Blame view
mm/hugetlb.c
107 KB
1da177e4c Linux-2.6.12-rc2 |
1 2 |
/* * Generic hugetlb support. |
6d49e352a propagate name ch... |
3 |
* (C) Nadia Yvette Chambers, April 2004 |
1da177e4c Linux-2.6.12-rc2 |
4 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
5 6 7 8 |
#include <linux/list.h> #include <linux/init.h> #include <linux/module.h> #include <linux/mm.h> |
e1759c215 proc: switch /pro... |
9 |
#include <linux/seq_file.h> |
1da177e4c Linux-2.6.12-rc2 |
10 11 |
#include <linux/sysctl.h> #include <linux/highmem.h> |
cddb8a5c1 mmu-notifiers: core |
12 |
#include <linux/mmu_notifier.h> |
1da177e4c Linux-2.6.12-rc2 |
13 |
#include <linux/nodemask.h> |
63551ae0f [PATCH] Hugepage ... |
14 |
#include <linux/pagemap.h> |
5da7ca860 [PATCH] Add NUMA ... |
15 |
#include <linux/mempolicy.h> |
3b32123d7 mm: use macros fr... |
16 |
#include <linux/compiler.h> |
aea47ff36 [PATCH] mm: make ... |
17 |
#include <linux/cpuset.h> |
3935baa9b [PATCH] hugepage:... |
18 |
#include <linux/mutex.h> |
aa888a749 hugetlb: support ... |
19 |
#include <linux/bootmem.h> |
a34378701 hugetlb: new sysf... |
20 |
#include <linux/sysfs.h> |
5a0e3ad6a include cleanup: ... |
21 |
#include <linux/slab.h> |
0fe6e20b9 hugetlb, rmap: ad... |
22 |
#include <linux/rmap.h> |
fd6a03edd HWPOISON, hugetlb... |
23 24 |
#include <linux/swap.h> #include <linux/swapops.h> |
c8721bbbd mm: memory-hotplu... |
25 |
#include <linux/page-isolation.h> |
8382d914e mm, hugetlb: impr... |
26 |
#include <linux/jhash.h> |
d6606683a Revert duplicate ... |
27 |
|
63551ae0f [PATCH] Hugepage ... |
28 29 |
#include <asm/page.h> #include <asm/pgtable.h> |
24669e584 hugetlb: use mmu_... |
30 |
#include <asm/tlb.h> |
63551ae0f [PATCH] Hugepage ... |
31 |
|
24669e584 hugetlb: use mmu_... |
32 |
#include <linux/io.h> |
63551ae0f [PATCH] Hugepage ... |
33 |
#include <linux/hugetlb.h> |
9dd540e23 hugetlb/cgroup: a... |
34 |
#include <linux/hugetlb_cgroup.h> |
9a3052306 hugetlb: add per ... |
35 |
#include <linux/node.h> |
7835e98b2 [PATCH] remove se... |
36 |
#include "internal.h" |
1da177e4c Linux-2.6.12-rc2 |
37 |
|
753162cd8 mm: hugetlb: fix ... |
38 |
int hugepages_treat_as_movable; |
a55164389 hugetlb: modular ... |
39 |
|
c3f38a387 hugetlb: make som... |
40 |
int hugetlb_max_hstate __read_mostly; |
e5ff21594 hugetlb: multiple... |
41 42 |
unsigned int default_hstate_idx; struct hstate hstates[HUGE_MAX_HSTATE]; |
641844f56 mm/hugetlb: intro... |
43 44 45 46 47 |
/* * Minimum page order among possible hugepage sizes, set to a proper value * at boot time. */ static unsigned int minimum_order __read_mostly = UINT_MAX; |
e5ff21594 hugetlb: multiple... |
48 |
|
53ba51d21 hugetlb: allow ar... |
49 |
__initdata LIST_HEAD(huge_boot_pages); |
e5ff21594 hugetlb: multiple... |
50 51 52 |
/* for command line parsing */ static struct hstate * __initdata parsed_hstate; static unsigned long __initdata default_hstate_max_huge_pages; |
e11bfbfcb hugetlb: override... |
53 |
static unsigned long __initdata default_hstate_size; |
e5ff21594 hugetlb: multiple... |
54 |
|
3935baa9b [PATCH] hugepage:... |
55 |
/* |
31caf665e mm: migrate: make... |
56 57 |
* Protects updates to hugepage_freelists, hugepage_activelist, nr_huge_pages, * free_huge_pages, and surplus_huge_pages. |
3935baa9b [PATCH] hugepage:... |
58 |
*/ |
c3f38a387 hugetlb: make som... |
59 |
DEFINE_SPINLOCK(hugetlb_lock); |
0bd0f9fb1 [PATCH] hugetlb: ... |
60 |
|
8382d914e mm, hugetlb: impr... |
61 62 63 64 65 66 |
/* * Serializes faults on the same logical page. This is used to * prevent spurious OOMs when the hugepage pool is fully utilized. */ static int num_fault_mutexes; static struct mutex *htlb_fault_mutex_table ____cacheline_aligned_in_smp; |
7ca02d0ae hugetlbfs: accept... |
67 68 |
/* Forward declaration */ static int hugetlb_acct_memory(struct hstate *h, long delta); |
90481622d hugepages: fix us... |
69 70 71 72 73 74 75 |
static inline void unlock_or_release_subpool(struct hugepage_subpool *spool) { bool free = (spool->count == 0) && (spool->used_hpages == 0); spin_unlock(&spool->lock); /* If no pages are used, and no other handles to the subpool |
7ca02d0ae hugetlbfs: accept... |
76 77 78 79 80 81 |
* remain, give up any reservations mased on minimum size and * free the subpool */ if (free) { if (spool->min_hpages != -1) hugetlb_acct_memory(spool->hstate, -spool->min_hpages); |
90481622d hugepages: fix us... |
82 |
kfree(spool); |
7ca02d0ae hugetlbfs: accept... |
83 |
} |
90481622d hugepages: fix us... |
84 |
} |
7ca02d0ae hugetlbfs: accept... |
85 86 |
struct hugepage_subpool *hugepage_new_subpool(struct hstate *h, long max_hpages, long min_hpages) |
90481622d hugepages: fix us... |
87 88 |
{ struct hugepage_subpool *spool; |
c6a918200 hugetlbfs: add mi... |
89 |
spool = kzalloc(sizeof(*spool), GFP_KERNEL); |
90481622d hugepages: fix us... |
90 91 92 93 94 |
if (!spool) return NULL; spin_lock_init(&spool->lock); spool->count = 1; |
7ca02d0ae hugetlbfs: accept... |
95 96 97 98 99 100 101 102 103 |
spool->max_hpages = max_hpages; spool->hstate = h; spool->min_hpages = min_hpages; if (min_hpages != -1 && hugetlb_acct_memory(h, min_hpages)) { kfree(spool); return NULL; } spool->rsv_hpages = min_hpages; |
90481622d hugepages: fix us... |
104 105 106 107 108 109 110 111 112 113 114 |
return spool; } void hugepage_put_subpool(struct hugepage_subpool *spool) { spin_lock(&spool->lock); BUG_ON(!spool->count); spool->count--; unlock_or_release_subpool(spool); } |
1c5ecae3a hugetlbfs: add mi... |
115 116 117 118 119 120 121 122 123 |
/* * Subpool accounting for allocating and reserving pages. * Return -ENOMEM if there are not enough resources to satisfy the * the request. Otherwise, return the number of pages by which the * global pools must be adjusted (upward). The returned value may * only be different than the passed value (delta) in the case where * a subpool minimum size must be manitained. */ static long hugepage_subpool_get_pages(struct hugepage_subpool *spool, |
90481622d hugepages: fix us... |
124 125 |
long delta) { |
1c5ecae3a hugetlbfs: add mi... |
126 |
long ret = delta; |
90481622d hugepages: fix us... |
127 128 |
if (!spool) |
1c5ecae3a hugetlbfs: add mi... |
129 |
return ret; |
90481622d hugepages: fix us... |
130 131 |
spin_lock(&spool->lock); |
1c5ecae3a hugetlbfs: add mi... |
132 133 134 135 136 137 138 139 |
if (spool->max_hpages != -1) { /* maximum size accounting */ if ((spool->used_hpages + delta) <= spool->max_hpages) spool->used_hpages += delta; else { ret = -ENOMEM; goto unlock_ret; } |
90481622d hugepages: fix us... |
140 |
} |
90481622d hugepages: fix us... |
141 |
|
1c5ecae3a hugetlbfs: add mi... |
142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 |
if (spool->min_hpages != -1) { /* minimum size accounting */ if (delta > spool->rsv_hpages) { /* * Asking for more reserves than those already taken on * behalf of subpool. Return difference. */ ret = delta - spool->rsv_hpages; spool->rsv_hpages = 0; } else { ret = 0; /* reserves already accounted for */ spool->rsv_hpages -= delta; } } unlock_ret: spin_unlock(&spool->lock); |
90481622d hugepages: fix us... |
158 159 |
return ret; } |
1c5ecae3a hugetlbfs: add mi... |
160 161 162 163 164 165 166 |
/* * Subpool accounting for freeing and unreserving pages. * Return the number of global page reservations that must be dropped. * The return value may only be different than the passed value (delta) * in the case where a subpool minimum size must be maintained. */ static long hugepage_subpool_put_pages(struct hugepage_subpool *spool, |
90481622d hugepages: fix us... |
167 168 |
long delta) { |
1c5ecae3a hugetlbfs: add mi... |
169 |
long ret = delta; |
90481622d hugepages: fix us... |
170 |
if (!spool) |
1c5ecae3a hugetlbfs: add mi... |
171 |
return delta; |
90481622d hugepages: fix us... |
172 173 |
spin_lock(&spool->lock); |
1c5ecae3a hugetlbfs: add mi... |
174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 |
if (spool->max_hpages != -1) /* maximum size accounting */ spool->used_hpages -= delta; if (spool->min_hpages != -1) { /* minimum size accounting */ if (spool->rsv_hpages + delta <= spool->min_hpages) ret = 0; else ret = spool->rsv_hpages + delta - spool->min_hpages; spool->rsv_hpages += delta; if (spool->rsv_hpages > spool->min_hpages) spool->rsv_hpages = spool->min_hpages; } /* * If hugetlbfs_put_super couldn't free spool due to an outstanding * quota reference, free it now. */ |
90481622d hugepages: fix us... |
193 |
unlock_or_release_subpool(spool); |
1c5ecae3a hugetlbfs: add mi... |
194 195 |
return ret; |
90481622d hugepages: fix us... |
196 197 198 199 200 201 202 203 204 |
} static inline struct hugepage_subpool *subpool_inode(struct inode *inode) { return HUGETLBFS_SB(inode->i_sb)->spool; } static inline struct hugepage_subpool *subpool_vma(struct vm_area_struct *vma) { |
496ad9aa8 new helper: file_... |
205 |
return subpool_inode(file_inode(vma->vm_file)); |
90481622d hugepages: fix us... |
206 |
} |
e7c4b0bfd huge page private... |
207 |
/* |
968229048 hugetlb: move res... |
208 209 |
* Region tracking -- allows tracking of reservations and instantiated pages * across the pages in a mapping. |
84afd99b8 hugetlb reservati... |
210 |
* |
1dd308a7b mm/hugetlb: docum... |
211 212 213 214 215 216 217 218 219 220 221 222 223 224 |
* The region data structures are embedded into a resv_map and protected * by a resv_map's lock. The set of regions within the resv_map represent * reservations for huge pages, or huge pages that have already been * instantiated within the map. The from and to elements are huge page * indicies into the associated mapping. from indicates the starting index * of the region. to represents the first index past the end of the region. * * For example, a file region structure with from == 0 and to == 4 represents * four huge pages in a mapping. It is important to note that the to element * represents the first element past the end of the region. This is used in * arithmetic as 4(to) - 0(from) = 4 huge pages in the region. * * Interval notation of the form [from, to) will be used to indicate that * the endpoint from is inclusive and to is exclusive. |
968229048 hugetlb: move res... |
225 226 227 228 229 230 |
*/ struct file_region { struct list_head link; long from; long to; }; |
1dd308a7b mm/hugetlb: docum... |
231 232 233 234 235 236 237 |
/* * Add the huge page range represented by [f, t) to the reserve * map. Existing regions will be expanded to accommodate the * specified range. We know only existing regions need to be * expanded, because region_add is only called after region_chg * with the same range. If a new file_region structure must * be allocated, it is done in region_chg. |
cf3ad20bf mm/hugetlb: compu... |
238 239 240 |
* * Return the number of new huge pages added to the map. This * number is greater than or equal to zero. |
1dd308a7b mm/hugetlb: docum... |
241 |
*/ |
1406ec9ba mm, hugetlb: impr... |
242 |
static long region_add(struct resv_map *resv, long f, long t) |
968229048 hugetlb: move res... |
243 |
{ |
1406ec9ba mm, hugetlb: impr... |
244 |
struct list_head *head = &resv->regions; |
968229048 hugetlb: move res... |
245 |
struct file_region *rg, *nrg, *trg; |
cf3ad20bf mm/hugetlb: compu... |
246 |
long add = 0; |
968229048 hugetlb: move res... |
247 |
|
7b24d8616 mm, hugetlb: fix ... |
248 |
spin_lock(&resv->lock); |
968229048 hugetlb: move res... |
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 |
/* Locate the region we are either in or before. */ list_for_each_entry(rg, head, link) if (f <= rg->to) break; /* Round our left edge to the current segment if it encloses us. */ if (f > rg->from) f = rg->from; /* Check for and consume any regions we now overlap with. */ nrg = rg; list_for_each_entry_safe(rg, trg, rg->link.prev, link) { if (&rg->link == head) break; if (rg->from > t) break; /* If this area reaches higher then extend our area to * include it completely. If this is not the first area * which we intend to reuse, free it. */ if (rg->to > t) t = rg->to; if (rg != nrg) { |
cf3ad20bf mm/hugetlb: compu... |
272 273 274 275 276 |
/* Decrement return value by the deleted range. * Another range will span this area so that by * end of routine add will be >= zero */ add -= (rg->to - rg->from); |
968229048 hugetlb: move res... |
277 278 279 280 |
list_del(&rg->link); kfree(rg); } } |
cf3ad20bf mm/hugetlb: compu... |
281 282 |
add += (nrg->from - f); /* Added to beginning of region */ |
968229048 hugetlb: move res... |
283 |
nrg->from = f; |
cf3ad20bf mm/hugetlb: compu... |
284 |
add += t - nrg->to; /* Added to end of region */ |
968229048 hugetlb: move res... |
285 |
nrg->to = t; |
cf3ad20bf mm/hugetlb: compu... |
286 |
|
7b24d8616 mm, hugetlb: fix ... |
287 |
spin_unlock(&resv->lock); |
cf3ad20bf mm/hugetlb: compu... |
288 289 |
VM_BUG_ON(add < 0); return add; |
968229048 hugetlb: move res... |
290 |
} |
1dd308a7b mm/hugetlb: docum... |
291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 |
/* * Examine the existing reserve map and determine how many * huge pages in the specified range [f, t) are NOT currently * represented. This routine is called before a subsequent * call to region_add that will actually modify the reserve * map to add the specified range [f, t). region_chg does * not change the number of huge pages represented by the * map. However, if the existing regions in the map can not * be expanded to represent the new range, a new file_region * structure is added to the map as a placeholder. This is * so that the subsequent region_add call will have all the * regions it needs and will not fail. * * Returns the number of huge pages that need to be added * to the existing reservation map for the range [f, t). * This number is greater or equal to zero. -ENOMEM is * returned if a new file_region structure is needed and can * not be allocated. */ |
1406ec9ba mm, hugetlb: impr... |
310 |
static long region_chg(struct resv_map *resv, long f, long t) |
968229048 hugetlb: move res... |
311 |
{ |
1406ec9ba mm, hugetlb: impr... |
312 |
struct list_head *head = &resv->regions; |
7b24d8616 mm, hugetlb: fix ... |
313 |
struct file_region *rg, *nrg = NULL; |
968229048 hugetlb: move res... |
314 |
long chg = 0; |
7b24d8616 mm, hugetlb: fix ... |
315 316 |
retry: spin_lock(&resv->lock); |
968229048 hugetlb: move res... |
317 318 319 320 321 322 323 324 325 |
/* Locate the region we are before or in. */ list_for_each_entry(rg, head, link) if (f <= rg->to) break; /* If we are below the current region then a new region is required. * Subtle, allocate a new region at the position but make it zero * size such that we can guarantee to record the reservation. */ if (&rg->link == head || t < rg->from) { |
7b24d8616 mm, hugetlb: fix ... |
326 327 328 329 330 331 332 333 334 335 336 |
if (!nrg) { spin_unlock(&resv->lock); nrg = kmalloc(sizeof(*nrg), GFP_KERNEL); if (!nrg) return -ENOMEM; nrg->from = f; nrg->to = f; INIT_LIST_HEAD(&nrg->link); goto retry; } |
968229048 hugetlb: move res... |
337 |
|
7b24d8616 mm, hugetlb: fix ... |
338 339 340 |
list_add(&nrg->link, rg->link.prev); chg = t - f; goto out_nrg; |
968229048 hugetlb: move res... |
341 342 343 344 345 346 347 348 349 350 351 352 |
