Re: [PATCH] address_space_operations unification

From: Roman V. Shaposhnick (roman@yellow.pdmi.ras.ru)
Date: Sun May 07 2000 - 13:10:34 EST


On Sat, May 06, 2000 at 07:00:44PM -0700, Linus Torvalds wrote:
>
>
> On Sat, 6 May 2000, Brian J. Murrell wrote:
> >
> > I would hate to think Linus would "rip anyone a new one" in
> > this fashion.

[ stoneage history skiped... ]

> So Roman, my apologies.

  Accepted.

> I won't promise it won't happen again, but let's
> discuss why you felt your patch was needed, ok?

  Ok. First of all, I would like to assure you that I've cooked that patch for
exactly two reasons[1]:

  1. I need consistent interface for pagecache to implement p9fs methods. You do
     not need to see the sources of it, because I have precisely the same set of
     problem as Trond has with NFS.

  2. Ironically, when I first told my story to fsdevel guys I was asking for
     just the same thing you proposed -- change the first parameter of each
     method of address_space_operations to 'struct file *'. But after a while
     we came to the conclusion that pagecache should be general but not a
     file oriented beast.
  
  Thus in my *very* humble opinion there was a problem -- pagecache should be
generic while it is not. What I was trying to do was to "fix" existing
interface and make it appropriate for the non-fs user. If you know better
way of of doing that, please describe that method and I would love to
implement it. In case may be vision of what pagecache should be differs from
yours, please, tell us the truth.

  Linus, please keep in mind, that I do not want to advocate the particular
patch. All I want is simple and rational solution for my problems. Thus if
for some reason you do not want to tackle with pagecache generalization than
would you accept a temporary solution that changes 'readpage' to something
like that:
           int (*readpage)(struct file *, struct page *);
and leaves all other things intact ?

 

Roman.

[1] I hope there is no need for lengthy explanations just because Al, Trond
and Steve mentioned everything I consider relevant for this topic.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/



This archive was generated by hypermail 2b29 : Sun May 07 2000 - 21:00:21 EST