slice.c 22.0 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31
/*
 * address space "slices" (meta-segments) support
 *
 * Copyright (C) 2007 Benjamin Herrenschmidt, IBM Corporation.
 *
 * Based on hugetlb implementation
 *
 * Copyright (C) 2003 David Gibson, IBM Corporation.
 *
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License as published by
 * the Free Software Foundation; either version 2 of the License, or
 * (at your option) any later version.
 *
 * This program is distributed in the hope that it will be useful,
 * but WITHOUT ANY WARRANTY; without even the implied warranty of
 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
 * GNU General Public License for more details.
 *
 * You should have received a copy of the GNU General Public License
 * along with this program; if not, write to the Free Software
 * Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA  02111-1307  USA
 */

#undef DEBUG

#include <linux/kernel.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/err.h>
#include <linux/spinlock.h>
32
#include <linux/export.h>
33
#include <linux/hugetlb.h>
34 35
#include <asm/mman.h>
#include <asm/mmu.h>
36
#include <asm/copro.h>
37
#include <asm/hugetlb.h>
38

R
Roel Kluin 已提交
39
static DEFINE_SPINLOCK(slice_convert_lock);
40 41 42 43 44 45 46 47 48
/*
 * One bit per slice. We have lower slices which cover 256MB segments
 * upto 4G range. That gets us 16 low slices. For the rest we track slices
 * in 1TB size.
 */
struct slice_mask {
	u64 low_slices;
	DECLARE_BITMAP(high_slices, SLICE_NUM_HIGH);
};
49 50 51 52 53 54 55 56

#ifdef DEBUG
int _slice_debug = 1;

static void slice_print_mask(const char *label, struct slice_mask mask)
{
	if (!_slice_debug)
		return;
57 58
	pr_devel("%s low_slice: %*pbl\n", label, (int)SLICE_NUM_LOW, &mask.low_slices);
	pr_devel("%s high_slice: %*pbl\n", label, (int)SLICE_NUM_HIGH, mask.high_slices);
59 60
}

61
#define slice_dbg(fmt...) do { if (_slice_debug) pr_devel(fmt); } while (0)
62 63 64 65 66 67 68 69

#else

static void slice_print_mask(const char *label, struct slice_mask mask) {}
#define slice_dbg(fmt...)

#endif

70 71
static void slice_range_to_mask(unsigned long start, unsigned long len,
				struct slice_mask *ret)
72 73
{
	unsigned long end = start + len - 1;
74

75
	ret->low_slices = 0;
76 77
	if (SLICE_NUM_HIGH)
		bitmap_zero(ret->high_slices, SLICE_NUM_HIGH);
78 79

	if (start < SLICE_LOW_TOP) {
80 81
		unsigned long mend = min(end,
					 (unsigned long)(SLICE_LOW_TOP - 1));
82

83
		ret->low_slices = (1u << (GET_LOW_SLICE_INDEX(mend) + 1))
84
			- (1u << GET_LOW_SLICE_INDEX(start));
85 86
	}

87 88 89 90
	if ((start + len) > SLICE_LOW_TOP) {
		unsigned long start_index = GET_HIGH_SLICE_INDEX(start);
		unsigned long align_end = ALIGN(end, (1UL << SLICE_HIGH_SHIFT));
		unsigned long count = GET_HIGH_SLICE_INDEX(align_end) - start_index;
91

92
		bitmap_set(ret->high_slices, start_index, count);
93
	}
94 95 96 97 98 99 100
}

static int slice_area_is_free(struct mm_struct *mm, unsigned long addr,
			      unsigned long len)
{
	struct vm_area_struct *vma;

