Techniques for accessing a shared resource using an improved synchronization mechanism

• US 7,644,409 B2
• Filed: 06/04/2004
• Issued: 01/05/2010
• Est. Priority Date: 06/04/2004
• Status: Active Grant

• Current Assignee: Oracle America, Inc. (Oracle Corporation)
• Original Assignee: Sun Microsystems Incorporated (Oracle Corporation)
• Inventors: Yang, Mingyao, Huang, Hui, Dice, David
• Primary Examiner(s): An; Meng-Ai
• Assistant Examiner(s): Kawsar; Abdullah Al
• Application Number:
  US10/861,795
• Publication Number:
  US 20060031844A1
• Time in Patent Office: 2,041 Days
• Field of Search: 718 1-108, 717140-151, 717/106, 717159-161, 711/100, 711/130, 711/147, 711152-155, 711/163, 711165-173, 719/312
• US Class Current: 718/104
• CPC Class Codes: G06F 9/52 Program synchronisation; Mutual exclusion, e.g. by means of semaphores

Abstract

A technique for accessing a shared resource of a computerized system involves running a first portion of a first thread within the computerized system, the first portion (i) requesting a lock on the shared resource and (ii) directing the computerized system to make operations of a second thread visible in a correct order. The technique further involves making operations of the second thread visible in the correct order in response to the first portion of the first thread running within the computerized system, and running a second portion of the first thread within the computerized system to determine whether the first thread has obtained the lock on the shared resource. Such a technique alleviates the need for using a MEMBAR instruction in the second thread.

Claims

1. A method for accessing a shared resource of a computerized system, the method comprising:

running a first portion of a first thread within the computerized system, the first portion (i) requesting a lock on the shared resource and (ii) directing the computerized system to make operations of a second thread visible in a correct order, the correct order preventing an exclusion failure, the first portion of the thread representing at least instructions that perform the steps of requesting and directing;

making operations of the second thread visible in the correct order in response to the first portion of the first thread running within the computerized system, the second thread visible to at least one other processor; and

running a second portion of the first thread within the computerized system to determine whether the first thread has obtained the lock on the shared resource, the second portion of the thread representing at least one instruction that performs the step of determining whether the first thread has obtained the lock on the shared resource;

wherein running the first portion of the first thread includes;

performing a store operation which stores a lock request value in a memory variable which is accessible by both the first and second threads;

performing a memory barrier operation which guarantees execution of the store operation before a next operation; and

performing, as the next operation, a system call which directs the computerized system to make operations of the second thread visible in the correct order;

wherein performing the system call includes;

executing operating system code to change protection of a page of memory of the computerized system.

2. The method of claim 1 wherein running the second portion of the first thread includes:

performing a load operation which loads contents of another memory variable which is accessible by both the first and second threads to enable the first thread to determine whether first thread has obtained the lock on the shared resource, the first thread obtaining the lock on the share resource when the contents of the other memory variable has a first value, and the first thread not obtaining the lock on the share resource when the contents of the other memory variable has a second value that is different than the first value.

3. The method of claim 1 wherein the second thread is configured to access the shared resource more frequently than the first thread, and wherein the method further comprises:

running the second thread without a memory barrier operation to reduce system overhead incurring during execution of the second thread.

4. The method of claim 1 wherein the computerized system is a Windows-based operating system platform, and wherein executing the operating system code includes:

carrying out a VirtualProtect() procedure within the Windows-based operating system platform.

5. The method of claim 1 wherein the computerized system is a UNIX-based operating system platform, and wherein executing the operating system code includes:

carrying out an mprotect() procedure within the UNIX-based operating system platform.

