1258079 九、發明說明： [發明所屬之技術領域】 本發明是有關於一種嵌入式 ，牯別θ 4匕仏〇 凡义貝抖配置管理方法 寸L曰-種易於更新、節省儲存 資料配置管理方法。 〕瓜入式尔統之 【先前技術】 嵌入式系統（Embedded System) $ + ¥ Ύ + ::的裝置，例如個人電腦一::： = =貧料存取介面等，或者如個人數位助 = 寺，㈣由細人緣㈣㈣嫩憶體二 内的程式，而達到完成特定動作的目的。 “如圖1所不’說明目前嵌入式系統5的資料備份管理 方式…般嵌人式“ 5主要是由—處理單元Η、—己情 體控制單元52及周邊裝置53構成資料處理的核心，而在 各個單元間係以一系統匯流排54來作為資料傳輸的通道， 亚錯由記憶體控制單元52來對_電氣可抹除可程式規割 55(以下簡稱EEPRqm)來進行資料存取的控制。 _在進行開機之初始化（InitiaIing)過程時，記憶體控制單 元52是將EEPROM(55)之資料備份區55ι巾的備份資料載 入一快取（CM6)記憶14 56中，由於快取記憶體56可使系 統5快速地進行資料讀取或寫入的動作，因此系統5在進 行-般資料的存取時’即可對絲記憶體％直接進行存取 動作。 EEPROM(55)中具有之資料備份區551，是以映射區 1258079 (Mirr〇r Bank)的方式重複寫入和附加一寫入順序之計數器來 達成保護資料之功效；因此，當系統重新啟動時，藉由比 對資料備份區551之資料是否異常，再利用該寫入順序之 計數器之資訊取得未毀損之上次資料。 §快取記憶體56的資料有毀損時，系統5可以 EEPR〇M(55)中的備份資料回復之，若快取記憶體兄有資 料寫入時，則記憶體控制單元52中的剖析器521(parser)會 將依據設定檔54(ConfigUration File)中的資料對應地將變更 的資料重複寫入EEPR〇M(55)中，且剖析器521亦會在寫入 資料時附加適當資訊，預防寫入時不慎斷電而在重新啟動 時可辨識出該資訊所代表之意義再作適當處理。 使用剖析器52 1作為資料備份管理的做法，是將所有 需要備份纟EEPROM(55)中的參數以人工方式將各參數的名 稱、索引參數名稱、所佔記憶體大小及初始值相關資料登 錄在設定檔5 4中。 目前資料備份管理方式存在有以下的缺點： 1 · EEPR〇M(55)須提供二倍的儲存空間予原始資料及映 射區備份資料，十分佔用記憶體空間。 2.隨著產品特性的不同，系統版本亦須不斷更新，則 必㈣時開發並維護剖析器521，則對應的設"μ中的 組態設定方式亦須大幅修改，費時費力。 3.設定檔54須以人工方式 同時使用備份資料時，則設定檔 知入，當許多不同開發者須 54將不易維護。 無法針對不 4· EEPR〇M(55)對於資料寫入的策略固定 同資料特性進行最 【發明内容】 土 。 口此，本發明 、省力，且可節小j立的，即在提供一種維護上更加省時 營理方法及其裝置。L肢儲存空間的嵌入式系統之記憶體 於是，本發明與A 4 式系統具有一電子:式系統之記憶體管理方法，該嵌入 ，eeprom内含、@式可袜除可規劃型唯讀記憶體（EEPR〇M) ，該方法包含下I數分別儲存有不同資料内容之資料表單 別標定一項目获Μ步驟：（A)將EEPR〇M之各該資料表單分 同資料内容在該==^態值’纟中各該項目代碼表示不 否正在修正該資料“早内之存放位址’該狀態值表示是 入一筆資料内☆ I今，其預設值為非寫入狀態；（B)欲寫 該頊曰二，谷時’變更該預設值使其呈寫入狀態；(C)依 六' .1查5旬在該存放位址之資料内容，並將該資料内 以更新’及（D)將該狀態值回復為非“㈣之該 值0 呈士：明嵌入式系統之記憶體管理方法，該嵌入式系統 -有兒子式可抹除可規劃型唯讀記憶體（EEpR〇M)及一隨 機存取記憶體（RAM)，EEpR〇M Μ含一主程式及複數儲存 不同貝料内容之資料表單，該方法包含下述步驟：（Α)將 EEPROM之各該資料表單分別標定_識別碼，各該識別碼 表不不同貪料表單在之存放位址；（Β)系統啟動時，由 EEPROM下載該等資料表單至Ram中，RAM並要求 EEPROM之主程式指定在RAM中的存放位址；及 1258079 (c)EEPR〇m之主裎今 料寫入RAM中。式依據各該識別碼分配該存放位址供資 本發明谈入式系试—^ 參數的方式達到管理的功::憶：管理方法是利用標定幾種 方式作為系統之寫入二二如以項目代碼及狀態值的 的方式，當識別碼變動=功“及利用標定識別碼 而給予適當的處理。—置有變動的情況 【實施方式】 '有關本發明之前述及其他技術内容、特點與功效，在 '下配σ|考圖式之—較佳實施例的詳細說明中，將可产 楚的呈現。 a 如圖2所示，為本發明嵌入式系統之記憶體管理方法 勺較4土貝％例。必須§兒明的是，由於該方法僅涉及記憶體 之使用’因此以下將針對嵌入式系統之記憶模、组】之相關 元件及原理作闡述。 本發明嵌入式系統之記憶模組丨包含一用以暫存資料 之隨機存取記憶體（以下簡稱RAM)n，及一耦接RAM(n) 之電子式可抹除可規劃型唯讀記憶體（以下簡稱eepr〇M) j 2 ，其中，EEPR〇M( 12)具有一主程式】〇及以不同儲存位址 之配置方式（容後再述）分割的複數資料區塊121。 其中’各資料區塊1 21分別儲存不同資料庫内容之資 料表單]3，各資料表單13分別標定有一識別碼（Magic ID)131、一項目代碼（ltem)]32、一 狀態值（Active Value)133 及多筆資料内容（Content) 134，而各識別碼131表示不同資 1258079 料表單13在EEPR〇M(12)中的存放位址；各項目代碼i32 表示不同資料内容】34在資料表單13内之存放位址；狀態 值133則表示是否正在處於修正資料内容的狀態，其預設 值為非舄入狀態（Non-Active)。 本發明嵌入式系統之識別碼131的原理即在於··由於 記憶體空間有⑯，針對具有大量資料的嵌人式系統，以數 位電視接收裝置的記憶體為例，原本全區的頻道為 CH14 CH69 ’各頻運各自儲存8筆資料内容134，此時全區 類別之識別碼131設定為〇χ〇〇。若各頻道須追加2筆成】〇 筆資料内各’但是EEPR〇M(12)中已無多餘空間，因此可利 用刀區頻道較少的特性，如掃描後得到僅有〔Η〕#〜CH35， 則剩餘的空間（CH14〜CH23、CH36〜CH69)便可重新分配給 各頻迢（CH24〜CH35)作為儲存追加資料内容134之用，此種 。己匕'版重新为配方式即是將此種分區類別之識別碼】3 1加 以《又疋，在系統開機時便可以比對識別碼13丨的方式來對 兄憶體配置作最佳化的設定。 另外，項目代碼132與狀態值133的使用方式，是欲 修正貝料内容134時，先去變更狀態值133使其呈寫入狀 心（Active) ’接著修改項目代碼132，並依項目代碼】修 正在存放位址之資料内容】34，並將資料内容]34加以更新 ，再將狀態值133回復為非寫入狀態之預設值。 疋若由於系統斷電或操作不慎使得資料内容1 3 4之 寫入值為異常時’纟重新開機後，系統便先去㈣狀態值 1 3 J疋否為禺入狀態？若是，再檢查其項目代碼1 3 2對應的 1258079 資料内谷1 34是否與系統預設值一致，若不—致時，即判 斷為操作異常狀態，再將系統預設值寫入資料内容134以 作為還原，還原後再將設定狀態值133回復為非寫入狀態 之預設值。 由於是内建主程式10進行處理，可省去剖析器的設置 ，將各資料區塊121 +的資料表# 13搭配各自的識別碼 來分配各頻道資料，以及使用項目代碼132及狀態值 133作為系統寫入中斷檢知與資料保全，可使系統更容易維 護’亦增加了程式設計上的彈性。 茲將本發明嵌入式系統之記憶體管理方法的流程說明 如下： 當系統一開機時（步驟101)，由eepr〇m〇2)下載該等 資料表單]3至RAM⑴）中，RAM⑴）並要求EEpR〇M⑽ 之主程式ίο指定的存放位址（步驟1〇2)，接著， EEPR0M(12)之主程式依據各識別* 131分配該存放位址供 資料寫入RAM(n)中（步驟1〇3)。藉此，當系統啟動時， EEPR〇M(I2)之主程式1G由該等資料區塊ΐ2ι下載資料至 RA>M(U)中，並要求主程式指定的存放位址，則可依據 各識別碼131分配存放位址供資料寫人ram⑴）中。 配置纪憶體之存放位址後（步驟1〇1〜]〇3)，檢查4 卿ROM⑽之各資料表$ π之狀態值133 {否呈寫^ 心：(乂驟1〇4) ’若呈非寫入狀態，顯示資料為正常狀態㈢ 矣又損）’即可進行一般開機程序（步驟105)。 