Blame view
mm/gup.c
84 KB
457c89965
|
1 |
// SPDX-License-Identifier: GPL-2.0-only |
4bbd4c776
|
2 3 4 5 |
#include <linux/kernel.h> #include <linux/errno.h> #include <linux/err.h> #include <linux/spinlock.h> |
4bbd4c776
|
6 |
#include <linux/mm.h> |
3565fce3a
|
7 |
#include <linux/memremap.h> |
4bbd4c776
|
8 9 10 11 |
#include <linux/pagemap.h> #include <linux/rmap.h> #include <linux/swap.h> #include <linux/swapops.h> |
174cd4b1e
|
12 |
#include <linux/sched/signal.h> |
2667f50e8
|
13 |
#include <linux/rwsem.h> |
f30c59e92
|
14 |
#include <linux/hugetlb.h> |
9a4e9f3b2
|
15 16 17 |
#include <linux/migrate.h> #include <linux/mm_inline.h> #include <linux/sched/mm.h> |
1027e4436
|
18 |
|
33a709b25
|
19 |
#include <asm/mmu_context.h> |
1027e4436
|
20 |
#include <asm/tlbflush.h> |
2667f50e8
|
21 |
|
4bbd4c776
|
22 |
#include "internal.h" |
df06b37ff
|
23 24 25 26 |
struct follow_page_context { struct dev_pagemap *pgmap; unsigned int page_mask; }; |
47e29d32a
|
27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 |
static void hpage_pincount_add(struct page *page, int refs) { VM_BUG_ON_PAGE(!hpage_pincount_available(page), page); VM_BUG_ON_PAGE(page != compound_head(page), page); atomic_add(refs, compound_pincount_ptr(page)); } static void hpage_pincount_sub(struct page *page, int refs) { VM_BUG_ON_PAGE(!hpage_pincount_available(page), page); VM_BUG_ON_PAGE(page != compound_head(page), page); atomic_sub(refs, compound_pincount_ptr(page)); } |
a707cdd55
|
42 43 44 45 46 47 48 49 50 51 52 53 54 55 |
/* * Return the compound head page with ref appropriately incremented, * or NULL if that failed. */ static inline struct page *try_get_compound_head(struct page *page, int refs) { struct page *head = compound_head(page); if (WARN_ON_ONCE(page_ref_count(head) < 0)) return NULL; if (unlikely(!page_cache_add_speculative(head, refs))) return NULL; return head; } |
3faa52c03
|
56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 |
/* * try_grab_compound_head() - attempt to elevate a page's refcount, by a * flags-dependent amount. * * "grab" names in this file mean, "look at flags to decide whether to use * FOLL_PIN or FOLL_GET behavior, when incrementing the page's refcount. * * Either FOLL_PIN or FOLL_GET (or neither) must be set, but not both at the * same time. (That's true throughout the get_user_pages*() and * pin_user_pages*() APIs.) Cases: * * FOLL_GET: page's refcount will be incremented by 1. * FOLL_PIN: page's refcount will be incremented by GUP_PIN_COUNTING_BIAS. * * Return: head page (with refcount appropriately incremented) for success, or * NULL upon failure. If neither FOLL_GET nor FOLL_PIN was set, that's * considered failure, and furthermore, a likely bug in the caller, so a warning * is also emitted. */ static __maybe_unused struct page *try_grab_compound_head(struct page *page, int refs, unsigned int flags) { if (flags & FOLL_GET) return try_get_compound_head(page, refs); else if (flags & FOLL_PIN) { |
1970dc6f5
|
82 |
int orig_refs = refs; |
47e29d32a
|
83 |
/* |
df3a0a21b
|
84 85 86 87 88 89 90 91 |
* Can't do FOLL_LONGTERM + FOLL_PIN with CMA in the gup fast * path, so fail and let the caller fall back to the slow path. */ if (unlikely(flags & FOLL_LONGTERM) && is_migrate_cma_page(page)) return NULL; /* |
47e29d32a
|
92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 |
* When pinning a compound page of order > 1 (which is what * hpage_pincount_available() checks for), use an exact count to * track it, via hpage_pincount_add/_sub(). * * However, be sure to *also* increment the normal page refcount * field at least once, so that the page really is pinned. */ if (!hpage_pincount_available(page)) refs *= GUP_PIN_COUNTING_BIAS; page = try_get_compound_head(page, refs); if (!page) return NULL; if (hpage_pincount_available(page)) hpage_pincount_add(page, refs); |
1970dc6f5
|
108 109 |
mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_ACQUIRED, orig_refs); |
47e29d32a
|
110 |
return page; |
3faa52c03
|
111 112 113 114 115 |
} WARN_ON_ONCE(1); return NULL; } |
cfde6c181
|
116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 |
static void put_compound_head(struct page *page, int refs, unsigned int flags) { if (flags & FOLL_PIN) { mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_RELEASED, refs); if (hpage_pincount_available(page)) hpage_pincount_sub(page, refs); else refs *= GUP_PIN_COUNTING_BIAS; } VM_BUG_ON_PAGE(page_ref_count(page) < refs, page); /* * Calling put_page() for each ref is unnecessarily slow. Only the last * ref needs a put_page(). */ if (refs > 1) page_ref_sub(page, refs - 1); put_page(page); } |
3faa52c03
|
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 |
/** * try_grab_page() - elevate a page's refcount by a flag-dependent amount * * This might not do anything at all, depending on the flags argument. * * "grab" names in this file mean, "look at flags to decide whether to use * FOLL_PIN or FOLL_GET behavior, when incrementing the page's refcount. * * @page: pointer to page to be grabbed * @flags: gup flags: these are the FOLL_* flag values. * * Either FOLL_PIN or FOLL_GET (or neither) may be set, but not both at the same * time. Cases: * * FOLL_GET: page's refcount will be incremented by 1. * FOLL_PIN: page's refcount will be incremented by GUP_PIN_COUNTING_BIAS. * * Return: true for success, or if no action was required (if neither FOLL_PIN * nor FOLL_GET was set, nothing is done). False for failure: FOLL_GET or * FOLL_PIN was set, but the page could not be grabbed. */ bool __must_check try_grab_page(struct page *page, unsigned int flags) { WARN_ON_ONCE((flags & (FOLL_GET | FOLL_PIN)) == (FOLL_GET | FOLL_PIN)); if (flags & FOLL_GET) return try_get_page(page); else if (flags & FOLL_PIN) { |
47e29d32a
|
165 |
int refs = 1; |
3faa52c03
|
166 167 168 169 |
page = compound_head(page); if (WARN_ON_ONCE(page_ref_count(page) <= 0)) return false; |
47e29d32a
|
170 171 172 173 174 175 176 177 178 179 180 181 |
if (hpage_pincount_available(page)) hpage_pincount_add(page, 1); else refs = GUP_PIN_COUNTING_BIAS; /* * Similar to try_grab_compound_head(): even if using the * hpage_pincount_add/_sub() routines, be sure to * *also* increment the normal page refcount field at least * once, so that the page really is pinned. */ page_ref_add(page, refs); |
1970dc6f5
|
182 183 |
mod_node_page_state(page_pgdat(page), NR_FOLL_PIN_ACQUIRED, 1); |
3faa52c03
|
184 185 186 187 |
} return true; } |
3faa52c03
|
188 189 190 191 192 193 194 195 196 197 198 |
/** * unpin_user_page() - release a dma-pinned page * @page: pointer to page to be released * * Pages that were pinned via pin_user_pages*() must be released via either * unpin_user_page(), or one of the unpin_user_pages*() routines. This is so * that such pages can be separately tracked and uniquely handled. In * particular, interactions with RDMA and filesystems need special handling. */ void unpin_user_page(struct page *page) { |
cfde6c181
|
199 |
put_compound_head(compound_head(page), 1, FOLL_PIN); |
3faa52c03
|
200 201 |
} EXPORT_SYMBOL(unpin_user_page); |
fc1d8e7cc
|
202 |
/** |
f1f6a7dd9
|
203 |
* unpin_user_pages_dirty_lock() - release and optionally dirty gup-pinned pages |
2d15eb31b
|
204 |
* @pages: array of pages to be maybe marked dirty, and definitely released. |
fc1d8e7cc
|
205 |
* @npages: number of pages in the @pages array. |
2d15eb31b
|
206 |
* @make_dirty: whether to mark the pages dirty |
fc1d8e7cc
|
207 208 209 210 211 |
* * "gup-pinned page" refers to a page that has had one of the get_user_pages() * variants called on that page. * * For each page in the @pages array, make that page (or its head page, if a |
2d15eb31b
|
212 |
* compound page) dirty, if @make_dirty is true, and if the page was previously |
f1f6a7dd9
|
213 214 |
* listed as clean. In any case, releases all pages using unpin_user_page(), * possibly via unpin_user_pages(), for the non-dirty case. |
fc1d8e7cc
|
215 |
* |
f1f6a7dd9
|
216 |
* Please see the unpin_user_page() documentation for details. |
fc1d8e7cc
|
217 |
* |
2d15eb31b
|
218 219 220 |
* set_page_dirty_lock() is used internally. If instead, set_page_dirty() is * required, then the caller should a) verify that this is really correct, * because _lock() is usually required, and b) hand code it: |
f1f6a7dd9
|
221 |
* set_page_dirty_lock(), unpin_user_page(). |
fc1d8e7cc
|
222 223 |
* */ |
f1f6a7dd9
|
224 225 |
void unpin_user_pages_dirty_lock(struct page **pages, unsigned long npages, bool make_dirty) |
fc1d8e7cc
|
226 |
{ |
2d15eb31b
|
227 |
unsigned long index; |
fc1d8e7cc
|
228 |
|
2d15eb31b
|
229 230 231 232 233 234 235 |
/* * TODO: this can be optimized for huge pages: if a series of pages is * physically contiguous and part of the same compound page, then a * single operation to the head page should suffice. */ if (!make_dirty) { |
f1f6a7dd9
|
236 |
unpin_user_pages(pages, npages); |
2d15eb31b
|
237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 |
return; } for (index = 0; index < npages; index++) { struct page *page = compound_head(pages[index]); /* * Checking PageDirty at this point may race with * clear_page_dirty_for_io(), but that's OK. Two key * cases: * * 1) This code sees the page as already dirty, so it * skips the call to set_page_dirty(). That could happen * because clear_page_dirty_for_io() called * page_mkclean(), followed by set_page_dirty(). * However, now the page is going to get written back, * which meets the original intention of setting it * dirty, so all is well: clear_page_dirty_for_io() goes * on to call TestClearPageDirty(), and write the page * back. * * 2) This code sees the page as clean, so it calls * set_page_dirty(). The page stays dirty, despite being * written back, so it gets written back again in the * next writeback cycle. This is harmless. */ if (!PageDirty(page)) set_page_dirty_lock(page); |
f1f6a7dd9
|
264 |
unpin_user_page(page); |
2d15eb31b
|
265 |
} |
fc1d8e7cc
|
266 |
} |
f1f6a7dd9
|
267 |
EXPORT_SYMBOL(unpin_user_pages_dirty_lock); |
fc1d8e7cc
|
268 269 |
/** |
f1f6a7dd9
|
270 |
* unpin_user_pages() - release an array of gup-pinned pages. |
fc1d8e7cc
|
271 272 273 |
* @pages: array of pages to be marked dirty and released. * @npages: number of pages in the @pages array. * |
f1f6a7dd9
|
274 |
* For each page in the @pages array, release the page using unpin_user_page(). |
fc1d8e7cc
|
275 |
* |
f1f6a7dd9
|
276 |
* Please see the unpin_user_page() documentation for details. |
fc1d8e7cc
|
277 |
*/ |
f1f6a7dd9
|
278 |
void unpin_user_pages(struct page **pages, unsigned long npages) |
fc1d8e7cc
|
279 280 281 282 |
{ unsigned long index; /* |
146608bb7
|
283 284 285 286 287 288 289 |
* If this WARN_ON() fires, then the system *might* be leaking pages (by * leaving them pinned), but probably not. More likely, gup/pup returned * a hard -ERRNO error to the caller, who erroneously passed it here. */ if (WARN_ON(IS_ERR_VALUE(npages))) return; /* |
fc1d8e7cc
|
290 291 292 293 294 |
* TODO: this can be optimized for huge pages: if a series of pages is * physically contiguous and part of the same compound page, then a * single operation to the head page should suffice. */ for (index = 0; index < npages; index++) |
f1f6a7dd9
|
295 |
unpin_user_page(pages[index]); |
fc1d8e7cc
|
296 |
} |
f1f6a7dd9
|
297 |
EXPORT_SYMBOL(unpin_user_pages); |
fc1d8e7cc
|
298 |
|
050a9adc6
|
299 |
#ifdef CONFIG_MMU |
69e68b4f0
|
300 301 |
static struct page *no_page_table(struct vm_area_struct *vma, unsigned int flags) |
4bbd4c776
|
302 |
{ |
69e68b4f0
|
303 304 305 306 307 308 309 310 |
/* * When core dumping an enormous anonymous area that nobody * has touched so far, we don't want to allocate unnecessary pages or * page tables. Return error instead of NULL to skip handle_mm_fault, * then get_dump_page() will return NULL to leave a hole in the dump. * But we can only make this optimization where a hole would surely * be zero-filled if handle_mm_fault() actually did handle it. */ |
a0137f16d
|
311 312 |
if ((flags & FOLL_DUMP) && (vma_is_anonymous(vma) || !vma->vm_ops->fault)) |
69e68b4f0
|
313 314 315 |
return ERR_PTR(-EFAULT); return NULL; } |
4bbd4c776
|
316 |
|
1027e4436
|
317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 |
static int follow_pfn_pte(struct vm_area_struct *vma, unsigned long address, pte_t *pte, unsigned int flags) { /* No page to get reference */ if (flags & FOLL_GET) return -EFAULT; if (flags & FOLL_TOUCH) { pte_t entry = *pte; if (flags & FOLL_WRITE) entry = pte_mkdirty(entry); entry = pte_mkyoung(entry); if (!pte_same(*pte, entry)) { set_pte_at(vma->vm_mm, address, pte, entry); update_mmu_cache(vma, address, pte); } } /* Proper page table entry exists, but no corresponding struct page */ return -EEXIST; } |
19be0eaff
|
340 |
/* |
a308c71bf
|
341 342 |
* FOLL_FORCE can write to even unwritable pte's, but only * after we've gone through a COW cycle and they are dirty. |
19be0eaff
|
343 344 345 |
*/ static inline bool can_follow_write_pte(pte_t pte, unsigned int flags) { |
a308c71bf
|
346 347 |
return pte_write(pte) || ((flags & FOLL_FORCE) && (flags & FOLL_COW) && pte_dirty(pte)); |
19be0eaff
|
348 |
} |
69e68b4f0
|
349 |
static struct page *follow_page_pte(struct vm_area_struct *vma, |
df06b37ff
|
350 351 |
unsigned long address, pmd_t *pmd, unsigned int flags, struct dev_pagemap **pgmap) |
69e68b4f0
|
352 353 354 355 356 |
