Blame view
mm/mprotect.c
9.59 KB
1da177e4c
|
1 2 3 4 5 6 |
/* * mm/mprotect.c * * (C) Copyright 1994 Linus Torvalds * (C) Copyright 2002 Christoph Hellwig * |
046c68842
|
7 |
* Address space accounting code <alan@lxorguk.ukuu.org.uk> |
1da177e4c
|
8 9 10 11 12 |
* (C) Copyright 2002 Red Hat Inc, All Rights Reserved */ #include <linux/mm.h> #include <linux/hugetlb.h> |
1da177e4c
|
13 14 15 16 17 18 19 20 |
#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> |
0697212a4
|
21 22 |
#include <linux/swap.h> #include <linux/swapops.h> |
cddb8a5c1
|
23 |
#include <linux/mmu_notifier.h> |
64cdd548f
|
24 |
#include <linux/migrate.h> |
cdd6c482c
|
25 |
#include <linux/perf_event.h> |
64a9a34e2
|
26 |
#include <linux/ksm.h> |
1da177e4c
|
27 28 29 30 |
#include <asm/uaccess.h> #include <asm/pgtable.h> #include <asm/cacheflush.h> #include <asm/tlbflush.h> |
1c12c4cf9
|
31 32 33 34 35 36 |
#ifndef pgprot_modify static inline pgprot_t pgprot_modify(pgprot_t oldprot, pgprot_t newprot) { return newprot; } #endif |
4b10e7d56
|
37 |
static unsigned long change_pte_range(struct vm_area_struct *vma, pmd_t *pmd, |
c1e6098b2
|
38 |
unsigned long addr, unsigned long end, pgprot_t newprot, |
0f19c1792
|
39 |
int dirty_accountable, int prot_numa) |
1da177e4c
|
40 |
{ |
4b10e7d56
|
41 |
struct mm_struct *mm = vma->vm_mm; |
0697212a4
|
42 |
pte_t *pte, oldpte; |
705e87c0c
|
43 |
spinlock_t *ptl; |
7da4d641c
|
44 |
unsigned long pages = 0; |
1da177e4c
|
45 |
|
705e87c0c
|
46 |
pte = pte_offset_map_lock(mm, pmd, addr, &ptl); |
6606c3e0d
|
47 |
arch_enter_lazy_mmu_mode(); |
1da177e4c
|
48 |
do { |
0697212a4
|
49 50 |
oldpte = *pte; if (pte_present(oldpte)) { |
1da177e4c
|
51 |
pte_t ptent; |
4b10e7d56
|
52 |
bool updated = false; |
1da177e4c
|
53 |
|
4b10e7d56
|
54 |
if (!prot_numa) { |
0c5f83c23
|
55 |
ptent = ptep_modify_prot_start(mm, addr, pte); |
1667918b6
|
56 57 |
if (pte_numa(ptent)) ptent = pte_mknonnuma(ptent); |
4b10e7d56
|
58 |
ptent = pte_modify(ptent, newprot); |
9d85d5863
|
59 60 61 62 63 64 65 |
/* * Avoid taking write faults for pages we * know to be dirty. */ if (dirty_accountable && pte_dirty(ptent)) ptent = pte_mkwrite(ptent); ptep_modify_prot_commit(mm, addr, pte, ptent); |
4b10e7d56
|
66 67 68 69 70 |
updated = true; } else { struct page *page; page = vm_normal_page(vma, addr, oldpte); |
64a9a34e2
|
71 |
if (page && !PageKsm(page)) { |
1bc115d87
|
72 |
if (!pte_numa(oldpte)) { |
56eecdb91
|
73 |
ptep_set_numa(mm, addr, pte); |
4b10e7d56
|
74 75 76 77 |
updated = true; } } } |
4b10e7d56
|
78 79 |
if (updated) pages++; |
ce1744f4e
|
80 |
} else if (IS_ENABLED(CONFIG_MIGRATION) && !pte_file(oldpte)) { |
0697212a4
|
81 82 83 |
swp_entry_t entry = pte_to_swp_entry(oldpte); if (is_write_migration_entry(entry)) { |
c3d16e165
|
84 |
pte_t newpte; |
0697212a4
|
85 86 87 88 89 |
/* * A protection check is difficult so * just be safe and disable write */ make_migration_entry_read(&entry); |
c3d16e165
|
90 91 92 93 |
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); |
e920e14ca
|
94 95 |
pages++; |
0697212a4
|
96 |
} |
1da177e4c
|
97 98 |
} } while (pte++, addr += PAGE_SIZE, addr != end); |
6606c3e0d
|
99 |
arch_leave_lazy_mmu_mode(); |
705e87c0c
|
100 |
pte_unmap_unlock(pte - 1, ptl); |
7da4d641c
|
101 102 |
return pages; |
1da177e4c
|
103 |
} |
7d12efaea
|
104 105 106 |
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) |
1da177e4c
|
107 108 109 |
{ pmd_t *pmd; unsigned long next; |
7da4d641c
|
110 |
unsigned long pages = 0; |
72403b4a0
|
111 |
unsigned long nr_huge_updates = 0; |
1da177e4c
|
112 113 114 |
pmd = pmd_offset(pud, addr); do { |