} /* Round our left edge to the current segment if it encloses us. */ if (f > rg->from) f = rg->from; chg = t - f; /* Check for and consume any regions we now overlap with. */ list_for_each_entry(rg, rg->link.prev, link) { if (&rg->link == head) break; if (rg->from > t) |
7b24d8616 mm, hugetlb: fix ... |
353 |
goto out; |
968229048 hugetlb: move res... |
354 |
|
25985edce Fix common misspe... |
355 |
/* We overlap with this area, if it extends further than |
968229048 hugetlb: move res... |
356 357 358 359 360 361 362 363 |
* us then we must extend ourselves. Account for its * existing reservation. */ if (rg->to > t) { chg += rg->to - t; t = rg->to; } chg -= rg->to - rg->from; } |
7b24d8616 mm, hugetlb: fix ... |
364 365 366 367 368 369 370 371 |
out: spin_unlock(&resv->lock); /* We already know we raced and no longer need the new region */ kfree(nrg); return chg; out_nrg: spin_unlock(&resv->lock); |
968229048 hugetlb: move res... |
372 373 |
return chg; } |
1dd308a7b mm/hugetlb: docum... |
374 375 376 377 378 |
/* * Truncate the reserve map at index 'end'. Modify/truncate any * region which contains end. Delete any regions past end. * Return the number of huge pages removed from the map. */ |
1406ec9ba mm, hugetlb: impr... |
379 |
static long region_truncate(struct resv_map *resv, long end) |
968229048 hugetlb: move res... |
380 |
{ |
1406ec9ba mm, hugetlb: impr... |
381 |
struct list_head *head = &resv->regions; |
968229048 hugetlb: move res... |
382 383 |
struct file_region *rg, *trg; long chg = 0; |
7b24d8616 mm, hugetlb: fix ... |
384 |
spin_lock(&resv->lock); |
968229048 hugetlb: move res... |
385 386 387 388 389 |
/* Locate the region we are either in or before. */ list_for_each_entry(rg, head, link) if (end <= rg->to) break; if (&rg->link == head) |
7b24d8616 mm, hugetlb: fix ... |
390 |
goto out; |
968229048 hugetlb: move res... |
391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 |
/* If we are in the middle of a region then adjust it. */ if (end > rg->from) { chg = rg->to - end; rg->to = end; rg = list_entry(rg->link.next, typeof(*rg), link); } /* Drop any remaining regions. */ list_for_each_entry_safe(rg, trg, rg->link.prev, link) { if (&rg->link == head) break; chg += rg->to - rg->from; list_del(&rg->link); kfree(rg); } |
7b24d8616 mm, hugetlb: fix ... |
407 408 409 |
out: spin_unlock(&resv->lock); |
968229048 hugetlb: move res... |
410 411 |
return chg; } |
1dd308a7b mm/hugetlb: docum... |
412 413 414 415 |
/* * Count and return the number of huge pages in the reserve map * that intersect with the range [f, t). */ |
1406ec9ba mm, hugetlb: impr... |
416 |
static long region_count(struct resv_map *resv, long f, long t) |
84afd99b8 hugetlb reservati... |
417 |
{ |
1406ec9ba mm, hugetlb: impr... |
418 |
struct list_head *head = &resv->regions; |
84afd99b8 hugetlb reservati... |
419 420 |
struct file_region *rg; long chg = 0; |
7b24d8616 mm, hugetlb: fix ... |
421 |
spin_lock(&resv->lock); |
84afd99b8 hugetlb reservati... |
422 423 |
/* Locate each segment we overlap with, and count that overlap. */ list_for_each_entry(rg, head, link) { |
f2135a4a5 mm/hugetlb.c: use... |
424 425 |
long seg_from; long seg_to; |
84afd99b8 hugetlb reservati... |
426 427 428 429 430 431 432 433 434 435 436 |
if (rg->to <= f) continue; if (rg->from >= t) break; seg_from = max(rg->from, f); seg_to = min(rg->to, t); chg += seg_to - seg_from; } |
7b24d8616 mm, hugetlb: fix ... |
437 |
spin_unlock(&resv->lock); |
84afd99b8 hugetlb reservati... |
438 439 440 |
return chg; } |
968229048 hugetlb: move res... |
441 |
/* |
e7c4b0bfd huge page private... |
442 |
* Convert the address within this vma to the page offset within |
e7c4b0bfd huge page private... |
443 444 |
* the mapping, in pagecache page units; huge pages here. */ |
a55164389 hugetlb: modular ... |
445 446 |
static pgoff_t vma_hugecache_offset(struct hstate *h, struct vm_area_struct *vma, unsigned long address) |
e7c4b0bfd huge page private... |
447 |
{ |
a55164389 hugetlb: modular ... |
448 449 |
return ((address - vma->vm_start) >> huge_page_shift(h)) + (vma->vm_pgoff >> huge_page_order(h)); |
e7c4b0bfd huge page private... |
450 |
} |
0fe6e20b9 hugetlb, rmap: ad... |
451 452 453 454 455 |
pgoff_t linear_hugepage_index(struct vm_area_struct *vma, unsigned long address) { return vma_hugecache_offset(hstate_vma(vma), vma, address); } |
84afd99b8 hugetlb reservati... |
456 |
/* |
08fba6998 mm: report the pa... |
457 458 459 460 461 462 463 464 465 466 467 |
* Return the size of the pages allocated when backing a VMA. In the majority * cases this will be same size as used by the page table entries. */ unsigned long vma_kernel_pagesize(struct vm_area_struct *vma) { struct hstate *hstate; if (!is_vm_hugetlb_page(vma)) return PAGE_SIZE; hstate = hstate_vma(vma); |
2415cf12e mm/hugetlb: use a... |
468 |
return 1UL << huge_page_shift(hstate); |
08fba6998 mm: report the pa... |
469 |
} |
f340ca0f0 hugetlbfs: export... |
470 |
EXPORT_SYMBOL_GPL(vma_kernel_pagesize); |
08fba6998 mm: report the pa... |
471 472 |
/* |
3340289dd mm: report the MM... |
473 474 475 476 477 478 479 480 481 482 483 484 485 |
* Return the page size being used by the MMU to back a VMA. In the majority * of cases, the page size used by the kernel matches the MMU size. On * architectures where it differs, an architecture-specific version of this * function is required. */ #ifndef vma_mmu_pagesize unsigned long vma_mmu_pagesize(struct vm_area_struct *vma) { return vma_kernel_pagesize(vma); } #endif /* |
84afd99b8 hugetlb reservati... |
486 487 488 489 490 491 |
* Flags for MAP_PRIVATE reservations. These are stored in the bottom * bits of the reservation map pointer, which are always clear due to * alignment. */ #define HPAGE_RESV_OWNER (1UL << 0) #define HPAGE_RESV_UNMAPPED (1UL << 1) |
04f2cbe35 hugetlb: guarante... |
492 |
#define HPAGE_RESV_MASK (HPAGE_RESV_OWNER | HPAGE_RESV_UNMAPPED) |
84afd99b8 hugetlb reservati... |
493 |
|
a1e78772d hugetlb: reserve ... |
494 495 496 497 498 499 500 501 502 |
/* * These helpers are used to track how many pages are reserved for * faults in a MAP_PRIVATE mapping. Only the process that called mmap() * is guaranteed to have their future faults succeed. * * With the exception of reset_vma_resv_huge_pages() which is called at fork(), * the reserve counters are updated with the hugetlb_lock held. It is safe * to reset the VMA at fork() time as it is not in use yet and there is no * chance of the global counters getting corrupted as a result of the values. |
84afd99b8 hugetlb reservati... |
503 504 505 506 507 508 509 510 511 |
* * The private mapping reservation is represented in a subtly different * manner to a shared mapping. A shared mapping has a region map associated * with the underlying file, this region map represents the backing file * pages which have ever had a reservation assigned which this persists even * after the page is instantiated. A private mapping has a region map * associated with the original mmap which is attached to all VMAs which * reference it, this region map represents those offsets which have consumed * reservation ie. where pages have been instantiated. |
a1e78772d hugetlb: reserve ... |
512 |
*/ |
e7c4b0bfd huge page private... |
513 514 515 516 517 518 519 520 521 522 |
static unsigned long get_vma_private_data(struct vm_area_struct *vma) { return (unsigned long)vma->vm_private_data; } static void set_vma_private_data(struct vm_area_struct *vma, unsigned long value) { vma->vm_private_data = (void *)value; } |
9119a41e9 mm, hugetlb: unif... |
523 |
struct resv_map *resv_map_alloc(void) |
84afd99b8 hugetlb reservati... |
524 525 526 527 528 529 |
{ struct resv_map *resv_map = kmalloc(sizeof(*resv_map), GFP_KERNEL); if (!resv_map) return NULL; kref_init(&resv_map->refs); |
7b24d8616 mm, hugetlb: fix ... |
530 |
spin_lock_init(&resv_map->lock); |
84afd99b8 hugetlb reservati... |
531 532 533 534 |
INIT_LIST_HEAD(&resv_map->regions); return resv_map; } |
9119a41e9 mm, hugetlb: unif... |
535 |
void resv_map_release(struct kref *ref) |
84afd99b8 hugetlb reservati... |
536 537 538 539 |
{ struct resv_map *resv_map = container_of(ref, struct resv_map, refs); /* Clear out any active regions before we release the map. */ |
1406ec9ba mm, hugetlb: impr... |
540 |
region_truncate(resv_map, 0); |
84afd99b8 hugetlb reservati... |
541 542 |
kfree(resv_map); } |
4e35f4838 mm, hugetlb: use ... |
543 544 545 546 |
static inline struct resv_map *inode_resv_map(struct inode *inode) { return inode->i_mapping->private_data; } |
84afd99b8 hugetlb reservati... |
547 |
static struct resv_map *vma_resv_map(struct vm_area_struct *vma) |
a1e78772d hugetlb: reserve ... |
548 |
{ |
81d1b09c6 mm: convert a few... |
549 |
VM_BUG_ON_VMA(!is_vm_hugetlb_page(vma), vma); |
4e35f4838 mm, hugetlb: use ... |
550 551 552 553 554 555 556 |
if (vma->vm_flags & VM_MAYSHARE) { struct address_space *mapping = vma->vm_file->f_mapping; struct inode *inode = mapping->host; return inode_resv_map(inode); } else { |
84afd99b8 hugetlb reservati... |
557 558 |
return (struct resv_map *)(get_vma_private_data(vma) & ~HPAGE_RESV_MASK); |
4e35f4838 mm, hugetlb: use ... |
559 |
} |
a1e78772d hugetlb: reserve ... |
560 |
} |
84afd99b8 hugetlb reservati... |
561 |
static void set_vma_resv_map(struct vm_area_struct *vma, struct resv_map *map) |
a1e78772d hugetlb: reserve ... |
562 |
{ |
81d1b09c6 mm: convert a few... |
563 564 |
VM_BUG_ON_VMA(!is_vm_hugetlb_page(vma), vma); VM_BUG_ON_VMA(vma->vm_flags & VM_MAYSHARE, vma); |
a1e78772d hugetlb: reserve ... |
565 |
|
84afd99b8 hugetlb reservati... |
566 567 |
set_vma_private_data(vma, (get_vma_private_data(vma) & HPAGE_RESV_MASK) | (unsigned long)map); |
04f2cbe35 hugetlb: guarante... |
568 569 570 571 |
} static void set_vma_resv_flags(struct vm_area_struct *vma, unsigned long flags) { |
81d1b09c6 mm: convert a few... |
572 573 |
VM_BUG_ON_VMA(!is_vm_hugetlb_page(vma), vma); VM_BUG_ON_VMA(vma->vm_flags & VM_MAYSHARE, vma); |
e7c4b0bfd huge page private... |
574 575 |
set_vma_private_data(vma, get_vma_private_data(vma) | flags); |
04f2cbe35 hugetlb: guarante... |
576 577 578 579 |
} static int is_vma_resv_set(struct vm_area_struct *vma, unsigned long flag) { |
81d1b09c6 mm: convert a few... |
580 |
VM_BUG_ON_VMA(!is_vm_hugetlb_page(vma), vma); |
e7c4b0bfd huge page private... |
581 582 |
return (get_vma_private_data(vma) & flag) != 0; |
a1e78772d hugetlb: reserve ... |
583 |
} |
04f2cbe35 hugetlb: guarante... |
584 |
/* Reset counters to 0 and clear all HPAGE_RESV_* flags */ |
a1e78772d hugetlb: reserve ... |
585 586 |
void reset_vma_resv_huge_pages(struct vm_area_struct *vma) { |
81d1b09c6 mm: convert a few... |
587 |
VM_BUG_ON_VMA(!is_vm_hugetlb_page(vma), vma); |
f83a275db mm: account for M... |
588 |
if (!(vma->vm_flags & VM_MAYSHARE)) |
a1e78772d hugetlb: reserve ... |
589 590 591 592 |
vma->vm_private_data = (void *)0; } /* Returns true if the VMA has associated reserve pages */ |
af0ed73e6 mm, hugetlb: decr... |
593 |
static int vma_has_reserves(struct vm_area_struct *vma, long chg) |
a1e78772d hugetlb: reserve ... |
594 |
{ |
af0ed73e6 mm, hugetlb: decr... |
595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 |
if (vma->vm_flags & VM_NORESERVE) { /* * This address is already reserved by other process(chg == 0), * so, we should decrement reserved count. Without decrementing, * reserve count remains after releasing inode, because this * allocated page will go into page cache and is regarded as * coming from reserved pool in releasing step. Currently, we * don't have any other solution to deal with this situation * properly, so add work-around here. */ if (vma->vm_flags & VM_MAYSHARE && chg == 0) return 1; else return 0; } |
a63884e92 mm, hugetlb: remo... |
610 611 |
/* Shared mappings always use reserves */ |
f83a275db mm: account for M... |
612 |
if (vma->vm_flags & VM_MAYSHARE) |
7f09ca51e hugetlb: fix a hu... |
613 |
return 1; |
a63884e92 mm, hugetlb: remo... |
614 615 616 617 618 |
/* * Only the process that called mmap() has reserves for * private mappings. */ |
7f09ca51e hugetlb: fix a hu... |
619 620 |
if (is_vma_resv_set(vma, HPAGE_RESV_OWNER)) return 1; |
a63884e92 mm, hugetlb: remo... |
621 |
|
7f09ca51e hugetlb: fix a hu... |
622 |
return 0; |
a1e78772d hugetlb: reserve ... |
623 |
} |
a55164389 hugetlb: modular ... |
624 |
static void enqueue_huge_page(struct hstate *h, struct page *page) |
1da177e4c Linux-2.6.12-rc2 |
625 626 |
{ int nid = page_to_nid(page); |
0edaecfab hugetlb: add a li... |
627 |
list_move(&page->lru, &h->hugepage_freelists[nid]); |
a55164389 hugetlb: modular ... |
628 629 |
h->free_huge_pages++; h->free_huge_pages_node[nid]++; |
1da177e4c Linux-2.6.12-rc2 |
630 |
} |
bf50bab2b hugetlb: add allo... |
631 632 633 |
static struct page *dequeue_huge_page_node(struct hstate *h, int nid) { struct page *page; |
c8721bbbd mm: memory-hotplu... |
634 635 636 637 638 639 640 641 |
list_for_each_entry(page, &h->hugepage_freelists[nid], lru) if (!is_migrate_isolate_page(page)) break; /* * if 'non-isolated free hugepage' not found on the list, * the allocation fails. */ if (&h->hugepage_freelists[nid] == &page->lru) |
bf50bab2b hugetlb: add allo... |
642 |
return NULL; |
0edaecfab hugetlb: add a li... |
643 |
list_move(&page->lru, &h->hugepage_activelist); |
a9869b837 hugetlb: move ref... |
644 |
set_page_refcounted(page); |
bf50bab2b hugetlb: add allo... |
645 646 647 648 |
h->free_huge_pages--; h->free_huge_pages_node[nid]--; return page; } |
86cdb465c mm: prepare to re... |
649 650 651 |
/* Movability of hugepages depends on migration support. */ static inline gfp_t htlb_alloc_mask(struct hstate *h) { |
100873d7a hugetlb: rename h... |
652 |
if (hugepages_treat_as_movable || hugepage_migration_supported(h)) |
86cdb465c mm: prepare to re... |
653 654 655 656 |
return GFP_HIGHUSER_MOVABLE; else return GFP_HIGHUSER; } |
a55164389 hugetlb: modular ... |
657 658 |
static struct page *dequeue_huge_page_vma(struct hstate *h, struct vm_area_struct *vma, |
af0ed73e6 mm, hugetlb: decr... |
659 660 |
unsigned long address, int avoid_reserve, long chg) |
1da177e4c Linux-2.6.12-rc2 |
661 |
{ |
b1c12cbcd mm/hugetlb: fix w... |
662 |
struct page *page = NULL; |
480eccf9a Fix NUMA Memory P... |
663 |
struct mempolicy *mpol; |
19770b326 mm: filter based ... |
664 |
nodemask_t *nodemask; |
c0ff7453b cpuset,mm: fix no... |
665 |
struct zonelist *zonelist; |
dd1a239f6 mm: have zonelist... |
666 667 |
struct zone *zone; struct zoneref *z; |
cc9a6c877 cpuset: mm: reduc... |
668 |
unsigned int cpuset_mems_cookie; |
1da177e4c Linux-2.6.12-rc2 |
669 |
|
a1e78772d hugetlb: reserve ... |
670 671 672 673 674 |
/* * A child process with MAP_PRIVATE mappings created by their parent * have no page reserves. This check ensures that reservations are * not "stolen". The child may still get SIGKILLed */ |
af0ed73e6 mm, hugetlb: decr... |
675 |
if (!vma_has_reserves(vma, chg) && |
a55164389 hugetlb: modular ... |
676 |
h->free_huge_pages - h->resv_huge_pages == 0) |
c0ff7453b cpuset,mm: fix no... |
677 |
goto err; |
a1e78772d hugetlb: reserve ... |
678 |
|
04f2cbe35 hugetlb: guarante... |
679 |
/* If reserves cannot be used, ensure enough pages are in the pool */ |
a55164389 hugetlb: modular ... |
680 |
if (avoid_reserve && h->free_huge_pages - h->resv_huge_pages == 0) |
6eab04a87 treewide: remove ... |
681 |
goto err; |
