JP2012043089A - Information processor and program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To allocate a shared memory to an OS having no virtual storage functions even if a free space in the shared memory of a virtual machine fragments.SOLUTION: An information processor specifies a free space allocated to none of processes within a shared memory 103A when a first OS 300 that controls a first process 310 requesting allocation of the shared memory 103A has no virtual storage functions. When the free space with consecutive physical addresses that has a capacity equal to or greater than a requested amount is not extracted, the information processor writes a memory content in an allocated memory area into another memory area within the shared memory so as to obtain the free space with consecutive physical addresses that has a capacity equal to or greater than the requested amount, notifies the first process 310 of a location of free space with consecutive physical addresses and notifies a process corresponding to a memory content written into another memory area of a memory location of the memory content.

Description

  The present invention relates to an information processing apparatus and a program.

There is known a virtual machine that uses one computer as a plurality of computers by operating a plurality of OSs (Operating Systems) in parallel on one computer. In the virtual machine, a hypervisor which is a basic program for realizing the virtual machine is operated, and a plurality of OSs are operated under the control of the hypervisor.
One of the roles of the hypervisor is to manage the storage area of the main storage device. For example, in the invention described in Patent Document 1, first, a first storage area is assigned to a first OS, a second storage area is assigned to a second OS, and a part of the first storage area is shared memory. Set to. The first OS and the second OS have a conversion table that associates the physical address and the virtual address of the storage area. When the second OS requests the hypervisor to use the shared memory, the hypervisor The top physical address of the second storage area stored in the OS conversion table is rewritten to the top physical address of the shared memory.

Japanese Patent No. 3874603

  An object of the present invention is to allocate a shared memory to an OS that does not have a virtual storage function even when a free area of the shared memory of a virtual machine is fragmented.

  The invention according to claim 1 is a receiving unit that receives a request to allocate a shared memory of the main storage device from the first process, and an OS that controls the first process when the request is received by the receiving unit. The operating system is assigned to any process in the shared memory when it is determined by the determining means that the OS does not have a virtual storage function. An unoccupied free area, and extracting means for extracting a continuous free area of physical addresses having a capacity greater than or equal to the requested amount from the first process; and an amount greater than the requested amount from the first process by the extracting means A physical address having a capacity equal to or greater than the requested amount from the first process when a continuous free area of the physical address having a capacity of A writing means for writing storage contents of a storage area allocated to a process other than the first process to another storage area in the shared memory, so that a continuous free area can be obtained. The obtained storage location of the storage content written by the writing means is notified to the first process of the location of the continuous free area of the physical address having a capacity larger than the requested amount from the first process. An information processing apparatus is provided that includes notification means for notifying a process corresponding to the stored content.

  According to a second aspect of the present invention, the extracting unit specifies a free area that is not allocated to any process in the shared memory when the determining unit determines that the OS does not have a virtual storage function. If the total capacity of the specified free areas is equal to or larger than the requested amount from the first process, continuous free areas of physical addresses having a capacity larger than the requested amount from the first process are extracted. The information processing apparatus according to claim 1 is provided.

  The invention according to claim 3 controls the first process when the computer accepts a request for allocation of the shared memory of the main storage device from the first process and the request by the accepting unit. A determination unit that determines whether or not an operating system (OS) to operate has a virtual storage function, and which process in the shared memory when the determination unit determines that the OS does not have a virtual storage function An unassigned free area, and extracting means for extracting continuous free areas of physical addresses having a capacity equal to or greater than the requested amount from the first process; and from the first process by the extracting means When a continuous free area of a physical address having a capacity larger than the requested amount is not extracted, the capacity larger than the requested amount from the first process Writing means for writing the storage contents of a storage area allocated to a process other than the first process to another storage area in a shared memory, so that a continuous free area of physical addresses having The storage contents obtained by the writing means are notified to the first process of the location of the continuous free area of the physical address having a capacity equal to or larger than the requested amount from the first process. A program for causing the storage location to function as notification means for notifying the process corresponding to the storage content is provided.

According to the first and third aspects of the invention, even when the free area of the shared memory of the virtual machine is fragmented, the shared memory can be allocated to an OS that does not have a virtual storage function.
According to the second aspect of the invention, useless processing can be omitted.

1 is a diagram conceptually illustrating a configuration of an information processing device 1. FIG. 2 is a diagram illustrating a configuration of hardware 100. FIG. 6 is a diagram illustrating a shared memory management table 201. FIG. It is a figure showing the procedure of the process which the hypervisor 200 allocates a storage area. It is a figure showing the mode of allocation of a shared memory. It is a figure showing the mode of allocation of a shared memory. It is a figure showing the mode of allocation of a shared memory. It is a figure showing the mode of allocation of a shared memory.

