Storage system with subdivisions
This disclosure provides for improvements in managing multi-drive, multi-die or multi-plane NAND flash memory. In one embodiment, the host directly assigns physical addresses and performs logical-to-physical address translation in a manner that reduces or eliminates the need for a memory controller to handle these functions, and initiates functions such as wear leveling in a manner that avoids competition with host data accesses. A memory controller optionally educates the host on array composition, capabilities and addressing restrictions. Host software can therefore interleave write and read requests across dies in a manner unencumbered by memory controller address translation. For multi-plane designs, the host writes related data in a manner consistent with multi-plane device addressing limitations. The host is therefore able to “plan ahead” in a manner supporting host issuance of true multi-plane read commands.
1 . A method comprising:
transmitting one or more NVMe query commands to a solid state storage drive (SSD), wherein the SSD comprises flash memory divided into non-overlapping subdivisions of the flash memory, and each subdivision is mapped to at least two erase units and the erase units are on at least two different planes of the flash memory; and responsively receiving:
one or more subdivision identifiers that identify subdivisions in the SSD; and
capacities for the subdivisions;
receiving information from the SSD about a specific one of the subdivisions; wherein the specific subdivision is identified by its identifier, and the received information includes an extent to which the erase units mapped to the specific subdivision are not yet written following an erase and can be written with data; and
in response to receiving the information from the SSD, transmitting one or more additional NVMe commands to the SSD, wherein the one or more additional NVMe commands identify the specific subdivision by its identifier.
2 . A storage system comprising:
a solid state storage drive (SSD) comprising:
flash memory divided into non-overlapping subdivisions of the flash memory, wherein each subdivision is mapped to at least two erase units and the erase units are on at least two different planes of the flash memory; and
a host interface configured to communicate with a host using commands from a Non-Volatile Memory Express standard (NVMe commands);
a host that communicates with the SSD using NVMe commands, wherein the host:
transmits one or more NVMe query commands to the SSD and responsively receives:
one or more subdivision identifiers that identify subdivisions in the SSD; and
capacities for the subdivisions;
receives information from the SSD about a specific one of the subdivisions; wherein the specific subdivision is identified by its subdivision identifier, and the received information includes an extent to which the erase units mapped to the specific subdivision are not yet written following an erase and can be written with data; and
in response to receiving the information from the SSD, transmits one or more additional NVMe commands to the SSD, wherein the one or more additional NVMe commands identify the specific subdivision by its subdivision identifier.
3 . The storage system of claim 2 , wherein the one or more additional NVMe commands are write commands to write data to the specific subdivision.
4 . The storage system of claim 3 , wherein the SSD updates metadata that indicates the extent to which the erase units mapped to the specific subdivision are not yet written following an erase, in response to receiving the one or more write commands.
5 . The storage system of claim 2 , wherein the one or more additional NVMe commands are maintenance commands, and, upon receipt of the maintenance commands, the extent to which the erase units mapped to the specific subdivision are not yet written following an erase is set equal to a capacity of the specific subdivision.
6 . The storage system of claim 2 , wherein the SSD is operable to, responsive to the one or more additional NVMe commands, erase the erase units mapped to the specific subdivision.
7 . The storage system of claim 2 , wherein the one or more additional NVMe commands comprise at least one write command to write to a logical address within the specific subdivision with data, and wherein the SSD is operable to:
track metadata representing utilization of the erase units mapped to the specific subdivision, wherein the utilization of the erase units indicates an extent to which data written to the erase units are still valid; and
responsive to the at least one write command: deallocate a physical storage location to which the logical address was previously mapped and, in association with the deallocation, automatically erase the erase units mapped to the specific subdivision.
8 . The storage system of claim 2 , wherein:
the SSD is further operable to store metadata relating to the specific subdivision; and
the information received from the SSD depends on the stored metadata relating to the specific subdivision.
9 . The storage system of claim 8 , wherein the received information further represents a time since data was written to the erase units mapped to the specific subdivision.
10 . The storage system of claim 2 , wherein a minimum amount of storage capacity in the specific subdivision that can be erased by the SSD is equal to the capacity of the specific subdivision.
