Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Subject: Re: [PATCH 03/34] teach move_mount(2) to work with OPEN_TREE_CLONE
 [ver #12]
To:     David Howells <dhowells@redhat.com>, viro@zeniv.linux.org.uk
Cc:     torvalds@linux-foundation.org, ebiederm@xmission.com,
        linux-fsdevel@vger.kernel.org, linux-kernel@vger.kernel.org,
        mszeredi@redhat.com
References: <de050902-f3d0-b2db-2627-983a36c87b3c@gmail.com>
 <153754740781.17872.7869536526927736855.stgit@warthog.procyon.org.uk>
 <153754743491.17872.12115848333103740766.stgit@warthog.procyon.org.uk>
 <15751.1539176549@warthog.procyon.org.uk>
From:   Alan Jenkins <alan.christopher.jenkins@gmail.com>
Message-ID: <63c8d073-8309-82cf-6089-ef805feb9fae@gmail.com>
Date:   Wed, 10 Oct 2018 14:06:22 +0100
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101
 Thunderbird/60.0
MIME-Version: 1.0
In-Reply-To: <15751.1539176549@warthog.procyon.org.uk>
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
Content-Language: en-US
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 10/10/2018 14:02, David Howells wrote:
> The attached change seems to fix the lazy-umount problem.
>
> David
> ---
> diff --git a/fs/namespace.c b/fs/namespace.c
> index 5adeeea2a4d9..d43f0fa152e9 100644
> --- a/fs/namespace.c
> +++ b/fs/namespace.c
> @@ -2472,7 +2472,7 @@ static int do_move_mount(struct path *old_path, struct path *new_path)
>   	if (old->mnt_ns && !attached)
>   		goto out1;
>   
> -	if (old->mnt.mnt_flags & MNT_LOCKED)
> +	if (old->mnt.mnt_flags & (MNT_LOCKED | MNT_UMOUNT))
>   		goto out1;
>   
>   	if (old_path->dentry != old_path->mnt->mnt_root)


I can't test any more at the moment, as my laptop died today :). But I 
have no objection to this.

It would be more fun if there was a way to support it :), but I don't 
have a genuine reason to want it.  And you couldn't use it for fully 
general purposes anyway, because umount2( , MNT_DETACH) is defined as 
separating all the child mounts.

P.S. Regarding the issue with the namespace loop.  My strawman solution 
would be for graft_tree() to silently detach any NS file mounts that 
have a sequence number less than or equal to the namespace they are 
being mounted into.