Blame view
mm/migrate.c
80.8 KB
b24413180 License cleanup: ... |
1 |
// SPDX-License-Identifier: GPL-2.0 |
b20a35035 [PATCH] page migr... |
2 |
/* |
14e0f9bcc mm: correct a cou... |
3 |
* Memory Migration functionality - linux/mm/migrate.c |
b20a35035 [PATCH] page migr... |
4 5 6 7 8 9 10 11 12 |
* * Copyright (C) 2006 Silicon Graphics, Inc., Christoph Lameter * * Page migration was first developed in the context of the memory hotplug * project. The main authors of the migration code are: * * IWAMOTO Toshihiro <iwamoto@valinux.co.jp> * Hirokazu Takahashi <taka@valinux.co.jp> * Dave Hansen <haveblue@us.ibm.com> |
cde535359 Christoph has moved |
13 |
* Christoph Lameter |
b20a35035 [PATCH] page migr... |
14 15 16 |
*/ #include <linux/migrate.h> |
b95f1b31b mm: Map most file... |
17 |
#include <linux/export.h> |
b20a35035 [PATCH] page migr... |
18 |
#include <linux/swap.h> |
0697212a4 [PATCH] Swapless ... |
19 |
#include <linux/swapops.h> |
b20a35035 [PATCH] page migr... |
20 |
#include <linux/pagemap.h> |
e23ca00bf [PATCH] Some page... |
21 |
#include <linux/buffer_head.h> |
b20a35035 [PATCH] page migr... |
22 |
#include <linux/mm_inline.h> |
b488893a3 pid namespaces: c... |
23 |
#include <linux/nsproxy.h> |
b20a35035 [PATCH] page migr... |
24 |
#include <linux/pagevec.h> |
e9995ef97 ksm: rmap_walk to... |
25 |
#include <linux/ksm.h> |
b20a35035 [PATCH] page migr... |
26 27 28 29 |
#include <linux/rmap.h> #include <linux/topology.h> #include <linux/cpu.h> #include <linux/cpuset.h> |
04e62a29b [PATCH] More page... |
30 |
#include <linux/writeback.h> |
742755a1d [PATCH] page migr... |
31 32 |
#include <linux/mempolicy.h> #include <linux/vmalloc.h> |
86c3a7645 [PATCH] SELinux: ... |
33 |
#include <linux/security.h> |
42cb14b11 mm: migrate dirty... |
34 |
#include <linux/backing-dev.h> |
bda807d44 mm: migrate: supp... |
35 |
#include <linux/compaction.h> |
4f5ca2657 mm/migrate.c shou... |
36 |
#include <linux/syscalls.h> |
7addf4438 mm: add kernel_mo... |
37 |
#include <linux/compat.h> |
290408d4a hugetlb: hugepage... |
38 |
#include <linux/hugetlb.h> |
8e6ac7fab hugetlb/cgroup: m... |
39 |
#include <linux/hugetlb_cgroup.h> |
5a0e3ad6a include cleanup: ... |
40 |
#include <linux/gfp.h> |
a520110e4 mm: split out a n... |
41 |
#include <linux/pagewalk.h> |
df6ad6983 mm/device-public-... |
42 |
#include <linux/pfn_t.h> |
a5430dda8 mm/migrate: suppo... |
43 |
#include <linux/memremap.h> |
8315ada7f mm/migrate: allow... |
44 |
#include <linux/userfaultfd_k.h> |
bf6bddf19 mm: introduce com... |
45 |
#include <linux/balloon_compaction.h> |
f714f4f20 mm: numa: call MM... |
46 |
#include <linux/mmu_notifier.h> |
33c3fc71c mm: introduce idl... |
47 |
#include <linux/page_idle.h> |
d435edca9 mm, page_owner: c... |
48 |
#include <linux/page_owner.h> |
6e84f3152 sched/headers: Pr... |
49 |
#include <linux/sched/mm.h> |
197e7e521 Sanitize 'move_pa... |
50 |
#include <linux/ptrace.h> |
34290e2c6 mm/migrate: add s... |
51 |
#include <linux/oom.h> |
b20a35035 [PATCH] page migr... |
52 |
|
0d1836c36 mm/migrate.c: fix... |
53 |
#include <asm/tlbflush.h> |
7b2a2d4a1 mm: migrate: Add ... |
54 55 |
#define CREATE_TRACE_POINTS #include <trace/events/migrate.h> |
b20a35035 [PATCH] page migr... |
56 |
#include "internal.h" |
b20a35035 [PATCH] page migr... |
57 |
/* |
742755a1d [PATCH] page migr... |
58 |
* migrate_prep() needs to be called before we start compiling a list of pages |
748446bb6 mm: compaction: m... |
59 60 |
* to be migrated using isolate_lru_page(). If scheduling work on other CPUs is * undesirable, use migrate_prep_local() |
b20a35035 [PATCH] page migr... |
61 62 63 |
*/ int migrate_prep(void) { |
b20a35035 [PATCH] page migr... |
64 65 66 67 68 69 70 71 72 73 |
/* * Clear the LRU lists so pages can be isolated. * Note that pages may be moved off the LRU after we have * drained them. Those pages will fail to migrate like other * pages that may be busy. */ lru_add_drain_all(); return 0; } |
748446bb6 mm: compaction: m... |
74 75 76 77 78 79 80 |
/* Do the necessary work of migrate_prep but not if it involves other CPUs */ int migrate_prep_local(void) { lru_add_drain(); return 0; } |
9e5bcd610 mm/migration: mak... |
81 |
int isolate_movable_page(struct page *page, isolate_mode_t mode) |
bda807d44 mm: migrate: supp... |
82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 |
{ struct address_space *mapping; /* * Avoid burning cycles with pages that are yet under __free_pages(), * or just got freed under us. * * In case we 'win' a race for a movable page being freed under us and * raise its refcount preventing __free_pages() from doing its job * the put_page() at the end of this block will take care of * release this page, thus avoiding a nasty leakage. */ if (unlikely(!get_page_unless_zero(page))) goto out; /* * Check PageMovable before holding a PG_lock because page's owner * assumes anybody doesn't touch PG_lock of newly allocated page |
8bb4e7a2e mm: fix some typo... |
100 |
* so unconditionally grabbing the lock ruins page's owner side. |
bda807d44 mm: migrate: supp... |
101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 |
*/ if (unlikely(!__PageMovable(page))) goto out_putpage; /* * As movable pages are not isolated from LRU lists, concurrent * compaction threads can race against page migration functions * as well as race against the releasing a page. * * In order to avoid having an already isolated movable page * being (wrongly) re-isolated while it is under migration, * or to avoid attempting to isolate pages being released, * lets be sure we have the page lock * before proceeding with the movable page isolation steps. */ if (unlikely(!trylock_page(page))) goto out_putpage; if (!PageMovable(page) || PageIsolated(page)) goto out_no_isolated; mapping = page_mapping(page); VM_BUG_ON_PAGE(!mapping, page); if (!mapping->a_ops->isolate_page(page, mode)) goto out_no_isolated; /* Driver shouldn't use PG_isolated bit of page->flags */ WARN_ON_ONCE(PageIsolated(page)); __SetPageIsolated(page); unlock_page(page); |
9e5bcd610 mm/migration: mak... |
131 |
return 0; |
bda807d44 mm: migrate: supp... |
132 133 134 135 136 137 |
out_no_isolated: unlock_page(page); out_putpage: put_page(page); out: |
9e5bcd610 mm/migration: mak... |
138 |
return -EBUSY; |
bda807d44 mm: migrate: supp... |
139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 |
} /* It should be called on page which is PG_movable */ void putback_movable_page(struct page *page) { struct address_space *mapping; VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(!PageMovable(page), page); VM_BUG_ON_PAGE(!PageIsolated(page), page); mapping = page_mapping(page); mapping->a_ops->putback_page(page); __ClearPageIsolated(page); } |
b20a35035 [PATCH] page migr... |
154 |
/* |
5733c7d11 mm: introduce put... |
155 156 157 |
* Put previously isolated pages back onto the appropriate lists * from where they were once taken off for compaction/migration. * |
59c82b70d mm/migrate: remov... |
158 159 160 |
* This function shall be used whenever the isolated pageset has been * built from lru, balloon, hugetlbfs page. See isolate_migratepages_range() * and isolate_huge_page(). |
5733c7d11 mm: introduce put... |
161 162 163 164 165 166 167 |
*/ void putback_movable_pages(struct list_head *l) { struct page *page; struct page *page2; list_for_each_entry_safe(page, page2, l, lru) { |
31caf665e mm: migrate: make... |
168 169 170 171 |
if (unlikely(PageHuge(page))) { putback_active_hugepage(page); continue; } |
5733c7d11 mm: introduce put... |
172 |
list_del(&page->lru); |
bda807d44 mm: migrate: supp... |
173 174 175 176 177 |
/* * We isolated non-lru movable page so here we can use * __PageMovable because LRU page's mapping cannot have * PAGE_MAPPING_MOVABLE. */ |
b1123ea6d mm: balloon: use ... |
178 |
if (unlikely(__PageMovable(page))) { |
bda807d44 mm: migrate: supp... |
179 180 181 182 183 184 185 186 187 |
VM_BUG_ON_PAGE(!PageIsolated(page), page); lock_page(page); if (PageMovable(page)) putback_movable_page(page); else __ClearPageIsolated(page); unlock_page(page); put_page(page); } else { |
e8db67eb0 mm: migrate: move... |
188 |
mod_node_page_state(page_pgdat(page), NR_ISOLATED_ANON + |
6c357848b mm: replace hpage... |
189 |
page_is_file_lru(page), -thp_nr_pages(page)); |
fc280fe87 mm: prevent NR_IS... |
190 |
putback_lru_page(page); |
bda807d44 mm: migrate: supp... |
191 |
} |
b20a35035 [PATCH] page migr... |
192 |
} |
b20a35035 [PATCH] page migr... |
193 |
} |
0697212a4 [PATCH] Swapless ... |
194 195 196 |
/* * Restore a potential migration pte to a working pte entry */ |
e4b822227 mm: make rmap_one... |
197 |
static bool remove_migration_pte(struct page *page, struct vm_area_struct *vma, |
e9995ef97 ksm: rmap_walk to... |
198 |
unsigned long addr, void *old) |
0697212a4 [PATCH] Swapless ... |
199 |
{ |
3fe87967c mm: convert remov... |
200 201 202 203 204 205 206 207 |
struct page_vma_mapped_walk pvmw = { .page = old, .vma = vma, .address = addr, .flags = PVMW_SYNC | PVMW_MIGRATION, }; struct page *new; pte_t pte; |
0697212a4 [PATCH] Swapless ... |
208 |
swp_entry_t entry; |
0697212a4 [PATCH] Swapless ... |
209 |
|
3fe87967c mm: convert remov... |
210 211 |
VM_BUG_ON_PAGE(PageTail(page), page); while (page_vma_mapped_walk(&pvmw)) { |
4b0ece6fa mm: migrate: fix ... |
212 213 214 215 216 |
if (PageKsm(page)) new = page; else new = page - pvmw.page->index + linear_page_index(vma, pvmw.address); |
0697212a4 [PATCH] Swapless ... |
217 |
|
616b83715 mm: thp: enable t... |
218 219 220 221 222 223 224 225 |
#ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION /* PMD-mapped THP migration entry */ if (!pvmw.pte) { VM_BUG_ON_PAGE(PageHuge(page) || !PageTransCompound(page), page); remove_migration_pmd(&pvmw, new); continue; } #endif |
3fe87967c mm: convert remov... |
226 227 228 229 |
get_page(new); pte = pte_mkold(mk_pte(new, READ_ONCE(vma->vm_page_prot))); if (pte_swp_soft_dirty(*pvmw.pte)) pte = pte_mksoft_dirty(pte); |
0697212a4 [PATCH] Swapless ... |
230 |
|
3fe87967c mm: convert remov... |
231 232 233 234 235 236 |
/* * Recheck VMA as permissions can change since migration started */ entry = pte_to_swp_entry(*pvmw.pte); if (is_write_migration_entry(entry)) pte = maybe_mkwrite(pte, vma); |
f45ec5ff1 userfaultfd: wp: ... |
237 238 |
else if (pte_swp_uffd_wp(*pvmw.pte)) pte = pte_mkuffd_wp(pte); |
d3cb8bf60 mm: migrate: Clos... |
239 |
|
6128763fc mm/migrate: remov... |
240 241 242 |
if (unlikely(is_device_private_page(new))) { entry = make_device_private_entry(new, pte_write(pte)); pte = swp_entry_to_pte(entry); |
3d321bf82 mm/migrate: prese... |
243 244 |
if (pte_swp_soft_dirty(*pvmw.pte)) pte = pte_swp_mksoft_dirty(pte); |
6128763fc mm/migrate: remov... |
245 246 |
if (pte_swp_uffd_wp(*pvmw.pte)) pte = pte_swp_mkuffd_wp(pte); |
d2b2c6dd2 mm/migrate.c: add... |
247 |
} |
a5430dda8 mm/migrate: suppo... |
248 |
|
3ef8fd7f7 Fix migration.c c... |
249 |
#ifdef CONFIG_HUGETLB_PAGE |
3fe87967c mm: convert remov... |
250 251 252 |
if (PageHuge(new)) { pte = pte_mkhuge(pte); pte = arch_make_huge_pte(pte, vma, new, 0); |
383321ab8 mm/hugetlb/migrat... |
253 |
set_huge_pte_at(vma->vm_mm, pvmw.address, pvmw.pte, pte); |
3fe87967c mm: convert remov... |
254 255 256 257 |
if (PageAnon(new)) hugepage_add_anon_rmap(new, vma, pvmw.address); else page_dup_rmap(new, true); |
383321ab8 mm/hugetlb/migrat... |
258 259 260 261 |
} else #endif { set_pte_at(vma->vm_mm, pvmw.address, pvmw.pte, pte); |
04e62a29b [PATCH] More page... |
262 |
|
383321ab8 mm/hugetlb/migrat... |
263 264 265 266 267 |
if (PageAnon(new)) page_add_anon_rmap(new, vma, pvmw.address, false); else page_add_file_rmap(new, false); } |
3fe87967c mm: convert remov... |
268 269 |
if (vma->vm_flags & VM_LOCKED && !PageTransCompound(new)) mlock_vma_page(new); |
e125fe405 mm, thp: fix mloc... |
270 271 |
if (PageTransHuge(page) && PageMlocked(page)) clear_page_mlock(page); |
3fe87967c mm: convert remov... |
272 273 274 |
/* No need to invalidate - it was non-present before */ update_mmu_cache(vma, pvmw.address, pvmw.pte); } |
51afb12ba mm: page migratio... |
275 |
|
e4b822227 mm: make rmap_one... |
276 |
return true; |
0697212a4 [PATCH] Swapless ... |
277 278 279 |
} /* |
04e62a29b [PATCH] More page... |
280 281 282 |
* Get rid of all migration entries and replace them by * references to the indicated page. */ |
e388466de mm: make remove_m... |
283 |
void remove_migration_ptes(struct page *old, struct page *new, bool locked) |
04e62a29b [PATCH] More page... |
284 |
{ |
051ac83ad mm/rmap: make rma... |
285 286 287 288 |
struct rmap_walk_control rwc = { .rmap_one = remove_migration_pte, .arg = old, }; |
e388466de mm: make remove_m... |
289 290 291 292 |
if (locked) rmap_walk_locked(new, &rwc); else rmap_walk(new, &rwc); |
04e62a29b [PATCH] More page... |
293 294 295 |
} /* |
0697212a4 [PATCH] Swapless ... |
296 297 298 |
* Something used the pte of a page under migration. We need to * get to the page and wait until migration is finished. * When we return from this function the fault will be retried. |
0697212a4 [PATCH] Swapless ... |
299 |
*/ |
e66f17ff7 mm/hugetlb: take ... |
300 |
void __migration_entry_wait(struct mm_struct *mm, pte_t *ptep, |
30dad3092 mm: migration: ad... |
301 |
spinlock_t *ptl) |
0697212a4 [PATCH] Swapless ... |
302 |
{ |
30dad3092 mm: migration: ad... |
303 |
pte_t pte; |
0697212a4 [PATCH] Swapless ... |
304 305 |
swp_entry_t entry; struct page *page; |
30dad3092 mm: migration: ad... |
306 |
spin_lock(ptl); |
0697212a4 [PATCH] Swapless ... |
307 308 309 310 311 312 313 314 315 |
pte = *ptep; if (!is_swap_pte(pte)) goto out; entry = pte_to_swp_entry(pte); if (!is_migration_entry(entry)) goto out; page = migration_entry_to_page(entry); |
e286781d5 mm: speculative p... |
316 |
/* |
89eb946a7 mm: Convert page ... |
317 |
* Once page cache replacement of page migration started, page_count |
9a1ea439b mm: put_and_wait_... |
318 319 |
* is zero; but we must not call put_and_wait_on_page_locked() without * a ref. Use get_page_unless_zero(), and just fault again if it fails. |
e286781d5 mm: speculative p... |
320 321 322 |
*/ if (!get_page_unless_zero(page)) goto out; |
0697212a4 [PATCH] Swapless ... |
323 |
pte_unmap_unlock(ptep, ptl); |
9a1ea439b mm: put_and_wait_... |
324 |
put_and_wait_on_page_locked(page); |
0697212a4 [PATCH] Swapless ... |
325 326 327 328 |
return; out: pte_unmap_unlock(ptep, ptl); } |
30dad3092 mm: migration: ad... |
329 330 331 332 333 334 335 |
void migration_entry_wait(struct mm_struct *mm, pmd_t *pmd, unsigned long address) { spinlock_t *ptl = pte_lockptr(mm, pmd); pte_t *ptep = pte_offset_map(pmd, address); __migration_entry_wait(mm, ptep, ptl); } |
cb900f412 mm, hugetlb: conv... |
336 337 |
void migration_entry_wait_huge(struct vm_area_struct *vma, struct mm_struct *mm, pte_t *pte) |
30dad3092 mm: migration: ad... |
338 |
{ |
cb900f412 mm, hugetlb: conv... |
339 |
spinlock_t *ptl = huge_pte_lockptr(hstate_vma(vma), mm, pte); |
30dad3092 mm: migration: ad... |
340 341 |
__migration_entry_wait(mm, pte, ptl); } |
616b83715 mm: thp: enable t... |
342 343 344 345 346 347 348 349 350 351 352 353 354 |
#ifdef CONFIG_ARCH_ENABLE_THP_MIGRATION void pmd_migration_entry_wait(struct mm_struct *mm, pmd_t *pmd) { spinlock_t *ptl; struct page *page; ptl = pmd_lock(mm, pmd); if (!is_pmd_migration_entry(*pmd)) goto unlock; page = migration_entry_to_page(pmd_to_swp_entry(*pmd)); if (!get_page_unless_zero(page)) goto unlock; spin_unlock(ptl); |
9a1ea439b mm: put_and_wait_... |
355 |
put_and_wait_on_page_locked(page); |
616b83715 mm: thp: enable t... |
356 357 358 359 360 |
return; unlock: spin_unlock(ptl); } #endif |
f900482da mm/migrate.c: cle... |
361 |
static int expected_page_refs(struct address_space *mapping, struct page *page) |
0b3901b38 mm: migration: fa... |
362 363 364 365 |
{ int expected_count = 1; /* |
f1f4f3ab5 mm/migrate: remov... |
366 |
* Device private pages have an extra refcount as they are |
0b3901b38 mm: migration: fa... |
367 368 369 |
* ZONE_DEVICE pages. */ expected_count += is_device_private_page(page); |
f900482da mm/migrate.c: cle... |
370 |
if (mapping) |
6c357848b mm: replace hpage... |
371 |
expected_count += thp_nr_pages(page) + page_has_private(page); |
0b3901b38 mm: migration: fa... |
372 373 374 |
return expected_count; } |
b20a35035 [PATCH] page migr... |
375 |
/* |
c3fcf8a5d [PATCH] page migr... |
376 |
* Replace the page in the mapping. |
5b5c7120e [PATCH] page migr... |
377 378 379 380 |
* * The number of remaining references must be: * 1 for anonymous pages without a mapping * 2 for pages with a mapping |
266cf658e FS-Cache: Recruit... |
381 |
