Multiprocessor synchronization using region locks
Disclosed is a method of synchronizing a plurality of processors accesses to at least one shared resource. One of a plurality of processors requests an exclusive region lock for a shared resource using a logical block address (LBA) of a dummy target. The LBA is defined in a region map that associates LBAs to shared resources. The exclusive region lock request is inserted as a node in a region lock tree of the dummy target. Access to the shared resource is granted based on a determination whether there is an existing region lock in the region lock tree that is overlaps with the new exclusive region lock request.
1. A method of synchronizing a plurality of processors' accesses to at least one shared resource, comprising:
requesting, by one of said plurality of processors, an exclusive region lock for a shared resource using a logical block address (LBA) of a dummy target, said LBA defined in a region map that contains at least one LBA, the at least one LBA in said region map associated with said shared resource;
inserting said exclusive region lock request as a node in a region lock tree of said dummy target; and,
granting one of said plurality of processors access to said shared resource based on a determination whether there is a first existing region lock in the region lock tree that is overlapping with said exclusive region lock request.
2. The method of claim 1 , wherein all of said plurality of processors request exclusive region locks for said shared resource using the same LBA of said dummy target.
3. The method of claim 1 , wherein said region map associates shared resources with corresponding lock ranges.
4. The method of claim 1 , further comprising:
requesting, by one of said plurality of processors, an exclusive region lock for a second shared resource using a second LBA of said dummy target, said second LBA defined in said region map that contains a second at least one LBA, the second at least one LBA in said region map associated with said second shared resource.
5. The method of claim 4 , further comprising:
inserting said second exclusive region lock request as a second node in said region lock tree of said dummy target.
6. The method of claim 5 , further comprising:
granting one of said plurality of processors access to said shared resource based on a determination whether there is a second existing region lock on the region lock tree that is overlapping with said second exclusive region lock request.
7. The method of claim 5 , further comprising:
granting one of said plurality of processors access to said shared resource based on a determination whether there is a first existing region lock on the region lock tree that is overlapping with said exclusive region lock request and whether there is a second existing region lock on the region lock tree that is overlapping with said second exclusive region lock request.
8. The method of claim 1 , wherein said shared resource comprises nonvolatile memory.
9. The method of claim 1 , wherein said shared resource comprises resources used by input/output (I/O) operations.
10. A computer system, comprising:
a plurality of processors configured to access at least one shared resource by requesting an exclusive region lock for a shared resource using a logical block address (LBA) of a dummy target;
a region map containing said LBA and configured to contain a plurality of LBAs, said region map configured to associate said LBA with said shared resource;
a region lock tree of said dummy target configured to receive said exclusive region lock request as a node in said region lock tree of said dummy target; and,
said plurality of processors configured to receive access to said shared resource based on a determination whether there is a first existing region lock in the region lock tree that is overlapping with said exclusive region lock request.
11. The computer system of claim 10 , wherein all of said plurality of processors request exclusive region locks for said shared resource using the same LBA of said dummy target.
12. The computer system of claim 10 , wherein said region map associates shared resources with corresponding LBA ranges.
13. A computer readable medium having instructions stored thereon for synchronizing a plurality of processors accesses to at least one shared resource that, when executed by a computer, at least instruct the computer to:
request, by one of said plurality of processors, an exclusive region lock for a shared resource using a logical block address (LBA) of a dummy target, said LBA defined in a region map that contains at least one LBA, the at least one LBA in said region map associated with said shared resource;
insert said exclusive region lock request as a node in a region lock tree of said dummy target; and,
grant one of said plurality of processors access to said shared resource based on a determination whether there is a first existing region lock on the region lock tree that is overlapping with said exclusive region lock request.
14. The computer readable medium of claim 13 , wherein all of said plurality of processors request exclusive region locks for said shared resource using the same LBA of said dummy target.
15. The computer readable medium of claim 13 , wherein said region map associates shared resources with corresponding LBA ranges.
16. The computer readable medium of claim 13 , wherein the computer is further instructed to:
request, by one of said plurality of processors, an exclusive region lock for a second shared resource using a second LBA of said dummy target, said second LBA defined in said region map that contains a second at least one LBA, the second at least one LBA in said region map associated with said second shared resource.
17. The computer readable medium of claim 16 , wherein the computer is further instructed to:
insert said second exclusive region lock request as a second node in said region lock tree of said dummy target.
18. The computer readable medium of claim 17 , wherein the computer is further instructed to:
grant one of said plurality of processors access to said shared resource based on a determination whether there is a second existing region lock on the region lock tree that is overlapping with said second exclusive region lock request.
19. The computer readable medium of claim 17 , wherein the computer is further instructed to:
grant one of said plurality of processors access to said shared resource based on a determination whether there is a first existing region lock on the region lock tree that is overlapping with said exclusive region lock request and whether there is a second existing region lock on the region lock tree that is overlapping with said second exclusive region lock request.