amba-pl08x.c 58.2 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
/*
 * Copyright (c) 2006 ARM Ltd.
 * Copyright (c) 2010 ST-Ericsson SA
 *
 * Author: Peter Pearse <peter.pearse@arm.com>
 * Author: Linus Walleij <linus.walleij@stericsson.com>
 *
 * 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.
 *
22 23
 * The full GNU General Public License is in this distribution in the file
 * called COPYING.
24 25
 *
 * Documentation: ARM DDI 0196G == PL080
26
 * Documentation: ARM DDI 0218E == PL081
27
 *
28 29
 * PL080 & PL081 both have 16 sets of DMA signals that can be routed to any
 * channel.
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
 *
 * The PL080 has 8 channels available for simultaneous use, and the PL081
 * has only two channels. So on these DMA controllers the number of channels
 * and the number of incoming DMA signals are two totally different things.
 * It is usually not possible to theoretically handle all physical signals,
 * so a multiplexing scheme with possible denial of use is necessary.
 *
 * The PL080 has a dual bus master, PL081 has a single master.
 *
 * Memory to peripheral transfer may be visualized as
 *	Get data from memory to DMAC
 *	Until no data left
 *		On burst request from peripheral
 *			Destination burst from DMAC to peripheral
 *			Clear burst request
 *	Raise terminal count interrupt
 *
 * For peripherals with a FIFO:
 * Source      burst size == half the depth of the peripheral FIFO
 * Destination burst size == the depth of the peripheral FIFO
 *
 * (Bursts are irrelevant for mem to mem transfers - there are no burst
 * signals, the DMA controller will simply facilitate its AHB master.)
 *
 * ASSUMES default (little) endianness for DMA transfers
 *
56 57 58 59 60 61 62 63 64 65 66 67 68
 * The PL08x has two flow control settings:
 *  - DMAC flow control: the transfer size defines the number of transfers
 *    which occur for the current LLI entry, and the DMAC raises TC at the
 *    end of every LLI entry.  Observed behaviour shows the DMAC listening
 *    to both the BREQ and SREQ signals (contrary to documented),
 *    transferring data if either is active.  The LBREQ and LSREQ signals
 *    are ignored.
 *
 *  - Peripheral flow control: the transfer size is ignored (and should be
 *    zero).  The data is transferred from the current LLI entry, until
 *    after the final transfer signalled by LBREQ or LSREQ.  The DMAC
 *    will then move to the next LLI entry.
 *
69 70 71
 * Global TODO:
 * - Break out common code from arch/arm/mach-s3c64xx and share
 */
72
#include <linux/amba/bus.h>
73 74
#include <linux/amba/pl08x.h>
#include <linux/debugfs.h>
75 76 77 78
#include <linux/delay.h>
#include <linux/device.h>
#include <linux/dmaengine.h>
#include <linux/dmapool.h>
V
Vinod Koul 已提交
79
#include <linux/dma-mapping.h>
80 81 82
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/module.h>
83
#include <linux/pm_runtime.h>
84
#include <linux/seq_file.h>
85
#include <linux/slab.h>
86 87
#include <asm/hardware/pl080.h>

88 89
#include "dmaengine.h"

90 91
#define DRIVER_NAME	"pl08xdmac"

92
static struct amba_driver pl08x_amba_driver;
93
struct pl08x_driver_data;
94

95
/**
96
 * struct vendor_data - vendor-specific config parameters for PL08x derivatives
97
 * @channels: the number of channels available in this variant
98
 * @dualmaster: whether this version supports dual AHB masters or not.
99 100 101
 * @nomadik: whether the channels have Nomadik security extension bits
 *	that need to be checked for permission before use and some registers are
 *	missing
102 103 104 105
 */
struct vendor_data {
	u8 channels;
	bool dualmaster;
106
	bool nomadik;
107 108 109 110
};

/*
 * PL08X private data structures
111
 * An LLI struct - see PL08x TRM.  Note that next uses bit[0] as a bus bit,
112 113
 * start & end do not - their bus bit info is in cctl.  Also note that these
 * are fixed 32-bit quantities.
114
 */
115
struct pl08x_lli {
116 117
	u32 src;
	u32 dst;
118
	u32 lli;
119 120 121
	u32 cctl;
};

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
/**
 * struct pl08x_bus_data - information of source or destination
 * busses for a transfer
 * @addr: current address
 * @maxwidth: the maximum width of a transfer on this bus
 * @buswidth: the width of this bus in bytes: 1, 2 or 4
 */
struct pl08x_bus_data {
	dma_addr_t addr;
	u8 maxwidth;
	u8 buswidth;
};

/**
 * struct pl08x_phy_chan - holder for the physical channels
 * @id: physical index to this channel
 * @lock: a lock to use when altering an instance of this struct
 * @signal: the physical signal (aka channel) serving this physical channel
 * right now
 * @serving: the virtual channel currently being served by this physical
 * channel
 */
struct pl08x_phy_chan {
	unsigned int id;
	void __iomem *base;
	spinlock_t lock;
	int signal;
	struct pl08x_dma_chan *serving;
};

/**
 * struct pl08x_sg - structure containing data per sg
 * @src_addr: src address of sg
 * @dst_addr: dst address of sg
 * @len: transfer len in bytes
 * @node: node for txd's dsg_list
 */
struct pl08x_sg {
	dma_addr_t src_addr;
	dma_addr_t dst_addr;
	size_t len;
	struct list_head node;
};

/**
 * struct pl08x_txd - wrapper for struct dma_async_tx_descriptor
 * @tx: async tx descriptor
 * @node: node for txd list for channels
 * @dsg_list: list of children sg's
 * @direction: direction of transfer
 * @llis_bus: DMA memory address (physical) start for the LLIs
 * @llis_va: virtual memory address start for the LLIs
 * @cctl: control reg values for current txd
 * @ccfg: config reg values for current txd
 */
struct pl08x_txd {
	struct dma_async_tx_descriptor tx;
	struct list_head node;
	struct list_head dsg_list;
	enum dma_transfer_direction direction;
	dma_addr_t llis_bus;
	struct pl08x_lli *llis_va;
	/* Default cctl value for LLIs */
	u32 cctl;
	/*
	 * Settings to be put into the physical channel when we
	 * trigger this txd.  Other registers are in llis_va[0].
	 */
	u32 ccfg;
};

/**
 * struct pl08x_dma_chan_state - holds the PL08x specific virtual channel
 * states
 * @PL08X_CHAN_IDLE: the channel is idle
 * @PL08X_CHAN_RUNNING: the channel has allocated a physical transport
 * channel and is running a transfer on it
 * @PL08X_CHAN_PAUSED: the channel has allocated a physical transport
 * channel, but the transfer is currently paused
 * @PL08X_CHAN_WAITING: the channel is waiting for a physical transport
 * channel to become available (only pertains to memcpy channels)
 */
enum pl08x_dma_chan_state {
	PL08X_CHAN_IDLE,
	PL08X_CHAN_RUNNING,
	PL08X_CHAN_PAUSED,
	PL08X_CHAN_WAITING,
};

/**
 * struct pl08x_dma_chan - this structure wraps a DMA ENGINE channel
 * @chan: wrappped abstract channel
 * @phychan: the physical channel utilized by this channel, if there is one
 * @phychan_hold: if non-zero, hold on to the physical channel even if we
 * have no pending entries
 * @tasklet: tasklet scheduled by the IRQ to handle actual work etc
 * @name: name of channel
 * @cd: channel platform data
 * @runtime_addr: address for RX/TX according to the runtime config
 * @runtime_direction: current direction of this channel according to
 * runtime config
 * @pend_list: queued transactions pending on this channel
 * @at: active transaction on this channel
 * @lock: a lock for this channel data
 * @host: a pointer to the host (internal use)
 * @state: whether the channel is idle, paused, running etc
 * @slave: whether this channel is a device (slave) or for memcpy
 * @waiting: a TX descriptor on this channel which is waiting for a physical
 * channel to become available
 */
struct pl08x_dma_chan {
	struct dma_chan chan;
	struct pl08x_phy_chan *phychan;
	int phychan_hold;
	struct tasklet_struct tasklet;
237
	const char *name;
238
	const struct pl08x_channel_data *cd;
239
	struct dma_slave_config cfg;
240 241 242 243 244 245 246 247 248 249 250 251
	u32 src_cctl;
	u32 dst_cctl;
	enum dma_transfer_direction runtime_direction;
	struct list_head pend_list;
	struct pl08x_txd *at;
	spinlock_t lock;
	struct pl08x_driver_data *host;
	enum pl08x_dma_chan_state state;
	bool slave;
	struct pl08x_txd *waiting;
};

252 253 254 255 256 257 258 259 260 261 262
/**
 * struct pl08x_driver_data - the local state holder for the PL08x
 * @slave: slave engine for this instance
 * @memcpy: memcpy engine for this instance
 * @base: virtual memory base (remapped) for the PL08x
 * @adev: the corresponding AMBA (PrimeCell) bus entry
 * @vd: vendor data for this PL08x variant
 * @pd: platform data passed in from the platform/machine
 * @phy_chans: array of data for the physical channels
 * @pool: a pool for the LLI descriptors
 * @pool_ctr: counter of LLIs in the pool
263 264
 * @lli_buses: bitmask to or in to LLI pointer selecting AHB port for LLI
 * fetches
265
 * @mem_buses: set to indicate memory transfers on AHB2.
266 267 268 269 270 271 272
 * @lock: a spinlock for this struct
 */
struct pl08x_driver_data {
	struct dma_device slave;
	struct dma_device memcpy;
	void __iomem *base;
	struct amba_device *adev;
273
	const struct vendor_data *vd;
274 275 276 277
	struct pl08x_platform_data *pd;
	struct pl08x_phy_chan *phy_chans;
	struct dma_pool *pool;
	int pool_ctr;
278 279
	u8 lli_buses;
	u8 mem_buses;
280 281 282 283 284 285 286 287 288
};

/*
 * PL08X specific defines
 */

/* Size (bytes) of each LLI buffer allocated for one transfer */
# define PL08X_LLI_TSFR_SIZE	0x2000

289
/* Maximum times we call dma_pool_alloc on this pool without freeing */
290
#define MAX_NUM_TSFR_LLIS	(PL08X_LLI_TSFR_SIZE/sizeof(struct pl08x_lli))
291 292 293 294 295 296 297
#define PL08X_ALIGN		8

static inline struct pl08x_dma_chan *to_pl08x_chan(struct dma_chan *chan)
{
	return container_of(chan, struct pl08x_dma_chan, chan);
}

