JPS583112A - Block assignment system in magnetic disc device - Google Patents

Abstract

PURPOSE:To keep an alternate link without interference of an upper-order device, even if one information is lost, by recording an alternate address to an ID section of a defective block and to an alternate map block. CONSTITUTION:An index signal and a sector signal are applied from a disc driver and given to a transfer control circuit 12d as the display of the head of the alternate map or the head of block. The circuit 12d controls a switching circuit 12e, an ID signal is given to an ID buffer memory (BM) 12b, a data signal is given to an upper device 11 via a data BM 12c or inversely to a disc driving device 13 via a serial-parallel conversion circuit 12h. A seek signal is transmitted and received via an interface control circuit 12f and a processor 12a. An ECC control circuit 12g forms a check code for read check at readout.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to block allocation method improvement 1 in a magnetic disk file device that writes and reads information in a fixed block format.

Conventionally, in magnetic disk file devices, there are two cases in which a block on the disk becomes unreadable due to, for example, a head crash. There is. In this case, in order to link the defective block and the replacement block, the address of the replacement block is written in the ID section of the defective block, and the address of the defective block is written in the ID section of the replacement block.

However, in a bad block, not only the data part but also the ID part itself may become unreadable. In that case, you can either search the ID part of the replacement block to find the replacement block, or search the ID part of the replacement block. A method is adopted in which the address of the alternate block is written by shifting the position of the D ID part.

However, in the method of searching the ID part of the replacement block, all replacement blocks on the replacement track are searched, so it is characterized by a large amount of time, especially when a read error is detected during the search. More time will be required for recovery. In addition, if a method is adopted in which the position of the ID part of the defective block is shifted,
Similarly, if the shifted position cannot be read, it cannot be linked.

In order to facilitate understanding of the present invention, the above-mentioned prior art will be explained in more detail.

. . . FIG. 1 is an explanatory diagram of the main parts of a magnetic disk device.

In the figure, DI r Dt + DH is a magnetic disk, HDl.

HD2 is a write/read head, B5 is a servo head, Tu is a user track, T- is a replacement track,
C indicates a head driving device.

FIG. 2 is a clear diagram showing the track format, and FIG. 3 is an explanatory diagram showing the format of one block in detail.

In Figure 2 (α), u-BO, w-1, u-B2-
・-w-B63 indicates a block in the user track, and in FIG. 2 (6), A-BO, A-B1. A
-B2...A-B65 indicates a block in the replacement track. What, 8p hasector pulse?

In Figure 3, Bit is an ID section, Bit is a data section,
Bidf is a flag, Bidm is a self ID, Birl
a Alternate ID (default 0'' when normal), Ride is checked.
Code and BcLtc respectively indicate check codes.

(3) Now, for example, block w-B1 in the user track
When the data section of the track becomes unreadable due to scratches or the like, an unused block such as A-B2 on the replacement track is allocated as a substitute for this block u-B1. At this time, set the flag indicating that the ID section of block w-B1 is defective, and set the cylinder number of w-B1 to the self ID.
Write head number, block number, and replacement ID
Write the cylinder number, head number, and block number of A-B'l in the block A-B'l, and set the use indication bit to the flag in the ID section of block A-B2, and write the self ID.
and the ID of the defective block w-B1.

By writing the ID part in this way, when a write/read request is made to block %-B1, the position is placed in the alternate block A-B2 depending on the contents of the ID part, and writing/reading to the data part is executed. However, for example, when the damage is caused by a head crash, not only the data section but also the ID section is often damaged, and in that case, the location of the replacement block cannot be known.

(4) Therefore, in such a case, in order to search for the alternately allocated light block of block u-B1 of the user track, the first block of the alternate track is positioned, and the ID part of each block of the track is sequentially read. I have to compare it. That is, in the case of the example described above, block A-
A replacement block is searched for by sequentially reading out the BOOID section, the ID section of blocks A-B1, and the ID section of blocks A-B2.

During this search operation for replacement blocks, for example, block A
- If the ID part of the EO cannot be read, it will take a considerable amount of time to retry several times (including rezero and normal seek rotation wait time).

If block A-BO cannot be read even after retrying, the search operation continues to read the ID section of the next block A-E1.

As is clear from the above, if the ID section in the block of the user track is unreadable,
A considerable amount of time must be spent searching for a replacement block.

In the alternate allocation of bad blocks in a magnetic disk device, the present invention not only writes a replacement address in the ID section of the bad block, but also writes the bad block in the area for the replacement map of the track where the bad block exists. This system relies on writing an ID section consisting of the same data as the data written in the ID section of the defective block, so that even if the ID section of the defective block cannot be read, the replacement link can be maintained. This will be explained in detail below.

FIG. 4 is an explanatory diagram showing a track format in one embodiment of the present invention.

In Fig. 4 (α), w'-BQ, u'-B1-.
u'-B65 is a block in the user track, w
'-EX indicates alternate map blocks, respectively.

In FIG. 4(b), A'-BO, A'-B1...
. . . A'-B63 indicates a block on the replacement track, and A'-EX indicates a replacement map block.

As you can see, this track format has 1
There are 4 blocks on the front of the track.

Then, alternate map blocks u'-EX, A'-E
X C Consists only of multiple ID sections. , rx represents an index signal indicating the presence of a replacement map block. It is necessary that this signal be clearly distinguished from the sector pulse.

