JP5085611B2 - Computer program that transitions between real and virtual environments - Google Patents

Description

  The present invention relates to a technology for causing a computer system to transition bi-directionally between a real environment and a virtual environment.

  In a personal computer or server, a virtual environment called a hypervisor or a virtual machine monitor (VMM) is introduced to form a virtual environment and a plurality of operating systems (OS) can be operated. It is possible to perform clustering of wear. As a result, for example, existing servers can be integrated in the server, OSs and application programs that are supposed to run on specific hardware can be run on the latest hardware, and the OS can be switched according to the situation. There is an advantage that can be. However, an OS that operates in a virtual environment has an increased overhead due to the virtualization program and has a lower performance or memory than an OS that operates in a real environment that is a normal OS environment in which the virtualization program is not loaded. There is also a drawback that extra resources such as CPU and CPU are consumed.

  Japanese Patent Laid-Open No. 2004-151867 describes a dependency relationship between a guest OS and a host OS for hardware access without restarting the guest OS when a failure occurs in the host OS in a virtual machine system in which the guest OS runs on the host OS. A technique for switching is disclosed. In this document, when the guest OS is switched to the host and the host OS is switched to the guest, the guest OS changes the memory of the host OS after switching based on the guest memory management information of the guest OS and the host memory management information of the host OS. It describes that new guest memory management information as management information and new host memory management information as memory management information of the guest OS 300 after switching are generated.

  Patent Document 2 enables expansion and reduction of memory usable in a virtual machine without restarting the OS in a virtual machine system in which a virtual machine environment and a plurality of virtual machines are constructed on a physical machine. Disclose technology. In this document, a memory area that can be used by each of a plurality of virtual machines operating in a virtual machine environment is associated with a physical memory that the physical machine has, and the memory area is determined according to the amount of used memory of each virtual machine. This predetermined area is described as a reserved area that can be allocated and released to each virtual machine.

  Patent Document 3 describes a technique for suspending an OS execution environment to a hard disk as a conventional technique for switching between multiple OSs. This document describes that the BIOS initializes the hardware and acquires the execution environment of the OS from the hard disk. Patent Document 4 discloses a technique for resuming immediately after shifting a computer to a suspended state and transferring control to the system BIOS.

JP 2007-34572 A JP 2008-225520 A JP 2008-524730 A JP 2008-26934 A

  The usage example of the virtual environment can be considered in addition to the above-described advantages. In recent years, malicious software (malware) such as computer viruses, worms, and spyware is frequently sent to a user's computer system via a network. Most of such software is discovered and removed by firewalls and anti-virus software. However, a recently developed software tool technology called a rootkit hides the presence of such malware and makes it difficult to detect. For example, a rootkit may intercept an API process used by an application program to obtain information and delete a malware program from the program list to hide it. Then, the virus removal program can no longer find the malware.

  Moreover, even if the OS itself tries to recognize the existence of the rootkit, it becomes difficult if the rootkit already lurks in the OS. In such a case, an inspection OS different from the OS to be inspected is operated in the same virtual environment, and the inspection OS examines the code expanded in the main memory of the OS to be inspected. The existence of a rootkit or malware can be detected by examining a file in a storage such as a hard disk or a nonvolatile memory in which the OS is stored. Also, even when the OS is partially repaired, it is safer to use another OS in the virtual environment than to use the physical environment. However, since maintenance work such as inspection and repair of the OS may be performed periodically or as necessary, the OS is operated in a virtual environment in which performance is constantly reduced or resources are excessively consumed. There is no need. Therefore, if the OS can be operated in the real environment, and if necessary, the operation can be performed after moving to the virtual environment and returning to the real environment while maintaining the system state before the migration, the OS can be maintained. Is convenient.

  Therefore, an object of the present invention is to provide a computer program that can be migrated from a real environment to a virtual environment while maintaining the system state in the real environment. A further object of the present invention is to provide a computer program that can be migrated from a virtual environment to a real environment while maintaining the system state in the virtual environment. Furthermore, an object of the present invention is to provide a computer program that can temporarily move to a virtual environment while maintaining the system state in the real environment, perform maintenance work, and move to the real environment again. Another object of the present invention is to provide a computer program capable of automatically performing a migration from a real environment to a virtual environment and a migration from a virtual environment to a real environment in a short time.

