From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from [203.197.255.76] (helo=mail.navayuga.co.in) by canuck.infradead.org with esmtps (Exim 4.42 #1 (Red Hat Linux)) id 1CeW5C-0001rl-QX for linux-mtd@lists.infradead.org; Wed, 15 Dec 2004 05:08:37 -0500 Message-ID: <005101c4e28e$558ec5b0$8500a8c0@nit37> From: "Subrahmanyam" To: , References: <200412150900.iBF90stl039451@mail.csie.nctu.edu.tw> Date: Wed, 15 Dec 2004 15:40:31 +0530 MIME-Version: 1.0 Content-Type: text/plain; format=flowed; charset="big5"; reply-type=original Content-Transfer-Encoding: 7bit Cc: Subject: Re: A problem of porting jffs2 and mtd in kernel 2.4.18 Reply-To: Subrahmanyam List-Id: Linux MTD discussion mailing list List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Hai jianhung, I have project similar to yours, wherein In my target too we have a file system and mtd drivers stored in the flash which also consists of ieee1394firewire and usb drivers. In the flash in the target system we need to store a particular key/data in the target. This data shall be dynamically modifiable by a user interface provided to the user in the host. Since you almost similar to the platform, henceforth I request you to suggest me a suitable approach. Thanks, Subrahmanyam. ----- Original Message ----- From: To: Sent: Wednesday, December 15, 2004 02:30 PM Subject: A problem of porting jffs2 and mtd in kernel 2.4.18 > Hello, > > This is my first posting. We have a project to porting jffs2 and mtd in > embedded system, the purpose is to use the NAND flash, and the kernel > is 2.4.18. We find there are jffs2 and mtd source code in the kernel > source code. So we just enable it ( use make menuconfig),and write a > mtd driver... > > The good news is this works, we can add, delete a file and after power > on again, the file still exist. Thanks for JFFS2 and MTD, but the > problem comes.... > > 1. the mount time is too slow, so we try to update the source code in > the kernel, but I don\\\'t know the version of jffs2 and mtd in the > kernel source code now is, how do I know it ? And I don\\\'t know > how > to update it, I hope someone can tell me, I would appreciate your > kindness. > > 2. If I use ECC function, and some thing wrong happens. After I trace > the source code, I find when I write a file, the jffs2 will write to > the same nand flash page (although in different position of the same > page) without erase first many times . So the ECC position in the > spare area will be write many times. It is reasonable when I read > this page next time, the ECC value is not equal to the value I > writed in the last time. of course the ecc will error, is this a bug > of this version in kernel 2.4.18? (if I don\\\'t use ECC function, > the > subsystem seems ok) > > 3. I didn\\\'t follow the instruction of \\\"The Linux MTD, JFFS HOW > TO\\\", so > I didn\\\'t use the utility erase and mkfs.jffs. But this still > works, > is this reasonable ? > > Best Regards, > Yoda > > ______________________________________________________ > Linux MTD discussion mailing list > http://lists.infradead.org/mailman/listinfo/linux-mtd/