[linux-2.6-block.git] / crypto / async_tx / async_pq.c

/*
 * Copyright(c) 2007 Yuri Tikhonov <yur@emcraft.com>
 * Copyright(c) 2009 Intel Corporation
 *
 * 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.
 */
#include <linux/kernel.h>
#include <linux/interrupt.h>
#include <linux/module.h>
#include <linux/dma-mapping.h>
#include <linux/raid/pq.h>
#include <linux/async_tx.h>
#include <linux/gfp.h>

/**
 * pq_scribble_page - space to hold throwaway P or Q buffer for
 * synchronous gen_syndrome
 */
static struct page *pq_scribble_page;

/* the struct page *blocks[] parameter passed to async_gen_syndrome()
 * and async_syndrome_val() contains the 'P' destination address at
 * blocks[disks-2] and the 'Q' destination address at blocks[disks-1]
 *
 * note: these are macros as they are used as lvalues
 */
#define P(b, d) (b[d-2])
#define Q(b, d) (b[d-1])

/**
 * do_async_gen_syndrome - asynchronously calculate P and/or Q
 */
static __async_inline struct dma_async_tx_descriptor *
do_async_gen_syndrome(struct dma_chan *chan,
		      const unsigned char *scfs, int disks,
		      struct dmaengine_unmap_data *unmap,
		      enum dma_ctrl_flags dma_flags,
		      struct async_submit_ctl *submit)
{
	struct dma_async_tx_descriptor *tx = NULL;
	struct dma_device *dma = chan->device;
	enum async_tx_flags flags_orig = submit->flags;
	dma_async_tx_callback cb_fn_orig = submit->cb_fn;
	dma_async_tx_callback cb_param_orig = submit->cb_param;
	int src_cnt = disks - 2;
	unsigned short pq_src_cnt;
	dma_addr_t dma_dest[2];
	int src_off = 0;

	if (submit->flags & ASYNC_TX_FENCE)
		dma_flags |= DMA_PREP_FENCE;

	while (src_cnt > 0) {
		submit->flags = flags_orig;
		pq_src_cnt = min(src_cnt, dma_maxpq(dma, dma_flags));
		/* if we are submitting additional pqs, leave the chain open,
		 * clear the callback parameters, and leave the destination
		 * buffers mapped
		 */
		if (src_cnt > pq_src_cnt) {
			submit->flags &= ~ASYNC_TX_ACK;
			submit->flags |= ASYNC_TX_FENCE;
			submit->cb_fn = NULL;
			submit->cb_param = NULL;
		} else {
			submit->cb_fn = cb_fn_orig;
			submit->cb_param = cb_param_orig;
			if (cb_fn_orig)
				dma_flags |= DMA_PREP_INTERRUPT;
		}

		/* Drivers force forward progress in case they can not provide
		 * a descriptor
		 */
		for (;;) {
			dma_dest[0] = unmap->addr[disks - 2];
			dma_dest[1] = unmap->addr[disks - 1];
			tx = dma->device_prep_dma_pq(chan, dma_dest,
						     &unmap->addr[src_off],
						     pq_src_cnt,
						     &scfs[src_off], unmap->len,
						     dma_flags);
			if (likely(tx))
				break;
			async_tx_quiesce(&submit->depend_tx);
			dma_async_issue_pending(chan);
		}

		dma_set_unmap(tx, unmap);
		async_tx_submit(chan, tx, submit);
		submit->depend_tx = tx;

		/* drop completed sources */
		src_cnt -= pq_src_cnt;
		src_off += pq_src_cnt;

		dma_flags |= DMA_PREP_CONTINUE;
	}

	return tx;
}

/**
 * do_sync_gen_syndrome - synchronously calculate a raid6 syndrome
 */
static void
do_sync_gen_syndrome(struct page **blocks, unsigned int offset, int disks,
		     size_t len, struct async_submit_ctl *submit)
{
	void **srcs;
	int i;
	int start = -1, stop = disks - 3;

	if (submit->scribble)
		srcs = submit->scribble;
	else
		srcs = (void **) blocks;

	for (i = 0; i < disks; i++) {
		if (blocks[i] == NULL) {
			BUG_ON(i > disks - 3); /* P or Q can't be zero */
			srcs[i] = (void*)raid6_empty_zero_page;
		} else {
			srcs[i] = page_address(blocks[i]) + offset;
			if (i < disks - 2) {
				stop = i;
				if (start == -1)
					start = i;
			}
		}
	}
	if (submit->flags & ASYNC_TX_PQ_XOR_DST) {
		BUG_ON(!raid6_call.xor_syndrome);
		if (start >= 0)
			raid6_call.xor_syndrome(disks, start, stop, len, srcs);
	} else
		raid6_call.gen_syndrome(disks, len, srcs);
	async_tx_sync_epilog(submit);
}