  The present invention can be applied to a computer system that can operate in either a real environment or a virtual environment such as a server or a client. First, the computer system executes the OS in a real environment. Then, after the computer system shifts from the power-on state to the power-saving state in which the system state can be maintained, the computer system starts to return from the power-saving state to the power-on state. As a trigger for the transition to the power saving state, an internally generated signal or a signal received from the outside can be used. The power saving state in which the system state can be maintained is an operating state that consumes less power than the power on state and can return to the system state prior to entering the power saving state when the power on state is restored. State and hibernation state are included.

  Subsequently, the virtualization program is loaded into a predetermined area secured in the main memory and executed until the computer system returns from the power saving state to the power on state. Then, resume the OS and move to the virtual environment. When such a procedure is executed, the computer system operating in the real environment can be automatically transferred to the virtual environment while maintaining the operating state of the real environment. When the power saving state is changed to the suspended state, the execution environment of the computer system can be shifted in a very short time if the power saving state is shifted in response to the transition to the suspended state. Responding to the transition to the suspend state means that the process for maintaining the system state is completed and the power state is changed to the power saving state defined by the computer system. It does not include waiting for an event based on other conditions to be generated after entering the power saving state, or waiting for a predetermined time to elapse with a timer. .

  When migrating from the virtual environment to the real environment, the computer system transitions from the power-on state to the power-saving state where the system state can be maintained, as in the case of the transition from the real environment to the virtual environment.・ Start returning to the ON state. Before the computer system returns from the power saving state to the power on state, the virtualization program is unloaded from a predetermined area of the main memory, and the OS is resumed to shift to the real environment. When such a procedure is executed, the computer system operating in the virtual environment can be transferred to the real environment while maintaining the system state in the virtual environment. Unloading the virtualization program means that the virtualization program cannot be executed on the main memory, and at the same time the execution environment of the virtualization program may be stored in a non-volatile memory. Alternatively, the code of the virtualization program may be overwritten and cleared so that it cannot be read.

  After the OS operates in the real environment, the OS can release a predetermined area of the main memory where the virtualization program is unloaded for its own use. By doing this, it is not necessary to allocate main memory to virtualized programs that are not necessary in the real environment, and memory resources can be used effectively. The main memory area for loading the virtualization program may be secured at all times while the OS is operating in the real environment, or may be secured using a part of the SMRAM area. May be.

  According to the above procedure, it is convenient to operate in a real environment that usually has good performance and does not consume more main memory than necessary, and can operate in a virtual environment only when necessary. In particular, it is convenient when performing virus inspection and repair of the OS in a virtual environment that cannot be performed sufficiently in the real environment and that is not always necessary in the real environment. The virtualization program can be loaded and unloaded by the BIOS that obtains control of the computer system during the transition from the power saving state to the power on state. The computer program according to the present invention can be composed of a BIOS, a hypervisor, a Domain0 guest OS, and the like. The computer program can be mounted on a computer system such as a server or a client to move the operating environment.

  According to the present invention, it is possible to provide a computer program that can be migrated from a real environment to a virtual environment while maintaining the system state in the real environment. Furthermore, according to the present invention, it is possible to provide a computer program that can be migrated from a virtual environment to a real environment while maintaining the system state in the virtual environment. Furthermore, according to the present invention, it is possible to provide a computer program capable of temporarily moving to a virtual environment while maintaining the system state in the real environment, performing maintenance work, and moving to the real environment again. Furthermore, according to the present invention, it is possible to provide a computer program capable of automatically performing a transition from a real environment to a virtual environment and a transition from a virtual environment to a real environment in a short time.

It is a functional block diagram which shows the structure required for description of this invention of a computer system. FIG. 2 is a diagram illustrating a software configuration when a computer system operates in a real environment and when it operates in a virtual environment. It is a flowchart which shows the procedure when a computer system transfers to a virtual environment from a real environment. It is a figure which shows a mode that the recording area of a main memory is ensured in order to load the module of the guest OS of a hypervisor and Domain0. It is a flowchart which shows the procedure when a computer system transfers to a real environment from a virtual environment.

