mr.c 20.7 KB
Newer Older
1 2
/*
 * Copyright (c) 2007 Cisco Systems, Inc. All rights reserved.
3
 * Copyright (c) 2007, 2008 Mellanox Technologies. All rights reserved.
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
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 */

34
#include <linux/slab.h>
35
#include <rdma/ib_user_verbs.h>
36

37 38 39 40 41 42 43 44
#include "mlx4_ib.h"

static u32 convert_access(int acc)
{
	return (acc & IB_ACCESS_REMOTE_ATOMIC ? MLX4_PERM_ATOMIC       : 0) |
	       (acc & IB_ACCESS_REMOTE_WRITE  ? MLX4_PERM_REMOTE_WRITE : 0) |
	       (acc & IB_ACCESS_REMOTE_READ   ? MLX4_PERM_REMOTE_READ  : 0) |
	       (acc & IB_ACCESS_LOCAL_WRITE   ? MLX4_PERM_LOCAL_WRITE  : 0) |
45
	       (acc & IB_ACCESS_MW_BIND	      ? MLX4_PERM_BIND_MW      : 0) |
46 47 48
	       MLX4_PERM_LOCAL_READ;
}

49 50 51 52 53 54 55 56 57
static enum mlx4_mw_type to_mlx4_type(enum ib_mw_type type)
{
	switch (type) {
	case IB_MW_TYPE_1:	return MLX4_MW_TYPE_1;
	case IB_MW_TYPE_2:	return MLX4_MW_TYPE_2;
	default:		return -1;
	}
}

58 59 60 61 62
struct ib_mr *mlx4_ib_get_dma_mr(struct ib_pd *pd, int acc)
{
	struct mlx4_ib_mr *mr;
	int err;

63
	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81
	if (!mr)
		return ERR_PTR(-ENOMEM);

	err = mlx4_mr_alloc(to_mdev(pd->device)->dev, to_mpd(pd)->pdn, 0,
			    ~0ull, convert_access(acc), 0, 0, &mr->mmr);
	if (err)
		goto err_free;

	err = mlx4_mr_enable(to_mdev(pd->device)->dev, &mr->mmr);
	if (err)
		goto err_mr;

	mr->ibmr.rkey = mr->ibmr.lkey = mr->mmr.key;
	mr->umem = NULL;

	return &mr->ibmr;

err_mr:
82
	(void) mlx4_mr_free(to_mdev(pd->device)->dev, &mr->mmr);
83 84 85 86 87 88 89

err_free:
	kfree(mr);

	return ERR_PTR(err);
}

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
enum {
	MLX4_MAX_MTT_SHIFT = 31
};

static int mlx4_ib_umem_write_mtt_block(struct mlx4_ib_dev *dev,
					struct mlx4_mtt *mtt,
					u64 mtt_size, u64 mtt_shift, u64 len,
					u64 cur_start_addr, u64 *pages,
					int *start_index, int *npages)
{
	u64 cur_end_addr = cur_start_addr + len;
	u64 cur_end_addr_aligned = 0;
	u64 mtt_entries;
	int err = 0;
	int k;

	len += (cur_start_addr & (mtt_size - 1ULL));
	cur_end_addr_aligned = round_up(cur_end_addr, mtt_size);
	len += (cur_end_addr_aligned - cur_end_addr);
	if (len & (mtt_size - 1ULL)) {
		pr_warn("write_block: len %llx is not aligned to mtt_size %llx\n",
			len, mtt_size);
		return -EINVAL;
	}

	mtt_entries = (len >> mtt_shift);

	/*
	 * Align the MTT start address to the mtt_size.
	 * Required to handle cases when the MR starts in the middle of an MTT
	 * record. Was not required in old code since the physical addresses
	 * provided by the dma subsystem were page aligned, which was also the
	 * MTT size.
	 */
	cur_start_addr = round_down(cur_start_addr, mtt_size);
	/* A new block is started ... */
	for (k = 0; k < mtt_entries; ++k) {
		pages[*npages] = cur_start_addr + (mtt_size * k);
		(*npages)++;
		/*
		 * Be friendly to mlx4_write_mtt() and pass it chunks of
		 * appropriate size.
		 */
		if (*npages == PAGE_SIZE / sizeof(u64)) {
			err = mlx4_write_mtt(dev->dev, mtt, *start_index,
					     *npages, pages);
			if (err)
				return err;

