JP4743529B2 - OS switching system, virtual machine system, OS switching method, and OS switching program - Google Patents

Description

  The present invention relates to an OS switching system, a virtual machine system, an OS switching method, and an OS switching program.

The virtual computer system is a technology that physically shows one computer to a plurality of computers. In other words, the virtual computer system allows a user to have multiple computers by simultaneously executing multiple OSs (eg, Windows (registered trademark), Linux (registered trademark) ) on one computer. It is a technique that makes it appear to exist.

  FIG. 1 is a configuration diagram showing an overview of a virtual machine system. The virtual machine system 10a includes a host OS 200a, a guest OS 300a, and hardware 100a. The host OS 200a is a program that controls the hardware 100a. On the host OS 200a, host OS applications 201-1 to 201-m (m is a natural number) that are applications of the host OS 200a operate. On the guest OS 300a, guest OS applications 301-1 to 301-n (n is a natural number) that are applications of the guest OS operate. The guest OS 300a is a kind of application of the host OS 200a. These are controlled by a CPU (not shown). The host OS 200a allocates CPU resources to the host OS applications 201-1 to 201-m and the guest OS 300a in a time division manner. The host OS applications 201-1 to 201-m and the guest OS 300a advance their processes using the allocated CPU time. The guest OS 300a has guest OS applications 301-1 to 301-n managed by the guest OS 300a itself. Then, the CPU time allocated from the host OS 300a is further divided and allocated to the guest OS applications 301-1 to 301-n. As described above, the host OS applications 201-1 to 201-m, the guest OS 300a, and the guest OS applications 301-1 to 301-n can be executed by assigning CPU resources to each application in a time division manner.

  Hereinafter, a mechanism for switching between the guest OS 300a and the host OS 200a will be described first, and then an input / output mechanism for the guest OS 300a and the host OS 200a will be described.

  First, a mechanism for switching between the guest OS 300a and the host OS 200a will be described. The guest OS 300a operates on the host OS 200a. In other words, the guest OS 300a can be considered as one of the host OS applications. In order to allocate hardware resources to applications in a time-sharing manner, it is necessary to replace programs that occupy hardware resources, that is, to switch applications. The host OS applications 201-1 to 201-m and the guest OS 300a are switched using a mechanism called context switching of the host OS 200a.

  FIG. 2 is a diagram showing a context switch of the host OS. Only one program can be executed on the CPU at any given moment. Therefore, in order to execute a plurality of programs in parallel, a CPU is given to each program in a time division manner. Various interrupt signals such as a timer interrupt give an opportunity to change the program being executed on the CPU. The CPU that has received the interrupt signal calls and executes an interrupt signal handler registered by the host OS. In the illustrated example, a timer is set first. Then, the host OS application 201-1 is executed. When the time set by the timer has elapsed, the timer sends a timer interrupt signal to the CPU. The CPU that has received the interrupt signal calls and executes a previously registered interrupt signal handler. The interrupt signal handler stores state information when the program is interrupted, such as register information and program counter of the host OS application 201-1, in a predetermined memory. Then, the state information (example: register information, program counter) of the guest OS 300a is read from a predetermined memory and set. Thereby, the guest OS 300a is executed. Thus, the host OS 200a can operate a plurality of programs in parallel by repeating the context switch.

  A normal context switch uses a timer interrupt signal. However, the context switch from the guest OS 300a to the host OS 200a uses an input / output interrupt signal to the host OS 200a. FIG. 3 is a diagram for explaining a mechanism of context switching from the guest OS 300a to the host OS 200a. Similar to the host OS 200a, the guest OS 300a also performs a context switch by timer interruption in order to execute a plurality of applications in parallel. One interrupt signal handler can be registered for one interrupt signal. The guest OS 300a uses the timer interrupt to perform context switching of the guest OS applications 301-1 to 301-n. Therefore, to return to the host OS 200a, it is necessary to execute an interrupt signal other than the timer interrupt. In other words, it is necessary to write a switching instruction to the host OS 200a in an interrupt signal handler other than the timer interrupt and execute it. That is an input / output interrupt to the host OS 200a. An input / output interrupt is a notification of processing completion from a device. Raise the hard disk as an example. The CPU requests the hard disk to write from a certain area to a certain area, but does not sequentially transfer data existing on the memory to the disk. When only the area is designated, the auxiliary arithmetic unit transfers the data on the memory to the disk for writing instead of the CPU. While the auxiliary arithmetic unit performs writing, the CPU performs another process. When the writing to the disk is completed, the auxiliary arithmetic unit transmits an input / output interrupt signal, and the CPU that receives the input / output interrupt signal notifies the application that the writing to the disk is completed. When the guest OS 300a receives an input / output interrupt signal issued when an input / output request issued in advance by the host OS 200a is completed or an input / output interrupt signal generated when a keyboard, mouse, or the like is operated, The input / output interrupt signal handler is executed to switch to the host OS 200a.

Next, the input / output mechanism of the guest OS 300a and the host OS 200a will be described. The host OS 200a receives an input / output request from the host OS application 201 and causes the hardware 100a to perform processing. The reason why the host OS application 201 mediates the host OS 200a for processing in the hardware 100a is to avoid competition between the host OS applications 201. The host OS 200a organizes requests from the plurality of host OS applications 201 and causes the hardware 100a to process them in order. A guest OS application 301 is running on the guest OS 300a. The guest OS 300a receives the input / output request from the guest OS application 301 and transfers the input / output request to the host OS 200a. The guest OS 300a is a kind of host OS application. Therefore, by executing a plurality of guest OSs 300a and causing the host OS 200a to process input / output requests, the virtual computer system 10 physically has a plurality of OSs (eg, Windows (registered trademark), Linux) on one computer. (Registered trademark) ) can be executed simultaneously. Examples of such virtual machines include Non-Patent Document 1 and Non-Patent Document 2.

  As described above, the processing in the hardware 100a of the guest OS 300a depends on the host OS 200a. Management of the hardware 100a performed by the host OS 200a is roughly divided into memory management and device management. The guest OS 300a depends on the host OS 200a for memory management and device management. Hereinafter, the memory management of the virtual machine system will be described with reference to FIGS. 4 to 6, and the device management of the virtual machine system will be described with reference to FIGS.

  First, memory management of the virtual machine system will be described. FIG. 4 is a diagram for explaining memory management in the virtual machine system. The hardware 100 a includes a physical memory 101, and the physical memory 101 includes a guest OS area 102. The guest OS area 102 is a physical memory area allocated by the host OS 200a when the guest OS 300a is activated, and only this area can be seen from the guest OS 300a.

  First, memory management on the host OS 200a side will be described. The host OS 200a includes a memory acquisition unit 206, a free area management unit 207, a free area list 208, a logical address table 203, a page information table 204, and an address conversion unit 205 for memory management.

The logical address table 203 is a table that stores sets of logical addresses and physical addresses. The logical address table is created for each host (guest) OS application. For each logical address, one physical address is allocated by the host OS 200a, and the host OS application 201 uses the logical address when reading and writing the memory. For example, consider a state in which a logical address A and a physical address A are associated. The host OS application 201 executes a data write command to the logical address A. This write command is received by the host OS 200a. The host OS 200a refers to the logical address table 203, converts the logical address A to the physical address A, and writes data to the physical address A. The page information table 204 is a table for managing physical memory usage. The physical memory is managed in units of pages (example: normally, 4 pages per page in Linux (registered trademark) ). The page information table 204 includes entries that pair a page number and a busy flag. When a certain page is used, the in-use flag is set. When the page is not used, the in-use flag is lowered.

  The address conversion unit 205 refers to the logical address table 203 and converts the logical address to a physical address. The free area management unit 207 refers to the page information table 204 and has free physical memory as a free area list. The memory acquisition unit 206 acquires a free area from the free area management unit 207 when the host (guest) OS application needs a new memory, and returns free area information to the user.

  These units and tables operate as follows. When the host OS application 201 accesses a logical address, the host OS 200a converts the logical address into a physical address using the address conversion unit 205. Therefore, the physical address is not directly referenced from the host OS 200a. By associating multiple logical addresses with one physical address, all host OS applications can utilize the same amount of physical memory.

  As an example, consider a case where the logical address 1 of the host OS application 201-1 and the logical address 1 of the host OS application 201-2 are associated with the physical address 1. Data of the host OS application 201-1 is stored at the physical address 1, and the data of the host OS application 201-2 is saved in another location (example: disk). In such a case, if there is an access from the host OS application 201-2, the data of the host OS application 201-1 stored in the physical address 1 is saved to another location (example: disk), and the host The data of the OS application 201-2 is copied from another location (example: disk) to the physical address 1, and the physical address 1 is accessed by the host OS application 201-2 after the copy is completed. FIG. 5 is a diagram illustrating the relationship between the logical address table, the physical memory, and the page information table in the host OS. The in-use flag of the page information table 204 is a flag indicating whether or not the page 107 is in use.

