Merge tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-nextgrafted

Pull networking updates from Jakub Kicinski:
 "Core:

   - Add dedicated kmem_cache for typical/small skb->head, avoid having
     to access struct page at kfree time, and improve memory use.

   - Introduce sysctl to set default RPS configuration for new netdevs.

   - Define Netlink protocol specification format which can be used to
     describe messages used by each family and auto-generate parsers.
     Add tools for generating kernel data structures and uAPI headers.

   - Expose all net/core sysctls inside netns.

   - Remove 4s sleep in netpoll if carrier is instantly detected on
     boot.

   - Add configurable limit of MDB entries per port, and port-vlan.

   - Continue populating drop reasons throughout the stack.

   - Retire a handful of legacy Qdiscs and classifiers.

  Protocols:

   - Support IPv4 big TCP (TSO frames larger than 64kB).

   - Add IP_LOCAL_PORT_RANGE socket option, to control local port range
     on socket by socket basis.

   - Track and report in procfs number of MPTCP sockets used.

   - Support mixing IPv4 and IPv6 flows in the in-kernel MPTCP path
     manager.

   - IPv6: don't check net.ipv6.route.max_size and rely on garbage
     collection to free memory (similarly to IPv4).

   - Support Penultimate Segment Pop (PSP) flavor in SRv6 (RFC8986).

   - ICMP: add per-rate limit counters.

   - Add support for user scanning requests in ieee802154.

   - Remove static WEP support.

   - Support minimal Wi-Fi 7 Extremely High Throughput (EHT) rate
     reporting.

   - WiFi 7 EHT channel puncturing support (client & AP).

  BPF:

   - Add a rbtree data structure following the "next-gen data structure"
     precedent set by recently added linked list, that is, by using
     kfunc + kptr instead of adding a new BPF map type.

   - Expose XDP hints via kfuncs with initial support for RX hash and
     timestamp metadata.

   - Add BPF_F_NO_TUNNEL_KEY extension to bpf_skb_set_tunnel_key to
     better support decap on GRE tunnel devices not operating in collect
     metadata.

   - Improve x86 JIT's codegen for PROBE_MEM runtime error checks.

   - Remove the need for trace_printk_lock for bpf_trace_printk and
     bpf_trace_vprintk helpers.

   - Extend libbpf's bpf_tracing.h support for tracing arguments of
     kprobes/uprobes and syscall as a special case.

   - Significantly reduce the search time for module symbols by
     livepatch and BPF.

   - Enable cpumasks to be used as kptrs, which is useful for tracing
     programs tracking which tasks end up running on which CPUs in
     different time intervals.

   - Add support for BPF trampoline on s390x and riscv64.

   - Add capability to export the XDP features supported by the NIC.

   - Add __bpf_kfunc tag for marking kernel functions as kfuncs.

   - Add cgroup.memory=nobpf kernel parameter option to disable BPF
     memory accounting for container environments.

  Netfilter:

   - Remove the CLUSTERIP target. It has been marked as obsolete for
     years, and we still have WARN splats wrt races of the out-of-band
     /proc interface installed by this target.

   - Add 'destroy' commands to nf_tables. They are identical to the
     existing 'delete' commands, but do not return an error if the
     referenced object (set, chain, rule...) did not exist.

  Driver API:

   - Improve cpumask_local_spread() locality to help NICs set the right
     IRQ affinity on AMD platforms.

   - Separate C22 and C45 MDIO bus transactions more clearly.

   - Introduce new DCB table to control DSCP rewrite on egress.

   - Support configuration of Physical Layer Collision Avoidance (PLCA)
     Reconciliation Sublayer (RS) (802.3cg-2019). Modern version of
     shared medium Ethernet.

   - Support for MAC Merge layer (IEEE 802.3-2018 clause 99). Allowing
     preemption of low priority frames by high priority frames.

   - Add support for controlling MACSec offload using netlink SET.

   - Rework devlink instance refcounts to allow registration and
     de-registration under the instance lock. Split the code into
     multiple files, drop some of the unnecessarily granular locks and
     factor out common parts of netlink operation handling.

   - Add TX frame aggregation parameters (for USB drivers).

   - Add a new attr TCA_EXT_WARN_MSG to report TC (offload) warning
     messages with notifications for debug.

   - Allow offloading of UDP NEW connections via act_ct.

   - Add support for per action HW stats in TC.

   - Support hardware miss to TC action (continue processing in SW from
     a specific point in the action chain).

   - Warn if old Wireless Extension user space interface is used with
     modern cfg80211/mac80211 drivers. Do not support Wireless
     Extensions for Wi-Fi 7 devices at all. Everyone should switch to
     using nl80211 interface instead.

   - Improve the CAN bit timing configuration. Use extack to return
     error messages directly to user space, update the SJW handling,
     including the definition of a new default value that will benefit
     CAN-FD controllers, by increasing their oscillator tolerance.

  New hardware / drivers:

   - Ethernet:
      - nVidia BlueField-3 support (control traffic driver)
      - Ethernet support for imx93 SoCs
      - Motorcomm yt8531 gigabit Ethernet PHY
      - onsemi NCN26000 10BASE-T1S PHY (with support for PLCA)
      - Microchip LAN8841 PHY (incl. cable diagnostics and PTP)
      - Amlogic gxl MDIO mux

   - WiFi:
      - RealTek RTL8188EU (rtl8xxxu)
      - Qualcomm Wi-Fi 7 devices (ath12k)

   - CAN:
      - Renesas R-Car V4H

  Drivers:

   - Bluetooth:
      - Set Per Platform Antenna Gain (PPAG) for Intel controllers.