			(*start_index) += *npages;
			*npages = 0;
		}
	}

	return 0;
}

static inline u64 alignment_of(u64 ptr)
{
	return ilog2(ptr & (~(ptr - 1)));
}

static int mlx4_ib_umem_calc_block_mtt(u64 next_block_start,
				       u64 current_block_end,
				       u64 block_shift)
{
	/* Check whether the alignment of the new block is aligned as well as
	 * the previous block.
	 * Block address must start with zeros till size of entity_size.
	 */
	if ((next_block_start & ((1ULL << block_shift) - 1ULL)) != 0)
		/*
		 * It is not as well aligned as the previous block-reduce the
		 * mtt size accordingly. Here we take the last right bit which
		 * is 1.
		 */
		block_shift = alignment_of(next_block_start);

	/*
	 * Check whether the alignment of the end of previous block - is it
	 * aligned as well as the start of the block
	 */
	if (((current_block_end) & ((1ULL << block_shift) - 1ULL)) != 0)
		/*
		 * It is not as well aligned as the start of the block -
		 * reduce the mtt size accordingly.
		 */
		block_shift = alignment_of(current_block_end);

	return block_shift;
}

182 183 184 185
int mlx4_ib_umem_write_mtt(struct mlx4_ib_dev *dev, struct mlx4_mtt *mtt,
			   struct ib_umem *umem)
{
	u64 *pages;
186
	u64 len = 0;
187
	int err = 0;
188 189 190 191 192
	u64 mtt_size;
	u64 cur_start_addr = 0;
	u64 mtt_shift;
	int start_index = 0;
	int npages = 0;
193
	struct scatterlist *sg;
194
	int i;
195 196 197 198 199

	pages = (u64 *) __get_free_page(GFP_KERNEL);
	if (!pages)
		return -ENOMEM;

200 201
	mtt_shift = mtt->page_shift;
	mtt_size = 1ULL << mtt_shift;
202

203 204 205 206 207
	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, i) {
		if (cur_start_addr + len == sg_dma_address(sg)) {
			/* still the same block */
			len += sg_dma_len(sg);
			continue;
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
		/*
		 * A new block is started ...
		 * If len is malaligned, write an extra mtt entry to cover the
		 * misaligned area (round up the division)
		 */
		err = mlx4_ib_umem_write_mtt_block(dev, mtt, mtt_size,
						   mtt_shift, len,
						   cur_start_addr,
						   pages, &start_index,
						   &npages);
		if (err)
			goto out;

		cur_start_addr = sg_dma_address(sg);
		len = sg_dma_len(sg);
	}

	/* Handle the last block */
	if (len > 0) {
		/*
		 * If len is malaligned, write an extra mtt entry to cover
		 * the misaligned area (round up the division)
		 */
		err = mlx4_ib_umem_write_mtt_block(dev, mtt, mtt_size,
						   mtt_shift, len,
						   cur_start_addr, pages,
						   &start_index, &npages);
		if (err)
			goto out;
238
	}
239

240 241
	if (npages)
		err = mlx4_write_mtt(dev->dev, mtt, start_index, npages, pages);
242 243 244 245 246 247

out:
	free_page((unsigned long) pages);
	return err;
}

248 249 250 251 252 253 254 255 256
/*
 * Calculate optimal mtt size based on contiguous pages.
 * Function will return also the number of pages that are not aligned to the
 * calculated mtt_size to be added to total number of pages. For that we should
 * check the first chunk length & last chunk length and if not aligned to
 * mtt_size we should increment the non_aligned_pages number. All chunks in the
 * middle already handled as part of mtt shift calculation for both their start
 * & end addresses.
 */
257 258
int mlx4_ib_umem_calc_optimal_mtt_size(struct ib_umem *umem, u64 start_va,
				       int *num_of_mtts)
259 260
{
	u64 block_shift = MLX4_MAX_MTT_SHIFT;
261
	u64 min_shift = PAGE_SHIFT;
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
	u64 last_block_aligned_end = 0;
	u64 current_block_start = 0;
	u64 first_block_start = 0;
	u64 current_block_len = 0;
	u64 last_block_end = 0;
	struct scatterlist *sg;
	u64 current_block_end;
	u64 misalignment_bits;
	u64 next_block_start;
	u64 total_len = 0;
	int i;