04f2cbe35 hugetlb: guarante... |
682 |
|
9966c4bbb mm, hugetlb: move... |
683 |
retry_cpuset: |
d26914d11 mm: optimize put_... |
684 |
cpuset_mems_cookie = read_mems_allowed_begin(); |
9966c4bbb mm, hugetlb: move... |
685 |
zonelist = huge_zonelist(vma, address, |
86cdb465c mm: prepare to re... |
686 |
htlb_alloc_mask(h), &mpol, &nodemask); |
9966c4bbb mm, hugetlb: move... |
687 |
|
19770b326 mm: filter based ... |
688 689 |
for_each_zone_zonelist_nodemask(zone, z, zonelist, MAX_NR_ZONES - 1, nodemask) { |
344736f29 cpuset: simplify ... |
690 |
if (cpuset_zone_allowed(zone, htlb_alloc_mask(h))) { |
bf50bab2b hugetlb: add allo... |
691 692 |
page = dequeue_huge_page_node(h, zone_to_nid(zone)); if (page) { |
af0ed73e6 mm, hugetlb: decr... |
693 694 695 696 |
if (avoid_reserve) break; if (!vma_has_reserves(vma, chg)) break; |
07443a85a mm, hugetlb: retu... |
697 |
SetPagePrivate(page); |
af0ed73e6 mm, hugetlb: decr... |
698 |
h->resv_huge_pages--; |
bf50bab2b hugetlb: add allo... |
699 700 |
break; } |
3abf7afd4 dequeue_huge_page... |
701 |
} |
1da177e4c Linux-2.6.12-rc2 |
702 |
} |
cc9a6c877 cpuset: mm: reduc... |
703 |
|
52cd3b074 mempolicy: rework... |
704 |
mpol_cond_put(mpol); |
d26914d11 mm: optimize put_... |
705 |
if (unlikely(!page && read_mems_allowed_retry(cpuset_mems_cookie))) |
cc9a6c877 cpuset: mm: reduc... |
706 |
goto retry_cpuset; |
1da177e4c Linux-2.6.12-rc2 |
707 |
return page; |
cc9a6c877 cpuset: mm: reduc... |
708 709 |
err: |
cc9a6c877 cpuset: mm: reduc... |
710 |
return NULL; |
1da177e4c Linux-2.6.12-rc2 |
711 |
} |
1cac6f2c0 hugetlb: move hel... |
712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 |
/* * common helper functions for hstate_next_node_to_{alloc|free}. * We may have allocated or freed a huge page based on a different * nodes_allowed previously, so h->next_node_to_{alloc|free} might * be outside of *nodes_allowed. Ensure that we use an allowed * node for alloc or free. */ static int next_node_allowed(int nid, nodemask_t *nodes_allowed) { nid = next_node(nid, *nodes_allowed); if (nid == MAX_NUMNODES) nid = first_node(*nodes_allowed); VM_BUG_ON(nid >= MAX_NUMNODES); return nid; } static int get_valid_node_allowed(int nid, nodemask_t *nodes_allowed) { if (!node_isset(nid, *nodes_allowed)) nid = next_node_allowed(nid, nodes_allowed); return nid; } /* * returns the previously saved node ["this node"] from which to * allocate a persistent huge page for the pool and advance the * next node from which to allocate, handling wrap at end of node * mask. */ static int hstate_next_node_to_alloc(struct hstate *h, nodemask_t *nodes_allowed) { int nid; VM_BUG_ON(!nodes_allowed); nid = get_valid_node_allowed(h->next_nid_to_alloc, nodes_allowed); h->next_nid_to_alloc = next_node_allowed(nid, nodes_allowed); return nid; } /* * helper for free_pool_huge_page() - return the previously saved * node ["this node"] from which to free a huge page. Advance the * next node id whether or not we find a free huge page to free so * that the next attempt to free addresses the next node. */ static int hstate_next_node_to_free(struct hstate *h, nodemask_t *nodes_allowed) { int nid; VM_BUG_ON(!nodes_allowed); nid = get_valid_node_allowed(h->next_nid_to_free, nodes_allowed); h->next_nid_to_free = next_node_allowed(nid, nodes_allowed); return nid; } #define for_each_node_mask_to_alloc(hs, nr_nodes, node, mask) \ for (nr_nodes = nodes_weight(*mask); \ nr_nodes > 0 && \ ((node = hstate_next_node_to_alloc(hs, mask)) || 1); \ nr_nodes--) #define for_each_node_mask_to_free(hs, nr_nodes, node, mask) \ for (nr_nodes = nodes_weight(*mask); \ nr_nodes > 0 && \ ((node = hstate_next_node_to_free(hs, mask)) || 1); \ nr_nodes--) |
944d9fec8 hugetlb: add supp... |
784 785 786 787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 |
#if defined(CONFIG_CMA) && defined(CONFIG_X86_64) static void destroy_compound_gigantic_page(struct page *page, unsigned long order) { int i; int nr_pages = 1 << order; struct page *p = page + 1; for (i = 1; i < nr_pages; i++, p = mem_map_next(p, page, i)) { __ClearPageTail(p); set_page_refcounted(p); p->first_page = NULL; } set_compound_order(page, 0); __ClearPageHead(page); } static void free_gigantic_page(struct page *page, unsigned order) { free_contig_range(page_to_pfn(page), 1 << order); } static int __alloc_gigantic_page(unsigned long start_pfn, unsigned long nr_pages) { unsigned long end_pfn = start_pfn + nr_pages; return alloc_contig_range(start_pfn, end_pfn, MIGRATE_MOVABLE); } static bool pfn_range_valid_gigantic(unsigned long start_pfn, unsigned long nr_pages) { unsigned long i, end_pfn = start_pfn + nr_pages; struct page *page; for (i = start_pfn; i < end_pfn; i++) { if (!pfn_valid(i)) return false; page = pfn_to_page(i); if (PageReserved(page)) return false; if (page_count(page) > 0) return false; if (PageHuge(page)) return false; } return true; } static bool zone_spans_last_pfn(const struct zone *zone, unsigned long start_pfn, unsigned long nr_pages) { unsigned long last_pfn = start_pfn + nr_pages - 1; return zone_spans_pfn(zone, last_pfn); } static struct page *alloc_gigantic_page(int nid, unsigned order) { unsigned long nr_pages = 1 << order; unsigned long ret, pfn, flags; struct zone *z; z = NODE_DATA(nid)->node_zones; for (; z - NODE_DATA(nid)->node_zones < MAX_NR_ZONES; z++) { spin_lock_irqsave(&z->lock, flags); pfn = ALIGN(z->zone_start_pfn, nr_pages); while (zone_spans_last_pfn(z, pfn, nr_pages)) { if (pfn_range_valid_gigantic(pfn, nr_pages)) { /* * We release the zone lock here because * alloc_contig_range() will also lock the zone * at some point. If there's an allocation * spinning on this lock, it may win the race * and cause alloc_contig_range() to fail... */ spin_unlock_irqrestore(&z->lock, flags); ret = __alloc_gigantic_page(pfn, nr_pages); if (!ret) return pfn_to_page(pfn); spin_lock_irqsave(&z->lock, flags); } pfn += nr_pages; } spin_unlock_irqrestore(&z->lock, flags); } return NULL; } static void prep_new_huge_page(struct hstate *h, struct page *page, int nid); static void prep_compound_gigantic_page(struct page *page, unsigned long order); static struct page *alloc_fresh_gigantic_page_node(struct hstate *h, int nid) { struct page *page; page = alloc_gigantic_page(nid, huge_page_order(h)); if (page) { prep_compound_gigantic_page(page, huge_page_order(h)); prep_new_huge_page(h, page, nid); } return page; } static int alloc_fresh_gigantic_page(struct hstate *h, nodemask_t *nodes_allowed) { struct page *page = NULL; int nr_nodes, node; for_each_node_mask_to_alloc(h, nr_nodes, node, nodes_allowed) { page = alloc_fresh_gigantic_page_node(h, node); if (page) return 1; } return 0; } static inline bool gigantic_page_supported(void) { return true; } #else static inline bool gigantic_page_supported(void) { return false; } static inline void free_gigantic_page(struct page *page, unsigned order) { } static inline void destroy_compound_gigantic_page(struct page *page, unsigned long order) { } static inline int alloc_fresh_gigantic_page(struct hstate *h, nodemask_t *nodes_allowed) { return 0; } #endif |
a55164389 hugetlb: modular ... |
921 |
static void update_and_free_page(struct hstate *h, struct page *page) |
6af2acb66 hugetlb: Move upd... |
922 923 |
{ int i; |
a55164389 hugetlb: modular ... |
924 |
|
944d9fec8 hugetlb: add supp... |
925 926 |
if (hstate_is_gigantic(h) && !gigantic_page_supported()) return; |
18229df5b hugetlb: pull gig... |
927 |
|
a55164389 hugetlb: modular ... |
928 929 930 |
h->nr_huge_pages--; h->nr_huge_pages_node[page_to_nid(page)]--; for (i = 0; i < pages_per_huge_page(h); i++) { |
32f84528f mm: hugetlb: fix ... |
931 932 |
page[i].flags &= ~(1 << PG_locked | 1 << PG_error | 1 << PG_referenced | 1 << PG_dirty | |
a7407a27c hugetlb: update_a... |
933 934 |
1 << PG_active | 1 << PG_private | 1 << PG_writeback); |
6af2acb66 hugetlb: Move upd... |
935 |
} |
309381fea mm: dump page whe... |
936 |
VM_BUG_ON_PAGE(hugetlb_cgroup_from_page(page), page); |
6af2acb66 hugetlb: Move upd... |
937 938 |
set_compound_page_dtor(page, NULL); set_page_refcounted(page); |
944d9fec8 hugetlb: add supp... |
939 940 941 942 |
if (hstate_is_gigantic(h)) { destroy_compound_gigantic_page(page, huge_page_order(h)); free_gigantic_page(page, huge_page_order(h)); } else { |
944d9fec8 hugetlb: add supp... |
943 944 |
__free_pages(page, huge_page_order(h)); } |
6af2acb66 hugetlb: Move upd... |
945 |
} |
e5ff21594 hugetlb: multiple... |
946 947 948 949 950 951 952 953 954 955 |
struct hstate *size_to_hstate(unsigned long size) { struct hstate *h; for_each_hstate(h) { if (huge_page_size(h) == size) return h; } return NULL; } |
bcc542223 mm: hugetlb: intr... |
956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 |
/* * Test to determine whether the hugepage is "active/in-use" (i.e. being linked * to hstate->hugepage_activelist.) * * This function can be called for tail pages, but never returns true for them. */ bool page_huge_active(struct page *page) { VM_BUG_ON_PAGE(!PageHuge(page), page); return PageHead(page) && PagePrivate(&page[1]); } /* never called for tail page */ static void set_page_huge_active(struct page *page) { VM_BUG_ON_PAGE(!PageHeadHuge(page), page); SetPagePrivate(&page[1]); } static void clear_page_huge_active(struct page *page) { VM_BUG_ON_PAGE(!PageHeadHuge(page), page); ClearPagePrivate(&page[1]); } |
8f1d26d0e kexec: export fre... |
980 |
void free_huge_page(struct page *page) |
27a85ef1b [PATCH] hugepage:... |
981 |
{ |
a55164389 hugetlb: modular ... |
982 983 984 985 |
/* * Can't pass hstate in here because it is called from the * compound page destructor. */ |
e5ff21594 hugetlb: multiple... |
986 |
struct hstate *h = page_hstate(page); |
7893d1d50 hugetlb: Try to g... |
987 |
int nid = page_to_nid(page); |
90481622d hugepages: fix us... |
988 989 |
struct hugepage_subpool *spool = (struct hugepage_subpool *)page_private(page); |
07443a85a mm, hugetlb: retu... |
990 |
bool restore_reserve; |
27a85ef1b [PATCH] hugepage:... |
991 |
|
e5df70ab1 hugetlb: ensure w... |
992 |
set_page_private(page, 0); |
23be7468e hugetlb: fix infi... |
993 |
page->mapping = NULL; |
7893d1d50 hugetlb: Try to g... |
994 |
BUG_ON(page_count(page)); |
0fe6e20b9 hugetlb, rmap: ad... |
995 |
BUG_ON(page_mapcount(page)); |
07443a85a mm, hugetlb: retu... |
996 |
restore_reserve = PagePrivate(page); |
16c794b4f mm/hugetlb.c: cor... |
997 |
ClearPagePrivate(page); |
27a85ef1b [PATCH] hugepage:... |
998 |
|
1c5ecae3a hugetlbfs: add mi... |
999 1000 1001 1002 1003 1004 1005 |
/* * A return code of zero implies that the subpool will be under its * minimum size if the reservation is not restored after page is free. * Therefore, force restore_reserve operation. */ if (hugepage_subpool_put_pages(spool, 1) == 0) restore_reserve = true; |
27a85ef1b [PATCH] hugepage:... |
1006 |
spin_lock(&hugetlb_lock); |
bcc542223 mm: hugetlb: intr... |
1007 |
clear_page_huge_active(page); |
6d76dcf40 hugetlb/cgroup: a... |
1008 1009 |
hugetlb_cgroup_uncharge_page(hstate_index(h), pages_per_huge_page(h), page); |
07443a85a mm, hugetlb: retu... |
1010 1011 |
if (restore_reserve) h->resv_huge_pages++; |
944d9fec8 hugetlb: add supp... |
1012 |
if (h->surplus_huge_pages_node[nid]) { |
0edaecfab hugetlb: add a li... |
1013 1014 |
/* remove the page from active list */ list_del(&page->lru); |
a55164389 hugetlb: modular ... |
1015 1016 1017 |
update_and_free_page(h, page); h->surplus_huge_pages--; h->surplus_huge_pages_node[nid]--; |
7893d1d50 hugetlb: Try to g... |
1018 |
} else { |
5d3a551c2 mm: hugetlb: add ... |
1019 |
arch_clear_hugepage_flags(page); |
a55164389 hugetlb: modular ... |
1020 |
enqueue_huge_page(h, page); |
7893d1d50 hugetlb: Try to g... |
1021 |
} |
27a85ef1b [PATCH] hugepage:... |
1022 1023 |
spin_unlock(&hugetlb_lock); } |
a55164389 hugetlb: modular ... |
1024 |
static void prep_new_huge_page(struct hstate *h, struct page *page, int nid) |
b7ba30c67 hugetlb: factor o... |
1025 |
{ |
0edaecfab hugetlb: add a li... |
1026 |
INIT_LIST_HEAD(&page->lru); |
b7ba30c67 hugetlb: factor o... |
1027 1028 |
set_compound_page_dtor(page, free_huge_page); spin_lock(&hugetlb_lock); |
9dd540e23 hugetlb/cgroup: a... |
1029 |
set_hugetlb_cgroup(page, NULL); |
a55164389 hugetlb: modular ... |
1030 1031 |
h->nr_huge_pages++; h->nr_huge_pages_node[nid]++; |
b7ba30c67 hugetlb: factor o... |
1032 1033 1034 |
spin_unlock(&hugetlb_lock); put_page(page); /* free it into the hugepage allocator */ } |
2906dd528 hugetlb: prep_com... |
1035 |
static void prep_compound_gigantic_page(struct page *page, unsigned long order) |
20a0307c0 mm: introduce Pag... |
1036 1037 1038 1039 1040 1041 1042 1043 |
{ int i; int nr_pages = 1 << order; struct page *p = page + 1; /* we rely on prep_new_huge_page to set the destructor */ set_compound_order(page, order); __SetPageHead(page); |
ef5a22be2 mm: hugetlb: init... |
1044 |
__ClearPageReserved(page); |
20a0307c0 mm: introduce Pag... |
1045 |
for (i = 1; i < nr_pages; i++, p = mem_map_next(p, page, i)) { |
ef5a22be2 mm: hugetlb: init... |
1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 |
/* * For gigantic hugepages allocated through bootmem at * boot, it's safer to be consistent with the not-gigantic * hugepages and clear the PG_reserved bit from all tail pages * too. Otherwse drivers using get_user_pages() to access tail * pages may get the reference counting wrong if they see * PG_reserved set on a tail page (despite the head page not * having PG_reserved set). Enforcing this consistency between * head and tail pages allows drivers to optimize away a check * on the head page when they need know if put_page() is needed * after get_user_pages(). */ __ClearPageReserved(p); |
58a84aa92 thp: set compound... |
1059 |
set_page_count(p, 0); |
20a0307c0 mm: introduce Pag... |
1060 |
p->first_page = page; |
44fc80573 mm, hugetlb: clos... |
1061 1062 1063 |
/* Make sure p->first_page is always valid for PageTail() */ smp_wmb(); __SetPageTail(p); |
20a0307c0 mm: introduce Pag... |
1064 1065 |
} } |
7795912c2 mm: document Page... |
1066 1067 1068 1069 1070 |
/* * PageHuge() only returns true for hugetlbfs pages, but not for normal or * transparent huge pages. See the PageTransHuge() documentation for more * details. */ |
20a0307c0 mm: introduce Pag... |
1071 1072 |
int PageHuge(struct page *page) { |
20a0307c0 mm: introduce Pag... |
1073 1074 1075 1076 |
if (!PageCompound(page)) return 0; page = compound_head(page); |
758f66a29 mm/hugetlb.c: sim... |
1077 |
return get_compound_page_dtor(page) == free_huge_page; |
20a0307c0 mm: introduce Pag... |
1078 |
} |
43131e141 HWPOISON, hugetlb... |
1079 |
EXPORT_SYMBOL_GPL(PageHuge); |
27c73ae75 mm: hugetlbfs: fi... |
1080 1081 1082 1083 1084 1085 |
/* * PageHeadHuge() only returns true for hugetlbfs head page, but not for * normal or transparent huge pages. */ int PageHeadHuge(struct page *page_head) { |
27c73ae75 mm: hugetlbfs: fi... |
1086 1087 |
if (!PageHead(page_head)) return 0; |
758f66a29 mm/hugetlb.c: sim... |
1088 |
return get_compound_page_dtor(page_head) == free_huge_page; |
27c73ae75 mm: hugetlbfs: fi... |
1089 |
} |
27c73ae75 mm: hugetlbfs: fi... |
1090 |
|
13d60f4b6 futex: Take hugep... |
1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 |