* 3 for pages with a mapping and PagePrivate/PagePrivate2 set. |
b20a35035 [PATCH] page migr... |
382 |
*/ |
36bc08cc0 fs/aio: Add suppo... |
383 |
int migrate_page_move_mapping(struct address_space *mapping, |
371096949 mm: migrate: remo... |
384 |
struct page *newpage, struct page *page, int extra_count) |
b20a35035 [PATCH] page migr... |
385 |
{ |
89eb946a7 mm: Convert page ... |
386 |
XA_STATE(xas, &mapping->i_pages, page_index(page)); |
42cb14b11 mm: migrate dirty... |
387 388 |
struct zone *oldzone, *newzone; int dirty; |
f900482da mm/migrate.c: cle... |
389 |
int expected_count = expected_page_refs(mapping, page) + extra_count; |
8763cb45a mm/migrate: new m... |
390 |
|
6c5240ae7 [PATCH] Swapless ... |
391 |
if (!mapping) { |
0e8c7d0fd page migration: f... |
392 |
/* Anonymous page without mapping */ |
8e321fefb aio/migratepages:... |
393 |
if (page_count(page) != expected_count) |
6c5240ae7 [PATCH] Swapless ... |
394 |
return -EAGAIN; |
cf4b769ab mm: page migratio... |
395 396 |
/* No turning back from here */ |
cf4b769ab mm: page migratio... |
397 398 399 |
newpage->index = page->index; newpage->mapping = page->mapping; if (PageSwapBacked(page)) |
fa9949da5 mm: use __SetPage... |
400 |
__SetPageSwapBacked(newpage); |
cf4b769ab mm: page migratio... |
401 |
|
78bd52097 mm: adjust addres... |
402 |
return MIGRATEPAGE_SUCCESS; |
6c5240ae7 [PATCH] Swapless ... |
403 |
} |
42cb14b11 mm: migrate dirty... |
404 405 |
oldzone = page_zone(page); newzone = page_zone(newpage); |
89eb946a7 mm: Convert page ... |
406 |
xas_lock_irq(&xas); |
89eb946a7 mm: Convert page ... |
407 408 |
if (page_count(page) != expected_count || xas_load(&xas) != page) { xas_unlock_irq(&xas); |
e23ca00bf [PATCH] Some page... |
409 |
return -EAGAIN; |
b20a35035 [PATCH] page migr... |
410 |
} |
fe896d187 mm: introduce pag... |
411 |
if (!page_ref_freeze(page, expected_count)) { |
89eb946a7 mm: Convert page ... |
412 |
xas_unlock_irq(&xas); |
e286781d5 mm: speculative p... |
413 414 |
return -EAGAIN; } |
b20a35035 [PATCH] page migr... |
415 |
/* |
cf4b769ab mm: page migratio... |
416 417 |
* Now we know that no one else is looking at the page: * no turning back from here. |
b20a35035 [PATCH] page migr... |
418 |
*/ |
cf4b769ab mm: page migratio... |
419 420 |
newpage->index = page->index; newpage->mapping = page->mapping; |
6c357848b mm: replace hpage... |
421 |
page_ref_add(newpage, thp_nr_pages(page)); /* add cache reference */ |
6326fec11 mm: Use owner_pri... |
422 423 424 425 426 427 428 429 |
if (PageSwapBacked(page)) { __SetPageSwapBacked(newpage); if (PageSwapCache(page)) { SetPageSwapCache(newpage); set_page_private(newpage, page_private(page)); } } else { VM_BUG_ON_PAGE(PageSwapCache(page), page); |
b20a35035 [PATCH] page migr... |
430 |
} |
42cb14b11 mm: migrate dirty... |
431 432 433 434 435 436 |
/* Move dirty while page refs frozen and newpage not yet exposed */ dirty = PageDirty(page); if (dirty) { ClearPageDirty(page); SetPageDirty(newpage); } |
89eb946a7 mm: Convert page ... |
437 |
xas_store(&xas, newpage); |
e71769ae5 mm: enable thp mi... |
438 439 |
if (PageTransHuge(page)) { int i; |
e71769ae5 mm: enable thp mi... |
440 |
|
013567be1 mm: migrate: fix ... |
441 |
for (i = 1; i < HPAGE_PMD_NR; i++) { |
89eb946a7 mm: Convert page ... |
442 |
xas_next(&xas); |
4101196b1 mm: page cache: s... |
443 |
xas_store(&xas, newpage); |
e71769ae5 mm: enable thp mi... |
444 |
} |
e71769ae5 mm: enable thp mi... |
445 |
} |
7cf9c2c76 [PATCH] radix-tre... |
446 447 |
/* |
937a94c9d mm: migrate: one ... |
448 449 |
* Drop cache reference from old page by unfreezing * to one less reference. |
7cf9c2c76 [PATCH] radix-tre... |
450 451 |
* We know this isn't the last reference. */ |
6c357848b mm: replace hpage... |
452 |
page_ref_unfreeze(page, expected_count - thp_nr_pages(page)); |
7cf9c2c76 [PATCH] radix-tre... |
453 |
|
89eb946a7 mm: Convert page ... |
454 |
xas_unlock(&xas); |
42cb14b11 mm: migrate dirty... |
455 |
/* Leave irq disabled to prevent preemption while updating stats */ |
0e8c7d0fd page migration: f... |
456 457 458 459 460 461 462 |
/* * If moved to a different zone then also account * the page for that zone. Other VM counters will be * taken care of when we establish references to the * new page and drop references to the old page. * * Note that anonymous pages are accounted for |
4b9d0fab7 mm: rename NR_ANO... |
463 |
* via NR_FILE_PAGES and NR_ANON_MAPPED if they |
0e8c7d0fd page migration: f... |
464 465 |
* are mapped to swap space. */ |
42cb14b11 mm: migrate dirty... |
466 |
if (newzone != oldzone) { |
0d1c20722 mm: memcontrol: s... |
467 468 469 470 471 472 473 474 475 |
struct lruvec *old_lruvec, *new_lruvec; struct mem_cgroup *memcg; memcg = page_memcg(page); old_lruvec = mem_cgroup_lruvec(memcg, oldzone->zone_pgdat); new_lruvec = mem_cgroup_lruvec(memcg, newzone->zone_pgdat); __dec_lruvec_state(old_lruvec, NR_FILE_PAGES); __inc_lruvec_state(new_lruvec, NR_FILE_PAGES); |
42cb14b11 mm: migrate dirty... |
476 |
if (PageSwapBacked(page) && !PageSwapCache(page)) { |
0d1c20722 mm: memcontrol: s... |
477 478 |
__dec_lruvec_state(old_lruvec, NR_SHMEM); __inc_lruvec_state(new_lruvec, NR_SHMEM); |
42cb14b11 mm: migrate dirty... |
479 |
} |
f56753ac2 bdi: replace BDI_... |
480 |
if (dirty && mapping_can_writeback(mapping)) { |
11fb99898 mm: move most fil... |
481 |
__dec_node_state(oldzone->zone_pgdat, NR_FILE_DIRTY); |
5a1c84b40 mm: remove reclai... |
482 |
__dec_zone_state(oldzone, NR_ZONE_WRITE_PENDING); |
11fb99898 mm: move most fil... |
483 |
__inc_node_state(newzone->zone_pgdat, NR_FILE_DIRTY); |
5a1c84b40 mm: remove reclai... |
484 |
__inc_zone_state(newzone, NR_ZONE_WRITE_PENDING); |
42cb14b11 mm: migrate dirty... |
485 |
} |
4b02108ac mm: oom analysis:... |
486 |
} |
42cb14b11 mm: migrate dirty... |
487 |
local_irq_enable(); |
b20a35035 [PATCH] page migr... |
488 |
|
78bd52097 mm: adjust addres... |
489 |
return MIGRATEPAGE_SUCCESS; |
b20a35035 [PATCH] page migr... |
490 |
} |
1118dce77 mm: Export migrat... |
491 |
EXPORT_SYMBOL(migrate_page_move_mapping); |
b20a35035 [PATCH] page migr... |
492 493 |
/* |
290408d4a hugetlb: hugepage... |
494 495 496 497 498 499 |
* The expected number of remaining references is the same as that * of migrate_page_move_mapping(). */ int migrate_huge_page_move_mapping(struct address_space *mapping, struct page *newpage, struct page *page) { |
89eb946a7 mm: Convert page ... |
500 |
XA_STATE(xas, &mapping->i_pages, page_index(page)); |
290408d4a hugetlb: hugepage... |
501 |
int expected_count; |
290408d4a hugetlb: hugepage... |
502 |
|
89eb946a7 mm: Convert page ... |
503 |
xas_lock_irq(&xas); |
290408d4a hugetlb: hugepage... |
504 |
expected_count = 2 + page_has_private(page); |
89eb946a7 mm: Convert page ... |
505 506 |
if (page_count(page) != expected_count || xas_load(&xas) != page) { xas_unlock_irq(&xas); |
290408d4a hugetlb: hugepage... |
507 508 |
return -EAGAIN; } |
fe896d187 mm: introduce pag... |
509 |
if (!page_ref_freeze(page, expected_count)) { |
89eb946a7 mm: Convert page ... |
510 |
xas_unlock_irq(&xas); |
290408d4a hugetlb: hugepage... |
511 512 |
return -EAGAIN; } |
cf4b769ab mm: page migratio... |
513 514 |
newpage->index = page->index; newpage->mapping = page->mapping; |
6a93ca8fd mm: migrate: do n... |
515 |
|
290408d4a hugetlb: hugepage... |
516 |
get_page(newpage); |
89eb946a7 mm: Convert page ... |
517 |
xas_store(&xas, newpage); |
290408d4a hugetlb: hugepage... |
518 |
|
fe896d187 mm: introduce pag... |
519 |
page_ref_unfreeze(page, expected_count - 1); |
290408d4a hugetlb: hugepage... |
520 |
|
89eb946a7 mm: Convert page ... |
521 |
xas_unlock_irq(&xas); |
6a93ca8fd mm: migrate: do n... |
522 |
|
78bd52097 mm: adjust addres... |
523 |
return MIGRATEPAGE_SUCCESS; |
290408d4a hugetlb: hugepage... |
524 525 526 |
} /* |
30b0a105d mm: thp: give tra... |
527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 |
* Gigantic pages are so large that we do not guarantee that page++ pointer * arithmetic will work across the entire page. We need something more * specialized. */ static void __copy_gigantic_page(struct page *dst, struct page *src, int nr_pages) { int i; struct page *dst_base = dst; struct page *src_base = src; for (i = 0; i < nr_pages; ) { cond_resched(); copy_highpage(dst, src); i++; dst = mem_map_next(dst, dst_base, i); src = mem_map_next(src, src_base, i); } } static void copy_huge_page(struct page *dst, struct page *src) { int i; int nr_pages; if (PageHuge(src)) { /* hugetlbfs page */ struct hstate *h = page_hstate(src); nr_pages = pages_per_huge_page(h); if (unlikely(nr_pages > MAX_ORDER_NR_PAGES)) { __copy_gigantic_page(dst, src, nr_pages); return; } } else { /* thp page */ BUG_ON(!PageTransHuge(src)); |
6c357848b mm: replace hpage... |
565 |
nr_pages = thp_nr_pages(src); |
30b0a105d mm: thp: give tra... |
566 567 568 569 570 571 572 573 574 |
} for (i = 0; i < nr_pages; i++) { cond_resched(); copy_highpage(dst + i, src + i); } } /* |
b20a35035 [PATCH] page migr... |
575 576 |
* Copy the page to its new location */ |
2916ecc0f mm/migrate: new m... |
577 |
void migrate_page_states(struct page *newpage, struct page *page) |
b20a35035 [PATCH] page migr... |
578 |
{ |
7851a45cd mm: numa: Copy cp... |
579 |
int cpupid; |
b20a35035 [PATCH] page migr... |
580 581 582 583 584 585 |
if (PageError(page)) SetPageError(newpage); if (PageReferenced(page)) SetPageReferenced(newpage); if (PageUptodate(page)) SetPageUptodate(newpage); |
894bc3104 Unevictable LRU I... |
586 |
if (TestClearPageActive(page)) { |
309381fea mm: dump page whe... |
587 |
VM_BUG_ON_PAGE(PageUnevictable(page), page); |
b20a35035 [PATCH] page migr... |
588 |
SetPageActive(newpage); |
418b27ef5 mm: remove unevic... |
589 590 |
} else if (TestClearPageUnevictable(page)) SetPageUnevictable(newpage); |
1899ad18c mm: workingset: t... |
591 592 |
if (PageWorkingset(page)) SetPageWorkingset(newpage); |
b20a35035 [PATCH] page migr... |
593 594 595 596 |
if (PageChecked(page)) SetPageChecked(newpage); if (PageMappedToDisk(page)) SetPageMappedToDisk(newpage); |
42cb14b11 mm: migrate dirty... |
597 598 599 |
/* Move dirty on pages not done by migrate_page_move_mapping() */ if (PageDirty(page)) SetPageDirty(newpage); |
b20a35035 [PATCH] page migr... |
600 |
|
33c3fc71c mm: introduce idl... |
601 602 603 604 |
if (page_is_young(page)) set_page_young(newpage); if (page_is_idle(page)) set_page_idle(newpage); |
7851a45cd mm: numa: Copy cp... |
605 606 607 608 609 610 |
/* * Copy NUMA information to the new page, to prevent over-eager * future migrations of this same page. */ cpupid = page_cpupid_xchg_last(page, -1); page_cpupid_xchg_last(newpage, cpupid); |
e9995ef97 ksm: rmap_walk to... |
611 |
ksm_migrate_page(newpage, page); |
c8d6553b9 ksm: make KSM pag... |
612 613 614 615 |
/* * Please do not reorder this without considering how mm/ksm.c's * get_ksm_page() depends upon ksm_migrate_page() and PageSwapCache(). */ |
b3b3a99c5 mm/migrate: check... |
616 617 |
if (PageSwapCache(page)) ClearPageSwapCache(page); |
b20a35035 [PATCH] page migr... |
618 619 |
ClearPagePrivate(page); set_page_private(page, 0); |
b20a35035 [PATCH] page migr... |
620 621 622 623 624 625 626 |
/* * If any waiters have accumulated on the new page then * wake them up. */ if (PageWriteback(newpage)) end_page_writeback(newpage); |
d435edca9 mm, page_owner: c... |
627 |
|
6aeff241f mm/migrate.c: mig... |
628 629 630 631 632 633 634 |
/* * PG_readahead shares the same bit with PG_reclaim. The above * end_page_writeback() may clear PG_readahead mistakenly, so set the * bit after that. */ if (PageReadahead(page)) SetPageReadahead(newpage); |
d435edca9 mm, page_owner: c... |
635 |
copy_page_owner(page, newpage); |
74485cf2b mm: migrate: cons... |
636 |
|
a333e3e73 mm: migration of ... |
637 638 |
if (!PageHuge(page)) mem_cgroup_migrate(page, newpage); |
b20a35035 [PATCH] page migr... |
639 |
} |
2916ecc0f mm/migrate: new m... |
640 641 642 643 644 645 646 647 648 649 650 |
EXPORT_SYMBOL(migrate_page_states); void migrate_page_copy(struct page *newpage, struct page *page) { if (PageHuge(page) || PageTransHuge(page)) copy_huge_page(newpage, page); else copy_highpage(newpage, page); migrate_page_states(newpage, page); } |
1118dce77 mm: Export migrat... |
651 |
EXPORT_SYMBOL(migrate_page_copy); |
b20a35035 [PATCH] page migr... |
652 |
|
1d8b85ccf [PATCH] page migr... |
653 654 655 |
/************************************************************ * Migration functions ***********************************************************/ |
b20a35035 [PATCH] page migr... |
656 |
/* |
bda807d44 mm: migrate: supp... |
657 |
* Common logic to directly migrate a single LRU page suitable for |
266cf658e FS-Cache: Recruit... |
658 |
* pages that do not use PagePrivate/PagePrivate2. |
b20a35035 [PATCH] page migr... |
659 660 661 |
* * Pages are locked upon entry and exit. */ |
2d1db3b11 [PATCH] page migr... |
662 |
int migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
663 664 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
b20a35035 [PATCH] page migr... |
665 666 667 668 |
{ int rc; BUG_ON(PageWriteback(page)); /* Writeback must be complete */ |
371096949 mm: migrate: remo... |
669 |
rc = migrate_page_move_mapping(mapping, newpage, page, 0); |
b20a35035 [PATCH] page migr... |
670 |
|
78bd52097 mm: adjust addres... |
671 |
if (rc != MIGRATEPAGE_SUCCESS) |
b20a35035 [PATCH] page migr... |
672 |
return rc; |
2916ecc0f mm/migrate: new m... |
673 674 675 676 |
if (mode != MIGRATE_SYNC_NO_COPY) migrate_page_copy(newpage, page); else migrate_page_states(newpage, page); |
78bd52097 mm: adjust addres... |
677 |
return MIGRATEPAGE_SUCCESS; |
b20a35035 [PATCH] page migr... |
678 679 |
} EXPORT_SYMBOL(migrate_page); |
9361401eb [PATCH] BLOCK: Ma... |
680 |
#ifdef CONFIG_BLOCK |
84ade7c15 mm: migrate: move... |
681 682 683 684 685 686 687 688 689 |
/* Returns true if all buffers are successfully locked */ static bool buffer_migrate_lock_buffers(struct buffer_head *head, enum migrate_mode mode) { struct buffer_head *bh = head; /* Simple case, sync compaction */ if (mode != MIGRATE_ASYNC) { do { |
84ade7c15 mm: migrate: move... |
690 691 692 693 694 695 696 697 698 699 |
lock_buffer(bh); bh = bh->b_this_page; } while (bh != head); return true; } /* async case, we cannot block on lock_buffer so use trylock_buffer */ do { |
84ade7c15 mm: migrate: move... |
700 701 702 703 704 705 |
if (!trylock_buffer(bh)) { /* * We failed to lock the buffer and cannot stall in * async migration. Release the taken locks */ struct buffer_head *failed_bh = bh; |
84ade7c15 mm: migrate: move... |
706 707 708 |
bh = head; while (bh != failed_bh) { unlock_buffer(bh); |
84ade7c15 mm: migrate: move... |
709 710 711 712 713 714 715 716 717 |
bh = bh->b_this_page; } return false; } bh = bh->b_this_page; } while (bh != head); return true; } |
89cb0888c mm: migrate: prov... |
718 719 720 |
static int __buffer_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page, enum migrate_mode mode, bool check_refs) |
1d8b85ccf [PATCH] page migr... |
721 |
{ |
1d8b85ccf [PATCH] page migr... |
722 723 |
struct buffer_head *bh, *head; int rc; |
cc4f11e69 mm: migrate: lock... |
724 |
int expected_count; |
1d8b85ccf [PATCH] page migr... |
725 |
|
1d8b85ccf [PATCH] page migr... |
726 |
if (!page_has_buffers(page)) |
a6bc32b89 mm: compaction: i... |
727 |
return migrate_page(mapping, newpage, page, mode); |
1d8b85ccf [PATCH] page migr... |
728 |
|
cc4f11e69 mm: migrate: lock... |
729 |
/* Check whether page does not have extra refs before we do more work */ |
f900482da mm/migrate.c: cle... |
730 |
expected_count = expected_page_refs(mapping, page); |
cc4f11e69 mm: migrate: lock... |
731 732 |
if (page_count(page) != expected_count) return -EAGAIN; |
1d8b85ccf [PATCH] page migr... |
733 |
|
cc4f11e69 mm: migrate: lock... |
734 735 736 |
head = page_buffers(page); if (!buffer_migrate_lock_buffers(head, mode)) return -EAGAIN; |
1d8b85ccf [PATCH] page migr... |
737 |
|
89cb0888c mm: migrate: prov... |
738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 |
if (check_refs) { bool busy; bool invalidated = false; recheck_buffers: busy = false; spin_lock(&mapping->private_lock); bh = head; do { if (atomic_read(&bh->b_count)) { busy = true; break; } bh = bh->b_this_page; } while (bh != head); |
89cb0888c mm: migrate: prov... |
753 754 755 756 757 |
if (busy) { if (invalidated) { rc = -EAGAIN; goto unlock_buffers; } |
ebdf4de56 mm: migrate: fix ... |
758 |
spin_unlock(&mapping->private_lock); |
89cb0888c mm: migrate: prov... |
759 760 761 762 763 |
invalidate_bh_lrus(); invalidated = true; goto recheck_buffers; } } |
