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, 759 insertions, 0 deletions

diff --git a/drivers/md/persistent-data/dm-btree-remove.c b/drivers/md/persistent-data/dm-btree-remove.c
new file mode 100644
index 000000000..4ead31e0d
--- /dev/null
+++ b/drivers/md/persistent-data/dm-btree-remove.c
@@ -0,0 +1,759 @@
+/*
+ * Copyright (C) 2011 Red Hat, Inc.
+ *
+ * This file is released under the GPL.
+ */
+
+#include "dm-btree.h"
+#include "dm-btree-internal.h"
+#include "dm-transaction-manager.h"
+
+#include <linux/export.h>
+#include <linux/device-mapper.h>
+
+#define DM_MSG_PREFIX "btree"
+
+/*
+ * Removing an entry from a btree
+ * ==============================
+ *
+ * A very important constraint for our btree is that no node, except the
+ * root, may have fewer than a certain number of entries.
+ * (MIN_ENTRIES <= nr_entries <= MAX_ENTRIES).
+ *
+ * Ensuring this is complicated by the way we want to only ever hold the
+ * locks on 2 nodes concurrently, and only change nodes in a top to bottom
+ * fashion.
+ *
+ * Each node may have a left or right sibling.  When decending the spine,
+ * if a node contains only MIN_ENTRIES then we try and increase this to at
+ * least MIN_ENTRIES + 1.  We do this in the following ways:
+ *
+ * [A] No siblings => this can only happen if the node is the root, in which
+ *     case we copy the childs contents over the root.
+ *
+ * [B] No left sibling
+ *     ==> rebalance(node, right sibling)
+ *
+ * [C] No right sibling
+ *     ==> rebalance(left sibling, node)
+ *
+ * [D] Both siblings, total_entries(left, node, right) <= DEL_THRESHOLD
+ *     ==> delete node adding it's contents to left and right
+ *
+ * [E] Both siblings, total_entries(left, node, right) > DEL_THRESHOLD
+ *     ==> rebalance(left, node, right)
+ *
+ * After these operations it's possible that the our original node no
+ * longer contains the desired sub tree.  For this reason this rebalancing
+ * is performed on the children of the current node.  This also avoids
+ * having a special case for the root.
+ *
+ * Once this rebalancing has occurred we can then step into the child node
+ * for internal nodes.  Or delete the entry for leaf nodes.
+ */
+
+/*
+ * Some little utilities for moving node data around.
+ */
+static void node_shift(struct btree_node *n, int shift)
+{
+	uint32_t nr_entries = le32_to_cpu(n->header.nr_entries);
+	uint32_t value_size = le32_to_cpu(n->header.value_size);
+
+	if (shift < 0) {
+		shift = -shift;
+		BUG_ON(shift > nr_entries);
+		BUG_ON((void *) key_ptr(n, shift) >= value_ptr(n, shift));
+		memmove(key_ptr(n, 0),
+			key_ptr(n, shift),
+			(nr_entries - shift) * sizeof(__le64));
+		memmove(value_ptr(n, 0),
+			value_ptr(n, shift),
+			(nr_entries - shift) * value_size);
+	} else {
+		BUG_ON(nr_entries + shift > le32_to_cpu(n->header.max_entries));
+		memmove(key_ptr(n, shift),
+			key_ptr(n, 0),
+			nr_entries * sizeof(__le64));
+		memmove(value_ptr(n, shift),
+			value_ptr(n, 0),
+			nr_entries * value_size);
+	}
+}
+
+static int node_copy(struct btree_node *left, struct btree_node *right, int shift)
+{
+	uint32_t nr_left = le32_to_cpu(left->header.nr_entries);
+	uint32_t value_size = le32_to_cpu(left->header.value_size);
+	if (value_size != le32_to_cpu(right->header.value_size)) {
+		DMERR("mismatched value size");
+		return -EILSEQ;
+	}
+
+	if (shift < 0) {
+		shift = -shift;
+
+		if (nr_left + shift > le32_to_cpu(left->header.max_entries)) {
+			DMERR("bad shift");
+			return -EINVAL;
+		}
+
+		memcpy(key_ptr(left, nr_left),
+		       key_ptr(right, 0),
+		       shift * sizeof(__le64));
+		memcpy(value_ptr(left, nr_left),
+		       value_ptr(right, 0),
+		       shift * value_size);
+	} else {
+		if (shift > le32_to_cpu(right->header.max_entries)) {
+			DMERR("bad shift");
+			return -EINVAL;
+		}
+
+		memcpy(key_ptr(right, 0),
+		       key_ptr(left, nr_left - shift),
+		       shift * sizeof(__le64));
+		memcpy(value_ptr(right, 0),
+		       value_ptr(left, nr_left - shift),
+		       shift * value_size);
+	}
+	return 0;
+}
+
+/*
+ * Delete a specific entry from a leaf node.
