Thread: N900 overclock tests/results and new versions
View Single Post
egoshin is offline egoshin
2010-04-10 , 17:28
Posts: 992 | Thanked: 995 times | Joined on Dec 2009 @ California
#73
Originally Posted by igor View Post
One word of warning to those experiencing instability: if you are triggering writes to the internal OneNAND memory (the one where most of the system files live), you are better off with reflashing the whole fiasco image.

Because of the nature of the medium, write operations cause shuffling around of random pages, to even out the wearing.
I doubt that.

In my best knowledge most of flash memory chips do not move pages on it's own behalf. And one of equipment known to me which moves pages (it was not a single chip) did it with attention to failures - if write fails it doesn't update page table.

Flash chips only select page location for new pages in some pseudo-random manner. So, if you fail - you fail with only page which is just written.

Of course, there is a problem of file system pointers writes but usualy it is not a big problem with modern FS - it has a lot of duplicates and so on. Not sure about UBIFS...
 

The Following User Says Thank You to egoshin For This Useful Post: