Electrical computers and digital processing systems: memory – Storage accessing and control – Control technique
Reexamination Certificate
1999-10-22
2001-12-18
Peikari, B. James (Department: 2186)
Electrical computers and digital processing systems: memory
Storage accessing and control
Control technique
C711S112000
Reexamination Certificate
active
06332182
ABSTRACT:
FIELD OF THE INVENTION
This invention relates generally to digital computer systems, particularly to mass storage subsystems. More particularly, the invention relates to layout of sectors within tracks of disk drives and to control of operations (such as format, write and read) that access such sectors.
BACKGROUND OF THE INVENTION
Computer systems often contain a storage device such as a disk drive.
FIG. 1
shows a typical architecture of a disk subsystem A disk drive subsystem can be described as a system of disk platter(s); analog read channel and write channel circuitry; an encoder/decoder; a servo engine including a mechanical mechanism to position the read/write heads on the disk platter(s) and an electronic mechanism to control the head position; a local microprocessor (sometimes referred to as a local microcontroller); local processor memory; buffer memory; and disk drive controller circuitry. It is the task of the disk drive controller to interface with the host central processor or microcomputer and correctly read and write user data onto the disk.
Disk Organization
In a magnetic disk storage system, each disk platter is organized in tracks and sectors (also known as logical sectors). Each platter of a disk drive contains many concentric tracks. A track is an annular region of a given radius (actually, of a small but finite radial width) where data is stored. If both sides of a platter are used for storage or if the disk drive contains multiple platters, then multiple heads are required. A cylinder is the set of tracks that can be accessed by choosing among the heads without moving the heads to a new track position. Usually, it is necessary to transfer digital data onto the disk drive in sectors of a predetermined size. A sector is the smallest unit of user data that the host computer system can access (format, write or read) independently. A sector is a predetermined contiguous arc portion of a track having its own identification and the capability of magnetically storing a predetermined amount of digital data, for example, 512 bytes of user data.
In hard disk drives, tracks and sectors are often physically pre-defined at the time of manufacture of the disk by the placement of servo burst signals embedded within the recording area of each track. These servo bursts are used to define, measure and control the running positions (inward versus outward) of the read/write heads on the disk drive. One servo burst per track is uniquely identified as the index servo burst, and these indices align across tracks. Each sector normally has a first part that contains the sector identification or ID fields. The ID field for each sector is typically only written each time a disk is actually formatted. One task of the disk controller is to initialize or format the disks prior to storing any user data. Formatting defines the area of a disk to be set up as zones, tracks, and sectors. An alternative to the above-described embedded servo systems is a dedicated servo system in which one track per cylinder is dedicated to containing only servo information.
Many bytes in addition to the user data bytes are actually necessary within a sector in order to reliably address the sector on the disk and to provide for detection and correction of errors that may occur in the read/storage/write process due to phenomena such as defects in the platter media, electrical noise, and variations in the rotational speed of the platter(s). Another task of the controller is to add error correction codes (ECC) and cycle redundancy checking (CRC) to the user data in order to detect and possibly correct data that has been corrupted because of disk damage or some other problem. Additionally, the controller meshes the slow read and write times of the disk with the high speed transfers to and from the host computer. Additionally, the controller must monitor and accurately control the timing of reading and writing of data off of and onto the disk. This has been accomplished by means of a high speed format microengine, or formatter microengine, that executes microinstructions contained in a writable control store (WCS) contained in the controller.
The format microengine provides the data and the control required to perform the transfer of parallel digital user data into formatted, addressed, and error-protected channel data that can be serially placed on a disk. The data that is normally written into a sector contains the following: an ID (identification) field that acts like an address, a data field, an error correction field, gaps that are necessary to allow the read/write channel electronics to switch from reading to writing; and predetermined bit and byte patterns that are required to recover the exact bit frequency (which varies somewhat due to variations in the rotational speed of the platter), bit phase, and alignment of bit-stream data into bytes. The ID field and the data field each make up a concatenated set of bytes within a sector.
One common hard disk format that has been extensively used is to follow each servo burst with a gap, then a series of synch pulses to establish bit-time synchronization across read/write speed variations, then a mark byte to establish byte alignment within the bit stream, then a sector ID field that contains track number and sector number, then a gap, then a series of synch pulses to establish bit-time synchronization across read/write speed variations, then a mark byte to establish byte alignment within the bit stream, then a 512-byte data field, then a 7-byte ECC (Error Correction Code) field. In this format each sector is preceded by the servo burst. Servo bursts are written only at the time the disk is manufactured-portions are written slightly off the center of the track in order to allow the servo engine to keep the head centered on the track. Thus, during servo burst time, the access operations of reading, writing and formatting must be suspended.
One problem arising in modern disk drive system designs is lack of flexibility to accommodate the ever more complicated formats required to increase density, decrease latency, and improve reliability.
In the prior art, magnetic disk drive sector ID information is generally provided once as part of a substantially larger preamble to the sector data, with error tolerance being provided by CRC bits appended to the sector ID information for error detection purposes. While the CRC keeps misreading errors very low, the ID information for a particular sector will be lost in the event of hard errors in the ID field reading. In the case of optical drives where error rates are much higher, it is common to repeat the entire ID field, including bit-synchronizing and byte-aligning information, a number of times so as to allow repeated ID field reading attempts during a single pass to greatly increase the likelihood of at least one good read of an ID field for each sector. This technique has not been adapted in the case of magnetic disk drives however, because of the relative amount of storage capacity that would have to be dedicated to the entire and repetitive ID fields.
Defect Management
If a sector is, or over time becomes, defective (i.e. unable to accurately reproduce the data written thereon), there must be a way to map out to another physical sector the data that the logical sector would otherwise contain, or alternatively to simply skip over the sector (called slipping the sector), which results in a track that is shorter than expected. There is a need for a way to handle defective sectors automatically, i.e. with minimal intervention from the local microprocessor to handle the exception conditions that occur when access is attempted to a slipped or mapped sector. Additionally, there is a need to handle such occurrences efficiently, (i.e. with minimal impact on access times).
An important consideration in modern disk drive systems is the amount of time it requires for the read/write heads to locate and move to a new track position or (access time). This time includes the seek time and the rotational latency. Normally the magnitude
Estakhrf Petro
Geldman John S.
Ho Son H.
Schadegg John J.
Blakely & Sokoloff, Taylor & Zafman
Cirrus Logic Inc.
Peikari B. James
LandOfFree
Disk drive control without identification fields does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Disk drive control without identification fields, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Disk drive control without identification fields will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2599251