From: John Feeney <jfeeney@redhat.com>
Date: Tue, 21 Apr 2009 19:02:36 -0400
Subject: [misc] I/O AT: new dmaengine_v3.c
Message-id: 49EE508C.4020307@redhat.com
O-Subject: [RHEL5.4 PATCH 6/8] Update I/O AT: new dmaengine_v3.c
Bugzilla: 436048
RH-Acked-by: Prarit Bhargava <prarit@redhat.com>
RH-Acked-by: John W. Linville <linville@redhat.com>

bz436048 Update I/O AT code to upstream

This patch adds Kconfig, Makefile and dmaengine_v3.c in drivers/dma_v3.

 Kconfig        |   34 +++
 Makefile       |    4
 dmaengine_v3.c |  634
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++
 3 files changed, 672 insertions(+)

The commits to each file and any modifications are listed below.

Note: upstream commits come from drivers/dma but patch
adds files to the new directory, drivers/dma_v3.

drivers/dma_v3/Kconfig
 Same as these commits with exception of adding "_V3" or
  "version 3" to comments in order to differentiate with dma/Kconfig:
    c13c8260da3155f2cefb63b0d1b7dcdcb405c644 Leech 5/24/06
    0bbd5f4e97ff9c057b385a1886b4aed1fb0300f1 Leech 5/24/06
    db21733488f84a596faaad0d05430b3f51804692 Leech 6/18/06

drivers/dma_v3/Makefile
  Same as commits below but files that are already in drivers/dma have
  "_v3" suffix added to help differentiate and "_V3" added to CONFIG_*
  names.
   c13c8260da3155f2cefb63b0d1b7dcdcb405c644 Leech 5/24/06
   0bbd5f4e97ff9c057b385a1886b4aed1fb0300f1 Leech 6/18/06
   de5506e155276d385712c2aa1c2d9a27cd4ed947 Leech 6/18/06
   1fda5f4e96225c3ed0baded942704c0ae399da23 Nelson 10/16/07
   8ab89567da0cea9bae2c1b5dad47b51c424479e4 Nelson 10/16/07
   2ed6dc34f9ed39bb8e4c81ea1056f0ba56315841 Nelson 10/16/07

drivers/dma_v3/dmaengine_3.c
 In order to differentiate from existing, the new dmaengine is
 named "dmaengine_v3.c".

 The commits found below make up this file with the following exceptions:
 All EXPORT function names have the "_v3" suffix (dma_sync_wait_v3(),
   dma_chan_cleanup_v3(), dma_async_client_register_v3(),
   dma_async_client_unregister_v3(), dma_async_client_chan_request_v3(),
   dma_async_device_register_v3(), dma_async_device_unregister_v3(),
   dma_async_memcpy_buf_to_buf_v3(), dma_async_memcpy_buf_to_pg_v3(),
   dma_async_memcpy_pg_to_pg_v3(), and dma_async_tx_descriptor_init_v3().
