JP2008117491A - Recording device, recording method and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a recording device to record data in a disk scheduling without needing any special command from a host device side. <P>SOLUTION: A table T1 shows a nonvolatile memory and a physical page number of an HDD associated with a logical page number. A block bit is designed to record "1" when the logical page number is converted into a physical page number by a block unit, and "0" when the logical page number is converted into a physical page number by a page unit. The invention is applied to a recording device for recording data in the HDD. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a recording device, a recording method, and a program, and in particular, realizes so-called disk scheduling recording in which data supplied from a host device is evenly recorded on a disk medium such as a hard disk to guarantee a constant transfer rate. The present invention relates to a recording apparatus, a recording method, and a program that can be used.

  Consider that data is output from a host device such as a personal computer or a video camera, and the output data is recorded in a recording device including a disk medium such as a hard disk.

  Until now, PC (Personal Computer) as a host device has requested the recording device to increase the number of command processing per unit time and the data transfer rate (average value for a relatively long time), and therefore data transfer Variations in speed and response speed were not considered much.

  On the other hand, the performance required by a consumer device for a storage device is not so high on average. For example, when high-definition video recording is performed by MPEG (Moving Picture Experts Group) 2, the data rate is about 25 Mbit / sec, which is not so high in view of the transfer rate of the HDD exceeding 100 Mbit / sec.

  However, there is a strong demand for reducing variations in transfer speed, increasing response speed, that is, real-time characteristics. In fact, if the transfer speed varies greatly and the response speed is slow, problems such as stopping of images and sound during reproduction or inability to record during recording or recording will occur. These problems are fatal for using a storage device as a consumer device.

  There are roughly two factors that cause such transfer speed variations. The first factor that slows down the transfer rate is that there is a seek error or an error in the read data, and this takes time to correct or retry. Further, the second factor that slows down the transfer speed is that there are many seek operations and waiting for rotation and there is a lot of wasted time, or an area where the transfer speed is slow is used for a long time.

  The first factor described above can be solved for AV data or the like whose data reliability may be low by ignoring errors or setting an upper limit for the command processing time. For example, in ATA (AT Attachment) -7 Streaming Feature Set (generally referred to as AV command), the processing time of each command can be directly specified.

  However, there is a problem that how to execute it is completely up to the storage side, and it is not guaranteed whether the data can actually be read and written correctly. However, the frequency of occurrence of such problems is relatively small.

  The second factor described above has a higher occurrence frequency than the first factor and is more serious as a problem. Conventionally, two solutions have been proposed: disk scheduling and device physical arrangement.

  Disk scheduling is a technique for changing the execution order of a plurality of I / O requests requested by an application so that the processing time is reduced. A lot of research has been done for a long time, and representative algorithms include SSTF (shortest-seek-time-first least seek order), SCAN, FCFS (first-come-first-served arrival order), EDF ( earliest-deadline-first) (see Non-Patent Document 1).

  For example, in SSTF, several to a hundred commands are queued, and the values calculated for the execution time are stored as a table. The values of the stored table and the execution time of the most recently entered command are stored. In comparison, it is normal to execute the smallest one. It has been proposed that the rotation waiting time is taken into account when calculating the execution time. SSTF has been mainly implemented in controllers inside storage devices.

  In SCAN, commands are queued in the same way as SSTF, but commands are executed in one direction from the inner circumference to the outer circumference of the disk (or from the outer circumference to the inner circumference). The direction is reversed. As an improvement, after executing the command, if there is no access request ahead, the reverse is Look Scan, and the command is executed cyclically, specifically, the command is executed at the outermost periphery. The one that executes commands on the innermost circumference is called C (Circular) Look Scan. SCAN has been implemented mainly on the host device side.

  However, these methods are not suitable for handling AV data. This is because the amount of data transfer per I / O request is large, and therefore the number of I / O requests per unit time is small. In the above algorithm, if the number of I / O requests to be compared is small, the effect is weak. However, if waiting until a large number of I / O requests are obtained, the real-time property is impaired.

  Therefore, when handling AV data, FCFS and EDF are often used. However, when the disk media is used at a constant angular velocity (CAV), the data transfer speed is large on the inner and outer circumferences of the disk. Because of the difference, “a region with a low transfer rate is used for a long time”, and even if the response is fast, there is a possibility that sufficient throughput cannot be obtained.

  On the other hand, there are many research and implementation methods for devising the physical layout of files. For example, in recent file systems, efforts are made on the logical block address (LBA) space by using the locality of data and placing related data in physical proximity. In addition, a method has been proposed in which a seek operation is increased in order to guarantee the minimum transfer rate and a combination of a low transfer rate region and a fast transfer region is used (see Patent Document 1).

  However, the host device cannot directly know information such as where the head currently exists in the HDD and where the data is arranged on the disk. For this reason, it is difficult to accurately know the time for seeking the head to a predetermined track, especially the rotation waiting time thereafter, and there is a limit to shortening it.

  On the other hand, a technique for further optimizing the rotation waiting time by limiting the file arrangement has been proposed (see Patent Document 2), but the seek operation increases as the block is divided, and the disk surface There are problems such as poor usage efficiency.

  In order to solve these problems, the component that manages the arrangement information in the file system is moved to the storage device side, and the storage area of the disk is divided into a plurality of areas from the outer periphery to the inner periphery, and a predetermined pattern is obtained. A technique has been proposed in which the lower limit of the data transfer rate is guaranteed by using each area almost at the same frequency according to (see Patent Document 3).

Japanese Patent Laid-Open No. 10-49996 JP-A-9-185864 JP 2005-228404 A Maekawa Mamoru “Iwanami Lecture Software Science Operating System”, Iwanami Shoten, September 6, 1988

  However, in a conventional system consisting of a host device and a recording device, in order to record and read data on a disk by the proposed method, a simple write command or read command from the host device side to the recording device side is It was necessary to output a different special command, and the burden on the host device was heavy.

  The present invention has been made in view of such a situation, and does not require a special command from the host device side, so that data can be recorded in a disk scheduling on a disk medium in accordance with an existing command output from the host device. It is to make.

