[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <CAMDZJNVuFqaR0oBBGAzijQQfqm+Ak8rgH8AZ2YMqD_2idu_3xw@mail.gmail.com>
Date: Sat, 12 Jan 2019 11:53:05 +0800
From: Tonghao Zhang <xiangxia.m.yue@...il.com>
To: Cong Wang <xiyou.wangcong@...il.com>
Cc: Saeed Mahameed <saeedm@...lanox.com>, leonro@...lanox.com,
Linux Kernel Network Developers <netdev@...r.kernel.org>,
Hadar Hen Zion <hadarh@...lanox.com>
Subject: Re: [PATCH] net: mlx5: update hw flows when encap source mac changed
On Sat, Jan 12, 2019 at 6:50 AM Cong Wang <xiyou.wangcong@...il.com> wrote:
>
> On Fri, Jan 11, 2019 at 6:19 AM <xiangxia.m.yue@...il.com> wrote:
> > @@ -622,17 +629,20 @@ static void mlx5e_rep_neigh_update(struct work_struct *work)
> > memcpy(ha, n->ha, ETH_ALEN);
> > nud_state = n->nud_state;
> > dead = n->dead;
> > + netdev = n->dev;
> > read_unlock_bh(&n->lock);
> >
> > neigh_connected = (nud_state & NUD_VALID) && !dead;
> > + source_mac_changed = (e->h_source == netdev) && !ether_addr_equal(e->h_source, netdev->dev_addr);
>
> I don't know why you want to compare ->h_source, which points to
> a mac address, with a net_device pointer here.
my fault, but it works, because it also update the source mac. I
simply the implement in v2.
Powered by blists - more mailing lists