Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next

Pull networking updates from Davic Miller:

 1) Support busy polling generically, for all NAPI drivers.  From Eric
    Dumazet.

 2) Add byte/packet counter support to nft_ct, from Floriani Westphal.

 3) Add RSS/XPS support to mvneta driver, from Gregory Clement.

 4) Implement IPV6_HDRINCL socket option for raw sockets, from Hannes
    Frederic Sowa.

 5) Add support for T6 adapter to cxgb4 driver, from Hariprasad Shenai.

 6) Add support for VLAN device bridging to mlxsw switch driver, from
    Ido Schimmel.

 7) Add driver for Netronome NFP4000/NFP6000, from Jakub Kicinski.

 8) Provide hwmon interface to mlxsw switch driver, from Jiri Pirko.

 9) Reorganize wireless drivers into per-vendor directories just like we
    do for ethernet drivers.  From Kalle Valo.

10) Provide a way for administrators "destroy" connected sockets via the
    SOCK_DESTROY socket netlink diag operation.  From Lorenzo Colitti.

11) Add support to add/remove multicast routes via netlink, from Nikolay
    Aleksandrov.

12) Make TCP keepalive settings per-namespace, from Nikolay Borisov.

13) Add forwarding and packet duplication facilities to nf_tables, from
    Pablo Neira Ayuso.

14) Dead route support in MPLS, from Roopa Prabhu.

15) TSO support for thunderx chips, from Sunil Goutham.

16) Add driver for IBM's System i/p VNIC protocol, from Thomas Falcon.

17) Rationalize, consolidate, and more completely document the checksum
    offloading facilities in the networking stack.  From Tom Herbert.

18) Support aborting an ongoing scan in mac80211/cfg80211, from
    Vidyullatha Kanchanapally.

19) Use per-bucket spinlock for bpf hash facility, from Tom Leiming.

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (1375 commits)
  net: bnxt: always return values from _bnxt_get_max_rings
  net: bpf: reject invalid shifts
  phonet: properly unshare skbs in phonet_rcv()
  dwc_eth_qos: Fix dma address for multi-fragment skbs
  phy: remove an unneeded condition
  mdio: remove an unneed condition
  mdio_bus: NULL dereference on allocation error
  net: Fix typo in netdev_intersect_features
  net: freescale: mac-fec: Fix build error from phy_device API change
  net: freescale: ucc_geth: Fix build error from phy_device API change
  bonding: Prevent IPv6 link local address on enslaved devices
  IB/mlx5: Add flow steering support
  net/mlx5_core: Export flow steering API
  net/mlx5_core: Make ipv4/ipv6 location more clear
  net/mlx5_core: Enable flow steering support for the IB driver
  net/mlx5_core: Initialize namespaces only when supported by device
  net/mlx5_core: Set priority attributes
  net/mlx5_core: Connect flow tables
  net/mlx5_core: Introduce modify flow table command
  net/mlx5_core: Managing root flow table
  ...

1 files changed, 31 insertions, 10 deletions

diff --git a/include/linux/filter.h b/include/linux/filter.h
index 5972ffe5719a..43aa1f8855c7 100644
--- a/include/linux/filter.h
+++ b/include/linux/filter.h
@@ -350,25 +350,43 @@ struct sk_filter {
 
 #define BPF_PROG_RUN(filter, ctx)  (*filter->bpf_func)(ctx, filter->insnsi)
 
+#define BPF_SKB_CB_LEN QDISC_CB_PRIV_LEN
+
+static inline u8 *bpf_skb_cb(struct sk_buff *skb)
+{
+	/* eBPF programs may read/write skb->cb[] area to transfer meta
+	 * data between tail calls. Since this also needs to work with
+	 * tc, that scratch memory is mapped to qdisc_skb_cb's data area.
+	 *
+	 * In some socket filter cases, the cb unfortunately needs to be
+	 * saved/restored so that protocol specific skb->cb[] data won't
+	 * be lost. In any case, due to unpriviledged eBPF programs
+	 * attached to sockets, we need to clear the bpf_skb_cb() area
+	 * to not leak previous contents to user space.
+	 */
+	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) != BPF_SKB_CB_LEN);
+	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) !=
+		     FIELD_SIZEOF(struct qdisc_skb_cb, data));
+
+	return qdisc_skb_cb(skb)->data;
+}
+
 static inline u32 bpf_prog_run_save_cb(const struct bpf_prog *prog,
 				       struct sk_buff *skb)
 {
-	u8 *cb_data = qdisc_skb_cb(skb)->data;
-	u8 saved_cb[QDISC_CB_PRIV_LEN];
+	u8 *cb_data = bpf_skb_cb(skb);
+	u8 cb_saved[BPF_SKB_CB_LEN];
 	u32 res;
 
-	BUILD_BUG_ON(FIELD_SIZEOF(struct __sk_buff, cb) !=
-		     QDISC_CB_PRIV_LEN);
-
 	if (unlikely(prog->cb_access)) {
-		memcpy(saved_cb, cb_data, sizeof(saved_cb));
-		memset(cb_data, 0, sizeof(saved_cb));
+		memcpy(cb_saved, cb_data, sizeof(cb_saved));
+		memset(cb_data, 0, sizeof(cb_saved));
 	}
 
 	res = BPF_PROG_RUN(prog, skb);
 
 	if (unlikely(prog->cb_access))
-		memcpy(cb_data, saved_cb, sizeof(saved_cb));
+		memcpy(cb_data, cb_saved, sizeof(cb_saved));
 
 	return res;
 }
@@ -376,10 +394,11 @@ static inline u32 bpf_prog_run_save_cb(const struct bpf_prog *prog,
 static inline u32 bpf_prog_run_clear_cb(const struct bpf_prog *prog,
 					struct sk_buff *skb)
 {
-	u8 *cb_data = qdisc_skb_cb(skb)->data;
+	u8 *cb_data = bpf_skb_cb(skb);
 
 	if (unlikely(prog->cb_access))
-		memset(cb_data, 0, QDISC_CB_PRIV_LEN);
+		memset(cb_data, 0, BPF_SKB_CB_LEN);
+
 	return BPF_PROG_RUN(prog, skb);
 }
 
@@ -447,6 +466,8 @@ void bpf_prog_destroy(struct bpf_prog *fp);
 
 int sk_attach_filter(struct sock_fprog *fprog, struct sock *sk);
 int sk_attach_bpf(u32 ufd, struct sock *sk);
+int sk_reuseport_attach_filter(struct sock_fprog *fprog, struct sock *sk);
+int sk_reuseport_attach_bpf(u32 ufd, struct sock *sk);
 int sk_detach_filter(struct sock *sk);
 int sk_get_filter(struct sock *sk, struct sock_filter __user *filter,
 		  unsigned int len);