Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.18 as permitted sender) client-ip=23.128.96.18;
MIME-Version: 1.0
References: <20210713123654.31174-1-yajun.deng@linux.dev> <20210713123654.31174-2-yajun.deng@linux.dev>
In-Reply-To: <20210713123654.31174-2-yajun.deng@linux.dev>
From:   Cong Wang <xiyou.wangcong@gmail.com>
Date:   Wed, 14 Jul 2021 21:21:52 -0700
Message-ID: <CAM_iQpUMxAOzMBJhxbYYYq2D2Fn_yYSxsWkxCDGVavGZhdKxHw@mail.gmail.com>
Subject: Re: [PATCH 1/2] rtnetlink: use nlmsg_{multicast, unicast} instead of netlink_{broadcast,unicast}
To:     Yajun Deng <yajun.deng@linux.dev>
Cc:     David Miller <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>,
        Jamal Hadi Salim <jhs@mojatatu.com>,
        Jiri Pirko <jiri@resnulli.us>,
        Johannes Berg <johannes.berg@intel.com>,
        ryazanov.s.a@gmail.com, Andrey Wagin <avagin@gmail.com>,
        vladimir.oltean@nxp.com, Roopa Prabhu <roopa@cumulusnetworks.com>,
        "zhudi (J)" <zhudi21@huawei.com>,
        LKML <linux-kernel@vger.kernel.org>,
        Linux Kernel Network Developers <netdev@vger.kernel.org>
Content-Type: text/plain; charset="UTF-8"
Precedence: bulk

On Tue, Jul 13, 2021 at 5:37 AM Yajun Deng <yajun.deng@linux.dev> wrote:
>
> It has a 'NETLINK_CB(' statement in nlmsg_multicast() and has 'if (err'
> in nlmsg_{multicast, unicast}, use nlmsg_{multicast, unicast} instead
> of netlink_{broadcast,unicast}. so the caller would not deal with the
> 'if (err >0 )' statement. Add the return value for nlmsg_multicast.
> As also, rename rtnetlink_send() to rtnl_send(), this makes style
> uniform.
>
> Signed-off-by: Yajun Deng <yajun.deng@linux.dev>
> ---
>  include/linux/rtnetlink.h |  2 +-
>  net/core/rtnetlink.c      | 13 +++++++------
>  2 files changed, 8 insertions(+), 7 deletions(-)
>
> diff --git a/include/linux/rtnetlink.h b/include/linux/rtnetlink.h
> index bb9cb84114c1..60bef82e42ab 100644
> --- a/include/linux/rtnetlink.h
> +++ b/include/linux/rtnetlink.h
> @@ -9,7 +9,7 @@
>  #include <linux/refcount.h>
>  #include <uapi/linux/rtnetlink.h>
>
> -extern int rtnetlink_send(struct sk_buff *skb, struct net *net, u32 pid, u32 group, int echo);
> +extern int rtnl_send(struct sk_buff *skb, struct net *net, u32 pid, u32 group, int echo);
>  extern int rtnl_unicast(struct sk_buff *skb, struct net *net, u32 pid);
>  extern void rtnl_notify(struct sk_buff *skb, struct net *net, u32 pid,
>                         u32 group, struct nlmsghdr *nlh, gfp_t flags);
> diff --git a/net/core/rtnetlink.c b/net/core/rtnetlink.c
> index f6af3e74fc44..c081d607bb69 100644
> --- a/net/core/rtnetlink.c
> +++ b/net/core/rtnetlink.c
> @@ -707,17 +707,18 @@ static int rtnl_link_fill(struct sk_buff *skb, const struct net_device *dev)
>         return err;
>  }
>
> -int rtnetlink_send(struct sk_buff *skb, struct net *net, u32 pid, unsigned int group, int echo)
> +int rtnl_send(struct sk_buff *skb, struct net *net, u32 pid, unsigned int group, int echo)
>  {
>         struct sock *rtnl = net->rtnl;
>         int err = 0;
>
> -       NETLINK_CB(skb).dst_group = group;
> -       if (echo)
> +       err = nlmsg_multicast(rtnl, skb, pid, group, GFP_KERNEL);
> +
> +       if (echo) {
>                 refcount_inc(&skb->users);

You also moved this refcount_inc() down after nlmsg_multicast().
Are you sure it is safe?

And the name rtnl_send() is bad given that rtnl_unicast() follows it...

Thanks.