Re: [PATCH v3 1/1] dcache: Translating dentry into pathname without taking rename_lock

[Ian Kent <raven@themaw.net>]

On Sat, 2013-09-07 at 18:32 +0100, Al Viro wrote:
> On Sat, Sep 07, 2013 at 04:01:10AM +0100, Al Viro wrote:
> > 	* plain seqretry loop (d_lookup(), is_subdir(), autofs4_getpath(),
> > ceph_misc_build_path(), [cifs] build_path_from_dentry(), nfs_path(),
>       _mds_, actually - sorry.
> > [audit] handle_path())
> > 	* try seqretry once, then switch to write_seqlock() (the things
> > that got unified into d_walk())
> > 	* try seqretry three times, then switch to write_seqlock() (d_path()
> > and friends)
> > 	* several pure write_seqlock() users (d_move(), d_set_mounted(),
> > d_materialize_unique())
> 
> BTW, autofs4_getpath() looks really odd:
> static int autofs4_getpath(struct autofs_sb_info *sbi,
>                            struct dentry *dentry, char **name)
> and *name is never modified in it.  Why not simply pass it by value?
> Moreover, I'm not sure I understand what do we need sbi->fs_lock in
> there.  Other than that, it's very close to dentry_path() (well, that
> and different calling conventions).  Ian?

Yes, it is close to dentry_path() but the path functions used to return
the full path to the root, although I don't see where dentry_path() get
the root, mmm ... 

autofs4_getpath() is supposed to return a path relative to the current
dentry.

That goes back to the pseudo direct mounts of autofs version 4 where the
keys could be a relative path.

The fs_lock probably isn't needed. This was a tree of directories and I
didn't want mount requests coming in while I was getting the path, and
didn't want an expire to happen either, although there should only be
one expire process anyway.

Ian