JPH0883151A - Magnetic disk device - Google Patents

Abstract

PURPOSE: To prevent the alteration of data from occurring by erasing unerased data due to the abnormal completion of a padding operation by executing reformating on a record starting from a first record when an error has occurred while the padding operation is being executed. CONSTITUTION: When the error has occurred while the padding operation is being executed, the error is held by an error holding means 50, and a cylinder value and a head value that is the value of a physical position where the error has occurred, are held in a cylinder value/head value holding means 62. When no sync/no data is detected by seeking the cylinder value and the head value and executing a read operation, the reformatting is executed on the record starting from the first record by a reformatting execution means 55A, and after that, the padding operation is started. Since the abnormal completion of the padding operation can be detected and the unerased data due to the abnormal completion of padding can be erased, abnormality due to the reading of the unerased data can be prevented from occurring, and the alteration of data can be prevented from occurring in the worst case.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk device which executes a format write release function in response to a command from a host device. The format write release function in the count key data format (hereinafter referred to as the CKD format) erases the remaining data of the track from the time when the writing of the record is completed when the command is issued by setting the length of the record. This is a function of performing an operation by the magnetic disk device alone, and is usually called a padding operation. With this function, the host device is released from the control of the magnetic disk device, so that the efficiency of the system is improved.

Further, in recent magnetic disk devices and the like, there is a further demand for high density recording of writing by shifting to a small diameter disk by downsizing. In addition, active maintenance is required to reduce maintenance time and replace parts without stopping the system in the event of a failure.

[0003]

2. Description of the Related Art As a conventional magnetic disk device, there is one in which a plurality of magnetic disk devices are stored in a disk unit and are controlled by a disk controller (CTRL) which is a higher-level device. The disk controller (CTRL) is connected to the CPU via a channel.

FIG. 13 shows an example of the CKD format used in such a magnetic disk device (DRV). All tracks start with an index mark (INDEX) and end with an index mark (INDEX). The track is composed of a home address (HA) and records 0 to N, which are separated by a gap. The record is composed of a count section (C), a key section (K), and a data section (D), which are also separated by a gap. An address mark (AM) is written in front of the count section (C) after record 1. Address mark (A
M) represents the start position of the record that can be used by the user.

Next, FIG. 14 shows an example of the contents of the home address (HA) and the count section (C). In FIG. 14, the SC section indicates a skip control (SKIP CONTROL) indicating the position of the scratch on the track, the SN section indicates the segment (SEGMENT) number, and the PA section indicates the physical address to which the count section (C) belongs. Value, head value, and flag byte (FLAG BY
TE), C, H and R parts are logical addresses to which the count part belongs, KL part is key length, DL part is data length,
Error correction code (ERROR CORR
Each of the ACTION CODEs) is stored.

Next, FIG. 15 shows an example of padding operation.
FIG. 15A shows a track format, and FIG.
(B) is a higher-level device (CTRL) and a magnetic disk device (D)
It shows the exchange of commands and responses with RV). In an example of formatting the record N, the host device has a format G3 instruction (writes the count portion), a format G2 instruction (writes a key and a data portion), an erase G3 instruction (writes gap data to the count portion), and an end. After the read / write command (end of the read / write operation from the host device) and each data (C data, K data, D data) is issued, the connection with the magnetic disk device is released.

The magnetic disk device executes the above-mentioned command, responds (C data reception, K data reception, D data reception, erase G3 command reception), and receives the end read / write command. The write operation according to the instruction is stopped, and the write operation is switched to the padding operation for writing the gap data by the magnetic disk device itself. This operation ends by detecting the index mark. A indicates a write operation according to an instruction from the host device, B indicates a write operation according to an instruction from the padding control circuit, and C indicates that the remaining data is erased.

Although not shown in FIG. 15, a pad (PAD) instruction is provided between each instruction (a gap
Write data) is included, but the description is omitted.
Next, FIG. 16 shows a padding control circuit section. The padding control circuit unit is included in the format control unit in the magnetic disk device.

When a read / write instruction is sent from the host device, read / write active (RWACT) is turned on. When the format command is received, the processor write gate (PWTGT) is turned on, and then the cue pad (CUEPAD) is turned on. When the processor write gate is turned off by the read / write end instruction, the AND circuit 1 is turned on, and the flip-flop 2 is set.
The padding (PADDING) is output.