	for_each_sg(umem->sg_head.sgl, sg, umem->nmap, i) {
		/*
		 * Initialization - save the first chunk start as the
		 * current_block_start - block means contiguous pages.
		 */
		if (current_block_len == 0 && current_block_start == 0) {
			current_block_start = sg_dma_address(sg);
			first_block_start = current_block_start;
			/*
			 * Find the bits that are different between the physical
			 * address and the virtual address for the start of the
			 * MR.
			 * umem_get aligned the start_va to a page boundary.
			 * Therefore, we need to align the start va to the same
			 * boundary.
			 * misalignment_bits is needed to handle the  case of a
			 * single memory region. In this case, the rest of the
			 * logic will not reduce the block size.  If we use a
			 * block size which is bigger than the alignment of the
			 * misalignment bits, we might use the virtual page
			 * number instead of the physical page number, resulting
			 * in access to the wrong data.
			 */
			misalignment_bits =
298 299
				(start_va & (~(((u64)(PAGE_SIZE)) - 1ULL))) ^
				current_block_start;
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
			block_shift = min(alignment_of(misalignment_bits),
					  block_shift);
		}

		/*
		 * Go over the scatter entries and check if they continue the
		 * previous scatter entry.
		 */
		next_block_start = sg_dma_address(sg);
		current_block_end = current_block_start	+ current_block_len;
		/* If we have a split (non-contig.) between two blocks */
		if (current_block_end != next_block_start) {
			block_shift = mlx4_ib_umem_calc_block_mtt
					(next_block_start,
					 current_block_end,
					 block_shift);

			/*
			 * If we reached the minimum shift for 4k page we stop
			 * the loop.
			 */
			if (block_shift <= min_shift)
				goto end;

			/*
			 * If not saved yet we are in first block - we save the
			 * length of first block to calculate the
			 * non_aligned_pages number at the end.
			 */
			total_len += current_block_len;

			/* Start a new block */
			current_block_start = next_block_start;
			current_block_len = sg_dma_len(sg);
			continue;
		}
		/* The scatter entry is another part of the current block,
		 * increase the block size.
		 * An entry in the scatter can be larger than 4k (page) as of
		 * dma mapping which merge some blocks together.
		 */
		current_block_len += sg_dma_len(sg);
	}

	/* Account for the last block in the total len */
	total_len += current_block_len;
	/* Add to the first block the misalignment that it suffers from. */
	total_len += (first_block_start & ((1ULL << block_shift) - 1ULL));
	last_block_end = current_block_start + current_block_len;
349
	last_block_aligned_end = round_up(last_block_end, 1ULL << block_shift);
350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369
	total_len += (last_block_aligned_end - last_block_end);

	if (total_len & ((1ULL << block_shift) - 1ULL))
		pr_warn("misaligned total length detected (%llu, %llu)!",
			total_len, block_shift);

	*num_of_mtts = total_len >> block_shift;
end:
	if (block_shift < min_shift) {
		/*
		 * If shift is less than the min we set a warning and return the
		 * min shift.
		 */
		pr_warn("umem_calc_optimal_mtt_size - unexpected shift %lld\n", block_shift);

		block_shift = min_shift;
	}
	return block_shift;
}

370
static struct ib_umem *mlx4_get_umem_mr(struct ib_udata *udata, u64 start,
371
					u64 length, int access_flags)
372 373 374 375 376 377 378 379
{
	/*
	 * Force registering the memory as writable if the underlying pages
	 * are writable.  This is so rereg can change the access permissions
	 * from readable to writable without having to run through ib_umem_get
	 * again
	 */
	if (!ib_access_writable(access_flags)) {
380
		unsigned long untagged_start = untagged_addr(start);
381 382 383 384 385 386 387 388
		struct vm_area_struct *vma;

