学步园

      又到了周末，今天是2010年7月31日，很纠结。。。。。。因为串口接收大数据包丢失问题，还没有清晰的办法。

      暂且放下它，还是把以前做过的比较有意思的东西重新整理下：

      很多处理器，都支持大容量的NAND，拿三星6410处理器来讲，从资料上看，支持大容量8G NANDFLASH，也说明了需要加片选来控制，这次做的是2GNAND的支持，需要2个片选。OK ～～～～～～～硬件上，NANDFLASH上增加一路片选信号，增加一路R/B控制。

我这个是由2个8G08构成的2GNAND

      软件上调整：

      1, 让系统判断出NAND的容量大小：在初始化NAND的时候，读2个片选对应的NAND的ID，读出两个NAND的相关信息，然后把这些信息告诉系统

     static DWORD ReadFlashID(void)<br /> {<br /> BYTE Mfg, Dev;<br /> int i;</p> <p> NF_nFCE_L(); // Deselect the flash chip.<br /> NF_CMD(CMD_READID); // Send flash ID read command.</p> <p> NF_ADDR(0);</p> <p> for (i=0; i<10; i++)<br /> {<br /> Mfg = NF_RDDATA_BYTE();<br /> if (Mfg == 0xEC || Mfg == 0x98) break;<br /> }</p> <p> Dev = NF_RDDATA_BYTE();</p> <p> NF_nFCE_H();</p> <p> return ((DWORD)(Mfg<<8)+Dev);<br /> }</p> <p>//增加一个读片选2的函数<br /> static DWORD ReadFlashID2(void)<br /> {<br /> BYTE Mfg, Dev;<br /> int i;</p> <p> NF_nFCE_L1(); // Deselect the flash chip.<br /> NF_CMD(CMD_READID); // Send flash ID read command.</p> <p> NF_ADDR(0);</p> <p> for (i=0; i<10; i++)<br /> {<br /> Mfg = NF_RDDATA_BYTE();<br /> if (Mfg == 0xEC || Mfg == 0x98) break;<br /> }</p> <p> Dev = NF_RDDATA_BYTE();</p> <p> NF_nFCE_H1();</p> <p> return ((DWORD)(Mfg<<8)+Dev);<br /> }

NF_nFCE_L1()为增加的那个片选的控制。

2, 把第1步中得到的2GNAND的信息，重新赋给全局变量，即告诉系统新的NAND信息。

在NAND初始化函数中 添加对第2快NAND的初始化

 //********************************初始化第2片NAND******************************************<br /> RETAILMSG(1, (TEXT("init the secend nand/n")));<br /> nMID = (UINT8)(nNandID2 >> 8);<br /> nDID = (UINT8)(nNandID2 & 0xff);</p> <p> RETAILMSG(1, (TEXT("[FMD:INF] FMD_Init() : Read ID = 0x%08x/n"), nNandID2));</p> <p> for (nCnt = 0; astNandSpec[nCnt].nMID != 0; nCnt++)<br /> {<br /> if (nDID == astNandSpec[nCnt].nDID)<br /> {<br /> bNandExt = TRUE;<br /> break;<br /> }<br /> }</p> <p> if (!bNandExt)<br /> {<br /> RETAILMSG(1, (TEXT("[FMD:ERR] FMD_Init() : Unknown ID = 0x%08x/n"), nNandID));<br /> return NULL;<br /> }</p> <p> int NUM_OF_BLOCKS2 = astNandSpec[nCnt].nNumOfBlks;<br /> int PAGES_PER_BLOCK2 = astNandSpec[nCnt].nPgsPerBlk;<br /> int SECTORS_PER_PAGE2 = astNandSpec[nCnt].nSctsPerPg;</p> <p> RETAILMSG(1, (TEXT("[FMD] FMD_Init() : NUM_OF_BLOCKS2 = %d/n"), NUM_OF_BLOCKS2));<br /> RETAILMSG(1, (TEXT("[FMD] FMD_Init() : PAGES_PER_BLOCK2 = %d/n"), PAGES_PER_BLOCK2));<br /> RETAILMSG(1, (TEXT("[FMD] FMD_Init() : SECTORS_PER_PAGE2 = %d/n"), SECTORS_PER_PAGE2));</p> <p> RETAILMSG(1, (TEXT("[FMD] --FMD_Init()2/n")));</p> <p>//相加<br /> NUM_OF_BLOCKS = NUM_OF_BLOCKS1+NUM_OF_BLOCKS1;<br /> PAGES_PER_BLOCK = astNandSpec[nCnt].nPgsPerBlk;<br /> SECTORS_PER_PAGE = astNandSpec[nCnt].nSctsPerPg;</p> <p> RETAILMSG(1, (TEXT("[FMD] FMD_Init() : NUM_OF_BLOCKS = %d/n"), NUM_OF_BLOCKS));<br /> RETAILMSG(1, (TEXT("[FMD] FMD_Init() : PAGES_PER_BLOCK = %d/n"), PAGES_PER_BLOCK));<br /> RETAILMSG(1, (TEXT("[FMD] FMD_Init() : SECTORS_PER_PAGE = %d/n"), SECTORS_PER_PAGE));</p> <p> return((PVOID)g_pNFConReg);<br /> }</p> <p>

