From mboxrd@z Thu Jan  1 00:00:00 1970
From: Remy Horton <remy.horton@intel.com>
Subject: Re: [PATCH v12 1/6] lib: add information metrics library
Date: Tue, 28 Mar 2017 16:00:32 +0100
Message-ID: <aafe450e-1422-b998-a27d-9d7eca4134eb@intel.com>
References: <1490646097-27803-1-git-send-email-remy.horton@intel.com>
 <1490646097-27803-2-git-send-email-remy.horton@intel.com>
 <3AEA2BF9852C6F48A459DA490692831F0112D6FE@irsmsx110.ger.corp.intel.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=windows-1252; format=flowed
Content-Transfer-Encoding: 7bit
Cc: Thomas Monjalon <thomas.monjalon@6wind.com>
To: "Pattan, Reshma" <reshma.pattan@intel.com>, "dev@dpdk.org" <dev@dpdk.org>
Return-path: <dev-bounces@dpdk.org>
Received: from mga06.intel.com (mga06.intel.com [134.134.136.31])
 by dpdk.org (Postfix) with ESMTP id 6ED0ECF8A
 for <dev@dpdk.org>; Tue, 28 Mar 2017 17:00:36 +0200 (CEST)
In-Reply-To: <3AEA2BF9852C6F48A459DA490692831F0112D6FE@irsmsx110.ger.corp.intel.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>


On 28/03/2017 15:14, Pattan, Reshma wrote:
[..]
>> +
>> +include $(RTE_SDK)/mk/rte.vars.mk
>> +
>> +# library name
>> +LIB = librte_metrics.a
>> +
>> +CFLAGS += $(WERROR_FLAGS) -I$(SRCDIR)
>
> How about adding -O3 to flags?

Off the top of my mind, I think such flags are added to CFLAGS 
separately. I think optimisation flags probably shouldn't be set at 
third-level makefiles such as this one.


>> +int
>> +rte_metrics_get_names(struct rte_metric_name *names,
>> +	uint16_t capacity)
>> +{
[..]
>> +	memzone = rte_memzone_lookup(RTE_METRICS_MEMZONE_NAME);
>> +	/* If not allocated, fail silently */
>> +	if (memzone == NULL)
>> +		return 0;
>> +
>> +	stats = memzone->addr;
>> +	rte_spinlock_lock(&stats->lock);
>> +	if (names != NULL) {
>
> You can check names == NULL at memzone check and return instead of doing that inside lock?

Since the names==NULL case still requires accessing of the protected 
memory to get the name count, don't see any gain in doing so.


>> +
>> +int
>> +rte_metrics_get_values(int port_id,
>> +	struct rte_metric_value *values,
>> +	uint16_t capacity)
>> +{
[..]
>> +	if (values != NULL) {
>
> Can we have values == NULL check  along with memzone and return? Then remove this if check here?

Same as rte_metrics_get_names() above.

Other changes noted. Will send out v12 either tonight or tomorrow.