298 299 300 301 302
static inline struct pl08x_txd *to_pl08x_txd(struct dma_async_tx_descriptor *tx)
{
	return container_of(tx, struct pl08x_txd, tx);
}

303 304 305 306 307 308 309 310 311 312 313 314 315 316 317
/*
 * Physical channel handling
 */

/* Whether a certain channel is busy or not */
static int pl08x_phy_channel_busy(struct pl08x_phy_chan *ch)
{
	unsigned int val;

	val = readl(ch->base + PL080_CH_CONFIG);
	return val & PL080_CONFIG_ACTIVE;
}

/*
 * Set the initial DMA register values i.e. those for the first LLI
318
 * The next LLI pointer and the configuration interrupt bit have
319 320
 * been set when the LLIs were constructed.  Poke them into the hardware
 * and start the transfer.
321
 */
322 323
static void pl08x_start_txd(struct pl08x_dma_chan *plchan,
	struct pl08x_txd *txd)
324
{
325
	struct pl08x_driver_data *pl08x = plchan->host;
326
	struct pl08x_phy_chan *phychan = plchan->phychan;
327
	struct pl08x_lli *lli = &txd->llis_va[0];
328
	u32 val;
329 330

	plchan->at = txd;
331

332 333 334
	/* Wait for channel inactive */
	while (pl08x_phy_channel_busy(phychan))
		cpu_relax();
335

336 337
	dev_vdbg(&pl08x->adev->dev,
		"WRITE channel %d: csrc=0x%08x, cdst=0x%08x, "
338 339
		"clli=0x%08x, cctl=0x%08x, ccfg=0x%08x\n",
		phychan->id, lli->src, lli->dst, lli->lli, lli->cctl,
340
		txd->ccfg);
341 342 343 344 345

	writel(lli->src, phychan->base + PL080_CH_SRC_ADDR);
	writel(lli->dst, phychan->base + PL080_CH_DST_ADDR);
	writel(lli->lli, phychan->base + PL080_CH_LLI);
	writel(lli->cctl, phychan->base + PL080_CH_CONTROL);
346
	writel(txd->ccfg, phychan->base + PL080_CH_CONFIG);
347 348 349 350

	/* Enable the DMA channel */
	/* Do not access config register until channel shows as disabled */
	while (readl(pl08x->base + PL080_EN_CHAN) & (1 << phychan->id))
351
		cpu_relax();
352

353 354
	/* Do not access config register until channel shows as inactive */
	val = readl(phychan->base + PL080_CH_CONFIG);
355
	while ((val & PL080_CONFIG_ACTIVE) || (val & PL080_CONFIG_ENABLE))
356
		val = readl(phychan->base + PL080_CH_CONFIG);
357

358
	writel(val | PL080_CONFIG_ENABLE, phychan->base + PL080_CH_CONFIG);
359 360 361
}

/*
362
 * Pause the channel by setting the HALT bit.
363
 *
364 365 366
 * For M->P transfers, pause the DMAC first and then stop the peripheral -
 * the FIFO can only drain if the peripheral is still requesting data.
 * (note: this can still timeout if the DMAC FIFO never drains of data.)
367
 *
368 369
 * For P->M transfers, disable the peripheral first to stop it filling
 * the DMAC FIFO, and then pause the DMAC.
370 371 372 373
 */
static void pl08x_pause_phy_chan(struct pl08x_phy_chan *ch)
{
	u32 val;
374
	int timeout;
375 376 377 378 379 380 381

	/* Set the HALT bit and wait for the FIFO to drain */
	val = readl(ch->base + PL080_CH_CONFIG);
	val |= PL080_CONFIG_HALT;
	writel(val, ch->base + PL080_CH_CONFIG);

	/* Wait for channel inactive */
382 383 384 385 386 387 388
	for (timeout = 1000; timeout; timeout--) {
		if (!pl08x_phy_channel_busy(ch))
			break;
		udelay(1);
	}
	if (pl08x_phy_channel_busy(ch))
		pr_err("pl08x: channel%u timeout waiting for pause\n", ch->id);
389 390 391 392 393 394 395 396 397 398 399 400
}

static void pl08x_resume_phy_chan(struct pl08x_phy_chan *ch)
{
	u32 val;

	/* Clear the HALT bit */
	val = readl(ch->base + PL080_CH_CONFIG);
	val &= ~PL080_CONFIG_HALT;
	writel(val, ch->base + PL080_CH_CONFIG);
}

401 402 403 404 405 406 407 408
/*
 * pl08x_terminate_phy_chan() stops the channel, clears the FIFO and
 * clears any pending interrupt status.  This should not be used for
 * an on-going transfer, but as a method of shutting down a channel
 * (eg, when it's no longer used) or terminating a transfer.
 */
static void pl08x_terminate_phy_chan(struct pl08x_driver_data *pl08x,
	struct pl08x_phy_chan *ch)
409
{
410
	u32 val = readl(ch->base + PL080_CH_CONFIG);
411

412 413
	val &= ~(PL080_CONFIG_ENABLE | PL080_CONFIG_ERR_IRQ_MASK |
	         PL080_CONFIG_TC_IRQ_MASK);
414 415

	writel(val, ch->base + PL080_CH_CONFIG);
416 417 418

	writel(1 << ch->id, pl08x->base + PL080_ERR_CLEAR);
	writel(1 << ch->id, pl08x->base + PL080_TC_CLEAR);
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
}

static inline u32 get_bytes_in_cctl(u32 cctl)
{
	/* The source width defines the number of bytes */
	u32 bytes = cctl & PL080_CONTROL_TRANSFER_SIZE_MASK;

	switch (cctl >> PL080_CONTROL_SWIDTH_SHIFT) {
	case PL080_WIDTH_8BIT:
		break;
	case PL080_WIDTH_16BIT:
		bytes *= 2;
		break;
	case PL080_WIDTH_32BIT:
		bytes *= 4;
		break;
	}
	return bytes;
}

/* The channel should be paused when calling this */
static u32 pl08x_getbytes_chan(struct pl08x_dma_chan *plchan)
{
	struct pl08x_phy_chan *ch;
	struct pl08x_txd *txd;
	unsigned long flags;
445
	size_t bytes = 0;
446 447 448 449 450 451

	spin_lock_irqsave(&plchan->lock, flags);
	ch = plchan->phychan;
	txd = plchan->at;

	/*
452 453
	 * Follow the LLIs to get the number of remaining
	 * bytes in the currently active transaction.
454 455
	 */
	if (ch && txd) {
456
		u32 clli = readl(ch->base + PL080_CH_LLI) & ~PL080_LLI_LM_AHB2;
457

458
		/* First get the remaining bytes in the active transfer */
459 460 461
		bytes = get_bytes_in_cctl(readl(ch->base + PL080_CH_CONTROL));

		if (clli) {
462 463 464 465 466 467
			struct pl08x_lli *llis_va = txd->llis_va;
			dma_addr_t llis_bus = txd->llis_bus;
			int index;

			BUG_ON(clli < llis_bus || clli >= llis_bus +
				sizeof(struct pl08x_lli) * MAX_NUM_TSFR_LLIS);
468

469 470 471 472 473 474 475 476
			/*
			 * Locate the next LLI - as this is an array,
			 * it's simple maths to find.
			 */
			index = (clli - llis_bus) / sizeof(struct pl08x_lli);

			for (; index < MAX_NUM_TSFR_LLIS; index++) {
				bytes += get_bytes_in_cctl(llis_va[index].cctl);
477 478

				/*
479
				 * A LLI pointer of 0 terminates the LLI list
480
				 */
481 482
				if (!llis_va[index].lli)
					break;
483 484 485 486 487
			}
		}
	}

	/* Sum up all queued transactions */
488
	if (!list_empty(&plchan->pend_list)) {
489
		struct pl08x_txd *txdi;
490
		list_for_each_entry(txdi, &plchan->pend_list, node) {
491 492 493
			struct pl08x_sg *dsg;
			list_for_each_entry(dsg, &txd->dsg_list, node)
				bytes += dsg->len;
494 495 496 497 498 499 500 501 502 503
		}
	}

	spin_unlock_irqrestore(&plchan->lock, flags);

	return bytes;
}

/*
 * Allocate a physical channel for a virtual channel
504 505 506 507
 *
 * Try to locate a physical channel to be used for this transfer. If all
 * are taken return NULL and the requester will have to cope by using
 * some fallback PIO mode or retrying later.
508 509 510 511 512 513 514 515 516 517 518 519 520 521
 */
static struct pl08x_phy_chan *
pl08x_get_phy_channel(struct pl08x_driver_data *pl08x,
		      struct pl08x_dma_chan *virt_chan)
{
	struct pl08x_phy_chan *ch = NULL;
	unsigned long flags;
	int i;

	for (i = 0; i < pl08x->vd->channels; i++) {
		ch = &pl08x->phy_chans[i];

		spin_lock_irqsave(&ch->lock, flags);

522
		if (!ch->locked && !ch->serving) {
523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544
			ch->serving = virt_chan;
			ch->signal = -1;
			spin_unlock_irqrestore(&ch->lock, flags);
			break;
		}

		spin_unlock_irqrestore(&ch->lock, flags);
	}

	if (i == pl08x->vd->channels) {
		/* No physical channel available, cope with it */
		return NULL;
	}

	return ch;
}

static inline void pl08x_put_phy_channel(struct pl08x_driver_data *pl08x,
					 struct pl08x_phy_chan *ch)
{
	unsigned long flags;

545 546
	spin_lock_irqsave(&ch->lock, flags);

547
	/* Stop the channel and clear its interrupts */
548
	pl08x_terminate_phy_chan(pl08x, ch);
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

	/* Mark it as free */
	ch->serving = NULL;
	spin_unlock_irqrestore(&ch->lock, flags);
}

/*
 * LLI handling
 */

static inline unsigned int pl08x_get_bytes_for_cctl(unsigned int coded)
{
	switch (coded) {
	case PL080_WIDTH_8BIT:
		return 1;
	case PL080_WIDTH_16BIT:
		return 2;
	case PL080_WIDTH_32BIT:
		return 4;
	default:
		break;
	}
	BUG();
	return 0;
}

static inline u32 pl08x_cctl_bits(u32 cctl, u8 srcwidth, u8 dstwidth,
576
				  size_t tsize)
577 578 579
{
	u32 retbits = cctl;

580
	/* Remove all src, dst and transfer size bits */
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
	retbits &= ~PL080_CONTROL_DWIDTH_MASK;
	retbits &= ~PL080_CONTROL_SWIDTH_MASK;
	retbits &= ~PL080_CONTROL_TRANSFER_SIZE_MASK;

