2006-05-24 08:18:44 +08:00
|
|
|
/*
|
|
|
|
* 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.
|
|
|
|
*
|
2007-07-10 02:56:42 +08:00
|
|
|
* Each client is responsible for keeping track of the channels it uses. See
|
|
|
|
* the definition of dma_event_callback in dmaengine.h.
|
2006-05-24 08:18:44 +08:00
|
|
|
*
|
|
|
|
* Each device has a kref, which is initialized to 1 when the device is
|
2007-09-25 08:03:03 +08:00
|
|
|
* registered. A kref_get is done for each device registered. When the
|
2008-04-22 06:38:45 +08:00
|
|
|
* device is released, the corresponding kref_put is done in the release
|
2006-05-24 08:18:44 +08:00
|
|
|
* method. Every time one of the device's channels is allocated to a client,
|
2008-04-22 06:38:45 +08:00
|
|
|
* a kref_get occurs. When the channel is freed, the corresponding kref_put
|
2006-05-24 08:18:44 +08:00
|
|
|
* happens. The device's release function does a completion, so
|
2007-09-25 08:03:03 +08:00
|
|
|
* unregister_device does a remove event, device_unregister, a kref_put
|
2006-05-24 08:18:44 +08:00
|
|
|
* 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,"
|
2007-07-10 02:56:42 +08:00
|
|
|
* 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
|
2008-04-22 06:38:45 +08:00
|
|
|
* a channel is removed or a client using it is unregistered. A client can
|
2007-07-10 02:56:42 +08:00
|
|
|
* 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
|
2006-05-24 08:18:44 +08:00
|
|
|
*/
|
|
|
|
|
|
|
|
#include <linux/init.h>
|
|
|
|
#include <linux/module.h>
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
#include <linux/mm.h>
|
2006-05-24 08:18:44 +08:00
|
|
|
#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>
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
#include <linux/jiffies.h>
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
static DEFINE_MUTEX(dma_list_mutex);
|
|
|
|
static LIST_HEAD(dma_device_list);
|
|
|
|
static LIST_HEAD(dma_client_list);
|
|
|
|
|
|
|
|
/* --- sysfs implementation --- */
|
|
|
|
|
2007-09-25 08:03:03 +08:00
|
|
|
static ssize_t show_memcpy_count(struct device *dev, struct device_attribute *attr, char *buf)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
2007-09-25 08:03:03 +08:00
|
|
|
struct dma_chan *chan = to_dma_chan(dev);
|
2006-05-24 08:18:44 +08:00
|
|
|
unsigned long count = 0;
|
|
|
|
int i;
|
|
|
|
|
2006-05-26 04:26:53 +08:00
|
|
|
for_each_possible_cpu(i)
|
2006-05-24 08:18:44 +08:00
|
|
|
count += per_cpu_ptr(chan->local, i)->memcpy_count;
|
|
|
|
|
|
|
|
return sprintf(buf, "%lu\n", count);
|
|
|
|
}
|
|
|
|
|
2007-09-25 08:03:03 +08:00
|
|
|
static ssize_t show_bytes_transferred(struct device *dev, struct device_attribute *attr,
|
|
|
|
char *buf)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
2007-09-25 08:03:03 +08:00
|
|
|
struct dma_chan *chan = to_dma_chan(dev);
|
2006-05-24 08:18:44 +08:00
|
|
|
unsigned long count = 0;
|
|
|
|
int i;
|
|
|
|
|
2006-05-26 04:26:53 +08:00
|
|
|
for_each_possible_cpu(i)
|
2006-05-24 08:18:44 +08:00
|
|
|
count += per_cpu_ptr(chan->local, i)->bytes_transferred;
|
|
|
|
|
|
|
|
return sprintf(buf, "%lu\n", count);
|
|
|
|
}
|
|
|
|
|
2007-09-25 08:03:03 +08:00
|
|
|
static ssize_t show_in_use(struct device *dev, struct device_attribute *attr, char *buf)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
2007-09-25 08:03:03 +08:00
|
|
|
struct dma_chan *chan = to_dma_chan(dev);
|
2007-07-10 02:56:42 +08:00
|
|
|
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();
|
|
|
|
}
|
2006-05-24 08:18:44 +08:00
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
