From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ferruh Yigit Subject: Re: [PATCH v1 0/6] ethdev: add min/max MTU to device info Date: Fri, 22 Mar 2019 14:08:18 +0000 Message-ID: References: <1551303948-19746-1-git-send-email-ian.stokes@intel.com> <5e86db4c-215d-ffea-29ee-df026c894627@intel.com> <1cce752b-09e4-abf1-3f1e-3afc6a4418fe@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: 8bit Cc: stephen@networkplumber.org, Thomas Monjalon , Andrew Rybchenko To: Ian Stokes , dev@dpdk.org Return-path: Received: from mga04.intel.com (mga04.intel.com [192.55.52.120]) by dpdk.org (Postfix) with ESMTP id A0B3F1B5C2 for ; Fri, 22 Mar 2019 15:08:22 +0100 (CET) In-Reply-To: <1cce752b-09e4-abf1-3f1e-3afc6a4418fe@intel.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 3/22/2019 1:05 PM, Ian Stokes wrote: > On 3/21/2019 1:03 PM, Ian Stokes wrote: >> On 3/19/2019 4:30 PM, Ferruh Yigit wrote: >>> On 2/27/2019 9:45 PM, Ian Stokes wrote: >>>> Building upon the discussion around [1], this series introduces MTU min >>>> and MTU max variables. It also provides updates to PMD implementations >>>> for ixgbe, i40e and IGB devices so that these variables are populated >>>> for use when retrieving device info. >>>> >>>> This series was tested with OVS DPDK and functions as expected for the >>>> drivers listed below. But a wider selection of PMD drivers would have to >>>> adopt this to ensure jumbo frames functionality remains for drivers not >>>> modified in the series. >>>> >>>> There is also ongoing discussion in [2] regarding overhead to be >>>> considered with MTU and how this may change from device to device, this >>>> series uses existing overhead assumptions. >>>> >>>> This series was previously posted as an RFC in [3], this revision >>>> removes RFC status and implements changes received in feedback. >>>> >>>> [1] http://mails.dpdk.org/archives/dev/2018-September/110959.html >>>> [2] http://mails.dpdk.org/archives/dev/2019-February/124457.html >>>> [3] http://mails.dpdk.org/archives/dev/2019-February/124938.html >>>> >>>> Ian Stokes (5): >>>>    net/i40e: set min and max MTU for i40e devices >>>>    net/i40e: set min and max MTU for i40e VF devices >>>>    net/ixgbe: set min and max MTU for ixgbe devices >>>>    net/ixgbe: set min and max MTU for ixgbe VF devices >>>>    net/e1000: set min and max MTU for igb devices >>>> >>>> Stephen Hemminger (1): >>>>    ethdev: add min/max MTU to device info >>> >>> Hi Ian, Stephen, >>> >>> API and driver updates are included in the patchset, but I believe it >>> would be >>> good to have some application code that uses it as well, I assume testpmd >>> already has some code to set MTU, can you please update it too >>> accordingly? >>> >> >> Thanks Ferruh, sure I had looked at this but held off in the v1 as I >> wasn't sure what best practice was, i.e.  introduce the change to sample >> app now or wait unitl all PMDs were on board. If it's preferred to >> introduce usage in a sample app then I can do this in the v2. >> >>> Also, what do you think starting a unit test (which has a long term >>> target to >>> verify all ethdev APIs) that tests 'rte_eth_dev_set_mtu()' API with >>> various values? >>> >> >> Sounds useful, I can take a look for the v2, first steps  might be basic >> but can look into it. >> >> Ian >>> In long term all vendors can run this unit test against their HW and >>> verify >>> ehtdev API implementation of their... >>> >> > > Hi Ferruh, > > I've posted a v2 of the patchset based on the feedback. > > http://mails.dpdk.org/archives/dev/2019-March/127344.html > > Unfortunately I did not have time to look at implementing the unit test > aspect. I don't think I'll have the bandwidth before the rc1 window next > week to implement this aspect but would be happy to look at it possibly > in the next 19.08 release if this is acceptable, is the unit test a > blocker for the rest of this work? Thanks for checking it, I believe it is not a blocker but I thought it may be a good start for verifying ethdev APIs, we can pursue this goal later.