25cbbef19
|
115 |
unsigned long this_pages; |
1da177e4c
|
116 |
next = pmd_addr_end(addr, end); |
cd7548ab3
|
117 118 |
if (pmd_trans_huge(*pmd)) { if (next - addr != HPAGE_PMD_SIZE) |
e180377f1
|
119 |
split_huge_page_pmd(vma, addr, pmd); |
f123d74ab
|
120 121 122 123 124 |
else { int nr_ptes = change_huge_pmd(vma, pmd, addr, newprot, prot_numa); if (nr_ptes) { |
72403b4a0
|
125 126 127 128 |
if (nr_ptes == HPAGE_PMD_NR) { pages += HPAGE_PMD_NR; nr_huge_updates++; } |
f123d74ab
|
129 130 |
continue; } |
7da4d641c
|
131 |
} |
cd7548ab3
|
132 133 |
/* fall through */ } |
1da177e4c
|
134 135 |
if (pmd_none_or_clear_bad(pmd)) continue; |
25cbbef19
|
136 |
this_pages = change_pte_range(vma, pmd, addr, next, newprot, |
0f19c1792
|
137 |
dirty_accountable, prot_numa); |
25cbbef19
|
138 |
pages += this_pages; |
1da177e4c
|
139 |
} while (pmd++, addr = next, addr != end); |
7da4d641c
|
140 |
|
72403b4a0
|
141 142 |
if (nr_huge_updates) count_vm_numa_events(NUMA_HUGE_PTE_UPDATES, nr_huge_updates); |
7da4d641c
|
143 |
return pages; |
1da177e4c
|
144 |
} |
7d12efaea
|
145 146 147 |
static inline unsigned long change_pud_range(struct vm_area_struct *vma, pgd_t *pgd, unsigned long addr, unsigned long end, pgprot_t newprot, int dirty_accountable, int prot_numa) |
1da177e4c
|
148 149 150 |
{ pud_t *pud; unsigned long next; |
7da4d641c
|
151 |
unsigned long pages = 0; |
1da177e4c
|
152 153 154 155 156 157 |
pud = pud_offset(pgd, addr); do { next = pud_addr_end(addr, end); if (pud_none_or_clear_bad(pud)) continue; |
7da4d641c
|
158 |
pages += change_pmd_range(vma, pud, addr, next, newprot, |
4b10e7d56
|
159 |
dirty_accountable, prot_numa); |
1da177e4c
|
160 |
} while (pud++, addr = next, addr != end); |
7da4d641c
|
161 162 |
return pages; |
1da177e4c
|
163 |
} |
7da4d641c
|
164 |
static unsigned long change_protection_range(struct vm_area_struct *vma, |
c1e6098b2
|
165 |
unsigned long addr, unsigned long end, pgprot_t newprot, |
4b10e7d56
|
166 |
int dirty_accountable, int prot_numa) |
1da177e4c
|
167 168 169 170 171 |
{ struct mm_struct *mm = vma->vm_mm; pgd_t *pgd; unsigned long next; unsigned long start = addr; |
7da4d641c
|
172 |
unsigned long pages = 0; |
1da177e4c
|
173 174 175 176 |
BUG_ON(addr >= end); pgd = pgd_offset(mm, addr); flush_cache_range(vma, addr, end); |
208414059
|
177 |
set_tlb_flush_pending(mm); |
1da177e4c
|
178 179 180 181 |
do { next = pgd_addr_end(addr, end); if (pgd_none_or_clear_bad(pgd)) continue; |
7da4d641c
|
182 |
pages += change_pud_range(vma, pgd, addr, next, newprot, |
4b10e7d56
|
183 |
dirty_accountable, prot_numa); |
1da177e4c
|
184 |
} while (pgd++, addr = next, addr != end); |
7da4d641c
|
185 |
|
1233d5882
|
186 187 188 |
/* Only flush the TLB if we actually modified any entries: */ if (pages) flush_tlb_range(vma, start, end); |
208414059
|
189 |
clear_tlb_flush_pending(mm); |
7da4d641c
|
190 191 192 193 194 195 |
return pages; } unsigned long change_protection(struct vm_area_struct *vma, unsigned long start, unsigned long end, pgprot_t newprot, |
4b10e7d56
|
196 |
int dirty_accountable, int prot_numa) |
7da4d641c
|
197 198 199 200 201 202 203 204 |
{ struct mm_struct *mm = vma->vm_mm; unsigned long pages; mmu_notifier_invalidate_range_start(mm, start, end); if (is_vm_hugetlb_page(vma)) pages = hugetlb_change_protection(vma, start, end, newprot); else |
4b10e7d56
|
205 |
pages = change_protection_range(vma, start, end, newprot, dirty_accountable, prot_numa); |
7da4d641c
|
206 207 208 |
mmu_notifier_invalidate_range_end(mm, start, end); return pages; |
1da177e4c
|
209 |
} |
b6a2fea39
|
210 |