  A recording apparatus according to an aspect of the present invention is a recording apparatus that records data to be recorded on a disk medium based on data to be recorded supplied from a host device and a logical block address indicating a recording position. Converting means for converting the logical block address supplied from the host device into a logical page number; holding means for holding a first table indicating the physical page number of the disk medium associated with the logical page number; Selection means for selecting a physical page number of the disk medium associated with the logical page number after the change by the conversion means, and the host device at the position of the physical page number of the disk medium selected by the selection means Recording means for recording the data to be recorded supplied from the recording means, and the recording means Corresponding to the fact that the data to be recorded has been recorded, the first table is created by associating the physical page number on which the data to be recorded is recorded with the logical page number after the change by the converting means. Updating means for updating.

  The first table can indicate the correspondence between the logical page number and the physical page number of the disk medium in block units that are integer multiples of the page unit.

  The holding unit may also hold a second table indicating whether or not a physical block address belonging to the AV area of the disk medium is already associated with the logical page number.

  The holding means is a non-volatile memory and can store the data to be recorded, and the first table also includes a physical page number of the non-volatile memory associated with the logical page number. Can be recorded.

  The disk medium is provided with an AV data area for recording AV data having a relatively large data amount and a non-AV data area for recording non-AV data having a relatively small data amount. Can be.

  The selection means determines the physical page number of the disk medium to be associated with the logical page number after the change by the conversion means based on the data amount of the data to be recorded from the AV area or the non-AV area. Can be selected.

  When the data amount of the data to be recorded is a predetermined data unit, the selecting unit sets the physical page number of the disk medium associated with the logical page number after the change by the converting unit for each predetermined data amount. The plurality of areas provided in the AV area can be switched and selected in a predetermined order.

  A recording method according to one aspect of the present invention is a recording method of a recording apparatus that records data to be recorded on a disk medium based on data to be recorded supplied from a host device and a logical block address indicating a recording position. A conversion step of converting the logical block address supplied from the host device into a logical page number, and a selection for selecting a physical page number of the disk medium associated with the logical page number after the change by the processing of the conversion step A recording step for recording the data to be recorded supplied from the host device at the position of the physical page number of the disk medium selected in the processing of the selection step; and Corresponding to the recording of data to be recorded, A first table indicating the physical page number of the disk medium associated with the logical page number by associating the physical page number on which the data to be recorded is recorded with the logical page number after the change. Updating step.

  A program according to an aspect of the present invention is a program for controlling a recording apparatus that records data to be recorded on a disk medium based on data to be recorded supplied from a host apparatus and a logical block address indicating a recording position. A conversion step of converting the logical block address supplied from the host device into a logical page number; and a physical page number of the disk medium associated with the logical page number after the change by the processing of the conversion step. A selection step for selecting, a recording step for recording the data to be recorded supplied from the host device at the position of the physical page number of the disk medium selected in the processing of the selection step, and In response to the recording of the data to be recorded in the processing, The physical page number of the disk medium associated with the logical page number is indicated by associating the physical page number on which the data to be recorded is associated with the logical page number after the change by the processing of the group. Causing the computer to execute a process including an update step of updating the first table.

  In one aspect of the present invention, a logical block address supplied from a host device is converted into a logical page number, a physical page number of a disk medium associated with the changed logical page number is selected, and the physical of the selected disk medium is selected. Data to be recorded supplied from the host device is recorded at the position of the page number. Then, in response to the recording of the data to be recorded, the disk medium associated with the logical page number by associating the physical page number on which the data to be recorded is recorded with the logical page number after the change. The first table indicating the physical page number is updated.

  According to one aspect of the present invention, it is possible to record data on a disk medium according to an existing command output from the host apparatus without requiring a special command from the host apparatus side.

  Embodiments of the present invention will be described below. Correspondences between constituent elements of the present invention and the embodiments described in the specification or the drawings are exemplified as follows. This description is intended to confirm that the embodiments supporting the present invention are described in the specification or the drawings. Therefore, even if there is an embodiment which is described in the specification or the drawings but is not described here as an embodiment corresponding to the constituent elements of the present invention, that is not the case. It does not mean that the form does not correspond to the constituent requirements. Conversely, even if an embodiment is described here as corresponding to a configuration requirement, that means that the embodiment does not correspond to a configuration requirement other than the configuration requirement. It's not something to do.

  A recording apparatus (for example, the recording apparatus 20 in FIG. 1) according to one aspect of the present invention includes data to be recorded supplied from a host apparatus (for example, the host apparatus 10 in FIG. 1) and a logical block address indicating a recording position. In the recording apparatus for recording the data to be recorded on a disk medium (for example, the disk medium 39 in FIG. 1) based on the above, conversion means (converting the logical block address supplied from the host apparatus into a logical page number) For example, the CPU 21 in FIG. 1 that executes step S2 in FIG. 19) and a first table (for example, the table T1 in FIG. 1) that indicates the physical page number of the disk medium that is associated with the logical page number. The holding means for holding (for example, the non-volatile memory 29 in FIG. 1) and the logical page number after the change by the converting means are associated with the logical page number Selection means for selecting a physical page number of the disk medium (for example, the CPU 21 of FIG. 1 that executes step S16 of FIG. 20), and the host at the position of the physical page number of the disk medium selected by the selection means Recording means for recording the data to be recorded supplied from the apparatus (for example, the HDD control unit 25 in FIG. 1 that executes step S17 in FIG. 20) and the data to be recorded are recorded by the recording means And updating means for updating the first table by associating the physical page number on which the data to be recorded is recorded with the logical page number after the change by the converting means (for example, FIG. 20). 1 for executing step S19.

  The holding unit also holds a second table (for example, the table T2 in FIG. 1) indicating whether a physical block address belonging to the AV area of the disk medium is already associated with the logical page number. To be able to.

