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

Pull networking updates from David Miller:
 "Another merge window, another pull full of stuff:

   1) Support alternative names for network devices, from Jiri Pirko.

   2) Introduce per-netns netdev notifiers, also from Jiri Pirko.

   3) Support MSG_PEEK in vsock/virtio, from Matias Ezequiel Vara
      Larsen.

   4) Allow compiling out the TLS TOE code, from Jakub Kicinski.

   5) Add several new tracepoints to the kTLS code, also from Jakub.

   6) Support set channels ethtool callback in ena driver, from Sameeh
      Jubran.

   7) New SCTP events SCTP_ADDR_ADDED, SCTP_ADDR_REMOVED,
      SCTP_ADDR_MADE_PRIM, and SCTP_SEND_FAILED_EVENT. From Xin Long.

   8) Add XDP support to mvneta driver, from Lorenzo Bianconi.

   9) Lots of netfilter hw offload fixes, cleanups and enhancements,
      from Pablo Neira Ayuso.

  10) PTP support for aquantia chips, from Egor Pomozov.

  11) Add UDP segmentation offload support to igb, ixgbe, and i40e. From
      Josh Hunt.

  12) Add smart nagle to tipc, from Jon Maloy.

  13) Support L2 field rewrite by TC offloads in bnxt_en, from Venkat
      Duvvuru.

  14) Add a flow mask cache to OVS, from Tonghao Zhang.

  15) Add XDP support to ice driver, from Maciej Fijalkowski.

  16) Add AF_XDP support to ice driver, from Krzysztof Kazimierczak.

  17) Support UDP GSO offload in atlantic driver, from Igor Russkikh.

  18) Support it in stmmac driver too, from Jose Abreu.

  19) Support TIPC encryption and auth, from Tuong Lien.

  20) Introduce BPF trampolines, from Alexei Starovoitov.

  21) Make page_pool API more numa friendly, from Saeed Mahameed.

  22) Introduce route hints to ipv4 and ipv6, from Paolo Abeni.

  23) Add UDP segmentation offload to cxgb4, Rahul Lakkireddy"

* git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1857 commits)
  libbpf: Fix usage of u32 in userspace code
  mm: Implement no-MMU variant of vmalloc_user_node_flags
  slip: Fix use-after-free Read in slip_open
  net: dsa: sja1105: fix sja1105_parse_rgmii_delays()
  macvlan: schedule bc_work even if error
  enetc: add support Credit Based Shaper(CBS) for hardware offload
  net: phy: add helpers phy_(un)lock_mdio_bus
  mdio_bus: don't use managed reset-controller
  ax88179_178a: add ethtool_op_get_ts_info()
  mlxsw: spectrum_router: Fix use of uninitialized adjacency index
  mlxsw: spectrum_router: After underlay moves, demote conflicting tunnels
  bpf: Simplify __bpf_arch_text_poke poke type handling
  bpf: Introduce BPF_TRACE_x helper for the tracing tests
  bpf: Add bpf_jit_blinding_enabled for !CONFIG_BPF_JIT
  bpf, testing: Add various tail call test cases
  bpf, x86: Emit patchable direct jump as tail call
  bpf: Constant map key tracking for prog array pokes
  bpf: Add poke dependency tracking for prog array maps
  bpf: Add initial poke descriptor table for jit images
  bpf: Move owner type, jited info into array auxiliary data
  ...

1 files changed, 134 insertions, 6 deletions

diff --git a/drivers/net/ethernet/intel/ice/ice_txrx.h b/drivers/net/ethernet/intel/ice/ice_txrx.h
index 94a9280193e2..a84cc0e6dd27 100644
--- a/drivers/net/ethernet/intel/ice/ice_txrx.h
+++ b/drivers/net/ethernet/intel/ice/ice_txrx.h
@@ -4,8 +4,12 @@
 #ifndef _ICE_TXRX_H_
 #define _ICE_TXRX_H_
 
