1364660 九、發明說明： 【發明所屬之技術領域】 本發明涉及一種靜態記憶體領域’特別涉及一種儲存 媒體管理方法。 【先前技術】 現有對儲存媒體的管理方法一般是建立物理位址與邏 〇輯位址的對照表，參照第一圖所示的邏輯塊與物理塊對照 關係示意圖，通過對照表可查找與邏輯塊對應的物理塊，'' 冨有物理塊成為壞塊時可將其剔除不用。然而隨著儲存媒 體的容量越來越大，對照表的資料量也越來越大，要在嵌 入式系統的記憶體中保存整個儲存媒體的對照表，幾乎是 不實際的。 現有一種分段管理技術，對儲存媒體的物理塊分段進 行官理。如第二圖所示的邏輯塊與物理塊對照關係示意 3圖，設儲存媒體有《»個物理塊，分為L·個區域，每個區域 N塊，分別對每個物理塊區域做物理位址與邏輯位址的對 照表。只有使用到相應邏輯位址區域的時候才在記憶體中 建立與該邏輯位址所在區域相應的對照表，這樣就在記憶 體空間和效率之間取得了一個折衷。但這也引起了另一個 問題，因為儲存空間被物理分割，各區域之間就被隔絕。 根據儲存媒體的實際使用特點，某些區域的物理塊經常被 使用，另外一些區域的物理塊卻很少使用，於是會導致使 用的不均衡，某些物理塊已經達到壽命的極限，而某些物 理位址，將寫命令要官 — 理位址中；將血寫上料禺到與交換塊對應的物 _ /、舄p 7扣疋迴輯位址對應的物理位址中不 f寫命令覆蓋的原有資料搬遷到與交換塊對應的物理位址 ’建立父換塊與寫命令指定邏輯位址對應的物理位址之 ::的對應關係，建立寫命令指定邏輯位址與交換塊對應的 位址之間的對應關係；擦除與交換塊對應的物理位 址0 土处根據至夕一區域邏輯位址與物理位址的對應關係 行訪門7 7的步驟還包括：更新至少一區域邏輯位址盥 物理位址的對應關係，更新對照表的步驟。 ’、 _込將Z、寫'卩令指定邏輯位址對應的物理位址中不被 寫命令覆蓋的原有資料搬遷到與交換塊對應的物理位址中 I括將與寫τ令指定邏輯位址對應的物理位址與交換塊 ’ΐ? 2物理位址比較，得到不被寫命令指定寫入新資料的 未才曰疋位址’將未指定位址中的資料複製到與交換塊對應 的物理位址中。 上述儲存媒體管理方法單獨使用一交換塊，或輪流使 用至少一交換塊。 本發明通過將儲存媒體的邏輯位址劃分多個區域，各 ί域分別組織對照表’實現佔用極少量的記憶體完成大容 里储存媒體的管理；本發明還利用交換塊實現各個區域之 ，的物理塊相互交換，從而消除了區域間磨損不均衡的問 題β同4避免了整個區域資料的搬_而導致的性能不穩定 問題，安全穩定，易於實現。 9 【實施方式】 本發明目的的麻招 丄 參照附圖做進一步：：功能特點及優點將結合實施例， 參照第三圖，+山 塊對照關係示专圖 '明第一實施例的邏輯塊與物理 Γ ΡίΛςίί . “㈤，本貧施例應用於快閃記憶體介質 對儲二媒體中作對照表，但僅在邏輯域 邏輯區域都可以對日在物理域不劃分區域，每個 邏輯位址的時候，=媒體的物理塊。當訪問到某 到記憶體中。、轉位址所在區域的對照表载入 現物置交換塊(swap)，通過共用交換塊實 見物理塊的父換，從而消除物理塊磨損不均衡的問題。 判八ir::’设儲存媒體共有m個物理塊，將邏輯位址 j刀為個區域，分別是區域〇、區域1……區域L，各 個區域N個邏輯塊，每個邏輯塊分別與一個物理塊建立： 照關係，例如區域〇的邏輯塊〇對應物理Μ，區 邏輯塊X對應物理塊y。再設置—邏輯交換塊， _； 物理塊Z。 嗯 ^_^_表1所示的邏對日g ,:1364660 IX. Description of the Invention: [Technical Field] The present invention relates to the field of static memory, and particularly relates to a storage medium management method. [Prior Art] The existing management method for the storage medium generally establishes a comparison table between the physical address and the logical address, and refers to the schematic diagram of the logical block and the physical block shown in the first figure, and can be searched and logical through the comparison table. The physical block corresponding to the block, ''when a physical block becomes a bad block, it can be eliminated. However, as the capacity of the storage medium increases, the amount of data in the look-up table is also increasing. It is almost impractical to store a comparison table of the entire storage medium in the memory of the embedded system. There is a segmentation management technique that implements the management of physical block segments of storage media. As shown in the second figure, the logical block and the physical block are in a relationship of 3, and the storage medium has "» physical blocks, which are divided into L areas, each area N blocks, and each physical block area is physically made. A comparison table of addresses and logical addresses. A comparison table corresponding to the area of the logical address is created in the memory only when the corresponding logical address area is used, thus achieving a compromise between memory space and efficiency. But this also raises another problem, because the storage space is physically divided and the areas are isolated. According to the actual use characteristics of the storage medium, physical blocks of certain areas are often used, and physical blocks of other areas are rarely used, which may result in uneven use, and some physical blocks have reached the limit of life, and some The physical address, the command will be written in the official address; the blood will be written to the object corresponding to the exchange block _ /, 