101
	if ((mm->context.slb_addr_limit - len) < addr)
102 103
		return 0;
	vma = find_vma(mm, addr);
104
	return (!vma || (addr + len) <= vm_start_gap(vma));
105 106 107 108 109 110 111 112 113 114 115 116 117
}

static int slice_low_has_vma(struct mm_struct *mm, unsigned long slice)
{
	return !slice_area_is_free(mm, slice << SLICE_LOW_SHIFT,
				   1ul << SLICE_LOW_SHIFT);
}

static int slice_high_has_vma(struct mm_struct *mm, unsigned long slice)
{
	unsigned long start = slice << SLICE_HIGH_SHIFT;
	unsigned long end = start + (1ul << SLICE_HIGH_SHIFT);

118
#ifdef CONFIG_PPC64
119 120 121 122 123
	/* Hack, so that each addresses is controlled by exactly one
	 * of the high or low area bitmaps, the first high area starts
	 * at 4GB, not 0 */
	if (start == 0)
		start = SLICE_LOW_TOP;
124
#endif
125 126 127 128

	return !slice_area_is_free(mm, start, end - start);
}

129 130
static void slice_mask_for_free(struct mm_struct *mm, struct slice_mask *ret,
				unsigned long high_limit)
131 132 133
{
	unsigned long i;

134
	ret->low_slices = 0;
135 136
	if (SLICE_NUM_HIGH)
		bitmap_zero(ret->high_slices, SLICE_NUM_HIGH);
137

138 139
	for (i = 0; i < SLICE_NUM_LOW; i++)
		if (!slice_low_has_vma(mm, i))
140
			ret->low_slices |= 1u << i;
141

142
	if (high_limit <= SLICE_LOW_TOP)
143
		return;
144

145
	for (i = 0; i < GET_HIGH_SLICE_INDEX(high_limit); i++)
146
		if (!slice_high_has_vma(mm, i))
147
			__set_bit(i, ret->high_slices);
148 149
}

150 151
static void slice_mask_for_size(struct mm_struct *mm, int psize, struct slice_mask *ret,
				unsigned long high_limit)
152
{
153 154
	unsigned char *hpsizes;
	int index, mask_index;
155
	unsigned long i;
156
	u64 lpsizes;
157

158
	ret->low_slices = 0;
159 160
	if (SLICE_NUM_HIGH)
		bitmap_zero(ret->high_slices, SLICE_NUM_HIGH);
161

162
	lpsizes = mm->context.low_slices_psize;
163
	for (i = 0; i < SLICE_NUM_LOW; i++)
164
		if (((lpsizes >> (i * 4)) & 0xf) == psize)
165
			ret->low_slices |= 1u << i;
166

167 168 169
	if (high_limit <= SLICE_LOW_TOP)
		return;

170
	hpsizes = mm->context.high_slices_psize;
171
	for (i = 0; i < GET_HIGH_SLICE_INDEX(high_limit); i++) {
172 173 174
		mask_index = i & 0x1;
		index = i >> 1;
		if (((hpsizes[index] >> (mask_index * 4)) & 0xf) == psize)
175
			__set_bit(i, ret->high_slices);
176
	}
177 178
}

179 180
static int slice_check_fit(struct mm_struct *mm,
			   struct slice_mask mask, struct slice_mask available)
181
{
182
	DECLARE_BITMAP(result, SLICE_NUM_HIGH);
183 184 185 186
	/*
	 * Make sure we just do bit compare only to the max
	 * addr limit and not the full bit map size.
	 */
187
	unsigned long slice_count = GET_HIGH_SLICE_INDEX(mm->context.slb_addr_limit);
188

189 190 191 192
	if (!SLICE_NUM_HIGH)
		return (mask.low_slices & available.low_slices) ==
		       mask.low_slices;

193
	bitmap_and(result, mask.high_slices,
194
		   available.high_slices, slice_count);
195

196
	return (mask.low_slices & available.low_slices) == mask.low_slices &&
197
		bitmap_equal(result, mask.high_slices, slice_count);
198 199 200 201
}

static void slice_flush_segments(void *parm)
{
202
#ifdef CONFIG_PPC64
203 204 205 206 207 208
	struct mm_struct *mm = parm;
	unsigned long flags;

	if (mm != current->active_mm)
		return;

