KR20110033066A - Fast speed computer system power-on & power-off method - Google Patents

Abstract

PURPOSE: A high computer system power on/off method is provided to drastically reduce the amount of a memory which is used in a system among hibernation, thereby increases the speed restoring to a full-speed operation. CONSTITUTION: A computer system divides a main memory into a clean and dirty pages when a computer system enters a hibernation. The clean pages swap outs without I/O access. The dirty pages is stored in a swap space and a file system of a secondary storing device and records in a hibernation file(S30). If the computer system reactivates, the computer system restores the hibernation file to the main memory from a swap space and a file system(S32).

Description

High-speed computer system power on and power off method {FAST SPEED COMPUTER SYSTEM POWER-ON & POWER-OFF METHOD}

The present invention relates to a computer system power-on and power-off method, and more particularly, to a high-speed computer system power-on and power-off method applicable when a computer system enters a power saving mode.

In general, the operational state of a typical computer system includes a power saving mode, in addition to the normal mode of running the operating system and various application programs. The purpose of the computer entering hibernation is to save energy on the one hand and to reduce the noise of the system on the one hand.

According to the design classification of the Advanced Configuration and Power Interface (ACIP), there are two types of hibernation for computer systems: one of them is in the Suspend-To-RAM (STR) state (also in the S3 state). The other is in the Spend-To-Disk (STD) state (also called the S4 state). In the S3 state, the computer system only continues to supply power to a portion of volatile memory including frame buffers, main memory, and the like, and cuts power to the remaining portions. There are two advantages to the S3 state: one of these is that it will be faster for the computer system to return to full speed operation; The other is that the S3 state is used for better security when the user is operating private and secret data and does not want to store it on the hard disk. Another type of hibernation is a suspended-to-disk (STD) state (also called an S4 state). In the S4 state, the data in operation is all written to non-volatile memory for storage, and then the whole computer system is powered off, so the advantage of the S4 state is better power savings.

However, the effects and performances of the above described states S3 and S4 of hibernation are not very satisfactory. Because when the computer system enters the S3 state, power must continue to be provided to volatile memory within the computer system to maintain the system storage ecology. If the S4 state is employed, the rate at which the computer system returns to full-speed operation is much slower than the speed of the S3 state, although more power is saved than the S3 state.

In view of the problems and drawbacks of the prior art, the present invention discloses a high speed computer system power on and power off method, such that when the computer system enters hibernation, the memory used is reduced so that the computer system runs at full speed operation. It is possible to increase the reaction rate and the efficiency of reactivation.

SUMMARY OF THE INVENTION The primary object of the present invention is to provide a high speed computer system power on and power off method such that the amount of memory used in the system can be drastically reduced when the computer system enters hibernation, thereby reducing the main memory to secondary storage device. By reducing the amount of data that is transferred, it increases the speed of the computer system returning to full speed operation.

It is another object of the present invention to provide a high speed computer system power on and power off method such that the degree of data loss and corruption can be reduced during the temporary powering of the computer system.

It is a further object of the present invention to provide a high speed computer system power on and power off method wherein data recording and data retrieval are performed in a random access manner such that the computer system enters hibernation, or is full speed. When reactivated with the diagram operation, the data access speed can be greatly increased.

In order to achieve the above object, the present invention provides a high-speed computer system power on and power off method, so that when the computer system enters hibernation, at least a portion of the main memory included in the computer system is a plurality of clean pages (clean pages) and a plurality of dirty pages, the clean pages are swapped out and discarded storage, the dirty pages are written to a hibernation file and swap at least of secondary storage devices Stored in space and file system. Therefore, when the computer system is reactivated, these dirty pages are reloaded from the swap space and file system into the hibernation file and vice versa into the main memory so that the computer system retrieves the necessary data from the secondary storage device. It can then be read and loaded back into the main memory. As such, the computer system may reduce the amount of data transferred and the amount of data stored from the main memory to the secondary storage device in speeding up computer system reactivation and return to maximum operating speed. Can be.

Further scope of applicability of the present invention will become apparent from the detailed description provided hereinafter. However, it should be understood that the detailed description and specific examples represent preferred embodiments of the invention, which are provided by way of example only, since various modifications and modifications within the spirit and scope of the invention will be apparent to those skilled in the art from this description. Because it will be clear.

Due to the configuration as described above, the present invention allows the amount of memory used in the system to be drastically reduced when the computer system enters hibernation, thereby reducing the amount of data transferred from the main memory to the secondary storage device. Reducing the speed of the computer system returning to full speed operation. In addition, the present invention can reduce the degree of data loss and corruption during the temporary powering of the computer system, and increase the data access speed when the computer system enters hibernation, or when reactivated at full speed operation. Can be.

The relevant drawings in connection with the following detailed description of the invention, which is made hereafter, are described as follows:
1 is a schematic diagram of a high speed computer system power on and power off device in accordance with the present invention; And
2 is a flow chart of the steps of a high speed computer system power on and power off method in accordance with the present invention.

The objects, configurations, features, functions and advantages of the present invention may be more thoroughly recognized and understood through the following detailed description with reference to the accompanying drawings.

