JPH0770192B2 - Disk device control method - Google Patents

Description

The present invention relates to a head repositioning device for a disk device.

〔Overview〕

According to the present invention, in the method of relocating the header of the disk device, a read error may be caused by reading the identification information of the record starting from the sector before the sector to be relocated and comparing it with the identification information of the desired record. The number of read operations can be reduced.

[Prior Art] In the prior art, a disk device is equipped with a reorient counter that is reset when an address mark is processed and has a count value for one disk revolution, and counts with reference to the zero signal of this reorient counter. The header of the disk device is repositioned by reading the part.

By the way, in case of repositioning to the first record after processing multiple records, when data transfer from the central processing unit was not in time for disk rotation during data writing (data overrun), or in the central processing unit If a disk cache mechanism is added to the disk controller to compensate for the access time gap between the main memory and the disk unit, and after successful search operations on this cache memory, write operations are subsequently processed through store-through. The head cannot be positioned using the Orient Counter.

Conventionally, when the header is positioned in such a situation, the disk control device sequentially reads the record identification information (segment number of the counter section) included in each record, starting from the index sent from the disk device. , The read operation was carried out exhaustively until the record identification information of the target record was found.

[Problems to be solved by the invention]

However, in such conventional means, since the read operation is repeatedly performed until the record identification information of the target record is found, the read operation is repeated for about half of all the records included in one track on average. Become. As a result, as the number of read operations that can cause a read error increases, the number of error correction processes for correcting the read error also increases. Therefore, the overhead in the read operation of the entire system increases.

It is an object of the present invention to eliminate such drawbacks and to provide a disk drive control method capable of reducing the number of read operations that may cause a read error.

[Means for solving problems]

The present invention provides a method for controlling a disk device, which comprises a rotational position sensing means for positioning a header on a sector obtained by dividing a track on a disk medium by an equal amount of data, and for writing / reading a record including unique identification information. The identification information of the read record is stored in the storage means of the disk control device every time the record is read in order to specify the record to be written, and when a request for re-execution of the record write is made during the record write, the storage means is provided. The sector position is obtained based on the identification information of the record stored in, the head is positioned at this sector position to read the record, and the identification information of the read record is compared with the identification information stored in the storage means. When it matches, the head is repositioned successfully and the next record of the matched record is determined. And executes the writing mode.

[Action]

A track of the disk device is divided into sectors at equal intervals, and a sector number is given to each record on the track.
When the disk controller sets the sector position in the disk device, the rotational position sensing means of the disk device detects the target sector position and sends an interrupt signal to the disk controller.

Now, the central processing unit obtains the record identification information (segment number) and the sector position of the repositioned record to which the field immediately preceding the head field requesting reaccess belongs. This record identification information is stored in the control memory in the disk controller, the rotational position sensing means is activated by this sector position, and the record of the repositioned record is detected after the disk controller detects an interrupt signal from the disk device. The identification information (segment number) is read. The read record identification information and the record identification information stored in the control memory in the disk controller are compared, and if they match, the positioning is considered to be successful, and if they do not match, the positioning is considered to be unsuccessful and abnormal termination is performed.

〔Example〕

 Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention. The central processing unit 10 has a channel interface 15
The disk controller 11 is instructed to write / read out the disk device 12 via. The disk control device 11 adopts a micro program control method and includes a processor, a micro program storage unit, a control memory, and the like. The disk control device 11 of this embodiment does not have a disk cache mechanism. The disk device 12 is
It includes a disk medium, a rotational position sensing unit 13 and a write / read head, and receives commands from the disk controller 11 via a disk interface 16. When the write / read head reaches the sector position on the disk medium set by the disk controller 11, the rotational position sensor 13 sends an interrupt signal to the disk controller 11 via the disk interface 16.

Fig. 2 shows a CK that is commonly used in large disc devices.
D Indicates the recording method. The disk device 12 of this embodiment employs this recording method. In the CKD recording method, index 20 is used as the starting point and the home address (hereinafter HAF)
Say. ) 21, a count section (hereinafter referred to as CF) 23, a key section (hereinafter referred to as KF) 24, and a data section (hereinafter referred to as DF)
Say. There are a plurality of records 22 consisting of CF23, KF24 and DF25 that are consecutive 25) on the track. Record numbers 0 to N (N is an integer of 0 or more) are given to these records 22 in the direction opposite to the rotation of the disk medium from the side closer to the index 20. Record 22 may not have KF24, but CF23 and DF
25 is always present, and the minimum field length of these fields is determined for control purposes. The unit of this minimum field length is called a segment, and by assigning a number of 0 to the segment immediately after the index 20, the position of each field on the medium can be uniquely expressed by this segment number. Also, the sector position is the same as the segment number except that the minimum unit is different, and the sector position can be easily obtained from the segment number.

FIG. 3 is a format showing in detail the information contained in the CF 23 shown in FIG. Here, the skip control (hereinafter referred to as SC) 30 displays at least the defect information of the medium after CF23. Further, a segment number (hereinafter, referred to as SN) 31 indicates the position on the track where the CF 23 exists as a segment number, and can be handled as record identification information. Also, a physical address (hereinafter, referred to as PA) 32, a flag byte (hereinafter, F)
Say. ) 33, logical cylinder number (hereinafter referred to as CC)
34 and the logical head number (hereinafter referred to as HH) 35 display the status of the track in which this record exists, where
PA32 displays the physical position of the track, F33 displays the usage status of the track, and CC34 and HH35 display the logical position of the track. Also, the logical record number (hereinafter R
Say. ) 36, key length (hereinafter referred to as KL) 37 and data length (hereinafter referred to as DL) 38 display the status of this record 22, where R36 is the logical number of the record and KL37 is DL38 is the effective information amount of KF24 in this record.
Displays the amount of effective information of DF25.

