Abstract: Methods, systems, and devices for commanded device states for a memory system are described. For example, a memory system may be configured with different device states that are each associated with a respective allocation of resources (e.g., feature sets) for operations of the memory system. Resource allocations corresponding to the different device states may be associated with different combinations of memory management configurations, error control configurations, trim parameters, degrees of parallelism, or endurance configurations, among other parameters of the memory system, which may support different tradeoffs between performance characteristics of the memory system. A host system may be configured to evaluate various parameters of operating the host system, and to transmit commands for a memory system to enter a desired device state of the memory system.

Abstract: Methods, systems, and devices for data defragmentation control are described. A memory system may include one or more regions of logical addresses and a plurality of memory cells arranged according to a plurality of physical addresses. In some instances, data may be stored to one or more discontinuous physical addresses and it may be desirable rearrange the data to be within continuous physical addresses (e.g., it may be desirable to defragment the data). Accordingly, the data stored to the one or more discontinuous physical addresses may be arranged (e.g., rearranged) to be within continuous physical addresses based at least in part on a value stored to one or more registers of the memory system.

Abstract: Methods, systems, and devices for detecting page fault traffic are described. A memory device may execute a self-learning algorithm to determine a priority size for read requests, such as a maximum readahead window size or other size related to page faults in a memory system. The memory device may determine the priority size based at least in part on by tracking how many read requests are received for different sizes of sets of data. Once the priority size is determined, the memory device may detect subsequent read requests for sets of data having the priority size, and the memory device may prioritize or other optimize the execution of such read requests.

Abstract: Methods, systems, and devices for memory device access techniques are described. Memory systems may be enabled to allow device-controlled access to a portion of volatile memory at a host system. By enabling the memory system to access volatile memory at the host system, the memory system may perform access operations which may reduce a quantity of messages exchanged between the memory system to the host system. The host system may allocate a list of memory resources in volatile memory associated with a first access command. The host system may allocate the same memory resources for a second access command. By allocating the same memory resources, the memory device may transmit a Ready To Transfer (RTT) message for multiple access commands, rather than for each command. In some cases, reducing the quantity of RTT messages may reduce latency and improve performance at the memory system.

Abstract: Methods, systems, and devices for dynamic power modes for boot-up procedures are described. A memory system may initiate a boot-up procedure according to a predefined first power mode that is associated with a first power consumption. The memory system may then determine whether to perform the boot-up procedure according to the first power mode or a second power mode associated with a different second power consumption. In cases that the memory system receives an indication of the second power mode from the host system, the memory system may perform the boot-up procedure according to the second power mode. Additionally, in cases that the memory system does not receive an indication of the second power mode from the host system, the memory system may perform the boot-up procedure according to the first power mode.

Abstract: Methods, systems, and devices supporting techniques for memory system configuration using a queue refill time are described. A memory system may receive a command from a host system and may add the command to a command queue including a set of commands to be executed by the memory system. The memory system may determine a queue refill time of the command queue using measurements for at least one queue tag of the command queue and may adjust at least one resource of the command queue in response to the determined queue refill time. In some examples, the memory system may reallocate processing or memory resources previously allocated to the command queue, deactivate processing or memory resources previously allocated to the command queue, adjust a threshold queue depth for the command queue, or any combination thereof, among other options, based on the queue refill time.

Abstract: Methods, systems, and devices for memory device access techniques are described. Memory systems may be enabled to allow device-controlled access to a portion of volatile memory at a host system. By enabling the memory system to access volatile memory at the host system, the memory system may perform access operations which may reduce a quantity of messages exchanged between the memory system to the host system. The host system may allocate a list of memory resources in volatile memory associated with a first access command. The host system may allocate the same memory resources for a second access command. By allocating the same memory resources, the memory device may transmit a Ready To Transfer (RTT) message for multiple access commands, rather than for each command. In some cases, reducing the quantity of RTT messages may reduce latency and improve performance at the memory system.

Abstract: Methods, systems, and devices for techniques for controlling command order are described. An entity of a host system, such as a file system, may insert a sequential identifier into commands generated by the entity to indicate an order of the commands. In some examples, the host system may specify a set of commands in a first sequence to be transmitted to the memory system. The host system may subsequently reorder the set of commands in to a second sequence and transmit the set of commands to the memory system. In some cases, following a power-on condition, the memory system may determine a latest valid command of the set of commands. The memory system may subsequently invalidate one or more logical addresses associated with commands having sequence identifiers after the sequence identifier of the latest valid command.

Abstract: Methods, systems, and devices for unmap operation techniques are described. A memory system may include a volatile memory device and a non-volatile memory device. The memory system may receive a set of unmap commands that each include a logical block address associated with unused data. The memory system may determine whether one or more parameters associated with the set of unmap commands satisfy a threshold. If the one or more parameters satisfy the threshold, the memory system may select a first procedure for performing the set of unmap commands different from a second procedure (e.g., a default procedure) for performing the set of unmap commands and may perform the set of unmap commands using the first procedure. If the one or more parameters do not satisfy the threshold, the memory system may perform the set of unmap commands using the second procedure.

Abstract: A processing device of a memory sub-system can monitor a plurality of received commands to identify a forced unit access command. The processing device can identify a metadata area of the memory device based on the forced unit access command. The processing device can also perform an action responsive to identifying a subsequent forced unit access command to the metadata area.

