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, 702 insertions, 0 deletions

diff --git a/drivers/infiniband/hw/mthca/mthca_mr.c b/drivers/infiniband/hw/mthca/mthca_mr.c
new file mode 100644
index 000000000..192f83fd7
--- /dev/null
+++ b/drivers/infiniband/hw/mthca/mthca_mr.c
@@ -0,0 +1,702 @@
+/*
+ * Copyright (c) 2004 Topspin Communications.  All rights reserved.
+ * Copyright (c) 2005 Mellanox Technologies. All rights reserved.
+ *
+ * This software is available to you under a choice of one of two
+ * licenses.  You may choose to be licensed under the terms of the GNU
+ * General Public License (GPL) Version 2, available from the file
+ * COPYING in the main directory of this source tree, or the
+ * OpenIB.org BSD license below:
+ *
+ *     Redistribution and use in source and binary forms, with or
+ *     without modification, are permitted provided that the following
+ *     conditions are met:
+ *
+ *      - Redistributions of source code must retain the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer.
+ *
+ *      - Redistributions in binary form must reproduce the above
+ *        copyright notice, this list of conditions and the following
+ *        disclaimer in the documentation and/or other materials
+ *        provided with the distribution.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
+ * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
+ * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
+ * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+ * SOFTWARE.
+ */
+
+#include <linux/slab.h>
+#include <linux/errno.h>
+
+#include "mthca_dev.h"
+#include "mthca_cmd.h"
+#include "mthca_memfree.h"
+
+struct mthca_mtt {
+	struct mthca_buddy *buddy;
+	int                 order;
+	u32                 first_seg;
+};
+
+/*
+ * Must be packed because mtt_seg is 64 bits but only aligned to 32 bits.
+ */
+struct mthca_mpt_entry {
+	__be32 flags;
+	__be32 page_size;
+	__be32 key;
+	__be32 pd;
+	__be64 start;
+	__be64 length;
+	__be32 lkey;
+	__be32 window_count;
+	__be32 window_count_limit;
+	__be64 mtt_seg;
+	__be32 mtt_sz;		/* Arbel only */
+	u32    reserved[2];
+} __packed;
+
+#define MTHCA_MPT_FLAG_SW_OWNS       (0xfUL << 28)
+#define MTHCA_MPT_FLAG_MIO           (1 << 17)
+#define MTHCA_MPT_FLAG_BIND_ENABLE   (1 << 15)
+#define MTHCA_MPT_FLAG_PHYSICAL      (1 <<  9)
+#define MTHCA_MPT_FLAG_REGION        (1 <<  8)
+
+#define MTHCA_MTT_FLAG_PRESENT       1
+
+#define MTHCA_MPT_STATUS_SW 0xF0
+#define MTHCA_MPT_STATUS_HW 0x00
+
+#define SINAI_FMR_KEY_INC 0x1000000
+
+/*
+ * Buddy allocator for MTT segments (currently not very efficient
+ * since it doesn't keep a free list and just searches linearly
+ * through the bitmaps)
+ */
+
+static u32 mthca_buddy_alloc(struct mthca_buddy *buddy, int order)
+{
+	int o;
+	int m;
+	u32 seg;
+
+	spin_lock(&buddy->lock);
+
+	for (o = order; o <= buddy->max_order; ++o)
+		if (buddy->num_free[o]) {
+			m = 1 << (buddy->max_order - o);
+			seg = find_first_bit(buddy->bits[o], m);
+			if (seg < m)
+				goto found;
+		}
+
+	spin_unlock(&buddy->lock);
+	return -1;
+
+ found:
+	__clear_bit(seg, buddy->bits[o]);
+	--buddy->num_free[o];
+
+	while (o > order) {
+		--o;
+		seg <<= 1;
+		__set_bit(seg ^ 1, buddy->bits[o]);
+		++buddy->num_free[o];
+	}
+
+	spin_unlock(&buddy->lock);
+
+	seg <<= order;
+
+	return seg;
+}
+
+static void mthca_buddy_free(struct mthca_buddy *buddy, u32 seg, int order)
+{
+	seg >>= order;
+
+	spin_lock(&buddy->lock);
+
+	while (test_bit(seg ^ 1, buddy->bits[order])) {