The embodiment will now be described.
(1) Configuration FIG. 1 is a diagram conceptually showing the configuration of the information processing apparatus 1 of the present embodiment. The information processing apparatus 1 is incorporated in an image forming apparatus as a control device that controls the image forming apparatus, for example.
FIG. 2 is a diagram illustrating the configuration of the hardware 100. The hardware 100 includes a CPU (Central Processing Unit) 101, a ROM (Read Only Memory) 102, a RAM (Random Access Memory) 103, a storage unit 104, an operation unit 105, and a display unit 106.

The RAM 103 is a main storage device, and is used as a work area for writing the program body and data to be processed when the CPU 101 executes various programs.
The storage unit 104 is an auxiliary storage device such as a hard disk storage device, and stores a hypervisor 200, a first OS (Operating System) 300, a second OS 400, application programs, and the like. The ROM 102 stores a program representing a procedure for reading and executing the hypervisor 200 and the OS.
The operation unit 105 is a keyboard and a mouse. The display unit 106 is a liquid crystal display device, for example.

The hypervisor 200 is a program that performs control for operating a plurality of OSs in parallel on the hardware 100. The first OS 300 and the second OS 400 operate in parallel under the control of the hypervisor, and the first process 310 and the second process 410 are under the control of the first OS 300 and the second OS 400, respectively. Operate.
The hypervisor 200 sets a part of the storage area of the RAM 103 as a shared memory accessible from a plurality of processes, and assigns the shared memory to each process. The hypervisor 200 has the shared memory management table 201 shown in FIG. 3, and stores the top address and capacity of the area allocated to each process in the shared memory in association with the process identifier in the shared memory management table 201. To do.

  Although an example in which two OSs are operated in parallel is shown here, any number of OSs may be operated in parallel. Further, the number of processes operating under the control of each OS is not limited. Each process may be a process as a unit of processing executed by each OS, or may be a process as a unit of processing executed by an application program operating under the control of each OS. Further, an OS other than the first OS 300 and the second OS 400 may be operated on the hardware 100, and the hypervisor 200 may be operated under the control of the OS.

  The first OS 300 is an OS that does not have a virtual storage function, and is, for example, a real-time OS or a device driver incorporated in the OS. In a virtual machine, a free area may become fragmented because a storage area allocated to a certain OS among a plurality of OSs is released. In this case, the physical address of the free area is discontinuous. Accordingly, when the first OS 300 requests allocation of a storage area having a capacity larger than each fragmented free area, the first OS 300 is free even if the fragmented free area is allocated to the first OS 300. The area cannot be accessed correctly.

  On the other hand, the second OS 400 is an OS having a virtual storage function. The second OS 400 has an address conversion table 401 that stores a physical address and a virtual address on the RAM 103 in association with each other. In the virtual address space, the head address to the tail address are continuous. When starting the second process 410 that operates under the control of the second OS 400, the hypervisor 200 allocates a free area on the RAM 103 to the second process 410. At this time, the physical address and virtual address of the free area are allocated. Addresses are written to the address conversion table 401 in a one-to-one correspondence from the lower address side. When accessing the RAM 103, the second process 410 converts the virtual address into a physical address based on the address conversion table 401, and accesses the storage area corresponding to the physical address.

(2) Operation FIG. 4 is a diagram illustrating a procedure of processing in which the hypervisor 200 allocates a storage area to a process. Here, the CPU 101 is executing the hypervisor 200, the first OS 300, and the second OS 400. When a process is newly activated under the control of any OS and the process requests the CPU 101 to allocate a certain amount of shared memory, the CPU 101 executes the following processing. That is, the processing shown below is processing realized by the CPU 101 executing the hypervisor 200, and the physical subject of the operation is the CPU 101. However, for convenience, the hypervisor 200 will be described as the subject of the operation. .

In step A01, the hypervisor 200 inquires of the process or the OS whether the OS that controls the process that requested the allocation of the shared memory has a virtual storage function. If the OS has a virtual storage function (step A01: YES), the process proceeds to step A08. If the OS does not have a virtual storage function (step A01: NO), the process proceeds to step A02.
In short, the CPU 101 receives a request for allocation of the shared memory of the main storage device from the first process, and an OS (Operating System) that controls the first process when the request is received by the receiving means. 3 is an example of a determination unit that determines whether or not has a virtual storage function.

  In step A02, the hypervisor 200 refers to the shared memory management table 201 and identifies the start address and capacity of an area (free area) that is not assigned to any process. When the free area is fragmented, a plurality of sets of head address and capacity are specified. Then, the capacities of the free areas are summed, and it is determined whether or not the sum of the free capacities is equal to or greater than the request amount from the process. If the total free capacity is equal to or greater than the requested amount (step A02: YES), the process proceeds to step A03. If not greater than the requested amount (step A02: NO), it is indicated that the shared memory cannot be allocated. The process is notified and the process ends.