 #ifndef __GENKSYMS__ added around linux/mm.h and linux/jiffies.h
   to prevent kabi errors in Brew
 .name in dma_devclass has "_v3" suffix
 All references to dma_device changed to dma_device_v3
 All references to chan->device changed to chan->device_v3
 All references to client->event_callback changed to
client->event_callback_v3
 Use DMA_STATE_RESOURCE_AVAILABLE and
  DMA_STATE_RESOURCE_REMOVED instead of
   DMA_RESOURCE_AVAILABLE and DMA_RESOURCE_REMOVED

  c13c8260da3155f2cefb63b0d1b7dcdcb405c644 Leech 5/24/06
  17f3ae08b6e7fd778371f2cafbd1c988a67ee343 Morton 5/25/06
  6508871eddbbd3e62799f3b6182a6a4fd3ef31d5 Dunlop 7/4/06
  765e3d8a71bbc1f3400667d5cfcfd7b03382d587 Brownell 3/16/07
  ff487fb773749124550a5ad2b7fbfe0376af6f0d Garzik 7/11/07
  7405f74badf46b5d023c5d2b670b4471525f6c91 Williams 07/13/07
  d379b01e9087a582d58f4b678208a4f8d8376fe7 Williams 7/13/07
  348badf1e825323c419dd118f65783db0f7d2ec8 Skinnemoen 11/15/07
  0036731c88fdb5bf4f04a796a30b5e445fc57f54 Williams 2/6/08
  9b941c6660bae673e27c207f1d20d98ef8ecd450 Zhang 3/13/08
  19242d7233df7d658405d4b7ee1758d21414cfaa Wiliams 4/17/08
  ce4d65a5db77e1568c82d5151a746f627c4f6ed5 Williams 4/17/08
  636bdeaa1243327501edfd2a597ed7443eb4239a Williams 4/17/08
  8a5703f846e2363fc466aff3f53608340a1ae33f Siewior 4/21/08
  7cc5bf9a3a84e5a02e23e5739fb894790b37c101 Williams 7/8/08
  848c536a37b8db4e461f14ca15fe29850151c822 Williams 7/8/08
  dc0ee6435cb92ccc81b14ff28d163fecc5a7f120 Skinnemoen 7/8/08

Note: The entire of the RHEL-5 source tree was checked to make
sure  the "_v3" EXPORTed functions replaced the existing functions,
outside of the drivers/dma directory.

diff --git a/drivers/dma_v3/Kconfig b/drivers/dma_v3/Kconfig
new file mode 100644
index 0000000..bfd2c11
--- /dev/null
+++ b/drivers/dma_v3/Kconfig
@@ -0,0 +1,34 @@
+#
+# DMA engine configuration
+#
+
+menu "DMA Engine support version 3"
+
+config DMA_ENGINE_V3
+	bool "Support for DMA engines version 3"
+	---help---
+	   DMA engines offload copy operations from the CPU to dedicated
+	   hardware, allowing the copies to happen asynchronously.
+
+comment "DMA Clients V3"
+
+config NET_DMA
+	bool "Network: TCP receive copy offload"
+	depends on DMA_ENGINE_V3 && NET
+	default y
+	---help---
+	 This enables the use of DMA engines in the network stack to
+	 offload receive copy-to-user operations, freeing CPU cycles.
+	 Since this is the main user of the DMA engine, it should be enabled;
+	 say Y here.
+
+comment "DMA Devices"
+
+config INTEL_IOATDMA_V3
+	tristate "Intel I/OAT DMA support version 3"
+	depends on DMA_ENGINE_V3 && PCI
+	default m
+	---help---
+	  Enable version 3 support for the Intel(R) I/OAT DMA engine.
+
+endmenu
diff --git a/drivers/dma_v3/Makefile b/drivers/dma_v3/Makefile
new file mode 100644
index 0000000..ada56b1
--- /dev/null
+++ b/drivers/dma_v3/Makefile
@@ -0,0 +1,4 @@
+obj-$(CONFIG_DMA_ENGINE_V3) += dmaengine_v3.o
+obj-$(CONFIG_NET_DMA) += iovlock_v3.o
+obj-$(CONFIG_INTEL_IOATDMA_V3) += ioatdma_v3.o
+ioatdma_v3-objs := ioat.o ioat_dma.o ioat_dca.o
diff --git a/drivers/dma_v3/dmaengine_v3.c b/drivers/dma_v3/dmaengine_v3.c
new file mode 100644
index 0000000..04db9a2
--- /dev/null
+++ b/drivers/dma_v3/dmaengine_v3.c
@@ -0,0 +1,634 @@
+/*
+ * Copyright(c) 2004 - 2006 Intel Corporation. All rights reserved.
+ *
+ * This program is free software; you can redistribute it and/or modify it
+ * under the terms of the GNU General Public License as published by the Free
+ * Software Foundation; either version 2 of the License, or (at your option)
+ * any later version.
