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

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * Copyright(c) 2007 Yuri Tikhonov <yur@emcraft.com>
 * Copyright(c) 2009 Intel Corporation
 */
#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>

/*
 * struct 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])

#define MAX_DISKS 255

/*
 * 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;

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

		/* 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 *offsets, 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]) + offsets[i];

			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);
}

static inline bool
is_dma_pq_aligned_offs(struct dma_device *dev, unsigned int *offs,
				     int src_cnt, size_t len)
{
	int i;

	for (i = 0; i < src_cnt; i++) {
		if (!is_dma_pq_aligned(dev, offs[i], 0, len))
			return false;
	}
	return true;
}

/**
 * async_gen_syndrome - asynchronously calculate a raid6 syndrome
 * @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
 * @offsets: offset array 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 *offsets, 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 > MAX_DISKS || !(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_offs(device, offsets, disks, len)) {
		struct dma_async_tx_descriptor *tx;
		enum dma_ctrl_flags dma_flags = 0;
		unsigned char coefs[MAX_DISKS];
		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],
						offsets[i], 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),
							P(offsets, disks),
							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),
							Q(offsets, disks),
							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;
		P(offsets, disks) = 0;
	}
	if (!Q(blocks, disks)) {
		Q(blocks, disks) = pq_scribble_page;
		Q(offsets, disks) = 0;
	}
	do_sync_gen_syndrome(blocks, offsets, 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
 * @offsets: 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
 * @s_off: spare buffer page offset
 * @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 *offsets, int disks,
		   size_t len, enum sum_check_flags *pqres, struct page *spare,
		   unsigned int s_off, 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[MAX_DISKS];
	enum dma_ctrl_flags dma_flags = submit->cb_fn ? DMA_PREP_INTERRUPT : 0;
	struct dmaengine_unmap_data *unmap = NULL;

	BUG_ON(disks < 4 || disks > MAX_DISKS);

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

	if (unmap && disks <= dma_maxpq(device, 0) &&
	    is_dma_pq_aligned_offs(device, offsets, disks, 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],
							      offsets[i], 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),
					     P(offsets, disks), 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),
					     Q(offsets, disks), 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);
		unsigned int p_off = P(offsets, disks);
		struct page *q_src = Q(blocks, disks);
		unsigned int q_off = Q(offsets, 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_offs(spare, s_off,
					blocks, offsets, disks-2, len, submit);
			async_tx_quiesce(&tx);
			p = page_address(p_src) + p_off;
			s = page_address(spare) + s_off;
			*pqres |= !!memcmp(p, s, len) << SUM_CHECK_P;
		}