		down_read(&current->mm->mmap_sem);
		/*
		 * FIXME: Ideally this would iterate over all the vmas that
		 * cover the memory, but for now it requires a single vma to
		 * entirely cover the MR to support RO mappings.
		 */
389 390 391
		vma = find_vma(current->mm, untagged_start);
		if (vma && vma->vm_end >= untagged_start + length &&
		    vma->vm_start <= untagged_start) {
392 393 394 395 396 397 398 399 400
			if (vma->vm_flags & VM_WRITE)
				access_flags |= IB_ACCESS_LOCAL_WRITE;
		} else {
			access_flags |= IB_ACCESS_LOCAL_WRITE;
		}

		up_read(&current->mm->mmap_sem);
	}

401
	return ib_umem_get(udata, start, length, access_flags, 0);
402 403
}

404 405 406 407 408 409 410 411 412 413
struct ib_mr *mlx4_ib_reg_user_mr(struct ib_pd *pd, u64 start, u64 length,
				  u64 virt_addr, int access_flags,
				  struct ib_udata *udata)
{
	struct mlx4_ib_dev *dev = to_mdev(pd->device);
	struct mlx4_ib_mr *mr;
	int shift;
	int err;
	int n;

414
	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
415 416 417
	if (!mr)
		return ERR_PTR(-ENOMEM);

418
	mr->umem = mlx4_get_umem_mr(udata, start, length, access_flags);
419 420 421 422 423 424
	if (IS_ERR(mr->umem)) {
		err = PTR_ERR(mr->umem);
		goto err_free;
	}

	n = ib_umem_page_count(mr->umem);
425
	shift = mlx4_ib_umem_calc_optimal_mtt_size(mr->umem, start, &n);
426 427 428 429 430 431 432 433 434 435 436 437 438 439 440

	err = mlx4_mr_alloc(dev->dev, to_mpd(pd)->pdn, virt_addr, length,
			    convert_access(access_flags), n, shift, &mr->mmr);
	if (err)
		goto err_umem;

	err = mlx4_ib_umem_write_mtt(dev, &mr->mmr.mtt, mr->umem);
	if (err)
		goto err_mr;

	err = mlx4_mr_enable(dev->dev, &mr->mmr);
	if (err)
		goto err_mr;

	mr->ibmr.rkey = mr->ibmr.lkey = mr->mmr.key;
441 442 443
	mr->ibmr.length = length;
	mr->ibmr.iova = virt_addr;
	mr->ibmr.page_size = 1U << shift;
444 445 446 447

	return &mr->ibmr;

err_mr:
448
	(void) mlx4_mr_free(to_mdev(pd->device)->dev, &mr->mmr);
449 450 451 452 453 454 455 456 457 458

err_umem:
	ib_umem_release(mr->umem);

err_free:
	kfree(mr);

	return ERR_PTR(err);
}

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
int mlx4_ib_rereg_user_mr(struct ib_mr *mr, int flags,
			  u64 start, u64 length, u64 virt_addr,
			  int mr_access_flags, struct ib_pd *pd,
			  struct ib_udata *udata)
{
	struct mlx4_ib_dev *dev = to_mdev(mr->device);
	struct mlx4_ib_mr *mmr = to_mmr(mr);
	struct mlx4_mpt_entry *mpt_entry;
	struct mlx4_mpt_entry **pmpt_entry = &mpt_entry;
	int err;

	/* Since we synchronize this call and mlx4_ib_dereg_mr via uverbs,
	 * we assume that the calls can't run concurrently. Otherwise, a
	 * race exists.
	 */
	err =  mlx4_mr_hw_get_mpt(dev->dev, &mmr->mmr, &pmpt_entry);

	if (err)
		return err;

	if (flags & IB_MR_REREG_PD) {
		err = mlx4_mr_hw_change_pd(dev->dev, *pmpt_entry,
					   to_mpd(pd)->pdn);

		if (err)
			goto release_mpt_entry;
	}

	if (flags & IB_MR_REREG_ACCESS) {
488 489 490 491 492
		if (ib_access_writable(mr_access_flags) &&
		    !mmr->umem->writable) {
			err = -EPERM;
			goto release_mpt_entry;
		}
493