{ struct mm_struct *mm = vma->vm_mm; struct page *page; spinlock_t *ptl; pte_t *ptep, pte; |
f28d43636
|
357 |
int ret; |
4bbd4c776
|
358 |
|
eddb1c228
|
359 360 361 362 |
/* FOLL_GET and FOLL_PIN are mutually exclusive. */ if (WARN_ON_ONCE((flags & (FOLL_PIN | FOLL_GET)) == (FOLL_PIN | FOLL_GET))) return ERR_PTR(-EINVAL); |
69e68b4f0
|
363 |
retry: |
4bbd4c776
|
364 |
if (unlikely(pmd_bad(*pmd))) |
69e68b4f0
|
365 |
return no_page_table(vma, flags); |
4bbd4c776
|
366 367 |
ptep = pte_offset_map_lock(mm, pmd, address, &ptl); |
4bbd4c776
|
368 369 370 371 372 373 374 375 376 377 |
pte = *ptep; if (!pte_present(pte)) { swp_entry_t entry; /* * KSM's break_ksm() relies upon recognizing a ksm page * even while it is being migrated, so for that case we * need migration_entry_wait(). */ if (likely(!(flags & FOLL_MIGRATION))) goto no_page; |
0661a3361
|
378 |
if (pte_none(pte)) |
4bbd4c776
|
379 380 381 382 383 384 |
goto no_page; entry = pte_to_swp_entry(pte); if (!is_migration_entry(entry)) goto no_page; pte_unmap_unlock(ptep, ptl); migration_entry_wait(mm, pmd, address); |
69e68b4f0
|
385 |
goto retry; |
4bbd4c776
|
386 |
} |
8a0516ed8
|
387 |
if ((flags & FOLL_NUMA) && pte_protnone(pte)) |
4bbd4c776
|
388 |
goto no_page; |
19be0eaff
|
389 |
if ((flags & FOLL_WRITE) && !can_follow_write_pte(pte, flags)) { |
69e68b4f0
|
390 391 392 |
pte_unmap_unlock(ptep, ptl); return NULL; } |
4bbd4c776
|
393 394 |
page = vm_normal_page(vma, address, pte); |
3faa52c03
|
395 |
if (!page && pte_devmap(pte) && (flags & (FOLL_GET | FOLL_PIN))) { |
3565fce3a
|
396 |
/* |
3faa52c03
|
397 398 399 |
* Only return device mapping pages in the FOLL_GET or FOLL_PIN * case since they are only valid while holding the pgmap * reference. |
3565fce3a
|
400 |
*/ |
df06b37ff
|
401 402 |
*pgmap = get_dev_pagemap(pte_pfn(pte), *pgmap); if (*pgmap) |
3565fce3a
|
403 404 405 406 |
page = pte_page(pte); else goto no_page; } else if (unlikely(!page)) { |
1027e4436
|
407 408 409 410 411 412 413 414 415 |
if (flags & FOLL_DUMP) { /* Avoid special (like zero) pages in core dumps */ page = ERR_PTR(-EFAULT); goto out; } if (is_zero_pfn(pte_pfn(pte))) { page = pte_page(pte); } else { |
1027e4436
|
416 417 418 419 |
ret = follow_pfn_pte(vma, address, ptep, flags); page = ERR_PTR(ret); goto out; } |
4bbd4c776
|
420 |
} |
6742d293c
|
421 |
if (flags & FOLL_SPLIT && PageTransCompound(page)) { |
6742d293c
|
422 423 424 425 426 427 428 429 430 431 |
get_page(page); pte_unmap_unlock(ptep, ptl); lock_page(page); ret = split_huge_page(page); unlock_page(page); put_page(page); if (ret) return ERR_PTR(ret); goto retry; } |
3faa52c03
|
432 433 434 435 |
/* try_grab_page() does nothing unless FOLL_GET or FOLL_PIN is set. */ if (unlikely(!try_grab_page(page, flags))) { page = ERR_PTR(-ENOMEM); goto out; |
8fde12ca7
|
436 |
} |
f28d43636
|
437 438 439 440 441 442 443 444 445 446 447 448 449 |
/* * We need to make the page accessible if and only if we are going * to access its content (the FOLL_PIN case). Please see * Documentation/core-api/pin_user_pages.rst for details. */ if (flags & FOLL_PIN) { ret = arch_make_page_accessible(page); if (ret) { unpin_user_page(page); page = ERR_PTR(ret); goto out; } } |
4bbd4c776
|
450 451 452 453 454 455 456 457 458 459 460 |
if (flags & FOLL_TOUCH) { if ((flags & FOLL_WRITE) && !pte_dirty(pte) && !PageDirty(page)) set_page_dirty(page); /* * pte_mkyoung() would be more correct here, but atomic care * is needed to avoid losing the dirty bit: it is easier to use * mark_page_accessed(). */ mark_page_accessed(page); } |
de60f5f10
|
461 |
if ((flags & FOLL_MLOCK) && (vma->vm_flags & VM_LOCKED)) { |
e90309c9f
|
462 463 464 |
/* Do not mlock pte-mapped THP */ if (PageTransCompound(page)) goto out; |
4bbd4c776
|
465 466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 |
/* * The preliminary mapping check is mainly to avoid the * pointless overhead of lock_page on the ZERO_PAGE * which might bounce very badly if there is contention. * * If the page is already locked, we don't need to * handle it now - vmscan will handle it later if and * when it attempts to reclaim the page. */ if (page->mapping && trylock_page(page)) { lru_add_drain(); /* push cached pages to LRU */ /* * Because we lock page here, and migration is * blocked by the pte's page reference, and we * know the page is still mapped, we don't even * need to check for file-cache page truncation. */ mlock_vma_page(page); unlock_page(page); } } |
1027e4436
|
486 |
out: |
4bbd4c776
|
487 |
pte_unmap_unlock(ptep, ptl); |
4bbd4c776
|
488 |
return page; |
4bbd4c776
|
489 490 491 |
no_page: pte_unmap_unlock(ptep, ptl); if (!pte_none(pte)) |
69e68b4f0
|
492 493 494 |
return NULL; return no_page_table(vma, flags); } |
080dbb618
|
495 496 |
static struct page *follow_pmd_mask(struct vm_area_struct *vma, unsigned long address, pud_t *pudp, |
df06b37ff
|
497 498 |
unsigned int flags, struct follow_page_context *ctx) |
69e68b4f0
|
499 |
{ |
688272809
|
500 |
pmd_t *pmd, pmdval; |
69e68b4f0
|
501 502 503 |
spinlock_t *ptl; struct page *page; struct mm_struct *mm = vma->vm_mm; |
080dbb618
|
504 |
pmd = pmd_offset(pudp, address); |
688272809
|
505 506 507 508 509 510 |
/* * The READ_ONCE() will stabilize the pmdval in a register or * on the stack so that it will stop changing under the code. */ pmdval = READ_ONCE(*pmd); if (pmd_none(pmdval)) |
69e68b4f0
|
511 |
return no_page_table(vma, flags); |
be9d30458
|
512 |
if (pmd_huge(pmdval) && is_vm_hugetlb_page(vma)) { |
e66f17ff7
|
513 514 515 516 |
page = follow_huge_pmd(mm, address, pmd, flags); if (page) return page; return no_page_table(vma, flags); |
69e68b4f0
|
517 |
} |
688272809
|
518 |
if (is_hugepd(__hugepd(pmd_val(pmdval)))) { |
4dc71451a
|
519 |
page = follow_huge_pd(vma, address, |
688272809
|
520 |
__hugepd(pmd_val(pmdval)), flags, |
4dc71451a
|
521 522 523 524 525 |
PMD_SHIFT); if (page) return page; return no_page_table(vma, flags); } |
84c3fc4e9
|
526 |
retry: |
688272809
|
527 |
if (!pmd_present(pmdval)) { |
84c3fc4e9
|
528 529 530 |
if (likely(!(flags & FOLL_MIGRATION))) return no_page_table(vma, flags); VM_BUG_ON(thp_migration_supported() && |
688272809
|
531 532 |
!is_pmd_migration_entry(pmdval)); if (is_pmd_migration_entry(pmdval)) |
84c3fc4e9
|
533 |
pmd_migration_entry_wait(mm, pmd); |
688272809
|
534 535 536 |
pmdval = READ_ONCE(*pmd); /* * MADV_DONTNEED may convert the pmd to null because |
c1e8d7c6a
|
537 |
* mmap_lock is held in read mode |
688272809
|
538 539 540 |
*/ if (pmd_none(pmdval)) return no_page_table(vma, flags); |
84c3fc4e9
|
541 542 |
goto retry; } |
688272809
|
543 |
if (pmd_devmap(pmdval)) { |
3565fce3a
|
544 |
ptl = pmd_lock(mm, pmd); |
df06b37ff
|
545 |
page = follow_devmap_pmd(vma, address, pmd, flags, &ctx->pgmap); |
3565fce3a
|
546 547 548 549 |
spin_unlock(ptl); if (page) return page; } |
688272809
|
550 |
if (likely(!pmd_trans_huge(pmdval))) |
df06b37ff
|
551 |
return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap); |
6742d293c
|
552 |
|
688272809
|
553 |
if ((flags & FOLL_NUMA) && pmd_protnone(pmdval)) |
db08f2030
|
554 |
return no_page_table(vma, flags); |
84c3fc4e9
|
555 |
retry_locked: |
6742d293c
|
556 |
ptl = pmd_lock(mm, pmd); |
688272809
|
557 558 559 560 |
if (unlikely(pmd_none(*pmd))) { spin_unlock(ptl); return no_page_table(vma, flags); } |
84c3fc4e9
|
561 562 563 564 565 566 567 |
if (unlikely(!pmd_present(*pmd))) { spin_unlock(ptl); if (likely(!(flags & FOLL_MIGRATION))) return no_page_table(vma, flags); pmd_migration_entry_wait(mm, pmd); goto retry_locked; } |
6742d293c
|
568 569 |
if (unlikely(!pmd_trans_huge(*pmd))) { spin_unlock(ptl); |
df06b37ff
|
570 |
return follow_page_pte(vma, address, pmd, flags, &ctx->pgmap); |
6742d293c
|
571 |
} |
bfe7b00de
|
572 |
if (flags & (FOLL_SPLIT | FOLL_SPLIT_PMD)) { |
6742d293c
|
573 574 575 576 577 |
int ret; page = pmd_page(*pmd); if (is_huge_zero_page(page)) { spin_unlock(ptl); ret = 0; |
78ddc5347
|
578 |
split_huge_pmd(vma, pmd, address); |
337d9abf1
|
579 580 |
if (pmd_trans_unstable(pmd)) ret = -EBUSY; |
bfe7b00de
|
581 |
} else if (flags & FOLL_SPLIT) { |
8fde12ca7
|
582 583 584 585 |
if (unlikely(!try_get_page(page))) { spin_unlock(ptl); return ERR_PTR(-ENOMEM); } |
69e68b4f0
|
586 |
spin_unlock(ptl); |
6742d293c
|
587 588 589 590 |
lock_page(page); ret = split_huge_page(page); unlock_page(page); put_page(page); |
baa355fd3
|
591 592 |
if (pmd_none(*pmd)) return no_page_table(vma, flags); |
bfe7b00de
|
593 594 595 596 |
} else { /* flags & FOLL_SPLIT_PMD */ spin_unlock(ptl); split_huge_pmd(vma, pmd, address); ret = pte_alloc(mm, pmd) ? -ENOMEM : 0; |
6742d293c
|
597 598 599 |
} return ret ? ERR_PTR(ret) : |
df06b37ff
|
600 |
follow_page_pte(vma, address, pmd, flags, &ctx->pgmap); |
69e68b4f0
|
601 |
} |
6742d293c
|
602 603 |
page = follow_trans_huge_pmd(vma, address, pmd, flags); spin_unlock(ptl); |
df06b37ff
|
604 |
ctx->page_mask = HPAGE_PMD_NR - 1; |
6742d293c
|
605 |
return page; |
4bbd4c776
|
606 |
} |
080dbb618
|
607 608 |
static struct page *follow_pud_mask(struct vm_area_struct *vma, unsigned long address, p4d_t *p4dp, |
df06b37ff
|
609 610 |
unsigned int flags, struct follow_page_context *ctx) |
080dbb618
|
611 612 613 614 615 616 617 618 619 |
{ pud_t *pud; spinlock_t *ptl; struct page *page; struct mm_struct *mm = vma->vm_mm; pud = pud_offset(p4dp, address); if (pud_none(*pud)) return no_page_table(vma, flags); |
be9d30458
|
620 |
if (pud_huge(*pud) && is_vm_hugetlb_page(vma)) { |
080dbb618
|
621 622 623 624 625 |
page = follow_huge_pud(mm, address, pud, flags); if (page) return page; return no_page_table(vma, flags); } |
4dc71451a
|
626 627 628 629 630 631 632 633 |
if (is_hugepd(__hugepd(pud_val(*pud)))) { page = follow_huge_pd(vma, address, __hugepd(pud_val(*pud)), flags, PUD_SHIFT); if (page) return page; return no_page_table(vma, flags); } |
080dbb618
|
634 635 |
if (pud_devmap(*pud)) { ptl = pud_lock(mm, pud); |
df06b37ff
|
636 |
page = follow_devmap_pud(vma, address, pud, flags, &ctx->pgmap); |
080dbb618
|
637 638 639 640 641 642 |
spin_unlock(ptl); if (page) return page; } if (unlikely(pud_bad(*pud))) return no_page_table(vma, flags); |
df06b37ff
|
643 |
return follow_pmd_mask(vma, address, pud, flags, ctx); |
080dbb618
|
644 |
} |
080dbb618
|
645 646 |
static struct page *follow_p4d_mask(struct vm_area_struct *vma, unsigned long address, pgd_t *pgdp, |
df06b37ff
|
647 648 |
unsigned int flags, struct follow_page_context *ctx) |
080dbb618
|
649 650 |
{ p4d_t *p4d; |
4dc71451a
|
651 |
struct page *page; |
080dbb618
|
652 653 654 655 656 657 658 |
p4d = p4d_offset(pgdp, address); if (p4d_none(*p4d)) return no_page_table(vma, flags); BUILD_BUG_ON(p4d_huge(*p4d)); if (unlikely(p4d_bad(*p4d))) return no_page_table(vma, flags); |
4dc71451a
|
659 660 661 662 663 664 665 666 |
if (is_hugepd(__hugepd(p4d_val(*p4d)))) { page = follow_huge_pd(vma, address, __hugepd(p4d_val(*p4d)), flags, P4D_SHIFT); if (page) return page; return no_page_table(vma, flags); } |
df06b37ff
|
667 |
return follow_pud_mask(vma, address, p4d, flags, ctx); |
080dbb618
|
668 669 670 671 672 673 674 |
} /** * follow_page_mask - look up a page descriptor from a user-virtual address * @vma: vm_area_struct mapping @address * @address: virtual address to look up * @flags: flags modifying lookup behaviour |
78179556e
|
675 676 |
* @ctx: contains dev_pagemap for %ZONE_DEVICE memory pinning and a * pointer to output page_mask |
080dbb618
|
677 678 679 |
* * @flags can have FOLL_ flags set, defined in <linux/mm.h> * |
78179556e
|
680 681 682 683 684 685 |
* When getting pages from ZONE_DEVICE memory, the @ctx->pgmap caches * the device's dev_pagemap metadata to avoid repeating expensive lookups. * * On output, the @ctx->page_mask is set according to the size of the page. * * Return: the mapped (struct page *), %NULL if no mapping exists, or |
080dbb618
|
686 687 688 |
* an error pointer if there is a mapping to something not represented * by a page descriptor (see also vm_normal_page()). */ |
a7030aea2
|
689 |
static struct page *follow_page_mask(struct vm_area_struct *vma, |
080dbb618
|
690 |
unsigned long address, unsigned int flags, |
df06b37ff
|
691 |
struct follow_page_context *ctx) |
080dbb618
|
692 693 694 695 |
{ pgd_t *pgd; struct page *page; struct mm_struct *mm = vma->vm_mm; |
df06b37ff
|
696 |
ctx->page_mask = 0; |
080dbb618
|
697 698 699 700 |
/* make this handle hugepd */ page = follow_huge_addr(mm, address, flags & FOLL_WRITE); if (!IS_ERR(page)) { |
3faa52c03
|
701 |
WARN_ON_ONCE(flags & (FOLL_GET | FOLL_PIN)); |
080dbb618
|
702 703 704 705 706 707 708 |
return page; } pgd = pgd_offset(mm, address); if (pgd_none(*pgd) || unlikely(pgd_bad(*pgd))) return no_page_table(vma, flags); |
faaa5b62d
|
709 710 711 712 713 714 |
if (pgd_huge(*pgd)) { page = follow_huge_pgd(mm, address, pgd, flags); if (page) return page; return no_page_table(vma, flags); } |
4dc71451a
|
715 716 717 718 719 720 721 722 |
if (is_hugepd(__hugepd(pgd_val(*pgd)))) { page = follow_huge_pd(vma, address, __hugepd(pgd_val(*pgd)), flags, PGDIR_SHIFT); if (page) return page; return no_page_table(vma, flags); } |
faaa5b62d
|
723 |
|
df06b37ff
|
724 725 726 727 728 729 730 731 732 733 734 735 736 |
return follow_p4d_mask(vma, address, pgd, flags, ctx); } struct page *follow_page(struct vm_area_struct *vma, unsigned long address, unsigned int foll_flags) { struct follow_page_context ctx = { NULL }; struct page *page; page = follow_page_mask(vma, address, foll_flags, &ctx); if (ctx.pgmap) put_dev_pagemap(ctx.pgmap); return page; |