+		__clear_bit(seg ^ 1, buddy->bits[order]);
+		--buddy->num_free[order];
+		seg >>= 1;
+		++order;
+	}
+
+	__set_bit(seg, buddy->bits[order]);
+	++buddy->num_free[order];
+
+	spin_unlock(&buddy->lock);
+}
+
+static int mthca_buddy_init(struct mthca_buddy *buddy, int max_order)
+{
+	int i;
+
+	buddy->max_order = max_order;
+	spin_lock_init(&buddy->lock);
+
+	buddy->bits = kcalloc(buddy->max_order + 1, sizeof(long *),
+			      GFP_KERNEL);
+	buddy->num_free = kcalloc((buddy->max_order + 1), sizeof *buddy->num_free,
+				  GFP_KERNEL);
+	if (!buddy->bits || !buddy->num_free)
+		goto err_out;
+
+	for (i = 0; i <= buddy->max_order; ++i) {
+		buddy->bits[i] = bitmap_zalloc(1 << (buddy->max_order - i),
+					       GFP_KERNEL);
+		if (!buddy->bits[i])
+			goto err_out_free;
+	}
+
+	__set_bit(0, buddy->bits[buddy->max_order]);
+	buddy->num_free[buddy->max_order] = 1;
+
+	return 0;
+
+err_out_free:
+	for (i = 0; i <= buddy->max_order; ++i)
+		bitmap_free(buddy->bits[i]);
+
+err_out:
+	kfree(buddy->bits);
+	kfree(buddy->num_free);
+
+	return -ENOMEM;
+}
+
+static void mthca_buddy_cleanup(struct mthca_buddy *buddy)
+{
+	int i;
+
+	for (i = 0; i <= buddy->max_order; ++i)
+		bitmap_free(buddy->bits[i]);
+
+	kfree(buddy->bits);
+	kfree(buddy->num_free);
+}
+
+static u32 mthca_alloc_mtt_range(struct mthca_dev *dev, int order,
+				 struct mthca_buddy *buddy)
+{
+	u32 seg = mthca_buddy_alloc(buddy, order);
+
+	if (seg == -1)
+		return -1;
+
+	if (mthca_is_memfree(dev))
+		if (mthca_table_get_range(dev, dev->mr_table.mtt_table, seg,
+					  seg + (1 << order) - 1)) {
+			mthca_buddy_free(buddy, seg, order);
+			seg = -1;
+		}
+
+	return seg;
+}
+
+static struct mthca_mtt *__mthca_alloc_mtt(struct mthca_dev *dev, int size,
+					   struct mthca_buddy *buddy)
+{
+	struct mthca_mtt *mtt;
+	int i;
+
+	if (size <= 0)
+		return ERR_PTR(-EINVAL);
+
+	mtt = kmalloc(sizeof *mtt, GFP_KERNEL);
+	if (!mtt)
+		return ERR_PTR(-ENOMEM);
+
+	mtt->buddy = buddy;
+	mtt->order = 0;
+	for (i = dev->limits.mtt_seg_size / 8; i < size; i <<= 1)
+		++mtt->order;
+
+	mtt->first_seg = mthca_alloc_mtt_range(dev, mtt->order, buddy);
+	if (mtt->first_seg == -1) {
+		kfree(mtt);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	return mtt;
+}
+
+struct mthca_mtt *mthca_alloc_mtt(struct mthca_dev *dev, int size)
+{
+	return __mthca_alloc_mtt(dev, size, &dev->mr_table.mtt_buddy);
+}
+
+void mthca_free_mtt(struct mthca_dev *dev, struct mthca_mtt *mtt)
+{
+	if (!mtt)
+		return;
+
+	mthca_buddy_free(mtt->buddy, mtt->first_seg, mtt->order);
+
+	mthca_table_put_range(dev, dev->mr_table.mtt_table,
+			      mtt->first_seg,
+			      mtt->first_seg + (1 << mtt->order) - 1);
+
+	kfree(mtt);
+}
+
+static int __mthca_write_mtt(struct mthca_dev *dev, struct mthca_mtt *mtt,
+			     int start_index, u64 *buffer_list, int list_len)
+{
+	struct mthca_mailbox *mailbox;
+	__be64 *mtt_entry;
+	int err = 0;
+	int i;
+
+	mailbox = mthca_alloc_mailbox(dev, GFP_KERNEL);
+	if (IS_ERR(mailbox))
+		return PTR_ERR(mailbox);
+	mtt_entry = mailbox->buf;
+
+	while (list_len > 0) {
+		mtt_entry[0] = cpu_to_be64(dev->mr_table.mtt_base +
+					   mtt->first_seg * dev->limits.mtt_seg_size +
+					   start_index * 8);
+		mtt_entry[1] = 0;
+		for (i = 0; i < list_len && i < MTHCA_MAILBOX_SIZE / 8 - 2; ++i)
+			mtt_entry[i + 2] = cpu_to_be64(buffer_list[i] |
+						       MTHCA_MTT_FLAG_PRESENT);
+
+		/*
+		 * If we have an odd number of entries to write, add
+		 * one more dummy entry for firmware efficiency.
