netdev - Re: [PATCH net-next v2 01/13] net: sched: support skb chain ext in tc classification path

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] [day] [month] [year] [list]

Date:   Fri, 24 Jan 2020 08:46:22 +0000
From:   Paul Blakey <paulb@...lanox.com>
To:     Jakub Kicinski <kuba@...nel.org>
CC:     Saeed Mahameed <saeedm@...lanox.com>,
        Oz Shlomo <ozsh@...lanox.com>,
        Jakub Kicinski <jakub.kicinski@...ronome.com>,
        Vlad Buslov <vladbu@...lanox.com>,
        David Miller <davem@...emloft.net>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        Jiri Pirko <jiri@...nulli.us>, Roi Dayan <roid@...lanox.com>
Subject: Re: [PATCH net-next v2 01/13] net: sched: support skb chain ext in tc
 classification path


On 22/01/2020 17:29, Jakub Kicinski wrote:
> On Wed, 22 Jan 2020 15:52:46 +0200, Paul Blakey wrote:
>> +int tcf_classify_ingress(struct sk_buff *skb,
>> +			 const struct tcf_block *ingress_block,
>> +			 const struct tcf_proto *tp, struct tcf_result *res,
>> +			 bool compat_mode)
>> +{
>> +	const struct tcf_proto *orig_tp = tp;
>> +
>> +#if IS_ENABLED(CONFIG_NET_TC_SKB_EXT)
>> +	{
>> +		struct tc_skb_ext *ext = skb_ext_find(skb, TC_SKB_EXT);
>> +
>> +		if (ext && ext->chain && ingress_block) {
>> +			struct tcf_chain *fchain;
>> +
>> +			fchain = tcf_chain_lookup_rcu(ingress_block,
>> +						      ext->chain);
>> +			if (!fchain)
>> +				return TC_ACT_UNSPEC;
>> +
>> +			tp = rcu_dereference_bh(fchain->filter_chain);
>> +		}
> Doesn't this skb ext have to be somehow "consumed" by the first lookup?
> What if the skb finds its way to an ingress of another device?
[Resending this reply again, as original reply was rejected]

Good point, This also applies to reclassify.
Simply consuming the skb extension will not handle the tc miss path to ovs.

I will post a v3 where the skb extension will be initialized with the last processed chain.

>
>> +	}
>> +#endif
>> +
>> +	return tcf_classify(skb, tp, orig_tp, res, compat_mode);
>> +}
>> +EXPORT_SYMBOL(tcf_classify_ingress);