JPH1083618A - Magnetic disk control device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To find a defective block in a short time when a faulty block is detected by repeatedly diagnosing a backward adjacent block or a forward adjacent block until the faulty block is detected. SOLUTION: Since blocks 2-7 in diagnostic order are not faulty, each patrol diagnosis is executed by moving a magnetic head every T time by K blocks. Then, when a block 8 in diagnostic order is judged to be faulty, the next diagnostic position is one block before the present diagnostic position, and when this block is further judged to be still faulty, the next diagnostic position is one block before the present diagnostic position (10 in diagnostic order in this case). Then, when no faulty block is decided here, the next diagnostic position is a block 11 in diagnostic order having its block number 7K+2, i.e., the first faulty block +1. When the faulty block is decided here, the next diagnostic position is a block of the present diagnostic position +1 = 12, and when this block is decided to be no faulty block, the next diagnostic position is a block of the present diagnostic position +K.

Description

DETAILED DESCRIPTION OF THE INVENTION

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a magnetic disk controller, and more particularly to a magnetic disk controller that periodically performs patrol diagnosis independently of a host device.

In recent computer systems, a huge amount of information is stored and held in a magnetic disk drive. This magnetic disk drive has an operating environment (for example, temperature, humidity, vibration,
(Shock, altitude, etc.), or if the frequency of access is low without long-term use, or if the total operation time is long, minute dust and water droplets on the disk may They adhere to and accumulate on the surface, causing head crashes and bad blocks.

[0002] Such a situation results in loss of important information stored in the magnetic disk drive, leading to a reduction in the overall operation rate and reliability of the computer system.

Therefore, there is a need for a magnetic disk controller capable of avoiding a head crash and a bad block in advance by performing regular patrol diagnosis independently of a host device, and quickly finding the bad block that has occurred. Have been.

[0004]

2. Description of the Related Art A conventional magnetic disk control device starts a magnetic disk device at a fixed interval time independently of an operating system and application programs of a host device, and keeps a magnetic head in the magnetic disk device constant. By moving the magnetic head over the disk surface at a distance interval of, the dust attached to the magnetic head and the disk surface is shaken off, thereby avoiding a head crash in advance.

Further, the magnetic disk control device reads the block at the position where the magnetic head has moved, and at the same time determines the quality of the block. This will be described more specifically.

FIG. 15 shows the operation principle of patrol diagnosis in a conventional magnetic disk control device.
0 indicates a magnetic head, 50 indicates a disk surface, and “block” in FIG. 1A indicates one block of the disk surface 50 to be inspected by one patrol diagnosis by “□”. The “block number” of 2) is a serial number assigned to the block (in FIG.
(K is a positive integer)), and the “diagnosis order” in FIG.
Indicates the order in which patrol diagnosis is performed on blocks, and “interval time, movement interval” in FIG. 4D indicates the time T until the next patrol diagnosis is performed and the distance interval K during movement (the unit is the number of blocks). Is represented.

The magnetic head 40 indicated by the solid line first determines the quality of the block of "diagnosis order" = 1, "block number" = 1, and then "interval time" = T (for example, 1 hour) "Movement interval" = magnetic head 40 moved by K (for example, 4) blocks later (shown by a broken line)
And move to the position of “diagnosis order” = 2 “block number”
= K + 1 is determined.

[0008] Similarly, K for each interval time T
The patrol diagnosis is sequentially performed by the number of blocks and the “diagnosis order” = 3 to 18 is executed.

The patrol diagnosis of the diagnosis number 19 is executed by returning to the block of the block number = 2. Further, the subsequent patrol diagnosis is executed by moving by K blocks every interval time T.

When the block number exceeds 18K, the process returns to the block of the youngest block number for which the patrol diagnosis has not been executed yet, and when the patrol diagnosis of all blocks is completed, the process returns to the block of block number 1 and repeats the same operation.

By the above operation, the magnetic head 40 moves evenly on the disk surface 50 to shake off dust adhering to the magnetic head 40 and the disk surface 50 and to judge the quality of all the blocks.

[0012]

FIG. 16 shows a procedure for finding a bad block by patrol diagnosis in a conventional disk controller. The operation of the patrol diagnosis is shown in FIG.
This is the same as the principle of No. 5.

However, in FIG. 16, for example, blocks (××) of consecutive block numbers 7K, 7K + 1, 7K + 2
The defective block group A is formed by the occurrence of a defect in the block (indicated by a symbol), and blocks with continuous block numbers 10K, 10K + 1, and 10K + 2 near the defective block group A (indicated by a symbol x). It is assumed that a defective block group B is formed due to the occurrence of a defect.

When the operating environment of the magnetic disk is bad,
It is possible that the bad block groups A and B having the same cause are formed close to each other.

First, consider the procedure and time when each bad block belonging to the bad block group A is found by the conventional patrol diagnosis.

The first block found in the bad block group A is "diagnosis order" = 8 and block number = 7K
+1 bad block. However, a defective block with a block number = 7K, 7K + 2 adjacent to this defective block is not found at this time.

"Diagnosis order" = 26 and "block number" =
When the defect of the block of 7K + 2 is found, the defect of the block of “block number” 7K is found (26−
8) After T time. For example, if the interval time T = 1 hour, it will be 18 hours later.

It is further later that a block defect of "block number" = 7K is found.

That is, it takes time to find all the continuous bad blocks belonging to the bad block group A.

Next, the procedure and time for finding a bad block belonging to the bad block group B after finding a bad block belonging to the bad block group A will be considered.

The block in which the first defect is found in the defective block group A is "diagnosis order" =
8 is a bad block of block number = 7K + 1.

The block in which the first defect is found in the defective block group B is "diagnosis order" = 11 and block number =
This is a 10K + 1 defective block.

