mxs-dma.c 26.4 KB
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016
/*
 * Copyright 2011-2015 Freescale Semiconductor, Inc. All Rights Reserved.
 * Copyright 2017 NXP
 *
 * Refer to drivers/dma/imx-sdma.c
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 */

#include <linux/init.h>
#include <linux/types.h>
#include <linux/mm.h>
#include <linux/interrupt.h>
#include <linux/clk.h>
#include <linux/wait.h>
#include <linux/sched.h>
#include <linux/semaphore.h>
#include <linux/device.h>
#include <linux/dma-mapping.h>
#include <linux/slab.h>
#include <linux/platform_device.h>
#include <linux/dmaengine.h>
#include <linux/delay.h>
#include <linux/module.h>
#include <linux/stmp_device.h>
#include <linux/of.h>
#include <linux/of_device.h>
#include <linux/of_dma.h>
#include <linux/list.h>
#include <asm/irq.h>
#include <linux/pm_runtime.h>
#include <linux/dmapool.h>

#include "dmaengine.h"

/*
 * NOTE: The term "PIO" throughout the mxs-dma implementation means
 * PIO mode of mxs apbh-dma and apbx-dma.  With this working mode,
 * dma can program the controller registers of peripheral devices.
 */

#define dma_is_apbh(mxs_dma)	((mxs_dma)->type == MXS_DMA_APBH)
#define apbh_is_old(mxs_dma)	((mxs_dma)->dev_id == IMX23_DMA)

#define MXS_DMA_RPM_TIMEOUT 50 /* ms */

#define HW_APBHX_CTRL0				0x000
#define BM_APBH_CTRL0_APB_BURST8_EN		(1 << 29)
#define BM_APBH_CTRL0_APB_BURST_EN		(1 << 28)
#define BP_APBH_CTRL0_RESET_CHANNEL		16
#define HW_APBHX_CTRL1				0x010
#define HW_APBHX_CTRL2				0x020
#define HW_APBHX_CHANNEL_CTRL			0x030
#define BP_APBHX_CHANNEL_CTRL_RESET_CHANNEL	16
/*
 * The offset of NXTCMDAR register is different per both dma type and version,
 * while stride for each channel is all the same 0x70.
 */
#define HW_APBHX_CHn_NXTCMDAR(d, n) \
	(((dma_is_apbh(d) && apbh_is_old(d)) ? 0x050 : 0x110) + (n) * 0x70)
#define HW_APBHX_CHn_SEMA(d, n) \
	(((dma_is_apbh(d) && apbh_is_old(d)) ? 0x080 : 0x140) + (n) * 0x70)
#define HW_APBHX_CHn_BAR(d, n) \
	(((dma_is_apbh(d) && apbh_is_old(d)) ? 0x070 : 0x130) + (n) * 0x70)
#define HW_APBX_CHn_DEBUG1(d, n) (0x150 + (n) * 0x70)

/*
 * ccw bits definitions
 *
 * COMMAND:		0..1	(2)
 * CHAIN:		2	(1)
 * IRQ:			3	(1)
 * NAND_LOCK:		4	(1) - not implemented
 * NAND_WAIT4READY:	5	(1) - not implemented
 * DEC_SEM:		6	(1)
 * WAIT4END:		7	(1)
 * HALT_ON_TERMINATE:	8	(1)
 * TERMINATE_FLUSH:	9	(1)
 * RESERVED:		10..11	(2)
 * PIO_NUM:		12..15	(4)
 */
#define BP_CCW_COMMAND		0
#define BM_CCW_COMMAND		(3 << 0)
#define CCW_CHAIN		(1 << 2)
#define CCW_IRQ			(1 << 3)
#define CCW_DEC_SEM		(1 << 6)
#define CCW_WAIT4END		(1 << 7)
#define CCW_HALT_ON_TERM	(1 << 8)
#define CCW_TERM_FLUSH		(1 << 9)
#define BP_CCW_PIO_NUM		12
#define BM_CCW_PIO_NUM		(0xf << 12)

#define BF_CCW(value, field)	(((value) << BP_CCW_##field) & BM_CCW_##field)

#define MXS_DMA_CMD_NO_XFER	0
#define MXS_DMA_CMD_WRITE	1
#define MXS_DMA_CMD_READ	2
#define MXS_DMA_CMD_DMA_SENSE	3	/* not implemented */

struct mxs_dma_ccw {
	u32		next;
	u16		bits;
	u16		xfer_bytes;
#define MAX_XFER_BYTES	0xff00
	u32		bufaddr;
#define MXS_PIO_WORDS	16
	u32		pio_words[MXS_PIO_WORDS];
};