494 495 496 497 498 499 500 501 502 503 504 505 506
		err = mlx4_mr_hw_change_access(dev->dev, *pmpt_entry,
					       convert_access(mr_access_flags));

		if (err)
			goto release_mpt_entry;
	}

	if (flags & IB_MR_REREG_TRANS) {
		int shift;
		int n;

		mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
		ib_umem_release(mmr->umem);
507
		mmr->umem = mlx4_get_umem_mr(udata, start, length,
508
					     mr_access_flags);
509 510
		if (IS_ERR(mmr->umem)) {
			err = PTR_ERR(mmr->umem);
511
			/* Prevent mlx4_ib_dereg_mr from free'ing invalid pointer */
512 513 514 515
			mmr->umem = NULL;
			goto release_mpt_entry;
		}
		n = ib_umem_page_count(mmr->umem);
516
		shift = PAGE_SHIFT;
517 518 519 520 521 522 523 524

		err = mlx4_mr_rereg_mem_write(dev->dev, &mmr->mmr,
					      virt_addr, length, n, shift,
					      *pmpt_entry);
		if (err) {
			ib_umem_release(mmr->umem);
			goto release_mpt_entry;
		}
525 526
		mmr->mmr.iova       = virt_addr;
		mmr->mmr.size       = length;
527 528 529 530 531 532 533 534 535 536 537 538 539

		err = mlx4_ib_umem_write_mtt(dev, &mmr->mmr.mtt, mmr->umem);
		if (err) {
			mlx4_mr_rereg_mem_cleanup(dev->dev, &mmr->mmr);
			ib_umem_release(mmr->umem);
			goto release_mpt_entry;
		}
	}

	/* If we couldn't transfer the MR to the HCA, just remember to
	 * return a failure. But dereg_mr will free the resources.
	 */
	err = mlx4_mr_hw_write_mpt(dev->dev, &mmr->mmr, pmpt_entry);
540 541
	if (!err && flags & IB_MR_REREG_ACCESS)
		mmr->mmr.access = mr_access_flags;
542 543 544 545 546 547 548

release_mpt_entry:
	mlx4_mr_hw_put_mpt(dev->dev, pmpt_entry);

	return err;
}

549 550 551 552 553 554 555
static int
mlx4_alloc_priv_pages(struct ib_device *device,
		      struct mlx4_ib_mr *mr,
		      int max_pages)
{
	int ret;

556 557 558 559 560 561 562
	/* Ensure that size is aligned to DMA cacheline
	 * requirements.
	 * max_pages is limited to MLX4_MAX_FAST_REG_PAGES
	 * so page_map_size will never cross PAGE_SIZE.
	 */
	mr->page_map_size = roundup(max_pages * sizeof(u64),
				    MLX4_MR_PAGES_ALIGN);
563

564 565 566
	/* Prevent cross page boundary allocation. */
	mr->pages = (__be64 *)get_zeroed_page(GFP_KERNEL);
	if (!mr->pages)
567 568
		return -ENOMEM;

569
	mr->page_map = dma_map_single(device->dev.parent, mr->pages,
570
				      mr->page_map_size, DMA_TO_DEVICE);
571

572
	if (dma_mapping_error(device->dev.parent, mr->page_map)) {
573 574 575 576 577 578
		ret = -ENOMEM;
		goto err;
	}

	return 0;

579 580
err:
	free_page((unsigned long)mr->pages);
581 582 583 584 585 586 587 588 589
	return ret;
}

static void
mlx4_free_priv_pages(struct mlx4_ib_mr *mr)
{
	if (mr->pages) {
		struct ib_device *device = mr->ibmr.device;

590
		dma_unmap_single(device->dev.parent, mr->page_map,
591 592
				 mr->page_map_size, DMA_TO_DEVICE);
		free_page((unsigned long)mr->pages);
593 594 595 596
		mr->pages = NULL;
	}
}

597
int mlx4_ib_dereg_mr(struct ib_mr *ibmr, struct ib_udata *udata)
598 599
{
	struct mlx4_ib_mr *mr = to_mmr(ibmr);
600
	int ret;
601