Therefore, the time from when a bad block belonging to the bad block group A is found to when a bad block belonging to the bad block group B is found is (11-8) T time. For example, if T = 1 hour, it will be 3 hours later.

Further, for example, when the bad block group B occurs between the blocks of the "diagnosis order" 11 and 12, the bad block group B cannot be found by the 11th and 12th patrol diagnosis of the patrol diagnosis. The discovery of B will be delayed further.

That is, there is a problem that the detection of a defect in the bad block group B caused by the same cause as that of the bad block group A is delayed.

In order to solve this problem, the number of patrol diagnoses per unit time is increased so that the interval time T
Is shorter, the possibility of collision between an input / output request from an application program or the like of a higher-level device and a start request for patrol diagnosis increases.

As a result, the response time of the application program, which is the primary task, is delayed, which leads to a decrease in the performance of the entire computer system.

Therefore, the present invention provides a magnetic disk control device for patrol diagnosis of block quality by moving a magnetic head of a magnetic disk device at a constant distance interval at a constant interval time. It is an object of the present invention to find a bad block in a short time and find a group of a plurality of nearby bad blocks in a short time.

[0029]

[Means for Solving the Problems]

[1] In order to achieve the above object, the magnetic disk control device according to the present invention, when detecting a bad block, returns or advances the magnetic head to a block position adjacent to the back or front of the bad block, respectively. It is characterized in that the rear or front adjacent diagnosis for diagnosing the quality of a block is repeatedly performed until a non-defective block is detected.

That is, when a bad block is found first, a backward or forward adjacent diagnosis is executed, and it is possible to sequentially detect all the bad blocks that are continuously adjacent to the back or forward of the first bad block.

[2] In the present invention, in the above [1],
When a non-defective block is detected in the rear or front adjacent diagnosis, the next patrol diagnosis moves the magnetic head to a block position adjacent to the front or rear of the first defective block, and checks the quality of the block, respectively. The front or rear adjacent diagnosis to be diagnosed can be repeated until a non-defective block is detected.

That is, after all the continuous bad blocks in the rear or the front are found, the adjacent diagnosis is performed in the front or the rear, respectively, so that all the defective blocks consecutively adjacent to the front or the rear of the first defective block are obtained. Can also be sequentially detected.

[3] In the present invention, in the above [2], if a block that is not defective is detected in the front or rear adjacent diagnosis, the next patrol diagnosis sets the distance interval to the first distance interval. Then, the magnetic head is moved at a second distance interval shorter than the first distance interval to perform a tracing diagnosis for diagnosing the quality of the block, and the tracing diagnosis does not detect a defective block until a certain time has elapsed. In this case, the diagnosis can be performed by returning the moving distance of the magnetic head to the first distance interval.

In other words, by performing a tracking diagnosis for a certain period of time on the block behind the bad block group found in the front and rear adjacent diagnosis modes, the moving distance of the patrol diagnosis is shortened, so that the block behind the bad block group is executed. A bad block group can be detected.

[4] Further, according to the present invention, in the above [1] or [2], the patrol diagnosis can be performed at an interval time shorter than the interval time at the time of the rear or front adjacent diagnosis.

That is, by performing the continuous backward or forward adjacent diagnosis in a short time, it is possible to expedite the finding of the adjacent bad block.

[5] In the present invention, in the above [3], the patrol diagnosis can be performed at an interval time shorter than the interval time at the time of the tracking diagnosis.

That is, by performing continuous tracking diagnosis at short time intervals, it is possible to quickly find a bad block near a bad block.

[0039]

FIG. 1 shows the operation principle <1> of patrol diagnosis of a magnetic disk control device according to the present invention with respect to FIGS. 15 and 16 showing a conventional example. The difference is that the “interval time” and the “movement interval” in (4) can be changed by the “diagnosis mode”. This "diagnosis mode" has the following 4
Classified into types.

(I) The same normal diagnostic mode as in the prior art: A dashed arrow pointing forward. “Interval time” = T time, “movement interval” = K blocks. (ii) Rear adjacent diagnostic mode: solid arrow pointing backward.
“Interval time” = τ, “movement interval” = 1 block. (iii) Front adjacent diagnosis mode: solid arrow pointing forward.
“Interval time” = τ, “movement interval” = 1 block. (iv) Follow-up diagnostic mode: One-dot chain line arrow pointing forward.
“Interval time” = t, “movement interval” = k blocks.

The "interval time" includes T>t>
It is assumed that τ and “movement interval” have a relationship of K>k> 1. As an example, “interval time”: T = 1 hour, t =
10 minutes, τ = 1 second, “movement interval”: K = 4 blocks, k
= 2 blocks, etc.

FIG. 2 shows the state transition of the diagnostic mode in the operating principle <1> of the magnetic disk controller according to the present invention and the causes of the transition (whether or not there is a bad block, and whether or not the count-up of a tracking counter to be described later is or not). ).

FIG. 4 also shows a “patrol diagnosis position (block number)” for executing the next patrol diagnosis.
"Interval time" and "tracking counter" settings,
It shows the calculation of the “tracking counter” and the designation of the storage data of the “bad block position”.

Note that "long distance" and "short distance" in the figure correspond to "movement intervals" K and k in FIG. 1, respectively, and "long time", "short time" and "moment" are respectively shown in FIG. 1 corresponds to “interval time” T, t, and τ.

The operation of the patrol diagnosis of FIG. 1 will be described with reference to FIG. Note that the diagnosis in the initial state is a normal diagnosis mode.

First, "diagnosis order" in FIG. 1 (3) is "1".
In FIG. 2B, a “block” having a “block number” of “1” (hereinafter sometimes referred to as a block having a diagnostic order “1”) may be referred to as a patrol diagnosis (hereinafter, simply referred to as a diagnosis). ), And is a normal block shown by □, so it is determined that there is no bad block.