  FIG. 1 is a functional block diagram showing a configuration necessary for explaining the present invention of the computer system 10. Connected to the north bridge 13 are a CPU 11, a main memory 15, a video controller 17, and a south bridge 19. The computer system 10 realizes complete virtualization that does not require modification to the OS by performing virtualization with hypervisor-type virtualization software supported by CPU called HVM (Hardware Virtual Machine). By using the HVM, an OS such as Windows (registered trademark) whose source code is not disclosed can be executed in a virtual environment.

  The CPU 11 has a virtualization support mechanism for realizing HVM. The virtualization support mechanism is provided by a technology called VT (Virtualization Technology) mounted on an Intel CPU or AMD-V mounted on an AMD CPU, for example. In the CPU 11 equipped with the virtualization support mechanism, the OS kernel operates at the privilege level of the ring 0 as usual, and the hypervisor operates at the privilege level corresponding to the virtual ring-1 added on the ring 0. . However, the present invention can also be applied to para-virtualization in which the hypervisor operates in the ring 0 and the OS kernel operates in the ring 1. Paravirtualization has less performance degradation than when the OS operates in a real environment, but it is necessary to modify the code to change the operating environment of the kernel.

  An LCD 19 is connected to the video controller 17. The south bridge 19 incorporates interface circuits for SATA (Serial Advanced Technology Attachment), HDA (High Definition Audio) bus, USB (Universal Serial Bus), PCI (Peripheral Component Interconnect) bus, LPC (Low Pin Count) bus, etc. It is. The HDD 21 is connected to the SATA interface. The HDD 21 stores Xen (trademark) as a hypervisor, Windows (registered commercial) and Linux (registered trademark) as guest OSs. An embedded controller (EC) 25 and a flash ROM 27 are connected to the LPC bus 23.

  The EC25 is a microcomputer composed of an 8- to 16-bit CPU, ROM, RAM, and the like, and further includes a multi-channel A / D input terminal, a D / A output terminal, a timer, and a digital input / output terminal. Yes. The EC 25 can execute a program related to management of an operating environment such as an internal temperature and power supply of the computer system 10 independently of the CPU 11. The EC 25 includes a keyboard controller, and a keyboard 29 is connected thereto. The computer system 10 conforms to the ACPI (Advanced Configuration and Power Interface) standard, and the flash ROM 27 stores a BIOS corresponding to the ACPI.

  In the ACPI standard, a plurality of power states are defined, and the computer system 10 writes the S0 state, which is in the power-on state, and the work contents to the main memory 15 and retains the storage in the main memory 15. S3 state, which is a suspend state for supplying power only to devices required for restart and devices required for restart, hibernation for writing power to the HDD 21 or other nonvolatile memory to supply power only to devices required for restart It is possible to transition to either the S4 state, which is a state, and the S5 state, which is a soft-off or power-off state in which power is supplied only to devices required for restart. In the suspended state, power to the south bridge 19, EC 25, and flash ROM 27 is maintained.

  FIG. 2 is a diagram illustrating a software configuration when the computer system 10 operates in a real environment and when it operates in a virtual environment. FIG. 2A shows a state when the computer system 10 operates in a real environment. The guest OS 101 operates on the hardware 105, and the application 102 operates on the guest OS. The guest OS 101 abstracts the hardware 105 mounted on the computer system 10 such as the CPU 11, HDD 25, main memory 13, and flash ROM 27 to the application. The BIOS 103 can directly access the hardware 105 by an instruction from the guest OS 101 or the application 102.

  The guest OS does not have the concept of “guest” when operating in a real environment, but when operating in a real environment, the module configuration is the same as when operating in a virtual environment. For convenience, it is expressed as a guest OS 101 when operating in a real environment. The guest OS 101 is assumed to be Windows (registered trademark) in this embodiment. FIG. 2B shows a state when the computer system 10 operates in a virtual environment. The hypervisor 109 exists as a virtualization layer on the hardware 105 and provides virtual machines for the guest OSs 101 and 107. To do. The guest OSs 101 and 107 operate on the hypervisor 109 while receiving the service in the same manner as an application. In this embodiment, the hypervisor 109 employs Xen provided as open source software, and the guest OS 107 employs Linux (registered trademark), but this selection is an example. The guest OS 107 is executed as Domain 0 of the hypervisor 107, is given special authority, and is allowed to access the hardware 105. The guest OS 101 is executed as a domain U of the hypervisor.