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

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

		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
e62d9491	1	// SPDX-License-Identifier: GPL-2.0-or-later
b2f46fd8 DW	2	/*
	3	* Copyright(c) 2007 Yuri Tikhonov <yur@emcraft.com>
	4	* Copyright(c) 2009 Intel Corporation
b2f46fd8 DW	5	*/
	6	#include <linux/kernel.h>
	7	#include <linux/interrupt.h>
4bb33cc8	8	#include <linux/module.h>
b2f46fd8 DW	9	#include <linux/dma-mapping.h>
	10	#include <linux/raid/pq.h>
	11	#include <linux/async_tx.h>
5a0e3ad6	12	#include <linux/gfp.h>
b2f46fd8	13
0dee6cd2 RD	14	/*
0dee6cd2 RD	15	* struct pq_scribble_page - space to hold throwaway P or Q buffer for
030b0772	16	* synchronous gen_syndrome
b2f46fd8	17	*/
030b0772	18	static struct page *pq_scribble_page;
b2f46fd8	19
b2f46fd8 DW	20	/* the struct page *blocks[] parameter passed to async_gen_syndrome()
	21	* and async_syndrome_val() contains the 'P' destination address at
	22	* blocks[disks-2] and the 'Q' destination address at blocks[disks-1]
	23	*
	24	* note: these are macros as they are used as lvalues
	25	*/
	26	#define P(b, d) (b[d-2])
	27	#define Q(b, d) (b[d-1])
	28
89a7e2f7 KS	29	#define MAX_DISKS 255
89a7e2f7 KS	30
0dee6cd2	31	/*
b2f46fd8 DW	32	* do_async_gen_syndrome - asynchronously calculate P and/or Q
	33	*/
	34	static __async_inline struct dma_async_tx_descriptor *
7476bd79 DW	35	do_async_gen_syndrome(struct dma_chan *chan,
	36	const unsigned char *scfs, int disks,
	37	struct dmaengine_unmap_data *unmap,
	38	enum dma_ctrl_flags dma_flags,
b2f46fd8 DW	39	struct async_submit_ctl *submit)
	40	{
	41	struct dma_async_tx_descriptor *tx = NULL;
	42	struct dma_device *dma = chan->device;
b2f46fd8 DW	43	enum async_tx_flags flags_orig = submit->flags;
	44	dma_async_tx_callback cb_fn_orig = submit->cb_fn;
	45	dma_async_tx_callback cb_param_orig = submit->cb_param;
	46	int src_cnt = disks - 2;
b2f46fd8 DW	47	unsigned short pq_src_cnt;
	48	dma_addr_t dma_dest[2];
	49	int src_off = 0;
b2f46fd8	50
b2f46fd8 DW	51	while (src_cnt > 0) {
	52	submit->flags = flags_orig;
	53	pq_src_cnt = min(src_cnt, dma_maxpq(dma, dma_flags));
	54	/* if we are submitting additional pqs, leave the chain open,
	55	* clear the callback parameters, and leave the destination
	56	* buffers mapped
	57	*/
	58	if (src_cnt > pq_src_cnt) {
	59	submit->flags &= ~ASYNC_TX_ACK;
0403e382	60	submit->flags \|= ASYNC_TX_FENCE;
b2f46fd8 DW	61	submit->cb_fn = NULL;
	62	submit->cb_param = NULL;
	63	} else {
b2f46fd8 DW	64	submit->cb_fn = cb_fn_orig;
	65	submit->cb_param = cb_param_orig;
	66	if (cb_fn_orig)
	67	dma_flags \|= DMA_PREP_INTERRUPT;
	68	}
baae03a0 AP	69	if (submit->flags & ASYNC_TX_FENCE)
baae03a0 AP	70	dma_flags \|= DMA_PREP_FENCE;
b2f46fd8	71
7476bd79 DW	72	/* Drivers force forward progress in case they can not provide
7476bd79 DW	73	* a descriptor
b2f46fd8 DW	74	*/
b2f46fd8 DW	75	for (;;) {
7476bd79 DW	76	dma_dest[0] = unmap->addr[disks - 2];
7476bd79 DW	77	dma_dest[1] = unmap->addr[disks - 1];
b2f46fd8	78	tx = dma->device_prep_dma_pq(chan, dma_dest,
7476bd79	79	&unmap->addr[src_off],
b2f46fd8	80	pq_src_cnt,
7476bd79	81	&scfs[src_off], unmap->len,
b2f46fd8 DW	82	dma_flags);
	83	if (likely(tx))
	84	break;
	85	async_tx_quiesce(&submit->depend_tx);
