linux-kernel - Re: [PATCH net-net] tun: Assign missing bpf_net

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]

Message-ID: <20240912-polite-wooden-jellyfish-acae2f@leitao>
Date: Thu, 12 Sep 2024 07:40:52 -0700
From: Breno Leitao <leitao@...ian.org>
To: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
Cc: Vadim Fedorenko <vadim.fedorenko@...ux.dev>,
	Jakub Kicinski <kuba@...nel.org>, andrii@...nel.org, ast@...nel.org,
	syzbot <syzbot+08811615f0e17bc6708b@...kaller.appspotmail.com>,
	bpf@...r.kernel.org, daniel@...earbox.net, davem@...emloft.net,
	eddyz87@...il.com, haoluo@...gle.com, hawk@...nel.org,
	john.fastabend@...il.com, jolsa@...nel.org, kpsingh@...nel.org,
	linux-kernel@...r.kernel.org, martin.lau@...ux.dev,
	netdev@...r.kernel.org, sdf@...ichev.me, song@...nel.org,
	syzkaller-bugs@...glegroups.com, yonghong.song@...ux.dev
Subject: Re: [PATCH net-net] tun: Assign missing bpf_net_context.

On Thu, Sep 12, 2024 at 04:30:29PM +0200, Sebastian Andrzej Siewior wrote:
> On 2024-09-12 07:19:54 [-0700], Breno Leitao wrote:
> > Hello Vadim,
> > 
> > On Thu, Sep 12, 2024 at 02:32:55PM +0100, Vadim Fedorenko wrote:
> > > On 12/09/2024 14:17, Breno Leitao wrote:
> > > > @@ -72,6 +73,7 @@ static netdev_tx_t netkit_xmit(struct sk_buff *skb, struct net_device *dev)
> > > >   	struct net_device *peer;
> > > >   	int len = skb->len;
> > > > +	bpf_net_ctx = bpf_net_ctx_set(&__bpf_net_ctx);
> > > >   	rcu_read_lock();
> > > 
> > > Hi Breno,
> > > 
> > > looks like bpf_net_ctx should be set under rcu read lock...
> > 
> > Why exactly?
> > 
> > I saw in some examples where bpf_net_ctx_set() was set inside the
> > rcu_read_lock(), but, I was not able to come up with justification to do
> > the same. Would you mind elaborating why this might be needed inside the
> > lock?
> 
> It might have been done due to simpler nesting or other reasons but
> there is no requirement to do this under RCU protection. The assignment
> and cleanup is always performed task-local.

Thanks. I will keep it out of the RCU lock then, as in the patch above.

--breno