+ *
+ * This program is distributed in the hope that it will be useful, but WITHOUT
+ * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
+ * FITNESS FOR A PARTICULAR PURPOSE.  See the GNU General Public License for
+ * more details.
+ *
+ * You should have received a copy of the GNU General Public License along with
+ * this program; if not, write to the Free Software Foundation, Inc., 59
+ * Temple Place - Suite 330, Boston, MA  02111-1307, USA.
+ *
+ * The full GNU General Public License is included in this distribution in the
+ * file called COPYING.
+ */
+
+/*
+ * This code implements the DMA subsystem. It provides a HW-neutral interface
+ * for other kernel code to use asynchronous memory copy capabilities,
+ * if present, and allows different HW DMA drivers to register as providing
+ * this capability.
+ *
+ * Due to the fact we are accelerating what is already a relatively fast
+ * operation, the code goes to great lengths to avoid additional overhead,
+ * such as locking.
+ *
+ * LOCKING:
+ *
+ * The subsystem keeps two global lists, dma_device_list and dma_client_list.
+ * Both of these are protected by a mutex, dma_list_mutex.
+ *
+ * Each device has a channels list, which runs unlocked but is never modified
+ * once the device is registered, it's just setup by the driver.
+ *
+ * Each client is responsible for keeping track of the channels it uses.  See
+ * the definition of dma_event_callback in dmaengine.h.
+ *
+ * Each device has a kref, which is initialized to 1 when the device is
+ * registered. A kref_get is done for each class_device registered.  When the
+ * class_device is released, the corresponding kref_put is done in the release
+ * method. Every time one of the device's channels is allocated to a client,
+ * a kref_get occurs.  When the channel is freed, the corresponding kref_put
+ * happens. The device's release function does a completion, so
+ * unregister_device does a remove event, class_device_unregister, a kref_put
+ * for the first reference, then waits on the completion for all other
+ * references to finish.
+ *
+ * Each channel has an open-coded implementation of Rusty Russell's "bigref,"
+ * with a kref and a per_cpu local_t.  A dma_chan_get is called when a client
+ * signals that it wants to use a channel, and dma_chan_put is called when
+ * a channel is removed or a client using it is unregistered.  A client can
+ * take extra references per outstanding transaction, as is the case with
+ * the NET DMA client.  The release function does a kref_put on the device.
+ *	-ChrisL, DanW
+ */
+
+#include <linux/init.h>
+#include <linux/module.h>
+#ifndef __GENKSYMS__
+#include <linux/mm.h>
+#endif
+#include <linux/device.h>
+#include <linux/dmaengine.h>
+#include <linux/hardirq.h>
+#include <linux/spinlock.h>
+#include <linux/percpu.h>
+#include <linux/rcupdate.h>
+#include <linux/mutex.h>
+#ifndef __GENKSYMS__
+#include <linux/jiffies.h>
+#endif
+
+static DEFINE_MUTEX(dma_list_mutex);
+static LIST_HEAD(dma_device_list);
+static LIST_HEAD(dma_client_list);
+
+/* --- sysfs implementation --- */
+
+static ssize_t show_memcpy_count(struct class_device *cd, char *buf)
+{