	86	dma_async_issue_pending(chan);
	87	}
	88
7476bd79	89	dma_set_unmap(tx, unmap);
b2f46fd8 DW	90	async_tx_submit(chan, tx, submit);
	91	submit->depend_tx = tx;
	92
	93	/* drop completed sources */
	94	src_cnt -= pq_src_cnt;
	95	src_off += pq_src_cnt;
	96
	97	dma_flags \|= DMA_PREP_CONTINUE;
	98	}
	99
	100	return tx;
	101	}
	102
0dee6cd2	103	/*
b2f46fd8 DW	104	* do_sync_gen_syndrome - synchronously calculate a raid6 syndrome
	105	*/
	106	static void
d69454bc	107	do_sync_gen_syndrome(struct page *blocks, unsigned int offsets, int disks,
b2f46fd8 DW	108	size_t len, struct async_submit_ctl *submit)
	109	{
	110	void **srcs;
	111	int i;
584acdd4	112	int start = -1, stop = disks - 3;
b2f46fd8 DW	113
	114	if (submit->scribble)
	115	srcs = submit->scribble;
	116	else
	117	srcs = (void **) blocks;
	118
	119	for (i = 0; i < disks; i++) {
5dd33c9a	120	if (blocks[i] == NULL) {
b2f46fd8	121	BUG_ON(i > disks - 3); /* P or Q can't be zero */
5dd33c9a	122	srcs[i] = (void*)raid6_empty_zero_page;
584acdd4	123	} else {
d69454bc YY	124	srcs[i] = page_address(blocks[i]) + offsets[i];
d69454bc YY	125
584acdd4 MS	126	if (i < disks - 2) {
	127	stop = i;
	128	if (start == -1)
	129	start = i;
	130	}
	131	}
b2f46fd8	132	}
584acdd4 MS	133	if (submit->flags & ASYNC_TX_PQ_XOR_DST) {
	134	BUG_ON(!raid6_call.xor_syndrome);
	135	if (start >= 0)
	136	raid6_call.xor_syndrome(disks, start, stop, len, srcs);
	137	} else
	138	raid6_call.gen_syndrome(disks, len, srcs);
b2f46fd8 DW	139	async_tx_sync_epilog(submit);
	140	}
	141
d69454bc YY	142	static inline bool
	143	is_dma_pq_aligned_offs(struct dma_device dev, unsigned int offs,
	144	int src_cnt, size_t len)
	145	{
	146	int i;
	147
	148	for (i = 0; i < src_cnt; i++) {
	149	if (!is_dma_pq_aligned(dev, offs[i], 0, len))
	150	return false;
	151	}
	152	return true;
	153	}
	154
b2f46fd8 DW	155	/**
	156	* async_gen_syndrome - asynchronously calculate a raid6 syndrome
	157	* @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
d69454bc	158	* @offsets: offset array into each block (src and dest) to start transaction
b2f46fd8 DW	159	* @disks: number of blocks (including missing P or Q, see below)
	160	* @len: length of operation in bytes
	161	* @submit: submission/completion modifiers
	162	*
	163	* General note: This routine assumes a field of GF(2^8) with a
	164	* primitive polynomial of 0x11d and a generator of {02}.
	165	*
	166	* 'disks' note: callers can optionally omit either P or Q (but not
	167	* both) from the calculation by setting blocks[disks-2] or
	168	* blocks[disks-1] to NULL. When P or Q is omitted 'len' must be <=
	169	* PAGE_SIZE as a temporary buffer of this size is used in the
	170	* synchronous path. 'disks' always accounts for both destination
5676470f DW	171	* buffers. If any source buffers (blocks[i] where i < disks - 2) are
	172	* set to NULL those buffers will be replaced with the raid6_zero_page
	173	* in the synchronous path and omitted in the hardware-asynchronous
	174	* path.
b2f46fd8 DW	175	*/
b2f46fd8 DW	176	struct dma_async_tx_descriptor *
d69454bc	177	async_gen_syndrome(struct page *blocks, unsigned int offsets, int disks,
b2f46fd8 DW	178	size_t len, struct async_submit_ctl *submit)
	179	{
	180	int src_cnt = disks - 2;
	181	struct dma_chan *chan = async_tx_find_channel(submit, DMA_PQ,