Here, referring to the state transition diagram of FIG. 2 based on (i) the normal diagnosis mode and “no bad block”, the next state is (i) the normal diagnosis mode, The next “patrol diagnosis position” is the (K + 1) th block obtained by adding the “long distance” K block to the current “patrol diagnosis position (block number“ 1 ”)”, and the “interval time” is “long”. The time is set to T time.

Therefore, the next diagnostic position, FIG.
In the block with the block number “K + 1” in (2), the “diagnosis order” in FIG. 3C is “2”, and the “diagnosis mode” in FIG. Is extended from the block of the diagnostic order "1" to the block of the diagnostic order "2" after K blocks, and the interval time T and the movement interval K (the number of blocks) are written below it.

Since the blocks in the diagnosis order "2" to "7" are not defective, (i) the normal diagnosis mode state does not change and each patrol diagnosis is executed by moving the magnetic head by K blocks every T time. .

The block in the diagnostic order "8" is determined to be "bad block" indicated by "x". Based on the determination result and the current state (i) the normal diagnostic mode, referring to FIG. (Ii) is in the rear adjacent diagnostic mode, the diagnostic position is one block before the current diagnostic position (patrol diagnostic position-1 block = "9"), and the interval time until the next diagnostic is instantaneous τ seconds. is there.

Further, a block number "7K + 1", which is a patrol diagnosis position where a defect is first detected, is stored in the "bad block position".

The block in the diagnostic order "9" is also determined to be "defective block", and referring to FIG. 2 using this determination result and (ii) the rear adjacent diagnostic mode as keys, the next state is (ii) the rear adjacent The diagnosis mode, the diagnosis position is one block before the current diagnosis position, the interval time is τ seconds, and the diagnosis order “10” is designated.

The block in the diagnostic order "10" is "no bad block" and the current state is (ii) the rear adjacent diagnostic mode, so the next state is (iii) the front adjacent diagnostic mode and "bad block position". "+1 block number" 7K +
2 "becomes a block in the diagnostic order" 11 ", and the interval time is τ seconds.

The block in the diagnosis order "11" is determined to be "defective block", and based on the result of this determination and the current state (iii) the front adjacent diagnosis mode, referring to FIG. 2, the diagnosis mode does not change. The next diagnosis position is the current diagnosis position + 1 block = “12”, and the interval time is τ seconds.

The block in the diagnostic order "12" is determined to be "no bad block". Based on the determination result and the current state (iii) the front adjacent diagnostic mode, referring to FIG. iv) The tracking diagnostic mode is set, and the next diagnostic position is shorter than the current moving distance interval to be the current diagnostic position + k blocks, and preferably the interval time is also t.
In minutes. Further, a tracking counter for counting the number of times of tracking diagnosis is cleared to "0".

Here, when the "tracking diagnosis" is repeatedly executed and a defective block cannot be found even after a certain period of time has elapsed, the routine shifts to the "normal diagnosis". Since the interval time is constant, the number of diagnoses is counted, and the determination is made based on whether or not the number of end times of the “tracking diagnosis” (for example, 12) has been exceeded.

In the normal block in the diagnosis order "13" which is diagnosed for the first time after the "tracking diagnosis" is started, the value of the tracking counter "0" is smaller than the tracking diagnosis end value "12" (FIG. 2). "YES"), the next diagnosis position is a block k blocks behind the current diagnosis position, and the interval time is t time. The tracking counter is incremented by one.

A normal block is also detected in the diagnosis order 14 to 16, the tracking diagnosis is executed, and the "tracking counter" is incremented by 1 for each diagnosis. 4 ".

In the bad block of the diagnostic order "17", (iv) tracking diagnostic mode and "bad block exist" are set, the next state is (ii) the rear adjacent diagnostic mode, and the next diagnostic position is the current diagnostic position. One block before, the interval time until the next diagnosis is τ seconds.

Further, the block number of the diagnosis order "17" which is the patrol diagnosis position where the failure is detected for the first time after the tracking diagnosis is stored in the "bad block position".

In the diagnostic order "17", the same change and setting of the diagnostic mode state as in the diagnostic order "8" are performed.

Similarly, the rear adjacent diagnosis is executed in the diagnosis order "18", the front adjacent diagnosis is executed in the diagnosis order "19" to "21", and the defective blocks are executed in the diagnosis order "19" and "20". Are found and the diagnostic order “22” ~
The tracking diagnosis is executed at “33”.

In this tracking diagnosis, the “tracking counter” is incremented by 1 for each diagnosis, and the “tracking counter” is “12” at the time of the diagnosis order 33.

In the block of the diagnostic order “33”,
(iv) Since the tracking diagnostic mode, “no bad block”, and “tracking counter” <end value (= 12) are not satisfied, the determination in the figure is “NO”, and the next diagnostic mode is (i) normal diagnosis. In the mode, the next diagnostic position is a block obtained by adding K blocks to the current diagnostic position, and the interval time is T time.

In the diagnostic order "33" and lower, the normal diagnostics are executed until a defective block is found in the same manner as in the diagnostic order "1" and lower.

The patrol diagnosis in FIG. 1 and FIG. 2 includes (i) a normal diagnosis mode (or (iv) a tracking diagnosis mode), (ii) a “rear adjacent diagnosis mode”, and (iii) a “forward adjacent diagnosis mode”. , (Iv) the diagnostic mode state was transited in the order of the tracking diagnostic mode and executed. However, (i) the normal diagnostic mode (or
(iv) Follow-up diagnosis mode), (iii) Forward adjacent diagnosis mode, (i
Even if the diagnostic mode state is shifted in the reverse direction in the order of i) the rear adjacent diagnostic mode and (iv) the tracking diagnostic mode, the defective blocks belonging to the defective block groups A and B can be detected in exactly the same manner. .