The padding (PADDING) and the processor write gate (PWTGT) are ORed by the OR circuit 3, and the write gate (WTGT) is sent to the read / write analog section to write on the medium. Also,
An index mark (INDEX) and an error (CHK) are input to the OR circuit 4, and the output resets the flip-flop 2.

Further, write data (WTDT) and gap data from the host device are input to the multiplex 5, and write data (WTDT) to be written on the medium when the padding (PADDING) is turned on is written from the host device. To gap data. Read / write active (RWACT), processor write gate (PWTGT), keypad (CUEPAD)
Is instructed by the processor unit.

Next, FIG. 17 shows an example in the case of abnormal termination of the padding operation. In FIG. 17, when the write operation of the record N is completed and an error occurs at the position of asterisk while executing the padding according to the instruction from the padding control circuit, the padding operation is completed. Causes of errors include abnormal write current flowing in the data head, abnormal head selection signal, and abnormal position control (the head moved to a position exceeding the write slice level due to on-track). This is to prevent data destruction other than the existing tracks.

As a result, the erasing operation of the remaining data due to the padding operation is stopped midway, and the old record (data) to be erased thereafter remains. In FIG. 17, D indicates an erase section and E indicates an area in which previous data remains.

[0014]

In such a conventional magnetic disk device, a search operation of an address mark is normally performed as a read operation start of the CKD format, and if detected, it is judged that there is a proper record after that. Therefore, the count key data is read, and therefore, when an address mark is detected in the above-mentioned erasure-missing record, the record is processed as normal data.

As a result, there is a problem in that the reading operation is abnormal or, in the worst case, the data is garbled. The present invention has been made in view of such a conventional problem, and by erasing unerased data due to abnormal termination of the padding operation, it is possible to prevent an abnormal reading operation or garbled data. An object is to provide a magnetic disk device.

[0016]

FIG. 1 is a diagram for explaining the principle of the present invention. The present invention relates to an error holding means 50 for holding an error when an error occurs during execution of a padding operation in a magnetic disk device 15 that performs format write release in a predetermined format in response to a command from a host device 13, When no-sync / no-data is detected by performing a read operation by seeking to the cylinder value / head value holding means 62 for holding the cylinder value / head value which is the physical position where the error has occurred, and the cylinder value / head value. The reformatting executing means 55A for reformatting from the first record is provided, and the padding operation is started thereafter.

Further, according to the present invention, in the magnetic disk device 15 for performing format write release in a predetermined format in accordance with an instruction from the host device 13, if an error occurs during execution of the padding operation, error holding is held. A means 50, a cylinder value head value holding means 62 for holding a cylinder value and a head value which are the physical position where the error has occurred, and a seek operation for the cylinder value and the head value to execute a read operation to obtain no sync / no data. When it is detected, it is provided with a positioning control means for controlling to position it at the end of the record before the no-sync / no-data is detected, and after that, the padding operation is started.

The present invention further comprises address mark detecting means for performing an address mark search operation after detecting the no-sync / no-data to detect whether or not there is an address mark before the index mark.
Further, according to the present invention, in the magnetic disk device 15 for performing format write release in a predetermined format in accordance with a command from the upper level device 13, when an error occurs during the padding operation, an error holding means 50 is provided. A cylinder value which is a physical position where the error has occurred, a head value, and a sector value obtained by the output of the error holding means, a head value sector value holding means, and a seek after the cylinder value and the head value Comparing means for performing a sector search operation and comparing the output of the sector counter with the target sector, which is the sector value, starts the padding operation when the sectors match.

[0019]

According to the magnetic disk drive of the present invention having such a configuration, when an error occurs during the padding operation, the error is held and the cylinder value and the head value, which are the physical position where the error occurred, are stored. Hold, seek to cylinder value, head value, execute read operation, and when no sync / no data is detected, reformatting is executed from the first record and padding operation is started after that. Since abnormal termination can be detected and unerased data due to abnormal termination of padding can be erased, it is possible to prevent abnormalities due to reading of unerased data, and in the worst case, prevent garbled data.

Further, seeking to the cylinder value and the head value,
When no-sync / no-data is detected by executing the read operation, it is controlled so that it is positioned at the end of the record before no-sync / no-data is detected, and the padding operation is started after that. Since the unerased data due to abnormal termination of padding can be erased in the same manner as in the case, it is possible to prevent the abnormality due to the reading of the unerased data and, in the worst case, prevent the garbled data.

