TWI475489B - Method of cloning data in a memory for a virtual machine, product of computer programs and computer system - Google Patents

Description

Method for copying memory material of virtual machine, computer program product thereof And computer system

The disclosure is directed to a method of copying data from a virtual machine memory.

System Virualization is the aggregation and abstraction of resources on a shared platform. This abstraction process separates the software from the hardware and enables multiple operating system image archives to run in parallel on a single physical platform without interfering with each other. Virtualization can be combined with workloads running on many physical machines into multiple virtual machines running on a single physical machine, which increases the utilization of computing resources. This combination can greatly reduce the power consumption and floor space required in the data center. A centralized management interface can be used to provision virtual machines, replicate and migrate them as needed.

In order to enable multiple operating systems to run on the same physical platform, a platform layer that operates in a software manner separates the operating system from the hardware. This layer is called a virtual machine manager (hypervisor) or a virtual machine monitor (VMM). Under the system virtualization architecture, the virtualized operating system is called the "Guest Operating System." In order to properly virtualize and isolate visitors, the virtual machine manager or virtual machine monitor (VMM) must control or regulate all privileged operations performed by the guest. The virtual machine manager can use various techniques to accomplish this control or adjustment.

In a virtualized environment or a platform for system virtualization, you need to pay attention to virtual Fast or near-instant copy of a machine (VM). This technology can be used for Load Balancing, Pre-deployment Testing, Speculative Execution, Data Miming, and virus scanning or malware analysis in large data centers. Execution of unknown software may be intercepted or interrupted by an antivirus program or agent when loading or preparing to execute unknown software. At this point, you may request a quick copy of the virtual machine in advance. The virtual machine being replicated has the same operating environment and state as the parent virtual machine or the original virtual machine. This agent can arbitrarily execute this unknown software while performing a virus scan or malicious software analysis. Even in the case where the execution of the software may cause the virtual machine to crash or crash, the original virtual machine is still unaffected because the damage only occurs on the virtual machine being copied.

In traditional technology, Live Migration technology is often used for fast or near-instant copying of virtual machines. Instant migration allows the server administrator to transfer a running virtual machine or application between different physical machines without stopping the connection with the client or stopping the application. For a successful instant migration, the virtual machine's memory, storage, and network connections need to be migrated to the destination. However, the copying of the virtual machine can only begin after the data transfer of the memory, storage device or disk is completed, and thus the virtual machine may not be able to meet the needs of fast or near-instant copying.

In the conventional art, a Snapshot mechanism is also often mentioned for quickly copying a virtual machine and initializing the copied virtual machine. However, performing this snapshot mechanism requires the original virtual machine to be shut down or left in place. In the quiesced state, the copy of the virtual machine in the execution state is thus achieved. This snapshot mechanism can be used to quickly replicate a large number of virtual machines, but does not support instant isolation environments for replication. In other conventional techniques, during the process of executing a copy source virtual machine, the method used to copy the material stored in the physical memory may copy all metadata of the source virtual machine to the target virtual machine. The metadata (eg, page table) in the target virtual machine will be redirected to the source virtual machine's memory. The information stored in all pages of the memory pointed to by the two meta-data will be marked as "read only". Copy-on-write (COW) will be used for access operations. However, in the Page Table Directory (PTD) structure, the only way to set the Page Table Entry (PTE) to read-only is to traverse the entire page table and put all the paging table items. The read flag (Flag) is changed to "read only". This type of operation takes almost the same amount of time as copying the entire page table. If the memory capacity increases as needed, the time spent on such operations will also increase. If the memory capacity reaches several millionths of a gigabyte (GB) or more, the time will increase to more than a few seconds or more, and this requirement will not be able to meet the fast or near-instantaneous virtual machine (VM). Copy technology.

The present disclosure provides one of the exemplary embodiments, including a method of copying a source virtual machine (VM) and at least one material in a memory of the copied virtual machine. The mapping relationship between the guest entity address from the source virtual machine or the copied virtual machine and the host entity address of the memory is determined by a plurality of page tables configured in a plurality of hierarchical levels Righteousness. The method includes: copying metadata of a page table in a highest level or a higher level of the plurality of hierarchical levels to the virtual machine; and selecting the highest level or higher of the plurality of hierarchical levels Remaining metadata of the page table in the level is correspondingly copied to the virtual machine according to the operation of the access; and accessing the corresponding address stored in the memory according to the metadata and the copied metadata data of.