In the present invention, a high speed computer power on and power off method is proposed, so that the amount of memory used decreases when the computer system enters hibernation, thereby reducing the amount of data written to the secondary storage device and entering hibernation. Speed up the computer system, or reactivate or recover to full speed operation. In the following, preferred embodiments are described to explain the technical features of the present invention. First, reference is made to FIG. 1 which shows a schematic diagram of a high speed computer system power on and power off device of the present invention. As shown in FIG. 1, computer system 10 includes a central processing unit (CPU) 12, a main memory 14, and a secondary storage device 16. Secondary storage device 16 may be a fast random access memory (RAM) device, such as flash memory, and is primarily for file system 18, swap space 20, and hibernation file 22. Is used.

Next, refer to FIG. 2 for a flow chart of the steps of the high speed computer system power on and power off method of the present invention, while referring to FIG. 1. As shown in FIG. 2, first, as shown in step S30, the operating system OS of the computer system 10 enters the hibernation mode and uses the section or page in units of main memory 14. ) Is divided into clean pages and dirty pages. Since the copy having the same contents as the clean pages is stored in the secondary storage device 16, the swapped out clean pages may be discarded storage. Therefore, a kernel function is used to allocate main memory 14 and to control the swap out of clean pages. The kernel function may allocate main memory 14 based on the amount of clean pages needed to swap out and will release main memory after swapping out the clean pages. Dirty pages in the machine state are stored in the hibernation file 22 and then in the swap space 20 and file system 18 of the secondary storage device 16. Here, dirty pages can be stored according to their addresses, and dirty pages of consecutive addresses can be merged into a single write command. Then, in step S32, when the computer system 10 is reactivated, the computer system 10 reads the hibernation file 22 from the swap space 20 and the file system 18 and stores the hibernation file in main memory 14. Will restore the system state. Finally, as shown in S34, computer system 10 will read data from secondary storage device 16 and load it back into main memory 14.

The speed at which a computer system performs a Suspend-To-Flash function and stops sending data to a secondary storage device 16, such as flash memory, is often recorded in nonvolatile memory. It depends on the amount of data you need. Swapping-Before-Hibernating is implemented by the operating system (OS) through its existing memory management technology, which allows most of the memory pages to be discarded directly without performing any write operations, which leads to suspend speed. Can be increased. As soon as the execute function is executed, the data can be retrieved in reverse from each of the three locations: the hibernation file 22, the swap space 20, and the file system 18. The main data and code of the system belonging to the kernel will be reversed immediately upon reactivation of the computer system, and the rest of the data will be paging-on-demand for the portion of the user's current needs. Is searched backwards. Since the amount of data required by the user is much less than all the amount of data in main memory before the suspension of the system, the reproduction speed can be increased.

In general, memory pages can be classified into three categories, respectively: as follows: free page, anonymous page, and named page. Here, the free pages are memory not currently used by the system, so the contents of the free pages are meaningless to the system. Anonymous pages are memory that is dynamically allocated during the execution of a program, and portions of this memory mainly include: the stack and the heap. A named page is a copy of a file in main memory 14, whose behavior is similar to that of cache memory in secondary storage device 16. A named page includes: an executable file and a dynamic-link library corresponding to specific locations in memory, and the program uses the memory mapping file to map the file into memory. Anonymous pages and named pages can both be clean pages or dirty pages. When dirty anonymous pages are swapped out, they must be written to swap space 20, and pages with dirty names are swapped out into file system 18. The operating system (OS) needs to swap out the clean page because the page must have a copy of the exact same contents as in the secondary storage device 16, so the operating system (OS) discards this page directly. will be.

When the system starts to perform a suspend-to-flash, most of the pages do not need to be written back to the secondary storage device 16 because most of these pages are clean pages. Some of these pages must be written back to swap space 20 and file system 18, and the remaining pages are non-swqppable memory, and these pages are primarily the operating system kernel or some of the pages Set by the program as non-swappable due to performance causes. Finally, non-swappable memory will be written to the hibernation file 22.

Upon restarting the computer system, upon completion of routine hardware initialization and loading into the OS loader, the system will determine whether to return to normal from hibernation or to perform a normal power-on procedure. In the former case, the data in the hibernation file 22 will first be loaded into main memory 14. As soon as this operation was completed, the OS completed the basic power-on procedure. Then, according to user requirements, the computer system reversely loads the data in file system 18 and swap space 20 into main memory 14 and returns to the original system state. The swapping non-pore hibernate function needs to be executed again because some of the dirty pages have been written to swap space 20 during the last swapping non-pore hiberating function and these pages become clean pages. Therefore, the CPU 12 will not perform the write operation for these pages, so these pages do not need to be written back to the swap space 20.

Flash memory serving as secondary storage device 16 may have performance close to continuous access in random access through the use of optimization techniques. Through the writing-combining function, small random writes that result during swapping out of memory pages are queued before they are actually written to flash memory, and these pages are first sorted according to their physical addresses. It can then be recognized whether contiguous range complete writes are present in these swapped out pages. If the response is affirmative, a large write request operation is used to speed up the swapping non-pore hibernation function by replacing a plurality of small write request operations.