In step A03, the hypervisor 200 extracts continuous free areas of physical addresses having a capacity equal to or larger than the requested amount based on the capacity specified in step A02. When a continuous free area of physical addresses having a capacity larger than the requested amount is extracted (step A03: YES), the process proceeds to step A04, and when such a free area is not extracted (step A03: NO). The process proceeds to step A05.
In short, when the CPU 101 determines that the OS does not have a virtual storage function by the determination unit, the CPU 101 specifies a free area that is not allocated to any process in the shared memory, and starts from the first process. It is an example of the extraction means which extracts the continuous empty area | region of the physical address which has the capacity | capacitance more than the required amount of.

  In step A04, the hypervisor 200 allocates, to the process, a free area with the lowest start address among consecutive free areas of physical addresses having a capacity larger than the requested amount. Specifically, the hypervisor 200 associates the start address and capacity of the free area with the identifier of the process, writes them in the shared memory management table 201, and notifies the process of the start address.

  On the other hand, in step A05, the hypervisor 200 adds the capacity of the free area from the lower address side until the sum exceeds the required amount. Specifically, the sum of the capacities of the empty area with the lowest start address and the empty area with the second lowest start address is obtained. If the calculated sum is not equal to or greater than the required amount, the capacity of the empty area whose head address is the third lowest is added to the sum, and this procedure is repeated until the sum of the capacities exceeds the required amount. Then, out of the areas already allocated to the process, an area having a head address lower than the empty area to which the capacity is added last is extracted. The extracted area is called a movement required area. Then, the process to which the required movement area is allocated is specified from the shared memory management table 201, and the movement of the storage contents of the required movement area is notified to the specified process.

  In step A06, the hypervisor 200 waits until it receives a notification of acceptance from the process that has been notified of the movement of the stored contents. If the notification is received (step A06: YES), the process proceeds to step A07. The process notified of the movement of the stored contents notifies the CPU 101 of the approval of the movement as soon as the stored contents can be moved.

In step A07, the hypervisor 200 writes a copy of the storage contents of the area to be moved continuously from the start address of the shared memory. When the writing is completed, the hypervisor 200 ends the process in step A05 from the next address at the end of the written area. The memory contents up to the last address of the free area plus the capacity are deleted, and the start address and capacity of the area where the memory contents are erased are associated with the identifier of the process that requested the allocation of shared memory, and the shared memory management Write to the table 201 and notify the process of the start address. Further, the start address of the destination area is notified to the process whose storage contents are to be moved, and the process is terminated.
In short, the CPU 101 has a capacity larger than the required amount from the first process when the continuous free area of the physical address having a larger capacity than the required amount from the first process is not extracted by the extracting unit. Writing means for writing the storage contents of a storage area allocated to a process other than the first process to another storage area in a shared memory, so that a continuous free area of physical addresses having The storage contents obtained by the writing means are notified to the first process of the location of the continuous free area of the physical address having a capacity equal to or larger than the requested amount from the first process. This is an example of notification means for notifying the process corresponding to the stored contents of the storage location.

In step A08, the hypervisor 200 refers to the shared memory management table 201, identifies the start address and capacity of the free area, and assigns the free area to the process from the lower address side. Specifically, the free space start address and capacity are written in the shared memory management table 201 in association with the process identifier, and the start address is notified to the process. If the free space is fragmented, the request amount may not be satisfied with only the free space with the lowest start address, but in this case, the request amount of the fragmented free regions from the low address side Assign in order until it meets. When the assignment is completed, the process is terminated.
The above is the procedure of the process in which the hypervisor 200 allocates a storage area to the process.

Next, a specific example of assignment performed according to the above procedure will be described.
FIG. 5 is a diagram illustrating how the shared memory 103A is allocated. In this example, the hypervisor 200 manages the storage area of the shared memory 103A as 4 kB (kilobyte) as one block, and the maximum usable capacity as the shared memory 103A is 2 MB (megabyte), that is, 512 blocks. This figure shows a state in which 256 blocks are allocated to the first process 310 from the top. On the RAM 103, a storage area different from the shared memory 103A is allocated as a dedicated storage area for each OS and each process, but only the storage area set in the shared memory 103A is shown here.

Next, it is assumed that the second process 410 requests the hypervisor 200 to allocate 16 blocks. Then, as shown in FIG. 6, 16 blocks are allocated with the next block at the end of the block already allocated to the first process 310 as the head.
Next, it is assumed that the first process 310 notifies the hypervisor 200 of the release of the allocated block. Then, as shown in FIG. 7, 256 blocks are released from the beginning, but the free space is fragmented into a portion of 256 blocks from the beginning and a portion of 240 blocks from the end by the block allocated to the second process 410. It will be converted.