	182	&P(blocks, disks), 2,
	183	blocks, src_cnt, len);
	184	struct dma_device *device = chan ? chan->device : NULL;
7476bd79	185	struct dmaengine_unmap_data *unmap = NULL;
b2f46fd8	186
89a7e2f7	187	BUG_ON(disks > MAX_DISKS \|\| !(P(blocks, disks) \|\| Q(blocks, disks)));
b2f46fd8	188
7476bd79	189	if (device)
b02bab6b	190	unmap = dmaengine_get_unmap_data(device->dev, disks, GFP_NOWAIT);
b2f46fd8	191
584acdd4 MS	192	/* XORing P/Q is only implemented in software */
584acdd4 MS	193	if (unmap && !(submit->flags & ASYNC_TX_PQ_XOR_DST) &&
b2f46fd8	194	(src_cnt <= dma_maxpq(device, 0) \|\|
83544ae9	195	dma_maxpq(device, DMA_PREP_CONTINUE) > 0) &&
d69454bc	196	is_dma_pq_aligned_offs(device, offsets, disks, len)) {
7476bd79 DW	197	struct dma_async_tx_descriptor *tx;
7476bd79 DW	198	enum dma_ctrl_flags dma_flags = 0;
89a7e2f7	199	unsigned char coefs[MAX_DISKS];
7476bd79 DW	200	int i, j;
7476bd79 DW	201
b2f46fd8 DW	202	/* run the p+q asynchronously */
	203	pr_debug("%s: (async) disks: %d len: %zu\n",
	204	__func__, disks, len);
7476bd79 DW	205
	206	/* convert source addresses being careful to collapse 'empty'
	207	* sources and update the coefficients accordingly
	208	*/
	209	unmap->len = len;
	210	for (i = 0, j = 0; i < src_cnt; i++) {
	211	if (blocks[i] == NULL)
	212	continue;
d69454bc YY	213	unmap->addr[j] = dma_map_page(device->dev, blocks[i],
d69454bc YY	214	offsets[i], len, DMA_TO_DEVICE);
7476bd79 DW	215	coefs[j] = raid6_gfexp[i];
	216	unmap->to_cnt++;
	217	j++;
	218	}
	219
	220	/*
	221	* DMAs use destinations as sources,
	222	* so use BIDIRECTIONAL mapping
	223	*/
	224	unmap->bidi_cnt++;
	225	if (P(blocks, disks))
	226	unmap->addr[j++] = dma_map_page(device->dev, P(blocks, disks),
d69454bc YY	227	P(offsets, disks),
d69454bc YY	228	len, DMA_BIDIRECTIONAL);
7476bd79 DW	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),
d69454bc YY	237	Q(offsets, disks),
d69454bc YY	238	len, DMA_BIDIRECTIONAL);
7476bd79 DW	239	else {
	240	unmap->addr[j++] = 0;
	241	dma_flags \|= DMA_PREP_PQ_DISABLE_Q;
	242	}
	243
	244	tx = do_async_gen_syndrome(chan, coefs, j, unmap, dma_flags, submit);
	245	dmaengine_unmap_put(unmap);
	246	return tx;
b2f46fd8 DW	247	}
b2f46fd8 DW	248
7476bd79 DW	249	dmaengine_unmap_put(unmap);
7476bd79 DW	250
b2f46fd8 DW	251	/* run the pq synchronously */
	252	pr_debug("%s: (sync) disks: %d len: %zu\n", __func__, disks, len);
	253
	254	/* wait for any prerequisite operations */
	255	async_tx_quiesce(&submit->depend_tx);
	256
	257	if (!P(blocks, disks)) {
030b0772	258	P(blocks, disks) = pq_scribble_page;
d69454bc	259	P(offsets, disks) = 0;
b2f46fd8 DW	260	}
b2f46fd8 DW	261	if (!Q(blocks, disks)) {
030b0772	262	Q(blocks, disks) = pq_scribble_page;
d69454bc	263	Q(offsets, disks) = 0;
b2f46fd8	264	}
d69454bc	265	do_sync_gen_syndrome(blocks, offsets, disks, len, submit);
b2f46fd8 DW	266
	267	return NULL;
	268	}
	269	EXPORT_SYMBOL_GPL(async_gen_syndrome);
	270
7b3cc2b1 DW	271	static inline struct dma_chan *
	272	pq_val_chan(struct async_submit_ctl submit, struct page *blocks, int disks, size_t len)
	273	{
	274	#ifdef CONFIG_ASYNC_TX_DISABLE_PQ_VAL_DMA
	275	return NULL;
	276	#endif
	277	return async_tx_find_channel(submit, DMA_PQ_VAL, NULL, 0, blocks,
	278	disks, len);
	279	}
	280