return sprintf(buf, "%d\n", in_use);
|
2006-05-24 08:18:44 +08:00
|
|
|
}
|
|
|
|
|
2007-09-25 08:03:03 +08:00
|
|
|
static struct device_attribute dma_attrs[] = {
|
2006-05-24 08:18:44 +08:00
|
|
|
__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);
|
|
|
|
|
2007-09-25 08:03:03 +08:00
|
|
|
static void dma_dev_release(struct device *dev)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
2007-09-25 08:03:03 +08:00
|
|
|
struct dma_chan *chan = to_dma_chan(dev);
|
2006-05-24 08:18:44 +08:00
|
|
|
kref_put(&chan->device->refcount, dma_async_device_cleanup);
|
|
|
|
}
|
|
|
|
|
|
|
|
static struct class dma_devclass = {
|
2007-09-25 08:03:03 +08:00
|
|
|
.name = "dma",
|
|
|
|
.dev_attrs = dma_attrs,
|
|
|
|
.dev_release = dma_dev_release,
|
2006-05-24 08:18:44 +08:00
|
|
|
};
|
|
|
|
|
|
|
|
/* --- client and device registration --- */
|
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
#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->cap_mask.bits,
|
|
|
|
DMA_TX_TYPE_END);
|
|
|
|
return bitmap_equal(want->bits, has.bits, DMA_TX_TYPE_END);
|
|
|
|
}
|
|
|
|
|
2006-05-24 08:18:44 +08:00
|
|
|
/**
|
2007-07-10 02:56:42 +08:00
|
|
|
* dma_client_chan_alloc - try to allocate channels to a client
|
2006-05-24 08:18:44 +08:00
|
|
|
* @client: &dma_client
|
|
|
|
*
|
|
|
|
* Called with dma_list_mutex held.
|
|
|
|
*/
|
2007-07-10 02:56:42 +08:00
|
|
|
static void dma_client_chan_alloc(struct dma_client *client)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
|
|
|
struct dma_device *device;
|
|
|
|
struct dma_chan *chan;
|
|
|
|
int desc; /* allocated descriptor count */
|
2007-07-10 02:56:42 +08:00
|
|
|
enum dma_state_client ack;
|
2006-05-24 08:18:44 +08:00
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
/* Find a channel */
|
|
|
|
list_for_each_entry(device, &dma_device_list, global_node)
|
2006-05-24 08:18:44 +08:00
|
|
|
list_for_each_entry(chan, &device->channels, device_node) {
|
2007-07-10 02:56:42 +08:00
|
|
|
if (!dma_chan_satisfies_mask(chan, client->cap_mask))
|
2006-05-24 08:18:44 +08:00
|
|
|
continue;
|
|
|
|
|
|
|
|
desc = chan->device->device_alloc_chan_resources(chan);
|
|
|
|
if (desc >= 0) {
|
2007-07-10 02:56:42 +08:00
|
|
|
ack = client->event_callback(client,
|
|
|
|
chan,
|
|
|
|
DMA_RESOURCE_AVAILABLE);
|
|
|
|
|
|
|
|
/* we are done once this client rejects
|
|
|
|
* an available resource
|
|
|
|
*/
|
2007-11-15 08:59:27 +08:00
|
|
|
if (ack == DMA_ACK)
|
2007-07-10 02:56:42 +08:00
|
|
|
dma_chan_get(chan);
|
2007-11-15 08:59:27 +08:00
|
|
|
else if (ack == DMA_NAK)
|
2007-07-10 02:56:42 +08:00
|
|
|
return;
|
2006-05-24 08:18:44 +08:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
enum dma_status dma_sync_wait(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);
|
|
|
|
|
2006-05-24 08:18:44 +08:00
|
|
|
/**
|
2006-07-04 10:45:31 +08:00
|
|
|
* dma_chan_cleanup - release a DMA channel's resources
|
|
|
|
* @kref: kernel reference structure that contains the DMA channel device
|
2006-05-24 08:18:44 +08:00
|
|
|
*/
|
|
|
|
void dma_chan_cleanup(struct kref *kref)
|
|
|
|
{
|
|
|
|
struct dma_chan *chan = container_of(kref, struct dma_chan, refcount);
|
|
|
|
chan->device->device_free_chan_resources(chan);
|
|
|
|
kref_put(&chan->device->refcount, dma_async_device_cleanup);
|
|
|
|
}
|
2007-03-17 05:38:05 +08:00
|
|
|
EXPORT_SYMBOL(dma_chan_cleanup);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
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;
|
2006-05-26 04:26:53 +08:00
|
|
|
for_each_possible_cpu(i)
|
