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, 30 insertions, 1 deletions

diff --git a/Documentation/networking/ip-sysctl.txt b/Documentation/networking/ip-sysctl.txt
index 2ea4c45cf1c8..ceb44a095a27 100644
--- a/Documentation/networking/ip-sysctl.txt
+++ b/Documentation/networking/ip-sysctl.txt
@@ -335,6 +335,14 @@ tcp_keepalive_intvl - INTEGER
 	after probes started. Default value: 75sec i.e. connection
 	will be aborted after ~11 minutes of retries.
 
+tcp_l3mdev_accept - BOOLEAN
+	Enables child sockets to inherit the L3 master device index.
+	Enabling this option allows a "global" listen socket to work
+	across L3 master domains (e.g., VRFs) with connected sockets
+	derived from the listen socket to be bound to the L3 domain in
+	which the packets originated. Only valid when the kernel was
+	compiled with CONFIG_NET_L3_MASTER_DEV.
+
 tcp_low_latency - BOOLEAN
 	If set, the TCP stack makes decisions that prefer lower
 	latency as opposed to higher throughput.  By default, this
@@ -1723,6 +1731,25 @@ addip_enable - BOOLEAN
 
 	Default: 0
 
+pf_enable - INTEGER
+	Enable or disable pf (pf is short for potentially failed) state. A value
+	of pf_retrans > path_max_retrans also disables pf state. That is, one of
+	both pf_enable and pf_retrans > path_max_retrans can disable pf state.
+	Since pf_retrans and path_max_retrans can be changed by userspace
+	application, sometimes user expects to disable pf state by the value of
+	pf_retrans > path_max_retrans, but occasionally the value of pf_retrans
+	or path_max_retrans is changed by the user application, this pf state is
+	enabled. As such, it is necessary to add this to dynamically enable
+	and disable pf state. See:
+	https://datatracker.ietf.org/doc/draft-ietf-tsvwg-sctp-failover for
+	details.
+
+	1: Enable pf.
+
+	0: Disable pf.
+
+	Default: 1
+
 addip_noauth_enable - BOOLEAN
 	Dynamic Address Reconfiguration (ADD-IP) requires the use of
 	authentication to protect the operations of adding or removing new
@@ -1799,7 +1826,9 @@ pf_retrans - INTEGER
 	having to reduce path_max_retrans to a very low value.  See:
 	http://www.ietf.org/id/draft-nishida-tsvwg-sctp-failover-05.txt
 	for details.  Note also that a value of pf_retrans > path_max_retrans
-	disables this feature
+	disables this feature. Since both pf_retrans and path_max_retrans can
+	be changed by userspace application, a variable pf_enable is used to
+	disable pf state.
 
 	Default: 0