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

Re: Major JFFS2 bug (?)

> Dang! Well, there must be another way. Suggestions?

If your app cares use file locking. If you need to do application level
transactions do transactions at the application level.

> Actually, I thought that write() was allowed to return before having
> written *all the data desired*- but this is *not* happening. write()
> always seems to write the entire data.

Writes can be cached. fsync and friends are write barriers. But they dont have
to be.  Also random power off is outside the posix definition

> write(fd, buf, 0 < size < infinity); /* NOW guaranteed to be atomic due
> to ioctl wrappers around it */
> ioctl(JFFS2_END_ATOMIC_WRITE); /* this guy causes the commit of all data

write versus write is atomic anyway. See the inode semaphore. for write versus
read use file locking

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