This will be briefly described with reference to FIGS.
Is the operation when the patrol diagnosis is executed in the order of (i) normal diagnosis mode (or (iv) tracking diagnosis mode), (iii) front adjacent diagnosis mode, (ii) rear adjacent diagnosis mode, and (iv) tracking diagnosis mode. The principle <2> and its state transition are shown.

In FIG. 3, after detecting a bad block in the diagnosis order 8, the front adjacent diagnosis is executed in the blocks in the order 9 and 10, and the rear adjacent diagnosis is executed in the blocks in the order 11 and 12, respectively. A bad block is detected in the diagnosis.

Similarly, after the detection of the bad block in the diagnosis order 17, the front adjacent diagnosis and the rear adjacent diagnosis are sequentially executed for the diagnoses in the orders 18 to 22, and the bad blocks are detected in the diagnoses in the orders 18 and 19. I have.

In FIG. 4, (i) the normal diagnosis mode and (i)
v) If the tracking diagnostic mode is "bad block",
The next state is (iii) the front adjacent diagnosis mode, and the next diagnosis position is specified by adding one block to the current diagnosis position and the front adjacent block.

In the case of (iii) "no defective block" in the front adjacent diagnostic mode, the next state is set to (ii) the rear adjacent diagnostic mode, and the next diagnostic position is a block at the position of bad block position minus 1 block. Is specified.

Further, in the case of (ii) "No bad block" in the rear adjacent diagnosis mode, the next state is set to (iv) the tracking diagnosis mode, and the next diagnosis position is added to the current diagnosis position by k blocks and forwardly added. The adjacent block is designated, and the “tracking counter” is cleared to “0”.

The other state transitions of the diagnostic mode, the diagnostic position, the interval time, the storage of the bad block position, and the like in the figure are the same as those in the case of FIG.

[0074]

FIG. 5 shows an embodiment of a magnetic disk control device according to the present invention in a functional configuration.
And the magnetic disk controller 2 are connected via a magnetic disk interface cable 3.

The magnetic disk device 1 includes a magnetic disk 4 indicating a medium (not shown) such as a magnetic head 40, a disk 50 (see FIGS. 1 and 3), and an access arm, and a magnetic disk control for driving and controlling the magnetic disk. Part 5,
It comprises a higher-level device interface controller 6 as an interface for transmitting and receiving data and control signals between the magnetic disk controller 5 and the magnetic disk controller 2.

The magnetic disk controller 2 has a higher-level device interface controller 6 of the magnetic disk device 1, a magnetic disk interface controller 7 as an interface for transmitting and receiving the above data and control signals, and a magnetic disk via the interface controller 7. An input / output instruction control unit 8 for controlling the disk device 1; an input / output interrupt receiving unit 9 for receiving input / output interrupts from an operating system or an application program; An interrupt control unit 10 starts a control circuit in the input / output instruction control unit 8 in response to an input / output request from a system or an application program.

Further, the magnetic disk control device 2 interrupts an interval time storage unit 11 for storing an interval time of the patrol diagnosis, and an interrupt signal requesting the patrol diagnosis when the interval time stored in the interval time storage unit 11 elapses. A patrol diagnosis start-up unit 12 to be sent to the control unit 10 and a control instruction for executing a patrol diagnosis in a predetermined diagnosis mode in response to a patrol diagnosis instruction from the interrupt control unit 10 receiving an interrupt signal of the patrol diagnosis request are input / output instructions. A bad block automatic tracking patrol diagnosis control unit 13 to be sent to the control unit 8, a patrol diagnosis position storage unit 14 for storing a patrol diagnosis position required for the control operation of the patrol diagnosis control unit 13, and a bad block for storing a bad block position The position storage unit 15 and an additional memory for counting the number of tracking diagnostics are provided. A counter storage unit 16, a start counter storage unit 17 for specifying a start block position to be returned when the diagnosis position specified by the patrol diagnosis position storage unit 14 exceeds the maximum number of blocks, and a diagnosis result of "no bad block / present". A diagnostic result storage unit 18 for storing, a diagnostic mode storage unit 19 for storing the current diagnostic mode state, and a state transition table storage unit 20 for storing a state transition table indicating a diagnostic mode state described later regarding presence or absence of a bad block. , It consists of.

FIGS. 6 and 7 are flow charts showing the operations of the interrupt control unit 10 and the bad block automatic tracking patrol diagnosis control unit 13, respectively.

FIG. 8 is a state transition table of the diagnostic mode set in the patrol diagnosis controller 13 shown in FIG. 7, and corresponds to the state transition table of FIG.

In FIG. 8, "normal diagnosis mode", "rear adjacent diagnosis mode", "forward diagnosis mode", and "tracking diagnosis mode" are assigned state numbers 0, 1, 2, and 3, respectively. “No bad block” and “having bad block” are assigned diagnosis result (or cause) numbers 0 and 1, respectively.

FIGS. 9 to 12 show processes corresponding to the normal diagnosis mode, the rear adjacent diagnosis mode, the front adjacent diagnosis mode, and the tracking diagnosis mode used during the operation of the patrol diagnosis controller 13 shown in FIG. It is a flowchart figure shown.

The flowcharts shown in FIGS. 9 to 12 correspond to (1) the processing in the case of "no bad block" and (2) in accordance with the presence / absence of the bad block which is the cause of the state transition.
The processing is divided into the processing in the case of “there is a bad block”.

The operation of the embodiment of FIG. 5 will now be described with reference to FIGS.
This will be described with reference to FIG. As an initial state, “0” which is a state number of “normal diagnosis mode” is set in the diagnosis mode storage unit 19, and an interval time T (for example, one hour) at the time of normal diagnosis is set in the interval time storage unit 11, and firstly, For example, “1” which is a block position of the magnetic disk to be patrol-diagnosed is set in the patrol diagnosis position storage unit 14 and the starting point counter storage unit 17.