#define CCW_BLOCK_SIZE	(4 * PAGE_SIZE)
#define NUM_CCW	(int)(CCW_BLOCK_SIZE / sizeof(struct mxs_dma_ccw))

struct mxs_dma_chan {
	struct mxs_dma_engine		*mxs_dma;
	struct dma_chan			chan;
	struct dma_async_tx_descriptor	desc;
	struct tasklet_struct		tasklet;
	unsigned int			chan_irq;
	struct mxs_dma_ccw		*ccw;
	dma_addr_t			ccw_phys;
	int				desc_count;
	enum dma_status			status;
	unsigned int			flags;
	bool				reset;
	struct dma_pool			*ccw_pool;
#define MXS_DMA_SG_LOOP			(1 << 0)
#define MXS_DMA_USE_SEMAPHORE		(1 << 1)
};

#define MXS_DMA_CHANNELS		16
#define MXS_DMA_CHANNELS_MASK		0xffff

enum mxs_dma_devtype {
	MXS_DMA_APBH,
	MXS_DMA_APBX,
};

enum mxs_dma_id {
	IMX23_DMA,
	IMX28_DMA,
	IMX7D_DMA,
	IMX8QXP_DMA,
};

struct mxs_dma_engine {
	enum mxs_dma_id			dev_id;
	enum mxs_dma_devtype		type;
	void __iomem			*base;
	struct clk			*clk;
	struct clk			*clk_io;
	struct dma_device		dma_device;
	struct device_dma_parameters	dma_parms;
	struct mxs_dma_chan		mxs_chans[MXS_DMA_CHANNELS];
	struct platform_device		*pdev;
	unsigned int			nr_channels;
};

struct mxs_dma_type {
	enum mxs_dma_id id;
	enum mxs_dma_devtype type;
};

static struct mxs_dma_type mxs_dma_types[] = {
	{
		.id = IMX23_DMA,
		.type = MXS_DMA_APBH,
	}, {
		.id = IMX23_DMA,
		.type = MXS_DMA_APBX,
	}, {
		.id = IMX28_DMA,
		.type = MXS_DMA_APBH,
	}, {
		.id = IMX28_DMA,
		.type = MXS_DMA_APBX,
	}, {
		.id = IMX7D_DMA,
		.type = MXS_DMA_APBH,
	}, {
		.id = IMX8QXP_DMA,
		.type = MXS_DMA_APBH,
	}
};

static const struct platform_device_id mxs_dma_ids[] = {
	{
		.name = "imx23-dma-apbh",
		.driver_data = (kernel_ulong_t) &mxs_dma_types[0],
	}, {
		.name = "imx23-dma-apbx",
		.driver_data = (kernel_ulong_t) &mxs_dma_types[1],
	}, {
		.name = "imx28-dma-apbh",
		.driver_data = (kernel_ulong_t) &mxs_dma_types[2],
	}, {
		.name = "imx28-dma-apbx",
		.driver_data = (kernel_ulong_t) &mxs_dma_types[3],
	}, {
		.name = "imx7d-dma-apbh",
		.driver_data = (kernel_ulong_t) &mxs_dma_types[4],
	}, {
		.name = "imx8qxp-dma-apbh",
		.driver_data = (kernel_ulong_t) &mxs_dma_types[5],
	}, {
		/* end of list */
	}
};

static const struct of_device_id mxs_dma_dt_ids[] = {
	{ .compatible = "fsl,imx23-dma-apbh", .data = &mxs_dma_ids[0], },
	{ .compatible = "fsl,imx23-dma-apbx", .data = &mxs_dma_ids[1], },
	{ .compatible = "fsl,imx28-dma-apbh", .data = &mxs_dma_ids[2], },
	{ .compatible = "fsl,imx28-dma-apbx", .data = &mxs_dma_ids[3], },
	{ .compatible = "fsl,imx7d-dma-apbh", .data = &mxs_dma_ids[4], },
	{ .compatible = "fsl,imx8qxp-dma-apbh", .data = &mxs_dma_ids[5], },
	{ /* sentinel */ }
};
MODULE_DEVICE_TABLE(of, mxs_dma_dt_ids);

static struct mxs_dma_chan *to_mxs_dma_chan(struct dma_chan *chan)
{
	return container_of(chan, struct mxs_dma_chan, chan);
}

static void mxs_dma_reset_chan(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	int chan_id = mxs_chan->chan.chan_id;