Further, since the address mark search operation is performed after no sync / no data is detected to detect whether or not there is an address mark up to the index, an abnormal record is detected only when the address mark is detected. Is judged to be unerased and will be erased.
In addition to the above effects, the connection time with the host device 13 can be shortened, and the efficiency of data transfer can be improved.

Further, in addition to the cylinder value and the head value, the sector value is held, after seeking to the cylinder value and the head value, the target sector, which is the sector value, is compared with the sector count value, and when they match, the padding operation is started. By doing so, not only the above-mentioned abnormal read operation and data garbling can be prevented, but also the connection time with the host device 13 can be shortened and the efficiency of data transfer can be improved. .

[0023]

Embodiments of the present invention will be described below with reference to the drawings. 2 to 6 are views showing a first embodiment of the present invention. FIG. 2 is an overall configuration diagram showing a first embodiment of the present invention. In FIG. 2, reference numeral 11 denotes a CPU, and a disk controller (CTRL) 13 which is a higher-level device is connected to the CPU 11 via a channel (CH) 12. A disk unit (DKU) 14 is connected to the disk control device 13, and a plurality of magnetic disk devices (DRV) 15 are stored in the disk unit 14.

Next, FIG. 3 is a block diagram of the host device (CTRL) 13 and the magnetic disk device (DRV) 15.
In FIG. 3, reference numeral 13 is the host device, 15 is the disk magnetic device, and the disk magnetic device 15 includes a driver receiver 16, an interface control unit 17, a processor unit 18, a program memory 19, a drive control unit 20, and a servo control unit 21. , Servo analog section 22, disk enclosure 23, format control section 24, VFO section 2
5 and a read / write analog section 26.

The disk enclosure 23 includes a plurality of media 27 and a spindle shaft 28 on which the plurality of media 27 are mounted.
A spindle motor 29 for rotating the spindle shaft 28; a data head 30; a servo head 31; a carriage 32 connecting the data head 30 and the servo head 31; and a voice coil motor for positioning a target cylinder. 33.

The read / write analog section 26 applies a current to the data head 30 at the time of writing to write data on the data surface of the medium 27, and at the time of reading, the data head 3
The data from the medium 27 is read by the current from 0. VF
The O unit 25 adjusts the clock frequency to the data when reading and the clock frequency to the servo information when writing. Further, the VFO section 25 has an address mark (AM).
Generate

The format controller 24 controls the format and generates padding (PADDING), pad check (PAD CHK), no sync, no data, and the like. The servo analog unit 22 controls the current to the voice coil motor 33, and the servo head 3
Positioning 1 is performed.

The servo control unit 21 performs servo control,
The drive control unit 20 performs control necessary for driving. The drive control unit 20 has a sector counter and outputs a sector count value. The interface unit 17 controls the interface, and the driver receiver 16 controls the host device 13.
Send and receive signals to and from.

The processor section 18 has a processor and an external register and controls each section. A control program is stored in the program memory 19. Host device 1
3 issues a selection command for the magnetic disk device 15 and a status request command for the magnetic disk device 15. If there is an error report, the content of the error is judged, and the error position information request command and the magnetic disk device 15 report. Cylinder value, head value seek command, all record read command, no sync / no data detection report from the magnetic disk unit 15 to issue reformatting command from the first record, end read / write command, etc. Deselect to.

Next, FIG. 4 is a block diagram of the format control unit 24 and the processor unit 18. In FIG. 4, 2
Reference numeral 4 denotes the format control unit, and in the format control unit 24, registers 41 and 42, an AND circuit 43,
44, OR circuits 45 and 46, flip-flops 47 and 4
8, 49, 50, knot circuits 51, 52, comparison circuit 53
And a multiplexer 54 is provided.

The clock input terminal of the register 41 is supplied from the processor unit 18 to the processor address (PRO AD
D) is input, data is input to the data input terminal from the processor unit 18 via the processor bus (PRO BUS), and read / write control (RWC) is output. The contents of read / write control (RWC) are
For example, write sync byte (WTSB), processor write gate (PWTGT), read / write act (RW)
ACT), cue pad (CUEPAD), check reset (CHECK RST), read gate (RDG)
T), which are output respectively.

