Apparatus including memory management control circuitry and related methods for allocation of a write block cluster
Memory system controllers can include non-volatile memory control circuitry including a plurality of channel control circuits. Each of the plurality of channel control circuits can be configured to be coupled to a respective number of logical units (LUNs). Memory management circuitry can be coupled to the non-volatile memory control circuitry and configured to allocate a write block cluster for host writes based on an information width of a host bus and a protocol of the host bus. The write block cluster can include one block from fewer than all of the LUNs.
1. An apparatus, comprising:
non-volatile memory control circuitry including a plurality of channel control circuits, wherein each of the plurality of channel control circuits is configured to be coupled to a respective number of logical units (LUNs) that each include a plurality of blocks; and
memory management circuitry coupled to the non-volatile memory control circuitry and configured to:
allocate a write block cluster for host writes such that blocks of the write block cluster are written in a particular sequence of LUNs, wherein the write block cluster is based on an information width of a host bus and a protocol of the host bus, and wherein the write block cluster comprises one of the plurality of blocks from fewer than all of the LUNs; and
allocate a remainder of the plurality of blocks from the LUNs for wear leveling operations; and
temporarily deallocate at least one of the allocated LUNs from wear leveling operations at least partially in response to the particular sequence of LUNs indicating that the at least one of the allocated LUNs is within a threshold number of LUNs of being next in the particular sequence of LUNs for a host write, wherein the threshold number is associated with a wear leveling read operation, a wear leveling write operation, or an erase operation.
2. The apparatus of claim 1 , wherein the memory management circuitry is configured to limit the size of the write block cluster to a minimum number of the LUNs used to support a maximum host bus bandwidth according to the information width of the host bus and the protocol of the host bus.
3. The apparatus of claim 1 , wherein the threshold number of LUNs is different for each of the wear leveling read operation, the wear leveling write operation, and the erase operation.
4. The apparatus of claim 1 , wherein the non-volatile memory control circuitry includes a number of queues associated with each of the LUNs; and
wherein the memory management circuitry is configured to limit a number of commands issued for wear leveling operations to a particular one of the LUNs at least partially in response to the number of queues associated with the particular LUN being within a threshold of being full.
5. The apparatus of claim 1 , wherein the memory management circuitry is configured to limit a number of commands issued for a wear leveling operation at least partially in response to an amount of power caused to be used by the apparatus exceeding a threshold amount of power.
6. The apparatus of claim 1 , wherein the memory management circuitry is configured to enforce at least a minimum time between issuance of a first command and a second command for a wear leveling operation.
7. The apparatus of claim 6 , wherein the memory management circuitry is configured to enforce the minimum time between issuance of the first and the second command for the wear leveling operation based on a total number of erased blocks in the LUNs.
8. The apparatus of claim 1 , wherein the memory management circuitry is configured to:
track a number of erased blocks for each of the LUNs; and
suspend wear leveling operations for a particular one of the LUNs at least partially in response to the number of erased blocks in the particular LUN exceeding a wear leveling threshold number.
9. The apparatus of claim 8 , wherein the memory management circuitry is configured to suspend host writes for the particular LUN at least partially in response to the number of erased blocks in the particular LUN falling below a host write threshold number.
10. The apparatus of claim 1 , wherein the plurality of channel control circuits include polling logic configured to issue a status read to the respective number of LUNs, and wherein the plurality of channel control circuits are configured to idle the polling logic for a particular time according to a type of operation indicated by a command issued to one of the number of LUNs.
11. A method, comprising:
allocating a write block cluster via memory management control circuitry for host writes such that a plurality of blocks of the write block cluster are written in a particular sequence of a plurality of logical units (LUNs), wherein the write block cluster is based on an information width of a host bus and a protocol of the host bus, and wherein the write block cluster comprises one block from fewer than all of the plurality of LUNs;
allocating a remainder of the plurality of blocks from the plurality of LUNs for wear leveling operations;
temporarily deallocating at least one of the allocated LUNs from wear leveling operations at least partially in response to the particular sequence of LUNs indicating that the at least one of the allocated LUNs is within a threshold number of LUNs of being next in the particular sequence of LUNs for a host write, wherein the threshold number is associated with a wear leveling read operation, a wear leveling write operation, or an erase operation; and
wherein each of a plurality of channel control circuits is configured to be coupled to a respective number of the plurality of LUNs.
12. The method of claim 11 , wherein the method includes limiting the size of the write block cluster to a minimum number of the plurality of LUNs used to support a maximum host bus bandwidth according to the information width of the host bus and the protocol of the host bus.
13. The method of claim 11 , wherein the method includes:
tracking a number of erased blocks for each of the plurality of LUNs; and
suspending wear leveling operations for a particular one of the plurality of LUNs at least partially in response to the number of erased blocks in the particular LUN exceeding a wear leveling threshold number.
14. A method, comprising:
allocating a write block cluster via memory management control circuitry for host writes such that blocks of the write block cluster are written in a particular sequence of a plurality of logical units (LUNs), wherein the write block cluster is based on an information width of a host bus and a protocol of the host bus, and wherein the write block cluster comprises one block from fewer than all of the plurality of LUNs that each include a plurality of blocks; and
allocating a remainder of the plurality of blocks from the plurality of LUNs for wear leveling operations via the memory management control circuitry; and
temporarily deallocating at least one of the allocated LUNs from wear leveling operations at least partially in response to the particular sequence of LUNs indicating that the at least one of the allocated LUNs is within a threshold number of LUNs of being next in the particular sequence of LUNs for a host write, wherein the threshold number is associated with a wear leveling read operation, a wear leveling write operation, or an erase operation.
15. The method of claim 14 , wherein the method includes limiting a number of commands issued for wear leveling operations to a particular one of the plurality of LUNs at least partially in response to a number of queues associated with the particular one of the plurality of LUNs being within a threshold of being full.