  A recording method and a program according to an aspect of the present invention include a conversion step (for example, step S2 in FIG. 19) for converting the logical block address supplied from the host device into a logical page number, and processing of the conversion step. A selection step (for example, step S16 in FIG. 20) for selecting the physical page number of the disk medium associated with the logical page number after the change, and the physical page number of the disk medium selected in the processing of the selection step Corresponding to the recording step (for example, step S17 in FIG. 20) for recording the data to be recorded supplied from the host device at the position and the data to be recorded recorded in the processing of the recording step. , The data to be recorded in the logical page number after the change in the process of the conversion step. An update step for updating the first table indicating the physical page number of the disk medium associated with the logical page number by associating the physical page number recorded with (for example, step S19 in FIG. 20) Including.

  Hereinafter, specific embodiments to which the present invention is applied will be described in detail with reference to the drawings.

  FIG. 1 shows a configuration example of a hybrid recording system to which the present invention is applied.

  The hybrid recording system 1 includes, for example, a host device 10 that outputs data such as a personal computer and a video camera, and a nonvolatile memory 29 including a NAND flash memory as a recording medium for recording data input from the host device 10 and The recording apparatus 20 includes a hard disk drive (hereinafter referred to as HDD) 30.

  The host device 10 executes a predetermined control program, thereby controlling a CPU (Central Processing Unit) 11 that controls the entire host device 10, and a CPU memory 13 that holds a control program and is used as a processing area of the CPU 11. And a recording device I / F (interface) 14 composed of IDE (Integrated Drive Electronics), SCSI (Small Computer System Interface), FC (Fibre Channel), USB (Universal Serial Bus), and the like. The CPU memory 13 and the recording device I / F 14 are connected to the CPU 11 via the bus 12.

  In the host device 10, when instructing the recording device 20 to record data, the CPU 11 performs a general write command, data to be recorded (for example, AV data, etc.), and a logical block address (hereinafter referred to as a recording block location). , LBA: Logical Block Address) is output to the recording device 20 via the recording device I / F 14.

  The recording device 20 controls the entire recording device 20 by executing a predetermined control program, holds the control program, and uses the CPU memory 23 used as a processing area of the CPU 21, the host device 10 and the recording device In order to keep constant the data transfer rate to the host device I / F 24 that communicates data to be transmitted, the HDD control unit 25 that controls the HDD 30, the memory control unit 26 that controls the buffer memory 27 to the nonvolatile memory 29, and the nonvolatile memory 29 Buffer memory 27, page buffer 28 for improving the efficiency of data transfer between the buffer memory 27 and the nonvolatile memory 29 (described later), and nonvolatile memory 29 in which data input from the host device 10 is recorded And the HDD 30 in which data input from the host device 10 is recorded

  The CPU memory 23 to the memory control unit 26 are connected to the CPU 21 via the bus 22.

  The buffer memory 27 is configured by DRAM (Dynamic Random Access Memory) or SRAM (Static RAM) in order to reduce the cost. The page buffer 28 is used not only to transfer data between the buffer memory 27 and the nonvolatile memory 29 but also to move data in the nonvolatile memory 29.

  Writing, reading, and erasing of data with respect to the nonvolatile memory 29 are executed by the memory control unit 26. The non-volatile memory 29 reads and writes data in units of, for example, 2 kilobytes, and erases data in units of blocks (for example, 128 kilobytes) that is an integral multiple of the page unit. When writing data to the nonvolatile memory 29, it is necessary to erase the data that has been recorded up to that point.

  The nonvolatile memory 29 stores data input from the host device 10 and holds tables T1, T2, and t2 (described later) described later.

  The HDD 30 is connected to the HDD control unit 25 to communicate data and the like, the control unit 32 that controls the entire HDD 30, a buffer memory 33 that temporarily holds data before recording on the disk medium 39, a recording A channel unit 34 that modulates data to be generated to generate an encoded signal and demodulates the encoded signal read from the disk medium 39; a head 35 that records and reads the encoded signal on the disk medium 39; and a spindle motor 37, a servo control unit 36 that controls the tracking motor 38, a spindle motor 37 that rotates the disk medium 39, and a tracking motor 38 that moves the head 35 to a predetermined position on the disk medium 39.

  Writing and reading of data with respect to the HDD 30 is executed by the HDD control unit 26. The HDD 30 reads and writes data in units of sectors, for example consisting of 512 bytes. When data is written to the HDD 30, it can be overwritten even if unnecessary data remains at the writing position.

  When writing data to the HDD 30, the storage area is divided into a plurality of areas from the outer peripheral side to the inner peripheral side in order to guarantee the lower limit of the transfer rate for the host device 10, and each area is divided according to a predetermined pattern. It is performed by disk scheduling recording that is used at an approximately equal frequency. Note that since the table T1 is used to assign the recording position to each area, the host device 10 does not need to output a special command.

  The number of areas is the number of zones in a technique called zone bid recording, which is adopted in general HDDs, and divides the storage area into a plurality of zones of about 10 to 20 to increase the capacity and changes the transfer speed for each zone. It is about 5 to 10 which is half of the above. If the data write size is too small, the overhead of the seek operation is conspicuous, so the write size should exceed the capacity for a plurality of tracks.

  Note that the size of the page, which is the recording unit of the nonvolatile memory 29, does not necessarily match the size of the sector, which is the recording unit of the HDD 30. However, since the page size (for example, 2 kilobytes) is an integral multiple of the sector size (for example, 512 bytes), the logical page number of the nonvolatile memory 29 and the logical block address space designated by the host device 10 are The logical page number of the HDD 30 can be easily associated.

  In the recording device 20, when data writing is instructed from the host device 10, the CPU 21 converts the LBA notified from the host device 10 acquired via the host device I / F 24 into a logical page number, and this logical page number. Is converted into a physical page number of the nonvolatile memory 29 or the HDD 30 with reference to the tables T1 and T2 held in the nonvolatile memory 29, and data is recorded in either the HDD control unit 25 or the memory control unit 26 Let me control.

  Hereinafter, it is assumed that the capacity of the nonvolatile memory 29 is 128 megabytes, and the capacity of the HDD 30 is 8 gigabytes or more. However, the capacities of the nonvolatile memory 29 and the HDD 30 are not limited to the values described above.

  Here, the garbage collection of the NAND flash memory using the tables t1 and t2, which is a prototype of the tables T1 and T2, will be described.