The above and other objects, features, and advantages of the present invention will become more apparent from the description of the appended claims.

The present disclosure provides a method of quickly copying data stored in a memory of a virtual machine (VM). One of the exemplary embodiments includes providing a copy-on-access (COA) copy metadata and copying data by copy-on-write (COW) for rapid copying of virtual machines. The method of memory data. In this way, isolation protection and/or independent execution of different levels of virtual machines can be achieved and applied.

In one of the exemplary embodiments, when the virtual machine wants to access the material, the access time copy (COA) technique is used to copy the remaining metadata corresponding to the access operation to perform the access operation. If the source virtual machine or the target virtual machine writes data to the physical memory, copy-on-write (COW) technology is used to copy the data stored in the corresponding address accessed in the physical memory. In the system virtualization architecture, the Virtual Address is the address that the program uses to access data and instructions. The virtual address contains a segment and an offset field. This section information is used to determine this The protection information of the section and the starting address. The operating system typically uses Flat Segmentation, where all segments are mapped to the entire physical address space. According to the smooth segmentation, the virtual address actually becomes a Linear Address. If paging is enabled, the processor address is converted to a physical address using Processor Paging Hardware. In order to use paging, the operating system builds and manages a set of page tables (Set). The address translation will use these page tables as well as the various bit fields in the linear address. Please refer to FIG. 1A, which schematically illustrates a new page-table structure diagram introduced herein. Under the system virtualization architecture, the virtualized operating system is called a "guest." To properly virtualize and isolate guests, a hypervisor or virtual machine monitor (VMM) can control or regulate all privileged operations performed by the guest. Under the control of the memory management hardware in the virtual machine management program or virtual machine monitor, this new page table structure is used to define the mapping relationship between the guest entity address and the host entity address. In this page table structure, the visitor has full control over his own visitor IA-32 page table 102. The guest linear address is mapped to the guest entity address by using Control Register 3 (CR3) under the control of the visitor. In one embodiment, the CR3 is often referred to as a Page Directory Base Register (PDBR). The aforementioned scratchpad contains the physical address of the page, which can be a Page Directory. If paging is initiated and CR3 is set to an invalid or uninitialized page directory, then this machine is likely to suffer an unrecoverable failure because all memory reference information will be removed from the machine.

Then, according to the control of the memory management hardware circuit, the guest entity address is mapped to the host entity address through the multi-level page table 104 and the base pointer (Base Pointer). The multi-level page table 104 can be a B-tree like structure and can have multiple levels to translate guest addresses into physical addresses. In a virtualization platform, central management units (CPUs), memory management units (MMUs), and input/output (I/O) devices create some overhead. Some of the latest x86 processors (for example, AMD® and Intel®) are beginning to offer hardware extensions to illustrate how to solve this performance problem. The two vendors introduced support for the first generation of hardware for x86 virtualization through AMD-Virtualization ^TM (AMD-V ^TM ) and Intel® VT-x technology. Intel introduced its hardware support with MMU virtualization, called the Extended Page Table (EPT). In one embodiment, whether it is AMD-Virtualization ^TM (AMD-V) ^TM ) technology is also Intel® VT-x technology, or other technologies that use hardware extensions to define the mapping between guest entity addresses and host entity addresses through a four-level page table.

