[PATCH 4.10 106/129] tcp: do not underestimate skb->truesize in tcp_trim_head()

From: Greg Kroah-Hartman
Date: Thu May 11 2017 - 11:25:00 EST


4.10-stable review patch. If anyone has any objections, please let me know.

------------------

From: Eric Dumazet <edumazet@xxxxxxxxxx>


[ Upstream commit 7162fb242cb8322beb558828fd26b33c3e9fc805 ]

Andrey found a way to trigger the WARN_ON_ONCE(delta < len) in
skb_try_coalesce() using syzkaller and a filter attached to a TCP
socket over loopback interface.

I believe one issue with looped skbs is that tcp_trim_head() can end up
producing skb with under estimated truesize.

It hardly matters for normal conditions, since packets sent over
loopback are never truncated.

Bytes trimmed from skb->head should not change skb truesize, since
skb->head is not reallocated.

Signed-off-by: Eric Dumazet <edumazet@xxxxxxxxxx>
Reported-by: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
Tested-by: Andrey Konovalov <andreyknvl@xxxxxxxxxx>
Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
net/ipv4/tcp_output.c | 19 ++++++++++++-------
1 file changed, 12 insertions(+), 7 deletions(-)

--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -1257,7 +1257,7 @@ int tcp_fragment(struct sock *sk, struct
* eventually). The difference is that pulled data not copied, but
* immediately discarded.
*/
-static void __pskb_trim_head(struct sk_buff *skb, int len)
+static int __pskb_trim_head(struct sk_buff *skb, int len)
{
struct skb_shared_info *shinfo;
int i, k, eat;
@@ -1267,7 +1267,7 @@ static void __pskb_trim_head(struct sk_b
__skb_pull(skb, eat);
len -= eat;
if (!len)
- return;
+ return 0;
}
eat = len;
k = 0;
@@ -1293,23 +1293,28 @@ static void __pskb_trim_head(struct sk_b
skb_reset_tail_pointer(skb);
skb->data_len -= len;
skb->len = skb->data_len;
+ return len;
}

/* Remove acked data from a packet in the transmit queue. */
int tcp_trim_head(struct sock *sk, struct sk_buff *skb, u32 len)
{
+ u32 delta_truesize;
+
if (skb_unclone(skb, GFP_ATOMIC))
return -ENOMEM;

- __pskb_trim_head(skb, len);
+ delta_truesize = __pskb_trim_head(skb, len);

TCP_SKB_CB(skb)->seq += len;
skb->ip_summed = CHECKSUM_PARTIAL;

- skb->truesize -= len;
- sk->sk_wmem_queued -= len;
- sk_mem_uncharge(sk, len);
- sock_set_flag(sk, SOCK_QUEUE_SHRUNK);
+ if (delta_truesize) {
+ skb->truesize -= delta_truesize;
+ sk->sk_wmem_queued -= delta_truesize;
+ sk_mem_uncharge(sk, delta_truesize);
+ sock_set_flag(sk, SOCK_QUEUE_SHRUNK);
+ }

/* Any change of skb->len requires recalculation of tso factor. */
if (tcp_skb_pcount(skb) > 1)