  Next, memory management on the guest OS 300a side will be described. Referring to FIG. 4, the guest OS 300a includes a memory acquisition unit 308, a free area management unit 309, a free area list 310, a logical address table 303, a page information table 304, and an address conversion unit 305 for memory management. These correspond to the memory acquisition unit 206, the free space management unit 207, the free space list 208, the logical address table 203, the page information table 204, and the address conversion unit 205, respectively. The guest OS 300a further includes a physical address conversion table 306 and a physical address conversion unit 307 for memory management.

  Here, before explaining the memory management of the guest OS 300a, how the guest OS 300a can allocate memory from the host OS 200a will be explained. The host OS application 201 uses the memory acquisition unit 206 to request the host OS 200a to allocate a new memory. The host OS 200a that has received the request refers to the free area list 208 and allocates a page that has not been allocated to the host OS application 201 that has issued the request. The host OS 200a returns the top logical address of the allocated area to the host OS application 201. Normally, memory is allocated in this way. However, the guest OS 300a has physical memory allocated directly from the host OS 200a using a method different from the above allocation method. When the guest OS 300a requests the host OS 200a to allocate memory, the guest OS 300a refers to the free area list 208 of the host OS 200a and allocates an unallocated page to the guest OS 300a. Thereafter, in the case of the normal host OS application 201, the top logical address of the allocated area is returned, but when the request is the guest OS 300a, the “physical address” of the top of the allocated area is returned. Further, the allocated area (hereinafter referred to as the guest OS area 102) always exists on the physical memory 101. As described in the memory management of the host OS 200a, the conversion between the logical address and the physical address is a necessary mechanism when not only the physical memory but also an area other than the physical memory such as a disk is used as the memory. When the allocated area always exists on the physical memory 101, conversion between a logical address and a physical address is not necessary. Therefore, the logical address table 203 for the guest OS 300a existing on the host OS 200a is not used, and a busy flag indicating that it is in use is set only in the page information table 204 of the host OS 200a.

  The physical address conversion table 306 is a table used for converting a virtual physical address into a physical address. The logical address table 303 in the guest OS 300a stores virtual physical addresses instead of physical addresses. The virtual physical address is a physical address for the guest OS 300a. In the OS, the memory address starts from address 0 throughout the memory management, and it is often an implicit understanding that it is a serial number. The physical memory addresses assigned by the host OS 200a are not necessarily serial numbers and do not start from address 0. Therefore, the above-mentioned implicit understanding matter is followed by using a serial number integer starting from address 0, which is a virtual physical address, as the physical address of the guest OS. The virtual physical address and the physical address are associated one to one. The physical address conversion unit 307 refers to the physical address conversion table 306 and converts the virtual physical address into a physical address.

  These tables and parts operate as follows. The conversion of the logical address and the virtual physical address in the guest OS 300a is the same as the conversion of the logical address and the physical address of the host OS 200a. In the guest OS 300a, when accessing the virtual physical address, the physical address conversion unit 307 converts the virtual physical address into a physical address based on the physical address conversion table 306. By this conversion, the guest OS application 301 can access the guest OS area 102 managed by the host OS 200a. FIG. 6 is a diagram illustrating a relationship among a logical address table, a page information table, a physical address conversion table, and a guest OS area in the guest OS.

  Next, device management of the virtual machine system will be described. FIG. 7 is a diagram for explaining device management in the virtual machine system. The hardware 100a includes a DMAC (Direct Memory Access Controller) 106, a device interface 103, a device controller 104, and devices 1 to N_105.

  First, device management on the host OS 200a side will be described. The host OS 200a uses a device list 209, a device information table 210, a device input / output receiving unit 211, a device abstraction unit 212, device drivers 1 to N_213, a resource structure 215, a resource management unit 216, and an I / O for device management. A port 214 is provided.

  The device list 209 is a list of devices 105 recognized by the host OS 200a. FIG. 8 is a table showing a device list. The device list 209 is a table and includes a device number 221, a device name 223, and a device operation function 222 as entries. Referring to FIG. 7, the device information table 210 is a table storing information on the device 105 being used. FIG. 9 is a table showing a device information table. The device information table 210 includes a device number 221, a usage counter 225, a device status 210, and user information 224 as entries. An entry in the device information table 210 is created every time the device 105 is used, and does not always exist. An entry is created when the host (guest) OS application uses the device 105, and the usage counter 225 is incremented. When the user of the device 105 becomes 0, the entry corresponding to the device 105 is deleted.

  Referring to FIG. 7, the device input / output receiving unit 211 receives a device input / output request from the guest OS 300a, and transfers the device input / output to the host OS 200a. The device abstraction unit 212 abstracts the access method to each device 105. The operation method of each device 105 differs depending on the application. The device abstraction unit 212 abstracts the access method to each device 105 by associating the operation function of each device 105 with an interface such as “read” or “write”. As an example, consider a case where the write function to the device 1 is “dev1_write”. A device operation function such as “dev1_write” is described in a device driver 1_213, which will be described later, and “dev1_write” is associated with the “write” function of the device abstraction unit by registration of the device driver 1_213. When writing to the device 1, the user instructs “write to device 1” and the host OS 200 a to execute “dev1_write” associated with the device driver registration. The device drivers 1 to N_213 are programs in which device operation functions for operating the devices 1 to N_105 are described. Upon receiving the input / output request from the user, the device drivers 1 to N_213 cause the devices 1 to N_105 to execute processing through the I / O port 214. The resource structure 215 is a structure for managing an I / O port 214 described later. The resource structure 215 includes a set of a device name and an I / O port as an entry. The resource management unit 216 assigns the I / O port 214 to each device. The I / O port 214 is an address assigned to each device 1 to N_105.

  The DMAC 106 is hardware, receives a request from the CPU, receives a processing result in the devices 1 to N_105 instead of the CPU, and stores the result in an appropriate physical memory 101. The device interface 103 is a hardware circuit that divides the input / output request received from the device driver in more detail. A specific example is disk access. The device driver 213 requests an input / output command with a granularity of “write to sector”. The device interface 103 converts a command from the device driver such as “write to sector” to a more specific command such as “move the disk arm to an appropriate position” or “write data to that position”. translate. The device controller 104 is a hardware circuit that executes an instruction sequence translated by the device interface 103. The devices 1 to N_105 are main devices as devices controlled by the device controller 104. In the example of disk access in the description of the device interface 103, it is a disk arm, that is, a disk to which data is written.

  These units and tables operate as follows. First, an operation in which the device drivers 1 to N_105 are registered in the host OS 200a will be described. Registration refers to adding an entry to the device list 209 and assigning the I / O port 214 by the resource management unit 216. Accordingly, the user can access the devices 1 to N_105 through the device abstraction unit 212. Next, a process until an input / output request from the user to the device (1 to N_) 105 is processed by the device 105 is described. Upon receiving the device input / output request, the device abstraction unit 212 confirms the device information table 210, and confirms whether or not the device 105 has already been accessed and the device information table 210 has an entry for the device 105. If there is an entry, the user who issued the input / output request is added to the user information in the device information table 210. If there is no entry, a new entry for the device 105 is registered in the device information table 210. After the user information is added or the entry is added, the device 105 is accessed using the device operation function of the device list 209.

  Next, device management on the guest OS 300a side will be described. The guest OS 300a includes a device list 311, a device information table 312, a device correspondence table 313, a device abstraction unit 314, device drivers 1 to M_315, and a device input / output transmission unit 316 for device management. The device list 311, the device information table 312, and the device abstraction unit 314 are the same as the device list 209, device information table 210, and device abstraction unit 212 of the host OS 200a. The device correspondence table 313 is a table representing the correspondence between the guest OS 300a device and the host OS 200a device. FIG. 10 is a table showing a device correspondence table of the guest OS. A host OS device name 322 and a guest OS device name 323 are associated with each other. Referring to FIG. 7, device operation functions are described in device drivers 1 to M_315 in the same manner as device drivers 1 to N_213 of the host OS. However, the device operation functions of the device drivers 1 to M_315 of the guest OS 300a describe processing for transferring an input / output request to the device input / output transmission unit 316 without requesting the device to perform processing. The device input / output transmission unit 316 transfers the device input / output request to the device input / output reception unit 211 of the host OS 200a.

  These units and tables operate as follows. First, an operation in which the device drivers 1 to M_315 are registered in the host OS 200a will be described. The device registration in the guest OS 300a is only addition of an entry to the device list 311 and the resource management unit 216 does not allocate the I / O port 214. Next, a description will be given of processing until an input / output request from the user to the devices 1 to M_105 (M ≦ N) is processed by the device. The device abstraction unit 314 that has received the device input / output request confirms the device information table 312 and confirms whether or not there is an access to the device 105 and the device information table 312 has an entry for the device 105. If there is an entry, the user who issued the input / output request is added to the user information in the device information table 312. If there is no entry, a new entry for the device 105 is registered in the device information table 312. After the user information is added or the entry is added, the device input / output request is transferred to the device 105 of the host OS 200a via the device input / output transmission unit 316. When transferring a device input / output request to the host OS 200a, the device correspondence table 313 is referred to in order to clarify “to which device 105 the host OS 200a is requested”, and the “host corresponding to the device name of the device 105” "OS device name" is added to the device input / output request.

  The above is the outline of memory management and device management in the virtual machine system 10.