Abstract: Methods, systems, and devices for identification and storage of boot information at a memory system are described to support transferring boot information to higher reliability memory storage. A memory system may identify boot information stored at a memory array based on a command received from a host system, which may identify the boot information for the memory system, or based on performing a boot procedure with the host system, in which the boot information may be requested from the memory system. After identifying the boot information stored at the memory array, the memory system may move or transfer the boot information from physical addresses associated with lower reliable memory storage to physical addresses associated with higher reliable memory storage.

Abstract: A processing device of a system receives a request to access a selected sector in a memory component. The selected sector is associated with a sector number. The processing device determines a virtual block corresponding to the selected sector. The virtual block is associated with a misalignment factor and a misalignment counter. The processing device determines if the misalignment counter satisfies a threshold criterion. In response to the misalignment counter satisfying the threshold criterion, the processing device generates an updated sector number by shifting the sector number by the misalignment factor and performs the access to the selected sector using the updated sector number. In response to the misalignment counter not satisfying the threshold criterion, the processing device updates the misalignment counter and performs the access to the selected sector using the sector number.

Abstract: Methods, systems, and devices for optimized command sequences are described. An apparatus includes a memory array and a controller coupled with the memory array. The controller may be configured to receive a first command indicating a start of a sequence of access commands to store at the controller, then receive a first set of access commands associated with the sequence of access commands, and then receive a second command indicating the end of the sequence of access commands. The controller may also receive a second set of access commands after the command. The controller may execute an operation associated with a third set of access commands of the sequence after receiving the second set of access commands and before receiving the third set of access commands based at least in part on identifying the second set of access commands as starting the sequence of access commands.

Abstract: Methods, systems, and devices for techniques for atomic write operations are described. A memory system may determine a set of pages for an atomic write operation in which data associated with a write command is linked together for writing to a non-volatile memory. The memory system may write, to the non-volatile memory, metadata that indicates the set of pages is associated with the atomic write operation. Based on the metadata, the memory system may determine whether each page of the set of pages has been written with data for the atomic write operation. The memory system may then communicate to a host system an indication of a completion status for the atomic write operation based on determining whether each page of the set of pages has been written with the data for the atomic write operation.

Abstract: Methods, systems, and devices for modes to extend life of memory systems are described. In some examples a host system and a memory system may support one or more vendor commands to configure one or more logical units of the memory system to be accessible in a read-only mode. For example, the host system may periodically transmit a command to retrieve health information of the memory system. The host system may compare the health information to one or more thresholds associated with one or more host entities. If the health information satisfies a threshold for a host entity, the host system may transmit a command to the memory system to initiate a read-only mode for the logical units associated with the host entity. Additionally or alternatively, the memory system may track the health information and compare the health information to the one or more thresholds.

Abstract: Methods, systems, and devices for memory-aligned access operations are described. A target packet size based on a quantity of physical pages addressable by individual first-level pages of a first-level page table for mapping logical address to respective physical pages may be indicated to a host system. A buffer may be configured based on the target packet size and data for an application at the host system and associated with the target packet size may be stored in the buffer. Based on a utilization threshold of the buffer being reached, a set of data stored in the buffer and having the target packet size may be written to a memory device, where a set of physical addresses for storing the set of data may be identified based on a second-level entry of a second-level page.

Abstract: Methods, systems, and devices for write booster buffer and hibernate are described. The memory system may initiate a first operation to enter a first power mode having a lower power consumption than a second power mode. In some cases, the memory system may determine whether a quantity of data stored in a buffer of single-level cells associated with write booster information satisfies a threshold based on initiating the first operation. The memory system may determine whether to perform a second operation to transfer the quantity of data stored in the buffer of single-level cells to a portion of memory comprising multiple level cells based on determining whether the quantity of data satisfies the threshold. The memory system may enter the first power mode based on determining to perform the second operation to transfer the quantity of data from the buffer to the portion of memory.

Abstract: Methods, systems, and devices for data defragmentation for a system boot procedure are described. The memory system may determine a write random index associated with a boot procedure. The write random index may indicate a relationship between a first quantity of sequential logical addresses accessed as part of the boot procedure and a second quantity of random logical addresses accessed as part of the boot procedure. The memory system may determine whether the write random index satisfies a threshold based on determining the write random index. In some cases, the memory system may transfer, to a second portion of the memory system, data stored in a first portion of the memory system based on determining that the write random index satisfies the threshold. The memory system may receive a request to perform the boot procedure after transferring the data and output, to the host system, the data transferred.

Abstract: Methods, systems, and devices are described to indicate, in an entry of logical to physical (L2P) mapping information stored at a host system, whether data associated with the entry is sequential to other data associated with a next entry or a previous entry. Each entry may have a third field, which may indicate whether the data is sequential. Based on the third field, the host system may determine whether data to be read from a memory system is sequential. The host system may transmit one read command to the memory system if the data is sequential, where the read command may include at least a portion of an L2P entry associated with the data. Similarly, based on the third field, the memory system may determine whether the data to be read is sequential, and may read additional, sequential data if the memory system determines that the data is sequential.

Abstract: Methods, systems, and devices for detecting page fault traffic are described. A memory device may execute a self-learning algorithm to determine a priority size for read requests, such as a maximum readahead window size or other size related to page faults in a memory system. The memory device may determine the priority size based at least in part on by tracking how many read requests are received for different sizes of sets of data. Once the priority size is determined, the memory device may detect subsequent read requests for sets of data having the priority size, and the memory device may prioritize or other optimize the execution of such read requests.