[linux-cifs-client] Re: [PATCH 00/12] cifs: fix string conversions from wire format to local charset (try #3)

[Jeff Layton]

On Thu, 30 Apr 2009 10:56:57 -0400
Jeff Layton <jlayton at redhat.com> wrote:

> 
> Don't forget about the duplicate code in CIFSSesSetup (or whatever the
> studly caps are for it). That adds another few hundred lines or so.
> Removing the NTLMSSP code would take a 1000 line chunk out of cifs, to
> which I say "hurrah!".
> 
> On the readlink thing, I delved back into old-2.6-bkcvs tree and found
> this:
> 
> commit 8199b6963e73ad209f40b4b3d4bc51a54818852d
> Author: torvalds <torvalds>
> Date:   Thu Sep 2 08:19:12 2004 +0000
> 
>     2004/09/01 23:35:11-05:00 stevef
>     [CIFS] Fix CIFS symlink regression when long symlink paths
>     
>     Signed-of-by: Steve French (sfrench at us.ibm.com)
>     
> ...at this point cifs_readlink was unhooked from cifs_symlink_inode_ops
> and generic_symlink was hooked in. The cifs_readlink code was never
> removed, however. If we've lived without this for 4+ years, I think it's
> safe to remove it.
> 

The bottom line is that those 2 patches remove code that is just dead
weight today. That code should never have been in the tree in the first
place.

I realize that you don't agree, but quite frankly you don't have a dog
in this fight. Those of us who have to maintain this code for
enterprise distributions *do*. Reducing code size and complexity is
very real concern for us. Removing dead codepaths is a huge win.

When the readlink and NTLMSSP code is ready to be re-merged, I'll be
happy to see it, but until then it has no business being in the tree.

-- 
Jeff Layton <jlayton at redhat.com>