2006-05-24 08:18:44 +08:00
|
|
|
bias -= local_read(&per_cpu_ptr(chan->local, i)->refcount);
|
|
|
|
atomic_sub(bias, &chan->refcount.refcount);
|
|
|
|
kref_put(&chan->refcount, dma_chan_cleanup);
|
|
|
|
}
|
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
static void dma_chan_release(struct dma_chan *chan)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
|
|
|
atomic_add(0x7FFFFFFF, &chan->refcount.refcount);
|
|
|
|
chan->slow_ref = 1;
|
|
|
|
call_rcu(&chan->rcu, dma_chan_free_rcu);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2007-07-10 02:56:42 +08:00
|
|
|
* dma_chans_notify_available - broadcast available channels to the clients
|
2006-05-24 08:18:44 +08:00
|
|
|
*/
|
2007-07-10 02:56:42 +08:00
|
|
|
static void dma_clients_notify_available(void)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
|
|
|
struct dma_client *client;
|
|
|
|
|
|
|
|
mutex_lock(&dma_list_mutex);
|
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
list_for_each_entry(client, &dma_client_list, global_node)
|
|
|
|
dma_client_chan_alloc(client);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
mutex_unlock(&dma_list_mutex);
|
|
|
|
}
|
|
|
|
|
|
|
|
/**
|
2007-07-10 02:56:42 +08:00
|
|
|
* dma_chans_notify_available - tell the clients that a channel is going away
|
|
|
|
* @chan: channel on its way out
|
2006-05-24 08:18:44 +08:00
|
|
|
*/
|
2007-07-10 02:56:42 +08:00
|
|
|
static void dma_clients_notify_removed(struct dma_chan *chan)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
|
|
|
struct dma_client *client;
|
2007-07-10 02:56:42 +08:00
|
|
|
enum dma_state_client ack;
|
2006-05-24 08:18:44 +08:00
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
mutex_lock(&dma_list_mutex);
|
|
|
|
|
|
|
|
list_for_each_entry(client, &dma_client_list, global_node) {
|
|
|
|
ack = client->event_callback(client, chan,
|
|
|
|
DMA_RESOURCE_REMOVED);
|
|
|
|
|
|
|
|
/* client was holding resources for this channel so
|
|
|
|
* free it
|
|
|
|
*/
|
2007-11-15 08:59:27 +08:00
|
|
|
if (ack == DMA_ACK)
|
2007-07-10 02:56:42 +08:00
|
|
|
dma_chan_put(chan);
|
|
|
|
}
|
2006-05-24 08:18:44 +08:00
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
mutex_unlock(&dma_list_mutex);
|
|
|
|
}
|
2006-05-24 08:18:44 +08:00
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
/**
|
|
|
|
* dma_async_client_register - register a &dma_client
|
|
|
|
* @client: ptr to a client structure with valid 'event_callback' and 'cap_mask'
|
|
|
|
*/
|
|
|
|
void dma_async_client_register(struct dma_client *client)
|
|
|
|
{
|
2006-05-24 08:18:44 +08:00
|
|
|
mutex_lock(&dma_list_mutex);
|
|
|
|
list_add_tail(&client->global_node, &dma_client_list);
|
|
|
|
mutex_unlock(&dma_list_mutex);
|
|
|
|
}
|
2007-03-17 05:38:05 +08:00
|
|
|
EXPORT_SYMBOL(dma_async_client_register);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
/**
|
|
|
|
* dma_async_client_unregister - unregister a client and free the &dma_client
|
2006-07-04 10:45:31 +08:00
|
|
|
* @client: &dma_client to free
|
2006-05-24 08:18:44 +08:00
|
|
|
*
|
|
|
|
* Force frees any allocated DMA channels, frees the &dma_client memory
|
|
|
|
*/
|
|
|
|
void dma_async_client_unregister(struct dma_client *client)
|
|
|
|
{
|
2007-07-10 02:56:42 +08:00
|
|
|
struct dma_device *device;
|
2006-05-24 08:18:44 +08:00
|
|
|
struct dma_chan *chan;
|
2007-07-10 02:56:42 +08:00
|
|
|
enum dma_state_client ack;
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
if (!client)
|
|
|
|
return;
|
|
|
|
|
|
|
|
mutex_lock(&dma_list_mutex);
|
2007-07-10 02:56:42 +08:00
|
|
|