若該狀態值133顯示EEPR〇M(12)呈寫入狀態 即顯示 10 1258079 寊料可能毀損而來不及將狀態值133回復成預設值，因此 需要依項目代碼132修正在存放位址之資料内容134，並將 貧料内容134加以更新（步驟1Q7);此外，再進行—般開機 程序（步驟105)後，欲再修正資料内容】34時，必須先變更 狀〜值1 3 j使其呈舄入狀態（步驟】〇6)，接著依項目代碼 132修正在存放位址之資料内容134，並將資料内容丨34加 以更新（步驟1〇7)，最後，再將狀態值n3回復為非寫入狀 態之預設值（步驟108)。 知納上述，本發明嵌入式系統之記憶體管理方法具有 下述優點： 、么田本犹貝料量很大時，以項目代碼及狀態值的方玉 作為尔，，先之寫入中斷檢知功㊣，無須作全部資料之映射名 伤作為保全，較為節省存取時間及記憶體空間。 —^ a有設置複雜的剖析器及設定播，而是直接以程3 设定麥數的方式存取記憶體，節省了以往剖 的维護工作。 久。又疋七 =標⑼別碼的方式，當識別碼變動時即顯示tt 匕，配置有變動的情況而給予適 純治 田版本一旦琴 可針對不同資料進行最佳化的記憶體配置。 惟以上所述者，僅為本發明之較佳實施例而已 旎以此限宏太八DD〜 0 田1 〜 本^明貫施之範圍，即大凡依本發明申咬直4 範圍及私日日…、口 α 甲。月專矛 ^明内容所作之簡單的等效變化與修飾 屬本舍明專利涵蓋之範圍内。 【圖式簡單說明】 1258079 圖1是一示意圖，說明目前嵌入式系統的資料備份管 理方式； 圖2是一示意圖，說明本發明嵌入式系統之記憶體管 理方法的一較佳實施例；及 圖3是一流程圖，說明本發明嵌入式系統之記憶體管 理方法的各步驟。 12 1258079 【主要元件符號說明】 1 嵌入式糸統之記 13 - 資料表單 憶模組 131 識別碼 10 主程式 13 2 * 項目代碼 11 RAM 133 狀態值 12' EEPROM 1 3 4… 資料内容 121 ^料£塊 101〜108 步驟 131258079 IX. Description of the invention: [Technical field to which the invention pertains] The present invention relates to an embedded, discriminating θ 4 匕仏〇 义 贝 贝 贝 配置 配置 配置 配置 种 种 种 种 种 种 种 种 种 种 种 易于 易于 易于 易于 易于 易于 易于 易于 易于 易于 易于. 〕Into the system [previous technology] Embedded System (Embedded System) $ + ¥ Ύ + :: devices, such as personal computer one::: = = poor material access interface, etc., or as personal digital help = The temple, (4), achieves the purpose of accomplishing a specific action by the program of the fine-minded (four) (four) reminiscent of the body. "As shown in Figure 1", the current data backup management method of the embedded system 5 is as follows: "The embedded system" is mainly composed of a processing unit, a modal control unit 52, and a peripheral device 53. A system bus 54 is used as a channel for data transmission between the units, and the error is stored by the memory control unit 52 for data access by the EEPROM (hereinafter referred to as EEPRqm). control. _In the initialization process (InitiaIing), the memory control unit 52 loads the backup data of the data backup area 55 of the EEPROM (55) into a cache (CM6) memory 14 56 due to the cache memory. 56 allows the system 5 to quickly perform data reading or writing operations, so that the system 5 can directly access the silk memory % when accessing the general data. The data backup area 551 in the EEPROM (55) is repeatedly written and appended with a write sequence counter in the manner of mapping area 1258079 (Mirr〇r Bank) to achieve the protection of data; therefore, when the system is restarted By comparing the data of the data backup area 551, the information of the counter of the writing sequence is used to obtain the uncorrupted data. § When the data of the cache memory 56 is damaged, the system 5 can reply the backup data in the EEPR〇M (55). If the cache memory has data to be written, the parser in the memory control unit 52 521 (parser) will repeatedly write the changed data in EEPR〇M(55) according to the data in the configuration file 54 (ConfigUration File), and the parser 521 will also add appropriate information when writing the data to prevent Inadvertent power-off during writing and recognition of the meaning of the information on the restart can be handled appropriately. Using the parser 52 1 as the data backup management method, all the parameters that need to be backed up in the EEPROM (55) are manually registered with the names of the parameters, the index parameter names, the occupied memory size, and the initial values. Set file 5 4 in. At present, the data backup management method has the following disadvantages: 1 · EEPR〇M(55) must provide twice the storage space for the original data and the backup area backup data, which is very occupying the memory space. 