602 603
	mlx4_free_priv_pages(mr);

604 605 606
	ret = mlx4_mr_free(to_mdev(ibmr->device)->dev, &mr->mmr);
	if (ret)
		return ret;
607 608 609 610 611 612
	if (mr->umem)
		ib_umem_release(mr->umem);
	kfree(mr);

	return 0;
}
Jack Morgenstein's avatar
Jack Morgenstein committed
613

614 615
struct ib_mw *mlx4_ib_alloc_mw(struct ib_pd *pd, enum ib_mw_type type,
			       struct ib_udata *udata)
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
{
	struct mlx4_ib_dev *dev = to_mdev(pd->device);
	struct mlx4_ib_mw *mw;
	int err;

	mw = kmalloc(sizeof(*mw), GFP_KERNEL);
	if (!mw)
		return ERR_PTR(-ENOMEM);

	err = mlx4_mw_alloc(dev->dev, to_mpd(pd)->pdn,
			    to_mlx4_type(type), &mw->mmw);
	if (err)
		goto err_free;

	err = mlx4_mw_enable(dev->dev, &mw->mmw);
	if (err)
		goto err_mw;

	mw->ibmw.rkey = mw->mmw.key;

	return &mw->ibmw;

err_mw:
	mlx4_mw_free(dev->dev, &mw->mmw);

err_free:
	kfree(mw);

	return ERR_PTR(err);
}

int mlx4_ib_dealloc_mw(struct ib_mw *ibmw)
{
	struct mlx4_ib_mw *mw = to_mmw(ibmw);

	mlx4_mw_free(to_mdev(ibmw->device)->dev, &mw->mmw);
	kfree(mw);

	return 0;
}

657 658
struct ib_mr *mlx4_ib_alloc_mr(struct ib_pd *pd, enum ib_mr_type mr_type,
			       u32 max_num_sg, struct ib_udata *udata)
659 660 661 662 663
{
	struct mlx4_ib_dev *dev = to_mdev(pd->device);
	struct mlx4_ib_mr *mr;
	int err;

664 665 666 667
	if (mr_type != IB_MR_TYPE_MEM_REG ||
	    max_num_sg > MLX4_MAX_FAST_REG_PAGES)
		return ERR_PTR(-EINVAL);

668
	mr = kzalloc(sizeof(*mr), GFP_KERNEL);
669 670 671 672
	if (!mr)
		return ERR_PTR(-ENOMEM);

	err = mlx4_mr_alloc(dev->dev, to_mpd(pd)->pdn, 0, 0, 0,
673
			    max_num_sg, 0, &mr->mmr);
674 675 676
	if (err)
		goto err_free;

677 678 679 680 681
	err = mlx4_alloc_priv_pages(pd->device, mr, max_num_sg);
	if (err)
		goto err_free_mr;

	mr->max_pages = max_num_sg;
682 683
	err = mlx4_mr_enable(dev->dev, &mr->mmr);
	if (err)
684
		goto err_free_pl;
685

686
	mr->ibmr.rkey = mr->ibmr.lkey = mr->mmr.key;
687
	mr->umem = NULL;
688

689 690
	return &mr->ibmr;

691
err_free_pl:
692
	mr->ibmr.device = pd->device;
693 694
	mlx4_free_priv_pages(mr);
err_free_mr:
695
	(void) mlx4_mr_free(dev->dev, &mr->mmr);
696 697 698 699 700
err_free:
	kfree(mr);
	return ERR_PTR(err);
}

Jack Morgenstein's avatar
Jack Morgenstein committed
701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717
struct ib_fmr *mlx4_ib_fmr_alloc(struct ib_pd *pd, int acc,
				 struct ib_fmr_attr *fmr_attr)
{
	struct mlx4_ib_dev *dev = to_mdev(pd->device);
	struct mlx4_ib_fmr *fmr;
	int err = -ENOMEM;

	fmr = kmalloc(sizeof *fmr, GFP_KERNEL);
	if (!fmr)
		return ERR_PTR(-ENOMEM);

	err = mlx4_fmr_alloc(dev->dev, to_mpd(pd)->pdn, convert_access(acc),
			     fmr_attr->max_pages, fmr_attr->max_maps,
			     fmr_attr->page_shift, &fmr->mfmr);
	if (err)
		goto err_free;

