drivers/net/ethernet/freescale/dpaa/dpaa_eth.c | 9 ++++++++- 1 file changed, 8 insertions(+), 1 deletion(-)
When sending packets under 60 bytes, up to three bytes of the buffer
following the data may be leaked. Avoid this by extending all packets to
ETH_ZLEN, ensuring nothing is leaked in the padding. This bug can be
reproduced by running
$ ping -s 11 destination
Fixes: 9ad1a3749333 ("dpaa_eth: add support for DPAA Ethernet")
Suggested-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: Sean Anderson <sean.anderson@linux.dev>
---
Changes in v2:
- Fix the nonlinear varable becoming out-of-sync with the skb's
linearization state by padding the packet before anything else.
drivers/net/ethernet/freescale/dpaa/dpaa_eth.c | 9 ++++++++-
1 file changed, 8 insertions(+), 1 deletion(-)
diff --git a/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c b/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
index cfe6b57b1da0..4a55e521c17e 100644
--- a/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
+++ b/drivers/net/ethernet/freescale/dpaa/dpaa_eth.c
@@ -2272,12 +2272,12 @@ static netdev_tx_t
dpaa_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
{
const int queue_mapping = skb_get_queue_mapping(skb);
- bool nonlinear = skb_is_nonlinear(skb);
struct rtnl_link_stats64 *percpu_stats;
struct dpaa_percpu_priv *percpu_priv;
struct netdev_queue *txq;
struct dpaa_priv *priv;
struct qm_fd fd;
+ bool nonlinear;
int offset = 0;
int err = 0;
@@ -2287,6 +2287,13 @@ dpaa_start_xmit(struct sk_buff *skb, struct net_device *net_dev)
qm_fd_clear_fd(&fd);
+ /* Packet data is always read as 32-bit words, so zero out any part of
+ * the skb which might be sent if we have to pad the packet
+ */
+ if (__skb_put_padto(skb, ETH_ZLEN, false))
+ goto enomem;
+
+ nonlinear = skb_is_nonlinear(skb);
if (!nonlinear) {
/* We're going to store the skb backpointer at the beginning
* of the data buffer, so we need a privately owned skb
--
2.35.1.1320.gc452695387.dirty
On Tue, Sep 10, 2024 at 4:32 PM Sean Anderson <sean.anderson@linux.dev> wrote: > > When sending packets under 60 bytes, up to three bytes of the buffer > following the data may be leaked. Avoid this by extending all packets to > ETH_ZLEN, ensuring nothing is leaked in the padding. This bug can be > reproduced by running > > $ ping -s 11 destination > > Fixes: 9ad1a3749333 ("dpaa_eth: add support for DPAA Ethernet") > Suggested-by: Eric Dumazet <edumazet@google.com> > Signed-off-by: Sean Anderson <sean.anderson@linux.dev> Reviewed-by: Eric Dumazet <edumazet@google.com> Thanks !
© 2016 - 2024 Red Hat, Inc.