Skip to content

Commit

Permalink
sch_cake: Return __NET_XMIT_STOLEN when consuming enqueued skb
Browse files Browse the repository at this point in the history
[ Upstream commit 90fabae ]

When the GSO splitting feature of sch_cake is enabled, GSO superpackets
will be broken up and the resulting segments enqueued in place of the
original skb. In this case, CAKE calls consume_skb() on the original skb,
but still returns NET_XMIT_SUCCESS. This can confuse parent qdiscs into
assuming the original skb still exists, when it really has been freed. Fix
this by adding the __NET_XMIT_STOLEN flag to the return value in this case.

Fixes: 0c85034 ("sch_cake: Conditionally split GSO segments")
Signed-off-by: Toke Høiland-Jørgensen <toke@toke.dk>
Reported-by: zdi-disclosures@trendmicro.com # ZDI-CAN-18231
Link: https://lore.kernel.org/r/20220831092103.442868-1-toke@toke.dk
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
  • Loading branch information
tohojo authored and gregkh committed Sep 8, 2022
1 parent 55fb8c3 commit 1c472d6
Showing 1 changed file with 3 additions and 1 deletion.
4 changes: 3 additions & 1 deletion net/sched/sch_cake.c
Expand Up @@ -1713,6 +1713,7 @@ static s32 cake_enqueue(struct sk_buff *skb, struct Qdisc *sch,
}
idx--;
flow = &b->flows[idx];
ret = NET_XMIT_SUCCESS;

/* ensure shaper state isn't stale */
if (!b->tin_backlog) {
Expand Down Expand Up @@ -1771,6 +1772,7 @@ static s32 cake_enqueue(struct sk_buff *skb, struct Qdisc *sch,

qdisc_tree_reduce_backlog(sch, 1-numsegs, len-slen);
consume_skb(skb);
ret |= __NET_XMIT_STOLEN;
} else {
/* not splitting */
cobalt_set_enqueue_time(skb, now);
Expand Down Expand Up @@ -1904,7 +1906,7 @@ static s32 cake_enqueue(struct sk_buff *skb, struct Qdisc *sch,
}
b->drop_overlimit += dropped;
}
return NET_XMIT_SUCCESS;
return ret;
}

static struct sk_buff *cake_dequeue_one(struct Qdisc *sch)
Expand Down

0 comments on commit 1c472d6

Please sign in to comment.