Re: cramfs and named-pipe

[Andrew Morton <akpm@linux-foundation.org>]

On Fri, 15 Aug 2008 15:52:20 -0700
Andrew Morton <akpm@linux-foundation.org> wrote:

> On Tue, 05 Aug 2008 12:09:09 +0900 (JST)
> Atsushi Nemoto <anemo@mba.ocn.ne.jp> wrote:
> 
> > On Mon, 4 Aug 2008 11:03:47 +0100, Al Viro <viro@ZenIV.linux.org.uk> wrote:
> > > Eeek...  I'd rather not play these games with directories and devices nodes
> > > as well.  Rationale for the original patch simply doesn't apply for those.
> > > 
> > > IOW, I think it would be much saner if we did the following: make ..._test()
> > > refuse to merge inodes with ->i_ino == 1, take inode setup back to
> > > get_cramfs_inode() and make ->drop_inode() evict ones with ->i_ino == 1
> > > immediately.  Comments?
> > > 
> > > Patch below is completely untested; it builds, but that's it.
> > 
> > Thanks, your patch works well for me.  But it looks a bit large for
> > stable tree (100 line rule).
> > 
> > With current code, I think no problem on empty directories and device
> > nodes.  So how about fixing only FIFO case first (and send it to
> > stable tree) and then go to your patch?
> > 
> 
> Nothing seems to have happened.  Al, do you think your (now tested) patch
> is good for 2.6.27 and 2.6.26.x?  And, it seems, 2.6.25.x.  (All the way
> down to 2.6.14.x!)

<crickets chirping>

Oh well, I'll send it in.