b2f46fd8 DW	281	/**
	282	* async_syndrome_val - asynchronously validate a raid6 syndrome
	283	* @blocks: source blocks from idx 0..disks-3, P @ disks-2 and Q @ disks-1
0dee6cd2	284	* @offsets: common offset into each block (src and dest) to start transaction
b2f46fd8 DW	285	* @disks: number of blocks (including missing P or Q, see below)
	286	* @len: length of operation in bytes
	287	* @pqres: on val failure SUM_CHECK_P_RESULT and/or SUM_CHECK_Q_RESULT are set
	288	* @spare: temporary result buffer for the synchronous case
d69454bc	289	* @s_off: spare buffer page offset
b2f46fd8 DW	290	* @submit: submission / completion modifiers
	291	*
	292	* The same notes from async_gen_syndrome apply to the 'blocks',
	293	* and 'disks' parameters of this routine. The synchronous path
	294	* requires a temporary result buffer and submit->scribble to be
	295	* specified.
	296	*/
	297	struct dma_async_tx_descriptor *
d69454bc	298	async_syndrome_val(struct page *blocks, unsigned int offsets, int disks,
b2f46fd8	299	size_t len, enum sum_check_flags pqres, struct page spare,
d69454bc	300	unsigned int s_off, struct async_submit_ctl *submit)
b2f46fd8	301	{
7b3cc2b1	302	struct dma_chan *chan = pq_val_chan(submit, blocks, disks, len);
b2f46fd8 DW	303	struct dma_device *device = chan ? chan->device : NULL;
b2f46fd8 DW	304	struct dma_async_tx_descriptor *tx;
89a7e2f7	305	unsigned char coefs[MAX_DISKS];
b2f46fd8	306	enum dma_ctrl_flags dma_flags = submit->cb_fn ? DMA_PREP_INTERRUPT : 0;
1786b943	307	struct dmaengine_unmap_data *unmap = NULL;
b2f46fd8	308
89a7e2f7	309	BUG_ON(disks < 4 \|\| disks > MAX_DISKS);
b2f46fd8	310
1786b943	311	if (device)
b02bab6b	312	unmap = dmaengine_get_unmap_data(device->dev, disks, GFP_NOWAIT);
b2f46fd8	313
1786b943	314	if (unmap && disks <= dma_maxpq(device, 0) &&
d69454bc	315	is_dma_pq_aligned_offs(device, offsets, disks, len)) {
b2f46fd8	316	struct device *dev = device->dev;
1786b943 DW	317	dma_addr_t pq[2];
1786b943 DW	318	int i, j = 0, src_cnt = 0;
b2f46fd8 DW	319
	320	pr_debug("%s: (async) disks: %d len: %zu\n",
	321	__func__, disks, len);
1786b943 DW	322
	323	unmap->len = len;
	324	for (i = 0; i < disks-2; i++)
	325	if (likely(blocks[i])) {
	326	unmap->addr[j] = dma_map_page(dev, blocks[i],
d69454bc	327	offsets[i], len,
1786b943 DW	328	DMA_TO_DEVICE);
	329	coefs[j] = raid6_gfexp[i];
	330	unmap->to_cnt++;
	331	src_cnt++;
	332	j++;
	333	}
	334
	335	if (!P(blocks, disks)) {
	336	pq[0] = 0;
b2f46fd8	337	dma_flags \|= DMA_PREP_PQ_DISABLE_P;
1786b943	338	} else {
5676470f	339	pq[0] = dma_map_page(dev, P(blocks, disks),
d69454bc	340	P(offsets, disks), len,
b2141e69	341	DMA_TO_DEVICE);
1786b943 DW	342	unmap->addr[j++] = pq[0];
	343	unmap->to_cnt++;
	344	}
	345	if (!Q(blocks, disks)) {
	346	pq[1] = 0;
b2f46fd8	347	dma_flags \|= DMA_PREP_PQ_DISABLE_Q;
1786b943	348	} else {
5676470f	349	pq[1] = dma_map_page(dev, Q(blocks, disks),
d69454bc	350	Q(offsets, disks), len,
b2141e69	351	DMA_TO_DEVICE);
1786b943 DW	352	unmap->addr[j++] = pq[1];
	353	unmap->to_cnt++;
	354	}
b2141e69	355
0403e382 DW	356	if (submit->flags & ASYNC_TX_FENCE)
0403e382 DW	357	dma_flags \|= DMA_PREP_FENCE;
b2f46fd8	358	for (;;) {
1786b943 DW	359	tx = device->device_prep_dma_pq_val(chan, pq,
1786b943 DW	360	unmap->addr,
b2141e69 N	361	src_cnt,