  Garbage collection is the nature of the NAND flash memory employed in the non-volatile memory 29, that is, “data is read and written in units of pages, and data is erased in units of blocks that are integer multiples of pages. This is a method for efficiently using the recording capacity of the NAND flash memory in response to the need to erase the data that has been recorded at the writing position.

  In the following description, it is assumed that the recording capacity of the NAND flash memory is 128 megabytes, the page of writing unit is 2 kilobytes, and the block of erasing unit is 128 kilobytes. In this case, the NAND flash memory is composed of 1024 blocks or 65536 pages.

  When representing the physical page number of the NAND flash memory, as shown in FIG. 2, 16 bits are used, and the block is represented by 10 bits (0x000 to 0x3FF) of the MSB (Most Significant Bit) side, LSB (Least Significant Bit) ) Side 6 bits (0x00 to 0x3F) represent the offset in the block. Hereinafter, not only the physical page number but also the logical page number uses a block number_offset notation such as 0x000_3F.

  Data is written to the NAND flash memory by write-once writing so that data can be written using each page as uniformly as possible, and each block can be erased as evenly as possible. For this write-once type writing, a table t1 shown in FIG. 3A, a table t2 shown in FIG. 3B, and a pointer shown in FIG. 3C are used.

  The table t1 shows the physical page number associated with the logical page number.

  In the table t2, a flag string indicating the state of each page in each physical block is recorded. This flag string is composed of decimal numbers having the same number of digits as the number of pages constituting each physical block, and the offset amount of the page in the block matches the number of digits from the upper side in the flag string. In this case, the flag string has 64 digits. For example, the flag of the physical page number 0x002_01 exists in the second digit of the flag string corresponding to the physical block number 0x002 of the table t2. When the flag is 0, 1, or 2, the page is not used (data is not written), valid (necessary data is written), or invalid (unnecessary data is written). Means.

  In the pointer, a physical page number to which data is next written is recorded.

  First, a process for newly writing data to the NAND flash memory will be described with reference to FIGS. First, the table t1 in FIG. 3A is referred to. According to this table t1, since the logical page number 0x001_00 has not been assigned, this is determined as the data recording position. Next, the pointer of FIG. 3C is referred. Since the pointer indicates the physical page number 0x002_01, the data is written to the page with the physical page number 0x002_01.

  Corresponding to the writing of this new data, as shown in FIG. 4, the physical page number 0x002_01 is recorded in association with the logical page number 0x001_00 of the table t1. Further, the flag corresponding to the physical page number 0x002_01 in the table t2, that is, the second digit from the top of the flag string corresponding to the physical block number 0x002 is updated from unused 0 to valid 1. Further, the pointer is incremented by one page and updated to the physical page number 0x002_02.

  Next, processing for rewriting data in the NAND flash memory will be described with reference to FIGS. When rewriting the data written in the logical page number 0x001_01 of the table t1 shown in FIG. 5A, first, the pointer in FIG. 5A is referred to. Since the pointer indicates the physical page number 0x003_02, the rewritten data is written to the page of the physical page number 0x003_02.

  Corresponding to the data writing after the rewriting, as shown in FIG. 6, the correspondence of the logical block numbers in the table t1 is changed from the physical page number 0x002_03 to the physical page number 0x003_02, and the physical page number in the table t2 The flag corresponding to 0x002_03, that is, the flag corresponding to the physical block number 0x002, the fourth digit from the top is updated from valid 1 to invalid 2, and the flag corresponding to the newly written physical page number 0x003_02, ie, physical The third digit from the top of the flag string corresponding to block number 0x003 is updated from unused 0 to valid 1. Further, the pointer is incremented by one page and updated to the physical page number 0x003_03.

  If such new data writing or data rewriting is continued, invalid pages gradually increase, and eventually there are no usable pages. For this purpose, valid data is moved to another block, all pages in one block are invalidated, and then erased to increase the usable area. This series of processing is called garbage collection.

  An example of garbage collection will be described with reference to FIGS. In order to simplify the description, it is assumed here that one block is composed of four pages. Therefore, the flag string in the table t2 is a 4-digit decimal number.

  FIG. 7 shows tables t1, t2 and pointers corresponding to the state in which data is written in the pages of logical page numbers 0x000_1 to 0x002_1. FIG. 8 shows the logical page numbers 0x000_0 and 0x000_0 from the state of FIG. Tables t1 and t2 and pointers corresponding to the state after the page data of 0x000_1 is rewritten and the page data of logical page numbers 0x001_2 and 0x002_3 are erased are shown. As can be seen from the table t2 shown in FIG. 8B, there are invalid pages in the physical block numbers 0x000 and 0x001, and these are the targets of garbage collection.

  Thereafter, garbage collection is performed in the order of physical block numbers 0x000 and 0x001.

  First, based on the table t2 shown in FIG. 9B, among the four pages that make up the physical block number 0x000, two effective pages are the physical page number 0x003_0 indicated by the pointer in FIG. 9C and the next physical page number 0x003_1. Copy to page. Corresponding to this copy, the tables t1, t2 and the pointer are updated as shown in FIG.

  Specifically, the logical page numbers 0x000_2 and 0x000_3 in the table t1 are updated to physical page numbers 0x003_0 and 0x003_1. In addition, the flags corresponding to the physical page numbers 0x000_2 and 0x000_3 in the table t2, that is, the third and fourth digits from the top of the flag column corresponding to the physical block number 0x000 are updated from valid 1 to invalid 2. In addition, the flags corresponding to the physical page numbers 0x003_0 and 0x003_1 in the table t2, that is, the first and second digits of the flag column corresponding to the physical block number 0x003 are updated from unused 0 to valid 1. Further, the pointer is incremented by two pages and updated to the physical page number 0x003_02.

  Thereafter, the data of the block having the physical block number 0x000 is erased, and as shown in FIG. 11B, the flag string of the physical block number 0x000 in the table t2 is updated to unused 0 corresponding to this erasure.