	/*
	 * mxs dma channel resets can cause a channel stall. To recover from a
	 * channel stall, we have to reset the whole DMA engine. To avoid this,
	 * we use cyclic DMA with semaphores, that are enhanced in
	 * mxs_dma_int_handler. To reset the channel, we can simply stop writing
	 * into the semaphore counter.
	 */
	if (mxs_chan->flags & MXS_DMA_USE_SEMAPHORE &&
			mxs_chan->flags & MXS_DMA_SG_LOOP) {
		mxs_chan->reset = true;
	} else if (dma_is_apbh(mxs_dma) && apbh_is_old(mxs_dma)) {
		writel(1 << (chan_id + BP_APBH_CTRL0_RESET_CHANNEL),
			mxs_dma->base + HW_APBHX_CTRL0 + STMP_OFFSET_REG_SET);
	} else {
		unsigned long elapsed = 0;
		const unsigned long max_wait = 50000; /* 50ms */
		void __iomem *reg_dbg1 = mxs_dma->base +
				HW_APBX_CHn_DEBUG1(mxs_dma, chan_id);

		/*
		 * On i.MX28 APBX, the DMA channel can stop working if we reset
		 * the channel while it is in READ_FLUSH (0x08) state.
		 * We wait here until we leave the state. Then we trigger the
		 * reset. Waiting a maximum of 50ms, the kernel shouldn't crash
		 * because of this.
		 */
		while ((readl(reg_dbg1) & 0xf) == 0x8 && elapsed < max_wait) {
			udelay(100);
			elapsed += 100;
		}

		if (elapsed >= max_wait)
			dev_err(&mxs_chan->mxs_dma->pdev->dev,
					"Failed waiting for the DMA channel %d to leave state READ_FLUSH, trying to reset channel in READ_FLUSH state now\n",
					chan_id);

		writel(1 << (chan_id + BP_APBHX_CHANNEL_CTRL_RESET_CHANNEL),
			mxs_dma->base + HW_APBHX_CHANNEL_CTRL + STMP_OFFSET_REG_SET);
	}

	mxs_chan->status = DMA_COMPLETE;
}

static void mxs_dma_enable_chan(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	int chan_id = mxs_chan->chan.chan_id;

	/* set cmd_addr up */
	writel(mxs_chan->ccw_phys,
		mxs_dma->base + HW_APBHX_CHn_NXTCMDAR(mxs_dma, chan_id));

	/* write 1 to SEMA to kick off the channel */
	if (mxs_chan->flags & MXS_DMA_USE_SEMAPHORE &&
			mxs_chan->flags & MXS_DMA_SG_LOOP) {
		/* A cyclic DMA consists of at least 2 segments, so initialize
		 * the semaphore with 2 so we have enough time to add 1 to the
		 * semaphore if we need to */
		writel(2, mxs_dma->base + HW_APBHX_CHn_SEMA(mxs_dma, chan_id));
	} else {
		writel(1, mxs_dma->base + HW_APBHX_CHn_SEMA(mxs_dma, chan_id));
	}
	mxs_chan->reset = false;
}

static void mxs_dma_disable_chan(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);

	mxs_chan->status = DMA_COMPLETE;
}

static int mxs_dma_pause_chan(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	int chan_id = mxs_chan->chan.chan_id;

	/* freeze the channel */
	if (dma_is_apbh(mxs_dma) && apbh_is_old(mxs_dma))
		writel(1 << chan_id,
			mxs_dma->base + HW_APBHX_CTRL0 + STMP_OFFSET_REG_SET);
	else
		writel(1 << chan_id,
			mxs_dma->base + HW_APBHX_CHANNEL_CTRL + STMP_OFFSET_REG_SET);

	mxs_chan->status = DMA_PAUSED;
	return 0;
}

static int mxs_dma_resume_chan(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	int chan_id = mxs_chan->chan.chan_id;

	/* unfreeze the channel */
	if (dma_is_apbh(mxs_dma) && apbh_is_old(mxs_dma))
		writel(1 << chan_id,
			mxs_dma->base + HW_APBHX_CTRL0 + STMP_OFFSET_REG_CLR);
	else
		writel(1 << chan_id,
			mxs_dma->base + HW_APBHX_CHANNEL_CTRL + STMP_OFFSET_REG_CLR);

	mxs_chan->status = DMA_IN_PROGRESS;
	return 0;
}

static dma_cookie_t mxs_dma_tx_submit(struct dma_async_tx_descriptor *tx)
{
	return dma_cookie_assign(tx);
}

static void mxs_dma_tasklet(unsigned long data)
{
	struct mxs_dma_chan *mxs_chan = (struct mxs_dma_chan *) data;

	dmaengine_desc_get_callback_invoke(&mxs_chan->desc, NULL);
}

static int mxs_dma_irq_to_chan(struct mxs_dma_engine *mxs_dma, int irq)
{
	int i;

	for (i = 0; i != mxs_dma->nr_channels; ++i)
		if (mxs_dma->mxs_chans[i].chan_irq == irq)
			return i;

	return -EINVAL;
}

static irqreturn_t mxs_dma_int_handler(int irq, void *dev_id)
{
	struct mxs_dma_engine *mxs_dma = dev_id;
	struct mxs_dma_chan *mxs_chan;
	u32 completed;
	u32 err;
	int chan = mxs_dma_irq_to_chan(mxs_dma, irq);

	if (chan < 0)
		return IRQ_NONE;

