From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: [PATCH v6 2/4] lib: add bitrate statistics library Date: Wed, 11 Jan 2017 08:15:02 -0800 Message-ID: <20170111081502.220f6824@xeon-e3> References: <1484150594-3758-1-git-send-email-remy.horton@intel.com> <1484150594-3758-3-git-send-email-remy.horton@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: dev@dpdk.org, Thomas Monjalon To: Remy Horton Return-path: Received: from mail-pg0-f50.google.com (mail-pg0-f50.google.com [74.125.83.50]) by dpdk.org (Postfix) with ESMTP id C9F92D592 for ; Wed, 11 Jan 2017 17:15:12 +0100 (CET) Received: by mail-pg0-f50.google.com with SMTP id 204so29346525pge.0 for ; Wed, 11 Jan 2017 08:15:12 -0800 (PST) In-Reply-To: <1484150594-3758-3-git-send-email-remy.horton@intel.com> List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On Thu, 12 Jan 2017 00:03:12 +0800 Remy Horton wrote: > + > +struct rte_stats_bitrates_s * > +rte_stats_bitrate_create(void) > +{ > + return rte_zmalloc(NULL, sizeof(struct rte_stats_bitrates_s), 0); > +} > + You want these statistics to be cache aligned (ie not 0) to avoid false cache sharing. Since these statistics are often associated with a device would it not make sense to have a NUMA socket associated with the alllocation?