209
	copy_mm_to_paca(current->active_mm);
210 211 212 213

	local_irq_save(flags);
	slb_flush_and_rebolt();
	local_irq_restore(flags);
214
#endif
215 216 217 218
}

static void slice_convert(struct mm_struct *mm, struct slice_mask mask, int psize)
{
219
	int index, mask_index;
220
	/* Write the new slice psize bits */
221 222
	unsigned char *hpsizes;
	u64 lpsizes;
223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238
	unsigned long i, flags;

	slice_dbg("slice_convert(mm=%p, psize=%d)\n", mm, psize);
	slice_print_mask(" mask", mask);

	/* We need to use a spinlock here to protect against
	 * concurrent 64k -> 4k demotion ...
	 */
	spin_lock_irqsave(&slice_convert_lock, flags);

	lpsizes = mm->context.low_slices_psize;
	for (i = 0; i < SLICE_NUM_LOW; i++)
		if (mask.low_slices & (1u << i))
			lpsizes = (lpsizes & ~(0xful << (i * 4))) |
				(((unsigned long)psize) << (i * 4));

239
	/* Assign the value back */
240
	mm->context.low_slices_psize = lpsizes;
241 242

	hpsizes = mm->context.high_slices_psize;
243
	for (i = 0; i < GET_HIGH_SLICE_INDEX(mm->context.slb_addr_limit); i++) {
244 245
		mask_index = i & 0x1;
		index = i >> 1;
246
		if (test_bit(i, mask.high_slices))
247 248 249 250
			hpsizes[index] = (hpsizes[index] &
					  ~(0xf << (mask_index * 4))) |
				(((unsigned long)psize) << (mask_index * 4));
	}
251 252

	slice_dbg(" lsps=%lx, hsps=%lx\n",
253 254
		  (unsigned long)mm->context.low_slices_psize,
		  (unsigned long)mm->context.high_slices_psize);
255 256 257

	spin_unlock_irqrestore(&slice_convert_lock, flags);

258
	copro_flush_all_slbs(mm);
259 260
}

261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281
/*
 * Compute which slice addr is part of;
 * set *boundary_addr to the start or end boundary of that slice
 * (depending on 'end' parameter);
 * return boolean indicating if the slice is marked as available in the
 * 'available' slice_mark.
 */
static bool slice_scan_available(unsigned long addr,
				 struct slice_mask available,
				 int end,
				 unsigned long *boundary_addr)
{
	unsigned long slice;
	if (addr < SLICE_LOW_TOP) {
		slice = GET_LOW_SLICE_INDEX(addr);
		*boundary_addr = (slice + end) << SLICE_LOW_SHIFT;
		return !!(available.low_slices & (1u << slice));
	} else {
		slice = GET_HIGH_SLICE_INDEX(addr);
		*boundary_addr = (slice + end) ?
			((slice + end) << SLICE_HIGH_SHIFT) : SLICE_LOW_TOP;
282
		return !!test_bit(slice, available.high_slices);
283 284 285
	}
}

286 287 288
static unsigned long slice_find_area_bottomup(struct mm_struct *mm,
					      unsigned long len,
					      struct slice_mask available,
289
					      int psize, unsigned long high_limit)
290 291
{
	int pshift = max_t(int, mmu_psize_defs[psize].shift, PAGE_SHIFT);
292 293
	unsigned long addr, found, next_end;
	struct vm_unmapped_area_info info;
294

295 296 297 298
	info.flags = 0;
	info.length = len;
	info.align_mask = PAGE_MASK & ((1ul << pshift) - 1);
	info.align_offset = 0;
299