+ */
+static void delete_at(struct btree_node *n, unsigned index)
+{
+	unsigned nr_entries = le32_to_cpu(n->header.nr_entries);
+	unsigned nr_to_copy = nr_entries - (index + 1);
+	uint32_t value_size = le32_to_cpu(n->header.value_size);
+	BUG_ON(index >= nr_entries);
+
+	if (nr_to_copy) {
+		memmove(key_ptr(n, index),
+			key_ptr(n, index + 1),
+			nr_to_copy * sizeof(__le64));
+
+		memmove(value_ptr(n, index),
+			value_ptr(n, index + 1),
+			nr_to_copy * value_size);
+	}
+
+	n->header.nr_entries = cpu_to_le32(nr_entries - 1);
+}
+
+static unsigned merge_threshold(struct btree_node *n)
+{
+	return le32_to_cpu(n->header.max_entries) / 3;
+}
+
+struct child {
+	unsigned index;
+	struct dm_block *block;
+	struct btree_node *n;
+};
+
+static int init_child(struct dm_btree_info *info, struct dm_btree_value_type *vt,
+		      struct btree_node *parent,
+		      unsigned index, struct child *result)
+{
+	int r, inc;
+	dm_block_t root;
+
+	result->index = index;
+	root = value64(parent, index);
+
+	r = dm_tm_shadow_block(info->tm, root, &btree_node_validator,
+			       &result->block, &inc);
+	if (r)
+		return r;
+
+	result->n = dm_block_data(result->block);
+
+	if (inc)
+		inc_children(info->tm, result->n, vt);
+
+	*((__le64 *) value_ptr(parent, index)) =
+		cpu_to_le64(dm_block_location(result->block));
+
+	return 0;
+}
+
+static void exit_child(struct dm_btree_info *info, struct child *c)
+{
+	dm_tm_unlock(info->tm, c->block);
+}
+
+static int shift(struct btree_node *left, struct btree_node *right, int count)
+{
+	int r;
+	uint32_t nr_left = le32_to_cpu(left->header.nr_entries);
+	uint32_t nr_right = le32_to_cpu(right->header.nr_entries);
+	uint32_t max_entries = le32_to_cpu(left->header.max_entries);
+	uint32_t r_max_entries = le32_to_cpu(right->header.max_entries);
+
+	if (max_entries != r_max_entries) {
+		DMERR("node max_entries mismatch");
+		return -EILSEQ;
+	}
+
+	if (nr_left - count > max_entries) {
+		DMERR("node shift out of bounds");
+		return -EINVAL;
+	}
+
+	if (nr_right + count > max_entries) {
+		DMERR("node shift out of bounds");
+		return -EINVAL;
+	}
+
+	if (!count)
+		return 0;
+
+	if (count > 0) {
+		node_shift(right, count);
+		r = node_copy(left, right, count);
+		if (r)
+			return r;
+	} else {
+		r = node_copy(left, right, count);
+		if (r)
+			return r;
+		node_shift(right, count);
+	}
+
+	left->header.nr_entries = cpu_to_le32(nr_left - count);
+	right->header.nr_entries = cpu_to_le32(nr_right + count);
+
+	return 0;
+}
+
+static int __rebalance2(struct dm_btree_info *info, struct btree_node *parent,
+			struct child *l, struct child *r)
+{
+	int ret;
+	struct btree_node *left = l->n;
+	struct btree_node *right = r->n;
+	uint32_t nr_left = le32_to_cpu(left->header.nr_entries);
+	uint32_t nr_right = le32_to_cpu(right->header.nr_entries);
+	/*
+	 * Ensure the number of entries in each child will be greater
+	 * than or equal to (max_entries / 3 + 1), so no matter which
+	 * child is used for removal, the number will still be not
+	 * less than (max_entries / 3).