371096949 mm: migrate: remo... |
764 |
rc = migrate_page_move_mapping(mapping, newpage, page, 0); |
78bd52097 mm: adjust addres... |
765 |
if (rc != MIGRATEPAGE_SUCCESS) |
cc4f11e69 mm: migrate: lock... |
766 |
goto unlock_buffers; |
1d8b85ccf [PATCH] page migr... |
767 |
|
cd0f37154 mm/migrate.c: cal... |
768 |
attach_page_private(newpage, detach_page_private(page)); |
1d8b85ccf [PATCH] page migr... |
769 770 771 772 773 774 775 |
bh = head; do { set_bh_page(bh, newpage, bh_offset(bh)); bh = bh->b_this_page; } while (bh != head); |
2916ecc0f mm/migrate: new m... |
776 777 778 779 |
if (mode != MIGRATE_SYNC_NO_COPY) migrate_page_copy(newpage, page); else migrate_page_states(newpage, page); |
1d8b85ccf [PATCH] page migr... |
780 |
|
cc4f11e69 mm: migrate: lock... |
781 782 |
rc = MIGRATEPAGE_SUCCESS; unlock_buffers: |
ebdf4de56 mm: migrate: fix ... |
783 784 |
if (check_refs) spin_unlock(&mapping->private_lock); |
1d8b85ccf [PATCH] page migr... |
785 786 787 |
bh = head; do { unlock_buffer(bh); |
1d8b85ccf [PATCH] page migr... |
788 789 790 |
bh = bh->b_this_page; } while (bh != head); |
cc4f11e69 mm: migrate: lock... |
791 |
return rc; |
1d8b85ccf [PATCH] page migr... |
792 |
} |
89cb0888c mm: migrate: prov... |
793 794 795 796 797 798 799 800 801 802 803 |
/* * Migration function for pages with buffers. This function can only be used * if the underlying filesystem guarantees that no other references to "page" * exist. For example attached buffer heads are accessed only under page lock. */ int buffer_migrate_page(struct address_space *mapping, struct page *newpage, struct page *page, enum migrate_mode mode) { return __buffer_migrate_page(mapping, newpage, page, mode, false); } |
1d8b85ccf [PATCH] page migr... |
804 |
EXPORT_SYMBOL(buffer_migrate_page); |
89cb0888c mm: migrate: prov... |
805 806 807 808 809 810 811 812 813 814 815 816 |
/* * Same as above except that this variant is more careful and checks that there * are also no buffer head references. This function is the right one for * mappings where buffer heads are directly looked up and referenced (such as * block device mappings). */ int buffer_migrate_page_norefs(struct address_space *mapping, struct page *newpage, struct page *page, enum migrate_mode mode) { return __buffer_migrate_page(mapping, newpage, page, mode, true); } |
9361401eb [PATCH] BLOCK: Ma... |
817 |
#endif |
1d8b85ccf [PATCH] page migr... |
818 |
|
04e62a29b [PATCH] More page... |
819 820 821 822 |
/* * Writeback a page to clean the dirty state */ static int writeout(struct address_space *mapping, struct page *page) |
8351a6e47 [PATCH] page migr... |
823 |
{ |
04e62a29b [PATCH] More page... |
824 825 826 827 828 |
struct writeback_control wbc = { .sync_mode = WB_SYNC_NONE, .nr_to_write = 1, .range_start = 0, .range_end = LLONG_MAX, |
04e62a29b [PATCH] More page... |
829 830 831 832 833 834 835 836 837 838 839 |
.for_reclaim = 1 }; int rc; if (!mapping->a_ops->writepage) /* No write method for the address space */ return -EINVAL; if (!clear_page_dirty_for_io(page)) /* Someone else already triggered a write */ return -EAGAIN; |
8351a6e47 [PATCH] page migr... |
840 |
/* |
04e62a29b [PATCH] More page... |
841 842 843 844 845 846 |
* A dirty page may imply that the underlying filesystem has * the page on some queue. So the page must be clean for * migration. Writeout may mean we loose the lock and the * page state is no longer what we checked for earlier. * At this point we know that the migration attempt cannot * be successful. |
8351a6e47 [PATCH] page migr... |
847 |
*/ |
e388466de mm: make remove_m... |
848 |
remove_migration_ptes(page, page, false); |
8351a6e47 [PATCH] page migr... |
849 |
|
04e62a29b [PATCH] More page... |
850 |
rc = mapping->a_ops->writepage(page, &wbc); |
8351a6e47 [PATCH] page migr... |
851 |
|
04e62a29b [PATCH] More page... |
852 853 854 |
if (rc != AOP_WRITEPAGE_ACTIVATE) /* unlocked. Relock */ lock_page(page); |
bda8550de migration: fix wr... |
855 |
return (rc < 0) ? -EIO : -EAGAIN; |
04e62a29b [PATCH] More page... |
856 857 858 859 860 861 |
} /* * Default handling if a filesystem does not provide a migration function. */ static int fallback_migrate_page(struct address_space *mapping, |
a6bc32b89 mm: compaction: i... |
862 |
struct page *newpage, struct page *page, enum migrate_mode mode) |
04e62a29b [PATCH] More page... |
863 |
{ |
b969c4ab9 mm: compaction: d... |
864 |
if (PageDirty(page)) { |
a6bc32b89 mm: compaction: i... |
865 |
/* Only writeback pages in full synchronous migration */ |
2916ecc0f mm/migrate: new m... |
866 867 868 869 870 |
switch (mode) { case MIGRATE_SYNC: case MIGRATE_SYNC_NO_COPY: break; default: |
b969c4ab9 mm: compaction: d... |
871 |
return -EBUSY; |
2916ecc0f mm/migrate: new m... |
872 |
} |
04e62a29b [PATCH] More page... |
873 |
return writeout(mapping, page); |
b969c4ab9 mm: compaction: d... |
874 |
} |
8351a6e47 [PATCH] page migr... |
875 876 877 878 879 |
/* * Buffers may be managed in a filesystem specific way. * We must have no buffers or drop them. */ |
266cf658e FS-Cache: Recruit... |
880 |
if (page_has_private(page) && |
8351a6e47 [PATCH] page migr... |
881 |
!try_to_release_page(page, GFP_KERNEL)) |
806031bb5 mm, migrate: imme... |
882 |
return mode == MIGRATE_SYNC ? -EAGAIN : -EBUSY; |
8351a6e47 [PATCH] page migr... |
883 |
|
a6bc32b89 mm: compaction: i... |
884 |
return migrate_page(mapping, newpage, page, mode); |
8351a6e47 [PATCH] page migr... |
885 |
} |
1d8b85ccf [PATCH] page migr... |
886 |
/* |
e24f0b8f7 [PATCH] page migr... |
887 888 889 890 891 |
* Move a page to a newly allocated page * The page is locked and all ptes have been successfully removed. * * The new page will have replaced the old page if this function * is successful. |
894bc3104 Unevictable LRU I... |
892 893 894 |
* * Return value: * < 0 - error code |
78bd52097 mm: adjust addres... |
895 |
* MIGRATEPAGE_SUCCESS - success |
e24f0b8f7 [PATCH] page migr... |
896 |
*/ |
3fe2011ff mm: migration: al... |
897 |
static int move_to_new_page(struct page *newpage, struct page *page, |
5c3f9a673 mm: page migratio... |
898 |
enum migrate_mode mode) |
e24f0b8f7 [PATCH] page migr... |
899 900 |
{ struct address_space *mapping; |
bda807d44 mm: migrate: supp... |
901 902 |
int rc = -EAGAIN; bool is_lru = !__PageMovable(page); |
e24f0b8f7 [PATCH] page migr... |
903 |
|
7db7671f8 mm: page migratio... |
904 905 |
VM_BUG_ON_PAGE(!PageLocked(page), page); VM_BUG_ON_PAGE(!PageLocked(newpage), newpage); |
e24f0b8f7 [PATCH] page migr... |
906 |
|
e24f0b8f7 [PATCH] page migr... |
907 |
mapping = page_mapping(page); |
bda807d44 mm: migrate: supp... |
908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 |
if (likely(is_lru)) { if (!mapping) rc = migrate_page(mapping, newpage, page, mode); else if (mapping->a_ops->migratepage) /* * Most pages have a mapping and most filesystems * provide a migratepage callback. Anonymous pages * are part of swap space which also has its own * migratepage callback. This is the most common path * for page migration. */ rc = mapping->a_ops->migratepage(mapping, newpage, page, mode); else rc = fallback_migrate_page(mapping, newpage, page, mode); } else { |
e24f0b8f7 [PATCH] page migr... |
926 |
/* |
bda807d44 mm: migrate: supp... |
927 928 |
* In case of non-lru page, it could be released after * isolation step. In that case, we shouldn't try migration. |
e24f0b8f7 [PATCH] page migr... |
929 |
*/ |
bda807d44 mm: migrate: supp... |
930 931 932 933 934 935 936 937 938 939 940 941 |
VM_BUG_ON_PAGE(!PageIsolated(page), page); if (!PageMovable(page)) { rc = MIGRATEPAGE_SUCCESS; __ClearPageIsolated(page); goto out; } rc = mapping->a_ops->migratepage(mapping, newpage, page, mode); WARN_ON_ONCE(rc == MIGRATEPAGE_SUCCESS && !PageIsolated(page)); } |
e24f0b8f7 [PATCH] page migr... |
942 |
|
5c3f9a673 mm: page migratio... |
943 944 945 946 947 |
/* * When successful, old pagecache page->mapping must be cleared before * page is freed; but stats require that PageAnon be left as PageAnon. */ if (rc == MIGRATEPAGE_SUCCESS) { |
bda807d44 mm: migrate: supp... |
948 949 950 951 952 953 954 955 956 957 958 |
if (__PageMovable(page)) { VM_BUG_ON_PAGE(!PageIsolated(page), page); /* * We clear PG_movable under page_lock so any compactor * cannot try to migrate this page. */ __ClearPageIsolated(page); } /* |
c23a0c997 mm/migrate: clean... |
959 |
* Anonymous and movable page->mapping will be cleared by |
bda807d44 mm: migrate: supp... |
960 961 962 963 |
* free_pages_prepare so don't reset it here for keeping * the type to work PageAnon, for example. */ if (!PageMappingFlags(page)) |
5c3f9a673 mm: page migratio... |
964 |
page->mapping = NULL; |
d2b2c6dd2 mm/migrate.c: add... |
965 |
|
25b2995a3 mm: remove MEMORY... |
966 |
if (likely(!is_zone_device_page(newpage))) |
d2b2c6dd2 mm/migrate.c: add... |
967 |
flush_dcache_page(newpage); |
3fe2011ff mm: migration: al... |
968 |
} |
bda807d44 mm: migrate: supp... |
969 |
out: |
e24f0b8f7 [PATCH] page migr... |
970 971 |
return rc; } |
0dabec93d mm: migration: cl... |
972 |
static int __unmap_and_move(struct page *page, struct page *newpage, |
9c620e2bc mm: remove offlin... |
973 |
int force, enum migrate_mode mode) |
e24f0b8f7 [PATCH] page migr... |
974 |
{ |
0dabec93d mm: migration: cl... |
975 |
int rc = -EAGAIN; |
2ebba6b7e mm: unmapped page... |
976 |
int page_was_mapped = 0; |
3f6c82728 mm: migration: ta... |
977 |
struct anon_vma *anon_vma = NULL; |
bda807d44 mm: migrate: supp... |
978 |
bool is_lru = !__PageMovable(page); |
95a402c38 [PATCH] page migr... |
979 |
|
529ae9aaa mm: rename page t... |
980 |
if (!trylock_page(page)) { |
a6bc32b89 mm: compaction: i... |
981 |
if (!force || mode == MIGRATE_ASYNC) |
0dabec93d mm: migration: cl... |
982 |
goto out; |
3e7d34497 mm: vmscan: recla... |
983 984 985 986 987 988 989 |
/* * It's not safe for direct compaction to call lock_page. * For example, during page readahead pages are added locked * to the LRU. Later, when the IO completes the pages are * marked uptodate and unlocked. However, the queueing * could be merging multiple pages for one bio (e.g. |
d4388340a fs: convert mpage... |
990 |
* mpage_readahead). If an allocation happens for the |
3e7d34497 mm: vmscan: recla... |
991 992 993 994 995 996 997 |
* second or third page, the process can end up locking * the same page twice and deadlocking. Rather than * trying to be clever about what pages can be locked, * avoid the use of lock_page for direct compaction * altogether. */ if (current->flags & PF_MEMALLOC) |
0dabec93d mm: migration: cl... |
998 |
goto out; |
3e7d34497 mm: vmscan: recla... |
999 |
|
e24f0b8f7 [PATCH] page migr... |
1000 1001 1002 1003 |
lock_page(page); } if (PageWriteback(page)) { |
11bc82d67 mm: compaction: U... |
1004 |
/* |
fed5b64a9 mm/migrate: fix c... |
1005 |
* Only in the case of a full synchronous migration is it |
a6bc32b89 mm: compaction: i... |
1006 1007 1008 |
* necessary to wait for PageWriteback. In the async case, * the retry loop is too short and in the sync-light case, * the overhead of stalling is too much |
11bc82d67 mm: compaction: U... |
1009 |
*/ |
2916ecc0f mm/migrate: new m... |
1010 1011 1012 1013 1014 |
switch (mode) { case MIGRATE_SYNC: case MIGRATE_SYNC_NO_COPY: break; default: |
11bc82d67 mm: compaction: U... |
1015 |
rc = -EBUSY; |
0a31bc97c mm: memcontrol: r... |
1016 |
goto out_unlock; |
11bc82d67 mm: compaction: U... |
1017 1018 |
} if (!force) |
0a31bc97c mm: memcontrol: r... |
1019 |
goto out_unlock; |
e24f0b8f7 [PATCH] page migr... |
1020 1021 |
wait_on_page_writeback(page); } |
03f15c86c mm: simplify page... |
1022 |
|
e24f0b8f7 [PATCH] page migr... |
1023 |
/* |
dc386d4d1 memory unplug: mi... |
1024 1025 |
* By try_to_unmap(), page->mapcount goes down to 0 here. In this case, * we cannot notice that anon_vma is freed while we migrates a page. |
1ce82b69e mm: fix migration... |
1026 |
* This get_anon_vma() delays freeing anon_vma pointer until the end |
dc386d4d1 memory unplug: mi... |
1027 |
* of migration. File cache pages are no problem because of page_lock() |
989f89c57 fix rcu_read_lock... |
1028 1029 |
* File Caches may use write_page() or lock_page() in migration, then, * just care Anon page here. |
03f15c86c mm: simplify page... |
1030 1031 1032 1033 1034 1035 |
* * Only page_get_anon_vma() understands the subtleties of * getting a hold on an anon_vma from outside one of its mms. * But if we cannot get anon_vma, then we won't need it anyway, * because that implies that the anon page is no longer mapped * (and cannot be remapped so long as we hold the page lock). |
dc386d4d1 memory unplug: mi... |
1036 |
*/ |
03f15c86c mm: simplify page... |
1037 |
if (PageAnon(page) && !PageKsm(page)) |
746b18d42 mm: use refcounts... |
1038 |
anon_vma = page_get_anon_vma(page); |
62e1c5530 page migraton: ha... |
1039 |
|
7db7671f8 mm: page migratio... |
1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 |
/* * Block others from accessing the new page when we get around to * establishing additional references. We are usually the only one * holding a reference to newpage at this point. We used to have a BUG * here if trylock_page(newpage) fails, but would like to allow for * cases where there might be a race with the previous use of newpage. * This is much like races on refcount of oldpage: just don't BUG(). */ if (unlikely(!trylock_page(newpage))) goto out_unlock; |
bda807d44 mm: migrate: supp... |
1050 1051 1052 1053 |
if (unlikely(!is_lru)) { rc = move_to_new_page(newpage, page, mode); goto out_unlock_both; } |
dc386d4d1 memory unplug: mi... |
1054 |
/* |
62e1c5530 page migraton: ha... |
1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 |
* Corner case handling: * 1. When a new swap-cache page is read into, it is added to the LRU * and treated as swapcache but it has no rmap yet. * Calling try_to_unmap() against a page->mapping==NULL page will * trigger a BUG. So handle it here. * 2. An orphaned page (see truncate_complete_page) might have * fs-private metadata. The page can be picked up due to memory * offlining. Everywhere else except page reclaim, the page is * invisible to the vm, so the page can not be migrated. So try to * free the metadata, so the page can be freed. |
e24f0b8f7 [PATCH] page migr... |
1065 |
*/ |
62e1c5530 page migraton: ha... |
1066 |
if (!page->mapping) { |
309381fea mm: dump page whe... |
1067 |
VM_BUG_ON_PAGE(PageAnon(page), page); |
1ce82b69e mm: fix migration... |
1068 |
if (page_has_private(page)) { |
62e1c5530 page migraton: ha... |
1069 |
try_to_free_buffers(page); |
7db7671f8 mm: page migratio... |
1070 |
goto out_unlock_both; |
62e1c5530 page migraton: ha... |
1071 |
} |
7db7671f8 mm: page migratio... |
1072 1073 |
} else if (page_mapped(page)) { /* Establish migration ptes */ |
03f15c86c mm: simplify page... |
1074 1075 |
VM_BUG_ON_PAGE(PageAnon(page) && !PageKsm(page) && !anon_vma, page); |
dd156e3fc mm/rmap: always d... |
1076 |
try_to_unmap(page, TTU_MIGRATION|TTU_IGNORE_MLOCK); |
2ebba6b7e mm: unmapped page... |
1077 1078 |
page_was_mapped = 1; } |
dc386d4d1 memory unplug: mi... |
1079 |
|
e6a1530d6 [PATCH] Allow mig... |
1080 |
if (!page_mapped(page)) |
5c3f9a673 mm: page migratio... |
1081 |
rc = move_to_new_page(newpage, page, mode); |
e24f0b8f7 [PATCH] page migr... |
1082 |
|
5c3f9a673 mm: page migratio... |
1083 1084 |
if (page_was_mapped) remove_migration_ptes(page, |
e388466de mm: make remove_m... |
1085 |
rc == MIGRATEPAGE_SUCCESS ? newpage : page, false); |
3f6c82728 mm: migration: ta... |
1086 |
|
7db7671f8 mm: page migratio... |
1087 1088 1089 |
out_unlock_both: unlock_page(newpage); out_unlock: |
3f6c82728 mm: migration: ta... |
1090 |
/* Drop an anon_vma reference if we took one */ |
76545066c mm: extend KSM re... |
1091 |
if (anon_vma) |
9e60109f1 mm: rename drop_a... |
1092 |
put_anon_vma(anon_vma); |
e24f0b8f7 [PATCH] page migr... |
1093 |
unlock_page(page); |
0dabec93d mm: migration: cl... |
1094 |
out: |
c6c919eb9 mm: use put_page(... |
1095 1096 1097 1098 |
/* * If migration is successful, decrease refcount of the newpage * which will not free the page because new page owner increased * refcounter. As well, if it is LRU page, add the page to LRU |
e0a352fab mm: migrate: don'... |
1099 1100 1101 1102 |
* list in here. Use the old state of the isolated source page to * determine if we migrated a LRU page. newpage was already unlocked * and possibly modified by its owner - don't rely on the page * state. |
c6c919eb9 mm: use put_page(... |
1103 1104 |
*/ if (rc == MIGRATEPAGE_SUCCESS) { |
e0a352fab mm: migrate: don'... |
1105 |
if (unlikely(!is_lru)) |
c6c919eb9 mm: use put_page(... |
