From mboxrd@z Thu Jan  1 00:00:00 1970
From: Thomas Monjalon <thomas@monjalon.net>
Subject: Re: [PATCH] net/mlx4: fix last Tx wqe stamping lack
Date: Sat, 11 Nov 2017 15:53:07 +0100
Message-ID: <11709391.GH4dFWTqEr@xps>
References: <1510302438-29138-1-git-send-email-matan@mellanox.com>
 <20171110150241.GI24849@6wind.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7Bit
Cc: dev@dpdk.org, Ophir Munk <ophirmu@mellanox.com>,
 Moti Haimovsky <motih@mellanox.com>, Olga Shern <olgas@mellanox.com>
To: Adrien Mazarguil <adrien.mazarguil@6wind.com>,
 Matan Azrad <matan@mellanox.com>
Return-path: <dev-bounces@dpdk.org>
Received: from out4-smtp.messagingengine.com (out4-smtp.messagingengine.com
 [66.111.4.28]) by dpdk.org (Postfix) with ESMTP id DCBEB1B1BC
 for <dev@dpdk.org>; Sat, 11 Nov 2017 15:53:09 +0100 (CET)
In-Reply-To: <20171110150241.GI24849@6wind.com>
List-Id: DPDK patches and discussions <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>

10/11/2017 16:02, Adrien Mazarguil:
> On Fri, Nov 10, 2017 at 08:27:18AM +0000, Matan Azrad wrote:
> > When Tx pakcet HW processing is done, SW should stamp all the completion
> > burst WQEs.
> > 
> > Stamp missed last completion burst WQE.
> > 
> > Fixes: c3c977bbecbd ("net/mlx4: add Tx bypassing Verbs")
> > 
> > Signed-off-by: Matan Azrad <matan@mellanox.com>
> 
> This reads like you were in a hurry :)
> 
> Took me a while to understand the problem and how you addressed it. So in
> short, wqe_index is consumed but its TXBBs aren't stamped because the loop
> stops at its index without processing it.
> 
> Patch looks good but could have been simpler by directly initializing
> nr_txbbs to sq->tail, not use sq->tail as an offset afterward and get rid of
> sq_tail. It's OK as this wouldn't have resulted in a smaller patch anyway.
> 
> Commit log rewording suggestion:
> 
>  net/mlx4: fix missing stamp during Tx completion
> 
>  After processing completed packets, the owner bit of each TXBB comprised
>  in its WQEs must be invalidated. The loop stops short of processing the
>  last WQE.
> 
> Other than that,
> 
> Acked-by: Adrien Mazarguil <adrien.mazarguil@6wind.com>

Applied with suggested commit log, thanks