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

Pull networking updates from David Miller:
 "Highlights:

   1) Maintain the TCP retransmit queue using an rbtree, with 1GB
      windows at 100Gb this really has become necessary. From Eric
      Dumazet.

   2) Multi-program support for cgroup+bpf, from Alexei Starovoitov.

   3) Perform broadcast flooding in hardware in mv88e6xxx, from Andrew
      Lunn.

   4) Add meter action support to openvswitch, from Andy Zhou.

   5) Add a data meta pointer for BPF accessible packets, from Daniel
      Borkmann.

   6) Namespace-ify almost all TCP sysctl knobs, from Eric Dumazet.

   7) Turn on Broadcom Tags in b53 driver, from Florian Fainelli.

   8) More work to move the RTNL mutex down, from Florian Westphal.

   9) Add 'bpftool' utility, to help with bpf program introspection.
      From Jakub Kicinski.

  10) Add new 'cpumap' type for XDP_REDIRECT action, from Jesper
      Dangaard Brouer.

  11) Support 'blocks' of transformations in the packet scheduler which
      can span multiple network devices, from Jiri Pirko.

  12) TC flower offload support in cxgb4, from Kumar Sanghvi.

  13) Priority based stream scheduler for SCTP, from Marcelo Ricardo
      Leitner.

  14) Thunderbolt networking driver, from Amir Levy and Mika Westerberg.

  15) Add RED qdisc offloadability, and use it in mlxsw driver. From
      Nogah Frankel.

  16) eBPF based device controller for cgroup v2, from Roman Gushchin.

  17) Add some fundamental tracepoints for TCP, from Song Liu.

  18) Remove garbage collection from ipv6 route layer, this is a
      significant accomplishment. From Wei Wang.

  19) Add multicast route offload support to mlxsw, from Yotam Gigi"

* git://git.kernel.org/pub/scm/linux/kernel/git/davem/net-next: (2177 commits)
  tcp: highest_sack fix
  geneve: fix fill_info when link down
  bpf: fix lockdep splat
  net: cdc_ncm: GetNtbFormat endian fix
  openvswitch: meter: fix NULL pointer dereference in ovs_meter_cmd_reply_start
  netem: remove unnecessary 64 bit modulus
  netem: use 64 bit divide by rate
  tcp: Namespace-ify sysctl_tcp_default_congestion_control
  net: Protect iterations over net::fib_notifier_ops in fib_seq_sum()
  ipv6: set all.accept_dad to 0 by default
  uapi: fix linux/tls.h userspace compilation error
  usbnet: ipheth: prevent TX queue timeouts when device not ready
  vhost_net: conditionally enable tx polling
  uapi: fix linux/rxrpc.h userspace compilation errors
  net: stmmac: fix LPI transitioning for dwmac4
  atm: horizon: Fix irq release error
  net-sysfs: trigger netlink notification on ifalias change via sysfs
  openvswitch: Using kfree_rcu() to simplify the code
  openvswitch: Make local function ovs_nsh_key_attr_size() static
  openvswitch: Fix return value check in ovs_meter_cmd_features()
  ...

1 files changed, 113 insertions, 165 deletions

diff --git a/drivers/net/ethernet/netronome/nfp/bpf/offload.c b/drivers/net/ethernet/netronome/nfp/bpf/offload.c
index a88bb5bc0082..b6cee71f49d3 100644
--- a/drivers/net/ethernet/netronome/nfp/bpf/offload.c
+++ b/drivers/net/ethernet/netronome/nfp/bpf/offload.c
@@ -51,112 +51,114 @@
 #include "../nfp_net_ctrl.h"
 #include "../nfp_net.h"
 
-void nfp_net_filter_stats_timer(unsigned long data)
+static int
+nfp_prog_prepare(struct nfp_prog *nfp_prog, const struct bpf_insn *prog,
+		 unsigned int cnt)
 {
-	struct nfp_net *nn = (void *)data;
-	struct nfp_net_bpf_priv *priv;
-	struct nfp_stat_pair latest;
-
-	priv = nn->app_priv;
-
-	spin_lock_bh(&priv->rx_filter_lock);
+	unsigned int i;
 
-	if (nn->dp.ctrl & NFP_NET_CFG_CTRL_BPF)
-		mod_timer(&priv->rx_filter_stats_timer,
-			  jiffies + NFP_NET_STAT_POLL_IVL);
+	for (i = 0; i < cnt; i++) {
+		struct nfp_insn_meta *meta;
 
-	spin_unlock_bh(&priv->rx_filter_lock);
+		meta = kzalloc(sizeof(*meta), GFP_KERNEL);
+		if (!meta)
+			return -ENOMEM;
 