   - Ethernet NICs:
      - Intel (1G, igc):
         - support TSN / Qbv / packet scheduling features of i226 model
      - Intel (100G, ice):
         - use GNSS subsystem instead of TTY
         - multi-buffer XDP support
         - extend support for GPIO pins to E823 devices
      - nVidia/Mellanox:
         - update the shared buffer configuration on PFC commands
         - implement PTP adjphase function for HW offset control
         - TC support for Geneve and GRE with VF tunnel offload
         - more efficient crypto key management method
         - multi-port eswitch support
      - Netronome/Corigine:
         - add DCB IEEE support
         - support IPsec offloading for NFP3800
      - Freescale/NXP (enetc):
         - support XDP_REDIRECT for XDP non-linear buffers
         - improve reconfig, avoid link flap and waiting for idle
         - support MAC Merge layer
      - Other NICs:
         - sfc/ef100: add basic devlink support for ef100
         - ionic: rx_push mode operation (writing descriptors via MMIO)
         - bnxt: use the auxiliary bus abstraction for RDMA
         - r8169: disable ASPM and reset bus in case of tx timeout
         - cpsw: support QSGMII mode for J721e CPSW9G
         - cpts: support pulse-per-second output
         - ngbe: add an mdio bus driver
         - usbnet: optimize usbnet_bh() by avoiding unnecessary queuing
         - r8152: handle devices with FW with NCM support
         - amd-xgbe: support 10Mbps, 2.5GbE speeds and rx-adaptation
         - virtio-net: support multi buffer XDP
         - virtio/vsock: replace virtio_vsock_pkt with sk_buff
         - tsnep: XDP support

   - Ethernet high-speed switches:
      - nVidia/Mellanox (mlxsw):
         - add support for latency TLV (in FW control messages)
      - Microchip (sparx5):
         - separate explicit and implicit traffic forwarding rules, make
           the implicit rules always active
         - add support for egress DSCP rewrite
         - IS0 VCAP support (Ingress Classification)
         - IS2 VCAP filters (protos, L3 addrs, L4 ports, flags, ToS
           etc.)
         - ES2 VCAP support (Egress Access Control)
         - support for Per-Stream Filtering and Policing (802.1Q,
           8.6.5.1)

   - Ethernet embedded switches:
      - Marvell (mv88e6xxx):
         - add MAB (port auth) offload support
         - enable PTP receive for mv88e6390
      - NXP (ocelot):
         - support MAC Merge layer
         - support for the the vsc7512 internal copper phys
      - Microchip:
         - lan9303: convert to PHYLINK
         - lan966x: support TC flower filter statistics
         - lan937x: PTP support for KSZ9563/KSZ8563 and LAN937x
         - lan937x: support Credit Based Shaper configuration
         - ksz9477: support Energy Efficient Ethernet
      - other:
         - qca8k: convert to regmap read/write API, use bulk operations
         - rswitch: Improve TX timestamp accuracy

   - Intel WiFi (iwlwifi):
      - EHT (Wi-Fi 7) rate reporting
      - STEP equalizer support: transfer some STEP (connection to radio
        on platforms with integrated wifi) related parameters from the
        BIOS to the firmware.

   - Qualcomm 802.11ax WiFi (ath11k):
      - IPQ5018 support
      - Fine Timing Measurement (FTM) responder role support
      - channel 177 support

   - MediaTek WiFi (mt76):
      - per-PHY LED support
      - mt7996: EHT (Wi-Fi 7) support
      - Wireless Ethernet Dispatch (WED) reset support
      - switch to using page pool allocator

   - RealTek WiFi (rtw89):
      - support new version of Bluetooth co-existance

   - Mobile:
      - rmnet: support TX aggregation"

* tag 'net-next-6.3' of git://git.kernel.org/pub/scm/linux/kernel/git/netdev/net-next: (1872 commits)
  page_pool: add a comment explaining the fragment counter usage
  net: ethtool: fix __ethtool_dev_mm_supported() implementation
  ethtool: pse-pd: Fix double word in comments
  xsk: add linux/vmalloc.h to xsk.c
  sefltests: netdevsim: wait for devlink instance after netns removal
  selftest: fib_tests: Always cleanup before exit
  net/mlx5e: Align IPsec ASO result memory to be as required by hardware
  net/mlx5e: TC, Set CT miss to the specific ct action instance
  net/mlx5e: Rename CHAIN_TO_REG to MAPPED_OBJ_TO_REG
  net/mlx5: Refactor tc miss handling to a single function
  net/mlx5: Kconfig: Make tc offload depend on tc skb extension
  net/sched: flower: Support hardware miss to tc action
  net/sched: flower: Move filter handle initialization earlier
  net/sched: cls_api: Support hardware miss to tc action
  net/sched: Rename user cookie and act cookie
  sfc: fix builds without CONFIG_RTC_LIB
  sfc: clean up some inconsistent indentings
  net/mlx4_en: Introduce flexible array to silence overflow warning
  net: lan966x: Fix possible deadlock inside PTP
  net/ulp: Remove redundant ->clone() test in inet_clone_ulp().
  ...