1B and FIG. 1C, which illustrates an embodiment of establishing a mapping between a guest entity address and a host entity address, in an embodiment of hierarchical storage management (Hierarchiecal Levels), for example, using four levels. Page table, but not limited to this. In this embodiment, the linear or virtual address 106 includes a number of Segment fields and an Offset field, for example, from the 12th bit to the 47th bit in the 64 address. Table offset fields from level one to level four. If paging is enabled, the processor will be available The page hardware converts the linear address 106 into a physical address. To use paging, the operating system builds and manages Set of Page Tables, for example, the fourth level to first level page tables 110, 112, 114, and 116 as shown. Base Pointer 108 is used for address translation. FIG. 1C illustrates a High Level algorithm for address translation. The content of the different levels of page table offset fields is converted separately by mapping to different levels of page tables (eg, fourth level to first level page tables 110, 112, 114, and 116). In tiered storage management, storage capacity can reach 512 million bytes (Giga Byte, GB). In one of a number of exemplary embodiments, it is proposed to quickly copy virtual machine memory state by utilizing a higher level algorithm for address translation in a hierarchical storage management structure. In the method of copying the state of the virtual machine memory, metadata about the highest level or higher level page table is copied, and the copied metadata is composed of two virtual machines (for example, the source virtual machine and the target virtual machine) ) Sharing at the same time. The target virtual machine is a virtual machine that is replicated from the source virtual machine. In one of the various exemplary embodiments, the highest level (fourth level) is employed in the method of copying the state of the virtual machine memory, but the disclosure is not limited thereto. For copying virtual machines, it is possible to adopt higher levels, for example, third and fourth, or second to fourth. In the embodiment, the metadata about the plurality of page tables in the highest level (fourth level) is copied and set to "invalid". The time spent on such replication is almost fixed. Access-time copying (COA) and/or copy-on-write (COW) are used for later operations. When the source virtual machine or the target virtual machine accesses the material, the access-time copy (COA) technology is used to copy other levels of remaining metadata for this access operation. If source The virtual machine or the target virtual machine writes the data to the storage space of the accessed address in the physical memory, and then uses the copy-on-write when storing the data in the corresponding accessed address in the physical memory ( COW) technology to replicate data. Please refer to FIG. 2, which is a flow diagram illustrating a method of copying virtual machine memory state in one of the exemplary embodiments. In the exemplary embodiment, the highest level is employed for copying, but the disclosure is not limited thereto. In step S202, the metadata in the higher level is copied. The mapping between the guest entity address and the host entity address is defined by four levels of page tables. For convenience of explanation, the four meta-data layers L4 to L1 are regarded as corresponding metadata in different levels, for example, the metadata layer L4 represents the highest level. In step S204, in response to the operation of the access, a copy-on-access (COA) technique is implemented to copy the remaining metadata. In step S206, copy-on-write (COW) technology is used to copy the data stored in the corresponding address accessed in the physical memory. The foregoing structure is similar to Cloning On-demand, in which if a virtual machine attempts to access data, an access operation is required, and then the corresponding one or more access paths are established by the following two methods. Include, in response to the operation of the access, copying the remaining metadata using copy-on-access (COA); and/or copying the corresponding address accessed in the physical memory using copy-on-write (COW) Stored information. Please refer to FIG. 3, which illustrates a relationship between a source virtual machine and a physical memory in a B-tree memory management algorithm. The mapping between the guest entity address and the host entity address is defined by using four levels of page tables. In the metadata layer 340, the four meta-data layers L4 to L1 are respectively regarded as corresponding elements of the page table in different levels. material. In blocks 341, 342, 343, and 344, the relationships between page tables in different page table directories (PTDs) are respectively indicated. Different access paths may be used to correlate the visited guest entity address with the host entity address in the physical memory. For example, in the source virtual machine 320, the guest virtual address for the access operation is mapped to the guest entity address by the guest page table 322. Corresponding to the guest entity address, an access path is established from blocks 344, 343, 342, and 341 between the guest entity address and a host entity address corresponding to a different page in the physical memory 360. .

Please refer to FIG. 4A and FIG. 4B , which are schematic diagrams showing a method for copying a state of a virtual machine memory without a counter, one of a plurality of exemplary embodiments of the present disclosure. In the described embodiment, if only one virtual machine needs to be copied to the source virtual machine, then it is not necessary to use a counter to record the common state of each page table entry (PTE) in the page table.

As shown in FIG. 4A, the source virtual machine 320 is copied and the target virtual machine 330 is generated accordingly. The source virtual machine 320 has a guest page table 322, and the target virtual machine 330 also has a guest page table 332 copied from the guest page table 322. In this exemplary embodiment, the highest level (fourth level) is employed in the method of copying the state of the virtual machine memory, but the disclosure is not limited thereto. For copying virtual machines, it is possible to adopt several higher levels, for example, third and fourth, or second to fourth. In this embodiment, the metadata of the page table in the highest level (fourth level) will be copied and then set to "invalid", for example, copying the metadata of the page table 402 of the source virtual machine and acting as the target virtual machine. The metadata of the highest level (fourth level) page table 404. And the page table directory (PTD) in the highest level (fourth level) Set to invalid.

In this page table data structure, for the page table 404 in the target virtual machine 330, two page table directories are also created, the two page table directories being in the page table 402 in the fourth level of the source virtual machine 320. Correlated between page tables 406 and 408 in the next lower level (third level). The page table directory in the lower level (third level to first level) remains unchanged, that is, the lower level metadata remains unchanged. If a guest virtual address is converted to the host entity address of page 362 in physical memory 360, then an access path will be established in page table 406 from the third level, the page table in the second level. 410. The host entity address of the page table 412 in the first level to the page 362 in the physical memory 360.