  As a related technique, an OS control method is disclosed in JP-A-2-297636. This OS control method is a bigy-back virtual computer system that includes a parent OS that operates on a real computer, a VM monitor that operates on the parent OS, and a child OS that operates on the parent OS via the VM monitor. When the parent OS is stopped for some reason, the child OS is prepared by preparing an environment so that it operates on the real computer under the control of the parent OS and one of the child OSs can operate on the real computer. It is characterized by having a parent / child OS switching means for transferring control to.

Dan Aloni, “Cooperative Linux”, Proceedings of the Linux Symposium, pp23-32, Ottawa Ontario Canada, July 21th-24th, 2004 Jeff Dike, “A user-mode port of the Linux kernel 1”, http: // user-mode-linux. sourceforge. net / als2000 / index. html Daniel P. Bovet, Marco Cesati. "Detailed Linux kernel" O'REILLY JAPAN, Feb, 2004 JP-A-2-297636

  The problem with the virtual machine system is that when the host OS is stopped, the guest OS must also be stopped. This is because the guest OS depends on the host OS for hardware access. As described above, access to the memory of the guest OS and access to the device are performed through the host OS. Therefore, when the host OS is stopped, the guest OS cannot access the memory and the device.

  Japanese Patent Laid-Open No. 2-297636 describes an OS control method for switching the relationship between a parent OS and a child OS. However, the specific method is not clearly described, and a specific method is described below. For example, it is unknown whether the environment can be prepared so that the child OS can operate on the real computer. For example, if the memory management information and device management information of the host OS and guest OS are simply replaced, the application used in the host OS and the application used by the guest OS are different. It is impossible.

  Accordingly, an object of the present invention is to provide an OS switching system, a virtual machine system, an OS switching method, and an OS switching program that can operate a guest OS even when a failure occurs in a host OS.

  Another object of the present invention is to provide an OS switching system, a virtual machine system, an OS switching method, and an OS switching program capable of switching the dependency relationship between the guest OS and the host OS for hardware access without restarting the guest OS. Is to provide.

  Hereinafter, means for solving the problem will be described using the numbers and symbols used in the best mode for carrying out the invention. These numbers and symbols are added in parentheses in order to clarify the correspondence between the description of the claims and the best mode for carrying out the invention. However, these numbers and symbols should not be used for interpreting the technical scope of the invention described in the claims.

  In order to solve the above problems, an OS switching system of the present invention includes a host OS (200) and a guest OS (300) operating on the host OS (200). When the guest OS (300) is switched to the host and the host OS (200) is switched to the guest, the guest OS (300) changes the guest memory management information (303, 304, 306) of the guest OS (300) and the host OS (200). ) Host memory management information (203, 204), new guest memory management information (203a, 611, 612) serving as memory management information of the host OS (200) after switching, and the guest OS (( 300), new host memory management information (303a, 304a) is generated. The guest OS (300) replaces the guest memory management information (303, 304, 306) with new host memory management information (303a, 304a). The host OS (200) replaces the host memory management information (203, 204) with new guest memory management information (203a, 611, 612). The host OS (200) and guest OS (300) switch execution instructions.

  In the OS switching system described above, the guest OS (300) is switched based on the guest device management information (311 and 313) of the guest OS (300) and the host device management information (209) of the host OS (200). New guest device management information (613, 614) serving as device management information of the host OS (200) of the host OS and new host device management information (311a) serving as device management information of the guest OS (300) after switching. The guest OS (300) replaces the guest device management information (311 and 313) with new guest device management information (613 and 614). The host OS (200) replaces the host device management information (209) with new host device management information (311a).

In the OS switching system, the guest OS (300) performs new guest memory management based on the host page information table (204) and the host logical address table (203) included in the host memory management information (203, 204). The memory information conversion unit 540 generates a new guest page information table (612), a new guest physical address conversion table (611), and a new guest logical address table (203a) included in the information (203a, 611, 612). The host page information table (204) shows the correspondence between the physical address and the usage status of the physical address. The host logical address table (203) shows the correspondence between logical addresses and physical addresses. The new guest page information table (612) indicates the correspondence between the usable physical addresses of the host OS (200) in the host page information table (204) and the usage status of the physical addresses. The new guest physical address conversion table (611) shows the correspondence between the physical address and the virtual physical address in the new guest page information table (612).
The new guest logical address table (203a) shows the correspondence between the logical addresses in the host logical address table (203) and the virtual physical addresses in the new guest physical address conversion table (611).

  In the OS switching system, the memory information conversion unit 540 further includes the host page information table (204), the guest logical address table (303) and the guest physical address included in the guest memory management information (303, 304, 306). Based on the conversion table (306), a new host logical address table (303a) and a new host page information table (304a) included in the new host memory management information (303a, 304a) are generated. The guest logical address table (303) shows the correspondence between logical addresses and virtual physical addresses in the guest OS (300). The guest physical address conversion table (306) shows the correspondence between virtual physical addresses and physical addresses in the guest OS (300). The new host logical address table (303a) shows the correspondence between the logical addresses in the guest logical address table (303) and the physical addresses in the guest physical address conversion table (306). The new host page information table (304a) is obtained by changing the host page information table (204) for the guest OS (300).

  In the OS switching system, the guest OS (300) further includes a guest device correspondence table (313) and a guest device list (311) included in the guest device management information (311 and 313), and host device management information (209). , 210) based on the host device list (209) included in the new guest device management information (613, 614), a device that generates the new guest device correspondence table (613) and the new guest device list (614) An information conversion unit 550 is provided. The guest device correspondence table (313) indicates the correspondence between devices recognized by the guest OS (300) and devices recognized by the host OS (200). The guest device list (311) indicates information on devices that can be used by the guest OS (300). The host device list (209) indicates information related to devices that can be used by the host OS (200). The new guest device list (614) is information (device name 223, device operation function 222) related to the device of the device in the guest device correspondence table (313) among the devices in the host device list (209). The guest device list (311) is updated, and information on the remaining devices of the host device list (209) is added to the updated guest device list (311). The new guest device correspondence table (613) is obtained by replacing the device recognized by the guest OS (300) and the device recognized by the host OS (200) in the guest device correspondence table (313).

  In the above OS switching system, the guest OS (300) uses the process information of the guest OS (300) process as the process information of the host OS (200), and uses the memory information of the process of the host OS (200) as a new guest. Rewrite based on the memory management information (203a, 611, 612), rewrite the register information of the host OS (200) process to the register information of the host OS (200), and register the process of the host OS (200) in the scheduling queue. A host process generation unit (570) is provided.

  A virtual computer system comprising the OS switching system (20) and hardware (100) used for the OS switching system (20).

  In order to solve the above problems, the OS switching method according to the present invention includes: (a) when the guest OS (300) is switched to the host and the host OS (200) is switched to the guest, the guest OS (300) is switched to the guest OS (300 ) And the new guest memory that becomes the memory management information of the host OS (200) after switching based on the guest memory management information (303, 304, 306) of the host OS (200) and the host memory management information (203, 204) of the host OS (200) Generating management information (203a, 611, 612) and new host memory management information (303a, 304a) as memory management information of the guest OS (300) after switching; (b) guest OS (300) Guest memory management information (303, 304, 306) is replaced with new host memory management information (303a, 304). ), (C) the host OS (200) replaces the host memory management information (203, 204) with new guest memory management information (203a, 611, 612), and (d) the host OS (200). And the guest OS (300) include a step of switching execution instructions.

  In the OS switching method described above, (e) the guest OS (300) is based on the guest device management information (311 and 313) of the guest OS (300) and the host device management information (209) of the host OS (200). New guest device management information (613, 614) that becomes device management information of the host OS (200) after switching, and new host device management information (311a) that becomes device management information of the guest OS (300) after switching Generating, (f) the guest OS (300) replacing the guest device management information (311 313) with new guest device management information (613, 614), and (g) the host OS (200) Replaced host device management information (209) with new host device management information (311a) Further comprising a step.

  In the OS switching method, the step (a) includes (a1) a new guest based on the host page information table (204) and the host logical address table (203) included in the host memory management information (203, 204). A step of generating a new guest page information table (612), a new guest physical address conversion table (611), and a new guest logical address table (203a) included in the memory management information (203a, 611, 612). The host logical address table (203) shows the correspondence between logical addresses and physical addresses. The new guest page information table (612) indicates the correspondence between the usable physical addresses of the host OS (200) in the host page information table (204) and the usage status of the physical addresses. The new guest physical address conversion table (611) shows the correspondence between the physical address and the virtual physical address in the new guest page information table (612). The new guest logical address table (203a) shows the correspondence between the logical addresses in the host logical address table (203) and the virtual physical addresses in the new guest physical address conversion table (611).

In the OS switching method, (a) step includes (a2) host page information table (204), guest logical address table (303) and guest physical address included in guest memory management information (303, 304, 306). A step of generating a new host logical address table (303a) and a new host page information table (304a) included in the new host memory management information (303a, 304a) based on the conversion table (306) is further provided. The guest logical address table (303) shows the correspondence between logical addresses and virtual physical addresses in the guest OS (300).
The guest physical address conversion table (306) shows the correspondence between virtual physical addresses and physical addresses in the guest OS (300). The new host logical address table (303a) shows the correspondence between the logical addresses in the guest logical address table (303) and the physical addresses in the guest physical address conversion table (306). The new host page information table (304a) is obtained by changing the host page information table (204) for the guest OS (300).

