|
Re: Yaffs and MTD: msg#00035linux.file-systems.yaffs
On Monday 12 December 2005 23:18, André Tousch wrote: > Hi, > > I'm currently using yaffs on a new linux embedded device, and I made my > tests by simply implementing the write / read / erase / init methods, and > setting it manualy in yaffs_internal_read_super. > My question is : What is the advantage of writing a new mtd driver for a > yaffs filesystem? (I don't plan to use another flash filesystem, and don't > plan to use already existing mtd drivers). Yes there is: performance. The structure of mtd is mainly embodied in nand.c which is a generic chunk of code that tries to do (almost) all things for (almost) all silicon. While I think the mtd folk have done a good job of trying to meet that goal, a chunk of code that is written to only work with one set of hardware obviously has potential to be a lot more streamlined and faster. Some people are forced down this track anyway because they have some very different hardware (eg. two NAND chips in parallel being treated as a single NAND chip). -- Charles |
|
| <Prev in Thread] | Current Thread | [Next in Thread> |
|---|---|---|
| Previous by Date: | Yaffs and MTD: 00035, André Tousch |
|---|---|
| Next by Date: | YAFFS2/MTD incompatibility isolated: 00035, Michael Arm |
| Previous by Thread: | Yaffs and MTDi: 00035, André Tousch |
| Next by Thread: | Re: Yaffs and MTD: 00035, André Tousch |
| Indexes: | [Date] [Thread] [Top] [All Lists] |
| News | FAQ | advertise |