+	struct dma_chan *chan = container_of(cd, struct dma_chan, class_dev);
+	unsigned long count = 0;
+	int i;
+
+	for_each_possible_cpu(i)
+		count += per_cpu_ptr(chan->local, i)->memcpy_count;
+
+	return sprintf(buf, "%lu\n", count);
+}
+
+static ssize_t show_bytes_transferred(struct class_device *cd, char *buf)
+{
+	struct dma_chan *chan = container_of(cd, struct dma_chan, class_dev);
+	unsigned long count = 0;
+	int i;
+
+	for_each_possible_cpu(i)
+		count += per_cpu_ptr(chan->local, i)->bytes_transferred;
+
+	return sprintf(buf, "%lu\n", count);
+}
+
+static ssize_t show_in_use(struct class_device *cd, char *buf)
+{
+	struct dma_chan *chan = container_of(cd, struct dma_chan, class_dev);
+	int in_use = 0;
+
+	if (unlikely(chan->slow_ref) &&
+		atomic_read(&chan->refcount.refcount) > 1)
+		in_use = 1;
+	else {
+		if (local_read(&(per_cpu_ptr(chan->local,
+			get_cpu())->refcount)) > 0)
+			in_use = 1;
+		put_cpu();
+	}
+
+	return sprintf(buf, "%d\n", in_use);
+}
+
+static struct class_device_attribute dma_class_attrs[] = {
+	__ATTR(memcpy_count, S_IRUGO, show_memcpy_count, NULL),
+	__ATTR(bytes_transferred, S_IRUGO, show_bytes_transferred, NULL),
+	__ATTR(in_use, S_IRUGO, show_in_use, NULL),
+	__ATTR_NULL
+};
+
+static void dma_async_device_cleanup(struct kref *kref);
+
+static void dma_class_dev_release(struct class_device *cd)
+{
+	struct dma_chan *chan = container_of(cd, struct dma_chan, class_dev);
+	kref_put(&chan->device_v3->refcount, dma_async_device_cleanup);
+}
+
+static struct class dma_devclass = {
+	.name		 = "dma_v3",
+	.class_dev_attrs = dma_class_attrs,
+	.release 	 = dma_class_dev_release,
+};
+
+/* --- client and device registration --- */
+
+#define dma_chan_satisfies_mask(chan, mask) \
+	__dma_chan_satisfies_mask((chan), &(mask))
+static int
+__dma_chan_satisfies_mask(struct dma_chan *chan, dma_cap_mask_t *want)
+{
+	dma_cap_mask_t has;
+
+	bitmap_and(has.bits, want->bits, chan->device_v3->cap_mask.bits,
+		DMA_TX_TYPE_END);
+	return bitmap_equal(want->bits, has.bits, DMA_TX_TYPE_END);
+}
+
+/**
+ * dma_client_chan_alloc - try to allocate channels to a client
+ * @client: &dma_client
+ *
+ * Called with dma_list_mutex held.
+ */
+static void dma_client_chan_alloc(struct dma_client *client)
+{
+	struct dma_device_v3 *device;
+	struct dma_chan *chan;
+	int desc;	/* allocated descriptor count */
+	enum dma_state_client ack;
+
+	/* Find a channel */
+	list_for_each_entry(device, &dma_device_list, global_node) {
+		/* Does the client require a specific DMA controller? */
+		if (client->slave && client->slave->dma_dev
+				&& client->slave->dma_dev != device->dev)
+			continue;
+
+		list_for_each_entry(chan, &device->channels, device_node) {
+			if (!dma_chan_satisfies_mask(chan, client->cap_mask))
+				continue;
+
+			desc = chan->device_v3->device_alloc_chan_resources(
+					chan, client);
+			if (desc >= 0) {
+				ack = client->event_callback_v3(client,
+						chan,
+						DMA_STATE_RESOURCE_AVAILABLE);
+
+				/* we are done once this client rejects
+				 * an available resource
+				 */
+				if (ack == DMA_ACK) {
+					dma_chan_get(chan);
+					chan->client_count++;
+				} else if (ack == DMA_NAK)