  In the OS switching method, the step (e) includes (e1) the guest device correspondence table (313) and guest device list (311) included in the guest device management information (311 and 313), and the host device management information (209). , 210) based on the host device list (209) included in the new guest device management information (613, 614), a new guest device correspondence table (613) and a new guest device list (614) are generated. Is provided. The guest device correspondence table (313) indicates the correspondence between devices recognized by the guest OS (300) and devices recognized by the host OS (200). The guest device list (311) indicates information on devices that can be used by the guest OS (300). The host device list (209) indicates information related to devices that can be used by the host OS (200). The new guest device list (614) is information (device name 223, device operation function 222) related to the device of the device in the guest device correspondence table (313) among the devices in the host device list (209). The guest device list (311) is updated, and information on the remaining devices of the host device list (209) is added to the updated guest device list (311). The new guest device correspondence table (613) is obtained by replacing the device recognized by the guest OS (300) and the device recognized by the host OS (200) in the guest device correspondence table (313).

  In the OS switching method, (d) step includes (d1) using the process information of the process of the guest OS (300) as the process information of the process of the host OS (200), and (d2) the host OS (200 ) Rewriting the memory information of the process based on the new guest memory management information (203a, 611, 612), and (d3) rewriting the register information of the process of the host OS (200) to the register information of the host OS (200). And (d4) registering the process of the host OS (200) in the scheduling queue.

  In order to solve the above-described problems, the program of the present invention provides: (a) When switching the guest OS (300) to the host and the host OS (200) to the guest, the guest OS (300) is the guest OS (300). New guest memory management information that becomes the memory management information of the host OS (200) after switching based on the guest memory management information (303, 304, 306) and the host memory management information (203, 204) of the host OS (200) (203a, 611, 612) and a step of generating new host memory management information (303a, 304a) as memory management information of the guest OS (300) after switching, and (b) the guest OS (300) Guest memory management information (303, 304, 306) is changed to new host memory management information (303a, 304a) (C) the host OS (200) replaces the host memory management information (203, 204) with new guest memory management information (203a, 611, 612), and (d) the host OS (200). And the guest OS (300) causes the computer to execute an OS switching method including a step of switching execution instructions.

  In the above program, (e) the guest OS (300) switches based on the guest device management information (311 and 313) of the guest OS (300) and the host device management information (209) of the host OS (200) New guest device management information (613, 614) that becomes device management information of the host OS (200) after that and new host device management information (311a) that becomes device management information of the guest OS (300) after the switching are generated. (F) the guest OS (300) replaces the guest device management information (311 313) with the new guest device management information (613, 614), and (g) the host OS (200) The step of replacing the management information (209) with the new host device management information (311a) Further comprising a flop.

  In the above program, the (a) step includes (a1) a new guest memory management based on the host page information table (204) and the host logical address table (203) included in the host memory management information (203, 204). A step of generating a new guest page information table (612), a new guest physical address conversion table (611), and a new guest logical address table (203a) included in the information (203a, 611, 612). The host logical address table (203) shows the correspondence between logical addresses and physical addresses. The new guest page information table (612) indicates the correspondence between the usable physical addresses of the host OS (200) in the host page information table (204) and the usage status of the physical addresses. The new guest physical address conversion table (611) shows the correspondence between the physical address and the virtual physical address in the new guest page information table (612). The new guest logical address table (203a) shows the correspondence between the logical addresses in the host logical address table (203) and the virtual physical addresses in the new guest physical address conversion table (611).

  In the above program, (a) step includes (a2) host page information table (204), guest logical address table (303) and guest physical address conversion table included in guest memory management information (303, 304, 306). Based on (306), a step of generating a new host logical address table (303a) and a new host page information table (304a) included in the new host memory management information (303a, 304a) is further provided. The guest logical address table (303) shows the correspondence between logical addresses and virtual physical addresses in the guest OS (300). The guest physical address conversion table (306) shows the correspondence between virtual physical addresses and physical addresses in the guest OS (300). The new host logical address table (303a) shows the correspondence between the logical addresses in the guest logical address table (303) and the physical addresses in the guest physical address conversion table (306). The new host page information table (304a) is obtained by changing the host page information table (204) for the guest OS (300).

  In the above program, (e) step includes (e1) Guest device correspondence table (313) and guest device list (311) included in guest device management information (311 and 313), and host device management information (209 and 210). ) Including a new guest device correspondence table (613) and a new guest device list (614) included in the new guest device management information (613, 614) based on the host device list (209) included in . The guest device correspondence table (313) indicates the correspondence between devices recognized by the guest OS (300) and devices recognized by the host OS (200). The guest device list (311) indicates information on devices that can be used by the guest OS (300). The host device list (209) indicates information related to devices that can be used by the host OS (200). The new guest device list (614) is information (device name 223, device operation function 222) related to the device of the device in the guest device correspondence table (313) among the devices in the host device list (209). The guest device list (311) is updated, and information on the remaining devices of the host device list (209) is added to the updated guest device list (311). The new guest device correspondence table (613) is obtained by replacing the device recognized by the guest OS (300) and the device recognized by the host OS (200) in the guest device correspondence table (313).

  In the above program, the step (d) includes the step (d1) using the process information of the process of the guest OS (300) as the process information of the process of the host OS (200), and (d2) the process information of the host OS (200). Rewriting the memory information of the process based on the new guest memory management information (203a, 611, 612); and (d3) rewriting the register information of the process of the host OS (200) to the register information of the host OS (200). (D4) registering the process of the host OS (200) in the scheduling queue.

  In the present invention, in the virtual machine system, when the guest OS (300) is the host and the host OS (200) is switched to the guest, the memory management information / device management before switching between the host OS (200) and the guest OS (300) Based on the information, memory management information / device management information to be used after switching is newly generated. Therefore, there is no inconsistency in memory management information / device management information before and after switching, and smooth switching can be performed without restarting the guest OS. Then, the dependency relationship between the guest OS and the host OS can be switched for hardware access without restarting the guest OS.

  Hereinafter, embodiments of an OS switching system, a virtual machine system, an OS switching method, and an OS switching program according to the present invention will be described with reference to the accompanying drawings.

  FIG. 11 is a schematic configuration diagram showing an embodiment of a virtual machine system to which the OS switching system of the present invention is applied. The virtual machine system 10 includes a host OS 200, a guest OS 300, and hardware 100. The host OS 200 is a program that controls the hardware 100. On the host OS 200, host OS applications 201-1 to 201-m that are applications of the host OS 200 operate. On the guest OS 300, guest OS applications 301-1 to 301-n, which are guest OS applications, operate. The guest OS 300 is a kind of application of the host OS 200. These are controlled by a CPU (not shown). The host OS 200, the guest OS 300, and the hardware 100 include the functions of the host OS 200a, the guest OS 300a, and the hardware 100a described with reference to FIGS.

  The OS switching system of the present invention switches the host OS 200 to the guest and the guest OS 300 to the host. That is, the host OS 200 is switched to the new guest OS, and the guest OS 300 is switched to the new host OS.

  FIG. 12 is a block diagram showing the configuration of the embodiment of the OS switching system of the present invention. The OS switching system 20 includes a memory information saving unit 500, a device information saving unit 510, a memory information acquisition unit 520, a device information acquisition unit 530, a memory information conversion unit 540, a device information conversion unit 550, an execution instruction on the guest OS 300 side. A switching unit 560, a host process generation unit 570, a context information acquisition unit 580, and a process information acquisition unit 590 are provided. In addition, on the host OS 200 side, a memory information saving unit 400, a device information saving unit 410, a memory information acquisition unit 420, a device information acquisition unit 430, a memory information replacement unit 440, a device information replacement unit 450, an execution instruction switching unit 460, A context information saving unit 470 and a process information acquisition unit 480 are provided. Each of these configurations can be implemented by hardware or a combination of software and hardware in addition to software. The OS switching system 20 further includes an OS state saving area 600 on the main storage device 110 of the hardware 100. The main storage device 110 is a physical memory, for example.

  The OS state saving area 600 is an area for storing the memory management information and device management information of the guest OS 300 and the host OS 200 and the context information of the host OS 200. FIG. 13 is a block diagram showing the OS state save area 600. The OS state saving area 600 is arranged on the physical memory. The OS state saving area 600 includes a guest OS state saving area 610 and a host OS state saving area 620. The guest OS state saving area 610 stores memory management information and device management information for a new guest OS created by the guest OS 300. Here, the memory management information includes a new guest OS physical address conversion table 611 and a new guest OS page information table 612. The device management information includes a new guest OS device correspondence table 613 and a new guest OS device list 614. The host OS state saving area 620 stores memory management information, device management information, and context information 227 used by the host OS 200. Here, the memory management information includes a page information table 204. The device management information includes a device information table 210 and a device list 209.

