Merge tag 'net-next-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next

Pull networking updates from Jakub Kicinski:
 "Core:

   - Enable memcg accounting for various networking objects.

  BPF:

   - Introduce bpf timers.

   - Add perf link and opaque bpf_cookie which the program can read out
     again, to be used in libbpf-based USDT library.

   - Add bpf_task_pt_regs() helper to access user space pt_regs in
     kprobes, to help user space stack unwinding.

   - Add support for UNIX sockets for BPF sockmap.

   - Extend BPF iterator support for UNIX domain sockets.

   - Allow BPF TCP congestion control progs and bpf iterators to call
     bpf_setsockopt(), e.g. to switch to another congestion control
     algorithm.

  Protocols:

   - Support IOAM Pre-allocated Trace with IPv6.

   - Support Management Component Transport Protocol.

   - bridge: multicast: add vlan support.

   - netfilter: add hooks for the SRv6 lightweight tunnel driver.

   - tcp:
       - enable mid-stream window clamping (by user space or BPF)
       - allow data-less, empty-cookie SYN with TFO_SERVER_COOKIE_NOT_REQD
       - more accurate DSACK processing for RACK-TLP

   - mptcp:
       - add full mesh path manager option
       - add partial support for MP_FAIL
       - improve use of backup subflows
       - optimize option processing

   - af_unix: add OOB notification support.

   - ipv6: add IFLA_INET6_RA_MTU to expose MTU value advertised by the
     router.

   - mac80211: Target Wake Time support in AP mode.

   - can: j1939: extend UAPI to notify about RX status.

  Driver APIs:

   - Add page frag support in page pool API.

   - Many improvements to the DSA (distributed switch) APIs.

   - ethtool: extend IRQ coalesce uAPI with timer reset modes.

   - devlink: control which auxiliary devices are created.

   - Support CAN PHYs via the generic PHY subsystem.

   - Proper cross-chip support for tag_8021q.

   - Allow TX forwarding for the software bridge data path to be
     offloaded to capable devices.

  Drivers:

   - veth: more flexible channels number configuration.

   - openvswitch: introduce per-cpu upcall dispatch.

   - Add internet mix (IMIX) mode to pktgen.

   - Transparently handle XDP operations in the bonding driver.

   - Add LiteETH network driver.

   - Renesas (ravb):
       - support Gigabit Ethernet IP

   - NXP Ethernet switch (sja1105):
       - fast aging support
       - support for "H" switch topologies
       - traffic termination for ports under VLAN-aware bridge

   - Intel 1G Ethernet
       - support getcrosststamp() with PCIe PTM (Precision Time
         Measurement) for better time sync
       - support Credit-Based Shaper (CBS) offload, enabling HW traffic
         prioritization and bandwidth reservation

   - Broadcom Ethernet (bnxt)
       - support pulse-per-second output
       - support larger Rx rings

   - Mellanox Ethernet (mlx5)
       - support ethtool RSS contexts and MQPRIO channel mode
       - support LAG offload with bridging
       - support devlink rate limit API
       - support packet sampling on tunnels

   - Huawei Ethernet (hns3):
       - basic devlink support
       - add extended IRQ coalescing support
       - report extended link state

   - Netronome Ethernet (nfp):
       - add conntrack offload support

   - Broadcom WiFi (brcmfmac):
       - add WPA3 Personal with FT to supported cipher suites
       - support 43752 SDIO device

   - Intel WiFi (iwlwifi):
       - support scanning hidden 6GHz networks
       - support for a new hardware family (Bz)

   - Xen pv driver:
       - harden netfront against malicious backends

   - Qualcomm mobile
       - ipa: refactor power management and enable automatic suspend
       - mhi: move MBIM to WWAN subsystem interfaces

  Refactor:

   - Ambient BPF run context and cgroup storage cleanup.

   - Compat rework for ndo_ioctl.

  Old code removal:

   - prism54 remove the obsoleted driver, deprecated by the p54 driver.

