Re: [PATCH 2/2] fat (exportfs): reconnect file handles to evictedinodes/dentries

[Steven J. Magnani]

On Tue, 2012-07-10 at 06:34 +0900, OGAWA Hirofumi wrote: 
> "Steven J. Magnani" <steve@xxxxxxxxxxxxxxx> writes:
> 
> >> Ah, i_ino. I was talking about i_pos. Well, so, what happens if the
> >> child was renamed to other parent on NFS server machine (not via nfs
> >> client)? The file handle would be including the old i_ino, and the old
> >> i_ino on file handle is still vaild as old parent. So, it returns the
> >> wrong parent?
> >
> > Yes, but I believe exportfs_decode_fh() handles that case:
> >
> >   /*
> >    * Now that we've got both a well-connected parent and a
> >    * dentry for the inode we're after, make sure that our
> >    * inode is actually connected to the parent.
> >    */
> >
> >
> > Really, the FAT NFS code will pretty much parallel that of ext2.
> 
> Hm, not really, if the file handle is including parent ino. ext2 will
> get the latest parent ino, because it checks parent of inode of file
> handle.

Can you point me to the code for this? The code I see looks pretty
congruent to what I think the FAT code would be.

> But if the file handle is including parent ino and we believe it is
> parent, I think NFS server can be return the old parent. The difference
> is the result of ->get_parent().

I'm a little confused about which function we're discussing here.
fat_get_parent() isn't called with a file handle. fat_fh_to_parent() is,
but it is only called by exportfs_decode_fh() and I am reasonably sure
that that function is handling the case you're concerned about.
------------------------------------------------------------------------
Steven J. Magnani               "I claim this network for MARS!
www.digidescorp.com              Earthling, return my space modulator!"

#include <standard.disclaimer>



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