  Referring to FIG. 12, the memory information saving unit 400 is arranged in the host OS 200, and the memory information saving unit 500 is arranged in the guest OS 300. The memory information saving unit 400 stores the page information table 204 used by the host OS 200 itself in the host OS status saving area 620 in the OS status saving area 600. The memory information saving unit 500 stores the new guest OS physical address conversion table 611 and the new guest OS page information table 612 created by the guest OS 300 in the guest OS state saving area 610 in the OS state saving area 600.

  The device information saving unit 410 is arranged in the host OS 200, and the device information saving unit 510 is arranged in the guest OS 300. The device information saving unit 410 stores the device information table 210 and the device list 209 used by the host OS 200 itself in the host OS status saving area 620 in the OS status saving area 600. The device information saving unit 510 stores the new guest OS device correspondence table 613 and the new guest OS device list 614 created by the guest OS 300 in the guest OS state saving area 610.

  The memory information acquisition unit 420 is arranged in the host OS 200, and the memory information acquisition unit 520 is arranged in the guest OS 300. The memory information acquisition unit 420 acquires the new guest OS physical address conversion table 611 and the new guest OS page information table 612 created by the guest OS 300 from the guest OS state save area 610 in the OS state save area 600. The memory information acquisition unit 520 acquires the page information table 204 stored by the host OS 200 from the host OS state save area 620 in the OS state save area 600.

  The device information acquisition unit 430 is arranged in the host OS 200, and the device information acquisition unit 530 is arranged in the guest OS 300. The device information acquisition unit 430 acquires the new guest OS device correspondence table 613 and the new guest OS device list 614 created by the guest OS 300 from the guest OS state save area 610. The device information acquisition unit 530 acquires the device information table 210 and the device list 209 stored by the host OS 200 from the host OS state save area 620 in the OS state save area 600.

  The memory information conversion unit 540 is arranged in the guest OS 300. Based on the page information table 204 of the host OS 200 acquired by the memory information acquisition unit 520, the page information table 304 is rewritten to a new host OS page information table 304a. Also, the logical address table 303 is rewritten to the new host OS logical address table 303a based on the new host OS page information table 304a, the logical address table 303, and the physical address conversion table 306. At the same time, a new guest OS page information table 612 is newly generated based on the new host OS page information table 304a. In addition, a new guest OS physical address conversion table 611 is created based on the new host OS page information table 304a and the logical address table 203, and the logical address table 203 is stored in the new guest OS logical address table 203a. rewrite. The new guest OS page information table 612 corresponds to the page information table 204 used before OS switching is deleted from the page used by the guest OS 300. The new guest OS physical address conversion table 611 is created by assigning serial numbers to physical memory addresses excluding the physical memory addresses used by the guest OS 300.

  The device information conversion unit 550 is arranged in the guest OS 300. Based on the device information table 210 and device list 209 of the host OS 200 acquired by the device information acquisition unit 530 and the device correspondence table 313 and device list 311 used by the guest OS 300, the device driver of the host OS 200 is transferred to the guest OS 300. Register and create a new host OS device list 311a. Similarly, the device driver of the guest OS 300 is registered in the host OS 200, and a new guest OS device list 614 is newly created. In addition, the host-side data 322 and the guest-side data 323 in the device correspondence table 313 are exchanged, and a new guest OS device correspondence table 613 is newly generated.

  The memory information exchanging unit 440 is arranged in the host OS 200. Using the new guest OS page information table 612 and the new guest OS physical address conversion table 611 acquired by the memory information acquisition unit 520 of the host OS 200, the physical address 218 described in the logical address table 203 of the host OS 200 is changed to the virtual physical. Replace with address 318. At the same time, the page information table 204 is replaced with a page information table 612 for new guest OS.

  The device information replacement unit 450 newly registers a device driver using the new guest OS device list 614 and the new guest OS device correspondence table 613 acquired by the device information acquisition unit 430 of the host OS 200.

  The context information saving unit 470 is arranged in the host OS 200. The register information of the host OS application is stored in the host OS state saving area 620 in the OS state saving area 600. As described in the context switch in the background art, when the host OS 200 is switched to the guest OS and the guest OS 300 is switched to the host OS, the register information of the host OS application that has been operated so far and the instruction to be executed next are stored. A memory address (hereinafter referred to as a program counter) is stored in a data area managed by the host OS 200 as an application execution suspended state (context information 227). That is, the context information saving unit 470 saves the execution suspension state (context information 227) such as a program counter and register information.

  The context information acquisition unit 580 is arranged in the guest OS 300. The context information 227 stored by the context information saving unit 470 is acquired. The host process generation unit 570 is arranged in the guest OS 300. A process that is the main body of the host OS 200 is generated on the guest OS 300.

  The execution instruction switching unit 460 is disposed in the host OS 200, and the execution switching unit 560 is disposed in the guest OS 300. The execution command switching units 460 and 560 switch the memory and device operation function used by the guest OS 300 and the host OS 200 when the operations of the above-described units are completed. Unlike the normal (host OS) memory management, the guest OS 300 has a physical address conversion table 306 and a physical address conversion unit 307 as a memory operation function. Since the guest OS 300 behaves as a host OS after executing the above-described units, it is necessary to switch to the memory operation function used by the host OS 200. The same applies to devices. However, these operation functions are assumed to be possessed by both the execution instruction switching units 460 and 560, respectively.

  The process information acquisition unit 480 is arranged in the host OS 200, and the process information acquisition unit 590 is arranged in the guest OS 300. In each of the host OS 200 and the guest OS 300, process information is acquired from the process information list, and the presence / absence of the device 105 in use is confirmed. If there is a device 105 in use, the request to the device 105 is canceled and an error is returned to the process.

  Next, the operation of the embodiment of the OS switching system of the present invention (embodiment of the OS switching method) will be described. 14 to 16 are flowcharts showing the operation of the embodiment of the OS switching system of the present invention.

  FIG. 14 particularly shows the operation of the host OS 200. First, the process information acquisition unit 480 acquires process information from the list of process information (S1). The process information represents the status of a running program created for each application program 201. The process information stores information about the device 105 to which the process in the application program 201 has issued an input / output request. That is, the process information associates a process in the application program with a device that has issued an input / output request. The process information list associates application programs with process information. The OS 200 holds a list of process information. At the end of the list, that is, when the acquisition of process information for all processes is completed, the next step (S6) is executed (S2: YES). If the acquisition of process information has not been completed for all processes (S2: NO), the process information acquisition unit 480 confirms the presence or absence of the in-use device 105 from the process information (S3). If there is a device 105 in use (S4: YES), the process information acquisition unit 480 cancels the request to the device 105 and returns an error to the process (S5). If there is no device 105 in use regarding the process (S4: NO), the process information acquisition unit 480 acquires the next process information. The work so far is preparation work so that the device driver 213 can be replaced at any time by eliminating the state during device input / output.

  Next, the memory information saving unit 400 copies the page information table 204 to the host OS state saving area 620 (S6). Subsequently, the device information saving unit 410 copies the device information table 210 to the host OS state saving area 620 (S7), and copies the device list 209 to the host OS state saving area 620 (S8). Then, the context information saving unit 470 copies the context information 227 to the OS state saving area 620 (S9). Thereafter, context switching to the guest OS 300 is performed in the same manner as before (S10).

  FIG. 15 is a continuation of FIG. 14 and particularly shows the operation of the guest OS 300. First, the process information acquisition unit 590 acquires process information from the process information list (S11). The process information represents the status of a running program created for each application program 301. The process information stores information about the device 105 from which the process in the application program 301 has issued an input / output request. That is, the process information associates a process in the application program with a device that has issued an input / output request. The process information list associates application programs with process information. The OS 300 holds a list of process information. At the end of the list, that is, when the acquisition of process information for all processes is completed, the next step (S16) is executed (S12: YES). If the acquisition of process information has not been completed for all processes (S12: NO), the process information acquisition unit 590 checks the presence / absence of the in-use device 105 from the process information (S13). If there is a device 105 in use (S14: YES), the process information acquisition unit 590 cancels the request to the device 105 and returns an error to the process (S15). If there is no device 105 in use regarding the process (S14: NO), the process information acquisition unit 590 acquires the next process information. The operations so far are preparations so that the device driver 315 can be replaced at any time by eliminating the device input / output state.

  Next, the memory information acquisition unit 520 saves the page information table 204 saved in S6, and the device information acquisition unit 530 saves the device information table 210 saved in S7 and the device list 209 saved in S8, respectively. Obtained from the host OS state saving area 620 of the area 600 (S16). Based on the information of the host OS 200 (the page information table 204, the device information table 210, and the device list 209), the memory information conversion unit 540 converts its own page information table 304 into the new host OS page information table 304a. The address table 303 is rewritten to the new host OS logical address table 303a (S17). Thereafter, the memory information conversion unit 540 creates a new guest OS page information table 612 and a new guest OS physical address conversion table 611 based on the new host OS page information table 304a. Then, the memory information saving unit 500 stores the new guest OS page information table 612 and the new guest OS physical address conversion table 611 in the guest OS state saving area 610 (S18). The detailed operation of the memory information conversion unit 540 in S17 and S18 will be described later.

