mprotect.c 13.9 KB
Newer Older
1
// SPDX-License-Identifier: GPL-2.0
Linus Torvalds's avatar
Linus Torvalds committed
2 3 4 5 6 7
/*
 *  mm/mprotect.c
 *
 *  (C) Copyright 1994 Linus Torvalds
 *  (C) Copyright 2002 Christoph Hellwig
 *
Alan Cox's avatar
Alan Cox committed
8
 *  Address space accounting code	<alan@lxorguk.ukuu.org.uk>
Linus Torvalds's avatar
Linus Torvalds committed
9 10 11 12 13 14 15 16 17 18 19 20 21
 *  (C) Copyright 2002 Red Hat Inc, All Rights Reserved
 */

#include <linux/mm.h>
#include <linux/hugetlb.h>
#include <linux/shm.h>
#include <linux/mman.h>
#include <linux/fs.h>
#include <linux/highmem.h>
#include <linux/security.h>
#include <linux/mempolicy.h>
#include <linux/personality.h>
#include <linux/syscalls.h>
22 23
#include <linux/swap.h>
#include <linux/swapops.h>
Andrea Arcangeli's avatar
Andrea Arcangeli committed
24
#include <linux/mmu_notifier.h>
25
#include <linux/migrate.h>
26
#include <linux/perf_event.h>
27
#include <linux/pkeys.h>
28
#include <linux/ksm.h>
29
#include <linux/uaccess.h>
Linus Torvalds's avatar
Linus Torvalds committed
30 31
#include <asm/pgtable.h>
#include <asm/cacheflush.h>
32
#include <asm/mmu_context.h>
Linus Torvalds's avatar
Linus Torvalds committed
33 34
#include <asm/tlbflush.h>

35 36
#include "internal.h"

37
static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
38
		unsigned long addr, unsigned long end, pgprot_t newprot,
39
		int dirty_accountable, int prot_numa)
Linus Torvalds's avatar
Linus Torvalds committed
40
{
41
	struct mm_struct *mm = vma->vm_mm;
42
	pte_t *pte, oldpte;
43
	spinlock_t *ptl;
44
	unsigned long pages = 0;
45
	int target_node = NUMA_NO_NODE;
Linus Torvalds's avatar
Linus Torvalds committed
46

47 48 49 50 51 52 53 54 55 56 57 58 59 60 61
	/*
	 * Can be called with only the mmap_sem for reading by
	 * prot_numa so we must check the pmd isn't constantly
	 * changing from under us from pmd_none to pmd_trans_huge
	 * and/or the other way around.
	 */
	if (pmd_trans_unstable(pmd))
		return 0;

	/*
	 * The pmd points to a regular pte so the pmd can't change
	 * from under us even if the mmap_sem is only hold for
	 * reading.
	 */
	pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
62

63 64 65 66 67
	/* Get target node for single threaded private VMAs */
	if (prot_numa && !(vma->vm_flags & VM_SHARED) &&
	    atomic_read(&vma->vm_mm->mm_users) == 1)
		target_node = numa_node_id();

68
	flush_tlb_batched_pending(vma->vm_mm);
69
	arch_enter_lazy_mmu_mode();
Linus Torvalds's avatar
Linus Torvalds committed
70
	do {
71 72
		oldpte = *pte;
		if (pte_present(oldpte)) {
Linus Torvalds's avatar
Linus Torvalds committed
73
			pte_t ptent;
74
			bool preserve_write = prot_numa && pte_write(oldpte);
Linus Torvalds's avatar
Linus Torvalds committed
75

76 77 78 79 80 81 82 83 84 85
			/*
			 * Avoid trapping faults against the zero or KSM
			 * pages. See similar comment in change_huge_pmd.
			 */
			if (prot_numa) {
				struct page *page;

				page = vm_normal_page(vma, addr, oldpte);
				if (!page || PageKsm(page))
					continue;
86 87 88 89