300
	addr = TASK_UNMAPPED_BASE;
301 302 303 304
	/*
	 * Check till the allow max value for this mmap request
	 */
	while (addr < high_limit) {
305 306
		info.low_limit = addr;
		if (!slice_scan_available(addr, available, 1, &addr))
307
			continue;
308 309 310 311 312 313 314 315

 next_slice:
		/*
		 * At this point [info.low_limit; addr) covers
		 * available slices only and ends at a slice boundary.
		 * Check if we need to reduce the range, or if we can
		 * extend it to cover the next available slice.
		 */
316 317
		if (addr >= high_limit)
			addr = high_limit;
318 319 320
		else if (slice_scan_available(addr, available, 1, &next_end)) {
			addr = next_end;
			goto next_slice;
321
		}
322 323 324 325 326
		info.high_limit = addr;

		found = vm_unmapped_area(&info);
		if (!(found & ~PAGE_MASK))
			return found;
327 328 329 330 331 332 333 334
	}

	return -ENOMEM;
}

static unsigned long slice_find_area_topdown(struct mm_struct *mm,
					     unsigned long len,
					     struct slice_mask available,
335
					     int psize, unsigned long high_limit)
336 337
{
	int pshift = max_t(int, mmu_psize_defs[psize].shift, PAGE_SHIFT);
338 339 340 341 342 343 344
	unsigned long addr, found, prev;
	struct vm_unmapped_area_info info;

	info.flags = VM_UNMAPPED_AREA_TOPDOWN;
	info.length = len;
	info.align_mask = PAGE_MASK & ((1ul << pshift) - 1);
	info.align_offset = 0;
345 346

	addr = mm->mmap_base;
347 348 349 350 351 352
	/*
	 * If we are trying to allocate above DEFAULT_MAP_WINDOW
	 * Add the different to the mmap_base.
	 * Only for that request for which high_limit is above
	 * DEFAULT_MAP_WINDOW we should apply this.
	 */
353 354
	if (high_limit > DEFAULT_MAP_WINDOW)
		addr += mm->context.slb_addr_limit - DEFAULT_MAP_WINDOW;
355

356 357 358
	while (addr > PAGE_SIZE) {
		info.high_limit = addr;
		if (!slice_scan_available(addr - 1, available, 0, &addr))
359 360
			continue;

361
 prev_slice:
362
		/*
363 364 365 366
		 * At this point [addr; info.high_limit) covers
		 * available slices only and starts at a slice boundary.
		 * Check if we need to reduce the range, or if we can
		 * extend it to cover the previous available slice.
367
		 */
368 369 370 371 372 373 374
		if (addr < PAGE_SIZE)
			addr = PAGE_SIZE;
		else if (slice_scan_available(addr - 1, available, 0, &prev)) {
			addr = prev;
			goto prev_slice;
		}
		info.low_limit = addr;
375

376 377 378
		found = vm_unmapped_area(&info);
		if (!(found & ~PAGE_MASK))
			return found;
379 380 381 382 383 384 385 386
	}

	/*
	 * A failed mmap() very likely causes application failure,
	 * so fall back to the bottom-up function here. This scenario
	 * can happen with large stack limits and large mmap()
	 * allocations.
	 */
387
	return slice_find_area_bottomup(mm, len, available, psize, high_limit);
388 389 390 391 392
}


static unsigned long slice_find_area(struct mm_struct *mm, unsigned long len,
				     struct slice_mask mask, int psize,
393
				     int topdown, unsigned long high_limit)
394 395
{
	if (topdown)
396
		return slice_find_area_topdown(mm, len, mask, psize, high_limit);
397
	else
398
		return slice_find_area_bottomup(mm, len, mask, psize, high_limit);
399 400
}

401 402 403
static inline void slice_or_mask(struct slice_mask *dst, struct slice_mask *src)
{
	dst->low_slices |= src->low_slices;
404 405
	if (!SLICE_NUM_HIGH)
		return;
406 407
	bitmap_or(dst->high_slices, dst->high_slices, src->high_slices,
		  SLICE_NUM_HIGH);
408 409 410 411 412 413
}

static inline void slice_andnot_mask(struct slice_mask *dst, struct slice_mask *src)
{
	dst->low_slices &= ~src->low_slices;

