Combining memory pages having identical content

• US 9,058,212 B2
• Filed: 03/21/2011
• Issued: 06/16/2015
• Est. Priority Date: 03/21/2011
• Status: Active Grant

• Current Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Original Assignee: Microsoft Technology Licensing LLC (Microsoft Corporation)
• Inventors: Wang, Landy, Iyigun, Mehmet, Bak, Yevgeniy, Sundaram, Vijay, Lebedynskiy, Pavlo L.
• Primary Examiner(s): BIRKHIMER, CHRISTOPHER D
• Application Number:
  US13/052,562
• Publication Number:
  US 20120246436A1
• Time in Patent Office: 1,548 Days
• Field of Search: None
• US Class Current: 1/1
• CPC Class Codes: G06F 12/109 for multiple virtual address spaces, e.g. segmentation G06F12/1036 takes precedence
• CPC Class Codes: G06F 9/5016 the resource being the memory

Abstract

In a device having a memory accessed as multiple pages, two or more pages of the multiple pages having identical content are identified. While the two or more pages are being identified, other processes running in the device are allowed to use the two or more pages, including being allowed to change cache attributes of each of the two or more pages. The two or more pages are combined into a single combined page (e.g., a newly allocated page of the multiple pages), and a process page record having multiple entries pointing to the multiple pages is updated so that entries that previously pointed to one of the two or more pages instead point to the single page.

Claims

1. A method in a device having a memory accessed as multiple pages, the method comprising:

identifying two or more pages of the multiple pages having identical content, the identical content including at least the two or more pages having the same page protections, the same page protections identifying how the two or more pages may be used including at least one of read, write, or execute, the identifying the two or more pages including at least generating a check value for each of the multiple pages;

allowing, while the two or more pages are being identified, other processes running on the device to use the two or more pages including allowing the other processes to change cache attributes of each of the two or more pages;

combining the two or more pages into a single combined page; and

updating a process page record having multiple entries pointing to the multiple pages so that entries that previously pointed to one of the two or more pages instead point to the single combined page, the other processes being unaware of the identifying, combining, and updating while the two or more pages are being identified, the process page record including at least an identifier for each of the multiple pages indicating when the generating the check value is being done.

2. A method as recited in claim 1, the allowing comprising:

ceasing analyzing a page to determine whether the page may have identical content as one or more other pages of the multiple pages in response to a request that can potentially result in a memory coherency or corruption situation being received while analyzing the page; and

allowing the request to be performed.

3. A method as recited in claim 1, further comprising performing the identifying, allowing, combining, and updating in response to an indication from a user mode page combine policy module to analyze the multiple pages.

4. A method as recited in claim 3, the user mode page combine policy module generating the indication based at least in part on various aspects of the device that are monitored by the user mode page combine policy module.

5. A method as recited in claim 1, the combining comprising allocating a new page of the memory as the single combined page, and the updating the process page record comprising updating the process page record so that each entry that previously pointed to one of the two or more pages instead points to the new page.

6. A method as recited in claim 5, the new page being allocated in a way that reduces physical memory fragmentation.

7. A method as recited in claim 1, the check value being a cyclic redundancy check (CRC) value, the identifying comprising:

identifying ones of the multiple pages having matching CRC values; and

comparing the ones of the multiple pages having matching CRC values to determine which of the ones, if any, have identical content.

8. A method as recited in claim 7, the multiple pages including previously combined pages.

9. A method as recited in claim 8, the previously combined pages including one or more previously combined pages that have been transferred to a storage device.

10. A method as recited in claim 1, further comprising, in response to a request from one of the other processes to write to a memory address included in one of the two or more pages:

allocating a new page in the memory;

copying the content of the single combined page to the new page; and

updating the process page record so that an entry for the process from which the request was received that previously pointed to the combined page instead points to the new page.

11. A method as recited in claim 10, further comprising

repeating the allocating, copying, and updating the process page record in response to each of multiple additional write requests to memory addresses included in one of the two or more pages;

and

de-allocating the combined page when a reference count indicates that no more process page record entries point to the combined page.

12. One or more computer storage memories having stored thereon multiple instructions that, when executed by one or more processors of a device having a memory accessed as multiple pages, implement a page combination procedure by causing the one or more processors to:

identify two or more pages of the multiple pages having identical content, the identical content including at least the two or more pages having the same page protections, the same page protections identifying how the two or more pages may be used including at least one of read, write, or execute, the identify the two or more pages including at least generating a check value for each of the multiple pages;

allocate a new page of the multiple pages, the new page having the identical content as the two or more pages;

update a process page record having multiple entries pointing to the multiple pages so that entries that previously pointed to one of the two or more pages instead point to the new page, the process page record including at least an identifier for each of the multiple pages indicating when the generating the check value is being done; and

cause one or more additional processes running on the device other than a process implementing the page combination procedure to be unaware of the identification, allocation, and updating.

13. One or more computer storage memories as recited in claim 12, wherein the check value is a cyclic redundancy check (CRC) value and to identify two or more pages is to:

identify ones of the multiple pages having matching CRC values; and

compare the ones of the multiple pages having matching CRC values to determine which of the ones, if any, have identical content.

14. One or more computer storage memories as recited in claim 13, the multiple pages including previously combined pages.

15. One or more computer storage memories as recited in claim 14, the previously combined pages including one or more previously combined pages that have been transferred to a storage device.

16. One or more computer storage memories as recited in claim 12, wherein the instructions further cause the one or more processors to, in response to a request to write to a memory address included in one of the two or more pages:

allocate an additional new page in the memory;

copy the content of the new page to the additional new page; and

update the process page record so that an entry for a process from which the request was received that previously pointed to the new page instead points to the additional new page.

17. One or more computer storage memories as recited in claim 12, the new page being allocated in a way that reduces physical memory fragmentation.

18. A method in a device having a memory accessed as multiple pages, the method comprising:

determining, in response to an indication received from a user mode page combine policy module, to perform a page combination procedure, the indication occurring in response to a user being away from the device for a threshold amount of time, an increase of a threshold amount of memory being used within a particular amount of time, or a rate at which new pages are being allocated exceeds a threshold amount, the threshold amount of memory based on a total number of pages comprising the multiple pages;

identifying, as part of the page combination procedure, two or more pages of the multiple pages having identical content, the identical content including at least the two or more pages having the same page protection, the multiple pages including previously combined pages that have been transferred to a storage device, the identifying the two or more pages including at least generating a check value for each of the multiple pages;

allowing, while the two or more pages are being identified, other processes running on the device to use the two or more pages including allowing the other processes to change cache attributes of each of the two or more pages;

combining, as part of the page combination procedure, the two or more pages into a single combined page by allocating a new page of the multiple pages and copying the content of one of the two or more pages to the new page; and

updating, as part of the page combination procedure, a process page record having multiple entries pointing to the multiple pages so that entries that previously pointed to one of the two or more pages instead point to the single combined page, the other processes being unaware of the identifying, combining, and updating while the two or more pages are being identified, the process page record including at least an identifier for each of the multiple pages indicating when the generating the check value is being done.

19. A method as recited in claim 18, wherein the new page comprises the same attributes and page protections as the two or more pages being combined.

20. A method as recited in claim 18, the new page being allocated in a way that reduces physical memory fragmentation.