080dbb618
|
737 |
} |
f2b495ca8
|
738 739 740 741 742 |
static int get_gate_page(struct mm_struct *mm, unsigned long address, unsigned int gup_flags, struct vm_area_struct **vma, struct page **page) { pgd_t *pgd; |
c2febafc6
|
743 |
p4d_t *p4d; |
f2b495ca8
|
744 745 746 747 748 749 750 751 752 753 754 755 |
pud_t *pud; pmd_t *pmd; pte_t *pte; int ret = -EFAULT; /* user gate pages are read-only */ if (gup_flags & FOLL_WRITE) return -EFAULT; if (address > TASK_SIZE) pgd = pgd_offset_k(address); else pgd = pgd_offset_gate(mm, address); |
b5d1c39f3
|
756 757 |
if (pgd_none(*pgd)) return -EFAULT; |
c2febafc6
|
758 |
p4d = p4d_offset(pgd, address); |
b5d1c39f3
|
759 760 |
if (p4d_none(*p4d)) return -EFAULT; |
c2febafc6
|
761 |
pud = pud_offset(p4d, address); |
b5d1c39f3
|
762 763 |
if (pud_none(*pud)) return -EFAULT; |
f2b495ca8
|
764 |
pmd = pmd_offset(pud, address); |
84c3fc4e9
|
765 |
if (!pmd_present(*pmd)) |
f2b495ca8
|
766 767 768 769 770 771 772 773 774 775 776 777 778 779 |
return -EFAULT; VM_BUG_ON(pmd_trans_huge(*pmd)); pte = pte_offset_map(pmd, address); if (pte_none(*pte)) goto unmap; *vma = get_gate_vma(mm); if (!page) goto out; *page = vm_normal_page(*vma, address, *pte); if (!*page) { if ((gup_flags & FOLL_DUMP) || !is_zero_pfn(pte_pfn(*pte))) goto unmap; *page = pte_page(*pte); } |
9fa2dd946
|
780 |
if (unlikely(!try_grab_page(*page, gup_flags))) { |
8fde12ca7
|
781 782 783 |
ret = -ENOMEM; goto unmap; } |
f2b495ca8
|
784 785 786 787 788 789 |
out: ret = 0; unmap: pte_unmap(pte); return ret; } |
9a95f3cf7
|
790 |
/* |
c1e8d7c6a
|
791 792 |
* mmap_lock must be held on entry. If @locked != NULL and *@flags * does not include FOLL_NOWAIT, the mmap_lock may be released. If it |
4f6da9341
|
793 |
* is, *@locked will be set to 0 and -EBUSY returned. |
9a95f3cf7
|
794 |
*/ |
64019a2e4
|
795 |
static int faultin_page(struct vm_area_struct *vma, |
4f6da9341
|
796 |
unsigned long address, unsigned int *flags, int *locked) |
167444834
|
797 |
{ |
167444834
|
798 |
unsigned int fault_flags = 0; |
2b7403035
|
799 |
vm_fault_t ret; |
167444834
|
800 |
|
de60f5f10
|
801 802 803 |
/* mlock all present pages, but do not fault in new pages */ if ((*flags & (FOLL_POPULATE | FOLL_MLOCK)) == FOLL_MLOCK) return -ENOENT; |
167444834
|
804 805 |
if (*flags & FOLL_WRITE) fault_flags |= FAULT_FLAG_WRITE; |
1b2ee1266
|
806 807 |
if (*flags & FOLL_REMOTE) fault_flags |= FAULT_FLAG_REMOTE; |
4f6da9341
|
808 |
if (locked) |
71335f37c
|
809 |
fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE; |
167444834
|
810 811 |
if (*flags & FOLL_NOWAIT) fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_RETRY_NOWAIT; |
234b239be
|
812 |
if (*flags & FOLL_TRIED) { |
4426e945d
|
813 814 815 816 |
/* * Note: FAULT_FLAG_ALLOW_RETRY and FAULT_FLAG_TRIED * can co-exist */ |
234b239be
|
817 818 |
fault_flags |= FAULT_FLAG_TRIED; } |
167444834
|
819 |
|
bce617ede
|
820 |
ret = handle_mm_fault(vma, address, fault_flags, NULL); |
167444834
|
821 |
if (ret & VM_FAULT_ERROR) { |
9a291a7c9
|
822 823 824 825 |
int err = vm_fault_to_errno(ret, *flags); if (err) return err; |
167444834
|
826 827 |
BUG(); } |
167444834
|
828 |
if (ret & VM_FAULT_RETRY) { |
4f6da9341
|
829 830 |
if (locked && !(fault_flags & FAULT_FLAG_RETRY_NOWAIT)) *locked = 0; |
167444834
|
831 832 833 834 835 836 837 838 839 840 841 842 843 |
return -EBUSY; } /* * The VM_FAULT_WRITE bit tells us that do_wp_page has broken COW when * necessary, even if maybe_mkwrite decided not to set pte_write. We * can thus safely do subsequent page lookups as if they were reads. * But only do so when looping for pte_write is futile: in some cases * userspace may also be wanting to write to the gotten user page, * which a read fault here might prevent (a readonly page might get * reCOWed by userspace write). */ if ((ret & VM_FAULT_WRITE) && !(vma->vm_flags & VM_WRITE)) |
2923117b7
|
844 |
*flags |= FOLL_COW; |
167444834
|
845 846 |
return 0; } |
fa5bb2093
|
847 848 849 |
static int check_vma_flags(struct vm_area_struct *vma, unsigned long gup_flags) { vm_flags_t vm_flags = vma->vm_flags; |
1b2ee1266
|
850 851 |
int write = (gup_flags & FOLL_WRITE); int foreign = (gup_flags & FOLL_REMOTE); |
fa5bb2093
|
852 853 854 |
if (vm_flags & (VM_IO | VM_PFNMAP)) return -EFAULT; |
7f7ccc2cc
|
855 856 |
if (gup_flags & FOLL_ANON && !vma_is_anonymous(vma)) return -EFAULT; |
1b2ee1266
|
857 |
if (write) { |
fa5bb2093
|
858 859 860 861 862 863 864 865 866 867 868 869 |
if (!(vm_flags & VM_WRITE)) { if (!(gup_flags & FOLL_FORCE)) return -EFAULT; /* * We used to let the write,force case do COW in a * VM_MAYWRITE VM_SHARED !VM_WRITE vma, so ptrace could * set a breakpoint in a read-only mapping of an * executable, without corrupting the file (yet only * when that file had been opened for writing!). * Anon pages in shared mappings are surprising: now * just reject it. */ |
464353647
|
870 |
if (!is_cow_mapping(vm_flags)) |
fa5bb2093
|
871 |
return -EFAULT; |
fa5bb2093
|
872 873 874 875 876 877 878 879 880 881 882 |
} } else if (!(vm_flags & VM_READ)) { if (!(gup_flags & FOLL_FORCE)) return -EFAULT; /* * Is there actually any vma we can reach here which does not * have VM_MAYREAD set? */ if (!(vm_flags & VM_MAYREAD)) return -EFAULT; } |
d61172b4b
|
883 884 885 886 887 |
/* * gups are always data accesses, not instruction * fetches, so execute=false here */ if (!arch_vma_access_permitted(vma, write, false, foreign)) |
33a709b25
|
888 |
return -EFAULT; |
fa5bb2093
|
889 890 |
return 0; } |
4bbd4c776
|
891 892 |
/** * __get_user_pages() - pin user pages in memory |
4bbd4c776
|
893 894 895 896 897 898 899 900 901 |
* @mm: mm_struct of target mm * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. Or NULL, if caller * only intends to ensure the pages are faulted in. * @vmas: array of pointers to vmas corresponding to each page. * Or NULL if the caller does not require them. |
c1e8d7c6a
|
902 |
* @locked: whether we're still with the mmap_lock held |
4bbd4c776
|
903 |
* |
d2dfbe47f
|
904 905 906 907 908 909 910 |
* Returns either number of pages pinned (which may be less than the * number requested), or an error. Details about the return value: * * -- If nr_pages is 0, returns 0. * -- If nr_pages is >0, but no pages were pinned, returns -errno. * -- If nr_pages is >0, and some pages were pinned, returns the number of * pages pinned. Again, this may be less than nr_pages. |
2d3a36a47
|
911 |
* -- 0 return value is possible when the fault would need to be retried. |
d2dfbe47f
|
912 913 914 |
* * The caller is responsible for releasing returned @pages, via put_page(). * |
c1e8d7c6a
|
915 |
* @vmas are valid only as long as mmap_lock is held. |
4bbd4c776
|
916 |
* |
c1e8d7c6a
|
917 |
* Must be called with mmap_lock held. It may be released. See below. |
4bbd4c776
|
918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 |
* * __get_user_pages walks a process's page tables and takes a reference to * each struct page that each user address corresponds to at a given * instant. That is, it takes the page that would be accessed if a user * thread accesses the given user virtual address at that instant. * * This does not guarantee that the page exists in the user mappings when * __get_user_pages returns, and there may even be a completely different * page there in some cases (eg. if mmapped pagecache has been invalidated * and subsequently re faulted). However it does guarantee that the page * won't be freed completely. And mostly callers simply care that the page * contains data that was valid *at some point in time*. Typically, an IO * or similar operation cannot guarantee anything stronger anyway because * locks can't be held over the syscall boundary. * * If @gup_flags & FOLL_WRITE == 0, the page must not be written to. If * the page is written to, set_page_dirty (or set_page_dirty_lock, as * appropriate) must be called after the page is finished with, and * before put_page is called. * |
c1e8d7c6a
|
938 |
* If @locked != NULL, *@locked will be set to 0 when mmap_lock is |
4f6da9341
|
939 940 |
* released by an up_read(). That can happen if @gup_flags does not * have FOLL_NOWAIT. |
9a95f3cf7
|
941 |
* |
4f6da9341
|
942 |
* A caller using such a combination of @locked and @gup_flags |
c1e8d7c6a
|
943 |
* must therefore hold the mmap_lock for reading only, and recognize |
9a95f3cf7
|
944 945 |
* when it's been released. Otherwise, it must be held for either * reading or writing and will not be released. |
4bbd4c776
|
946 947 948 949 950 |
* * In most cases, get_user_pages or get_user_pages_fast should be used * instead of __get_user_pages. __get_user_pages should be used only if * you need some special @gup_flags. */ |
64019a2e4
|
951 |
static long __get_user_pages(struct mm_struct *mm, |
4bbd4c776
|
952 953 |
unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, |
4f6da9341
|
954 |
struct vm_area_struct **vmas, int *locked) |
4bbd4c776
|
955 |
{ |
df06b37ff
|
956 |
long ret = 0, i = 0; |
fa5bb2093
|
957 |
struct vm_area_struct *vma = NULL; |
df06b37ff
|
958 |
struct follow_page_context ctx = { NULL }; |
4bbd4c776
|
959 960 961 |
if (!nr_pages) return 0; |
f96525941
|
962 |
start = untagged_addr(start); |
eddb1c228
|
963 |
VM_BUG_ON(!!pages != !!(gup_flags & (FOLL_GET | FOLL_PIN))); |
4bbd4c776
|
964 965 966 967 968 969 970 971 |
/* * If FOLL_FORCE is set then do not force a full fault as the hinting * fault information is unrelated to the reference behaviour of a task * using the address space */ if (!(gup_flags & FOLL_FORCE)) gup_flags |= FOLL_NUMA; |
4bbd4c776
|
972 |
do { |
fa5bb2093
|
973 974 975 976 977 978 979 980 |
struct page *page; unsigned int foll_flags = gup_flags; unsigned int page_increm; /* first iteration or cross vma bound */ if (!vma || start >= vma->vm_end) { vma = find_extend_vma(mm, start); if (!vma && in_gate_area(mm, start)) { |
fa5bb2093
|
981 982 983 984 |
ret = get_gate_page(mm, start & PAGE_MASK, gup_flags, &vma, pages ? &pages[i] : NULL); if (ret) |
08be37b79
|
985 |
goto out; |
df06b37ff
|
986 |
ctx.page_mask = 0; |
fa5bb2093
|
987 988 |
goto next_page; } |
4bbd4c776
|
989 |
|
df06b37ff
|
990 991 992 993 |
if (!vma || check_vma_flags(vma, gup_flags)) { ret = -EFAULT; goto out; } |
fa5bb2093
|
994 995 996 |
if (is_vm_hugetlb_page(vma)) { i = follow_hugetlb_page(mm, vma, pages, vmas, &start, &nr_pages, i, |
a308c71bf
|
997 |
gup_flags, locked); |
ad415db81
|
998 999 1000 |
if (locked && *locked == 0) { /* * We've got a VM_FAULT_RETRY |
c1e8d7c6a
|
1001 |
* and we've lost mmap_lock. |
ad415db81
|
1002 1003 1004 1005 1006 1007 |
* We must stop here. */ BUG_ON(gup_flags & FOLL_NOWAIT); BUG_ON(ret != 0); goto out; } |
fa5bb2093
|
1008 |
continue; |
4bbd4c776
|
1009 |
} |
fa5bb2093
|
1010 1011 1012 1013 1014 1015 |
} retry: /* * If we have a pending SIGKILL, don't keep faulting pages and * potentially allocating memory. */ |
fa45f1162
|
1016 |
if (fatal_signal_pending(current)) { |
d180870d8
|
1017 |
ret = -EINTR; |
df06b37ff
|
1018 1019 |
goto out; } |
fa5bb2093
|
1020 |
cond_resched(); |
df06b37ff
|
1021 1022 |
page = follow_page_mask(vma, start, foll_flags, &ctx); |
fa5bb2093
|
1023 |
if (!page) { |
64019a2e4
|
1024 |
ret = faultin_page(vma, start, &foll_flags, locked); |
fa5bb2093
|
1025 1026 1027 |
switch (ret) { case 0: goto retry; |
df06b37ff
|
1028 1029 |
case -EBUSY: ret = 0; |
e4a9bc589
|
1030 |
fallthrough; |
fa5bb2093
|
1031 1032 1033 |
case -EFAULT: case -ENOMEM: case -EHWPOISON: |
df06b37ff
|
1034 |
goto out; |
fa5bb2093
|
1035 1036 |
case -ENOENT: goto next_page; |
4bbd4c776
|
1037 |
} |
fa5bb2093
|
1038 |
BUG(); |
1027e4436
|
1039 1040 1041 1042 1043 1044 1045 |
} else if (PTR_ERR(page) == -EEXIST) { /* * Proper page table entry exists, but no corresponding * struct page. */ goto next_page; } else if (IS_ERR(page)) { |
df06b37ff
|
1046 1047 |
ret = PTR_ERR(page); goto out; |
1027e4436
|
1048 |
} |
fa5bb2093
|
1049 1050 1051 1052 |
if (pages) { pages[i] = page; flush_anon_page(vma, page, start); flush_dcache_page(page); |
df06b37ff
|
1053 |
ctx.page_mask = 0; |
4bbd4c776
|
1054 |
} |
4bbd4c776
|
1055 |
next_page: |
fa5bb2093
|
1056 1057 |
if (vmas) { vmas[i] = vma; |
df06b37ff
|
1058 |
ctx.page_mask = 0; |
fa5bb2093
|
1059 |
} |
df06b37ff
|
1060 |
page_increm = 1 + (~(start >> PAGE_SHIFT) & ctx.page_mask); |
fa5bb2093
|
1061 1062 1063 1064 1065 |
if (page_increm > nr_pages) page_increm = nr_pages; i += page_increm; start += page_increm * PAGE_SIZE; nr_pages -= page_increm; |
4bbd4c776
|
1066 |
} while (nr_pages); |
df06b37ff
|
1067 1068 1069 1070 |
out: if (ctx.pgmap) put_dev_pagemap(ctx.pgmap); return i ? i : ret; |
4bbd4c776
|
1071 |
} |
4bbd4c776
|
1072 |
|
771ab4302
|
1073 1074 |
static bool vma_permits_fault(struct vm_area_struct *vma, unsigned int fault_flags) |
d4925e00d
|
1075 |
{ |
1b2ee1266
|
1076 1077 |
bool write = !!(fault_flags & FAULT_FLAG_WRITE); bool foreign = !!(fault_flags & FAULT_FLAG_REMOTE); |
33a709b25
|
1078 |
vm_flags_t vm_flags = write ? VM_WRITE : VM_READ; |
d4925e00d
|
1079 1080 1081 |
if (!(vm_flags & vma->vm_flags)) return false; |
33a709b25
|
1082 1083 |
/* * The architecture might have a hardware protection |
1b2ee1266
|
1084 |
* mechanism other than read/write that can deny access. |
d61172b4b
|
1085 1086 1087 |