11 . The storage system of claim 2 , wherein the SSD is further operable to:
store metadata indicating utilization of the erase units mapped to the specific subdivision, wherein the utilization of the erase units indicates an extent to which data written to the erase units are still valid;
compare the metadata to at least one utilization criterion for low utilization of the erase units; and
upon detection that the metadata satisfies the at least one utilization criterion: copy data from the erase units mapped to the specific subdivision to a different one or more erase units, and map the different one or more erase units uniquely to the specific subdivision and not to any other subdivisions.
12 . The storage system of claim 2 , wherein the host:
tracks release status information for individual logical block addresses (LBAs) associated with the specific subdivision; and
issues one or more NVMe commands addressed to the specific subdivision as a function of the tracked release status information, wherein the SSD is operable to perform at least one maintenance operation that is directed to the specific subdivision and not to any other subdivisions.
13 . The storage system of claim 2 , wherein the SSD is further operable to identify one of the erase units which is fully released and to responsively control erasure of the identified erase unit.
14 . The storage system of claim 2 , wherein the SSD also includes dies of flash memory, and wherein the at least two different planes of the flash memory are located on at least two different dies.
15 . The storage system of claim 2 , wherein receiving one or more subdivision identifiers comprises: receiving a total number of subdivision identifiers, and the host determines the subdivision identifiers based on the total number of subdivision identifiers.
16 . The storage system of claim 2 , wherein the subdivisions all have a same capacity, and receiving capacities for the subdivisions comprises receiving an amount for the same capacity.
17 . The storage system of claim 2 , wherein the one or more additional NVMe commands are write commands to write data to the specific subdivision, each write command accompanied with a logical block address, wherein the SSD stores the logical block address as metadata for the specific subdivision.
18 . The storage system of claim 2 , wherein the host comprises a storage manager operable to:
map single non-overlapping segments of logical address ranges to the subdivisions of the flash memory;
maintain an address translation table that associates individual writes with a specific address within a specific segment;
track release status information of data written to respective segments;
identify at least one garbage collection candidate segment by comparing the release status information to a criterion for low utilization of segments;
perform segment cleaning wherein the host issues one or more NVMe commands to the SSD to copy data from the identified segment to a different segment; and
update the address translation table to indicate new addresses of the data copied.
19 . A host device comprising:
at least one processor; and
a non-transitory machine readable media storing instructions that, when executed by the at least one processor, cause the host device to:
transmit one or more NVMe query commands to a solid state storage drive (SSD), wherein the SSD comprises flash memory divided into non-overlapping subdivisions of the flash memory, and each subdivision is mapped to at least two erase units and the erase units are on at least two different planes of the flash memory; and the host device responsively receives:
one or more subdivision identifiers that identify subdivisions in the SSD; and
capacities for the subdivisions;
receive information from the SSD about a specific one of the subdivisions; wherein the specific subdivision is identified by its subdivision identifier, and the received information includes an extent to which the erase units mapped to the specific subdivision are not yet written following an erase and can be written with data; and
in response to receiving the information from the SSD, transmit one or more additional NVMe commands to the SSD, wherein the one or more additional NVMe commands identify the specific subdivision by its identifier.
20 . The host device of claim 19 , wherein the one or more additional NVMe commands comprise at least one write command to write to a logical address within the specific subdivision with data, and wherein the SSD is operable to:
track metadata representing utilization of the erase units mapped to the specific subdivision, wherein the utilization of the erase units indicates an extent to which data written to the erase units are still valid; and
responsive to the at least one write command: deallocate a physical storage location to which the logical address was previously mapped and, in association with the deallocation, automatically erase the erase units mapped to the specific subdivision.
21 . The host device of claim 19 , wherein the SSD is further operable to:
store metadata indicating utilization of the erase units mapped to the specific subdivision, wherein the utilization of the erase units indicates an extent to which data written to the erase units are still valid;
compare the metadata to at least one utilization criterion for low utilization of the erase units; and
upon detection that the metadata satisfies the at least one utilization criterion: copy data from the erase units mapped to the specific subdivision to a different one or more erase units, and map the different one or more erase units to the specific subdivision and not to any other subdivisions.
22 . The host device of claim 19 , wherein the SSD is further operable to identify one of the erase units which is fully released and to responsively control erasure of the identified erase unit.