The defective block position storage unit 15, the tracking counter storage unit 16, and the diagnosis result storage unit 18 are not initialized.

The patrol diagnosis starting unit 12 measures the time, and sends an interrupt signal for requesting the patrol diagnosis to the interrupt control unit 10 when one hour set in the interval time storage unit 11 has elapsed.

When an input / output instruction interrupt has occurred from the input / output interrupt receiving section 9 (“Yes” in step S 1 in FIG. 6), the interrupt control section 10 activates the control section corresponding to the input / output instruction interrupt. The input / output command controller 8 sends a command to the input / output command controller 8 to execute the predetermined input / output operation in response to the start command (step S2).

Then, after activating the input / output control unit 8, the interrupt control unit 10 returns to the operation of monitoring the occurrence of the input / output interrupt (step S1).

If no input / output instruction interrupt has occurred (“No” in step S1), the interrupt control unit 10
Indicates that no patrol diagnosis interrupt has occurred from the patrol diagnosis activation unit 11 (“N” in step S3).
o ") returns to the operation for confirming the occurrence of the input / output instruction interrupt (step S1).

If the patrol diagnosis interrupt has occurred (“Yes” in step S3), the interrupt control unit 1
0 is a bad block automatic tracking patrol diagnosis control unit 13
(Hereinafter, also referred to as the patrol diagnosis control unit 13) to start the patrol diagnosis control unit 13 (step S4).

After activating the patrol diagnosis control unit 13, the interrupt control unit 10 returns to the operation of monitoring the occurrence of an input / output interrupt (step S1).

The patrol diagnosis control unit 13 started in response to the start command from the interrupt control unit 10 (step S in FIG. 7)
4) Read the block position "1" for executing patrol diagnosis from the patrol diagnosis position storage unit 14, send an instruction to read data of block "1" of the disk to the input / output instruction control unit 8, and receive this instruction. The input / output command controller 8 executes an operation of receiving the data of the block "1" from the magnetic disk device 1 via the magnetic disk interface controller 7, and sends the received data to the patrol diagnosis controller 13 (step S5). ).

The patrol diagnosis controller 13 having received the data of the block "1" checks whether or not an error has occurred in this data (step S6). For example, if no error has occurred (step S6) “No” in S6)
The diagnosis result storage unit 18 is set to “0” (step S
7).

The patrol diagnosis control unit 13 having executed the above-described process executes a process of shifting to the next diagnosis mode with reference to the state transition table storage unit 20 storing the state transition table of FIG.

That is, the patrol diagnosis control unit 13
“0” indicating “normal diagnosis mode” is read from the diagnosis mode storage unit 19, “0” indicating “no bad block” is read from the diagnosis result storage unit 18, and “processing” is referred to with reference to the state transition table of FIG. 00 ”.

The contents of "Process 00" are shown in FIG. 9 (1). In order to execute the next diagnostic mode, the stored value "1" of the patrol diagnostic position storage unit 14 contains the number of moving blocks K
= “5” obtained by adding “4” is set as a new storage value of the patrol diagnosis position storage unit 14 ((1) in FIG.
0), the interval time storage unit 11 is set to one hour (step S21), the diagnostic mode storage unit 18 is set to “0”, which is the normal diagnostic mode (step S22), and the process ends (step S22). Step S23).

The patrol diagnosis control unit 13 that has completed the above “Process 00” compares the stored value “5” of the patrol diagnosis position storage unit 14 with the maximum block number “40” (step S11 in FIG. 7). Since the stored value is smaller than the maximum block number value ("Yes" in step S11), the process ends (step S14).

On the other hand, one hour after the time when the interval time storage unit 11 is set to one hour (step S21 in FIG. 9), the patrol diagnosis starting unit 12 requests a patrol diagnosis interrupt from the interrupt control unit 10, and receives this. The interrupt control unit 10 activates the patrol diagnosis control unit 13 (Step S4 in FIG. 6).

The patrol diagnosis control unit 13 (step S4 in FIG. 7) reads the block position “5” from the patrol diagnosis position storage unit 14 and receives the data of the block “5” from the magnetic disk device 1 (step S4). S
5).

The patrol diagnosis control unit 13 having received the data of the block "5" checks whether or not an error has occurred in this data (step S6). For example, if an error has occurred (step S6) “Yes” in S6)
The diagnosis result storage unit 18 is set to "1" (step S8),
Further, processing for notifying the host device of the disk recovery process and the error is executed (step S9).

Then, the patrol diagnosis control unit 13 reads “0” from the diagnosis mode storage unit 19, reads “1” indicating “there is a bad block” from the diagnosis result storage unit 18, and reads the state transition table of FIG. Refer to “Process 01”
Will be executed.

In “Process 01” of FIG. 9B, the patrol diagnosis control unit 13 sets the storage value “5” of the patrol diagnosis position storage unit 14 as the number of the first block determined to be defective to determine the bad block position. It is stored in the storage unit 15 (step S24 in FIG. 2B), and “4” obtained by subtracting 1 from the storage value “5” of the patrol diagnosis position storage unit 14 is set as a new storage value of the patrol diagnosis position storage unit 14 (see FIG. (Step S25), the interval time storage unit 11 is set to one second (Step S26), and the diagnostic mode storage unit 18 is set to “1” which is the rear adjacent diagnostic mode (Step S26).
27), and terminate the process (step S28).

The patrol diagnosis control unit 13 that has completed the “process 01” compares the storage value “4” of the patrol diagnosis position storage unit 14 with the maximum block number “40” (FIG. 7).
Since the stored value is smaller than the maximum block number value ("Yes" in step S11), the process ends (step S14).