1 files changed, 816 insertions, 0 deletions

diff --git a/net/rxrpc/conn_client.c b/net/rxrpc/conn_client.c
new file mode 100644
index 000000000..981ca5b98
--- /dev/null
+++ b/net/rxrpc/conn_client.c
@@ -0,0 +1,816 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/* Client connection-specific management code.
+ *
+ * Copyright (C) 2016, 2020 Red Hat, Inc. All Rights Reserved.
+ * Written by David Howells (dhowells@redhat.com)
+ *
+ * Client connections need to be cached for a little while after they've made a
+ * call so as to handle retransmitted DATA packets in case the server didn't
+ * receive the final ACK or terminating ABORT we sent it.
+ *
+ * There are flags of relevance to the cache:
+ *
+ *  (2) DONT_REUSE - The connection should be discarded as soon as possible and
+ *      should not be reused.  This is set when an exclusive connection is used
+ *      or a call ID counter overflows.
+ *
+ * The caching state may only be changed if the cache lock is held.
+ *
+ * There are two idle client connection expiry durations.  If the total number
+ * of connections is below the reap threshold, we use the normal duration; if
+ * it's above, we use the fast duration.
+ */
+
+#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt
+
+#include <linux/slab.h>
+#include <linux/idr.h>
+#include <linux/timer.h>
+#include <linux/sched/signal.h>
+
+#include "ar-internal.h"
+
+__read_mostly unsigned int rxrpc_reap_client_connections = 900;
+__read_mostly unsigned long rxrpc_conn_idle_client_expiry = 2 * 60 * HZ;
+__read_mostly unsigned long rxrpc_conn_idle_client_fast_expiry = 2 * HZ;
+
+static void rxrpc_activate_bundle(struct rxrpc_bundle *bundle)
+{
+	atomic_inc(&bundle->active);
+}
+
+/*
+ * Release a connection ID for a client connection.
+ */
+static void rxrpc_put_client_connection_id(struct rxrpc_local *local,
+					   struct rxrpc_connection *conn)
+{
+	idr_remove(&local->conn_ids, conn->proto.cid >> RXRPC_CIDSHIFT);
+}
+
+/*
+ * Destroy the client connection ID tree.
+ */
+static void rxrpc_destroy_client_conn_ids(struct rxrpc_local *local)
+{
+	struct rxrpc_connection *conn;
+	int id;
+
+	if (!idr_is_empty(&local->conn_ids)) {
+		idr_for_each_entry(&local->conn_ids, conn, id) {
+			pr_err("AF_RXRPC: Leaked client conn %p {%d}\n",
+			       conn, refcount_read(&conn->ref));
+		}
+		BUG();
+	}
+
+	idr_destroy(&local->conn_ids);
+}
+
+/*
+ * Allocate a connection bundle.
+ */
+static struct rxrpc_bundle *rxrpc_alloc_bundle(struct rxrpc_call *call,
+					       gfp_t gfp)
+{
+	struct rxrpc_bundle *bundle;
+
+	bundle = kzalloc(sizeof(*bundle), gfp);
+	if (bundle) {
+		bundle->local		= call->local;
+		bundle->peer		= rxrpc_get_peer(call->peer, rxrpc_peer_get_bundle);
+		bundle->key		= key_get(call->key);
+		bundle->security	= call->security;
+		bundle->exclusive	= test_bit(RXRPC_CALL_EXCLUSIVE, &call->flags);
+		bundle->upgrade		= test_bit(RXRPC_CALL_UPGRADE, &call->flags);
+		bundle->service_id	= call->dest_srx.srx_service;
+		bundle->security_level	= call->security_level;
+		refcount_set(&bundle->ref, 1);
+		atomic_set(&bundle->active, 1);
+		INIT_LIST_HEAD(&bundle->waiting_calls);
+		trace_rxrpc_bundle(bundle->debug_id, 1, rxrpc_bundle_new);
+	}
+	return bundle;
+}
+
+struct rxrpc_bundle *rxrpc_get_bundle(struct rxrpc_bundle *bundle,
+				      enum rxrpc_bundle_trace why)
+{
+	int r;
+
+	__refcount_inc(&bundle->ref, &r);
+	trace_rxrpc_bundle(bundle->debug_id, r + 1, why);
+	return bundle;
+}
+
+static void rxrpc_free_bundle(struct rxrpc_bundle *bundle)
+{
+	trace_rxrpc_bundle(bundle->debug_id, 1, rxrpc_bundle_free);
+	rxrpc_put_peer(bundle->peer, rxrpc_peer_put_bundle);
+	key_put(bundle->key);
+	kfree(bundle);
+}
+
+void rxrpc_put_bundle(struct rxrpc_bundle *bundle, enum rxrpc_bundle_trace why)
+{
+	unsigned int id;
+	bool dead;
+	int r;
+
+	if (bundle) {
+		id = bundle->debug_id;
+		dead = __refcount_dec_and_test(&bundle->ref, &r);
+		trace_rxrpc_bundle(id, r - 1, why);
+		if (dead)