  Next, the device information conversion unit 550 generates a new host OS device list 311a based on the device information table 210 and device list 209 of the host OS 200 and the device correspondence table 313 and device list 311 of the guest OS 300 ( S19). Similarly, the device information conversion unit 550 registers the device driver of the guest OS 300 in the host OS 200 and newly creates a new guest OS device list 614. In addition, the host OS device name 322 on the host side and the guest OS device name 323 on the guest side in the device correspondence table 313 are exchanged, and a new guest OS device correspondence table 613 is newly generated. The device information saving unit 510 stores the new guest OS device correspondence table 613 and the new guest OS device list 614 in the guest OS state saving area 610 (S20). The detailed operation of the device information conversion unit 550 in S19 and S20 will be described later.

  Next, the host process generation unit 570 newly generates a process that becomes the host OS (S21). Details of generation of the host process will be described later. Then, the execution command switching unit 560 switches the execution command (S22). With the operations so far, the guest OS 300 is executed as a new host OS. That is, when the guest OS 300 performs the function of the original host OS 200, the guest OS 300 can operate as a new host OS. Thereafter, a context switch is finally performed to the host OS 200 (S23).

  FIG. 16 particularly shows the operation of the host OS 200 (new guest OS). First, the memory information acquisition unit 420 includes a new guest OS physical address conversion table 611 and a new guest OS page information table 612, and the device information acquisition unit 430 includes a new guest OS device correspondence table 613 and a new guest OS device list 614. Are acquired from the guest OS state saving area 610 (S24). Next, the memory information switching unit 440 adds a new guest OS physical address conversion table 611, replaces the page information table 204 with the new guest OS page information table 612, and rewrites the logical address table 203 (S25). Specifically, it refers to the work of rewriting the physical address 218 described in the logical address table 203 to a virtual physical address to make a new guest OS logical address table 203a. A new guest OS physical address conversion table 611 is used for this work. Then, the page information table 204 used so far is replaced with the newly acquired page information table 612 for new guest OS.

  Next, the device information replacement unit 450 replaces the device 105 using the information for the new guest OS (S26). Specifically, a new guest OS device correspondence table 613 is newly provided in the host OS 200, and the device list 209 is replaced with a new guest OS device list 614, and is described in the new guest OS device list 614. Refers to the process of newly registering a device driver for a device. Device drivers that are no longer used will be deregistered as a result of this work. Finally, the execution instruction switching unit 460 switches the execution instruction of the host OS 200. By the operation so far, the host OS 200 is executed as a new guest OS.

  The above is the overall operation of the embodiment of the OS switching system of the present invention.

Next, details of the operation of the memory information conversion unit 540 will be described.
FIG. 17 is a flowchart showing the operation of the memory information conversion unit in S17 and S18. FIG. 20 is an example showing a process of generating the new host OS page information table 304a in S17. FIG. 21 shows an example of the process of generating the new host OS logical address table 303a in S17. FIG. 22 is an example showing a process of generating a new guest OS page information table in S18. FIG. 23 is an example of the new guest OS physical address conversion table 611 generated in S18. FIG. 24 shows an example of the process of generating the new guest OS logical address table 203a in S18.

  Referring to FIG. 17, the memory information conversion unit 540 copies the page information table 204 acquired from the host OS 200 (S31). Referring to the page information table 204 in FIG. 20A, the page information table 204 associates the page number 219 with the in-use flag 220. As described above, the page number 219 is associated with a physical address, and in the following description, for the sake of simplicity, it is assumed that the physical address and the page number correspond to 1: 1. The in-use flag 220 indicates whether the physical memory at the physical address indicated by the corresponding page number 219 is in use or unused as described above. “Used” / “unused” (example: page number 2 / page number 3) in white indicates use / unused in the host OS 200. The color background “in use” / “unused” (example: page number 1 / page number m) indicates use / unused in the guest OS 300.

  Referring to FIG. 17, the memory information conversion unit 540 replaces “used” of the white background and “used” of the color background in the page information table 204, and “unused” of the white background and the color background. Replaced with “unused” to generate a new new host OS page information table 304a (S32). In other words, the page information secured by the host OS 200 (use permitted) is replaced with the page information secured by the guest OS 300 (use permitted). FIG. 20B shows a new new host OS page information table 304a. Then, the page information table 304 is replaced with a new host OS page information table 304a.

  17, the memory information conversion unit 540 converts the virtual physical address 318 of the logical address table 303 into the physical address (host OS 200) of the physical address conversion table 306 based on the logical address table 303 and the physical address conversion table 306. (S33). FIGS. 21A and 21B show a logical address table 303 and a physical address conversion table 306, respectively. By rewriting the virtual physical address 318 to a physical address, a new host OS logical address table 303a (FIG. 21C) is generated.

  As a result, a new host OS page information table 304a and a new host OS logical address table 303a for the new host (guest OS 300) are generated (corresponding to S17).

  Next, with reference to FIG. 17, the memory information conversion unit 540 extracts an entry for the host OS 200 (new guest OS) from the new host OS page information table 304a, and generates a new guest OS page information table 612. (S34). Referring to FIG. 22, in the new host OS page information table 304a of FIG. 22 (a), “used” / “unused” (page numbers 1, 6, 8, ..., m) are extracted and arranged. As a result, a new guest OS page information table 612 (FIG. 22B) is generated.

  Referring to FIG. 17, the memory information conversion unit 540 generates a new guest OS physical address conversion table 611 by associating the sequential virtual physical addresses with the physical addresses of the new guest OS page information table 612. (S37). Referring to FIG. 23, serial number virtual physical address A0, corresponding to physical addresses (in this example, directly corresponding to page numbers) 1, 6, 8,..., M of page information table 612 for new guest OS. A new guest OS physical address conversion table 611 is generated in association with A1, A2,..., An.

  As described above, the new guest OS physical address conversion table 611 and the new guest OS page information table 612 for the new guest (host OS 200) are completed (corresponding to S18).

  Here, the step of rewriting the logical address table 203 in S25 to obtain the new guest OS logical address table 203a will be described. The memory information exchanging unit 440 rewrites the physical address 218 described in the logical address table 203 to a virtual physical address described in the new guest OS physical address conversion table 611. 24A and 24B show a logical address table 203 and a new guest OS physical address conversion table 611, respectively. Here, by rewriting the physical addresses B0, B1,..., Bn in the logical address table 203 to the virtual physical addresses A0, A1,..., An in the new guest OS physical address conversion table 611, a new guest OS logical address is obtained. A table 203a (FIG. 24C) is generated.

Next, details of the operation of the device information conversion unit 550 will be described.
FIG. 18 is a flowchart showing the operation of the device information conversion unit in S19. FIG. 25 shows an example of the process of generating the new host OS device list 311a in S19. FIG. 26 shows an example of the new guest OS device list 614 generated in S20.

  Referring to FIG. 18, the device information conversion unit 550 acquires one entry from the device list 209 (FIG. 25A) of the host OS 200 acquired in S16 (S41). The entry includes a device number 221, a device name 223, and a device operation function 222. When the entry can be acquired (S42: YES), the host OS device name 322 that matches the device name 223 in the entry is searched from the device correspondence table 313 (FIG. 25B) (S43). If a match is found (S44: YES), a device name 328 equal to the guest OS device name 323 corresponding to the host OS device name 322 and a device operation function 327 in the device list 311 (FIG. 25C) Is replaced with the device name 221 and the device operation function 222 in the device list 209. In the example of FIG. 25, the device names (223) Lh2 and Lh4 in the device list 209 match the host OS device names (322) Lh2 and Lh4 in the device correspondence table 313. Accordingly, in the device list 311, the guest OS device names (323) Lg2 and Lg3 corresponding to the host OS device names (322) Lh2 and Lh4, device names (328) Lg2 and Lg3, and device operation functions (327) Fg2 , Fg3 are replaced with the device names (223) Lh2, Lh4 and device operation functions (222) Fh2, Fh4. At the same time, the device driver 315 of the guest OS 300 with the matching device name is deleted, and the device driver 215 of the host OS 200 is newly registered as the device driver 315 (S45).

  If there is no device name matching among the devices of the host OS 200 (S44: NO), the device name 221 and the device operation function 222 having no matching device name in the device list 209 in the device list 311, respectively, The corresponding device operation function 327 is registered. In the example of FIG. 25, there is no device name (223) Lh1, Lh3, Lhq in the device list 209 that matches the device correspondence table 313. Therefore, in the device list 311, the device names (223) Lh 1, Lh 3, Lhq and the device operation functions (222) Fh 1, Fh 3, Fhq that do not match are registered in the device name 328 a and the corresponding device operation function 327 a, respectively. . At the same time, the device driver 215 of the host OS 200 is newly registered as the device driver 315 (S46). When this operation is performed for all entries (S42: NO), the device list 311 for the guest OS 300 is rewritten to the device list 311a for the new host OS (corresponding to S19).

