From mboxrd@z Thu Jan  1 00:00:00 1970
From: "Liang, Cunming" <cunming.liang@intel.com>
Subject: Re: [PATCH v13 02/14] eal/linux: add rte_epoll_wait/ctl
	support
Date: Fri, 17 Jul 2015 05:47:11 +0000
Message-ID: <D0158A423229094DA7ABF71CF2FA0DA31198BD68@shsmsx102.ccr.corp.intel.com>
References: <1433741351-27005-1-git-send-email-cunming.liang@intel.com>
 <1434686442-578-1-git-send-email-cunming.liang@intel.com>
 <1434686442-578-3-git-send-email-cunming.liang@intel.com>
 <3479933.5lm7YXT529@xps13>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Cc: "shemming@brocade.com" <shemming@brocade.com>,
 "dev@dpdk.org" <dev@dpdk.org>, "Wang,
 Liang-min" <liang-min.wang@intel.com>
To: Thomas Monjalon <thomas.monjalon@6wind.com>
Return-path: <dev-bounces@dpdk.org>
Received: from mga03.intel.com (mga03.intel.com [134.134.136.65])
 by dpdk.org (Postfix) with ESMTP id 7D3F75921
 for <dev@dpdk.org>; Fri, 17 Jul 2015 07:48:25 +0200 (CEST)
In-Reply-To: <3479933.5lm7YXT529@xps13>
Content-Language: en-US
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>



> -----Original Message-----
> From: Thomas Monjalon [mailto:thomas.monjalon@6wind.com]
> Sent: Tuesday, July 14, 2015 12:56 AM
> To: Liang, Cunming
> Cc: dev@dpdk.org; shemming@brocade.com; david.marchand@6wind.com;
> Zhou, Danny; Wang, Liang-min; Richardson, Bruce; Liu, Yong;
> nhorman@tuxdriver.com
> Subject: Re: [PATCH v13 02/14] eal/linux: add rte_epoll_wait/ctl support
>=20
> 2015-06-19 12:00, Cunming Liang:
> > +int
> > +rte_epoll_wait(int epfd, struct rte_epoll_event *events,
> > +	       int maxevents, int timeout)
> > +{
> > +	struct epoll_event evs[maxevents];
> > +	int rc;
> > +
> > +	if (!events) {
> > +		RTE_LOG(ERR, EAL, "rte_epoll_event can't be NULL\n");
> > +		return -1;
> > +	}
> > +
> > +	/* using per thread epoll fd */
> > +	if (epfd =3D=3D RTE_EPOLL_PER_THREAD)
> > +		epfd =3D rte_intr_tls_epfd();
> > +
> > +	while (1) {
> > +		rc =3D epoll_wait(epfd, evs, maxevents, timeout);
> > +		if (likely(rc > 0)) {
> > +			/* epoll_wait has at least one fd ready to read */
> > +			rc =3D eal_epoll_process_event(evs, rc, events);
> > +			break;
> > +		} else if (rc < 0) {
> > +			if (errno =3D=3D EINTR)
> > +				continue;
> > +			/* epoll_wait fail */
> > +			RTE_LOG(ERR, EAL, "epoll_wait returns with fail %s\n",
> > +				strerror(errno));
> > +			rc =3D -1;
> > +			break;
> > +		}
> > +	}
> > +
> > +	return rc;
> > +}
>=20
> In general, such loop is application-level.
> What is the added value of rte_epoll_wait()?
> Do we need some wrappers to libc in DPDK?
Some motivations to do it,
1) 'epoll_event' takes either fd or a data point. However we require more t=
o cover both rx interrupt and other user's events.
2) Some errno processing can be addressed commonly, it's helpful to focus o=
n the real event we're interested in.
3) Usually there's one epoll instance per lcore to serve the events. Here g=
ives a default one if it isn't assigned.