* * gup always represents data access, not instruction * fetches, so execute=false here: |
33a709b25
|
1088 |
*/ |
d61172b4b
|
1089 |
if (!arch_vma_access_permitted(vma, write, false, foreign)) |
33a709b25
|
1090 |
return false; |
d4925e00d
|
1091 1092 |
return true; } |
adc8cb406
|
1093 |
/** |
4bbd4c776
|
1094 |
* fixup_user_fault() - manually resolve a user page fault |
4bbd4c776
|
1095 1096 1097 |
* @mm: mm_struct of target mm * @address: user address * @fault_flags:flags to pass down to handle_mm_fault() |
c1e8d7c6a
|
1098 |
* @unlocked: did we unlock the mmap_lock while retrying, maybe NULL if caller |
548b6a1e5
|
1099 1100 |
* does not allow retry. If NULL, the caller must guarantee * that fault_flags does not contain FAULT_FLAG_ALLOW_RETRY. |
4bbd4c776
|
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 |
* * This is meant to be called in the specific scenario where for locking reasons * we try to access user memory in atomic context (within a pagefault_disable() * section), this returns -EFAULT, and we want to resolve the user fault before * trying again. * * Typically this is meant to be used by the futex code. * * The main difference with get_user_pages() is that this function will * unconditionally call handle_mm_fault() which will in turn perform all the * necessary SW fixup of the dirty and young bits in the PTE, while |
4a9e1cda2
|
1112 |
* get_user_pages() only guarantees to update these in the struct page. |
4bbd4c776
|
1113 1114 1115 1116 1117 1118 |
* * This is important for some architectures where those bits also gate the * access permission to the page because they are maintained in software. On * such architectures, gup() will not be enough to make a subsequent access * succeed. * |
c1e8d7c6a
|
1119 1120 |
* This function will not return with an unlocked mmap_lock. So it has not the * same semantics wrt the @mm->mmap_lock as does filemap_fault(). |
4bbd4c776
|
1121 |
*/ |
64019a2e4
|
1122 |
int fixup_user_fault(struct mm_struct *mm, |
4a9e1cda2
|
1123 1124 |
unsigned long address, unsigned int fault_flags, bool *unlocked) |
4bbd4c776
|
1125 1126 |
{ struct vm_area_struct *vma; |
2b7403035
|
1127 |
vm_fault_t ret, major = 0; |
4a9e1cda2
|
1128 |
|
f96525941
|
1129 |
address = untagged_addr(address); |
4a9e1cda2
|
1130 |
if (unlocked) |
71335f37c
|
1131 |
fault_flags |= FAULT_FLAG_ALLOW_RETRY | FAULT_FLAG_KILLABLE; |
4bbd4c776
|
1132 |
|
4a9e1cda2
|
1133 |
retry: |
4bbd4c776
|
1134 1135 1136 |
vma = find_extend_vma(mm, address); if (!vma || address < vma->vm_start) return -EFAULT; |
d4925e00d
|
1137 |
if (!vma_permits_fault(vma, fault_flags)) |
4bbd4c776
|
1138 |
return -EFAULT; |
475f4dfc0
|
1139 1140 1141 |
if ((fault_flags & FAULT_FLAG_KILLABLE) && fatal_signal_pending(current)) return -EINTR; |
bce617ede
|
1142 |
ret = handle_mm_fault(vma, address, fault_flags, NULL); |
4a9e1cda2
|
1143 |
major |= ret & VM_FAULT_MAJOR; |
4bbd4c776
|
1144 |
if (ret & VM_FAULT_ERROR) { |
9a291a7c9
|
1145 1146 1147 1148 |
int err = vm_fault_to_errno(ret, 0); if (err) return err; |
4bbd4c776
|
1149 1150 |
BUG(); } |
4a9e1cda2
|
1151 1152 |
if (ret & VM_FAULT_RETRY) { |
d8ed45c5d
|
1153 |
mmap_read_lock(mm); |
475f4dfc0
|
1154 1155 1156 |
*unlocked = true; fault_flags |= FAULT_FLAG_TRIED; goto retry; |
4a9e1cda2
|
1157 |
} |
4bbd4c776
|
1158 1159 |
return 0; } |
add6a0cd1
|
1160 |
EXPORT_SYMBOL_GPL(fixup_user_fault); |
4bbd4c776
|
1161 |
|
2d3a36a47
|
1162 1163 1164 1165 |
/* * Please note that this function, unlike __get_user_pages will not * return 0 for nr_pages > 0 without FOLL_NOWAIT */ |
64019a2e4
|
1166 |
static __always_inline long __get_user_pages_locked(struct mm_struct *mm, |
f0818f472
|
1167 1168 |
unsigned long start, unsigned long nr_pages, |
f0818f472
|
1169 1170 |
struct page **pages, struct vm_area_struct **vmas, |
e716712f8
|
1171 |
int *locked, |
0fd71a56f
|
1172 |
unsigned int flags) |
f0818f472
|
1173 |
{ |
f0818f472
|
1174 1175 1176 1177 1178 1179 1180 1181 1182 |
long ret, pages_done; bool lock_dropped; if (locked) { /* if VM_FAULT_RETRY can be returned, vmas become invalid */ BUG_ON(vmas); /* check caller initialized locked */ BUG_ON(*locked != 1); } |
008cfe441
|
1183 |
if (flags & FOLL_PIN) |
a4d63c373
|
1184 |
atomic_set(&mm->has_pinned, 1); |
008cfe441
|
1185 |
|
eddb1c228
|
1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 |
/* * FOLL_PIN and FOLL_GET are mutually exclusive. Traditional behavior * is to set FOLL_GET if the caller wants pages[] filled in (but has * carelessly failed to specify FOLL_GET), so keep doing that, but only * for FOLL_GET, not for the newer FOLL_PIN. * * FOLL_PIN always expects pages to be non-null, but no need to assert * that here, as any failures will be obvious enough. */ if (pages && !(flags & FOLL_PIN)) |
f0818f472
|
1196 |
flags |= FOLL_GET; |
f0818f472
|
1197 1198 1199 1200 |
pages_done = 0; lock_dropped = false; for (;;) { |
64019a2e4
|
1201 |
ret = __get_user_pages(mm, start, nr_pages, flags, pages, |
f0818f472
|
1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 |
vmas, locked); if (!locked) /* VM_FAULT_RETRY couldn't trigger, bypass */ return ret; /* VM_FAULT_RETRY cannot return errors */ if (!*locked) { BUG_ON(ret < 0); BUG_ON(ret >= nr_pages); } |
f0818f472
|
1212 1213 1214 1215 1216 1217 1218 |
if (ret > 0) { nr_pages -= ret; pages_done += ret; if (!nr_pages) break; } if (*locked) { |
96312e612
|
1219 1220 1221 1222 |
/* * VM_FAULT_RETRY didn't trigger or it was a * FOLL_NOWAIT. */ |
f0818f472
|
1223 1224 1225 1226 |
if (!pages_done) pages_done = ret; break; } |
df17277b2
|
1227 1228 1229 1230 1231 1232 |
/* * VM_FAULT_RETRY triggered, so seek to the faulting offset. * For the prefault case (!pages) we only update counts. */ if (likely(pages)) pages += ret; |
f0818f472
|
1233 |
start += ret << PAGE_SHIFT; |
4426e945d
|
1234 |
lock_dropped = true; |
f0818f472
|
1235 |
|
4426e945d
|
1236 |
retry: |
f0818f472
|
1237 1238 |
/* * Repeat on the address that fired VM_FAULT_RETRY |
4426e945d
|
1239 1240 1241 1242 |
* with both FAULT_FLAG_ALLOW_RETRY and * FAULT_FLAG_TRIED. Note that GUP can be interrupted * by fatal signals, so we need to check it before we * start trying again otherwise it can loop forever. |
f0818f472
|
1243 |
*/ |
4426e945d
|
1244 |
|
ae46d2aa6
|
1245 1246 1247 |
if (fatal_signal_pending(current)) { if (!pages_done) pages_done = -EINTR; |
4426e945d
|
1248 |
break; |
ae46d2aa6
|
1249 |
} |
4426e945d
|
1250 |
|
d8ed45c5d
|
1251 |
ret = mmap_read_lock_killable(mm); |
71335f37c
|
1252 1253 1254 1255 1256 1257 |
if (ret) { BUG_ON(ret > 0); if (!pages_done) pages_done = ret; break; } |
4426e945d
|
1258 |
|
c7b6a566b
|
1259 |
*locked = 1; |
64019a2e4
|
1260 |
ret = __get_user_pages(mm, start, 1, flags | FOLL_TRIED, |
4426e945d
|
1261 1262 1263 1264 1265 1266 |
pages, NULL, locked); if (!*locked) { /* Continue to retry until we succeeded */ BUG_ON(ret != 0); goto retry; } |
f0818f472
|
1267 1268 1269 1270 1271 1272 1273 1274 1275 1276 |
if (ret != 1) { BUG_ON(ret > 1); if (!pages_done) pages_done = ret; break; } nr_pages--; pages_done++; if (!nr_pages) break; |
df17277b2
|
1277 1278 |
if (likely(pages)) pages++; |
f0818f472
|
1279 1280 |
start += PAGE_SIZE; } |
e716712f8
|
1281 |
if (lock_dropped && *locked) { |
f0818f472
|
1282 1283 1284 1285 |
/* * We must let the caller know we temporarily dropped the lock * and so the critical section protected by it was lost. */ |
d8ed45c5d
|
1286 |
mmap_read_unlock(mm); |
f0818f472
|
1287 1288 1289 1290 |
*locked = 0; } return pages_done; } |
d3649f68b
|
1291 1292 1293 1294 1295 |
/** * populate_vma_page_range() - populate a range of pages in the vma. * @vma: target vma * @start: start address * @end: end address |
c1e8d7c6a
|
1296 |
* @locked: whether the mmap_lock is still held |
d3649f68b
|
1297 1298 1299 |
* * This takes care of mlocking the pages too if VM_LOCKED is set. * |
0a36f7f85
|
1300 1301 |
* Return either number of pages pinned in the vma, or a negative error * code on error. |
d3649f68b
|
1302 |
* |
c1e8d7c6a
|
1303 |
* vma->vm_mm->mmap_lock must be held. |
d3649f68b
|
1304 |
* |
4f6da9341
|
1305 |
* If @locked is NULL, it may be held for read or write and will |
d3649f68b
|
1306 1307 |
* be unperturbed. * |
4f6da9341
|
1308 1309 |
* If @locked is non-NULL, it must held for read only and may be * released. If it's released, *@locked will be set to 0. |
d3649f68b
|
1310 1311 |
*/ long populate_vma_page_range(struct vm_area_struct *vma, |
4f6da9341
|
1312 |
unsigned long start, unsigned long end, int *locked) |
d3649f68b
|
1313 1314 1315 1316 1317 1318 1319 1320 1321 |
{ struct mm_struct *mm = vma->vm_mm; unsigned long nr_pages = (end - start) / PAGE_SIZE; int gup_flags; VM_BUG_ON(start & ~PAGE_MASK); VM_BUG_ON(end & ~PAGE_MASK); VM_BUG_ON_VMA(start < vma->vm_start, vma); VM_BUG_ON_VMA(end > vma->vm_end, vma); |
42fc54140
|
1322 |
mmap_assert_locked(mm); |
d3649f68b
|
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 |
gup_flags = FOLL_TOUCH | FOLL_POPULATE | FOLL_MLOCK; if (vma->vm_flags & VM_LOCKONFAULT) gup_flags &= ~FOLL_POPULATE; /* * We want to touch writable mappings with a write fault in order * to break COW, except for shared mappings because these don't COW * and we would not want to dirty them for nothing. */ if ((vma->vm_flags & (VM_WRITE | VM_SHARED)) == VM_WRITE) gup_flags |= FOLL_WRITE; /* * We want mlock to succeed for regions that have any permissions * other than PROT_NONE. */ |
3122e80ef
|
1339 |
if (vma_is_accessible(vma)) |
d3649f68b
|
1340 1341 1342 1343 1344 1345 |
gup_flags |= FOLL_FORCE; /* * We made sure addr is within a VMA, so the following will * not result in a stack expansion that recurses back here. */ |
64019a2e4
|
1346 |
return __get_user_pages(mm, start, nr_pages, gup_flags, |
4f6da9341
|
1347 |
NULL, NULL, locked); |
d3649f68b
|
1348 1349 1350 1351 1352 1353 1354 |
} /* * __mm_populate - populate and/or mlock pages within a range of address space. * * This is used to implement mlock() and the MAP_POPULATE / MAP_LOCKED mmap * flags. VMAs must be already marked with the desired vm_flags, and |
c1e8d7c6a
|
1355 |
* mmap_lock must not be held. |
d3649f68b
|
1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 |
*/ int __mm_populate(unsigned long start, unsigned long len, int ignore_errors) { struct mm_struct *mm = current->mm; unsigned long end, nstart, nend; struct vm_area_struct *vma = NULL; int locked = 0; long ret = 0; end = start + len; for (nstart = start; nstart < end; nstart = nend) { /* * We want to fault in pages for [nstart; end) address range. * Find first corresponding VMA. */ if (!locked) { locked = 1; |
d8ed45c5d
|
1374 |
mmap_read_lock(mm); |
d3649f68b
|
1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 |
vma = find_vma(mm, nstart); } else if (nstart >= vma->vm_end) vma = vma->vm_next; if (!vma || vma->vm_start >= end) break; /* * Set [nstart; nend) to intersection of desired address * range with the first VMA. Also, skip undesirable VMA types. */ nend = min(end, vma->vm_end); if (vma->vm_flags & (VM_IO | VM_PFNMAP)) continue; if (nstart < vma->vm_start) nstart = vma->vm_start; /* * Now fault in a range of pages. populate_vma_page_range() * double checks the vma flags, so that it won't mlock pages * if the vma was already munlocked. */ ret = populate_vma_page_range(vma, nstart, nend, &locked); if (ret < 0) { if (ignore_errors) { ret = 0; continue; /* continue at next VMA */ } break; } nend = nstart + ret * PAGE_SIZE; ret = 0; } if (locked) |
d8ed45c5d
|
1406 |
mmap_read_unlock(mm); |
d3649f68b
|
1407 1408 |
return ret; /* 0 or negative error code */ } |
050a9adc6
|
1409 |
#else /* CONFIG_MMU */ |
64019a2e4
|
1410 |
static long __get_user_pages_locked(struct mm_struct *mm, unsigned long start, |
050a9adc6
|
1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 |
unsigned long nr_pages, struct page **pages, struct vm_area_struct **vmas, int *locked, unsigned int foll_flags) { struct vm_area_struct *vma; unsigned long vm_flags; int i; /* calculate required read or write permissions. * If FOLL_FORCE is set, we only require the "MAY" flags. */ vm_flags = (foll_flags & FOLL_WRITE) ? (VM_WRITE | VM_MAYWRITE) : (VM_READ | VM_MAYREAD); vm_flags &= (foll_flags & FOLL_FORCE) ? (VM_MAYREAD | VM_MAYWRITE) : (VM_READ | VM_WRITE); for (i = 0; i < nr_pages; i++) { vma = find_vma(mm, start); if (!vma) goto finish_or_fault; /* protect what we can, including chardevs */ if ((vma->vm_flags & (VM_IO | VM_PFNMAP)) || !(vm_flags & vma->vm_flags)) goto finish_or_fault; if (pages) { pages[i] = virt_to_page(start); if (pages[i]) get_page(pages[i]); } if (vmas) vmas[i] = vma; start = (start + PAGE_SIZE) & PAGE_MASK; } return i; finish_or_fault: return i ? : -EFAULT; } #endif /* !CONFIG_MMU */ |
d3649f68b
|
1453 |
|
8f942eea1
|
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 |
/** * get_dump_page() - pin user page in memory while writing it to core dump * @addr: user address * * Returns struct page pointer of user page pinned for dump, * to be freed afterwards by put_page(). * * Returns NULL on any kind of failure - a hole must then be inserted into * the corefile, to preserve alignment with its headers; and also returns * NULL wherever the ZERO_PAGE, or an anonymous pte_none, has been found - * allowing a hole to be left in the corefile to save diskspace. * |
7f3bfab52
|
1466 |
* Called without mmap_lock (takes and releases the mmap_lock by itself). |