On the other hand, one second after the time when the interval time storage unit 11 is set to one second (step S21 in FIG. 9), the patrol diagnosis starting unit 12 starts the patrol diagnosis control unit 13 via the interrupt control unit 10. (Step S4 in FIG. 6).

The patrol diagnosis control unit 13 reads the stored value (block position) “4” from the patrol diagnosis position storage unit 14 (step S 4 in FIG. 7) and receives the data of the block “4” from the magnetic disk device 1. Then, it is checked whether or not an error has occurred in the data (step S6). For example, when an error has occurred (step S7).
"Yes"), the diagnosis result storage unit 18 is set to "1" (step S8), and a recovery process or the like is executed (step S9).

Then, referring to the state transition table storage unit 20,
The value “1” of the diagnosis mode storage unit 19 and the diagnosis result storage unit 1
The patrol diagnosis control unit 13 selects and calls “process 11” based on the value “1” of 8 (step S1).
0).

In the "Process 11", "3" is obtained by subtracting "1" from the value "4" of the patrol diagnosis position storage unit 14.
(Step S34 in FIG. 10 (2)), the interval time storage unit 11 is set to one second (Step S35), and the diagnosis mode storage unit 19 is set to “1”, and the processing ends (Step S34).
36, 37).

Thereafter, the patrol diagnosis starting unit 12, the interrupt control unit 10, and the patrol diagnosis control unit 13 execute the same operations as described above, and the data of the block "3" of the magnetic disk is read (step S5 in FIG. 7). For example, it is determined that there is no bad block, and “0” is written in the diagnosis result storage unit 18 (step S8).

The patrol diagnosis control unit 13 refers to the state transition table storage unit 20 and converts the value “1” in the diagnosis mode storage unit 19 and the value “0” in the diagnosis result storage unit 18 into “processing 1”.
"0" (step S10 in FIG. 7).

In the "Process 10", the value of the patrol diagnosis position storage unit 14 is set to "6" which is obtained by adding "1" to the storage value "5" of the bad block position storage unit 15 (FIG. 10).
(1) Step S30), interval time storage unit 11
Is set to 1 second (step S31), and the diagnostic mode storage unit 1
9 as "2" (the same steps S32 and S3).
3).

Thereafter, the patrol diagnosis starting unit 12, the interrupt control unit 10, and the patrol diagnosis control unit 13 execute the same operation as described above. For example, the block "6" is determined to be "defective block" and the diagnosis result is stored. “1” is written into the unit 18 (step S8 in FIG. 7).

The patrol diagnosis control unit 13 refers to the state transition table storage unit 20 and selects “processing 21” from the value “2” in the diagnosis mode storage unit 19 and the value “1” in the diagnosis result storage unit 18. (Step S10 in FIG. 7).

In this "process 21", the value "6" of the patrol diagnosis position storage unit 14 becomes "7" obtained by adding "1" (step S45 in FIG. 11B), and the interval time storage unit 11 becomes 1 second. (Step S46), the diagnostic mode storage unit 19 becomes “2”.

Thereafter, the patrol diagnosis start-up unit 12, the interrupt control unit 10, and the patrol diagnosis control unit 13 execute the same operation as described above, and the block "7" is determined to be, for example, "no bad block", and the diagnosis result is stored. “0” is written in the unit 18 (step S7 in FIG. 7).

The patrol diagnosis control unit 13 refers to the state transition table storage unit 20 and stores the value “2” in the diagnosis mode storage unit 19.
Then, “processing 20” is selected from the value “0” in the diagnosis result storage unit 18 (step S10 in FIG. 7).

In this "process 20", the value "7" of the patrol diagnosis position storage unit 14 is added with "2" to become "9" (step S40 in FIG. 11 (1)), and the interval time storage unit 11 stores 10 minutes. (Step S41), the tracking counter storage unit 16 is cleared to "0" (step S4).
2), the diagnostic mode storage unit 19 becomes "3" which is the tracking diagnostic mode.

Thereafter, the patrol diagnosis start-up unit 12, the interrupt control unit 10, and the patrol diagnosis control unit 13 execute the same operation as described above, and the block "9" is determined to be, for example, "no bad block", and the diagnosis result is stored. “0” is written in the unit 18 (step S7 in FIG. 7).

The patrol diagnosis control unit 13 refers to the state transition table storage unit 20 and stores the value “3” in the diagnosis mode storage unit 19.
Then, “processing 30” is selected from the value “0” in the diagnosis result storage unit 18 (step S10 in FIG. 7).

In this “process 30”, the stored value “0” of the tracking counter storage unit 16 and the maximum number of times of tracking diagnosis “1”
2 (step S50 in FIG. 12 (1)). Since the stored value “0” is smaller than the maximum number of times of tracing diagnosis “12” (“Yes” in step S50), the value “ “11” is obtained by adding “2” to “9” (step S51), and the interval time storage unit 11 is set to “1”.
0 minutes (step S52), the tracking counter storage unit 1
6 is set to "1" which is obtained by incrementing the stored value "0" by "1" (step S53), and the diagnostic mode storage unit 1
9 is set to “3”.

Thereafter, the patrol diagnosis starting unit 12, the interrupt control unit 10, and the patrol diagnosis control unit 13 execute the same operation as described above, and the block "11" is determined to be, for example, "no defective block", and the diagnosis result is stored. "0" in part 18
Is written (step S8 in FIG. 7).

The patrol diagnosis control unit 13 refers to the state transition table storage unit 20 and stores the value “3” in the diagnosis mode storage unit 19.
"Process 30" is selected from the value "0" of the diagnosis result storage unit 18 (step S10 in FIG. 7).

Block 11 in the above "Process 30"
Blocks 13, 15, 17, 19, 21,
Assuming that 23, 25, 27, 29, 31, and 33 are not bad blocks, each block is diagnosed in the tracking diagnostic mode every 10 minutes. In the block 33, the tracking counter storage unit 16 stores the stored value for each diagnosis. Is incremented by “1” to “12”, the value stored in the diagnosis mode storage unit 19 is “3”, and the diagnosis result storage unit 18 is “0”.

