Persistent volatile memory fault tracking using entries in the non-volatile memory of a fault storage unit
One embodiment disclosed relates to a method for persistently tracking volatile memory faults. A memory error is detected in relation to at least one dynamic random access memory (DRAM) unit on a particular memory module. An entry pertaining to the memory error is written in non-volatile memory of a fault storage unit on that particular memory module. Another embodiment disclosed relates to a memory module that persistently tracks volatile memory faults. The memory module includes a plurality of dynamic random access memories (DRAMs) and a fault storage unit. The fault storage unit includes non-volatile memory configured to store entries pertaining to faults in the plurality of DRAMs on that memory module.
1. A circuit board of a system, the circuit board comprising:
a plurality of connectors, each connector connects to a memory module which includes multiple volatile memory units and a non-volatile fault storage unit;
a memory controller configured to read and write data into the volatile memory units of memory modules;
a memory error interface configured to provide read and write access to the non-volatile fault storage units of the memory modules; and
a computer-readable medium including computer-readable error handling code configured to write entries relating to detected memory errors into the non-volatile fault storage unit and to read said entries from the non-volatile fault storage unit,
wherein each said entry comprises a memory unit identifier, a start bit of a memory error, an end bit of the memory error, and tag bits indicating time of last failure and number of occurrences of failure.
2. The circuit board of claim 1 , further comprising:
a processor dependent hardware (PDH) interface communicatively coupled between a central processing unit and the memory error interface.
3. The circuit board of claim 2 , further comprising:
a processor dependent code (PDC) unit accessible via the PDH interface,
wherein the PDC unit includes said computer-readable medium, and
wherein said computer-readable medium further includes computer-readable boot code.
4. The circuit board of claim 3 , wherein the computer-readable boot code is configured to read the entries from the non-volatile fault storage unit and to remove memory bits associated with the entries from a set of usable memory.
5. The circuit board of claim 1 , wherein the volatile memory units comprise dynamic random access memory, and wherein the plurality of memory modules comprise dual in-line memory modules (DIMMs).