pgoff_t __basepage_index(struct page *page) { struct page *page_head = compound_head(page); pgoff_t index = page_index(page_head); unsigned long compound_idx; if (!PageHuge(page_head)) return page_index(page); if (compound_order(page_head) >= MAX_ORDER) compound_idx = page_to_pfn(page) - page_to_pfn(page_head); else compound_idx = page - page_head; return (index << compound_order(page_head)) + compound_idx; } |
a55164389 hugetlb: modular ... |
1107 |
static struct page *alloc_fresh_huge_page_node(struct hstate *h, int nid) |
1da177e4c Linux-2.6.12-rc2 |
1108 |
{ |
1da177e4c Linux-2.6.12-rc2 |
1109 |
struct page *page; |
f96efd585 hugetlb: fix race... |
1110 |
|
6484eb3e2 page allocator: d... |
1111 |
page = alloc_pages_exact_node(nid, |
86cdb465c mm: prepare to re... |
1112 |
htlb_alloc_mask(h)|__GFP_COMP|__GFP_THISNODE| |
551883ae8 page allocator: e... |
1113 |
__GFP_REPEAT|__GFP_NOWARN, |
a55164389 hugetlb: modular ... |
1114 |
huge_page_order(h)); |
1da177e4c Linux-2.6.12-rc2 |
1115 |
if (page) { |
a55164389 hugetlb: modular ... |
1116 |
prep_new_huge_page(h, page, nid); |
1da177e4c Linux-2.6.12-rc2 |
1117 |
} |
63b4613c3 hugetlb: fix huge... |
1118 1119 1120 |
return page; } |
b22610268 mm, hugetlb: fix ... |
1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 |
static int alloc_fresh_huge_page(struct hstate *h, nodemask_t *nodes_allowed) { struct page *page; int nr_nodes, node; int ret = 0; for_each_node_mask_to_alloc(h, nr_nodes, node, nodes_allowed) { page = alloc_fresh_huge_page_node(h, node); if (page) { ret = 1; break; } } if (ret) count_vm_event(HTLB_BUDDY_PGALLOC); else count_vm_event(HTLB_BUDDY_PGALLOC_FAIL); return ret; } |
e8c5c8249 hugetlb: balance ... |
1142 1143 1144 1145 1146 1147 |
/* * Free huge page from pool from next node to free. * Attempt to keep persistent huge pages more or less * balanced over allowed nodes. * Called with hugetlb_lock locked. */ |
6ae11b278 hugetlb: add node... |
1148 1149 |
static int free_pool_huge_page(struct hstate *h, nodemask_t *nodes_allowed, bool acct_surplus) |
e8c5c8249 hugetlb: balance ... |
1150 |
{ |
b22610268 mm, hugetlb: fix ... |
1151 |
int nr_nodes, node; |
e8c5c8249 hugetlb: balance ... |
1152 |
int ret = 0; |
b22610268 mm, hugetlb: fix ... |
1153 |
for_each_node_mask_to_free(h, nr_nodes, node, nodes_allowed) { |
685f34570 hugetlb: use free... |
1154 1155 1156 1157 |
/* * If we're returning unused surplus pages, only examine * nodes with surplus pages. */ |
b22610268 mm, hugetlb: fix ... |
1158 1159 |
if ((!acct_surplus || h->surplus_huge_pages_node[node]) && !list_empty(&h->hugepage_freelists[node])) { |
e8c5c8249 hugetlb: balance ... |
1160 |
struct page *page = |
b22610268 mm, hugetlb: fix ... |
1161 |
list_entry(h->hugepage_freelists[node].next, |
e8c5c8249 hugetlb: balance ... |
1162 1163 1164 |
struct page, lru); list_del(&page->lru); h->free_huge_pages--; |
b22610268 mm, hugetlb: fix ... |
1165 |
h->free_huge_pages_node[node]--; |
685f34570 hugetlb: use free... |
1166 1167 |
if (acct_surplus) { h->surplus_huge_pages--; |
b22610268 mm, hugetlb: fix ... |
1168 |
h->surplus_huge_pages_node[node]--; |
685f34570 hugetlb: use free... |
1169 |
} |
e8c5c8249 hugetlb: balance ... |
1170 1171 |
update_and_free_page(h, page); ret = 1; |
9a76db099 hugetlb: rework h... |
1172 |
break; |
e8c5c8249 hugetlb: balance ... |
1173 |
} |
b22610268 mm, hugetlb: fix ... |
1174 |
} |
e8c5c8249 hugetlb: balance ... |
1175 1176 1177 |
return ret; } |
c8721bbbd mm: memory-hotplu... |
1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 |
/* * Dissolve a given free hugepage into free buddy pages. This function does * nothing for in-use (including surplus) hugepages. */ static void dissolve_free_huge_page(struct page *page) { spin_lock(&hugetlb_lock); if (PageHuge(page) && !page_count(page)) { struct hstate *h = page_hstate(page); int nid = page_to_nid(page); list_del(&page->lru); h->free_huge_pages--; h->free_huge_pages_node[nid]--; update_and_free_page(h, page); } spin_unlock(&hugetlb_lock); } /* * Dissolve free hugepages in a given pfn range. Used by memory hotplug to * make specified memory blocks removable from the system. * Note that start_pfn should aligned with (minimum) hugepage size. */ void dissolve_free_huge_pages(unsigned long start_pfn, unsigned long end_pfn) { |
c8721bbbd mm: memory-hotplu... |
1203 |
unsigned long pfn; |
c8721bbbd mm: memory-hotplu... |
1204 |
|
d01776393 mm: fix potential... |
1205 1206 |
if (!hugepages_supported()) return; |
641844f56 mm/hugetlb: intro... |
1207 1208 |
VM_BUG_ON(!IS_ALIGNED(start_pfn, 1 << minimum_order)); for (pfn = start_pfn; pfn < end_pfn; pfn += 1 << minimum_order) |
c8721bbbd mm: memory-hotplu... |
1209 1210 |
dissolve_free_huge_page(pfn_to_page(pfn)); } |
bf50bab2b hugetlb: add allo... |
1211 |
static struct page *alloc_buddy_huge_page(struct hstate *h, int nid) |
7893d1d50 hugetlb: Try to g... |
1212 1213 |
{ struct page *page; |
bf50bab2b hugetlb: add allo... |
1214 |
unsigned int r_nid; |
7893d1d50 hugetlb: Try to g... |
1215 |
|
bae7f4ae1 hugetlb: add hsta... |
1216 |
if (hstate_is_gigantic(h)) |
aa888a749 hugetlb: support ... |
1217 |
return NULL; |
d1c3fb1f8 hugetlb: introduc... |
1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240 1241 |
/* * Assume we will successfully allocate the surplus page to * prevent racing processes from causing the surplus to exceed * overcommit * * This however introduces a different race, where a process B * tries to grow the static hugepage pool while alloc_pages() is * called by process A. B will only examine the per-node * counters in determining if surplus huge pages can be * converted to normal huge pages in adjust_pool_surplus(). A * won't be able to increment the per-node counter, until the * lock is dropped by B, but B doesn't drop hugetlb_lock until * no more huge pages can be converted from surplus to normal * state (and doesn't try to convert again). Thus, we have a * case where a surplus huge page exists, the pool is grown, and * the surplus huge page still exists after, even though it * should just have been converted to a normal huge page. This * does not leak memory, though, as the hugepage will be freed * once it is out of use. It also does not allow the counters to * go out of whack in adjust_pool_surplus() as we don't modify * the node values until we've gotten the hugepage and only the * per-node value is checked there. */ spin_lock(&hugetlb_lock); |
a55164389 hugetlb: modular ... |
1242 |
if (h->surplus_huge_pages >= h->nr_overcommit_huge_pages) { |
d1c3fb1f8 hugetlb: introduc... |
1243 1244 1245 |
spin_unlock(&hugetlb_lock); return NULL; } else { |
a55164389 hugetlb: modular ... |
1246 1247 |
h->nr_huge_pages++; h->surplus_huge_pages++; |
d1c3fb1f8 hugetlb: introduc... |
1248 1249 |
} spin_unlock(&hugetlb_lock); |
bf50bab2b hugetlb: add allo... |
1250 |
if (nid == NUMA_NO_NODE) |
86cdb465c mm: prepare to re... |
1251 |
page = alloc_pages(htlb_alloc_mask(h)|__GFP_COMP| |
bf50bab2b hugetlb: add allo... |
1252 1253 1254 1255 |
__GFP_REPEAT|__GFP_NOWARN, huge_page_order(h)); else page = alloc_pages_exact_node(nid, |
86cdb465c mm: prepare to re... |
1256 |
htlb_alloc_mask(h)|__GFP_COMP|__GFP_THISNODE| |
bf50bab2b hugetlb: add allo... |
1257 |
__GFP_REPEAT|__GFP_NOWARN, huge_page_order(h)); |
d1c3fb1f8 hugetlb: introduc... |
1258 1259 |
spin_lock(&hugetlb_lock); |
7893d1d50 hugetlb: Try to g... |
1260 |
if (page) { |
0edaecfab hugetlb: add a li... |
1261 |
INIT_LIST_HEAD(&page->lru); |
bf50bab2b hugetlb: add allo... |
1262 |
r_nid = page_to_nid(page); |
7893d1d50 hugetlb: Try to g... |
1263 |
set_compound_page_dtor(page, free_huge_page); |
9dd540e23 hugetlb/cgroup: a... |
1264 |
set_hugetlb_cgroup(page, NULL); |
d1c3fb1f8 hugetlb: introduc... |
1265 1266 1267 |
/* * We incremented the global counters already */ |
bf50bab2b hugetlb: add allo... |
1268 1269 |
h->nr_huge_pages_node[r_nid]++; h->surplus_huge_pages_node[r_nid]++; |
3b1163006 Subject: [PATCH] ... |
1270 |
__count_vm_event(HTLB_BUDDY_PGALLOC); |
d1c3fb1f8 hugetlb: introduc... |
1271 |
} else { |
a55164389 hugetlb: modular ... |
1272 1273 |
h->nr_huge_pages--; h->surplus_huge_pages--; |
3b1163006 Subject: [PATCH] ... |
1274 |
__count_vm_event(HTLB_BUDDY_PGALLOC_FAIL); |
7893d1d50 hugetlb: Try to g... |
1275 |
} |
d1c3fb1f8 hugetlb: introduc... |
1276 |
spin_unlock(&hugetlb_lock); |
7893d1d50 hugetlb: Try to g... |
1277 1278 1279 |
return page; } |
e4e574b76 hugetlb: Try to g... |
1280 |
/* |
bf50bab2b hugetlb: add allo... |
1281 1282 1283 1284 1285 1286 |
* This allocation function is useful in the context where vma is irrelevant. * E.g. soft-offlining uses this function because it only cares physical * address of error page. */ struct page *alloc_huge_page_node(struct hstate *h, int nid) { |
4ef918480 mm, hugetlb: prot... |
1287 |
struct page *page = NULL; |
bf50bab2b hugetlb: add allo... |
1288 1289 |
spin_lock(&hugetlb_lock); |
4ef918480 mm, hugetlb: prot... |
1290 1291 |
if (h->free_huge_pages - h->resv_huge_pages > 0) page = dequeue_huge_page_node(h, nid); |
bf50bab2b hugetlb: add allo... |
1292 |
spin_unlock(&hugetlb_lock); |
94ae8ba71 hugetlb/cgroup: a... |
1293 |
if (!page) |
bf50bab2b hugetlb: add allo... |
1294 1295 1296 1297 1298 1299 |
page = alloc_buddy_huge_page(h, nid); return page; } /* |
25985edce Fix common misspe... |
1300 |
* Increase the hugetlb pool such that it can accommodate a reservation |
e4e574b76 hugetlb: Try to g... |
1301 1302 |
* of size 'delta'. */ |
a55164389 hugetlb: modular ... |
1303 |
static int gather_surplus_pages(struct hstate *h, int delta) |
e4e574b76 hugetlb: Try to g... |
1304 1305 1306 1307 1308 |
{ struct list_head surplus_list; struct page *page, *tmp; int ret, i; int needed, allocated; |
28073b02b mm: hugetlb: defe... |
1309 |
bool alloc_ok = true; |
e4e574b76 hugetlb: Try to g... |
1310 |
|
a55164389 hugetlb: modular ... |
1311 |
needed = (h->resv_huge_pages + delta) - h->free_huge_pages; |
ac09b3a15 hugetlb: close a ... |
1312 |
if (needed <= 0) { |
a55164389 hugetlb: modular ... |
1313 |
h->resv_huge_pages += delta; |
e4e574b76 hugetlb: Try to g... |
1314 |
return 0; |
ac09b3a15 hugetlb: close a ... |
1315 |
} |
e4e574b76 hugetlb: Try to g... |
1316 1317 1318 1319 1320 1321 1322 1323 |
allocated = 0; INIT_LIST_HEAD(&surplus_list); ret = -ENOMEM; retry: spin_unlock(&hugetlb_lock); for (i = 0; i < needed; i++) { |
bf50bab2b hugetlb: add allo... |
1324 |
page = alloc_buddy_huge_page(h, NUMA_NO_NODE); |
28073b02b mm: hugetlb: defe... |
1325 1326 1327 1328 |
if (!page) { alloc_ok = false; break; } |
e4e574b76 hugetlb: Try to g... |
1329 1330 |
list_add(&page->lru, &surplus_list); } |
28073b02b mm: hugetlb: defe... |
1331 |
allocated += i; |
e4e574b76 hugetlb: Try to g... |
1332 1333 1334 1335 1336 1337 |
/* * After retaking hugetlb_lock, we need to recalculate 'needed' * because either resv_huge_pages or free_huge_pages may have changed. */ spin_lock(&hugetlb_lock); |
a55164389 hugetlb: modular ... |
1338 1339 |
needed = (h->resv_huge_pages + delta) - (h->free_huge_pages + allocated); |
28073b02b mm: hugetlb: defe... |
1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 |
if (needed > 0) { if (alloc_ok) goto retry; /* * We were not able to allocate enough pages to * satisfy the entire reservation so we free what * we've allocated so far. */ goto free; } |
e4e574b76 hugetlb: Try to g... |
1350 1351 |
/* * The surplus_list now contains _at_least_ the number of extra pages |
25985edce Fix common misspe... |
1352 |
* needed to accommodate the reservation. Add the appropriate number |
e4e574b76 hugetlb: Try to g... |
1353 |
* of pages to the hugetlb pool and free the extras back to the buddy |
ac09b3a15 hugetlb: close a ... |
1354 1355 1356 |
* allocator. Commit the entire reservation here to prevent another * process from stealing the pages as they are added to the pool but * before they are reserved. |
e4e574b76 hugetlb: Try to g... |
1357 1358 |
*/ needed += allocated; |
a55164389 hugetlb: modular ... |
1359 |
h->resv_huge_pages += delta; |
e4e574b76 hugetlb: Try to g... |
1360 |
ret = 0; |
a9869b837 hugetlb: move ref... |
1361 |
|
19fc3f0ac hugetlb: decrease... |
1362 |
/* Free the needed pages to the hugetlb pool */ |
e4e574b76 hugetlb: Try to g... |
1363 |
list_for_each_entry_safe(page, tmp, &surplus_list, lru) { |
19fc3f0ac hugetlb: decrease... |
1364 1365 |
if ((--needed) < 0) break; |
a9869b837 hugetlb: move ref... |
1366 1367 1368 1369 1370 |
/* * This page is now managed by the hugetlb allocator and has * no users -- drop the buddy allocator's reference. */ put_page_testzero(page); |
309381fea mm: dump page whe... |
1371 |
VM_BUG_ON_PAGE(page_count(page), page); |
a55164389 hugetlb: modular ... |
1372 |
enqueue_huge_page(h, page); |
19fc3f0ac hugetlb: decrease... |
1373 |
} |
28073b02b mm: hugetlb: defe... |
1374 |
free: |
b0365c8d0 mm: hugetlb: fix ... |
1375 |
spin_unlock(&hugetlb_lock); |
19fc3f0ac hugetlb: decrease... |
1376 1377 |
/* Free unnecessary surplus pages to the buddy allocator */ |
c0d934ba2 mm, hugetlb: remo... |
1378 1379 |
list_for_each_entry_safe(page, tmp, &surplus_list, lru) put_page(page); |
a9869b837 hugetlb: move ref... |
1380 |
spin_lock(&hugetlb_lock); |
e4e574b76 hugetlb: Try to g... |
1381 1382 1383 1384 1385 1386 1387 1388 |
return ret; } /* * When releasing a hugetlb pool reservation, any surplus pages that were * allocated to satisfy the reservation must be explicitly freed if they were * never used. |
685f34570 hugetlb: use free... |
1389 |
* Called with hugetlb_lock held. |
e4e574b76 hugetlb: Try to g... |
1390 |
*/ |
a55164389 hugetlb: modular ... |
1391 1392 |
static void return_unused_surplus_pages(struct hstate *h, unsigned long unused_resv_pages) |
e4e574b76 hugetlb: Try to g... |
1393 |
{ |
e4e574b76 hugetlb: Try to g... |
1394 |
unsigned long nr_pages; |
ac09b3a15 hugetlb: close a ... |
1395 |
/* Uncommit the reservation */ |
a55164389 hugetlb: modular ... |
1396 |
h->resv_huge_pages -= unused_resv_pages; |
ac09b3a15 hugetlb: close a ... |
1397 |
|
aa888a749 hugetlb: support ... |
1398 |
/* Cannot return gigantic pages currently */ |
bae7f4ae1 hugetlb: add hsta... |
1399 |
if (hstate_is_gigantic(h)) |
aa888a749 hugetlb: support ... |
1400 |
return; |
a55164389 hugetlb: modular ... |
1401 |
nr_pages = min(unused_resv_pages, h->surplus_huge_pages); |
e4e574b76 hugetlb: Try to g... |
1402 |
|
685f34570 hugetlb: use free... |
1403 1404 |
/* * We want to release as many surplus pages as possible, spread |
9b5e5d0fd hugetlb: use only... |
1405 1406 1407 1408 1409 |
* evenly across all nodes with memory. Iterate across these nodes * until we can no longer free unreserved surplus pages. This occurs * when the nodes with surplus pages have no free pages. * free_pool_huge_page() will balance the the freed pages across the * on-line nodes with memory and will handle the hstate accounting. |
685f34570 hugetlb: use free... |
1410 1411 |
*/ while (nr_pages--) { |
8cebfcd07 hugetlb: use N_ME... |
1412 |
if (!free_pool_huge_page(h, &node_states[N_MEMORY], 1)) |
685f34570 hugetlb: use free... |
1413 |
break; |
7848a4bf5 mm/hugetlb.c: add... |
1414 |
cond_resched_lock(&hugetlb_lock); |
e4e574b76 hugetlb: Try to g... |
1415 1416 |
} } |
c37f9fb11 hugetlb: allow hu... |
1417 |
/* |
cf3ad20bf mm/hugetlb: compu... |
1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 |