1106 1107 1108 1109 |
put_page(newpage); else putback_lru_page(newpage); } |
0dabec93d mm: migration: cl... |
1110 1111 |
return rc; } |
95a402c38 [PATCH] page migr... |
1112 |
|
0dabec93d mm: migration: cl... |
1113 1114 1115 1116 |
/* * Obtain the lock on page, remove all ptes and migrate the page * to the newly allocated page in newpage. */ |
6ec4476ac Raise gcc version... |
1117 |
static int unmap_and_move(new_page_t get_new_page, |
ef2a5153b mm/migrate: mark ... |
1118 1119 |
free_page_t put_new_page, unsigned long private, struct page *page, |
add05cece mm: soft-offline:... |
1120 1121 |
int force, enum migrate_mode mode, enum migrate_reason reason) |
0dabec93d mm: migration: cl... |
1122 |
{ |
2def7424c mm: page migratio... |
1123 |
int rc = MIGRATEPAGE_SUCCESS; |
74d4a5797 mm/migrate.c: han... |
1124 |
struct page *newpage = NULL; |
0dabec93d mm: migration: cl... |
1125 |
|
94723aafb mm: unclutter THP... |
1126 1127 |
if (!thp_migration_supported() && PageTransHuge(page)) return -ENOMEM; |
0dabec93d mm: migration: cl... |
1128 1129 |
if (page_count(page) == 1) { /* page was freed from under us. So we are done. */ |
c6c919eb9 mm: use put_page(... |
1130 1131 |
ClearPageActive(page); ClearPageUnevictable(page); |
bda807d44 mm: migrate: supp... |
1132 1133 1134 1135 1136 1137 |
if (unlikely(__PageMovable(page))) { lock_page(page); if (!PageMovable(page)) __ClearPageIsolated(page); unlock_page(page); } |
0dabec93d mm: migration: cl... |
1138 1139 |
goto out; } |
74d4a5797 mm/migrate.c: han... |
1140 1141 1142 |
newpage = get_new_page(page, private); if (!newpage) return -ENOMEM; |
9c620e2bc mm: remove offlin... |
1143 |
rc = __unmap_and_move(page, newpage, force, mode); |
c6c919eb9 mm: use put_page(... |
1144 |
if (rc == MIGRATEPAGE_SUCCESS) |
7cd12b4ab mm, page_owner: t... |
1145 |
set_page_owner_migrate_reason(newpage, reason); |
bf6bddf19 mm: introduce com... |
1146 |
|
0dabec93d mm: migration: cl... |
1147 |
out: |
e24f0b8f7 [PATCH] page migr... |
1148 |
if (rc != -EAGAIN) { |
0dabec93d mm: migration: cl... |
1149 1150 1151 |
/* * A page that has been migrated has all references * removed and will be freed. A page that has not been |
c23a0c997 mm/migrate: clean... |
1152 |
* migrated will have kept its references and be restored. |
0dabec93d mm: migration: cl... |
1153 1154 |
*/ list_del(&page->lru); |
6afcf8ef0 mm, compaction: f... |
1155 1156 1157 1158 1159 1160 1161 |
/* * Compaction can migrate also non-LRU pages which are * not accounted to NR_ISOLATED_*. They can be recognized * as __PageMovable */ if (likely(!__PageMovable(page))) |
e8db67eb0 mm: migrate: move... |
1162 |
mod_node_page_state(page_pgdat(page), NR_ISOLATED_ANON + |
6c357848b mm: replace hpage... |
1163 |
page_is_file_lru(page), -thp_nr_pages(page)); |
c6c919eb9 mm: use put_page(... |
1164 1165 1166 1167 1168 1169 1170 1171 |
} /* * If migration is successful, releases reference grabbed during * isolation. Otherwise, restore the page to right list unless * we want to retry. */ if (rc == MIGRATEPAGE_SUCCESS) { |
79f5f8fab mm,hwpoison: rewo... |
1172 |
if (reason != MR_MEMORY_FAILURE) |
d7e69488b mm/hwpoison: fix ... |
1173 |
/* |
79f5f8fab mm,hwpoison: rewo... |
1174 |
* We release the page in page_handle_poison. |
d7e69488b mm/hwpoison: fix ... |
1175 |
*/ |
79f5f8fab mm,hwpoison: rewo... |
1176 |
put_page(page); |
c6c919eb9 mm: use put_page(... |
1177 |
} else { |
bda807d44 mm: migrate: supp... |
1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 |
if (rc != -EAGAIN) { if (likely(!__PageMovable(page))) { putback_lru_page(page); goto put_new; } lock_page(page); if (PageMovable(page)) putback_movable_page(page); else __ClearPageIsolated(page); unlock_page(page); put_page(page); } put_new: |
c6c919eb9 mm: use put_page(... |
1193 1194 1195 1196 |
if (put_new_page) put_new_page(newpage, private); else put_page(newpage); |
e24f0b8f7 [PATCH] page migr... |
1197 |
} |
68711a746 mm, migration: ad... |
1198 |
|
e24f0b8f7 [PATCH] page migr... |
1199 1200 1201 1202 |
return rc; } /* |
290408d4a hugetlb: hugepage... |
1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 |
* Counterpart of unmap_and_move_page() for hugepage migration. * * This function doesn't wait the completion of hugepage I/O * because there is no race between I/O and migration for hugepage. * Note that currently hugepage I/O occurs only in direct I/O * where no lock is held and PG_writeback is irrelevant, * and writeback status of all subpages are counted in the reference * count of the head page (i.e. if all subpages of a 2MB hugepage are * under direct I/O, the reference of the head page is 512 and a bit more.) * This means that when we try to migrate hugepage whose subpages are * doing direct I/O, some references remain after try_to_unmap() and * hugepage migration fails without data corruption. * * There is also no race when direct I/O is issued on the page under migration, * because then pte is replaced with migration swap entry and direct I/O code * will wait in the page fault for migration to complete. */ static int unmap_and_move_huge_page(new_page_t get_new_page, |
68711a746 mm, migration: ad... |
1221 1222 |
free_page_t put_new_page, unsigned long private, struct page *hpage, int force, |
7cd12b4ab mm, page_owner: t... |
1223 |
enum migrate_mode mode, int reason) |
290408d4a hugetlb: hugepage... |
1224 |
{ |
2def7424c mm: page migratio... |
1225 |
int rc = -EAGAIN; |
2ebba6b7e mm: unmapped page... |
1226 |
int page_was_mapped = 0; |
32665f2bb mm/migrate: corre... |
1227 |
struct page *new_hpage; |
290408d4a hugetlb: hugepage... |
1228 |
struct anon_vma *anon_vma = NULL; |
c0d0381ad hugetlbfs: use i_... |
1229 |
struct address_space *mapping = NULL; |
290408d4a hugetlb: hugepage... |
1230 |
|
83467efbd mm: migrate: chec... |
1231 |
/* |
7ed2c31da mm/hugetlb: disti... |
1232 |
* Migratability of hugepages depends on architectures and their size. |
83467efbd mm: migrate: chec... |
1233 1234 1235 1236 1237 |
* This check is necessary because some callers of hugepage migration * like soft offline and memory hotremove don't walk through page * tables or check whether the hugepage is pmd-based or not before * kicking migration. */ |
100873d7a hugetlb: rename h... |
1238 |
if (!hugepage_migration_supported(page_hstate(hpage))) { |
32665f2bb mm/migrate: corre... |
1239 |
putback_active_hugepage(hpage); |
83467efbd mm: migrate: chec... |
1240 |
return -ENOSYS; |
32665f2bb mm/migrate: corre... |
1241 |
} |
83467efbd mm: migrate: chec... |
1242 |
|
666feb21a mm, migrate: remo... |
1243 |
new_hpage = get_new_page(hpage, private); |
290408d4a hugetlb: hugepage... |
1244 1245 |
if (!new_hpage) return -ENOMEM; |
290408d4a hugetlb: hugepage... |
1246 |
if (!trylock_page(hpage)) { |
2916ecc0f mm/migrate: new m... |
1247 |
if (!force) |
290408d4a hugetlb: hugepage... |
1248 |
goto out; |
2916ecc0f mm/migrate: new m... |
1249 1250 1251 1252 1253 1254 1255 |
switch (mode) { case MIGRATE_SYNC: case MIGRATE_SYNC_NO_COPY: break; default: goto out; } |
290408d4a hugetlb: hugepage... |
1256 1257 |
lock_page(hpage); } |
cb6acd01e hugetlbfs: fix ra... |
1258 1259 1260 1261 1262 1263 1264 1265 1266 |
/* * Check for pages which are in the process of being freed. Without * page_mapping() set, hugetlbfs specific move page routine will not * be called and we could leak usage counts for subpools. */ if (page_private(hpage) && !page_mapping(hpage)) { rc = -EBUSY; goto out_unlock; } |
746b18d42 mm: use refcounts... |
1267 1268 |
if (PageAnon(hpage)) anon_vma = page_get_anon_vma(hpage); |
290408d4a hugetlb: hugepage... |
1269 |
|
7db7671f8 mm: page migratio... |
1270 1271 |
if (unlikely(!trylock_page(new_hpage))) goto put_anon; |
2ebba6b7e mm: unmapped page... |
1272 |
if (page_mapped(hpage)) { |
336bf30eb hugetlbfs: fix an... |
1273 |
bool mapping_locked = false; |
dd156e3fc mm/rmap: always d... |
1274 |
enum ttu_flags ttu = TTU_MIGRATION|TTU_IGNORE_MLOCK; |
336bf30eb hugetlbfs: fix an... |
1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 |
if (!PageAnon(hpage)) { /* * In shared mappings, try_to_unmap could potentially * call huge_pmd_unshare. Because of this, take * semaphore in write mode here and set TTU_RMAP_LOCKED * to let lower levels know we have taken the lock. */ mapping = hugetlb_page_mapping_lock_write(hpage); if (unlikely(!mapping)) goto unlock_put_anon; mapping_locked = true; ttu |= TTU_RMAP_LOCKED; } |
c0d0381ad hugetlbfs: use i_... |
1290 |
|
336bf30eb hugetlbfs: fix an... |
1291 |
try_to_unmap(hpage, ttu); |
2ebba6b7e mm: unmapped page... |
1292 |
page_was_mapped = 1; |
336bf30eb hugetlbfs: fix an... |
1293 1294 1295 |
if (mapping_locked) i_mmap_unlock_write(mapping); |
2ebba6b7e mm: unmapped page... |
1296 |
} |
290408d4a hugetlb: hugepage... |
1297 1298 |
if (!page_mapped(hpage)) |
5c3f9a673 mm: page migratio... |
1299 |
rc = move_to_new_page(new_hpage, hpage, mode); |
290408d4a hugetlb: hugepage... |
1300 |
|
336bf30eb hugetlbfs: fix an... |
1301 |
if (page_was_mapped) |
5c3f9a673 mm: page migratio... |
1302 |
remove_migration_ptes(hpage, |
336bf30eb hugetlbfs: fix an... |
1303 |
rc == MIGRATEPAGE_SUCCESS ? new_hpage : hpage, false); |
290408d4a hugetlb: hugepage... |
1304 |
|
c0d0381ad hugetlbfs: use i_... |
1305 |
unlock_put_anon: |
7db7671f8 mm: page migratio... |
1306 1307 1308 |
unlock_page(new_hpage); put_anon: |
fd4a4663d mm: fix hugepage ... |
1309 |
if (anon_vma) |
9e60109f1 mm: rename drop_a... |
1310 |
put_anon_vma(anon_vma); |
8e6ac7fab hugetlb/cgroup: m... |
1311 |
|
2def7424c mm: page migratio... |
1312 |
if (rc == MIGRATEPAGE_SUCCESS) { |
ab5ac90ae mm, hugetlb: do n... |
1313 |
move_hugetlb_state(hpage, new_hpage, reason); |
2def7424c mm: page migratio... |
1314 1315 |
put_new_page = NULL; } |
8e6ac7fab hugetlb/cgroup: m... |
1316 |
|
cb6acd01e hugetlbfs: fix ra... |
1317 |
out_unlock: |
290408d4a hugetlb: hugepage... |
1318 |
unlock_page(hpage); |
09761333e mm/migrate.c: pai... |
1319 |
out: |
b8ec1cee5 mm: soft-offline:... |
1320 1321 |
if (rc != -EAGAIN) putback_active_hugepage(hpage); |
68711a746 mm, migration: ad... |
1322 1323 1324 1325 1326 1327 |
/* * If migration was not successful and there's a freeing callback, use * it. Otherwise, put_page() will drop the reference grabbed during * isolation. */ |
2def7424c mm: page migratio... |
1328 |
if (put_new_page) |
68711a746 mm, migration: ad... |
1329 1330 |
put_new_page(new_hpage, private); else |
3aaa76e12 mm: migrate: huge... |
1331 |
putback_active_hugepage(new_hpage); |
68711a746 mm, migration: ad... |
1332 |
|
290408d4a hugetlb: hugepage... |
1333 1334 1335 1336 |
return rc; } /* |
c73e5c9c5 mm: rewrite the c... |
1337 1338 |
* migrate_pages - migrate the pages specified in a list, to the free pages * supplied as the target for the page migration |
b20a35035 [PATCH] page migr... |
1339 |
* |
c73e5c9c5 mm: rewrite the c... |
1340 1341 1342 |
* @from: The list of pages to be migrated. * @get_new_page: The function used to allocate free pages to be used * as the target of the page migration. |
68711a746 mm, migration: ad... |
1343 1344 |
* @put_new_page: The function used to free target pages if migration * fails, or NULL if no special handling is necessary. |
c73e5c9c5 mm: rewrite the c... |
1345 1346 1347 1348 |
* @private: Private data to be passed on to get_new_page() * @mode: The migration mode that specifies the constraints for * page migration, if any. * @reason: The reason for page migration. |
b20a35035 [PATCH] page migr... |
1349 |
* |
c73e5c9c5 mm: rewrite the c... |
1350 1351 |
* The function returns after 10 attempts or if no pages are movable any more * because the list has become empty or no retryable pages exist any more. |
14e0f9bcc mm: correct a cou... |
1352 |
* The caller should call putback_movable_pages() to return pages to the LRU |
28bd65781 mm: migration: cl... |
1353 |
* or free list only if ret != 0. |
b20a35035 [PATCH] page migr... |
1354 |
* |
c73e5c9c5 mm: rewrite the c... |
1355 |
* Returns the number of pages that were not migrated, or an error code. |
b20a35035 [PATCH] page migr... |
1356 |
*/ |
9c620e2bc mm: remove offlin... |
1357 |
int migrate_pages(struct list_head *from, new_page_t get_new_page, |
68711a746 mm, migration: ad... |
1358 1359 |
free_page_t put_new_page, unsigned long private, enum migrate_mode mode, int reason) |
b20a35035 [PATCH] page migr... |
1360 |
{ |
e24f0b8f7 [PATCH] page migr... |
1361 |
int retry = 1; |
1a5bae25e mm/vmstat: add ev... |
1362 |
int thp_retry = 1; |
b20a35035 [PATCH] page migr... |
1363 |
int nr_failed = 0; |
5647bc293 mm: compaction: M... |
1364 |
int nr_succeeded = 0; |
1a5bae25e mm/vmstat: add ev... |
1365 1366 1367 |
int nr_thp_succeeded = 0; int nr_thp_failed = 0; int nr_thp_split = 0; |
b20a35035 [PATCH] page migr... |
1368 |
int pass = 0; |
1a5bae25e mm/vmstat: add ev... |
1369 |
bool is_thp = false; |
b20a35035 [PATCH] page migr... |
1370 1371 1372 |
struct page *page; struct page *page2; int swapwrite = current->flags & PF_SWAPWRITE; |
1a5bae25e mm/vmstat: add ev... |
1373 |
int rc, nr_subpages; |
b20a35035 [PATCH] page migr... |
1374 1375 1376 |
if (!swapwrite) current->flags |= PF_SWAPWRITE; |
1a5bae25e mm/vmstat: add ev... |
1377 |
for (pass = 0; pass < 10 && (retry || thp_retry); pass++) { |
e24f0b8f7 [PATCH] page migr... |
1378 |
retry = 0; |
1a5bae25e mm/vmstat: add ev... |
1379 |
thp_retry = 0; |
b20a35035 [PATCH] page migr... |
1380 |
|
e24f0b8f7 [PATCH] page migr... |
1381 |
list_for_each_entry_safe(page, page2, from, lru) { |
94723aafb mm: unclutter THP... |
1382 |
retry: |
1a5bae25e mm/vmstat: add ev... |
1383 1384 1385 1386 1387 |
/* * THP statistics is based on the source huge page. * Capture required information that might get lost * during migration. */ |
6c5c7b9f3 mm/migrate: corre... |
1388 |
is_thp = PageTransHuge(page) && !PageHuge(page); |
6c357848b mm: replace hpage... |
1389 |
nr_subpages = thp_nr_pages(page); |
e24f0b8f7 [PATCH] page migr... |
1390 |
cond_resched(); |
2d1db3b11 [PATCH] page migr... |
1391 |
|
31caf665e mm: migrate: make... |
1392 1393 |
if (PageHuge(page)) rc = unmap_and_move_huge_page(get_new_page, |
68711a746 mm, migration: ad... |
1394 |
put_new_page, private, page, |
7cd12b4ab mm, page_owner: t... |
1395 |
pass > 2, mode, reason); |
31caf665e mm: migrate: make... |
1396 |
else |
68711a746 mm, migration: ad... |
1397 |
rc = unmap_and_move(get_new_page, put_new_page, |
add05cece mm: soft-offline:... |
1398 1399 |
private, page, pass > 2, mode, reason); |
2d1db3b11 [PATCH] page migr... |
1400 |
|
e24f0b8f7 [PATCH] page migr... |
1401 |
switch(rc) { |
95a402c38 [PATCH] page migr... |
1402 |
case -ENOMEM: |
94723aafb mm: unclutter THP... |
1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413 |
/* * THP migration might be unsupported or the * allocation could've failed so we should * retry on the same page with the THP split * to base pages. * * Head page is retried immediately and tail * pages are added to the tail of the list so * we encounter them after the rest of the list * is processed. */ |
6c5c7b9f3 mm/migrate: corre... |
1414 |
if (is_thp) { |
94723aafb mm: unclutter THP... |
1415 1416 1417 1418 1419 |
lock_page(page); rc = split_huge_page_to_list(page, from); unlock_page(page); if (!rc) { list_safe_reset_next(page, page2, lru); |
1a5bae25e mm/vmstat: add ev... |
1420 |
nr_thp_split++; |
94723aafb mm: unclutter THP... |
1421 1422 |
goto retry; } |
6c5c7b9f3 mm/migrate: corre... |
1423 |
|
1a5bae25e mm/vmstat: add ev... |
1424 1425 1426 1427 |
nr_thp_failed++; nr_failed += nr_subpages; goto out; } |
dfef2ef40 mm, migrate: incr... |
1428 |
nr_failed++; |
95a402c38 [PATCH] page migr... |
1429 |
goto out; |
e24f0b8f7 [PATCH] page migr... |
1430 |
case -EAGAIN: |
1a5bae25e mm/vmstat: add ev... |
1431 1432 1433 1434 |
if (is_thp) { thp_retry++; break; } |
2d1db3b11 [PATCH] page migr... |
1435 |
retry++; |
e24f0b8f7 [PATCH] page migr... |
1436 |
break; |
78bd52097 mm: adjust addres... |
1437 |
case MIGRATEPAGE_SUCCESS: |
1a5bae25e mm/vmstat: add ev... |
1438 1439 1440 1441 1442 |
if (is_thp) { nr_thp_succeeded++; nr_succeeded += nr_subpages; break; } |
5647bc293 mm: compaction: M... |
1443 |
nr_succeeded++; |
e24f0b8f7 [PATCH] page migr... |
1444 1445 |
break; default: |
354a33633 mm/migrate: add c... |
1446 1447 1448 1449 1450 1451 |
/* * Permanent failure (-EBUSY, -ENOSYS, etc.): * unlike -EAGAIN case, the failed page is * removed from migration page list and not * retried in the next outer loop. */ |
1a5bae25e mm/vmstat: add ev... |
1452 1453 1454 1455 1456 |
if (is_thp) { nr_thp_failed++; nr_failed += nr_subpages; break; } |
2d1db3b11 [PATCH] page migr... |