  Next, it is assumed that the first process 310 requests the hypervisor 200 to allocate 384 blocks. In this case, since the first OS that controls the first process 310 does not have a virtual storage function, when two free areas fragmented as shown in FIG. 7 are allocated, the physical addresses of the two free areas are consecutive. Therefore, the first process 310 cannot correctly access the shared memory 103A. Therefore, as shown in FIG. 8, the hypervisor 200 moves the storage contents of 16 blocks allocated to the second process 410 to the first 16 blocks of the shared memory 103A, and as a result, 496 with physical addresses consecutive. Get free space in the block. Then, 384 blocks from the head of this empty area are allocated to the first process 310. In the allocated area, physical addresses are continuous.

(3) Modifications Modifications shown below may be combined with each other.
(3.1) Modification 1
In the embodiment, the second OS 400 has the virtual storage function. However, all the OSs operated under the control of the hypervisor 200 may have a configuration that does not have the virtual storage function.
(3.2) Modification 2
In the embodiment, the stored contents are moved as much as necessary to secure the requested amount from the process. However, all the stored contents in the areas already allocated to other processes may be moved. In the embodiment, the storage content of the required movement amount area is moved to the lower side of the physical address, but may be moved to the higher side. In short, the storage contents of an area already allocated to another process may be moved so that a continuous storage area of physical addresses having a capacity larger than the amount required by the process can be obtained.

(3.3) Modification 3
In the embodiment, the head address and the capacity of the area allocated to the process are stored in the shared memory management table 201. However, the head address and the tail address may be stored.
(3.4) Modification 4
You may abbreviate | omit the process of step A02 of embodiment.
(3.5) Modification 5
In the embodiment, the hypervisor 200 inquires of the process whether or not the OS that controls the process has the virtual memory function. However, when the process requests allocation of the shared memory, the OS has the virtual memory function. Information indicating whether or not the request may be included in the request.
(3.6) Modification 6
In the embodiment, the example in which the CPU executes the process related to the allocation of the shared memory by executing the program has been described. However, the same function may be mounted on the electronic circuit.

DESCRIPTION OF SYMBOLS 1 ... Information processing apparatus, 100 ... Hardware, 101 ... CPU, 102 ... ROM, 103 ... RAM, 103A ... Shared memory, 104 ... Memory | storage part, 105 ... Operation part, 106 ... Display part, 200 ... Hypervisor, 300 ... 1st OS, 400 ... 2nd OS, 310 ... 1st process, 410 ... 2nd process, 201 ... Shared memory management table, 401 ... Address conversion table

Claims (3)

Receiving means for receiving a request for allocation of the shared memory of the main storage device from the first process;
A determination unit that determines whether an OS (Operating System) that controls the first process has a virtual storage function when a request is received by the reception unit;
When the determination unit determines that the OS does not have a virtual storage function, it identifies a free area that is not allocated to any process in the shared memory, and exceeds the amount requested from the first process. Extraction means for extracting continuous free space of physical addresses having a capacity;
When a continuous free area of a physical address having a capacity larger than the requested amount from the first process is not extracted by the extracting unit, a physical address having a larger capacity than the requested amount from the first process is extracted. Writing means for writing storage contents of a storage area allocated to a process other than the first process to another storage area in the shared memory so that a continuous free area can be obtained;
Notifying the first process of the location of consecutive free areas of physical addresses having a capacity equal to or greater than the requested amount from the first process, obtained by the writing means, and written by the writing means An information processing apparatus comprising: a notification unit configured to notify a storage location of stored content to a process corresponding to the stored content. The extraction unit specifies a free area that is not allocated to any process in the shared memory when the determination unit determines that the OS does not have a virtual storage function, and The continuous free area of physical addresses having a capacity equal to or greater than the requested amount from the first process is extracted when the total capacity is equal to or greater than the requested amount from the first process. The information processing apparatus according to 1. Computer
Receiving means for receiving a request for allocation of the shared memory of the main storage device from the first process;
A determination unit that determines whether an OS (Operating System) that controls the first process has a virtual storage function when a request is received by the reception unit;
When the determination unit determines that the OS does not have a virtual storage function, it identifies a free area that is not allocated to any process in the shared memory, and exceeds the amount requested from the first process. Extraction means for extracting continuous free space of physical addresses having a capacity;
When a continuous free area of a physical address having a capacity larger than the requested amount from the first process is not extracted by the extracting unit, a physical address having a larger capacity than the requested amount from the first process is extracted. Writing means for writing storage contents of a storage area allocated to a process other than the first process to another storage area in the shared memory so that a continuous free area can be obtained;
Notifying the first process of the location of consecutive free areas of physical addresses having a capacity equal to or greater than the requested amount from the first process, obtained by the writing means, and written by the writing means A program for functioning as a notification means for notifying a process corresponding to a storage content of the storage content.

Images