The clock input terminal of the register 42 is connected to the processor address (PRO AD
D) is input, the data from the processor unit 18 is input to the data input terminal via the processor bus (PRO BUS), and the sync byte pattern (SBPAT) is output. The AND circuit 43 has a read / write act (RW
ACT), cue pad (CUEPAD), and processor write gate (PWTGT) are input, and the processor write gate (PWTGT) is turned off, the AND circuit 4
3 turns on and sets flip-flop 47.
As a result, the flip-flop 47 is padded (PAD
DING) is output. An index mark (INDEX) and an error (CHK) from the drive control unit are input to the OR circuit 45, a logical sum is obtained, and the output thereof resets the flip-flop 47. Thus, the padding operation is interrupted when an error occurs.

The OR circuit 46 has padding (PADDI
NG) and the processor write gate (PWTGT) are input, the logical sum is calculated, and the write gate (WTGT) is VF.
It is output to the O section 25. When the padding (PADDING) is input to the multiplexer 54, the gap data is output as write data to the VFO unit 25, and when the padding (PADDING) is off, the host device (CTR
The write data from (L) 13 is output to the VFO unit 25, and the sync byte pattern (SBPAT) is output to the VFO unit 25 as write data (WTDT) by the input of the write sync byte (WTSB).

The AND circuit 44 has padding (PADD
ING) and an error (CHE) are input, a logical product is calculated, and the output is input to the clock input terminal of the flip-flop 48. Further, “1” is input to the data input terminal of the flip-flop 48,
8 is a pad check (P) indicating that an error has occurred during padding due to the input of "1" from the AND circuit 44.
AD CHK) is output and check reset (CHEC)
K RST) is reset.

The flip-flop 48 has a function as error holding means for holding an error when an error occurs during the padding operation. The read data (RDDT) from the VFO unit 25 and the sync byte pattern (SBPAT) are input to the comparison circuit 53. When the two data match, "1" is output, and when they do not match, "0" is output.

A read clock (RDCL from the VFO section 25 is supplied to the clock input terminal of the flip-flop 49.
K), the output from the comparison circuit 53 is input to the J input terminal, “0” is input to the K input terminal, and the read gate (RDGT) inverted by the knot circuit 51 is input to the reset terminal. When it is "0", the sync byte fund is output and is reset when the read gate (RDGT) is turned off.

A cell (CELL) (output of the sector counter) from the drive controller 20 is connected to the clock input terminal of the flip-flop 50, and the knot circuit 5 is connected to the data input terminal.
The sync byte fund inverted by 2 and the read gate (RDG) inverted by the knot circuit 51 at the reset terminal.
T) is input and no sync / no data is output.
That is, when the signals compared by the comparison circuit 53 do not match, no sync / no data is output.

Reference numeral 18 denotes the processor section. The processor section 18 is provided with a processor 55, a decoder 56, a non-volatile memory 57, registers 58 and 59, and a multiplexer 60. A write select (WS) and a processor address (PRO ADD) are input to the decoder 56 from the processor 55, and after being decoded, they are stored in the registers 41, 42, 58, 59, the register 61 of the interface control unit 17, and the non-volatile memory 57. It is output respectively. The cylinder value (CAR), which is the physical position where the error has occurred, is read from the non-volatile memory 57 by the processor 55 via the processor bus (PRO BUS) and set in the register 58.

Further, the head value (HAR), which is the physical position where the error has occurred, is read from the non-volatile memory 57 by the processor 55 via the processor bus (PRO BUS) and set in the register 59. Registers 58 and 59 are a cylinder value and a head value holding means 6 for holding a cylinder value and a head value which are physical positions where an error has occurred.
It has the function of 2.

Various signals are written in the nonvolatile memory 57 by the processor address (PRO ADD) and the write select (WS), and the processor address (PRO
Various signals are sent from the nonvolatile memory 57 to the processor bus (PRO) by the ADD) and the output enable (OE).
BUS). A pad check (PAD CHK) is applied to the input terminal (I0) of the multiplexer 60.
No sync / no data is input to the input terminal (I1), a cylinder value is input to the input terminal (I2), and a head value is input to the input terminal (I3).
Output enable (O
When E) is on, the output becomes valid and various signals are sent to the processor bus (PRO BUS).

Processor 55, format control unit 2
4, the VFO unit 25, the read / write analog unit 26, etc. seek the cylinder value and the head value, execute the read operation of each record, and when no sync / no data is detected, a reformatting command from the host device 13 is performed. Thus, it has a function as a reformatting executing means 55A for reformatting from the first record.