1457 |
nr_failed++; |
e24f0b8f7 [PATCH] page migr... |
1458 |
break; |
2d1db3b11 [PATCH] page migr... |
1459 |
} |
b20a35035 [PATCH] page migr... |
1460 1461 |
} } |
1a5bae25e mm/vmstat: add ev... |
1462 1463 |
nr_failed += retry + thp_retry; nr_thp_failed += thp_retry; |
f2f81fb2b mm, migrate: coun... |
1464 |
rc = nr_failed; |
95a402c38 [PATCH] page migr... |
1465 |
out: |
1a5bae25e mm/vmstat: add ev... |
1466 1467 1468 1469 1470 1471 1472 |
count_vm_events(PGMIGRATE_SUCCESS, nr_succeeded); count_vm_events(PGMIGRATE_FAIL, nr_failed); count_vm_events(THP_MIGRATION_SUCCESS, nr_thp_succeeded); count_vm_events(THP_MIGRATION_FAIL, nr_thp_failed); count_vm_events(THP_MIGRATION_SPLIT, nr_thp_split); trace_mm_migrate_pages(nr_succeeded, nr_failed, nr_thp_succeeded, nr_thp_failed, nr_thp_split, mode, reason); |
7b2a2d4a1 mm: migrate: Add ... |
1473 |
|
b20a35035 [PATCH] page migr... |
1474 1475 |
if (!swapwrite) current->flags &= ~PF_SWAPWRITE; |
78bd52097 mm: adjust addres... |
1476 |
return rc; |
b20a35035 [PATCH] page migr... |
1477 |
} |
95a402c38 [PATCH] page migr... |
1478 |
|
19fc7bed2 mm/migrate: intro... |
1479 |
struct page *alloc_migration_target(struct page *page, unsigned long private) |
b4b382238 mm/migrate: move ... |
1480 |
{ |
19fc7bed2 mm/migrate: intro... |
1481 1482 |
struct migration_target_control *mtc; gfp_t gfp_mask; |
b4b382238 mm/migrate: move ... |
1483 1484 |
unsigned int order = 0; struct page *new_page = NULL; |
19fc7bed2 mm/migrate: intro... |
1485 1486 1487 1488 1489 1490 1491 1492 |
int nid; int zidx; mtc = (struct migration_target_control *)private; gfp_mask = mtc->gfp_mask; nid = mtc->nid; if (nid == NUMA_NO_NODE) nid = page_to_nid(page); |
b4b382238 mm/migrate: move ... |
1493 |
|
d92bbc271 mm/hugetlb: unify... |
1494 1495 |
if (PageHuge(page)) { struct hstate *h = page_hstate(compound_head(page)); |
19fc7bed2 mm/migrate: intro... |
1496 1497 |
gfp_mask = htlb_modify_alloc_mask(h, gfp_mask); return alloc_huge_page_nodemask(h, nid, mtc->nmask, gfp_mask); |
d92bbc271 mm/hugetlb: unify... |
1498 |
} |
b4b382238 mm/migrate: move ... |
1499 1500 |
if (PageTransHuge(page)) { |
9933a0c8a mm/migrate: clear... |
1501 1502 1503 1504 1505 |
/* * clear __GFP_RECLAIM to make the migration callback * consistent with regular THP allocations. */ gfp_mask &= ~__GFP_RECLAIM; |
b4b382238 mm/migrate: move ... |
1506 1507 1508 |
gfp_mask |= GFP_TRANSHUGE; order = HPAGE_PMD_ORDER; } |
19fc7bed2 mm/migrate: intro... |
1509 1510 |
zidx = zone_idx(page_zone(page)); if (is_highmem_idx(zidx) || zidx == ZONE_MOVABLE) |
b4b382238 mm/migrate: move ... |
1511 |
gfp_mask |= __GFP_HIGHMEM; |
19fc7bed2 mm/migrate: intro... |
1512 |
new_page = __alloc_pages_nodemask(gfp_mask, order, nid, mtc->nmask); |
b4b382238 mm/migrate: move ... |
1513 1514 1515 1516 1517 1518 |
if (new_page && PageTransHuge(new_page)) prep_transhuge_page(new_page); return new_page; } |
742755a1d [PATCH] page migr... |
1519 |
#ifdef CONFIG_NUMA |
742755a1d [PATCH] page migr... |
1520 |
|
a49bd4d71 mm, numa: rework ... |
1521 |
static int store_status(int __user *status, int start, int value, int nr) |
742755a1d [PATCH] page migr... |
1522 |
{ |
a49bd4d71 mm, numa: rework ... |
1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 |
while (nr-- > 0) { if (put_user(value, status + start)) return -EFAULT; start++; } return 0; } static int do_move_pages_to_node(struct mm_struct *mm, struct list_head *pagelist, int node) { int err; |
a09763116 mm/mempolicy: use... |
1536 1537 1538 1539 |
struct migration_target_control mtc = { .nid = node, .gfp_mask = GFP_HIGHUSER_MOVABLE | __GFP_THISNODE, }; |
a49bd4d71 mm, numa: rework ... |
1540 |
|
a09763116 mm/mempolicy: use... |
1541 1542 |
err = migrate_pages(pagelist, alloc_migration_target, NULL, (unsigned long)&mtc, MIGRATE_SYNC, MR_SYSCALL); |
a49bd4d71 mm, numa: rework ... |
1543 1544 1545 |
if (err) putback_movable_pages(pagelist); return err; |
742755a1d [PATCH] page migr... |
1546 1547 1548 |
} /* |
a49bd4d71 mm, numa: rework ... |
1549 1550 |
* Resolves the given address to a struct page, isolates it from the LRU and * puts it to the given pagelist. |
e0153fc2c mm: move_pages: r... |
1551 1552 1553 1554 1555 |
* Returns: * errno - if the page cannot be found/isolated * 0 - when it doesn't have to be migrated because it is already on the * target node * 1 - when it has been queued |
742755a1d [PATCH] page migr... |
1556 |
*/ |
a49bd4d71 mm, numa: rework ... |
1557 1558 |
static int add_page_for_migration(struct mm_struct *mm, unsigned long addr, int node, struct list_head *pagelist, bool migrate_all) |
742755a1d [PATCH] page migr... |
1559 |
{ |
a49bd4d71 mm, numa: rework ... |
1560 1561 1562 |
struct vm_area_struct *vma; struct page *page; unsigned int follflags; |
742755a1d [PATCH] page migr... |
1563 |
int err; |
742755a1d [PATCH] page migr... |
1564 |
|
d8ed45c5d mmap locking API:... |
1565 |
mmap_read_lock(mm); |
a49bd4d71 mm, numa: rework ... |
1566 1567 1568 1569 |
err = -EFAULT; vma = find_vma(mm, addr); if (!vma || addr < vma->vm_start || !vma_migratable(vma)) goto out; |
742755a1d [PATCH] page migr... |
1570 |
|
a49bd4d71 mm, numa: rework ... |
1571 1572 |
/* FOLL_DUMP to ignore special (like zero) pages */ follflags = FOLL_GET | FOLL_DUMP; |
a49bd4d71 mm, numa: rework ... |
1573 |
page = follow_page(vma, addr, follflags); |
89f5b7da2 Reinstate ZERO_PA... |
1574 |
|
a49bd4d71 mm, numa: rework ... |
1575 1576 1577 |
err = PTR_ERR(page); if (IS_ERR(page)) goto out; |
89f5b7da2 Reinstate ZERO_PA... |
1578 |
|
a49bd4d71 mm, numa: rework ... |
1579 1580 1581 |
err = -ENOENT; if (!page) goto out; |
742755a1d [PATCH] page migr... |
1582 |
|
a49bd4d71 mm, numa: rework ... |
1583 1584 1585 |
err = 0; if (page_to_nid(page) == node) goto out_putpage; |
742755a1d [PATCH] page migr... |
1586 |
|
a49bd4d71 mm, numa: rework ... |
1587 1588 1589 |
err = -EACCES; if (page_mapcount(page) > 1 && !migrate_all) goto out_putpage; |
742755a1d [PATCH] page migr... |
1590 |
|
a49bd4d71 mm, numa: rework ... |
1591 1592 1593 |
if (PageHuge(page)) { if (PageHead(page)) { isolate_huge_page(page, pagelist); |
e0153fc2c mm: move_pages: r... |
1594 |
err = 1; |
e632a938d mm: migrate: add ... |
1595 |
} |
a49bd4d71 mm, numa: rework ... |
1596 1597 |
} else { struct page *head; |
e632a938d mm: migrate: add ... |
1598 |
|
e8db67eb0 mm: migrate: move... |
1599 1600 |
head = compound_head(page); err = isolate_lru_page(head); |
cf608ac19 mm: compaction: f... |
1601 |
if (err) |
a49bd4d71 mm, numa: rework ... |
1602 |
goto out_putpage; |
742755a1d [PATCH] page migr... |
1603 |
|
e0153fc2c mm: move_pages: r... |
1604 |
err = 1; |
a49bd4d71 mm, numa: rework ... |
1605 1606 |
list_add_tail(&head->lru, pagelist); mod_node_page_state(page_pgdat(head), |
9de4f22a6 mm: code cleanup ... |
1607 |
NR_ISOLATED_ANON + page_is_file_lru(head), |
6c357848b mm: replace hpage... |
1608 |
thp_nr_pages(head)); |
a49bd4d71 mm, numa: rework ... |
1609 1610 1611 1612 1613 1614 1615 1616 1617 |
} out_putpage: /* * Either remove the duplicate refcount from * isolate_lru_page() or drop the page ref if it was * not isolated. */ put_page(page); out: |
d8ed45c5d mmap locking API:... |
1618 |
mmap_read_unlock(mm); |
742755a1d [PATCH] page migr... |
1619 1620 |
return err; } |
7ca8783ad mm/migrate.c: wra... |
1621 1622 1623 1624 1625 |
static int move_pages_and_store_status(struct mm_struct *mm, int node, struct list_head *pagelist, int __user *status, int start, int i, unsigned long nr_pages) { int err; |
5d7ae891c mm/migrate.c: che... |
1626 1627 |
if (list_empty(pagelist)) return 0; |
7ca8783ad mm/migrate.c: wra... |
1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642 1643 |
err = do_move_pages_to_node(mm, pagelist, node); if (err) { /* * Positive err means the number of failed * pages to migrate. Since we are going to * abort and return the number of non-migrated * pages, so need to incude the rest of the * nr_pages that have not been attempted as * well. */ if (err > 0) err += nr_pages - i - 1; return err; } return store_status(status, start, node, i - start); } |
742755a1d [PATCH] page migr... |
1644 |
/* |
5e9a0f023 mm: extract do_pa... |
1645 1646 1647 |
* Migrate an array of page address onto an array of nodes and fill * the corresponding array of status. */ |
3268c63ed mm: fix move/migr... |
1648 |
static int do_pages_move(struct mm_struct *mm, nodemask_t task_nodes, |
5e9a0f023 mm: extract do_pa... |
1649 1650 1651 1652 1653 |
unsigned long nr_pages, const void __user * __user *pages, const int __user *nodes, int __user *status, int flags) { |
a49bd4d71 mm, numa: rework ... |
1654 1655 1656 1657 |
int current_node = NUMA_NO_NODE; LIST_HEAD(pagelist); int start, i; int err = 0, err1; |
35282a2de migration: only m... |
1658 1659 |
migrate_prep(); |
a49bd4d71 mm, numa: rework ... |
1660 1661 1662 1663 |
for (i = start = 0; i < nr_pages; i++) { const void __user *p; unsigned long addr; int node; |
3140a2273 mm: rework do_pag... |
1664 |
|
a49bd4d71 mm, numa: rework ... |
1665 1666 1667 1668 1669 |
err = -EFAULT; if (get_user(p, pages + i)) goto out_flush; if (get_user(node, nodes + i)) goto out_flush; |
057d33891 mm: untag user po... |
1670 |
addr = (unsigned long)untagged_addr(p); |
a49bd4d71 mm, numa: rework ... |
1671 1672 1673 1674 1675 1676 |
err = -ENODEV; if (node < 0 || node >= MAX_NUMNODES) goto out_flush; if (!node_state(node, N_MEMORY)) goto out_flush; |
5e9a0f023 mm: extract do_pa... |
1677 |
|
a49bd4d71 mm, numa: rework ... |
1678 1679 1680 1681 1682 1683 1684 1685 |
err = -EACCES; if (!node_isset(node, task_nodes)) goto out_flush; if (current_node == NUMA_NO_NODE) { current_node = node; start = i; } else if (node != current_node) { |
7ca8783ad mm/migrate.c: wra... |
1686 1687 |
err = move_pages_and_store_status(mm, current_node, &pagelist, status, start, i, nr_pages); |
a49bd4d71 mm, numa: rework ... |
1688 1689 1690 1691 |
if (err) goto out; start = i; current_node = node; |
3140a2273 mm: rework do_pag... |
1692 |
} |
a49bd4d71 mm, numa: rework ... |
1693 1694 1695 1696 1697 1698 |
/* * Errors in the page lookup or isolation are not fatal and we simply * report them via status */ err = add_page_for_migration(mm, addr, current_node, &pagelist, flags & MPOL_MF_MOVE_ALL); |
e0153fc2c mm: move_pages: r... |
1699 |
|
d08221a08 mm/migrate.c: uni... |
1700 |
if (err > 0) { |
e0153fc2c mm: move_pages: r... |
1701 1702 1703 |
/* The page is successfully queued for migration */ continue; } |
3140a2273 mm: rework do_pag... |
1704 |
|
d08221a08 mm/migrate.c: uni... |
1705 1706 1707 1708 1709 |
/* * If the page is already on the target node (!err), store the * node, otherwise, store the err. */ err = store_status(status, i, err ? : current_node, 1); |
a49bd4d71 mm, numa: rework ... |
1710 1711 |
if (err) goto out_flush; |
5e9a0f023 mm: extract do_pa... |
1712 |
|
7ca8783ad mm/migrate.c: wra... |
1713 1714 |
err = move_pages_and_store_status(mm, current_node, &pagelist, status, start, i, nr_pages); |
4afdacec2 mm/migrate.c: no ... |
1715 1716 |
if (err) goto out; |
a49bd4d71 mm, numa: rework ... |
1717 |
current_node = NUMA_NO_NODE; |
3140a2273 mm: rework do_pag... |
1718 |
} |
a49bd4d71 mm, numa: rework ... |
1719 1720 |
out_flush: /* Make sure we do not overwrite the existing error */ |
7ca8783ad mm/migrate.c: wra... |
1721 1722 |
err1 = move_pages_and_store_status(mm, current_node, &pagelist, status, start, i, nr_pages); |
dfe9aa23c mm/migrate.c: als... |
1723 |
if (err >= 0) |
a49bd4d71 mm, numa: rework ... |
1724 |
err = err1; |
5e9a0f023 mm: extract do_pa... |
1725 1726 1727 1728 1729 |
out: return err; } /* |
2f007e74b mm: don't vmalloc... |
1730 |
* Determine the nodes of an array of pages and store it in an array of status. |
742755a1d [PATCH] page migr... |
1731 |
*/ |
80bba1290 mm: no get_user/p... |
1732 1733 |
static void do_pages_stat_array(struct mm_struct *mm, unsigned long nr_pages, const void __user **pages, int *status) |
742755a1d [PATCH] page migr... |
1734 |
{ |
2f007e74b mm: don't vmalloc... |
1735 |
unsigned long i; |
2f007e74b mm: don't vmalloc... |
1736 |
|
d8ed45c5d mmap locking API:... |
1737 |
mmap_read_lock(mm); |
742755a1d [PATCH] page migr... |
1738 |
|
2f007e74b mm: don't vmalloc... |
1739 |
for (i = 0; i < nr_pages; i++) { |
80bba1290 mm: no get_user/p... |
1740 |
unsigned long addr = (unsigned long)(*pages); |
742755a1d [PATCH] page migr... |
1741 1742 |
struct vm_area_struct *vma; struct page *page; |
c095adbc2 mm: Don't touch u... |
1743 |
int err = -EFAULT; |
2f007e74b mm: don't vmalloc... |
1744 1745 |
vma = find_vma(mm, addr); |
70384dc6d mm: fix error rep... |
1746 |
if (!vma || addr < vma->vm_start) |
742755a1d [PATCH] page migr... |
1747 |
goto set_status; |
d899844e9 mm: fix status co... |
1748 1749 |
/* FOLL_DUMP to ignore special (like zero) pages */ page = follow_page(vma, addr, FOLL_DUMP); |
89f5b7da2 Reinstate ZERO_PA... |
1750 1751 1752 1753 |
err = PTR_ERR(page); if (IS_ERR(page)) goto set_status; |
d899844e9 mm: fix status co... |
1754 |
err = page ? page_to_nid(page) : -ENOENT; |
742755a1d [PATCH] page migr... |
1755 |
set_status: |
80bba1290 mm: no get_user/p... |
1756 1757 1758 1759 1760 |
*status = err; pages++; status++; } |
d8ed45c5d mmap locking API:... |
1761 |
mmap_read_unlock(mm); |
80bba1290 mm: no get_user/p... |
1762 1763 1764 1765 1766 1767 1768 1769 1770 1771 1772 1773 1774 |
} /* * Determine the nodes of a user array of pages and store it in * a user array of status. */ static int do_pages_stat(struct mm_struct *mm, unsigned long nr_pages, const void __user * __user *pages, int __user *status) { #define DO_PAGES_STAT_CHUNK_NR 16 const void __user *chunk_pages[DO_PAGES_STAT_CHUNK_NR]; int chunk_status[DO_PAGES_STAT_CHUNK_NR]; |
80bba1290 mm: no get_user/p... |
1775 |
|
87b8d1ade mm: Make copy_fro... |
1776 1777 |
while (nr_pages) { unsigned long chunk_nr; |
80bba1290 mm: no get_user/p... |
1778 |
|
87b8d1ade mm: Make copy_fro... |
1779 1780 1781 1782 1783 1784 |
chunk_nr = nr_pages; if (chunk_nr > DO_PAGES_STAT_CHUNK_NR) chunk_nr = DO_PAGES_STAT_CHUNK_NR; if (copy_from_user(chunk_pages, pages, chunk_nr * sizeof(*chunk_pages))) break; |
80bba1290 mm: no get_user/p... |
1785 1786 |
do_pages_stat_array(mm, chunk_nr, chunk_pages, chunk_status); |
87b8d1ade mm: Make copy_fro... |
1787 1788 |
if (copy_to_user(status, chunk_status, chunk_nr * sizeof(*status))) break; |
742755a1d [PATCH] page migr... |
1789 |
|
87b8d1ade mm: Make copy_fro... |
1790 1791 1792 1793 1794 |
pages += chunk_nr; status += chunk_nr; nr_pages -= chunk_nr; } return nr_pages ? -EFAULT : 0; |
742755a1d [PATCH] page migr... |
1795 |
} |
4dc200cee mm/migrate: avoid... |
1796 |
static struct mm_struct *find_mm_struct(pid_t pid, nodemask_t *mem_nodes) |
742755a1d [PATCH] page migr... |
1797 |
{ |
742755a1d [PATCH] page migr... |
1798 |
struct task_struct *task; |
742755a1d [PATCH] page migr... |
1799 |
struct mm_struct *mm; |
742755a1d [PATCH] page migr... |
1800 |
|
4dc200cee mm/migrate: avoid... |
1801 1802 1803 1804 1805 1806 1807 1808 1809 |
/* * There is no need to check if current process has the right to modify * the specified process when they are same. */ if (!pid) { mmget(current->mm); *mem_nodes = cpuset_mems_allowed(current); return current->mm; } |
742755a1d [PATCH] page migr... |
1810 1811 |
/* Find the mm_struct */ |
a879bf582 mm: grab rcu read... |
1812 |
rcu_read_lock(); |
4dc200cee mm/migrate: avoid... |
1813 |
task = find_task_by_vpid(pid); |
742755a1d [PATCH] page migr... |
1814 |
if (!task) { |
a879bf582 mm: grab rcu read... |
1815 |
rcu_read_unlock(); |
4dc200cee mm/migrate: avoid... |
1816 |
return ERR_PTR(-ESRCH); |
742755a1d [PATCH] page migr... |
1817 |
} |
3268c63ed mm: fix move/migr... |
1818 |
get_task_struct(task); |
742755a1d [PATCH] page migr... |
1819 1820 1821 |
/* * Check if this process has the right to modify the specified |
197e7e521 Sanitize 'move_pa... |
1822 |
* process. Use the regular "ptrace_may_access()" checks. |
742755a1d [PATCH] page migr... |
1823 |
*/ |
197e7e521 Sanitize 'move_pa... |
1824 |
if (!ptrace_may_access(task, PTRACE_MODE_READ_REALCREDS)) { |
c69e8d9c0 CRED: Use RCU to ... |
1825 |
rcu_read_unlock(); |
4dc200cee mm/migrate: avoid... |
1826 |
mm = ERR_PTR(-EPERM); |
5e9a0f023 mm: extract do_pa... |
1827 |
goto out; |
742755a1d [PATCH] page migr... |
1828 |
} |
c69e8d9c0 CRED: Use RCU to ... |
1829 |
rcu_read_unlock(); |
742755a1d [PATCH] page migr... |
1830 |
|
4dc200cee mm/migrate: avoid... |
1831 1832 |
mm = ERR_PTR(security_task_movememory(task)); if (IS_ERR(mm)) |
5e9a0f023 mm: extract do_pa... |
1833 |
goto out; |
4dc200cee mm/migrate: avoid... |