	/* completion status */
	completed = readl(mxs_dma->base + HW_APBHX_CTRL1);
	completed = (completed >> chan) & 0x1;

	/* Clear interrupt */
	writel((1 << chan),
			mxs_dma->base + HW_APBHX_CTRL1 + STMP_OFFSET_REG_CLR);

	/* error status */
	err = readl(mxs_dma->base + HW_APBHX_CTRL2);
	err &= (1 << (MXS_DMA_CHANNELS + chan)) | (1 << chan);

	/*
	 * error status bit is in the upper 16 bits, error irq bit in the lower
	 * 16 bits. We transform it into a simpler error code:
	 * err: 0x00 = no error, 0x01 = TERMINATION, 0x02 = BUS_ERROR
	 */
	err = (err >> (MXS_DMA_CHANNELS + chan)) + (err >> chan);

	/* Clear error irq */
	writel((1 << chan),
			mxs_dma->base + HW_APBHX_CTRL2 + STMP_OFFSET_REG_CLR);

	/*
	 * When both completion and error of termination bits set at the
	 * same time, we do not take it as an error.  IOW, it only becomes
	 * an error we need to handle here in case of either it's a bus
	 * error or a termination error with no completion. 0x01 is termination
	 * error, so we can subtract err & completed to get the real error case.
	 */
	err -= err & completed;

	mxs_chan = &mxs_dma->mxs_chans[chan];

	if (err) {
		dev_dbg(mxs_dma->dma_device.dev,
			"%s: error in channel %d\n", __func__,
			chan);
		mxs_chan->status = DMA_ERROR;
		mxs_dma_reset_chan(&mxs_chan->chan);
	} else if (mxs_chan->status != DMA_COMPLETE) {
		if (mxs_chan->flags & MXS_DMA_SG_LOOP) {
			mxs_chan->status = DMA_IN_PROGRESS;
			if (mxs_chan->flags & MXS_DMA_USE_SEMAPHORE)
				writel(1, mxs_dma->base +
					HW_APBHX_CHn_SEMA(mxs_dma, chan));
		} else {
			mxs_chan->status = DMA_COMPLETE;
		}
	}

	if (mxs_chan->status == DMA_COMPLETE) {
		if (mxs_chan->reset)
			return IRQ_HANDLED;
		dma_cookie_complete(&mxs_chan->desc);
	}

	/* schedule tasklet on this channel */
	tasklet_schedule(&mxs_chan->tasklet);

	return IRQ_HANDLED;
}

static int mxs_dma_alloc_chan_resources(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	struct device *dev = &mxs_dma->pdev->dev;
	int ret;

	mxs_chan->ccw = dma_pool_zalloc(mxs_chan->ccw_pool,
				       GFP_ATOMIC,
				       &mxs_chan->ccw_phys);
	if (!mxs_chan->ccw) {
		ret = -ENOMEM;
		goto err_alloc;
	}

	ret = request_irq(mxs_chan->chan_irq, mxs_dma_int_handler,
			  0, "mxs-dma", mxs_dma);
	if (ret)
		goto err_irq;

	ret = pm_runtime_get_sync(dev);
	if (ret < 0) {
		dev_err(dev, "Failed to enable clock\n");
		goto err_clk;
	}

	mxs_dma_reset_chan(chan);

	dma_async_tx_descriptor_init(&mxs_chan->desc, chan);
	mxs_chan->desc.tx_submit = mxs_dma_tx_submit;

	/* the descriptor is ready */
	async_tx_ack(&mxs_chan->desc);

	return 0;

err_clk:
	free_irq(mxs_chan->chan_irq, mxs_dma);
err_irq:
	dma_pool_free(mxs_chan->ccw_pool, mxs_chan->ccw,
		      mxs_chan->ccw_phys);
err_alloc:
	return ret;
}

static void mxs_dma_free_chan_resources(struct dma_chan *chan)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	struct device *dev = &mxs_dma->pdev->dev;

	mxs_dma_disable_chan(chan);

	free_irq(mxs_chan->chan_irq, mxs_dma);

	dma_pool_free(mxs_chan->ccw_pool, mxs_chan->ccw,
		      mxs_chan->ccw_phys);

	pm_runtime_mark_last_busy(dev);
	pm_runtime_put_autosuspend(dev);