The register 61 provided in the interface controller 17 is a processor address (PRO ADD).
Various signals selected via the processor bus (PRO BUS) are set, and the output is sent to the host device 13 via the driver receiver 16 as a bus in (BUS IN). Next, FIG. 5 is a time chart showing an operation example of the format control unit 24.

In FIG. 5, (A) shows medium information,
Here, the address mark (AM), sync byte (S
B), a count section is shown.
(B) shows an index mark (INDEX), which shows the start and end of a track. (C) is a cell (CEL
L), which is the output of the sector counter. This is a signal that divides the track into equal parts.

(D) is a read / write act (RWAC).
T), the read / write act (RWACT) is turned on when a read / write command is sent from the higher-level device 13, and the read / write act (RWACT) is turned off when the read / write command is turned off. In (E), as will be described later, an address mark (AM) is created by an address mark write (AM write), and an address mark fund is detected by an address mark search (AM search) as shown in (F). It

The read / write act (RWACT) is turned on, the cue pad (CUEPAD) shown in (H) is turned on, and the processor write gate (PWTG) shown in (G).
When T) is turned off, padding (PAD) shown in (I)
DING) turns on. The write gate (WTGT) shown in (J) is a processor write gate (PWTG).
T) turns on and padding (PADDING)
When turns off, it turns off. Padding (PADDI
NG) is turned off by an index mark (INDEX) or an error (CHK).

(K) shows a read gate (RDGT), and when the read gate (RDGT) is turned off, (L)
The sync byte fund shown in is turned off. (M) indicates write data, and the gap and address mark (A
M), gap, sync byte (SB) are written,
Subsequently, the data from the higher-level device 13 is written, and then the padding by turning on the padding (PADDING) is written.

Next, FIG. 6 is an operation explanatory view for explaining the operation of this embodiment. In FIG. 6, first, the host device 13
Issues a selection command to the magnetic disk device 15 to select one of the magnetic disk devices 15. The magnetic disk device 15 receives the selection command at the processor unit 18 via the driver receiver 16 and the interface control unit 17, and the corresponding magnetic disk device 15 responds to the host device 13.

Next, the higher-level device 13 issues a status request command to the magnetic disk device 15 that responded. The magnetic disk device 15 receives the status request command and reports an error if an error has occurred.
If an error occurs during the padding operation, the flip-flop 48 is set and at the same time the padding operation is stopped.

The pad check (PAD CHK) output from the flip-flop 48 is selected by the processor address (PRO ADD), and when the output enable (OE) is turned on, it is sent from the multiplexer 60 to the processor bus (PRO BUS). , The processor address (PRO AD
D) and write select (WS). Pad check written in the non-volatile memory 57 (PA
D CHK is the processor address (PRO AD
D) and output enable (OE), it is read to the processor bus (PRO BUS) and sent to the host device 13 as an error report.

When the host device 13 determines the error content and determines that the pad check (PAD CHK) is detected, it issues a request command for error position information to the magnetic disk device 15. Magnetic disk drive 15
Then, when the error position information request command is received, the cylinder value and the head value, which are the physical positions where the error has occurred, are reported to the host device 13. When an error occurs during the padding operation, the cylinder value is held in the flip-flop 58 and the head value is held in the flip-flop 59 and then stored in the nonvolatile memory 57 via the multiplexer 60.

The cylinder value and the head value stored in the non-volatile memory 57 are sent to the host device 13 in response to an error position information request command from the host device 13. Upon receiving the report of the cylinder value and the head value, the upper device 13 issues a seek command to the cylinder value and the head value to the magnetic disk device 15. Upon receiving the seek command, the magnetic disk device 15 executes the seek and reports the seek completion to the higher-level device 13. The seek instruction is the driver receiver 1
6, servo controller 3 via interface controller 17, processor 18, drive controller 20, servo controller 21, servo analog unit 22, voice coil motor 33
It is executed in 1.

When the higher-level device 13 receives the seek completion report, it issues a read command for all records to the magnetic disk device 15.
Issue to Upon receiving the read command, the magnetic disk device 15 executes the read operation of each record. If no sync / no data is detected during execution of this read operation, this is reported to the host device 13.

The no-sync / no-data is an inverted input of the sync byte fund output from the flip-flop 49 when the output of the comparison circuit 53 for comparing the read data (RDDT) and the sync byte pattern (SBPAT) is "0". Is detected by the flip-flop 50.
Upon receiving the no-sync / no-data detection report, the higher-level device 13 issues to the magnetic disk device 15 a reformatting command for reformatting from the first record of the track.

