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, 721 insertions, 0 deletions

diff --git a/drivers/firmware/tegra/ivc.c b/drivers/firmware/tegra/ivc.c
new file mode 100644
index 000000000..8c9aff980
--- /dev/null
+++ b/drivers/firmware/tegra/ivc.c
@@ -0,0 +1,721 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (c) 2014-2016, NVIDIA CORPORATION.  All rights reserved.
+ */
+
+#include <soc/tegra/ivc.h>
+
+#define TEGRA_IVC_ALIGN 64
+
+/*
+ * IVC channel reset protocol.
+ *
+ * Each end uses its tx_channel.state to indicate its synchronization state.
+ */
+enum tegra_ivc_state {
+	/*
+	 * This value is zero for backwards compatibility with services that
+	 * assume channels to be initially zeroed. Such channels are in an
+	 * initially valid state, but cannot be asynchronously reset, and must
+	 * maintain a valid state at all times.
+	 *
+	 * The transmitting end can enter the established state from the sync or
+	 * ack state when it observes the receiving endpoint in the ack or
+	 * established state, indicating that has cleared the counters in our
+	 * rx_channel.
+	 */
+	TEGRA_IVC_STATE_ESTABLISHED = 0,
+
+	/*
+	 * If an endpoint is observed in the sync state, the remote endpoint is
+	 * allowed to clear the counters it owns asynchronously with respect to
+	 * the current endpoint. Therefore, the current endpoint is no longer
+	 * allowed to communicate.
+	 */
+	TEGRA_IVC_STATE_SYNC,
+
+	/*
+	 * When the transmitting end observes the receiving end in the sync
+	 * state, it can clear the w_count and r_count and transition to the ack
+	 * state. If the remote endpoint observes us in the ack state, it can
+	 * return to the established state once it has cleared its counters.
+	 */
+	TEGRA_IVC_STATE_ACK
+};
+
+/*
+ * This structure is divided into two-cache aligned parts, the first is only
+ * written through the tx.channel pointer, while the second is only written
+ * through the rx.channel pointer. This delineates ownership of the cache
+ * lines, which is critical to performance and necessary in non-cache coherent
+ * implementations.
+ */
+struct tegra_ivc_header {
+	union {
+		struct {
+			/* fields owned by the transmitting end */
+			u32 count;
+			u32 state;
+		};
+
+		u8 pad[TEGRA_IVC_ALIGN];
+	} tx;
+
+	union {
+		/* fields owned by the receiving end */
+		u32 count;
+		u8 pad[TEGRA_IVC_ALIGN];
+	} rx;
+};
+
+#define tegra_ivc_header_read_field(hdr, field) \
+	iosys_map_rd_field(hdr, 0, struct tegra_ivc_header, field)
+
+#define tegra_ivc_header_write_field(hdr, field, value) \
+	iosys_map_wr_field(hdr, 0, struct tegra_ivc_header, field, value)
+
+static inline void tegra_ivc_invalidate(struct tegra_ivc *ivc, dma_addr_t phys)
+{
+	if (!ivc->peer)
+		return;
+
+	dma_sync_single_for_cpu(ivc->peer, phys, TEGRA_IVC_ALIGN,
+				DMA_FROM_DEVICE);
+}
+
+static inline void tegra_ivc_flush(struct tegra_ivc *ivc, dma_addr_t phys)
+{
+	if (!ivc->peer)
+		return;
+
+	dma_sync_single_for_device(ivc->peer, phys, TEGRA_IVC_ALIGN,
+				   DMA_TO_DEVICE);
+}
+
+static inline bool tegra_ivc_empty(struct tegra_ivc *ivc, struct iosys_map *map)
+{
+	/*
+	 * This function performs multiple checks on the same values with
+	 * security implications, so create snapshots with READ_ONCE() to
+	 * ensure that these checks use the same values.