414 415
	if (!SLICE_NUM_HIGH)
		return;
416 417
	bitmap_andnot(dst->high_slices, dst->high_slices, src->high_slices,
		      SLICE_NUM_HIGH);
418
}
419 420 421 422 423 424 425

#ifdef CONFIG_PPC_64K_PAGES
#define MMU_PAGE_BASE	MMU_PAGE_64K
#else
#define MMU_PAGE_BASE	MMU_PAGE_4K
#endif

426 427
unsigned long slice_get_unmapped_area(unsigned long addr, unsigned long len,
				      unsigned long flags, unsigned int psize,
428
				      int topdown)
429
{
430
	struct slice_mask mask;
431
	struct slice_mask good_mask;
432 433
	struct slice_mask potential_mask;
	struct slice_mask compat_mask;
434 435
	int fixed = (flags & MAP_FIXED);
	int pshift = max_t(int, mmu_psize_defs[psize].shift, PAGE_SHIFT);
436
	unsigned long page_size = 1UL << pshift;
437
	struct mm_struct *mm = current->mm;
438
	unsigned long newaddr;
439
	unsigned long high_limit;
440

441
	high_limit = DEFAULT_MAP_WINDOW;
442
	if (addr >= high_limit || (fixed && (addr + len > high_limit)))
443 444 445 446 447 448 449 450 451 452 453 454 455
		high_limit = TASK_SIZE;

	if (len > high_limit)
		return -ENOMEM;
	if (len & (page_size - 1))
		return -EINVAL;
	if (fixed) {
		if (addr & (page_size - 1))
			return -EINVAL;
		if (addr > high_limit - len)
			return -ENOMEM;
	}

456 457
	if (high_limit > mm->context.slb_addr_limit) {
		mm->context.slb_addr_limit = high_limit;
458 459
		on_each_cpu(slice_flush_segments, mm, 1);
	}
460

461 462 463 464 465 466 467 468 469
	/*
	 * init different masks
	 */
	mask.low_slices = 0;

	/* silence stupid warning */;
	potential_mask.low_slices = 0;

	compat_mask.low_slices = 0;
470 471 472 473 474 475

	if (SLICE_NUM_HIGH) {
		bitmap_zero(mask.high_slices, SLICE_NUM_HIGH);
		bitmap_zero(potential_mask.high_slices, SLICE_NUM_HIGH);
		bitmap_zero(compat_mask.high_slices, SLICE_NUM_HIGH);
	}
476

477 478
	/* Sanity checks */
	BUG_ON(mm->task_size == 0);
479
	BUG_ON(mm->context.slb_addr_limit == 0);
480
	VM_BUG_ON(radix_enabled());
481 482

	slice_dbg("slice_get_unmapped_area(mm=%p, psize=%d...\n", mm, psize);
483 484
	slice_dbg(" addr=%lx, len=%lx, flags=%lx, topdown=%d\n",
		  addr, len, flags, topdown);
485 486 487

	/* If hint, make sure it matches our alignment restrictions */
	if (!fixed && addr) {
488
		addr = _ALIGN_UP(addr, page_size);
489
		slice_dbg(" aligned addr=%lx\n", addr);
490
		/* Ignore hint if it's too large or overlaps a VMA */
491
		if (addr > high_limit - len ||
492 493
		    !slice_area_is_free(mm, addr, len))
			addr = 0;
494 495
	}

496
	/* First make up a "good" mask of slices that have the right size
497 498
	 * already
	 */
499
	slice_mask_for_size(mm, psize, &good_mask, high_limit);
500 501
	slice_print_mask(" good_mask", good_mask);