-	latest.pkts = nn_readq(nn, NFP_NET_CFG_STATS_APP1_FRAMES);
-	latest.bytes = nn_readq(nn, NFP_NET_CFG_STATS_APP1_BYTES);
+		meta->insn = prog[i];
+		meta->n = i;
 
-	if (latest.pkts != priv->rx_filter.pkts)
-		priv->rx_filter_change = jiffies;
+		list_add_tail(&meta->l, &nfp_prog->insns);
+	}
 
-	priv->rx_filter = latest;
+	return 0;
 }
 
-static void nfp_net_bpf_stats_reset(struct nfp_net *nn)
+static void nfp_prog_free(struct nfp_prog *nfp_prog)
 {
-	struct nfp_net_bpf_priv *priv = nn->app_priv;
+	struct nfp_insn_meta *meta, *tmp;
 
-	priv->rx_filter.pkts = nn_readq(nn, NFP_NET_CFG_STATS_APP1_FRAMES);
-	priv->rx_filter.bytes = nn_readq(nn, NFP_NET_CFG_STATS_APP1_BYTES);
-	priv->rx_filter_prev = priv->rx_filter;
-	priv->rx_filter_change = jiffies;
+	list_for_each_entry_safe(meta, tmp, &nfp_prog->insns, l) {
+		list_del(&meta->l);
+		kfree(meta);
+	}
+	kfree(nfp_prog);
 }
 
-static int
-nfp_net_bpf_stats_update(struct nfp_net *nn, struct tc_cls_bpf_offload *cls_bpf)
+int nfp_bpf_verifier_prep(struct nfp_app *app, struct nfp_net *nn,
+			  struct netdev_bpf *bpf)
 {
-	struct nfp_net_bpf_priv *priv = nn->app_priv;
-	u64 bytes, pkts;
+	struct bpf_prog *prog = bpf->verifier.prog;
+	struct nfp_prog *nfp_prog;
+	int ret;
 
-	pkts = priv->rx_filter.pkts - priv->rx_filter_prev.pkts;
-	bytes = priv->rx_filter.bytes - priv->rx_filter_prev.bytes;
-	bytes -= pkts * ETH_HLEN;
+	nfp_prog = kzalloc(sizeof(*nfp_prog), GFP_KERNEL);
+	if (!nfp_prog)
+		return -ENOMEM;
+	prog->aux->offload->dev_priv = nfp_prog;
 
-	priv->rx_filter_prev = priv->rx_filter;
+	INIT_LIST_HEAD(&nfp_prog->insns);
+	nfp_prog->type = prog->type;
 
-	tcf_exts_stats_update(cls_bpf->exts,
-			      bytes, pkts, priv->rx_filter_change);
+	ret = nfp_prog_prepare(nfp_prog, prog->insnsi, prog->len);
+	if (ret)
+		goto err_free;
 
-	return 0;
-}
+	nfp_prog->verifier_meta = nfp_prog_first_meta(nfp_prog);
+	bpf->verifier.ops = &nfp_bpf_analyzer_ops;
 
-static int
-nfp_net_bpf_get_act(struct nfp_net *nn, struct tc_cls_bpf_offload *cls_bpf)
-{
-	const struct tc_action *a;
-	LIST_HEAD(actions);
+	return 0;
 
-	if (!cls_bpf->exts)
-		return NN_ACT_XDP;
+err_free:
+	nfp_prog_free(nfp_prog);
 
-	/* TC direct action */
-	if (cls_bpf->exts_integrated) {
-		if (!tcf_exts_has_actions(cls_bpf->exts))
-			return NN_ACT_DIRECT;
+	return ret;
+}
 
+int nfp_bpf_translate(struct nfp_app *app, struct nfp_net *nn,
+		      struct bpf_prog *prog)
+{
+	struct nfp_prog *nfp_prog = prog->aux->offload->dev_priv;
+	unsigned int stack_size;
+	unsigned int max_instr;
+
+	stack_size = nn_readb(nn, NFP_NET_CFG_BPF_STACK_SZ) * 64;
+	if (prog->aux->stack_depth > stack_size) {
+		nn_info(nn, "stack too large: program %dB > FW stack %dB\n",
+			prog->aux->stack_depth, stack_size);
 		return -EOPNOTSUPP;
 	}
 
-	/* TC legacy mode */
-	if (!tcf_exts_has_one_action(cls_bpf->exts))
-		return -EOPNOTSUPP;
+	nfp_prog->stack_depth = prog->aux->stack_depth;
+	nfp_prog->start_off = nn_readw(nn, NFP_NET_CFG_BPF_START);
+	nfp_prog->tgt_done = nn_readw(nn, NFP_NET_CFG_BPF_DONE);
 
-	tcf_exts_to_list(cls_bpf->exts, &actions);
-	list_for_each_entry(a, &actions, list) {
-		if (is_tcf_gact_shot(a))
-			return NN_ACT_TC_DROP;
+	max_instr = nn_readw(nn, NFP_NET_CFG_BPF_MAX_LEN);
+	nfp_prog->__prog_alloc_len = max_instr * sizeof(u64);
 
