JP2009223933A - Disk storage and method for managing sector group of disk storage - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a disk storage and a method for managing a sector group of the disk storage which can efficiently register defect sectors caused by a large defect area existing across a plurality of tracks and which can prevent a size of a defect management table from becoming too large. <P>SOLUTION: Physical block addresses of a start sector and an end sector of an area are stored, a physical LBA of the sector group in the area is specified from the stored physical LBA of the start sector and the end sector, after a logical LBA of the specified sector, a difference between the logical LBA and the physical LBA of the specified sector and a minimum logical LBA in the area are stored, whether or not a plurality of areas overlap in the same track is determined from these, and, when the areas overlap, this area is changed from the logical LBA to the physical LBA. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a disk storage device such as a hard disk drive and a sector group management method for the disk storage device.

  As is well known, a hard disk drive (hereinafter referred to as HDD) is a large-capacity and high-reliability information recording medium. In recent years, it has been widely used for recording computer data, video data, audio data, and the like. ing. In addition, HDDs have become smaller in size as they are mounted on portable electronic devices.

  A normal HDD system may have a sector that cannot be read / written normally due to a scratch or the like formed on the medium. Those sectors are managed as defective sectors by the HDD system. A defective sector discovered at the time of HDD manufacture is managed as defective sector information at the time of manufacture, and a logical block address (LBA) is assigned so as to avoid the defective sector.

  On the other hand, a defective sector generated during use by the user is managed as subsequent defective sector information, and data is replaced with a spare sector prepared in advance. When accessing the data of the subsequent defective sector, the address is converted by the defect management table and the alternative sector is accessed.

2. Description of the Related Art Conventionally, as information for managing defective sectors at the time of manufacturing, a technique for managing information that can specify a start / end position for each slip so as to realize slip processing of a logical block address has been proposed (Patent Document 1). ). However, this registration method has a problem that the size of the management table becomes enormous when the number of slips is large.
JP 2000-57698 A

  An object of the present invention is to provide a disk storage device and a sector group of the disk storage device capable of efficiently registering a defective sector caused by a huge defective area extending over a plurality of tracks and preventing the defect management table from becoming too large. To provide a management method.

  In order to achieve the above object, a disk storage device of the present invention has a medium on which tracks are concentrically formed, and has a rectangular area with one side in the track circumferential direction and the radial direction in the plane of the medium. A disk storage device having a sector group management unit that manages a defective sector group included therein, a storage unit that stores a physical block address of a start sector and an end sector of the area, and the start sector stored in the storage unit And specifying means for specifying a physical address of a sector group in the area from the physical block address of the end sector, a logical block address of the sector after specifying the area by the specifying means, and a logical block address of the sector after specifying And the physical block address difference and the smallest logical block address in the area are stored in the storage unit. In addition, from the logical block address of the specified sector stored in the storage means, the difference between the ethical block address and the physical block address of the sector after the specification, and the logical block address that is the smallest in the area, A determination unit configured to determine whether or not a plurality of areas overlap in the same track; and an address conversion unit configured to convert the logical block address to a physical block address when the determination unit overlaps a plurality of areas in the same track. It is characterized by that.

  Further, the sector group management method for a disk storage device of the present invention has a medium in which tracks are formed concentrically, and in a rectangular area with each side in the track circumferential direction and the radial direction in the plane of the medium. A sector group management method of a disk storage device having a sector group management unit for managing a defective sector group included therein, wherein physical block addresses of a start sector and an end sector of the area are stored in a storage unit, and the storage unit stores the physical block address by a specifying unit. The physical address of the sector group in the area is specified by the physical block address of the start sector and the end sector stored in the section, the logical block address of the sector after specifying the area by the specifying means, and after the specifying The difference between the logical block address and physical block address of the sector and the minimum logical block address in the area. In the storage unit, the logical block address of the sector after the identification stored in the storage unit by the determination unit, the difference between the ethical block address and the physical block address of the sector after the identification, and the area It is determined whether or not a plurality of areas are overlapped in the same track from the logical block address that is the smallest within the same track, and when a plurality of areas overlap in the same track by the determination means by the address conversion means, It is converted into a block address.

  In the disk storage device and the sector group management method of the disk storage device according to the present invention, defective sectors caused by a huge defective area extending over a plurality of tracks can be registered efficiently, and the size of the defect management table is prevented from becoming too large. It becomes possible to do.