As shown in FIG. 4B, if a guest virtual address for an access operation is mapped to a guest entity address through the guest page table 322, then corresponding to the guest entity address, the guest entity address corresponds to the physical memory. An access path is established between the host entity addresses of page 362 in volume 360. Based on this access path, the page table for the page table directory along the access path in the lower level will be restored first. That is, the page table 406a in the third level, the page table 410a in the second level, and the page table 412a in the first level are restored. Other page tables that are not related to the access path remain in an "invalid" state. Based on the access path to page 362 of physical memory 360, the index (Flags) of the page table entry (PTE) accessed in page table 412a is marked as read-only, which indicates that copy-on-write (COW) is ready. operating. If a write operation is required, copy-on-write (COW) techniques are used to copy the corresponding address accessed in page 362 in physical memory 360. Information stored.

5A and 5B illustrate a method of copying a state of a virtual machine memory using a counter, one of a plurality of exemplary embodiments of the present disclosure. 5A and 5B use the same reference numerals in FIGS. 4A and 4B to denote the same elements or structures.

As shown in FIG. 5A, the source virtual machine 320 is copied and the target virtual machine 330 is generated accordingly. The source virtual machine 320 has a guest page table 322, and the target virtual machine 330 also has a guest page table 332 copied from the guest page table 322. In this exemplary embodiment, the highest level (fourth level) is employed in the method of copying the state of the virtual machine memory, but the disclosure is not limited thereto. For replicating virtual machines, it is possible to deploy higher levels, such as third and fourth, or second to fourth. In this embodiment, the metadata in the highest level (fourth level) is copied and then marked as "read only". The page table directory (PTD) in the highest level (fourth level) is also set to Read Only.

The counter is used to count the common state of each page table in different levels. If the number corresponding to a page table is "1", then this means that only one link (Link) or one path (Path) to the page table is established. The page table does not have to be copied unless a write operation (ie, copy-on-write, COW) is performed. If two or more links or paths to the page table are established, then the number corresponding to the page table in the next write operation will be greater than "1", which means that the page table is A copy operation is necessary when performing a write operation.

If there is a page from the source virtual machine 320 or by the target virtual machine Write, then the number of the counter corresponding to the page table of the access path will be incremented by one, so for example, it becomes "2". The page table will be copied in the write operation and a new page table will be created to restore the page table as needed (On Demand). In this page table data structure, for the page table 504 in the target virtual machine 330, two page table directories are also created, the two page table directories being in the page table 502 in the fourth level of the source virtual machine 320. There is an association between page tables 506 and 508 in the next lower level (third level). The page table directory in the lower level (third level to first level) remains unchanged, ie, the lower level metadata remains unchanged. In the next write operation of the page 362 in the physical memory 360, the number of counters of the page table 506 becomes "1". As shown in FIG. 5B, from the page table 506 in the third level, the page table 510 in the second level, the page table 512 in the first level, to the host entity address of the page 362 in the physical memory 360. An access path. Along this access path, page tables 506, 510, and 512 will be restored. Other page tables that are not related to the access path remain "read only". Based on the access path to page 362 of physical memory 360, the flags of the page table entries (PTEs) accessed in page tables 504, 506a, 510a, and 512a of the access path are labeled as Read only, this means that you are ready for copy-on-write (COW) operations.

The use of the numbers "1" or "2" in Figures 5A and 5B is only for explaining the common state of each page. This counter can also be configured in the page table data structure or elsewhere. For example, in the Extended Page Table (EPT) structure of Intel® VT-x technology, some reserved bits are available for the counter. If you create counters in other locations, you can use a Hash Table to associate with each page table. In some other implementations In the example, the entire page table can also be completely unshared.

For the case of copying multiple virtual machines, please refer to FIG. 6A and FIG. 6B, which illustrate one of the exemplary embodiments of the present disclosure, and the method for copying the virtual machine memory state of the virtual machine in the presence of a counter. .

As shown in FIG. 6A, the source virtual machine 320 will be replicated and the first target virtual machine 330a and the second target virtual machine 330b will be generated accordingly. The source virtual machine 320 has a guest page table 322, while the first target virtual machine 330a and the second target virtual machine 330b also have their respective guest page tables 332a and 332b copied from the guest page table 322. In this exemplary embodiment, the highest level (fourth level) is employed in the method of copying the state of the virtual machine memory, but the disclosure is not limited thereto. For replicating virtual machines, it is possible to deploy higher levels, such as third and fourth, or second to fourth. In the described embodiment, the metadata in the higher level (fourth level) is copied and then marked as "read only".