In the process of writing swappable pages to secondary storage device 16, a large amount of memory is stored in the operating system in order to prevent the OS's optimization operation from affecting the amount of memory actually allocated when allocating main memory. (OS) may be requested. Therefore, once a page is allocated, data bytes are written into the page by causing the OS to immediately allocate physical memory to the program. This forced action causes the OS to swap out a large portion of memory. Subsequently, free operations are performed to return all memory requested and obtained from the system back to the OS.

Through implementations of the two steps described above, since most pages in the operating system (OS) will be free pages, these free pages do not need to be written back to the hibernation file 22.

In addition, the request can be made directly from the operating system (OS) kernel to allocate memory, so that the requested pages can be allocated immediately via the kernel function while requiring additional write operations to actually allow the OS to allocate memory. Do not do it. As soon as the allocation of memory is completed, free operation will be performed immediately and the system will generate a large amount of free pages. Since memory allocation is done through kernel functionality, no additional write operations are needed. Therefore, since the number of swapped out pages can be calculated and obtained in advance, it can calculate how much memory is needed for allocation.

Moreover, the swapper mechanism in the operating system (OS) can be manipulated directly to allocate the required memory. In this case, the Linux operating system (OS) may be employed as an example, and its kernel is provided with a memory management function, that is, a Shrink_All_Memory function. This function is used to retrieve memory pages, and as soon as they are called, they can be loaded with the amount of pages to be retrieved, so that memory pages can be freed and released without affecting the stability of the system and programs used by users. Can be. The mechanism of the Shrink_All_Memory function is implemented through two least recently used (Least-Recently-Used) lists in the kernel: active list and inactive list. Wherein the active list includes pages that have not been accessed most recently, and the inactive list includes pages that have not been accessed for a certain period of time. In practical applications, the Shrink_All_Memory function will first start searching pages from the inactive list and then retrieve the pages from the active list. Therefore, through the application of the Shrink_All_Memory function, pages can be swapped as needed without affecting the operation of the kernel of the systems and programs used by users.

From the kernel's point of view in the Linux operating system (OS), the secondary storage device 16 is considered a normal input / output device and called when memory pages are swapped out in the swap space 20. Parameters received by the Submit_Bio function that are being received include: sector number, read / write commands, memory address, memory length, etc., so that pages that are swapped out can be further processed by intercepting the Submit_Bio function. .

As discussed above, pages that are swapped out are determined by the Shrink_All_Memory function as they are based on the usage state of the page and not based on its physical address in main memory 14. In the application, first queue the write requests generated by the Shrink_All_Memory function, then sort these pages according to physical memory address, reducing the number of input / output accesses and increasing the size of the input / output request. The larger the requested size, the better the write performance of the flash memory serving as secondary storage device 16. When a memory page is swapped out, the kernel must write the location of the page in swap space 20, so the kernel can reload this page from secondary storage device 16 to main memory 24. Each time the write combining function rewrites the Shrink_All_Memory function to perform the Shrink_All_Memory function, the position of each write page in the swap space 20 must be authenticated, so this information becomes the swapped out page identifier in the page table.

The foregoing description of the preferred embodiment is intended to more clearly describe the features and advantages of the present invention. However, the disclosed preferred embodiments are not intended with any limitations on the scope of the invention. On the contrary, it is an object of the present invention to include various changes and equivalents falling within the scope of the appended claims.

Claims (9)

In a high speed computer system power on and power off method:
When the computer system enters hibernation, the computer system splits main memory into a plurality of clean pages and a plurality of dirty pages, the clean pages are swapped out without I / O access, and the dirty pages are hibernation files. Written into and stored in the swap space and file system of at least the secondary storage device;
When the computer system is reactivated, the computer system reading the hibernation file from the swap space and the file system and restoring the hibernation file to the main memory; And
And the computer system reading data from the secondary storage device and loading the data into the main memory. The method of claim 1,
And wherein said clean pages are swapped out and said main memory is released when said computer system enters said hibernation. The method of claim 2,
In the step of the computer system entering the hibernation, the main memory is provided with a kernel function, the kernel function is used to control the swap out of the clean pages, and after completion of the swap out of the clean pages, the main memory. A method of powering on and powering off a high speed computer system, comprising: releasing. The method of claim 3, wherein
And said kernel function is capable of calculating the number of said clean pages to be swapped out and allocating said main memory based on said number of said clean pages. The method of claim 1,
And said secondary storage device is a flash memory. The method of claim 1,
And said secondary storage device is a fast random access memory device. The method of claim 1,
In the step of entering the hibernation of the computer system, the computer system using a section as a unit in dividing the main memory. The method of claim 1,
And said secondary storage device is provided with coffee having its own content equal to that of said clean page. The method of claim 1,
When the computer system enters the hibernation, the swapped out dirty pages are classified according to their addresses, and the dirty pages of consecutive addresses are incorporated into the single write function and write combing to the secondary storage device. A high-speed computer system power on and power off method, characterized in that the writing (combing).

Images