission queue element recorded in the queue management module, obtains information of the physical area page list corresponding to the host by parsing the submission queue element, and exchanges data with the host according to the information of the submission queue element and the physical area page list; The solid state drive sends a completion queue element to the queue management module, and the queue management module updates the completion queue element according to a queue mapping relationship between the virtual hard disk and the physical hard disk and sends the update to the host.
2. The method for implementing solid state drive data channel virtualization according to claim 1, characterized in that: Before the submission queue element processing module obtains the submission queue element to be processed from the host: The host updates the submission queue doorbell state of the virtual hard disk and triggers the virtual hard disk queue scheduler, and the virtual hard disk queue scheduler sends a scheduling result to the submission queue element processing module.
3. The method for implementing solid state drive data channel virtualization according to claim 2, characterized in that: After the queue management module modifies the submission queue element recorded in the queue management module, the submission queue element processing module updates the submission queue doorbell status recorded in the doorbell register of the solid-state drive in a point-to-point manner: the submission queue element processing module modifies the submission queue doorbell register of the solid-state drive, thereby notifying the solid-state drive that the current submission queue element has changed.
4. The method for implementing solid state drive data channel virtualization according to claim 1, characterized in that: After the queue management module updates the completion queue element, the queue management module also updates the completion queue doorbell status of the solid state drive in a point-to-point manner.
5. The method for implementing solid state drive data channel virtualization according to claim 4, characterized in that: After the queue management module updates the completion queue doorbell status of the solid state drive, the host updates the completion queue doorbell status recorded by the queue management module after receiving interrupt request information or polling the completion queue and determining that the current IO operation is completed.
6. The method for implementing solid state drive data channel virtualization according to any one of claims 1 to 5, characterized in that: The method also includes virtual device initialization operations: After the host is started, all interface devices are enumerated, and a physical driver is loaded into the solid state drive, a submission queue and a completion queue are created in the memory of the host, and when creating the submission queue and the completion queue, the base addresses of the submission queue and the completion queue are configured as the base address register space of the computer interface device virtualized by the hardware acceleration control chip; A virtual driver is loaded into the solid-state drive, and all base address register space read and write access operations except the IO queue are transmitted from the computer interface to the controller of the hardware acceleration control chip for processing. The controller returns completion TLP information according to the base address register information and transmits it to the host through the computer interface.
7. The method for implementing solid state drive data channel virtualization according to claim 6, characterized in that: When the host enumerates all interface devices, the hardware acceleration control chip transmits the request information for configuring TLP information to the processor, and the processor encapsulates the saved configuration space information and forms TLP information and returns it to the computer interface, and then returns it to the host through the computer interface.
8. A device for implementing solid state hard disk data channel virtualization, characterized in that: It includes a host, a solid state drive and a hardware acceleration control chip, wherein the hardware acceleration control chip has a queue management module; The queue management module has a submission queue element processing module, which is used to obtain the submission queue element to be processed from the host, and modify the submission queue element recorded in the queue management module according to the mapping relationship between the submission queue of the virtual hard disk and the submission queue of the physical hard disk; The solid state drive is used to read the submission queue element recorded in the queue management module, and obtain the information of the physical area page list corresponding to the host by parsing the submission queue element, and exchange data with the host according to the information of the submission queue element and the physical area page list; The solid state drive sends a completion queue element to the queue management module, and the queue management module updates the completion queue element according to a queue mapping relationship between the virtual hard disk and the physical hard disk and sends the update to the host.
9. The device for implementing solid state hard disk data channel virtualization according to claim 8, characterized in that: The hardware acceleration control chip also includes a computer interface and an embedded controller. The computer interface is connected to the host, and the computer interface and the controller perform data exchange.
10. The device for implementing solid state hard disk data channel virtualization according to claim 9, characterized in that: The host comprises two processor cores, and the computer interfaces are connected to the two processor cores.
Citation Information
Patent Citations
Data path construction method, equipment and medium
CN115840620A
System-on-chip diagnosis method based on NVMe disk
CN115221070A
NVMe acceleration method and device based on multi-level cache and feedback algorithm
CN117785063A
Architecture for flash storage server
US10452279B1
Shared memory mechanism to support fast transport of sq / CQ pair communication between SSD device driver in virtualization environment and physical SSD
US20220365729A1