+			rxrpc_free_bundle(bundle);
+	}
+}
+
+/*
+ * Get rid of outstanding client connection preallocations when a local
+ * endpoint is destroyed.
+ */
+void rxrpc_purge_client_connections(struct rxrpc_local *local)
+{
+	rxrpc_destroy_client_conn_ids(local);
+}
+
+/*
+ * Allocate a client connection.
+ */
+static struct rxrpc_connection *
+rxrpc_alloc_client_connection(struct rxrpc_bundle *bundle)
+{
+	struct rxrpc_connection *conn;
+	struct rxrpc_local *local = bundle->local;
+	struct rxrpc_net *rxnet = local->rxnet;
+	int id;
+
+	_enter("");
+
+	conn = rxrpc_alloc_connection(rxnet, GFP_ATOMIC | __GFP_NOWARN);
+	if (!conn)
+		return ERR_PTR(-ENOMEM);
+
+	id = idr_alloc_cyclic(&local->conn_ids, conn, 1, 0x40000000,
+			      GFP_ATOMIC | __GFP_NOWARN);
+	if (id < 0) {
+		kfree(conn);
+		return ERR_PTR(id);
+	}
+
+	refcount_set(&conn->ref, 1);
+	conn->proto.cid		= id << RXRPC_CIDSHIFT;
+	conn->proto.epoch	= local->rxnet->epoch;
+	conn->out_clientflag	= RXRPC_CLIENT_INITIATED;
+	conn->bundle		= rxrpc_get_bundle(bundle, rxrpc_bundle_get_client_conn);
+	conn->local		= rxrpc_get_local(bundle->local, rxrpc_local_get_client_conn);
+	conn->peer		= rxrpc_get_peer(bundle->peer, rxrpc_peer_get_client_conn);
+	conn->key		= key_get(bundle->key);
+	conn->security		= bundle->security;
+	conn->exclusive		= bundle->exclusive;
+	conn->upgrade		= bundle->upgrade;
+	conn->orig_service_id	= bundle->service_id;
+	conn->security_level	= bundle->security_level;
+	conn->state		= RXRPC_CONN_CLIENT_UNSECURED;
+	conn->service_id	= conn->orig_service_id;
+
+	if (conn->security == &rxrpc_no_security)
+		conn->state	= RXRPC_CONN_CLIENT;
+
+	atomic_inc(&rxnet->nr_conns);
+	write_lock(&rxnet->conn_lock);
+	list_add_tail(&conn->proc_link, &rxnet->conn_proc_list);
+	write_unlock(&rxnet->conn_lock);
+
+	rxrpc_see_connection(conn, rxrpc_conn_new_client);
+
+	atomic_inc(&rxnet->nr_client_conns);
+	trace_rxrpc_client(conn, -1, rxrpc_client_alloc);
+	return conn;
+}
+
+/*
+ * Determine if a connection may be reused.
+ */
+static bool rxrpc_may_reuse_conn(struct rxrpc_connection *conn)
+{
+	struct rxrpc_net *rxnet;
+	int id_cursor, id, distance, limit;
+
+	if (!conn)
+		goto dont_reuse;
+
+	rxnet = conn->rxnet;
+	if (test_bit(RXRPC_CONN_DONT_REUSE, &conn->flags))
+		goto dont_reuse;
+
+	if ((conn->state != RXRPC_CONN_CLIENT_UNSECURED &&
+	     conn->state != RXRPC_CONN_CLIENT) ||
+	    conn->proto.epoch != rxnet->epoch)
+		goto mark_dont_reuse;
+
+	/* The IDR tree gets very expensive on memory if the connection IDs are
+	 * widely scattered throughout the number space, so we shall want to
+	 * kill off connections that, say, have an ID more than about four
+	 * times the maximum number of client conns away from the current
+	 * allocation point to try and keep the IDs concentrated.
+	 */
+	id_cursor = idr_get_cursor(&conn->local->conn_ids);
+	id = conn->proto.cid >> RXRPC_CIDSHIFT;
+	distance = id - id_cursor;
+	if (distance < 0)
+		distance = -distance;
+	limit = max_t(unsigned long, atomic_read(&rxnet->nr_conns) * 4, 1024);
+	if (distance > limit)
+		goto mark_dont_reuse;
+
+	return true;
+
+mark_dont_reuse:
+	set_bit(RXRPC_CONN_DONT_REUSE, &conn->flags);
+dont_reuse:
+	return false;
+}
+
+/*
+ * Look up the conn bundle that matches the connection parameters, adding it if
+ * it doesn't yet exist.
+ */
+int rxrpc_look_up_bundle(struct rxrpc_call *call, gfp_t gfp)
+{
+	static atomic_t rxrpc_bundle_id;
+	struct rxrpc_bundle *bundle, *candidate;
+	struct rxrpc_local *local = call->local;
+	struct rb_node *p, **pp, *parent;
+	long diff;
+	bool upgrade = test_bit(RXRPC_CALL_UPGRADE, &call->flags);
+
+	_enter("{%px,%x,%u,%u}",
+	       call->peer, key_serial(call->key), call->security_level,
+	       upgrade);
+
+	if (test_bit(RXRPC_CALL_EXCLUSIVE, &call->flags)) {
+		call->bundle = rxrpc_alloc_bundle(call, gfp);
+		return call->bundle ? 0 : -ENOMEM;
+	}
+
+	/* First, see if the bundle is already there. */
+	_debug("search 1");
+	spin_lock(&local->client_bundles_lock);
