2022-11-23 07:29:06

by Yue Haibing

[permalink] [raw]
Subject: [PATCH net] net: hsr: Fix potential use-after-free

The skb is delivered to netif_rx() which may free it, after calling this,
dereferencing skb may trigger use-after-free.

Fixes: f266a683a480 ("net/hsr: Better frame dispatch")
Signed-off-by: YueHaibing <[email protected]>
---
net/hsr/hsr_forward.c | 5 +++--
1 file changed, 3 insertions(+), 2 deletions(-)

diff --git a/net/hsr/hsr_forward.c b/net/hsr/hsr_forward.c
index a50429a62f74..56bb27d67a2e 100644
--- a/net/hsr/hsr_forward.c
+++ b/net/hsr/hsr_forward.c
@@ -351,17 +351,18 @@ static void hsr_deliver_master(struct sk_buff *skb, struct net_device *dev,
struct hsr_node *node_src)
{
bool was_multicast_frame;
- int res;
+ int res, recv_len;

was_multicast_frame = (skb->pkt_type == PACKET_MULTICAST);
hsr_addr_subst_source(node_src, skb);
skb_pull(skb, ETH_HLEN);
+ recv_len = skb->len;
res = netif_rx(skb);
if (res == NET_RX_DROP) {
dev->stats.rx_dropped++;
} else {
dev->stats.rx_packets++;
- dev->stats.rx_bytes += skb->len;
+ dev->stats.rx_bytes += recv_len;
if (was_multicast_frame)
dev->stats.multicast++;
}
--
2.17.1


2022-11-24 09:11:03

by Paolo Abeni

[permalink] [raw]
Subject: Re: [PATCH net] net: hsr: Fix potential use-after-free

Hello,

On Wed, 2022-11-23 at 14:30 +0800, YueHaibing wrote:
> The skb is delivered to netif_rx() which may free it, after calling this,
> dereferencing skb may trigger use-after-free.
>
> Fixes: f266a683a480 ("net/hsr: Better frame dispatch")
> Signed-off-by: YueHaibing <[email protected]>

The code looks good, but the above is not the commit introducing the
issue, it just move the netif_rx() and later skb access from somewhere
else.

Please go deeper in git history and find the change that originated the
issue.

Thanks,

Paolo

2022-11-25 08:16:39

by Yue Haibing

[permalink] [raw]
Subject: Re: [PATCH net] net: hsr: Fix potential use-after-free

On 2022/11/24 16:53, Paolo Abeni wrote:
> Hello,
>
> On Wed, 2022-11-23 at 14:30 +0800, YueHaibing wrote:
>> The skb is delivered to netif_rx() which may free it, after calling this,
>> dereferencing skb may trigger use-after-free.
>>
>> Fixes: f266a683a480 ("net/hsr: Better frame dispatch")
>> Signed-off-by: YueHaibing <[email protected]>
>
> The code looks good, but the above is not the commit introducing the
> issue, it just move the netif_rx() and later skb access from somewhere
> else.
>
> Please go deeper in git history and find the change that originated the> issue.

Ok, will dig it.

>
> Thanks,
>
> Paolo
>
> .
>