+	 */
+	unsigned int threshold = 2 * (merge_threshold(left) + 1);
+
+	if (nr_left + nr_right < threshold) {
+		/*
+		 * Merge
+		 */
+		node_copy(left, right, -nr_right);
+		left->header.nr_entries = cpu_to_le32(nr_left + nr_right);
+		delete_at(parent, r->index);
+
+		/*
+		 * We need to decrement the right block, but not it's
+		 * children, since they're still referenced by left.
+		 */
+		dm_tm_dec(info->tm, dm_block_location(r->block));
+	} else {
+		/*
+		 * Rebalance.
+		 */
+		unsigned target_left = (nr_left + nr_right) / 2;
+		ret = shift(left, right, nr_left - target_left);
+		if (ret)
+			return ret;
+		*key_ptr(parent, r->index) = right->keys[0];
+	}
+	return 0;
+}
+
+static int rebalance2(struct shadow_spine *s, struct dm_btree_info *info,
+		      struct dm_btree_value_type *vt, unsigned left_index)
+{
+	int r;
+	struct btree_node *parent;
+	struct child left, right;
+
+	parent = dm_block_data(shadow_current(s));
+
+	r = init_child(info, vt, parent, left_index, &left);
+	if (r)
+		return r;
+
+	r = init_child(info, vt, parent, left_index + 1, &right);
+	if (r) {
+		exit_child(info, &left);
+		return r;
+	}
+
+	r = __rebalance2(info, parent, &left, &right);
+
+	exit_child(info, &left);
+	exit_child(info, &right);
+
+	return r;
+}
+
+/*
+ * We dump as many entries from center as possible into left, then the rest
+ * in right, then rebalance2.  This wastes some cpu, but I want something
+ * simple atm.
+ */
+static int delete_center_node(struct dm_btree_info *info, struct btree_node *parent,
+			      struct child *l, struct child *c, struct child *r,
+			      struct btree_node *left, struct btree_node *center, struct btree_node *right,
+			      uint32_t nr_left, uint32_t nr_center, uint32_t nr_right)
+{
+	uint32_t max_entries = le32_to_cpu(left->header.max_entries);
+	unsigned shift = min(max_entries - nr_left, nr_center);
+
+	if (nr_left + shift > max_entries) {
+		DMERR("node shift out of bounds");
+		return -EINVAL;
+	}
+
+	node_copy(left, center, -shift);
+	left->header.nr_entries = cpu_to_le32(nr_left + shift);
+
+	if (shift != nr_center) {
+		shift = nr_center - shift;
+
+		if ((nr_right + shift) > max_entries) {
+			DMERR("node shift out of bounds");
+			return -EINVAL;
+		}
+
+		node_shift(right, shift);
+		node_copy(center, right, shift);
+		right->header.nr_entries = cpu_to_le32(nr_right + shift);
+	}
+	*key_ptr(parent, r->index) = right->keys[0];
+
+	delete_at(parent, c->index);
+	r->index--;
+
+	dm_tm_dec(info->tm, dm_block_location(c->block));
+	return __rebalance2(info, parent, l, r);
+}
+
+/*
+ * Redistributes entries among 3 sibling nodes.
