Re: [PATCH 6.1] cifs: use origin fullpath for automounts

[Greg KH <gregkh@linuxfoundation.org>]

On Mon, Jul 15, 2024 at 01:48:44PM +0200, Greg KH wrote:
> On Sat, Jul 13, 2024 at 03:11:47AM +0000, Andrew Paniakin wrote:
> > From: Paulo Alcantara <pc@cjr.nz>
> > 
> > commit 7ad54b98fc1f141cfb70cfe2a3d6def5a85169ff upstream.
> > 
> > Use TCP_Server_Info::origin_fullpath instead of cifs_tcon::tree_name
> > when building source paths for automounts as it will be useful for
> > domain-based DFS referrals where the connections and referrals would
> > get either re-used from the cache or re-created when chasing the dfs
> > link.
> > 
> > Signed-off-by: Paulo Alcantara (SUSE) <pc@cjr.nz>
> > Signed-off-by: Steve French <stfrench@microsoft.com>
> > [apanyaki: backport to v6.1-stable]
> > Signed-off-by: Andrew Paniakin <apanyaki@amazon.com>
> > ---
> > This patch fixes issue reported in
> > https://lore.kernel.org/regressions/ZnMkNzmitQdP9OIC@3c06303d853a.ant.amazon.com
> > 
> > 1. The set_dest_addr function gets ip address differntly. In kernel 6.1
> > the dns_resolve_server_name_to_ip function returns string instead of
> > struct sockaddr, this string needs to be converted with
> > cifs_convert_address then.
> > 
> > 2. There's no tmp.leaf_fullpath field in kernel 6.1, it was introduced
> > later in a1c0d00572fc ("cifs: share dfs connections and supers")
> > 
> > 3. __build_path_from_dentry_optional_prefix and
> > dfs_get_automount_devname were added to fs/smb/client/cifsproto.h
> > instead of fs/cifs/dfs.h which doesn't exist in 6.1
> 
> Now queued up, thanks.

Nope, now dropped.  This required me to apply d5a863a153e9 ("cifs: avoid
dup prefix path in dfs_get_automount_devname()") to the tree (and
attempt to backport it myself), but that didn't work, as it caused
reported build errors as seen here:
	https://lore.kernel.org/r/aaccd8cc-2bfe-4b2e-b690-be50540f9965@gmail.com

So I'm dropping this, and the fix-up patch.  Please be careful when
sending backports that you do NOT miss additional patches that are
required as well.  I'll be glad to take backports, as long as they are
correct :)

thanks,

greg k-h