}

/*
 * How to use the flags for ->device_prep_slave_sg() :
 *    [1] If there is only one DMA command in the DMA chain, the code should be:
 *            ......
 *            ->device_prep_slave_sg(DMA_CTRL_ACK);
 *            ......
 *    [2] If there are two DMA commands in the DMA chain, the code should be
 *            ......
 *            ->device_prep_slave_sg(0);
 *            ......
 *            ->device_prep_slave_sg(DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
 *            ......
 *    [3] If there are more than two DMA commands in the DMA chain, the code
 *        should be:
 *            ......
 *            ->device_prep_slave_sg(0);                                // First
 *            ......
 *            ->device_prep_slave_sg(DMA_PREP_INTERRUPT [| DMA_CTRL_ACK]);
 *            ......
 *            ->device_prep_slave_sg(DMA_PREP_INTERRUPT | DMA_CTRL_ACK); // Last
 *            ......
 */
static struct dma_async_tx_descriptor *mxs_dma_prep_slave_sg(
		struct dma_chan *chan, struct scatterlist *sgl,
		unsigned int sg_len, enum dma_transfer_direction direction,
		unsigned long flags, void *context)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	struct mxs_dma_ccw *ccw;
	struct scatterlist *sg;
	u32 i, j;
	u32 *pio;
	bool append = flags & DMA_PREP_INTERRUPT;
	int idx = append ? mxs_chan->desc_count : 0;

	if (mxs_chan->status == DMA_IN_PROGRESS && !append)
		return NULL;

	if (sg_len + (append ? idx : 0) > NUM_CCW) {
		dev_err(mxs_dma->dma_device.dev,
				"maximum number of sg exceeded: %d > %d\n",
				sg_len, NUM_CCW);
		goto err_out;
	}

	mxs_chan->status = DMA_IN_PROGRESS;
	mxs_chan->flags = 0;

	/*
	 * If the sg is prepared with append flag set, the sg
	 * will be appended to the last prepared sg.
	 */
	if (append) {
		BUG_ON(idx < 1);
		ccw = &mxs_chan->ccw[idx - 1];
		ccw->next = mxs_chan->ccw_phys + sizeof(*ccw) * idx;
		ccw->bits |= CCW_CHAIN;
		ccw->bits &= ~CCW_IRQ;
		ccw->bits &= ~CCW_DEC_SEM;
	} else {
		idx = 0;
	}

	if (direction == DMA_TRANS_NONE) {
		ccw = &mxs_chan->ccw[idx++];
		pio = (u32 *) sgl;

		for (j = 0; j < sg_len;)
			ccw->pio_words[j++] = *pio++;

		ccw->bits = 0;
		ccw->bits |= CCW_IRQ;
		ccw->bits |= CCW_DEC_SEM;
		if (flags & DMA_CTRL_ACK)
			ccw->bits |= CCW_WAIT4END;
		ccw->bits |= CCW_HALT_ON_TERM;
		ccw->bits |= CCW_TERM_FLUSH;
		ccw->bits |= BF_CCW(sg_len, PIO_NUM);
		ccw->bits |= BF_CCW(MXS_DMA_CMD_NO_XFER, COMMAND);
	} else {
		for_each_sg(sgl, sg, sg_len, i) {
			if (sg_dma_len(sg) > MAX_XFER_BYTES) {
				dev_err(mxs_dma->dma_device.dev, "maximum bytes for sg entry exceeded: %d > %d\n",
						sg_dma_len(sg), MAX_XFER_BYTES);
				goto err_out;
			}

			ccw = &mxs_chan->ccw[idx++];

			ccw->next = mxs_chan->ccw_phys + sizeof(*ccw) * idx;
			ccw->bufaddr = sg->dma_address;
			ccw->xfer_bytes = sg_dma_len(sg);

			ccw->bits = 0;
			ccw->bits |= CCW_CHAIN;
			ccw->bits |= CCW_HALT_ON_TERM;
			ccw->bits |= CCW_TERM_FLUSH;
			ccw->bits |= BF_CCW(direction == DMA_DEV_TO_MEM ?
					MXS_DMA_CMD_WRITE : MXS_DMA_CMD_READ,
					COMMAND);

			if (i + 1 == sg_len) {
				ccw->bits &= ~CCW_CHAIN;
				ccw->bits |= CCW_IRQ;
				ccw->bits |= CCW_DEC_SEM;
				if (flags & DMA_CTRL_ACK)
					ccw->bits |= CCW_WAIT4END;
			}
		}
	}
	mxs_chan->desc_count = idx;

	return &mxs_chan->desc;