+	p = local->client_bundles.rb_node;
+	while (p) {
+		bundle = rb_entry(p, struct rxrpc_bundle, local_node);
+
+#define cmp(X, Y) ((long)(X) - (long)(Y))
+		diff = (cmp(bundle->peer, call->peer) ?:
+			cmp(bundle->key, call->key) ?:
+			cmp(bundle->security_level, call->security_level) ?:
+			cmp(bundle->upgrade, upgrade));
+#undef cmp
+		if (diff < 0)
+			p = p->rb_left;
+		else if (diff > 0)
+			p = p->rb_right;
+		else
+			goto found_bundle;
+	}
+	spin_unlock(&local->client_bundles_lock);
+	_debug("not found");
+
+	/* It wasn't.  We need to add one. */
+	candidate = rxrpc_alloc_bundle(call, gfp);
+	if (!candidate)
+		return -ENOMEM;
+
+	_debug("search 2");
+	spin_lock(&local->client_bundles_lock);
+	pp = &local->client_bundles.rb_node;
+	parent = NULL;
+	while (*pp) {
+		parent = *pp;
+		bundle = rb_entry(parent, struct rxrpc_bundle, local_node);
+
+#define cmp(X, Y) ((long)(X) - (long)(Y))
+		diff = (cmp(bundle->peer, call->peer) ?:
+			cmp(bundle->key, call->key) ?:
+			cmp(bundle->security_level, call->security_level) ?:
+			cmp(bundle->upgrade, upgrade));
+#undef cmp
+		if (diff < 0)
+			pp = &(*pp)->rb_left;
+		else if (diff > 0)
+			pp = &(*pp)->rb_right;
+		else
+			goto found_bundle_free;
+	}
+
+	_debug("new bundle");
+	candidate->debug_id = atomic_inc_return(&rxrpc_bundle_id);
+	rb_link_node(&candidate->local_node, parent, pp);
+	rb_insert_color(&candidate->local_node, &local->client_bundles);
+	call->bundle = rxrpc_get_bundle(candidate, rxrpc_bundle_get_client_call);
+	spin_unlock(&local->client_bundles_lock);
+	_leave(" = B=%u [new]", call->bundle->debug_id);
+	return 0;
+
+found_bundle_free:
+	rxrpc_free_bundle(candidate);
+found_bundle:
+	call->bundle = rxrpc_get_bundle(bundle, rxrpc_bundle_get_client_call);
+	rxrpc_activate_bundle(bundle);
+	spin_unlock(&local->client_bundles_lock);
+	_leave(" = B=%u [found]", call->bundle->debug_id);
+	return 0;
+}
+
+/*
+ * Allocate a new connection and add it into a bundle.
+ */
+static bool rxrpc_add_conn_to_bundle(struct rxrpc_bundle *bundle,
+				     unsigned int slot)
+{
+	struct rxrpc_connection *conn, *old;
+	unsigned int shift = slot * RXRPC_MAXCALLS;
+	unsigned int i;
+
+	old = bundle->conns[slot];
+	if (old) {
+		bundle->conns[slot] = NULL;
+		trace_rxrpc_client(old, -1, rxrpc_client_replace);
+		rxrpc_put_connection(old, rxrpc_conn_put_noreuse);
+	}
+
+	conn = rxrpc_alloc_client_connection(bundle);
+	if (IS_ERR(conn)) {
+		bundle->alloc_error = PTR_ERR(conn);
+		return false;
+	}
+
+	rxrpc_activate_bundle(bundle);
+	conn->bundle_shift = shift;
+	bundle->conns[slot] = conn;
+	for (i = 0; i < RXRPC_MAXCALLS; i++)
+		set_bit(shift + i, &bundle->avail_chans);
+	return true;
+}
+
+/*
+ * Add a connection to a bundle if there are no usable connections or we have
+ * connections waiting for extra capacity.
+ */
+static bool rxrpc_bundle_has_space(struct rxrpc_bundle *bundle)
+{
+	int slot = -1, i, usable;
+
+	_enter("");
+
+	bundle->alloc_error = 0;
+
+	/* See if there are any usable connections. */
+	usable = 0;
+	for (i = 0; i < ARRAY_SIZE(bundle->conns); i++) {
+		if (rxrpc_may_reuse_conn(bundle->conns[i]))
+			usable++;
+		else if (slot == -1)
+			slot = i;
+	}
+
+	if (!usable && bundle->upgrade)
+		bundle->try_upgrade = true;
+
+	if (!usable)
+		goto alloc_conn;
+
+	if (!bundle->avail_chans &&
+	    !bundle->try_upgrade &&
+	    usable < ARRAY_SIZE(bundle->conns))
+		goto alloc_conn;
+
+	_leave("");
+	return usable;
+
+alloc_conn:
+	return slot >= 0 ? rxrpc_add_conn_to_bundle(bundle, slot) : false;
+}
+
+/*
+ * Assign a channel to the call at the front of the queue and wake the call up.
+ * We don't increment the callNumber counter until this number has been exposed
+ * to the world.
+ */
+static void rxrpc_activate_one_channel(struct rxrpc_connection *conn,
+				       unsigned int channel)
+{
+	struct rxrpc_channel *chan = &conn->channels[channel];
+	struct rxrpc_bundle *bundle = conn->bundle;
