Western Digital Technologies, Inc.

Sandisk Technologies Inc.

Amir Rozen

Shay Benisty

Amir Segev

David Chaim Brief

Arseniy Aharonov

Asaf Gueta

Tomer Spector

Vitali Linkovsky

When entering low power mode, any commands that are fetched and waiting to be executed are completed. When the data storage device exits low power mode, new commands are fetched. The completing and fetching takes time. Rather than completing all of the commands, some selected commands can be stored in a persistent memory region (PMR) when entering low power mode. The storing in PMR is faster than completing the commands. Furthermore, retrieving the commands from PMR is faster than retrieving new commands from a host device on exiting low power mode. Thus, using PMR results in more efficient use of data storage device bandwidth.

Pipeline optimization in low power mode using PMR

Systems and methods for data migration via a peer communication channel between data storage devices are disclosed. The data storage devices include a host interface configured to connect to at least one host system and a peer interface to connect to the peer communication channel, where the host interface and the peer interface and separate physical interfaces. A source data storage device establishes peer communication with a destination data storage device over the peer communication channel, determines a set of host data, and sends the set of host data to the destination data storage device, while continuing to receive and process host storage operations through the host interface.

Data migration via data storage device peer channel

A data storage device includes a memory device and a controller coupled to the memory device. The controller is configured to receive a command, wherein the command comprises a plurality of logical block addresses (LBAs), determine that one or more LBAs of the plurality of LBAs are not aligned to a transfer layer packet (TLP) boundary, determine whether the one or more LBAs that are not aligned to a TLP boundary has a head that is unaligned that matches a previously stored tail that is unaligned, and merge and transfer the head that is unaligned with a previously stored tail that is unaligned when the head that is unaligned matches the previously stored tail that is unaligned.

Semi and cached TLP coalescing

The present disclosure generally relates to effective transport layer packet (TLP) utilization. When the controller of the data storage device generates a request for transferring data to or from the storage device, the request is stored in a merging buffer. The merging buffer may include previously generated requests, where the previously generated requests and the new requests are merged. A timeout counter is initialized for the requests stored in the merging buffer. The timeout counter has a configurable threshold value that corresponds to a weight value, adjusted for latency or bandwidth considerations. When the merged request is greater than the maximum TLP size, the merged request is partitioned, where at least one partition is in the size of the maximum TLP size. The request is sent from the buffer when the request is in the size of the maximum TLP size or when the threshold value is exceeded.

Effective PCIe utilization by PCIe TLP coalescing

A method and apparatus for allocation of back-end (BE) logic resources between NVM sets. When a controller detects that an NVM set is in an idle state, it deallocates the BE logic from the originally assigned NVM set and provides the BE logic resource to another NVM set. An NVM set controller matrix maps interconnections between the BE logic resource and the new NVM set to enable use of the BE logic resource and the new NVM set. When a new command arrives for the originally assigned NVM set, the BE logic resources is re-allocated to the originally assigned NVM set.

NVM endurance group controller using shared resource architecture

A system and method for time stamp synchronization are disclosed. In one embodiment, first and second devices are provided. The second device receives a first time stamp of the first device, wherein the first time stamp was generated in response to a time stamp synchronization event common to the first and second devices; generates a second time stamp of the second device in response to the time stamp synchronization event, wherein the first and second time stamps are in different time domains; and correlates the first and second time stamps, wherein correlating the first and second time stamps provide a relationship between the time domains because the first and second time stamps were both generated with respect to the same time stamp synchronization event common to the first and second devices.

System and method for time stamp synchronization

A system on a chip or storage device has a dynamic process for handling system events that are transmitted at varying transmission rates. This dynamic process is a hybrid mode of operation that tailors the use of time stamp information according to the dynamic flow of events that are submitted in the system. Relative time stamps can be used along with explicit time stamps. Periodic wrap around events which use relative time stamps based on the periodic wrap events may be suppressed when there were no events between consecutive wrap around events. When an asynchronous event occurs during the suppression, the event is identified with a high precision time stamp (HPTS) rather than a relative time stamp. The periodic wrap around events can be re-initiated after the HPTS event is stamped.

Growing community of inventors

Rishon Lezion, Israel

Amir Rozen