				/* Avoid TLB flush if possible */
				if (pte_protnone(oldpte))
					continue;
90 91 92 93 94 95 96

				/*
				 * Don't mess with PTEs if page is already on the node
				 * a single-threaded process is running on.
				 */
				if (target_node == page_to_nid(page))
					continue;
97 98
			}

99 100
			ptent = ptep_modify_prot_start(mm, addr, pte);
			ptent = pte_modify(ptent, newprot);
101
			if (preserve_write)
102
				ptent = pte_mk_savedwrite(ptent);
103

104 105 106 107 108
			/* Avoid taking write faults for known dirty pages */
			if (dirty_accountable && pte_dirty(ptent) &&
					(pte_soft_dirty(ptent) ||
					 !(vma->vm_flags & VM_SOFTDIRTY))) {
				ptent = pte_mkwrite(ptent);
109
			}
110 111
			ptep_modify_prot_commit(mm, addr, pte, ptent);
			pages++;
112
		} else if (IS_ENABLED(CONFIG_MIGRATION)) {
113 114 115
			swp_entry_t entry = pte_to_swp_entry(oldpte);

			if (is_write_migration_entry(entry)) {
116
				pte_t newpte;
117 118 119 120 121
				/*
				 * A protection check is difficult so
				 * just be safe and disable write
				 */
				make_migration_entry_read(&entry);
122 123 124 125
				newpte = swp_entry_to_pte(entry);
				if (pte_swp_soft_dirty(oldpte))
					newpte = pte_swp_mksoft_dirty(newpte);
				set_pte_at(mm, addr, pte, newpte);
126 127

				pages++;
128
			}
129 130 131 132 133 134 135 136 137 138 139 140 141 142

			if (is_write_device_private_entry(entry)) {
				pte_t newpte;

				/*
				 * We do not preserve soft-dirtiness. See
				 * copy_one_pte() for explanation.
				 */
				make_device_private_entry_read(&entry);
				newpte = swp_entry_to_pte(entry);
				set_pte_at(mm, addr, pte, newpte);

				pages++;
			}
Linus Torvalds's avatar
Linus Torvalds committed
143 144
		}
	} while (pte++, addr += PAGE_SIZE, addr != end);
145
	arch_leave_lazy_mmu_mode();
146
	pte_unmap_unlock(pte - 1, ptl);
147 148

	return pages;
Linus Torvalds's avatar
Linus Torvalds committed
149 150
}

151 152 153
static inline unsigned long change_pmd_range(struct vm_area_struct *vma,
		pud_t *pud, unsigned long addr, unsigned long end,
		pgprot_t newprot, int dirty_accountable, int prot_numa)
Linus Torvalds's avatar
Linus Torvalds committed
154 155
{
	pmd_t *pmd;
156
	struct mm_struct *mm = vma->vm_mm;
Linus Torvalds's avatar
Linus Torvalds committed
157
	unsigned long next;
158
	unsigned long pages = 0;
159
	unsigned long nr_huge_updates = 0;
160
	unsigned long mni_start = 0;
Linus Torvalds's avatar
Linus Torvalds committed
161 162 163

	pmd = pmd_offset(pud, addr);
	do {
164 165
		unsigned long this_pages;

Linus Torvalds's avatar
Linus Torvalds committed
166
		next = pmd_addr_end(addr, end);
167
		if (!is_swap_pmd(*pmd) && !pmd_trans_huge(*pmd) && !pmd_devmap(*pmd)
168
				&& pmd_none_or_clear_bad(pmd))
169
			continue;
170 171 172 173 174 175 176

		/* invoke the mmu notifier if the pmd is populated */
		if (!mni_start) {
			mni_start = addr;
			mmu_notifier_invalidate_range_start(mm, mni_start, end);
		}

