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:   Fri, 19 Aug 2022 16:45:19 -0700
From:   Jakub Kicinski <kuba@kernel.org>
To:     Sean Anderson <sean.anderson@seco.com>,
        Andrew Lunn <andrew@lunn.ch>,
        Heiner Kallweit <hkallweit1@gmail.com>,
        Russell King <linux@armlinux.org.uk>
Cc:     netdev@vger.kernel.org, Paolo Abeni <pabeni@redhat.com>,
        "David S . Miller" <davem@davemloft.net>,
        linux-kernel@vger.kernel.org, Eric Dumazet <edumazet@google.com>,
        Vladimir Oltean <olteanv@gmail.com>
Subject: Re: [PATCH net] net: phy: Warn if phy is attached when removing
Message-ID: <20220819164519.2c71823e@kernel.org>
In-Reply-To: <20220816163701.1578850-1-sean.anderson@seco.com>
References: <20220816163701.1578850-1-sean.anderson@seco.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit
Precedence: bulk

On Tue, 16 Aug 2022 12:37:01 -0400 Sean Anderson wrote:
> netdevs using phylib can be oopsed from userspace in the following
> manner:
> 
> $ ip link set $iface up
> $ echo $(basename $(readlink /sys/class/net/$iface/phydev)) > \
>       /sys/class/net/$iface/phydev/driver/unbind
> $ ip link set $iface down
> 
> However, the traceback provided is a bit too late, since it does not
> capture the root of the problem (unbinding the driver). It's also
> possible that the memory has been reallocated if sufficient time passes
> between when the phy is detached and when the netdev touches the phy
> (which could result in silent memory corruption). Add a warning at the
> source of the problem. A future patch could make this more robust by
> calling dev_close.

FWIW, I think DaveM marked this patch as changes requested.

I don't really know enough to have an opinion.

PHY maintainers, anyone willing to cast a vote?