舄 p 7 deducted from the physical address corresponding to the address, not f write command The original data that is overwritten is relocated to the physical address corresponding to the exchange block. The correspondence between the physical address corresponding to the logical address specified by the parent block and the write command is established. The write command specifies the logical address corresponding to the exchange block. Corresponding relationship between the addresses; erasing the physical address corresponding to the exchange block 0. The step of accessing the gate 7 7 according to the corresponding relationship between the logical address and the physical address of the area is further included: updating at least one area The correspondence between the logical address and the physical address, and the step of updating the comparison table. ', _ 込 Z Z, write '卩 order the physical address corresponding to the specified logical address is not covered by the write command to overwrite the original data to the physical address corresponding to the exchange block I will be written with the τ command specified logic The physical address corresponding to the address is compared with the physical address of the exchange block 'ΐ? 2, and the unaddressed address that is not written to the new data by the write command is obtained. 'Copy the data in the unspecified address to the exchange block. Corresponding physical address. The above storage medium management method uses a single exchange block alone or uses at least one exchange block in turn. The invention divides the logical address of the storage medium into a plurality of regions, and each ί domain organizes a comparison table to realize management of a storage medium in a large capacity by using a very small amount of memory; the present invention also implements each region by using a switching block, The physical blocks are exchanged, thereby eliminating the problem of uneven wear between regions. β and 4 avoid the problem of unstable performance caused by the movement of the entire region, which is safe, stable, and easy to implement. [Embodiment] The object of the present invention is further described with reference to the accompanying drawings: functional features and advantages will be combined with the embodiment, with reference to the third figure, and the comparison diagram of the + mountain block relationship shows the logic block of the first embodiment. Physics Γ ΛςίΛςίί . "(5), this poor example is applied to the flash memory medium as a comparison table in the storage media, but only in the logical domain logical region can be divided into regions in the physical domain, each logical address At the time, = physical block of the media. When accessing a certain memory, the comparison table of the area where the address is transferred is loaded into the swap block, and the parent of the physical block is seen through the shared exchange block, thereby Eliminate the problem of uneven wear of physical blocks. Judging eight ir:: 'The storage medium has a total of m physical blocks, and the logical address j is a region, which is region 区域, region 1 ... region L, and each region N Logical blocks, each logical block is respectively established with a physical block: according to the relationship, for example, the logical block of the area 〇 corresponds to the physical Μ, the area logical block X corresponds to the physical block y. Then, the logical switching block, _; physical block Z. Um, ^ _^_ The log pair day g shown in Table 1, :
