Re: [RFC] Unbork fs.h, 1 of 4

[Daniel Phillips <phillips@bonn-fries.net>]

On January 6, 2002 11:37 pm, Jeff Garzik wrote:
> I share the opinion of a couple others, to use generic_ip and
> generic_sbp.  That's what they are there for.

That approach is workable but inferior.

> May I suggest the course of action whereby you convert the code to use
> ext2_sb(sb) and ext2_i(inode) first, without changing fs.h at all.  That
> lays the groundwork for arguing out the final solution, and the code is
> much cleaner either way.
> 
> I have not seen anyone arguing -against- ext2_sb() and ext2_i()
> cleanups... they seem to make the code obviously more clean.
> 
> After those cleanups go in, your further "unbork fs.h" patches will be
> smaller and only show the core changes you are interested in.

Patches (2) and (3), which implement the 'ext2_i' and 'ext2_s' changes 
respectively, apply independently of the patch (1), the VFS changes.  Did you 
read my posting?

--
Daniel