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, 1254 insertions, 0 deletions

diff --git a/arch/powerpc/platforms/ps3/mm.c b/arch/powerpc/platforms/ps3/mm.c
new file mode 100644
index 000000000..1326de55f
--- /dev/null
+++ b/arch/powerpc/platforms/ps3/mm.c
@@ -0,0 +1,1254 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ *  PS3 address space management.
+ *
+ *  Copyright (C) 2006 Sony Computer Entertainment Inc.
+ *  Copyright 2006 Sony Corp.
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/kernel.h>
+#include <linux/export.h>
+#include <linux/memblock.h>
+#include <linux/slab.h>
+
+#include <asm/cell-regs.h>
+#include <asm/firmware.h>
+#include <asm/udbg.h>
+#include <asm/lv1call.h>
+#include <asm/setup.h>
+
+#include "platform.h"
+
+#if defined(DEBUG)
+#define DBG udbg_printf
+#else
+#define DBG pr_devel
+#endif
+
+enum {
+#if defined(CONFIG_PS3_DYNAMIC_DMA)
+	USE_DYNAMIC_DMA = 1,
+#else
+	USE_DYNAMIC_DMA = 0,
+#endif
+};
+
+enum {
+	PAGE_SHIFT_4K = 12U,
+	PAGE_SHIFT_64K = 16U,
+	PAGE_SHIFT_16M = 24U,
+};
+
+static unsigned long __init make_page_sizes(unsigned long a, unsigned long b)
+{
+	return (a << 56) | (b << 48);
+}
+
+enum {
+	ALLOCATE_MEMORY_TRY_ALT_UNIT = 0X04,
+	ALLOCATE_MEMORY_ADDR_ZERO = 0X08,
+};
+
+/* valid htab sizes are {18,19,20} = 256K, 512K, 1M */
+
+enum {
+	HTAB_SIZE_MAX = 20U, /* HV limit of 1MB */
+	HTAB_SIZE_MIN = 18U, /* CPU limit of 256KB */
+};
+
+/*============================================================================*/
+/* virtual address space routines                                             */
+/*============================================================================*/
+
+/**
+ * struct mem_region - memory region structure
+ * @base: base address
+ * @size: size in bytes
+ * @offset: difference between base and rm.size
+ * @destroy: flag if region should be destroyed upon shutdown
+ */
+
+struct mem_region {
+	u64 base;
+	u64 size;
+	unsigned long offset;
+	int destroy;
+};
+
+/**
+ * struct map - address space state variables holder
+ * @total: total memory available as reported by HV
+ * @vas_id - HV virtual address space id
+ * @htab_size: htab size in bytes
+ *
+ * The HV virtual address space (vas) allows for hotplug memory regions.
+ * Memory regions can be created and destroyed in the vas at runtime.
+ * @rm: real mode (bootmem) region
+ * @r1: highmem region(s)
+ *
+ * ps3 addresses
+ * virt_addr: a cpu 'translated' effective address
+ * phys_addr: an address in what Linux thinks is the physical address space
+ * lpar_addr: an address in the HV virtual address space
+ * bus_addr: an io controller 'translated' address on a device bus
+ */
+
+struct map {
+	u64 total;
+	u64 vas_id;
+	u64 htab_size;
+	struct mem_region rm;
+	struct mem_region r1;
+};
+
+#define debug_dump_map(x) _debug_dump_map(x, __func__, __LINE__)
+static void __maybe_unused _debug_dump_map(const struct map *m,
+	const char *func, int line)
+{
+	DBG("%s:%d: map.total     = %llxh\n", func, line, m->total);
+	DBG("%s:%d: map.rm.size   = %llxh\n", func, line, m->rm.size);
+	DBG("%s:%d: map.vas_id    = %llu\n", func, line, m->vas_id);
+	DBG("%s:%d: map.htab_size = %llxh\n", func, line, m->htab_size);
+	DBG("%s:%d: map.r1.base   = %llxh\n", func, line, m->r1.base);
+	DBG("%s:%d: map.r1.offset = %lxh\n", func, line, m->r1.offset);
+	DBG("%s:%d: map.r1.size   = %llxh\n", func, line, m->r1.size);
+}
+
+static struct map map;
+
+/**
+ * ps3_mm_phys_to_lpar - translate a linux physical address to lpar address
+ * @phys_addr: linux physical address
+ */
+
+unsigned long ps3_mm_phys_to_lpar(unsigned long phys_addr)
+{
+	BUG_ON(is_kernel_addr(phys_addr));
+	return (phys_addr < map.rm.size || phys_addr >= map.total)
+		? phys_addr : phys_addr + map.r1.offset;
+}
+
+EXPORT_SYMBOL(ps3_mm_phys_to_lpar);
+
+/**
+ * ps3_mm_vas_create - create the virtual address space
+ */
+
+void __init ps3_mm_vas_create(unsigned long* htab_size)
+{
+	int result;
+	u64 start_address;
+	u64 size;
+	u64 access_right;
+	u64 max_page_size;
+	u64 flags;
+
+	result = lv1_query_logical_partition_address_region_info(0,
+		&start_address, &size, &access_right, &max_page_size,
+		&flags);
+
+	if (result) {
+		DBG("%s:%d: lv1_query_logical_partition_address_region_info "
+			"failed: %s\n", __func__, __LINE__,
+			ps3_result(result));
+		goto fail;
+	}
+
+	if (max_page_size < PAGE_SHIFT_16M) {
+		DBG("%s:%d: bad max_page_size %llxh\n", __func__, __LINE__,
+			max_page_size);
+		goto fail;
+	}
+
+	BUILD_BUG_ON(CONFIG_PS3_HTAB_SIZE > HTAB_SIZE_MAX);
+	BUILD_BUG_ON(CONFIG_PS3_HTAB_SIZE < HTAB_SIZE_MIN);
+
+	result = lv1_construct_virtual_address_space(CONFIG_PS3_HTAB_SIZE,
+			2, make_page_sizes(PAGE_SHIFT_16M, PAGE_SHIFT_64K),
+			&map.vas_id, &map.htab_size);
+
+	if (result) {
+		DBG("%s:%d: lv1_construct_virtual_address_space failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+		goto fail;
+	}
+
+	result = lv1_select_virtual_address_space(map.vas_id);
+
+	if (result) {
+		DBG("%s:%d: lv1_select_virtual_address_space failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+		goto fail;
+	}
+
+	*htab_size = map.htab_size;
+
+	debug_dump_map(&map);
+
+	return;
+
+fail:
+	panic("ps3_mm_vas_create failed");
+}
+
+/**
+ * ps3_mm_vas_destroy -
+ *
+ * called during kexec sequence with MMU off.