  Next, based on the table t2 shown in FIG. 11B, the valid two pages of the four pages constituting the physical block number 0x001 are the physical page number 0x003_2 indicated by the pointer in FIG. 11C and the next physical page number 0x003_3. Copy to 2 pages. Corresponding to this copy, the tables t1, t2 and the pointer are updated as shown in FIG.

  Specifically, the logical page numbers 0x001_0 and 0x001_1 in the table t1 are updated to the physical page numbers 0x003_2 and 0x003_2. In addition, the flags corresponding to the physical page numbers 0x001_0 and 0x001_1 in the table t2, that is, the first and second digits of the flag column corresponding to the physical block number 0x001 are updated from valid 1 to invalid 2. In addition, the flags corresponding to the physical page numbers 0x003_2 and 0x003_3 in the table t2, that is, the third and fourth digits from the top of the flag string corresponding to the physical block number 0x003 are updated from unused 0 to valid 1. Further, the pointer is incremented by two pages and updated to the physical page number 0x004_0.

  Thereafter, the data of the block having the physical block number 0x001 is erased, and the flag string of the physical block number 0x001 in the table t2 is updated to unused 0 as shown in FIG. 13B corresponding to this erasure.

  This is the end of the description of the NAND flash memory garbage collection using the tables t1 and t2.

  Next, an LBA specified when the host apparatus 10 instructs the recording apparatus 20 to write and read data and a FAT (File Allocation Tables) file system employed in the hybrid recording system 1 will be described.

  FIG. 14 shows an LBA space designated by the host device 10 for the recording device 20. In this LBA space, a sector (for example, 512 bytes) which is the minimum access unit of the HDD 30 is handled as one unit (a block in the LBA space). As described above, the sector of the HDD 30 and the page which is the recording unit of the nonvolatile memory 29 do not necessarily have the same size, but the page size is an integral multiple of the sector size. The pages of the memory 29 can be easily associated.

  The FAT file system employs a cluster composed of a plurality of sectors as the minimum unit for managing file data. Clusters are assigned cluster numbers in order from the smallest LBA. In the following description, it is assumed that the offset (where the cluster with cluster number 0 is placed in the LBA space) is ignored, and the cluster number is simply divided by the number of sectors (for example, 16) constituting one cluster (for example, 16). In the current example, one cluster is 8 kilobytes).

  The FAT stored in the LBA space is a table in which concatenation information indicating how the files recorded by being divided into a plurality of clusters are arranged is written. Therefore, the cluster number is used instead of the LBA in the FAT. The storage area managed by the FAT file system includes a reserved area, a FAT area, and a file and directory data area.

  For example, in the directory item Da stored in the file and directory data area, a file name (in this case, f1) and its start cluster number (in this case, 12340h) are described. When accessing the file f1, the host device 10 reads the start cluster number 12340h from the directory item Da, and refers to the FAT entry (FAT Entry) corresponding thereto, and the next cluster in which the file f1 is recorded is the cluster number. Get to be at 12341h. Subsequently, the next cluster number is obtained by sequentially referring to the FAT, and the arrangement information is read until the FAT item of the cluster number 1237Fh is EOF (End of File). The host device 10 accesses the file based on the read arrangement information.

  In the recording device 20, the LBA instructed from the host device 10 is converted into a logical page number, and this logical page number is referred to the tables T 1, T 2, t 2 held in the nonvolatile memory 29, and the nonvolatile memory 29. The data is converted into a physical page number or a physical block address of the HDD 30, and either the HDD control unit 25 or the memory control unit 26 controls data recording.

  Next, a correspondence relationship between the LBA, the logical page number, the physical page number of the nonvolatile memory 29, and the physical page number of the HDD 30 will be described with reference to FIGS.

  FIG. 15 shows the correspondence between the LBA space and the logical page number space. The LBA space is divided into three areas, a user area UA1 corresponding to the reserved area and the FAT area in FIG. 14, and user areas UA2 and UA3 corresponding to the file and directory data areas in FIG. A system area is provided at the head of the logical page number space, followed by user areas UA1, UA3, and UA2. However, the order of arrangement of the user areas UA1 to UA3 may be other orders, for example, the user areas UA1, UA2, UA3.

  Note that the correspondence between the LBA space and the logical page number space is managed by firmware executed by the CPU 21. Hereinafter, it is assumed that the user area UA1 is 32 megabytes and the user area UA2 is 8 gigabytes.

  FIG. 16 shows the correspondence between the logical page number space, the physical page number of the nonvolatile memory 29, and the physical block address of the HDD 30. The system area, user area UA 1, and user area UA 3 in the logical page number space are associated with physical page numbers in the nonvolatile memory 29. Furthermore, the user area UA1 and the user area UA3 are also associated with physical block addresses 0x00000000 to 0x0000FFFF of the HDD 30 for backup. A system area accessible only by the CPU 21 of the recording device 20 is provided at the top of the physical page number of the nonvolatile memory 29. In this system area, block access and random access in byte units are possible, and tables T1 and T2 described in detail below are held.

  The user area UA2 occupying the logical page numbers 0x00400_00 to 0x103FF_3F in the logical page number space is associated with the physical block addresses 0x00010000 to 0x0100FFFF of the HDD 30. In the user area UA2, logical page numbers 0x00400_00 to 0x004FF_3F are associated with the physical block addresses 0x00010000 to 0x0001FFF of the HDD 30, and are used for recording non-AV data that has a relatively small data amount and is not subject to disk scheduling recording. It is said to be non. The remaining logical page numbers 0x00500_00 to 0x103FF_3F are associated with the physical block addresses 0x00020000 to 0x0100FFFF of the HDD 30, and are used for recording AV data having a relatively large data amount and subject to disk scheduling recording.

  In the illustrated description, the recording area of the HDD 30 is described using a physical block address. However, when the HDD control unit 25 controls the HDD 30, an LBA is used.

  As is apparent from FIGS. 15 and 16, since the head part of the physical block address of the HDD 30 is used for backup of the user areas UA1 and UA3, the LBA space accessed by the host device 10 and the physical block address of the HDD 30 are The range is consistent.

  The correspondence described above is summarized as follows (1) to (4).

  (1) The physical block addresses 0x00000000 to 0x000FFFF of the HDD 30 are associated with the logical page number assigned to the user area UA1. The association between the logical page number and the physical block address is not changed.