int |
1da177e4c
|
211 212 213 214 215 216 217 |
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; |
1da177e4c
|
218 219 |
pgoff_t pgoff; int error; |
c1e6098b2
|
220 |
int dirty_accountable = 0; |
1da177e4c
|
221 222 223 224 225 226 227 228 229 |
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 |
5a6fe1259
|
230 231 |
* make it unwritable again. hugetlb mapping were accounted for * even if read-only so there is no need to account for them here |
1da177e4c
|
232 233 |
*/ if (newflags & VM_WRITE) { |
5a6fe1259
|
234 |
if (!(oldflags & (VM_ACCOUNT|VM_WRITE|VM_HUGETLB| |
cdfd4325c
|
235 |
VM_SHARED|VM_NORESERVE))) { |
1da177e4c
|
236 |
charged = nrpages; |
191c54244
|
237 |
if (security_vm_enough_memory_mm(mm, charged)) |
1da177e4c
|
238 239 240 241 |
return -ENOMEM; newflags |= VM_ACCOUNT; } } |
1da177e4c
|
242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 |
/* * 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, vma->anon_vma, vma->vm_file, pgoff, vma_policy(vma)); if (*pprev) { vma = *pprev; 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; |
1c12c4cf9
|
273 274 |
vma->vm_page_prot = pgprot_modify(vma->vm_page_prot, vm_get_page_prot(newflags)); |
c1e6098b2
|
275 |
if (vma_wants_writenotify(vma)) { |
1ddd439ef
|
276 |
vma->vm_page_prot = vm_get_page_prot(newflags & ~VM_SHARED); |
c1e6098b2
|
277 278 |
dirty_accountable = 1; } |
d08b3851d
|
279 |
|
7d12efaea
|
280 281 |
change_protection(vma, start, end, vma->vm_page_prot, dirty_accountable, 0); |
7da4d641c
|
282 |
|
ab50b8ed8
|
283 284 |
vm_stat_account(mm, oldflags, vma->vm_file, -nrpages); vm_stat_account(mm, newflags, vma->vm_file, nrpages); |
63bfd7384
|
285 |
perf_event_mmap(vma); |
1da177e4c
|
286 287 288 289 290 291 |
return 0; fail: vm_unacct_memory(charged); return error; } |
6a6160a7b
|
292 293 |
SYSCALL_DEFINE3(mprotect, unsigned long, start, size_t, len, unsigned long, prot) |
1da177e4c
|
294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 |
{ unsigned long vm_flags, nstart, end, tmp, reqprot; struct vm_area_struct *vma, *prev; int error = -EINVAL; const int grows = prot & (PROT_GROWSDOWN|PROT_GROWSUP); 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; |
b845f313d
|
311 |
if (!arch_validate_prot(prot)) |
1da177e4c
|
312 313 314 315 316 317 |
return -EINVAL; reqprot = prot; /* * Does the application expect PROT_READ to imply PROT_EXEC: */ |
b344e05c5
|
318 |
if ((prot & PROT_READ) && (current->personality & READ_IMPLIES_EXEC)) |
1da177e4c
|
319 320 321 322 323 |
prot |= PROT_EXEC; vm_flags = calc_vm_prot_bits(prot); down_write(¤t->mm->mmap_sem); |
097d59106
|
324 |
vma = find_vma(current->mm, start); |
1da177e4c
|
325 326 327 |
error = -ENOMEM; if (!vma) goto out; |
097d59106
|
328 |
prev = vma->vm_prev; |
1da177e4c
|
329 330 331 332 333 334 335 |
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; |
7d12efaea
|
336 |
} else { |
1da177e4c
|
337 338 339 340 341 342 343 344 345 346 347 348 349 350 |
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 ; ; ) { unsigned long newflags; |
7d12efaea
|
351 |
/* Here we know that vma->vm_start <= nstart < vma->vm_end. */ |
1da177e4c
|
352 |
|
7d12efaea
|
353 354 |
newflags = vm_flags; newflags |= (vma->vm_flags & ~(VM_READ | VM_WRITE | VM_EXEC)); |
1da177e4c
|
355 |
|
7e2cff42c
|
356 357 |
/* newflags >> 4 shift VM_MAY% in place of VM_% */ if ((newflags & ~(newflags >> 4)) & (VM_READ | VM_WRITE | VM_EXEC)) { |
1da177e4c
|
358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 |
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; } } out: up_write(¤t->mm->mmap_sem); return error; } |