2. With the different product characteristics, the system version must be updated continuously. If the parser 521 is developed and maintained, the configuration setting method of the corresponding setting "μ must be greatly modified, which is time-consuming and labor-intensive. 3. When the profile 54 has to be manually used to back up the data at the same time, the profile is known, and many different developers must be difficult to maintain. It is impossible to fix the data for the data writing of the EEPR 〇M (55). In view of this, the present invention is labor-saving and can be reduced in size, that is, it provides a more time-saving method and apparatus for maintenance. The memory of the embedded system of the L limb storage space, the present invention and the A 4 type system have an electronic: system memory management method, the embedding, the eeprom inclusion, the @式可袜, in addition to the programmable read only memory Body (EEPR〇M), the method includes the following information, the data sheet stored separately with different data content, and the calibration step of the item: (A) the information content of each of the EEPR〇M data forms in the == ^ State value '纟 each item code indicates whether the data is being corrected. "Pre-storage address" This status value indicates that it is in a data ☆ I, its default value is non-write status; (B ) To write the second, the valley 'change the preset value to make it write state; (C) check the contents of the storage address according to the six '.1 and update the data 'And (D) return the status value to a value other than "(4). 0: The memory management method of the embedded system, the embedded system - has a son-style erasable planable read-only memory ( EEpR〇M) and a random access memory (RAM), EEpR〇M Μ contains a main program and multiple storage The data form of the beaker content, the method comprises the following steps: (Α) calibrating each of the data forms of the EEPROM with an _identification code, each of which is not different from the gracious form in which the address is stored; (Β) system At startup, the data form is downloaded from the EEPROM to the RAM, and the RAM is required to specify the storage address in the RAM by the main program of the EEPROM; and 1258079 (c) the main component of the EEPR〇m is written into the RAM. According to each identification code, the storage address is allocated for the capital invention to enter the system test-^ parameter to achieve the management work:: the management method is to use the calibration method as the system writes the second and second The way of the code and the status value, when the identification code changes = the work "and the appropriate processing is given by using the calibration identification code. - The case where there is a change [Embodiment] 'The foregoing and other technical contents, features and effects of the present invention In the detailed description of the preferred embodiment of the present invention, the present invention will be presented in the detailed description of the preferred embodiment. As shown in FIG. 2, the memory management method of the embedded system of the present invention is more than four soils. It is necessary to clarify that since this method only involves the use of memory, the following will explain the related components and principles of the memory module and group of the embedded system. The group includes a random access memory (hereinafter referred to as RAM) n for temporarily storing data, and an electronic erasable programmable read only memory (hereinafter referred to as eepr〇M) coupled to the RAM (n). j 2 , where EEPR M(12) has a main program 〇 and a plurality of data blocks 121 divided by different storage address configurations (described later). wherein each data block 1 21 stores data forms of different database contents respectively. 