/**
 * async_gen_syndrome - asynchronously calculate a raid6 syndrome
 * @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
 * @offset: common offset into each block (src and dest) to start transaction
 * @disks: number of blocks (including missing P or Q, see below)
 * @len: length of operation in bytes
 * @submit: submission/completion modifiers
 *
 * General note: This routine assumes a field of GF(2^8) with a
 * primitive polynomial of 0x11d and a generator of {02}.
 *
 * 'disks' note: callers can optionally omit either P or Q (but not
 * both) from the calculation by setting blocks[disks-2] or
 * blocks[disks-1] to NULL.  When P or Q is omitted 'len' must be <=
 * PAGE_SIZE as a temporary buffer of this size is used in the
 * synchronous path.  'disks' always accounts for both destination
 * buffers.  If any source buffers (blocks[i] where i < disks - 2) are
 * set to NULL those buffers will be replaced with the raid6_zero_page
 * in the synchronous path and omitted in the hardware-asynchronous
 * path.
 */
struct dma_async_tx_descriptor *
async_gen_syndrome(struct page **blocks, unsigned int offset, int disks,
		   size_t len, struct async_submit_ctl *submit)
{
	int src_cnt = disks - 2;
	struct dma_chan *chan = async_tx_find_channel(submit, DMA_PQ,
						      &P(blocks, disks), 2,
						      blocks, src_cnt, len);
	struct dma_device *device = chan ? chan->device : NULL;
	struct dmaengine_unmap_data *unmap = NULL;

	BUG_ON(disks > 255 || !(P(blocks, disks) || Q(blocks, disks)));

	if (device)
		unmap = dmaengine_get_unmap_data(device->dev, disks, GFP_NOWAIT);

	/* XORing P/Q is only implemented in software */
	if (unmap && !(submit->flags & ASYNC_TX_PQ_XOR_DST) &&
	    (src_cnt <= dma_maxpq(device, 0) ||
	     dma_maxpq(device, DMA_PREP_CONTINUE) > 0) &&
	    is_dma_pq_aligned(device, offset, 0, len)) {
		struct dma_async_tx_descriptor *tx;
		enum dma_ctrl_flags dma_flags = 0;
		unsigned char coefs[src_cnt];
		int i, j;

		/* run the p+q asynchronously */
		pr_debug("%s: (async) disks: %d len: %zu\n",
			 __func__, disks, len);

		/* convert source addresses being careful to collapse 'empty'
		 * sources and update the coefficients accordingly
		 */
		unmap->len = len;
		for (i = 0, j = 0; i < src_cnt; i++) {
			if (blocks[i] == NULL)
				continue;
			unmap->addr[j] = dma_map_page(device->dev, blocks[i], offset,
						      len, DMA_TO_DEVICE);
			coefs[j] = raid6_gfexp[i];
			unmap->to_cnt++;
			j++;
		}

		/*
		 * DMAs use destinations as sources,
		 * so use BIDIRECTIONAL mapping
		 */
		unmap->bidi_cnt++;
		if (P(blocks, disks))
			unmap->addr[j++] = dma_map_page(device->dev, P(blocks, disks),
							offset, len, DMA_BIDIRECTIONAL);
		else {
			unmap->addr[j++] = 0;
			dma_flags |= DMA_PREP_PQ_DISABLE_P;
		}

		unmap->bidi_cnt++;
		if (Q(blocks, disks))
			unmap->addr[j++] = dma_map_page(device->dev, Q(blocks, disks),
						       offset, len, DMA_BIDIRECTIONAL);
		else {
			unmap->addr[j++] = 0;
			dma_flags |= DMA_PREP_PQ_DISABLE_Q;
		}

		tx = do_async_gen_syndrome(chan, coefs, j, unmap, dma_flags, submit);
		dmaengine_unmap_put(unmap);
		return tx;
	}

	dmaengine_unmap_put(unmap);

	/* run the pq synchronously */
	pr_debug("%s: (sync) disks: %d len: %zu\n", __func__, disks, len);

	/* wait for any prerequisite operations */
	async_tx_quiesce(&submit->depend_tx);

	if (!P(blocks, disks)) {
		P(blocks, disks) = pq_scribble_page;
		BUG_ON(len + offset > PAGE_SIZE);
	}
	if (!Q(blocks, disks)) {
		Q(blocks, disks) = pq_scribble_page;
		BUG_ON(len + offset > PAGE_SIZE);
	}
	do_sync_gen_syndrome(blocks, offset, disks, len, submit);

	return NULL;
}
EXPORT_SYMBOL_GPL(async_gen_syndrome);

static inline struct dma_chan *
pq_val_chan(struct async_submit_ctl *submit, struct page **blocks, int disks, size_t len)
{
	#ifdef CONFIG_ASYNC_TX_DISABLE_PQ_VAL_DMA
	return NULL;
	#endif
	return async_tx_find_channel(submit, DMA_PQ_VAL, NULL, 0,  blocks,
				     disks, len);
}

