Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
Date:   Tue, 1 Jun 2021 13:26:02 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     Christian Brauner <christian.brauner@ubuntu.com>
Cc:     Changbin Du <changbin.du@gmail.com>,
        Alexander Viro <viro@zeniv.linux.org.uk>,
        "David S. Miller" <davem@davemloft.net>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org, linux-fsdevel@vger.kernel.org,
        stable@vger.kernel.org, Cong Wang <xiyou.wangcong@gmail.com>,
        David Laight <David.Laight@ACULAB.COM>
Subject: Re: [PATCH] nsfs: fix oops when ns->ops is not provided
Message-ID: <20210601132602.02e92678@kicinski-fedora-PC1C0HJN.hsd1.ca.comcast.net>
In-Reply-To: <20210601080654.cl7caplm7rsagl6u@wittgenstein>
References: <20210531153410.93150-1-changbin.du@gmail.com>
        <20210531220128.26c0cb36@kicinski-fedora-PC1C0HJN.hsd1.ca.comcast.net>
        <20210601080654.cl7caplm7rsagl6u@wittgenstein>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Tue, 1 Jun 2021 10:06:54 +0200 Christian Brauner wrote:
> > I'm not sure why we'd pick runtime checks for something that can be
> > perfectly easily solved at compilation time. Networking should not
> > be asking for FDs for objects which don't exist.  
> 
> Agreed!
> This should be fixable by sm like:
> 
> diff --git a/net/socket.c b/net/socket.c
> index 27e3e7d53f8e..2484466d96ad 100644
> --- a/net/socket.c
> +++ b/net/socket.c
> @@ -1150,10 +1150,12 @@ static long sock_ioctl(struct file *file, unsigned cmd, unsigned long arg)
>                         break;
>                 case SIOCGSKNS:
>                         err = -EPERM;
> +#ifdef CONFIG_NET_NS
>                         if (!ns_capable(net->user_ns, CAP_NET_ADMIN))
>                                 break;
> 
>                         err = open_related_ns(&net->ns, get_net_ns);
> +#endif
>                         break;
>                 case SIOCGSTAMP_OLD:
>                 case SIOCGSTAMPNS_OLD:

Thanks! You weren't CCed on v1, so FWIW I was suggesting
checking in get_net_ns(), to catch other callers:

diff --git a/net/socket.c b/net/socket.c
index 27e3e7d53f8e..3b44f2700e0c 100644
--- a/net/socket.c
+++ b/net/socket.c
@@ -1081,6 +1081,8 @@ static long sock_do_ioctl(struct net *net, struct socket *sock,
 
 struct ns_common *get_net_ns(struct ns_common *ns)
 {
+       if (!IS_ENABLED(CONFIG_NET_NS))
+               return ERR_PTR(-EOPNOTSUPP);
        return &get_net(container_of(ns, struct net, ns))->ns;
 }
 EXPORT_SYMBOL_GPL(get_net_ns);