As shown in FIGS. 5A and 5B, the counter is used to count the shared state of each page table in different levels. If one of the pages is written by the first target virtual machine 330a, then the page table will be copied in the write operation and a new page table will be created as needed to restore the page table. As shown in FIG. 6A, if the first target virtual machine 330a accesses the page 362 in the physical memory 360, the access path will be established in the page table 608 from the third level, the page table 610 in the second level, The page table 612 in the first level is the host entity address of the page 362 in the physical memory 360. Page tables 606, 610, and 612 along this access path will be restored. Other pages not related to the access path The table remains "read only". Based on the access path to page 362 of physical memory 360, the flags of the page table entries (PTEs) accessed in page tables 604, 608a, 610a, and 612a of the access path are marked as read-only. This will be ready for copy-on-write (COW) operations.

In another case, as shown in FIG. 6B, if the second target virtual machine 330b accesses the page 362 in the physical memory 360, then the other access path will be established in the page table 606 from the first level, The page table 608a in the third level, the page table 610a in the second level, the page table 612a in the first level, and the host entity address of the page 362 in the physical memory 360. The page tables 608a, 610a, and 612a along this access path are restored to the page tables 608b, 610b, and 612b, respectively. Other page tables that are not related to the access path remain in an "invalid" state. Based on the access path to page 362 of physical memory 360, the flags of the page table entries (PTEs) accessed in page tables 604, 608b, 610b, and 612b of the access path are marked as read-only, This will indicate that a copy-on-write (COW) operation is ready.

The present disclosure provides a method of quickly copying data stored in a memory of a virtual machine (VM). One of the various exemplary embodiments includes providing a method of copying metadata by means of copy-on-access (COA) and copying data via copy-on-write (COW) to quickly copy a virtual machine. In one of the various exemplary embodiments, when the virtual machine wants to access the data, the access time copy (COA) technique is used to copy the remaining metadata to perform the access operation. If the source virtual machine or the target virtual machine writes data to the physical memory, copy-on-write (COW) technology is used to copy the data stored in the corresponding address accessed in the physical memory.

The method of rapid replication in a virtual memory data storage machine (VM) may be applied in different virtualization configuration, ^{e.g., AMD-Virtualization TM (AMD-} V TM) and Intel® VT-x technology. In the Intel® VT-x technology, an extended page table (EPT) structure is proposed. In ^{AMD-Virtualization TM (AMD-V} TM) technology, it is proposed nested tab (Nested Paging) structure. The Extended Page Table (EPT) is Intel's second-generation x86 virtualization technology for memory management units (MMUs). When this feature is available, the normal IA-32 page table (referenced by control register CR3) is converted from a linear address to a guest entity address. A separate page table (EPT table) set is converted from the guest entity address to the host entity address used to access the memory. The result will be to allow the guest software to modify its own IA-32 page table and handle page faults directly. This will allow the Virtual Machine Monitor (VMM) to avoid virtual machine exits associated with page table virtualization, which would be the primary source of virtualization overhead when there is no EPT.

The different uses of the present disclosure can be implemented in the form of a fully hardware embodiment, a fully software embodiment, or an embodiment containing both hardware and software components. In a preferred embodiment, the present disclosure is implemented in software, including but not limited to firmware, resident software, microcode, and the like.

In addition, the different uses of the present disclosure may be in the form of a computer program product accessible from a computer-usable or computer-readable medium, providing a program for use by a computer or any instruction execution system or in conjunction with a computer or any instruction execution system. code. In accordance with the present disclosure, a computer-usable or computer-readable medium can be used, stored, communicated, propagated, or transmitted for use by or in connection with an instruction execution system, apparatus, or device. Any tangible device of the program.

The medium can be an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system (or device, or device) or a propagation medium. Examples of computer readable media include semiconductor or solid state memory, magnetic tape, removable computer magnetic disk, random access memory (RAM), read-only memory (ROM), hard magnetic disk Or a disc. Current examples of optical discs include compact disk-read only memory (CD-ROM), compact disk-read/write (CD-R/W), and digital versatile disc (DVD) ).