+		 */
+		if (i & 1)
+			mtt_entry[i + 2] = 0;
+
+		err = mthca_WRITE_MTT(dev, mailbox, (i + 1) & ~1);
+		if (err) {
+			mthca_warn(dev, "WRITE_MTT failed (%d)\n", err);
+			goto out;
+		}
+
+		list_len    -= i;
+		start_index += i;
+		buffer_list += i;
+	}
+
+out:
+	mthca_free_mailbox(dev, mailbox);
+	return err;
+}
+
+int mthca_write_mtt_size(struct mthca_dev *dev)
+{
+	if (dev->mr_table.fmr_mtt_buddy != &dev->mr_table.mtt_buddy ||
+	    !(dev->mthca_flags & MTHCA_FLAG_FMR))
+		/*
+		 * Be friendly to WRITE_MTT command
+		 * and leave two empty slots for the
+		 * index and reserved fields of the
+		 * mailbox.
+		 */
+		return PAGE_SIZE / sizeof (u64) - 2;
+
+	/* For Arbel, all MTTs must fit in the same page. */
+	return mthca_is_memfree(dev) ? (PAGE_SIZE / sizeof (u64)) : 0x7ffffff;
+}
+
+static void mthca_tavor_write_mtt_seg(struct mthca_dev *dev,
+				      struct mthca_mtt *mtt, int start_index,
+				      u64 *buffer_list, int list_len)
+{
+	u64 __iomem *mtts;
+	int i;
+
+	mtts = dev->mr_table.tavor_fmr.mtt_base + mtt->first_seg * dev->limits.mtt_seg_size +
+		start_index * sizeof (u64);
+	for (i = 0; i < list_len; ++i)
+		mthca_write64_raw(cpu_to_be64(buffer_list[i] | MTHCA_MTT_FLAG_PRESENT),
+				  mtts + i);
+}
+
+static void mthca_arbel_write_mtt_seg(struct mthca_dev *dev,
+				      struct mthca_mtt *mtt, int start_index,
+				      u64 *buffer_list, int list_len)
+{
+	__be64 *mtts;
+	dma_addr_t dma_handle;
+	int i;
+	int s = start_index * sizeof (u64);
+
+	/* For Arbel, all MTTs must fit in the same page. */
+	BUG_ON(s / PAGE_SIZE != (s + list_len * sizeof(u64) - 1) / PAGE_SIZE);
+	/* Require full segments */
+	BUG_ON(s % dev->limits.mtt_seg_size);
+
+	mtts = mthca_table_find(dev->mr_table.mtt_table, mtt->first_seg +
+				s / dev->limits.mtt_seg_size, &dma_handle);
+
+	BUG_ON(!mtts);
+
+	dma_sync_single_for_cpu(&dev->pdev->dev, dma_handle,
+				list_len * sizeof (u64), DMA_TO_DEVICE);
+
+	for (i = 0; i < list_len; ++i)
+		mtts[i] = cpu_to_be64(buffer_list[i] | MTHCA_MTT_FLAG_PRESENT);
+
+	dma_sync_single_for_device(&dev->pdev->dev, dma_handle,
+				   list_len * sizeof (u64), DMA_TO_DEVICE);
+}
+
+int mthca_write_mtt(struct mthca_dev *dev, struct mthca_mtt *mtt,
+		    int start_index, u64 *buffer_list, int list_len)
+{
+	int size = mthca_write_mtt_size(dev);
+	int chunk;
+
+	if (dev->mr_table.fmr_mtt_buddy != &dev->mr_table.mtt_buddy ||
+	    !(dev->mthca_flags & MTHCA_FLAG_FMR))
+		return __mthca_write_mtt(dev, mtt, start_index, buffer_list, list_len);
+
+	while (list_len > 0) {
+		chunk = min(size, list_len);
+		if (mthca_is_memfree(dev))
+			mthca_arbel_write_mtt_seg(dev, mtt, start_index,
+						  buffer_list, chunk);
+		else
+			mthca_tavor_write_mtt_seg(dev, mtt, start_index,
