Request to own chaining in multi-socketed systems

A non-transitory computer readable storage medium storing instructions for managing a cache line, the instructions comprising functionality to:

• receive, from a plurality of local threads executing on a local multithreaded processor, a plurality of local requests to access the cache line;
• insert, into an address chain associated with the cache line, a plurality of entries corresponding to the plurality of local requests;
• identify a first entry of the plurality of entries at a head of the address chain;
• initiate, in response to the first entry being indentified at the head and in response to the first entry corresponding to a request to own (RTO) the cache line, a traversal of the address chain;
• set, during the traversal of the address chain, a state element identified in a second entry of the plurality of entries, wherein the state element corresponds to a serialization acknowledgement from a central ordering unit for a local thread corresponding to the second entry;
• receive, from a foreign thread executing on a foreign multithreaded processor, a foreign request to access the cache line;
• insert, in response to the state element being set, a third entry corresponding to the foreign request into the address chain after the second entry;
• and
• relinquish, in response to the third entry being inserted after the second entry in the address chain, the cache line to the foreign thread after executing the plurality of local requests.

A system for managing a cache line, comprising:

• a plurality of local threads executing on a local multithreaded processor and issuing a plurality of local requests to access the cache line;
• a foreign thread executing on a foreign multithreaded processor and issuing a foreign request to access the cache line;
• and
• a local cache storing the cache line and comprising an address chain storing a plurality of local entries corresponding to the plurality of local requests and a foreign entry corresponding to the foreign request,
• wherein the local cache is configured to set, during a traversal of the address chain, a state element identified in the plurality of local entries and corresponding to a serialization acknowledgement from a central ordering unit,
• wherein the foreign entry is inserted into the address chain after the plurality of local entries in response to the state element being set, and
• wherein the cache line is relinquished to the foreign thread after the plurality of local requests are executed in response to the foreign entry being inserted into the address chain after the plurality of local entries.