	/* Then set the bits according to the parameters */
	switch (srcwidth) {
	case 1:
		retbits |= PL080_WIDTH_8BIT << PL080_CONTROL_SWIDTH_SHIFT;
		break;
	case 2:
		retbits |= PL080_WIDTH_16BIT << PL080_CONTROL_SWIDTH_SHIFT;
		break;
	case 4:
		retbits |= PL080_WIDTH_32BIT << PL080_CONTROL_SWIDTH_SHIFT;
		break;
	default:
		BUG();
		break;
	}

	switch (dstwidth) {
	case 1:
		retbits |= PL080_WIDTH_8BIT << PL080_CONTROL_DWIDTH_SHIFT;
		break;
	case 2:
		retbits |= PL080_WIDTH_16BIT << PL080_CONTROL_DWIDTH_SHIFT;
		break;
	case 4:
		retbits |= PL080_WIDTH_32BIT << PL080_CONTROL_DWIDTH_SHIFT;
		break;
	default:
		BUG();
		break;
	}

	retbits |= tsize << PL080_CONTROL_TRANSFER_SIZE_SHIFT;
	return retbits;
}

620 621 622 623 624
struct pl08x_lli_build_data {
	struct pl08x_txd *txd;
	struct pl08x_bus_data srcbus;
	struct pl08x_bus_data dstbus;
	size_t remainder;
625
	u32 lli_bus;
626 627
};

628
/*
629 630 631 632 633 634
 * Autoselect a master bus to use for the transfer. Slave will be the chosen as
 * victim in case src & dest are not similarly aligned. i.e. If after aligning
 * masters address with width requirements of transfer (by sending few byte by
 * byte data), slave is still not aligned, then its width will be reduced to
 * BYTE.
 * - prefers the destination bus if both available
635
 * - prefers bus with fixed address (i.e. peripheral)
636
 */
637 638
static void pl08x_choose_master_bus(struct pl08x_lli_build_data *bd,
	struct pl08x_bus_data **mbus, struct pl08x_bus_data **sbus, u32 cctl)
639 640
{
	if (!(cctl & PL080_CONTROL_DST_INCR)) {
641 642
		*mbus = &bd->dstbus;
		*sbus = &bd->srcbus;
643 644 645
	} else if (!(cctl & PL080_CONTROL_SRC_INCR)) {
		*mbus = &bd->srcbus;
		*sbus = &bd->dstbus;
646
	} else {
647
		if (bd->dstbus.buswidth >= bd->srcbus.buswidth) {
648 649
			*mbus = &bd->dstbus;
			*sbus = &bd->srcbus;
650
		} else {
651 652
			*mbus = &bd->srcbus;
			*sbus = &bd->dstbus;
653 654 655 656 657
		}
	}
}

/*
658
 * Fills in one LLI for a certain transfer descriptor and advance the counter
659
 */
660 661
static void pl08x_fill_lli_for_desc(struct pl08x_lli_build_data *bd,
	int num_llis, int len, u32 cctl)
662
{
663 664
	struct pl08x_lli *llis_va = bd->txd->llis_va;
	dma_addr_t llis_bus = bd->txd->llis_bus;
665 666 667

	BUG_ON(num_llis >= MAX_NUM_TSFR_LLIS);

668
	llis_va[num_llis].cctl = cctl;
669 670
	llis_va[num_llis].src = bd->srcbus.addr;
	llis_va[num_llis].dst = bd->dstbus.addr;
671 672
	llis_va[num_llis].lli = llis_bus + (num_llis + 1) *
		sizeof(struct pl08x_lli);
673
	llis_va[num_llis].lli |= bd->lli_bus;
674 675

	if (cctl & PL080_CONTROL_SRC_INCR)
676
		bd->srcbus.addr += len;
677
	if (cctl & PL080_CONTROL_DST_INCR)
678
		bd->dstbus.addr += len;
679

680
	BUG_ON(bd->remainder < len);
681

682
	bd->remainder -= len;
683 684
}

685 686
static inline void prep_byte_width_lli(struct pl08x_lli_build_data *bd,
		u32 *cctl, u32 len, int num_llis, size_t *total_bytes)
687
{
688 689 690
	*cctl = pl08x_cctl_bits(*cctl, 1, 1, len);
	pl08x_fill_lli_for_desc(bd, num_llis, len, *cctl);
	(*total_bytes) += len;
691 692 693 694 695 696 697 698 699 700 701
}

/*
 * This fills in the table of LLIs for the transfer descriptor
 * Note that we assume we never have to change the burst sizes
 * Return 0 for error
 */
static int pl08x_fill_llis_for_desc(struct pl08x_driver_data *pl08x,
			      struct pl08x_txd *txd)
{
	struct pl08x_bus_data *mbus, *sbus;
702
	struct pl08x_lli_build_data bd;
703
	int num_llis = 0;
704
	u32 cctl, early_bytes = 0;
705
	size_t max_bytes_per_lli, total_bytes;
706
	struct pl08x_lli *llis_va;
707
	struct pl08x_sg *dsg;
708

709
	txd->llis_va = dma_pool_alloc(pl08x->pool, GFP_NOWAIT, &txd->llis_bus);
710 711 712 713 714 715 716
	if (!txd->llis_va) {
		dev_err(&pl08x->adev->dev, "%s no memory for llis\n", __func__);
		return 0;
	}

	pl08x->pool_ctr++;

717
	bd.txd = txd;
718
	bd.lli_bus = (pl08x->lli_buses & PL08X_AHB2) ? PL080_LLI_LM_AHB2 : 0;
719
	cctl = txd->cctl;
720

721
	/* Find maximum width of the source bus */
722
	bd.srcbus.maxwidth =
723 724 725 726
		pl08x_get_bytes_for_cctl((cctl & PL080_CONTROL_SWIDTH_MASK) >>
				       PL080_CONTROL_SWIDTH_SHIFT);

	/* Find maximum width of the destination bus */
727
	bd.dstbus.maxwidth =
728 729 730
		pl08x_get_bytes_for_cctl((cctl & PL080_CONTROL_DWIDTH_MASK) >>
				       PL080_CONTROL_DWIDTH_SHIFT);

731 732 733
	list_for_each_entry(dsg, &txd->dsg_list, node) {
		total_bytes = 0;
		cctl = txd->cctl;
734

735 736 737 738 739
		bd.srcbus.addr = dsg->src_addr;
		bd.dstbus.addr = dsg->dst_addr;
		bd.remainder = dsg->len;
		bd.srcbus.buswidth = bd.srcbus.maxwidth;
		bd.dstbus.buswidth = bd.dstbus.maxwidth;
740

741
		pl08x_choose_master_bus(&bd, &mbus, &sbus, cctl);
742

743 744 745 746 747 748 749 750 751
		dev_vdbg(&pl08x->adev->dev, "src=0x%08x%s/%u dst=0x%08x%s/%u len=%zu\n",
			bd.srcbus.addr, cctl & PL080_CONTROL_SRC_INCR ? "+" : "",
			bd.srcbus.buswidth,
			bd.dstbus.addr, cctl & PL080_CONTROL_DST_INCR ? "+" : "",
			bd.dstbus.buswidth,
			bd.remainder);
		dev_vdbg(&pl08x->adev->dev, "mbus=%s sbus=%s\n",
			mbus == &bd.srcbus ? "src" : "dst",
			sbus == &bd.srcbus ? "src" : "dst");
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
		/*
		 * Zero length is only allowed if all these requirements are
		 * met:
		 * - flow controller is peripheral.
		 * - src.addr is aligned to src.width
		 * - dst.addr is aligned to dst.width
		 *
		 * sg_len == 1 should be true, as there can be two cases here:
		 *
		 * - Memory addresses are contiguous and are not scattered.
		 *   Here, Only one sg will be passed by user driver, with
		 *   memory address and zero length. We pass this to controller
		 *   and after the transfer it will receive the last burst
		 *   request from peripheral and so transfer finishes.
		 *
		 * - Memory addresses are scattered and are not contiguous.
		 *   Here, Obviously as DMA controller doesn't know when a lli's
		 *   transfer gets over, it can't load next lli. So in this
		 *   case, there has to be an assumption that only one lli is
		 *   supported. Thus, we can't have scattered addresses.
		 */
		if (!bd.remainder) {
			u32 fc = (txd->ccfg & PL080_CONFIG_FLOW_CONTROL_MASK) >>
				PL080_CONFIG_FLOW_CONTROL_SHIFT;
			if (!((fc >= PL080_FLOW_SRC2DST_DST) &&
778
					(fc <= PL080_FLOW_SRC2DST_SRC))) {
779 780 781 782
				dev_err(&pl08x->adev->dev, "%s sg len can't be zero",
					__func__);
				return 0;
			}
783

784
			if ((bd.srcbus.addr % bd.srcbus.buswidth) ||
785
					(bd.dstbus.addr % bd.dstbus.buswidth)) {
786 787 788 789 790 791
				dev_err(&pl08x->adev->dev,
					"%s src & dst address must be aligned to src"
					" & dst width if peripheral is flow controller",
					__func__);
				return 0;
			}
792

793 794 795 796 797
			cctl = pl08x_cctl_bits(cctl, bd.srcbus.buswidth,
					bd.dstbus.buswidth, 0);
			pl08x_fill_lli_for_desc(&bd, num_llis++, 0, cctl);
			break;
		}
798 799

		/*
800 801 802
		 * Send byte by byte for following cases
		 * - Less than a bus width available
		 * - until master bus is aligned
803
		 */
804 805 806 807 808 809 810 811
		if (bd.remainder < mbus->buswidth)
			early_bytes = bd.remainder;
		else if ((mbus->addr) % (mbus->buswidth)) {
			early_bytes = mbus->buswidth - (mbus->addr) %
				(mbus->buswidth);
			if ((bd.remainder - early_bytes) < mbus->buswidth)
				early_bytes = bd.remainder;
		}
812

813 814 815 816 817 818
		if (early_bytes) {
			dev_vdbg(&pl08x->adev->dev,
				"%s byte width LLIs (remain 0x%08x)\n",
				__func__, bd.remainder);
			prep_byte_width_lli(&bd, &cctl, early_bytes, num_llis++,
				&total_bytes);
819 820
		}

821 822 823 824 825 826 827 828 829
		if (bd.remainder) {
			/*
			 * Master now aligned
			 * - if slave is not then we must set its width down
			 */
			if (sbus->addr % sbus->buswidth) {
				dev_dbg(&pl08x->adev->dev,
					"%s set down bus width to one byte\n",
					__func__);
830