+#include "ice_type.h"
+
 #define ICE_DFLT_IRQ_WORK	256
+#define ICE_RXBUF_3072		3072
 #define ICE_RXBUF_2048		2048
+#define ICE_RXBUF_1536		1536
 #define ICE_MAX_CHAINED_RX_BUFS	5
 #define ICE_MAX_BUF_TXD		8
 #define ICE_MIN_TX_LEN		17
@@ -22,6 +26,71 @@
 #define ICE_RX_BUF_WRITE	16	/* Must be power of 2 */
 #define ICE_MAX_TXQ_PER_TXQG	128
 
+/* Attempt to maximize the headroom available for incoming frames. We use a 2K
+ * buffer for MTUs <= 1500 and need 1536/1534 to store the data for the frame.
+ * This leaves us with 512 bytes of room.  From that we need to deduct the
+ * space needed for the shared info and the padding needed to IP align the
+ * frame.
+ *
+ * Note: For cache line sizes 256 or larger this value is going to end
+ *       up negative.  In these cases we should fall back to the legacy
+ *       receive path.
+ */
+#if (PAGE_SIZE < 8192)
+#define ICE_2K_TOO_SMALL_WITH_PADDING \
+((NET_SKB_PAD + ICE_RXBUF_1536) > SKB_WITH_OVERHEAD(ICE_RXBUF_2048))
+
+/**
+ * ice_compute_pad - compute the padding
+ * rx_buf_len: buffer length
+ *
+ * Figure out the size of half page based on given buffer length and
+ * then subtract the skb_shared_info followed by subtraction of the
+ * actual buffer length; this in turn results in the actual space that
+ * is left for padding usage
+ */
+static inline int ice_compute_pad(int rx_buf_len)
+{
+	int half_page_size;
+
+	half_page_size = ALIGN(rx_buf_len, PAGE_SIZE / 2);
+	return SKB_WITH_OVERHEAD(half_page_size) - rx_buf_len;
+}
+
+/**
+ * ice_skb_pad - determine the padding that we can supply
+ *
+ * Figure out the right Rx buffer size and based on that calculate the
+ * padding
+ */
+static inline int ice_skb_pad(void)
+{
+	int rx_buf_len;
+
+	/* If a 2K buffer cannot handle a standard Ethernet frame then
+	 * optimize padding for a 3K buffer instead of a 1.5K buffer.
+	 *
+	 * For a 3K buffer we need to add enough padding to allow for
+	 * tailroom due to NET_IP_ALIGN possibly shifting us out of
+	 * cache-line alignment.
+	 */
+	if (ICE_2K_TOO_SMALL_WITH_PADDING)
+		rx_buf_len = ICE_RXBUF_3072 + SKB_DATA_ALIGN(NET_IP_ALIGN);
+	else
+		rx_buf_len = ICE_RXBUF_1536;
+
+	/* if needed make room for NET_IP_ALIGN */
+	rx_buf_len -= NET_IP_ALIGN;
+
+	return ice_compute_pad(rx_buf_len);
+}
+
+#define ICE_SKB_PAD ice_skb_pad()
+#else
+#define ICE_2K_TOO_SMALL_WITH_PADDING false
+#define ICE_SKB_PAD (NET_SKB_PAD + NET_IP_ALIGN)
+#endif
+
 /* We are assuming that the cache line is always 64 Bytes here for ice.
  * In order to make sure that is a correct assumption there is a check in probe
  * to print a warning if the read from GLPCI_CNF2 tells us that the cache line
@@ -49,12 +118,24 @@
 #define ICE_TX_FLAGS_VLAN_PR_S	29
 #define ICE_TX_FLAGS_VLAN_S	16
 
+#define ICE_XDP_PASS		0
+#define ICE_XDP_CONSUMED	BIT(0)
+#define ICE_XDP_TX		BIT(1)
+#define ICE_XDP_REDIR		BIT(2)
+
 #define ICE_RX_DMA_ATTR \
 	(DMA_ATTR_SKIP_CPU_SYNC | DMA_ATTR_WEAK_ORDERING)
 