邏輯位 區起 10 區域1 • · · 址 0 1 • · · N 0 «· · X • · · N • · · 土舍 物理位 址 1 6 • ·» 0 4 »·« y -—— • · · 2 • · · 〜 — Z 1364660 • < 表 1 接收到對儲存媒體的訪人 的處理過程袁昭第i η ρ 7後，本實施例對寫命令 ^…、第五圖包括如下步驟： 步驟S11 ，收到對邏輯 步⑽，計算邏輯塊入新資料的寫命令； 牛驄ςη # 屬於哪個邏輯區域； 乂驟813,將該邏輯區域 步驟S14，查對照表 ：至“己憶體； 到對應的物理塊7 ; ㈣康避輯塊X的邏輯位址查找 步驟S1.5，取預先定義的交換塊； 步驟S16，將寫命令要寫入 換塊對應的物理塊z中； <輯塊X的新貧料寫到交 步驟S17，將物理塊7中不被寫命 搬遷到物理塊z中； 是息的原有貝枓 步驟S18，將交換塊改為與物 X改為與物理塊z對應；、物理”對應，將邏輯塊 束。步驟S19，擦除交換塊對應的物理塊y，寫過程結 參照第四圖所示的經過上述寫過程後，邏輯塊與物理 亚對照關係變化示意圖，邏輯塊x中的原有資料和寫命人 要馬入的新資料都寫入物理塊z中’而將交換塊從物理^ =物理塊y。同理’對每一個寫命令，都僅载入一個 =域的對照表，使用交換塊寫入資料，既實現了分區域 理儲存媒體’減少佔用記憶體，.又實現了物理塊的使用 衡’還無需對整個區域資料進行搬遷。 1364660 上述步驟S12，計算邏輯塊X屬於哪個邏輯區域，具 體來說在對邏輯位址分區時，每個邏輯區域的大小都是預 先a又疋的’確定下來一般無需改變，如一個區域包括1024 個邏輯塊。當收到訪問命令後，本實施例可以通過邏輯位 址來推算要訪問的塊屬於哪個邏輯區域。一個區域包括 1024個邏輯塊為例’則可算出邏輯塊0到邏輯塊1023都屬 於區域0 ’而邏輯塊1〇24至邏輯塊2〇47都屬於區域1，依 此類推。Logic bit area from 10 area 1 • · · address 0 1 • · · N 0 «· · X • · · N • · · Earthwork physical address 1 6 • ·» 0 4 »·« y -—— • · 2 • · · ~ — Z 1364660 • < Table 1 After receiving the processing procedure for the interviewer of the storage medium, Yuan Zhaodi i η ρ 7 , the write command ^..., the fifth figure in this embodiment includes the following steps: Step S11, receiving a logic command (10), calculating a write command of the logic block into the new data; which logical area belongs to the 骢ς 骢ς ; 乂; 乂 813, the logic area step S14, checking the comparison table: to the “remembered body; Go to the corresponding physical block 7; (4) logical address search step S1.5 of the escrow block X, take a predefined exchange block; step S16, write the write command to the physical block z corresponding to the block change; < The new poor material of the block X is written to the intersection step S17, and the physical block 7 is not relocated to the physical block z; the original block of the interest is step S18, and the exchange block is changed to the object X. The physical block z corresponds to; physical" corresponds to the logical block bundle. Step S19, erasing the physical block y corresponding to the exchange block, and referring to the change of the relationship between the logical block and the physical sub-control relationship after the above-mentioned writing process shown in the fourth figure, the original data and the write life in the logical block x The new data that the person wants to enter is written into the physical block z' and the exchange block is from the physical ^ = physical block y. Similarly, for each write command, only one = domain comparison table is loaded, and the data is written using the exchange block, which realizes the sub-regional storage medium 'reducing the occupied memory, and realizing the use of the physical block. 'There is no need to relocate the entire area. 1364660 In the above step S12, it is calculated which logical area the logical block X belongs to, specifically, when the logical address is partitioned, the size of each logical area is pre-a and 疋 'determined generally does not need to be changed, such as an area including 1024 Logical blocks. After receiving the access command, the embodiment can estimate which logical area the block to be accessed belongs to by the logical address. An area including 1024 logical blocks is taken as an example, and it can be calculated that logical block 0 to logical block 1023 belong to area 0 ' and logical block 1 〇 24 to logical block 2 〇 47 belong to area 1, and so on.