177
		if (is_swap_pmd(*pmd) || pmd_trans_huge(*pmd) || pmd_devmap(*pmd)) {
178
			if (next - addr != HPAGE_PMD_SIZE) {
179
				__split_huge_pmd(vma, pmd, addr, false, NULL);
180
			} else {
181
				int nr_ptes = change_huge_pmd(vma, pmd, addr,
182
						newprot, prot_numa);
183 184

				if (nr_ptes) {
185 186 187 188
					if (nr_ptes == HPAGE_PMD_NR) {
						pages += HPAGE_PMD_NR;
						nr_huge_updates++;
					}
189 190

					/* huge pmd was handled */
191 192
					continue;
				}
193
			}
194
			/* fall through, the trans huge pmd just split */
195
		}
196
		this_pages = change_pte_range(vma, pmd, addr, next, newprot,
197
				 dirty_accountable, prot_numa);
198
		pages += this_pages;
Linus Torvalds's avatar
Linus Torvalds committed
199
	} while (pmd++, addr = next, addr != end);
200

201 202 203
	if (mni_start)
		mmu_notifier_invalidate_range_end(mm, mni_start, end);

204 205
	if (nr_huge_updates)
		count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates);
206
	return pages;
Linus Torvalds's avatar
Linus Torvalds committed
207 208
}

209
static inline unsigned long change_pud_range(struct vm_area_struct *vma,
210
		p4d_t *p4d, unsigned long addr, unsigned long end,
211
		pgprot_t newprot, int dirty_accountable, int prot_numa)
Linus Torvalds's avatar
Linus Torvalds committed
212 213 214
{
	pud_t *pud;
	unsigned long next;
215
	unsigned long pages = 0;
Linus Torvalds's avatar
Linus Torvalds committed
216

217
	pud = pud_offset(p4d, addr);
Linus Torvalds's avatar
Linus Torvalds committed
218 219 220 221
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
222
		pages += change_pmd_range(vma, pud, addr, next, newprot,
223
				 dirty_accountable, prot_numa);
Linus Torvalds's avatar
Linus Torvalds committed
224
	} while (pud++, addr = next, addr != end);
225 226

	return pages;
Linus Torvalds's avatar
Linus Torvalds committed
227 228
}

229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248
static inline unsigned long change_p4d_range(struct vm_area_struct *vma,
		pgd_t *pgd, unsigned long addr, unsigned long end,
		pgprot_t newprot, int dirty_accountable, int prot_numa)
{
	p4d_t *p4d;
	unsigned long next;
	unsigned long pages = 0;

	p4d = p4d_offset(pgd, addr);
	do {
		next = p4d_addr_end(addr, end);
		if (p4d_none_or_clear_bad(p4d))
			continue;
		pages += change_pud_range(vma, p4d, addr, next, newprot,
				 dirty_accountable, prot_numa);
	} while (p4d++, addr = next, addr != end);

	return pages;
}

249
static unsigned long change_protection_range(struct vm_area_struct *vma,
250
		unsigned long addr, unsigned long end, pgprot_t newprot,
251
		int dirty_accountable, int prot_numa)
Linus Torvalds's avatar
Linus Torvalds committed
252 253 254 255 256
{
	struct mm_struct *mm = vma->vm_mm;
	pgd_t *pgd;
	unsigned long next;
	unsigned long start = addr;
257
	unsigned long pages = 0;
Linus Torvalds's avatar
Linus Torvalds committed
258 259 260 261

	BUG_ON(addr >= end);
	pgd = pgd_offset(mm, addr);
	flush_cache_range(vma, addr, end);
262
	inc_tlb_flush_pending(mm);
Linus Torvalds's avatar
Linus Torvalds committed
263 264 265 266
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
267
		pages += change_p4d_range(vma, pgd, addr, next, newprot,
268
				 dirty_accountable, prot_numa);
Linus Torvalds's avatar
Linus Torvalds committed
269
	} while (pgd++, addr = next, addr != end);
270

