From mboxrd@z Thu Jan  1 00:00:00 1970
From: Song Liu <songliubraving@fb.com>
Subject: Re: [PATCH net 00/15] netpoll: avoid capture effects for NAPI drivers
Date: Mon, 24 Sep 2018 21:18:13 +0000
Message-ID: <ACD08072-B5A0-4A5D-8E02-30E5F763124D@fb.com>
References: <20180921222752.101307-1-edumazet@google.com>
 <20180923.220440.434897583339366571.davem@davemloft.net>
 <CANn89i+8mngtTLs8gzmEFjreW9ie64MPmZWaUDV+KRR7wgNfMA@mail.gmail.com>
 <B352B219-D053-432C-AEB7-C0A74D2F6613@fb.com>
 <CANn89iLdMaW6rcK3+JT3J56M3xT2pgPPEpQRhngBL4DocMNChw@mail.gmail.com>
 <ED77932E-558E-4422-AA7C-2DEED4F2954C@fb.com>
 <CANn89iL46aONnTP6+Mb=tSq2-ywntAMwasLEuCs8VNBd5jqitg@mail.gmail.com>
 <CANn89i+dri9_d8XMcUd=+-ncFYmgDQRd4TWB93QwtoAk+ooX6g@mail.gmail.com>
Mime-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: quoted-printable
Cc: David Miller <davem@davemloft.net>,
        netdev <netdev@vger.kernel.org>,
        Michael Chan <michael.chan@broadcom.com>,
        Ariel Elior <ariel.elior@cavium.com>,
        Eric Dumazet <eric.dumazet@gmail.com>,
        Tariq Toukan <tariqt@mellanox.com>,
        Saeed Mahameed <saeedm@mellanox.com>,
        Jeff Kirsher <jeffrey.t.kirsher@intel.com>,
        Jakub Kicinski <jakub.kicinski@netronome.com>,
        Jay Vosburgh <j.vosburgh@gmail.com>,
        Veaceslav Falico <vfalico@gmail.com>,
        Andy Gospodarek <andy@greyhouse.net>
To: Eric Dumazet <edumazet@google.com>
Return-path: <netdev-owner@vger.kernel.org>
Received: from mx0a-00082601.pphosted.com ([67.231.145.42]:47882 "EHLO
        mx0a-00082601.pphosted.com" rhost-flags-OK-OK-OK-OK)
        by vger.kernel.org with ESMTP id S1727531AbeIYDWh (ORCPT
        <rfc822;netdev@vger.kernel.org>); Mon, 24 Sep 2018 23:22:37 -0400
In-Reply-To: <CANn89i+dri9_d8XMcUd=+-ncFYmgDQRd4TWB93QwtoAk+ooX6g@mail.gmail.com>
Content-Language: en-US
Content-ID: <AAB6B56D8163C341850B9781280DED48@namprd15.prod.outlook.com>
Sender: netdev-owner@vger.kernel.org
List-ID: <netdev.vger.kernel.org>



> On Sep 24, 2018, at 2:05 PM, Eric Dumazet <edumazet@google.com> wrote:
>=20
>>=20
>> Interesting, maybe a bnxt specific issue.
>>=20
>> It seems their model is to process TX/RX notification in the same queue,
>> they throw away RX events if budget =3D=3D 0
>>=20
>> It means commit e7b9569102995ebc26821789628eef45bd9840d8 is wrong and
>> must be reverted.
>>=20
>> Otherwise, we have a possibility of blocking a queue under netpoll press=
ure.
>=20
> Hmm, actually a revert might not be enough, since code at lines 2030-2031
> would fire and we might not call napi_complete_done() anyway.
>=20
> Unfortunately this driver logic is quite complex.
>=20
> Could you test on other NIC eventually ?
>=20

It actually runs OK on ixgbe.=20

@Michael, could you please help us with this?=20

Thanks,
Song