+	 */
+	u32 tx = tegra_ivc_header_read_field(map, tx.count);
+	u32 rx = tegra_ivc_header_read_field(map, rx.count);
+
+	/*
+	 * Perform an over-full check to prevent denial of service attacks
+	 * where a server could be easily fooled into believing that there's
+	 * an extremely large number of frames ready, since receivers are not
+	 * expected to check for full or over-full conditions.
+	 *
+	 * Although the channel isn't empty, this is an invalid case caused by
+	 * a potentially malicious peer, so returning empty is safer, because
+	 * it gives the impression that the channel has gone silent.
+	 */
+	if (tx - rx > ivc->num_frames)
+		return true;
+
+	return tx == rx;
+}
+
+static inline bool tegra_ivc_full(struct tegra_ivc *ivc, struct iosys_map *map)
+{
+	u32 tx = tegra_ivc_header_read_field(map, tx.count);
+	u32 rx = tegra_ivc_header_read_field(map, rx.count);
+
+	/*
+	 * Invalid cases where the counters indicate that the queue is over
+	 * capacity also appear full.
+	 */
+	return tx - rx >= ivc->num_frames;
+}
+
+static inline u32 tegra_ivc_available(struct tegra_ivc *ivc, struct iosys_map *map)
+{
+	u32 tx = tegra_ivc_header_read_field(map, tx.count);
+	u32 rx = tegra_ivc_header_read_field(map, rx.count);
+
+	/*
+	 * This function isn't expected to be used in scenarios where an
+	 * over-full situation can lead to denial of service attacks. See the
+	 * comment in tegra_ivc_empty() for an explanation about special
+	 * over-full considerations.
+	 */
+	return tx - rx;
+}
+
+static inline void tegra_ivc_advance_tx(struct tegra_ivc *ivc)
+{
+	unsigned int count = tegra_ivc_header_read_field(&ivc->tx.map, tx.count);
+
+	tegra_ivc_header_write_field(&ivc->tx.map, tx.count, count + 1);
+
+	if (ivc->tx.position == ivc->num_frames - 1)
+		ivc->tx.position = 0;
+	else
+		ivc->tx.position++;
+}
+
+static inline void tegra_ivc_advance_rx(struct tegra_ivc *ivc)
+{
+	unsigned int count = tegra_ivc_header_read_field(&ivc->rx.map, rx.count);
+
+	tegra_ivc_header_write_field(&ivc->rx.map, rx.count, count + 1);
+
+	if (ivc->rx.position == ivc->num_frames - 1)
+		ivc->rx.position = 0;
+	else
+		ivc->rx.position++;
+}
+
+static inline int tegra_ivc_check_read(struct tegra_ivc *ivc)
+{
+	unsigned int offset = offsetof(struct tegra_ivc_header, tx.count);
+	unsigned int state;
+
+	/*
+	 * tx.channel->state is set locally, so it is not synchronized with
+	 * state from the remote peer. The remote peer cannot reset its
+	 * transmit counters until we've acknowledged its synchronization
+	 * request, so no additional synchronization is required because an
+	 * asynchronous transition of rx.channel->state to
+	 * TEGRA_IVC_STATE_ACK is not allowed.
+	 */
+	state = tegra_ivc_header_read_field(&ivc->tx.map, tx.state);
+	if (state != TEGRA_IVC_STATE_ESTABLISHED)
+		return -ECONNRESET;
+
+	/*
+	 * Avoid unnecessary invalidations when performing repeated accesses
+	 * to an IVC channel by checking the old queue pointers first.
+	 *
+	 * Synchronization is only necessary when these pointers indicate
+	 * empty or full.