Embodiments in which the present invention is applied to a magnetic disk apparatus will be described below with reference to the drawings.
<Registration method of rectangular area in media and slip calculation>
In this embodiment, as shown in FIG. 1, when there is a signal failure area extending over a plurality of tracks, a sector group of an area surrounded by a rectangle including the area is registered as a defect area. This makes it possible to collectively manage the sector group included in this area. Since tracks on actual media are arranged concentrically, they are actually shaped like a part of a ring, but in this embodiment, they are treated as rectangular areas for simplicity.

  In the conventional registration method, it is necessary to hold information for each of the four slips as shown in FIG. In the method of the present invention, it is not necessary to hold all slip information by calculating and obtaining information on each slip when necessary from the start position and end position of the registered area.

  Note that when the same area as in FIG. 2 is registered in a system having an address arrangement with skew as shown in FIG. 4, the addresses of the sector groups included in the registered area are different from those in FIG. When FIG. 4 is rearranged in the physical LBA order as shown in FIG. 5, the sector addresses included in the registration area are 10 to 12, 17 to 19, 22 to 27, 30 to 32, 37 to 39, and 42 to 47. It can be seen that the same track may be divided into two. The state of slipping in this region is shown in FIG. In this case, since six slips exist, the conventional registration method has to hold six slip information.

  In the present embodiment, the slip information can be correctly performed by obtaining the address information of the defective sector as shown in FIG. 6 from several pieces of information for defining the rectangular area as shown in FIG.

Hereinafter, a specific rectangular area registration method and slip information calculation method will be described.
Consider a case where areas A and B as shown in FIG. 7 are registered and managed.
For simplicity, it is assumed that the number of sectors included in each track (represented by Track_Size) is equal, and that each adjacent track has the same length of skew (represented by Skew).

  Normally, a track is designated by a cylinder number and a head number. However, it is assumed that a track is designated by a cylinder number with only one head. The addresses are assigned in order from the beginning of the track (circled in the figure) within the track, and the subsequent numbers are assigned from the beginning of the next track (address direction).

Information managed in the present embodiment is the management table of FIG.
This table includes “logical block address after region end (logical LBA)”, “LBA difference after region end”, “physical block address (physical LBA) of end sector of rectangular region”, “physical of first sector of rectangular region” It consists of five attributes: “block address (physical LBA)” and “minimum logical block address in area (minimum logical LBA)”.

These relationships will be described with reference to FIG.
In FIG. 11, the logical block address (logical LBA) after the end of the area is A_LBA, the LBA difference after the end of the area is A_DIFF (difference of two numbers), the physical block address (physical LBA) of the end sector of the rectangular area is A_END, The physical block address (physical LBA) of the first sector in the rectangular area is A_START, and the minimum logical block address (minimum logical LBA) in the area is A_S_LBA.

  A_S_LBA is defined to be equal to the value of A_LBA when all the areas (sectors of all physical block addresses (physical LBA) from A_START to A_END) are slips. A_START and A_END are physical addresses of sectors located at the upper left and lower right of the rectangle. A_LBA is the logical block address (logical LBA) of the sector next to the largest physical block address (physical LBA) sector in the area. A_DIFF is a difference between the logical block address (logical LBA) and the physical block address (physical LBA).

FIG. 8 shows a state of slipping with respect to regions A and B in FIG.
Each slip is indicated by an arrow. Assume that the cylinder number at the end of area A is A_Cyl and the cylinder number at the end of area B is B_Cyl. The partial area specified by the cylinder number A_Cyl of area A includes the head position of the track. Therefore, the slip in this area is divided into two areas, the beginning and end of the track. Similarly, the slip of the cylinder B_Cyl in the region B is also divided.

  In order to make this state easy to understand, FIG. 9 shows a state in which the beginning of the track is aligned with the left end. It is necessary to be able to identify slips scattered in each track from the information in the management table of FIG. The slip information calculation procedure will be described with reference to the flowchart of FIG.

FIG. 16 is a flow for calculating the slip included in the region A.
First, the cylinder number of the track to which the end sector of the area belongs and the head physical block address (physical LBA) of the track are obtained. The value of A_Cyl is an absolute value obtained by dividing A_END by the track size (Track_Size). The first physical address (physical LBA) of the track of the cylinder A_Cyl is 1st_Sec (A_Cyl) = (Track_Size) × A_Cyl.

  Next, a circumferential length A_Wide and a radial length A_Long of the region are obtained. Since the area start cylinder A_Start_Cyl is obtained by the absolute value of the value obtained by dividing the defect start physical block address (physical LBA) by the track size, A_Long is obtained by A_Cyl-A_Start_Cyl + 1.