+	struct rxrpc_call *call = list_entry(bundle->waiting_calls.next,
+					     struct rxrpc_call, wait_link);
+	u32 call_id = chan->call_counter + 1;
+
+	_enter("C=%x,%u", conn->debug_id, channel);
+
+	list_del_init(&call->wait_link);
+
+	trace_rxrpc_client(conn, channel, rxrpc_client_chan_activate);
+
+	/* Cancel the final ACK on the previous call if it hasn't been sent yet
+	 * as the DATA packet will implicitly ACK it.
+	 */
+	clear_bit(RXRPC_CONN_FINAL_ACK_0 + channel, &conn->flags);
+	clear_bit(conn->bundle_shift + channel, &bundle->avail_chans);
+
+	rxrpc_see_call(call, rxrpc_call_see_activate_client);
+	call->conn	= rxrpc_get_connection(conn, rxrpc_conn_get_activate_call);
+	call->cid	= conn->proto.cid | channel;
+	call->call_id	= call_id;
+	call->dest_srx.srx_service = conn->service_id;
+	call->cong_ssthresh = call->peer->cong_ssthresh;
+	if (call->cong_cwnd >= call->cong_ssthresh)
+		call->cong_mode = RXRPC_CALL_CONGEST_AVOIDANCE;
+	else
+		call->cong_mode = RXRPC_CALL_SLOW_START;
+
+	chan->call_id		= call_id;
+	chan->call_debug_id	= call->debug_id;
+	chan->call		= call;
+
+	rxrpc_see_call(call, rxrpc_call_see_connected);
+	trace_rxrpc_connect_call(call);
+	call->tx_last_sent = ktime_get_real();
+	rxrpc_start_call_timer(call);
+	rxrpc_set_call_state(call, RXRPC_CALL_CLIENT_SEND_REQUEST);
+	wake_up(&call->waitq);
+}
+
+/*
+ * Remove a connection from the idle list if it's on it.
+ */
+static void rxrpc_unidle_conn(struct rxrpc_connection *conn)
+{
+	if (!list_empty(&conn->cache_link)) {
+		list_del_init(&conn->cache_link);
+		rxrpc_put_connection(conn, rxrpc_conn_put_unidle);
+	}
+}
+
+/*
+ * Assign channels and callNumbers to waiting calls.
+ */
+static void rxrpc_activate_channels(struct rxrpc_bundle *bundle)
+{
+	struct rxrpc_connection *conn;
+	unsigned long avail, mask;
+	unsigned int channel, slot;
+
+	trace_rxrpc_client(NULL, -1, rxrpc_client_activate_chans);
+
+	if (bundle->try_upgrade)
+		mask = 1;
+	else
+		mask = ULONG_MAX;
+
+	while (!list_empty(&bundle->waiting_calls)) {
+		avail = bundle->avail_chans & mask;
+		if (!avail)
+			break;
+		channel = __ffs(avail);
+		clear_bit(channel, &bundle->avail_chans);
+
+		slot = channel / RXRPC_MAXCALLS;
+		conn = bundle->conns[slot];
+		if (!conn)
+			break;
+
+		if (bundle->try_upgrade)
+			set_bit(RXRPC_CONN_PROBING_FOR_UPGRADE, &conn->flags);
+		rxrpc_unidle_conn(conn);
+
+		channel &= (RXRPC_MAXCALLS - 1);
+		conn->act_chans	|= 1 << channel;
+		rxrpc_activate_one_channel(conn, channel);
+	}
+}
+
+/*
+ * Connect waiting channels (called from the I/O thread).
+ */
+void rxrpc_connect_client_calls(struct rxrpc_local *local)
+{
+	struct rxrpc_call *call;
+
+	while ((call = list_first_entry_or_null(&local->new_client_calls,
+						struct rxrpc_call, wait_link))
+	       ) {
+		struct rxrpc_bundle *bundle = call->bundle;
+
+		spin_lock(&local->client_call_lock);
+		list_move_tail(&call->wait_link, &bundle->waiting_calls);
+		spin_unlock(&local->client_call_lock);
+
+		if (rxrpc_bundle_has_space(bundle))
+			rxrpc_activate_channels(bundle);
+	}
+}
+
+/*
+ * Note that a call, and thus a connection, is about to be exposed to the
+ * world.
+ */
+void rxrpc_expose_client_call(struct rxrpc_call *call)
+{
+	unsigned int channel = call->cid & RXRPC_CHANNELMASK;
+	struct rxrpc_connection *conn = call->conn;
+	struct rxrpc_channel *chan = &conn->channels[channel];
+
+	if (!test_and_set_bit(RXRPC_CALL_EXPOSED, &call->flags)) {
+		/* Mark the call ID as being used.  If the callNumber counter
+		 * exceeds ~2 billion, we kill the connection after its
+		 * outstanding calls have finished so that the counter doesn't
+		 * wrap.
+		 */
+		chan->call_counter++;
+		if (chan->call_counter >= INT_MAX)
+			set_bit(RXRPC_CONN_DONT_REUSE, &conn->flags);
+		trace_rxrpc_client(conn, channel, rxrpc_client_exposed);
+
+		spin_lock(&call->peer->lock);
+		hlist_add_head(&call->error_link, &call->peer->error_targets);
+		spin_unlock(&call->peer->lock);
+	}
+}
+
+/*
+ * Set the reap timer.
+ */