  In S20, if the guest and the host are exchanged and the same operation as in FIG. 18 is performed, a new guest OS device list 614 as shown in FIG. 26 can be created. That is, the device information conversion unit 550 matches the device names (328) Lg2 and Lg3 in the device list 311 with the guest OS device names (323) Lg2 and Lg3 in the device correspondence table 313. Accordingly, in the device list 209, the host OS device names (322) Lh2 and Lh4 corresponding to the guest OS device names (323) Lg2 and Lg3, device names (223) Lh2 and Lh4, and device operation functions (222) Fh2 , Fh4 are replaced with the device names (328) Lg2, Lg3 and device operation functions (327) Fg2, Fg3. In this manner, by replacing the device list 209 acquired from the host OS state saving area 620 in S16, a new guest OS device list 614 is generated. At the same time, the device driver 215 of the host OS 200 with the matching device name is deleted, and the device driver 315 of the guest OS 300 is newly registered as the device driver 215.

  Further, as described above, a new guest OS device correspondence table 613 is newly generated by exchanging the host-side data 322 and the guest-side data 323 of the device correspondence table 313.

Next, details of the operation of the host process generation unit 570 will be described.
FIG. 19 is a flowchart showing the operation of the host process generation unit in S21. The host process generation unit 570 copies the process information of the parent process of the guest OS 300 as a new host as the process information of the child process of the host OS 200 as a new guest (S51). Next, the memory information (page table) of the child process is based on the memory information of the host OS 200 (new guest OS logical address table 203a, new guest OS physical address conversion table 611, new guest OS page information table 612). To rewrite (S52). Next, the register information of the child process is rewritten based on the context information 227 acquired by the context information acquisition unit 580 (S53). Finally, the process is registered in the scheduling queue (S54).

  In the normal process generation procedure, the memory space (memory information) of the parent process is also copied as the memory space (memory information) of the child process together with S51. This operation creates a child process that performs exactly the same operation as the parent process, but the parent process and child process are different because the parent process and child process (newly created process) differ in part of the process information. It can be recognized as a process. After the process is generated, the child process can execute a new program by loading the program on the memory.

  However, the host process generation unit 570 according to the present invention makes it appear that a new program is loaded by mapping the memory of the existing host OS instead of loading a new program onto the memory. This is different from generation.

  According to the present invention, the present invention can be applied to the use of applying a security patch to the host OS without stopping the guest OS. When a security hole is found in the host OS, a patch for closing the security hole must be applied. On the other hand, an important application is being executed on the guest OS, and it is desirable to avoid stopping the application on the guest OS. In such a case, by switching the hardware dependency between the guest OS and the host OS using the present invention, the guest OS can continue to operate and can be restarted after applying the patch to the host OS.

FIG. 1 is a configuration diagram showing an overview of a virtual machine system. FIG. 2 is a diagram showing a context switch of the host OS. FIG. 3 is a diagram for explaining a mechanism of context switching from the guest OS to the host OS. FIG. 4 is a diagram for explaining memory management in the virtual machine system. FIG. 5 is a diagram illustrating the relationship between the logical address table, the physical memory, and the page information table in the host OS. FIG. 6 is a diagram showing the relationship among the logical address table, page information table, physical address conversion table, and guest OS area in the guest OS. FIG. 7 is a diagram for explaining device management in the virtual machine system. FIG. 8 is a table showing a device list. FIG. 9 is a table showing a device information table. FIG. 10 is a table showing a device correspondence table of the guest OS. FIG. 11 is a schematic configuration diagram showing an embodiment of a virtual machine system to which the OS switching system of the present invention is applied. FIG. 12 is a block diagram showing the configuration of the embodiment of the OS switching system of the present invention. FIG. 13 is a block diagram showing the OS state saving area. FIG. 14 is a flowchart showing the operation of the embodiment of the OS switching system of the present invention. FIG. 15 is a flowchart showing the operation of the embodiment of the OS switching system of the present invention. FIG. 16 is a flowchart showing the operation of the embodiment of the OS switching system of the present invention. FIG. 17 is a flowchart showing the operation of the memory information conversion unit in S17 and S18. FIG. 18 is a flowchart showing the operation of the device information conversion unit in S19 and S20. FIG. 19 is a flowchart showing the operation of the host process generation unit in S21. FIG. 20 is an example showing a process of generating a new host OS page information table in S17. FIG. 21 is an example showing a process of generating a new host OS logical address table in S17. FIG. 22 is an example showing a process of generating a new guest OS page information table in S18. FIG. 23 is an example of a new guest OS physical address conversion table generated in S18. FIG. 24 is an example showing the process of generating a new guest OS logical address table in S18. FIG. 25 is an example showing a process of creating a new host OS device list in S19. FIG. 26 shows an example of the new guest OS device list generated in S20.

Explanation of symbols

100, 100a Hardware 200, 200a Host OS
300, 300a Guest OS
301-1 to 301-n Guest OS application 201-1 to 201-m Host OS application 303 Logical address table (guest OS)
303a Logical address table for new host OS 304 Page information table (guest OS)
304a Page information table for new host OS 305 Address conversion unit (guest OS)
306 Physical address conversion table (Guest OS)
307 Physical address translation unit (guest OS)
308 Memory acquisition unit (guest OS)
309 Free space management unit (guest OS)
310 Free space list (Guest OS)
203 Logical address table (host OS)
203a Logical address table for new guest OS 204 Page information table (host OS)
205 Address converter (host OS)
206 Memory acquisition unit (host OS)
207 Free space management unit (host OS)
208 Free space list (host OS)
101 Physical Memory 102 Guest OS Area 311 Device List (Guest OS)
311a Device list for new host OS 312 Device information table (guest OS)
313 Device correspondence table (Guest OS)
314 Device Abstraction Unit (Guest OS)
315 Device driver (guest OS)
316 Device input / output transmitter (guest OS)
209 Device list (host OS)
210 Device information table (host OS)
211 Device input / output receiver (host OS)
212 Device abstraction unit (host OS)
213 Device driver (host OS)
214 I / O port (host OS)
215 Resource structure (host OS)
216 Resource management unit (host OS)
103 Device Interface 104 Device Controller 105 Device 106 DMAC
217 Logical address (host OS)
218 Physical address (host OS)
219 Page number (host OS)
220 In-use flag (host OS)
Page 107 317 Logical Address (Guest OS)
318 Virtual physical address (guest OS)
319 Conversion table entry (guest OS)
320 Page number (Guest OS)
321 In-use flag (guest OS)
221 Device number 222 Device operation function 223 Device name 224 User information 225 Usage counter 226 Device status 322 Host OS device name 323 Guest OS device name 500 Memory information saving unit 510 Device information saving unit 520 Memory information acquisition unit 530 Device information Acquisition unit 540 Memory information conversion unit 550 Device information conversion unit 560 Execution instruction switching unit 570 Context switch unit 400 Memory information saving unit 410 Device information saving unit 420 Memory information acquisition unit 430 Device information acquisition unit 440 Memory information replacement unit 450 Device information replacement Section 460 Context switch section 470 Execution instruction switching section 600 OS state saving section 610 Guest OS state saving area 620 Host OS state saving area 6 1 new guest OS for the physical address conversion table 612 new guest OS for page information table 613 new guest OS device correspondence table 614 new guest OS for the device list

Claims (19)