err_out:
	mxs_chan->status = DMA_ERROR;
	return NULL;
}

static struct dma_async_tx_descriptor *mxs_dma_prep_dma_cyclic(
		struct dma_chan *chan, dma_addr_t dma_addr, size_t buf_len,
		size_t period_len, enum dma_transfer_direction direction,
		unsigned long flags)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	u32 num_periods = buf_len / period_len;
	u32 i = 0, buf = 0;

	if (mxs_chan->status == DMA_IN_PROGRESS)
		return NULL;

	mxs_chan->status = DMA_IN_PROGRESS;
	mxs_chan->flags |= MXS_DMA_SG_LOOP;
	mxs_chan->flags |= MXS_DMA_USE_SEMAPHORE;

	if (num_periods > NUM_CCW) {
		dev_err(mxs_dma->dma_device.dev,
				"maximum number of sg exceeded: %d > %d\n",
				num_periods, NUM_CCW);
		goto err_out;
	}

	if (period_len > MAX_XFER_BYTES) {
		dev_err(mxs_dma->dma_device.dev,
				"maximum period size exceeded: %zu > %d\n",
				period_len, MAX_XFER_BYTES);
		goto err_out;
	}

	while (buf < buf_len) {
		struct mxs_dma_ccw *ccw = &mxs_chan->ccw[i];

		if (i + 1 == num_periods)
			ccw->next = mxs_chan->ccw_phys;
		else
			ccw->next = mxs_chan->ccw_phys + sizeof(*ccw) * (i + 1);

		ccw->bufaddr = dma_addr;
		ccw->xfer_bytes = period_len;

		ccw->bits = 0;
		ccw->bits |= CCW_CHAIN;
		ccw->bits |= CCW_IRQ;
		ccw->bits |= CCW_HALT_ON_TERM;
		ccw->bits |= CCW_TERM_FLUSH;
		ccw->bits |= CCW_DEC_SEM;
		ccw->bits |= BF_CCW(direction == DMA_DEV_TO_MEM ?
				MXS_DMA_CMD_WRITE : MXS_DMA_CMD_READ, COMMAND);

		dma_addr += period_len;
		buf += period_len;

		i++;
	}
	mxs_chan->desc_count = i;

	return &mxs_chan->desc;

err_out:
	mxs_chan->status = DMA_ERROR;
	return NULL;
}

static int mxs_dma_terminate_all(struct dma_chan *chan)
{
	mxs_dma_reset_chan(chan);
	mxs_dma_disable_chan(chan);

	return 0;
}

static enum dma_status mxs_dma_tx_status(struct dma_chan *chan,
			dma_cookie_t cookie, struct dma_tx_state *txstate)
{
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	u32 residue = 0;

	if (mxs_chan->status == DMA_IN_PROGRESS &&
			mxs_chan->flags & MXS_DMA_SG_LOOP) {
		struct mxs_dma_ccw *last_ccw;
		u32 bar;

		last_ccw = &mxs_chan->ccw[mxs_chan->desc_count - 1];
		residue = last_ccw->xfer_bytes + last_ccw->bufaddr;

		bar = readl(mxs_dma->base +
				HW_APBHX_CHn_BAR(mxs_dma, chan->chan_id));
		residue -= bar;
	}

	dma_set_tx_state(txstate, chan->completed_cookie, chan->cookie,
			residue);

	return mxs_chan->status;
}

static int mxs_dma_init_rpm(struct mxs_dma_engine *mxs_dma)
{
	struct device *dev = &mxs_dma->pdev->dev;

	pm_runtime_enable(dev);
	pm_runtime_set_autosuspend_delay(dev, MXS_DMA_RPM_TIMEOUT);
	pm_runtime_use_autosuspend(dev);

	return 0;
}

static int mxs_dma_init(struct mxs_dma_engine *mxs_dma)
{
	struct device *dev = &mxs_dma->pdev->dev;
	int ret;

	ret = mxs_dma_init_rpm(mxs_dma);
	if (ret)
		return ret;

	ret = pm_runtime_get_sync(dev);
	if (ret < 0) {
		dev_err(dev, "Failed to enable clock\n");
		return ret;
	}

	ret = stmp_reset_block(mxs_dma->base);
	if (ret)
		goto err_clk;

	/* enable apbh burst */
	if (dma_is_apbh(mxs_dma)) {
		writel(BM_APBH_CTRL0_APB_BURST_EN,
			mxs_dma->base + HW_APBHX_CTRL0 + STMP_OFFSET_REG_SET);
		writel(BM_APBH_CTRL0_APB_BURST8_EN,
			mxs_dma->base + HW_APBHX_CTRL0 + STMP_OFFSET_REG_SET);
	}