   - wan: remove sbni/granch driver"

* tag 'net-next-5.15' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1715 commits)
  net: Add depends on OF_NET for LiteX's LiteETH
  ipv6: seg6: remove duplicated include
  net: hns3: remove unnecessary spaces
  net: hns3: add some required spaces
  net: hns3: clean up a type mismatch warning
  net: hns3: refine function hns3_set_default_feature()
  ipv6: remove duplicated 'net/lwtunnel.h' include
  net: w5100: check return value after calling platform_get_resource()
  net/mlxbf_gige: Make use of devm_platform_ioremap_resourcexxx()
  net: mdio: mscc-miim: Make use of the helper function devm_platform_ioremap_resource()
  net: mdio-ipq4019: Make use of devm_platform_ioremap_resource()
  fou: remove sparse errors
  ipv4: fix endianness issue in inet_rtm_getroute_build_skb()
  octeontx2-af: Set proper errorcode for IPv4 checksum errors
  octeontx2-af: Fix static code analyzer reported issues
  octeontx2-af: Fix mailbox errors in nix_rss_flowkey_cfg
  octeontx2-af: Fix loop in free and unmap counter
  af_unix: fix potential NULL deref in unix_dgram_connect()
  dpaa2-eth: Replace strlcpy with strscpy
  octeontx2-af: Use NDC TX for transmit packet data
  ...

1 files changed, 102 insertions, 10 deletions

diff --git a/net/dsa/dsa2.c b/net/dsa/dsa2.c
index 185629f27f80..1b2b25d7bd02 100644
--- a/net/dsa/dsa2.c
+++ b/net/dsa/dsa2.c
@@ -21,6 +21,9 @@
 static DEFINE_MUTEX(dsa2_mutex);
 LIST_HEAD(dsa_tree_list);
 
+/* Track the bridges with forwarding offload enabled */
+static unsigned long dsa_fwd_offloading_bridges;
+
 /**
  * dsa_tree_notify - Execute code for all switches in a DSA switch tree.
  * @dst: collection of struct dsa_switch devices to notify.
@@ -49,6 +52,9 @@ int dsa_tree_notify(struct dsa_switch_tree *dst, unsigned long e, void *v)
  * Can be used to notify the switching fabric of events such as cross-chip
  * bridging between disjoint trees (such as islands of tagger-compatible
  * switches bridged by an incompatible middle switch).
+ *
+ * WARNING: this function is not reliable during probe time, because probing
+ * between trees is asynchronous and not all DSA trees might have probed.
  */
 int dsa_broadcast(unsigned long e, void *v)
 {
@@ -123,6 +129,51 @@ void dsa_lag_unmap(struct dsa_switch_tree *dst, struct net_device *lag)
 	}
 }
 
+static int dsa_bridge_num_find(const struct net_device *bridge_dev)
+{
+	struct dsa_switch_tree *dst;
+	struct dsa_port *dp;
+
+	/* When preparing the offload for a port, it will have a valid
+	 * dp->bridge_dev pointer but a not yet valid dp->bridge_num.
+	 * However there might be other ports having the same dp->bridge_dev
+	 * and a valid dp->bridge_num, so just ignore this port.
+	 */
+	list_for_each_entry(dst, &dsa_tree_list, list)
+		list_for_each_entry(dp, &dst->ports, list)
+			if (dp->bridge_dev == bridge_dev &&
+			    dp->bridge_num != -1)
+				return dp->bridge_num;
+
+	return -1;
+}
+
+int dsa_bridge_num_get(const struct net_device *bridge_dev, int max)
+{
+	int bridge_num = dsa_bridge_num_find(bridge_dev);
+
+	if (bridge_num < 0) {
+		/* First port that offloads TX forwarding for this bridge */
+		bridge_num = find_first_zero_bit(&dsa_fwd_offloading_bridges,
+						 DSA_MAX_NUM_OFFLOADING_BRIDGES);
+		if (bridge_num >= max)
+			return -1;
+
+		set_bit(bridge_num, &dsa_fwd_offloading_bridges);
+	}
+
+	return bridge_num;
+}
+
+void dsa_bridge_num_put(const struct net_device *bridge_dev, int bridge_num)
+{
+	/* Check if the bridge is still in use, otherwise it is time
+	 * to clean it up so we can reuse this bridge_num later.
+	 */
+	if (!dsa_bridge_num_find(bridge_dev))
+		clear_bit(bridge_num, &dsa_fwd_offloading_bridges);
+}
+
 struct dsa_switch *dsa_switch_find(int tree_index, int sw_index)
 {
 	struct dsa_switch_tree *dst;
@@ -311,6 +362,9 @@ static struct dsa_port *dsa_tree_find_first_cpu(struct dsa_switch_tree *dst)
 	return NULL;
 }
 
+/* Assign the default CPU port (the first one in the tree) to all ports of the
+ * fabric which don't already have one as part of their own switch.
+ */
 static int dsa_tree_setup_default_cpu(struct dsa_switch_tree *dst)
 {
 	struct dsa_port *cpu_dp, *dp;
@@ -321,15 +375,48 @@ static int dsa_tree_setup_default_cpu(struct dsa_switch_tree *dst)
 		return -EINVAL;
 	}
 