+	 */
+	if (!tegra_ivc_empty(ivc, &ivc->rx.map))
+		return 0;
+
+	tegra_ivc_invalidate(ivc, ivc->rx.phys + offset);
+
+	if (tegra_ivc_empty(ivc, &ivc->rx.map))
+		return -ENOSPC;
+
+	return 0;
+}
+
+static inline int tegra_ivc_check_write(struct tegra_ivc *ivc)
+{
+	unsigned int offset = offsetof(struct tegra_ivc_header, rx.count);
+	unsigned int state;
+
+	state = tegra_ivc_header_read_field(&ivc->tx.map, tx.state);
+	if (state != TEGRA_IVC_STATE_ESTABLISHED)
+		return -ECONNRESET;
+
+	if (!tegra_ivc_full(ivc, &ivc->tx.map))
+		return 0;
+
+	tegra_ivc_invalidate(ivc, ivc->tx.phys + offset);
+
+	if (tegra_ivc_full(ivc, &ivc->tx.map))
+		return -ENOSPC;
+
+	return 0;
+}
+
+static int tegra_ivc_frame_virt(struct tegra_ivc *ivc, const struct iosys_map *header,
+				unsigned int frame, struct iosys_map *map)
+{
+	size_t offset = sizeof(struct tegra_ivc_header) + ivc->frame_size * frame;
+
+	if (WARN_ON(frame >= ivc->num_frames))
+		return -EINVAL;
+
+	*map = IOSYS_MAP_INIT_OFFSET(header, offset);
+
+	return 0;
+}
+
+static inline dma_addr_t tegra_ivc_frame_phys(struct tegra_ivc *ivc,
+					      dma_addr_t phys,
+					      unsigned int frame)
+{
+	unsigned long offset;
+
+	offset = sizeof(struct tegra_ivc_header) + ivc->frame_size * frame;
+
+	return phys + offset;
+}
+
+static inline void tegra_ivc_invalidate_frame(struct tegra_ivc *ivc,
+					      dma_addr_t phys,
+					      unsigned int frame,
+					      unsigned int offset,
+					      size_t size)
+{
+	if (!ivc->peer || WARN_ON(frame >= ivc->num_frames))
+		return;
+
+	phys = tegra_ivc_frame_phys(ivc, phys, frame) + offset;
+
+	dma_sync_single_for_cpu(ivc->peer, phys, size, DMA_FROM_DEVICE);
+}
+
+static inline void tegra_ivc_flush_frame(struct tegra_ivc *ivc,
+					 dma_addr_t phys,
+					 unsigned int frame,
+					 unsigned int offset,
+					 size_t size)
+{
+	if (!ivc->peer || WARN_ON(frame >= ivc->num_frames))
+		return;
+
+	phys = tegra_ivc_frame_phys(ivc, phys, frame) + offset;
+
+	dma_sync_single_for_device(ivc->peer, phys, size, DMA_TO_DEVICE);
+}
+
+/* directly peek at the next frame rx'ed */
+int tegra_ivc_read_get_next_frame(struct tegra_ivc *ivc, struct iosys_map *map)
+{
+	int err;
+
+	if (WARN_ON(ivc == NULL))
+		return -EINVAL;
+
+	err = tegra_ivc_check_read(ivc);
+	if (err < 0)
+		return err;
+
+	/*
+	 * Order observation of ivc->rx.position potentially indicating new
+	 * data before data read.
+	 */
+	smp_rmb();
+
+	tegra_ivc_invalidate_frame(ivc, ivc->rx.phys, ivc->rx.position, 0,
+				   ivc->frame_size);
+
+	return tegra_ivc_frame_virt(ivc, &ivc->rx.map, ivc->rx.position, map);
+}
+EXPORT_SYMBOL(tegra_ivc_read_get_next_frame);
+
+int tegra_ivc_read_advance(struct tegra_ivc *ivc)
+{
+	unsigned int rx = offsetof(struct tegra_ivc_header, rx.count);
+	unsigned int tx = offsetof(struct tegra_ivc_header, tx.count);
+	int err;
+
+	/*
+	 * No read barriers or synchronization here: the caller is expected to
+	 * have already observed the channel non-empty. This check is just to
+	 * catch programming errors.
