Marvell International Limited

Unisys Corporation

Marvellworld Trade Ltd.

Qlogic Corporation

Kha Nguyen

William C Wong

Huy Tu Nguyen

Gregory T Elkins

Arie L Krantz

Theodore C White

Jayesh V Sheth

Yehua Yang

William W Dennin, Iii

David M Kalish

Leon A Krantz

Craig Weaver Harris

Gary C Whitlock

Roger Baldwin

Chung W Wong

Russell Lee Marrash

Michael J North

Dan T Tran

Richard A Cowgill

Mouluan Jang

Jane X Wang

Bruce Edward Moolenaar

Methods and systems associated with re-transferring data that was unsuccessfully transmitted to a host are described. According to one embodiment a method includes transferring data from a buffer to a host; receiving an unsuccessful status from the host indicating a transmission error; identifying a block of data being transferred when the transmission error occurred; and re-transferring data in the identified block from the buffer to the host without re-transferring successfully transferred blocks.

Restart operation with logical blocks in queued commands

A system and method for verifying an identifier of a command. The method includes receiving an incoming command and sending a first alert to auto-logging hardware, wherein the auto-logging hardware sends a fetch instruction in response to receiving the first alert; retrieving an identifier of the incoming command in response to the fetch instruction and sending a second alert to the auto-logging hardware, wherein the auto-logging hardware sends a search instruction in response to receiving the second alert; and searching for the identifier of the incoming command in a table in response to the search instruction, the table storing identifiers previously assigned to other commands, wherein the incoming command is logged into the search table and marked as a searched command after the search for the first identifier in the table has completed successfully.

Auto-logging of read/write commands in a storage network

A method and system for command queuing in disk drives may improve performance by queuing multiple commands and sequentially executing them automatically without firmware intervention. The method may use a number of queues, e.g., a staging queue for commands to be executed, an execution queue for commands currently being executed, and a holding queue for commands which have been executed but have not received a status report from a host. With the pipelined nature of queued commands, when data requested by one command are being sent to the host, the queue logic may already be fetching data for the next command. If an error occurs in the transmission, commands in the queues may backtrack and restart from the point where data were last known to have been successfully sent to the host.

Command queuing in disk drives

The transmission of buffered data is coordinated between a storage medium and a host in response to a request from the host. One or more blocks of data are transferred from the storage medium to a buffer memory. One or more frames of data are transmitted from the buffer memory to the host, wherein the number of blocks ending in the frame is recorded in a blocks/frame register, and possibly also in a block count accumulator register. Buffer release pulses for releasing buffer space in memory are sent to the buffer memory, based on the number of blocks in the blocks/frame register, or the number of blocks accumulated in the block count accumulator register when a signal is received from the host. A pointer which points to the last block of data successfully transferred is updated in accordance with the buffer release pulses.

Method and apparatus for coordinating transmission of data between a storage medium and a host

A storage controller including a processor and a memory controller. The processor is configured to generate a command corresponding to a first write operation and a second write operation, in which the first write operation is contiguous to the second write operation, and the first write operation is received prior to the second write operation. The command arranges the second write operation prior to the first write operation. The memory controller is configured to, in response to the command, execute each of the first write operation and the second write operation. The second write operation is executed by the memory controller prior to the first write operation.

Disk controller configured to perform out of order execution of write operations

Methods and systems associated with re-transferring data that was unsuccessfully transmitted to a host are described. According to one embodiment method includes receiving a first command to transfer data to a host, wherein the data is arranged in blocks. The data is transferred to the host. When an unsuccessful status is received from the host indicating a transmission error occurred for the first command, a block being transferred when the transmission error occurred is identified. The data in the identified block is re-transferred to the host without re-transferring successfully transferred blocks.

A method and system for command queuing in disk drives may improve performance by queuing multiple commands and sequentially executing them automatically without firmware intervention. The method may use a number of queues, e.g., a staging queue for commands to be executed, an execution queue for commands currently being executed, or a holding queue for commands which may have been executed but have not received a status report from a host.

Hard disk controller which coordinates transmission of buffered data with a host

Methods and systems for performing a restart operation in a storage device include in response to receiving an unsuccessful status from the host indicating a transmission error occurred for the first command, determining in which one of the logical blocks the transmission error occurred and from which data needs to be resent, thereby identifying a restart logical block. The number of blocks in the restart logical block that were sent successfully are calculated, and it is determined that the block immediately following the blocks sent successfully is the block in which the transmission error occurred and from which the data needs to be resent, thereby identifying a restart block. An amount of data successfully sent in the restart block is determined; and an indication is provided to resend at least a portion of the data in the restart block to the host.

A FIFO memory has integrated error management to react to different errors according to the current state of operation of the input and output as well as internal conditions such as buffer memory status. The FIFO memory completes or aborts current operations according to state and leaves the FIFO memory in known condition following error handling. Thus, data sent to a host avoids data gaps or data overlaps because the FIFO memory leaves operations in a known state before reporting the error to a controller.

Concurrent input/output control and integrated error management in FIFO

Method and system for command queuing in disk drives

A controller for a disk drive including a buffer memory controller and a processor. The buffer memory controller is configured to receive a first write operation and a second write operation in a first order, store the first write operation and the second write operation in a buffer memory, and transfer write data corresponding to the first write operation and the second write operation in a second order that is different than the first order in response to a single command. The processor is configured to arrange the first write operation and the second write operation in the second order and issue the single command according to the second order.

A method for maintaining flow control in a buffer memory coupled to a storage controller is provided. The storage controller includes, first and second counters that are used to monitor when data is read from a buffer memory and when data is transferred from the buffer memory to the host. The method includes, incrementing first and second counter values when data is placed in the buffer memory; decrementing a first counter value when data is read from the buffer memory; and decrementing the second counter value when data is sent to a host. The method further includes, pausing a first channel logic between a transport module and a storage disk when there is no data in the buffer memory; and pausing a second channel logic between a disk and the buffer if there is no space in the buffer memory.

System and method for controlling buffer memory overflow and underflow conditions in storage controllers

A controller including an interface module and an index module. The interface module is configured to connect devices. The index module is configured to include, in a table stored in memory, an entry for each of the devices. Each entry includes an address field. The index module is configured to: receive a frame of data including an address of one of the devices; compare the address to the address fields associated with the entries in the table; in response to the address matching one of the address fields, access an index value identifying an entry of the table when the address matches one of the address fields; and in response to the address not matching one of the address fields, generate the index value. The index value is used to connect the device associated with the matching one of the address fields with the one of the devices.

Growing community of inventors

Anaheim, CA, United States of America

Kha Nguyen