  The guest OS 107 incorporates a maintenance system that inspects the virus of the guest OS 101 and repairs the guest OS 101. The maintenance system may be configured as a part of the hypervisor 109 module. In the present embodiment, since the virtual environment is realized by using the virtualization support mechanism of the CPU 11, it is not necessary to modify the guest OS 101. Further, it is not necessary to modify the hardware 105 other than changing the setting of the virtualization support mechanism as necessary. However, the hypervisor 109 and the BIOS 103 need to incorporate a function for executing the procedures of FIGS.

  FIG. 3 is a flowchart showing a procedure when the computer system 10 shifts from the real environment to the virtual environment. In block 301, the computer system 10 operates in the power-on state in the real environment, and the program is executed in the state shown in FIG. In block 303, an event for migrating to the virtual environment is generated in the computer system 10 and received by the guest OS 101. The migration to the virtual environment in the present embodiment is temporarily performed in order to inspect and maintain the guest OS 101. An event is generated when the BIOS 103 detects that the user has operated a special key assigned to a key on the keyboard 29, or an event for schedule management executed on the guest OS 101 is generated periodically. can do.

  The event includes transitioning the computer system 10 to suspend. In the computer system 10, transition to suspend is performed in two ways: normal suspend and suspend for the purpose of transition of the operation mode between the real environment and the virtual environment. The guest OS 101 and the BIOS 103 can recognize whether the event is intended for the normal suspend or the transition of the operating environment. However, when an event for shifting to the normal suspend state occurs in the computer system 10, it may be set in the BIOS 103 in advance so as to always execute the procedure of FIG.

  In block 305, the guest OS 101 dynamically reduces the storage area allocated to the main memory 15 for the guest OS 101 using the plug-and-play function of ACPI, and the hypervisor 109 And the guest OS 107 are allocated for loading. FIG. 4 is a diagram showing a state in which a recording area of the main memory 15 is secured in order to load the modules of the hypervisor 109 and the guest OS 107 of Domain0. In this example, the capacity of the main memory 15 is 2 GB. On the other hand, since the CPU 11 can process data with 32 bits, the program can use a 4 GB logical address space. FIG. 4 shows a 2 GB physical address space 201 corresponding to the main memory 15, a linear address space 203 corresponding to the hypervisor 109 and the guest OS 107, and a linear dress space 205 corresponding to the guest OS 101. .

  While the guest OS 101 is operating in the block 301, the virtual address areas 213 and 215 are allocated to the linear address space 205, and the physical address areas 209 and 211 of the physical address space 201 are mapped to them to perform paging. The guest OS 109 is executed using the function. Here, the guest OS 101 reduces the physical address area 211 to be assigned to the hypervisor 109 and the guest OS 107, assigns the virtual address area 207 to the hypervisor 109 and the guest OS 107, and reduces its own virtual address area to 213. Then, the hypervisor 109 and the guest OS 107 are assigned a virtual address area 207 mapped to the physical address area 211, and the guest OS 213 is assigned a virtual address area 213 mapped to the physical address area 209. Can be loaded.

  In block 307, when the BIOS 103 inputs information indicating that the computer system 10 has shifted to the suspended state to the EC 25, the computer 103 shifts the computer system 10 to the power-on state. Set the instant wakeup. This setting may be performed using RTC (Real Time Clock). Then, the BIOS 103 notifies the guest OS 101 when the setting of the instantaneous wakeup to the EC 25 is completed. However, the transition of the operating environment in the present invention can be performed even when wakeup is not performed instantaneously.