23 . A storage system comprising:
a solid state storage drive (SSD) comprising:
flash memory divided into non-overlapping subdivisions of the flash memory, wherein each subdivision is mapped to at least two erase units and the erase units are on at least two different planes of the flash memory; and
a host interface configured to communicate with a host using commands from a Non-Volatile Memory Express standard (NVMe commands);
a host that communicates with the SSD using NVMe commands, wherein the host:
transmits one or more NVMe query commands to the SSD and responsively receives:
one or more subdivision identifiers that identify subdivisions in the SSD; and
capacities for the subdivisions;
receives information from the SSD about a specific one of the subdivisions; wherein the specific subdivision is identified by its subdivision identifier, and the received information includes an extent to which the erase units mapped to the specific subdivision are not yet written following an erase and can be written with data;
in response to receiving the information from the SSD, transmits one or more additional NVMe commands to the SSD for the specific subdivision, wherein the one or more additional NVMe commands identify the specific subdivision by its subdivision identifier; and
the host comprises a storage manager operable to:
track release status information for data written to the subdivisions of the flash memory;
compare the release status information to a criterion for low utilization of segments;
identify a subdivision based upon the comparison; and
copy data from the identified subdivision to a different location.
24 . The storage system of claim 23 , wherein:
the one or more additional NVMe commands are write commands to write data to the specific subdivision; and
the SSD updates metadata that indicates the extent to which the erase units mapped to the specific subdivision are not yet written following an erase, in response to receiving the one or more write commands.
25 . The storage system of claim 23 , wherein the one or more additional NVMe commands are maintenance commands, and, upon receipt of the maintenance commands, the extent to which the erase units mapped to the specific subdivision are not yet written following an erase is set equal to a capacity of the specific subdivision.
26 . The storage system of claim 23 , wherein the SSD is operable to, responsive to the one or more additional NVMe commands, erase the erase units mapped to the specific subdivision.
27 . The storage system of claim 23 , wherein the one or more additional NVMe commands comprise at least one write command to write to a logical address within the specific subdivision with data, and wherein the SSD is operable to:
track metadata representing utilization of the erase units mapped to the specific subdivision, wherein the utilization of the erase units indicates an extent to which data written to the erase units are still valid; and
responsive to the at least one write command: deallocate a physical storage location to which the logical address was previously mapped and, in association with the deallocation, automatically erase the erase units mapped to the specific subdivision.
28 . The storage system of claim 23 , wherein:
the SSD is further operable to store metadata relating to the specific subdivision;
the information received from the SSD depends on the stored metadata relating to the specific subdivision; and
the received information further represents a time since data was written to the erase units mapped to the specific subdivision.
29 . The storage system of claim 23 , wherein a minimum amount of storage capacity in the specific subdivision that can be erased by the SSD is equal to the capacity of the specific subdivision.
30 . The storage system of claim 23 , wherein the SSD is further operable to:
store metadata indicating utilization of the erase units mapped to the specific subdivision, wherein the utilization of the erase units indicates an extent to which data written to the erase units are still valid;
compare the metadata to at least one utilization criterion for low utilization of the erase units; and
upon detection that the metadata satisfies the at least one utilization criterion: copy data from the erase units mapped to the specific subdivision to a different one or more erase units, and map the different one or more erase units to the specific subdivision and not to any other subdivisions.
31 . The storage system of claim 23 , wherein the host:
tracks release status information for individual logical block addresses (LBAs) associated with the specific subdivision; and
issues one or more NVMe commands addressed to the specific subdivision as a function of the tracked release status information, wherein the SSD is operable to perform at least one maintenance operation that is directed to the specific subdivision and not to any other subdivisions.
32 . The storage system of claim 23 , wherein the SSD is further operable to identify one of the erase units which is fully released and to responsively control erasure of the identified erase unit.
33 . The storage system of claim 23 , wherein the SSD also includes dies of flash memory, and wherein the at least two different planes of the flash memory are located on at least two different dies.
34 . The storage system of claim 23 , wherein the one or more additional NVMe commands are write commands to write data to the specific subdivision, each write command accompanied with a logical block address, wherein the SSD stores the logical block address as metadata for the specific subdivision.