+static void rxrpc_set_client_reap_timer(struct rxrpc_local *local)
+{
+	if (!local->kill_all_client_conns) {
+		unsigned long now = jiffies;
+		unsigned long reap_at = now + rxrpc_conn_idle_client_expiry;
+
+		if (local->rxnet->live)
+			timer_reduce(&local->client_conn_reap_timer, reap_at);
+	}
+}
+
+/*
+ * Disconnect a client call.
+ */
+void rxrpc_disconnect_client_call(struct rxrpc_bundle *bundle, struct rxrpc_call *call)
+{
+	struct rxrpc_connection *conn;
+	struct rxrpc_channel *chan = NULL;
+	struct rxrpc_local *local = bundle->local;
+	unsigned int channel;
+	bool may_reuse;
+	u32 cid;
+
+	_enter("c=%x", call->debug_id);
+
+	/* Calls that have never actually been assigned a channel can simply be
+	 * discarded.
+	 */
+	conn = call->conn;
+	if (!conn) {
+		_debug("call is waiting");
+		ASSERTCMP(call->call_id, ==, 0);
+		ASSERT(!test_bit(RXRPC_CALL_EXPOSED, &call->flags));
+		list_del_init(&call->wait_link);
+		return;
+	}
+
+	cid = call->cid;
+	channel = cid & RXRPC_CHANNELMASK;
+	chan = &conn->channels[channel];
+	trace_rxrpc_client(conn, channel, rxrpc_client_chan_disconnect);
+
+	if (WARN_ON(chan->call != call))
+		return;
+
+	may_reuse = rxrpc_may_reuse_conn(conn);
+
+	/* If a client call was exposed to the world, we save the result for
+	 * retransmission.
+	 *
+	 * We use a barrier here so that the call number and abort code can be
+	 * read without needing to take a lock.
+	 *
+	 * TODO: Make the incoming packet handler check this and handle
+	 * terminal retransmission without requiring access to the call.
+	 */
+	if (test_bit(RXRPC_CALL_EXPOSED, &call->flags)) {
+		_debug("exposed %u,%u", call->call_id, call->abort_code);
+		__rxrpc_disconnect_call(conn, call);
+
+		if (test_and_clear_bit(RXRPC_CONN_PROBING_FOR_UPGRADE, &conn->flags)) {
+			trace_rxrpc_client(conn, channel, rxrpc_client_to_active);
+			bundle->try_upgrade = false;
+			if (may_reuse)
+				rxrpc_activate_channels(bundle);
+		}
+	}
+
+	/* See if we can pass the channel directly to another call. */
+	if (may_reuse && !list_empty(&bundle->waiting_calls)) {
+		trace_rxrpc_client(conn, channel, rxrpc_client_chan_pass);
+		rxrpc_activate_one_channel(conn, channel);
+		return;
+	}
+
+	/* Schedule the final ACK to be transmitted in a short while so that it
+	 * can be skipped if we find a follow-on call.  The first DATA packet
+	 * of the follow on call will implicitly ACK this call.
+	 */
+	if (call->completion == RXRPC_CALL_SUCCEEDED &&
+	    test_bit(RXRPC_CALL_EXPOSED, &call->flags)) {
+		unsigned long final_ack_at = jiffies + 2;
+
+		WRITE_ONCE(chan->final_ack_at, final_ack_at);
+		smp_wmb(); /* vs rxrpc_process_delayed_final_acks() */
+		set_bit(RXRPC_CONN_FINAL_ACK_0 + channel, &conn->flags);
+		rxrpc_reduce_conn_timer(conn, final_ack_at);
+	}
+
+	/* Deactivate the channel. */
+	chan->call = NULL;
+	set_bit(conn->bundle_shift + channel, &conn->bundle->avail_chans);
+	conn->act_chans	&= ~(1 << channel);
+
+	/* If no channels remain active, then put the connection on the idle
+	 * list for a short while.  Give it a ref to stop it going away if it
+	 * becomes unbundled.
+	 */
+	if (!conn->act_chans) {
+		trace_rxrpc_client(conn, channel, rxrpc_client_to_idle);
+		conn->idle_timestamp = jiffies;
+
+		rxrpc_get_connection(conn, rxrpc_conn_get_idle);
+		list_move_tail(&conn->cache_link, &local->idle_client_conns);
+
+		rxrpc_set_client_reap_timer(local);
+	}
+}
+
+/*
+ * Remove a connection from a bundle.
+ */
+static void rxrpc_unbundle_conn(struct rxrpc_connection *conn)
+{
+	struct rxrpc_bundle *bundle = conn->bundle;
+	unsigned int bindex;
+	int i;
+
+	_enter("C=%x", conn->debug_id);
+
+	if (conn->flags & RXRPC_CONN_FINAL_ACK_MASK)
+		rxrpc_process_delayed_final_acks(conn, true);
+
+	bindex = conn->bundle_shift / RXRPC_MAXCALLS;
+	if (bundle->conns[bindex] == conn) {
+		_debug("clear slot %u", bindex);
+		bundle->conns[bindex] = NULL;
+		for (i = 0; i < RXRPC_MAXCALLS; i++)
+			clear_bit(conn->bundle_shift + i, &bundle->avail_chans);