+ */
+
+notrace void ps3_mm_vas_destroy(void)
+{
+	int result;
+
+	if (map.vas_id) {
+		result = lv1_select_virtual_address_space(0);
+		result += lv1_destruct_virtual_address_space(map.vas_id);
+
+		if (result) {
+			lv1_panic(0);
+		}
+
+		map.vas_id = 0;
+	}
+}
+
+static int __init ps3_mm_get_repository_highmem(struct mem_region *r)
+{
+	int result;
+
+	/* Assume a single highmem region. */
+
+	result = ps3_repository_read_highmem_info(0, &r->base, &r->size);
+
+	if (result)
+		goto zero_region;
+
+	if (!r->base || !r->size) {
+		result = -1;
+		goto zero_region;
+	}
+
+	r->offset = r->base - map.rm.size;
+
+	DBG("%s:%d: Found high region in repository: %llxh %llxh\n",
+	    __func__, __LINE__, r->base, r->size);
+
+	return 0;
+
+zero_region:
+	DBG("%s:%d: No high region in repository.\n", __func__, __LINE__);
+
+	r->size = r->base = r->offset = 0;
+	return result;
+}
+
+static int ps3_mm_set_repository_highmem(const struct mem_region *r)
+{
+	/* Assume a single highmem region. */
+
+	return r ? ps3_repository_write_highmem_info(0, r->base, r->size) :
+		ps3_repository_write_highmem_info(0, 0, 0);
+}
+
+/**
+ * ps3_mm_region_create - create a memory region in the vas
+ * @r: pointer to a struct mem_region to accept initialized values
+ * @size: requested region size
+ *
+ * This implementation creates the region with the vas large page size.
+ * @size is rounded down to a multiple of the vas large page size.
+ */
+
+static int ps3_mm_region_create(struct mem_region *r, unsigned long size)
+{
+	int result;
+	u64 muid;
+
+	r->size = ALIGN_DOWN(size, 1 << PAGE_SHIFT_16M);
+
+	DBG("%s:%d requested  %lxh\n", __func__, __LINE__, size);
+	DBG("%s:%d actual     %llxh\n", __func__, __LINE__, r->size);
+	DBG("%s:%d difference %llxh (%lluMB)\n", __func__, __LINE__,
+		size - r->size, (size - r->size) / 1024 / 1024);
+
+	if (r->size == 0) {
+		DBG("%s:%d: size == 0\n", __func__, __LINE__);
+		result = -1;
+		goto zero_region;
+	}
+
+	result = lv1_allocate_memory(r->size, PAGE_SHIFT_16M, 0,
+		ALLOCATE_MEMORY_TRY_ALT_UNIT, &r->base, &muid);
+
+	if (result || r->base < map.rm.size) {
+		DBG("%s:%d: lv1_allocate_memory failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+		goto zero_region;
+	}
+
+	r->destroy = 1;
+	r->offset = r->base - map.rm.size;
+	return result;
+
+zero_region:
+	r->size = r->base = r->offset = 0;
+	return result;
+}
+
+/**
+ * ps3_mm_region_destroy - destroy a memory region
+ * @r: pointer to struct mem_region
+ */
+
+static void ps3_mm_region_destroy(struct mem_region *r)
+{
+	int result;
+
+	if (!r->destroy) {
+		return;
+	}
+
+	if (r->base) {
+		result = lv1_release_memory(r->base);
+
+		if (result) {
+			lv1_panic(0);
+		}
+
+		r->size = r->base = r->offset = 0;
+		map.total = map.rm.size;
+	}
+
+	ps3_mm_set_repository_highmem(NULL);
+}
+
+/*============================================================================*/
+/* dma routines                                                               */
+/*============================================================================*/
+
+/**
+ * dma_sb_lpar_to_bus - Translate an lpar address to ioc mapped bus address.
