From mboxrd@z Thu Jan  1 00:00:00 1970
From: Olivier MATZ <olivier.matz@6wind.com>
Subject: Re: [PATCH v3] mempool: reduce rte_mempool structure size
Date: Mon, 15 Feb 2016 11:15:22 +0100
Message-ID: <56C1A53A.8010607@6wind.com>
References: <1455039006-86816-1-git-send-email-keith.wiles@intel.com>
 <56BDEA08.9040402@redhat.com>
 <09D5A01F-7205-49E8-9A27-95161235963E@intel.com> <10558998.3znIRhOpQL@xps13>
 <56BDFF2B.5090401@6wind.com> <56C1A15C.8060603@intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=windows-1252
Content-Transfer-Encoding: 7bit
Cc: dev@dpdk.org
To: "Hunt, David" <david.hunt@intel.com>,
 Thomas Monjalon <thomas.monjalon@6wind.com>,
 "Wiles, Keith" <keith.wiles@intel.com>
Return-path: <dev-bounces@dpdk.org>
Received: from mail.droids-corp.org (zoll.droids-corp.org [94.23.50.67])
 by dpdk.org (Postfix) with ESMTP id CA7025938
 for <dev@dpdk.org>; Mon, 15 Feb 2016 11:15:30 +0100 (CET)
In-Reply-To: <56C1A15C.8060603@intel.com>
List-Id: patches and discussions about DPDK <dev.dpdk.org>
List-Unsubscribe: <http://dpdk.org/ml/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://dpdk.org/ml/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <http://dpdk.org/ml/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>

Hi David,

On 02/15/2016 10:58 AM, Hunt, David wrote:
> On 12/02/2016 15:50, Olivier MATZ wrote:
>> - NEXT_ABI does make the code harder to read in this case, and I'm
>>    thinking about the patchset from David Hunt (external mempool handler)
>>    that will be in the same situation, and maybe also another patchset
>>    I'm working on.
> 
> Olivier,
>     I'm working on that at the moment with the external mempool handler
> code. However, it crossed my mind that we have a choice to use symbol
> versioning OR use NEXT_ABI. Would one method be preferred over the other?

I think symbol versioning should always be preferred when possible.

In your case, as far as I remember, your are updating the rte_mempool
structure, which is accessed by static inline functions. I don't think
it is easily manageable with symbol versioning. Moreover, the ABI will
already be broken by Keith's patch, so I think it's less problematic
to have other patches breaking the ABI at the same time.

Regards,
Olivier