From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hemant Agrawal Subject: Re: [PATCHv6 00/33] NXP DPAA2 PMD Date: Thu, 26 Jan 2017 17:48:50 +0530 Message-ID: <99c7a139-6098-b7ce-25a1-ee3efb045ff0@nxp.com> References: <1484832240-2048-1-git-send-email-hemant.agrawal@nxp.com> <1485172803-17288-1-git-send-email-hemant.agrawal@nxp.com> <5a72bfd0-7120-4486-9246-03eb3c5c6bf1@intel.com> <67f91414-0439-952d-e07f-23556db110d9@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8"; format=flowed Content-Transfer-Encoding: 7bit Cc: , , , , To: Ferruh Yigit , Return-path: Received: from NAM02-CY1-obe.outbound.protection.outlook.com (mail-cys01nam02on0056.outbound.protection.outlook.com [104.47.37.56]) by dpdk.org (Postfix) with ESMTP id 02CAADE5 for ; Thu, 26 Jan 2017 13:19:00 +0100 (CET) In-Reply-To: <67f91414-0439-952d-e07f-23556db110d9@intel.com> 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 1/26/2017 5:25 PM, Ferruh Yigit wrote: > On 1/23/2017 5:56 PM, Ferruh Yigit wrote: >> On 1/23/2017 11:59 AM, Hemant Agrawal wrote: >> <...> >> >>> >>> Hemant Agrawal (33): >>> mk/dpaa2: add the crc support to the machine type >>> drivers/common/dpaa2: adding qbman driver >>> bus/fslmc: introducing fsl-mc bus driver >>> bus/fslmc: introduce mc object functions >>> bus/fslmc: add mc dpni object support >>> bus/fslmc: add mc dpio object support >>> bus/fslmc: add mc dpbp object support >>> bus/fslmc: add mc dpseci object support >>> eal/vfio: adding vfio utility functions in map file >>> bus/fslmc: add vfio support >>> bus/fslmc: scan for net and sec devices >>> net/dpaa2: introducing NXP dpaa2 pmd driver >>> doc: add dpaa2 nic details >>> bus/fslmc: add debug log message support >>> drivers/common/dpaa2: dpio portal driver >>> drivers/pool/dpaa2: adding hw offloaded mempool >>> drivers/common/dpaa2: dpio routine to affine to crypto threads >>> net/dpaa2: adding eth ops to dpaa2 >>> net/dpaa2: add rss flow distribution >>> net/dpaa2: configure mac address at init >>> net/dpaa2: attach the buffer pool to dpni >>> net/dpaa2: add support for l3 and l4 checksum offload >>> net/dpaa2: add support for promiscuous mode >>> net/dpaa2: add mtu config support >>> net/dpaa2: add packet rx and tx support >>> net/dpaa2: rx packet parsing and packet type support >>> net/dpaa2: link status update >>> net/dpaa2: basic stats support >>> net/dpaa2: enable stashing for LS2088A devices >>> net/dpaa2: add support for non hw buffer pool packet transmit >>> net/dpaa2: enabling the use of physical addresses >>> bus/fslmc: add support for dmamap to ARM SMMU >>> drivers/common/dpaa2: frame queue based dq storage alloc >>> >> <...> >>> 66 files changed, 15984 insertions(+), 5 deletions(-) >> >> I have some concerns about this PMD, >> >> - This is a big one, as seen above, and it is hard to review it all, I >> don't feel confident about the amount of review done, more reviewers are >> welcome. And we are already post RC1. >> >> - Although this driver introduces a new bus type, in some parts, driver >> still has virtual devices like usage, perhaps this is not because of >> this PMD but mostly because of overall dpdk bus structure. Still I have >> concerns about getting driver like this, and would like to hear more >> comments. > > As a result of above concerns, I propose postponing this PMD to 17.05 > release. > > The dependent rte_bus just get into main repo less than two weeks ago, > also this driver comes with a few new things first of its kind, and it > matters to make first samples correct. > > I believe it is good to let the PMD be around a little more to give > chance to both PMD and rte_bus to become more mature. > I agree that this driver is coming with few new thing and it is taking time to come up with agreeable and good solution for some of this new stuff. Finalizing the right framework for adding SoC based drivers took a little longer than expected. But thanks to Thomas help in the last that we got the basic bus framework integrated. Thomos- is it possible to integrate it early in 17.05 cycle, rather than waiting till end? > Thanks, > ferruh > >