+						  buffer_list, chunk);
+
+		list_len    -= chunk;
+		start_index += chunk;
+		buffer_list += chunk;
+	}
+
+	return 0;
+}
+
+static inline u32 tavor_hw_index_to_key(u32 ind)
+{
+	return ind;
+}
+
+static inline u32 tavor_key_to_hw_index(u32 key)
+{
+	return key;
+}
+
+static inline u32 arbel_hw_index_to_key(u32 ind)
+{
+	return (ind >> 24) | (ind << 8);
+}
+
+static inline u32 arbel_key_to_hw_index(u32 key)
+{
+	return (key << 24) | (key >> 8);
+}
+
+static inline u32 hw_index_to_key(struct mthca_dev *dev, u32 ind)
+{
+	if (mthca_is_memfree(dev))
+		return arbel_hw_index_to_key(ind);
+	else
+		return tavor_hw_index_to_key(ind);
+}
+
+static inline u32 key_to_hw_index(struct mthca_dev *dev, u32 key)
+{
+	if (mthca_is_memfree(dev))
+		return arbel_key_to_hw_index(key);
+	else
+		return tavor_key_to_hw_index(key);
+}
+
+static inline u32 adjust_key(struct mthca_dev *dev, u32 key)
+{
+	if (dev->mthca_flags & MTHCA_FLAG_SINAI_OPT)
+		return ((key << 20) & 0x800000) | (key & 0x7fffff);
+	else
+		return key;
+}
+
+int mthca_mr_alloc(struct mthca_dev *dev, u32 pd, int buffer_size_shift,
+		   u64 iova, u64 total_size, u32 access, struct mthca_mr *mr)
+{
+	struct mthca_mailbox *mailbox;
+	struct mthca_mpt_entry *mpt_entry;
+	u32 key;
+	int i;
+	int err;
+
+	WARN_ON(buffer_size_shift >= 32);
+
+	key = mthca_alloc(&dev->mr_table.mpt_alloc);
+	if (key == -1)
+		return -ENOMEM;
+	key = adjust_key(dev, key);
+	mr->ibmr.rkey = mr->ibmr.lkey = hw_index_to_key(dev, key);
+
+	if (mthca_is_memfree(dev)) {
+		err = mthca_table_get(dev, dev->mr_table.mpt_table, key);
+		if (err)
+			goto err_out_mpt_free;
+	}
+
+	mailbox = mthca_alloc_mailbox(dev, GFP_KERNEL);
+	if (IS_ERR(mailbox)) {
+		err = PTR_ERR(mailbox);
+		goto err_out_table;
+	}
+	mpt_entry = mailbox->buf;
+
+	mpt_entry->flags = cpu_to_be32(MTHCA_MPT_FLAG_SW_OWNS     |
+				       MTHCA_MPT_FLAG_MIO         |
+				       MTHCA_MPT_FLAG_REGION      |
+				       access);
+	if (!mr->mtt)
+		mpt_entry->flags |= cpu_to_be32(MTHCA_MPT_FLAG_PHYSICAL);
+
+	mpt_entry->page_size = cpu_to_be32(buffer_size_shift - 12);
+	mpt_entry->key       = cpu_to_be32(key);
+	mpt_entry->pd        = cpu_to_be32(pd);
+	mpt_entry->start     = cpu_to_be64(iova);
+	mpt_entry->length    = cpu_to_be64(total_size);
+
+	memset_startat(mpt_entry, 0, lkey);
+
+	if (mr->mtt)
+		mpt_entry->mtt_seg =
+			cpu_to_be64(dev->mr_table.mtt_base +
+				    mr->mtt->first_seg * dev->limits.mtt_seg_size);
+
+	if (0) {
+		mthca_dbg(dev, "Dumping MPT entry %08x:\n", mr->ibmr.lkey);
+		for (i = 0; i < sizeof (struct mthca_mpt_entry) / 4; ++i) {
+			if (i % 4 == 0)
+				printk("[%02x] ", i * 4);
+			printk(" %08x", be32_to_cpu(((__be32 *) mpt_entry)[i]));
+			if ((i + 1) % 4 == 0)
+				printk("\n");
+		}
+	}
+