+ * @r: pointer to dma region structure
+ * @lpar_addr: HV lpar address
+ */
+
+static unsigned long dma_sb_lpar_to_bus(struct ps3_dma_region *r,
+	unsigned long lpar_addr)
+{
+	if (lpar_addr >= map.rm.size)
+		lpar_addr -= map.r1.offset;
+	BUG_ON(lpar_addr < r->offset);
+	BUG_ON(lpar_addr >= r->offset + r->len);
+	return r->bus_addr + lpar_addr - r->offset;
+}
+
+#define dma_dump_region(_a) _dma_dump_region(_a, __func__, __LINE__)
+static void  __maybe_unused _dma_dump_region(const struct ps3_dma_region *r,
+	const char *func, int line)
+{
+	DBG("%s:%d: dev        %llu:%llu\n", func, line, r->dev->bus_id,
+		r->dev->dev_id);
+	DBG("%s:%d: page_size  %u\n", func, line, r->page_size);
+	DBG("%s:%d: bus_addr   %lxh\n", func, line, r->bus_addr);
+	DBG("%s:%d: len        %lxh\n", func, line, r->len);
+	DBG("%s:%d: offset     %lxh\n", func, line, r->offset);
+}
+
+  /**
+ * dma_chunk - A chunk of dma pages mapped by the io controller.
+ * @region - The dma region that owns this chunk.
+ * @lpar_addr: Starting lpar address of the area to map.
+ * @bus_addr: Starting ioc bus address of the area to map.
+ * @len: Length in bytes of the area to map.
+ * @link: A struct list_head used with struct ps3_dma_region.chunk_list, the
+ * list of all chunks owned by the region.
+ *
+ * This implementation uses a very simple dma page manager
+ * based on the dma_chunk structure.  This scheme assumes
+ * that all drivers use very well behaved dma ops.
+ */
+
+struct dma_chunk {
+	struct ps3_dma_region *region;
+	unsigned long lpar_addr;
+	unsigned long bus_addr;
+	unsigned long len;
+	struct list_head link;
+	unsigned int usage_count;
+};
+
+#define dma_dump_chunk(_a) _dma_dump_chunk(_a, __func__, __LINE__)
+static void _dma_dump_chunk (const struct dma_chunk* c, const char* func,
+	int line)
+{
+	DBG("%s:%d: r.dev        %llu:%llu\n", func, line,
+		c->region->dev->bus_id, c->region->dev->dev_id);
+	DBG("%s:%d: r.bus_addr   %lxh\n", func, line, c->region->bus_addr);
+	DBG("%s:%d: r.page_size  %u\n", func, line, c->region->page_size);
+	DBG("%s:%d: r.len        %lxh\n", func, line, c->region->len);
+	DBG("%s:%d: r.offset     %lxh\n", func, line, c->region->offset);
+	DBG("%s:%d: c.lpar_addr  %lxh\n", func, line, c->lpar_addr);
+	DBG("%s:%d: c.bus_addr   %lxh\n", func, line, c->bus_addr);
+	DBG("%s:%d: c.len        %lxh\n", func, line, c->len);
+}
+
+static struct dma_chunk * dma_find_chunk(struct ps3_dma_region *r,
+	unsigned long bus_addr, unsigned long len)
+{
+	struct dma_chunk *c;
+	unsigned long aligned_bus = ALIGN_DOWN(bus_addr, 1 << r->page_size);
+	unsigned long aligned_len = ALIGN(len+bus_addr-aligned_bus,
+					      1 << r->page_size);
+
+	list_for_each_entry(c, &r->chunk_list.head, link) {
+		/* intersection */
+		if (aligned_bus >= c->bus_addr &&
+		    aligned_bus + aligned_len <= c->bus_addr + c->len)
+			return c;
+
+		/* below */
+		if (aligned_bus + aligned_len <= c->bus_addr)
+			continue;
+
+		/* above */
+		if (aligned_bus >= c->bus_addr + c->len)
+			continue;
+
+		/* we don't handle the multi-chunk case for now */
+		dma_dump_chunk(c);
+		BUG();
+	}
+	return NULL;
+}
+
+static struct dma_chunk *dma_find_chunk_lpar(struct ps3_dma_region *r,
+	unsigned long lpar_addr, unsigned long len)
+{
+	struct dma_chunk *c;
+	unsigned long aligned_lpar = ALIGN_DOWN(lpar_addr, 1 << r->page_size);
+	unsigned long aligned_len = ALIGN(len + lpar_addr - aligned_lpar,
+					      1 << r->page_size);
+
+	list_for_each_entry(c, &r->chunk_list.head, link) {
+		/* intersection */
+		if (c->lpar_addr <= aligned_lpar &&
+		    aligned_lpar < c->lpar_addr + c->len) {
+			if (aligned_lpar + aligned_len <= c->lpar_addr + c->len)
+				return c;
+			else {
+				dma_dump_chunk(c);
+				BUG();
+			}
+		}
+		/* below */
+		if (aligned_lpar + aligned_len <= c->lpar_addr) {
+			continue;
+		}
+		/* above */
+		if (c->lpar_addr + c->len <= aligned_lpar) {
+			continue;
+		}
+	}
+	return NULL;
+}
+
+static int dma_sb_free_chunk(struct dma_chunk *c)
+{
+	int result = 0;
+
+	if (c->bus_addr) {
+		result = lv1_unmap_device_dma_region(c->region->dev->bus_id,
+			c->region->dev->dev_id, c->bus_addr, c->len);
+		BUG_ON(result);
+	}
+
+	kfree(c);
+	return result;
+}
+
+static int dma_ioc0_free_chunk(struct dma_chunk *c)
+{
+	int result = 0;
+	int iopage;
+	unsigned long offset;
+	struct ps3_dma_region *r = c->region;
+
+	DBG("%s:start\n", __func__);
+	for (iopage = 0; iopage < (c->len >> r->page_size); iopage++) {
+		offset = (1 << r->page_size) * iopage;
+		/* put INVALID entry */
+		result = lv1_put_iopte(0,
+				       c->bus_addr + offset,
+				       c->lpar_addr + offset,
+				       r->ioid,
+				       0);
+		DBG("%s: bus=%#lx, lpar=%#lx, ioid=%d\n", __func__,
+		    c->bus_addr + offset,
+		    c->lpar_addr + offset,
+		    r->ioid);
+
+		if (result) {
+			DBG("%s:%d: lv1_put_iopte failed: %s\n", __func__,
+			    __LINE__, ps3_result(result));
+		}
+	}
+	kfree(c);
+	DBG("%s:end\n", __func__);
+	return result;
+}
+
+/**
+ * dma_sb_map_pages - Maps dma pages into the io controller bus address space.