/**
 * async_syndrome_val - asynchronously validate a raid6 syndrome
 * @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
 * @offset: common offset into each block (src and dest) to start transaction
 * @disks: number of blocks (including missing P or Q, see below)
 * @len: length of operation in bytes
 * @pqres: on val failure SUM_CHECK_P_RESULT and/or SUM_CHECK_Q_RESULT are set
 * @spare: temporary result buffer for the synchronous case
 * @submit: submission / completion modifiers
 *
 * The same notes from async_gen_syndrome apply to the 'blocks',
 * and 'disks' parameters of this routine.  The synchronous path
 * requires a temporary result buffer and submit->scribble to be
 * specified.
 */
struct dma_async_tx_descriptor *
async_syndrome_val(struct page **blocks, unsigned int offset, int disks,
		   size_t len, enum sum_check_flags *pqres, struct page *spare,
		   struct async_submit_ctl *submit)
{
	struct dma_chan *chan = pq_val_chan(submit, blocks, disks, len);
	struct dma_device *device = chan ? chan->device : NULL;
	struct dma_async_tx_descriptor *tx;
	unsigned char coefs[disks-2];
	enum dma_ctrl_flags dma_flags = submit->cb_fn ? DMA_PREP_INTERRUPT : 0;
	struct dmaengine_unmap_data *unmap = NULL;

	BUG_ON(disks < 4);

	if (device)
		unmap = dmaengine_get_unmap_data(device->dev, disks, GFP_NOWAIT);

	if (unmap && disks <= dma_maxpq(device, 0) &&
	    is_dma_pq_aligned(device, offset, 0, len)) {
		struct device *dev = device->dev;
		dma_addr_t pq[2];
		int i, j = 0, src_cnt = 0;

		pr_debug("%s: (async) disks: %d len: %zu\n",
			 __func__, disks, len);

		unmap->len = len;
		for (i = 0; i < disks-2; i++)
			if (likely(blocks[i])) {
				unmap->addr[j] = dma_map_page(dev, blocks[i],
							      offset, len,
							      DMA_TO_DEVICE);
				coefs[j] = raid6_gfexp[i];
				unmap->to_cnt++;
				src_cnt++;
				j++;
			}

		if (!P(blocks, disks)) {
			pq[0] = 0;
			dma_flags |= DMA_PREP_PQ_DISABLE_P;
		} else {
			pq[0] = dma_map_page(dev, P(blocks, disks),
					     offset, len,
					     DMA_TO_DEVICE);
			unmap->addr[j++] = pq[0];
			unmap->to_cnt++;
		}
		if (!Q(blocks, disks)) {
			pq[1] = 0;
			dma_flags |= DMA_PREP_PQ_DISABLE_Q;
		} else {
			pq[1] = dma_map_page(dev, Q(blocks, disks),
					     offset, len,
					     DMA_TO_DEVICE);
			unmap->addr[j++] = pq[1];
			unmap->to_cnt++;
		}

		if (submit->flags & ASYNC_TX_FENCE)
			dma_flags |= DMA_PREP_FENCE;
		for (;;) {
			tx = device->device_prep_dma_pq_val(chan, pq,
							    unmap->addr,
							    src_cnt,
							    coefs,
							    len, pqres,
							    dma_flags);
			if (likely(tx))
				break;
			async_tx_quiesce(&submit->depend_tx);
			dma_async_issue_pending(chan);
		}

		dma_set_unmap(tx, unmap);
		async_tx_submit(chan, tx, submit);
	} else {
		struct page *p_src = P(blocks, disks);
		struct page *q_src = Q(blocks, disks);
		enum async_tx_flags flags_orig = submit->flags;
		dma_async_tx_callback cb_fn_orig = submit->cb_fn;
		void *scribble = submit->scribble;
		void *cb_param_orig = submit->cb_param;
		void *p, *q, *s;

		pr_debug("%s: (sync) disks: %d len: %zu\n",
			 __func__, disks, len);

		/* caller must provide a temporary result buffer and
		 * allow the input parameters to be preserved
		 */
		BUG_ON(!spare || !scribble);

		/* wait for any prerequisite operations */
		async_tx_quiesce(&submit->depend_tx);