  (2) The physical block addresses 0x00010000 to 0x0001FFFF of the HDD 30 are associated with non-assigned logical page numbers 0x00400_00 to 0x004FF_3F in the user area UA2. The association between the logical page number and the physical block address is not changed.

  (3) The physical block addresses 0x00020000 to 0x0100FFFF of the HDD 30 are dynamically associated with the logical page numbers 0x00500_00 to 0x103FF_3F assigned to the user area UA2. That is, for example, as shown in FIG. 16, logical page numbers 0x00600_00 to 0x00600_3F are dynamically associated with physical block addresses 0x00030000 to 0x000300FF belonging to the range of physical block addresses 0x00020000 to 0x0100FFFF. This association is performed in principle in units of blocks or super clusters (a super cluster is an integer multiple of a block). However, there is an exceptional case where association is performed in units of pages.

  (4) The physical block addresses 0x01010000 to 0x0000FFFF of the HDD 30 are associated with the logical page number assigned to the user area UA3. The association between the logical page number and the physical block address is not changed.

  This is the end of the description of the correspondence relationship between the LBA, the logical page number, and the physical page number of the nonvolatile memory 29 or the physical page number of the HDD 30.

  Next, the tables T1, T2, and t2 used in the recording apparatus 20 will be described.

  FIG. 17 shows an example of the table T1. This table T1 shows the physical page number associated with the logical page number, and is expanded by adding a block bit to the table t1 used in the above-described garbage collection.

  The added block bit records 1 when the logical page number is converted into a physical page number in block units, and 0 when the logical page number can be converted into a physical page number in page units. . For example, in the example of FIG. 17, the block bit corresponding to (except for a part) is set to 1. For the part where the block bit is set to 1, the description of the page unit is omitted in the table T1, and the size of the table T1 can be reduced.

  However, in the example of FIG. 17, the corresponding block bit is set to 0 so that a part of can be used for recording of non-AV data. For the part where the block bit is changed to 0, the page unit description cannot be omitted in the table T1.

  FIG. 18 shows an example of the table T2. In this table T2, an address conversion flag is recorded in association with each physical block address belonging to the AV data area of the HDD 30, which is dynamically associated with the logical page number. The address translation flag indicates whether or not the physical block address has already been associated with the logical page number. When the address translation flag is 0, it means no association, and when the address translation flag is 1. Means matched.

  The example of the table T2 illustrated in FIG. 18 corresponds to a state in which the physical block addresses 0x00030000 to 0x000300FF of the HDD 30 are dynamically associated with the logical page numbers 0x00600_00 to 0x00600_3F, as illustrated in FIG. . Therefore, in the table T2 of FIG. 18, the address conversion flag corresponding to the physical block address 0x00030000 is set to 1.

  The table t2 is the same as that described in the garbage collection described above.

Next, data recording processing using the tables T1, T2, and t2 will be described with reference to the flowcharts of FIGS. However, it is assumed that the disk scheduling recording area used first, the pattern for changing the disk scheduling recording area, and the data size used when writing AV data are set in advance.
For example, since the disk scheduling recording area of the user area UA2 is the physical block address 0x00020000 to 0x0100FFFF of the HDD 30, it is divided into four areas having almost the same capacity, area A1: 0x00020000 to 0x003FFFFF, area A2: 0x00400000 to 0x007FFFFF Area A3: 0x00800000 to 0x00BFFFFF and area A4: 0x00C00000 to 0x0100FFFF.

  In step S 1, when the host apparatus 10 outputs a general write command, data to be recorded, and LBA indicating a recording position, the host apparatus I / F 24 acquires these and notifies the CPU 21 via the bus 22. .

  In step S 2, the CPU 21 (more specifically, firmware executed by the CPU 21) converts the LBA notified from the host device I / F 24 into a logical page number, and the converted logical page number is stored in the nonvolatile memory 29. It is determined whether it belongs to the corresponding user area UA1 or UA31. If it is determined that the converted logical page number belongs to the user area UA1 or UA31, the process proceeds to step S3.

  In step S3, the CPU 21 refers to the tables T1 and t2 and associates the converted logical page number with the physical page number of the nonvolatile memory 29 in consideration of wear leveling and the like, as in the previous stage of the garbage collection description. . In step S <b> 4, the memory control unit 26 records the data to be recorded supplied from the host device 24 in the physical page number of the associated nonvolatile memory 29 in accordance with the control from the CPU 21.

  In step S <b> 5, the memory control unit 26 determines whether or not the data recording to the nonvolatile memory 29 is successful, and notifies the CPU 21 of the determination result. If the data recording to the nonvolatile memory 29 is successful, the process proceeds to step S6, and the CPU 21 records the recorded physical page number in association with the logical page number of the table T1. In the table t2, the flag corresponding to the physical page number is updated to valid 1.

  On the other hand, if the data recording to the nonvolatile memory 29 is not successful, the process proceeds to step S7, and the CPU 21 notifies the host device 10 of the occurrence of an error via the host device I / F 24, and this data. The recording process ends.

  If it is determined in step S2 that the converted logical page number does not belong to the user area UA1 or UA31, that is, the converted logical page number belongs to the user area UA2 corresponding to the HDD 30, the process proceeds to step S8. It is done.

  In step S8, the CPU 21 determines whether or not the converted logical page number is a non-AV area in the user area UA2. If it is determined that the converted logical page number is a non-AV area, the process proceeds to step S9. In step S9, the CPU 21 determines the physical block address of the HDD 30 that is fixedly associated with the converted logical page number.

  In step S 10, the HDD control unit 25 controls the HDD 30 according to the control from the CPU 21, and the data to be recorded supplied from the host device I / F 24 via the bus 22 to the determined physical block address of the HDD 30. To record.

  In step S <b> 11, the HDD control unit 25 determines whether or not the data recording to the HDD 30 is successful, and notifies the CPU 21 of the determination result. If the data recording to the HDD 30 is not successful, the process proceeds to step S12, and the CPU 21 notifies the host apparatus 10 of the occurrence of an error via the host apparatus I / F 24 and ends the recording process. On the other hand, if the data recording to the HDD 30 is successful, step S12 is skipped and the data recording process is terminated.