8f942eea1
|
1467 1468 1469 1470 |
*/ #ifdef CONFIG_ELF_CORE struct page *get_dump_page(unsigned long addr) { |
7f3bfab52
|
1471 |
struct mm_struct *mm = current->mm; |
8f942eea1
|
1472 |
struct page *page; |
7f3bfab52
|
1473 1474 |
int locked = 1; int ret; |
8f942eea1
|
1475 |
|
7f3bfab52
|
1476 |
if (mmap_read_lock_killable(mm)) |
8f942eea1
|
1477 |
return NULL; |
7f3bfab52
|
1478 1479 1480 1481 1482 |
ret = __get_user_pages_locked(mm, addr, 1, &page, NULL, &locked, FOLL_FORCE | FOLL_DUMP | FOLL_GET); if (locked) mmap_read_unlock(mm); return (ret == 1) ? page : NULL; |
8f942eea1
|
1483 1484 |
} #endif /* CONFIG_ELF_CORE */ |
9a4e9f3b2
|
1485 |
#if defined(CONFIG_FS_DAX) || defined (CONFIG_CMA) |
9a4e9f3b2
|
1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 |
static bool check_dax_vmas(struct vm_area_struct **vmas, long nr_pages) { long i; struct vm_area_struct *vma_prev = NULL; for (i = 0; i < nr_pages; i++) { struct vm_area_struct *vma = vmas[i]; if (vma == vma_prev) continue; vma_prev = vma; if (vma_is_fsdax(vma)) return true; } return false; } |
9a4e9f3b2
|
1504 1505 |
#ifdef CONFIG_CMA |
64019a2e4
|
1506 |
static long check_and_migrate_cma_pages(struct mm_struct *mm, |
932f4a630
|
1507 1508 |
unsigned long start, unsigned long nr_pages, |
9a4e9f3b2
|
1509 |
struct page **pages, |
932f4a630
|
1510 1511 |
struct vm_area_struct **vmas, unsigned int gup_flags) |
9a4e9f3b2
|
1512 |
{ |
aa712399c
|
1513 1514 |
unsigned long i; unsigned long step; |
9a4e9f3b2
|
1515 1516 1517 |
bool drain_allow = true; bool migrate_allow = true; LIST_HEAD(cma_page_list); |
b96cc6551
|
1518 |
long ret = nr_pages; |
ed03d9245
|
1519 1520 1521 1522 |
struct migration_target_control mtc = { .nid = NUMA_NO_NODE, .gfp_mask = GFP_USER | __GFP_MOVABLE | __GFP_NOWARN, }; |
9a4e9f3b2
|
1523 1524 |
check_again: |
aa712399c
|
1525 1526 1527 1528 1529 1530 1531 1532 |
for (i = 0; i < nr_pages;) { struct page *head = compound_head(pages[i]); /* * gup may start from a tail page. Advance step by the left * part. */ |
d8c6546b1
|
1533 |
step = compound_nr(head) - (pages[i] - head); |
9a4e9f3b2
|
1534 1535 1536 1537 1538 |
/* * If we get a page from the CMA zone, since we are going to * be pinning these entries, we might as well move them out * of the CMA zone if possible. */ |
aa712399c
|
1539 1540 |
if (is_migrate_cma_page(head)) { if (PageHuge(head)) |
9a4e9f3b2
|
1541 |
isolate_huge_page(head, &cma_page_list); |
aa712399c
|
1542 |
else { |
9a4e9f3b2
|
1543 1544 1545 1546 1547 1548 1549 1550 1551 |
if (!PageLRU(head) && drain_allow) { lru_add_drain_all(); drain_allow = false; } if (!isolate_lru_page(head)) { list_add_tail(&head->lru, &cma_page_list); mod_node_page_state(page_pgdat(head), NR_ISOLATED_ANON + |
9de4f22a6
|
1552 |
page_is_file_lru(head), |
6c357848b
|
1553 |
thp_nr_pages(head)); |
9a4e9f3b2
|
1554 1555 1556 |
} } } |
aa712399c
|
1557 1558 |
i += step; |
9a4e9f3b2
|
1559 1560 1561 1562 1563 1564 |
} if (!list_empty(&cma_page_list)) { /* * drop the above get_user_pages reference. */ |
96e1fac16
|
1565 1566 1567 1568 1569 |
if (gup_flags & FOLL_PIN) unpin_user_pages(pages, nr_pages); else for (i = 0; i < nr_pages; i++) put_page(pages[i]); |
9a4e9f3b2
|
1570 |
|
ed03d9245
|
1571 1572 |
if (migrate_pages(&cma_page_list, alloc_migration_target, NULL, (unsigned long)&mtc, MIGRATE_SYNC, MR_CONTIG_RANGE)) { |
9a4e9f3b2
|
1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 |
/* * some of the pages failed migration. Do get_user_pages * without migration. */ migrate_allow = false; if (!list_empty(&cma_page_list)) putback_movable_pages(&cma_page_list); } /* |
932f4a630
|
1583 1584 1585 |
* We did migrate all the pages, Try to get the page references * again migrating any new CMA pages which we failed to isolate * earlier. |
9a4e9f3b2
|
1586 |
*/ |
64019a2e4
|
1587 |
ret = __get_user_pages_locked(mm, start, nr_pages, |
932f4a630
|
1588 1589 |
pages, vmas, NULL, gup_flags); |
b96cc6551
|
1590 1591 |
if ((ret > 0) && migrate_allow) { nr_pages = ret; |
9a4e9f3b2
|
1592 1593 1594 1595 |
drain_allow = true; goto check_again; } } |
b96cc6551
|
1596 |
return ret; |
9a4e9f3b2
|
1597 1598 |
} #else |
64019a2e4
|
1599 |
static long check_and_migrate_cma_pages(struct mm_struct *mm, |
932f4a630
|
1600 1601 1602 1603 1604 |
unsigned long start, unsigned long nr_pages, struct page **pages, struct vm_area_struct **vmas, unsigned int gup_flags) |
9a4e9f3b2
|
1605 1606 1607 |
{ return nr_pages; } |
050a9adc6
|
1608 |
#endif /* CONFIG_CMA */ |
9a4e9f3b2
|
1609 |
|
2bb6d2837
|
1610 |
/* |
932f4a630
|
1611 1612 |
* __gup_longterm_locked() is a wrapper for __get_user_pages_locked which * allows us to process the FOLL_LONGTERM flag. |
2bb6d2837
|
1613 |
*/ |
64019a2e4
|
1614 |
static long __gup_longterm_locked(struct mm_struct *mm, |
932f4a630
|
1615 1616 1617 1618 1619 |
unsigned long start, unsigned long nr_pages, struct page **pages, struct vm_area_struct **vmas, unsigned int gup_flags) |
2bb6d2837
|
1620 |
{ |
932f4a630
|
1621 1622 |
struct vm_area_struct **vmas_tmp = vmas; unsigned long flags = 0; |
2bb6d2837
|
1623 |
long rc, i; |
932f4a630
|
1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 |
if (gup_flags & FOLL_LONGTERM) { if (!pages) return -EINVAL; if (!vmas_tmp) { vmas_tmp = kcalloc(nr_pages, sizeof(struct vm_area_struct *), GFP_KERNEL); if (!vmas_tmp) return -ENOMEM; } flags = memalloc_nocma_save(); |
2bb6d2837
|
1636 |
} |
64019a2e4
|
1637 |
rc = __get_user_pages_locked(mm, start, nr_pages, pages, |
932f4a630
|
1638 |
vmas_tmp, NULL, gup_flags); |
2bb6d2837
|
1639 |
|
932f4a630
|
1640 |
if (gup_flags & FOLL_LONGTERM) { |
932f4a630
|
1641 1642 1643 1644 |
if (rc < 0) goto out; if (check_dax_vmas(vmas_tmp, rc)) { |
96e1fac16
|
1645 1646 1647 1648 1649 |
if (gup_flags & FOLL_PIN) unpin_user_pages(pages, rc); else for (i = 0; i < rc; i++) put_page(pages[i]); |
932f4a630
|
1650 1651 1652 |
rc = -EOPNOTSUPP; goto out; } |
64019a2e4
|
1653 |
rc = check_and_migrate_cma_pages(mm, start, rc, pages, |
932f4a630
|
1654 |
vmas_tmp, gup_flags); |
41b4dc14e
|
1655 1656 |
out: memalloc_nocma_restore(flags); |
9a4e9f3b2
|
1657 |
} |
2bb6d2837
|
1658 |
|
932f4a630
|
1659 1660 |
if (vmas_tmp != vmas) kfree(vmas_tmp); |
2bb6d2837
|
1661 1662 |
return rc; } |
932f4a630
|
1663 |
#else /* !CONFIG_FS_DAX && !CONFIG_CMA */ |
64019a2e4
|
1664 |
static __always_inline long __gup_longterm_locked(struct mm_struct *mm, |
932f4a630
|
1665 1666 1667 1668 1669 1670 |
unsigned long start, unsigned long nr_pages, struct page **pages, struct vm_area_struct **vmas, unsigned int flags) { |
64019a2e4
|
1671 |
return __get_user_pages_locked(mm, start, nr_pages, pages, vmas, |
932f4a630
|
1672 1673 1674 |
NULL, flags); } #endif /* CONFIG_FS_DAX || CONFIG_CMA */ |
447f3e45c
|
1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685 1686 1687 1688 1689 1690 1691 1692 |
static bool is_valid_gup_flags(unsigned int gup_flags) { /* * FOLL_PIN must only be set internally by the pin_user_pages*() APIs, * never directly by the caller, so enforce that with an assertion: */ if (WARN_ON_ONCE(gup_flags & FOLL_PIN)) return false; /* * FOLL_PIN is a prerequisite to FOLL_LONGTERM. Another way of saying * that is, FOLL_LONGTERM is a specific case, more restrictive case of * FOLL_PIN. */ if (WARN_ON_ONCE(gup_flags & FOLL_LONGTERM)) return false; return true; } |
22bf29b67
|
1693 |
#ifdef CONFIG_MMU |
64019a2e4
|
1694 |
static long __get_user_pages_remote(struct mm_struct *mm, |
22bf29b67
|
1695 1696 1697 1698 1699 1700 1701 1702 1703 1704 1705 1706 1707 1708 1709 1710 1711 1712 |
unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { /* * Parts of FOLL_LONGTERM behavior are incompatible with * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on * vmas. However, this only comes up if locked is set, and there are * callers that do request FOLL_LONGTERM, but do not set locked. So, * allow what we can. */ if (gup_flags & FOLL_LONGTERM) { if (WARN_ON_ONCE(locked)) return -EINVAL; /* * This will check the vmas (even if our vmas arg is NULL) * and return -ENOTSUPP if DAX isn't allowed in this case: */ |
64019a2e4
|
1713 |
return __gup_longterm_locked(mm, start, nr_pages, pages, |
22bf29b67
|
1714 1715 1716 |
vmas, gup_flags | FOLL_TOUCH | FOLL_REMOTE); } |
64019a2e4
|
1717 |
return __get_user_pages_locked(mm, start, nr_pages, pages, vmas, |
22bf29b67
|
1718 1719 1720 |
locked, gup_flags | FOLL_TOUCH | FOLL_REMOTE); } |
adc8cb406
|
1721 |
/** |
c4237f8b1
|
1722 |
* get_user_pages_remote() - pin user pages in memory |
c4237f8b1
|
1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740 1741 1742 1743 1744 1745 |
* @mm: mm_struct of target mm * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. Or NULL, if caller * only intends to ensure the pages are faulted in. * @vmas: array of pointers to vmas corresponding to each page. * Or NULL if the caller does not require them. * @locked: pointer to lock flag indicating whether lock is held and * subsequently whether VM_FAULT_RETRY functionality can be * utilised. Lock must initially be held. * * Returns either number of pages pinned (which may be less than the * number requested), or an error. Details about the return value: * * -- If nr_pages is 0, returns 0. * -- If nr_pages is >0, but no pages were pinned, returns -errno. * -- If nr_pages is >0, and some pages were pinned, returns the number of * pages pinned. Again, this may be less than nr_pages. * * The caller is responsible for releasing returned @pages, via put_page(). * |
c1e8d7c6a
|
1746 |
* @vmas are valid only as long as mmap_lock is held. |
c4237f8b1
|
1747 |
* |
c1e8d7c6a
|
1748 |
* Must be called with mmap_lock held for read or write. |
c4237f8b1
|
1749 |
* |
adc8cb406
|
1750 1751 |
* get_user_pages_remote walks a process's page tables and takes a reference * to each struct page that each user address corresponds to at a given |
c4237f8b1
|
1752 1753 1754 1755 |
* instant. That is, it takes the page that would be accessed if a user * thread accesses the given user virtual address at that instant. * * This does not guarantee that the page exists in the user mappings when |
adc8cb406
|
1756 |
* get_user_pages_remote returns, and there may even be a completely different |
c4237f8b1
|
1757 1758 1759 1760 1761 1762 1763 1764 1765 1766 1767 |
* page there in some cases (eg. if mmapped pagecache has been invalidated * and subsequently re faulted). However it does guarantee that the page * won't be freed completely. And mostly callers simply care that the page * contains data that was valid *at some point in time*. Typically, an IO * or similar operation cannot guarantee anything stronger anyway because * locks can't be held over the syscall boundary. * * If gup_flags & FOLL_WRITE == 0, the page must not be written to. If the page * is written to, set_page_dirty (or set_page_dirty_lock, as appropriate) must * be called after the page is finished with, and before put_page is called. * |
adc8cb406
|
1768 1769 1770 1771 1772 |
* get_user_pages_remote is typically used for fewer-copy IO operations, * to get a handle on the memory by some means other than accesses * via the user virtual addresses. The pages may be submitted for * DMA to devices or accessed via their kernel linear mapping (via the * kmap APIs). Care should be taken to use the correct cache flushing APIs. |
c4237f8b1
|
1773 1774 1775 |
* * See also get_user_pages_fast, for performance critical applications. * |
adc8cb406
|
1776 |
* get_user_pages_remote should be phased out in favor of |
c4237f8b1
|
1777 |
* get_user_pages_locked|unlocked or get_user_pages_fast. Nothing |
adc8cb406
|
1778 |
* should use get_user_pages_remote because it cannot pass |
c4237f8b1
|
1779 1780 |
* FAULT_FLAG_ALLOW_RETRY to handle_mm_fault. */ |
64019a2e4
|
1781 |
long get_user_pages_remote(struct mm_struct *mm, |
c4237f8b1
|
1782 1783 1784 1785 |
unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { |
447f3e45c
|
1786 |
if (!is_valid_gup_flags(gup_flags)) |
eddb1c228
|
1787 |
return -EINVAL; |
64019a2e4
|
1788 |
return __get_user_pages_remote(mm, start, nr_pages, gup_flags, |
22bf29b67
|
1789 |
pages, vmas, locked); |
c4237f8b1
|
1790 1791 |
} EXPORT_SYMBOL(get_user_pages_remote); |
eddb1c228
|
1792 |
#else /* CONFIG_MMU */ |
64019a2e4
|
1793 |
long get_user_pages_remote(struct mm_struct *mm, |
eddb1c228
|
1794 1795 1796 1797 1798 1799 |
unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { return 0; } |
3faa52c03
|
1800 |
|
64019a2e4
|
1801 |
static long __get_user_pages_remote(struct mm_struct *mm, |
3faa52c03
|
1802 1803 1804 1805 1806 1807 |
unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { return 0; } |
eddb1c228
|
1808 |
#endif /* !CONFIG_MMU */ |
adc8cb406
|
1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 |
/** * get_user_pages() - pin user pages in memory * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. Or NULL, if caller * only intends to ensure the pages are faulted in. * @vmas: array of pointers to vmas corresponding to each page. * Or NULL if the caller does not require them. * |
64019a2e4
|
1820 1821 1822 1823 |
* This is the same as get_user_pages_remote(), just with a less-flexible * calling convention where we assume that the mm being operated on belongs to * the current task, and doesn't allow passing of a locked parameter. We also * obviously don't pass FOLL_REMOTE in here. |
932f4a630
|
