[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: MTD control at JFFS2 level?

> For example, can it be guaranteed that MTD allocates exactly one node
> for a JFFS2 write(), even if the data to be written is large (say,
> several kilobytes)? 

No. JFFS2 will allocate at least one node for every page (usually 4KiB) of 
data, occasionally more.

> Also, can it be guaranteed that a write() operation be atomic? 

No. write() calls from userspace to a JFFS2 file will not be atomic w.r.t 
power failure, which I suspect is what you're after. They will be atomic 
w.r.t other reads and writes though.

Given that the write() call may always return successfully but having 
written fewer data than you asked it to, what possible use would atomic 
writes be anyway?

It may be possible to expose some kind of start_transaction and
end_transaction functionality to userspace, but nobody's looked seriously at
implementing that.


To unsubscribe from this list: send the line "unsubscribe jffs-dev" in
the body of a message to majordomo@xxxxxxx.com