Any combination of one or more computer readable media may be utilized for different uses of the present disclosure. The computer readable medium can be a computer readable signal medium or a computer readable storage medium. For example, a computer readable storage medium can be, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any suitable combination of the foregoing systems, devices, or devices. More specific examples (non-exhaustive list) of computer readable storage media include the following: electrical connections with one or more wires, portable computer diskettes, hard disks, random access memory (RAM), only Read memory (ROM), erasable programmable read-only memory (EPROM; or flash memory), optical fiber, portable optical disk read-only memory (CD-ROM), optical storage A device, a magnetic storage device, or any suitable combination of the foregoing. In the context of the present disclosure, a computer readable storage medium may be any tangible medium that can contain or store a program for use by or in connection with an instruction execution system, apparatus, or device.

The computer readable signal medium can include a propagated data signal having computer readable code embodied therein (e.g., in a baseband or as part of a carrier). Such propagated signals can take any of a variety of forms including, but not limited to, electromagnetic, optical, or any suitable combination thereof. A computer readable signal medium may be any computer readable medium that is not a computer readable storage medium but that can convey, propagate, or transport a program for use by or in connection with an instruction execution system, apparatus, or device.

The code embodied on a computer readable medium can be transmitted using any suitable medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

In various uses of the present disclosure, computer program code for performing the operations of several aspects of the present disclosure may be written in any combination of one or more programming languages, including, for example, Java, Smalltalk, C++. Or object-oriented programming language of or similar and a conventional programming language such as a "C" programming language or a similar programming language. The code may be executed entirely on the user's computer, partly on the user's computer, as a stand-alone software package, partly on the user's computer and partly on the remote computer, or entirely at Executed on a remote computer or server. In the case of full remote computer or server execution, the remote computer can be connected to the user's computer through any type of network (including local area network (LAN) or wide area network (WAN)), or can be connected to an external computer (For example, using an Internet service provider to connect via the Internet).

The various aspects of the present disclosure are described with reference to flowchart illustrations and/or block diagrams of the method, apparatus (system), and computer program product of the disclosed embodiments. It will be understood that each block of the flowchart illustrations and/or FIG. The computer program instructions can be provided to a processor of a general purpose computer, a special purpose computer, or other programmable data processing device to produce a machine such that instructions executed by the processor of the computer or other programmable data processing device are generated A means for implementing the functions/acts specified in one or more of the blocks of the flowcharts and/or block diagrams.

The computer program instructions can also be stored in a computer readable medium that can direct a computer, other programmable data processing device, or other device to function in a particular manner, such that the instructions stored on the computer readable medium produce An article of manufacture that implements the instructions of the functions/acts specified in one or more of the blocks of the flowcharts and/or block diagrams.

The computer program instructions can also be loaded onto a computer, other programmable data processing device, or other device to cause a series of operational steps to be performed on the computer, other programmable device, or other device to effect computer implementation. The processes that are executed on the computer or other programmable device provide a process for implementing the functions/acts specified in one or more of the blocks of the flowcharts and/or block diagrams. The flowchart of the present disclosure is intended to illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products in accordance with various embodiments of the present disclosure. Thus, each block in the flowchart can represent a module, segment or portion of the code, and its package One or more executable instructions for implementing the specified logic functions. It should also be noted that in some alternative implementations, the functions noted in the blocks may occur out of the order noted in the figures. For example, two blocks shown in succession may be executed substantially concurrently, or the blocks may sometimes be performed in the reverse order, depending on the functionality involved. It should also be noted that each block and/or flowchart description in the block and/or flowchart illustrations can be implemented by a special hardware-based system or a combination of dedicated hardware and computer instructions that perform the specified function or action. The combination of blocks. The terminology used herein is for the purpose of describing the particular embodiments, As used herein, the singular and " It is to be understood that the phrase "comprise" or "an" or "an" , steps, operations, components, components, and/or groups thereof. A data processing system suitable for storing and/or executing code includes at least one processor coupled directly or indirectly to a memory component through a system bus. The memory component can include a local register, a large storage device, and a cache memory applied during actual execution of the code, the cache memory providing temporary storage of at least some code to reduce The number of accesses must be retrieved from a large storage device during execution.

The present disclosure has been disclosed in the above preferred embodiments. However, it is not intended to limit the scope of the disclosure, and the invention may be modified and modified without departing from the spirit and scope of the disclosure. The scope of protection is subject to the definition of the scope of the patent application.