1824 1825 1826 1827 1828 |
*/ long get_user_pages(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas) { |
447f3e45c
|
1829 |
if (!is_valid_gup_flags(gup_flags)) |
eddb1c228
|
1830 |
return -EINVAL; |
64019a2e4
|
1831 |
return __gup_longterm_locked(current->mm, start, nr_pages, |
932f4a630
|
1832 1833 1834 |
pages, vmas, gup_flags | FOLL_TOUCH); } EXPORT_SYMBOL(get_user_pages); |
2bb6d2837
|
1835 |
|
adc8cb406
|
1836 |
/** |
d3649f68b
|
1837 |
* get_user_pages_locked() is suitable to replace the form: |
acc3c8d15
|
1838 |
* |
3e4e28c5a
|
1839 |
* mmap_read_lock(mm); |
d3649f68b
|
1840 |
* do_something() |
64019a2e4
|
1841 |
* get_user_pages(mm, ..., pages, NULL); |
3e4e28c5a
|
1842 |
* mmap_read_unlock(mm); |
acc3c8d15
|
1843 |
* |
d3649f68b
|
1844 |
* to: |
acc3c8d15
|
1845 |
* |
d3649f68b
|
1846 |
* int locked = 1; |
3e4e28c5a
|
1847 |
* mmap_read_lock(mm); |
d3649f68b
|
1848 |
* do_something() |
64019a2e4
|
1849 |
* get_user_pages_locked(mm, ..., pages, &locked); |
d3649f68b
|
1850 |
* if (locked) |
3e4e28c5a
|
1851 |
* mmap_read_unlock(mm); |
adc8cb406
|
1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 |
* * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. Or NULL, if caller * only intends to ensure the pages are faulted in. * @locked: pointer to lock flag indicating whether lock is held and * subsequently whether VM_FAULT_RETRY functionality can be * utilised. Lock must initially be held. * * We can leverage the VM_FAULT_RETRY functionality in the page fault * paths better by using either get_user_pages_locked() or * get_user_pages_unlocked(). * |
acc3c8d15
|
1867 |
*/ |
d3649f68b
|
1868 1869 1870 |
long get_user_pages_locked(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, int *locked) |
acc3c8d15
|
1871 |
{ |
acc3c8d15
|
1872 |
/* |
d3649f68b
|
1873 1874 1875 1876 |
* FIXME: Current FOLL_LONGTERM behavior is incompatible with * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on * vmas. As there are no users of this flag in this call we simply * disallow this option for now. |
acc3c8d15
|
1877 |
*/ |
d3649f68b
|
1878 1879 |
if (WARN_ON_ONCE(gup_flags & FOLL_LONGTERM)) return -EINVAL; |
420c2091b
|
1880 1881 1882 1883 1884 1885 |
/* * FOLL_PIN must only be set internally by the pin_user_pages*() APIs, * never directly by the caller, so enforce that: */ if (WARN_ON_ONCE(gup_flags & FOLL_PIN)) return -EINVAL; |
acc3c8d15
|
1886 |
|
64019a2e4
|
1887 |
return __get_user_pages_locked(current->mm, start, nr_pages, |
d3649f68b
|
1888 1889 |
pages, NULL, locked, gup_flags | FOLL_TOUCH); |
acc3c8d15
|
1890 |
} |
d3649f68b
|
1891 |
EXPORT_SYMBOL(get_user_pages_locked); |
acc3c8d15
|
1892 1893 |
/* |
d3649f68b
|
1894 |
* get_user_pages_unlocked() is suitable to replace the form: |
acc3c8d15
|
1895 |
* |
3e4e28c5a
|
1896 |
* mmap_read_lock(mm); |
64019a2e4
|
1897 |
* get_user_pages(mm, ..., pages, NULL); |
3e4e28c5a
|
1898 |
* mmap_read_unlock(mm); |
d3649f68b
|
1899 1900 1901 |
* * with: * |
64019a2e4
|
1902 |
* get_user_pages_unlocked(mm, ..., pages); |
d3649f68b
|
1903 1904 1905 1906 |
* * It is functionally equivalent to get_user_pages_fast so * get_user_pages_fast should be used instead if specific gup_flags * (e.g. FOLL_FORCE) are not required. |
acc3c8d15
|
1907 |
*/ |
d3649f68b
|
1908 1909 |
long get_user_pages_unlocked(unsigned long start, unsigned long nr_pages, struct page **pages, unsigned int gup_flags) |
acc3c8d15
|
1910 1911 |
{ struct mm_struct *mm = current->mm; |
d3649f68b
|
1912 1913 |
int locked = 1; long ret; |
acc3c8d15
|
1914 |
|
d3649f68b
|
1915 1916 1917 1918 1919 1920 1921 1922 |
/* * FIXME: Current FOLL_LONGTERM behavior is incompatible with * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on * vmas. As there are no users of this flag in this call we simply * disallow this option for now. */ if (WARN_ON_ONCE(gup_flags & FOLL_LONGTERM)) return -EINVAL; |
acc3c8d15
|
1923 |
|
d8ed45c5d
|
1924 |
mmap_read_lock(mm); |
64019a2e4
|
1925 |
ret = __get_user_pages_locked(mm, start, nr_pages, pages, NULL, |
d3649f68b
|
1926 |
&locked, gup_flags | FOLL_TOUCH); |
acc3c8d15
|
1927 |
if (locked) |
d8ed45c5d
|
1928 |
mmap_read_unlock(mm); |
d3649f68b
|
1929 |
return ret; |
4bbd4c776
|
1930 |
} |
d3649f68b
|
1931 |
EXPORT_SYMBOL(get_user_pages_unlocked); |
2667f50e8
|
1932 1933 |
/* |
67a929e09
|
1934 |
* Fast GUP |
2667f50e8
|
1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 |
* * get_user_pages_fast attempts to pin user pages by walking the page * tables directly and avoids taking locks. Thus the walker needs to be * protected from page table pages being freed from under it, and should * block any THP splits. * * One way to achieve this is to have the walker disable interrupts, and * rely on IPIs from the TLB flushing code blocking before the page table * pages are freed. This is unsuitable for architectures that do not need * to broadcast an IPI when invalidating TLBs. * * Another way to achieve this is to batch up page table containing pages * belonging to more than one mm_user, then rcu_sched a callback to free those * pages. Disabling interrupts will allow the fast_gup walker to both block * the rcu_sched callback, and an IPI that we broadcast for splitting THPs * (which is a relatively rare event). The code below adopts this strategy. * * Before activating this code, please be aware that the following assumptions * are currently made: * |
ff2e6d725
|
1955 |
* *) Either MMU_GATHER_RCU_TABLE_FREE is enabled, and tlb_remove_table() is used to |
e585513b7
|
1956 |
* free pages containing page tables or TLB flushing requires IPI broadcast. |
2667f50e8
|
1957 |
* |
2667f50e8
|
1958 1959 1960 1961 1962 1963 1964 1965 |
* *) ptes can be read atomically by the architecture. * * *) access_ok is sufficient to validate userspace address ranges. * * The last two assumptions can be relaxed by the addition of helper functions. * * This code is based heavily on the PowerPC implementation by Nick Piggin. */ |
67a929e09
|
1966 |
#ifdef CONFIG_HAVE_FAST_GUP |
39656e83d
|
1967 |
#ifdef CONFIG_GUP_GET_PTE_LOW_HIGH |
3faa52c03
|
1968 |
|
39656e83d
|
1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 |
/* * WARNING: only to be used in the get_user_pages_fast() implementation. * * With get_user_pages_fast(), we walk down the pagetables without taking any * locks. For this we would like to load the pointers atomically, but sometimes * that is not possible (e.g. without expensive cmpxchg8b on x86_32 PAE). What * we do have is the guarantee that a PTE will only either go from not present * to present, or present to not present or both -- it will not switch to a * completely different present page without a TLB flush in between; something * that we are blocking by holding interrupts off. * * Setting ptes from not present to present goes: * * ptep->pte_high = h; * smp_wmb(); * ptep->pte_low = l; * * And present to not present goes: * * ptep->pte_low = 0; * smp_wmb(); * ptep->pte_high = 0; * * We must ensure here that the load of pte_low sees 'l' IFF pte_high sees 'h'. * We load pte_high *after* loading pte_low, which ensures we don't see an older * value of pte_high. *Then* we recheck pte_low, which ensures that we haven't * picked up a changed pte high. We might have gotten rubbish values from * pte_low and pte_high, but we are guaranteed that pte_low will not have the * present bit set *unless* it is 'l'. Because get_user_pages_fast() only * operates on present ptes we're safe. */ static inline pte_t gup_get_pte(pte_t *ptep) { pte_t pte; |
2667f50e8
|
2003 |
|
39656e83d
|
2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 |
do { pte.pte_low = ptep->pte_low; smp_rmb(); pte.pte_high = ptep->pte_high; smp_rmb(); } while (unlikely(pte.pte_low != ptep->pte_low)); return pte; } #else /* CONFIG_GUP_GET_PTE_LOW_HIGH */ |
0005d20b2
|
2014 |
/* |
39656e83d
|
2015 |
* We require that the PTE can be read atomically. |
0005d20b2
|
2016 2017 2018 |
*/ static inline pte_t gup_get_pte(pte_t *ptep) { |
481e980a7
|
2019 |
return ptep_get(ptep); |
0005d20b2
|
2020 |
} |
39656e83d
|
2021 |
#endif /* CONFIG_GUP_GET_PTE_LOW_HIGH */ |
0005d20b2
|
2022 |
|
790c73690
|
2023 |
static void __maybe_unused undo_dev_pagemap(int *nr, int nr_start, |
3b78d8347
|
2024 |
unsigned int flags, |
790c73690
|
2025 |
struct page **pages) |
b59f65fa0
|
2026 2027 2028 2029 2030 |
{ while ((*nr) - nr_start) { struct page *page = pages[--(*nr)]; ClearPageReferenced(page); |
3faa52c03
|
2031 2032 2033 2034 |
if (flags & FOLL_PIN) unpin_user_page(page); else put_page(page); |
b59f65fa0
|
2035 2036 |
} } |
3010a5ea6
|
2037 |
#ifdef CONFIG_ARCH_HAS_PTE_SPECIAL |
2667f50e8
|
2038 |
static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end, |
b798bec47
|
2039 |
unsigned int flags, struct page **pages, int *nr) |
2667f50e8
|
2040 |
{ |
b59f65fa0
|
2041 2042 |
struct dev_pagemap *pgmap = NULL; int nr_start = *nr, ret = 0; |
2667f50e8
|
2043 |
pte_t *ptep, *ptem; |
2667f50e8
|
2044 2045 2046 |
ptem = ptep = pte_offset_map(&pmd, addr); do { |
0005d20b2
|
2047 |
pte_t pte = gup_get_pte(ptep); |
7aef4172c
|
2048 |
struct page *head, *page; |
2667f50e8
|
2049 2050 2051 |
/* * Similar to the PMD case below, NUMA hinting must take slow |
8a0516ed8
|
2052 |
* path using the pte_protnone check. |
2667f50e8
|
2053 |
*/ |
e7884f8ea
|
2054 2055 |
if (pte_protnone(pte)) goto pte_unmap; |
b798bec47
|
2056 |
if (!pte_access_permitted(pte, flags & FOLL_WRITE)) |
e7884f8ea
|
2057 |
goto pte_unmap; |
b59f65fa0
|
2058 |
if (pte_devmap(pte)) { |
7af75561e
|
2059 2060 |
if (unlikely(flags & FOLL_LONGTERM)) goto pte_unmap; |
b59f65fa0
|
2061 2062 |
pgmap = get_dev_pagemap(pte_pfn(pte), pgmap); if (unlikely(!pgmap)) { |
3b78d8347
|
2063 |
undo_dev_pagemap(nr, nr_start, flags, pages); |
b59f65fa0
|
2064 2065 2066 |
goto pte_unmap; } } else if (pte_special(pte)) |
2667f50e8
|
2067 2068 2069 2070 |
goto pte_unmap; VM_BUG_ON(!pfn_valid(pte_pfn(pte))); page = pte_page(pte); |
3faa52c03
|
2071 |
head = try_grab_compound_head(page, 1, flags); |
8fde12ca7
|
2072 |
if (!head) |
2667f50e8
|
2073 2074 2075 |
goto pte_unmap; if (unlikely(pte_val(pte) != pte_val(*ptep))) { |
3faa52c03
|
2076 |
put_compound_head(head, 1, flags); |
2667f50e8
|
2077 2078 |
goto pte_unmap; } |
7aef4172c
|
2079 |
VM_BUG_ON_PAGE(compound_head(page) != head, page); |
e93480537
|
2080 |
|
f28d43636
|
2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 |
/* * We need to make the page accessible if and only if we are * going to access its content (the FOLL_PIN case). Please * see Documentation/core-api/pin_user_pages.rst for * details. */ if (flags & FOLL_PIN) { ret = arch_make_page_accessible(page); if (ret) { unpin_user_page(page); goto pte_unmap; } } |
e93480537
|
2094 |
SetPageReferenced(page); |
2667f50e8
|
2095 2096 2097 2098 2099 2100 2101 2102 |
pages[*nr] = page; (*nr)++; } while (ptep++, addr += PAGE_SIZE, addr != end); ret = 1; pte_unmap: |
832d7aa05
|
2103 2104 |
if (pgmap) put_dev_pagemap(pgmap); |
2667f50e8
|
2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 |
pte_unmap(ptem); return ret; } #else /* * If we can't determine whether or not a pte is special, then fail immediately * for ptes. Note, we can still pin HugeTLB and THP as these are guaranteed not * to be special. * * For a futex to be placed on a THP tail page, get_futex_key requires a |
dadbb612f
|
2116 |
* get_user_pages_fast_only implementation that can pin pages. Thus it's still |
2667f50e8
|
2117 2118 2119 |
* useful to have gup_huge_pmd even if we can't operate on ptes. */ static int gup_pte_range(pmd_t pmd, unsigned long addr, unsigned long end, |
b798bec47
|
2120 |
unsigned int flags, struct page **pages, int *nr) |
2667f50e8
|
2121 2122 2123 |
{ return 0; } |
3010a5ea6
|
2124 |
#endif /* CONFIG_ARCH_HAS_PTE_SPECIAL */ |
2667f50e8
|
2125 |
|
175967318
|
2126 |
#if defined(CONFIG_ARCH_HAS_PTE_DEVMAP) && defined(CONFIG_TRANSPARENT_HUGEPAGE) |
b59f65fa0
|
2127 |
static int __gup_device_huge(unsigned long pfn, unsigned long addr, |
86dfbed49
|
2128 2129 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
b59f65fa0
|
2130 2131 2132 2133 2134 2135 2136 2137 2138 |
{ int nr_start = *nr; struct dev_pagemap *pgmap = NULL; do { struct page *page = pfn_to_page(pfn); pgmap = get_dev_pagemap(pfn, pgmap); if (unlikely(!pgmap)) { |
3b78d8347
|
2139 |
undo_dev_pagemap(nr, nr_start, flags, pages); |
b59f65fa0
|
2140 2141 2142 2143 |
return 0; } SetPageReferenced(page); pages[*nr] = page; |
3faa52c03
|
2144 2145 2146 2147 |
if (unlikely(!try_grab_page(page, flags))) { undo_dev_pagemap(nr, nr_start, flags, pages); return 0; } |
b59f65fa0
|
2148 2149 2150 |
(*nr)++; pfn++; } while (addr += PAGE_SIZE, addr != end); |
832d7aa05
|
2151 2152 2153 |
if (pgmap) put_dev_pagemap(pgmap); |
b59f65fa0
|
2154 2155 |
return 1; } |
a9b6de77b
|
2156 |
static int __gup_device_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, |
86dfbed49
|
2157 2158 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
b59f65fa0
|
2159 2160 |
{ unsigned long fault_pfn; |
a9b6de77b
|
2161 2162 2163 |
int nr_start = *nr; fault_pfn = pmd_pfn(orig) + ((addr & ~PMD_MASK) >> PAGE_SHIFT); |
86dfbed49
|
2164 |
if (!__gup_device_huge(fault_pfn, addr, end, flags, pages, nr)) |
a9b6de77b
|
2165 |
return 0; |
b59f65fa0
|
2166 |
|
a9b6de77b
|
2167 |
if (unlikely(pmd_val(orig) != pmd_val(*pmdp))) { |
3b78d8347
|
2168 |
undo_dev_pagemap(nr, nr_start, flags, pages); |
