linux-stable/include/linux/async_tx.h
Dan Williams b2f46fd8ef async_tx: add support for asynchronous GF multiplication
[ Based on an original patch by Yuri Tikhonov ]

This adds support for doing asynchronous GF multiplication by adding
two additional functions to the async_tx API:

 async_gen_syndrome() does simultaneous XOR and Galois field
    multiplication of sources.

 async_syndrome_val() validates the given source buffers against known P
    and Q values.

When a request is made to run async_pq against more than the hardware
maximum number of supported sources we need to reuse the previous
generated P and Q values as sources into the next operation.  Care must
be taken to remove Q from P' and P from Q'.  For example to perform a 5
source pq op with hardware that only supports 4 sources at a time the
following approach is taken:

p, q = PQ(src0, src1, src2, src3, COEF({01}, {02}, {04}, {08}))
p', q' = PQ(p, q, q, src4, COEF({00}, {01}, {00}, {10}))

p' = p + q + q + src4 = p + src4
q' = {00}*p + {01}*q + {00}*q + {10}*src4 = q + {10}*src4

Note: 4 is the minimum acceptable maxpq otherwise we punt to
synchronous-software path.

The DMA_PREP_CONTINUE flag indicates to the driver to reuse p and q as
sources (in the above manner) and fill the remaining slots up to maxpq
with the new sources/coefficients.

Note1: Some devices have native support for P+Q continuation and can skip
this extra work.  Devices with this capability can advertise it with
dma_set_maxpq.  It is up to each driver how to handle the
DMA_PREP_CONTINUE flag.

Note2: The api supports disabling the generation of P when generating Q,
this is ignored by the synchronous path but is implemented by some dma
devices to save unnecessary writes.  In this case the continuation
algorithm is simplified to only reuse Q as a source.

Cc: H. Peter Anvin <hpa@zytor.com>
Cc: David Woodhouse <David.Woodhouse@intel.com>
Signed-off-by: Yuri Tikhonov <yur@emcraft.com>
Signed-off-by: Ilya Yanok <yanok@emcraft.com>
Reviewed-by: Andre Noll <maan@systemlinux.org>
Acked-by: Maciej Sosnowski <maciej.sosnowski@intel.com>
Signed-off-by: Dan Williams <dan.j.williams@intel.com>
2009-08-29 19:09:27 -07:00