+ * @r: Pointer to a struct ps3_dma_region.
+ * @phys_addr: Starting physical address of the area to map.
+ * @len: Length in bytes of the area to map.
+ * c_out: A pointer to receive an allocated struct dma_chunk for this area.
+ *
+ * This is the lowest level dma mapping routine, and is the one that will
+ * make the HV call to add the pages into the io controller address space.
+ */
+
+static int dma_sb_map_pages(struct ps3_dma_region *r, unsigned long phys_addr,
+	    unsigned long len, struct dma_chunk **c_out, u64 iopte_flag)
+{
+	int result;
+	struct dma_chunk *c;
+
+	c = kzalloc(sizeof(*c), GFP_ATOMIC);
+	if (!c) {
+		result = -ENOMEM;
+		goto fail_alloc;
+	}
+
+	c->region = r;
+	c->lpar_addr = ps3_mm_phys_to_lpar(phys_addr);
+	c->bus_addr = dma_sb_lpar_to_bus(r, c->lpar_addr);
+	c->len = len;
+
+	BUG_ON(iopte_flag != 0xf800000000000000UL);
+	result = lv1_map_device_dma_region(c->region->dev->bus_id,
+					   c->region->dev->dev_id, c->lpar_addr,
+					   c->bus_addr, c->len, iopte_flag);
+	if (result) {
+		DBG("%s:%d: lv1_map_device_dma_region failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+		goto fail_map;
+	}
+
+	list_add(&c->link, &r->chunk_list.head);
+
+	*c_out = c;
+	return 0;
+
+fail_map:
+	kfree(c);
+fail_alloc:
+	*c_out = NULL;
+	DBG(" <- %s:%d\n", __func__, __LINE__);
+	return result;
+}
+
+static int dma_ioc0_map_pages(struct ps3_dma_region *r, unsigned long phys_addr,
+			      unsigned long len, struct dma_chunk **c_out,
+			      u64 iopte_flag)
+{
+	int result;
+	struct dma_chunk *c, *last;
+	int iopage, pages;
+	unsigned long offset;
+
+	DBG(KERN_ERR "%s: phy=%#lx, lpar%#lx, len=%#lx\n", __func__,
+	    phys_addr, ps3_mm_phys_to_lpar(phys_addr), len);
+	c = kzalloc(sizeof(*c), GFP_ATOMIC);
+	if (!c) {
+		result = -ENOMEM;
+		goto fail_alloc;
+	}
+
+	c->region = r;
+	c->len = len;
+	c->lpar_addr = ps3_mm_phys_to_lpar(phys_addr);
+	/* allocate IO address */
+	if (list_empty(&r->chunk_list.head)) {
+		/* first one */
+		c->bus_addr = r->bus_addr;
+	} else {
+		/* derive from last bus addr*/
+		last  = list_entry(r->chunk_list.head.next,
+				   struct dma_chunk, link);
+		c->bus_addr = last->bus_addr + last->len;
+		DBG("%s: last bus=%#lx, len=%#lx\n", __func__,
+		    last->bus_addr, last->len);
+	}
+
+	/* FIXME: check whether length exceeds region size */
+
+	/* build ioptes for the area */
+	pages = len >> r->page_size;
+	DBG("%s: pgsize=%#x len=%#lx pages=%#x iopteflag=%#llx\n", __func__,
+	    r->page_size, r->len, pages, iopte_flag);
+	for (iopage = 0; iopage < pages; iopage++) {
+		offset = (1 << r->page_size) * iopage;
+		result = lv1_put_iopte(0,
+				       c->bus_addr + offset,
+				       c->lpar_addr + offset,
+				       r->ioid,
+				       iopte_flag);
+		if (result) {
+			pr_warn("%s:%d: lv1_put_iopte failed: %s\n",
+				__func__, __LINE__, ps3_result(result));
+			goto fail_map;
+		}
+		DBG("%s: pg=%d bus=%#lx, lpar=%#lx, ioid=%#x\n", __func__,
+		    iopage, c->bus_addr + offset, c->lpar_addr + offset,
+		    r->ioid);
+	}
+
+	/* be sure that last allocated one is inserted at head */
+	list_add(&c->link, &r->chunk_list.head);
+
+	*c_out = c;
+	DBG("%s: end\n", __func__);
+	return 0;
+
+fail_map:
+	for (iopage--; 0 <= iopage; iopage--) {
+		lv1_put_iopte(0,
+			      c->bus_addr + offset,
+			      c->lpar_addr + offset,
+			      r->ioid,
+			      0);
+	}
+	kfree(c);
+fail_alloc:
+	*c_out = NULL;
+	return result;
+}
+
+/**
+ * dma_sb_region_create - Create a device dma region.
+ * @r: Pointer to a struct ps3_dma_region.
+ *
+ * This is the lowest level dma region create routine, and is the one that
+ * will make the HV call to create the region.