		/* recompute p and/or q into the temporary buffer and then
		 * check to see the result matches the current value
		 */
		tx = NULL;
		*pqres = 0;
		if (p_src) {
			init_async_submit(submit, ASYNC_TX_XOR_ZERO_DST, NULL,
					  NULL, NULL, scribble);
			tx = async_xor(spare, blocks, offset, disks-2, len, submit);
			async_tx_quiesce(&tx);
			p = page_address(p_src) + offset;
			s = page_address(spare) + offset;
			*pqres |= !!memcmp(p, s, len) << SUM_CHECK_P;
		}

		if (q_src) {
			P(blocks, disks) = NULL;
			Q(blocks, disks) = spare;
			init_async_submit(submit, 0, NULL, NULL, NULL, scribble);
			tx = async_gen_syndrome(blocks, offset, disks, len, submit);
			async_tx_quiesce(&tx);
			q = page_address(q_src) + offset;
			s = page_address(spare) + offset;
			*pqres |= !!memcmp(q, s, len) << SUM_CHECK_Q;
		}

		/* restore P, Q and submit */
		P(blocks, disks) = p_src;
		Q(blocks, disks) = q_src;

		submit->cb_fn = cb_fn_orig;
		submit->cb_param = cb_param_orig;
		submit->flags = flags_orig;
		async_tx_sync_epilog(submit);
		tx = NULL;
	}
	dmaengine_unmap_put(unmap);

	return tx;
}
EXPORT_SYMBOL_GPL(async_syndrome_val);

static int __init async_pq_init(void)
{
	pq_scribble_page = alloc_page(GFP_KERNEL);

	if (pq_scribble_page)
		return 0;

	pr_err("%s: failed to allocate required spare page\n", __func__);

	return -ENOMEM;
}

static void __exit async_pq_exit(void)
{
	__free_page(pq_scribble_page);
}

module_init(async_pq_init);
module_exit(async_pq_exit);