+	 */
+	err = tegra_ivc_check_read(ivc);
+	if (err < 0)
+		return err;
+
+	tegra_ivc_advance_rx(ivc);
+
+	tegra_ivc_flush(ivc, ivc->rx.phys + rx);
+
+	/*
+	 * Ensure our write to ivc->rx.position occurs before our read from
+	 * ivc->tx.position.
+	 */
+	smp_mb();
+
+	/*
+	 * Notify only upon transition from full to non-full. The available
+	 * count can only asynchronously increase, so the worst possible
+	 * side-effect will be a spurious notification.
+	 */
+	tegra_ivc_invalidate(ivc, ivc->rx.phys + tx);
+
+	if (tegra_ivc_available(ivc, &ivc->rx.map) == ivc->num_frames - 1)
+		ivc->notify(ivc, ivc->notify_data);
+
+	return 0;
+}
+EXPORT_SYMBOL(tegra_ivc_read_advance);
+
+/* directly poke at the next frame to be tx'ed */
+int tegra_ivc_write_get_next_frame(struct tegra_ivc *ivc, struct iosys_map *map)
+{
+	int err;
+
+	err = tegra_ivc_check_write(ivc);
+	if (err < 0)
+		return err;
+
+	return tegra_ivc_frame_virt(ivc, &ivc->tx.map, ivc->tx.position, map);
+}
+EXPORT_SYMBOL(tegra_ivc_write_get_next_frame);
+
+/* advance the tx buffer */
+int tegra_ivc_write_advance(struct tegra_ivc *ivc)
+{
+	unsigned int tx = offsetof(struct tegra_ivc_header, tx.count);
+	unsigned int rx = offsetof(struct tegra_ivc_header, rx.count);
+	int err;
+
+	err = tegra_ivc_check_write(ivc);
+	if (err < 0)
+		return err;
+
+	tegra_ivc_flush_frame(ivc, ivc->tx.phys, ivc->tx.position, 0,
+			      ivc->frame_size);
+
+	/*
+	 * Order any possible stores to the frame before update of
+	 * ivc->tx.position.
+	 */
+	smp_wmb();
+
+	tegra_ivc_advance_tx(ivc);
+	tegra_ivc_flush(ivc, ivc->tx.phys + tx);
+
+	/*
+	 * Ensure our write to ivc->tx.position occurs before our read from
+	 * ivc->rx.position.
+	 */
+	smp_mb();
+
+	/*
+	 * Notify only upon transition from empty to non-empty. The available
+	 * count can only asynchronously decrease, so the worst possible
+	 * side-effect will be a spurious notification.
+	 */
+	tegra_ivc_invalidate(ivc, ivc->tx.phys + rx);
+
+	if (tegra_ivc_available(ivc, &ivc->tx.map) == 1)
+		ivc->notify(ivc, ivc->notify_data);
+
+	return 0;
+}
+EXPORT_SYMBOL(tegra_ivc_write_advance);
+
+void tegra_ivc_reset(struct tegra_ivc *ivc)
+{
+	unsigned int offset = offsetof(struct tegra_ivc_header, tx.count);
+
+	tegra_ivc_header_write_field(&ivc->tx.map, tx.state, TEGRA_IVC_STATE_SYNC);
+	tegra_ivc_flush(ivc, ivc->tx.phys + offset);
+	ivc->notify(ivc, ivc->notify_data);
+}
+EXPORT_SYMBOL(tegra_ivc_reset);
+
+/*
+ * =======================================================
+ *  IVC State Transition Table - see tegra_ivc_notified()
+ * =======================================================
+ *
+ *	local	remote	action
+ *	-----	------	-----------------------------------
+ *	SYNC	EST	<none>
+ *	SYNC	ACK	reset counters; move to EST; notify
+ *	SYNC	SYNC	reset counters; move to ACK; notify
+ *	ACK	EST	move to EST; notify
