Hitachi Computer Products (america), Inc.

Unisys Corporation

Hitachi Computer Products, Inc.

Ruben Yomtoubian

The present invention is a method and apparatus for organizing memory buffers between an i/o subsystem and a computer system. The present invention organizes the memory buffers that exist between a computer system and an associated i/o subsystem into a memory hierarchy. An example of such a hierarchy includes, in order, a disk buffer, a disk cache, a controller memory and a computer system memory. Each datablock in the memory hierarchy is given an associated index word. The index word contains the address of the datablock within the media of the disk drive and contains index bits which indicate where the datablock may be found in the memory hierarchy. As a datablock moves between the disk drive and the computer system, a mirror copy of the datablock is maintained in a memory buffer that is in the next adjacent level of the memory hierarchy. In this way, the mirror copy may be used if the datablock becomes inaccessible.

Method for multiple verticle and/or horizontal levels of memory

The present invention is a method and apparatus for predicting the future failure of a disk drive. Structurally, the present invention is implemented as a series of diagnostic and statistical routines contained in a ROM on the logic board of a disk drive. The routines detect errors that occur during normal operation of the disk drive and issue a warning if the number of errors or the rate at which errors occur has exceeded predetermined levels. In addition to monitoring errors that occur during normal operation, the diagnostic routines may be activated to test for the presence of recoverable errors. Recoverable errors are detected by running a series of diagnostic tests after enabling error detection and disabling error correcting codes as well as device retries. Once again, a warning is issued if the number of errors or the rate at which errors occur has exceeded predetermined levels. Finally, the present invention may be used in combination with a list of suspected sectors that has been previously compiled by an internal disk tester. In such cases, the suspected sectors are periodically re-tested and a warning issued if the number of errors, or the rate of errors has exceeded predetermined levels. At the conclusion of testing, the drive is restored to its normal operating configuration with error correcting codes and device retries enabled.

Method and apparatus for protecting disk drive failure

The present invention is an inline tester for analyzing the integrity of a disk drive. Structurally, the inline tester includes a set of statistical and diagnostic routines included in the ROM of a disk drive and a diagnostic logic section included in the drive's logic board. In initial use, the inline tester reconfigures the drive, making the drive less tolerant of flaws within the drive's media. The drive is also reconfigured to enable error detection and disable error correcting codes as well as device retries. Each sector of the media is then tested with multiple test patterns and failing sectors are added to a list of suspected sectors. At the conclusion of testing, the drive is restored to its normal operating configuration with error correcting codes and device retries enabled. Later, each sector in the list of suspected sectors is re-tested. Sectors that fail are added to the drive's list of bad sectors and are replaced by an alternate sector allocated from a reserved portion of the drive.

Inline disk tester

Techniques for certifying recording sites (bit sites) on magnetic recording disks wherein test recording patterns are recorded along disk tracks: first in an analog mode, then in a digital mode; concluding in a full digital test that simulates customer operation.

Growing community of inventors

Saratoga, CA, United States of America

Ruben Yomtoubian