* vma_needs_reservation and vma_commit_reservation are used by the huge * page allocation routines to manage reservations. * * vma_needs_reservation is called to determine if the huge page at addr * within the vma has an associated reservation. If a reservation is * needed, the value 1 is returned. The caller is then responsible for * managing the global reservation and subpool usage counts. After * the huge page has been allocated, vma_commit_reservation is called * to add the page to the reservation map. * * In the normal case, vma_commit_reservation returns the same value * as the preceding vma_needs_reservation call. The only time this * is not the case is if a reserve map was changed between calls. It * is the responsibility of the caller to notice the difference and * take appropriate action. |
c37f9fb11 hugetlb: allow hu... |
1433 |
*/ |
cf3ad20bf mm/hugetlb: compu... |
1434 1435 1436 |
static long __vma_reservation_common(struct hstate *h, struct vm_area_struct *vma, unsigned long addr, bool commit) |
c37f9fb11 hugetlb: allow hu... |
1437 |
{ |
4e35f4838 mm, hugetlb: use ... |
1438 1439 |
struct resv_map *resv; pgoff_t idx; |
cf3ad20bf mm/hugetlb: compu... |
1440 |
long ret; |
c37f9fb11 hugetlb: allow hu... |
1441 |
|
4e35f4838 mm, hugetlb: use ... |
1442 1443 |
resv = vma_resv_map(vma); if (!resv) |
84afd99b8 hugetlb reservati... |
1444 |
return 1; |
c37f9fb11 hugetlb: allow hu... |
1445 |
|
4e35f4838 mm, hugetlb: use ... |
1446 |
idx = vma_hugecache_offset(h, vma, addr); |
cf3ad20bf mm/hugetlb: compu... |
1447 1448 1449 1450 |
if (commit) ret = region_add(resv, idx, idx + 1); else ret = region_chg(resv, idx, idx + 1); |
84afd99b8 hugetlb reservati... |
1451 |
|
4e35f4838 mm, hugetlb: use ... |
1452 |
if (vma->vm_flags & VM_MAYSHARE) |
cf3ad20bf mm/hugetlb: compu... |
1453 |
return ret; |
4e35f4838 mm, hugetlb: use ... |
1454 |
else |
cf3ad20bf mm/hugetlb: compu... |
1455 |
return ret < 0 ? ret : 0; |
c37f9fb11 hugetlb: allow hu... |
1456 |
} |
cf3ad20bf mm/hugetlb: compu... |
1457 1458 |
static long vma_needs_reservation(struct hstate *h, |
a55164389 hugetlb: modular ... |
1459 |
struct vm_area_struct *vma, unsigned long addr) |
c37f9fb11 hugetlb: allow hu... |
1460 |
{ |
cf3ad20bf mm/hugetlb: compu... |
1461 1462 |
return __vma_reservation_common(h, vma, addr, false); } |
84afd99b8 hugetlb reservati... |
1463 |
|
cf3ad20bf mm/hugetlb: compu... |
1464 1465 1466 1467 |
static long vma_commit_reservation(struct hstate *h, struct vm_area_struct *vma, unsigned long addr) { return __vma_reservation_common(h, vma, addr, true); |
c37f9fb11 hugetlb: allow hu... |
1468 |
} |
a1e78772d hugetlb: reserve ... |
1469 |
static struct page *alloc_huge_page(struct vm_area_struct *vma, |
04f2cbe35 hugetlb: guarante... |
1470 |
unsigned long addr, int avoid_reserve) |
1da177e4c Linux-2.6.12-rc2 |
1471 |
{ |
90481622d hugepages: fix us... |
1472 |
struct hugepage_subpool *spool = subpool_vma(vma); |
a55164389 hugetlb: modular ... |
1473 |
struct hstate *h = hstate_vma(vma); |
348ea204c hugetlb: split al... |
1474 |
struct page *page; |
33039678c mm/hugetlb: handl... |
1475 |
long chg, commit; |
6d76dcf40 hugetlb/cgroup: a... |
1476 1477 |
int ret, idx; struct hugetlb_cgroup *h_cg; |
a1e78772d hugetlb: reserve ... |
1478 |
|
6d76dcf40 hugetlb/cgroup: a... |
1479 |
idx = hstate_index(h); |
a1e78772d hugetlb: reserve ... |
1480 |
/* |
90481622d hugepages: fix us... |
1481 1482 1483 1484 1485 1486 |
* Processes that did not create the mapping will have no * reserves and will not have accounted against subpool * limit. Check that the subpool limit can be made before * satisfying the allocation MAP_NORESERVE mappings may also * need pages and subpool limit allocated allocated if no reserve * mapping overlaps. |
a1e78772d hugetlb: reserve ... |
1487 |
*/ |
a55164389 hugetlb: modular ... |
1488 |
chg = vma_needs_reservation(h, vma, addr); |
c37f9fb11 hugetlb: allow hu... |
1489 |
if (chg < 0) |
76dcee75c hugetlb: don't us... |
1490 |
return ERR_PTR(-ENOMEM); |
8bb3f12e7 mm, hugetlb: fix ... |
1491 |
if (chg || avoid_reserve) |
1c5ecae3a hugetlbfs: add mi... |
1492 |
if (hugepage_subpool_get_pages(spool, 1) < 0) |
76dcee75c hugetlb: don't us... |
1493 |
return ERR_PTR(-ENOSPC); |
1da177e4c Linux-2.6.12-rc2 |
1494 |
|
6d76dcf40 hugetlb/cgroup: a... |
1495 |
ret = hugetlb_cgroup_charge_cgroup(idx, pages_per_huge_page(h), &h_cg); |
8f34af6f9 mm, hugetlb: move... |
1496 1497 |
if (ret) goto out_subpool_put; |
1da177e4c Linux-2.6.12-rc2 |
1498 |
spin_lock(&hugetlb_lock); |
af0ed73e6 mm, hugetlb: decr... |
1499 |
page = dequeue_huge_page_vma(h, vma, addr, avoid_reserve, chg); |
81a6fcae3 mm, hugetlb: clea... |
1500 |
if (!page) { |
94ae8ba71 hugetlb/cgroup: a... |
1501 |
spin_unlock(&hugetlb_lock); |
bf50bab2b hugetlb: add allo... |
1502 |
page = alloc_buddy_huge_page(h, NUMA_NO_NODE); |
8f34af6f9 mm, hugetlb: move... |
1503 1504 |
if (!page) goto out_uncharge_cgroup; |
79dbb2368 hugetlb: move all... |
1505 1506 |
spin_lock(&hugetlb_lock); list_move(&page->lru, &h->hugepage_activelist); |
81a6fcae3 mm, hugetlb: clea... |
1507 |
/* Fall through */ |
68842c9b9 hugetlbfs: fix qu... |
1508 |
} |
81a6fcae3 mm, hugetlb: clea... |
1509 1510 |
hugetlb_cgroup_commit_charge(idx, pages_per_huge_page(h), h_cg, page); spin_unlock(&hugetlb_lock); |
348ea204c hugetlb: split al... |
1511 |
|
90481622d hugepages: fix us... |
1512 |
set_page_private(page, (unsigned long)spool); |
90d8b7e61 hugetlb: enforce ... |
1513 |
|
33039678c mm/hugetlb: handl... |
1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 |
commit = vma_commit_reservation(h, vma, addr); if (unlikely(chg > commit)) { /* * The page was added to the reservation map between * vma_needs_reservation and vma_commit_reservation. * This indicates a race with hugetlb_reserve_pages. * Adjust for the subpool count incremented above AND * in hugetlb_reserve_pages for the same page. Also, * the reservation count added in hugetlb_reserve_pages * no longer applies. */ long rsv_adjust; rsv_adjust = hugepage_subpool_put_pages(spool, 1); hugetlb_acct_memory(h, -rsv_adjust); } |
90d8b7e61 hugetlb: enforce ... |
1530 |
return page; |
8f34af6f9 mm, hugetlb: move... |
1531 1532 1533 1534 1535 1536 1537 |
out_uncharge_cgroup: hugetlb_cgroup_uncharge_cgroup(idx, pages_per_huge_page(h), h_cg); out_subpool_put: if (chg || avoid_reserve) hugepage_subpool_put_pages(spool, 1); return ERR_PTR(-ENOSPC); |
b45b5bd65 [PATCH] hugepage:... |
1538 |
} |
74060e4d7 mm: mbind: add hu... |
1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 |
/* * alloc_huge_page()'s wrapper which simply returns the page if allocation * succeeds, otherwise NULL. This function is called from new_vma_page(), * where no ERR_VALUE is expected to be returned. */ struct page *alloc_huge_page_noerr(struct vm_area_struct *vma, unsigned long addr, int avoid_reserve) { struct page *page = alloc_huge_page(vma, addr, avoid_reserve); if (IS_ERR(page)) page = NULL; return page; } |
91f47662d mm: hugetlb: remo... |
1552 |
int __weak alloc_bootmem_huge_page(struct hstate *h) |
aa888a749 hugetlb: support ... |
1553 1554 |
{ struct huge_bootmem_page *m; |
b22610268 mm, hugetlb: fix ... |
1555 |
int nr_nodes, node; |
aa888a749 hugetlb: support ... |
1556 |
|
b22610268 mm, hugetlb: fix ... |
1557 |
for_each_node_mask_to_alloc(h, nr_nodes, node, &node_states[N_MEMORY]) { |
aa888a749 hugetlb: support ... |
1558 |
void *addr; |
8b89a1169 mm/hugetlb.c: use... |
1559 1560 1561 |
addr = memblock_virt_alloc_try_nid_nopanic( huge_page_size(h), huge_page_size(h), 0, BOOTMEM_ALLOC_ACCESSIBLE, node); |
aa888a749 hugetlb: support ... |
1562 1563 1564 1565 1566 1567 1568 |
if (addr) { /* * Use the beginning of the huge page to store the * huge_bootmem_page struct (until gather_bootmem * puts them into the mem_map). */ m = addr; |
91f47662d mm: hugetlb: remo... |
1569 |
goto found; |
aa888a749 hugetlb: support ... |
1570 |
} |
aa888a749 hugetlb: support ... |
1571 1572 1573 1574 |
} return 0; found: |
df994ead5 hugetlb: alloc_bo... |
1575 |
BUG_ON(!IS_ALIGNED(virt_to_phys(m), huge_page_size(h))); |
aa888a749 hugetlb: support ... |
1576 1577 1578 1579 1580 |
/* Put them into a private list first because mem_map is not up yet */ list_add(&m->list, &huge_boot_pages); m->hstate = h; return 1; } |
f412c97ab mm, hugetlb: mark... |
1581 |
static void __init prep_compound_huge_page(struct page *page, int order) |
18229df5b hugetlb: pull gig... |
1582 1583 1584 1585 1586 1587 |
{ if (unlikely(order > (MAX_ORDER - 1))) prep_compound_gigantic_page(page, order); else prep_compound_page(page, order); } |
aa888a749 hugetlb: support ... |
1588 1589 1590 1591 1592 1593 |
/* Put bootmem huge pages into the standard lists after mem_map is up */ static void __init gather_bootmem_prealloc(void) { struct huge_bootmem_page *m; list_for_each_entry(m, &huge_boot_pages, list) { |
aa888a749 hugetlb: support ... |
1594 |
struct hstate *h = m->hstate; |
ee8f248d2 hugetlb: add phys... |
1595 1596 1597 1598 |
struct page *page; #ifdef CONFIG_HIGHMEM page = pfn_to_page(m->phys >> PAGE_SHIFT); |
8b89a1169 mm/hugetlb.c: use... |
1599 1600 |
memblock_free_late(__pa(m), sizeof(struct huge_bootmem_page)); |
ee8f248d2 hugetlb: add phys... |
1601 1602 1603 |
#else page = virt_to_page(m); #endif |
aa888a749 hugetlb: support ... |
1604 |
WARN_ON(page_count(page) != 1); |
18229df5b hugetlb: pull gig... |
1605 |
prep_compound_huge_page(page, h->order); |
ef5a22be2 mm: hugetlb: init... |
1606 |
WARN_ON(PageReserved(page)); |
aa888a749 hugetlb: support ... |
1607 |
prep_new_huge_page(h, page, page_to_nid(page)); |
b0320c7b7 mm: fix negative ... |
1608 1609 1610 1611 1612 1613 |
/* * If we had gigantic hugepages allocated at boot time, we need * to restore the 'stolen' pages to totalram_pages in order to * fix confusing memory reports from free(1) and another * side-effects, like CommitLimit going negative. */ |
bae7f4ae1 hugetlb: add hsta... |
1614 |
if (hstate_is_gigantic(h)) |
3dcc0571c mm: correctly upd... |
1615 |
adjust_managed_page_count(page, 1 << h->order); |
aa888a749 hugetlb: support ... |
1616 1617 |
} } |
8faa8b077 hugetlb: support ... |
1618 |
static void __init hugetlb_hstate_alloc_pages(struct hstate *h) |
1da177e4c Linux-2.6.12-rc2 |
1619 1620 |
{ unsigned long i; |
a55164389 hugetlb: modular ... |
1621 |
|
e5ff21594 hugetlb: multiple... |
1622 |
for (i = 0; i < h->max_huge_pages; ++i) { |
bae7f4ae1 hugetlb: add hsta... |
1623 |
if (hstate_is_gigantic(h)) { |
aa888a749 hugetlb: support ... |
1624 1625 |
if (!alloc_bootmem_huge_page(h)) break; |
9b5e5d0fd hugetlb: use only... |
1626 |
} else if (!alloc_fresh_huge_page(h, |
8cebfcd07 hugetlb: use N_ME... |
1627 |
&node_states[N_MEMORY])) |
1da177e4c Linux-2.6.12-rc2 |
1628 |
break; |
1da177e4c Linux-2.6.12-rc2 |
1629 |
} |
8faa8b077 hugetlb: support ... |
1630 |
h->max_huge_pages = i; |
e5ff21594 hugetlb: multiple... |
1631 1632 1633 1634 1635 1636 1637 |
} static void __init hugetlb_init_hstates(void) { struct hstate *h; for_each_hstate(h) { |
641844f56 mm/hugetlb: intro... |
1638 1639 |
if (minimum_order > huge_page_order(h)) minimum_order = huge_page_order(h); |
8faa8b077 hugetlb: support ... |
1640 |
/* oversize hugepages were init'ed in early boot */ |
bae7f4ae1 hugetlb: add hsta... |
1641 |
if (!hstate_is_gigantic(h)) |
8faa8b077 hugetlb: support ... |
1642 |
hugetlb_hstate_alloc_pages(h); |
e5ff21594 hugetlb: multiple... |
1643 |
} |
641844f56 mm/hugetlb: intro... |
1644 |
VM_BUG_ON(minimum_order == UINT_MAX); |
e5ff21594 hugetlb: multiple... |
1645 |
} |
4abd32dba hugetlb: printk c... |
1646 1647 1648 1649 1650 1651 1652 1653 1654 1655 |
static char * __init memfmt(char *buf, unsigned long n) { if (n >= (1UL << 30)) sprintf(buf, "%lu GB", n >> 30); else if (n >= (1UL << 20)) sprintf(buf, "%lu MB", n >> 20); else sprintf(buf, "%lu KB", n >> 10); return buf; } |
e5ff21594 hugetlb: multiple... |
1656 1657 1658 1659 1660 |
static void __init report_hugepages(void) { struct hstate *h; for_each_hstate(h) { |
4abd32dba hugetlb: printk c... |
1661 |
char buf[32]; |
ffb22af5b mm/hugetlb.c: con... |
1662 1663 |
pr_info("HugeTLB registered %s page size, pre-allocated %ld pages ", |
4abd32dba hugetlb: printk c... |
1664 1665 |
memfmt(buf, huge_page_size(h)), h->free_huge_pages); |
e5ff21594 hugetlb: multiple... |
1666 1667 |
} } |
1da177e4c Linux-2.6.12-rc2 |
1668 |
#ifdef CONFIG_HIGHMEM |
6ae11b278 hugetlb: add node... |
1669 1670 |
static void try_to_free_low(struct hstate *h, unsigned long count, nodemask_t *nodes_allowed) |
1da177e4c Linux-2.6.12-rc2 |
1671 |
{ |
4415cc8df [PATCH] Hugepages... |
1672 |
int i; |
bae7f4ae1 hugetlb: add hsta... |
1673 |
if (hstate_is_gigantic(h)) |
aa888a749 hugetlb: support ... |
1674 |
return; |
6ae11b278 hugetlb: add node... |
1675 |
for_each_node_mask(i, *nodes_allowed) { |
1da177e4c Linux-2.6.12-rc2 |
1676 |
struct page *page, *next; |
a55164389 hugetlb: modular ... |
1677 1678 1679 |
struct list_head *freel = &h->hugepage_freelists[i]; list_for_each_entry_safe(page, next, freel, lru) { if (count >= h->nr_huge_pages) |
6b0c880df hugetlb: fix pool... |
1680 |
return; |
1da177e4c Linux-2.6.12-rc2 |
1681 1682 1683 |
if (PageHighMem(page)) continue; list_del(&page->lru); |
e5ff21594 hugetlb: multiple... |
1684 |
update_and_free_page(h, page); |
a55164389 hugetlb: modular ... |
1685 1686 |
h->free_huge_pages--; h->free_huge_pages_node[page_to_nid(page)]--; |
1da177e4c Linux-2.6.12-rc2 |
1687 1688 1689 1690 |
} } } #else |
6ae11b278 hugetlb: add node... |
1691 1692 |
static inline void try_to_free_low(struct hstate *h, unsigned long count, nodemask_t *nodes_allowed) |
1da177e4c Linux-2.6.12-rc2 |
1693 1694 1695 |
{ } #endif |
20a0307c0 mm: introduce Pag... |
1696 1697 1698 1699 1700 |
/* * Increment or decrement surplus_huge_pages. Keep node-specific counters * balanced by operating on them in a round-robin fashion. * Returns 1 if an adjustment was made. */ |
6ae11b278 hugetlb: add node... |
1701 1702 |
static int adjust_pool_surplus(struct hstate *h, nodemask_t *nodes_allowed, int delta) |
20a0307c0 mm: introduce Pag... |
1703 |
{ |
b22610268 mm, hugetlb: fix ... |
1704 |
int nr_nodes, node; |
20a0307c0 mm: introduce Pag... |
1705 1706 |
VM_BUG_ON(delta != -1 && delta != 1); |
20a0307c0 mm: introduce Pag... |
1707 |
|
b22610268 mm, hugetlb: fix ... |
1708 1709 1710 1711 |
if (delta < 0) { for_each_node_mask_to_alloc(h, nr_nodes, node, nodes_allowed) { if (h->surplus_huge_pages_node[node]) goto found; |
e8c5c8249 hugetlb: balance ... |
1712 |
} |
b22610268 mm, hugetlb: fix ... |
1713 1714 1715 1716 1717 |
} else { for_each_node_mask_to_free(h, nr_nodes, node, nodes_allowed) { if (h->surplus_huge_pages_node[node] < h->nr_huge_pages_node[node]) goto found; |
e8c5c8249 hugetlb: balance ... |
1718 |
} |
b22610268 mm, hugetlb: fix ... |
1719 1720 |
} return 0; |
20a0307c0 mm: introduce Pag... |
1721 |
|
b22610268 mm, hugetlb: fix ... |
1722 1723 1724 1725 |
found: h->surplus_huge_pages += delta; h->surplus_huge_pages_node[node] += delta; return 1; |
20a0307c0 mm: introduce Pag... |
1726 |
} |
a55164389 hugetlb: modular ... |
1727 |
#define persistent_huge_pages(h) (h->nr_huge_pages - h->surplus_huge_pages) |
6ae11b278 hugetlb: add node... |
1728 1729 |
static unsigned long set_max_huge_pages(struct hstate *h, unsigned long count, nodemask_t *nodes_allowed) |