502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519
	/*
	 * Here "good" means slices that are already the right page size,
	 * "compat" means slices that have a compatible page size (i.e.
	 * 4k in a 64k pagesize kernel), and "free" means slices without
	 * any VMAs.
	 *
	 * If MAP_FIXED:
	 *	check if fits in good | compat => OK
	 *	check if fits in good | compat | free => convert free
	 *	else bad
	 * If have hint:
	 *	check if hint fits in good => OK
	 *	check if hint fits in good | free => convert free
	 * Otherwise:
	 *	search in good, found => OK
	 *	search in good | free, found => convert free
	 *	search in good | compat | free, found => convert free.
	 */
520

521 522 523
#ifdef CONFIG_PPC_64K_PAGES
	/* If we support combo pages, we can allow 64k pages in 4k slices */
	if (psize == MMU_PAGE_64K) {
524
		slice_mask_for_size(mm, MMU_PAGE_4K, &compat_mask, high_limit);
525
		if (fixed)
526
			slice_or_mask(&good_mask, &compat_mask);
527 528
	}
#endif
529

530 531
	/* First check hint if it's valid or if we have MAP_FIXED */
	if (addr != 0 || fixed) {
532
		/* Build a mask for the requested range */
533
		slice_range_to_mask(addr, len, &mask);
534 535 536 537 538
		slice_print_mask(" mask", mask);

		/* Check if we fit in the good mask. If we do, we just return,
		 * nothing else to do
		 */
539
		if (slice_check_fit(mm, mask, good_mask)) {
540 541 542
			slice_dbg(" fits good !\n");
			return addr;
		}
543 544 545
	} else {
		/* Now let's see if we can find something in the existing
		 * slices for that size
546
		 */
547 548
		newaddr = slice_find_area(mm, len, good_mask,
					  psize, topdown, high_limit);
549 550 551 552 553 554
		if (newaddr != -ENOMEM) {
			/* Found within the good mask, we don't have to setup,
			 * we thus return directly
			 */
			slice_dbg(" found area at 0x%lx\n", newaddr);
			return newaddr;
555 556
		}
	}
557 558
	/*
	 * We don't fit in the good mask, check what other slices are
559 560
	 * empty and thus can be converted
	 */
561
	slice_mask_for_free(mm, &potential_mask, high_limit);
562
	slice_or_mask(&potential_mask, &good_mask);
563 564
	slice_print_mask(" potential", potential_mask);

565
	if ((addr != 0 || fixed) && slice_check_fit(mm, mask, potential_mask)) {
566 567 568 569 570
		slice_dbg(" fits potential !\n");
		goto convert;
	}

	/* If we have MAP_FIXED and failed the above steps, then error out */
571 572 573 574 575
	if (fixed)
		return -EBUSY;

	slice_dbg(" search...\n");

576 577
	/* If we had a hint that didn't work out, see if we can fit
	 * anywhere in the good area.
578
	 */
579
	if (addr) {
580 581
		addr = slice_find_area(mm, len, good_mask,
				       psize, topdown, high_limit);
582 583 584 585
		if (addr != -ENOMEM) {
			slice_dbg(" found area at 0x%lx\n", addr);
			return addr;
		}
586 587 588
	}

	/* Now let's see if we can find something in the existing slices
589
	 * for that size plus free slices
590
	 */
591 592
	addr = slice_find_area(mm, len, potential_mask,
			       psize, topdown, high_limit);
593 594 595 596

#ifdef CONFIG_PPC_64K_PAGES
	if (addr == -ENOMEM && psize == MMU_PAGE_64K) {
		/* retry the search with 4k-page slices included */
597
		slice_or_mask(&potential_mask, &compat_mask);
598 599
		addr = slice_find_area(mm, len, potential_mask,
				       psize, topdown, high_limit);
600 601 602
	}
#endif

603 604 605
	if (addr == -ENOMEM)
		return -ENOMEM;

