Non-volatile memory module architecture to support memory error correction
Apparatus and methods are provided for operating a non-volatile memory module. In an example, a method can include filling a first plurality of pages of a first non-volatile memory with first data from a first data lane that includes a first volatile memory device, and filling a second plurality of pages of the first non-volatile memory device with second data from a second data lane that includes a second volatile memory device. In certain examples, the first plurality of pages does not include data from the second data lane.
1. A method of operating a dual inline memory module (DIMM), comprising:
receiving a first command at a register of the DIMM, the first command from a host device, the first command enabling saving of contents of dynamic random access memory (DRAM) of the memory module to non-volatile memory (NVM) of the (DIMM) in an event of loss of power to the DIMM;
in response to loss of power to the DIMM and to enabled saving of contents of the DRAM in the register, transferring contents of the DRAM to the NVM through use of a power source in communication with a controller of the DIMM;
receiving at the register a second command from the host device, the second command based at least in part on power to the DIMM being re-established;
in response to the second command, restoring the saved contents from the NVM to the DRAM; and
identifying an error in restoring the contents from the NVM to the DRAM.
2. The method of claim 1 , further comprising operating the DIMM in accordance with the identified error.
3. The method of claim 2 , wherein operating the DIMM in accordance with the identified error comprises correcting corrupted data causing the error.
4. The method of claim 3 , wherein correcting the corrupted data comprises implementing a chip kill algorithm on a failing memory chip.
5. The method of claim 4 , or wherein the chip kill algorithm is applied to a memory chip of the NVM.
6. The method of claim 1 , wherein the power source is mounted on the DIMM.
7. The method of claim 1 , wherein transferring content of the DRAM to the NVM is further in response to receiving a further signal from the host identifying a power loss event.
8. The method of claim 1 , wherein the DRAM comprises multiple DRAM memory devices.
9. The method of claim 1 , wherein the NVM comprises multiple flash memory devices.
10. A dual inline memory module (DIMM), comprising:
dynamic random access memory (DRAM);
non-volatile memory (NVM);
a power source; and
a controller and a register configured to perform operations, including:
receiving, from a host device, a first control signal establishing a value in the register enabling saving of data in the DRAM to the NVM through use of the power source in response to power loss to the DIMM;
receiving, from a host device, a second control signal to restore previously-stored data from the NVM to the DRAM the second control signal received in response to restoring of power to the DIMM; and
receiving from a host a signal indicating an error in restoring the data from the NVM to the DRAM.
11. The DIMM of claim 10 , wherein the operations further comprise reconstructing data resulting in the indicated error in restoring the data from the NVM to the DRAM.
12. The DIMM of claim 11 , wherein the operations further comprise in response to the first control signal initiating transfer of data from DRAM of the DIMM to non-volatile memory (NVM) of the DIMM.
13. The DIMM of claim 10 , wherein the operations further comprise identifying power loss to the DIMM and initiating saving of data in the DRAM to the NVM.
14. The DIMM of claim 13 , wherein identifying the power loss to the DIMM comprises receiving a signal from the host indicating the power loss to the DIMM.
15. The DIMM of claim 10 , wherein saving of data in the DRAM to the NVM comprises saving ECC data to the NVM; and
wherein restoring previously stored data from the NVM to the DRAM comprises restoring ECC data from the NVM to the DRAM.
16. A method comprising:
receiving from a host device, at a dual in-line memory module (DIMM) a first command enabling saving of data from dynamic random access memory (DRAM) of the DIMM to non-volatile memory (NVM) of the DIMM in response to a loss of power to the DIMM;
transferring, in response to the first command, data from DRAM of the DIMM to non-volatile memory NVM of the DIMM;
receiving, at the DIMM from the host device, a second command based at least in part on restoration of power to the DIMM;
in response to the second command, restoring data from NVM to DRAM;
receiving, at the DIMM, a third command from the host device identifying an error in restoring data from the NVM to the DRAM, the error including corrupted data in the DRAM; and
operating the DIMM based at least in part on the corrupted data and according to instructions from the host device.
17. The method of claim 16 , further comprising:
storing error correcting code (ECC) data on the DRAM, wherein the data transferred to the NVM and restored to DRAM comprises the ECC data; and
operating the DIMM based at least in part on the ECC data.
18. The method of claim 17 , wherein ECC data is saved in a separate page of the NVM from data related to the ECC data.
19. The method of claim 16 , wherein operating the DIMM comprises restoration of the corrupted data.
20. The method of claim 16 , wherein the first command further initiates saving of data from the DRAM to the NVM.