1da177e4c Linux-2.6.12-rc2 |
1730 |
{ |
7893d1d50 hugetlb: Try to g... |
1731 |
unsigned long min_count, ret; |
1da177e4c Linux-2.6.12-rc2 |
1732 |
|
944d9fec8 hugetlb: add supp... |
1733 |
if (hstate_is_gigantic(h) && !gigantic_page_supported()) |
aa888a749 hugetlb: support ... |
1734 |
return h->max_huge_pages; |
7893d1d50 hugetlb: Try to g... |
1735 1736 1737 1738 |
/* * Increase the pool size * First take pages out of surplus state. Then make up the * remaining difference by allocating fresh huge pages. |
d1c3fb1f8 hugetlb: introduc... |
1739 1740 1741 1742 1743 1744 |
* * We might race with alloc_buddy_huge_page() here and be unable * to convert a surplus huge page to a normal huge page. That is * not critical, though, it just means the overall size of the * pool might be one hugepage larger than it needs to be, but * within all the constraints specified by the sysctls. |
7893d1d50 hugetlb: Try to g... |
1745 |
*/ |
1da177e4c Linux-2.6.12-rc2 |
1746 |
spin_lock(&hugetlb_lock); |
a55164389 hugetlb: modular ... |
1747 |
while (h->surplus_huge_pages && count > persistent_huge_pages(h)) { |
6ae11b278 hugetlb: add node... |
1748 |
if (!adjust_pool_surplus(h, nodes_allowed, -1)) |
7893d1d50 hugetlb: Try to g... |
1749 1750 |
break; } |
a55164389 hugetlb: modular ... |
1751 |
while (count > persistent_huge_pages(h)) { |
7893d1d50 hugetlb: Try to g... |
1752 1753 1754 1755 1756 1757 |
/* * If this allocation races such that we no longer need the * page, free_huge_page will handle it by freeing the page * and reducing the surplus. */ spin_unlock(&hugetlb_lock); |
944d9fec8 hugetlb: add supp... |
1758 1759 1760 1761 |
if (hstate_is_gigantic(h)) ret = alloc_fresh_gigantic_page(h, nodes_allowed); else ret = alloc_fresh_huge_page(h, nodes_allowed); |
7893d1d50 hugetlb: Try to g... |
1762 1763 1764 |
spin_lock(&hugetlb_lock); if (!ret) goto out; |
536240f2b hugetlb: abort a ... |
1765 1766 1767 |
/* Bail for signals. Probably ctrl-c from user */ if (signal_pending(current)) goto out; |
7893d1d50 hugetlb: Try to g... |
1768 |
} |
7893d1d50 hugetlb: Try to g... |
1769 1770 1771 1772 1773 1774 1775 |
/* * Decrease the pool size * First return free pages to the buddy allocator (being careful * to keep enough around to satisfy reservations). Then place * pages into surplus state as needed so the pool will shrink * to the desired size as pages become free. |
d1c3fb1f8 hugetlb: introduc... |
1776 1777 1778 1779 1780 1781 1782 1783 |
* * By placing pages into the surplus state independent of the * overcommit value, we are allowing the surplus pool size to * exceed overcommit. There are few sane options here. Since * alloc_buddy_huge_page() is checking the global counter, * though, we'll note that we're not allowed to exceed surplus * and won't grow the pool anywhere else. Not until one of the * sysctls are changed, or the surplus pages go out of use. |
7893d1d50 hugetlb: Try to g... |
1784 |
*/ |
a55164389 hugetlb: modular ... |
1785 |
min_count = h->resv_huge_pages + h->nr_huge_pages - h->free_huge_pages; |
6b0c880df hugetlb: fix pool... |
1786 |
min_count = max(count, min_count); |
6ae11b278 hugetlb: add node... |
1787 |
try_to_free_low(h, min_count, nodes_allowed); |
a55164389 hugetlb: modular ... |
1788 |
while (min_count < persistent_huge_pages(h)) { |
6ae11b278 hugetlb: add node... |
1789 |
if (!free_pool_huge_page(h, nodes_allowed, 0)) |
1da177e4c Linux-2.6.12-rc2 |
1790 |
break; |
55f67141a mm: hugetlb: fix ... |
1791 |
cond_resched_lock(&hugetlb_lock); |
1da177e4c Linux-2.6.12-rc2 |
1792 |
} |
a55164389 hugetlb: modular ... |
1793 |
while (count < persistent_huge_pages(h)) { |
6ae11b278 hugetlb: add node... |
1794 |
if (!adjust_pool_surplus(h, nodes_allowed, 1)) |
7893d1d50 hugetlb: Try to g... |
1795 1796 1797 |
break; } out: |
a55164389 hugetlb: modular ... |
1798 |
ret = persistent_huge_pages(h); |
1da177e4c Linux-2.6.12-rc2 |
1799 |
spin_unlock(&hugetlb_lock); |
7893d1d50 hugetlb: Try to g... |
1800 |
return ret; |
1da177e4c Linux-2.6.12-rc2 |
1801 |
} |
a34378701 hugetlb: new sysf... |
1802 1803 1804 1805 1806 1807 1808 1809 1810 |
#define HSTATE_ATTR_RO(_name) \ static struct kobj_attribute _name##_attr = __ATTR_RO(_name) #define HSTATE_ATTR(_name) \ static struct kobj_attribute _name##_attr = \ __ATTR(_name, 0644, _name##_show, _name##_store) static struct kobject *hugepages_kobj; static struct kobject *hstate_kobjs[HUGE_MAX_HSTATE]; |
9a3052306 hugetlb: add per ... |
1811 1812 1813 |
static struct hstate *kobj_to_node_hstate(struct kobject *kobj, int *nidp); static struct hstate *kobj_to_hstate(struct kobject *kobj, int *nidp) |
a34378701 hugetlb: new sysf... |
1814 1815 |
{ int i; |
9a3052306 hugetlb: add per ... |
1816 |
|
a34378701 hugetlb: new sysf... |
1817 |
for (i = 0; i < HUGE_MAX_HSTATE; i++) |
9a3052306 hugetlb: add per ... |
1818 1819 1820 |
if (hstate_kobjs[i] == kobj) { if (nidp) *nidp = NUMA_NO_NODE; |
a34378701 hugetlb: new sysf... |
1821 |
return &hstates[i]; |
9a3052306 hugetlb: add per ... |
1822 1823 1824 |
} return kobj_to_node_hstate(kobj, nidp); |
a34378701 hugetlb: new sysf... |
1825 |
} |
06808b082 hugetlb: derive h... |
1826 |
static ssize_t nr_hugepages_show_common(struct kobject *kobj, |
a34378701 hugetlb: new sysf... |
1827 1828 |
struct kobj_attribute *attr, char *buf) { |
9a3052306 hugetlb: add per ... |
1829 1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840 |
struct hstate *h; unsigned long nr_huge_pages; int nid; h = kobj_to_hstate(kobj, &nid); if (nid == NUMA_NO_NODE) nr_huge_pages = h->nr_huge_pages; else nr_huge_pages = h->nr_huge_pages_node[nid]; return sprintf(buf, "%lu ", nr_huge_pages); |
a34378701 hugetlb: new sysf... |
1841 |
} |
adbe8726d hugetlb: do not a... |
1842 |
|
238d3c13f mm, hugetlb: gene... |
1843 1844 1845 |
static ssize_t __nr_hugepages_store_common(bool obey_mempolicy, struct hstate *h, int nid, unsigned long count, size_t len) |
a34378701 hugetlb: new sysf... |
1846 1847 |
{ int err; |
bad44b5be mm: add gfp flags... |
1848 |
NODEMASK_ALLOC(nodemask_t, nodes_allowed, GFP_KERNEL | __GFP_NORETRY); |
a34378701 hugetlb: new sysf... |
1849 |
|
944d9fec8 hugetlb: add supp... |
1850 |
if (hstate_is_gigantic(h) && !gigantic_page_supported()) { |
adbe8726d hugetlb: do not a... |
1851 1852 1853 |
err = -EINVAL; goto out; } |
9a3052306 hugetlb: add per ... |
1854 1855 1856 1857 1858 1859 1860 |
if (nid == NUMA_NO_NODE) { /* * global hstate attribute */ if (!(obey_mempolicy && init_nodemask_of_mempolicy(nodes_allowed))) { NODEMASK_FREE(nodes_allowed); |
8cebfcd07 hugetlb: use N_ME... |
1861 |
nodes_allowed = &node_states[N_MEMORY]; |
9a3052306 hugetlb: add per ... |
1862 1863 1864 1865 1866 1867 1868 1869 1870 |
} } else if (nodes_allowed) { /* * per node hstate attribute: adjust count to global, * but restrict alloc/free to the specified node. */ count += h->nr_huge_pages - h->nr_huge_pages_node[nid]; init_nodemask_of_node(nodes_allowed, nid); } else |
8cebfcd07 hugetlb: use N_ME... |
1871 |
nodes_allowed = &node_states[N_MEMORY]; |
9a3052306 hugetlb: add per ... |
1872 |
|
06808b082 hugetlb: derive h... |
1873 |
h->max_huge_pages = set_max_huge_pages(h, count, nodes_allowed); |
a34378701 hugetlb: new sysf... |
1874 |
|
8cebfcd07 hugetlb: use N_ME... |
1875 |
if (nodes_allowed != &node_states[N_MEMORY]) |
06808b082 hugetlb: derive h... |
1876 1877 1878 |
NODEMASK_FREE(nodes_allowed); return len; |
adbe8726d hugetlb: do not a... |
1879 1880 1881 |
out: NODEMASK_FREE(nodes_allowed); return err; |
06808b082 hugetlb: derive h... |
1882 |
} |
238d3c13f mm, hugetlb: gene... |
1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894 1895 1896 1897 1898 |
static ssize_t nr_hugepages_store_common(bool obey_mempolicy, struct kobject *kobj, const char *buf, size_t len) { struct hstate *h; unsigned long count; int nid; int err; err = kstrtoul(buf, 10, &count); if (err) return err; h = kobj_to_hstate(kobj, &nid); return __nr_hugepages_store_common(obey_mempolicy, h, nid, count, len); } |
06808b082 hugetlb: derive h... |
1899 1900 1901 1902 1903 1904 1905 1906 1907 |
static ssize_t nr_hugepages_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return nr_hugepages_show_common(kobj, attr, buf); } static ssize_t nr_hugepages_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t len) { |
238d3c13f mm, hugetlb: gene... |
1908 |
return nr_hugepages_store_common(false, kobj, buf, len); |
a34378701 hugetlb: new sysf... |
1909 1910 |
} HSTATE_ATTR(nr_hugepages); |
06808b082 hugetlb: derive h... |
1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 |
#ifdef CONFIG_NUMA /* * hstate attribute for optionally mempolicy-based constraint on persistent * huge page alloc/free. */ static ssize_t nr_hugepages_mempolicy_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { return nr_hugepages_show_common(kobj, attr, buf); } static ssize_t nr_hugepages_mempolicy_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t len) { |
238d3c13f mm, hugetlb: gene... |
1926 |
return nr_hugepages_store_common(true, kobj, buf, len); |
06808b082 hugetlb: derive h... |
1927 1928 1929 |
} HSTATE_ATTR(nr_hugepages_mempolicy); #endif |
a34378701 hugetlb: new sysf... |
1930 1931 1932 |
static ssize_t nr_overcommit_hugepages_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { |
9a3052306 hugetlb: add per ... |
1933 |
struct hstate *h = kobj_to_hstate(kobj, NULL); |
a34378701 hugetlb: new sysf... |
1934 1935 1936 |
return sprintf(buf, "%lu ", h->nr_overcommit_huge_pages); } |
adbe8726d hugetlb: do not a... |
1937 |
|
a34378701 hugetlb: new sysf... |
1938 1939 1940 1941 1942 |
static ssize_t nr_overcommit_hugepages_store(struct kobject *kobj, struct kobj_attribute *attr, const char *buf, size_t count) { int err; unsigned long input; |
9a3052306 hugetlb: add per ... |
1943 |
struct hstate *h = kobj_to_hstate(kobj, NULL); |
a34378701 hugetlb: new sysf... |
1944 |
|
bae7f4ae1 hugetlb: add hsta... |
1945 |
if (hstate_is_gigantic(h)) |
adbe8726d hugetlb: do not a... |
1946 |
return -EINVAL; |
3dbb95f78 mm: replace stric... |
1947 |
err = kstrtoul(buf, 10, &input); |
a34378701 hugetlb: new sysf... |
1948 |
if (err) |
73ae31e59 hugetlb: fix hand... |
1949 |
return err; |
a34378701 hugetlb: new sysf... |
1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 |
spin_lock(&hugetlb_lock); h->nr_overcommit_huge_pages = input; spin_unlock(&hugetlb_lock); return count; } HSTATE_ATTR(nr_overcommit_hugepages); static ssize_t free_hugepages_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { |
9a3052306 hugetlb: add per ... |
1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 |
struct hstate *h; unsigned long free_huge_pages; int nid; h = kobj_to_hstate(kobj, &nid); if (nid == NUMA_NO_NODE) free_huge_pages = h->free_huge_pages; else free_huge_pages = h->free_huge_pages_node[nid]; return sprintf(buf, "%lu ", free_huge_pages); |
a34378701 hugetlb: new sysf... |
1974 1975 1976 1977 1978 1979 |
} HSTATE_ATTR_RO(free_hugepages); static ssize_t resv_hugepages_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { |
9a3052306 hugetlb: add per ... |
1980 |
struct hstate *h = kobj_to_hstate(kobj, NULL); |
a34378701 hugetlb: new sysf... |
1981 1982 1983 1984 1985 1986 1987 1988 |
return sprintf(buf, "%lu ", h->resv_huge_pages); } HSTATE_ATTR_RO(resv_hugepages); static ssize_t surplus_hugepages_show(struct kobject *kobj, struct kobj_attribute *attr, char *buf) { |
9a3052306 hugetlb: add per ... |
1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 |
struct hstate *h; unsigned long surplus_huge_pages; int nid; h = kobj_to_hstate(kobj, &nid); if (nid == NUMA_NO_NODE) surplus_huge_pages = h->surplus_huge_pages; else surplus_huge_pages = h->surplus_huge_pages_node[nid]; return sprintf(buf, "%lu ", surplus_huge_pages); |
a34378701 hugetlb: new sysf... |
2001 2002 2003 2004 2005 2006 2007 2008 2009 |
} HSTATE_ATTR_RO(surplus_hugepages); static struct attribute *hstate_attrs[] = { &nr_hugepages_attr.attr, &nr_overcommit_hugepages_attr.attr, &free_hugepages_attr.attr, &resv_hugepages_attr.attr, &surplus_hugepages_attr.attr, |
06808b082 hugetlb: derive h... |
2010 2011 2012 |
#ifdef CONFIG_NUMA &nr_hugepages_mempolicy_attr.attr, #endif |
a34378701 hugetlb: new sysf... |
2013 2014 2015 2016 2017 2018 |
NULL, }; static struct attribute_group hstate_attr_group = { .attrs = hstate_attrs, }; |
094e9539b hugetlb: fix sect... |
2019 2020 2021 |
static int hugetlb_sysfs_add_hstate(struct hstate *h, struct kobject *parent, struct kobject **hstate_kobjs, struct attribute_group *hstate_attr_group) |
a34378701 hugetlb: new sysf... |
2022 2023 |
{ int retval; |
972dc4de1 hugetlb: add an i... |
2024 |
int hi = hstate_index(h); |
a34378701 hugetlb: new sysf... |
2025 |
|
9a3052306 hugetlb: add per ... |
2026 2027 |
hstate_kobjs[hi] = kobject_create_and_add(h->name, parent); if (!hstate_kobjs[hi]) |
a34378701 hugetlb: new sysf... |
2028 |
return -ENOMEM; |
9a3052306 hugetlb: add per ... |
2029 |
retval = sysfs_create_group(hstate_kobjs[hi], hstate_attr_group); |
a34378701 hugetlb: new sysf... |
2030 |
if (retval) |
9a3052306 hugetlb: add per ... |
2031 |
kobject_put(hstate_kobjs[hi]); |
a34378701 hugetlb: new sysf... |
2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 |
return retval; } static void __init hugetlb_sysfs_init(void) { struct hstate *h; int err; hugepages_kobj = kobject_create_and_add("hugepages", mm_kobj); if (!hugepages_kobj) return; for_each_hstate(h) { |
9a3052306 hugetlb: add per ... |
2046 2047 |
err = hugetlb_sysfs_add_hstate(h, hugepages_kobj, hstate_kobjs, &hstate_attr_group); |
a34378701 hugetlb: new sysf... |
2048 |
if (err) |
ffb22af5b mm/hugetlb.c: con... |
2049 |
pr_err("Hugetlb: Unable to add hstate %s", h->name); |
a34378701 hugetlb: new sysf... |
2050 2051 |
} } |
9a3052306 hugetlb: add per ... |
2052 2053 2054 2055 |
#ifdef CONFIG_NUMA /* * node_hstate/s - associate per node hstate attributes, via their kobjects, |
10fbcf4c6 convert 'memory' ... |
2056 2057 2058 |
* with node devices in node_devices[] using a parallel array. The array * index of a node device or _hstate == node id. * This is here to avoid any static dependency of the node device driver, in |
9a3052306 hugetlb: add per ... |
2059 2060 2061 2062 2063 2064 2065 2066 2067 |
* the base kernel, on the hugetlb module. */ struct node_hstate { struct kobject *hugepages_kobj; struct kobject *hstate_kobjs[HUGE_MAX_HSTATE]; }; struct node_hstate node_hstates[MAX_NUMNODES]; /* |
10fbcf4c6 convert 'memory' ... |
2068 |
* A subset of global hstate attributes for node devices |
9a3052306 hugetlb: add per ... |
2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 |
*/ static struct attribute *per_node_hstate_attrs[] = { &nr_hugepages_attr.attr, &free_hugepages_attr.attr, &surplus_hugepages_attr.attr, NULL, }; static struct attribute_group per_node_hstate_attr_group = { .attrs = per_node_hstate_attrs, }; /* |
10fbcf4c6 convert 'memory' ... |
2082 |
* kobj_to_node_hstate - lookup global hstate for node device hstate attr kobj. |
9a3052306 hugetlb: add per ... |
2083 2084 2085 2086 2087 2088 2089 2090 2091 2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 |