6. A computerized system, comprising:

an interface configured to provide external access to the computerized system;

a shared resource; and

control circuitry coupled to the interface and the share resource, the control circuitry, when accessing the shared resource, being configured to;

run a first portion of a first thread within the computerized system, the first portion (i) requesting a lock on the shared resource and (ii) directing the computerized system to make operations of a second thread visible in a correct order, the correct order preventing an exclusion failure, the first portion of the thread representing at least instructions that perform the steps of requesting and directing;

make operations of the second thread visible in the correct order in response to the first portion of the first thread running within the computerized system, the second thread visible to at least one other processor; and

run a second portion of the first thread within the computerized system to determine whether the first thread has obtained the lock on the shared resource, the second portion of the thread representing at least one instruction that performs the step of determining whether the first thread has obtained the lock on the shared resource;

wherein the control circuitry, when running the first portion of the first thread, is configured to;

perform a store operation which stores a lock request value in a memory variable which is accessible by both the first and second threads;

perform a memory baffler operation which guarantees execution of the store operation before a next operation; and

perform, as the next operation, the system call which directs the computerized system to make operations of the second thread visible in the correct order;

wherein the control circuitry, when performing the system call, is configured to;

execute operating system code to change protection of a page of memory of the computerized system.

7. The computerized system of claim 6 wherein the control circuitry, when running the second portion of the first thread, is configured to:

perform a load operation which loads contents of another memory variable which is accessible by both the first and second threads to enable the first thread to determine whether first thread has obtained the lock on the shared resource, the first thread obtaining the lock on the share resource when the contents of the other memory variable has a first value, and the first thread not obtaining the lock on the share resource when the contents of the other memory variable has a second value that is different than the first value.

8. The computerized system of claim 6 wherein the second thread is configured to access the shared resource more frequently than the first thread;

and wherein the control circuitry is further configured to;

run the second thread without a memory baffler operation to reduce system overhead incurred during execution of the second thread.

9. The computerized system of claim 6 wherein the computerized system is a Windows-based operating system platform;

and wherein the control circuitry, when executing the operating system code, is configured to;

carry out a VirtualProtect() procedure within the Windows-based operating system platform.

10. The computerized system of claim 6 wherein the computerized system is a UNIX-based operating system platform;

and wherein the control circuitry, when executing the operating system code, is configured to;

carry out an mprotect() procedure within the UNIX-based operating system platform.

11. A computer program product that includes a computer readable medium having instructions stored thereon for accessing a shared resource of a computerized system, such that the instructions, when carried out by the computerized system, causes the computerized system to:

run a first portion of a first thread within the computerized system, the first portion (i) requesting a lock on the shared resource and (ii) directing the computerized system to make operations of a second thread visible in a correct order, the correct order preventing an exclusion failure, the first portion of the thread representing at least instructions that perform the steps of requesting and directing;

make operations of the second thread visible in the correct order in response to the first portion of the first thread running within the computerized system, the second thread visible to at least one other processor; and

run a second portion of the first thread within the computerized system to determine whether the first thread has obtained the lock on the shared resource, the second portion of the thread representing at least one instruction that performs the step of determining whether the first thread has obtained the lock on the shared resource;

wherein running the first portion of the first thread includes;

performing a store operation which stores a lock request value in a memory variable which is accessible by both the first and second threads;

performing a memory barrier operation which guarantees execution of the store operation before a next operation; and

performing, as the next operation, the system call which directs the computerized system to make operations of the second thread visible in the correct order;

wherein performing the system call includes;

executing operating system code to change protection of a page of memory of the computerized system.

12. The computer program product of claim 11 wherein running the second portion of the first thread includes:

performing a load operation which loads contents of another memory variable which is accessible by both the first and second threads to enable the first thread to determine whether first thread has obtained the lock on the shared resource, the first thread obtaining the lock on the share resource when the contents of the other memory variable has a first value, and the first thread not obtaining the lock on the share resource when the contents of the other memory variable has a second value that is different than the first value.

13. The computer program product of claim 11 wherein the second thread is configured to access the shared resource more frequently than the first thread, and wherein the computer program product when carrying out the instructions:

runs the second thread without a memory baffler operation to reduce system overhead incurred during execution of the second thread.

14. The computer program product of claim 11 wherein the computerized system is a Windows-based operating system platform, and wherein executing the operating system code includes:

carrying out a VirtualProtect( ) procedure within the Windows-based operating system platform.

15. The computer program product of claim 11 wherein the computerized system is a UNIX-based operating system platform, and wherein executing the operating system code includes:

carrying out an mprotect() procedure within the UNIX-based operating system platform.