MODULE_DESCRIPTION("asynchronous raid6 syndrome generation/validation");
MODULE_LICENSE("GPL");
Commit	Line	Data
b2f46fd8 DW	1	/*
	2	* Copyright(c) 2007 Yuri Tikhonov <yur@emcraft.com>
	3	* Copyright(c) 2009 Intel Corporation
	4	*
	5	* This program is free software; you can redistribute it and/or modify it
	6	* under the terms of the GNU General Public License as published by the Free
	7	* Software Foundation; either version 2 of the License, or (at your option)
	8	* any later version.
	9	*
	10	* This program is distributed in the hope that it will be useful, but WITHOUT
	11	* ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
	12	* FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
	13	* more details.
	14	*
	15	* You should have received a copy of the GNU General Public License along with
	16	* this program; if not, write to the Free Software Foundation, Inc., 59
	17	* Temple Place - Suite 330, Boston, MA 02111-1307, USA.
	18	*
	19	* The full GNU General Public License is included in this distribution in the
	20	* file called COPYING.
	21	*/
	22	#include <linux/kernel.h>
	23	#include <linux/interrupt.h>
4bb33cc8	24	#include <linux/module.h>
b2f46fd8 DW	25	#include <linux/dma-mapping.h>
	26	#include <linux/raid/pq.h>
	27	#include <linux/async_tx.h>
5a0e3ad6	28	#include <linux/gfp.h>
b2f46fd8 DW	29
b2f46fd8 DW	30	/**
030b0772 DW	31	* pq_scribble_page - space to hold throwaway P or Q buffer for
030b0772 DW	32	* synchronous gen_syndrome
b2f46fd8	33	*/
030b0772	34	static struct page *pq_scribble_page;
b2f46fd8	35
b2f46fd8 DW	36	/* the struct page *blocks[] parameter passed to async_gen_syndrome()
	37	* and async_syndrome_val() contains the 'P' destination address at
	38	* blocks[disks-2] and the 'Q' destination address at blocks[disks-1]
	39	*
	40	* note: these are macros as they are used as lvalues
	41	*/
	42	#define P(b, d) (b[d-2])
	43	#define Q(b, d) (b[d-1])
	44
	45	/**
	46	* do_async_gen_syndrome - asynchronously calculate P and/or Q
	47	*/
	48	static __async_inline struct dma_async_tx_descriptor *
7476bd79 DW	49	do_async_gen_syndrome(struct dma_chan *chan,
	50	const unsigned char *scfs, int disks,
	51	struct dmaengine_unmap_data *unmap,
	52	enum dma_ctrl_flags dma_flags,
b2f46fd8 DW	53	struct async_submit_ctl *submit)
	54	{
	55	struct dma_async_tx_descriptor *tx = NULL;
	56	struct dma_device *dma = chan->device;
b2f46fd8 DW	57	enum async_tx_flags flags_orig = submit->flags;
	58	dma_async_tx_callback cb_fn_orig = submit->cb_fn;
	59	dma_async_tx_callback cb_param_orig = submit->cb_param;
	60	int src_cnt = disks - 2;
b2f46fd8 DW	61	unsigned short pq_src_cnt;
	62	dma_addr_t dma_dest[2];
	63	int src_off = 0;
b2f46fd8	64
7476bd79 DW	65	if (submit->flags & ASYNC_TX_FENCE)
7476bd79 DW	66	dma_flags \|= DMA_PREP_FENCE;
b2f46fd8 DW	67
	68	while (src_cnt > 0) {
	69	submit->flags = flags_orig;
	70	pq_src_cnt = min(src_cnt, dma_maxpq(dma, dma_flags));
	71	/* if we are submitting additional pqs, leave the chain open,
	72	* clear the callback parameters, and leave the destination
	73	* buffers mapped
	74	*/
	75	if (src_cnt > pq_src_cnt) {
	76	submit->flags &= ~ASYNC_TX_ACK;
0403e382	77	submit->flags \|= ASYNC_TX_FENCE;
b2f46fd8 DW	78	submit->cb_fn = NULL;
	79	submit->cb_param = NULL;
	80	} else {
b2f46fd8 DW	81	submit->cb_fn = cb_fn_orig;
	82	submit->cb_param = cb_param_orig;
	83	if (cb_fn_orig)
	84	dma_flags \|= DMA_PREP_INTERRUPT;
	85	}
	86
7476bd79 DW	87	/* Drivers force forward progress in case they can not provide
7476bd79 DW	88	* a descriptor
b2f46fd8 DW	89	*/
b2f46fd8 DW	90	for (;;) {
7476bd79 DW	91	dma_dest[0] = unmap->addr[disks - 2];
7476bd79 DW	92	dma_dest[1] = unmap->addr[disks - 1];
b2f46fd8	93	tx = dma->device_prep_dma_pq(chan, dma_dest,
7476bd79	94	&unmap->addr[src_off],
b2f46fd8	95	pq_src_cnt,
7476bd79	96	&scfs[src_off], unmap->len,
b2f46fd8 DW	97	dma_flags);
	98	if (likely(tx))
	99	break;
	100	async_tx_quiesce(&submit->depend_tx);
	101	dma_async_issue_pending(chan);
	102	}
