RE: Subject: [PATCH V7 1/4] mm: frontswap: swap data structure changes

[Dan Magenheimer <dan.magenheimer@oracle.com>]

> From: KAMEZAWA Hiroyuki [mailto:kamezawa.hiroyu@jp.fujitsu.com]
> Subject: Re: Subject: [PATCH V7 1/4] mm: frontswap: swap data structure changes
> 
> > > From: KAMEZAWA Hiroyuki [mailto:kamezawa.hiroyu@jp.fujitsu.com]
> > > Subject: Re: Subject: [PATCH V7 1/4] mm: frontswap: swap data structure changes
> >
> > Hi Kamezawa-san --
> >
> > Domo arigato for the review and feedback!
> >
> > > Hmm....could you modify mm/swapfile.c and remove 'static' in the same patch ?
> >
> > I separated out this header patch because I thought it would
> > make the key swap data structure changes more visible.  Are you
> > saying that it is more confusing?
> 
> Yes. I know you add a new header file which is not included but..
> 
> At reviewing patch, I check whether all required changes are done.
> In this case, you turned out the function to be externed but you
> leave the function definition as 'static'. This unbalance confues me.
> 
> I always read patches from 1 to END. When I found an incomplete change
> in patch 1, I remember it and need to find missng part from patch 2->End.
> This makes my review confused a little.
> 
> In another case, when a patch adds a new file, I check Makefile change.
> Considering dependency, the patch order should be
> 
> 	[patch 1] Documentaion/Config
> 	[patch 2] Makefile + add new file.
> 
> But plesse note: This is my thought. Other guys may have other idea.

I think that is probably a good approach.  I will try to use it
for future patches.  But since this frontswap patchset is already
on V7, I hope it is OK if I continue to organize it for V8 the same
as it has been, as it might be confusing to previous reviewers
to change the organization now.

Thanks,
Dan