271 272 273
	/* Only flush the TLB if we actually modified any entries: */
	if (pages)
		flush_tlb_range(vma, start, end);
274
	dec_tlb_flush_pending(mm);
275 276 277 278 279 280

	return pages;
}

unsigned long change_protection(struct vm_area_struct *vma, unsigned long start,
		       unsigned long end, pgprot_t newprot,
281
		       int dirty_accountable, int prot_numa)
282 283 284 285 286 287
{
	unsigned long pages;

	if (is_vm_hugetlb_page(vma))
		pages = hugetlb_change_protection(vma, start, end, newprot);
	else
288
		pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa);
289 290

	return pages;
Linus Torvalds's avatar
Linus Torvalds committed
291 292
}

293
int
Linus Torvalds's avatar
Linus Torvalds committed
294 295 296 297 298 299 300 301 302
mprotect_fixup(struct vm_area_struct *vma, struct vm_area_struct **pprev,
	unsigned long start, unsigned long end, unsigned long newflags)
{
	struct mm_struct *mm = vma->vm_mm;
	unsigned long oldflags = vma->vm_flags;
	long nrpages = (end - start) >> PAGE_SHIFT;
	unsigned long charged = 0;
	pgoff_t pgoff;
	int error;
303
	int dirty_accountable = 0;
Linus Torvalds's avatar
Linus Torvalds committed
304 305 306 307 308 309 310 311 312

	if (newflags == oldflags) {
		*pprev = vma;
		return 0;
	}

	/*
	 * If we make a private mapping writable we increase our commit;
	 * but (without finer accounting) cannot reduce our commit if we
313 314
	 * make it unwritable again. hugetlb mapping were accounted for
	 * even if read-only so there is no need to account for them here
Linus Torvalds's avatar
Linus Torvalds committed
315 316
	 */
	if (newflags & VM_WRITE) {
317 318 319 320
		/* Check space limits when area turns into data. */
		if (!may_expand_vm(mm, newflags, nrpages) &&
				may_expand_vm(mm, oldflags, nrpages))
			return -ENOMEM;
321
		if (!(oldflags & (VM_ACCOUNT|VM_WRITE|VM_HUGETLB|
322
						VM_SHARED|VM_NORESERVE))) {
Linus Torvalds's avatar
Linus Torvalds committed
323
			charged = nrpages;
324
			if (security_vm_enough_memory_mm(mm, charged))
Linus Torvalds's avatar
Linus Torvalds committed
325 326 327 328 329 330 331 332 333 334
				return -ENOMEM;
			newflags |= VM_ACCOUNT;
		}
	}

	/*
	 * First try to merge with previous and/or next vma.
	 */
	pgoff = vma->vm_pgoff + ((start - vma->vm_start) >> PAGE_SHIFT);
	*pprev = vma_merge(mm, *pprev, start, end, newflags,
335 336
			   vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma),
			   vma->vm_userfaultfd_ctx);
Linus Torvalds's avatar
Linus Torvalds committed
337 338
	if (*pprev) {
		vma = *pprev;
339
		VM_WARN_ON((vma->vm_flags ^ newflags) & ~VM_SOFTDIRTY);
Linus Torvalds's avatar
Linus Torvalds committed
340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
		goto success;
	}

	*pprev = vma;

	if (start != vma->vm_start) {
		error = split_vma(mm, vma, start, 1);
		if (error)
			goto fail;
	}

	if (end != vma->vm_end) {
		error = split_vma(mm, vma, end, 0);
		if (error)
			goto fail;
	}

success:
	/*
	 * vm_flags and vm_page_prot are protected by the mmap_sem
	 * held in write mode.
	 */
	vma->vm_flags = newflags;
363
	dirty_accountable = vma_wants_writenotify(vma, vma->vm_page_prot);
364
	vma_set_page_prot(vma);
365

366 367
	change_protection(vma, start, end, vma->vm_page_prot,
			  dirty_accountable, 0);
368