1834 |
*mem_nodes = cpuset_mems_allowed(task); |
3268c63ed mm: fix move/migr... |
1835 |
mm = get_task_mm(task); |
4dc200cee mm/migrate: avoid... |
1836 |
out: |
3268c63ed mm: fix move/migr... |
1837 |
put_task_struct(task); |
6e8b09eaf mm: fix NULL ptr ... |
1838 |
if (!mm) |
4dc200cee mm/migrate: avoid... |
1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 |
mm = ERR_PTR(-EINVAL); return mm; } /* * Move a list of pages in the address space of the currently executing * process. */ static int kernel_move_pages(pid_t pid, unsigned long nr_pages, const void __user * __user *pages, const int __user *nodes, int __user *status, int flags) { struct mm_struct *mm; int err; nodemask_t task_nodes; /* Check flags */ if (flags & ~(MPOL_MF_MOVE|MPOL_MF_MOVE_ALL)) |
6e8b09eaf mm: fix NULL ptr ... |
1858 |
return -EINVAL; |
4dc200cee mm/migrate: avoid... |
1859 1860 1861 1862 1863 1864 |
if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE)) return -EPERM; mm = find_mm_struct(pid, &task_nodes); if (IS_ERR(mm)) return PTR_ERR(mm); |
6e8b09eaf mm: fix NULL ptr ... |
1865 1866 1867 1868 1869 |
if (nodes) err = do_pages_move(mm, task_nodes, nr_pages, pages, nodes, status, flags); else err = do_pages_stat(mm, nr_pages, pages, status); |
742755a1d [PATCH] page migr... |
1870 |
|
742755a1d [PATCH] page migr... |
1871 1872 1873 |
mmput(mm); return err; } |
742755a1d [PATCH] page migr... |
1874 |
|
7addf4438 mm: add kernel_mo... |
1875 1876 1877 1878 1879 1880 1881 1882 1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 |
SYSCALL_DEFINE6(move_pages, pid_t, pid, unsigned long, nr_pages, const void __user * __user *, pages, const int __user *, nodes, int __user *, status, int, flags) { return kernel_move_pages(pid, nr_pages, pages, nodes, status, flags); } #ifdef CONFIG_COMPAT COMPAT_SYSCALL_DEFINE6(move_pages, pid_t, pid, compat_ulong_t, nr_pages, compat_uptr_t __user *, pages32, const int __user *, nodes, int __user *, status, int, flags) { const void __user * __user *pages; int i; pages = compat_alloc_user_space(nr_pages * sizeof(void *)); for (i = 0; i < nr_pages; i++) { compat_uptr_t p; if (get_user(p, pages32 + i) || put_user(compat_ptr(p), pages + i)) return -EFAULT; } return kernel_move_pages(pid, nr_pages, pages, nodes, status, flags); } #endif /* CONFIG_COMPAT */ |
7039e1dbe mm: migrate: Intr... |
1904 1905 1906 1907 1908 1909 |
#ifdef CONFIG_NUMA_BALANCING /* * Returns true if this is a safe migration target node for misplaced NUMA * pages. Currently it only checks the watermarks which crude */ static bool migrate_balanced_pgdat(struct pglist_data *pgdat, |
3abef4e6c mm: numa: take TH... |
1910 |
unsigned long nr_migrate_pages) |
7039e1dbe mm: migrate: Intr... |
1911 1912 |
{ int z; |
599d0c954 mm, vmscan: move ... |
1913 |
|
7039e1dbe mm: migrate: Intr... |
1914 1915 1916 1917 1918 |
for (z = pgdat->nr_zones - 1; z >= 0; z--) { struct zone *zone = pgdat->node_zones + z; if (!populated_zone(zone)) continue; |
7039e1dbe mm: migrate: Intr... |
1919 1920 1921 1922 |
/* Avoid waking kswapd by allocating pages_to_migrate pages. */ if (!zone_watermark_ok(zone, 0, high_wmark_pages(zone) + nr_migrate_pages, |
bfe9d006c autonuma: fix wat... |
1923 |
ZONE_MOVABLE, 0)) |
7039e1dbe mm: migrate: Intr... |
1924 1925 1926 1927 1928 1929 1930 |
continue; return true; } return false; } static struct page *alloc_misplaced_dst_page(struct page *page, |
666feb21a mm, migrate: remo... |
1931 |
unsigned long data) |
7039e1dbe mm: migrate: Intr... |
1932 1933 1934 |
{ int nid = (int) data; struct page *newpage; |
96db800f5 mm: rename alloc_... |
1935 |
newpage = __alloc_pages_node(nid, |
e97ca8e5b mm: fix GFP_THISN... |
1936 1937 1938 |
(GFP_HIGHUSER_MOVABLE | __GFP_THISNODE | __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN) & |
8479eba77 mm: numa: quickly... |
1939 |
~__GFP_RECLAIM, 0); |
bac0382c6 mm: numa: migrate... |
1940 |
|
7039e1dbe mm: migrate: Intr... |
1941 1942 |
return newpage; } |
1c30e0177 mm: numa: make NU... |
1943 |
static int numamigrate_isolate_page(pg_data_t *pgdat, struct page *page) |
b32967ff1 mm: numa: Add THP... |
1944 |
{ |
340ef3902 mm: numa: cleanup... |
1945 |
int page_lru; |
a8f607721 mm: numa: Rate li... |
1946 |
|
309381fea mm: dump page whe... |
1947 |
VM_BUG_ON_PAGE(compound_order(page) && !PageTransHuge(page), page); |
3abef4e6c mm: numa: take TH... |
1948 |
|
7039e1dbe mm: migrate: Intr... |
1949 |
/* Avoid migrating to a node that is nearly full */ |
d8c6546b1 mm: introduce com... |
1950 |
if (!migrate_balanced_pgdat(pgdat, compound_nr(page))) |
340ef3902 mm: numa: cleanup... |
1951 |
return 0; |
7039e1dbe mm: migrate: Intr... |
1952 |
|
340ef3902 mm: numa: cleanup... |
1953 1954 |
if (isolate_lru_page(page)) return 0; |
7039e1dbe mm: migrate: Intr... |
1955 |
|
340ef3902 mm: numa: cleanup... |
1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 |
/* * migrate_misplaced_transhuge_page() skips page migration's usual * check on page_count(), so we must do it here, now that the page * has been isolated: a GUP pin, or any other pin, prevents migration. * The expected page count is 3: 1 for page's mapcount and 1 for the * caller's pin and 1 for the reference taken by isolate_lru_page(). */ if (PageTransHuge(page) && page_count(page) != 3) { putback_lru_page(page); return 0; |
7039e1dbe mm: migrate: Intr... |
1966 |
} |
9de4f22a6 mm: code cleanup ... |
1967 |
page_lru = page_is_file_lru(page); |
599d0c954 mm, vmscan: move ... |
1968 |
mod_node_page_state(page_pgdat(page), NR_ISOLATED_ANON + page_lru, |
6c357848b mm: replace hpage... |
1969 |
thp_nr_pages(page)); |
340ef3902 mm: numa: cleanup... |
1970 |
|
149c33e1c mm: migrate: Drop... |
1971 |
/* |
340ef3902 mm: numa: cleanup... |
1972 1973 1974 |
* Isolating the page has taken another reference, so the * caller's reference can be safely dropped without the page * disappearing underneath us during migration. |
149c33e1c mm: migrate: Drop... |
1975 1976 |
*/ put_page(page); |
340ef3902 mm: numa: cleanup... |
1977 |
return 1; |
b32967ff1 mm: numa: Add THP... |
1978 |
} |
de466bd62 mm: numa: avoid u... |
1979 1980 1981 1982 1983 |
bool pmd_trans_migrating(pmd_t pmd) { struct page *page = pmd_page(pmd); return PageLocked(page); } |
b32967ff1 mm: numa: Add THP... |
1984 1985 1986 1987 1988 |
/* * Attempt to migrate a misplaced page to the specified destination * node. Caller is expected to have an elevated reference count on * the page that will be dropped by this function before returning. */ |
1bc115d87 mm: numa: Scan pa... |
1989 1990 |
int migrate_misplaced_page(struct page *page, struct vm_area_struct *vma, int node) |
b32967ff1 mm: numa: Add THP... |
1991 1992 |
{ pg_data_t *pgdat = NODE_DATA(node); |
340ef3902 mm: numa: cleanup... |
1993 |
int isolated; |
b32967ff1 mm: numa: Add THP... |
1994 1995 1996 1997 |
int nr_remaining; LIST_HEAD(migratepages); /* |
1bc115d87 mm: numa: Scan pa... |
1998 1999 |
* Don't migrate file pages that are mapped in multiple processes * with execute permissions as they are probably shared libraries. |
b32967ff1 mm: numa: Add THP... |
2000 |
*/ |
9de4f22a6 mm: code cleanup ... |
2001 |
if (page_mapcount(page) != 1 && page_is_file_lru(page) && |
1bc115d87 mm: numa: Scan pa... |
2002 |
(vma->vm_flags & VM_EXEC)) |
b32967ff1 mm: numa: Add THP... |
2003 |
goto out; |
b32967ff1 mm: numa: Add THP... |
2004 2005 |
/* |
09a913a7a sched/numa: avoid... |
2006 2007 2008 |
* Also do not migrate dirty pages as not all filesystems can move * dirty pages in MIGRATE_ASYNC mode which is a waste of cycles. */ |
9de4f22a6 mm: code cleanup ... |
2009 |
if (page_is_file_lru(page) && PageDirty(page)) |
09a913a7a sched/numa: avoid... |
2010 |
goto out; |
b32967ff1 mm: numa: Add THP... |
2011 2012 2013 2014 2015 |
isolated = numamigrate_isolate_page(pgdat, page); if (!isolated) goto out; list_add(&page->lru, &migratepages); |
9c620e2bc mm: remove offlin... |
2016 |
nr_remaining = migrate_pages(&migratepages, alloc_misplaced_dst_page, |
68711a746 mm, migration: ad... |
2017 2018 |
NULL, node, MIGRATE_ASYNC, MR_NUMA_MISPLACED); |
b32967ff1 mm: numa: Add THP... |
2019 |
if (nr_remaining) { |
59c82b70d mm/migrate: remov... |
2020 2021 |
if (!list_empty(&migratepages)) { list_del(&page->lru); |
599d0c954 mm, vmscan: move ... |
2022 |
dec_node_page_state(page, NR_ISOLATED_ANON + |
9de4f22a6 mm: code cleanup ... |
2023 |
page_is_file_lru(page)); |
59c82b70d mm/migrate: remov... |
2024 2025 |
putback_lru_page(page); } |
b32967ff1 mm: numa: Add THP... |
2026 2027 2028 |
isolated = 0; } else count_vm_numa_event(NUMA_PAGE_MIGRATE); |
7039e1dbe mm: migrate: Intr... |
2029 |
BUG_ON(!list_empty(&migratepages)); |
7039e1dbe mm: migrate: Intr... |
2030 |
return isolated; |
340ef3902 mm: numa: cleanup... |
2031 2032 2033 2034 |
out: put_page(page); return 0; |
7039e1dbe mm: migrate: Intr... |
2035 |
} |
220018d38 mm: numa: Add THP... |
2036 |
#endif /* CONFIG_NUMA_BALANCING */ |
b32967ff1 mm: numa: Add THP... |
2037 |
|
220018d38 mm: numa: Add THP... |
2038 |
#if defined(CONFIG_NUMA_BALANCING) && defined(CONFIG_TRANSPARENT_HUGEPAGE) |
340ef3902 mm: numa: cleanup... |
2039 2040 2041 2042 |
/* * Migrates a THP to a given target node. page must be locked and is unlocked * before returning. */ |
b32967ff1 mm: numa: Add THP... |
2043 2044 2045 2046 2047 2048 |
int migrate_misplaced_transhuge_page(struct mm_struct *mm, struct vm_area_struct *vma, pmd_t *pmd, pmd_t entry, unsigned long address, struct page *page, int node) { |
c4088ebdc mm: convert the r... |
2049 |
spinlock_t *ptl; |
b32967ff1 mm: numa: Add THP... |
2050 2051 2052 |
pg_data_t *pgdat = NODE_DATA(node); int isolated = 0; struct page *new_page = NULL; |
9de4f22a6 mm: code cleanup ... |
2053 |
int page_lru = page_is_file_lru(page); |
7066f0f93 mm: thp: fix mmu_... |
2054 |
unsigned long start = address & HPAGE_PMD_MASK; |
b32967ff1 mm: numa: Add THP... |
2055 |
|
b32967ff1 mm: numa: Add THP... |
2056 |
new_page = alloc_pages_node(node, |
251603549 mm, thp: remove _... |
2057 |
(GFP_TRANSHUGE_LIGHT | __GFP_THISNODE), |
e97ca8e5b mm: fix GFP_THISN... |
2058 |
HPAGE_PMD_ORDER); |
340ef3902 mm: numa: cleanup... |
2059 2060 |
if (!new_page) goto out_fail; |
9a982250f thp: introduce de... |
2061 |
prep_transhuge_page(new_page); |
340ef3902 mm: numa: cleanup... |
2062 |
|
b32967ff1 mm: numa: Add THP... |
2063 |
isolated = numamigrate_isolate_page(pgdat, page); |
340ef3902 mm: numa: cleanup... |
2064 |
if (!isolated) { |
b32967ff1 mm: numa: Add THP... |
2065 |
put_page(new_page); |
340ef3902 mm: numa: cleanup... |
2066 |
goto out_fail; |
b32967ff1 mm: numa: Add THP... |
2067 |
} |
b0943d61b mm: numa: defer T... |
2068 |
|
b32967ff1 mm: numa: Add THP... |
2069 |
/* Prepare a page as a migration target */ |
48c935ad8 page-flags: defin... |
2070 |
__SetPageLocked(new_page); |
d44d363f6 mm: don't assume ... |
2071 2072 |
if (PageSwapBacked(page)) __SetPageSwapBacked(new_page); |
b32967ff1 mm: numa: Add THP... |
2073 2074 2075 2076 |
/* anon mapping, we can simply copy page->mapping to the new page: */ new_page->mapping = page->mapping; new_page->index = page->index; |
7eef5f97c mm: thp: relocate... |
2077 2078 |
/* flush the cache before copying using the kernel virtual address */ flush_cache_range(vma, start, start + HPAGE_PMD_SIZE); |
b32967ff1 mm: numa: Add THP... |
2079 2080 2081 2082 |
migrate_page_copy(new_page, page); WARN_ON(PageLRU(new_page)); /* Recheck the target PMD */ |
c4088ebdc mm: convert the r... |
2083 |
ptl = pmd_lock(mm, pmd); |
f4e177d12 mm/migrate.c: sta... |
2084 |
if (unlikely(!pmd_same(*pmd, entry) || !page_ref_freeze(page, 2))) { |
c4088ebdc mm: convert the r... |
2085 |
spin_unlock(ptl); |
b32967ff1 mm: numa: Add THP... |
2086 2087 2088 2089 2090 2091 |
/* Reverse changes made by migrate_page_copy() */ if (TestClearPageActive(new_page)) SetPageActive(page); if (TestClearPageUnevictable(new_page)) SetPageUnevictable(page); |
b32967ff1 mm: numa: Add THP... |
2092 2093 2094 |
unlock_page(new_page); put_page(new_page); /* Free it */ |
a54a407fb mm: Close races b... |
2095 2096 |
/* Retake the callers reference and putback on LRU */ get_page(page); |
b32967ff1 mm: numa: Add THP... |
2097 |
putback_lru_page(page); |
599d0c954 mm, vmscan: move ... |
2098 |
mod_node_page_state(page_pgdat(page), |
a54a407fb mm: Close races b... |
2099 |
NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR); |
eb4489f69 mm: numa: avoid u... |
2100 2101 |
goto out_unlock; |
b32967ff1 mm: numa: Add THP... |
2102 |
} |
101024596 mm: introduce do_... |
2103 |
entry = mk_huge_pmd(new_page, vma->vm_page_prot); |
f55e1014f Revert "mm, thp: ... |
2104 |
entry = maybe_pmd_mkwrite(pmd_mkdirty(entry), vma); |
b32967ff1 mm: numa: Add THP... |
2105 |
|
2b4847e73 mm: numa: seriali... |
2106 |
/* |
d7c339341 mm: thp: fix MADV... |
2107 2108 2109 2110 2111 2112 |
* Overwrite the old entry under pagetable lock and establish * the new PTE. Any parallel GUP will either observe the old * page blocking on the page lock, block on the page table * lock or observe the new page. The SetPageUptodate on the * new page and page_add_new_anon_rmap guarantee the copy is * visible before the pagetable update. |
2b4847e73 mm: numa: seriali... |
2113 |
*/ |
7066f0f93 mm: thp: fix mmu_... |
2114 |
page_add_anon_rmap(new_page, vma, start, true); |
d7c339341 mm: thp: fix MADV... |
2115 2116 2117 2118 2119 2120 2121 |
/* * At this point the pmd is numa/protnone (i.e. non present) and the TLB * has already been flushed globally. So no TLB can be currently * caching this non present pmd mapping. There's no need to clear the * pmd before doing set_pmd_at(), nor to flush the TLB after * set_pmd_at(). Clearing the pmd here would introduce a race * condition against MADV_DONTNEED, because MADV_DONTNEED only holds the |
c1e8d7c6a mmap locking API:... |
2122 |
* mmap_lock for reading. If the pmd is set to NULL at any given time, |
d7c339341 mm: thp: fix MADV... |
2123 2124 2125 |
* MADV_DONTNEED won't wait on the pmd lock and it'll skip clearing this * pmd. */ |
7066f0f93 mm: thp: fix mmu_... |
2126 |
set_pmd_at(mm, start, pmd, entry); |
ce4a9cc57 mm,numa: fix upda... |
2127 |
update_mmu_cache_pmd(vma, address, &entry); |
2b4847e73 mm: numa: seriali... |
2128 |
|
f4e177d12 mm/migrate.c: sta... |
2129 |
page_ref_unfreeze(page, 2); |
51afb12ba mm: page migratio... |
2130 |
mlock_migrate_page(new_page, page); |
d281ee614 rmap: add argumen... |
2131 |
page_remove_rmap(page, true); |
7cd12b4ab mm, page_owner: t... |
2132 |
set_page_owner_migrate_reason(new_page, MR_NUMA_MISPLACED); |
2b4847e73 mm: numa: seriali... |
2133 |
|
c4088ebdc mm: convert the r... |
2134 |
spin_unlock(ptl); |
b32967ff1 mm: numa: Add THP... |
2135 |
|
11de9927f mm: numa: add mig... |
2136 2137 2138 |
/* Take an "isolate" reference and put new page on the LRU. */ get_page(new_page); putback_lru_page(new_page); |
b32967ff1 mm: numa: Add THP... |
2139 2140 2141 2142 2143 2144 2145 |
unlock_page(new_page); unlock_page(page); put_page(page); /* Drop the rmap reference */ put_page(page); /* Drop the LRU isolation reference */ count_vm_events(PGMIGRATE_SUCCESS, HPAGE_PMD_NR); count_vm_numa_events(NUMA_PAGE_MIGRATE, HPAGE_PMD_NR); |
599d0c954 mm, vmscan: move ... |
2146 |
mod_node_page_state(page_pgdat(page), |
b32967ff1 mm: numa: Add THP... |
2147 2148 2149 |
NR_ISOLATED_ANON + page_lru, -HPAGE_PMD_NR); return isolated; |
340ef3902 mm: numa: cleanup... |
2150 2151 |
out_fail: count_vm_events(PGMIGRATE_FAIL, HPAGE_PMD_NR); |
2b4847e73 mm: numa: seriali... |
2152 2153 |
ptl = pmd_lock(mm, pmd); if (pmd_same(*pmd, entry)) { |
4d9424669 mm: convert p[te|... |
2154 |
entry = pmd_modify(entry, vma->vm_page_prot); |
7066f0f93 mm: thp: fix mmu_... |
2155 |
set_pmd_at(mm, start, pmd, entry); |
2b4847e73 mm: numa: seriali... |
2156 2157 2158 |
update_mmu_cache_pmd(vma, address, &entry); } spin_unlock(ptl); |
a54a407fb mm: Close races b... |
2159 |
|
eb4489f69 mm: numa: avoid u... |
2160 |
out_unlock: |
340ef3902 mm: numa: cleanup... |
2161 |
unlock_page(page); |
b32967ff1 mm: numa: Add THP... |
2162 |
put_page(page); |
b32967ff1 mm: numa: Add THP... |
2163 2164 |
return 0; } |
7039e1dbe mm: migrate: Intr... |
2165 2166 2167 |
#endif /* CONFIG_NUMA_BALANCING */ #endif /* CONFIG_NUMA */ |
8763cb45a mm/migrate: new m... |
2168 |
|
9b2ed9cb9 mm: remove CONFIG... |
2169 |
#ifdef CONFIG_DEVICE_PRIVATE |
8763cb45a mm/migrate: new m... |
2170 2171 |
static int migrate_vma_collect_hole(unsigned long start, unsigned long end, |
b7a16c7ad mm: pagewalk: add... |
2172 |
__always_unused int depth, |
8763cb45a mm/migrate: new m... |
2173 2174 2175 2176 |
