From mboxrd@z Thu Jan 1 00:00:00 1970 From: Vlad Zolotarov Subject: Re: RFC: i40e xmit path HW limitation Date: Thu, 30 Jul 2015 19:50:27 +0300 Message-ID: <55BA55D3.2070105@cloudius-systems.com> References: <55BA3B5D.4020402@cloudius-systems.com> <20150730091753.1af6cc67@urahara> <55BA4EC6.3030301@cloudius-systems.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: quoted-printable Cc: "dev@dpdk.org" To: Avi Kivity , Stephen Hemminger Return-path: Received: from mail-wi0-f179.google.com (mail-wi0-f179.google.com [209.85.212.179]) by dpdk.org (Postfix) with ESMTP id 4ABBFC688 for ; Thu, 30 Jul 2015 18:50:29 +0200 (CEST) Received: by wibxm9 with SMTP id xm9so76720028wib.1 for ; Thu, 30 Jul 2015 09:50:29 -0700 (PDT) In-Reply-To: <55BA4EC6.3030301@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 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 fro= m >>> 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 giv= e >>> 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 woul= d >>> do). Number of elements in the pool should be as Tx ring length >>> multiplied by "64KB/(linear data length of the buffer in the poo= l >>> above)". Here I use 64KB as a maximum packet length and not taki= ng >>> into an account esoteric things like "Giant" TSO mentioned in th= e >>> spec above. Then we may actually go and linearize the cluster if >>> needed on top of the buffers from the pool above, post the buffe= r >>> 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 th= e >>> 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. +1 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. ;) > >