369 370 371 372 373 374 375 376 377
	/*
	 * Private VM_LOCKED VMA becoming writable: trigger COW to avoid major
	 * fault on access.
	 */
	if ((oldflags & (VM_WRITE | VM_SHARED | VM_LOCKED)) == VM_LOCKED &&
			(newflags & VM_WRITE)) {
		populate_vma_page_range(vma, start, end, NULL);
	}

378 379
	vm_stat_account(mm, oldflags, -nrpages);
	vm_stat_account(mm, newflags, nrpages);
380
	perf_event_mmap(vma);
Linus Torvalds's avatar
Linus Torvalds committed
381 382 383 384 385 386 387
	return 0;

fail:
	vm_unacct_memory(charged);
	return error;
}

388 389 390 391 392
/*
 * pkey==-1 when doing a legacy mprotect()
 */
static int do_mprotect_pkey(unsigned long start, size_t len,
		unsigned long prot, int pkey)
Linus Torvalds's avatar
Linus Torvalds committed
393
{
394
	unsigned long nstart, end, tmp, reqprot;
Linus Torvalds's avatar
Linus Torvalds committed
395 396 397
	struct vm_area_struct *vma, *prev;
	int error = -EINVAL;
	const int grows = prot & (PROT_GROWSDOWN|PROT_GROWSUP);
398 399 400
	const bool rier = (current->personality & READ_IMPLIES_EXEC) &&
				(prot & PROT_READ);

Linus Torvalds's avatar
Linus Torvalds committed
401 402 403 404 405 406 407 408 409 410 411 412
	prot &= ~(PROT_GROWSDOWN|PROT_GROWSUP);
	if (grows == (PROT_GROWSDOWN|PROT_GROWSUP)) /* can't be both */
		return -EINVAL;

	if (start & ~PAGE_MASK)
		return -EINVAL;
	if (!len)
		return 0;
	len = PAGE_ALIGN(len);
	end = start + len;
	if (end <= start)
		return -ENOMEM;
413
	if (!arch_validate_prot(prot))
Linus Torvalds's avatar
Linus Torvalds committed
414 415 416 417
		return -EINVAL;

	reqprot = prot;

418 419
	if (down_write_killable(&current->mm->mmap_sem))
		return -EINTR;
Linus Torvalds's avatar
Linus Torvalds committed
420

421 422 423 424 425 426 427 428
	/*
	 * If userspace did not allocate the pkey, do not let
	 * them use it here.
	 */
	error = -EINVAL;
	if ((pkey != -1) && !mm_pkey_is_allocated(current->mm, pkey))
		goto out;

429
	vma = find_vma(current->mm, start);
Linus Torvalds's avatar
Linus Torvalds committed
430 431 432
	error = -ENOMEM;
	if (!vma)
		goto out;
433
	prev = vma->vm_prev;
Linus Torvalds's avatar
Linus Torvalds committed
434 435 436 437 438 439 440
	if (unlikely(grows & PROT_GROWSDOWN)) {
		if (vma->vm_start >= end)
			goto out;
		start = vma->vm_start;
		error = -EINVAL;
		if (!(vma->vm_flags & VM_GROWSDOWN))
			goto out;
441
	} else {
Linus Torvalds's avatar
Linus Torvalds committed
442 443 444 445 446 447 448 449 450 451 452 453 454
		if (vma->vm_start > start)
			goto out;
		if (unlikely(grows & PROT_GROWSUP)) {
			end = vma->vm_end;
			error = -EINVAL;
			if (!(vma->vm_flags & VM_GROWSUP))
				goto out;
		}
	}
	if (start > vma->vm_start)
		prev = vma;

	for (nstart = start ; ; ) {
455
		unsigned long mask_off_old_flags;
Linus Torvalds's avatar
Linus Torvalds committed
456
		unsigned long newflags;
457
		int new_vma_pkey;
Linus Torvalds's avatar
Linus Torvalds committed
458