-		if (is_tcf_mirred_egress_redirect(a) &&
-		    tcf_mirred_ifindex(a) == nn->dp.netdev->ifindex)
-			return NN_ACT_TC_REDIR;
-	}
+	nfp_prog->prog = kmalloc(nfp_prog->__prog_alloc_len, GFP_KERNEL);
+	if (!nfp_prog->prog)
+		return -ENOMEM;
 
-	return -EOPNOTSUPP;
+	return nfp_bpf_jit(nfp_prog);
 }
 
-static int
-nfp_net_bpf_offload_prepare(struct nfp_net *nn,
-			    struct tc_cls_bpf_offload *cls_bpf,
-			    struct nfp_bpf_result *res,
-			    void **code, dma_addr_t *dma_addr, u16 max_instr)
+int nfp_bpf_destroy(struct nfp_app *app, struct nfp_net *nn,
+		    struct bpf_prog *prog)
 {
-	unsigned int code_sz = max_instr * sizeof(u64);
-	enum nfp_bpf_action_type act;
-	u16 start_off, done_off;
-	unsigned int max_mtu;
-	int ret;
+	struct nfp_prog *nfp_prog = prog->aux->offload->dev_priv;
 
-	if (!IS_ENABLED(CONFIG_BPF_SYSCALL))
-		return -EOPNOTSUPP;
+	kfree(nfp_prog->prog);
+	nfp_prog_free(nfp_prog);
 
-	ret = nfp_net_bpf_get_act(nn, cls_bpf);
-	if (ret < 0)
-		return ret;
-	act = ret;
+	return 0;
+}
+
+static int nfp_net_bpf_load(struct nfp_net *nn, struct bpf_prog *prog)
+{
+	struct nfp_prog *nfp_prog = prog->aux->offload->dev_priv;
+	unsigned int max_mtu;
+	dma_addr_t dma_addr;
+	int err;
 
 	max_mtu = nn_readb(nn, NFP_NET_CFG_BPF_INL_MTU) * 64 - 32;
 	if (max_mtu < nn->dp.netdev->mtu) {
@@ -164,134 +166,80 @@ nfp_net_bpf_offload_prepare(struct nfp_net *nn,
 		return -EOPNOTSUPP;
 	}
 
-	start_off = nn_readw(nn, NFP_NET_CFG_BPF_START);
-	done_off = nn_readw(nn, NFP_NET_CFG_BPF_DONE);
-
-	*code = dma_zalloc_coherent(nn->dp.dev, code_sz, dma_addr, GFP_KERNEL);
-	if (!*code)
+	dma_addr = dma_map_single(nn->dp.dev, nfp_prog->prog,
+				  nfp_prog->prog_len * sizeof(u64),
+				  DMA_TO_DEVICE);
+	if (dma_mapping_error(nn->dp.dev, dma_addr))
 		return -ENOMEM;
 
-	ret = nfp_bpf_jit(cls_bpf->prog, *code, act, start_off, done_off,
-			  max_instr, res);
-	if (ret)
-		goto out;
+	nn_writew(nn, NFP_NET_CFG_BPF_SIZE, nfp_prog->prog_len);
+	nn_writeq(nn, NFP_NET_CFG_BPF_ADDR, dma_addr);
 
-	return 0;
+	/* Load up the JITed code */
+	err = nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_BPF);
+	if (err)
+		nn_err(nn, "FW command error while loading BPF: %d\n", err);
 
-out:
-	dma_free_coherent(nn->dp.dev, code_sz, *code, *dma_addr);
-	return ret;
+	dma_unmap_single(nn->dp.dev, dma_addr, nfp_prog->prog_len * sizeof(u64),
+			 DMA_TO_DEVICE);
+
+	return err;
 }
 
-static void
-nfp_net_bpf_load_and_start(struct nfp_net *nn, u32 tc_flags,
-			   void *code, dma_addr_t dma_addr,
-			   unsigned int code_sz, unsigned int n_instr,
-			   bool dense_mode)
+static void nfp_net_bpf_start(struct nfp_net *nn)
 {
-	struct nfp_net_bpf_priv *priv = nn->app_priv;
-	u64 bpf_addr = dma_addr;
 	int err;
 