3，增加读写操作中的地址转换：

例如现在第1个NAND的大小是8192个BLOCK，那么当我们想去读/写第16000个BLOCK进行操作的时候，我们要做个判断，让系统去操作第2片  NAND的第（16000-8192）个BLOCK。

接下来就是细心的完全驱动相关函数的修改；

由于NAND的驱动相关函数很多，我这个就不全贴出，思路是一样的，注意这个地址转换就可以了（加个判断就OK了），仅列举读NAND函数

BOOL FMD_ReadSector(SECTOR_ADDR startSectorAddr, LPBYTE pSectorBuff, PSectorInfo pSectorInfoBuff, DWORD dwNumSectors)<br /> {<br /> BOOL bRet;</p> <p> //RETAILMSG(1, (TEXT("[R:0x%08x] /n"), startSectorAddr));<br /> #if (NAND_DEBUG)<br /> RETAILMSG(1, (TEXT("[FMD] ++FMD_ReadSector(0x%08x) /n"), startSectorAddr));<br /> #endif</p> <p>#ifdef SYNC_OP<br /> EnterCriticalSection(&g_csNandFlash);<br /> #endif</p> <p> if ( IS_LB )<br /> {<br /> if(8192*64<=startSectorAddr)<br /> {<br /> startSectorAddr=(startSectorAddr-8192*64);//转换为第2块NAND的地址，调用第2块NAND的函数<br /> bRet = FMD_LB_ReadSector2(startSectorAddr, pSectorBuff, pSectorInfoBuff, dwNumSectors, USE_NFCE);//！！！！</p> <p> }<br /> else<br /> bRet = FMD_LB_ReadSector(startSectorAddr, pSectorBuff, pSectorInfoBuff, dwNumSectors, USE_NFCE);<br /> }<br /> else<br /> {<br /> if(8192*64<=startSectorAddr)<br /> {<br /> startSectorAddr=(startSectorAddr-8192*64);//转换为第2块NAND的地址，调用第2块NAND的函数<br /> bRet = FMD_SB_ReadSector2(startSectorAddr, pSectorBuff, pSectorInfoBuff, dwNumSectors, USE_NFCE);//!!!!<br /> }<br /> else<br /> bRet = FMD_SB_ReadSector(startSectorAddr, pSectorBuff, pSectorInfoBuff, dwNumSectors, USE_NFCE);<br /> }</p> <p>#ifdef SYNC_OP<br /> LeaveCriticalSection(&g_csNandFlash);<br /> #endif</p> <p>#if (NAND_DEBUG)<br /> RETAILMSG(1, (TEXT("[FMD] --FMD_ReadSector()/n")));<br /> #endif</p> <p> return bRet;<br /> }</p> <p>

这样编译后，重启，擦出/低格/高格信息：

Enter your selection: a<br /> All block(16384) Erase...<br /> LB######## Error Erasing block 1462!<br /> LB######## Error Erasing block 2011!<br /> LB######## Error Erasing block 5027!<br /> LB######## Error Erasing block 5687!<br /> LB######## Error Erasing block 8048!<br /> 2LB######## Error Erasing block 1005!//转换后的第2块NAND<br /> 2LB######## Error Erasing block 1366!<br /> 2LB######## Error Erasing block 3396!<br /> 2LB######## Error Erasing block 5239!<br /> 2LB######## Error Erasing block 7557!<br /> //低格</p> <p>Reserving Blocks [0x0 - 0x5] ...<br /> ...reserve complete.<br /> Low-level format Blocks [0x6 - 0x3fff] ...<br /> LB######## Error Erasing block 1462!<br /> LB######## Error Erasing block 2011!<br /> LB######## Error Erasing block 5027!<br /> LB######## Error Erasing block 5687!<br /> LB######## Error Erasing block 8048!<br /> 2LB######## Error Erasing block 1005!<br /> 2LB######## Error Erasing block 1366!<br /> 2LB######## Error Erasing block 3396!<br /> 2LB######## Error Erasing block 5239!<br /> 2LB######## Error Erasing block 7557!<br /> ...erase complete.<br /> //高格</p> <p>Enter LowLevelFormat [0x6, 0x3fff].<br /> Erasing flash block(s) [0x6, 0x3fff] (please wait): EraseBlocks: found a bad block (0x5b6) - skipping...<br /> EraseBlocks: found a bad block (0x7db) - skipping...<br /> EraseBlocks: found a bad block (0x13a3) - skipping...<br /> EraseBlocks: found a bad block (0x1637) - skipping...<br /> EraseBlocks: found a bad block (0x1f70) - skipping...<br /> EraseBlocks: found a bad block (0x23ed) - skipping...<br /> EraseBlocks: found a bad block (0x2556) - skipping...<br /> EraseBlocks: found a bad block (0x2d44) - skipping...<br /> EraseBlocks: found a bad block (0x3477) - skipping...<br /> EraseBlocks: found a bad block (0x3d85) - skipping...<br /> Done.<br /> WriteMBR: MBR block = 0x6.<br /> Done.<br />

这样就可以了，做的时间久了 ，可以有些细节会忘记了，但是思路就是这样，很简单。其他4G/8G如果需要片选，则也可以用这个思路做，只不过驱动可能代码又长了很多～呵呵