struct mm_walk *walk) { struct migrate_vma *migrate = walk->private; unsigned long addr; |
0744f2807 mm/migrate: optim... |
2177 2178 2179 2180 2181 2182 2183 2184 2185 |
/* Only allow populating anonymous memory. */ if (!vma_is_anonymous(walk->vma)) { for (addr = start; addr < end; addr += PAGE_SIZE) { migrate->src[migrate->npages] = 0; migrate->dst[migrate->npages] = 0; migrate->npages++; } return 0; } |
872ea7075 mm/migrate: remov... |
2186 |
for (addr = start; addr < end; addr += PAGE_SIZE) { |
e20d103b6 mm/migrate: fix i... |
2187 |
migrate->src[migrate->npages] = MIGRATE_PFN_MIGRATE; |
8315ada7f mm/migrate: allow... |
2188 |
migrate->dst[migrate->npages] = 0; |
e20d103b6 mm/migrate: fix i... |
2189 |
migrate->npages++; |
8315ada7f mm/migrate: allow... |
2190 2191 2192 2193 2194 2195 2196 2197 2198 2199 2200 2201 |
migrate->cpages++; } return 0; } static int migrate_vma_collect_skip(unsigned long start, unsigned long end, struct mm_walk *walk) { struct migrate_vma *migrate = walk->private; unsigned long addr; |
872ea7075 mm/migrate: remov... |
2202 |
for (addr = start; addr < end; addr += PAGE_SIZE) { |
8763cb45a mm/migrate: new m... |
2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 2216 2217 |
migrate->dst[migrate->npages] = 0; migrate->src[migrate->npages++] = 0; } return 0; } static int migrate_vma_collect_pmd(pmd_t *pmdp, unsigned long start, unsigned long end, struct mm_walk *walk) { struct migrate_vma *migrate = walk->private; struct vm_area_struct *vma = walk->vma; struct mm_struct *mm = vma->vm_mm; |
8c3328f1f mm/migrate: migra... |
2218 |
unsigned long addr = start, unmapped = 0; |
8763cb45a mm/migrate: new m... |
2219 2220 2221 2222 2223 |
spinlock_t *ptl; pte_t *ptep; again: if (pmd_none(*pmdp)) |
b7a16c7ad mm: pagewalk: add... |
2224 |
return migrate_vma_collect_hole(start, end, -1, walk); |
8763cb45a mm/migrate: new m... |
2225 2226 2227 2228 2229 2230 2231 2232 2233 2234 2235 2236 2237 2238 2239 |
if (pmd_trans_huge(*pmdp)) { struct page *page; ptl = pmd_lock(mm, pmdp); if (unlikely(!pmd_trans_huge(*pmdp))) { spin_unlock(ptl); goto again; } page = pmd_page(*pmdp); if (is_huge_zero_page(page)) { spin_unlock(ptl); split_huge_pmd(vma, pmdp, addr); if (pmd_trans_unstable(pmdp)) |
8315ada7f mm/migrate: allow... |
2240 |
return migrate_vma_collect_skip(start, end, |
8763cb45a mm/migrate: new m... |
2241 2242 2243 2244 2245 2246 2247 |
walk); } else { int ret; get_page(page); spin_unlock(ptl); if (unlikely(!trylock_page(page))) |
8315ada7f mm/migrate: allow... |
2248 |
return migrate_vma_collect_skip(start, end, |
8763cb45a mm/migrate: new m... |
2249 2250 2251 2252 |
walk); ret = split_huge_page(page); unlock_page(page); put_page(page); |
8315ada7f mm/migrate: allow... |
2253 2254 2255 2256 |
if (ret) return migrate_vma_collect_skip(start, end, walk); if (pmd_none(*pmdp)) |
b7a16c7ad mm: pagewalk: add... |
2257 |
return migrate_vma_collect_hole(start, end, -1, |
8763cb45a mm/migrate: new m... |
2258 2259 2260 2261 2262 |
walk); } } if (unlikely(pmd_bad(*pmdp))) |
8315ada7f mm/migrate: allow... |
2263 |
return migrate_vma_collect_skip(start, end, walk); |
8763cb45a mm/migrate: new m... |
2264 2265 |
ptep = pte_offset_map_lock(mm, pmdp, addr, &ptl); |
8c3328f1f mm/migrate: migra... |
2266 |
arch_enter_lazy_mmu_mode(); |
8763cb45a mm/migrate: new m... |
2267 |
for (; addr < end; addr += PAGE_SIZE, ptep++) { |
800bb1c8d mm: handle multip... |
2268 |
unsigned long mpfn = 0, pfn; |
8763cb45a mm/migrate: new m... |
2269 |
struct page *page; |
8c3328f1f mm/migrate: migra... |
2270 |
swp_entry_t entry; |
8763cb45a mm/migrate: new m... |
2271 2272 2273 |
pte_t pte; pte = *ptep; |
8763cb45a mm/migrate: new m... |
2274 |
|
a5430dda8 mm/migrate: suppo... |
2275 |
if (pte_none(pte)) { |
0744f2807 mm/migrate: optim... |
2276 2277 2278 2279 |
if (vma_is_anonymous(vma)) { mpfn = MIGRATE_PFN_MIGRATE; migrate->cpages++; } |
8763cb45a mm/migrate: new m... |
2280 2281 |
goto next; } |
a5430dda8 mm/migrate: suppo... |
2282 |
if (!pte_present(pte)) { |
a5430dda8 mm/migrate: suppo... |
2283 2284 2285 2286 2287 2288 2289 2290 2291 2292 |
/* * Only care about unaddressable device page special * page table entry. Other special swap entries are not * migratable, and we ignore regular swapped page. */ entry = pte_to_swp_entry(pte); if (!is_device_private_entry(entry)) goto next; page = device_private_entry_to_page(entry); |
5143192cd mm/migrate: add a... |
2293 2294 2295 |
if (!(migrate->flags & MIGRATE_VMA_SELECT_DEVICE_PRIVATE) || page->pgmap->owner != migrate->pgmap_owner) |
800bb1c8d mm: handle multip... |
2296 |
goto next; |
06d462beb mm: remove the un... |
2297 2298 |
mpfn = migrate_pfn(page_to_pfn(page)) | MIGRATE_PFN_MIGRATE; |
a5430dda8 mm/migrate: suppo... |
2299 2300 2301 |
if (is_write_device_private_entry(entry)) mpfn |= MIGRATE_PFN_WRITE; } else { |
5143192cd mm/migrate: add a... |
2302 |
if (!(migrate->flags & MIGRATE_VMA_SELECT_SYSTEM)) |
800bb1c8d mm: handle multip... |
2303 |
goto next; |
276f756d7 mm/migrate.c: cle... |
2304 |
pfn = pte_pfn(pte); |
8315ada7f mm/migrate: allow... |
2305 2306 2307 |
if (is_zero_pfn(pfn)) { mpfn = MIGRATE_PFN_MIGRATE; migrate->cpages++; |
8315ada7f mm/migrate: allow... |
2308 2309 |
goto next; } |
25b2995a3 mm: remove MEMORY... |
2310 |
page = vm_normal_page(migrate->vma, addr, pte); |
a5430dda8 mm/migrate: suppo... |
2311 2312 2313 |
mpfn = migrate_pfn(pfn) | MIGRATE_PFN_MIGRATE; mpfn |= pte_write(pte) ? MIGRATE_PFN_WRITE : 0; } |
8763cb45a mm/migrate: new m... |
2314 |
/* FIXME support THP */ |
8763cb45a mm/migrate: new m... |
2315 |
if (!page || !page->mapping || PageTransCompound(page)) { |
276f756d7 mm/migrate.c: cle... |
2316 |
mpfn = 0; |
8763cb45a mm/migrate: new m... |
2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 |
goto next; } /* * By getting a reference on the page we pin it and that blocks * any kind of migration. Side effect is that it "freezes" the * pte. * * We drop this reference after isolating the page from the lru * for non device page (device page are not on the lru and thus * can't be dropped from it). */ get_page(page); migrate->cpages++; |
8763cb45a mm/migrate: new m... |
2331 |
|
8c3328f1f mm/migrate: migra... |
2332 2333 2334 2335 2336 2337 2338 2339 2340 2341 2342 2343 |
/* * Optimize for the common case where page is only mapped once * in one process. If we can lock the page, then we can safely * set up a special migration page table entry now. */ if (trylock_page(page)) { pte_t swp_pte; mpfn |= MIGRATE_PFN_LOCKED; ptep_get_and_clear(mm, addr, ptep); /* Setup special migration page table entry */ |
07707125a mm/migrate: prope... |
2344 2345 |
entry = make_migration_entry(page, mpfn & MIGRATE_PFN_WRITE); |
8c3328f1f mm/migrate: migra... |
2346 |
swp_pte = swp_entry_to_pte(entry); |
ad7df764b mm/rmap: fixup co... |
2347 2348 2349 2350 2351 2352 2353 2354 2355 2356 2357 |
if (pte_present(pte)) { if (pte_soft_dirty(pte)) swp_pte = pte_swp_mksoft_dirty(swp_pte); if (pte_uffd_wp(pte)) swp_pte = pte_swp_mkuffd_wp(swp_pte); } else { if (pte_swp_soft_dirty(pte)) swp_pte = pte_swp_mksoft_dirty(swp_pte); if (pte_swp_uffd_wp(pte)) swp_pte = pte_swp_mkuffd_wp(swp_pte); } |
8c3328f1f mm/migrate: migra... |
2358 2359 2360 2361 2362 2363 2364 2365 2366 |
set_pte_at(mm, addr, ptep, swp_pte); /* * This is like regular unmap: we remove the rmap and * drop page refcount. Page won't be freed, as we took * a reference just above. */ page_remove_rmap(page, false); put_page(page); |
a5430dda8 mm/migrate: suppo... |
2367 2368 2369 |
if (pte_present(pte)) unmapped++; |
8c3328f1f mm/migrate: migra... |
2370 |
} |
8763cb45a mm/migrate: new m... |
2371 |
next: |
a5430dda8 mm/migrate: suppo... |
2372 |
migrate->dst[migrate->npages] = 0; |
8763cb45a mm/migrate: new m... |
2373 2374 |
migrate->src[migrate->npages++] = mpfn; } |
8c3328f1f mm/migrate: migra... |
2375 |
arch_leave_lazy_mmu_mode(); |
8763cb45a mm/migrate: new m... |
2376 |
pte_unmap_unlock(ptep - 1, ptl); |
8c3328f1f mm/migrate: migra... |
2377 2378 2379 |
/* Only flush the TLB if we actually modified any entries */ if (unmapped) flush_tlb_range(walk->vma, start, end); |
8763cb45a mm/migrate: new m... |
2380 2381 |
return 0; } |
7b86ac337 pagewalk: separat... |
2382 2383 2384 2385 |
static const struct mm_walk_ops migrate_vma_walk_ops = { .pmd_entry = migrate_vma_collect_pmd, .pte_hole = migrate_vma_collect_hole, }; |
8763cb45a mm/migrate: new m... |
2386 2387 2388 2389 2390 2391 2392 2393 2394 2395 |
/* * migrate_vma_collect() - collect pages over a range of virtual addresses * @migrate: migrate struct containing all migration information * * This will walk the CPU page table. For each virtual address backed by a * valid page, it updates the src array and takes a reference on the page, in * order to pin the page until we lock it and unmap it. */ static void migrate_vma_collect(struct migrate_vma *migrate) { |
ac46d4f3c mm/mmu_notifier: ... |
2396 |
struct mmu_notifier_range range; |
8763cb45a mm/migrate: new m... |
2397 |
|
998427b3a mm/notifier: add ... |
2398 2399 2400 2401 2402 |
/* * Note that the pgmap_owner is passed to the mmu notifier callback so * that the registered device driver can skip invalidating device * private page mappings that won't be migrated. */ |
c1a06df6e mm/migrate: fix m... |
2403 2404 2405 |
mmu_notifier_range_init_migrate(&range, 0, migrate->vma, migrate->vma->vm_mm, migrate->start, migrate->end, migrate->pgmap_owner); |
ac46d4f3c mm/mmu_notifier: ... |
2406 |
mmu_notifier_invalidate_range_start(&range); |
8763cb45a mm/migrate: new m... |
2407 |
|
7b86ac337 pagewalk: separat... |
2408 2409 2410 2411 |
walk_page_range(migrate->vma->vm_mm, migrate->start, migrate->end, &migrate_vma_walk_ops, migrate); mmu_notifier_invalidate_range_end(&range); |
8763cb45a mm/migrate: new m... |
2412 2413 2414 2415 2416 2417 2418 2419 2420 2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 |
migrate->end = migrate->start + (migrate->npages << PAGE_SHIFT); } /* * migrate_vma_check_page() - check if page is pinned or not * @page: struct page to check * * Pinned pages cannot be migrated. This is the same test as in * migrate_page_move_mapping(), except that here we allow migration of a * ZONE_DEVICE page. */ static bool migrate_vma_check_page(struct page *page) { /* * One extra ref because caller holds an extra reference, either from * isolate_lru_page() for a regular page, or migrate_vma_collect() for * a device page. */ int extra = 1; /* * FIXME support THP (transparent huge page), it is bit more complex to * check them than regular pages, because they can be mapped with a pmd * or with a pte (split pte mapping). */ if (PageCompound(page)) return false; |
a5430dda8 mm/migrate: suppo... |
2439 2440 2441 2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 2452 2453 |
/* Page from ZONE_DEVICE have one extra reference */ if (is_zone_device_page(page)) { /* * Private page can never be pin as they have no valid pte and * GUP will fail for those. Yet if there is a pending migration * a thread might try to wait on the pte migration entry and * will bump the page reference count. Sadly there is no way to * differentiate a regular pin from migration wait. Hence to * avoid 2 racing thread trying to migrate back to CPU to enter * infinite loop (one stoping migration because the other is * waiting on pte migration entry). We always return true here. * * FIXME proper solution is to rework migration_entry_wait() so * it does not need to take a reference on page. */ |
25b2995a3 mm: remove MEMORY... |
2454 |
return is_device_private_page(page); |
a5430dda8 mm/migrate: suppo... |
2455 |
} |
df6ad6983 mm/device-public-... |
2456 2457 2458 |
/* For file back page */ if (page_mapping(page)) extra += 1 + page_has_private(page); |
8763cb45a mm/migrate: new m... |
2459 2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 |
if ((page_count(page) - extra) > page_mapcount(page)) return false; return true; } /* * migrate_vma_prepare() - lock pages and isolate them from the lru * @migrate: migrate struct containing all migration information * * This locks pages that have been collected by migrate_vma_collect(). Once each * page is locked it is isolated from the lru (for non-device pages). Finally, * the ref taken by migrate_vma_collect() is dropped, as locked pages cannot be * migrated by concurrent kernel threads. */ static void migrate_vma_prepare(struct migrate_vma *migrate) { const unsigned long npages = migrate->npages; |
8c3328f1f mm/migrate: migra... |
2477 2478 |
const unsigned long start = migrate->start; unsigned long addr, i, restore = 0; |
8763cb45a mm/migrate: new m... |
2479 |
bool allow_drain = true; |
8763cb45a mm/migrate: new m... |
2480 2481 2482 2483 2484 |
lru_add_drain(); for (i = 0; (i < npages) && migrate->cpages; i++) { struct page *page = migrate_pfn_to_page(migrate->src[i]); |
8c3328f1f mm/migrate: migra... |
2485 |
bool remap = true; |
8763cb45a mm/migrate: new m... |
2486 2487 2488 |
if (!page) continue; |
8c3328f1f mm/migrate: migra... |
2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 2504 2505 |
if (!(migrate->src[i] & MIGRATE_PFN_LOCKED)) { /* * Because we are migrating several pages there can be * a deadlock between 2 concurrent migration where each * are waiting on each other page lock. * * Make migrate_vma() a best effort thing and backoff * for any page we can not lock right away. */ if (!trylock_page(page)) { migrate->src[i] = 0; migrate->cpages--; put_page(page); continue; } remap = false; migrate->src[i] |= MIGRATE_PFN_LOCKED; |
8763cb45a mm/migrate: new m... |
2506 |
} |
8763cb45a mm/migrate: new m... |
2507 |
|
a5430dda8 mm/migrate: suppo... |
2508 2509 2510 2511 2512 2513 2514 |
/* ZONE_DEVICE pages are not on LRU */ if (!is_zone_device_page(page)) { if (!PageLRU(page) && allow_drain) { /* Drain CPU's pagevec */ lru_add_drain_all(); allow_drain = false; } |
8763cb45a mm/migrate: new m... |
2515 |
|
a5430dda8 mm/migrate: suppo... |
2516 2517 2518 2519 2520 2521 2522 2523 2524 2525 2526 2527 |
if (isolate_lru_page(page)) { if (remap) { migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; migrate->cpages--; restore++; } else { migrate->src[i] = 0; unlock_page(page); migrate->cpages--; put_page(page); } continue; |
8c3328f1f mm/migrate: migra... |
2528 |
} |
a5430dda8 mm/migrate: suppo... |
2529 2530 2531 |
/* Drop the reference we took in collect */ put_page(page); |
8763cb45a mm/migrate: new m... |
2532 2533 2534 |
} if (!migrate_vma_check_page(page)) { |
8c3328f1f mm/migrate: migra... |
2535 2536 2537 2538 |
if (remap) { migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; migrate->cpages--; restore++; |
8763cb45a mm/migrate: new m... |
2539 |
|
a5430dda8 mm/migrate: suppo... |
2540 2541 2542 2543 |
if (!is_zone_device_page(page)) { get_page(page); putback_lru_page(page); } |
8c3328f1f mm/migrate: migra... |
2544 2545 2546 2547 |
} else { migrate->src[i] = 0; unlock_page(page); migrate->cpages--; |
a5430dda8 mm/migrate: suppo... |
2548 2549 2550 2551 |
if (!is_zone_device_page(page)) putback_lru_page(page); else put_page(page); |
8c3328f1f mm/migrate: migra... |
2552 |
} |
8763cb45a mm/migrate: new m... |
2553 2554 |
} } |
8c3328f1f mm/migrate: migra... |
2555 2556 2557 2558 2559 2560 2561 2562 2563 2564 2565 2566 2567 2568 |
for (i = 0, addr = start; i < npages && restore; i++, addr += PAGE_SIZE) { struct page *page = migrate_pfn_to_page(migrate->src[i]); if (!page || (migrate->src[i] & MIGRATE_PFN_MIGRATE)) continue; remove_migration_pte(page, migrate->vma, addr, page); migrate->src[i] = 0; unlock_page(page); put_page(page); restore--; } |
8763cb45a mm/migrate: new m... |
2569 2570 2571 2572 2573 2574 2575 2576 2577 2578 2579 2580 2581 2582 2583 |
} /* * migrate_vma_unmap() - replace page mapping with special migration pte entry * @migrate: migrate struct containing all migration information * * Replace page mapping (CPU page table pte) with a special migration pte entry * and check again if it has been pinned. Pinned pages are restored because we * cannot migrate them. * * This is the last step before we call the device driver callback to allocate * destination memory and copy contents of original page over to new page. */ static void migrate_vma_unmap(struct migrate_vma *migrate) { |
dd156e3fc mm/rmap: always d... |
2584 |
int flags = TTU_MIGRATION | TTU_IGNORE_MLOCK; |
8763cb45a mm/migrate: new m... |
2585 2586 2587 2588 2589 2590 2591 2592 2593 |
const unsigned long npages = migrate->npages; const unsigned long start = migrate->start; unsigned long addr, i, restore = 0; for (i = 0; i < npages; i++) { struct page *page = migrate_pfn_to_page(migrate->src[i]); if (!page || !(migrate->src[i] & MIGRATE_PFN_MIGRATE)) continue; |