459
		/* Here we know that vma->vm_start <= nstart < vma->vm_end. */
Linus Torvalds's avatar
Linus Torvalds committed
460

461 462 463 464
		/* Does the application expect PROT_READ to imply PROT_EXEC */
		if (rier && (vma->vm_flags & VM_MAYEXEC))
			prot |= PROT_EXEC;

465 466 467 468 469 470 471 472
		/*
		 * Each mprotect() call explicitly passes r/w/x permissions.
		 * If a permission is not passed to mprotect(), it must be
		 * cleared from the VMA.
		 */
		mask_off_old_flags = VM_READ | VM_WRITE | VM_EXEC |
					ARCH_VM_PKEY_FLAGS;

473 474
		new_vma_pkey = arch_override_mprotect_pkey(vma, prot, pkey);
		newflags = calc_vm_prot_bits(prot, new_vma_pkey);
475
		newflags |= (vma->vm_flags & ~mask_off_old_flags);
Linus Torvalds's avatar
Linus Torvalds committed
476

477 478
		/* newflags >> 4 shift VM_MAY% in place of VM_% */
		if ((newflags & ~(newflags >> 4)) & (VM_READ | VM_WRITE | VM_EXEC)) {
Linus Torvalds's avatar
Linus Torvalds committed
479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504
			error = -EACCES;
			goto out;
		}

		error = security_file_mprotect(vma, reqprot, prot);
		if (error)
			goto out;

		tmp = vma->vm_end;
		if (tmp > end)
			tmp = end;
		error = mprotect_fixup(vma, &prev, nstart, tmp, newflags);
		if (error)
			goto out;
		nstart = tmp;

		if (nstart < prev->vm_end)
			nstart = prev->vm_end;
		if (nstart >= end)
			goto out;

		vma = prev->vm_next;
		if (!vma || vma->vm_start != nstart) {
			error = -ENOMEM;
			goto out;
		}
505
		prot = reqprot;
Linus Torvalds's avatar
Linus Torvalds committed
506 507 508 509 510
	}
out:
	up_write(&current->mm->mmap_sem);
	return error;
}
511 512 513 514 515 516 517

SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len,
		unsigned long, prot)
{
	return do_mprotect_pkey(start, len, prot, -1);
}

518 519
#ifdef CONFIG_ARCH_HAS_PKEYS

520 521 522 523 524
SYSCALL_DEFINE4(pkey_mprotect, unsigned long, start, size_t, len,
		unsigned long, prot, int, pkey)
{
	return do_mprotect_pkey(start, len, prot, pkey);
}
525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569

SYSCALL_DEFINE2(pkey_alloc, unsigned long, flags, unsigned long, init_val)
{
	int pkey;
	int ret;

	/* No flags supported yet. */
	if (flags)
		return -EINVAL;
	/* check for unsupported init values */
	if (init_val & ~PKEY_ACCESS_MASK)
		return -EINVAL;

	down_write(&current->mm->mmap_sem);
	pkey = mm_pkey_alloc(current->mm);

	ret = -ENOSPC;
	if (pkey == -1)
		goto out;

	ret = arch_set_user_pkey_access(current, pkey, init_val);
	if (ret) {
		mm_pkey_free(current->mm, pkey);
		goto out;
	}
	ret = pkey;
out:
	up_write(&current->mm->mmap_sem);
	return ret;
}

SYSCALL_DEFINE1(pkey_free, int, pkey)
{
	int ret;

	down_write(&current->mm->mmap_sem);
	ret = mm_pkey_free(current->mm, pkey);
	up_write(&current->mm->mmap_sem);

	/*
	 * We could provie warnings or errors if any VMA still
	 * has the pkey set here.
	 */
	return ret;
}
570 571

#endif /* CONFIG_ARCH_HAS_PKEYS */