[EMAIL PROTECTED] writes: > Can you program the flash chip in your 860 board using mtd?
Yes, but I had to patch the drivers, and as that was the very first thing I did, I haven't had time to get my changes into the core mtd tree. > I'm trying to do this for my 815ep board but I fail to do this. > when i load physmap.o, i got following message and fail to load it. > -- > Dec 2 22:59:27 hcyun kernel: physmap flash device: 0 at 0 > -- > > As you know, both 860 and 815ep use same FWH. So i think if you can do > this, you can tell me how to do. > if you are successful for this, plz tell me how to configure mtd. The code may already be in the archives somewhere. If not I'll get back at this in a couple of days. I want to resolve the whole bootloader thing before I go further.... > And one more question, does your "cache as ram" works fine on PIII ? > If it has no problem, i think it's much better to rewrite my 815ep codes > using C. > Wring code in no stack environment is terribly complicate. Definentily. The "cache as ram" code works with one big exception on a PIII. If you do a write outside of your "cache as ram" block it is treated as a cache conflict and your cache block is invalidated. And since you need to write outside the "cache as ram" area for initializing ECC sdram, and also for writing to the SDRAM config registers (on the 440[BG]X) it is a fatal issue. I believe with MMX and SSE there were some instructions introduced that bypass the cache. It is probably worth testing those, to see if they give the same problem. Eric