b2141e69 N	362	coefs,
b2f46fd8 DW	363	len, pqres,
	364	dma_flags);
	365	if (likely(tx))
	366	break;
	367	async_tx_quiesce(&submit->depend_tx);
	368	dma_async_issue_pending(chan);
	369	}
1786b943 DW	370
1786b943 DW	371	dma_set_unmap(tx, unmap);
b2f46fd8	372	async_tx_submit(chan, tx, submit);
b2f46fd8 DW	373	} else {
b2f46fd8 DW	374	struct page *p_src = P(blocks, disks);
d69454bc	375	unsigned int p_off = P(offsets, disks);
b2f46fd8	376	struct page *q_src = Q(blocks, disks);
d69454bc	377	unsigned int q_off = Q(offsets, disks);
b2f46fd8 DW	378	enum async_tx_flags flags_orig = submit->flags;
	379	dma_async_tx_callback cb_fn_orig = submit->cb_fn;
	380	void *scribble = submit->scribble;
	381	void *cb_param_orig = submit->cb_param;
	382	void p, q, *s;
	383
	384	pr_debug("%s: (sync) disks: %d len: %zu\n",
	385	__func__, disks, len);
	386
	387	/* caller must provide a temporary result buffer and
	388	* allow the input parameters to be preserved
	389	*/
	390	BUG_ON(!spare \|\| !scribble);
	391
	392	/* wait for any prerequisite operations */
	393	async_tx_quiesce(&submit->depend_tx);
	394
	395	/* recompute p and/or q into the temporary buffer and then
	396	* check to see the result matches the current value
	397	*/
	398	tx = NULL;
	399	*pqres = 0;
	400	if (p_src) {
	401	init_async_submit(submit, ASYNC_TX_XOR_ZERO_DST, NULL,
	402	NULL, NULL, scribble);
d69454bc YY	403	tx = async_xor_offs(spare, s_off,
d69454bc YY	404	blocks, offsets, disks-2, len, submit);
b2f46fd8	405	async_tx_quiesce(&tx);
d69454bc YY	406	p = page_address(p_src) + p_off;
d69454bc YY	407	s = page_address(spare) + s_off;
b2f46fd8 DW	408	*pqres \|= !!memcmp(p, s, len) << SUM_CHECK_P;
	409	}
	410
	411	if (q_src) {
	412	P(blocks, disks) = NULL;
	413	Q(blocks, disks) = spare;
d69454bc	414	Q(offsets, disks) = s_off;
b2f46fd8	415	init_async_submit(submit, 0, NULL, NULL, NULL, scribble);
d69454bc YY	416	tx = async_gen_syndrome(blocks, offsets, disks,
d69454bc YY	417	len, submit);
b2f46fd8	418	async_tx_quiesce(&tx);
d69454bc YY	419	q = page_address(q_src) + q_off;
d69454bc YY	420	s = page_address(spare) + s_off;
b2f46fd8 DW	421	*pqres \|= !!memcmp(q, s, len) << SUM_CHECK_Q;
	422	}
	423
	424	/* restore P, Q and submit */
	425	P(blocks, disks) = p_src;
d69454bc	426	P(offsets, disks) = p_off;
b2f46fd8	427	Q(blocks, disks) = q_src;
d69454bc	428	Q(offsets, disks) = q_off;
b2f46fd8 DW	429
	430	submit->cb_fn = cb_fn_orig;
	431	submit->cb_param = cb_param_orig;
	432	submit->flags = flags_orig;
	433	async_tx_sync_epilog(submit);
c8475090	434	tx = NULL;
b2f46fd8	435	}
c8475090 JM	436	dmaengine_unmap_put(unmap);
	437
	438	return tx;
b2f46fd8 DW	439	}
	440	EXPORT_SYMBOL_GPL(async_syndrome_val);
	441
	442	static int __init async_pq_init(void)
	443	{
030b0772	444	pq_scribble_page = alloc_page(GFP_KERNEL);
b2f46fd8	445
030b0772	446	if (pq_scribble_page)
b2f46fd8 DW	447	return 0;
	448
	449	pr_err("%s: failed to allocate required spare page\n", __func__);
	450
	451	return -ENOMEM;
	452	}
	453
	454	static void __exit async_pq_exit(void)
	455	{
95813b8f	456	__free_page(pq_scribble_page);
b2f46fd8 DW	457	}
	458
	459	module_init(async_pq_init);
	460	module_exit(async_pq_exit);
	461
	462	MODULE_DESCRIPTION("asynchronous raid6 syndrome generation/validation");
	463	MODULE_LICENSE("GPL");