From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Hemminger Subject: Re: RFC: i40e xmit path HW limitation Date: Thu, 30 Jul 2015 10:01:58 -0700 Message-ID: <20150730100158.1516dab3@urahara> References: <55BA3B5D.4020402@cloudius-systems.com> <20150730091753.1af6cc67@urahara> <55BA4EC6.3030301@cloudius-systems.com> <55BA55D3.2070105@cloudius-systems.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: quoted-printable Cc: "dev@dpdk.org" To: Vlad Zolotarov Return-path: Received: from mail-pa0-f47.google.com (mail-pa0-f47.google.com [209.85.220.47]) by dpdk.org (Postfix) with ESMTP id 3E4B0C690 for ; Thu, 30 Jul 2015 19:01:51 +0200 (CEST) Received: by pabkd10 with SMTP id kd10so26250224pab.2 for ; Thu, 30 Jul 2015 10:01:50 -0700 (PDT) In-Reply-To: <55BA55D3.2070105@cloudius-systems.com> List-Id: patches and discussions about DPDK List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On Thu, 30 Jul 2015 19:50:27 +0300 Vlad Zolotarov wrote: >=20 >=20 > On 07/30/15 19:20, Avi Kivity wrote: > > > > > > On 07/30/2015 07:17 PM, Stephen Hemminger wrote: > >> On Thu, 30 Jul 2015 17:57:33 +0300 > >> Vlad Zolotarov wrote: > >> > >>> Hi, Konstantin, Helin, > >>> there is a documented limitation of xl710 controllers (i40e driver) > >>> which is not handled in any way by a DPDK driver. > >>> From the datasheet chapter 8.4.1: > >>> > >>> "=E2=80=A2 A single transmit packet may span up to 8 buffers (up to 8= data=20 > >>> descriptors per packet including > >>> both the header and payload buffers). > >>> =E2=80=A2 The total number of data descriptors for the whole TSO (exp= lained=20 > >>> later on in this chapter) is > >>> unlimited as long as each segment within the TSO obeys the previous=20 > >>> rule (up to 8 data descriptors > >>> per segment for both the TSO header and the segment payload buffers)." > >>> > >>> This means that, for instance, long cluster with small fragments has = to > >>> be linearized before it may be placed on the HW ring. > >>> In more standard environments like Linux or FreeBSD drivers the=20 > >>> solution > >>> is straight forward - call skb_linearize()/m_collapse() corresponding. > >>> In the non-conformist environment like DPDK life is not that easy - > >>> there is no easy way to collapse the cluster into a linear buffer from > >>> inside the device driver > >>> since device driver doesn't allocate memory in a fast path and utiliz= es > >>> the user allocated pools only. > >>> > >>> Here are two proposals for a solution: > >>> > >>> 1. We may provide a callback that would return a user TRUE if a give > >>> cluster has to be linearized and it should always be called befo= re > >>> rte_eth_tx_burst(). Alternatively it may be called from inside t= he > >>> rte_eth_tx_burst() and rte_eth_tx_burst() is changed to return=20 > >>> some > >>> error code for a case when one of the clusters it's given has=20 > >>> to be > >>> linearized. > >>> 2. Another option is to allocate a mempool in the driver with the > >>> elements consuming a single page each (standard 2KB buffers would > >>> do). Number of elements in the pool should be as Tx ring length > >>> multiplied by "64KB/(linear data length of the buffer in the pool > >>> above)". Here I use 64KB as a maximum packet length and not taki= ng > >>> into an account esoteric things like "Giant" TSO mentioned in the > >>> spec above. Then we may actually go and linearize the cluster if > >>> needed on top of the buffers from the pool above, post the buffer > >>> from the mempool above on the HW ring, link the original=20 > >>> cluster to > >>> that new cluster (using the private data) and release it when the > >>> send is done. > >> Or just silently drop heavily scattered packets (and increment oerrors) > >> with a PMD_TX_LOG debug message. > >> > >> I think a DPDK driver doesn't have to accept all possible mbufs and do > >> extra work. It seems reasonable to expect caller to be well behaved > >> in this restricted ecosystem. > >> > > > > How can the caller know what's well behaved? It's device dependent. >=20 > +1 >=20 > Stephen, how do you imagine this well-behaved application? Having switch= =20 > case by an underlying device type and then "well-behaving" correspondingl= y? > Not to mention that to "well-behave" the application writer has to read=20 > HW specs and understand them, which would limit the amount of DPDK=20 > developers to a very small amount of people... ;) Not to mention that=20 > the mentioned above switch-case would be a super ugly thing to be found=20 > in an application that would raise a big question about the=20 > justification of a DPDK existence as as SDK providing device drivers=20 > interface. ;) Either have a RTE_MAX_MBUF_SEGMENTS that is global or a mbuf_linearize function? Driver already can stash the mbuf pool used for Rx and reuse it for the transient Tx buffers.