When the magnetic disk device 15 receives the reformatting command, the write HA → write ROC → format ROD → format G3 → format G2.
→ The reformatting command is executed by the command chain consisting of erase G3. In this way, the magnetic disk device 15 executes the full reformatting from the first record.

Next, the host device 13 issues an end read / write command to the magnetic disk device 15, and the magnetic disk device 15 starts the padding operation. Then, the host device 13 cancels the selection for the magnetic disk device 15. In this manner, since it is possible to detect that an error has occurred during the padding operation and erase the unerased data when the padding ends abnormally, there is no abnormality caused by reading the unerased data.

In the worst case, it is possible to prevent garbled data from occurring. Next, FIGS. 7 and 8 are views showing a second embodiment of the present invention. In FIG. 7, 55 is a processor of the processor unit 18, and the processor 55
When receiving a positioning command from the host device 13 to position the end of the record before the no-sync no-data is detected, the positioning is performed at the end of the record before the no-sync no-data is detected. It has a function as a positioning control means 55B for controlling. In addition,
The other structure is the same as that of the first embodiment, and the description thereof is omitted.

FIG. 8 is a diagram for explaining the operation of the second embodiment, and is the same as in the first embodiment up to the report of no-sync / no-data detection, and the description of each operation will be omitted and the characteristic part will be described. Upon receiving the no-sync / no-data detection report from the magnetic disk device 15, the host device 13 issues a positioning command to the magnetic disk device 15 to position it at the end of the record before the no-sync / no-data detection.

The method for positioning the record at the end before the no-sync no-data is detected is as follows: the physical segment number (SN) and the key data length representing the physical position of the record in the track of the last record read. (KL / DL), the host device 13 converts the key data length from the number of bytes to the number of segments, and the value obtained by adding it to the number of physical segments becomes the last position of the record. Position the number by performing a space operation (moving the specified number of cells).

As another method, the physical segment number is converted into a sector value, the sector search operation is executed with an arbitrarily smaller value than that value, and if a match with the sector is obtained, an address mark search is performed. Positioned on the record. Upon receiving the positioning command, the magnetic disk device 15 executes the positioning command and reports the completion of the positioning operation to the host device 13.

Upon receipt of the positioning operation end, the host device 13 issues an erase G3 command to the magnetic disk device 15. The magnetic disk device 15 executes the erase G3 operation, and reports the end of the erase G3 operation to the host device 13.
Upon receiving the erase G3 operation end, the host device 13 issues an end read / write command to the magnetic disk device 15.

The magnetic disk device 15 starts the padding operation in response to the end read / write command. Thereafter, the host device 13 cancels the selection for the magnetic disk device 15. Also in this embodiment, when an error occurs during the padding operation, unerased data due to abnormal end of the padding operation can be erased, so that the read operation is prevented from being abnormal or the data being garbled in the worst case. You can

Next, FIGS. 9 and 10 are diagrams showing a third embodiment of the present invention. FIG. 9 shows a VF showing the main part of this embodiment.
It is a block diagram of an O section 25. An address mark detection unit 71, an address mark creation unit 72, a serial / parallel conversion unit 73, AND circuits 74 and 75, and an OR circuit 76 are provided in the VFO unit 25.

Serial read data and the address mark search from the processor unit 18 are input to the address mark detecting unit 71 as the address mark detecting means, and it is detected whether or not there is an address mark (AM) up to the index. The address mark fund is output (see FIGS. (E) and (F)). This address mark fund is used for the processor bus (PROB) of the processor unit 18.
US).

When the address mark (AM) is not detected, it is determined that there is no data to be erased, and the subsequent operation is stopped. The address mark creation unit 72 creates address mark data by address mark writing and outputs it. The AND circuit 74 receives the address mark data and the address mark write, and the AND circuit 75 receives the address mark write and the serial write data obtained by converting the write data (WTDT) from the format controller 24 by the serial / parallel converter 73. To do.

When the address mark write is "1", the output of the AND circuit 74 becomes "1", while the output of the AND circuit 75 becomes "0". The OR circuit 76 ORs the output of the AND circuit 74 and the output of the AND circuit 75 and outputs serial write data to the read / write analog section 26. That is, when the address mark write is output, the address mark data is output, and when the address mark write is off, the write data is output from the format controller 24 (see FIGS. 5E and 5M). ). Note that the other configurations are similar to those in FIGS.