The patrol diagnosis control unit 13 refers to the state transition table storage unit 20 and stores the value “3” in the diagnosis mode storage unit 19.
“Process 30” is selected from the value “0” in the diagnosis result storage unit 18 (Step S10 in FIG. 7).

In this "process 30", the value "12" of the tracking counter storage section 16 is compared with the maximum number of times of tracking diagnosis "12" (step S50 in FIG. 12 (1)). Since it is not smaller than “12” (“No” in step S50), the patrol diagnosis position storage unit 14
Is set to "37" by adding "4" to the value "33" (step S56), the interval time storage unit 11 is set to one hour (step S57), and the diagnostic mode storage unit 19 is set to "0" which is the normal diagnostic mode. ".

That is, the patrol diagnosis in block 33 returns to the normal diagnosis mode.

During the operation of the "process 30", for example, if the block 15 is "bad block", the diagnosis result storage section 18 becomes "1" indicating "bad block" (step S8 in FIG. 7). Since "3" is stored in the diagnostic mode storage unit 19, "process 31" is selected (step S10).

In this "process 31", the patrol diagnosis control unit 13 stores the storage value "15" of the patrol diagnosis position storage unit 14 in the bad block position storage unit 15 (step S60 in FIG. 12 (2)). “14” obtained by subtracting 1 from the storage value “15” of the diagnostic position storage unit 14
(Step S61), the interval time storage unit 11
Is set to 1 second (step S62), the diagnostic mode storage unit 18 is set to “1” which is the rear adjacent diagnostic mode (step S63), and the process is terminated (step S63).
64).

Further, for example, the patrol diagnosis control unit 13 which has completed the “process 30”, makes the patrol diagnosis position storage unit 1
If the storage value of No. 4 is not equal to or less than the maximum block number “40” (“No” in step S11 of FIG. 7), the storage value “1” of the starting point counter storage unit 17 is incremented by “1” to “2” ( In the same step S12), the stored value "2" of the starting point counter storage unit 17 is further stored in the patrol diagnosis position storage unit 14.
Is stored (the same step S13), and the block position where the patrol diagnosis is performed is set to an adjacent block after the first diagnosed block and for which the diagnosis has not been executed.

In the same manner, the diagnosis of the remaining blocks for which the patrol diagnosis has not been performed is sequentially performed.

[0129]

As described above, according to the magnetic disk control apparatus of the present invention, when a defective block is detected, the magnetic head is returned to a block position adjacent to the rear of or in front of the defective block. The backward or forward adjacent diagnosis for diagnosing the pass / fail is repeatedly performed in a short interval until a non-bad block is detected. Preferably, when a non-bad block is detected in the backward or forward adjacent diagnosis, the next patrol diagnosis is performed. The magnetic head is moved to a block position adjacent to the front or rear of the first defective block, and the front or rear adjacent diagnosis for diagnosing the quality of the block is repeatedly performed in a short interval time until a non-bad block is detected. Preferably, non-defective blocks are detected in the front or rear adjacent diagnosis In this case, the next patrol diagnosis is configured so that the magnetic head is moved at a short distance interval and a short interval time to repeatedly execute a tracking diagnosis for diagnosing the quality of the block. This makes it possible to find all the defective blocks in a short time, and to find a plurality of nearby bad blocks in a short time.

This will be described with reference to the patrol diagnosis according to the related art and the present invention by measuring the time of finding a bad block.

FIG. 13A shows a simple model of the disk surface 50 of the magnetic disk device 1 in the conventional patrol diagnosis in an array, and one square in the array corresponds to one block. The total number of blocks is 40, of which 6 shaded blocks are defective blocks.

As shown in FIG. 13B, in the conventional patrol diagnosis, (Specification 1) maximum block number: 40 blocks, (Specification 2) interval time: 1 hour, (Specification 3) magnetic head moving distance: It was done with the specifications of 4 blocks.

In FIG. 1A, the patrol diagnosis is started from a block indicated by an arrow as a starting point. If the maximum block number value is exceeded while the patrol diagnosis is being performed, the starting point is set to one block. It was done just back to the starting point shifted backwards.

The numbers of the diagnosis order of the patrol diagnosis performed as described above are shown in squares representing each block.

FIG. 3C shows the arrangement of the disk surface 50 of the magnetic disk controller 1 according to the present invention with respect to FIG. 1A showing a conventional example. Same as the example.

As shown in FIG. 4D, the patrol diagnosis of the present invention is as follows: (Specification 1) Maximum block number: 40 blocks, (Specification 2) Interval time: One hour in the normal diagnosis mode, adjacent rear and front (Specification 3) Magnetic head moving distance: 4 blocks in the normal diagnosis mode, 1 block in the rear and front adjacent diagnosis modes, 2 blocks in the tracking diagnosis mode, (Specification 4) The number of times of the follow-up diagnosis mode was 12 times (2 hours).

If the patrol diagnosis position exceeds the maximum number of block numbers, the starting point is moved in the same procedure as in the conventional patrol diagnosis.

The numbers of the diagnosis order of the patrol diagnosis performed as described above are shown in the same manner as in the case of the conventional patrol diagnosis of FIG.

FIG. 14 shows the time required for performing the patrol diagnosis according to the related art and the present invention under the above-described conditions until all six bad blocks are found.

In the conventional patrol diagnosis, the order in which adjacent bad blocks belonging to the same bad block group are found is the first (diagnosis order 4), the third (same 14), and the fifth (same 33). ) And the second (10) and the fourth (2)
0) and the sixth (39), it can be seen that the time until discovery is long by one hour or more.