+#define ICE_ETH_PKT_HDR_PAD	(ETH_HLEN + ETH_FCS_LEN + (VLAN_HLEN * 2))
+
+#define ICE_TXD_LAST_DESC_CMD (ICE_TX_DESC_CMD_EOP | ICE_TX_DESC_CMD_RS)
+
 struct ice_tx_buf {
 	struct ice_tx_desc *next_to_watch;
-	struct sk_buff *skb;
+	union {
+		struct sk_buff *skb;
+		void *raw_buf; /* used for XDP */
+	};
 	unsigned int bytecount;
 	unsigned short gso_segs;
 	u32 tx_flags;
@@ -76,9 +157,17 @@ struct ice_tx_offload_params {
 struct ice_rx_buf {
 	struct sk_buff *skb;
 	dma_addr_t dma;
-	struct page *page;
-	unsigned int page_offset;
-	u16 pagecnt_bias;
+	union {
+		struct {
+			struct page *page;
+			unsigned int page_offset;
+			u16 pagecnt_bias;
+		};
+		struct {
+			void *addr;
+			u64 handle;
+		};
+	};
 };
 
 struct ice_q_stats {
@@ -198,18 +287,44 @@ struct ice_ring {
 	};
 
 	struct rcu_head rcu;		/* to avoid race on free */
+	struct bpf_prog *xdp_prog;
+	struct xdp_umem *xsk_umem;
+	struct zero_copy_allocator zca;
+	/* CL3 - 3rd cacheline starts here */
+	struct xdp_rxq_info xdp_rxq;
 	/* CLX - the below items are only accessed infrequently and should be
 	 * in their own cache line if possible
 	 */
+#define ICE_TX_FLAGS_RING_XDP		BIT(0)
+#define ICE_RX_FLAGS_RING_BUILD_SKB	BIT(1)
+	u8 flags;
 	dma_addr_t dma;			/* physical address of ring */
 	unsigned int size;		/* length of descriptor ring in bytes */
 	u32 txq_teid;			/* Added Tx queue TEID */
 	u16 rx_buf_len;
-#ifdef CONFIG_DCB
 	u8 dcb_tc;			/* Traffic class of ring */
-#endif /* CONFIG_DCB */
 } ____cacheline_internodealigned_in_smp;
 
+static inline bool ice_ring_uses_build_skb(struct ice_ring *ring)
+{
+	return !!(ring->flags & ICE_RX_FLAGS_RING_BUILD_SKB);
+}
+
+static inline void ice_set_ring_build_skb_ena(struct ice_ring *ring)
+{
+	ring->flags |= ICE_RX_FLAGS_RING_BUILD_SKB;
+}
+
+static inline void ice_clear_ring_build_skb_ena(struct ice_ring *ring)
+{
+	ring->flags &= ~ICE_RX_FLAGS_RING_BUILD_SKB;
+}
+
+static inline bool ice_ring_is_xdp(struct ice_ring *ring)
+{
+	return !!(ring->flags & ICE_TX_FLAGS_RING_XDP);
+}
+
 struct ice_ring_container {
 	/* head of linked-list of rings */
 	struct ice_ring *ring;
@@ -230,6 +345,19 @@ struct ice_ring_container {
 #define ice_for_each_ring(pos, head) \
 	for (pos = (head).ring; pos; pos = pos->next)
 
+static inline unsigned int ice_rx_pg_order(struct ice_ring *ring)
+{
+#if (PAGE_SIZE < 8192)
+	if (ring->rx_buf_len > (PAGE_SIZE / 2))
+		return 1;
+#endif
+	return 0;
+}
+
+#define ice_rx_pg_size(_ring) (PAGE_SIZE << ice_rx_pg_order(_ring))
+
+union ice_32b_rx_flex_desc;
+
 bool ice_alloc_rx_bufs(struct ice_ring *rxr, u16 cleaned_count);
 netdev_tx_t ice_start_xmit(struct sk_buff *skb, struct net_device *netdev);
 void ice_clean_tx_ring(struct ice_ring *tx_ring);