* Returns node id via non-NULL nidp. */ static struct hstate *kobj_to_node_hstate(struct kobject *kobj, int *nidp) { int nid; for (nid = 0; nid < nr_node_ids; nid++) { struct node_hstate *nhs = &node_hstates[nid]; int i; for (i = 0; i < HUGE_MAX_HSTATE; i++) if (nhs->hstate_kobjs[i] == kobj) { if (nidp) *nidp = nid; return &hstates[i]; } } BUG(); return NULL; } /* |
10fbcf4c6 convert 'memory' ... |
2105 |
* Unregister hstate attributes from a single node device. |
9a3052306 hugetlb: add per ... |
2106 2107 |
* No-op if no hstate attributes attached. */ |
3cd8b44fa hugetlb: fix spar... |
2108 |
static void hugetlb_unregister_node(struct node *node) |
9a3052306 hugetlb: add per ... |
2109 2110 |
{ struct hstate *h; |
10fbcf4c6 convert 'memory' ... |
2111 |
struct node_hstate *nhs = &node_hstates[node->dev.id]; |
9a3052306 hugetlb: add per ... |
2112 2113 |
if (!nhs->hugepages_kobj) |
9b5e5d0fd hugetlb: use only... |
2114 |
return; /* no hstate attributes */ |
9a3052306 hugetlb: add per ... |
2115 |
|
972dc4de1 hugetlb: add an i... |
2116 2117 2118 2119 2120 |
for_each_hstate(h) { int idx = hstate_index(h); if (nhs->hstate_kobjs[idx]) { kobject_put(nhs->hstate_kobjs[idx]); nhs->hstate_kobjs[idx] = NULL; |
9a3052306 hugetlb: add per ... |
2121 |
} |
972dc4de1 hugetlb: add an i... |
2122 |
} |
9a3052306 hugetlb: add per ... |
2123 2124 2125 2126 2127 2128 |
kobject_put(nhs->hugepages_kobj); nhs->hugepages_kobj = NULL; } /* |
10fbcf4c6 convert 'memory' ... |
2129 |
* hugetlb module exit: unregister hstate attributes from node devices |
9a3052306 hugetlb: add per ... |
2130 2131 2132 2133 2134 2135 2136 |
* that have them. */ static void hugetlb_unregister_all_nodes(void) { int nid; /* |
10fbcf4c6 convert 'memory' ... |
2137 |
* disable node device registrations. |
9a3052306 hugetlb: add per ... |
2138 2139 2140 2141 2142 2143 2144 |
*/ register_hugetlbfs_with_node(NULL, NULL); /* * remove hstate attributes from any nodes that have them. */ for (nid = 0; nid < nr_node_ids; nid++) |
8732794b1 numa: convert sta... |
2145 |
hugetlb_unregister_node(node_devices[nid]); |
9a3052306 hugetlb: add per ... |
2146 2147 2148 |
} /* |
10fbcf4c6 convert 'memory' ... |
2149 |
* Register hstate attributes for a single node device. |
9a3052306 hugetlb: add per ... |
2150 2151 |
* No-op if attributes already registered. */ |
3cd8b44fa hugetlb: fix spar... |
2152 |
static void hugetlb_register_node(struct node *node) |
9a3052306 hugetlb: add per ... |
2153 2154 |
{ struct hstate *h; |
10fbcf4c6 convert 'memory' ... |
2155 |
struct node_hstate *nhs = &node_hstates[node->dev.id]; |
9a3052306 hugetlb: add per ... |
2156 2157 2158 2159 2160 2161 |
int err; if (nhs->hugepages_kobj) return; /* already allocated */ nhs->hugepages_kobj = kobject_create_and_add("hugepages", |
10fbcf4c6 convert 'memory' ... |
2162 |
&node->dev.kobj); |
9a3052306 hugetlb: add per ... |
2163 2164 2165 2166 2167 2168 2169 2170 |
if (!nhs->hugepages_kobj) return; for_each_hstate(h) { err = hugetlb_sysfs_add_hstate(h, nhs->hugepages_kobj, nhs->hstate_kobjs, &per_node_hstate_attr_group); if (err) { |
ffb22af5b mm/hugetlb.c: con... |
2171 2172 2173 |
pr_err("Hugetlb: Unable to add hstate %s for node %d ", h->name, node->dev.id); |
9a3052306 hugetlb: add per ... |
2174 2175 2176 2177 2178 2179 2180 |
hugetlb_unregister_node(node); break; } } } /* |
9b5e5d0fd hugetlb: use only... |
2181 |
* hugetlb init time: register hstate attributes for all registered node |
10fbcf4c6 convert 'memory' ... |
2182 2183 |
* devices of nodes that have memory. All on-line nodes should have * registered their associated device by this time. |
9a3052306 hugetlb: add per ... |
2184 |
*/ |
7d9ca0004 hugetlb: hugetlb_... |
2185 |
static void __init hugetlb_register_all_nodes(void) |
9a3052306 hugetlb: add per ... |
2186 2187 |
{ int nid; |
8cebfcd07 hugetlb: use N_ME... |
2188 |
for_each_node_state(nid, N_MEMORY) { |
8732794b1 numa: convert sta... |
2189 |
struct node *node = node_devices[nid]; |
10fbcf4c6 convert 'memory' ... |
2190 |
if (node->dev.id == nid) |
9a3052306 hugetlb: add per ... |
2191 2192 2193 2194 |
hugetlb_register_node(node); } /* |
10fbcf4c6 convert 'memory' ... |
2195 |
* Let the node device driver know we're here so it can |
9a3052306 hugetlb: add per ... |
2196 2197 2198 2199 2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 2210 2211 2212 2213 2214 2215 |
* [un]register hstate attributes on node hotplug. */ register_hugetlbfs_with_node(hugetlb_register_node, hugetlb_unregister_node); } #else /* !CONFIG_NUMA */ static struct hstate *kobj_to_node_hstate(struct kobject *kobj, int *nidp) { BUG(); if (nidp) *nidp = -1; return NULL; } static void hugetlb_unregister_all_nodes(void) { } static void hugetlb_register_all_nodes(void) { } #endif |
a34378701 hugetlb: new sysf... |
2216 2217 2218 |
static void __exit hugetlb_exit(void) { struct hstate *h; |
9a3052306 hugetlb: add per ... |
2219 |
hugetlb_unregister_all_nodes(); |
a34378701 hugetlb: new sysf... |
2220 |
for_each_hstate(h) { |
972dc4de1 hugetlb: add an i... |
2221 |
kobject_put(hstate_kobjs[hstate_index(h)]); |
a34378701 hugetlb: new sysf... |
2222 2223 2224 |
} kobject_put(hugepages_kobj); |
8382d914e mm, hugetlb: impr... |
2225 |
kfree(htlb_fault_mutex_table); |
a34378701 hugetlb: new sysf... |
2226 2227 2228 2229 2230 |
} module_exit(hugetlb_exit); static int __init hugetlb_init(void) { |
8382d914e mm, hugetlb: impr... |
2231 |
int i; |
457c1b27e hugetlb: ensure h... |
2232 |
if (!hugepages_supported()) |
0ef89d25d mm/hugetlb: don't... |
2233 |
return 0; |
a34378701 hugetlb: new sysf... |
2234 |
|
e11bfbfcb hugetlb: override... |
2235 2236 2237 2238 |
if (!size_to_hstate(default_hstate_size)) { default_hstate_size = HPAGE_SIZE; if (!size_to_hstate(default_hstate_size)) hugetlb_add_hstate(HUGETLB_PAGE_ORDER); |
a34378701 hugetlb: new sysf... |
2239 |
} |
972dc4de1 hugetlb: add an i... |
2240 |
default_hstate_idx = hstate_index(size_to_hstate(default_hstate_size)); |
e11bfbfcb hugetlb: override... |
2241 2242 |
if (default_hstate_max_huge_pages) default_hstate.max_huge_pages = default_hstate_max_huge_pages; |
a34378701 hugetlb: new sysf... |
2243 2244 |
hugetlb_init_hstates(); |
aa888a749 hugetlb: support ... |
2245 |
gather_bootmem_prealloc(); |
a34378701 hugetlb: new sysf... |
2246 2247 2248 |
report_hugepages(); hugetlb_sysfs_init(); |
9a3052306 hugetlb: add per ... |
2249 |
hugetlb_register_all_nodes(); |
7179e7bf4 mm/hugetlb: creat... |
2250 |
hugetlb_cgroup_file_init(); |
9a3052306 hugetlb: add per ... |
2251 |
|
8382d914e mm, hugetlb: impr... |
2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 |
#ifdef CONFIG_SMP num_fault_mutexes = roundup_pow_of_two(8 * num_possible_cpus()); #else num_fault_mutexes = 1; #endif htlb_fault_mutex_table = kmalloc(sizeof(struct mutex) * num_fault_mutexes, GFP_KERNEL); BUG_ON(!htlb_fault_mutex_table); for (i = 0; i < num_fault_mutexes; i++) mutex_init(&htlb_fault_mutex_table[i]); |
a34378701 hugetlb: new sysf... |
2263 2264 2265 2266 2267 2268 2269 2270 |
return 0; } module_init(hugetlb_init); /* Should be called on processing a hugepagesz=... option */ void __init hugetlb_add_hstate(unsigned order) { struct hstate *h; |
8faa8b077 hugetlb: support ... |
2271 |
unsigned long i; |
a34378701 hugetlb: new sysf... |
2272 |
if (size_to_hstate(PAGE_SIZE << order)) { |
ffb22af5b mm/hugetlb.c: con... |
2273 2274 |
pr_warning("hugepagesz= specified twice, ignoring "); |
a34378701 hugetlb: new sysf... |
2275 2276 |
return; } |
47d38344a hugetlb: rename m... |
2277 |
BUG_ON(hugetlb_max_hstate >= HUGE_MAX_HSTATE); |
a34378701 hugetlb: new sysf... |
2278 |
BUG_ON(order == 0); |
47d38344a hugetlb: rename m... |
2279 |
h = &hstates[hugetlb_max_hstate++]; |
a34378701 hugetlb: new sysf... |
2280 2281 |
h->order = order; h->mask = ~((1ULL << (order + PAGE_SHIFT)) - 1); |
8faa8b077 hugetlb: support ... |
2282 2283 2284 2285 |
h->nr_huge_pages = 0; h->free_huge_pages = 0; for (i = 0; i < MAX_NUMNODES; ++i) INIT_LIST_HEAD(&h->hugepage_freelists[i]); |
0edaecfab hugetlb: add a li... |
2286 |
INIT_LIST_HEAD(&h->hugepage_activelist); |
8cebfcd07 hugetlb: use N_ME... |
2287 2288 |
h->next_nid_to_alloc = first_node(node_states[N_MEMORY]); h->next_nid_to_free = first_node(node_states[N_MEMORY]); |
a34378701 hugetlb: new sysf... |
2289 2290 |
snprintf(h->name, HSTATE_NAME_LEN, "hugepages-%lukB", huge_page_size(h)/1024); |
8faa8b077 hugetlb: support ... |
2291 |
|
a34378701 hugetlb: new sysf... |
2292 2293 |
parsed_hstate = h; } |
e11bfbfcb hugetlb: override... |
2294 |
static int __init hugetlb_nrpages_setup(char *s) |
a34378701 hugetlb: new sysf... |
2295 2296 |
{ unsigned long *mhp; |
8faa8b077 hugetlb: support ... |
2297 |
static unsigned long *last_mhp; |
a34378701 hugetlb: new sysf... |
2298 2299 |
/* |
47d38344a hugetlb: rename m... |
2300 |
* !hugetlb_max_hstate means we haven't parsed a hugepagesz= parameter yet, |
a34378701 hugetlb: new sysf... |
2301 2302 |
* so this hugepages= parameter goes to the "default hstate". */ |
47d38344a hugetlb: rename m... |
2303 |
if (!hugetlb_max_hstate) |
a34378701 hugetlb: new sysf... |
2304 2305 2306 |
mhp = &default_hstate_max_huge_pages; else mhp = &parsed_hstate->max_huge_pages; |
8faa8b077 hugetlb: support ... |
2307 |
if (mhp == last_mhp) { |
ffb22af5b mm/hugetlb.c: con... |
2308 2309 2310 |
pr_warning("hugepages= specified twice without " "interleaving hugepagesz=, ignoring "); |
8faa8b077 hugetlb: support ... |
2311 2312 |
return 1; } |
a34378701 hugetlb: new sysf... |
2313 2314 |
if (sscanf(s, "%lu", mhp) <= 0) *mhp = 0; |
8faa8b077 hugetlb: support ... |
2315 2316 2317 2318 2319 |
/* * Global state is always initialized later in hugetlb_init. * But we need to allocate >= MAX_ORDER hstates here early to still * use the bootmem allocator. */ |
47d38344a hugetlb: rename m... |
2320 |
if (hugetlb_max_hstate && parsed_hstate->order >= MAX_ORDER) |
8faa8b077 hugetlb: support ... |
2321 2322 2323 |
hugetlb_hstate_alloc_pages(parsed_hstate); last_mhp = mhp; |
a34378701 hugetlb: new sysf... |
2324 2325 |
return 1; } |
e11bfbfcb hugetlb: override... |
2326 2327 2328 2329 2330 2331 2332 2333 |
__setup("hugepages=", hugetlb_nrpages_setup); static int __init hugetlb_default_setup(char *s) { default_hstate_size = memparse(s, &s); return 1; } __setup("default_hugepagesz=", hugetlb_default_setup); |
a34378701 hugetlb: new sysf... |
2334 |
|
8a2134605 hugetlb: fix CONF... |
2335 2336 2337 2338 2339 2340 2341 2342 2343 2344 2345 2346 |
static unsigned int cpuset_mems_nr(unsigned int *array) { int node; unsigned int nr = 0; for_each_node_mask(node, cpuset_current_mems_allowed) nr += array[node]; return nr; } #ifdef CONFIG_SYSCTL |
06808b082 hugetlb: derive h... |
2347 2348 2349 |
static int hugetlb_sysctl_handler_common(bool obey_mempolicy, struct ctl_table *table, int write, void __user *buffer, size_t *length, loff_t *ppos) |
1da177e4c Linux-2.6.12-rc2 |
2350 |
{ |
e5ff21594 hugetlb: multiple... |
2351 |
struct hstate *h = &default_hstate; |
238d3c13f mm, hugetlb: gene... |
2352 |
unsigned long tmp = h->max_huge_pages; |
08d4a2465 hugetlb: check th... |
2353 |
int ret; |
e5ff21594 hugetlb: multiple... |
2354 |
|
457c1b27e hugetlb: ensure h... |
2355 2356 |
if (!hugepages_supported()) return -ENOTSUPP; |
e5ff21594 hugetlb: multiple... |
2357 2358 |
table->data = &tmp; table->maxlen = sizeof(unsigned long); |
08d4a2465 hugetlb: check th... |
2359 2360 2361 |
ret = proc_doulongvec_minmax(table, write, buffer, length, ppos); if (ret) goto out; |
e5ff21594 hugetlb: multiple... |
2362 |
|
238d3c13f mm, hugetlb: gene... |
2363 2364 2365 |
if (write) ret = __nr_hugepages_store_common(obey_mempolicy, h, NUMA_NO_NODE, tmp, *length); |
08d4a2465 hugetlb: check th... |
2366 2367 |
out: return ret; |
1da177e4c Linux-2.6.12-rc2 |
2368 |
} |
396faf030 Allow huge page a... |
2369 |
|
06808b082 hugetlb: derive h... |
2370 2371 2372 2373 2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 |
int hugetlb_sysctl_handler(struct ctl_table *table, int write, void __user *buffer, size_t *length, loff_t *ppos) { return hugetlb_sysctl_handler_common(false, table, write, buffer, length, ppos); } #ifdef CONFIG_NUMA int hugetlb_mempolicy_sysctl_handler(struct ctl_table *table, int write, void __user *buffer, size_t *length, loff_t *ppos) { return hugetlb_sysctl_handler_common(true, table, write, buffer, length, ppos); } #endif /* CONFIG_NUMA */ |
a3d0c6aa1 hugetlb: add lock... |
2386 |
int hugetlb_overcommit_handler(struct ctl_table *table, int write, |
8d65af789 sysctl: remove "s... |
2387 |
void __user *buffer, |
a3d0c6aa1 hugetlb: add lock... |
2388 2389 |
size_t *length, loff_t *ppos) { |
a55164389 hugetlb: modular ... |
2390 |
struct hstate *h = &default_hstate; |
e5ff21594 hugetlb: multiple... |
2391 |
unsigned long tmp; |
08d4a2465 hugetlb: check th... |
2392 |
int ret; |
e5ff21594 hugetlb: multiple... |
2393 |
|
457c1b27e hugetlb: ensure h... |
2394 2395 |
if (!hugepages_supported()) return -ENOTSUPP; |
c033a93c0 hugetlbfs: correc... |
2396 |
tmp = h->nr_overcommit_huge_pages; |
e5ff21594 hugetlb: multiple... |
2397 |
|
bae7f4ae1 hugetlb: add hsta... |
2398 |
if (write && hstate_is_gigantic(h)) |
adbe8726d hugetlb: do not a... |
2399 |
return -EINVAL; |
e5ff21594 hugetlb: multiple... |
2400 2401 |
table->data = &tmp; table->maxlen = sizeof(unsigned long); |
08d4a2465 hugetlb: check th... |
2402 2403 2404 |
ret = proc_doulongvec_minmax(table, write, buffer, length, ppos); if (ret) goto out; |
e5ff21594 hugetlb: multiple... |
2405 2406 2407 2408 2409 2410 |
if (write) { spin_lock(&hugetlb_lock); h->nr_overcommit_huge_pages = tmp; spin_unlock(&hugetlb_lock); } |
08d4a2465 hugetlb: check th... |
2411 2412 |
out: return ret; |
a3d0c6aa1 hugetlb: add lock... |
2413 |
} |
1da177e4c Linux-2.6.12-rc2 |
2414 |
#endif /* CONFIG_SYSCTL */ |
e1759c215 proc: switch /pro... |
2415 |
void hugetlb_report_meminfo(struct seq_file *m) |
1da177e4c Linux-2.6.12-rc2 |
2416 |
{ |
a55164389 hugetlb: modular ... |
2417 |
struct hstate *h = &default_hstate; |
457c1b27e hugetlb: ensure h... |
2418 2419 |
if (!hugepages_supported()) return; |
e1759c215 proc: switch /pro... |
2420 |
seq_printf(m, |
4f98a2fee vmscan: split LRU... |
2421 2422 2423 2424 2425 2426 2427 2428 2429 2430 |
"HugePages_Total: %5lu " "HugePages_Free: %5lu " "HugePages_Rsvd: %5lu " "HugePages_Surp: %5lu " "Hugepagesize: %8lu kB ", |
a55164389 hugetlb: modular ... |
2431 2432 2433 2434 2435 |
h->nr_huge_pages, h->free_huge_pages, h->resv_huge_pages, h->surplus_huge_pages, 1UL << (huge_page_order(h) + PAGE_SHIFT - 10)); |
1da177e4c Linux-2.6.12-rc2 |
2436 2437 2438 2439 |
} int hugetlb_report_node_meminfo(int nid, char *buf) { |
a55164389 hugetlb: modular ... |
2440 |
struct hstate *h = &default_hstate; |
457c1b27e hugetlb: ensure h... |
2441 2442 |
if (!hugepages_supported()) return 0; |
1da177e4c Linux-2.6.12-rc2 |
2443 2444 2445 |
return sprintf(buf, "Node %d HugePages_Total: %5u " |
a1de09195 hugetlb: indicate... |
2446 2447 2448 2449 |
"Node %d HugePages_Free: %5u " "Node %d HugePages_Surp: %5u ", |
a55164389 hugetlb: modular ... |
2450 2451 2452 |
nid, h->nr_huge_pages_node[nid], nid, h->free_huge_pages_node[nid], nid, h->surplus_huge_pages_node[nid]); |
1da177e4c Linux-2.6.12-rc2 |
2453 |
} |
949f7ec57 mm, hugetlb: incl... |
2454 2455 2456 2457 |
void hugetlb_show_meminfo(void) { struct hstate *h; int nid; |
457c1b27e hugetlb: ensure h... |