+					return;
+			}
+		}
+	}
+}
+
+enum dma_status dma_sync_wait_v3(struct dma_chan *chan, dma_cookie_t cookie)
+{
+	enum dma_status status;
+	unsigned long dma_sync_wait_timeout = jiffies + msecs_to_jiffies(5000);
+
+	dma_async_issue_pending(chan);
+	do {
+		status = dma_async_is_tx_complete(chan, cookie, NULL, NULL);
+		if (time_after_eq(jiffies, dma_sync_wait_timeout)) {
+			printk(KERN_ERR "dma_sync_wait_timeout!\n");
+			return DMA_ERROR;
+		}
+	} while (status == DMA_IN_PROGRESS);
+
+	return status;
+}
+EXPORT_SYMBOL(dma_sync_wait_v3);
+
+/**
+ * dma_chan_cleanup_v3 - release a DMA channel's resources
+ * @kref: kernel reference structure that contains the DMA channel device
+ */
+void dma_chan_cleanup_v3(struct kref *kref)
+{
+	struct dma_chan *chan = container_of(kref, struct dma_chan, refcount);
+	chan->device_v3->device_free_chan_resources(chan);
+	kref_put(&chan->device_v3->refcount, dma_async_device_cleanup);
+}
+EXPORT_SYMBOL(dma_chan_cleanup_v3);
+
+static void dma_chan_free_rcu(struct rcu_head *rcu)
+{
+	struct dma_chan *chan = container_of(rcu, struct dma_chan, rcu);
+	int bias = 0x7FFFFFFF;
+	int i;
+	for_each_possible_cpu(i)
+		bias -= local_read(&per_cpu_ptr(chan->local, i)->refcount);
+	atomic_sub(bias, &chan->refcount.refcount);
+	kref_put(&chan->refcount, dma_chan_cleanup_v3);
+}
+
+static void dma_chan_release(struct dma_chan *chan)
+{
+	atomic_add(0x7FFFFFFF, &chan->refcount.refcount);
+	chan->slow_ref = 1;
+	call_rcu(&chan->rcu, dma_chan_free_rcu);
+}
+
+/**
+ * dma_chans_notify_available - broadcast available channels to the clients
+ */
+static void dma_clients_notify_available(void)
+{
+	struct dma_client *client;
+
+	mutex_lock(&dma_list_mutex);
+
+	list_for_each_entry(client, &dma_client_list, global_node)
+		dma_client_chan_alloc(client);
+
+	mutex_unlock(&dma_list_mutex);
+}
+
+/**
+ * dma_chans_notify_available - tell the clients that a channel is going away
+ * @chan: channel on its way out
+ */
+static void dma_clients_notify_removed(struct dma_chan *chan)
+{
+	struct dma_client *client;
+	enum dma_state_client ack;
+
+	mutex_lock(&dma_list_mutex);
+
+	list_for_each_entry(client, &dma_client_list, global_node) {
+		ack = client->event_callback_v3(client, chan,
+				DMA_STATE_RESOURCE_REMOVED);
+
+		/* client was holding resources for this channel so
+		 * free it
+		 */
+		if (ack == DMA_ACK) {
+			dma_chan_put(chan);
+			chan->client_count--;
+		}
+	}
+
+	mutex_unlock(&dma_list_mutex);
+}
+
+/**
+ * dma_async_client_register_v3 - register a &dma_client
+ * @client: ptr to a client structure with valid 'event_callback' and 'cap_mask'
+ */
+void dma_async_client_register_v3(struct dma_client *client)
+{
+	/* validate client data */
+	BUG_ON(dma_has_cap(DMA_SLAVE, client->cap_mask) &&
+		!client->slave);
+
+	mutex_lock(&dma_list_mutex);
+	list_add_tail(&client->global_node, &dma_client_list);
+	mutex_unlock(&dma_list_mutex);
+}
+EXPORT_SYMBOL(dma_async_client_register_v3);
+
+/**
+ * dma_async_client_unregister_v3 - unregister a client and free the &dma_client
+ * @client: &dma_client to free
+ *
+ * Force frees any allocated DMA channels, frees the &dma_client memory