FIG. 4 shows a part of the control storage in the disk controller 11, and is a gap indicator (hereinafter referred to as GI).
40, current segment number (hereinafter referred to as CSN) 4
1, including a transit segment number (hereinafter referred to as PSN) 41, a reorient segment number (hereinafter referred to as RSN) 43 and a command register (hereinafter referred to as CMD) 44.
The disk controller 11 uses the GI40, CSN41, PSN42, RSN4
Update 3 and CMD44 as follows: Disk controller
When 11 receives a command from the central processing unit 10, the command code is stored in CMD44. Each time HAF21, CF23, KF24 and DF25 are processed, information indicating which field is being processed is stored in GI40. After processing any CF23, this
The SN31 included in CF23 is stored in CSN41. After processing any DF25, the information in CNS41 is stored in PSN42.

Next, the repositioning of the head will be described with reference to FIG. Central processor 10 to disk controller 11 CF of R ₂ 22
23, KF24 and DF25 write commands are issued, and the data sent from the central processing unit 10 is delayed during the DF25 write process by the disk controller 12 and writing to the disk unit 12 cannot be performed normally. , If there is a positioning request to retry this write command (step 5
0), the disk controller 11 performs the following processing under microprogram control.

Based on the information in GI40 and CMD44, the following relocation procedure is requested. If CMD44 is a write command for CF23, KF24, DF25, the positioning sector value is obtained from the PSN42 value regardless of GI40 (step 51), and the PSN42 value is stored in RSN43 (step 52). If CMD44 is a DF25 write command, the positioning sector value is calculated from the PSN42 value if the GI40 value indicates CF23, or the CSN41 value if it indicates KF24 or DF25 (step 51), and then stored in RSN43. (Step 52).

Next, the positioning sector value obtained in step 51 is set in the rotational position sensing unit 13 to activate the rotational position sensing means (step
53), and waits for an interrupt from the rotational position detector 13 (step 54). CF23 after the head position at the time of interrupt detection is read (step 55), and SN31 and RS43 in this CF23 are compared (step 56). If the two match as a result of the comparison, it is determined that the positioning is successful (step 57), and the CMD44 is re-executed. If they do not match, it is regarded as a positioning failure and abnormally terminated (step 58), and the central processing unit 10 is notified that this write command could not be processed normally.

〔The invention's effect〕

As described above, according to the present invention, the repositioning of the header when the reorient counter existing in the disk device cannot be used is changed from the passing segment number stored in the control memory by the disk controller to the record immediately before the target record. Obtain the corresponding physical sector position information and the Orient segment number, and the segment number of the counter part actually read from the previous record and the Orient segment number matched based on this physical sector position information. At this time, it is determined that the head repositioning has succeeded, so that it is only necessary to actually read the minimum necessary fields included in the record, and the number of read operations that may cause a read error can be reduced and the number of error correction processes can be reduced. As a result, there is an effect that the overhead in the read operation of the entire system can be improved.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of the embodiment of the present invention. FIG. 2 is an explanatory diagram showing a track format of the disk device. Figure 3 is a detailed diagram of the information contained in the CF in Figure 2. FIG. 4 is a diagram showing a part of the control memory in the disk controller. FIG. 5 is a control flow chart showing the operation of the embodiment apparatus. 10 ... Central processing unit, 11 ... Disk control unit, 12 ...
Disk device, 13 ... Rotational position sensor, 15 ... Channel interface, 16 ... Disk interface, 20 ...
… Index, 21 …… Home address part (HAF), 22
…… Record (R _{0 to} R _n ), 23 …… Count section (CF), 24
...... Key part (KF), 25 …… Data part (DF), 30 …… Skip control (SC), 31 …… Segment number (S
N), 32 ... Physical address (PA), 33 ... Flag byte (F), 34 ... Logical cylinder number (CC), 35 ... Logical header number (HH), 36 ... Logical record number (R) , 37 ...
… Key length (KL), 38… Data length (DL), 40… Cap indicator (GI), 41… Current segment number (CSN), 42… Passing segment number (PSN), 43… Reorient segment number (RSN), 44 ... Command register (CMD).

Claims (1)

[Claims] 1. A method of controlling a disk device, comprising: a rotation position sensing means for locating a header on a sector obtained by dividing a track on a disk medium by an equal amount of data; and writing and reading a record containing unique identification information. , The identification information of the read record is stored in the storage means of the disk controller every time the record is read in order to specify the record to be written, and when the record write re-execution request is issued during the record write, the above-mentioned storage is performed. The sector position is obtained based on the identification information of the record stored in the means, the head is positioned at this sector position to read the record, and the identification information of the read record and the identification information stored in the storage means are stored. When comparing and matching, it is judged that the head repositioning is successful, and the next record of the matching record is determined. Control method for a disk device and executes the writing of code.