Back to results
US 10,649,829 B2Granted
Tracking errors associated with memory access operations
AssigneeHewlett Packard Enterprise Development LP
InventorsDerek Alan Sherlock, Shawn Walker, Paolo Faraboschi
Life of this patent
Estimating the term…
Abstract
In some examples, a controller includes a counter to track errors associated with a group of memory access operations, and processing logic to detect an error associated with the group of memory access operations, determine whether the detected error causes an error state change of the group of memory access operations, and cause advancing of the counter responsive to determining that the detected error causes the error state change of the group of memory access operations.
Claims20
A controller comprising:
- a counter to track errors associated with a group of memory access operations;
- and
- processing logic to:
- detect an error associated with the group of memory access operations,
- determine whether the detected error causes an error state of the group of memory access operations to change by comparing an error type of the detected error to a previously detected error type of the group of memory access operations, wherein the detected error causes the error state of the group of memory operations to change if the error type of the detected error is different from the previously detected error type, and wherein the error type comprises one or more of:
- a device fault, a communication link fault, and a poisoned flush from a cache memory;
- cause advancing of the counter responsive to determining that the detected error causes the error state change of the group of memory access operations;
- and
- leave the counter unchanged in response to determining that the detected error corresponds to the same error state as the current error state of the group of memory access operations.
6 more elements
A system comprising:
- media controllers to control access to memory devices;
- a bridge to route memory requests from a requester to the media controllers, the bridge comprising a controller to:
- detect an error associated with a group of memory access operations of the requester,
- determine whether the detected error causes an error state of the group of memory access operations to change by comparing an error type of the detected error to a previously detected error type of the group of memory access operations, wherein the detected error causes the error state of the group of memory operations to change if the error type of the detected error is different from the previously detected error type, and wherein the error type comprises one or more of:
- a device fault, a communication link fault, and a poisoned flush from a cache memory;
- and
- update an error tracking value responsive to determining that the detected error causes the error state change of the group of memory access operations and leave the error tracking value unchanged responsive to determining that the detected error does not cause the error state change of the group of memory access operations.
4 more elements
An apparatus comprising:
- a processor;
- and
- a non-transitory storage medium storing instructions executable on the processor to:
- flush cache data from a cache memory to a network to which media controllers are coupled, the flushing of the cache data comprising write operations that are part of a group of memory access operations;
- send, to a bridge in the network, a request seeking a response once the write operations have successfully completed and are persisted;
- receive, in response to the request, an error counter value that is updated in response to a newly detected error associated with the group of memory access operations causing an error state change of the group of memory access operations and leave the error counter value unchanged when the newly detected error does not cause the error state change of the group of memory access operations, wherein the newly detected error causes the error state of the group of memory operations to change if an error type of the newly detected error is different from a previously detected error type of the group of memory operations, and wherein the error type comprises one or more of:
- a device fault, a communication link fault, and a poisoned flush from a cache memory;
- in response to determining that the received error counter value is different from a current error counter value, take action to address the error.
5 more elements
Dependencies are inferred from the claim text.