netdev - Re: [PATCH net-next v8 5/6] openvswitch: Sample action without side effects

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 for Android: free password hash cracker in your pocket

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

Date:	Wed, 23 Jul 2014 18:02:17 -0700
From:	Pravin Shelar <pshelar@...ira.com>
To:	David Miller <davem@...emloft.net>
Cc:	netdev <netdev@...r.kernel.org>, Simon Horman <horms@...ge.net.au>,
	Jesse Gross <jesse@...ira.com>
Subject: Re: [PATCH net-next v8 5/6] openvswitch: Sample action without side effects

On Wed, Jul 23, 2014 at 2:47 PM, David Miller <davem@...emloft.net> wrote:
> From: Pravin B Shelar <pshelar@...ira.com>
> Date: Tue, 22 Jul 2014 11:35:16 -0700
>
>> * As sample() provides a cloned skb in the unlikely case where there are
>>   nested actions other than a single userspace action it is no longer
>>   necessary to clone the skb in do_execute_actions() when executing a
>>   recirculation action just because the keep_skb parameter is set:
>
> This commit log message talks about recirculation, which has been removed
> from this series.
>
ok, I will update msg.

> And in fact, I think, for my suggested implementation of recirculation
> by just returning and re-running the stack, this whole skb clone issue
> either goes away or needs to happen in a different place.
>
> If you recirculate by just returning and re-processing the packet at
> the top-level entry point into openvswitch, then that's where you
> would clone.

This patch is about sample action.
sample action needs skb_clone() in some case. OVS sample action has
list of actions which can be one or more output actions or another
(nested) sample action. A port output consumes skb, therefore
depending on action list we need to clone skb before it is sent to a
port. This patch try to optimize this as you have pointed out. It try
to avoid skb_clone() in case where sample action is last action.
same applies to recirculation action.
--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html