+ */
+static int redistribute3(struct dm_btree_info *info, struct btree_node *parent,
+			 struct child *l, struct child *c, struct child *r,
+			 struct btree_node *left, struct btree_node *center, struct btree_node *right,
+			 uint32_t nr_left, uint32_t nr_center, uint32_t nr_right)
+{
+	int s, ret;
+	uint32_t max_entries = le32_to_cpu(left->header.max_entries);
+	unsigned total = nr_left + nr_center + nr_right;
+	unsigned target_right = total / 3;
+	unsigned remainder = (target_right * 3) != total;
+	unsigned target_left = target_right + remainder;
+
+	BUG_ON(target_left > max_entries);
+	BUG_ON(target_right > max_entries);
+
+	if (nr_left < nr_right) {
+		s = nr_left - target_left;
+
+		if (s < 0 && nr_center < -s) {
+			/* not enough in central node */
+			ret = shift(left, center, -nr_center);
+			if (ret)
+				return ret;
+
+			s += nr_center;
+			ret = shift(left, right, s);
+			if (ret)
+				return ret;
+
+			nr_right += s;
+		} else {
+			ret = shift(left, center, s);
+			if (ret)
+				return ret;
+		}
+
+		ret = shift(center, right, target_right - nr_right);
+		if (ret)
+			return ret;
+	} else {
+		s = target_right - nr_right;
+		if (s > 0 && nr_center < s) {
+			/* not enough in central node */
+			ret = shift(center, right, nr_center);
+			if (ret)
+				return ret;
+			s -= nr_center;
+			ret = shift(left, right, s);
+			if (ret)
+				return ret;
+			nr_left -= s;
+		} else {
+			ret = shift(center, right, s);
+			if (ret)
+				return ret;
+		}
+
+		ret = shift(left, center, nr_left - target_left);
+		if (ret)
+			return ret;
+	}
+
+	*key_ptr(parent, c->index) = center->keys[0];
+	*key_ptr(parent, r->index) = right->keys[0];
+	return 0;
+}
+
+static int __rebalance3(struct dm_btree_info *info, struct btree_node *parent,
+			struct child *l, struct child *c, struct child *r)
+{
+	struct btree_node *left = l->n;
+	struct btree_node *center = c->n;
+	struct btree_node *right = r->n;
+
+	uint32_t nr_left = le32_to_cpu(left->header.nr_entries);
+	uint32_t nr_center = le32_to_cpu(center->header.nr_entries);
+	uint32_t nr_right = le32_to_cpu(right->header.nr_entries);
+
+	unsigned threshold = merge_threshold(left) * 4 + 1;
+
+	if ((left->header.max_entries != center->header.max_entries) ||
+	    (center->header.max_entries != right->header.max_entries)) {
+		DMERR("bad btree metadata, max_entries differ");
+		return -EILSEQ;
+	}
+
+	if ((nr_left + nr_center + nr_right) < threshold) {
+		return delete_center_node(info, parent, l, c, r, left, center, right,
+					  nr_left, nr_center, nr_right);
+	}
+
+	return redistribute3(info, parent, l, c, r, left, center, right,
+			     nr_left, nr_center, nr_right);
+}
+
+static int rebalance3(struct shadow_spine *s, struct dm_btree_info *info,
+		      struct dm_btree_value_type *vt, unsigned left_index)
+{
+	int r;
+	struct btree_node *parent = dm_block_data(shadow_current(s));
+	struct child left, center, right;
+
+	/*
+	 * FIXME: fill out an array?