+ */
+
+static int dma_sb_region_create(struct ps3_dma_region *r)
+{
+	int result;
+	u64 bus_addr;
+
+	DBG(" -> %s:%d:\n", __func__, __LINE__);
+
+	BUG_ON(!r);
+
+	if (!r->dev->bus_id) {
+		pr_info("%s:%d: %llu:%llu no dma\n", __func__, __LINE__,
+			r->dev->bus_id, r->dev->dev_id);
+		return 0;
+	}
+
+	DBG("%s:%u: len = 0x%lx, page_size = %u, offset = 0x%lx\n", __func__,
+	    __LINE__, r->len, r->page_size, r->offset);
+
+	BUG_ON(!r->len);
+	BUG_ON(!r->page_size);
+	BUG_ON(!r->region_ops);
+
+	INIT_LIST_HEAD(&r->chunk_list.head);
+	spin_lock_init(&r->chunk_list.lock);
+
+	result = lv1_allocate_device_dma_region(r->dev->bus_id, r->dev->dev_id,
+		roundup_pow_of_two(r->len), r->page_size, r->region_type,
+		&bus_addr);
+	r->bus_addr = bus_addr;
+
+	if (result) {
+		DBG("%s:%d: lv1_allocate_device_dma_region failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+		r->len = r->bus_addr = 0;
+	}
+
+	return result;
+}
+
+static int dma_ioc0_region_create(struct ps3_dma_region *r)
+{
+	int result;
+	u64 bus_addr;
+
+	INIT_LIST_HEAD(&r->chunk_list.head);
+	spin_lock_init(&r->chunk_list.lock);
+
+	result = lv1_allocate_io_segment(0,
+					 r->len,
+					 r->page_size,
+					 &bus_addr);
+	r->bus_addr = bus_addr;
+	if (result) {
+		DBG("%s:%d: lv1_allocate_io_segment failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+		r->len = r->bus_addr = 0;
+	}
+	DBG("%s: len=%#lx, pg=%d, bus=%#lx\n", __func__,
+	    r->len, r->page_size, r->bus_addr);
+	return result;
+}
+
+/**
+ * dma_region_free - Free a device dma region.
+ * @r: Pointer to a struct ps3_dma_region.
+ *
+ * This is the lowest level dma region free routine, and is the one that
+ * will make the HV call to free the region.
+ */
+
+static int dma_sb_region_free(struct ps3_dma_region *r)
+{
+	int result;
+	struct dma_chunk *c;
+	struct dma_chunk *tmp;
+
+	BUG_ON(!r);
+
+	if (!r->dev->bus_id) {
+		pr_info("%s:%d: %llu:%llu no dma\n", __func__, __LINE__,
+			r->dev->bus_id, r->dev->dev_id);
+		return 0;
+	}
+
+	list_for_each_entry_safe(c, tmp, &r->chunk_list.head, link) {
+		list_del(&c->link);
+		dma_sb_free_chunk(c);
+	}
+
+	result = lv1_free_device_dma_region(r->dev->bus_id, r->dev->dev_id,
+		r->bus_addr);
+
+	if (result)
+		DBG("%s:%d: lv1_free_device_dma_region failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+
+	r->bus_addr = 0;
+
+	return result;
+}
+
+static int dma_ioc0_region_free(struct ps3_dma_region *r)
+{
+	int result;
+	struct dma_chunk *c, *n;
+
+	DBG("%s: start\n", __func__);
+	list_for_each_entry_safe(c, n, &r->chunk_list.head, link) {
+		list_del(&c->link);
+		dma_ioc0_free_chunk(c);
+	}
+
+	result = lv1_release_io_segment(0, r->bus_addr);
+
+	if (result)
+		DBG("%s:%d: lv1_free_device_dma_region failed: %s\n",
+			__func__, __LINE__, ps3_result(result));
+
+	r->bus_addr = 0;
+	DBG("%s: end\n", __func__);
+
+	return result;
+}
+
+/**
+ * dma_sb_map_area - Map an area of memory into a device dma region.
+ * @r: Pointer to a struct ps3_dma_region.
+ * @virt_addr: Starting virtual address of the area to map.
+ * @len: Length in bytes of the area to map.
+ * @bus_addr: A pointer to return the starting ioc bus address of the area to
+ * map.
+ *
+ * This is the common dma mapping routine.
