Re: [PATCH v3] can: ensure an initialized headroom in outgoing CAN sk_buffs

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 





On 10/12/2019 11.31, Marc Kleine-Budde wrote:
From: Oliver Hartkopp <socketcan@xxxxxxxxxxxx>

KMSAN sysbot detected a read access to an untinitialized value in the
headroom of an outgoing CAN related sk_buff. When using CAN sockets this
area is filled appropriately - but when using a packet socket this
initialization is missing.

The problematic read access occurs in the CAN receive path which can
only be triggered when the sk_buff is sent through a (virtual) CAN
interface. So we check in the sending path whether we need to perform
the missing initializations.

Fixes: d3b58c47d330d ("can: replace timestamp as unique skb attribute")
Reported-by: syzbot+b02ff0707a97e4e79ebb@xxxxxxxxxxxxxxxxxxxxxxxxx
Signed-off-by: Oliver Hartkopp <socketcan@xxxxxxxxxxxx>
Signed-off-by: Marc Kleine-Budde <mkl@xxxxxxxxxxxxxx>

When do you want to push it upstream?

The patch is ok for me.
Do you need another Acked-by or Tested-by tag?

Best regards,
Oliver

---
Changes since v2:
- move can_skb_headroom_valid() to header file and make it inline again
   to fix compilation if CAN_DEV is not enabled.

  include/linux/can/dev.h | 34 ++++++++++++++++++++++++++++++++++
  1 file changed, 34 insertions(+)

diff --git a/include/linux/can/dev.h b/include/linux/can/dev.h
index 9b3c720a31b1..5e3d45525bd3 100644
--- a/include/linux/can/dev.h
+++ b/include/linux/can/dev.h
@@ -18,6 +18,7 @@
  #include <linux/can/error.h>
  #include <linux/can/led.h>
  #include <linux/can/netlink.h>
+#include <linux/can/skb.h>
  #include <linux/netdevice.h>
/*
@@ -91,6 +92,36 @@ struct can_priv {
  #define get_can_dlc(i)		(min_t(__u8, (i), CAN_MAX_DLC))
  #define get_canfd_dlc(i)	(min_t(__u8, (i), CANFD_MAX_DLC))
+/* Check for outgoing skbs that have not been created by the CAN subsystem */
+static inline bool can_skb_headroom_valid(struct net_device *dev,
+					  struct sk_buff *skb)
+{
+	/* af_packet creates a headroom of HH_DATA_MOD bytes which is fine */
+	if (WARN_ON_ONCE(skb_headroom(skb) < sizeof(struct can_skb_priv)))
+		return false;
+
+	/* af_packet does not apply CAN skb specific settings */
+	if (skb->ip_summed == CHECKSUM_NONE) {
+		/* init headroom */
+		can_skb_prv(skb)->ifindex = dev->ifindex;
+		can_skb_prv(skb)->skbcnt = 0;
+
+		skb->ip_summed = CHECKSUM_UNNECESSARY;
+
+		/* preform proper loopback on capable devices */
+		if (dev->flags & IFF_ECHO)
+			skb->pkt_type = PACKET_LOOPBACK;
+		else
+			skb->pkt_type = PACKET_HOST;
+
+		skb_reset_mac_header(skb);
+		skb_reset_network_header(skb);
+		skb_reset_transport_header(skb);
+	}
+
+	return true;
+}
+
  /* Drop a given socketbuffer if it does not contain a valid CAN frame. */
  static inline bool can_dropped_invalid_skb(struct net_device *dev,
  					  struct sk_buff *skb)
@@ -108,6 +139,9 @@ static inline bool can_dropped_invalid_skb(struct net_device *dev,
  	} else
  		goto inval_skb;
+ if (!can_skb_headroom_valid(dev, skb))
+		goto inval_skb;
+
  	return false;
inval_skb:




[Index of Archives]     [Automotive Discussions]     [Linux ARM Kernel]     [Linux ARM]     [Linux Omap]     [Fedora ARM]     [IETF Annouce]     [Security]     [Bugtraq]     [Linux]     [Linux OMAP]     [Linux MIPS]     [eCos]     [Asterisk Internet PBX]     [Linux API]     [CAN Bus]

  Powered by Linux