+ */
+void dma_async_client_unregister_v3(struct dma_client *client)
+{
+	struct dma_device_v3 *device;
+	struct dma_chan *chan;
+	enum dma_state_client ack;
+
+	if (!client)
+		return;
+
+	mutex_lock(&dma_list_mutex);
+	/* free all channels the client is holding */
+	list_for_each_entry(device, &dma_device_list, global_node)
+		list_for_each_entry(chan, &device->channels, device_node) {
+			ack = client->event_callback_v3(client, chan,
+				DMA_STATE_RESOURCE_REMOVED);
+
+			if (ack == DMA_ACK) {
+				dma_chan_put(chan);
+				chan->client_count--;
+			}
+		}
+
+	list_del(&client->global_node);
+	mutex_unlock(&dma_list_mutex);
+}
+EXPORT_SYMBOL(dma_async_client_unregister_v3);
+
+/**
+ * dma_async_client_chan_request_v3 - send all available channels to the
+ * client that satisfy the capability mask
+ * @client - requester
+ */
+void dma_async_client_chan_request_v3(struct dma_client *client)
+{
+	mutex_lock(&dma_list_mutex);
+	dma_client_chan_alloc(client);
+	mutex_unlock(&dma_list_mutex);
+}
+EXPORT_SYMBOL(dma_async_client_chan_request_v3);
+
+/**
+ * dma_async_device_register_v3 - registers DMA devices found
+ * @device: &dma_device_v3
+ */
+int dma_async_device_register_v3(struct dma_device_v3 *device)
+{
+	static int id;
+	int chancnt = 0, rc;
+	struct dma_chan* chan;
+
+	if (!device)
+		return -ENODEV;
+
+	/* validate device routines */
+	BUG_ON(dma_has_cap(DMA_MEMCPY, device->cap_mask) &&
+		!device->device_prep_dma_memcpy);
+	BUG_ON(dma_has_cap(DMA_XOR, device->cap_mask) &&
+		!device->device_prep_dma_xor);
+	BUG_ON(dma_has_cap(DMA_ZERO_SUM, device->cap_mask) &&
+		!device->device_prep_dma_zero_sum);
+	BUG_ON(dma_has_cap(DMA_MEMSET, device->cap_mask) &&
+		!device->device_prep_dma_memset);
+	BUG_ON(dma_has_cap(DMA_INTERRUPT, device->cap_mask) &&
+		!device->device_prep_dma_interrupt);
+	BUG_ON(dma_has_cap(DMA_SLAVE, device->cap_mask) &&
+		!device->device_prep_slave_sg);
+	BUG_ON(dma_has_cap(DMA_SLAVE, device->cap_mask) &&
+		!device->device_terminate_all);
+
+	BUG_ON(!device->device_alloc_chan_resources);
+	BUG_ON(!device->device_free_chan_resources);
+	BUG_ON(!device->device_is_tx_complete);
+	BUG_ON(!device->device_issue_pending);
+	BUG_ON(!device->dev);
+
+	init_completion(&device->done);
+	kref_init(&device->refcount);
+	device->dev_id = id++;
+
+	/* represent channels in sysfs. Probably want devs too */
+	list_for_each_entry(chan, &device->channels, device_node) {
+		chan->local = alloc_percpu(typeof(*chan->local));
+		if (chan->local == NULL)
+			continue;
+
+		chan->chan_id = chancnt++;
+		chan->class_dev.class = &dma_devclass;
+		chan->class_dev.dev = NULL;
+		snprintf(chan->class_dev.class_id, BUS_ID_SIZE, "dma%dchan%d",
+		         device->dev_id, chan->chan_id);
+
+		rc = class_device_register(&chan->class_dev);
+		if (rc) {
+			chancnt--;
+			free_percpu(chan->local);
+			chan->local = NULL;
+			goto err_out;
+		}
+
+		/* One for the channel, one of the class device */
+		kref_get(&device->refcount);
+		kref_get(&device->refcount);
+		kref_init(&chan->refcount);
+		chan->client_count = 0;
+		chan->slow_ref = 0;
+		INIT_RCU_HEAD(&chan->rcu);
+	}
+
+	mutex_lock(&dma_list_mutex);