  In block 309, the guest OS 101 stores all information necessary for maintaining the operating state in the block 301, such as information stored in the registers of the CPU 11 and other devices, in the main memory 15, and then The bridge 19 is caused to process the power supply for shifting the computer system 10 to the suspended state. Subsequently, the state machine of the computer system 10 notifies the EC 25 that the power state of the computer system 10 has shifted to the suspended state. When the computer system 10 shifts to the suspended state, the power supplied to the CPU 11 is stopped and reset. Upon receiving the notification, the EC 25 wakes up the computer system 10 and starts shifting to the power-on state. At this time, the module of the guest OS 101 is loaded in the physical address area 209 mapped to the virtual address area 213.

  When the CPU 11 starts to operate after being reset, an address is designated so that the BIOS 103 is executed first. In block 311, the BIOS 103 loads the modules of the hypervisor 109 and the guest OS 107 from the HDD 25 to the physical address area 211 of the main memory 15 mapped to the virtual address area 207. Then, the virtualization support mechanism of the CPU 11 is enabled or the setting of the memory controller included in the north bridge 13 is changed as necessary. In block 313, the code for resuming in the BIOS 103 is executed to set SMI and necessary hardware. Subsequently, control is transferred to the hypervisor 109, and the hypervisor 109 resumes and performs necessary processing. Subsequently, when the hypervisor 109 transfers control to a predetermined address of the guest OS 107, the guest OS 107 resumes.

  In block 315, when the guest OS 107 transfers control to the hypervisor 109, and the hypervisor 109 transfers control to the return address of the guest OS 101, the guest OS 101 resumes and stores the register information stored in the main memory 15. Return to register. As a result, the computer system 10 returns to the system state of the block 301 in a state where the virtual environment is realized. In block 315, the computer system 10 operates in the virtual environment shown in FIG. 2B, and the maintenance system incorporated in the guest OS 107 inspects the guest OS 101 for viruses and repairs the guest OS 101.

  In the above procedure, the physical address area 211 of the main memory 15 for loading the hypervisor 109 and the guest OS 107 is secured in the procedure of block 305, but other methods are available for securing the physical address area of the main memory. There is also. The CPU 11 can operate in an SMM (System Management Mode), which is an operation mode for system management, by asserting an SMI (System Management Interrupt) input pin (SMI #). In order to implement SMM, the main memory 15 has an SMRAM area that can be exclusively used by the CPU 11 when the computer system 10 operates in SMM. In the main memory 15, in addition to the SMRAM capacity necessary for executing the original SMM, a capacity necessary for loading the hypervisor 109 and the guest OS 107 is secured in advance as the SMRAM.

  When the computer system 10 operates in a real environment, a memory controller incorporated in the north bridge 13 is set so that the entire SMRAM area cannot be accessed from the guest OS 101. The BIOS 103 releases the memory controller so that a part of the SMRAM area can be accessed by the guest OS 107 by the physical address area 211. This method is executed by the BIOS 103 when control is transferred from the suspended state to the BIOS 103. The physical address area secured in the main memory 15 may be secured not in the SMRAM area but in a part of the normal storage area of the main memory 15 accessible from the OS 107.

  FIG. 5 is a flowchart showing a procedure when the computer system 10 shifts from the virtual environment to the real environment. In block 401, the computer system 10 is operating in the power-on state in the virtual environment shown in FIG. 2B through the procedure shown in FIG. In block 303, an event for moving to the real environment is generated in the computer system 10, and the guest OS 101 receives it. Blocks 403, 405, and 407 are executed in the same procedure as described in blocks 303, 307, and 309 in FIG. However, the transition to the suspended state is performed not only by the guest OS 101 but also by the hypervisor 109 and the guest OS 107.

  In block 407, the guest OS 101 collects necessary information with reference to the ACPI table included in the BIOS 103 in order to shift to the suspended state. After the state transition in the OS environment is completed, the guest OS 101 issues an I / O instruction to the virtual hardware I / O address obtained from the ACPI table, and performs a transition operation to the suspended state. Similarly, the guest OS 107 also issues an I / O command to the virtual hardware and performs a transition operation to the suspended state. When the hypervisor 109 confirms that the transition to the suspended state by the guest OS 101 and the guest OS 107 has been completed, the hypervisor 109 issues an I / O command to the actual hardware 105 and transitions to the suspended state. If necessary, the BIOS 103 traps an I / O instruction and sets the south bridge 19 in order to shift to the suspended state.