	103
7476bd79	104	dma_set_unmap(tx, unmap);
b2f46fd8 DW	105	async_tx_submit(chan, tx, submit);
	106	submit->depend_tx = tx;
	107
	108	/* drop completed sources */
	109	src_cnt -= pq_src_cnt;
	110	src_off += pq_src_cnt;
	111
	112	dma_flags \|= DMA_PREP_CONTINUE;
	113	}
	114
	115	return tx;
	116	}
	117
	118	/**
	119	* do_sync_gen_syndrome - synchronously calculate a raid6 syndrome
	120	*/
	121	static void
	122	do_sync_gen_syndrome(struct page **blocks, unsigned int offset, int disks,
	123	size_t len, struct async_submit_ctl *submit)
	124	{
	125	void **srcs;
	126	int i;
584acdd4	127	int start = -1, stop = disks - 3;
b2f46fd8 DW	128
	129	if (submit->scribble)
	130	srcs = submit->scribble;
	131	else
	132	srcs = (void **) blocks;
	133
	134	for (i = 0; i < disks; i++) {
5dd33c9a	135	if (blocks[i] == NULL) {
b2f46fd8	136	BUG_ON(i > disks - 3); /* P or Q can't be zero */
5dd33c9a	137	srcs[i] = (void*)raid6_empty_zero_page;
584acdd4	138	} else {
b2f46fd8	139	srcs[i] = page_address(blocks[i]) + offset;
584acdd4 MS	140	if (i < disks - 2) {
	141	stop = i;
	142	if (start == -1)
	143	start = i;
	144	}
	145	}
b2f46fd8	146	}
584acdd4 MS	147	if (submit->flags & ASYNC_TX_PQ_XOR_DST) {
	148	BUG_ON(!raid6_call.xor_syndrome);
	149	if (start >= 0)
	150	raid6_call.xor_syndrome(disks, start, stop, len, srcs);
	151	} else
	152	raid6_call.gen_syndrome(disks, len, srcs);
b2f46fd8 DW	153	async_tx_sync_epilog(submit);
	154	}
	155
	156	/**
	157	* async_gen_syndrome - asynchronously calculate a raid6 syndrome
	158	* @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
	159	* @offset: common offset into each block (src and dest) to start transaction
	160	* @disks: number of blocks (including missing P or Q, see below)
	161	* @len: length of operation in bytes
	162	* @submit: submission/completion modifiers
	163	*
	164	* General note: This routine assumes a field of GF(2^8) with a
	165	* primitive polynomial of 0x11d and a generator of {02}.
	166	*
	167	* 'disks' note: callers can optionally omit either P or Q (but not
	168	* both) from the calculation by setting blocks[disks-2] or
	169	* blocks[disks-1] to NULL. When P or Q is omitted 'len' must be <=
	170	* PAGE_SIZE as a temporary buffer of this size is used in the
	171	* synchronous path. 'disks' always accounts for both destination
5676470f DW	172	* buffers. If any source buffers (blocks[i] where i < disks - 2) are
	173	* set to NULL those buffers will be replaced with the raid6_zero_page
	174	* in the synchronous path and omitted in the hardware-asynchronous
	175	* path.
b2f46fd8 DW	176	*/
	177	struct dma_async_tx_descriptor *
	178	async_gen_syndrome(struct page **blocks, unsigned int offset, int disks,
	179	size_t len, struct async_submit_ctl *submit)
	180	{
	181	int src_cnt = disks - 2;
	182	struct dma_chan *chan = async_tx_find_channel(submit, DMA_PQ,
	183	&P(blocks, disks), 2,
	184	blocks, src_cnt, len);
	185	struct dma_device *device = chan ? chan->device : NULL;
7476bd79	186	struct dmaengine_unmap_data *unmap = NULL;
b2f46fd8 DW	187
	188	BUG_ON(disks > 255 \|\| !(P(blocks, disks) \|\| Q(blocks, disks)));
	189
7476bd79	190	if (device)
b02bab6b	191	unmap = dmaengine_get_unmap_data(device->dev, disks, GFP_NOWAIT);
b2f46fd8	192
584acdd4 MS	193	/* XORing P/Q is only implemented in software */
584acdd4 MS	194	if (unmap && !(submit->flags & ASYNC_TX_PQ_XOR_DST) &&
b2f46fd8	195	(src_cnt <= dma_maxpq(device, 0) \|\|
83544ae9 DW	196	dma_maxpq(device, DMA_PREP_CONTINUE) > 0) &&
83544ae9 DW	197	is_dma_pq_aligned(device, offset, 0, len)) {
7476bd79 DW	198	struct dma_async_tx_descriptor *tx;
	199	enum dma_ctrl_flags dma_flags = 0;
	200	unsigned char coefs[src_cnt];
	201	int i, j;
	202
b2f46fd8 DW	203	/* run the p+q asynchronously */
	204	pr_debug("%s: (async) disks: %d len: %zu\n",
	205	__func__, disks, len);
7476bd79 DW	206
	207	/* convert source addresses being careful to collapse 'empty'
	208	* sources and update the coefficients accordingly
	209	*/
	210	unmap->len = len;