-	nn->dp.bpf_offload_skip_sw = !!(tc_flags & TCA_CLS_FLAGS_SKIP_SW);
-
-	if (dense_mode)
-		bpf_addr |= NFP_NET_CFG_BPF_CFG_8CTX;
-
-	nn_writew(nn, NFP_NET_CFG_BPF_SIZE, n_instr);
-	nn_writeq(nn, NFP_NET_CFG_BPF_ADDR, bpf_addr);
-
-	/* Load up the JITed code */
-	err = nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_BPF);
-	if (err)
-		nn_err(nn, "FW command error while loading BPF: %d\n", err);
-
 	/* Enable passing packets through BPF function */
 	nn->dp.ctrl |= NFP_NET_CFG_CTRL_BPF;
 	nn_writel(nn, NFP_NET_CFG_CTRL, nn->dp.ctrl);
 	err = nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_GEN);
 	if (err)
 		nn_err(nn, "FW command error while enabling BPF: %d\n", err);
-
-	dma_free_coherent(nn->dp.dev, code_sz, code, dma_addr);
-
-	nfp_net_bpf_stats_reset(nn);
-	mod_timer(&priv->rx_filter_stats_timer,
-		  jiffies + NFP_NET_STAT_POLL_IVL);
 }
 
 static int nfp_net_bpf_stop(struct nfp_net *nn)
 {
-	struct nfp_net_bpf_priv *priv = nn->app_priv;
-
 	if (!(nn->dp.ctrl & NFP_NET_CFG_CTRL_BPF))
 		return 0;
 
-	spin_lock_bh(&priv->rx_filter_lock);
 	nn->dp.ctrl &= ~NFP_NET_CFG_CTRL_BPF;
-	spin_unlock_bh(&priv->rx_filter_lock);
 	nn_writel(nn, NFP_NET_CFG_CTRL, nn->dp.ctrl);
 
-	del_timer_sync(&priv->rx_filter_stats_timer);
-	nn->dp.bpf_offload_skip_sw = 0;
-
 	return nfp_net_reconfig(nn, NFP_NET_CFG_UPDATE_GEN);
 }
 
-int nfp_net_bpf_offload(struct nfp_net *nn, struct tc_cls_bpf_offload *cls_bpf)
+int nfp_net_bpf_offload(struct nfp_net *nn, struct bpf_prog *prog,
+			bool old_prog)
 {
-	struct nfp_bpf_result res;
-	dma_addr_t dma_addr;
-	u16 max_instr;
-	void *code;
 	int err;
 
-	max_instr = nn_readw(nn, NFP_NET_CFG_BPF_MAX_LEN);
+	if (prog && !prog->aux->offload)
+		return -EINVAL;
 
-	switch (cls_bpf->command) {
-	case TC_CLSBPF_REPLACE:
-		/* There is nothing stopping us from implementing seamless
-		 * replace but the simple method of loading I adopted in
-		 * the firmware does not handle atomic replace (i.e. we have to
-		 * stop the BPF offload and re-enable it).  Leaking-in a few
-		 * frames which didn't have BPF applied in the hardware should
-		 * be fine if software fallback is available, though.
-		 */
-		if (nn->dp.bpf_offload_skip_sw)
-			return -EBUSY;
-
-		err = nfp_net_bpf_offload_prepare(nn, cls_bpf, &res, &code,
-						  &dma_addr, max_instr);
-		if (err)
-			return err;
-
-		nfp_net_bpf_stop(nn);
-		nfp_net_bpf_load_and_start(nn, cls_bpf->gen_flags, code,
-					   dma_addr, max_instr * sizeof(u64),
-					   res.n_instr, res.dense_mode);
-		return 0;
+	if (prog && old_prog) {
+		u8 cap;
 
-	case TC_CLSBPF_ADD:
-		if (nn->dp.ctrl & NFP_NET_CFG_CTRL_BPF)
+		cap = nn_readb(nn, NFP_NET_CFG_BPF_CAP);
+		if (!(cap & NFP_NET_BPF_CAP_RELO)) {
+			nn_err(nn, "FW does not support live reload\n");
 			return -EBUSY;
+		}
+	}
 
-		err = nfp_net_bpf_offload_prepare(nn, cls_bpf, &res, &code,
-						  &dma_addr, max_instr);
-		if (err)
-			return err;
-
-		nfp_net_bpf_load_and_start(nn, cls_bpf->gen_flags, code,
-					   dma_addr, max_instr * sizeof(u64),
-					   res.n_instr, res.dense_mode);
-		return 0;
+	/* Something else is loaded, different program type? */
+	if (!old_prog && nn->dp.ctrl & NFP_NET_CFG_CTRL_BPF)
+		return -EBUSY;
 
-	case TC_CLSBPF_DESTROY:
+	if (old_prog && !prog)
 		return nfp_net_bpf_stop(nn);
 
-	case TC_CLSBPF_STATS:
-		return nfp_net_bpf_stats_update(nn, cls_bpf);
+	err = nfp_net_bpf_load(nn, prog);
+	if (err)
+		return err;
+
+	if (!old_prog)
+		nfp_net_bpf_start(nn);
 
-	default:
-		return -EOPNOTSUPP;
-	}
+	return 0;
 }