831 832
				sbus->buswidth = 1;
			}
833 834

			/*
835 836
			 * Bytes transferred = tsize * src width, not
			 * MIN(buswidths)
837
			 */
838 839 840 841 842
			max_bytes_per_lli = bd.srcbus.buswidth *
				PL080_CONTROL_TRANSFER_SIZE_MASK;
			dev_vdbg(&pl08x->adev->dev,
				"%s max bytes per lli = %zu\n",
				__func__, max_bytes_per_lli);
843 844

			/*
845 846
			 * Make largest possible LLIs until less than one bus
			 * width left
847
			 */
848 849
			while (bd.remainder > (mbus->buswidth - 1)) {
				size_t lli_len, tsize, width;
850

851 852 853 854 855
				/*
				 * If enough left try to send max possible,
				 * otherwise try to send the remainder
				 */
				lli_len = min(bd.remainder, max_bytes_per_lli);
856

857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872
				/*
				 * Check against maximum bus alignment:
				 * Calculate actual transfer size in relation to
				 * bus width an get a maximum remainder of the
				 * highest bus width - 1
				 */
				width = max(mbus->buswidth, sbus->buswidth);
				lli_len = (lli_len / width) * width;
				tsize = lli_len / bd.srcbus.buswidth;

				dev_vdbg(&pl08x->adev->dev,
					"%s fill lli with single lli chunk of "
					"size 0x%08zx (remainder 0x%08zx)\n",
					__func__, lli_len, bd.remainder);

				cctl = pl08x_cctl_bits(cctl, bd.srcbus.buswidth,
873
					bd.dstbus.buswidth, tsize);
874 875 876 877
				pl08x_fill_lli_for_desc(&bd, num_llis++,
						lli_len, cctl);
				total_bytes += lli_len;
			}
878

879 880 881 882 883 884 885 886 887 888
			/*
			 * Send any odd bytes
			 */
			if (bd.remainder) {
				dev_vdbg(&pl08x->adev->dev,
					"%s align with boundary, send odd bytes (remain %zu)\n",
					__func__, bd.remainder);
				prep_byte_width_lli(&bd, &cctl, bd.remainder,
						num_llis++, &total_bytes);
			}
889
		}
890

891 892 893 894 895 896
		if (total_bytes != dsg->len) {
			dev_err(&pl08x->adev->dev,
				"%s size of encoded lli:s don't match total txd, transferred 0x%08zx from size 0x%08zx\n",
				__func__, total_bytes, dsg->len);
			return 0;
		}
897

898 899 900 901 902 903
		if (num_llis >= MAX_NUM_TSFR_LLIS) {
			dev_err(&pl08x->adev->dev,
				"%s need to increase MAX_NUM_TSFR_LLIS from 0x%08x\n",
				__func__, (u32) MAX_NUM_TSFR_LLIS);
			return 0;
		}
904
	}
905 906

	llis_va = txd->llis_va;
907
	/* The final LLI terminates the LLI. */
908
	llis_va[num_llis - 1].lli = 0;
909
	/* The final LLI element shall also fire an interrupt. */
910
	llis_va[num_llis - 1].cctl |= PL080_CONTROL_TC_IRQ_EN;
911 912 913 914 915

#ifdef VERBOSE_DEBUG
	{
		int i;

916 917 918
		dev_vdbg(&pl08x->adev->dev,
			 "%-3s %-9s  %-10s %-10s %-10s %s\n",
			 "lli", "", "csrc", "cdst", "clli", "cctl");
919 920
		for (i = 0; i < num_llis; i++) {
			dev_vdbg(&pl08x->adev->dev,
921 922 923
				 "%3d @%p: 0x%08x 0x%08x 0x%08x 0x%08x\n",
				 i, &llis_va[i], llis_va[i].src,
				 llis_va[i].dst, llis_va[i].lli, llis_va[i].cctl
924 925 926 927 928 929 930 931 932 933 934 935
				);
		}
	}
#endif

	return num_llis;
}

/* You should call this with the struct pl08x lock held */
static void pl08x_free_txd(struct pl08x_driver_data *pl08x,
			   struct pl08x_txd *txd)
{
936 937
	struct pl08x_sg *dsg, *_dsg;

938
	/* Free the LLI */
939 940
	if (txd->llis_va)
		dma_pool_free(pl08x->pool, txd->llis_va, txd->llis_bus);
941 942 943

	pl08x->pool_ctr--;

944 945 946 947 948
	list_for_each_entry_safe(dsg, _dsg, &txd->dsg_list, node) {
		list_del(&dsg->node);
		kfree(dsg);
	}

949 950 951 952 953 954 955 956 957
	kfree(txd);
}

static void pl08x_free_txd_list(struct pl08x_driver_data *pl08x,
				struct pl08x_dma_chan *plchan)
{
	struct pl08x_txd *txdi = NULL;
	struct pl08x_txd *next;

958
	if (!list_empty(&plchan->pend_list)) {
959
		list_for_each_entry_safe(txdi,
960
					 next, &plchan->pend_list, node) {
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
			list_del(&txdi->node);
			pl08x_free_txd(pl08x, txdi);
		}
	}
}

/*
 * The DMA ENGINE API
 */
static int pl08x_alloc_chan_resources(struct dma_chan *chan)
{
	return 0;
}

static void pl08x_free_chan_resources(struct dma_chan *chan)
{
}

/*
 * This should be called with the channel plchan->lock held
 */
static int prep_phy_channel(struct pl08x_dma_chan *plchan,
			    struct pl08x_txd *txd)
{
	struct pl08x_driver_data *pl08x = plchan->host;
	struct pl08x_phy_chan *ch;
	int ret;

	/* Check if we already have a channel */
990 991 992 993
	if (plchan->phychan) {
		ch = plchan->phychan;
		goto got_channel;
	}
994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006

	ch = pl08x_get_phy_channel(pl08x, plchan);
	if (!ch) {
		/* No physical channel available, cope with it */
		dev_dbg(&pl08x->adev->dev, "no physical channel available for xfer on %s\n", plchan->name);
		return -EBUSY;
	}

	/*
	 * OK we have a physical channel: for memcpy() this is all we
	 * need, but for slaves the physical signals may be muxed!
	 * Can the platform allow us to use this channel?
	 */
1007
	if (plchan->slave && pl08x->pd->get_signal) {
1008
		ret = pl08x->pd->get_signal(plchan->cd);
1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019
		if (ret < 0) {
			dev_dbg(&pl08x->adev->dev,
				"unable to use physical channel %d for transfer on %s due to platform restrictions\n",
				ch->id, plchan->name);
			/* Release physical channel & return */
			pl08x_put_phy_channel(pl08x, ch);
			return -EBUSY;
		}
		ch->signal = ret;
	}

1020
	plchan->phychan = ch;
1021 1022 1023 1024 1025
	dev_dbg(&pl08x->adev->dev, "allocated physical channel %d and signal %d for xfer on %s\n",
		 ch->id,
		 ch->signal,
		 plchan->name);

1026 1027 1028 1029 1030 1031 1032
got_channel:
	/* Assign the flow control signal to this channel */
	if (txd->direction == DMA_MEM_TO_DEV)
		txd->ccfg |= ch->signal << PL080_CONFIG_DST_SEL_SHIFT;
	else if (txd->direction == DMA_DEV_TO_MEM)
		txd->ccfg |= ch->signal << PL080_CONFIG_SRC_SEL_SHIFT;

1033
	plchan->phychan_hold++;
1034 1035 1036 1037

	return 0;
}

1038 1039 1040 1041 1042
static void release_phy_channel(struct pl08x_dma_chan *plchan)
{
	struct pl08x_driver_data *pl08x = plchan->host;

	if ((plchan->phychan->signal >= 0) && pl08x->pd->put_signal) {
1043
		pl08x->pd->put_signal(plchan->cd, plchan->phychan->signal);
1044 1045 1046 1047 1048 1049
		plchan->phychan->signal = -1;
	}
	pl08x_put_phy_channel(pl08x, plchan->phychan);
	plchan->phychan = NULL;
}

1050 1051 1052
static dma_cookie_t pl08x_tx_submit(struct dma_async_tx_descriptor *tx)
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(tx->chan);
1053
	struct pl08x_txd *txd = to_pl08x_txd(tx);
1054
	unsigned long flags;
1055
	dma_cookie_t cookie;
1056 1057

	spin_lock_irqsave(&plchan->lock, flags);
1058
	cookie = dma_cookie_assign(tx);
1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071

	/* Put this onto the pending list */
	list_add_tail(&txd->node, &plchan->pend_list);

	/*
	 * If there was no physical channel available for this memcpy,
	 * stack the request up and indicate that the channel is waiting
	 * for a free physical channel.
	 */
	if (!plchan->slave && !plchan->phychan) {
		/* Do this memcpy whenever there is a channel ready */
		plchan->state = PL08X_CHAN_WAITING;
		plchan->waiting = txd;
1072 1073
	} else {
		plchan->phychan_hold--;
1074 1075
	}

1076
	spin_unlock_irqrestore(&plchan->lock, flags);
1077

1078
	return cookie;
1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
}

static struct dma_async_tx_descriptor *pl08x_prep_dma_interrupt(
		struct dma_chan *chan, unsigned long flags)
{
	struct dma_async_tx_descriptor *retval = NULL;

	return retval;
}

/*
1090 1091 1092
 * Code accessing dma_async_is_complete() in a tight loop may give problems.
 * If slaves are relying on interrupts to signal completion this function
 * must not be called with interrupts disabled.
1093
 */
1094 1095
static enum dma_status pl08x_dma_tx_status(struct dma_chan *chan,
		dma_cookie_t cookie, struct dma_tx_state *txstate)
1096 1097 1098 1099
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
	enum dma_status ret;

1100 1101
	ret = dma_cookie_status(chan, cookie, txstate);
	if (ret == DMA_SUCCESS)
1102 1103 1104 1105
		return ret;

	/*
	 * This cookie not complete yet
1106
	 * Get number of bytes left in the active transactions and queue
1107
	 */
1108
	dma_set_residue(txstate, pl08x_getbytes_chan(plchan));
1109 1110 1111 1112 1113 1114 1115 1116 1117 1118

	if (plchan->state == PL08X_CHAN_PAUSED)
		return DMA_PAUSED;