FIG. 10 is a diagram for explaining the operation of this embodiment. The description of the portions that overlap with those of the first and second embodiments will be omitted, and the characteristic portions of this embodiment will be described. In FIG. 10, the higher-level device 1
3 receives the no sync no data detection report,
An address mark (AM) search command is issued to the magnetic disk device 15.

When the magnetic disk device 15 receives the address mark search command, the index mark (INDE
The address mark detection unit 71 detects whether or not there is an address mark (AM) before the address mark (A).
When M) is detected, the address mark fund is output, and when the address mark (AM) is not detected, it is determined that there is no data to be erased, and the operation ends.

When the address mark (AM) is detected, it is determined that the unauthorized data has not been erased, and the magnetic disk device 15 instructs the host device 13 to write the address mark (A).
M) Report detection. When the host device 13 receives the address mark (AM) detection, the same operation as in the above embodiment is performed thereafter.

Also in this embodiment, the same effect as that of the above embodiment can be obtained, and since the connection time with the host device 13 is shortened, the efficiency of data transfer can be improved. Next, FIGS. 11 and 12 are diagrams showing a fourth embodiment of the present invention. FIG. 11 is a block diagram showing a part of the drive control unit 20 and the processor unit 18 showing the main part of this embodiment.

In FIG. 11, the drive control section 20 is provided with a sector counter 81, a comparison circuit 82, and an AND circuit 83, and the processor section 18 is provided with a register 84. A pad check (PAD CHK) is input to the clock input terminal of the register 84, a sector count value (SECTOR COUNT) is input to the data input terminal, and the sector value (SCT) is input to the processor bus (PRO B).
US).

The register 84 is the register 58, 59.
At the same time, it constitutes a cylinder value head value sector value holding means 85 for holding the cylinder value head value sector value.
Servo clock (SERVO CLOCK) from the servo control unit 21 is input to the clock input terminal of the sector counter 81.
Is input and counted, the sector count value (SECT
OR COUNT) is output, and the index mark (INDEX) is input from the servo control unit 21 to the reset terminal to be reset.

The comparison circuit 82 serving as a comparison means includes a sector count value (SECTOR COUNT) and a target sector (TARGET SE) from the processor unit 18.
(CTOR) is input and “1” is output when they match and “0” is output when they do not match. As the target sector (TARGET SECTOR), the sector value (SCT) held in the register 84 is used. Normally, the length of the gap is longer than the sector value (SCT). Therefore, even if padding is performed from the sector value (SCT) where the pad check (PAD CHT) occurred, the normal data immediately before the sector value (SCT) of the track is written. Is not erased.

The AND circuit 83 outputs the output of the comparison circuit 82 and the sector search (SECTOR) from the processor section 18.
SEARCH) is input and sector matching is output to the processor bus (PRO BUS). FIG. 12 is an operation explanatory diagram of the present embodiment. The description of the parts that are the same as those of the above-described embodiment will be omitted, and the characteristic parts of the present invention will be described.

In FIG. 12, when the magnetic disk device 15 receives the error position information request command from the host device 13, it reports the cylinder value, head value, and sector value to the host device 13. The host device 13 issues a seek command to the received cylinder value and head value to the magnetic disk device 15.

When the magnetic disk device 15 receives the seek command, the magnetic disk device 15 executes the seek and reports the seek completion to the host device 13. The higher-level device 13 issues a target padding (TGT PADDING) command with the received sector value to the magnetic disk device 15. When the magnetic disk device 15 receives the target padding command, it starts the search operation, and the processor 18 performs sector search (SECTO).
R SEARCH) to the AND circuit 83.

Target sector (TARGET SEC
TOR) and sector count value (SECTOR COUNT)
When T) coincides and the comparison circuit 82 outputs "1", the AND circuit 83 outputs sector coincidence. Sector matching is
The data is sent to the processor bus (PRO BUS) and the padding operation is started. Also in this embodiment, unerased data due to abnormal termination of padding can be erased, so that an abnormal read operation of unerased data or, in the worst case, garbled data can be prevented and the host device 13 can be prevented.
Since the connection time with and can be shortened, the efficiency of data transfer can be improved.