+ *	ACK	ACK	move to EST; notify
+ *	ACK	SYNC	reset counters; move to ACK; notify
+ *	EST	EST	<none>
+ *	EST	ACK	<none>
+ *	EST	SYNC	reset counters; move to ACK; notify
+ *
+ * ===============================================================
+ */
+
+int tegra_ivc_notified(struct tegra_ivc *ivc)
+{
+	unsigned int offset = offsetof(struct tegra_ivc_header, tx.count);
+	enum tegra_ivc_state rx_state, tx_state;
+
+	/* Copy the receiver's state out of shared memory. */
+	tegra_ivc_invalidate(ivc, ivc->rx.phys + offset);
+	rx_state = tegra_ivc_header_read_field(&ivc->rx.map, tx.state);
+	tx_state = tegra_ivc_header_read_field(&ivc->tx.map, tx.state);
+
+	if (rx_state == TEGRA_IVC_STATE_SYNC) {
+		offset = offsetof(struct tegra_ivc_header, tx.count);
+
+		/*
+		 * Order observation of TEGRA_IVC_STATE_SYNC before stores
+		 * clearing tx.channel.
+		 */
+		smp_rmb();
+
+		/*
+		 * Reset tx.channel counters. The remote end is in the SYNC
+		 * state and won't make progress until we change our state,
+		 * so the counters are not in use at this time.
+		 */
+		tegra_ivc_header_write_field(&ivc->tx.map, tx.count, 0);
+		tegra_ivc_header_write_field(&ivc->rx.map, rx.count, 0);
+
+		ivc->tx.position = 0;
+		ivc->rx.position = 0;
+
+		/*
+		 * Ensure that counters appear cleared before new state can be
+		 * observed.
+		 */
+		smp_wmb();
+
+		/*
+		 * Move to ACK state. We have just cleared our counters, so it
+		 * is now safe for the remote end to start using these values.
+		 */
+		tegra_ivc_header_write_field(&ivc->tx.map, tx.state, TEGRA_IVC_STATE_ACK);
+		tegra_ivc_flush(ivc, ivc->tx.phys + offset);
+
+		/*
+		 * Notify remote end to observe state transition.
+		 */
+		ivc->notify(ivc, ivc->notify_data);
+
+	} else if (tx_state == TEGRA_IVC_STATE_SYNC &&
+		   rx_state == TEGRA_IVC_STATE_ACK) {
+		offset = offsetof(struct tegra_ivc_header, tx.count);
+
+		/*
+		 * Order observation of ivc_state_sync before stores clearing
+		 * tx_channel.
+		 */
+		smp_rmb();
+
+		/*
+		 * Reset tx.channel counters. The remote end is in the ACK
+		 * state and won't make progress until we change our state,
+		 * so the counters are not in use at this time.
+		 */
+		tegra_ivc_header_write_field(&ivc->tx.map, tx.count, 0);
+		tegra_ivc_header_write_field(&ivc->rx.map, rx.count, 0);
+
+		ivc->tx.position = 0;
+		ivc->rx.position = 0;
+
+		/*
+		 * Ensure that counters appear cleared before new state can be
+		 * observed.
+		 */
+		smp_wmb();
+
+		/*
+		 * Move to ESTABLISHED state. We know that the remote end has
+		 * already cleared its counters, so it is safe to start
+		 * writing/reading on this channel.
+		 */
+		tegra_ivc_header_write_field(&ivc->tx.map, tx.state, TEGRA_IVC_STATE_ESTABLISHED);
+		tegra_ivc_flush(ivc, ivc->tx.phys + offset);
+
+		/*
+		 * Notify remote end to observe state transition.