102‧‧‧ Visitors IA-32 page

104‧‧‧Multi-level page table

106‧‧‧Linear or virtual address

108‧‧‧Base Indicators (Base Pointer)

110, 112, 114 and 116‧‧‧ pages

Blocks 341, 342, 343 and 344‧‧

320‧‧‧ source virtual machine

322, 332, 332a, 332b‧‧‧ visitor page

330, 330a, 330b‧‧‧ target virtual machine

340‧‧‧ metadata layer

360‧‧‧ physical memory

362‧‧‧

402, 404, 406, 408, 410, 412‧‧‧ pages

406a, 410a, 412a‧‧‧ page

411‧‧‧ Path

502, 504, 506, 508, 510, 512‧‧‧ pages

506a, 510a, 512a‧‧‧ page

602, 604, 606, 610, 612‧‧‧ pages

606a, 610a, 612a‧‧‧ page

606b, 610b, 612b‧‧‧ page

FIG. 1A is a schematic diagram showing a new page table structure of one of a plurality of exemplary embodiments.

FIG. 1B and FIG. 1C are diagrams showing an embodiment of establishing a mapping between a guest entity address and a host entity address through a four-level page table.

2 is a flow chart showing a method of copying a state of a virtual machine memory in one of a plurality of exemplary embodiments.

FIG. 3 is a schematic diagram showing the relationship between a source virtual machine and a physical memory in a B-tree memory management algorithm.

4A and 4B are schematic diagrams showing a method of copying a virtual machine memory state without a counter in one of the exemplary embodiments of the present disclosure.

5A and 5B are schematic diagrams showing a method of copying a state of a virtual machine memory in the presence of a counter, in one of various exemplary embodiments of the present disclosure.

6A and 6B are diagrams illustrating a method of copying a virtual machine memory state of a virtual machine in the presence of a counter, in one of various exemplary embodiments of the present disclosure.

320‧‧‧ source virtual machine

322, 332‧‧‧ visitor page

330‧‧‧Target virtual machine

360‧‧‧ physical memory

362‧‧‧

402, 404, 406, 408, 410, 412‧‧‧ pages

406a, 410a, 412a‧‧‧ page

411‧‧‧ Path

Claims (28)