上述步驟S13，將該邏輯區域的對照表載入到記憶 體，其中載入是指把對照表讀取到記憶體中，具體載入方 式見下文說明。 上述步驟S16 ，將寫命令要寫入邏輯塊X的新資料寫 到交換塊對應的物理塊z中是採用現有的寫操作技術完成 的，完成一個寫操作可以分為兩個步驟，一個是設備從主In the above step S13, the comparison table of the logical area is loaded into the memory, wherein loading refers to reading the comparison table into the memory, and the specific loading mode is described below. In the above step S16, the new data written to the logical block X by the write command is written into the physical block z corresponding to the exchange block, which is completed by using the existing write operation technology. The completion of a write operation can be divided into two steps, one is a device. From the Lord
機接收資料，存到設備的緩衝區’然後從緩衝區把資料寫 入flash。本貫施例從寫命令中得知要對邏輯塊X寫入資 料’而邏輯塊X對應物理塊y，物理塊7本身是存有原有 資料的，根據快閃記憶體的特性，物理塊 料的’必須把物理塊⑼除才可以再寫入資料。但= 除物理塊y，原有㈣就會錢。因此不能直接把緩衝區 的貧料Μ物理塊y ’而只能寫到交換塊對應的物理塊z 完成將緩衝區資料寫入交換塊的寫操作後，物 和交換塊對應的物理塊〇各保存了邏輯Μ的―部分資 ι^〇4〇ου 料在以後的麵作令是無法使用的，因此還需要把物理塊 ^的原有資料搬遷到物理塊ζ巾。至此，邏輯塊X的所 ::料都保存在物理塊ζ中了’物理塊y中雖然存有原有 貝已^:無用了’就可以把物理塊y擦除。當然也可 :在物理塊y擦除以後，把物理塊Z中的所有資料重新拷 、回物理塊y ’但這樣明顯效率很低。因此本實施例只 把物理塊y與物理塊Z與邏輯塊的對應關係交換一下，把 塊y對應交換塊，物理塊2對應邏輯塊χ，節省了 _ 次貧料搬遷的操作，提高了效率。 -料2步驟Μ7 ’將物理塊7中不被寫命令覆蓋的原有 包括步驟： 祝百的貝枓搬遷技術，具體 步驟sm，從寫命令中取得要寫人新資料的指定位 iit 9 ，將指定位址與物理塊y的位址比較，得到 不，寫:1定寫人新資料的未指定位址，這些未指定位 址中的貝料是不被新資料覆蓋的原有資料； 步驟S173 ’將未覆蓋位址中的搭士 ^ z的相應位址中，流程結束的原有㈣複製到物理塊 直』的特性是要先擦除才能寫資料，如果 料的物理塊寫入新資料，很可能造 成原有貝丢失’故本實施例在每 都使用交換塊，以保證原有㈣㈣全。τ寫貝心^ 本貫鈀例對於讀命令的處理過程如下. 13 丄叶oou 收到讀取邏輯塊X資料的讀命令； 。十异邏輯塊X屬於哪個邏輯區域； 將該邏輯區域的對照表载入到記憶體； 查對照表，根據邏輯塊χ的邏輯位址查找 y ; 步驟S25，讀取物理塊y中的指定資料； 步驟S26 ’返回讀到的資料，讀過程結束。The machine receives the data and stores it in the buffer of the device' and then writes the data from the buffer to the flash. In the present example, it is known from the write command that the data block X is to be written to the logical block X and the logical block X corresponds to the physical block y. The physical block 7 itself has the original data. According to the characteristics of the flash memory, the physical block The material 'must be divided by the physical block (9) before the data can be written. But = In addition to the physical block y, the original (four) will be money. Therefore, the physical block of the buffer block can not be directly written to the physical block of the swap block. Only after the write operation of writing the buffer data to the swap block, the physical block corresponding to the object and the swap block The part of the Μ ι ι 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存 保存At this point, the logical block X is stored in the physical block ’. In the physical block y, the physical block y can be erased although the original suffix is present: useless. Of course, after the physical block y is erased, all the data in the physical block Z is copied back to the physical block y ' but this is obviously inefficient. Therefore, in this embodiment, only the correspondence between the physical block y and the physical block Z and the logical block is exchanged, and the block y corresponds to the exchange block, and the physical block 2 corresponds to the logical block χ, which saves the operation of the _ poor material relocation and improves the efficiency. . - Item 2 Step Μ 7 'Including the original block that is not covered by the write command in the physical block 7: Zhu Bai's