+ */
+
+static int dma_sb_map_area(struct ps3_dma_region *r, unsigned long virt_addr,
+	   unsigned long len, dma_addr_t *bus_addr,
+	   u64 iopte_flag)
+{
+	int result;
+	unsigned long flags;
+	struct dma_chunk *c;
+	unsigned long phys_addr = is_kernel_addr(virt_addr) ? __pa(virt_addr)
+		: virt_addr;
+	unsigned long aligned_phys = ALIGN_DOWN(phys_addr, 1 << r->page_size);
+	unsigned long aligned_len = ALIGN(len + phys_addr - aligned_phys,
+					      1 << r->page_size);
+	*bus_addr = dma_sb_lpar_to_bus(r, ps3_mm_phys_to_lpar(phys_addr));
+
+	if (!USE_DYNAMIC_DMA) {
+		unsigned long lpar_addr = ps3_mm_phys_to_lpar(phys_addr);
+		DBG(" -> %s:%d\n", __func__, __LINE__);
+		DBG("%s:%d virt_addr %lxh\n", __func__, __LINE__,
+			virt_addr);
+		DBG("%s:%d phys_addr %lxh\n", __func__, __LINE__,
+			phys_addr);
+		DBG("%s:%d lpar_addr %lxh\n", __func__, __LINE__,
+			lpar_addr);
+		DBG("%s:%d len       %lxh\n", __func__, __LINE__, len);
+		DBG("%s:%d bus_addr  %llxh (%lxh)\n", __func__, __LINE__,
+		*bus_addr, len);
+	}
+
+	spin_lock_irqsave(&r->chunk_list.lock, flags);
+	c = dma_find_chunk(r, *bus_addr, len);
+
+	if (c) {
+		DBG("%s:%d: reusing mapped chunk", __func__, __LINE__);
+		dma_dump_chunk(c);
+		c->usage_count++;
+		spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+		return 0;
+	}
+
+	result = dma_sb_map_pages(r, aligned_phys, aligned_len, &c, iopte_flag);
+
+	if (result) {
+		*bus_addr = 0;
+		DBG("%s:%d: dma_sb_map_pages failed (%d)\n",
+			__func__, __LINE__, result);
+		spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+		return result;
+	}
+
+	c->usage_count = 1;
+
+	spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+	return result;
+}
+
+static int dma_ioc0_map_area(struct ps3_dma_region *r, unsigned long virt_addr,
+	     unsigned long len, dma_addr_t *bus_addr,
+	     u64 iopte_flag)
+{
+	int result;
+	unsigned long flags;
+	struct dma_chunk *c;
+	unsigned long phys_addr = is_kernel_addr(virt_addr) ? __pa(virt_addr)
+		: virt_addr;
+	unsigned long aligned_phys = ALIGN_DOWN(phys_addr, 1 << r->page_size);
+	unsigned long aligned_len = ALIGN(len + phys_addr - aligned_phys,
+					      1 << r->page_size);
+
+	DBG(KERN_ERR "%s: vaddr=%#lx, len=%#lx\n", __func__,
+	    virt_addr, len);
+	DBG(KERN_ERR "%s: ph=%#lx a_ph=%#lx a_l=%#lx\n", __func__,
+	    phys_addr, aligned_phys, aligned_len);
+
+	spin_lock_irqsave(&r->chunk_list.lock, flags);
+	c = dma_find_chunk_lpar(r, ps3_mm_phys_to_lpar(phys_addr), len);
+
+	if (c) {
+		/* FIXME */
+		BUG();
+		*bus_addr = c->bus_addr + phys_addr - aligned_phys;
+		c->usage_count++;
+		spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+		return 0;
+	}
+
+	result = dma_ioc0_map_pages(r, aligned_phys, aligned_len, &c,
+				    iopte_flag);
+
+	if (result) {
+		*bus_addr = 0;
+		DBG("%s:%d: dma_ioc0_map_pages failed (%d)\n",
+			__func__, __LINE__, result);
+		spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+		return result;
+	}
+	*bus_addr = c->bus_addr + phys_addr - aligned_phys;
+	DBG("%s: va=%#lx pa=%#lx a_pa=%#lx bus=%#llx\n", __func__,
+	    virt_addr, phys_addr, aligned_phys, *bus_addr);
+	c->usage_count = 1;
+
+	spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+	return result;
+}
+
+/**
+ * dma_sb_unmap_area - Unmap an area of memory from a device dma region.
+ * @r: Pointer to a struct ps3_dma_region.
+ * @bus_addr: The starting ioc bus address of the area to unmap.
+ * @len: Length in bytes of the area to unmap.
+ *
+ * This is the common dma unmap routine.
+ */
+
+static int dma_sb_unmap_area(struct ps3_dma_region *r, dma_addr_t bus_addr,
+	unsigned long len)
+{
+	unsigned long flags;
+	struct dma_chunk *c;
+
+	spin_lock_irqsave(&r->chunk_list.lock, flags);
+	c = dma_find_chunk(r, bus_addr, len);
+
+	if (!c) {
+		unsigned long aligned_bus = ALIGN_DOWN(bus_addr,
+			1 << r->page_size);
+		unsigned long aligned_len = ALIGN(len + bus_addr
+			- aligned_bus, 1 << r->page_size);
+		DBG("%s:%d: not found: bus_addr %llxh\n",
+			__func__, __LINE__, bus_addr);
+		DBG("%s:%d: not found: len %lxh\n",
+			__func__, __LINE__, len);
+		DBG("%s:%d: not found: aligned_bus %lxh\n",
+			__func__, __LINE__, aligned_bus);
+		DBG("%s:%d: not found: aligned_len %lxh\n",
+			__func__, __LINE__, aligned_len);
+		BUG();
+	}
+
+	c->usage_count--;
+
+	if (!c->usage_count) {
+		list_del(&c->link);
+		dma_sb_free_chunk(c);
+	}
+
+	spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+	return 0;
+}
+
+static int dma_ioc0_unmap_area(struct ps3_dma_region *r,
+			dma_addr_t bus_addr, unsigned long len)
+{
+	unsigned long flags;
+	struct dma_chunk *c;
+
+	DBG("%s: start a=%#llx l=%#lx\n", __func__, bus_addr, len);
+	spin_lock_irqsave(&r->chunk_list.lock, flags);
+	c = dma_find_chunk(r, bus_addr, len);
+
+	if (!c) {
+		unsigned long aligned_bus = ALIGN_DOWN(bus_addr,
+							1 << r->page_size);
+		unsigned long aligned_len = ALIGN(len + bus_addr
+						      - aligned_bus,
+						      1 << r->page_size);
+		DBG("%s:%d: not found: bus_addr %llxh\n",
+		    __func__, __LINE__, bus_addr);
+		DBG("%s:%d: not found: len %lxh\n",
+		    __func__, __LINE__, len);
+		DBG("%s:%d: not found: aligned_bus %lxh\n",
+		    __func__, __LINE__, aligned_bus);
+		DBG("%s:%d: not found: aligned_len %lxh\n",
+		    __func__, __LINE__, aligned_len);
+		BUG();
+	}
+
+	c->usage_count--;
+
+	if (!c->usage_count) {
+		list_del(&c->link);
+		dma_ioc0_free_chunk(c);
+	}
+
+	spin_unlock_irqrestore(&r->chunk_list.lock, flags);
+	DBG("%s: end\n", __func__);
+	return 0;
+}
+
+/**
+ * dma_sb_region_create_linear - Setup a linear dma mapping for a device.