198 lines
6.3 KiB
C

/*
* Copyright © 2006, Intel Corporation.
*
* This program is free software; you can redistribute it and/or modify it
* under the terms and conditions of the GNU General Public License,
* version 2, as published by the Free Software Foundation.
*
* This program is distributed in the hope 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.,
* 51 Franklin St - Fifth Floor, Boston, MA 02110-1301 USA.
*
*/
#ifndef _ASYNC_TX_H_
#define _ASYNC_TX_H_
#include <linux/dmaengine.h>
#include <linux/spinlock.h>
#include <linux/interrupt.h>
/* on architectures without dma-mapping capabilities we need to ensure
* that the asynchronous path compiles away
*/
#ifdef CONFIG_HAS_DMA
#define __async_inline
#else
#define __async_inline __always_inline
#endif
/**
* dma_chan_ref - object used to manage dma channels received from the
* dmaengine core.
* @chan - the channel being tracked
* @node - node for the channel to be placed on async_tx_master_list
* @rcu - for list_del_rcu
* @count - number of times this channel is listed in the pool
* (for channels with multiple capabiities)
*/
struct dma_chan_ref {
struct dma_chan *chan;
struct list_head node;
struct rcu_head rcu;
atomic_t count;
};
/**
* async_tx_flags - modifiers for the async_* calls
* @ASYNC_TX_XOR_ZERO_DST: this flag must be used for xor operations where the
* the destination address is not a source. The asynchronous case handles this
* implicitly, the synchronous case needs to zero the destination block.
* @ASYNC_TX_XOR_DROP_DST: this flag must be used if the destination address is
* also one of the source addresses. In the synchronous case the destination
* address is an implied source, whereas the asynchronous case it must be listed
* as a source. The destination address must be the first address in the source
* array.
* @ASYNC_TX_ACK: immediately ack the descriptor, precludes setting up a
* dependency chain
*/
enum async_tx_flags {
ASYNC_TX_XOR_ZERO_DST = (1 << 0),
ASYNC_TX_XOR_DROP_DST = (1 << 1),
ASYNC_TX_ACK = (1 << 2),
};
/**
* struct async_submit_ctl - async_tx submission/completion modifiers
* @flags: submission modifiers
* @depend_tx: parent dependency of the current operation being submitted
* @cb_fn: callback routine to run at operation completion
* @cb_param: parameter for the callback routine
* @scribble: caller provided space for dma/page address conversions
*/
struct async_submit_ctl {
enum async_tx_flags flags;
struct dma_async_tx_descriptor *depend_tx;
dma_async_tx_callback cb_fn;
void *cb_param;
void *scribble;
};
#ifdef CONFIG_DMA_ENGINE
#define async_tx_issue_pending_all dma_issue_pending_all
/**
* async_tx_issue_pending - send pending descriptor to the hardware channel
* @tx: descriptor handle to retrieve hardware context
*
* Note: any dependent operations will have already been issued by
* async_tx_channel_switch, or (in the case of no channel switch) will
* be already pending on this channel.
*/
static inline void async_tx_issue_pending(struct dma_async_tx_descriptor *tx)
{
if (likely(tx)) {
struct dma_chan *chan = tx->chan;
struct dma_device *dma = chan->device;
dma->device_issue_pending(chan);
}
}
#ifdef CONFIG_ARCH_HAS_ASYNC_TX_FIND_CHANNEL
#include <asm/async_tx.h>
#else
#define async_tx_find_channel(dep, type, dst, dst_count, src, src_count, len) \
__async_tx_find_channel(dep, type)
struct dma_chan *
__async_tx_find_channel(struct async_submit_ctl *submit,
enum dma_transaction_type tx_type);
#endif /* CONFIG_ARCH_HAS_ASYNC_TX_FIND_CHANNEL */
#else
static inline void async_tx_issue_pending_all(void)
{
do { } while (0);
}
static inline void async_tx_issue_pending(struct dma_async_tx_descriptor *tx)
{
do { } while (0);
}
static inline struct dma_chan *
async_tx_find_channel(struct async_submit_ctl *submit,
enum dma_transaction_type tx_type, struct page **dst,
int dst_count, struct page **src, int src_count,
size_t len)
{
return NULL;
}
#endif
/**
* async_tx_sync_epilog - actions to take if an operation is run synchronously
* @cb_fn: function to call when the transaction completes
* @cb_fn_param: parameter to pass to the callback routine
*/
static inline void
async_tx_sync_epilog(struct async_submit_ctl *submit)
{
if (submit->cb_fn)
submit->cb_fn(submit->cb_param);
}
typedef union {
unsigned long addr;
struct page *page;
dma_addr_t dma;
} addr_conv_t;
static inline void
init_async_submit(struct async_submit_ctl *args, enum async_tx_flags flags,
struct dma_async_tx_descriptor *tx,
dma_async_tx_callback cb_fn, void *cb_param,
addr_conv_t *scribble)
{
args->flags = flags;
args->depend_tx = tx;
args->cb_fn = cb_fn;
args->cb_param = cb_param;
args->scribble = scribble;
}
void async_tx_submit(struct dma_chan *chan, struct dma_async_tx_descriptor *tx,
struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *
async_xor(struct page *dest, struct page **src_list, unsigned int offset,
int src_cnt, size_t len, struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *
async_xor_val(struct page *dest, struct page **src_list, unsigned int offset,
int src_cnt, size_t len, enum sum_check_flags *result,
struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *
async_memcpy(struct page *dest, struct page *src, unsigned int dest_offset,
unsigned int src_offset, size_t len,
struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *
async_memset(struct page *dest, int val, unsigned int offset,
size_t len, struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *async_trigger_callback(struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *
async_gen_syndrome(struct page **blocks, unsigned int offset, int src_cnt,
size_t len, struct async_submit_ctl *submit);
struct dma_async_tx_descriptor *
async_syndrome_val(struct page **blocks, unsigned int offset, int src_cnt,
size_t len, enum sum_check_flags *pqres, struct page *spare,
struct async_submit_ctl *submit);
void async_tx_quiesce(struct dma_async_tx_descriptor **tx);
#endif /* _ASYNC_TX_H_ */