  In the computer system 10 that has transitioned to the suspended state, the hypervisor 109 and guest OS 107 modules are stored in the physical address area 211 of the main memory 15, and the guest OS 101 module is stored in the physical address area 209. In block 409, the BIOS 103 unloads the hypervisor 109 and guest OS 107 modules from the main memory 15, and further clears the stored contents so that the information stored by other programs cannot be read. . In block 411, the module of the guest OS 101 resumes. At this time, the physical address area 209 is still allocated to the guest OS 101.

  In block 413, the guest OS 101 adds a virtual address area 215 to the linear address space 205 so that the physical address area 211 allocated to the hypervisor 109 and the guest OS 107 is released and used as its own physical address area. To do. Therefore, since the guest OS 101 can expand the address area that can be used in the main memory 15 when moving from the virtual environment to the real environment, memory resources are reduced for temporary use of the virtual environment. Can be prevented. If a part of the SMRAM area is used as a physical address area in the procedure of FIG. 3, the BIOS 103 sets the memory controller so that the OS 101 and applications cannot access the address area.

  In the procedure of FIGS. 3 and 5, the method of realizing the bidirectional transition between the real environment and the virtual environment through the transition to the suspend state has been described. However, the present invention can be implemented through the hibernation state. You can also. When waking up from normal hibernation, a wake-up signal is supplied to the computer system 10 through a keyboard 29, a switch on the chassis, or a wake on LAN function from the network. In order to realize the present invention via hibernation, the method may be waked up by such a method. However, the schedule management mechanism managed in the computer system 10 generates an event to shift the operating environment. You may make it carry out forcibly.

  3 and 5, the blocks 309 and 407 start the wake-up in response to the transition to the suspended state. However, the present invention is not limited to the transition to the suspended state. You may make it wake up based on the signal input from a keyboard, the wake on run, or the event of a schedule management mechanism. Furthermore, the transition of the operating environment in both directions may be performed by employing both the suspended state and the hibernation state. In the present invention, regardless of whether the system goes through the suspended state or the hibernation state, the present invention performs the transition between the real environment and the virtual environment while maintaining the state of the system before the migration. It can return to the state.

  Although the present invention has been described with the specific embodiments shown in the drawings, the present invention is not limited to the embodiments shown in the drawings, and is known so far as long as the effects of the present invention are achieved. It goes without saying that any configuration can be adopted.

201 ... physical address space 203, 205 ... linear address space 207, 213, 215 ... virtual address area 209, 211 ... physical address area

Claims (10)

To computer systems that can operate in real and virtual environments,
Running an operating system in the real environment;
Initiating a return from the power saving state to the power on state after the computer system transitions from a power on state to a power saving state capable of maintaining the system state;
Loading the virtualization program into a predetermined area secured in main memory until the computer system returns from the power saving state to the power on state;
Executing the virtualization program;
Resuming the operating system and migrating to the virtual environment. The process is
Starting the return from the power saving state to the power on state after the computer system operating in the virtual environment has transitioned from the power on state to a power saving state capable of maintaining the system state;
Unloading the virtualization program from the predetermined area before the computer system returns from the power saving state to the power on state;
The computer program according to claim 1, further comprising: resuming the operating system and moving to the real environment. The computer program according to claim 2, wherein, after the processing is executed in the real environment, the operating system includes a step of releasing the predetermined area in which the virtualization program is unloaded to the operating system. The process is
2. The method according to claim 1, further comprising a step of allocating a part of the area of the main memory used by the operating system as the predetermined area while the operating system is operating in a real environment. 4. The computer program according to any one of 3. The process is
4. The computer program according to claim 1, further comprising the step of providing a part of an SMRAM area as the predetermined area while the operating system is operating in a real environment.   The power saving state is a suspend state, and the step of starting the return includes the step of transitioning to the power-on state in response to transition to the suspend state. Computer programs.   The computer program according to claim 1, wherein the power saving state is a hibernation state.   The computer program according to claim 1, wherein the loading step is executed by a BIOS. The process is
The computer program according to claim 1, further comprising a step of checking for a virus in the operating system in the virtual environment.   A computer system on which the computer program according to any one of claims 1 to 9 is mounted.

Images