/* 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(client, chan,
|
|
|
|
DMA_RESOURCE_REMOVED);
|
|
|
|
|
2007-11-15 08:59:27 +08:00
|
|
|
if (ack == DMA_ACK)
|
2007-07-10 02:56:42 +08:00
|
|
|
dma_chan_put(chan);
|
|
|
|
}
|
|
|
|
|
2006-05-24 08:18:44 +08:00
|
|
|
list_del(&client->global_node);
|
|
|
|
mutex_unlock(&dma_list_mutex);
|
|
|
|
}
|
2007-03-17 05:38:05 +08:00
|
|
|
EXPORT_SYMBOL(dma_async_client_unregister);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
/**
|
2007-07-10 02:56:42 +08:00
|
|
|
* dma_async_client_chan_request - send all available channels to the
|
|
|
|
* client that satisfy the capability mask
|
|
|
|
* @client - requester
|
2006-05-24 08:18:44 +08:00
|
|
|
*/
|
2007-07-10 02:56:42 +08:00
|
|
|
void dma_async_client_chan_request(struct dma_client *client)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
2007-07-10 02:56:42 +08:00
|
|
|
mutex_lock(&dma_list_mutex);
|
|
|
|
dma_client_chan_alloc(client);
|
|
|
|
mutex_unlock(&dma_list_mutex);
|
2006-05-24 08:18:44 +08:00
|
|
|
}
|
2007-03-17 05:38:05 +08:00
|
|
|
EXPORT_SYMBOL(dma_async_client_chan_request);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
/**
|
2006-07-04 10:45:31 +08:00
|
|
|
* dma_async_device_register - registers DMA devices found
|
2006-05-24 08:18:44 +08:00
|
|
|
* @device: &dma_device
|
|
|
|
*/
|
|
|
|
int dma_async_device_register(struct dma_device *device)
|
|
|
|
{
|
|
|
|
static int id;
|
2007-03-09 01:57:34 +08:00
|
|
|
int chancnt = 0, rc;
|
2006-05-24 08:18:44 +08:00
|
|
|
struct dma_chan* chan;
|
|
|
|
|
|
|
|
if (!device)
|
|
|
|
return -ENODEV;
|
|
|
|
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
/* 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);
|
2008-03-14 08:45:28 +08:00
|
|
|
BUG_ON(dma_has_cap(DMA_INTERRUPT, device->cap_mask) &&
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
!device->device_prep_dma_interrupt);
|
|
|
|
|
|
|
|
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);
|
|
|
|
|
2006-05-24 08:18:44 +08:00
|
|
|
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++;
|
2007-09-25 08:03:03 +08:00
|
|
|
chan->dev.class = &dma_devclass;
|
|
|
|
chan->dev.parent = NULL;
|
|
|
|
snprintf(chan->dev.bus_id, BUS_ID_SIZE, "dma%dchan%d",
|
2006-05-24 08:18:44 +08:00
|
|
|
device->dev_id, chan->chan_id);
|
|
|
|
|
2007-09-25 08:03:03 +08:00
|
|
|
rc = device_register(&chan->dev);
|
2007-03-09 01:57:34 +08:00
|
|
|
if (rc) {
|
|
|
|
chancnt--;
|
|
|
|
free_percpu(chan->local);
|
|
|
|
chan->local = NULL;
|
|
|
|
goto err_out;
|
|
|
|
}
|
|
|
|
|
2007-11-15 08:59:27 +08:00
|
|
|
/* One for the channel, one of the class device */
|
|
|
|
kref_get(&device->refcount);
|
2006-05-24 08:18:44 +08:00
|
|
|
kref_get(&device->refcount);
|
2007-07-10 02:56:42 +08:00
|
|
|
kref_init(&chan->refcount);
|
|
|
|
chan->slow_ref = 0;
|
|
|
|
INIT_RCU_HEAD(&chan->rcu);
|
2006-05-24 08:18:44 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
mutex_lock(&dma_list_mutex);
|
|
|
|
list_add_tail(&device->global_node, &dma_device_list);
|
|
|
|
mutex_unlock(&dma_list_mutex);
|
|
|
|
|
2007-07-10 02:56:42 +08:00
|
|
|
dma_clients_notify_available();
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
return 0;
|
2007-03-09 01:57:34 +08:00
|
|
|
|
|
|
|
err_out:
|
|
|
|
list_for_each_entry(chan, &device->channels, device_node) {
|
|
|
|
if (chan->local == NULL)
|
|
|
|
continue;
|
|
|
|
kref_put(&device->refcount, dma_async_device_cleanup);
|
2007-09-25 08:03:03 +08:00
|
|
|
device_unregister(&chan->dev);
|
2007-03-09 01:57:34 +08:00
|
|
|
chancnt--;
|
|
|
|
free_percpu(chan->local);