  If it is determined in step S8 that the converted logical page number is not a non-AV area, that is, an AV data area, the process proceeds to step S13. The processing after step S13 is disk scheduling recording.

  In step S13, the CPU 21 refers to the table T1, and determines whether or not the physical block number of the HDD 30 has already been assigned to the logical page number after conversion. If it is determined that a physical block number is assigned (that is, data is rewritten), the process proceeds to step S14.

  In step S14, the HDD control unit 25 controls the HDD 30 according to the control from the CPU 21 and records the physical block address of the assigned HDD 30 supplied from the host device I / F 24 via the bus 22. Overwrite the data. Thereafter, as in the processing from step S11 onward, if data recording is not successful, an error is notified to the host device 10, and if successful, this data recording processing is terminated.

  If it is determined in step S13 that the physical block number of the HDD 30 is not assigned to the converted logical page number as a result of referring to the table T1, the process proceeds to step S15 in FIG.

  In step S15, the CPU 21 determines whether or not the size of the data to be recorded is a super cluster unit. If it is determined that the size is the super cluster unit, the process proceeds to step S16. In step S16, the CPU 21 determines an area to be used next to the area used in the previous disk scheduling recording based on a predetermined pattern, refers to the table T2, and randomly selects an empty block in the determined area. Select

  In step S <b> 17, the HDD control unit 25 controls the HDD 30 according to the control from the CPU 21, and the data to be recorded supplied to the physical block address of the selected HDD 30 from the host device I / F 24 via the bus 22. To record.

  In step S <b> 18, the HDD control unit 25 determines whether or not the data recording to the HDD 30 is successful, and notifies the CPU 21 of the determination result. If the data recording to the HDD 30 is successful, the process proceeds to step S19, and the CPU 21 records a physical page number corresponding to the physical block address where the data is recorded in association with the logical page number of the table T1. In the table T2, the address conversion flag corresponding to the physical page number is updated to 1. Further, the area used in the current disk scheduling recording is recorded in the system area, and this data recording process is terminated.

  On the other hand, if the data recording to the HDD 30 is not successful, the process proceeds to step S20, and the CPU 21 notifies the host apparatus 10 of the occurrence of an error via the host apparatus I / F 24 and performs this data recording process. finish.

  If it is determined in step S15 that the size of the data to be recorded is not a super cluster unit, the process proceeds to step S21. In this case, the data to be recorded has a data unit that is less than the super cluster unit, and relatively small data is recorded in the AV area.

  In step S21, the CPU 21 refers to the table T1 and detects a logical page number belonging to the AV data area of the user area UA2 that has a block bit of 0 and no physical page number assigned thereto. In step 22, it is determined whether or not it has been detected. If not detected, the process proceeds to step S23. On the other hand, if it can be detected, the process of step S23 is skipped.

  In step S23, the CPU 21 refers to the table T2 to detect an empty physical block address in the user area UA2 (with an address translation flag of 0). In the table T1, the CPU 21 detects the detected physical block address. The block bit of the corresponding logical page number is updated to 0.

  In step S24, the HDD control unit 25 controls the HDD 30 according to the control from the CPU 21, and the physical block address corresponding to the logical page number detected in the process of step S21 or the physical block detected in the process of step S23. Data to be recorded supplied from the host apparatus I / F 24 via the bus 22 is recorded in the block address.

  In step S <b> 25, the HDD control unit 25 determines whether or not the data recording to the HDD 30 is successful, and notifies the CPU 21 of the determination result. If the data recording to the HDD 30 is successful, the process proceeds to step S26.

  In step S26, the CPU 21 records a physical page number corresponding to the physical block address in which data is recorded in association with the logical page number of the table T1. In the table T2, the address conversion flag corresponding to the physical page number is updated to 1.

  In step S27, the CPU 21 determines whether or not the entire data to be recorded has been recorded. If it is determined that the entire data to be recorded is not recorded, that is, it is determined that the data to be recorded is still unrecorded, the process returns to step S21 and the subsequent processes are repeated. If it is determined in step S27 that the entire data to be recorded has been recorded, the data recording process is terminated.

  If it is determined in step S25 that data recording to the HDD 30 has not been successful, the process proceeds to step S28, and the CPU 21 notifies the host device 10 of the occurrence of an error via the host device I / F 24. The data recording process ends.

  This is the end of the description of the data recording process using the tables T1, T2, and t2.

Next, an example of the tables T1 and T2 updated by the data recording process described above is shown.

  FIG. 21 shows the case where 1 megabyte of AV data is recorded in the order of areas A1, A2, A3, A4, A1, A2,... Table T1 is shown. FIG. 22 shows a corresponding table T2. However, the size of the super cluster is made equal to the block size.

  First, 128 kilobytes of data is written after physical block address 0x00020000 in area A1. Corresponding to this writing, the physical page number 0x00020000 is recorded corresponding to the logical page number 0x00500_00 in the table T1. Further, the address conversion flag corresponding to the physical block address 0x00020000 in the table T2 is updated to 1.

  Next, the head 35 of the HDD 30 is sought to the area A2, and 128 kilobytes of data are written after the physical block address 0x00400000. Corresponding to this writing, the physical page number 0x00040000 is recorded corresponding to the logical page number 0x00501_00 in the table T1. In addition, the address conversion flag corresponding to the physical block address 0x00040000 in the table T2 is updated to 1.

  Thereafter, data is written in the order of areas A3, A4, A1,..., And tables T1 and T2 are updated correspondingly. Above, description of the example of table T1, T2 updated by a data recording process is complete | finished.

  In the above description, the size handled as AV data is one type of super cluster unit. However, it is also possible to increase the number of AV data and to perform scheduling recording for each type.

  In the above description, it is assumed that the nonvolatile memory 29 is not used for writing to the user area UA2. However, if a table indicating the correspondence between the logical block address and the logical page number is provided separately, the nonvolatile memory 29 is provided. The system area can be used as a temporary writing area. This is an effective method when writing in time cannot be performed for reasons such as vibration. Further, if a function for moving the written data from the nonvolatile memory 29 to the HDD 30 is provided, the nonvolatile memory 29 can be used effectively.

