Method for detecting and recovering an inoperable drive

A method in a storage device for recovering from a firmware failure that places the storage device in an undetectable state, the storage device comprises a controller to perform the method comprising:

• receiving, by a watchdog timer, periodic signals from firmware threads;
• identifying a firmware failure when one of a periodic signal is not received by a failure detector module from a firmware thread, an initialization counter value is greater than an initialization threshold, and a notification is received of a predefined number of power cycle events occurring with a given time frame;
• triggering an interrupt signal to the controller to execute a recovery routine when the failure detector module identifies the firmware failure;
• updating a boot address;
• obtaining a light mount of recovery firmware from a memory device based on an updated boot address;
• recovering the storage device in a recovery mode with the light mount of the recovery firmware;
• and
• performing phased recovery actions to restart the storage device.

A method in a storage device for recovering from a firmware failure that places the storage device in an undetectable state, the storage device comprises a controller to perform the method comprising:

• identifying a firmware failure when one of a periodic signal is not received by a failure detector module from a firmware thread, an initialization counter value is greater than an initialization threshold, and a notification is received of a predefined number of power cycle events occurring with a given time frame;
• updating a boot address;
• obtaining recovery firmware from a memory device based on an updated boot address;
• recovering the storage device in a recovery mode with the recovery firmware;
• executing a light mount of the recovery firmware to restart the storage device and enter a temporary read-only mode to enable a host to back-up data stored on the storage device;
• and
• executing a first recovery attempt by exiting a read-only mode to allow read and write operations, and if the first recovery attempt fails, executing a second recovery attempt by sending logical-to-physical tables to a host, erasing the logical-to-physical tables on the storage device, and retrieving the logical-to-physical tables from the host to allow read and write operations.