  Next, in order to obtain A_Wide, first, the leftmost sector A_BTM in the cylinder A_Cyl is obtained. Using A_START as a reference, the value of the location moved by A_Long-1 tracks is obtained. The address of the adjacent sector in the next track of a certain sector is the sum of the difference between the track size and the skew size, but when the accumulated skew reaches one track, it returns to the original position (one round It is necessary to consider.

  Therefore, A_BTM is obtained from the value of A_START and the value of A_Long. A_Wide is the difference between A_END and A_BTM. Considering the case where the end of the sector is in the middle of the area, after adding the track size, A_Wide is obtained by performing a remainder calculation with the track size.

  Next, the slip start physical address: A_H (N) and the number of sectors A_H_Size (N) to be slipped are obtained for each cylinder (cylinder N). When the slip is divided into the beginning and end of the track, the start slip address is A_H (N), the end slip address is A_T (N), and the length is A_T_Size (N).

  When the cylinder N is the final cylinder A_Cyl, A_H (N) is obtained from the values of A_END and A_Wide. In other cases, A_H (N) is obtained from the value of A_H (N + 1) in consideration of the skew value. When the value of A_H (N) obtained here is smaller than the head sector address of the cylinder N track or larger than the end sector address, the track size is added / subtracted and adjusted. If the partial region in the cylinder N is not split, the slip length is A_Wide.

  On the other hand, when the partial area in the cylinder N is split at the head / end, the value of A_T (N) becomes the value of A_H (N) obtained here, and the length A_T_Size (N) of this slip is This is the number from A_T (N) to the end sector. A_H (N) is the head sector address of this track, and the length is A_Wide minus A_T_Size (N).

By performing the above calculation for all the cylinders in the area A, the addresses and lengths of all slips included in the area A can be calculated.
<Method of calculating physical block address (physical LBA) for access block address (access LBA) from host and obtaining all slip information of access target track>
In the example of FIG. 9, in the cylinders A_Cyl and A_Cyl-1, not only the slip in the region A but also the slip in the region B exists. If there is no mixture of tracks in a plurality of areas as described above, the defect management table includes the physical block address (physical LBA) of the terminal sector of the rectangular area and the minimum logical block address (in the area) of the information shown in FIG. It can be calculated without the minimum logical LBA). Here, a procedure for calculating the physical block address (physical LBA) for the access block address (access LBA) from the host and calculating all slip information of the track to be accessed is assumed using the information in FIG. explain.

  A method for obtaining a physical block address (physical LBA) for an access block address (access LBA) from the host will be described with reference to FIG. FIG. 10 shows an example of how each area registered in the management table is distributed. The right axis shows the direction of the logical block address (logical LBA). The left axis indicates the direction in which the difference value between the logical block address (logical LBA) and the physical block address (physical LBA) increases.

  Arrows distributed on the left side indicate slips in each region. Each area is distributed in an address space sandwiched between the minimum logical block address (minimum logical LBA) in the area and the logical block address (logical LBA) after the end of the area. There may be a case where the distribution range matches the region (all slip) as in the region A. The slips in each region may be alternately arranged as shown in the figure. In order to obtain the physical block address (physical LBA) for an arbitrary logical block address (logical LBA), for the slip immediately before the target logical block address (logical LBA), the logical / physical LBA difference after the slip If the value is known.

  For the last slip in each region, the logical / physical LBA difference after the slip is the value of “LBA difference after region end” in the table of FIG. Further, the logical / physical LBA difference of a certain slip can be calculated by subtracting the slip size from the logical / physical LBA difference of the slip immediately after that.

  For example, the logical / physical LBA difference value C1_DIFF of the slip C1 is obtained by B2_DIFF−B2_SIZE. Here, B2_DIFF represents the logical / physical LBA difference value of the slip B2, and B2_SIZE represents the slip length. Since the slip B2 is the final slip in the region B2, B2_DIFF = B_DIFF. Here, B_DIFF is the value of “LBA difference after region end” in region B. Therefore, C1_DIFF = B_DIFF−B2_SIZE. Thus, if there is all slip information, the difference value between the logical and physical LBA can be obtained for an arbitrary logical block address (logical LBA).

  However, since it is costly to perform the calculation of FIG. 16 for all the registered areas, the areas that need to be calculated are narrowed down. In the case of the access target LBA1 of FIG. 10, it is sufficient to know only the information of the slip A1, and the information of the areas B to E is not necessary for the physical address calculation.

  In the case of the access target LBA2, information on the slip D1 is necessary. In order to calculate this, in addition to the slip information on the area D, slip information on the areas B, C, and E where the address range overlaps with the area D is calculated. To do. Actually, information on the area E is not necessary, but slip information is calculated for all areas overlapping the area D for simplicity.