Bellow relocation technology, the specific step sm, from the write command to obtain the designated bit iit 9 to write new data, Compare the specified address with the address of the physical block y to get no, write: 1 to specify the unspecified address of the new data of the person, the bedding in these unspecified addresses is the original data not covered by the new data; Step S173 'In the unaddressed address of the corresponding address in the address ^ z, the original (4) of the end of the flow is copied to the physical block straight" is characterized by being erased before writing the data, if the physical block of the material is written The new data is likely to cause the original shell to be lost. Therefore, this embodiment uses the exchange block every time to ensure the original (four) (four) all. τ写贝心 ^ The process of the palladium case for the read command is as follows. 13 丄叶oou Receive a read command to read the logic block X data; Which logical region X belongs to the logical block; loading the comparison table of the logical region into the memory; checking the comparison table, searching for y according to the logical address of the logical block ;; and reading the specified data in the physical block y at step S25 Step S26 'Return the read data and the reading process ends.
步驟S21， 步驟S22， 步驟S23， 步驟S24， 到對應的物理塊 矣知月提出第一貫施例’在第一實施例基礎上對對照 表和父換塊的管理提出改進。 本實施例至少在儲存媒體量產的時候建立一完整對照 :完整對照表 '可使所有邏輯位址都與好的物理塊 起來’而不對應壞物理塊’將壞的物理塊忽略不使 用’建立邏輯位址與所有好塊的物理位址之間的對應關 係0 例如儲存媒體共有1G個物理塊，其中物理塊2是壞 =建立完整對照表的做法是··從邏輯塊Q和物理塊〇開 二从立對^，先為邏輯塊〇找一個物理塊與之對應，檢測 故塊0是好塊就把物理塊〇對應到邏輯塊〇;再順序 幻找一個物理塊與之對應，因為物理塊〇已經有 以順序檢測物理塊1，物理塊u好塊就把 對到邏輯塊1。再為邏輯塊2找一値物理塊與之 二' 因為物理塊丨之前的塊都操作過了，就從物理塊2 、=秘發現物理塊2是壞塊，不能將物理塊2與任何一個 ^ 對應，於是再往後找到物理塊3，發現物理塊3是Step S21, step S22, step S23, step S24, to the corresponding physical block, the first embodiment is proposed. On the basis of the first embodiment, the management of the comparison table and the parent replacement block is proposed. In this embodiment, a complete comparison is established at least when the storage medium is mass-produced: the complete comparison table 'can make all logical addresses get good physical blocks' without corresponding bad physical blocks 'ignoring bad physical blocks without using' Establish a correspondence between a logical address and a physical address of all good blocks. For example, the storage medium has a total of 1G physical blocks, wherein the physical block 2 is bad. The practice of establishing a complete lookup table is from the logical block Q and the physical block. Open the two slave pairs ^, first find a physical block for the logic block, and detect that the block 0 is a good block and the physical block 〇 corresponds to the logical block 〇; then the phantom finds a physical block corresponding to it. Since the physical block 〇 already has the physical block 1 detected in order, the physical block u is a good block and is paired to the logical block 1. Then find a physical block for logic block 2 and two. 'Because the blocks before the physical block are all operated, from physical block 2, = secret, physical block 2 is a bad block, and physical block 2 cannot be combined with any one. ^ Correspondence, then find physical block 3 later, and find that physical block 3 is