+ * @r: Pointer to a struct ps3_dma_region.
+ *
+ * This routine creates an HV dma region for the device and maps all available
+ * ram into the io controller bus address space.
+ */
+
+static int dma_sb_region_create_linear(struct ps3_dma_region *r)
+{
+	int result;
+	unsigned long virt_addr, len;
+	dma_addr_t tmp;
+
+	if (r->len > 16*1024*1024) {	/* FIXME: need proper fix */
+		/* force 16M dma pages for linear mapping */
+		if (r->page_size != PS3_DMA_16M) {
+			pr_info("%s:%d: forcing 16M pages for linear map\n",
+				__func__, __LINE__);
+			r->page_size = PS3_DMA_16M;
+			r->len = ALIGN(r->len, 1 << r->page_size);
+		}
+	}
+
+	result = dma_sb_region_create(r);
+	BUG_ON(result);
+
+	if (r->offset < map.rm.size) {
+		/* Map (part of) 1st RAM chunk */
+		virt_addr = map.rm.base + r->offset;
+		len = map.rm.size - r->offset;
+		if (len > r->len)
+			len = r->len;
+		result = dma_sb_map_area(r, virt_addr, len, &tmp,
+			CBE_IOPTE_PP_W | CBE_IOPTE_PP_R | CBE_IOPTE_SO_RW |
+			CBE_IOPTE_M);
+		BUG_ON(result);
+	}
+
+	if (r->offset + r->len > map.rm.size) {
+		/* Map (part of) 2nd RAM chunk */
+		virt_addr = map.rm.size;
+		len = r->len;
+		if (r->offset >= map.rm.size)
+			virt_addr += r->offset - map.rm.size;
+		else
+			len -= map.rm.size - r->offset;
+		result = dma_sb_map_area(r, virt_addr, len, &tmp,
+			CBE_IOPTE_PP_W | CBE_IOPTE_PP_R | CBE_IOPTE_SO_RW |
+			CBE_IOPTE_M);
+		BUG_ON(result);
+	}
+
+	return result;
+}
+
+/**
+ * dma_sb_region_free_linear - Free a linear dma mapping for a device.
+ * @r: Pointer to a struct ps3_dma_region.
+ *
+ * This routine will unmap all mapped areas and free the HV dma region.
+ */
+
+static int dma_sb_region_free_linear(struct ps3_dma_region *r)
+{
+	int result;
+	dma_addr_t bus_addr;
+	unsigned long len, lpar_addr;
+
+	if (r->offset < map.rm.size) {
+		/* Unmap (part of) 1st RAM chunk */
+		lpar_addr = map.rm.base + r->offset;
+		len = map.rm.size - r->offset;
+		if (len > r->len)
+			len = r->len;
+		bus_addr = dma_sb_lpar_to_bus(r, lpar_addr);
+		result = dma_sb_unmap_area(r, bus_addr, len);
+		BUG_ON(result);
+	}
+
+	if (r->offset + r->len > map.rm.size) {
+		/* Unmap (part of) 2nd RAM chunk */
+		lpar_addr = map.r1.base;
+		len = r->len;
+		if (r->offset >= map.rm.size)
+			lpar_addr += r->offset - map.rm.size;
+		else
+			len -= map.rm.size - r->offset;
+		bus_addr = dma_sb_lpar_to_bus(r, lpar_addr);
+		result = dma_sb_unmap_area(r, bus_addr, len);
+		BUG_ON(result);
+	}
+
+	result = dma_sb_region_free(r);
+	BUG_ON(result);
+
+	return result;
+}
+
+/**
+ * dma_sb_map_area_linear - Map an area of memory into a device dma region.
+ * @r: Pointer to a struct ps3_dma_region.
+ * @virt_addr: Starting virtual address of the area to map.
+ * @len: Length in bytes of the area to map.
+ * @bus_addr: A pointer to return the starting ioc bus address of the area to
+ * map.
+ *
+ * This routine just returns the corresponding bus address.  Actual mapping
+ * occurs in dma_region_create_linear().
+ */
+
+static int dma_sb_map_area_linear(struct ps3_dma_region *r,
+	unsigned long virt_addr, unsigned long len, dma_addr_t *bus_addr,
+	u64 iopte_flag)
+{
+	unsigned long phys_addr = is_kernel_addr(virt_addr) ? __pa(virt_addr)
+		: virt_addr;
+	*bus_addr = dma_sb_lpar_to_bus(r, ps3_mm_phys_to_lpar(phys_addr));
+	return 0;
+}
+
+/**
+ * dma_unmap_area_linear - Unmap an area of memory from a device dma region.
+ * @r: Pointer to a struct ps3_dma_region.
+ * @bus_addr: The starting ioc bus address of the area to unmap.
+ * @len: Length in bytes of the area to unmap.
+ *
+ * This routine does nothing.  Unmapping occurs in dma_sb_region_free_linear().