  As described above, according to the hybrid recording system 1 to which the present invention is applied, the host apparatus 10 can perform disk scheduling recording in the recording apparatus 20 only by using a general write command. At this time, the condition imposed on the host device 10 is only to change the data write size between AV data and non-AV data and share it with the storage device 20 in advance. As a result, scheduling recording is performed by the recording device 20 regardless of whether the data is continuous in the LBA space or is randomly distributed. In any of these cases, the tables T1 and T2, which are information necessary for countermeasures against power interruption, are regularly recorded and updated by the recording apparatus 20 without the host apparatus 10 being involved. It becomes possible.

  In addition, since the non-AV data and the AV data area are divided on the storage space, and the non-AV data having a relatively small data amount can be recorded in units of pages, the use efficiency is reduced due to the fragmentation of the free area. And a decrease in writing speed can be suppressed.

  Furthermore, if the system area of the nonvolatile memory 29 is used as a buffer, it is possible to avoid a situation where the processing time limit cannot be temporarily satisfied.

  By the way, in this specification, the steps executed based on the program are executed in parallel or individually even if they are not necessarily processed in time series, as well as processes performed in time series according to the described order. It also includes processing.

  Further, in this specification, the system represents the entire apparatus constituted by a plurality of apparatuses.

  The embodiment of the present invention is not limited to the above-described embodiment, and various modifications can be made without departing from the gist of the present invention.

It is a block diagram which shows the structural example of the hybrid recording system to which this invention is applied. It is a figure which shows the physical page number of a non-volatile memory. It is a figure which shows table t1, t2 utilized for writing to a non-volatile memory, and a pointer. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure for demonstrating the garbage collection of a non-volatile memory. It is a figure which shows LBA space. It is a figure which shows the correspondence of LBA space and a logical page number space. It is a figure which shows the correspondence of a logical page number space and the physical page number space of a non-volatile memory or HDD. It is a figure which shows an example of table T1 which shows the correspondence of a logical page number and a physical page number. It is a figure which shows an example of the table T2 which shows the allocation condition of each physical block of the area | region for AV data of HDD. It is a flowchart explaining a data recording process. It is a flowchart explaining a data recording process. It is a figure which shows an example of table T1 which shows the correspondence of a logical page number and a physical page number. It is a figure which shows an example of the table T2 which shows the allocation condition of each physical block of the area | region for AV data of HDD.

Explanation of symbols

  1 hybrid recording system, 10 host device, 11 CPU, 20 recording device, 21 CPU, 24 host device I / F, 25 HDD control unit, 26 memory control unit, 29 nonvolatile memory, 30 HDD

Claims (9)

In a recording apparatus for recording data to be recorded on a disk medium based on data to be recorded supplied from a host device and a logical block address indicating a recording position,
Conversion means for converting the logical block address supplied from the host device into a logical page number;
Holding means for holding a first table indicating a physical page number of the disk medium associated with the logical page number;
Selecting means for selecting a physical page number of the disk medium associated with the logical page number after the change by the converting means;
Recording means for recording the data to be recorded supplied from the host device at the position of the physical page number of the disk medium selected by the selection means;
Corresponding to the fact that the data to be recorded is recorded by the recording means, the logical page number after the change by the converting means is associated with the physical page number where the data to be recorded is recorded. A recording device comprising: updating means for updating one table. The recording apparatus according to claim 1, wherein the first table indicates a correspondence relationship between the logical page number and the physical page number of the disk medium in a block unit that is an integral multiple of a page unit. The recording apparatus according to claim 1, wherein the holding unit also holds a second table indicating whether or not a physical block address belonging to the AV area of the disk medium is already associated with the logical page number. . The holding means is a nonvolatile memory and can store the data to be recorded,
The recording apparatus according to claim 1, wherein the first table also records a physical page number of the nonvolatile memory associated with the logical page number. The disk medium is provided with an AV data area for recording AV data having a relatively large amount of data and a non-AV data area for recording non-AV data having a relatively small amount of data. Item 2. The recording device according to Item 1. The selection means determines the physical page number of the disk medium to be associated with the logical page number after the change by the conversion means based on the data amount of the data to be recorded from the AV area or the non-AV area. The recording apparatus according to claim 5. When the data amount of the data to be recorded is a predetermined data unit, the selecting unit sets the physical page number of the disk medium associated with the logical page number after the change by the converting unit for each predetermined data amount. The recording apparatus according to claim 5, wherein a plurality of areas provided in the AV area are selected by switching in a predetermined order. In a recording method of a recording apparatus for recording data to be recorded on a disk medium based on data to be recorded supplied from a host device and a logical block address indicating a recording position,
A conversion step of converting the logical block address supplied from the host device into a logical page number;
A selection step of selecting a physical page number of the disk medium to be associated with the logical page number after the change by the process of the conversion step;
A recording step of recording the data to be recorded supplied from the host device at the position of the physical page number of the disk medium selected in the processing of the selection step;
Corresponding to the fact that the data to be recorded has been recorded in the process of the recording step, the physical page number on which the data to be recorded is associated with the logical page number after the change by the process of the converting step. And an update step of updating the first table indicating the physical page number of the disk medium associated with the logical page number. A program for controlling a recording device that records data to be recorded on a disk medium based on data to be recorded supplied from a host device and a logical block address indicating a recording position,
A conversion step of converting the logical block address supplied from the host device into a logical page number;
A selection step of selecting a physical page number of the disk medium to be associated with the logical page number after the change by the process of the conversion step;
A recording step of recording the data to be recorded supplied from the host device at the position of the physical page number of the disk medium selected in the processing of the selection step;
Corresponding to the fact that the data to be recorded has been recorded in the process of the recording step, the physical page number on which the data to be recorded is associated with the logical page number after the change by the process of the converting step. By this, a program for causing a computer to execute a process including an update step of updating a first table indicating a physical page number of the disk medium associated with the logical page number.

Images