718
	err = mlx4_fmr_enable(to_mdev(pd->device)->dev, &fmr->mfmr);
Jack Morgenstein's avatar
Jack Morgenstein committed
719 720 721 722 723 724 725 726
	if (err)
		goto err_mr;

	fmr->ibfmr.rkey = fmr->ibfmr.lkey = fmr->mfmr.mr.key;

	return &fmr->ibfmr;

err_mr:
727
	(void) mlx4_mr_free(to_mdev(pd->device)->dev, &fmr->mfmr.mr);
Jack Morgenstein's avatar
Jack Morgenstein committed
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

err_free:
	kfree(fmr);

	return ERR_PTR(err);
}

int mlx4_ib_map_phys_fmr(struct ib_fmr *ibfmr, u64 *page_list,
		      int npages, u64 iova)
{
	struct mlx4_ib_fmr *ifmr = to_mfmr(ibfmr);
	struct mlx4_ib_dev *dev = to_mdev(ifmr->ibfmr.device);

	return mlx4_map_phys_fmr(dev->dev, &ifmr->mfmr, page_list, npages, iova,
				 &ifmr->ibfmr.lkey, &ifmr->ibfmr.rkey);
}

int mlx4_ib_unmap_fmr(struct list_head *fmr_list)
{
	struct ib_fmr *ibfmr;
	int err;
	struct mlx4_dev *mdev = NULL;

	list_for_each_entry(ibfmr, fmr_list, list) {
		if (mdev && to_mdev(ibfmr->device)->dev != mdev)
			return -EINVAL;
		mdev = to_mdev(ibfmr->device)->dev;
	}

	if (!mdev)
		return 0;

	list_for_each_entry(ibfmr, fmr_list, list) {
		struct mlx4_ib_fmr *ifmr = to_mfmr(ibfmr);

		mlx4_fmr_unmap(mdev, &ifmr->mfmr, &ifmr->ibfmr.lkey, &ifmr->ibfmr.rkey);
	}

	/*
	 * Make sure all MPT status updates are visible before issuing
	 * SYNC_TPT firmware command.
	 */
	wmb();

	err = mlx4_SYNC_TPT(mdev);
	if (err)
774
		pr_warn("SYNC_TPT error %d when "
Jack Morgenstein's avatar
Jack Morgenstein committed
775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792
		       "unmapping FMRs\n", err);

	return 0;
}

int mlx4_ib_fmr_dealloc(struct ib_fmr *ibfmr)
{
	struct mlx4_ib_fmr *ifmr = to_mfmr(ibfmr);
	struct mlx4_ib_dev *dev = to_mdev(ibfmr->device);
	int err;

	err = mlx4_fmr_free(dev->dev, &ifmr->mfmr);

	if (!err)
		kfree(ifmr);

	return err;
}
793 794 795 796 797 798 799 800 801 802 803 804 805

static int mlx4_set_page(struct ib_mr *ibmr, u64 addr)
{
	struct mlx4_ib_mr *mr = to_mmr(ibmr);

	if (unlikely(mr->npages == mr->max_pages))
		return -ENOMEM;

	mr->pages[mr->npages++] = cpu_to_be64(addr | MLX4_MTT_FLAG_PRESENT);

	return 0;
}

806
int mlx4_ib_map_mr_sg(struct ib_mr *ibmr, struct scatterlist *sg, int sg_nents,
807
		      unsigned int *sg_offset)
808 809 810 811 812 813 814
{
	struct mlx4_ib_mr *mr = to_mmr(ibmr);
	int rc;

	mr->npages = 0;

	ib_dma_sync_single_for_cpu(ibmr->device, mr->page_map,
815
				   mr->page_map_size, DMA_TO_DEVICE);
816

817
	rc = ib_sg_to_pages(ibmr, sg, sg_nents, sg_offset, mlx4_set_page);
818 819

	ib_dma_sync_single_for_device(ibmr->device, mr->page_map,
820
				      mr->page_map_size, DMA_TO_DEVICE);
821 822 823

	return rc;
}