+	 */
+	r = init_child(info, vt, parent, left_index, &left);
+	if (r)
+		return r;
+
+	r = init_child(info, vt, parent, left_index + 1, &center);
+	if (r) {
+		exit_child(info, &left);
+		return r;
+	}
+
+	r = init_child(info, vt, parent, left_index + 2, &right);
+	if (r) {
+		exit_child(info, &left);
+		exit_child(info, &center);
+		return r;
+	}
+
+	r = __rebalance3(info, parent, &left, &center, &right);
+
+	exit_child(info, &left);
+	exit_child(info, &center);
+	exit_child(info, &right);
+
+	return r;
+}
+
+static int rebalance_children(struct shadow_spine *s,
+			      struct dm_btree_info *info,
+			      struct dm_btree_value_type *vt, uint64_t key)
+{
+	int i, r, has_left_sibling, has_right_sibling;
+	struct btree_node *n;
+
+	n = dm_block_data(shadow_current(s));
+
+	if (le32_to_cpu(n->header.nr_entries) == 1) {
+		struct dm_block *child;
+		dm_block_t b = value64(n, 0);
+
+		r = dm_tm_read_lock(info->tm, b, &btree_node_validator, &child);
+		if (r)
+			return r;
+
+		memcpy(n, dm_block_data(child),
+		       dm_bm_block_size(dm_tm_get_bm(info->tm)));
+
+		dm_tm_dec(info->tm, dm_block_location(child));
+		dm_tm_unlock(info->tm, child);
+		return 0;
+	}
+
+	i = lower_bound(n, key);
+	if (i < 0)
+		return -ENODATA;
+
+	has_left_sibling = i > 0;
+	has_right_sibling = i < (le32_to_cpu(n->header.nr_entries) - 1);
+
+	if (!has_left_sibling)
+		r = rebalance2(s, info, vt, i);
+
+	else if (!has_right_sibling)
+		r = rebalance2(s, info, vt, i - 1);
+
+	else
+		r = rebalance3(s, info, vt, i - 1);
+
+	return r;
+}
+
+static int do_leaf(struct btree_node *n, uint64_t key, unsigned *index)
+{
+	int i = lower_bound(n, key);
+
+	if ((i < 0) ||
+	    (i >= le32_to_cpu(n->header.nr_entries)) ||
+	    (le64_to_cpu(n->keys[i]) != key))
+		return -ENODATA;
+
+	*index = i;
+
+	return 0;
+}
+
+/*
+ * Prepares for removal from one level of the hierarchy.  The caller must
+ * call delete_at() to remove the entry at index.
+ */
+static int remove_raw(struct shadow_spine *s, struct dm_btree_info *info,
+		      struct dm_btree_value_type *vt, dm_block_t root,
+		      uint64_t key, unsigned *index)
+{
+	int i = *index, r;
+	struct btree_node *n;
+
+	for (;;) {
+		r = shadow_step(s, root, vt);
+		if (r < 0)
+			break;
+
+		/*
+		 * We have to patch up the parent node, ugly, but I don't
+		 * see a way to do this automatically as part of the spine
+		 * op.
+		 */
+		if (shadow_has_parent(s)) {
+			__le64 location = cpu_to_le64(dm_block_location(shadow_current(s)));
+			memcpy(value_ptr(dm_block_data(shadow_parent(s)), i),
+			       &location, sizeof(__le64));
+		}
+
+		n = dm_block_data(shadow_current(s));
+
+		if (le32_to_cpu(n->header.flags) & LEAF_NODE)
+			return do_leaf(n, key, index);
+
+		r = rebalance_children(s, info, vt, key);
+		if (r)
+			break;
+
+		n = dm_block_data(shadow_current(s));
+		if (le32_to_cpu(n->header.flags) & LEAF_NODE)
+			return do_leaf(n, key, index);
+
+		i = lower_bound(n, key);
+
+		/*
+		 * We know the key is present, or else
+		 * rebalance_children would have returned
+		 * -ENODATA
+		 */
+		root = value64(n, i);
+	}
+
+	return r;
+}
+
+int dm_btree_remove(struct dm_btree_info *info, dm_block_t root,
+		    uint64_t *keys, dm_block_t *new_root)
+{
+	unsigned level, last_level = info->levels - 1;
+	int index = 0, r = 0;
+	struct shadow_spine spine;
+	struct btree_node *n;
+	struct dm_btree_value_type le64_vt;
+
+	init_le64_type(info->tm, &le64_vt);
+	init_shadow_spine(&spine, info);
+	for (level = 0; level < info->levels; level++) {
+		r = remove_raw(&spine, info,
+			       (level == last_level ?