	/* enable irq for all the channels */
	writel(MXS_DMA_CHANNELS_MASK << MXS_DMA_CHANNELS,
		mxs_dma->base + HW_APBHX_CTRL1 + STMP_OFFSET_REG_SET);

	pm_runtime_mark_last_busy(dev);
	pm_runtime_put_autosuspend(dev);

err_clk:
	return ret;
}

struct mxs_dma_filter_param {
	struct device_node *of_node;
	unsigned int chan_id;
};

static bool mxs_dma_filter_fn(struct dma_chan *chan, void *fn_param)
{
	struct mxs_dma_filter_param *param = fn_param;
	struct mxs_dma_chan *mxs_chan = to_mxs_dma_chan(chan);
	struct mxs_dma_engine *mxs_dma = mxs_chan->mxs_dma;
	int chan_irq;

	if (strcmp(chan->device->dev->driver->name, "mxs-dma"))
		return false;

	if (!mxs_dma)
		return false;

	if (mxs_dma->dma_device.dev->of_node != param->of_node)
		return false;

	if (chan->chan_id != param->chan_id)
		return false;

	chan_irq = platform_get_irq(mxs_dma->pdev, param->chan_id);
	if (chan_irq < 0)
		return false;

	mxs_chan->chan_irq = chan_irq;

	return true;
}

static struct dma_chan *mxs_dma_xlate(struct of_phandle_args *dma_spec,
			       struct of_dma *ofdma)
{
	struct mxs_dma_engine *mxs_dma = ofdma->of_dma_data;
	dma_cap_mask_t mask = mxs_dma->dma_device.cap_mask;
	struct mxs_dma_filter_param param;

	if (dma_spec->args_count != 1)
		return NULL;

	param.of_node = ofdma->of_node;
	param.chan_id = dma_spec->args[0];

	if (param.chan_id >= mxs_dma->nr_channels)
		return NULL;

	return dma_request_channel(mask, mxs_dma_filter_fn, &param);
}

static int mxs_dma_probe(struct platform_device *pdev)
{
	struct device_node *np = pdev->dev.of_node;
	const struct platform_device_id *id_entry;
	const struct of_device_id *of_id;
	const struct mxs_dma_type *dma_type;
	struct mxs_dma_engine *mxs_dma;
	struct resource *iores;
	struct dma_pool *ccw_pool;
	int ret, i;

	mxs_dma = devm_kzalloc(&pdev->dev, sizeof(*mxs_dma), GFP_KERNEL);
	if (!mxs_dma)
		return -ENOMEM;

	ret = of_property_read_u32(np, "dma-channels", &mxs_dma->nr_channels);
	if (ret) {
		dev_err(&pdev->dev, "failed to read dma-channels\n");
		return ret;
	}

	of_id = of_match_device(mxs_dma_dt_ids, &pdev->dev);
	if (of_id)
		id_entry = of_id->data;
	else
		id_entry = platform_get_device_id(pdev);

	dma_type = (struct mxs_dma_type *)id_entry->driver_data;
	mxs_dma->type = dma_type->type;
	mxs_dma->dev_id = dma_type->id;

	iores = platform_get_resource(pdev, IORESOURCE_MEM, 0);
	mxs_dma->base = devm_ioremap_resource(&pdev->dev, iores);
	if (IS_ERR(mxs_dma->base))
		return PTR_ERR(mxs_dma->base);

	mxs_dma->clk = devm_clk_get(&pdev->dev, NULL);
	if (IS_ERR(mxs_dma->clk))
		return PTR_ERR(mxs_dma->clk);

	dma_cap_set(DMA_SLAVE, mxs_dma->dma_device.cap_mask);
	dma_cap_set(DMA_CYCLIC, mxs_dma->dma_device.cap_mask);

	INIT_LIST_HEAD(&mxs_dma->dma_device.channels);

	/* Initialize channel parameters */
	for (i = 0; i < MXS_DMA_CHANNELS; i++) {
		struct mxs_dma_chan *mxs_chan = &mxs_dma->mxs_chans[i];

		mxs_chan->mxs_dma = mxs_dma;
		mxs_chan->chan.device = &mxs_dma->dma_device;
		dma_cookie_init(&mxs_chan->chan);

		tasklet_init(&mxs_chan->tasklet, mxs_dma_tasklet,
			     (unsigned long) mxs_chan);

		/* Add the channel to mxs_chan list */
		list_add_tail(&mxs_chan->chan.device_node,
			&mxs_dma->dma_device.channels);
	}

	platform_set_drvdata(pdev, mxs_dma);
	mxs_dma->pdev = pdev;

	ret = mxs_dma_init(mxs_dma);
	if (ret)
		return ret;

	mxs_dma->dma_device.dev = &pdev->dev;
	dev_set_drvdata(&pdev->dev, mxs_dma);

