From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Burakov, Anatoly" Subject: Re: [PATCH 31/41] ethdev: use contiguous allocation for DMA memory Date: Mon, 5 Mar 2018 10:00:41 +0000 Message-ID: References: <405c4b97-4561-6ef7-5844-fdc0cdfaabc5@solarflare.com> <28a175c1-b306-ec7b-e69b-003b40968d2e@intel.com> <62f57c50-3e5a-7202-7a40-6d0e66c68b43@solarflare.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: 7bit Cc: Thomas Monjalon , keith.wiles@intel.com, jianfeng.tan@intel.com, andras.kovacs@ericsson.com, laszlo.vadkeri@ericsson.com, benjamin.walker@intel.com, bruce.richardson@intel.com, konstantin.ananyev@intel.com, kuralamudhan.ramakrishnan@intel.com, louise.m.daly@intel.com, nelio.laranjeiro@6wind.com, yskoh@mellanox.com, pepperjo@japf.ch, jerin.jacob@caviumnetworks.com, hemant.agrawal@nxp.com, olivier.matz@6wind.com To: Andrew Rybchenko , dev@dpdk.org Return-path: Received: from mga04.intel.com (mga04.intel.com [192.55.52.120]) by dpdk.org (Postfix) with ESMTP id BB11D5F17 for ; Mon, 5 Mar 2018 11:00:49 +0100 (CET) In-Reply-To: <62f57c50-3e5a-7202-7a40-6d0e66c68b43@solarflare.com> Content-Language: en-US List-Id: DPDK patches and discussions List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: dev-bounces@dpdk.org Sender: "dev" On 05-Mar-18 9:15 AM, Andrew Rybchenko wrote: > On 03/05/2018 12:08 PM, Burakov, Anatoly wrote: >> On 03-Mar-18 2:05 PM, Andrew Rybchenko wrote: >>> On 03/03/2018 04:46 PM, Anatoly Burakov wrote: >>>> This fixes the following drivers in one go: >>> >>> Does it mean that these drivers are broken in the middle of patch set >>> and fixed now? >>> If so, it would be good to avoid it. It breaks bisect. >>> >> >> Depends on the definition of "broken". Legacy memory mode will still >> work for all drivers throughout the patchset. As for new memory mode, >> yes, it will be "broken in the middle of the patchset", but due to the >> fact that there's enormous amount of code to review between fbarray >> changes, malloc changes, contiguous allocation changes and adding new >> rte_memzone API's, i favored ease of code review over bisect. >> >> I can of course reorder and roll up several different patchset and all >> driver updates into one giant patch, but do you really want to be the >> one reviewing such a patch? > > Is it possible to: > 1. Introduce _contig function > 2. Switch users of the contiguous allocation to it as you do now > 3. Make the old function to allocate possibly non-contiguous memory > Good point. I'll see if i can shuffle patches around for v2. Thanks! -- Thanks, Anatoly