+		rxrpc_put_client_connection_id(bundle->local, conn);
+		rxrpc_deactivate_bundle(bundle);
+		rxrpc_put_connection(conn, rxrpc_conn_put_unbundle);
+	}
+}
+
+/*
+ * Drop the active count on a bundle.
+ */
+void rxrpc_deactivate_bundle(struct rxrpc_bundle *bundle)
+{
+	struct rxrpc_local *local;
+	bool need_put = false;
+
+	if (!bundle)
+		return;
+
+	local = bundle->local;
+	if (atomic_dec_and_lock(&bundle->active, &local->client_bundles_lock)) {
+		if (!bundle->exclusive) {
+			_debug("erase bundle");
+			rb_erase(&bundle->local_node, &local->client_bundles);
+			need_put = true;
+		}
+
+		spin_unlock(&local->client_bundles_lock);
+		if (need_put)
+			rxrpc_put_bundle(bundle, rxrpc_bundle_put_discard);
+	}
+}
+
+/*
+ * Clean up a dead client connection.
+ */
+void rxrpc_kill_client_conn(struct rxrpc_connection *conn)
+{
+	struct rxrpc_local *local = conn->local;
+	struct rxrpc_net *rxnet = local->rxnet;
+
+	_enter("C=%x", conn->debug_id);
+
+	trace_rxrpc_client(conn, -1, rxrpc_client_cleanup);
+	atomic_dec(&rxnet->nr_client_conns);
+
+	rxrpc_put_client_connection_id(local, conn);
+}
+
+/*
+ * Discard expired client connections from the idle list.  Each conn in the
+ * idle list has been exposed and holds an extra ref because of that.
+ *
+ * This may be called from conn setup or from a work item so cannot be
+ * considered non-reentrant.
+ */
+void rxrpc_discard_expired_client_conns(struct rxrpc_local *local)
+{
+	struct rxrpc_connection *conn;
+	unsigned long expiry, conn_expires_at, now;
+	unsigned int nr_conns;
+
+	_enter("");
+
+	/* We keep an estimate of what the number of conns ought to be after
+	 * we've discarded some so that we don't overdo the discarding.
+	 */
+	nr_conns = atomic_read(&local->rxnet->nr_client_conns);
+
+next:
+	conn = list_first_entry_or_null(&local->idle_client_conns,
+					struct rxrpc_connection, cache_link);
+	if (!conn)
+		return;
+
+	if (!local->kill_all_client_conns) {
+		/* If the number of connections is over the reap limit, we
+		 * expedite discard by reducing the expiry timeout.  We must,
+		 * however, have at least a short grace period to be able to do
+		 * final-ACK or ABORT retransmission.
+		 */
+		expiry = rxrpc_conn_idle_client_expiry;
+		if (nr_conns > rxrpc_reap_client_connections)
+			expiry = rxrpc_conn_idle_client_fast_expiry;
+		if (conn->local->service_closed)
+			expiry = rxrpc_closed_conn_expiry * HZ;
+
+		conn_expires_at = conn->idle_timestamp + expiry;
+
+		now = READ_ONCE(jiffies);
+		if (time_after(conn_expires_at, now))
+			goto not_yet_expired;
+	}
+
+	atomic_dec(&conn->active);
+	trace_rxrpc_client(conn, -1, rxrpc_client_discard);
+	list_del_init(&conn->cache_link);
+
+	rxrpc_unbundle_conn(conn);
+	/* Drop the ->cache_link ref */
+	rxrpc_put_connection(conn, rxrpc_conn_put_discard_idle);
+
+	nr_conns--;
+	goto next;
+
+not_yet_expired:
+	/* The connection at the front of the queue hasn't yet expired, so
+	 * schedule the work item for that point if we discarded something.
+	 *
+	 * We don't worry if the work item is already scheduled - it can look
+	 * after rescheduling itself at a later time.  We could cancel it, but
+	 * then things get messier.
+	 */
+	_debug("not yet");
+	if (!local->kill_all_client_conns)
+		timer_reduce(&local->client_conn_reap_timer, conn_expires_at);
+
+	_leave("");
+}
+
+/*
+ * Clean up the client connections on a local endpoint.
+ */
+void rxrpc_clean_up_local_conns(struct rxrpc_local *local)
+{
+	struct rxrpc_connection *conn;
+
+	_enter("");
+
+	local->kill_all_client_conns = true;
+
+	del_timer_sync(&local->client_conn_reap_timer);
+
+	while ((conn = list_first_entry_or_null(&local->idle_client_conns,
+						struct rxrpc_connection, cache_link))) {
+		list_del_init(&conn->cache_link);
+		atomic_dec(&conn->active);
+		trace_rxrpc_client(conn, -1, rxrpc_client_discard);
+		rxrpc_unbundle_conn(conn);
+		rxrpc_put_connection(conn, rxrpc_conn_put_local_dead);
+	}
+
+	_leave(" [culled]");
+}