+		 */
+		ivc->notify(ivc, ivc->notify_data);
+
+	} else if (tx_state == TEGRA_IVC_STATE_ACK) {
+		offset = offsetof(struct tegra_ivc_header, tx.count);
+
+		/*
+		 * At this point, we have observed the peer to be in either
+		 * the ACK or ESTABLISHED state. Next, order observation of
+		 * peer state before storing to tx.channel.
+		 */
+		smp_rmb();
+
+		/*
+		 * Move to ESTABLISHED state. We know that we have previously
+		 * cleared our counters, and we know that the remote end has
+		 * cleared its counters, so it is safe to start writing/reading
+		 * on this channel.
+		 */
+		tegra_ivc_header_write_field(&ivc->tx.map, tx.state, TEGRA_IVC_STATE_ESTABLISHED);
+		tegra_ivc_flush(ivc, ivc->tx.phys + offset);
+
+		/*
+		 * Notify remote end to observe state transition.
+		 */
+		ivc->notify(ivc, ivc->notify_data);
+
+	} else {
+		/*
+		 * There is no need to handle any further action. Either the
+		 * channel is already fully established, or we are waiting for
+		 * the remote end to catch up with our current state. Refer
+		 * to the diagram in "IVC State Transition Table" above.
+		 */
+	}
+
+	if (tx_state != TEGRA_IVC_STATE_ESTABLISHED)
+		return -EAGAIN;
+
+	return 0;
+}
+EXPORT_SYMBOL(tegra_ivc_notified);
+
+size_t tegra_ivc_align(size_t size)
+{
+	return ALIGN(size, TEGRA_IVC_ALIGN);
+}
+EXPORT_SYMBOL(tegra_ivc_align);
+
+unsigned tegra_ivc_total_queue_size(unsigned queue_size)
+{
+	if (!IS_ALIGNED(queue_size, TEGRA_IVC_ALIGN)) {
+		pr_err("%s: queue_size (%u) must be %u-byte aligned\n",
+		       __func__, queue_size, TEGRA_IVC_ALIGN);
+		return 0;
+	}
+
+	return queue_size + sizeof(struct tegra_ivc_header);
+}
+EXPORT_SYMBOL(tegra_ivc_total_queue_size);
+
+static int tegra_ivc_check_params(unsigned long rx, unsigned long tx,
+				  unsigned int num_frames, size_t frame_size)
+{
+	BUILD_BUG_ON(!IS_ALIGNED(offsetof(struct tegra_ivc_header, tx.count),
+				 TEGRA_IVC_ALIGN));
+	BUILD_BUG_ON(!IS_ALIGNED(offsetof(struct tegra_ivc_header, rx.count),
+				 TEGRA_IVC_ALIGN));
+	BUILD_BUG_ON(!IS_ALIGNED(sizeof(struct tegra_ivc_header),
+				 TEGRA_IVC_ALIGN));
+
+	if ((uint64_t)num_frames * (uint64_t)frame_size >= 0x100000000UL) {
+		pr_err("num_frames * frame_size overflows\n");
+		return -EINVAL;
+	}
+
+	if (!IS_ALIGNED(frame_size, TEGRA_IVC_ALIGN)) {
+		pr_err("frame size not adequately aligned: %zu\n", frame_size);
+		return -EINVAL;
+	}
+
+	/*
+	 * The headers must at least be aligned enough for counters
+	 * to be accessed atomically.