3, each data form 13 is respectively labeled with an identification code (Magic ID) 131, an item code (ltem) 32, an status value (Active Value) 133, and a plurality of data contents (Content) 134, and each identification code 131 Indicates the storage address of the different resources 1258079 material form 13 in EEPR〇M(12); each item code i32 represents the different data content] 34 in the data form 13 storage address; the status value 133 indicates whether it is in the correction data The state of the content, the preset value is non-intrusion state (Non-Active). The principle of the identification code 131 of the embedded system of the present invention lies in that the memory space has 16 for an embedded system with a large amount of data. Taking the memory of the digital TV receiving device as an example, the original channel of the whole area is CH14 CH69 'each frequency store stores 8 pieces of data content 134, and the identification code 131 of the whole area category is set to 〇χ〇〇. Channel Add 2 pens into the data] but there is no extra space in the EEPR〇M(12), so you can use the characteristics of less knife channel, such as only [Η]#~CH35 after scanning, then the remaining The space (CH14~CH23, CH36~CH69) can be reassigned to each frequency (CH24~CH35) as the storage additional data content 134. This type of re-distribution method is to divide the partition. The identification code of the category] 3 1 is added, and when the system is turned on, the identification of the brother-remembering configuration can be optimized by comparing the identification code 13丨. In addition, when the item code 132 and the status value 133 are used, when the material content 134 is to be corrected, the status value 133 is first changed to be written to the active state ("After modifying the item code 132, and according to the item code] Correct the data content of the storage address] 34, and update the data content 34, and then return the status value 133 to the default value of the non-write state.疋If the data content of the data content is abnormal due to system power failure or inadvertent operation, the system will go to (4) status value 1 3 J疋 Is it the intrusion status? If yes, check whether the 1258079 data valley 1 34 corresponding to the project code 1 3 2 is consistent with the system preset value. If not, the operation abnormal state is determined, and then the system preset value is written into the data content 134. After the restoration, the set state value 133 is returned to the preset value of the non-write state. Since the built-in main program 10 performs processing, the parser settings can be omitted, and the data table #13 of each data block 121+ is assigned with the respective identification codes to allocate each channel data, and the project code 132 and the status value 133 are used. As system write interrupt detection and data preservation, the system can be more easily maintained' also increases the flexibility of programming. The flow of the memory management method of the embedded system of the present invention is explained as follows: When the system is powered on (step 101), the data form]3 