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;
Date:   Mon, 5 Dec 2022 10:37:25 +0100
From:   Piergiorgio Beruto <piergiorgio.beruto@gmail.com>
To:     "Russell King (Oracle)" <linux@armlinux.org.uk>
Cc:     Andrew Lunn <andrew@lunn.ch>,
        Heiner Kallweit <hkallweit1@gmail.com>,
        "David S. Miller" <davem@davemloft.net>,
        Eric Dumazet <edumazet@google.com>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>, linux-kernel@vger.kernel.org,
        netdev@vger.kernel.org, Oleksij Rempel <o.rempel@pengutronix.de>
Subject: Re: [PATCH v2 net-next 1/4] net/ethtool: add netlink interface for
 the PLCA RS
Message-ID: <Y4271Y9TtdiEgjn2@gvm01>
References: <fc3ac4f2d0c28d9c24b909e97791d1f784502a4a.1670204277.git.piergiorgio.beruto@gmail.com>
 <Y42509XkIN1S73Er@shell.armlinux.org.uk>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <Y42509XkIN1S73Er@shell.armlinux.org.uk>
Precedence: bulk

On Mon, Dec 05, 2022 at 09:28:51AM +0000, Russell King (Oracle) wrote:
> On Mon, Dec 05, 2022 at 02:41:35AM +0100, Piergiorgio Beruto wrote:
> > +int ethnl_set_plca_cfg(struct sk_buff *skb, struct genl_info *info)
> > +{
> > +	struct ethnl_req_info req_info = {};
> > +	struct nlattr **tb = info->attrs;
> > +	const struct ethtool_phy_ops *ops;
> > +	struct phy_plca_cfg plca_cfg;
> > +	struct net_device *dev;
> > +
> > +	bool mod = false;
> > +	int ret;
> > +
> > +	ret = ethnl_parse_header_dev_get(&req_info,
> > +					 tb[ETHTOOL_A_PLCA_HEADER],
> > +					 genl_info_net(info), info->extack,
> > +					 true);
> > +	if (ret < 0)
> > +		return ret;
> > +
> > +	dev = req_info.dev;
> > +
> > +	// check that the PHY device is available and connected
> > +	if (!dev->phydev) {
> > +		ret = -EOPNOTSUPP;
> > +		goto out;
> > +	}
> 
> This check should really be done under the RTNL lock. phydevs can come
> and go with SFP cages.
> 
> > +
> > +	rtnl_lock();
Good point Russell, I'll fix that. And I wish to seize the opportunity
to remark that the same problem may be present in cabletest.c
(see below).

Maybe we should post a patch to fix that?

Thanks,
Piergiorgio

int ethnl_act_cable_test(struct sk_buff *skb, struct genl_info *info)
{
	struct ethnl_req_info req_info = {};
	const struct ethtool_phy_ops *ops;
	struct nlattr **tb = info->attrs;
	struct net_device *dev;
	int ret;

	ret = ethnl_parse_header_dev_get(&req_info,
					 tb[ETHTOOL_A_CABLE_TEST_HEADER],
					 genl_info_net(info), info->extack,
					 true);
	if (ret < 0)
		return ret;

	dev = req_info.dev;
	if (!dev->phydev) {
		ret = -EOPNOTSUPP;
		goto out_dev_put;
	}

	rtnl_lock();