	/* Whether waiting or running, we're in progress */
	return DMA_IN_PROGRESS;
}

/* PrimeCell DMA extension */
struct burst_table {
1119
	u32 burstwords;
1120 1121 1122 1123 1124 1125
	u32 reg;
};

static const struct burst_table burst_sizes[] = {
	{
		.burstwords = 256,
1126
		.reg = PL080_BSIZE_256,
1127 1128 1129
	},
	{
		.burstwords = 128,
1130
		.reg = PL080_BSIZE_128,
1131 1132 1133
	},
	{
		.burstwords = 64,
1134
		.reg = PL080_BSIZE_64,
1135 1136 1137
	},
	{
		.burstwords = 32,
1138
		.reg = PL080_BSIZE_32,
1139 1140 1141
	},
	{
		.burstwords = 16,
1142
		.reg = PL080_BSIZE_16,
1143 1144 1145
	},
	{
		.burstwords = 8,
1146
		.reg = PL080_BSIZE_8,
1147 1148 1149
	},
	{
		.burstwords = 4,
1150
		.reg = PL080_BSIZE_4,
1151 1152
	},
	{
1153 1154
		.burstwords = 0,
		.reg = PL080_BSIZE_1,
1155 1156 1157
	},
};

1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
/*
 * Given the source and destination available bus masks, select which
 * will be routed to each port.  We try to have source and destination
 * on separate ports, but always respect the allowable settings.
 */
static u32 pl08x_select_bus(u8 src, u8 dst)
{
	u32 cctl = 0;

	if (!(dst & PL08X_AHB1) || ((dst & PL08X_AHB2) && (src & PL08X_AHB1)))
		cctl |= PL080_CONTROL_DST_AHB2;
	if (!(src & PL08X_AHB1) || ((src & PL08X_AHB2) && !(dst & PL08X_AHB2)))
		cctl |= PL080_CONTROL_SRC_AHB2;

	return cctl;
}

1175 1176 1177 1178 1179 1180 1181 1182 1183 1184
static u32 pl08x_cctl(u32 cctl)
{
	cctl &= ~(PL080_CONTROL_SRC_AHB2 | PL080_CONTROL_DST_AHB2 |
		  PL080_CONTROL_SRC_INCR | PL080_CONTROL_DST_INCR |
		  PL080_CONTROL_PROT_MASK);

	/* Access the cell in privileged mode, non-bufferable, non-cacheable */
	return cctl | PL080_CONTROL_PROT_SYS;
}

1185 1186 1187 1188 1189 1190 1191 1192 1193
static u32 pl08x_width(enum dma_slave_buswidth width)
{
	switch (width) {
	case DMA_SLAVE_BUSWIDTH_1_BYTE:
		return PL080_WIDTH_8BIT;
	case DMA_SLAVE_BUSWIDTH_2_BYTES:
		return PL080_WIDTH_16BIT;
	case DMA_SLAVE_BUSWIDTH_4_BYTES:
		return PL080_WIDTH_32BIT;
1194 1195
	default:
		return ~0;
1196 1197 1198
	}
}

1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209
static u32 pl08x_burst(u32 maxburst)
{
	int i;

	for (i = 0; i < ARRAY_SIZE(burst_sizes); i++)
		if (burst_sizes[i].burstwords <= maxburst)
			break;

	return burst_sizes[i].reg;
}

1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236
static u32 pl08x_get_cctl(struct pl08x_dma_chan *plchan,
	enum dma_slave_buswidth addr_width, u32 maxburst)
{
	u32 width, burst, cctl = 0;

	width = pl08x_width(addr_width);
	if (width == ~0)
		return ~0;

	cctl |= width << PL080_CONTROL_SWIDTH_SHIFT;
	cctl |= width << PL080_CONTROL_DWIDTH_SHIFT;

	/*
	 * If this channel will only request single transfers, set this
	 * down to ONE element.  Also select one element if no maxburst
	 * is specified.
	 */
	if (plchan->cd->single)
		maxburst = 1;

	burst = pl08x_burst(maxburst);
	cctl |= burst << PL080_CONTROL_SB_SIZE_SHIFT;
	cctl |= burst << PL080_CONTROL_DB_SIZE_SHIFT;

	return pl08x_cctl(cctl);
}

1237 1238
static int dma_set_runtime_config(struct dma_chan *chan,
				  struct dma_slave_config *config)
1239 1240 1241 1242
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
	struct pl08x_driver_data *pl08x = plchan->host;
	enum dma_slave_buswidth addr_width;
1243
	u32 maxburst, cctl = 0;
1244 1245 1246

	if (!plchan->slave)
		return -EINVAL;
1247 1248 1249

	/* Transfer direction */
	plchan->runtime_direction = config->direction;
1250
	if (config->direction == DMA_MEM_TO_DEV) {
1251 1252
		addr_width = config->dst_addr_width;
		maxburst = config->dst_maxburst;
1253
	} else if (config->direction == DMA_DEV_TO_MEM) {
1254 1255 1256 1257 1258
		addr_width = config->src_addr_width;
		maxburst = config->src_maxburst;
	} else {
		dev_err(&pl08x->adev->dev,
			"bad runtime_config: alien transfer direction\n");
1259
		return -EINVAL;
1260 1261
	}

1262 1263
	cctl = pl08x_get_cctl(plchan, addr_width, maxburst);
	if (cctl == ~0) {
1264 1265
		dev_err(&pl08x->adev->dev,
			"bad runtime_config: alien address width\n");
1266
		return -EINVAL;
1267 1268
	}

1269 1270
	plchan->cfg = *config;

1271
	if (plchan->runtime_direction == DMA_DEV_TO_MEM) {
1272
		plchan->src_cctl = cctl;
1273
	} else {
1274
		plchan->dst_cctl = cctl;
1275
	}
1276

1277 1278
	dev_dbg(&pl08x->adev->dev,
		"configured channel %s (%s) for %s, data width %d, "
1279
		"maxburst %d words, LE, CCTL=0x%08x\n",
1280
		dma_chan_name(chan), plchan->name,
1281
		(config->direction == DMA_DEV_TO_MEM) ? "RX" : "TX",
1282 1283
		addr_width,
		maxburst,
1284
		cctl);
1285 1286

	return 0;
1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298
}

/*
 * Slave transactions callback to the slave device to allow
 * synchronization of slave DMA signals with the DMAC enable
 */
static void pl08x_issue_pending(struct dma_chan *chan)
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
	unsigned long flags;

	spin_lock_irqsave(&plchan->lock, flags);
1299 1300 1301
	/* Something is already active, or we're waiting for a channel... */
	if (plchan->at || plchan->state == PL08X_CHAN_WAITING) {
		spin_unlock_irqrestore(&plchan->lock, flags);
1302
		return;
1303
	}
1304 1305

	/* Take the first element in the queue and execute it */
1306
	if (!list_empty(&plchan->pend_list)) {
1307 1308
		struct pl08x_txd *next;

1309
		next = list_first_entry(&plchan->pend_list,
1310 1311 1312 1313 1314
					struct pl08x_txd,
					node);
		list_del(&next->node);
		plchan->state = PL08X_CHAN_RUNNING;

1315
		pl08x_start_txd(plchan, next);
1316 1317 1318 1319 1320 1321 1322 1323 1324
	}

	spin_unlock_irqrestore(&plchan->lock, flags);
}

static int pl08x_prep_channel_resources(struct pl08x_dma_chan *plchan,
					struct pl08x_txd *txd)
{
	struct pl08x_driver_data *pl08x = plchan->host;
1325 1326
	unsigned long flags;
	int num_llis, ret;
1327 1328

	num_llis = pl08x_fill_llis_for_desc(pl08x, txd);
1329
	if (!num_llis) {
1330 1331 1332
		spin_lock_irqsave(&plchan->lock, flags);
		pl08x_free_txd(pl08x, txd);
		spin_unlock_irqrestore(&plchan->lock, flags);
1333
		return -EINVAL;
1334
	}
1335

1336
	spin_lock_irqsave(&plchan->lock, flags);
1337 1338 1339 1340 1341 1342 1343 1344

	/*
	 * See if we already have a physical channel allocated,
	 * else this is the time to try to get one.
	 */
	ret = prep_phy_channel(plchan, txd);
	if (ret) {
		/*
1345 1346 1347 1348 1349 1350 1351 1352 1353 1354
		 * No physical channel was available.
		 *
		 * memcpy transfers can be sorted out at submission time.
		 *
		 * Slave transfers may have been denied due to platform
		 * channel muxing restrictions.  Since there is no guarantee
		 * that this will ever be resolved, and the signal must be
		 * acquired AFTER acquiring the physical channel, we will let
		 * them be NACK:ed with -EBUSY here. The drivers can retry
		 * the prep() call if they are eager on doing this using DMA.
1355 1356 1357
		 */
		if (plchan->slave) {
			pl08x_free_txd_list(pl08x, plchan);
1358
			pl08x_free_txd(pl08x, txd);
1359
			spin_unlock_irqrestore(&plchan->lock, flags);
1360 1361 1362 1363
			return -EBUSY;
		}
	} else
		/*
1364 1365 1366 1367
		 * Else we're all set, paused and ready to roll, status
		 * will switch to PL08X_CHAN_RUNNING when we call
		 * issue_pending(). If there is something running on the
		 * channel already we don't change its state.
1368 1369 1370 1371
		 */
		if (plchan->state == PL08X_CHAN_IDLE)
			plchan->state = PL08X_CHAN_PAUSED;

1372
	spin_unlock_irqrestore(&plchan->lock, flags);
1373 1374 1375 1376

	return 0;
}

1377 1378
static struct pl08x_txd *pl08x_get_txd(struct pl08x_dma_chan *plchan,
	unsigned long flags)
1379
{
1380
	struct pl08x_txd *txd = kzalloc(sizeof(*txd), GFP_NOWAIT);
1381 1382 1383

	if (txd) {
		dma_async_tx_descriptor_init(&txd->tx, &plchan->chan);
1384
		txd->tx.flags = flags;
1385 1386
		txd->tx.tx_submit = pl08x_tx_submit;
		INIT_LIST_HEAD(&txd->node);
1387
		INIT_LIST_HEAD(&txd->dsg_list);
1388 1389 1390 1391

		/* Always enable error and terminal interrupts */
		txd->ccfg = PL080_CONFIG_ERR_IRQ_MASK |
			    PL080_CONFIG_TC_IRQ_MASK;
1392 1393 1394 1395
	}
	return txd;
}