A host OS,
A guest OS running on the host OS,
When switching the guest OS to a host and the host OS to a guest,
New guest memory management information that becomes memory management information of the new guest OS after the host OS is switched based on the guest memory management information of the guest OS and the host memory management information of the host OS; Generating new host memory management information to be memory management information of the new host OS after switching the guest OS,
The guest OS replaces the guest memory management information with the new host memory management information;
The host OS replaces the host memory management information with the new guest memory management information;
The host OS and the guest OS is, you switch the execution instruction,
When the guest OS generates the new guest memory management information and the new host memory management information,
The host OS saves the host memory management information in a memory save area,
The guest OS generates the new guest memory management information based on the guest memory management information and the host memory management information of the save area, stores the new guest memory management information in the save area, and generates the new host memory management information And
When the host OS replaces the host memory management information with the new guest memory management information,
The host OS replaces the host memory management information with the new guest memory management information in the save area,
Thereafter, an OS switching system in which the host OS and the guest OS switch execution commands . The OS switching system according to claim 1 ,
Based on the guest device management information of the guest OS and the host device management information of the host OS, the guest OS has new guest device management information that becomes device management information of the new guest OS, and a device of the new host OS Generate new host device management information to be management information,
The guest OS replaces the guest device management information with the new host device management information;
The OS switching system in which the host OS replaces the host device management information with the new guest device management information. The OS switching system according to claim 1 ,
The guest OS is
Based on the host page information table and host logical address table included in the host memory management information, a new guest page information table, a new guest physical address conversion table, and a new guest logical address table included in the new guest memory management information. A memory information conversion unit for generating,
The host page information table indicates the correspondence between the physical address and the usage status of the physical address,
The host logical address table indicates correspondence between logical addresses and physical addresses;
The new guest page information table indicates a correspondence between the physical address that can be used by the host OS in the host page information table and the usage status of the physical address.
The new guest physical address conversion table shows the correspondence between the physical address and the virtual physical address in the new guest page information table,
The OS switching system, wherein the new guest logical address table indicates correspondence between the logical address in the host logical address table and the virtual physical address in the new guest physical address conversion table. In the OS switching system according to claim 3 ,
The memory information conversion unit further includes a new host included in the new host memory management information based on the host page information table and a guest logical address table and a guest physical address conversion table included in the guest memory management information. Create a logical address table and a new host page information table,
The guest logical address table indicates a correspondence between a logical address and a virtual physical address in the guest OS,
The guest physical address conversion table shows the correspondence between the virtual physical address and the physical address in the guest OS,
The new host logical address table indicates a correspondence between the logical address in the guest logical address table and the physical address in the guest physical address conversion table,
The new host page information table is obtained by changing the host page information table for the guest OS. OS switching system. In the OS switching system according to claim 2 ,
The guest OS further includes
Based on the guest device correspondence table and guest device list included in the guest device management information and the host device list included in the host device management information, the new guest device correspondence table and new It has a device information conversion unit that generates a guest device list,
The guest device correspondence table indicates correspondence between devices recognized by the guest OS and devices recognized by the host OS,
The guest device list shows information on devices that can be used by the guest OS,
The host device list shows information on devices that can be used by the host OS,
The new guest device list updates the guest device list with information about the device of the host device list among the devices in the host device list, and the remaining devices in the host device list With additional information about the device
The new guest device correspondence table is obtained by replacing a device recognized by the guest OS and a device recognized by the host OS in the guest device correspondence table. The OS switching system according to claim 1 ,
The guest OS uses the process information of the first process of the guest OS as the process information of the second process of the host OS, rewrites the memory information of the second process based on the new guest memory management information, and the second process An OS switching system comprising a host process generation unit that rewrites the register information of the second OS to register information of the host OS and registers the second process in a scheduling queue. An OS switching system according to any one of claims 1 to 6 ,
A virtual computer system comprising: hardware used for the OS switching system. (A) When switching a guest OS to a host and a host OS to a guest, the guest OS switches the host OS based on guest memory management information of the guest OS and host memory management information of the host OS Generating new guest memory management information to be memory management information of the new guest OS after, and new host memory management information to be memory management information of the new host OS after switching the guest OS;
(B) the guest OS replacing the guest memory management information with the new host memory management information;
(C) the host OS replacing the host memory management information with the new guest memory management information;
; (D) the host OS and the guest OS, and ingredients Bei the step of switching the execution instruction,
The step (a) includes:
When switching the guest OS to the host and the host OS to the guest, the host OS saves the host memory management information in a memory save area;
The guest OS generates the new guest memory management information based on the guest memory management information and the host memory management information of the save area, stores the new guest memory management information in the save area, and generates the new host memory management information Step and
With
The step (c) includes:
(C1) An OS switching method comprising a step in which the host OS replaces the host memory management information with the new guest memory management information in the save area . The OS switching method according to claim 8 , wherein
(E) The guest OS includes new guest device management information that becomes device management information of the new guest OS based on guest device management information of the guest OS and host device management information of the host OS, and the new host Generating new host device management information to be OS device management information;
(F) the guest OS replacing the guest device management information with the new host device management information;
(G) The host OS further comprising the step of replacing the host device management information with the new guest device management information. The OS switching method according to claim 8 , wherein
The step (a) includes:
(A1) Based on the host page information table and host logical address table included in the host memory management information, a new guest page information table, a new guest physical address conversion table, and a new guest logic included in the new guest memory management information A step of generating an address table,
The host logical address table indicates correspondence between logical addresses and physical addresses;
The new guest page information table indicates a correspondence between the physical address that can be used by the host OS in the host page information table and the usage status of the physical address.
The new guest physical address conversion table shows the correspondence between the physical address and the virtual physical address in the new guest page information table,
The new guest logical address table indicates an association between the logical address in the host logical address table and the virtual physical address in the new guest physical address conversion table. The OS switching method according to claim 10 ,
The step (a) includes:
(A2) A new host logical address table and a new host included in the new host memory management information based on the host page information table, a guest logical address table and a guest physical address conversion table included in the guest memory management information A step of generating a page information table;
The guest logical address table indicates a correspondence between a logical address and a virtual physical address in the guest OS,
The guest physical address conversion table shows the correspondence between the virtual physical address and the physical address in the guest OS,
The new host logical address table indicates a correspondence between the logical address in the guest logical address table and the physical address in the guest physical address conversion table,
The new host page information table is obtained by changing the host page information table for the guest OS. OS switching method. The OS switching method according to claim 9 , wherein
The step (e) includes:
(E1) Based on a guest device correspondence table and guest device list included in the guest device management information and a host device list included in the host device management information, a new guest device correspondence included in the new guest device management information Creating a table and a list of new guest devices,
The guest device correspondence table indicates correspondence between devices recognized by the guest OS and devices recognized by the host OS,
The guest device list shows information on devices that can be used by the guest OS,
The host device list shows information on devices that can be used by the host OS,
The new guest device list updates the guest device list with information about the device of the host device list among the devices in the host device list, and the remaining devices in the host device list With additional information about the device
The new guest device correspondence table is obtained by replacing the device recognized by the guest OS and the device recognized by the host OS in the guest device correspondence table. The OS switching method according to claim 8 , wherein
The step (d) includes:
(D1) setting the process information of the first process of the guest OS as process information of the second process of the host OS;
(D2) rewriting the memory information of the second process based on new guest memory management information;
(D3) rewriting the register information of the second process with the register information of the host OS;
(D4) An OS switching method comprising: registering the second process in a scheduling queue. (A) When switching a guest OS to a host and a host OS to a guest, the guest OS switches the host OS based on guest memory management information of the guest OS and host memory management information of the host OS Generating new guest memory management information to be memory management information of the new guest OS after, and new host memory management information to be memory management information of the new host OS after switching the guest OS;
(B) the guest OS replacing the guest memory management information with the new host memory management information;
(C) the host OS replacing the host memory management information with the new guest memory management information;
; (D) the host OS and the guest OS, and ingredients Bei the step of switching the execution instruction,
The step (a) includes:
When switching the guest OS to the host and the host OS to the guest, the host OS saves the host memory management information in a memory save area;
The guest OS generates the new guest memory management information based on the guest memory management information and the host memory management information of the save area, stores the new guest memory management information in the save area, and generates the new host memory management information Step and
With
The step (c) includes:
(C1) A program that causes a computer to execute an OS switching method including the step of replacing the host memory management information with the new guest memory management information in the save area by the host OS . The program according to claim 14 , wherein
(E) The guest OS includes new guest device management information that becomes device management information of the new guest OS based on guest device management information of the guest OS and host device management information of the host OS, and the new host Generating new host device management information to be OS device management information;
(F) the guest OS replacing the guest device management information with the new host device management information;
(G) The host OS further comprising the step of replacing the host device management information with the new guest device management information. The program according to claim 14 , wherein
The step (a) includes:
(A1) Based on the host page information table and host logical address table included in the host memory management information, a new guest page information table, a new guest physical address conversion table, and a new guest logic included in the new guest memory management information A step of generating an address table,
The host logical address table indicates correspondence between logical addresses and physical addresses;
The new guest page information table indicates a correspondence between the physical address that can be used by the host OS in the host page information table and the usage status of the physical address.
The new guest physical address conversion table shows the correspondence between the physical address and the virtual physical address in the new guest page information table,
The new guest logical address table is a program indicating correspondence between the logical address in the host logical address table and the virtual physical address in the new guest physical address conversion table. The program according to claim 16 , wherein
The step (a) includes:
(A2) A new host logical address table and a new host included in the new host memory management information based on the host page information table, a guest logical address table and a guest physical address conversion table included in the guest memory management information A step of generating a page information table;
The guest logical address table indicates a correspondence between a logical address and a virtual physical address in the guest OS,
The guest physical address conversion table shows the correspondence between the virtual physical address and the physical address in the guest OS,
The new host logical address table indicates a correspondence between the logical address in the guest logical address table and the physical address in the guest physical address conversion table,
The new host page information table is a program obtained by changing the host page information table for the guest OS. The program according to claim 15 , wherein
The step (e) includes:
(E1) Based on a guest device correspondence table and guest device list included in the guest device management information and a host device list included in the host device management information, a new guest device correspondence included in the new guest device management information Creating a table and a list of new guest devices,
The guest device correspondence table indicates correspondence between devices recognized by the guest OS and devices recognized by the host OS,
The guest device list shows information on devices that can be used by the guest OS,
The host device list shows information on devices that can be used by the host OS,
The new guest device list updates the guest device list with information about the device of the host device list among the devices in the host device list, and the remaining devices in the host device list With additional information about the device
The new guest device correspondence table is a program in which a device recognized by the guest OS and a device recognized by the host OS in the guest device correspondence table are exchanged. The program according to claim 14 , wherein
The step (d) includes:
(D1) setting the process information of the first process of the guest OS as process information of the second process of the host OS;
(D2) rewriting the memory information of the second process based on new guest memory management information;
(D3) rewriting the register information of the second process with the register information of the host OS;
(D4) A program comprising: registering the second process in a scheduling queue.

Images