606
	slice_range_to_mask(addr, len, &mask);
607 608 609 610
	slice_dbg(" found potential area at 0x%lx\n", addr);
	slice_print_mask(" mask", mask);

 convert:
611 612
	slice_andnot_mask(&mask, &good_mask);
	slice_andnot_mask(&mask, &compat_mask);
613 614 615
	if (mask.low_slices ||
	    (SLICE_NUM_HIGH &&
	     !bitmap_empty(mask.high_slices, SLICE_NUM_HIGH))) {
616 617
		slice_convert(mm, mask, psize);
		if (psize > MMU_PAGE_BASE)
618
			on_each_cpu(slice_flush_segments, mm, 1);
619
	}
620 621 622 623 624 625 626 627 628 629 630 631
	return addr;

}
EXPORT_SYMBOL_GPL(slice_get_unmapped_area);

unsigned long arch_get_unmapped_area(struct file *filp,
				     unsigned long addr,
				     unsigned long len,
				     unsigned long pgoff,
				     unsigned long flags)
{
	return slice_get_unmapped_area(addr, len, flags,
632
				       current->mm->context.user_psize, 0);
633 634 635 636 637 638 639 640 641
}

unsigned long arch_get_unmapped_area_topdown(struct file *filp,
					     const unsigned long addr0,
					     const unsigned long len,
					     const unsigned long pgoff,
					     const unsigned long flags)
{
	return slice_get_unmapped_area(addr0, len, flags,
642
				       current->mm->context.user_psize, 1);
643 644 645 646
}

unsigned int get_slice_psize(struct mm_struct *mm, unsigned long addr)
{
647 648
	unsigned char *hpsizes;
	int index, mask_index;
649

650 651 652 653 654 655 656 657 658 659
	/*
	 * Radix doesn't use slice, but can get enabled along with MMU_SLICE
	 */
	if (radix_enabled()) {
#ifdef CONFIG_PPC_64K_PAGES
		return MMU_PAGE_64K;
#else
		return MMU_PAGE_4K;
#endif
	}
660
	if (addr < SLICE_LOW_TOP) {
661 662
		u64 lpsizes;
		lpsizes = mm->context.low_slices_psize;
663
		index = GET_LOW_SLICE_INDEX(addr);
664
		return (lpsizes >> (index * 4)) & 0xf;
665
	}
666 667 668 669
	hpsizes = mm->context.high_slices_psize;
	index = GET_HIGH_SLICE_INDEX(addr);
	mask_index = index & 0x1;
	return (hpsizes[index >> 1] >> (mask_index * 4)) & 0xf;
670 671 672 673 674 675 676 677 678 679 680
}
EXPORT_SYMBOL_GPL(get_slice_psize);

/*
 * This is called by hash_page when it needs to do a lazy conversion of
 * an address space from real 64K pages to combo 4K pages (typically
 * when hitting a non cacheable mapping on a processor or hypervisor
 * that won't allow them for 64K pages).
 *
 * This is also called in init_new_context() to change back the user
 * psize from whatever the parent context had it set to
681
 * N.B. This may be called before mm->context.id has been set.
682 683 684 685 686 687 688
 *
 * This function will only change the content of the {low,high)_slice_psize
 * masks, it will not flush SLBs as this shall be handled lazily by the
 * caller.
 */
void slice_set_user_psize(struct mm_struct *mm, unsigned int psize)
{
689 690 691
	int index, mask_index;
	unsigned char *hpsizes;
	unsigned long flags, lpsizes;
692 693 694 695 696
	unsigned int old_psize;
	int i;

	slice_dbg("slice_set_user_psize(mm=%p, psize=%d)\n", mm, psize);