After the host device 13 executes the error position information request command, the information in the non-volatile memory 57 in the magnetic disk device 15 is cleared. You may go by the command from.

[0078]

As described above, according to the present invention, the abnormal end of padding can be detected and the unerased data due to the abnormal end of padding can be erased. In this case, the data can be prevented from being corrupted. In addition, in another embodiment, in addition to the above effects, the connection time with the host device can be shortened, so that the efficiency of data transfer can be improved.

[Brief description of drawings]

FIG. 1 is an explanatory view of the principle of the present invention.

FIG. 2 is an overall configuration diagram showing a first embodiment of the present invention.

FIG. 3 is a block diagram of a magnetic disk device.

FIG. 4 is a block diagram of a format control unit and a processor unit.

FIG. 5 is a time chart explaining the operation of a format control unit.

FIG. 6 is an operation explanatory diagram

FIG. 7 is a block diagram showing a second embodiment of the present invention.

FIG. 8 is an operation explanatory diagram of the second embodiment.

FIG. 9 is a block diagram showing a third embodiment of the present invention.

FIG. 10 is an operation explanatory diagram of the third embodiment.

FIG. 11 is a block diagram showing a fourth embodiment of the present invention.

FIG. 12 is an operation explanatory diagram of the fourth embodiment.

FIG. 13 is a diagram showing an example of a conventional format.

FIG. 14 is a diagram showing an example of contents of a conventional home address counting section.

FIG. 15 is an explanatory diagram of an operation example of padding.

FIG. 16 is a block diagram showing a padding control circuit.

FIG. 17 is an explanatory diagram of abnormal termination of padding.

[Explanation of symbols]

11: CPU 12: Channel 13: Host device 14: Disk unit 15: Magnetic disk device 16: Driver receiver 17: Interface control unit 18: Processor unit 19: Program memory 20: Drive control unit 21: Servo control unit 22: Servo analog Part 23: Disk Enclosure 24: Format Control Part 25: VFO Part 26: Read / Write Analog Part 27: Medium 28: Spindle Axis 29: Spindle Motor 30: Data Head 31: Servo Head 32: Carriage 33: Voice Coil Motor 41, 42 , 58, 59, 61: register 43, 44: AND circuit 45, 46: OR circuit 47, 48, 49, 50: flip-flop 51, 52: knot circuit 53: comparison circuit 54, 60: multiplexer 55: process 55A: Reformat execution means 55B: Positioning control means 56: Decoder 57: Non-volatile memory 62: Cylinder value head value holding means 71: Address mark detection unit (address mark detection means) 72: Address mark creation unit 73: Serial parallel Converters 74, 75: AND circuit 76: OR circuit 81: Sector counter 82: Comparison circuit (comparison means) 83: AND circuit 84: Register 85: Cylinder value head value Sector value holding means

Claims (4)

[Claims] 1. A magnetic disk device for performing format write release in a predetermined format according to an instruction from a host device, and an error holding unit for holding an error when an error occurs during execution of a padding operation, and the error A cylinder value head value holding means for holding a cylinder value and a head value which are physical positions that have occurred, and a seek operation to the cylinder value and the head value to execute a read operation to detect a no-sync / no-data. A magnetic disk device comprising reformatting means for reformatting a record, and then starting a padding operation. 2. A magnetic disk device for performing format write release in a predetermined format in response to a command from a host device, and an error holding means for holding an error when an error occurs during execution of a padding operation, and the error A cylinder value head value holding means for holding a cylinder value and a head value, which are physical positions that have occurred, and a no-sync / no-data detection when no-sync / no-data is detected by performing a read operation by seeking to the cylinder / head value. A magnetic disk drive comprising positioning control means for controlling to position at the end of a record before no data is detected, and then starting a padding operation. 3. An address mark detecting means for detecting whether or not there is an address mark before the index mark by performing an address mark search operation after detecting the no-sync no-data. Item 2. The magnetic disk device according to item 2. 4. A magnetic disk device for performing format write release in a predetermined format in response to a command from a higher-level device, and error holding means for holding an error when an error occurs during execution of a padding operation, and the error Cylinder value head value sector value holding means for holding the generated cylinder position, head value and sector value obtained by the output of the error holding means, and a sector search operation after seeking to the cylinder value, head value The magnetic disk device is characterized by further comprising: comparing means for comparing the output of the sector counter with the target sector which is the sector value, and when the sectors match, the padding operation is started.