+				&info->value_type : &le64_vt),
+			       root, keys[level], (unsigned *)&index);
+		if (r < 0)
+			break;
+
+		n = dm_block_data(shadow_current(&spine));
+		if (level != last_level) {
+			root = value64(n, index);
+			continue;
+		}
+
+		BUG_ON(index < 0 || index >= le32_to_cpu(n->header.nr_entries));
+
+		if (info->value_type.dec)
+			info->value_type.dec(info->value_type.context,
+					     value_ptr(n, index), 1);
+
+		delete_at(n, index);
+	}
+
+	if (!r)
+		*new_root = shadow_root(&spine);
+	exit_shadow_spine(&spine);
+
+	return r;
+}
+EXPORT_SYMBOL_GPL(dm_btree_remove);
+
+/*----------------------------------------------------------------*/
+
+static int remove_nearest(struct shadow_spine *s, struct dm_btree_info *info,
+			  struct dm_btree_value_type *vt, dm_block_t root,
+			  uint64_t key, int *index)
+{
+	int i = *index, r;
+	struct btree_node *n;
+
+	for (;;) {
+		r = shadow_step(s, root, vt);
+		if (r < 0)
+			break;
+
+		/*
+		 * We have to patch up the parent node, ugly, but I don't
+		 * see a way to do this automatically as part of the spine
+		 * op.
+		 */
+		if (shadow_has_parent(s)) {
+			__le64 location = cpu_to_le64(dm_block_location(shadow_current(s)));
+			memcpy(value_ptr(dm_block_data(shadow_parent(s)), i),
+			       &location, sizeof(__le64));
+		}
+
+		n = dm_block_data(shadow_current(s));
+
+		if (le32_to_cpu(n->header.flags) & LEAF_NODE) {
+			*index = lower_bound(n, key);
+			return 0;
+		}
+
+		r = rebalance_children(s, info, vt, key);
+		if (r)
+			break;
+
+		n = dm_block_data(shadow_current(s));
+		if (le32_to_cpu(n->header.flags) & LEAF_NODE) {
+			*index = lower_bound(n, key);
+			return 0;
+		}
+
+		i = lower_bound(n, key);
+
+		/*
+		 * We know the key is present, or else
+		 * rebalance_children would have returned
+		 * -ENODATA
+		 */
+		root = value64(n, i);
+	}
+
+	return r;
+}
+
+static int remove_one(struct dm_btree_info *info, dm_block_t root,
+		      uint64_t *keys, uint64_t end_key,
+		      dm_block_t *new_root, unsigned *nr_removed)
+{
+	unsigned level, last_level = info->levels - 1;
+	int index = 0, r = 0;
+	struct shadow_spine spine;
+	struct btree_node *n;
+	struct dm_btree_value_type le64_vt;
+	uint64_t k;
+
+	init_le64_type(info->tm, &le64_vt);
+	init_shadow_spine(&spine, info);
+	for (level = 0; level < last_level; level++) {
+		r = remove_raw(&spine, info, &le64_vt,
+			       root, keys[level], (unsigned *) &index);
+		if (r < 0)
+			goto out;
+
+		n = dm_block_data(shadow_current(&spine));
+		root = value64(n, index);
+	}
+
+	r = remove_nearest(&spine, info, &info->value_type,
+			   root, keys[last_level], &index);
+	if (r < 0)
+		goto out;
+
+	n = dm_block_data(shadow_current(&spine));
+
+	if (index < 0)
+		index = 0;
+
+	if (index >= le32_to_cpu(n->header.nr_entries)) {
+		r = -ENODATA;
+		goto out;
+	}
+
+	k = le64_to_cpu(n->keys[index]);
+	if (k >= keys[last_level] && k < end_key) {
+		if (info->value_type.dec)
+			info->value_type.dec(info->value_type.context,
+					     value_ptr(n, index), 1);
+
+		delete_at(n, index);
+		keys[last_level] = k + 1ull;
+
+	} else
+		r = -ENODATA;
+
+out:
+	*new_root = shadow_root(&spine);
+	exit_shadow_spine(&spine);
+
+	return r;
+}
+
+int dm_btree_remove_leaves(struct dm_btree_info *info, dm_block_t root,
+			   uint64_t *first_key, uint64_t end_key,
+			   dm_block_t *new_root, unsigned *nr_removed)
+{
+	int r;
+
+	*nr_removed = 0;
+	do {
+		r = remove_one(info, root, first_key, end_key, &root, nr_removed);
+		if (!r)
+			(*nr_removed)++;
+	} while (!r);
+
+	*new_root = root;
+	return r == -ENODATA ? 0 : r;
+}
+EXPORT_SYMBOL_GPL(dm_btree_remove_leaves);