+	err = mthca_SW2HW_MPT(dev, mailbox,
+			      key & (dev->limits.num_mpts - 1));
+	if (err) {
+		mthca_warn(dev, "SW2HW_MPT failed (%d)\n", err);
+		goto err_out_mailbox;
+	}
+
+	mthca_free_mailbox(dev, mailbox);
+	return err;
+
+err_out_mailbox:
+	mthca_free_mailbox(dev, mailbox);
+
+err_out_table:
+	mthca_table_put(dev, dev->mr_table.mpt_table, key);
+
+err_out_mpt_free:
+	mthca_free(&dev->mr_table.mpt_alloc, key);
+	return err;
+}
+
+int mthca_mr_alloc_notrans(struct mthca_dev *dev, u32 pd,
+			   u32 access, struct mthca_mr *mr)
+{
+	mr->mtt = NULL;
+	return mthca_mr_alloc(dev, pd, 12, 0, ~0ULL, access, mr);
+}
+
+int mthca_mr_alloc_phys(struct mthca_dev *dev, u32 pd,
+			u64 *buffer_list, int buffer_size_shift,
+			int list_len, u64 iova, u64 total_size,
+			u32 access, struct mthca_mr *mr)
+{
+	int err;
+
+	mr->mtt = mthca_alloc_mtt(dev, list_len);
+	if (IS_ERR(mr->mtt))
+		return PTR_ERR(mr->mtt);
+
+	err = mthca_write_mtt(dev, mr->mtt, 0, buffer_list, list_len);
+	if (err) {
+		mthca_free_mtt(dev, mr->mtt);
+		return err;
+	}
+
+	err = mthca_mr_alloc(dev, pd, buffer_size_shift, iova,
+			     total_size, access, mr);
+	if (err)
+		mthca_free_mtt(dev, mr->mtt);
+
+	return err;
+}
+
+/* Free mr */
+static void mthca_free_region(struct mthca_dev *dev, u32 lkey)
+{
+	mthca_table_put(dev, dev->mr_table.mpt_table,
+			key_to_hw_index(dev, lkey));
+
+	mthca_free(&dev->mr_table.mpt_alloc, key_to_hw_index(dev, lkey));
+}
+
+void mthca_free_mr(struct mthca_dev *dev, struct mthca_mr *mr)
+{
+	int err;
+
+	err = mthca_HW2SW_MPT(dev, NULL,
+			      key_to_hw_index(dev, mr->ibmr.lkey) &
+			      (dev->limits.num_mpts - 1));
+	if (err)
+		mthca_warn(dev, "HW2SW_MPT failed (%d)\n", err);
+
+	mthca_free_region(dev, mr->ibmr.lkey);
+	mthca_free_mtt(dev, mr->mtt);
+}
+
+int mthca_init_mr_table(struct mthca_dev *dev)
+{
+	phys_addr_t addr;
+	int mpts, mtts, err, i;
+
+	err = mthca_alloc_init(&dev->mr_table.mpt_alloc,
+			       dev->limits.num_mpts,
+			       ~0, dev->limits.reserved_mrws);
+	if (err)
+		return err;
+
+	if (!mthca_is_memfree(dev) &&
+	    (dev->mthca_flags & MTHCA_FLAG_DDR_HIDDEN))
+		dev->limits.fmr_reserved_mtts = 0;
+	else
+		dev->mthca_flags |= MTHCA_FLAG_FMR;
+
+	if (dev->mthca_flags & MTHCA_FLAG_SINAI_OPT)
+		mthca_dbg(dev, "Memory key throughput optimization activated.\n");
+
+	err = mthca_buddy_init(&dev->mr_table.mtt_buddy,
+			       fls(dev->limits.num_mtt_segs - 1));
+
+	if (err)
+		goto err_mtt_buddy;
+
+	dev->mr_table.tavor_fmr.mpt_base = NULL;
+	dev->mr_table.tavor_fmr.mtt_base = NULL;
+
+	if (dev->limits.fmr_reserved_mtts) {
+		i = fls(dev->limits.fmr_reserved_mtts - 1);
+
+		if (i >= 31) {
+			mthca_warn(dev, "Unable to reserve 2^31 FMR MTTs.\n");
+			err = -EINVAL;
+			goto err_fmr_mpt;
+		}
