From mboxrd@z Thu Jan  1 00:00:00 1970
From: Jon Maloy <maloy@donjonn.com>
Subject: Re: [PATCH net-next 0/7] tipc: some optimizations
	and impovements
Date: Fri, 13 Mar 2015 13:09:54 -0400
Message-ID: <550319E2.2020108@donjonn.com>
References: <1426256860-31208-1-git-send-email-jon.maloy@ericsson.com>
	<20150313.125227.1523044445570081158.davem@davemloft.net>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
Cc: netdev@vger.kernel.org, tipc-discussion@lists.sourceforge.net,
	paul.gortmaker@windriver.com
To: David Miller <davem@davemloft.net>, jon.maloy@ericsson.com
Return-path: <tipc-discussion-bounces@lists.sourceforge.net>
In-Reply-To: <20150313.125227.1523044445570081158.davem@davemloft.net>
List-Unsubscribe: <https://lists.sourceforge.net/lists/listinfo/tipc-discussion>,
	<mailto:tipc-discussion-request@lists.sourceforge.net?subject=unsubscribe>
List-Archive: <http://sourceforge.net/mailarchive/forum.php?forum_name=tipc-discussion>
List-Post: <mailto:tipc-discussion@lists.sourceforge.net>
List-Help: <mailto:tipc-discussion-request@lists.sourceforge.net?subject=help>
List-Subscribe: <https://lists.sourceforge.net/lists/listinfo/tipc-discussion>,
	<mailto:tipc-discussion-request@lists.sourceforge.net?subject=subscribe>
Errors-To: tipc-discussion-bounces@lists.sourceforge.net
List-Id: netdev.vger.kernel.org


On 15-03-13 12:52 PM, David Miller wrote:
> From: Jon Maloy <jon.maloy@ericsson.com>
> Date: Fri, 13 Mar 2015 10:27:33 -0400
>
>> The commits in this series contain some relatively simple changes
>> that lead to better throughput across TIPC connections. We also make
>> changes to the implementation of link transmission queueing and priority
>> handling, in order to make the code more comprehensible and maintainable.
> I think you should do the SKB linearization changes properly.
>
> Simply do pskb_may_pull() before you inspect any header, and you can
> do this everwhere, even your bundling extraction code path.
Ok, I'll try that.

>
> I also don't think you even need to clone the SKB there either.

I don't don't see how we can not. We may extract many
buffers from different positions in the data area, which will
be sent in different directions, and be freed at different
moments of time.

///jon
>
> Thanks.


------------------------------------------------------------------------------
Dive into the World of Parallel Programming The Go Parallel Website, sponsored
by Intel and developed in partnership with Slashdot Media, is your hub for all
things parallel software development, from weekly thought leadership blogs to
news, videos, case studies, tutorials and more. Take a look and join the 
conversation now. http://goparallel.sourceforge.net/