|
|
|
|
}
|
|
|
|
return rc;
|
2006-05-24 08:18:44 +08:00
|
|
|
}
|
2007-03-17 05:38:05 +08:00
|
|
|
EXPORT_SYMBOL(dma_async_device_register);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
|
|
|
/**
|
2006-07-04 10:45:31 +08:00
|
|
|
* dma_async_device_cleanup - function called when all references are released
|
|
|
|
* @kref: kernel reference object
|
2006-05-24 08:18:44 +08:00
|
|
|
*/
|
|
|
|
static void dma_async_device_cleanup(struct kref *kref)
|
|
|
|
{
|
|
|
|
struct dma_device *device;
|
|
|
|
|
|
|
|
device = container_of(kref, struct dma_device, refcount);
|
|
|
|
complete(&device->done);
|
|
|
|
}
|
|
|
|
|
2006-07-04 10:45:31 +08:00
|
|
|
/**
|
|
|
|
* dma_async_device_unregister - unregisters DMA devices
|
|
|
|
* @device: &dma_device
|
|
|
|
*/
|
|
|
|
void dma_async_device_unregister(struct dma_device *device)
|
2006-05-24 08:18:44 +08:00
|
|
|
{
|
|
|
|
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) {
|
2007-07-10 02:56:42 +08:00
|
|
|
dma_clients_notify_removed(chan);
|
2007-09-25 08:03:03 +08:00
|
|
|
device_unregister(&chan->dev);
|
2007-07-10 02:56:42 +08:00
|
|
|
dma_chan_release(chan);
|
2006-05-24 08:18:44 +08:00
|
|
|
}
|
|
|
|
|
|
|
|
kref_put(&device->refcount, dma_async_device_cleanup);
|
|
|
|
wait_for_completion(&device->done);
|
|
|
|
}
|
2007-03-17 05:38:05 +08:00
|
|
|
EXPORT_SYMBOL(dma_async_device_unregister);
|
2006-05-24 08:18:44 +08:00
|
|
|
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
/**
|
|
|
|
* dma_async_memcpy_buf_to_buf - 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(struct dma_chan *chan, void *dest,
|
|
|
|
void *src, size_t len)
|
|
|
|
{
|
|
|
|
struct dma_device *dev = chan->device;
|
|
|
|
struct dma_async_tx_descriptor *tx;
|
2008-02-03 10:49:57 +08:00
|
|
|
dma_addr_t dma_dest, dma_src;
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
dma_cookie_t cookie;
|
|
|
|
int cpu;
|
|
|
|
|
2008-02-03 10:49:57 +08:00
|
|
|
dma_src = dma_map_single(dev->dev, src, len, DMA_TO_DEVICE);
|
|
|
|
dma_dest = dma_map_single(dev->dev, dest, len, DMA_FROM_DEVICE);
|
2008-04-18 11:17:26 +08:00
|
|
|
tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
|
|
|
|
DMA_CTRL_ACK);
|
2008-02-03 10:49:57 +08:00
|
|
|
|
|
|
|
if (!tx) {
|
|
|
|
dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
|
|
|
|
dma_unmap_single(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
return -ENOMEM;
|
2008-02-03 10:49:57 +08:00
|
|
|
}
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
|
|
|
|
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);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* dma_async_memcpy_buf_to_pg - 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(struct dma_chan *chan, struct page *page,
|
|
|
|
unsigned int offset, void *kdata, size_t len)
|
|
|
|
{
|
|
|
|
struct dma_device *dev = chan->device;
|
|
|
|
struct dma_async_tx_descriptor *tx;
|
2008-02-03 10:49:57 +08:00
|
|
|
dma_addr_t dma_dest, dma_src;
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
dma_cookie_t cookie;
|
|
|
|
int cpu;
|
|
|
|
|
2008-02-03 10:49:57 +08:00
|
|
|
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);
|
2008-04-18 11:17:26 +08:00
|
|
|
tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
|
|
|
|
DMA_CTRL_ACK);
|
2008-02-03 10:49:57 +08:00
|
|
|
|
|
|
|
if (!tx) {
|
|
|
|
dma_unmap_single(dev->dev, dma_src, len, DMA_TO_DEVICE);
|
|
|
|
dma_unmap_page(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
return -ENOMEM;
|
2008-02-03 10:49:57 +08:00
|
|
|
}
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
|
|
|
|
