netdev - Re: [PATCH 2/2] mac80211_hwsim: Allow managing radios from non-initial namespaces

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 03 May 2016 21:16:32 +0200
From:	Johannes Berg <johannes@...solutions.net>
To:	Martin Willi <martin@...ongswan.org>
Cc:	linux-wireless@...r.kernel.org, netdev@...r.kernel.org
Subject: Re: [PATCH 2/2] mac80211_hwsim: Allow managing radios from
 non-initial namespaces

On Tue, 2016-05-03 at 08:53 +0200, Martin Willi wrote:
> 
> +static __net_init int hwsim_init_net(struct net *net)
> +{
> +	struct mac80211_hwsim_data *data;
> +	bool exists = true;
> +	int netgroup = 0;
> +
> +	spin_lock_bh(&hwsim_radio_lock);
> +	while (exists) {
> +		exists = false;
> +		list_for_each_entry(data, &hwsim_radios, list) {
> +			if (netgroup == data->netgroup) {
> +				exists = true;
> +				netgroup++;
> +				break;
> +			}
> +		}
> +	}
> +	spin_unlock_bh(&hwsim_radio_lock);
> +
> +	*(int *)net_generic(net, hwsim_net_id) = netgroup;

This seems somewhat awkward. Why not just take the maximum of all the
netgroup IDs + 1? We'd run out of memory and radio IDs long before
netgroup IDs even that way, and they're not actually visible anywhere
so it doesn't matter.

Actually though, *both* your approach and my suggestion don't seem
safe: consider a new netns that doesn't have any hwsim radios yet. Now
you create *another* one, but it would get the same netgroup.

IOW, you should simply use a global counter. Surprising (net)
namespaces don't have an index like that already, but I don't see one.

> +static void __net_exit hwsim_exit_net(struct net *net)
> +{
> +	struct mac80211_hwsim_data *entry, *tmp;
> +
> +	spin_lock_bh(&hwsim_radio_lock);
> +	list_for_each_entry_safe(entry, tmp, &hwsim_radios, list) {
> +		if (net_eq(wiphy_net(entry->hw->wiphy), net)) {
> +			list_del(&entry->list);
> +			INIT_WORK(&entry->destroy_work,
> destroy_radio);
> +			schedule_work(&entry->destroy_work);
> +		}
> +	}
> +	spin_unlock_bh(&hwsim_radio_lock);
> +}

This changes today's default behaviour of moving the wiphys to the
default namespace. Did you intend to destroy them based on the
netgroup, i.e. based on the namespace that created them? Actually,
maybe they should move back to the namespace that created them, if the
namespace they are in is destroyed? But that's difficult, I don't mind
this behaviour, but I'm not sure it's what we want by default for
radios created in the init_net.

johannes