1396 1397 1398 1399 1400 1401 1402 1403 1404 1405
/*
 * Initialize a descriptor to be used by memcpy submit
 */
static struct dma_async_tx_descriptor *pl08x_prep_dma_memcpy(
		struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
		size_t len, unsigned long flags)
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
	struct pl08x_driver_data *pl08x = plchan->host;
	struct pl08x_txd *txd;
1406
	struct pl08x_sg *dsg;
1407 1408
	int ret;

1409
	txd = pl08x_get_txd(plchan, flags);
1410 1411 1412 1413 1414 1415
	if (!txd) {
		dev_err(&pl08x->adev->dev,
			"%s no memory for descriptor\n", __func__);
		return NULL;
	}

1416 1417 1418 1419 1420 1421 1422 1423 1424
	dsg = kzalloc(sizeof(struct pl08x_sg), GFP_NOWAIT);
	if (!dsg) {
		pl08x_free_txd(pl08x, txd);
		dev_err(&pl08x->adev->dev, "%s no memory for pl080 sg\n",
				__func__);
		return NULL;
	}
	list_add_tail(&dsg->node, &txd->dsg_list);

1425
	txd->direction = DMA_MEM_TO_MEM;
1426 1427 1428
	dsg->src_addr = src;
	dsg->dst_addr = dest;
	dsg->len = len;
1429 1430

	/* Set platform data for m2m */
1431
	txd->ccfg |= PL080_FLOW_MEM2MEM << PL080_CONFIG_FLOW_CONTROL_SHIFT;
1432 1433
	txd->cctl = pl08x->pd->memcpy_channel.cctl &
			~(PL080_CONTROL_DST_AHB2 | PL080_CONTROL_SRC_AHB2);
1434

1435
	/* Both to be incremented or the code will break */
1436
	txd->cctl |= PL080_CONTROL_SRC_INCR | PL080_CONTROL_DST_INCR;
1437 1438

	if (pl08x->vd->dualmaster)
1439 1440
		txd->cctl |= pl08x_select_bus(pl08x->mem_buses,
					      pl08x->mem_buses);
1441 1442 1443 1444 1445 1446 1447 1448

	ret = pl08x_prep_channel_resources(plchan, txd);
	if (ret)
		return NULL;

	return &txd->tx;
}

1449
static struct dma_async_tx_descriptor *pl08x_prep_slave_sg(
1450
		struct dma_chan *chan, struct scatterlist *sgl,
1451
		unsigned int sg_len, enum dma_transfer_direction direction,
1452
		unsigned long flags, void *context)
1453 1454 1455 1456
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
	struct pl08x_driver_data *pl08x = plchan->host;
	struct pl08x_txd *txd;
1457 1458 1459
	struct pl08x_sg *dsg;
	struct scatterlist *sg;
	dma_addr_t slave_addr;
1460
	int ret, tmp;
1461 1462
	u8 src_buses, dst_buses;
	u32 cctl;
1463 1464

	dev_dbg(&pl08x->adev->dev, "%s prepare transaction of %d bytes from %s\n",
1465
			__func__, sg_dma_len(sgl), plchan->name);
1466

1467
	txd = pl08x_get_txd(plchan, flags);
1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483
	if (!txd) {
		dev_err(&pl08x->adev->dev, "%s no txd\n", __func__);
		return NULL;
	}

	if (direction != plchan->runtime_direction)
		dev_err(&pl08x->adev->dev, "%s DMA setup does not match "
			"the direction configured for the PrimeCell\n",
			__func__);

	/*
	 * Set up addresses, the PrimeCell configured address
	 * will take precedence since this may configure the
	 * channel target address dynamically at runtime.
	 */
	txd->direction = direction;
1484

1485
	if (direction == DMA_MEM_TO_DEV) {
1486
		cctl = plchan->dst_cctl | PL080_CONTROL_SRC_INCR;
1487
		slave_addr = plchan->cfg.dst_addr;
1488 1489
		src_buses = pl08x->mem_buses;
		dst_buses = plchan->cd->periph_buses;
1490
	} else if (direction == DMA_DEV_TO_MEM) {
1491
		cctl = plchan->src_cctl | PL080_CONTROL_DST_INCR;
1492
		slave_addr = plchan->cfg.src_addr;
1493 1494
		src_buses = plchan->cd->periph_buses;
		dst_buses = pl08x->mem_buses;
1495
	} else {
1496
		pl08x_free_txd(pl08x, txd);
1497 1498 1499 1500 1501
		dev_err(&pl08x->adev->dev,
			"%s direction unsupported\n", __func__);
		return NULL;
	}

1502 1503
	txd->cctl = cctl | pl08x_select_bus(src_buses, dst_buses);

1504
	if (plchan->cfg.device_fc)
1505
		tmp = (direction == DMA_MEM_TO_DEV) ? PL080_FLOW_MEM2PER_PER :
1506 1507
			PL080_FLOW_PER2MEM_PER;
	else
1508
		tmp = (direction == DMA_MEM_TO_DEV) ? PL080_FLOW_MEM2PER :
1509 1510 1511 1512
			PL080_FLOW_PER2MEM;

	txd->ccfg |= tmp << PL080_CONFIG_FLOW_CONTROL_SHIFT;

1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523
	for_each_sg(sgl, sg, sg_len, tmp) {
		dsg = kzalloc(sizeof(struct pl08x_sg), GFP_NOWAIT);
		if (!dsg) {
			pl08x_free_txd(pl08x, txd);
			dev_err(&pl08x->adev->dev, "%s no mem for pl080 sg\n",
					__func__);
			return NULL;
		}
		list_add_tail(&dsg->node, &txd->dsg_list);

		dsg->len = sg_dma_len(sg);
1524
		if (direction == DMA_MEM_TO_DEV) {
1525
			dsg->src_addr = sg_dma_address(sg);
1526 1527 1528
			dsg->dst_addr = slave_addr;
		} else {
			dsg->src_addr = slave_addr;
1529
			dsg->dst_addr = sg_dma_address(sg);
1530 1531 1532
		}
	}

1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549
	ret = pl08x_prep_channel_resources(plchan, txd);
	if (ret)
		return NULL;

	return &txd->tx;
}

static int pl08x_control(struct dma_chan *chan, enum dma_ctrl_cmd cmd,
			 unsigned long arg)
{
	struct pl08x_dma_chan *plchan = to_pl08x_chan(chan);
	struct pl08x_driver_data *pl08x = plchan->host;
	unsigned long flags;
	int ret = 0;

	/* Controls applicable to inactive channels */
	if (cmd == DMA_SLAVE_CONFIG) {
1550 1551
		return dma_set_runtime_config(chan,
					      (struct dma_slave_config *)arg);
1552 1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568
	}

	/*
	 * Anything succeeds on channels with no physical allocation and
	 * no queued transfers.
	 */
	spin_lock_irqsave(&plchan->lock, flags);
	if (!plchan->phychan && !plchan->at) {
		spin_unlock_irqrestore(&plchan->lock, flags);
		return 0;
	}

	switch (cmd) {
	case DMA_TERMINATE_ALL:
		plchan->state = PL08X_CHAN_IDLE;

		if (plchan->phychan) {
1569
			pl08x_terminate_phy_chan(pl08x, plchan->phychan);
1570 1571 1572 1573 1574

			/*
			 * Mark physical channel as free and free any slave
			 * signal
			 */
1575
			release_phy_channel(plchan);
1576
			plchan->phychan_hold = 0;
1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606
		}
		/* Dequeue jobs and free LLIs */
		if (plchan->at) {
			pl08x_free_txd(pl08x, plchan->at);
			plchan->at = NULL;
		}
		/* Dequeue jobs not yet fired as well */
		pl08x_free_txd_list(pl08x, plchan);
		break;
	case DMA_PAUSE:
		pl08x_pause_phy_chan(plchan->phychan);
		plchan->state = PL08X_CHAN_PAUSED;
		break;
	case DMA_RESUME:
		pl08x_resume_phy_chan(plchan->phychan);
		plchan->state = PL08X_CHAN_RUNNING;
		break;
	default:
		/* Unknown command */
		ret = -ENXIO;
		break;
	}

	spin_unlock_irqrestore(&plchan->lock, flags);

	return ret;
}

bool pl08x_filter_id(struct dma_chan *chan, void *chan_id)
{
1607
	struct pl08x_dma_chan *plchan;
1608 1609
	char *name = chan_id;

1610 1611 1612 1613 1614 1615
	/* Reject channels for devices not bound to this driver */
	if (chan->device->dev->driver != &pl08x_amba_driver.drv)
		return false;

	plchan = to_pl08x_chan(chan);

1616 1617 1618 1619 1620 1621 1622 1623 1624
	/* Check that the channel is not taken! */
	if (!strcmp(plchan->name, name))
		return true;

	return false;
}

/*
 * Just check that the device is there and active
1625 1626 1627
 * TODO: turn this bit on/off depending on the number of physical channels
 * actually used, if it is zero... well shut it off. That will save some
 * power. Cut the clock at the same time.
1628 1629 1630
 */
static void pl08x_ensure_on(struct pl08x_driver_data *pl08x)
{
1631 1632 1633
	/* The Nomadik variant does not have the config register */
	if (pl08x->vd->nomadik)
		return;
1634
	writel(PL080_CONFIG_ENABLE, pl08x->base + PL080_CONFIG);
1635 1636
}

1637 1638 1639
static void pl08x_unmap_buffers(struct pl08x_txd *txd)
{
	struct device *dev = txd->tx.chan->device->dev;
1640
	struct pl08x_sg *dsg;
1641 1642 1643

	if (!(txd->tx.flags & DMA_COMPL_SKIP_SRC_UNMAP)) {
		if (txd->tx.flags & DMA_COMPL_SRC_UNMAP_SINGLE)
1644 1645 1646 1647 1648 1649 1650 1651
			list_for_each_entry(dsg, &txd->dsg_list, node)
				dma_unmap_single(dev, dsg->src_addr, dsg->len,
						DMA_TO_DEVICE);
		else {
			list_for_each_entry(dsg, &txd->dsg_list, node)
				dma_unmap_page(dev, dsg->src_addr, dsg->len,
						DMA_TO_DEVICE);
		}
1652 1653 1654
	}
	if (!(txd->tx.flags & DMA_COMPL_SKIP_DEST_UNMAP)) {
		if (txd->tx.flags & DMA_COMPL_DEST_UNMAP_SINGLE)
1655 1656 1657
			list_for_each_entry(dsg, &txd->dsg_list, node)
				dma_unmap_single(dev, dsg->dst_addr, dsg->len,
						DMA_FROM_DEVICE);
1658
		else
1659 1660 1661
			list_for_each_entry(dsg, &txd->dsg_list, node)
				dma_unmap_page(dev, dsg->dst_addr, dsg->len,
						DMA_FROM_DEVICE);
1662 1663 1664
	}
}