	211	for (i = 0, j = 0; i < src_cnt; i++) {
	212	if (blocks[i] == NULL)
	213	continue;
	214	unmap->addr[j] = dma_map_page(device->dev, blocks[i], offset,
	215	len, DMA_TO_DEVICE);
	216	coefs[j] = raid6_gfexp[i];
	217	unmap->to_cnt++;
	218	j++;
	219	}
	220
	221	/*
	222	* DMAs use destinations as sources,
	223	* so use BIDIRECTIONAL mapping
	224	*/
	225	unmap->bidi_cnt++;
	226	if (P(blocks, disks))
	227	unmap->addr[j++] = dma_map_page(device->dev, P(blocks, disks),
	228	offset, len, DMA_BIDIRECTIONAL);
	229	else {
	230	unmap->addr[j++] = 0;
	231	dma_flags \|= DMA_PREP_PQ_DISABLE_P;
	232	}
	233
	234	unmap->bidi_cnt++;
	235	if (Q(blocks, disks))
	236	unmap->addr[j++] = dma_map_page(device->dev, Q(blocks, disks),
	237	offset, len, DMA_BIDIRECTIONAL);
	238	else {
	239	unmap->addr[j++] = 0;
	240	dma_flags \|= DMA_PREP_PQ_DISABLE_Q;
	241	}
	242
	243	tx = do_async_gen_syndrome(chan, coefs, j, unmap, dma_flags, submit);
	244	dmaengine_unmap_put(unmap);
	245	return tx;
b2f46fd8 DW	246	}
b2f46fd8 DW	247
7476bd79 DW	248	dmaengine_unmap_put(unmap);
7476bd79 DW	249
b2f46fd8 DW	250	/* run the pq synchronously */
	251	pr_debug("%s: (sync) disks: %d len: %zu\n", __func__, disks, len);
	252
	253	/* wait for any prerequisite operations */
	254	async_tx_quiesce(&submit->depend_tx);
	255
	256	if (!P(blocks, disks)) {
030b0772	257	P(blocks, disks) = pq_scribble_page;
b2f46fd8 DW	258	BUG_ON(len + offset > PAGE_SIZE);
	259	}
	260	if (!Q(blocks, disks)) {
030b0772	261	Q(blocks, disks) = pq_scribble_page;
b2f46fd8 DW	262	BUG_ON(len + offset > PAGE_SIZE);
	263	}
	264	do_sync_gen_syndrome(blocks, offset, disks, len, submit);
	265
	266	return NULL;
	267	}
	268	EXPORT_SYMBOL_GPL(async_gen_syndrome);
	269
7b3cc2b1 DW	270	static inline struct dma_chan *
	271	pq_val_chan(struct async_submit_ctl submit, struct page *blocks, int disks, size_t len)
	272	{
	273	#ifdef CONFIG_ASYNC_TX_DISABLE_PQ_VAL_DMA
	274	return NULL;
	275	#endif
	276	return async_tx_find_channel(submit, DMA_PQ_VAL, NULL, 0, blocks,
	277	disks, len);
	278	}
	279
b2f46fd8 DW	280	/**
	281	* async_syndrome_val - asynchronously validate a raid6 syndrome
	282	* @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
	283	* @offset: common offset into each block (src and dest) to start transaction
	284	* @disks: number of blocks (including missing P or Q, see below)
	285	* @len: length of operation in bytes
	286	* @pqres: on val failure SUM_CHECK_P_RESULT and/or SUM_CHECK_Q_RESULT are set
	287	* @spare: temporary result buffer for the synchronous case
	288	* @submit: submission / completion modifiers
	289	*
	290	* The same notes from async_gen_syndrome apply to the 'blocks',
	291	* and 'disks' parameters of this routine. The synchronous path
	292	* requires a temporary result buffer and submit->scribble to be
	293	* specified.
	294	*/
	295	struct dma_async_tx_descriptor *
	296	async_syndrome_val(struct page **blocks, unsigned int offset, int disks,
	297	size_t len, enum sum_check_flags pqres, struct page spare,
	298	struct async_submit_ctl *submit)
	299	{
7b3cc2b1	300	struct dma_chan *chan = pq_val_chan(submit, blocks, disks, len);
b2f46fd8 DW	301	struct dma_device *device = chan ? chan->device : NULL;
b2f46fd8 DW	302	struct dma_async_tx_descriptor *tx;
b2141e69	303	unsigned char coefs[disks-2];
b2f46fd8	304	enum dma_ctrl_flags dma_flags = submit->cb_fn ? DMA_PREP_INTERRUPT : 0;
1786b943	305	struct dmaengine_unmap_data *unmap = NULL;
b2f46fd8 DW	306
	307	BUG_ON(disks < 4);
	308
1786b943	309	if (device)
b02bab6b	310	unmap = dmaengine_get_unmap_data(device->dev, disks, GFP_NOWAIT);
b2f46fd8	311
1786b943	312	if (unmap && disks <= dma_maxpq(device, 0) &&
83544ae9	313	is_dma_pq_aligned(device, offset, 0, len)) {
b2f46fd8	314	struct device *dev = device->dev;
1786b943 DW	315	dma_addr_t pq[2];
1786b943 DW	316	int i, j = 0, src_cnt = 0;
b2f46fd8 DW	317
	318	pr_debug("%s: (async) disks: %d len: %zu\n",
	319	__func__, disks, len);
1786b943 DW	320
	321	unmap->len = len;