Next, assuming that block u'-B1 is now defective, the process of allocating it to a replacement block and the process after the allocation will be described.

1) Position the head on the replacement track and read the ID section of the first block A'-EO.

If the ID section of block A'-BO is unreadable or the use indication bit is set in the flag, the ID section of the next block A'-B1 is read. In this way, we sequentially search for unused and normal replacement blocks,
If it is found, it is read into the ID buffer memory in the disk controller, and in this example, blocks A'-B2 are used as replacement blocks.

2) Position the head on the user track where the bad block u'-B1 exists, read the contents of the alternate map block u'-EX in this track (7), and find the ID of the bad block u'-B1. It is checked whether the ID section already exists and whether an unused ID section exists. If there is no unused ID section, the process is stopped and abnormal termination is reported to the CPU.

6) In the disk control device, data to be written to block w'-B1 and block s'-EX1 is prepared from the address of allocated replacement block A'-B2.

Also, a defective bit is set in the flag.

4) ID of block u'-E1 and block u'-EX
Write the data prepared in 6) above into the section.

However, the I of u'-B1 is already in the replacement map block.
If there is a D section, it is rewritten to the same ID section.

5) Position the head on the replacement track again and write to the ID section of blocks A'-B2.

1) to 5) above are processes for allocating alternate blocks.

When the higher-level device issues a read/write request to the data section of block u'-B1 of the user track alternately allocated (8) in this way, the following processing is performed.

6) Position the head at user block u'-B 1, read the ID section into the ID buffer memory, compare it with the ID section data prepared by the disk controller, and if they match, process the data section. but in this case,
Since there is a defective bit in the flag, there will be a mismatch and the processing of the data part will not be executed.

7) Use the ID of the replacement in the ID section to create the replacement block A.
'-Position the head at B2.

8) Execute the data section read/write process for one block requested by the host device for blocks A'-B2.

9) If the number of requested blocks from the host device is plural, read/write processing continues from the next user block u'-B2.

By the way, in the operation 6) above, I of block u'-B1
If part D cannot be read due to scratches, etc., the replacement map
Read block w'-BX to the ID buffer memory, search for data that is the same as the ID part of block w'-E1 in the disk controller using its own ID, and use the alternate ID to access the alternate block. Execute. Thereafter, the process continues to step 8) above.

It goes without saying that the method described above can be applied to a method in which a replacement block address is written by shifting the position of the normal ID section because there is a flaw in that position.

The replacement map block explained above is a normal block.
Unlike the format, since it is composed of only a plurality of ID parts as shown in FIG. 5, there is no need to take the usual one block length, and it can be realized with almost no decrease in space efficiency.

As for alternate map blocks, there are two methods: one in which there is a gap between each ID section as shown in FIG. 5(α), and the other in which there is no gap as shown in FIG. 5(b). As a unit of transfer between control devices, several I
There is a method in which part D is combined into one block. Also,
The number of ID sections in the alternate map block is determined by the quality of the disk medium and the space efficiency of the disk. Similarly, when the number of bad blocks in the same track increases beyond the number that can be stored in the alternate map block, the disk medium should be made unusable from the standpoint of system performance.

FIG. 6 is a block diagram of the main parts of an apparatus implementing the present invention.

In the figure, 11 is a host device, 12 is a disk control device, 13 is a disk drive device, and 12α is a processor. , 12
6U ID buffer memory, 12C is data buffer memory, 12d is transfer control circuit, 12g is switching circuit, 12f is interface control circuit, 12g is FCC (
12A is a serial/parallel conversion circuit, and 12i is an index (IX)/sector control circuit.

As is clear from the figure, the index signal and sector signal are supplied from the disk drive and sent to the transfer control circuit 12d as signals indicating the beginning of the alternation map or the beginning of the block (11).

The transfer control circuit 12d controls the switching circuit 126, and sends the ID signal to the ID buffer memory 12b and the data signal to the host device 11 via the data buffer memory 12C, and vice versa. 12
A is sent to the disk device 16 via A.

Seek signals and the like are sent and received via the interface control circuit 12f and processor 12α.

The ECC control circuit 12g creates a check code and performs a read check when reading.

As can be seen from the above explanation, according to the present invention, since the replacement address is recorded in the ID section of the defective block and the replacement map block, even if one of the pieces of information is lost, the replacement link can be sent to the higher-level device. can be maintained without intervention. Furthermore, since the alternating map block is composed of only the ID section, it is shorter than a normal block, and has the advantage that space efficiency does not deteriorate significantly.

(12)

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of the main parts of a magnetic disk device, FIGS. 2 and 3 are explanatory diagrams of a conventional track format, and FIGS. 4 and 5 are explanatory diagrams of a track format in the present invention. FIG. 6 is a block diagram of essential parts illustrating an apparatus for implementing the present invention. In the figure, 11 is a host device, 12 is a control device, 13 is a drive device, IX is an index, and x'-EX. A'-EX indicate alternate map blocks, respectively. Patent applicant Fujitsu Limited

Claims (1)

[Claims] In a magnetic disk control device, an alternate map block consisting of only a plurality of ID sections is allocated with an index signal at the beginning at least at the position of a user track on a magnetic disk, and a user's map block is allocated to the ID section of the alternate map block. A method of block allocation in a magnetic disk device characterized by writing data identical to the replacement address in the ID section of a defective block of a track for use in a magnetic disk drive.