a9b6de77b
|
2169 2170 2171 |
return 0; } return 1; |
b59f65fa0
|
2172 |
} |
a9b6de77b
|
2173 |
static int __gup_device_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr, |
86dfbed49
|
2174 2175 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
b59f65fa0
|
2176 2177 |
{ unsigned long fault_pfn; |
a9b6de77b
|
2178 2179 2180 |
int nr_start = *nr; fault_pfn = pud_pfn(orig) + ((addr & ~PUD_MASK) >> PAGE_SHIFT); |
86dfbed49
|
2181 |
if (!__gup_device_huge(fault_pfn, addr, end, flags, pages, nr)) |
a9b6de77b
|
2182 |
return 0; |
b59f65fa0
|
2183 |
|
a9b6de77b
|
2184 |
if (unlikely(pud_val(orig) != pud_val(*pudp))) { |
3b78d8347
|
2185 |
undo_dev_pagemap(nr, nr_start, flags, pages); |
a9b6de77b
|
2186 2187 2188 |
return 0; } return 1; |
b59f65fa0
|
2189 2190 |
} #else |
a9b6de77b
|
2191 |
static int __gup_device_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, |
86dfbed49
|
2192 2193 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
b59f65fa0
|
2194 2195 2196 2197 |
{ BUILD_BUG(); return 0; } |
a9b6de77b
|
2198 |
static int __gup_device_huge_pud(pud_t pud, pud_t *pudp, unsigned long addr, |
86dfbed49
|
2199 2200 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
b59f65fa0
|
2201 2202 2203 2204 2205 |
{ BUILD_BUG(); return 0; } #endif |
a43e98208
|
2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 |
static int record_subpages(struct page *page, unsigned long addr, unsigned long end, struct page **pages) { int nr; for (nr = 0; addr != end; addr += PAGE_SIZE) pages[nr++] = page++; return nr; } |
cbd34da7d
|
2216 2217 2218 2219 2220 2221 2222 2223 2224 |
#ifdef CONFIG_ARCH_HAS_HUGEPD static unsigned long hugepte_addr_end(unsigned long addr, unsigned long end, unsigned long sz) { unsigned long __boundary = (addr + sz) & ~(sz-1); return (__boundary - 1 < end - 1) ? __boundary : end; } static int gup_hugepte(pte_t *ptep, unsigned long sz, unsigned long addr, |
0cd22afdc
|
2225 2226 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
cbd34da7d
|
2227 2228 2229 2230 2231 2232 2233 2234 2235 |
{ unsigned long pte_end; struct page *head, *page; pte_t pte; int refs; pte_end = (addr + sz) & ~(sz-1); if (pte_end < end) end = pte_end; |
55ca22633
|
2236 |
pte = huge_ptep_get(ptep); |
cbd34da7d
|
2237 |
|
0cd22afdc
|
2238 |
if (!pte_access_permitted(pte, flags & FOLL_WRITE)) |
cbd34da7d
|
2239 2240 2241 2242 |
return 0; /* hugepages are never "special" */ VM_BUG_ON(!pfn_valid(pte_pfn(pte))); |
cbd34da7d
|
2243 |
head = pte_page(pte); |
cbd34da7d
|
2244 |
page = head + ((addr & (sz-1)) >> PAGE_SHIFT); |
a43e98208
|
2245 |
refs = record_subpages(page, addr, end, pages + *nr); |
cbd34da7d
|
2246 |
|
3faa52c03
|
2247 |
head = try_grab_compound_head(head, refs, flags); |
a43e98208
|
2248 |
if (!head) |
cbd34da7d
|
2249 |
return 0; |
cbd34da7d
|
2250 2251 |
if (unlikely(pte_val(pte) != pte_val(*ptep))) { |
3b78d8347
|
2252 |
put_compound_head(head, refs, flags); |
cbd34da7d
|
2253 2254 |
return 0; } |
a43e98208
|
2255 |
*nr += refs; |
520b4a449
|
2256 |
SetPageReferenced(head); |
cbd34da7d
|
2257 2258 2259 2260 |
return 1; } static int gup_huge_pd(hugepd_t hugepd, unsigned long addr, |
0cd22afdc
|
2261 |
unsigned int pdshift, unsigned long end, unsigned int flags, |
cbd34da7d
|
2262 2263 2264 2265 2266 2267 2268 2269 2270 |
struct page **pages, int *nr) { pte_t *ptep; unsigned long sz = 1UL << hugepd_shift(hugepd); unsigned long next; ptep = hugepte_offset(hugepd, addr, pdshift); do { next = hugepte_addr_end(addr, end, sz); |
0cd22afdc
|
2271 |
if (!gup_hugepte(ptep, sz, addr, end, flags, pages, nr)) |
cbd34da7d
|
2272 2273 2274 2275 2276 2277 2278 |
return 0; } while (ptep++, addr = next, addr != end); return 1; } #else static inline int gup_huge_pd(hugepd_t hugepd, unsigned long addr, |
0cd22afdc
|
2279 |
unsigned int pdshift, unsigned long end, unsigned int flags, |
cbd34da7d
|
2280 2281 2282 2283 2284 |
struct page **pages, int *nr) { return 0; } #endif /* CONFIG_ARCH_HAS_HUGEPD */ |
2667f50e8
|
2285 |
static int gup_huge_pmd(pmd_t orig, pmd_t *pmdp, unsigned long addr, |
0cd22afdc
|
2286 2287 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
2667f50e8
|
2288 |
{ |
ddc58f27f
|
2289 |
struct page *head, *page; |
2667f50e8
|
2290 |
int refs; |
b798bec47
|
2291 |
if (!pmd_access_permitted(orig, flags & FOLL_WRITE)) |
2667f50e8
|
2292 |
return 0; |
7af75561e
|
2293 2294 2295 |
if (pmd_devmap(orig)) { if (unlikely(flags & FOLL_LONGTERM)) return 0; |
86dfbed49
|
2296 2297 |
return __gup_device_huge_pmd(orig, pmdp, addr, end, flags, pages, nr); |
7af75561e
|
2298 |
} |
b59f65fa0
|
2299 |
|
d63206ee3
|
2300 |
page = pmd_page(orig) + ((addr & ~PMD_MASK) >> PAGE_SHIFT); |
a43e98208
|
2301 |
refs = record_subpages(page, addr, end, pages + *nr); |
2667f50e8
|
2302 |
|
3faa52c03
|
2303 |
head = try_grab_compound_head(pmd_page(orig), refs, flags); |
a43e98208
|
2304 |
if (!head) |
2667f50e8
|
2305 |
return 0; |
2667f50e8
|
2306 2307 |
if (unlikely(pmd_val(orig) != pmd_val(*pmdp))) { |
3b78d8347
|
2308 |
put_compound_head(head, refs, flags); |
2667f50e8
|
2309 2310 |
return 0; } |
a43e98208
|
2311 |
*nr += refs; |
e93480537
|
2312 |
SetPageReferenced(head); |
2667f50e8
|
2313 2314 2315 2316 |
return 1; } static int gup_huge_pud(pud_t orig, pud_t *pudp, unsigned long addr, |
86dfbed49
|
2317 2318 |
unsigned long end, unsigned int flags, struct page **pages, int *nr) |
2667f50e8
|
2319 |
{ |
ddc58f27f
|
2320 |
struct page *head, *page; |
2667f50e8
|
2321 |
int refs; |
b798bec47
|
2322 |
if (!pud_access_permitted(orig, flags & FOLL_WRITE)) |
2667f50e8
|
2323 |
return 0; |
7af75561e
|
2324 2325 2326 |
if (pud_devmap(orig)) { if (unlikely(flags & FOLL_LONGTERM)) return 0; |
86dfbed49
|
2327 2328 |
return __gup_device_huge_pud(orig, pudp, addr, end, flags, pages, nr); |
7af75561e
|
2329 |
} |
b59f65fa0
|
2330 |
|
d63206ee3
|
2331 |
page = pud_page(orig) + ((addr & ~PUD_MASK) >> PAGE_SHIFT); |
a43e98208
|
2332 |
refs = record_subpages(page, addr, end, pages + *nr); |
2667f50e8
|
2333 |
|
3faa52c03
|
2334 |
head = try_grab_compound_head(pud_page(orig), refs, flags); |
a43e98208
|
2335 |
if (!head) |
2667f50e8
|
2336 |
return 0; |
2667f50e8
|
2337 2338 |
if (unlikely(pud_val(orig) != pud_val(*pudp))) { |
3b78d8347
|
2339 |
put_compound_head(head, refs, flags); |
2667f50e8
|
2340 2341 |
return 0; } |
a43e98208
|
2342 |
*nr += refs; |
e93480537
|
2343 |
SetPageReferenced(head); |
2667f50e8
|
2344 2345 |
return 1; } |
f30c59e92
|
2346 |
static int gup_huge_pgd(pgd_t orig, pgd_t *pgdp, unsigned long addr, |
b798bec47
|
2347 |
unsigned long end, unsigned int flags, |
f30c59e92
|
2348 2349 2350 |
struct page **pages, int *nr) { int refs; |
ddc58f27f
|
2351 |
struct page *head, *page; |
f30c59e92
|
2352 |
|
b798bec47
|
2353 |
if (!pgd_access_permitted(orig, flags & FOLL_WRITE)) |
f30c59e92
|
2354 |
return 0; |
b59f65fa0
|
2355 |
BUILD_BUG_ON(pgd_devmap(orig)); |
a43e98208
|
2356 |
|
d63206ee3
|
2357 |
page = pgd_page(orig) + ((addr & ~PGDIR_MASK) >> PAGE_SHIFT); |
a43e98208
|
2358 |
refs = record_subpages(page, addr, end, pages + *nr); |
f30c59e92
|
2359 |
|
3faa52c03
|
2360 |
head = try_grab_compound_head(pgd_page(orig), refs, flags); |
a43e98208
|
2361 |
if (!head) |
f30c59e92
|
2362 |
return 0; |
f30c59e92
|
2363 2364 |
if (unlikely(pgd_val(orig) != pgd_val(*pgdp))) { |
3b78d8347
|
2365 |
put_compound_head(head, refs, flags); |
f30c59e92
|
2366 2367 |
return 0; } |
a43e98208
|
2368 |
*nr += refs; |
e93480537
|
2369 |
SetPageReferenced(head); |
f30c59e92
|
2370 2371 |
return 1; } |
d3f7b1bb2
|
2372 |
static int gup_pmd_range(pud_t *pudp, pud_t pud, unsigned long addr, unsigned long end, |
b798bec47
|
2373 |
unsigned int flags, struct page **pages, int *nr) |
2667f50e8
|
2374 2375 2376 |
{ unsigned long next; pmd_t *pmdp; |
d3f7b1bb2
|
2377 |
pmdp = pmd_offset_lockless(pudp, pud, addr); |
2667f50e8
|
2378 |
do { |
38c5ce936
|
2379 |
pmd_t pmd = READ_ONCE(*pmdp); |
2667f50e8
|
2380 2381 |
next = pmd_addr_end(addr, end); |
84c3fc4e9
|
2382 |
if (!pmd_present(pmd)) |
2667f50e8
|
2383 |
return 0; |
414fd080d
|
2384 2385 |
if (unlikely(pmd_trans_huge(pmd) || pmd_huge(pmd) || pmd_devmap(pmd))) { |
2667f50e8
|
2386 2387 2388 2389 2390 |
/* * NUMA hinting faults need to be handled in the GUP * slowpath for accounting purposes and so that they * can be serialised against THP migration. */ |
8a0516ed8
|
2391 |
if (pmd_protnone(pmd)) |
2667f50e8
|
2392 |
return 0; |
b798bec47
|
2393 |
if (!gup_huge_pmd(pmd, pmdp, addr, next, flags, |
2667f50e8
|
2394 2395 |
pages, nr)) return 0; |
f30c59e92
|
2396 2397 2398 2399 2400 2401 |
} else if (unlikely(is_hugepd(__hugepd(pmd_val(pmd))))) { /* * architecture have different format for hugetlbfs * pmd format and THP pmd format */ if (!gup_huge_pd(__hugepd(pmd_val(pmd)), addr, |
b798bec47
|
2402 |
PMD_SHIFT, next, flags, pages, nr)) |
f30c59e92
|
2403 |
return 0; |
b798bec47
|
2404 |
} else if (!gup_pte_range(pmd, addr, next, flags, pages, nr)) |
2923117b7
|
2405 |
return 0; |
2667f50e8
|
2406 2407 2408 2409 |
} while (pmdp++, addr = next, addr != end); return 1; } |
d3f7b1bb2
|
2410 |
static int gup_pud_range(p4d_t *p4dp, p4d_t p4d, unsigned long addr, unsigned long end, |
b798bec47
|
2411 |
unsigned int flags, struct page **pages, int *nr) |
2667f50e8
|
2412 2413 2414 |
{ unsigned long next; pud_t *pudp; |
d3f7b1bb2
|
2415 |
pudp = pud_offset_lockless(p4dp, p4d, addr); |
2667f50e8
|
2416 |
do { |
e37c69827
|
2417 |
pud_t pud = READ_ONCE(*pudp); |
2667f50e8
|
2418 2419 |
next = pud_addr_end(addr, end); |
15494520b
|
2420 |
if (unlikely(!pud_present(pud))) |
2667f50e8
|
2421 |
return 0; |
f30c59e92
|
2422 |
if (unlikely(pud_huge(pud))) { |
b798bec47
|
2423 |
if (!gup_huge_pud(pud, pudp, addr, next, flags, |
f30c59e92
|
2424 2425 2426 2427 |
pages, nr)) return 0; } else if (unlikely(is_hugepd(__hugepd(pud_val(pud))))) { if (!gup_huge_pd(__hugepd(pud_val(pud)), addr, |
b798bec47
|
2428 |
PUD_SHIFT, next, flags, pages, nr)) |
2667f50e8
|
2429 |
return 0; |
d3f7b1bb2
|
2430 |
} else if (!gup_pmd_range(pudp, pud, addr, next, flags, pages, nr)) |
2667f50e8
|
2431 2432 2433 2434 2435 |
return 0; } while (pudp++, addr = next, addr != end); return 1; } |
d3f7b1bb2
|
2436 |
static int gup_p4d_range(pgd_t *pgdp, pgd_t pgd, unsigned long addr, unsigned long end, |
b798bec47
|
2437 |
unsigned int flags, struct page **pages, int *nr) |
c2febafc6
|
2438 2439 2440 |
{ unsigned long next; p4d_t *p4dp; |
d3f7b1bb2
|
2441 |
p4dp = p4d_offset_lockless(pgdp, pgd, addr); |
c2febafc6
|
2442 2443 2444 2445 2446 2447 2448 2449 2450 |
do { p4d_t p4d = READ_ONCE(*p4dp); next = p4d_addr_end(addr, end); if (p4d_none(p4d)) return 0; BUILD_BUG_ON(p4d_huge(p4d)); if (unlikely(is_hugepd(__hugepd(p4d_val(p4d))))) { if (!gup_huge_pd(__hugepd(p4d_val(p4d)), addr, |
b798bec47
|
2451 |
P4D_SHIFT, next, flags, pages, nr)) |
c2febafc6
|
2452 |
return 0; |
d3f7b1bb2
|
2453 |
} else if (!gup_pud_range(p4dp, p4d, addr, next, flags, pages, nr)) |
c2febafc6
|
2454 2455 2456 2457 2458 |
return 0; } while (p4dp++, addr = next, addr != end); return 1; } |
5b65c4677
|
2459 |
static void gup_pgd_range(unsigned long addr, unsigned long end, |
b798bec47
|
2460 |
unsigned int flags, struct page **pages, int *nr) |
5b65c4677
|
2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 |
{ unsigned long next; pgd_t *pgdp; pgdp = pgd_offset(current->mm, addr); do { pgd_t pgd = READ_ONCE(*pgdp); next = pgd_addr_end(addr, end); if (pgd_none(pgd)) return; if (unlikely(pgd_huge(pgd))) { |
b798bec47
|
2473 |
if (!gup_huge_pgd(pgd, pgdp, addr, next, flags, |
5b65c4677
|
2474 2475 2476 2477 |
pages, nr)) return; } else if (unlikely(is_hugepd(__hugepd(pgd_val(pgd))))) { if (!gup_huge_pd(__hugepd(pgd_val(pgd)), addr, |
b798bec47
|
2478 |
PGDIR_SHIFT, next, flags, pages, nr)) |
5b65c4677
|
2479 |
return; |
d3f7b1bb2
|
2480 |
} else if (!gup_p4d_range(pgdp, pgd, addr, next, flags, pages, nr)) |
5b65c4677
|
2481 2482 2483 |
return; } while (pgdp++, addr = next, addr != end); } |
050a9adc6
|
2484 2485 2486 2487 2488 2489 |
#else static inline void gup_pgd_range(unsigned long addr, unsigned long end, unsigned int flags, struct page **pages, int *nr) { } #endif /* CONFIG_HAVE_FAST_GUP */ |
5b65c4677
|
2490 2491 2492 |
#ifndef gup_fast_permitted /* |
dadbb612f
|
2493 |
* Check if it's allowed to use get_user_pages_fast_only() for the range, or |
5b65c4677
|
2494 2495 |
* we need to fall back to the slow version: */ |
26f4c3280
|
2496 |
static bool gup_fast_permitted(unsigned long start, unsigned long end) |
5b65c4677
|
2497 |
{ |
26f4c3280
|
2498 |
return true; |
5b65c4677
|
2499 2500 |
} #endif |
7af75561e
|
2501 2502 2503 2504 2505 2506 2507 2508 2509 2510 |
static int __gup_longterm_unlocked(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { int ret; /* * FIXME: FOLL_LONGTERM does not work with * get_user_pages_unlocked() (see comments in that function) */ if (gup_flags & FOLL_LONGTERM) { |
d8ed45c5d
|
2511 |
mmap_read_lock(current->mm); |
64019a2e4
|
2512 |
ret = __gup_longterm_locked(current->mm, |
7af75561e
|
2513 2514 |
start, nr_pages, pages, NULL, gup_flags); |
d8ed45c5d
|
2515 |
mmap_read_unlock(current->mm); |
7af75561e
|
2516 2517 2518 2519 2520 2521 2522 |
} else { ret = get_user_pages_unlocked(start, nr_pages, pages, gup_flags); } return ret; } |
bcb0f647c
|
2523 2524 2525 2526 2527 2528 2529 |