Similarly, in the case of the access target LBA3, information on the slip B1 is necessary, so slip information on the area B and the areas C and D overlapping therewith is calculated.
A method of performing this procedure using the table information of FIG. 14 is shown in FIG.
A flowchart of this process is shown in FIG.
First, as shown in the list 1 on the left side of FIG. 12, a list arranged in ascending order by the value of “logical block address after completion of the area (logical LBA)” is prepared. From here, an entry (area M) having an equal value immediately before the LBA to be accessed is searched. The area of this entry is a hit area 1.

  Next, entries larger than the access block address (access LBA) are rearranged in ascending order according to the value of “minimum logical block address (minimum logical LBA) in the area” (list 2). When the value of the top entry (area X) in list 2 is smaller than the access block address (access LBA), the access block address (access LBA) is included in area X.

  Further, in order to obtain an area where the address overlaps with the area X, an entry larger than the “logical block address after completion (logical LBA)” of the area X is extracted from the list 2 and set as the hit area 3. When the value of the first entry (area X) in the list 2 is larger than the access block address (accelerator LBA), it is the same as in the case of the access target LBA1 in FIG.

  The slip information in the area is calculated for all the hit areas 1 to 3 obtained above. When there are no hit areas 2 and 3, only the hit area 1 is obtained. When all necessary slip information has been obtained, the access block is calculated by calculating the logical / physical LBA difference values of the adjacent slips one after another using the end slip of a certain region as shown in FIG. The slip logical / physical LBA difference value immediately before the address (access LBA) can be obtained.

  When actually accessing an address, it is necessary to know all slip information of the target track. When the physical block address (physical LBA) to be accessed is obtained, it can be understood which track should be accessed. Therefore, all areas existing in the same track as the track to be accessed are extracted from the list 1 and the list 2 in FIG. If there is a set other than hit areas 1 to 3 in the set obtained by this extraction, slip information is calculated using that as hit area 4.

  The address information obtained by the slip calculation has a configuration as shown in FIG. 13 (upper stage). If only the slip information of the track to be accessed (cylinder number N) is extracted from the address information obtained by the calculation of the areas 1 to 4 and arranged in ascending order by sector number, the result is as shown in FIG. Since this is all slip information included in the access track, slip processing can be realized using this information.

<Second Embodiment: Refresh Target Sector Group Management Method by In-Plane Area Management Method>
The area management method of the present invention can be used not only for defect management but also for other purposes. In the present embodiment, the registration area is managed as a refresh target area.
The state at the time of writing to the track will be described with reference to FIG. When the data B is written to the track N + 2, a part of the data A signal already existing on the track N + 1 is erased by the data B signal. In this way, when the distance between the tracks is narrow, the influence on the adjacent track cannot be ignored. The greater the number of writes to the adjacent track, the greater the extent to which data is erased. In the worst case, data cannot be read.

  Assume that there is a region with poor signal quality as shown in FIG. In FIG. 1, the sector in this area is registered as a defective sector, but it is not necessary to make it a defective sector as long as the error rate is somewhat reduced. However, there is a possibility that the influence on the adjacent track is larger than in other areas. Therefore, by limiting the number of adjacent writes to a certain number or less, the risk that data cannot be read can be minimized and the quality can be improved.

FIG. 15 shows a table for managing the refresh target area.
As in FIG. 14, in order to calculate a physical address, it has attributes of a physical block address (physical LBA) at the beginning and end of a rectangular area. Furthermore, it has the “number of times of writing to area” attribute, manages the number of times of writing to sectors in the area, and controls so that the number of times of writing does not reach a certain threshold or more. Unlike the defect management, since the address is not slipped in the refresh target area, the information on the logical block address (physical LBA) and the difference information are unnecessary. This table is not used for calculation of a physical block address (physical LBA).

  FIG. 19 shows a flow of processing performed at the time of write access to the refresh target area. First, in the case of a system performing defect management, an access block address (access LBA) is converted into a physical block address (physical LBA) according to the flow of FIG. 17, and slip information in the track is obtained. Thereafter, all the refresh target areas including the same track as the track to be accessed are extracted. For these areas, the sector address included in the area is calculated using the table data of FIG. 15 and the same flow as in FIG.

  The address information obtained here is used to monitor whether there is a write to a sector in the area. When writing, the write count attribute value of the management table is incremented. As a result, when the number of writes reaches the specified number or more, the data is temporarily stored in a safe save area without writing the data to the original physical address position.

