Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1657906

[PATCH 4.11 009/115] netem: fix skb_orphan_partial()

Show all headers | View raw

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

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

From: Eric Dumazet <edumazet@google.com>


[ Upstream commit f6ba8d33cfbb46df569972e64dbb5bb7e929bfd9 ]

I should have known that lowering skb->truesize was dangerous :/

In case packets are not leaving the host via a standard Ethernet device,
but looped back to local sockets, bad things can happen, as reported
by Michael Madsen ( https://bugzilla.kernel.org/show_bug.cgi?id=195713 )

So instead of tweaking skb->truesize, lets change skb->destructor
and keep a reference on the owner socket via its sk_refcnt.

Fixes: f2f872f9272a ("netem: Introduce skb_orphan_partial() helper")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Michael Madsen <mkm@nabto.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
---
 net/core/sock.c |   20 ++++++++------------
 1 file changed, 8 insertions(+), 12 deletions(-)

--- a/net/core/sock.c
+++ b/net/core/sock.c
@@ -1699,28 +1699,24 @@ EXPORT_SYMBOL(skb_set_owner_w);
  * delay queue. We want to allow the owner socket to send more
  * packets, as if they were already TX completed by a typical driver.
  * But we also want to keep skb->sk set because some packet schedulers
- * rely on it (sch_fq for example). So we set skb->truesize to a small
- * amount (1) and decrease sk_wmem_alloc accordingly.
+ * rely on it (sch_fq for example).
  */
 void skb_orphan_partial(struct sk_buff *skb)
 {
-	/* If this skb is a TCP pure ACK or already went here,
-	 * we have nothing to do. 2 is already a very small truesize.
-	 */
-	if (skb->truesize <= 2)
+	if (skb_is_tcp_pure_ack(skb))
 		return;
 
-	/* TCP stack sets skb->ooo_okay based on sk_wmem_alloc,
-	 * so we do not completely orphan skb, but transfert all
-	 * accounted bytes but one, to avoid unexpected reorders.
-	 */
 	if (skb->destructor == sock_wfree
 #ifdef CONFIG_INET
 	    || skb->destructor == tcp_wfree
 #endif
 		) {
-		atomic_sub(skb->truesize - 1, &skb->sk->sk_wmem_alloc);
-		skb->truesize = 1;
+		struct sock *sk = skb->sk;
+
+		if (atomic_inc_not_zero(&sk->sk_refcnt)) {
+			atomic_sub(skb->truesize, &sk->sk_wmem_alloc);
+			skb->destructor = sock_efree;
+		}
 	} else {
 		skb_orphan(skb);
 	}

Back to linux.kernel | Previous | Next — Previous in thread | Next in thread | Find similar | Unroll thread

Thread

[PATCH 4.11 000/115] 4.11.4-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 033/115] virtio-net: enable TSO/checksum offloads for Q-in-Q vlans Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 011/115] tcp: avoid fragmenting peculiar skbs in SACK Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 003/115] ipv6/dccp: do not inherit ipv6_mc_list from parent Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 009/115] netem: fix skb_orphan_partial() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 023/115] ipv6: Check ip6_find_1stfragopt() return value properly. Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 001/115] dccp/tcp: do not inherit mc_list from parent Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 018/115] tcp: eliminate negative reordering in tcp_clean_rtx_queue Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 024/115] bridge: netlink: check vlan_default_pvid range Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 027/115] ipv6: fix out of bound writes in __ip6_append_data() Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 017/115] net/mlx5e: Fix ethtool pause support and advertise reporting Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 029/115] net/mlx5: Avoid using pending command interface slots Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 021/115] net: Improve handling of failures on link and route dumps Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 008/115] bpf, arm64: fix faulty emission of map access in tail calls Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 012/115] tipc: make macro tipc_wait_for_cond() smp safe Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 043/115] sparc: Fix -Wstringop-overflow warning Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 005/115] s390/qeth: unbreak OSM and OSN support Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 007/115] s390/qeth: add missing hash table initializations Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 004/115] s390/qeth: handle sysfs error during initialization Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  [PATCH 4.11 020/115] net/smc: Add warning about remote memory exposure Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-05 19:00 +0200
  Re: [PATCH 4.11 000/115] 4.11.4-stable review Shuah Khan <shuahkh@osg.samsung.com> - 2017-06-05 22:40 +0200
    Re: [PATCH 4.11 000/115] 4.11.4-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-06 09:30 +0200
  Re: [PATCH 4.11 000/115] 4.11.4-stable review Guenter Roeck <linux@roeck-us.net> - 2017-06-06 00:30 +0200
    Re: [PATCH 4.11 000/115] 4.11.4-stable review Greg Kroah-Hartman <gregkh@linuxfoundation.org> - 2017-06-06 09:30 +0200

csiph-web