-	/* Assign the default CPU port to all ports of the fabric */
-	list_for_each_entry(dp, &dst->ports, list)
+	list_for_each_entry(dp, &dst->ports, list) {
+		if (dp->cpu_dp)
+			continue;
+
 		if (dsa_port_is_user(dp) || dsa_port_is_dsa(dp))
 			dp->cpu_dp = cpu_dp;
+	}
 
 	return 0;
 }
 
-static void dsa_tree_teardown_default_cpu(struct dsa_switch_tree *dst)
+/* Perform initial assignment of CPU ports to user ports and DSA links in the
+ * fabric, giving preference to CPU ports local to each switch. Default to
+ * using the first CPU port in the switch tree if the port does not have a CPU
+ * port local to this switch.
+ */
+static int dsa_tree_setup_cpu_ports(struct dsa_switch_tree *dst)
+{
+	struct dsa_port *cpu_dp, *dp;
+
+	list_for_each_entry(cpu_dp, &dst->ports, list) {
+		if (!dsa_port_is_cpu(cpu_dp))
+			continue;
+
+		list_for_each_entry(dp, &dst->ports, list) {
+			/* Prefer a local CPU port */
+			if (dp->ds != cpu_dp->ds)
+				continue;
+
+			/* Prefer the first local CPU port found */
+			if (dp->cpu_dp)
+				continue;
+
+			if (dsa_port_is_user(dp) || dsa_port_is_dsa(dp))
+				dp->cpu_dp = cpu_dp;
+		}
+	}
+
+	return dsa_tree_setup_default_cpu(dst);
+}
+
+static void dsa_tree_teardown_cpu_ports(struct dsa_switch_tree *dst)
 {
 	struct dsa_port *dp;
 
@@ -710,13 +797,14 @@ static int dsa_switch_setup(struct dsa_switch *ds)
 	/* Add the switch to devlink before calling setup, so that setup can
 	 * add dpipe tables
 	 */
-	ds->devlink = devlink_alloc(&dsa_devlink_ops, sizeof(*dl_priv));
+	ds->devlink =
+		devlink_alloc(&dsa_devlink_ops, sizeof(*dl_priv), ds->dev);
 	if (!ds->devlink)
 		return -ENOMEM;
 	dl_priv = devlink_priv(ds->devlink);
 	dl_priv->ds = ds;
 
-	err = devlink_register(ds->devlink, ds->dev);
+	err = devlink_register(ds->devlink);
 	if (err)
 		goto free_devlink;
 
@@ -921,13 +1009,13 @@ static int dsa_tree_setup(struct dsa_switch_tree *dst)
 	if (!complete)
 		return 0;
 
-	err = dsa_tree_setup_default_cpu(dst);
+	err = dsa_tree_setup_cpu_ports(dst);
 	if (err)
 		return err;
 
 	err = dsa_tree_setup_switches(dst);
 	if (err)
-		goto teardown_default_cpu;
+		goto teardown_cpu_ports;
 
 	err = dsa_tree_setup_master(dst);
 	if (err)
@@ -947,8 +1035,8 @@ teardown_master:
 	dsa_tree_teardown_master(dst);
 teardown_switches:
 	dsa_tree_teardown_switches(dst);
-teardown_default_cpu:
-	dsa_tree_teardown_default_cpu(dst);
+teardown_cpu_ports:
+	dsa_tree_teardown_cpu_ports(dst);
 
 	return err;
 }
@@ -966,7 +1054,7 @@ static void dsa_tree_teardown(struct dsa_switch_tree *dst)
 
 	dsa_tree_teardown_switches(dst);
 
-	dsa_tree_teardown_default_cpu(dst);
+	dsa_tree_teardown_cpu_ports(dst);
 
 	list_for_each_entry_safe(dl, next, &dst->rtable, list) {
 		list_del(&dl->list);
@@ -1044,6 +1132,7 @@ static struct dsa_port *dsa_port_touch(struct dsa_switch *ds, int index)
 
 	dp->ds = ds;
 	dp->index = index;
+	dp->bridge_num = -1;
 
 	INIT_LIST_HEAD(&dp->list);
 	list_add_tail(&dp->list, &dst->ports);
@@ -1265,6 +1354,9 @@ static int dsa_switch_parse_member_of(struct dsa_switch *ds,
 		return -EEXIST;
 	}
 
+	if (ds->dst->last_switch < ds->index)
+		ds->dst->last_switch = ds->index;
+
 	return 0;
 }