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

Re: JFFS2 endianness.




alan@xxxxxxx.uk said:
>  Sounds like a job for a user space tool, or even an optional fixer in
> the file system mount stage. At boot the log gets walked, is it that
> infeasible to have an extra option to include code to switch the fs to
> the right endian during the mount.

Sounds like a good plan. In fact, depending on just where I'd need to put 
the byteswaps, we may even be able to go one better. 

I don't want conditional byteswaps in to make it 'just work' because they'd
remove the need to convert and the problem would never go away. But it's
going to be temporary, we could do it for a while. 

What we could do is put the conditional byteswaps in but make it detect
endianness not per-fs, but per-node. That's fairly easy because you just
look whether the magic number is 0x1985 or 0x8519. You just make sure all
new nodes written out are BE, and the filesystem gradually gets converted.
The wear levelling will ensure that we do convert the whole filesystem over
time, and in a year or two I can take out the detection and conditional
byteswaps. 

On the other hand, that's probably overkill, and a userspace tool or 
pre-converter in the mount code would be sufficient.

--
dwmw2



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