697
	VM_BUG_ON(radix_enabled());
698 699 700 701 702 703 704 705 706 707 708 709 710 711 712
	spin_lock_irqsave(&slice_convert_lock, flags);

	old_psize = mm->context.user_psize;
	slice_dbg(" old_psize=%d\n", old_psize);
	if (old_psize == psize)
		goto bail;

	mm->context.user_psize = psize;
	wmb();

	lpsizes = mm->context.low_slices_psize;
	for (i = 0; i < SLICE_NUM_LOW; i++)
		if (((lpsizes >> (i * 4)) & 0xf) == old_psize)
			lpsizes = (lpsizes & ~(0xful << (i * 4))) |
				(((unsigned long)psize) << (i * 4));
713 714
	/* Assign the value back */
	mm->context.low_slices_psize = lpsizes;
715 716

	hpsizes = mm->context.high_slices_psize;
717 718 719 720 721 722 723 724 725 726
	for (i = 0; i < SLICE_NUM_HIGH; i++) {
		mask_index = i & 0x1;
		index = i >> 1;
		if (((hpsizes[index] >> (mask_index * 4)) & 0xf) == old_psize)
			hpsizes[index] = (hpsizes[index] &
					  ~(0xf << (mask_index * 4))) |
				(((unsigned long)psize) << (mask_index * 4));
	}


727 728 729


	slice_dbg(" lsps=%lx, hsps=%lx\n",
730 731
		  (unsigned long)mm->context.low_slices_psize,
		  (unsigned long)mm->context.high_slices_psize);
732 733 734 735 736

 bail:
	spin_unlock_irqrestore(&slice_convert_lock, flags);
}

737 738 739
void slice_set_range_psize(struct mm_struct *mm, unsigned long start,
			   unsigned long len, unsigned int psize)
{
740
	struct slice_mask mask;
741

742
	VM_BUG_ON(radix_enabled());
743 744

	slice_range_to_mask(start, len, &mask);
745 746 747
	slice_convert(mm, mask, psize);
}

748
#ifdef CONFIG_HUGETLB_PAGE
749
/*
750
 * is_hugepage_only_range() is used by generic code to verify whether
751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771
 * a normal mmap mapping (non hugetlbfs) is valid on a given area.
 *
 * until the generic code provides a more generic hook and/or starts
 * calling arch get_unmapped_area for MAP_FIXED (which our implementation
 * here knows how to deal with), we hijack it to keep standard mappings
 * away from us.
 *
 * because of that generic code limitation, MAP_FIXED mapping cannot
 * "convert" back a slice with no VMAs to the standard page size, only
 * get_unmapped_area() can. It would be possible to fix it here but I
 * prefer working on fixing the generic code instead.
 *
 * WARNING: This will not work if hugetlbfs isn't enabled since the
 * generic code will redefine that function as 0 in that. This is ok
 * for now as we only use slices with hugetlbfs enabled. This should
 * be fixed as the generic code gets fixed.
 */
int is_hugepage_only_range(struct mm_struct *mm, unsigned long addr,
			   unsigned long len)
{
	struct slice_mask mask, available;
772
	unsigned int psize = mm->context.user_psize;
773
	unsigned long high_limit = mm->context.slb_addr_limit;
774

775 776 777
	if (radix_enabled())
		return 0;

778
	slice_range_to_mask(addr, len, &mask);
779
	slice_mask_for_size(mm, psize, &available, high_limit);
780 781 782 783
#ifdef CONFIG_PPC_64K_PAGES
	/* We need to account for 4k slices too */
	if (psize == MMU_PAGE_64K) {
		struct slice_mask compat_mask;
784
		slice_mask_for_size(mm, MMU_PAGE_4K, &compat_mask, high_limit);
785
		slice_or_mask(&available, &compat_mask);
786 787
	}
#endif
788 789 790 791 792 793 794

#if 0 /* too verbose */
	slice_dbg("is_hugepage_only_range(mm=%p, addr=%lx, len=%lx)\n",
		 mm, addr, len);
	slice_print_mask(" mask", mask);
	slice_print_mask(" available", available);
#endif
795
	return !slice_check_fit(mm, mask, available);
796
}
797
#endif