From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Monjalon Subject: Re: [PATCH] net/i40e: add a specific API to control the LLDP agent Date: Mon, 21 May 2018 12:10:07 +0200 Message-ID: <15428557.l5iPiH4DZL@xps> References: <1526638447-31246-1-git-send-email-zijie.pan@6wind.com> <1752366.FEJfUWy9xN@xps> <6fc1b039-f79b-e82a-0e22-26f02a196697@6wind.com> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: quoted-printable Cc: Zijie Pan , dev , "beilei.xing" , "qi.z.zhang" , "helin.zhang" To: Laurent Hardy Return-path: Received: from out1-smtp.messagingengine.com (out1-smtp.messagingengine.com [66.111.4.25]) by dpdk.org (Postfix) with ESMTP id 54920AAEB for ; Mon, 21 May 2018 12:10:12 +0200 (CEST) In-Reply-To: <6fc1b039-f79b-e82a-0e22-26f02a196697@6wind.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" 21/05/2018 10:58, Laurent Hardy: > Hello Thomas, >=20 > On 05/21/2018 10:39 AM, Thomas Monjalon wrote: > > 21/05/2018 05:35, Zijie Pan: > >>>> Add a new API rte_pmd_i40e_set_lldp_cmd to control LLDP agent for i4= 0e. > >>>> It supports the following i40e debug lldp commands: > >>>> - start/stop of the LLDP agent. > >>>> - get local/remote of the LLDP MIB (Management Information Base= ). > >>>> > >>>> Signed-off-by: Laurent Hardy > >>>> Signed-off-by: Zijie Pan > >>> Does it mean that this feature is enabled by default in i40e devices? > >>> Should it be disabled by default? > >> The new API is just an interface to control the LLDP agent. No functio= ns of i40e call it by default. It is only used to debug LLDP of i40e device= s. For example we can add a new command for test-pmd to debug lldp of i40e = by calling rte_pmd_i40e_set_lldp_cmd(). > > There is a command to stop the LLDP agent. > > My question is to know whether this agent is running by default > > and should be disabled? > This agent is running by default. > Beside, disabling LLDP agent when using old fw (NVM 4.53 or 5.x) could=20 > have some side effects. >=20 > https://www.intel.com/content/dam/www/public/us/en/documents/specificatio= n-updates/xl710-10-40-controller-spec-update.pdf >=20 > Extracted from x710 intel spec: >=20 > "70. LLDP Disable Can Result in Incorrect Configuration of the Receive Pa= cket Buffer >=20 > Problem: > --------- > The default operation of the device is LLDP enabled. > LLDP can be disabled by an AQ command or by a custom NVM obtained from In= tel. > Disabling LLDP when using NVM 4.53 or 5.x can result in the device firmwa= re not configuring the Receive Packet Buffer according to the link mode and= flow control settings. > Implication: > ------------ > If flow control is disabled, a hang of the receive data path could occur = wherein all received packets are dropped. > If flow control is enabled, the incorrect settings could cause packet dro= ps despite the flow control, making the flow control ineffective. > Also, throughput might be sub-optimal in some cases. > Workaround: > ----------- > When using NVM 4.53 or 5.x, do one of the following: > =E2=80=A2 Do not disable LLDP or DCBx. Leave the device in default opera= tion mode. > =E2=80=A2 Use the Set Local LLDP MIB admin queue command to apply the (d= efault) DCB configuration during driver initialization and following any li= nk-up event. > When using NVM 6.01 or later, the receive data path does not hang even if= the FW LLDP is disabled." OK, interesting, thank you. I think you should add a comment in the code to explain why it is not disab= led. And later, we could consider disabling it with new firmwares. > > I mean i40e is generating LLDP packets by default? So the answer is yes, i40e generates packets silently :)