+	list_add_tail(&device->global_node, &dma_device_list);
+	mutex_unlock(&dma_list_mutex);
+
+	dma_clients_notify_available();
+
+	return 0;
+
+err_out:
+	list_for_each_entry(chan, &device->channels, device_node) {
+		if (chan->local == NULL)
+			continue;
+		kref_put(&device->refcount, dma_async_device_cleanup);
+		class_device_unregister(&chan->class_dev);
+		chancnt--;
+		free_percpu(chan->local);
+	}
+	return rc;
+}
+EXPORT_SYMBOL(dma_async_device_register_v3);
+
+/**
+ * dma_async_device_cleanup - function called when all references are released
+ * @kref: kernel reference object
+ */
+static void dma_async_device_cleanup(struct kref *kref)
+{
+	struct dma_device_v3 *device;
+
+	device = container_of(kref, struct dma_device_v3, refcount);
+	complete(&device->done);
+}
+
+/**
+ * dma_async_device_unregister_v3 - unregisters DMA devices
+ * @device: &dma_device
+ */
+void dma_async_device_unregister_v3(struct dma_device_v3 *device)
+{
+	struct dma_chan *chan;
+
+	mutex_lock(&dma_list_mutex);
+	list_del(&device->global_node);
+	mutex_unlock(&dma_list_mutex);
+
+	list_for_each_entry(chan, &device->channels, device_node) {
+		dma_clients_notify_removed(chan);
+		class_device_unregister(&chan->class_dev);
+		dma_chan_release(chan);
+	}
+
+	kref_put(&device->refcount, dma_async_device_cleanup);
+	wait_for_completion(&device->done);
+}
+EXPORT_SYMBOL(dma_async_device_unregister_v3);
+
+/**
+ * dma_async_memcpy_buf_to_buf_v3 - offloaded copy between virtual addresses
+ * @chan: DMA channel to offload copy to
+ * @dest: destination address (virtual)
+ * @src: source address (virtual)
+ * @len: length
+ *
+ * Both @dest and @src must be mappable to a bus address according to the
+ * DMA mapping API rules for streaming mappings.
+ * Both @dest and @src must stay memory resident (kernel memory or locked
+ * user space pages).
+ */
+dma_cookie_t
+dma_async_memcpy_buf_to_buf_v3(struct dma_chan *chan, void *dest,
+			void *src, size_t len)
+{
+	struct dma_device_v3 *dev = chan->device_v3;
+	struct dma_async_tx_descriptor *tx;
+	dma_addr_t dma_dest, dma_src;
+	dma_cookie_t cookie;
+	int cpu;
+
+	dma_src = dma_map_single(dev->dev, src, len, DMA_TO_DEVICE);
+	dma_dest = dma_map_single(dev->dev, dest, len, DMA_FROM_DEVICE);
+	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
+					 DMA_CTRL_ACK);
+
+	if (!tx) {
+		dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
+		dma_unmap_single(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
+		return -ENOMEM;
+	}
+
+	tx->callback = NULL;
+	cookie = tx->tx_submit(tx);
+
+	cpu = get_cpu();
+	per_cpu_ptr(chan->local, cpu)->bytes_transferred += len;
+	per_cpu_ptr(chan->local, cpu)->memcpy_count++;
+	put_cpu();
+
+	return cookie;
+}
+EXPORT_SYMBOL(dma_async_memcpy_buf_to_buf_v3);
+
+/**
+ * dma_async_memcpy_buf_to_pg_v3 - offloaded copy from address to page
+ * @chan: DMA channel to offload copy to
+ * @page: destination page
+ * @offset: offset in page to copy to
+ * @kdata: source address (virtual)
+ * @len: length
+ *
+ * Both @page/@offset and @kdata must be mappable to a bus address according
+ * to the DMA mapping API rules for streaming mappings.