1665 1666 1667 1668
static void pl08x_tasklet(unsigned long data)
{
	struct pl08x_dma_chan *plchan = (struct pl08x_dma_chan *) data;
	struct pl08x_driver_data *pl08x = plchan->host;
1669
	struct pl08x_txd *txd;
1670
	unsigned long flags;
1671

1672
	spin_lock_irqsave(&plchan->lock, flags);
1673

1674 1675
	txd = plchan->at;
	plchan->at = NULL;
1676

1677
	if (txd) {
1678
		/* Update last completed */
1679
		dma_cookie_complete(&txd->tx);
1680
	}
1681

1682
	/* If a new descriptor is queued, set it up plchan->at is NULL here */
1683
	if (!list_empty(&plchan->pend_list)) {
1684 1685
		struct pl08x_txd *next;

1686
		next = list_first_entry(&plchan->pend_list,
1687 1688 1689
					struct pl08x_txd,
					node);
		list_del(&next->node);
1690 1691

		pl08x_start_txd(plchan, next);
1692 1693 1694 1695 1696 1697
	} else if (plchan->phychan_hold) {
		/*
		 * This channel is still in use - we have a new txd being
		 * prepared and will soon be queued.  Don't give up the
		 * physical channel.
		 */
1698 1699 1700 1701 1702 1703 1704
	} else {
		struct pl08x_dma_chan *waiting = NULL;

		/*
		 * No more jobs, so free up the physical channel
		 * Free any allocated signal on slave transfers too
		 */
1705
		release_phy_channel(plchan);
1706 1707 1708
		plchan->state = PL08X_CHAN_IDLE;

		/*
1709 1710 1711 1712
		 * And NOW before anyone else can grab that free:d up
		 * physical channel, see if there is some memcpy pending
		 * that seriously needs to start because of being stacked
		 * up while we were choking the physical channels with data.
1713 1714 1715
		 */
		list_for_each_entry(waiting, &pl08x->memcpy.channels,
				    chan.device_node) {
1716 1717
			if (waiting->state == PL08X_CHAN_WAITING &&
				waiting->waiting != NULL) {
1718 1719 1720 1721 1722 1723
				int ret;

				/* This should REALLY not fail now */
				ret = prep_phy_channel(waiting,
						       waiting->waiting);
				BUG_ON(ret);
1724
				waiting->phychan_hold--;
1725 1726 1727 1728 1729 1730 1731 1732
				waiting->state = PL08X_CHAN_RUNNING;
				waiting->waiting = NULL;
				pl08x_issue_pending(&waiting->chan);
				break;
			}
		}
	}

1733
	spin_unlock_irqrestore(&plchan->lock, flags);
1734

1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751
	if (txd) {
		dma_async_tx_callback callback = txd->tx.callback;
		void *callback_param = txd->tx.callback_param;

		/* Don't try to unmap buffers on slave channels */
		if (!plchan->slave)
			pl08x_unmap_buffers(txd);

		/* Free the descriptor */
		spin_lock_irqsave(&plchan->lock, flags);
		pl08x_free_txd(pl08x, txd);
		spin_unlock_irqrestore(&plchan->lock, flags);

		/* Callback to signal completion */
		if (callback)
			callback(callback_param);
	}
1752 1753 1754 1755 1756
}

static irqreturn_t pl08x_irq(int irq, void *dev)
{
	struct pl08x_driver_data *pl08x = dev;
1757 1758 1759 1760 1761 1762 1763 1764
	u32 mask = 0, err, tc, i;

	/* check & clear - ERR & TC interrupts */
	err = readl(pl08x->base + PL080_ERR_STATUS);
	if (err) {
		dev_err(&pl08x->adev->dev, "%s error interrupt, register value 0x%08x\n",
			__func__, err);
		writel(err, pl08x->base + PL080_ERR_CLEAR);
1765
	}
1766
	tc = readl(pl08x->base + PL080_TC_STATUS);
1767 1768 1769 1770 1771 1772
	if (tc)
		writel(tc, pl08x->base + PL080_TC_CLEAR);

	if (!err && !tc)
		return IRQ_NONE;

1773
	for (i = 0; i < pl08x->vd->channels; i++) {
1774
		if (((1 << i) & err) || ((1 << i) & tc)) {
1775 1776 1777 1778
			/* Locate physical channel */
			struct pl08x_phy_chan *phychan = &pl08x->phy_chans[i];
			struct pl08x_dma_chan *plchan = phychan->serving;

1779 1780 1781 1782 1783 1784 1785
			if (!plchan) {
				dev_err(&pl08x->adev->dev,
					"%s Error TC interrupt on unused channel: 0x%08x\n",
					__func__, i);
				continue;
			}

1786 1787 1788 1789 1790 1791 1792 1793 1794
			/* Schedule tasklet on this channel */
			tasklet_schedule(&plchan->tasklet);
			mask |= (1 << i);
		}
	}

	return mask ? IRQ_HANDLED : IRQ_NONE;
}

1795 1796 1797 1798 1799 1800
static void pl08x_dma_slave_init(struct pl08x_dma_chan *chan)
{
	u32 cctl = pl08x_cctl(chan->cd->cctl);

	chan->slave = true;
	chan->name = chan->cd->bus_id;
1801 1802
	chan->cfg.src_addr = chan->cd->addr;
	chan->cfg.dst_addr = chan->cd->addr;
1803 1804
	chan->src_cctl = cctl;
	chan->dst_cctl = cctl;
1805 1806
}

1807 1808 1809 1810 1811
/*
 * Initialise the DMAC memcpy/slave channels.
 * Make a local wrapper to hold required data
 */
static int pl08x_dma_init_virtual_channels(struct pl08x_driver_data *pl08x,
1812
		struct dma_device *dmadev, unsigned int channels, bool slave)
1813 1814 1815 1816 1817
{
	struct pl08x_dma_chan *chan;
	int i;

	INIT_LIST_HEAD(&dmadev->channels);
1818

1819 1820 1821 1822 1823 1824
	/*
	 * Register as many many memcpy as we have physical channels,
	 * we won't always be able to use all but the code will have
	 * to cope with that situation.
	 */
	for (i = 0; i < channels; i++) {
1825
		chan = kzalloc(sizeof(*chan), GFP_KERNEL);
1826 1827 1828 1829 1830 1831 1832 1833 1834 1835 1836
		if (!chan) {
			dev_err(&pl08x->adev->dev,
				"%s no memory for channel\n", __func__);
			return -ENOMEM;
		}

		chan->host = pl08x;
		chan->state = PL08X_CHAN_IDLE;

		if (slave) {
			chan->cd = &pl08x->pd->slave_channels[i];
1837
			pl08x_dma_slave_init(chan);
1838 1839 1840 1841 1842 1843 1844 1845
		} else {
			chan->cd = &pl08x->pd->memcpy_channel;
			chan->name = kasprintf(GFP_KERNEL, "memcpy%d", i);
			if (!chan->name) {
				kfree(chan);
				return -ENOMEM;
			}
		}
1846
		dev_dbg(&pl08x->adev->dev,
1847 1848 1849 1850
			 "initialize virtual channel \"%s\"\n",
			 chan->name);

		chan->chan.device = dmadev;
1851
		dma_cookie_init(&chan->chan);
1852 1853

		spin_lock_init(&chan->lock);
1854
		INIT_LIST_HEAD(&chan->pend_list);
1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871 1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913
		tasklet_init(&chan->tasklet, pl08x_tasklet,
			     (unsigned long) chan);

		list_add_tail(&chan->chan.device_node, &dmadev->channels);
	}
	dev_info(&pl08x->adev->dev, "initialized %d virtual %s channels\n",
		 i, slave ? "slave" : "memcpy");
	return i;
}

static void pl08x_free_virtual_channels(struct dma_device *dmadev)
{
	struct pl08x_dma_chan *chan = NULL;
	struct pl08x_dma_chan *next;

	list_for_each_entry_safe(chan,
				 next, &dmadev->channels, chan.device_node) {
		list_del(&chan->chan.device_node);
		kfree(chan);
	}
}

#ifdef CONFIG_DEBUG_FS
static const char *pl08x_state_str(enum pl08x_dma_chan_state state)
{
	switch (state) {
	case PL08X_CHAN_IDLE:
		return "idle";
	case PL08X_CHAN_RUNNING:
		return "running";
	case PL08X_CHAN_PAUSED:
		return "paused";
	case PL08X_CHAN_WAITING:
		return "waiting";
	default:
		break;
	}
	return "UNKNOWN STATE";
}

static int pl08x_debugfs_show(struct seq_file *s, void *data)
{
	struct pl08x_driver_data *pl08x = s->private;
	struct pl08x_dma_chan *chan;
	struct pl08x_phy_chan *ch;
	unsigned long flags;
	int i;

	seq_printf(s, "PL08x physical channels:\n");
	seq_printf(s, "CHANNEL:\tUSER:\n");
	seq_printf(s, "--------\t-----\n");
	for (i = 0; i < pl08x->vd->channels; i++) {
		struct pl08x_dma_chan *virt_chan;

		ch = &pl08x->phy_chans[i];

		spin_lock_irqsave(&ch->lock, flags);
		virt_chan = ch->serving;

1914 1915 1916 1917
		seq_printf(s, "%d\t\t%s%s\n",
			   ch->id,
			   virt_chan ? virt_chan->name : "(none)",
			   ch->locked ? " LOCKED" : "");
1918 1919 1920 1921 1922 1923 1924 1925

		spin_unlock_irqrestore(&ch->lock, flags);
	}

	seq_printf(s, "\nPL08x virtual memcpy channels:\n");
	seq_printf(s, "CHANNEL:\tSTATE:\n");
	seq_printf(s, "--------\t------\n");
	list_for_each_entry(chan, &pl08x->memcpy.channels, chan.device_node) {
1926
		seq_printf(s, "%s\t\t%s\n", chan->name,
1927 1928 1929 1930 1931 1932 1933
			   pl08x_state_str(chan->state));
	}

	seq_printf(s, "\nPL08x virtual slave channels:\n");
	seq_printf(s, "CHANNEL:\tSTATE:\n");
	seq_printf(s, "--------\t------\n");
	list_for_each_entry(chan, &pl08x->slave.channels, chan.device_node) {
1934
		seq_printf(s, "%s\t\t%s\n", chan->name,
1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955
			   pl08x_state_str(chan->state));
	}

	return 0;
}

static int pl08x_debugfs_open(struct inode *inode, struct file *file)
{
	return single_open(file, pl08x_debugfs_show, inode->i_private);
}

static const struct file_operations pl08x_debugfs_operations = {
	.open		= pl08x_debugfs_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
};

static void init_pl08x_debugfs(struct pl08x_driver_data *pl08x)
{
	/* Expose a simple debugfs interface to view all clocks */
1956 1957 1958
	(void) debugfs_create_file(dev_name(&pl08x->adev->dev),
			S_IFREG | S_IRUGO, NULL, pl08x,
			&pl08x_debugfs_operations);
1959 1960 1961 1962 1963 1964 1965 1966
}