to RAM(1)), RAM(1)) are requested by eepr〇m〇2) The main program of EEpR〇M(10) ίο specifies the storage address (step 1〇2). Then, the main program of EEPR0M(12) allocates the storage address according to each identification*131 for data to be written into RAM(n) (step 1) 〇 3). Therefore, when the system is started, the main program 1G of the EEPR〇M (I2) downloads the data from the data block ΐ2 to the RA>M(U), and requests the storage address specified by the main program, according to each The identification code 131 allocates the storage address for the data writer ram(1)). After configuring the storage address of the memory system (steps 1〇1~]〇3), check the status value of each data table of the 4th ROM (10) $ π {No presentation ^ Heart: (乂1〇4) 'If In the non-write state, the display data is in the normal state (3) 矣 and loss), and the general boot process can be performed (step 105). If the status value 133 indicates that the EEPR 〇 M (12) is in the write state, the display 10 1258079 may be damaged and the status value 133 may not be restored to the preset value. Therefore, the data content at the storage address needs to be corrected according to the item code 132. 134, and the poor content 134 is updated (step 1Q7); in addition, after the general boot process (step 105), to modify the data content] 34, must first change the value ~ value 1 3 j to present Intrusion state (step 〇6), then the data content 134 at the storage address is corrected according to the item code 132, and the data content 丨34 is updated (step 1〇7), and finally, the status value n3 is returned to non- The preset value of the write state (step 108). Knowing the above, the memory management method of the embedded system of the present invention has the following advantages: When the amount of the money is large, the square of the project code and the state value is used as the er, and the write interrupt is first detected. Knowing the merits, there is no need to map all the data as a security, which saves access time and memory space. —^ a has a complicated parser and setting broadcast. Instead, it directly accesses the memory by setting the number of wheat in the path 3, which saves the maintenance work of the previous section. Long. In addition, the method of 疋7 = (9) is different. When the identification code changes, tt 匕 is displayed, and the configuration is changed to give the appropriate version of the memory. Once the piano can optimize the memory configuration for different data. However, the above description is only for the preferred embodiment of the present invention, and has been limited to the scope of the macro-eight DD~0 field 1~ this is the scope of the application, that is, according to the invention, the application of the bite straight 4 range and private day Day..., mouth α A. The simple equivalent changes and modifications made by the content of the Moon are within the scope of this patent. BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a schematic diagram showing a data backup management method of an embedded system; FIG. 2 is a schematic diagram showing a preferred embodiment of a memory management method of the embedded system of the present invention; 3 is a flow chart illustrating the steps of the memory management method of the embedded system of the present invention. 12 1258079 [Explanation of main component symbols] 1 Embedded system note 13 - Data form memory module 131 Identification code 10 Main program 13 2 * Item code 11 RAM 133 Status value 12' EEPROM 1 3 4... Data content 121 £blocks 101~108 Step 13