+		mpts = mtts = 1 << i;
+	} else {
+		mtts = dev->limits.num_mtt_segs;
+		mpts = dev->limits.num_mpts;
+	}
+
+	if (!mthca_is_memfree(dev) &&
+	    (dev->mthca_flags & MTHCA_FLAG_FMR)) {
+
+		addr = pci_resource_start(dev->pdev, 4) +
+			((pci_resource_len(dev->pdev, 4) - 1) &
+			 dev->mr_table.mpt_base);
+
+		dev->mr_table.tavor_fmr.mpt_base =
+			ioremap(addr, mpts * sizeof(struct mthca_mpt_entry));
+
+		if (!dev->mr_table.tavor_fmr.mpt_base) {
+			mthca_warn(dev, "MPT ioremap for FMR failed.\n");
+			err = -ENOMEM;
+			goto err_fmr_mpt;
+		}
+
+		addr = pci_resource_start(dev->pdev, 4) +
+			((pci_resource_len(dev->pdev, 4) - 1) &
+			 dev->mr_table.mtt_base);
+
+		dev->mr_table.tavor_fmr.mtt_base =
+			ioremap(addr, mtts * dev->limits.mtt_seg_size);
+		if (!dev->mr_table.tavor_fmr.mtt_base) {
+			mthca_warn(dev, "MTT ioremap for FMR failed.\n");
+			err = -ENOMEM;
+			goto err_fmr_mtt;
+		}
+	}
+
+	if (dev->limits.fmr_reserved_mtts) {
+		err = mthca_buddy_init(&dev->mr_table.tavor_fmr.mtt_buddy, fls(mtts - 1));
+		if (err)
+			goto err_fmr_mtt_buddy;
+
+		/* Prevent regular MRs from using FMR keys */
+		err = mthca_buddy_alloc(&dev->mr_table.mtt_buddy, fls(mtts - 1));
+		if (err)
+			goto err_reserve_fmr;
+
+		dev->mr_table.fmr_mtt_buddy =
+			&dev->mr_table.tavor_fmr.mtt_buddy;
+	} else
+		dev->mr_table.fmr_mtt_buddy = &dev->mr_table.mtt_buddy;
+
+	/* FMR table is always the first, take reserved MTTs out of there */
+	if (dev->limits.reserved_mtts) {
+		i = fls(dev->limits.reserved_mtts - 1);
+
+		if (mthca_alloc_mtt_range(dev, i,
+					  dev->mr_table.fmr_mtt_buddy) == -1) {
+			mthca_warn(dev, "MTT table of order %d is too small.\n",
+				  dev->mr_table.fmr_mtt_buddy->max_order);
+			err = -ENOMEM;
+			goto err_reserve_mtts;
+		}
+	}
+
+	return 0;
+
+err_reserve_mtts:
+err_reserve_fmr:
+	if (dev->limits.fmr_reserved_mtts)
+		mthca_buddy_cleanup(&dev->mr_table.tavor_fmr.mtt_buddy);
+
+err_fmr_mtt_buddy:
+	if (dev->mr_table.tavor_fmr.mtt_base)
+		iounmap(dev->mr_table.tavor_fmr.mtt_base);
+
+err_fmr_mtt:
+	if (dev->mr_table.tavor_fmr.mpt_base)
+		iounmap(dev->mr_table.tavor_fmr.mpt_base);
+
+err_fmr_mpt:
+	mthca_buddy_cleanup(&dev->mr_table.mtt_buddy);
+
+err_mtt_buddy:
+	mthca_alloc_cleanup(&dev->mr_table.mpt_alloc);
+
+	return err;
+}
+
+void mthca_cleanup_mr_table(struct mthca_dev *dev)
+{
+	/* XXX check if any MRs are still allocated? */
+	if (dev->limits.fmr_reserved_mtts)
+		mthca_buddy_cleanup(&dev->mr_table.tavor_fmr.mtt_buddy);
+
+	mthca_buddy_cleanup(&dev->mr_table.mtt_buddy);
+
+	if (dev->mr_table.tavor_fmr.mtt_base)
+		iounmap(dev->mr_table.tavor_fmr.mtt_base);
+	if (dev->mr_table.tavor_fmr.mpt_base)
+		iounmap(dev->mr_table.tavor_fmr.mpt_base);
+
+	mthca_alloc_cleanup(&dev->mr_table.mpt_alloc);
+}