+	 */
+	if (!IS_ALIGNED(rx, TEGRA_IVC_ALIGN)) {
+		pr_err("IVC channel start not aligned: %#lx\n", rx);
+		return -EINVAL;
+	}
+
+	if (!IS_ALIGNED(tx, TEGRA_IVC_ALIGN)) {
+		pr_err("IVC channel start not aligned: %#lx\n", tx);
+		return -EINVAL;
+	}
+
+	if (rx < tx) {
+		if (rx + frame_size * num_frames > tx) {
+			pr_err("queue regions overlap: %#lx + %zx > %#lx\n",
+			       rx, frame_size * num_frames, tx);
+			return -EINVAL;
+		}
+	} else {
+		if (tx + frame_size * num_frames > rx) {
+			pr_err("queue regions overlap: %#lx + %zx > %#lx\n",
+			       tx, frame_size * num_frames, rx);
+			return -EINVAL;
+		}
+	}
+
+	return 0;
+}
+
+static inline void iosys_map_copy(struct iosys_map *dst, const struct iosys_map *src)
+{
+	*dst = *src;
+}
+
+static inline unsigned long iosys_map_get_address(const struct iosys_map *map)
+{
+	if (map->is_iomem)
+		return (unsigned long)map->vaddr_iomem;
+
+	return (unsigned long)map->vaddr;
+}
+
+static inline void *iosys_map_get_vaddr(const struct iosys_map *map)
+{
+	if (WARN_ON(map->is_iomem))
+		return NULL;
+
+	return map->vaddr;
+}
+
+int tegra_ivc_init(struct tegra_ivc *ivc, struct device *peer, const struct iosys_map *rx,
+		   dma_addr_t rx_phys, const struct iosys_map *tx, dma_addr_t tx_phys,
+		   unsigned int num_frames, size_t frame_size,
+		   void (*notify)(struct tegra_ivc *ivc, void *data),
+		   void *data)
+{
+	size_t queue_size;
+	int err;
+
+	if (WARN_ON(!ivc || !notify))
+		return -EINVAL;
+
+	/*
+	 * All sizes that can be returned by communication functions should
+	 * fit in an int.
+	 */
+	if (frame_size > INT_MAX)
+		return -E2BIG;
+
+	err = tegra_ivc_check_params(iosys_map_get_address(rx), iosys_map_get_address(tx),
+				     num_frames, frame_size);
+	if (err < 0)
+		return err;
+
+	queue_size = tegra_ivc_total_queue_size(num_frames * frame_size);
+
+	if (peer) {
+		ivc->rx.phys = dma_map_single(peer, iosys_map_get_vaddr(rx), queue_size,
+					      DMA_BIDIRECTIONAL);
+		if (dma_mapping_error(peer, ivc->rx.phys))
+			return -ENOMEM;
+
+		ivc->tx.phys = dma_map_single(peer, iosys_map_get_vaddr(tx), queue_size,
+					      DMA_BIDIRECTIONAL);
+		if (dma_mapping_error(peer, ivc->tx.phys)) {
+			dma_unmap_single(peer, ivc->rx.phys, queue_size,
+					 DMA_BIDIRECTIONAL);
+			return -ENOMEM;
+		}
+	} else {
+		ivc->rx.phys = rx_phys;
+		ivc->tx.phys = tx_phys;
+	}
+
+	iosys_map_copy(&ivc->rx.map, rx);
+	iosys_map_copy(&ivc->tx.map, tx);
+	ivc->peer = peer;
+	ivc->notify = notify;
+	ivc->notify_data = data;
+	ivc->frame_size = frame_size;
+	ivc->num_frames = num_frames;
+
+	/*
+	 * These values aren't necessarily correct until the channel has been
+	 * reset.
+	 */
+	ivc->tx.position = 0;
+	ivc->rx.position = 0;
+
+	return 0;
+}
+EXPORT_SYMBOL(tegra_ivc_init);
+
+void tegra_ivc_cleanup(struct tegra_ivc *ivc)
+{
+	if (ivc->peer) {
+		size_t size = tegra_ivc_total_queue_size(ivc->num_frames *
+							 ivc->frame_size);
+
+		dma_unmap_single(ivc->peer, ivc->rx.phys, size,
+				 DMA_BIDIRECTIONAL);
+		dma_unmap_single(ivc->peer, ivc->tx.phys, size,
+				 DMA_BIDIRECTIONAL);
+	}
+}
+EXPORT_SYMBOL(tegra_ivc_cleanup);