[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <20250919-fanatiker-ethik-7a9bb32ee334@brauner>
Date: Fri, 19 Sep 2025 10:08:33 +0200
From: Christian Brauner <brauner@...nel.org>
To: Jan Kara <jack@...e.cz>
Cc: linux-fsdevel@...r.kernel.org, Amir Goldstein <amir73il@...il.com>,
Josef Bacik <josef@...icpanda.com>, Jeff Layton <jlayton@...nel.org>, Mike Yuan <me@...dnzj.com>,
Zbigniew Jędrzejewski-Szmek <zbyszek@...waw.pl>, Lennart Poettering <mzxreary@...inter.de>,
Daan De Meyer <daan.j.demeyer@...il.com>, Aleksa Sarai <cyphar@...har.com>,
Alexander Viro <viro@...iv.linux.org.uk>, Tejun Heo <tj@...nel.org>, Johannes Weiner <hannes@...xchg.org>,
Michal Koutný <mkoutny@...e.com>, Jakub Kicinski <kuba@...nel.org>,
Anna-Maria Behnsen <anna-maria@...utronix.de>, Frederic Weisbecker <frederic@...nel.org>,
Thomas Gleixner <tglx@...utronix.de>, cgroups@...r.kernel.org, linux-kernel@...r.kernel.org,
netdev@...r.kernel.org
Subject: Re: [PATCH 7/9] net: centralize ns_common initialization
On Thu, Sep 18, 2025 at 11:42:38AM +0200, Jan Kara wrote:
> On Wed 17-09-25 12:28:06, Christian Brauner wrote:
> > Centralize ns_common initialization.
> >
> > Signed-off-by: Christian Brauner <brauner@...nel.org>
> > ---
> > net/core/net_namespace.c | 23 +++--------------------
> > 1 file changed, 3 insertions(+), 20 deletions(-)
> >
> > diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
> > index a57b3cda8dbc..897f4927df9e 100644
> > --- a/net/core/net_namespace.c
> > +++ b/net/core/net_namespace.c
> > @@ -409,7 +409,7 @@ static __net_init int preinit_net(struct net *net, struct user_namespace *user_n
> > ns_ops = NULL;
> > #endif
> >
> > - ret = ns_common_init(&net->ns, ns_ops, false);
> > + ret = ns_common_init(&net->ns, ns_ops, true);
> > if (ret)
> > return ret;
> >
> > @@ -597,6 +597,7 @@ struct net *copy_net_ns(unsigned long flags,
> > net_passive_dec(net);
> > dec_ucounts:
> > dec_net_namespaces(ucounts);
> > + ns_free_inum(&net->ns);
>
> This looks like a wrong place to put it? dec_ucounts also gets called when we
> failed to create 'net' and thus net == NULL.
>
> > return ERR_PTR(rv);
> > }
> > return net;
> > @@ -718,6 +719,7 @@ static void cleanup_net(struct work_struct *work)
> > #endif
> > put_user_ns(net->user_ns);
> > net_passive_dec(net);
> > + ns_free_inum(&net->ns);
>
> The calling of ns_free_inum() after we've dropped our reference
> (net_passive_dec()) looks suspicious. Given how 'net' freeing works I don't
> think this can lead to actual UAF issues but it is in my opinion a bad
> coding pattern and for no good reason AFAICT.
All good points. I can't say I'm fond of the complexity in this specific
instance in general.
diff --git a/net/core/net_namespace.c b/net/core/net_namespace.c
index 897f4927df9e..9df236811454 100644
--- a/net/core/net_namespace.c
+++ b/net/core/net_namespace.c
@@ -590,6 +590,7 @@ struct net *copy_net_ns(unsigned long flags,
if (rv < 0) {
put_userns:
+ ns_free_inum(&net->ns);
#ifdef CONFIG_KEYS
key_remove_domain(net->key_domain);
#endif
@@ -597,7 +598,6 @@ struct net *copy_net_ns(unsigned long flags,
net_passive_dec(net);
dec_ucounts:
dec_net_namespaces(ucounts);
- ns_free_inum(&net->ns);
return ERR_PTR(rv);
}
return net;
@@ -713,13 +713,13 @@ static void cleanup_net(struct work_struct *work)
/* Finally it is safe to free my network namespace structure */
list_for_each_entry_safe(net, tmp, &net_exit_list, exit_list) {
list_del_init(&net->exit_list);
+ ns_free_inum(&net->ns);
dec_net_namespaces(net->ucounts);
#ifdef CONFIG_KEYS
key_remove_domain(net->key_domain);
#endif
put_user_ns(net->user_ns);
net_passive_dec(net);
- ns_free_inum(&net->ns);
}
WRITE_ONCE(cleanup_net_task, NULL);
}
Powered by blists - more mailing lists