+ */
+
+static int dma_sb_unmap_area_linear(struct ps3_dma_region *r,
+	dma_addr_t bus_addr, unsigned long len)
+{
+	return 0;
+};
+
+static const struct ps3_dma_region_ops ps3_dma_sb_region_ops =  {
+	.create = dma_sb_region_create,
+	.free = dma_sb_region_free,
+	.map = dma_sb_map_area,
+	.unmap = dma_sb_unmap_area
+};
+
+static const struct ps3_dma_region_ops ps3_dma_sb_region_linear_ops = {
+	.create = dma_sb_region_create_linear,
+	.free = dma_sb_region_free_linear,
+	.map = dma_sb_map_area_linear,
+	.unmap = dma_sb_unmap_area_linear
+};
+
+static const struct ps3_dma_region_ops ps3_dma_ioc0_region_ops = {
+	.create = dma_ioc0_region_create,
+	.free = dma_ioc0_region_free,
+	.map = dma_ioc0_map_area,
+	.unmap = dma_ioc0_unmap_area
+};
+
+int ps3_dma_region_init(struct ps3_system_bus_device *dev,
+	struct ps3_dma_region *r, enum ps3_dma_page_size page_size,
+	enum ps3_dma_region_type region_type, void *addr, unsigned long len)
+{
+	unsigned long lpar_addr;
+	int result;
+
+	lpar_addr = addr ? ps3_mm_phys_to_lpar(__pa(addr)) : 0;
+
+	r->dev = dev;
+	r->page_size = page_size;
+	r->region_type = region_type;
+	r->offset = lpar_addr;
+	if (r->offset >= map.rm.size)
+		r->offset -= map.r1.offset;
+	r->len = len ? len : ALIGN(map.total, 1 << r->page_size);
+
+	dev->core.dma_mask = &r->dma_mask;
+
+	result = dma_set_mask_and_coherent(&dev->core, DMA_BIT_MASK(32));
+
+	if (result < 0) {
+		dev_err(&dev->core, "%s:%d: dma_set_mask_and_coherent failed: %d\n",
+			__func__, __LINE__, result);
+		return result;
+	}
+
+	switch (dev->dev_type) {
+	case PS3_DEVICE_TYPE_SB:
+		r->region_ops =  (USE_DYNAMIC_DMA)
+			? &ps3_dma_sb_region_ops
+			: &ps3_dma_sb_region_linear_ops;
+		break;
+	case PS3_DEVICE_TYPE_IOC0:
+		r->region_ops = &ps3_dma_ioc0_region_ops;
+		break;
+	default:
+		BUG();
+		return -EINVAL;
+	}
+	return 0;
+}
+EXPORT_SYMBOL(ps3_dma_region_init);
+
+int ps3_dma_region_create(struct ps3_dma_region *r)
+{
+	BUG_ON(!r);
+	BUG_ON(!r->region_ops);
+	BUG_ON(!r->region_ops->create);
+	return r->region_ops->create(r);
+}
+EXPORT_SYMBOL(ps3_dma_region_create);
+
+int ps3_dma_region_free(struct ps3_dma_region *r)
+{
+	BUG_ON(!r);
+	BUG_ON(!r->region_ops);
+	BUG_ON(!r->region_ops->free);
+	return r->region_ops->free(r);
+}
+EXPORT_SYMBOL(ps3_dma_region_free);
+
+int ps3_dma_map(struct ps3_dma_region *r, unsigned long virt_addr,
+	unsigned long len, dma_addr_t *bus_addr,
+	u64 iopte_flag)
+{
+	return r->region_ops->map(r, virt_addr, len, bus_addr, iopte_flag);
+}
+
+int ps3_dma_unmap(struct ps3_dma_region *r, dma_addr_t bus_addr,
+	unsigned long len)
+{
+	return r->region_ops->unmap(r, bus_addr, len);
+}
+
+/*============================================================================*/
+/* system startup routines                                                    */
+/*============================================================================*/
+
+/**
+ * ps3_mm_init - initialize the address space state variables
+ */
+
+void __init ps3_mm_init(void)
+{
+	int result;
+
+	DBG(" -> %s:%d\n", __func__, __LINE__);
+
+	result = ps3_repository_read_mm_info(&map.rm.base, &map.rm.size,
+		&map.total);
+
+	if (result)
+		panic("ps3_repository_read_mm_info() failed");
+
+	map.rm.offset = map.rm.base;
+	map.vas_id = map.htab_size = 0;
+
+	/* this implementation assumes map.rm.base is zero */
+
+	BUG_ON(map.rm.base);
+	BUG_ON(!map.rm.size);
+
+	/* Check if we got the highmem region from an earlier boot step */
+
+	if (ps3_mm_get_repository_highmem(&map.r1)) {
+		result = ps3_mm_region_create(&map.r1, map.total - map.rm.size);
+
+		if (!result)
+			ps3_mm_set_repository_highmem(&map.r1);
+	}
+
+	/* correct map.total for the real total amount of memory we use */
+	map.total = map.rm.size + map.r1.size;
+
+	if (!map.r1.size) {
+		DBG("%s:%d: No highmem region found\n", __func__, __LINE__);
+	} else {
+		DBG("%s:%d: Adding highmem region: %llxh %llxh\n",
+			__func__, __LINE__, map.rm.size,
+			map.total - map.rm.size);
+		memblock_add(map.rm.size, map.total - map.rm.size);
+	}
+
+	DBG(" <- %s:%d\n", __func__, __LINE__);
+}
+
+/**
+ * ps3_mm_shutdown - final cleanup of address space
+ *
+ * called during kexec sequence with MMU off.
+ */
+
+notrace void ps3_mm_shutdown(void)
+{
+	ps3_mm_region_destroy(&map.r1);
+}