+ * Both @page/@offset and @kdata must stay memory resident (kernel memory or
+ * locked user space pages)
+ */
+dma_cookie_t
+dma_async_memcpy_buf_to_pg_v3(struct dma_chan *chan, struct page *page,
+			unsigned int offset, void *kdata, size_t len)
+{
+	struct dma_device_v3 *dev = chan->device_v3;
+	struct dma_async_tx_descriptor *tx;
+	dma_addr_t dma_dest, dma_src;
+	dma_cookie_t cookie;
+	int cpu;
+
+	dma_src = dma_map_single(dev->dev, kdata, len, DMA_TO_DEVICE);
+	dma_dest = dma_map_page(dev->dev, page, offset, len, DMA_FROM_DEVICE);
+	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
+					 DMA_CTRL_ACK);
+
+	if (!tx) {
+		dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
+		dma_unmap_page(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
+		return -ENOMEM;
+	}
+
+	tx->callback = NULL;
+	cookie = tx->tx_submit(tx);
+
+	cpu = get_cpu();
+	per_cpu_ptr(chan->local, cpu)->bytes_transferred += len;
+	per_cpu_ptr(chan->local, cpu)->memcpy_count++;
+	put_cpu();
+
+	return cookie;
+}
+EXPORT_SYMBOL(dma_async_memcpy_buf_to_pg_v3);
+
+/**
+ * dma_async_memcpy_pg_to_pg_v3 - offloaded copy from page to page
+ * @chan: DMA channel to offload copy to
+ * @dest_pg: destination page
+ * @dest_off: offset in page to copy to
+ * @src_pg: source page
+ * @src_off: offset in page to copy from
+ * @len: length
+ *
+ * Both @dest_page/@dest_off and @src_page/@src_off must be mappable to a bus
+ * address according to the DMA mapping API rules for streaming mappings.
+ * Both @dest_page/@dest_off and @src_page/@src_off must stay memory resident
+ * (kernel memory or locked user space pages).
+ */
+dma_cookie_t
+dma_async_memcpy_pg_to_pg_v3(struct dma_chan *chan, struct page *dest_pg,
+	unsigned int dest_off, struct page *src_pg, unsigned int src_off,
+	size_t len)
+{
+	struct dma_device_v3 *dev = chan->device_v3;
+	struct dma_async_tx_descriptor *tx;
+	dma_addr_t dma_dest, dma_src;
+	dma_cookie_t cookie;
+	int cpu;
+
+	dma_src = dma_map_page(dev->dev, src_pg, src_off, len, DMA_TO_DEVICE);
+	dma_dest = dma_map_page(dev->dev, dest_pg, dest_off, len,
+				DMA_FROM_DEVICE);
+	tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
+					 DMA_CTRL_ACK);
+
+	if (!tx) {
+		dma_unmap_page(dev->dev, dma_src, len, DMA_TO_DEVICE);
+		dma_unmap_page(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
+		return -ENOMEM;
+	}
+
+	tx->callback = NULL;
+	cookie = tx->tx_submit(tx);
+
+	cpu = get_cpu();
+	per_cpu_ptr(chan->local, cpu)->bytes_transferred += len;
+	per_cpu_ptr(chan->local, cpu)->memcpy_count++;
+	put_cpu();
+
+	return cookie;
+}
+EXPORT_SYMBOL(dma_async_memcpy_pg_to_pg_v3);
+
+void dma_async_tx_descriptor_init_v3(struct dma_async_tx_descriptor *tx,
+	struct dma_chan *chan)
+{
+	tx->chan = chan;
+	spin_lock_init(&tx->lock);
+}
+EXPORT_SYMBOL(dma_async_tx_descriptor_init_v3);
+
+static int __init dma_bus_init(void)
+{
+	mutex_init(&dma_list_mutex);
+	return class_register(&dma_devclass);
+}
+subsys_initcall(dma_bus_init);