	322	for (i = 0; i < disks-2; i++)
	323	if (likely(blocks[i])) {
	324	unmap->addr[j] = dma_map_page(dev, blocks[i],
	325	offset, len,
	326	DMA_TO_DEVICE);
	327	coefs[j] = raid6_gfexp[i];
	328	unmap->to_cnt++;
	329	src_cnt++;
	330	j++;
	331	}
	332
	333	if (!P(blocks, disks)) {
	334	pq[0] = 0;
b2f46fd8	335	dma_flags \|= DMA_PREP_PQ_DISABLE_P;
1786b943	336	} else {
5676470f	337	pq[0] = dma_map_page(dev, P(blocks, disks),
b2141e69 N	338	offset, len,
b2141e69 N	339	DMA_TO_DEVICE);
1786b943 DW	340	unmap->addr[j++] = pq[0];
	341	unmap->to_cnt++;
	342	}
	343	if (!Q(blocks, disks)) {
	344	pq[1] = 0;
b2f46fd8	345	dma_flags \|= DMA_PREP_PQ_DISABLE_Q;
1786b943	346	} else {
5676470f	347	pq[1] = dma_map_page(dev, Q(blocks, disks),
b2141e69 N	348	offset, len,
b2141e69 N	349	DMA_TO_DEVICE);
1786b943 DW	350	unmap->addr[j++] = pq[1];
	351	unmap->to_cnt++;
	352	}
b2141e69	353
0403e382 DW	354	if (submit->flags & ASYNC_TX_FENCE)
0403e382 DW	355	dma_flags \|= DMA_PREP_FENCE;
b2f46fd8	356	for (;;) {
1786b943 DW	357	tx = device->device_prep_dma_pq_val(chan, pq,
1786b943 DW	358	unmap->addr,
b2141e69 N	359	src_cnt,
b2141e69 N	360	coefs,
b2f46fd8 DW	361	len, pqres,
	362	dma_flags);
	363	if (likely(tx))
	364	break;
	365	async_tx_quiesce(&submit->depend_tx);
	366	dma_async_issue_pending(chan);
	367	}
1786b943 DW	368
1786b943 DW	369	dma_set_unmap(tx, unmap);
b2f46fd8	370	async_tx_submit(chan, tx, submit);
b2f46fd8 DW	371	} else {
	372	struct page *p_src = P(blocks, disks);
	373	struct page *q_src = Q(blocks, disks);
	374	enum async_tx_flags flags_orig = submit->flags;
	375	dma_async_tx_callback cb_fn_orig = submit->cb_fn;
	376	void *scribble = submit->scribble;
	377	void *cb_param_orig = submit->cb_param;
	378	void p, q, *s;
	379
	380	pr_debug("%s: (sync) disks: %d len: %zu\n",
	381	__func__, disks, len);
	382
	383	/* caller must provide a temporary result buffer and
	384	* allow the input parameters to be preserved
	385	*/
	386	BUG_ON(!spare \|\| !scribble);
	387
	388	/* wait for any prerequisite operations */
	389	async_tx_quiesce(&submit->depend_tx);
	390
	391	/* recompute p and/or q into the temporary buffer and then
	392	* check to see the result matches the current value
	393	*/
	394	tx = NULL;
	395	*pqres = 0;
	396	if (p_src) {
	397	init_async_submit(submit, ASYNC_TX_XOR_ZERO_DST, NULL,
	398	NULL, NULL, scribble);
	399	tx = async_xor(spare, blocks, offset, disks-2, len, submit);
	400	async_tx_quiesce(&tx);
	401	p = page_address(p_src) + offset;
	402	s = page_address(spare) + offset;
	403	*pqres \|= !!memcmp(p, s, len) << SUM_CHECK_P;
	404	}
	405
	406	if (q_src) {
	407	P(blocks, disks) = NULL;
	408	Q(blocks, disks) = spare;
	409	init_async_submit(submit, 0, NULL, NULL, NULL, scribble);
	410	tx = async_gen_syndrome(blocks, offset, disks, len, submit);
	411	async_tx_quiesce(&tx);
	412	q = page_address(q_src) + offset;
	413	s = page_address(spare) + offset;
	414	*pqres \|= !!memcmp(q, s, len) << SUM_CHECK_Q;
	415	}
	416
	417	/* restore P, Q and submit */
	418	P(blocks, disks) = p_src;
	419	Q(blocks, disks) = q_src;
	420
	421	submit->cb_fn = cb_fn_orig;
	422	submit->cb_param = cb_param_orig;
	423	submit->flags = flags_orig;
	424	async_tx_sync_epilog(submit);
c8475090	425	tx = NULL;
b2f46fd8	426	}
c8475090 JM	427	dmaengine_unmap_put(unmap);
	428
	429	return tx;
b2f46fd8 DW	430	}
	431	EXPORT_SYMBOL_GPL(async_syndrome_val);
	432
	433	static int __init async_pq_init(void)
	434	{
030b0772	435	pq_scribble_page = alloc_page(GFP_KERNEL);
b2f46fd8	436
030b0772	437	if (pq_scribble_page)
b2f46fd8 DW	438	return 0;
	439
	440	pr_err("%s: failed to allocate required spare page\n", __func__);
	441
	442	return -ENOMEM;
	443	}
	444
	445	static void __exit async_pq_exit(void)
	446	{
95813b8f	447	__free_page(pq_scribble_page);
b2f46fd8 DW	448	}
	449
	450	module_init(async_pq_init);
	451	module_exit(async_pq_exit);
	452
	453	MODULE_DESCRIPTION("asynchronous raid6 syndrome generation/validation");
	454	MODULE_LICENSE("GPL");