Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <20221213073801.361500-1-decot+git@google.com> <20221214204851.2102ba31@kernel.org>
 <CAF2d9jh_O0-uceNq=AP5rqPm9xcn=9y8bVxMD-2EiJ3bD_mZsQ@mail.gmail.com>
 <CAG88wWbZ3eXCFJBZ8mrfvddKiVihF-GfEOYAOmT_7VX_AeOoqQ@mail.gmail.com> <20221215110544.7e832e41@kernel.org>
In-Reply-To: <20221215110544.7e832e41@kernel.org>
From:   David Decotigny <ddecotig@google.com>
Date:   Thu, 15 Dec 2022 12:36:32 -0800
Message-ID: <CAG88wWYA72sij4iaWowLpawzM7tJdYdHCKQnE0bjndGO74vROw@mail.gmail.com>
Subject: Re: [PATCH v1 1/1] net: neigh: persist proxy config across link flaps
To:     Jakub Kicinski <kuba@kernel.org>
Cc:     David Ahern <dsahern@kernel.org>,
        =?UTF-8?B?TWFoZXNoIEJhbmRld2FyICjgpK7gpLngpYfgpLYg4KSs4KSC4KSh4KWH4KS14KS+4KSwKQ==?= 
        <maheshb@google.com>, David Decotigny <decot+git@google.com>,
        "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Paolo Abeni <pabeni@redhat.com>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org,
        Nikolay Aleksandrov <razor@blackwall.org>,
        "Denis V. Lunev" <den@openvz.org>,
        Daniel Borkmann <daniel@iogearbox.net>,
        Chen Zhongjin <chenzhongjin@huawei.com>,
        Yuwei Wang <wangyuweihx@gmail.com>,
        Alexander Mikhalitsyn <alexander.mikhalitsyn@virtuozzo.com>,
        Thomas Zeitlhofer <thomas.zeitlhofer+lkml@ze-it.at>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

(answer below)

On Thu, Dec 15, 2022 at 11:05 AM Jakub Kicinski <kuba@kernel.org> wrote:
>
> On Wed, 14 Dec 2022 22:18:04 -0800 David Decotigny wrote:
> > I don't think this patch is changing that part of the behavior: we still
> > flush the cached nd entries when the link flaps. What we don't remove are
> > the pneigh_entry-es (ip neigh add proxy ...) attached to the device where
> > the link flaps: those are configured once and this patch ensures that they
> > survive the link flaps as long as the netdev stays admin-up. When
> > the netdev is brought admin-down, we keep the behavior we had before the
> > patch.
>
> Makes sense. This is not urgent, tho, right?

Not that kind of urgent.

FTR, in the v2 you suggested to use NUD_PERMANENT, I can try to see
how this would look like. Note that this will make the patch larger
and more intrusive, and with potentially a behavior change for whoever
uses the netlink API directly instead of the iproute2 implementation
for ip neigh X proxy things.

>
> David A, do you agree and should we treat this as a fix with
>
> Fixes: 859bd2ef1fc1 ("net: Evict neighbor entries on carrier down")

Thanks.

>
> added?
>
> Reminder: please bottom post on the list