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;
From:   Vincent Mailhol <mailhol.vincent@wanadoo.fr>
To:     linux-can@vger.kernel.org, Wolfgang Grandegger <wg@grandegger.com>,
        Marc Kleine-Budde <mkl@pengutronix.de>,
        "David S . Miller" <davem@davemloft.net>
Cc:     Vincent Mailhol <mailhol.vincent@wanadoo.fr>,
        Jakub Kicinski <kuba@kernel.org>, netdev@vger.kernel.org,
        linux-kernel@vger.kernel.org
Subject: [PATCH 1/6] can: dev: can_get_echo_skb(): prevent call to kfree_skb() in hard IRQ context
Date:   Sun, 27 Sep 2020 02:45:31 +0900
Message-Id: <20200926174542.278166-2-mailhol.vincent@wanadoo.fr>
In-Reply-To: <20200926174542.278166-1-mailhol.vincent@wanadoo.fr>
References: <20200926174542.278166-1-mailhol.vincent@wanadoo.fr>
MIME-Version: 1.0
Content-Transfer-Encoding: 8bit
Precedence: bulk

If a driver calls can_get_echo_skb() during a hardware IRQ (which is
often, but not always, the case), the 'WARN_ON(in_irq)' in
net/core/skbuff.c#skb_release_head_state() might be triggered, under
network congestion circumstances, together with the potential risk of
a NULL pointer dereference.

The root cause of this issue is the call to kfree_skb() instead of
dev_kfree_skb_irq() in net/core/dev.c#enqueue_to_backlog().

This patch prevents the skb to be freed within the call to netif_rx()
by incrementing its reference count with skb_get(). The skb is finally
freed by one of the in-irq-context safe functions:
dev_consume_skb_any() or dev_kfree_skb_any().  The "any" version is
used because some drivers might call can_get_echo_skb() in a normal
context.

The reason for this issue to occur is that initially, in the core
network stack, loopback skb were not supposed to be received in
hardware IRQ context. The CAN stack is an exeption.

This bug is exactly what is described in
https://patchwork.ozlabs.org/patch/835236/

While above link proposes a patch that directly modifies
net/core/dev.c, we try to propose here a smoother modification local
to CAN network stack (the assumption behind is that only CAN devices
are affected by this issue).

Signed-off-by: Vincent Mailhol <mailhol.vincent@wanadoo.fr>
---
 drivers/net/can/dev.c | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

diff --git a/drivers/net/can/dev.c b/drivers/net/can/dev.c
index 68834a2853c9..e291fda395a0 100644
--- a/drivers/net/can/dev.c
+++ b/drivers/net/can/dev.c
@@ -512,7 +512,11 @@ unsigned int can_get_echo_skb(struct net_device *dev, unsigned int idx)
 	if (!skb)
 		return 0;
 
-	netif_rx(skb);
+	skb_get(skb);
+	if (netif_rx(skb) == NET_RX_SUCCESS)
+		dev_consume_skb_any(skb);
+	else
+		dev_kfree_skb_any(skb);
 
 	return len;
 }
-- 
2.26.2