On the other hand, in the patrol diagnosis of the present invention, the order in which adjacent bad blocks belonging to the same bad block group are found is the first (diagnosis order 4),
The third (5), the third (7) and the fourth (1)
9) Blocks (2nd and 3rd) belonging to the same bad block group when the first bad block (first) is found, the fifth (same 20) and the sixth (same 22) Is found within 3 seconds.

Further, a block (fourth block) belonging to another bad block group is found in about 1 hour, and the remaining blocks (fifth and sixth blocks) belonging to the same bad block group as this block are detected within 3 seconds. Have been discovered.

The time required for finding all six bad blocks is 39 hours in the conventional diagnosis, and it takes only 5 hours, 50 minutes and 7 seconds in the diagnosis of the present invention.

FIG. 16 is a graph summarizing the measurement results of FIG. 15. The time for finding the first bad block in the patrol diagnosis according to the prior art and the present invention is the same, but for the second and subsequent patrol diagnosis, the present invention is used. This indicates that the remaining bad blocks are found more efficiently and quickly than the conventional diagnosis.

[Brief description of the drawings]

FIG. 1 shows the operation principle of a magnetic disk control device according to the present invention.
It is a figure explaining <1>.

FIG. 2 shows the operation principle of the magnetic disk control device according to the present invention.
It is a state transition diagram of <1>.

FIG. 3 shows the operation principle of the magnetic disk control device according to the present invention.
It is a figure explaining <2>.

FIG. 4 is an operation principle of the magnetic disk control device according to the present invention;
It is a state transition diagram of <2>.

FIG. 5 is a block diagram showing a functional configuration in an embodiment of the magnetic disk control device according to the present invention.

FIG. 6 is a flowchart illustrating an operation of an interrupt control unit in the embodiment of the magnetic disk control device according to the present invention.

FIG. 7 is a flowchart illustrating an operation of a bad block automatic tracking patrol diagnosis control unit in the embodiment of the magnetic disk control device according to the present invention.

FIG. 8 is a diagram showing a state transition table in a diagnostic mode in the embodiment of the magnetic disk control device according to the present invention.

FIG. 9 is a processing flowchart in a normal diagnosis mode in the embodiment of the magnetic disk control device according to the present invention.

FIG. 10 is a processing flowchart in a rear adjacent diagnosis mode in the embodiment of the magnetic disk control device according to the present invention.

FIG. 11 is a processing flowchart in a front adjacent diagnosis mode in the embodiment of the magnetic disk control device according to the present invention.

FIG. 12 is a processing flowchart in a tracking diagnostic mode in the embodiment of the magnetic disk control device according to the present invention.

FIG. 13 is a diagram showing an example of an execution result of patrol diagnosis in the magnetic disk control devices according to the related art and the present invention.

FIG. 14 is a graph chart comparing examples of measurement time results of execution of patrol diagnosis in the magnetic disk control devices according to the related art and the present invention.

FIG. 15 is a principle diagram showing an operation of a patrol diagnosis in a conventional magnetic disk control device.

FIG. 16 is a principle diagram showing an operation of a patrol diagnosis in which a defective block exists in a conventional magnetic disk control device.

[Explanation of symbols]

 DESCRIPTION OF SYMBOLS 1 Magnetic disk device 2 Magnetic disk control device 3 Magnetic disk interface cable 4 Magnetic disk 5 Magnetic disk control unit 6 Host device interface control unit 7 Magnetic disk interface control unit 8 I / O command control unit 9 I / O interrupt reception unit 10 Interrupt control unit 11 Interval time storage unit 12 Patrol diagnosis activation unit 13 Bad block automatic tracking patrol diagnosis control unit 14 Patrol diagnosis position storage unit 15 Bad block position storage unit 16 Tracking counter storage unit 17 Starting counter storage unit 18 Diagnosis result storage unit 19 Diagnostic mode storage Unit 20 State transition table storage unit 40 Magnetic head 50 Disk surface In the drawings, the same reference numerals indicate the same or corresponding parts.

──────────────────────────────────────────────────の Continued on the front page (51) Int.Cl. ⁶ Identification number Agency reference number FI Technical display location G11B 20/18 501 G11B 20/18 501C 572 572B 572F

Claims (5)

[Claims] 1. A magnetic disk control device for patrol diagnosis of the quality of a block by moving a magnetic head of a magnetic disk device at a fixed distance interval at a fixed interval time, and when a bad block is detected, a position behind the bad block is detected. Alternatively, the magnetic disk control device is characterized in that the rear or front adjacent diagnosis for diagnosing the quality of the block by returning or advancing the magnetic head to the position of the block adjacent to the front is repeated until a non-defective block is detected. 2. The system according to claim 1, wherein, when a non-defective block is detected in the rear or front adjacent diagnosis, the next patrol diagnosis is performed in a position of a block adjacent to the front or rear of the first defective block. A magnetic disk control device wherein a front or rear adjacent diagnosis for diagnosing the quality of a block by moving a head is repeatedly performed until a non-defective block is detected. 3. The patrol diagnosis according to claim 2, wherein when a block that is not defective is detected in the front or rear adjacent diagnosis, the next patrol diagnosis is performed when the distance interval is set to the first distance interval. The magnetic head is moved at a second distance interval shorter than the interval to repeatedly perform a tracking diagnosis for diagnosing the quality of the block. If the tracking diagnosis fails to detect a bad block after a certain period of time, the magnetic head is moved. A magnetic disk control device wherein the diagnosis is performed by returning the moving distance to the first distance interval. 4. The magnetic disk control device according to claim 1, wherein a patrol diagnosis is performed in an interval time shorter than the interval time in the rear or front adjacent diagnosis. 5. The magnetic disk control device according to claim 3, wherein a patrol diagnosis is performed at an interval time shorter than the interval time at the time of the tracking diagnosis.