8c3328f1f mm/migrate: migra... |
2594 2595 2596 2597 |
if (page_mapped(page)) { try_to_unmap(page, flags); if (page_mapped(page)) goto restore; |
8763cb45a mm/migrate: new m... |
2598 |
} |
8c3328f1f mm/migrate: migra... |
2599 2600 2601 2602 2603 2604 2605 2606 |
if (migrate_vma_check_page(page)) continue; restore: migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; migrate->cpages--; restore++; |
8763cb45a mm/migrate: new m... |
2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 |
} for (addr = start, i = 0; i < npages && restore; addr += PAGE_SIZE, i++) { struct page *page = migrate_pfn_to_page(migrate->src[i]); if (!page || (migrate->src[i] & MIGRATE_PFN_MIGRATE)) continue; remove_migration_ptes(page, page, false); migrate->src[i] = 0; unlock_page(page); restore--; |
a5430dda8 mm/migrate: suppo... |
2620 2621 2622 2623 |
if (is_zone_device_page(page)) put_page(page); else putback_lru_page(page); |
8763cb45a mm/migrate: new m... |
2624 2625 |
} } |
a7d1f22bb mm: turn migrate_... |
2626 2627 |
/** * migrate_vma_setup() - prepare to migrate a range of memory |
eaf444dee mm/migrate.c: del... |
2628 |
* @args: contains the vma, start, and pfns arrays for the migration |
a7d1f22bb mm: turn migrate_... |
2629 2630 2631 2632 2633 2634 2635 2636 2637 2638 2639 2640 2641 2642 2643 2644 2645 2646 2647 2648 2649 2650 2651 2652 2653 2654 2655 2656 2657 2658 2659 2660 2661 2662 2663 2664 2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 |
* * Returns: negative errno on failures, 0 when 0 or more pages were migrated * without an error. * * Prepare to migrate a range of memory virtual address range by collecting all * the pages backing each virtual address in the range, saving them inside the * src array. Then lock those pages and unmap them. Once the pages are locked * and unmapped, check whether each page is pinned or not. Pages that aren't * pinned have the MIGRATE_PFN_MIGRATE flag set (by this function) in the * corresponding src array entry. Then restores any pages that are pinned, by * remapping and unlocking those pages. * * The caller should then allocate destination memory and copy source memory to * it for all those entries (ie with MIGRATE_PFN_VALID and MIGRATE_PFN_MIGRATE * flag set). Once these are allocated and copied, the caller must update each * corresponding entry in the dst array with the pfn value of the destination * page and with the MIGRATE_PFN_VALID and MIGRATE_PFN_LOCKED flags set * (destination pages must have their struct pages locked, via lock_page()). * * Note that the caller does not have to migrate all the pages that are marked * with MIGRATE_PFN_MIGRATE flag in src array unless this is a migration from * device memory to system memory. If the caller cannot migrate a device page * back to system memory, then it must return VM_FAULT_SIGBUS, which has severe * consequences for the userspace process, so it must be avoided if at all * possible. * * For empty entries inside CPU page table (pte_none() or pmd_none() is true) we * do set MIGRATE_PFN_MIGRATE flag inside the corresponding source array thus * allowing the caller to allocate device memory for those unback virtual * address. For this the caller simply has to allocate device memory and * properly set the destination entry like for regular migration. Note that * this can still fails and thus inside the device driver must check if the * migration was successful for those entries after calling migrate_vma_pages() * just like for regular migration. * * After that, the callers must call migrate_vma_pages() to go over each entry * in the src array that has the MIGRATE_PFN_VALID and MIGRATE_PFN_MIGRATE flag * set. If the corresponding entry in dst array has MIGRATE_PFN_VALID flag set, * then migrate_vma_pages() to migrate struct page information from the source * struct page to the destination struct page. If it fails to migrate the * struct page information, then it clears the MIGRATE_PFN_MIGRATE flag in the * src array. * * At this point all successfully migrated pages have an entry in the src * array with MIGRATE_PFN_VALID and MIGRATE_PFN_MIGRATE flag set and the dst * array entry with MIGRATE_PFN_VALID flag set. * * Once migrate_vma_pages() returns the caller may inspect which pages were * successfully migrated, and which were not. Successfully migrated pages will * have the MIGRATE_PFN_MIGRATE flag set for their src array entry. * * It is safe to update device page table after migrate_vma_pages() because |
c1e8d7c6a mmap locking API:... |
2681 |
* both destination and source page are still locked, and the mmap_lock is held |
a7d1f22bb mm: turn migrate_... |
2682 2683 2684 2685 2686 2687 2688 2689 2690 2691 2692 2693 2694 2695 2696 2697 2698 2699 2700 2701 2702 2703 2704 2705 2706 2707 2708 2709 2710 2711 2712 2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 2726 2727 2728 |
* in read mode (hence no one can unmap the range being migrated). * * Once the caller is done cleaning up things and updating its page table (if it * chose to do so, this is not an obligation) it finally calls * migrate_vma_finalize() to update the CPU page table to point to new pages * for successfully migrated pages or otherwise restore the CPU page table to * point to the original source pages. */ int migrate_vma_setup(struct migrate_vma *args) { long nr_pages = (args->end - args->start) >> PAGE_SHIFT; args->start &= PAGE_MASK; args->end &= PAGE_MASK; if (!args->vma || is_vm_hugetlb_page(args->vma) || (args->vma->vm_flags & VM_SPECIAL) || vma_is_dax(args->vma)) return -EINVAL; if (nr_pages <= 0) return -EINVAL; if (args->start < args->vma->vm_start || args->start >= args->vma->vm_end) return -EINVAL; if (args->end <= args->vma->vm_start || args->end > args->vma->vm_end) return -EINVAL; if (!args->src || !args->dst) return -EINVAL; memset(args->src, 0, sizeof(*args->src) * nr_pages); args->cpages = 0; args->npages = 0; migrate_vma_collect(args); if (args->cpages) migrate_vma_prepare(args); if (args->cpages) migrate_vma_unmap(args); /* * At this point pages are locked and unmapped, and thus they have * stable content and can safely be copied to destination memory that * is allocated by the drivers. */ return 0; } EXPORT_SYMBOL(migrate_vma_setup); |
34290e2c6 mm/migrate: add s... |
2729 2730 2731 2732 2733 2734 2735 2736 |
/* * This code closely matches the code in: * __handle_mm_fault() * handle_pte_fault() * do_anonymous_page() * to map in an anonymous zero page but the struct page will be a ZONE_DEVICE * private page. */ |
8315ada7f mm/migrate: allow... |
2737 2738 2739 2740 2741 2742 2743 2744 |
static void migrate_vma_insert_page(struct migrate_vma *migrate, unsigned long addr, struct page *page, unsigned long *src, unsigned long *dst) { struct vm_area_struct *vma = migrate->vma; struct mm_struct *mm = vma->vm_mm; |
8315ada7f mm/migrate: allow... |
2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 |
bool flush = false; spinlock_t *ptl; pte_t entry; pgd_t *pgdp; p4d_t *p4dp; pud_t *pudp; pmd_t *pmdp; pte_t *ptep; /* Only allow populating anonymous memory */ if (!vma_is_anonymous(vma)) goto abort; pgdp = pgd_offset(mm, addr); p4dp = p4d_alloc(mm, pgdp, addr); if (!p4dp) goto abort; pudp = pud_alloc(mm, p4dp, addr); if (!pudp) goto abort; pmdp = pmd_alloc(mm, pudp, addr); if (!pmdp) goto abort; if (pmd_trans_huge(*pmdp) || pmd_devmap(*pmdp)) goto abort; /* * Use pte_alloc() instead of pte_alloc_map(). We can't run * pte_offset_map() on pmds where a huge pmd might be created * from a different thread. * |
3e4e28c5a mmap locking API:... |
2777 |
* pte_alloc_map() is safe to use under mmap_write_lock(mm) or when |
8315ada7f mm/migrate: allow... |
2778 2779 |
* parallel threads are excluded by other means. * |
3e4e28c5a mmap locking API:... |
2780 |
* Here we only have mmap_read_lock(mm). |
8315ada7f mm/migrate: allow... |
2781 |
*/ |
4cf589249 mm: treewide: rem... |
2782 |
if (pte_alloc(mm, pmdp)) |
8315ada7f mm/migrate: allow... |
2783 2784 2785 2786 2787 2788 2789 2790 |
goto abort; /* See the comment in pte_alloc_one_map() */ if (unlikely(pmd_trans_unstable(pmdp))) goto abort; if (unlikely(anon_vma_prepare(vma))) goto abort; |
d9eb1ea2b mm: memcontrol: d... |
2791 |
if (mem_cgroup_charge(page, vma->vm_mm, GFP_KERNEL)) |
8315ada7f mm/migrate: allow... |
2792 2793 2794 2795 2796 2797 2798 2799 |
goto abort; /* * The memory barrier inside __SetPageUptodate makes sure that * preceding stores to the page contents become visible before * the set_pte_at() write. */ __SetPageUptodate(page); |
df6ad6983 mm/device-public-... |
2800 2801 2802 2803 2804 2805 |
if (is_zone_device_page(page)) { if (is_device_private_page(page)) { swp_entry_t swp_entry; swp_entry = make_device_private_entry(page, vma->vm_flags & VM_WRITE); entry = swp_entry_to_pte(swp_entry); |
df6ad6983 mm/device-public-... |
2806 |
} |
8315ada7f mm/migrate: allow... |
2807 2808 2809 2810 2811 2812 2813 |
} else { entry = mk_pte(page, vma->vm_page_prot); if (vma->vm_flags & VM_WRITE) entry = pte_mkwrite(pte_mkdirty(entry)); } ptep = pte_offset_map_lock(mm, pmdp, addr, &ptl); |
34290e2c6 mm/migrate: add s... |
2814 2815 |
if (check_stable_address_space(mm)) goto unlock_abort; |
8315ada7f mm/migrate: allow... |
2816 2817 |
if (pte_present(*ptep)) { unsigned long pfn = pte_pfn(*ptep); |
c23a0c997 mm/migrate: clean... |
2818 2819 |
if (!is_zero_pfn(pfn)) goto unlock_abort; |
8315ada7f mm/migrate: allow... |
2820 |
flush = true; |
c23a0c997 mm/migrate: clean... |
2821 2822 |
} else if (!pte_none(*ptep)) goto unlock_abort; |
8315ada7f mm/migrate: allow... |
2823 2824 |
/* |
c23a0c997 mm/migrate: clean... |
2825 |
* Check for userfaultfd but do not deliver the fault. Instead, |
8315ada7f mm/migrate: allow... |
2826 2827 |
* just back off. */ |
c23a0c997 mm/migrate: clean... |
2828 2829 |
if (userfaultfd_missing(vma)) goto unlock_abort; |
8315ada7f mm/migrate: allow... |
2830 2831 |
inc_mm_counter(mm, MM_ANONPAGES); |
be5d0a74c mm: memcontrol: s... |
2832 |
page_add_new_anon_rmap(page, vma, addr, false); |
8315ada7f mm/migrate: allow... |
2833 |
if (!is_zone_device_page(page)) |
b518154e5 mm/vmscan: protec... |
2834 |
lru_cache_add_inactive_or_unevictable(page, vma); |
8315ada7f mm/migrate: allow... |
2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 |
get_page(page); if (flush) { flush_cache_page(vma, addr, pte_pfn(*ptep)); ptep_clear_flush_notify(vma, addr, ptep); set_pte_at_notify(mm, addr, ptep, entry); update_mmu_cache(vma, addr, ptep); } else { /* No need to invalidate - it was non-present before */ set_pte_at(mm, addr, ptep, entry); update_mmu_cache(vma, addr, ptep); } pte_unmap_unlock(ptep, ptl); *src = MIGRATE_PFN_MIGRATE; return; |
c23a0c997 mm/migrate: clean... |
2851 2852 |
unlock_abort: pte_unmap_unlock(ptep, ptl); |
8315ada7f mm/migrate: allow... |
2853 2854 2855 |
abort: *src &= ~MIGRATE_PFN_MIGRATE; } |
a7d1f22bb mm: turn migrate_... |
2856 |
/** |
8763cb45a mm/migrate: new m... |
2857 2858 2859 2860 2861 2862 2863 |
* migrate_vma_pages() - migrate meta-data from src page to dst page * @migrate: migrate struct containing all migration information * * This migrates struct page meta-data from source struct page to destination * struct page. This effectively finishes the migration from source page to the * destination page. */ |
a7d1f22bb mm: turn migrate_... |
2864 |
void migrate_vma_pages(struct migrate_vma *migrate) |
8763cb45a mm/migrate: new m... |
2865 2866 2867 |
{ const unsigned long npages = migrate->npages; const unsigned long start = migrate->start; |
ac46d4f3c mm/mmu_notifier: ... |
2868 2869 |
struct mmu_notifier_range range; unsigned long addr, i; |
8315ada7f mm/migrate: allow... |
2870 |
bool notified = false; |
8763cb45a mm/migrate: new m... |
2871 2872 2873 2874 2875 2876 |
for (i = 0, addr = start; i < npages; addr += PAGE_SIZE, i++) { struct page *newpage = migrate_pfn_to_page(migrate->dst[i]); struct page *page = migrate_pfn_to_page(migrate->src[i]); struct address_space *mapping; int r; |
8315ada7f mm/migrate: allow... |
2877 2878 |
if (!newpage) { migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; |
8763cb45a mm/migrate: new m... |
2879 |
continue; |
8315ada7f mm/migrate: allow... |
2880 2881 2882 |
} if (!page) { |
c23a0c997 mm/migrate: clean... |
2883 |
if (!(migrate->src[i] & MIGRATE_PFN_MIGRATE)) |
8315ada7f mm/migrate: allow... |
2884 |
continue; |
8315ada7f mm/migrate: allow... |
2885 |
if (!notified) { |
8315ada7f mm/migrate: allow... |
2886 |
notified = true; |
ac46d4f3c mm/mmu_notifier: ... |
2887 2888 |
mmu_notifier_range_init(&range, |
7269f9999 mm/mmu_notifier: ... |
2889 |
MMU_NOTIFY_CLEAR, 0, |
6f4f13e8d mm/mmu_notifier: ... |
2890 |
NULL, |
ac46d4f3c mm/mmu_notifier: ... |
2891 2892 2893 |
migrate->vma->vm_mm, addr, migrate->end); mmu_notifier_invalidate_range_start(&range); |
8315ada7f mm/migrate: allow... |
2894 2895 2896 2897 |
} migrate_vma_insert_page(migrate, addr, newpage, &migrate->src[i], &migrate->dst[i]); |
8763cb45a mm/migrate: new m... |
2898 |
continue; |
8315ada7f mm/migrate: allow... |
2899 |
} |
8763cb45a mm/migrate: new m... |
2900 2901 |
mapping = page_mapping(page); |
a5430dda8 mm/migrate: suppo... |
2902 2903 2904 2905 2906 2907 2908 2909 2910 2911 |
if (is_zone_device_page(newpage)) { if (is_device_private_page(newpage)) { /* * For now only support private anonymous when * migrating to un-addressable device memory. */ if (mapping) { migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; continue; } |
25b2995a3 mm: remove MEMORY... |
2912 |
} else { |
a5430dda8 mm/migrate: suppo... |
2913 2914 2915 2916 2917 2918 2919 2920 |
/* * Other types of ZONE_DEVICE page are not * supported. */ migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; continue; } } |
8763cb45a mm/migrate: new m... |
2921 2922 2923 2924 |
r = migrate_page(mapping, newpage, page, MIGRATE_SYNC_NO_COPY); if (r != MIGRATEPAGE_SUCCESS) migrate->src[i] &= ~MIGRATE_PFN_MIGRATE; } |
8315ada7f mm/migrate: allow... |
2925 |
|
4645b9fe8 mm/mmu_notifier: ... |
2926 2927 2928 2929 2930 |
/* * No need to double call mmu_notifier->invalidate_range() callback as * the above ptep_clear_flush_notify() inside migrate_vma_insert_page() * did already call it. */ |
8315ada7f mm/migrate: allow... |
2931 |
if (notified) |
ac46d4f3c mm/mmu_notifier: ... |
2932 |
mmu_notifier_invalidate_range_only_end(&range); |
8763cb45a mm/migrate: new m... |
2933 |
} |
a7d1f22bb mm: turn migrate_... |
2934 |
EXPORT_SYMBOL(migrate_vma_pages); |
8763cb45a mm/migrate: new m... |
2935 |
|
a7d1f22bb mm: turn migrate_... |
2936 |
/** |
8763cb45a mm/migrate: new m... |
2937 2938 2939 2940 2941 2942 2943 2944 2945 2946 |
* migrate_vma_finalize() - restore CPU page table entry * @migrate: migrate struct containing all migration information * * This replaces the special migration pte entry with either a mapping to the * new page if migration was successful for that page, or to the original page * otherwise. * * This also unlocks the pages and puts them back on the lru, or drops the extra * refcount, for device pages. */ |
a7d1f22bb mm: turn migrate_... |
2947 |
void migrate_vma_finalize(struct migrate_vma *migrate) |
8763cb45a mm/migrate: new m... |
2948 2949 2950 2951 2952 2953 2954 |
{ const unsigned long npages = migrate->npages; unsigned long i; for (i = 0; i < npages; i++) { struct page *newpage = migrate_pfn_to_page(migrate->dst[i]); struct page *page = migrate_pfn_to_page(migrate->src[i]); |
8315ada7f mm/migrate: allow... |
2955 2956 2957 2958 2959 |
if (!page) { if (newpage) { unlock_page(newpage); put_page(newpage); } |
8763cb45a mm/migrate: new m... |
2960 |
continue; |
8315ada7f mm/migrate: allow... |
2961 |
} |
8763cb45a mm/migrate: new m... |
2962 2963 2964 2965 2966 2967 2968 2969 2970 2971 |
if (!(migrate->src[i] & MIGRATE_PFN_MIGRATE) || !newpage) { if (newpage) { unlock_page(newpage); put_page(newpage); } newpage = page; } remove_migration_ptes(page, newpage, false); unlock_page(page); |
8763cb45a mm/migrate: new m... |
2972 |
|
a5430dda8 mm/migrate: suppo... |
2973 2974 2975 2976 |
if (is_zone_device_page(page)) put_page(page); else putback_lru_page(page); |
8763cb45a mm/migrate: new m... |
2977 2978 2979 |
if (newpage != page) { unlock_page(newpage); |
a5430dda8 mm/migrate: suppo... |
2980 2981 2982 2983 |
if (is_zone_device_page(newpage)) put_page(newpage); else putback_lru_page(newpage); |
8763cb45a mm/migrate: new m... |
2984 2985 2986 |
} } } |
a7d1f22bb mm: turn migrate_... |
2987 |
EXPORT_SYMBOL(migrate_vma_finalize); |
9b2ed9cb9 mm: remove CONFIG... |
2988 |
#endif /* CONFIG_DEVICE_PRIVATE */ |