From mboxrd@z Thu Jan  1 00:00:00 1970
From: =?UTF-8?Q?Mattias_R=c3=b6nnblom?= <mattias.ronnblom@ericsson.com>
Subject: Re: DSW eventdev is getting "stuck"?
Date: Wed, 3 Apr 2019 20:33:58 +0200
Message-ID: <8510f66d-a63e-4ef9-44e1-071758fec6fd@ericsson.com>
References: <3353B035-E890-4753-8863-5647DFE61E23@paloaltonetworks.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=utf-8; format=flowed
Content-Transfer-Encoding: 8bit
To: Venky Venkatesh <vvenkatesh@paloaltonetworks.com>,
 "dev@dpdk.org" <dev@dpdk.org>
Return-path: <dev-bounces@dpdk.org>
Received: from mail.lysator.liu.se (mail.lysator.liu.se [130.236.254.3])
 by dpdk.org (Postfix) with ESMTP id 4A6B15398
 for <dev@dpdk.org>; Wed,  3 Apr 2019 20:34:00 +0200 (CEST)
Received: from mail.lysator.liu.se (localhost [127.0.0.1])
 by mail.lysator.liu.se (Postfix) with ESMTP id ABA2B40007
 for <dev@dpdk.org>; Wed,  3 Apr 2019 20:33:59 +0200 (CEST)
In-Reply-To: <3353B035-E890-4753-8863-5647DFE61E23@paloaltonetworks.com>
Content-Language: en-US
List-Id: DPDK patches and discussions <dev.dpdk.org>
List-Unsubscribe: <https://mails.dpdk.org/options/dev>,
 <mailto:dev-request@dpdk.org?subject=unsubscribe>
List-Archive: <http://mails.dpdk.org/archives/dev/>
List-Post: <mailto:dev@dpdk.org>
List-Help: <mailto:dev-request@dpdk.org?subject=help>
List-Subscribe: <https://mails.dpdk.org/listinfo/dev>,
 <mailto:dev-request@dpdk.org?subject=subscribe>
Errors-To: dev-bounces@dpdk.org
Sender: "dev" <dev-bounces@dpdk.org>

On 2019-04-03 20:17, Venky Venkatesh wrote:
> Hi,
> I am using the DSW code from 18.11 with the default settings for all the #defines. Here are some more details:
> I have an 8 port system with 1 queue.
> All ports can inject events. Port 0 and 7 inject events rarely.
> Ports 1-6 are linked to the queue and hence dequeue events.
> 
> I see that in steady state the total number enqueues into the system is much greater than the total number of dequeues.
> eventdev deq#: 1948491 enq#: 1949007
> 

If you have more enqueues than dequeues, it means there are events in 
flight in the scheduler. In the above case, it's roughly 516 events, 
which sounds perfectly healthy and normal. I say roughly, because 
there's no way to take a consistent, global snapshot of all xstats counters.

You will see a higher value than this on dev_credits_on_loan, because 
the ports will borrow (and return) credits in chunks of 64 to the shared 
pool, for efficiently reasons.

> However when I dump some of the data structure counters it doesn’t appear to be stuck. What may be going on?

I'm guessing it doesn't appear stuck because it's not stuck.

/M