	/* create the dma pool */
	ccw_pool = dma_pool_create("ccw_pool",
					     mxs_dma->dma_device.dev,
					     CCW_BLOCK_SIZE, 32, 0);

	for (i = 0; i < MXS_DMA_CHANNELS; i++) {
		struct mxs_dma_chan *mxs_chan = &mxs_dma->mxs_chans[i];
		mxs_chan->ccw_pool = ccw_pool;
	}

	/* mxs_dma gets 65535 bytes maximum sg size */
	mxs_dma->dma_device.dev->dma_parms = &mxs_dma->dma_parms;
	dma_set_max_seg_size(mxs_dma->dma_device.dev, MAX_XFER_BYTES);

	mxs_dma->dma_device.device_alloc_chan_resources = mxs_dma_alloc_chan_resources;
	mxs_dma->dma_device.device_free_chan_resources = mxs_dma_free_chan_resources;
	mxs_dma->dma_device.device_tx_status = mxs_dma_tx_status;
	mxs_dma->dma_device.device_prep_slave_sg = mxs_dma_prep_slave_sg;
	mxs_dma->dma_device.device_prep_dma_cyclic = mxs_dma_prep_dma_cyclic;
	mxs_dma->dma_device.device_pause = mxs_dma_pause_chan;
	mxs_dma->dma_device.device_resume = mxs_dma_resume_chan;
	mxs_dma->dma_device.device_terminate_all = mxs_dma_terminate_all;
	mxs_dma->dma_device.src_addr_widths = BIT(DMA_SLAVE_BUSWIDTH_4_BYTES);
	mxs_dma->dma_device.dst_addr_widths = BIT(DMA_SLAVE_BUSWIDTH_4_BYTES);
	mxs_dma->dma_device.directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
	mxs_dma->dma_device.residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
	mxs_dma->dma_device.device_issue_pending = mxs_dma_enable_chan;

	ret = dma_async_device_register(&mxs_dma->dma_device);
	if (ret) {
		dev_err(mxs_dma->dma_device.dev, "unable to register\n");
		return ret;
	}

	ret = of_dma_controller_register(np, mxs_dma_xlate, mxs_dma);
	if (ret) {
		dev_err(mxs_dma->dma_device.dev,
			"failed to register controller\n");
		dma_async_device_unregister(&mxs_dma->dma_device);
	}

	dev_info(mxs_dma->dma_device.dev, "initialized\n");

	return 0;
}

static int mxs_dma_remove(struct platform_device *pdev)
{
	struct mxs_dma_engine *mxs_dma = platform_get_drvdata(pdev);
	int i;

	dma_async_device_unregister(&mxs_dma->dma_device);
	dma_pool_destroy(mxs_dma->mxs_chans[0].ccw_pool);

	for (i = 0; i < MXS_DMA_CHANNELS; i++) {
		struct mxs_dma_chan *mxs_chan = &mxs_dma->mxs_chans[i];
		tasklet_kill(&mxs_chan->tasklet);
		mxs_chan->ccw_pool = NULL;
	}

	return 0;
}

static int mxs_dma_pm_suspend(struct device *dev)
{
	int ret;

	ret = pm_runtime_force_suspend(dev);

	return ret;
}

static int mxs_dma_pm_resume(struct device *dev)
{
	struct mxs_dma_engine *mxs_dma = dev_get_drvdata(dev);
	int ret;

	ret = mxs_dma_init(mxs_dma);
	if (ret)
		return ret;

	return 0;
}

int mxs_dma_runtime_suspend(struct device *dev)
{
	struct mxs_dma_engine *mxs_dma = dev_get_drvdata(dev);

	clk_disable_unprepare(mxs_dma->clk);

	return 0;
}

int mxs_dma_runtime_resume(struct device *dev)
{
	struct mxs_dma_engine *mxs_dma = dev_get_drvdata(dev);
	int ret;

	ret = clk_prepare_enable(mxs_dma->clk);
	if (ret) {
		dev_err(&mxs_dma->pdev->dev, "failed to enable the clock\n");
		return ret;
	}

	return 0;
}

static const struct dev_pm_ops mxs_dma_pm_ops = {
	SET_RUNTIME_PM_OPS(mxs_dma_runtime_suspend, mxs_dma_runtime_resume, NULL)
	SET_SYSTEM_SLEEP_PM_OPS(mxs_dma_pm_suspend, mxs_dma_pm_resume)
};

static struct platform_driver mxs_dma_driver = {
	.driver		= {
		.name	= "mxs-dma",
		.pm = &mxs_dma_pm_ops,
		.of_match_table = mxs_dma_dt_ids,
	},
	.id_table	= mxs_dma_ids,
	.remove		= mxs_dma_remove,
	.probe		= mxs_dma_probe,
};

module_platform_driver(mxs_dma_driver);