A method of copying virtual machine memory data, suitable for accessing a memory material by a source virtual machine (VM) and at least one copied virtual machine, wherein the virtual machine or the copied virtual machine is from the source virtual machine The mapping relationship between a guest entity address and a host entity address of the memory is defined by a plurality of page tables configured in a plurality of hierarchical levels, and the method for copying virtual machine memory data includes: Metadata of the page tables corresponding to the highest level or higher level of the hierarchical levels are copied to the virtual machine; according to the operation of the access, the highest level or the higher level of the hierarchical levels The page tables in the file, copying the metadata of the operation to be accessed to the virtual machine; and accessing the metadata stored in the memory, the metadata and the metadata copied according to the operation of the access The data in the corresponding address. The method for copying virtual machine memory data according to claim 1, wherein after copying the metadata of the page tables corresponding to the highest level or higher level of the hierarchical levels, The method further includes: setting a plurality of page table directories corresponding to the highest level or the higher level to be invalid. A method of copying virtual machine memory data as described in claim 1 wherein the highest level or the or higher levels of the or Copying the remaining metadata of the page tables in the levels to the virtual machine includes: corresponding to the access operation, restoring the highest level or the higher level or the higher level The page tables in the level. The method for copying virtual machine memory data according to claim 1, wherein accessing the data stored in the corresponding address of the memory comprises: according to the metadata and the operation according to the access Copying the meta-data to establish one or more access paths for accessing the corresponding address stored in the memory according to the meta-data and the meta-data copied according to the access operation The information. The method for copying virtual machine memory data according to claim 1, wherein accessing the data stored in the corresponding address of the memory comprises: writing the data according to copy-on-write (COW) technology The corresponding address to the memory. The method for copying virtual machine memory data according to claim 5, wherein the metadata is written according to the copy-on-write (COW) technology before the corresponding address of the memory is written. The metadata copied according to the operation of the access is set to read only. A method of copying virtual machine memory data as described in claim 1 wherein the number of the plurality of hierarchical levels is four. The method for copying virtual machine memory data as described in claim 1, wherein the guest entity address is compatible with an extended page table (EPT) structure, and the page tables are at a plurality of hierarchical levels. Configured to translate the guest entity address. The method for copying virtual machine memory data as described in claim 1, wherein the guest entity address is compatible with a Nested Paging structure, and the page tables are in the plurality of layers The level is configured to translate the guest entity address. The method of copying virtual machine memory data as described in claim 1, wherein the remaining metadata of the page tables in the level or levels other than the highest level or the higher level are copied. Previously, a common state count for each of the page tables configured in the one or the other levels other than the highest level or the higher levels, and copying each of the page tables based on the result of the counting Remaining metadata. The method of copying virtual machine memory data as described in claim 10, wherein the result of each of the counts in each of the page tables is stored in a corresponding page table of the count. A method of copying virtual machine memory data as described in claim 10, wherein the result of each of the counts in each of the page tables is stored in a hash table. Computer program product stored in computer accessible media, package A computer readable set of instructions for implementing the method of copying virtual machine memory data in accordance with claim 1 of the patent application scope in one or more computer systems. A computer system comprising: a host comprising: a busbar system; a memory module coupled to the busbar system, comprising a computer executable instruction set; and a processing unit coupled to the busbar system, wherein The processing unit executes the set of computer executable instructions for implementing the method of copying virtual machine memory data according to claim 1 of the scope of the patent application. A method of copying virtual machine memory data, suitable for accessing a memory material by a source virtual machine (VM) and at least one copied virtual machine, wherein the virtual machine or the copied virtual machine is from the source virtual machine The mapping relationship between a guest entity address and a host entity address of the memory is defined by a plurality of page tables configured in a plurality of hierarchical levels, and the method for copying virtual machine memory data includes: Metadata of the page tables corresponding to the highest level of the hierarchical levels are copied to the virtual machine; according to the operation of accessing, the page tables in the levels other than the highest level are correspondingly accessed The metadata of the operation is copied to the virtual machine; Accessing the metadata stored in the memory, the highest level corresponds to the metadata of the page tables and the corresponding addresses of the meta-data copied according to the operation of the access. The method for copying virtual machine memory data according to claim 15, wherein after the copying the metadata according to the highest level corresponding to the highest level, the method further comprises: corresponding to the highest Multiple page table directories for the level are set to invalid. The method for copying virtual machine memory data according to claim 15, wherein copying the remaining metadata of the page tables in the levels other than the highest level to the virtual machine comprises: corresponding access Operation, On demand, restores those page tables in those levels beyond the highest level. The method for copying virtual machine memory data according to claim 15, wherein accessing the data stored in the corresponding address of the memory comprises: according to the metadata and the operation according to the access Copying the meta-data to establish one or more access paths for accessing the corresponding address stored in the memory according to the meta-data and the meta-data copied according to the access operation The information. The method for copying virtual machine memory data according to claim 15, wherein the data stored in the corresponding address of the memory is accessed. The method includes: writing the data to the corresponding address of the memory according to a copy-on-write (COW) technique. The method for copying virtual machine memory data according to claim 19, wherein the metadata is written according to the copy-on-write (COW) technology before the corresponding address of the memory is written. The metadata copied according to the operation of the access is set to read only. A method of copying virtual machine memory data as described in claim 15 wherein the number of the plurality of hierarchical levels is four. A method for copying virtual machine memory data as described in claim 15 wherein the guest entity address is compatible with an extended page table (EPT) structure, and the page tables are at a plurality of hierarchical levels. Configured to translate the guest entity address. A method of copying virtual machine memory data as described in claim 15 wherein the guest entity address is compatible with a Nested Paging structure, and the page tables are in the plurality of layers The level is configured to translate the guest entity address. A method for copying virtual machine memory data as described in claim 15 wherein the copy of the remaining metadata of the page tables in the levels other than the highest level is prior to the highest level The shared state count for each of the page tables configured in the levels, and the result based on the count To copy the remaining metadata of each page table. A method of copying virtual machine memory data as described in claim 24, wherein the result of each of the counts in each of the page tables is stored in a corresponding page table of the count. A method of copying virtual machine memory data as described in claim 24, wherein the result of each of the counts in each of the page tables is stored in a hash table. A computer program product stored in a computer-accessible medium, comprising a computer-readable program set for implementing a method of copying virtual machine memory data according to claim 15 in one or more computer systems . A computer system comprising: a host comprising: a busbar system; a memory module coupled to the busbar system, comprising a computer executable instruction set; and a processing unit coupled to the busbar system, wherein The processing unit executes the set of computer executable instructions for implementing the method of copying virtual machine memory data according to claim 15 of the scope of the patent application.