FIG. 20 shows a flow of processing for refreshing data in the area.
This processing is performed at idle time. If there is an area where the number of writes exceeds the specified number, the area is refreshed. If a part of the data has already been stored in the save area by the processing of FIG. 19, other data is copied to the save area. Then, the saved data is written again in the area. At this time, writing is performed sequentially (all sectors in a track with an area are written and then sequentially moved to the adjacent track) so that interference with adjacent tracks described with reference to FIG. 18 does not occur as much as possible. Then, the management table write count is reset to zero.

  As described above, in the disk storage device and the sector group management method of the disk storage device according to the present embodiment, it is possible to efficiently register defective sectors due to a huge defective area across a plurality of tracks, and the size of the defect management table. Can be prevented from becoming too large.

The figure for demonstrating the signal defect area | region and defect registration area | region on the medium which concerns on embodiment of this invention. The figure for demonstrating physical LBA which concerns on this embodiment. The figure for demonstrating logical LBA which concerns on this embodiment. The figure for demonstrating the address arrangement | positioning when the skew concerning this embodiment is attached. The figure for demonstrating a mode that the address arrangement | positioning with the skew which concerns on this embodiment was rearranged in order of physical LBA. The figure for demonstrating a mode that the address arrangement | positioning with the skew which concerns on this embodiment was rearranged in order of logical LBA. FIG. 5 is a view for explaining a method for specifying a rectangular area in a media plane according to the embodiment. FIG. 5 is a view for explaining a method for specifying a rectangular area in a media plane according to the embodiment. FIG. 5 is a view for explaining a method for specifying a rectangular area in a media plane according to the embodiment. The figure for demonstrating the method of calculating the physical LBA of access LBA and calculating the slip information of a track for access. The figure for demonstrating the registration information with respect to the area | region A which concerns on this embodiment. The figure for demonstrating the method of calculating the physical LBA of the access LBA which concerns on this embodiment, and calculating the slip information of an access object track. Slip address information expanded in memory. 6 is a defect area management table according to the present embodiment. The refresh area management table which concerns on this embodiment. The flow of the physical address calculation of the slip in the registration area which concerns on this embodiment. The flow which calculates the physical LBA of the access LBA which concerns on this embodiment, and calculates the slip information of an access object track. The figure explaining the write state to the adjacent track which concerns on this embodiment. 10 is a flow at the time of write access to a refresh target area according to the present embodiment. The flow of the refresh process which concerns on this embodiment.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 1 ... Signal defect area | region over 2 tracks, 2 ... Defect registration area, 3 ... Area head, 4 ... Area end, 5 ... Area head, 6 ... Area end, 7 ... Logical address slip, 8 ... Track start address, 9 ... sector, 10 ... skew, 11 ... track, 12 ... track size (Track_Size), 13 ... address direction, 14 ... area A, 15 ... area B, 16 ... slip

Claims (2)

A disk having a medium in which tracks are formed concentrically, and having a sector group management unit for managing defective sector groups included in a rectangular area having one side each in the track circumferential direction and the radial direction in the plane of the medium A storage device,
A storage unit for storing physical block addresses of a start sector and an end sector of the area;
A specifying unit for specifying a physical address of a sector group in the area by a physical block address of the start sector and the end sector stored in the storage unit;
The storage unit stores the logical block address of the sector after specifying the area by the specifying means, the difference between the logical block address and physical block address of the sector after specifying, and the minimum logical block address in the area. And
From the logical block address of the specified sector stored in the storage means, the difference between the ethical block address and physical block address of the specified sector, and the smallest logical block address in the area, the same track Determination means for determining whether or not a plurality of regions overlap,
A disk storage device comprising: an address conversion unit that converts the logical block address to a physical block address when a plurality of areas overlap in the same track by the determination unit. A disk having a medium in which tracks are formed concentrically, and having a sector group management unit for managing defective sector groups included in a rectangular area having one side each in the track circumferential direction and the radial direction in the plane of the medium A storage device sector group management method comprising:
Store the physical block addresses of the start sector and end sector of the area in the storage unit,
Specifying the physical address of the sector group in the area by the physical block address of the start sector and the end sector stored in the storage unit by the specifying unit;
The storage unit stores the logical block address of the sector after specifying the area by the specifying means, the difference between the logical block address and physical block address of the sector after specifying, and the minimum logical block address in the area. And
From the logical block address of the sector after specification stored in the storage unit by the determination unit, the difference between the ethical block address and physical block address of the sector after specification, and the minimum logical block address in the area, Determine whether multiple areas overlap in the same track,
A sector group management method for a disk storage device, comprising: converting a logical block address to a physical block address when a plurality of areas are overlapped in the same track by the determination unit by an address conversion unit.

Images