static unsigned long lockless_pages_from_mm(unsigned long start, unsigned long end, unsigned int gup_flags, struct page **pages) { unsigned long flags; int nr_pinned = 0; |
537946556
|
2530 |
unsigned seq; |
bcb0f647c
|
2531 2532 2533 2534 |
if (!IS_ENABLED(CONFIG_HAVE_FAST_GUP) || !gup_fast_permitted(start, end)) return 0; |
537946556
|
2535 2536 2537 2538 2539 |
if (gup_flags & FOLL_PIN) { seq = raw_read_seqcount(¤t->mm->write_protect_seq); if (seq & 1) return 0; } |
bcb0f647c
|
2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 |
/* * Disable interrupts. The nested form is used, in order to allow full, * general purpose use of this routine. * * With interrupts disabled, we block page table pages from being freed * from under us. See struct mmu_table_batch comments in * include/asm-generic/tlb.h for more details. * * We do not adopt an rcu_read_lock() here as we also want to block IPIs * that come from THPs splitting. */ local_irq_save(flags); gup_pgd_range(start, end, gup_flags, pages, &nr_pinned); local_irq_restore(flags); |
537946556
|
2554 2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 |
/* * When pinning pages for DMA there could be a concurrent write protect * from fork() via copy_page_range(), in this case always fail fast GUP. */ if (gup_flags & FOLL_PIN) { if (read_seqcount_retry(¤t->mm->write_protect_seq, seq)) { unpin_user_pages(pages, nr_pinned); return 0; } } |
bcb0f647c
|
2565 2566 2567 2568 2569 |
return nr_pinned; } static int internal_get_user_pages_fast(unsigned long start, unsigned long nr_pages, |
eddb1c228
|
2570 2571 |
unsigned int gup_flags, struct page **pages) |
2667f50e8
|
2572 |
{ |
bcb0f647c
|
2573 2574 2575 |
unsigned long len, end; unsigned long nr_pinned; int ret; |
2667f50e8
|
2576 |
|
f4000fdf4
|
2577 |
if (WARN_ON_ONCE(gup_flags & ~(FOLL_WRITE | FOLL_LONGTERM | |
376a34efa
|
2578 2579 |
FOLL_FORCE | FOLL_PIN | FOLL_GET | FOLL_FAST_ONLY))) |
817be129e
|
2580 |
return -EINVAL; |
008cfe441
|
2581 2582 |
if (gup_flags & FOLL_PIN) atomic_set(¤t->mm->has_pinned, 1); |
f81cd178e
|
2583 |
if (!(gup_flags & FOLL_FAST_ONLY)) |
da1c55f1b
|
2584 |
might_lock_read(¤t->mm->mmap_lock); |
f81cd178e
|
2585 |
|
f455c8548
|
2586 |
start = untagged_addr(start) & PAGE_MASK; |
bcb0f647c
|
2587 2588 |
len = nr_pages << PAGE_SHIFT; if (check_add_overflow(start, len, &end)) |
c61611f70
|
2589 |
return 0; |
96d4f267e
|
2590 |
if (unlikely(!access_ok((void __user *)start, len))) |
c61611f70
|
2591 |
return -EFAULT; |
73e10a618
|
2592 |
|
bcb0f647c
|
2593 2594 2595 |
nr_pinned = lockless_pages_from_mm(start, end, gup_flags, pages); if (nr_pinned == nr_pages || gup_flags & FOLL_FAST_ONLY) return nr_pinned; |
2667f50e8
|
2596 |
|
bcb0f647c
|
2597 2598 2599 2600 2601 2602 2603 2604 2605 2606 2607 2608 2609 |
/* Slow path: try to get the remaining pages with get_user_pages */ start += nr_pinned << PAGE_SHIFT; pages += nr_pinned; ret = __gup_longterm_unlocked(start, nr_pages - nr_pinned, gup_flags, pages); if (ret < 0) { /* * The caller has to unpin the pages we already pinned so * returning -errno is not an option */ if (nr_pinned) return nr_pinned; return ret; |
2667f50e8
|
2610 |
} |
bcb0f647c
|
2611 |
return ret + nr_pinned; |
2667f50e8
|
2612 |
} |
bcb0f647c
|
2613 |
|
dadbb612f
|
2614 2615 2616 2617 2618 2619 2620 2621 |
/** * get_user_pages_fast_only() - pin user pages in memory * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * |
9e1f0580d
|
2622 2623 2624 2625 2626 2627 2628 2629 2630 2631 2632 2633 |
* Like get_user_pages_fast() except it's IRQ-safe in that it won't fall back to * the regular GUP. * Note a difference with get_user_pages_fast: this always returns the * number of pages pinned, 0 if no pages were pinned. * * If the architecture does not support this function, simply return with no * pages pinned. * * Careful, careful! COW breaking can go either way, so a non-write * access can get ambiguous page results. If you call this function without * 'write' set, you'd better be sure that you're ok with that ambiguity. */ |
dadbb612f
|
2634 2635 |
int get_user_pages_fast_only(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) |
9e1f0580d
|
2636 |
{ |
376a34efa
|
2637 |
int nr_pinned; |
9e1f0580d
|
2638 2639 2640 |
/* * Internally (within mm/gup.c), gup fast variants must set FOLL_GET, * because gup fast is always a "pin with a +1 page refcount" request. |
376a34efa
|
2641 2642 2643 |
* * FOLL_FAST_ONLY is required in order to match the API description of * this routine: no fall back to regular ("slow") GUP. |
9e1f0580d
|
2644 |
*/ |
dadbb612f
|
2645 |
gup_flags |= FOLL_GET | FOLL_FAST_ONLY; |
9e1f0580d
|
2646 |
|
376a34efa
|
2647 2648 |
nr_pinned = internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); |
9e1f0580d
|
2649 2650 |
/* |
376a34efa
|
2651 2652 2653 2654 |
* As specified in the API description above, this routine is not * allowed to return negative values. However, the common core * routine internal_get_user_pages_fast() *can* return -errno. * Therefore, correct for that here: |
9e1f0580d
|
2655 |
*/ |
376a34efa
|
2656 2657 |
if (nr_pinned < 0) nr_pinned = 0; |
9e1f0580d
|
2658 2659 2660 |
return nr_pinned; } |
dadbb612f
|
2661 |
EXPORT_SYMBOL_GPL(get_user_pages_fast_only); |
9e1f0580d
|
2662 |
|
eddb1c228
|
2663 2664 |
/** * get_user_pages_fast() - pin user pages in memory |
3faa52c03
|
2665 2666 2667 2668 2669 |
* @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. |
eddb1c228
|
2670 |
* |
c1e8d7c6a
|
2671 |
* Attempt to pin user pages in memory without taking mm->mmap_lock. |
eddb1c228
|
2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 |
* If not successful, it will fall back to taking the lock and * calling get_user_pages(). * * Returns number of pages pinned. This may be fewer than the number requested. * If nr_pages is 0 or negative, returns 0. If no pages were pinned, returns * -errno. */ int get_user_pages_fast(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { |
447f3e45c
|
2682 |
if (!is_valid_gup_flags(gup_flags)) |
eddb1c228
|
2683 |
return -EINVAL; |
94202f126
|
2684 2685 2686 2687 2688 2689 2690 |
/* * The caller may or may not have explicitly set FOLL_GET; either way is * OK. However, internally (within mm/gup.c), gup fast variants must set * FOLL_GET, because gup fast is always a "pin with a +1 page refcount" * request. */ gup_flags |= FOLL_GET; |
eddb1c228
|
2691 2692 |
return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); } |
050a9adc6
|
2693 |
EXPORT_SYMBOL_GPL(get_user_pages_fast); |
eddb1c228
|
2694 2695 2696 2697 |
/** * pin_user_pages_fast() - pin user pages in memory without taking locks * |
3faa52c03
|
2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 |
* @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying pin behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. * * Nearly the same as get_user_pages_fast(), except that FOLL_PIN is set. See * get_user_pages_fast() for documentation on the function arguments, because * the arguments here are identical. * * FOLL_PIN means that the pages must be released via unpin_user_page(). Please |
72ef5e52b
|
2709 |
* see Documentation/core-api/pin_user_pages.rst for further details. |
eddb1c228
|
2710 2711 2712 2713 |
*/ int pin_user_pages_fast(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { |
3faa52c03
|
2714 2715 2716 2717 2718 2719 |
/* FOLL_GET and FOLL_PIN are mutually exclusive. */ if (WARN_ON_ONCE(gup_flags & FOLL_GET)) return -EINVAL; gup_flags |= FOLL_PIN; return internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); |
eddb1c228
|
2720 2721 |
} EXPORT_SYMBOL_GPL(pin_user_pages_fast); |
104acc327
|
2722 |
/* |
dadbb612f
|
2723 2724 |
* This is the FOLL_PIN equivalent of get_user_pages_fast_only(). Behavior * is the same, except that this one sets FOLL_PIN instead of FOLL_GET. |
104acc327
|
2725 2726 2727 2728 2729 2730 2731 2732 2733 2734 2735 2736 2737 2738 2739 2740 2741 2742 2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 |
* * The API rules are the same, too: no negative values may be returned. */ int pin_user_pages_fast_only(unsigned long start, int nr_pages, unsigned int gup_flags, struct page **pages) { int nr_pinned; /* * FOLL_GET and FOLL_PIN are mutually exclusive. Note that the API * rules require returning 0, rather than -errno: */ if (WARN_ON_ONCE(gup_flags & FOLL_GET)) return 0; /* * FOLL_FAST_ONLY is required in order to match the API description of * this routine: no fall back to regular ("slow") GUP. */ gup_flags |= (FOLL_PIN | FOLL_FAST_ONLY); nr_pinned = internal_get_user_pages_fast(start, nr_pages, gup_flags, pages); /* * This routine is not allowed to return negative values. However, * internal_get_user_pages_fast() *can* return -errno. Therefore, * correct for that here: */ if (nr_pinned < 0) nr_pinned = 0; return nr_pinned; } EXPORT_SYMBOL_GPL(pin_user_pages_fast_only); |
eddb1c228
|
2757 |
/** |
64019a2e4
|
2758 |
* pin_user_pages_remote() - pin pages of a remote process |
eddb1c228
|
2759 |
* |
3faa52c03
|
2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 |
* @mm: mm_struct of target mm * @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. Or NULL, if caller * only intends to ensure the pages are faulted in. * @vmas: array of pointers to vmas corresponding to each page. * Or NULL if the caller does not require them. * @locked: pointer to lock flag indicating whether lock is held and * subsequently whether VM_FAULT_RETRY functionality can be * utilised. Lock must initially be held. * * Nearly the same as get_user_pages_remote(), except that FOLL_PIN is set. See * get_user_pages_remote() for documentation on the function arguments, because * the arguments here are identical. * * FOLL_PIN means that the pages must be released via unpin_user_page(). Please |
72ef5e52b
|
2778 |
* see Documentation/core-api/pin_user_pages.rst for details. |
eddb1c228
|
2779 |
*/ |
64019a2e4
|
2780 |
long pin_user_pages_remote(struct mm_struct *mm, |
eddb1c228
|
2781 2782 2783 2784 |
unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas, int *locked) { |
3faa52c03
|
2785 2786 2787 2788 2789 |
/* FOLL_GET and FOLL_PIN are mutually exclusive. */ if (WARN_ON_ONCE(gup_flags & FOLL_GET)) return -EINVAL; gup_flags |= FOLL_PIN; |
64019a2e4
|
2790 |
return __get_user_pages_remote(mm, start, nr_pages, gup_flags, |
3faa52c03
|
2791 |
pages, vmas, locked); |
eddb1c228
|
2792 2793 2794 2795 2796 2797 |
} EXPORT_SYMBOL(pin_user_pages_remote); /** * pin_user_pages() - pin user pages in memory for use by other devices * |
3faa52c03
|
2798 2799 2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 |
* @start: starting user address * @nr_pages: number of pages from start to pin * @gup_flags: flags modifying lookup behaviour * @pages: array that receives pointers to the pages pinned. * Should be at least nr_pages long. Or NULL, if caller * only intends to ensure the pages are faulted in. * @vmas: array of pointers to vmas corresponding to each page. * Or NULL if the caller does not require them. * * Nearly the same as get_user_pages(), except that FOLL_TOUCH is not set, and * FOLL_PIN is set. * * FOLL_PIN means that the pages must be released via unpin_user_page(). Please |
72ef5e52b
|
2811 |
* see Documentation/core-api/pin_user_pages.rst for details. |
eddb1c228
|
2812 2813 2814 2815 2816 |
*/ long pin_user_pages(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, struct vm_area_struct **vmas) { |
3faa52c03
|
2817 2818 2819 2820 2821 |
/* FOLL_GET and FOLL_PIN are mutually exclusive. */ if (WARN_ON_ONCE(gup_flags & FOLL_GET)) return -EINVAL; gup_flags |= FOLL_PIN; |
64019a2e4
|
2822 |
return __gup_longterm_locked(current->mm, start, nr_pages, |
3faa52c03
|
2823 |
pages, vmas, gup_flags); |
eddb1c228
|
2824 2825 |
} EXPORT_SYMBOL(pin_user_pages); |
914290233
|
2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 |
/* * pin_user_pages_unlocked() is the FOLL_PIN variant of * get_user_pages_unlocked(). Behavior is the same, except that this one sets * FOLL_PIN and rejects FOLL_GET. */ long pin_user_pages_unlocked(unsigned long start, unsigned long nr_pages, struct page **pages, unsigned int gup_flags) { /* FOLL_GET and FOLL_PIN are mutually exclusive. */ if (WARN_ON_ONCE(gup_flags & FOLL_GET)) return -EINVAL; gup_flags |= FOLL_PIN; return get_user_pages_unlocked(start, nr_pages, pages, gup_flags); } EXPORT_SYMBOL(pin_user_pages_unlocked); |
420c2091b
|
2843 2844 2845 2846 2847 2848 2849 2850 2851 2852 2853 2854 2855 2856 2857 2858 2859 2860 2861 2862 2863 2864 2865 2866 |
/* * pin_user_pages_locked() is the FOLL_PIN variant of get_user_pages_locked(). * Behavior is the same, except that this one sets FOLL_PIN and rejects * FOLL_GET. */ long pin_user_pages_locked(unsigned long start, unsigned long nr_pages, unsigned int gup_flags, struct page **pages, int *locked) { /* * FIXME: Current FOLL_LONGTERM behavior is incompatible with * FAULT_FLAG_ALLOW_RETRY because of the FS DAX check requirement on * vmas. As there are no users of this flag in this call we simply * disallow this option for now. */ if (WARN_ON_ONCE(gup_flags & FOLL_LONGTERM)) return -EINVAL; /* FOLL_GET and FOLL_PIN are mutually exclusive. */ if (WARN_ON_ONCE(gup_flags & FOLL_GET)) return -EINVAL; gup_flags |= FOLL_PIN; |
64019a2e4
|
2867 |
return __get_user_pages_locked(current->mm, start, nr_pages, |
420c2091b
|
2868 2869 2870 2871 |
pages, NULL, locked, gup_flags | FOLL_TOUCH); } EXPORT_SYMBOL(pin_user_pages_locked); |