#else
static inline void init_pl08x_debugfs(struct pl08x_driver_data *pl08x)
{
}
#endif

1967
static int pl08x_probe(struct amba_device *adev, const struct amba_id *id)
1968 1969
{
	struct pl08x_driver_data *pl08x;
1970
	const struct vendor_data *vd = id->data;
1971 1972 1973 1974 1975 1976 1977 1978
	int ret = 0;
	int i;

	ret = amba_request_regions(adev, NULL);
	if (ret)
		return ret;

	/* Create the driver state holder */
1979
	pl08x = kzalloc(sizeof(*pl08x), GFP_KERNEL);
1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017
	if (!pl08x) {
		ret = -ENOMEM;
		goto out_no_pl08x;
	}

	/* Initialize memcpy engine */
	dma_cap_set(DMA_MEMCPY, pl08x->memcpy.cap_mask);
	pl08x->memcpy.dev = &adev->dev;
	pl08x->memcpy.device_alloc_chan_resources = pl08x_alloc_chan_resources;
	pl08x->memcpy.device_free_chan_resources = pl08x_free_chan_resources;
	pl08x->memcpy.device_prep_dma_memcpy = pl08x_prep_dma_memcpy;
	pl08x->memcpy.device_prep_dma_interrupt = pl08x_prep_dma_interrupt;
	pl08x->memcpy.device_tx_status = pl08x_dma_tx_status;
	pl08x->memcpy.device_issue_pending = pl08x_issue_pending;
	pl08x->memcpy.device_control = pl08x_control;

	/* Initialize slave engine */
	dma_cap_set(DMA_SLAVE, pl08x->slave.cap_mask);
	pl08x->slave.dev = &adev->dev;
	pl08x->slave.device_alloc_chan_resources = pl08x_alloc_chan_resources;
	pl08x->slave.device_free_chan_resources = pl08x_free_chan_resources;
	pl08x->slave.device_prep_dma_interrupt = pl08x_prep_dma_interrupt;
	pl08x->slave.device_tx_status = pl08x_dma_tx_status;
	pl08x->slave.device_issue_pending = pl08x_issue_pending;
	pl08x->slave.device_prep_slave_sg = pl08x_prep_slave_sg;
	pl08x->slave.device_control = pl08x_control;

	/* Get the platform data */
	pl08x->pd = dev_get_platdata(&adev->dev);
	if (!pl08x->pd) {
		dev_err(&adev->dev, "no platform data supplied\n");
		goto out_no_platdata;
	}

	/* Assign useful pointers to the driver state */
	pl08x->adev = adev;
	pl08x->vd = vd;

2018 2019 2020 2021 2022 2023 2024 2025
	/* By default, AHB1 only.  If dualmaster, from platform */
	pl08x->lli_buses = PL08X_AHB1;
	pl08x->mem_buses = PL08X_AHB1;
	if (pl08x->vd->dualmaster) {
		pl08x->lli_buses = pl08x->pd->lli_buses;
		pl08x->mem_buses = pl08x->pd->mem_buses;
	}

2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042
	/* A DMA memory pool for LLIs, align on 1-byte boundary */
	pl08x->pool = dma_pool_create(DRIVER_NAME, &pl08x->adev->dev,
			PL08X_LLI_TSFR_SIZE, PL08X_ALIGN, 0);
	if (!pl08x->pool) {
		ret = -ENOMEM;
		goto out_no_lli_pool;
	}

	pl08x->base = ioremap(adev->res.start, resource_size(&adev->res));
	if (!pl08x->base) {
		ret = -ENOMEM;
		goto out_no_ioremap;
	}

	/* Turn on the PL08x */
	pl08x_ensure_on(pl08x);

2043
	/* Attach the interrupt handler */
2044 2045 2046 2047
	writel(0x000000FF, pl08x->base + PL080_ERR_CLEAR);
	writel(0x000000FF, pl08x->base + PL080_TC_CLEAR);

	ret = request_irq(adev->irq[0], pl08x_irq, IRQF_DISABLED,
2048
			  DRIVER_NAME, pl08x);
2049 2050 2051 2052 2053 2054 2055
	if (ret) {
		dev_err(&adev->dev, "%s failed to request interrupt %d\n",
			__func__, adev->irq[0]);
		goto out_no_irq;
	}

	/* Initialize physical channels */
2056
	pl08x->phy_chans = kzalloc((vd->channels * sizeof(*pl08x->phy_chans)),
2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071
			GFP_KERNEL);
	if (!pl08x->phy_chans) {
		dev_err(&adev->dev, "%s failed to allocate "
			"physical channel holders\n",
			__func__);
		goto out_no_phychans;
	}

	for (i = 0; i < vd->channels; i++) {
		struct pl08x_phy_chan *ch = &pl08x->phy_chans[i];

		ch->id = i;
		ch->base = pl08x->base + PL080_Cx_BASE(i);
		spin_lock_init(&ch->lock);
		ch->signal = -1;
2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087

		/*
		 * Nomadik variants can have channels that are locked
		 * down for the secure world only. Lock up these channels
		 * by perpetually serving a dummy virtual channel.
		 */
		if (vd->nomadik) {
			u32 val;

			val = readl(ch->base + PL080_CH_CONFIG);
			if (val & (PL080N_CONFIG_ITPROT | PL080N_CONFIG_SECPROT)) {
				dev_info(&adev->dev, "physical channel %d reserved for secure access only\n", i);
				ch->locked = true;
			}
		}

2088 2089
		dev_dbg(&adev->dev, "physical channel %d is %s\n",
			i, pl08x_phy_channel_busy(ch) ? "BUSY" : "FREE");
2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104
	}

	/* Register as many memcpy channels as there are physical channels */
	ret = pl08x_dma_init_virtual_channels(pl08x, &pl08x->memcpy,
					      pl08x->vd->channels, false);
	if (ret <= 0) {
		dev_warn(&pl08x->adev->dev,
			 "%s failed to enumerate memcpy channels - %d\n",
			 __func__, ret);
		goto out_no_memcpy;
	}
	pl08x->memcpy.chancnt = ret;

	/* Register slave channels */
	ret = pl08x_dma_init_virtual_channels(pl08x, &pl08x->slave,
2105
			pl08x->pd->num_slave_channels, true);
2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121 2122 2123 2124 2125 2126 2127 2128 2129 2130 2131
	if (ret <= 0) {
		dev_warn(&pl08x->adev->dev,
			"%s failed to enumerate slave channels - %d\n",
				__func__, ret);
		goto out_no_slave;
	}
	pl08x->slave.chancnt = ret;

	ret = dma_async_device_register(&pl08x->memcpy);
	if (ret) {
		dev_warn(&pl08x->adev->dev,
			"%s failed to register memcpy as an async device - %d\n",
			__func__, ret);
		goto out_no_memcpy_reg;
	}

	ret = dma_async_device_register(&pl08x->slave);
	if (ret) {
		dev_warn(&pl08x->adev->dev,
			"%s failed to register slave as an async device - %d\n",
			__func__, ret);
		goto out_no_slave_reg;
	}

	amba_set_drvdata(adev, pl08x);
	init_pl08x_debugfs(pl08x);
2132 2133 2134
	dev_info(&pl08x->adev->dev, "DMA: PL%03x rev%u at 0x%08llx irq %d\n",
		 amba_part(adev), amba_rev(adev),
		 (unsigned long long)adev->res.start, adev->irq[0]);
2135

2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165
	return 0;

out_no_slave_reg:
	dma_async_device_unregister(&pl08x->memcpy);
out_no_memcpy_reg:
	pl08x_free_virtual_channels(&pl08x->slave);
out_no_slave:
	pl08x_free_virtual_channels(&pl08x->memcpy);
out_no_memcpy:
	kfree(pl08x->phy_chans);
out_no_phychans:
	free_irq(adev->irq[0], pl08x);
out_no_irq:
	iounmap(pl08x->base);
out_no_ioremap:
	dma_pool_destroy(pl08x->pool);
out_no_lli_pool:
out_no_platdata:
	kfree(pl08x);
out_no_pl08x:
	amba_release_regions(adev);
	return ret;
}

/* PL080 has 8 channels and the PL080 have just 2 */
static struct vendor_data vendor_pl080 = {
	.channels = 8,
	.dualmaster = true,
};

2166 2167 2168 2169 2170 2171
static struct vendor_data vendor_nomadik = {
	.channels = 8,
	.dualmaster = true,
	.nomadik = true,
};

2172 2173 2174 2175 2176 2177 2178 2179 2180 2181 2182 2183 2184 2185 2186 2187 2188 2189 2190 2191
static struct vendor_data vendor_pl081 = {
	.channels = 2,
	.dualmaster = false,
};

static struct amba_id pl08x_ids[] = {
	/* PL080 */
	{
		.id	= 0x00041080,
		.mask	= 0x000fffff,
		.data	= &vendor_pl080,
	},
	/* PL081 */
	{
		.id	= 0x00041081,
		.mask	= 0x000fffff,
		.data	= &vendor_pl081,
	},
	/* Nomadik 8815 PL080 variant */
	{
2192
		.id	= 0x00280080,
2193
		.mask	= 0x00ffffff,
2194
		.data	= &vendor_nomadik,
2195 2196 2197 2198
	},
	{ 0, 0 },
};

2199 2200
MODULE_DEVICE_TABLE(amba, pl08x_ids);

2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212
static struct amba_driver pl08x_amba_driver = {
	.drv.name	= DRIVER_NAME,
	.id_table	= pl08x_ids,
	.probe		= pl08x_probe,
};

static int __init pl08x_init(void)
{
	int retval;
	retval = amba_driver_register(&pl08x_amba_driver);
	if (retval)
		printk(KERN_WARNING DRIVER_NAME
2213
		       "failed to register as an AMBA device (%d)\n",
2214 2215 2216 2217
		       retval);
	return retval;
}
subsys_initcall(pl08x_init);