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);
|
|
|
|
|
|
|
|
/**
|
|
|
|
* dma_async_memcpy_pg_to_pg - 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(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 *dev = chan->device;
|
|
|
|
struct dma_async_tx_descriptor *tx;
|
2008-02-03 10:49:57 +08:00
|
|
|
dma_addr_t dma_dest, dma_src;
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
dma_cookie_t cookie;
|
|
|
|
int cpu;
|
|
|
|
|
2008-02-03 10:49:57 +08:00
|
|
|
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);
|
2008-04-18 11:17:26 +08:00
|
|
|
tx = dev->device_prep_dma_memcpy(chan, dma_dest, dma_src, len,
|
|
|
|
DMA_CTRL_ACK);
|
2008-02-03 10:49:57 +08:00
|
|
|
|
|
|
|
if (!tx) {
|
|
|
|
dma_unmap_page(dev->dev, dma_src, len, DMA_TO_DEVICE);
|
|
|
|
dma_unmap_page(dev->dev, dma_dest, len, DMA_FROM_DEVICE);
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
return -ENOMEM;
|
2008-02-03 10:49:57 +08:00
|
|
|
}
|
dmaengine: refactor dmaengine around dma_async_tx_descriptor
The current dmaengine interface defines mutliple routines per operation,
i.e. dma_async_memcpy_buf_to_buf, dma_async_memcpy_buf_to_page etc. Adding
more operation types (xor, crc, etc) to this model would result in an
unmanageable number of method permutations.
Are we really going to add a set of hooks for each DMA engine
whizbang feature?
- Jeff Garzik
The descriptor creation process is refactored using the new common
dma_async_tx_descriptor structure. Instead of per driver
do_<operation>_<dest>_to_<src> methods, drivers integrate
dma_async_tx_descriptor into their private software descriptor and then
define a 'prep' routine per operation. The prep routine allocates a
descriptor and ensures that the tx_set_src, tx_set_dest, tx_submit routines
are valid. Descriptor creation and submission becomes:
struct dma_device *dev;
struct dma_chan *chan;
struct dma_async_tx_descriptor *tx;
tx = dev->device_prep_dma_<operation>(chan, len, int_flag)
tx->tx_set_src(dma_addr_t, tx, index /* for multi-source ops */)
tx->tx_set_dest(dma_addr_t, tx, index)
tx->tx_submit(tx)
In addition to the refactoring, dma_async_tx_descriptor also lays the
groundwork for definining cross-channel-operation dependencies, and a
callback facility for asynchronous notification of operation completion.
Changelog:
* drop dma mapping methods, suggested by Chris Leech
* fix ioat_dma_dependency_added, also caught by Andrew Morton
* fix dma_sync_wait, change from Andrew Morton
* uninline large functions, change from Andrew Morton
* add tx->callback = NULL to dmaengine calls to interoperate with async_tx
calls
* hookup ioat_tx_submit
* convert channel capabilities to a 'cpumask_t like' bitmap
* removed DMA_TX_ARRAY_INIT, no longer needed
* checkpatch.pl fixes
* make set_src, set_dest, and tx_submit descriptor specific methods
* fixup git-ioat merge
* move group_list and phys to dma_async_tx_descriptor
Cc: Jeff Garzik <jeff@garzik.org>
Cc: Chris Leech <christopher.leech@intel.com>
Signed-off-by: Shannon Nelson <shannon.nelson@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
Acked-by: David S. Miller <davem@davemloft.net>
2007-01-03 02:10:43 +08:00
|
|
|
|
|
|
|
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);
|
|
|
|
|
|
|
|
void dma_async_tx_descriptor_init(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);
|
|
|
|
|
2006-05-24 08:18:44 +08:00
|
|
|
static int __init dma_bus_init(void)
|
|
|
|
{
|
|
|
|
mutex_init(&dma_list_mutex);
|
|
|
|
return class_register(&dma_devclass);
|
|
|
|
}
|
|
|
|
subsys_initcall(dma_bus_init);
|
|
|
|
|