2458 2459 |
if (!hugepages_supported()) return; |
949f7ec57 mm, hugetlb: incl... |
2460 2461 2462 2463 2464 2465 2466 2467 2468 2469 |
for_each_node_state(nid, N_MEMORY) for_each_hstate(h) pr_info("Node %d hugepages_total=%u hugepages_free=%u hugepages_surp=%u hugepages_size=%lukB ", nid, h->nr_huge_pages_node[nid], h->free_huge_pages_node[nid], h->surplus_huge_pages_node[nid], 1UL << (huge_page_order(h) + PAGE_SHIFT - 10)); } |
1da177e4c Linux-2.6.12-rc2 |
2470 2471 2472 |
/* Return the number pages of memory we physically have, in PAGE_SIZE units. */ unsigned long hugetlb_total_pages(void) { |
d00285884 mm/hugetlb: fix t... |
2473 2474 2475 2476 2477 2478 |
struct hstate *h; unsigned long nr_total_pages = 0; for_each_hstate(h) nr_total_pages += h->nr_huge_pages * pages_per_huge_page(h); return nr_total_pages; |
1da177e4c Linux-2.6.12-rc2 |
2479 |
} |
1da177e4c Linux-2.6.12-rc2 |
2480 |
|
a55164389 hugetlb: modular ... |
2481 |
static int hugetlb_acct_memory(struct hstate *h, long delta) |
fc1b8a73d hugetlb: move hug... |
2482 2483 2484 2485 2486 2487 2488 2489 2490 2491 2492 2493 2494 2495 2496 2497 2498 2499 2500 2501 2502 2503 |
{ int ret = -ENOMEM; spin_lock(&hugetlb_lock); /* * When cpuset is configured, it breaks the strict hugetlb page * reservation as the accounting is done on a global variable. Such * reservation is completely rubbish in the presence of cpuset because * the reservation is not checked against page availability for the * current cpuset. Application can still potentially OOM'ed by kernel * with lack of free htlb page in cpuset that the task is in. * Attempt to enforce strict accounting with cpuset is almost * impossible (or too ugly) because cpuset is too fluid that * task or memory node can be dynamically moved between cpusets. * * The change of semantics for shared hugetlb mapping with cpuset is * undesirable. However, in order to preserve some of the semantics, * we fall back to check against current free page availability as * a best attempt and hopefully to minimize the impact of changing * semantics that cpuset has. */ if (delta > 0) { |
a55164389 hugetlb: modular ... |
2504 |
if (gather_surplus_pages(h, delta) < 0) |
fc1b8a73d hugetlb: move hug... |
2505 |
goto out; |
a55164389 hugetlb: modular ... |
2506 2507 |
if (delta > cpuset_mems_nr(h->free_huge_pages_node)) { return_unused_surplus_pages(h, delta); |
fc1b8a73d hugetlb: move hug... |
2508 2509 2510 2511 2512 2513 |
goto out; } } ret = 0; if (delta < 0) |
a55164389 hugetlb: modular ... |
2514 |
return_unused_surplus_pages(h, (unsigned long) -delta); |
fc1b8a73d hugetlb: move hug... |
2515 2516 2517 2518 2519 |
out: spin_unlock(&hugetlb_lock); return ret; } |
84afd99b8 hugetlb reservati... |
2520 2521 |
static void hugetlb_vm_op_open(struct vm_area_struct *vma) { |
f522c3ac0 mm, hugetlb: chan... |
2522 |
struct resv_map *resv = vma_resv_map(vma); |
84afd99b8 hugetlb reservati... |
2523 2524 2525 2526 2527 |
/* * This new VMA should share its siblings reservation map if present. * The VMA will only ever have a valid reservation map pointer where * it is being copied for another still existing VMA. As that VMA |
25985edce Fix common misspe... |
2528 |
* has a reference to the reservation map it cannot disappear until |
84afd99b8 hugetlb reservati... |
2529 2530 2531 |
* after this open call completes. It is therefore safe to take a * new reference here without additional locking. */ |
4e35f4838 mm, hugetlb: use ... |
2532 |
if (resv && is_vma_resv_set(vma, HPAGE_RESV_OWNER)) |
f522c3ac0 mm, hugetlb: chan... |
2533 |
kref_get(&resv->refs); |
84afd99b8 hugetlb reservati... |
2534 |
} |
a1e78772d hugetlb: reserve ... |
2535 2536 |
static void hugetlb_vm_op_close(struct vm_area_struct *vma) { |
a55164389 hugetlb: modular ... |
2537 |
struct hstate *h = hstate_vma(vma); |
f522c3ac0 mm, hugetlb: chan... |
2538 |
struct resv_map *resv = vma_resv_map(vma); |
90481622d hugepages: fix us... |
2539 |
struct hugepage_subpool *spool = subpool_vma(vma); |
4e35f4838 mm, hugetlb: use ... |
2540 |
unsigned long reserve, start, end; |
1c5ecae3a hugetlbfs: add mi... |
2541 |
long gbl_reserve; |
84afd99b8 hugetlb reservati... |
2542 |
|
4e35f4838 mm, hugetlb: use ... |
2543 2544 |
if (!resv || !is_vma_resv_set(vma, HPAGE_RESV_OWNER)) return; |
84afd99b8 hugetlb reservati... |
2545 |
|
4e35f4838 mm, hugetlb: use ... |
2546 2547 |
start = vma_hugecache_offset(h, vma, vma->vm_start); end = vma_hugecache_offset(h, vma, vma->vm_end); |
84afd99b8 hugetlb reservati... |
2548 |
|
4e35f4838 mm, hugetlb: use ... |
2549 |
reserve = (end - start) - region_count(resv, start, end); |
84afd99b8 hugetlb reservati... |
2550 |
|
4e35f4838 mm, hugetlb: use ... |
2551 2552 2553 |
kref_put(&resv->refs, resv_map_release); if (reserve) { |
1c5ecae3a hugetlbfs: add mi... |
2554 2555 2556 2557 2558 2559 |
/* * Decrement reserve counts. The global reserve count may be * adjusted if the subpool has a minimum size. */ gbl_reserve = hugepage_subpool_put_pages(spool, reserve); hugetlb_acct_memory(h, -gbl_reserve); |
84afd99b8 hugetlb reservati... |
2560 |
} |
a1e78772d hugetlb: reserve ... |
2561 |
} |
1da177e4c Linux-2.6.12-rc2 |
2562 2563 2564 2565 2566 2567 |
/* * We cannot handle pagefaults against hugetlb pages at all. They cause * handle_mm_fault() to try to instantiate regular-sized pages in the * hugegpage VMA. do_page_fault() is supposed to trap this, so BUG is we get * this far. */ |
d0217ac04 mm: fault feedbac... |
2568 |
static int hugetlb_vm_op_fault(struct vm_area_struct *vma, struct vm_fault *vmf) |
1da177e4c Linux-2.6.12-rc2 |
2569 2570 |
{ BUG(); |
d0217ac04 mm: fault feedbac... |
2571 |
return 0; |
1da177e4c Linux-2.6.12-rc2 |
2572 |
} |
f0f37e2f7 const: mark struc... |
2573 |
const struct vm_operations_struct hugetlb_vm_ops = { |
d0217ac04 mm: fault feedbac... |
2574 |
.fault = hugetlb_vm_op_fault, |
84afd99b8 hugetlb reservati... |
2575 |
.open = hugetlb_vm_op_open, |
a1e78772d hugetlb: reserve ... |
2576 |
.close = hugetlb_vm_op_close, |
1da177e4c Linux-2.6.12-rc2 |
2577 |
}; |
1e8f889b1 [PATCH] Hugetlb: ... |
2578 2579 |
static pte_t make_huge_pte(struct vm_area_struct *vma, struct page *page, int writable) |
63551ae0f [PATCH] Hugepage ... |
2580 2581 |
{ pte_t entry; |
1e8f889b1 [PATCH] Hugetlb: ... |
2582 |
if (writable) { |
106c992a5 mm/hugetlb: add m... |
2583 2584 |
entry = huge_pte_mkwrite(huge_pte_mkdirty(mk_huge_pte(page, vma->vm_page_prot))); |
63551ae0f [PATCH] Hugepage ... |
2585 |
} else { |
106c992a5 mm/hugetlb: add m... |
2586 2587 |
entry = huge_pte_wrprotect(mk_huge_pte(page, vma->vm_page_prot)); |
63551ae0f [PATCH] Hugepage ... |
2588 2589 2590 |
} entry = pte_mkyoung(entry); entry = pte_mkhuge(entry); |
d9ed9faac mm: add new arch_... |
2591 |
entry = arch_make_huge_pte(entry, vma, page, writable); |
63551ae0f [PATCH] Hugepage ... |
2592 2593 2594 |
return entry; } |
1e8f889b1 [PATCH] Hugetlb: ... |
2595 2596 2597 2598 |
static void set_huge_ptep_writable(struct vm_area_struct *vma, unsigned long address, pte_t *ptep) { pte_t entry; |
106c992a5 mm/hugetlb: add m... |
2599 |
entry = huge_pte_mkwrite(huge_pte_mkdirty(huge_ptep_get(ptep))); |
32f84528f mm: hugetlb: fix ... |
2600 |
if (huge_ptep_set_access_flags(vma, address, ptep, entry, 1)) |
4b3073e1c MM: Pass a PTE po... |
2601 |
update_mmu_cache(vma, address, ptep); |
1e8f889b1 [PATCH] Hugetlb: ... |
2602 |
} |
4a705fef9 hugetlb: fix copy... |
2603 2604 2605 2606 2607 2608 2609 2610 2611 2612 2613 2614 2615 2616 2617 2618 2619 2620 2621 2622 2623 2624 2625 2626 2627 |
static int is_hugetlb_entry_migration(pte_t pte) { swp_entry_t swp; if (huge_pte_none(pte) || pte_present(pte)) return 0; swp = pte_to_swp_entry(pte); if (non_swap_entry(swp) && is_migration_entry(swp)) return 1; else return 0; } static int is_hugetlb_entry_hwpoisoned(pte_t pte) { swp_entry_t swp; if (huge_pte_none(pte) || pte_present(pte)) return 0; swp = pte_to_swp_entry(pte); if (non_swap_entry(swp) && is_hwpoison_entry(swp)) return 1; else return 0; } |
1e8f889b1 [PATCH] Hugetlb: ... |
2628 |
|
63551ae0f [PATCH] Hugepage ... |
2629 2630 2631 2632 2633 |
int copy_hugetlb_page_range(struct mm_struct *dst, struct mm_struct *src, struct vm_area_struct *vma) { pte_t *src_pte, *dst_pte, entry; struct page *ptepage; |
1c59827d1 [PATCH] mm: huget... |
2634 |
unsigned long addr; |
1e8f889b1 [PATCH] Hugetlb: ... |
2635 |
int cow; |
a55164389 hugetlb: modular ... |
2636 2637 |
struct hstate *h = hstate_vma(vma); unsigned long sz = huge_page_size(h); |
e8569dd29 mm/hugetlb.c: cal... |
2638 2639 2640 |
unsigned long mmun_start; /* For mmu_notifiers */ unsigned long mmun_end; /* For mmu_notifiers */ int ret = 0; |
1e8f889b1 [PATCH] Hugetlb: ... |
2641 2642 |
cow = (vma->vm_flags & (VM_SHARED | VM_MAYWRITE)) == VM_MAYWRITE; |
63551ae0f [PATCH] Hugepage ... |
2643 |
|
e8569dd29 mm/hugetlb.c: cal... |
2644 2645 2646 2647 |
mmun_start = vma->vm_start; mmun_end = vma->vm_end; if (cow) mmu_notifier_invalidate_range_start(src, mmun_start, mmun_end); |
a55164389 hugetlb: modular ... |
2648 |
for (addr = vma->vm_start; addr < vma->vm_end; addr += sz) { |
cb900f412 mm, hugetlb: conv... |
2649 |
spinlock_t *src_ptl, *dst_ptl; |
c74df32c7 [PATCH] mm: ptd_a... |
2650 2651 2652 |
src_pte = huge_pte_offset(src, addr); if (!src_pte) continue; |
a55164389 hugetlb: modular ... |
2653 |
dst_pte = huge_pte_alloc(dst, addr, sz); |
e8569dd29 mm/hugetlb.c: cal... |
2654 2655 2656 2657 |
if (!dst_pte) { ret = -ENOMEM; break; } |
c5c99429f fix hugepages lea... |
2658 2659 2660 2661 |
/* If the pagetables are shared don't copy or take references */ if (dst_pte == src_pte) continue; |
cb900f412 mm, hugetlb: conv... |
2662 2663 2664 |
dst_ptl = huge_pte_lock(h, dst, dst_pte); src_ptl = huge_pte_lockptr(h, src, src_pte); spin_lock_nested(src_ptl, SINGLE_DEPTH_NESTING); |
4a705fef9 hugetlb: fix copy... |
2665 2666 2667 2668 2669 2670 2671 2672 2673 2674 2675 2676 2677 2678 2679 2680 2681 2682 |
entry = huge_ptep_get(src_pte); if (huge_pte_none(entry)) { /* skip none entry */ ; } else if (unlikely(is_hugetlb_entry_migration(entry) || is_hugetlb_entry_hwpoisoned(entry))) { swp_entry_t swp_entry = pte_to_swp_entry(entry); if (is_write_migration_entry(swp_entry) && cow) { /* * COW mappings require pages in both * parent and child to be set to read. */ make_migration_entry_read(&swp_entry); entry = swp_entry_to_pte(swp_entry); set_huge_pte_at(src, addr, src_pte, entry); } set_huge_pte_at(dst, addr, dst_pte, entry); } else { |
34ee645e8 mmu_notifier: cal... |
2683 |
if (cow) { |
7f2e9525b hugetlbfs: common... |
2684 |
huge_ptep_set_wrprotect(src, addr, src_pte); |
34ee645e8 mmu_notifier: cal... |
2685 2686 2687 |
mmu_notifier_invalidate_range(src, mmun_start, mmun_end); } |
0253d634e mm: hugetlb: fix ... |
2688 |
entry = huge_ptep_get(src_pte); |
1c59827d1 [PATCH] mm: huget... |
2689 2690 |
ptepage = pte_page(entry); get_page(ptepage); |
0fe6e20b9 hugetlb, rmap: ad... |
2691 |
page_dup_rmap(ptepage); |
1c59827d1 [PATCH] mm: huget... |
2692 2693 |
set_huge_pte_at(dst, addr, dst_pte, entry); } |
cb900f412 mm, hugetlb: conv... |
2694 2695 |
spin_unlock(src_ptl); spin_unlock(dst_ptl); |
63551ae0f [PATCH] Hugepage ... |
2696 |
} |
63551ae0f [PATCH] Hugepage ... |
2697 |
|
e8569dd29 mm/hugetlb.c: cal... |
2698 2699 2700 2701 |
if (cow) mmu_notifier_invalidate_range_end(src, mmun_start, mmun_end); return ret; |
63551ae0f [PATCH] Hugepage ... |
2702 |
} |
24669e584 hugetlb: use mmu_... |
2703 2704 2705 |
void __unmap_hugepage_range(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long start, unsigned long end, struct page *ref_page) |
63551ae0f [PATCH] Hugepage ... |
2706 |
{ |
24669e584 hugetlb: use mmu_... |
2707 |
int force_flush = 0; |
63551ae0f [PATCH] Hugepage ... |
2708 2709 |
struct mm_struct *mm = vma->vm_mm; unsigned long address; |
c7546f8f0 [PATCH] Fix hugep... |
2710 |
pte_t *ptep; |
63551ae0f [PATCH] Hugepage ... |
2711 |
pte_t pte; |
cb900f412 mm, hugetlb: conv... |
2712 |
spinlock_t *ptl; |
63551ae0f [PATCH] Hugepage ... |
2713 |
struct page *page; |
a55164389 hugetlb: modular ... |
2714 2715 |
struct hstate *h = hstate_vma(vma); unsigned long sz = huge_page_size(h); |
2ec74c3ef mm: move all mmu ... |
2716 2717 |
const unsigned long mmun_start = start; /* For mmu_notifiers */ const unsigned long mmun_end = end; /* For mmu_notifiers */ |
a55164389 hugetlb: modular ... |
2718 |
|
63551ae0f [PATCH] Hugepage ... |
2719 |
WARN_ON(!is_vm_hugetlb_page(vma)); |
a55164389 hugetlb: modular ... |
2720 2721 |
BUG_ON(start & ~huge_page_mask(h)); BUG_ON(end & ~huge_page_mask(h)); |
63551ae0f [PATCH] Hugepage ... |
2722 |
|
24669e584 hugetlb: use mmu_... |
2723 |
tlb_start_vma(tlb, vma); |
2ec74c3ef mm: move all mmu ... |
2724 |
mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end); |
569f48b85 mm: hugetlb: fix ... |
2725 |
address = start; |
24669e584 hugetlb: use mmu_... |
2726 |
again: |
569f48b85 mm: hugetlb: fix ... |
2727 |
for (; address < end; address += sz) { |
c7546f8f0 [PATCH] Fix hugep... |
2728 |
ptep = huge_pte_offset(mm, address); |
4c8872659 [PATCH] hugetlb: ... |
2729 |
if (!ptep) |
c7546f8f0 [PATCH] Fix hugep... |
2730 |
continue; |
cb900f412 mm, hugetlb: conv... |
2731 |
ptl = huge_pte_lock(h, mm, ptep); |
39dde65c9 [PATCH] shared pa... |
2732 |
if (huge_pmd_unshare(mm, &address, ptep)) |
cb900f412 mm, hugetlb: conv... |
2733 |
goto unlock; |
39dde65c9 [PATCH] shared pa... |
2734 |
|
6629326b8 mm: hugetlb: clea... |
2735 2736 |
pte = huge_ptep_get(ptep); if (huge_pte_none(pte)) |
cb900f412 mm, hugetlb: conv... |
2737 |
goto unlock; |
6629326b8 mm: hugetlb: clea... |
2738 2739 |
/* |
9fbc1f635 mm/hugetlb: add m... |
2740 2741 |
* Migrating hugepage or HWPoisoned hugepage is already * unmapped and its refcount is dropped, so just clear pte here. |
6629326b8 mm: hugetlb: clea... |
2742 |
*/ |
9fbc1f635 mm/hugetlb: add m... |
2743 |
if (unlikely(!pte_present(pte))) { |
106c992a5 mm/hugetlb: add m... |
2744 |
huge_pte_clear(mm, address, ptep); |
cb900f412 mm, hugetlb: conv... |
2745 |
goto unlock; |
8c4894c6b hwpoison, hugetlb... |
2746 |
} |
6629326b8 mm: hugetlb: clea... |
2747 2748 |
page = pte_page(pte); |
04f2cbe35 hugetlb: guarante... |
2749 2750 2751 2752 2753 2754 |
/* * If a reference page is supplied, it is because a specific * page is being unmapped, not a range. Ensure the page we * are about to unmap is the actual page of interest. */ if (ref_page) { |
04f2cbe35 hugetlb: guarante... |
2755 |
if (page != ref_page) |
cb900f412 mm, hugetlb: conv... |
2756 |
goto unlock; |
04f2cbe35 hugetlb: guarante... |
2757 2758 2759 2760 2761 2762 2763 2764 |
/* * Mark the VMA as having unmapped its page so that * future faults in this VMA will fail rather than * looking like data was lost */ set_vma_resv_flags(vma, HPAGE_RESV_UNMAPPED); } |
c7546f8f0 [PATCH] Fix hugep... |
2765 |
pte = huge_ptep_get_and_clear(mm, address, ptep); |
24669e584 hugetlb: use mmu_... |
2766 |
tlb_remove_tlb_entry(tlb, ptep, address); |
106c992a5 mm/hugetlb: add m... |
2767 |
if (huge_pte_dirty(pte)) |
6649a3863 [PATCH] hugetlb: ... |
2768 |
set_page_dirty(page); |
9e81130b7 mm: hugetlb: bail... |
2769 |
|
24669e584 hugetlb: use mmu_... |
2770 2771 |
page_remove_rmap(page); force_flush = !__tlb_remove_page(tlb, page); |
cb900f412 mm, hugetlb: conv... |
2772 |
if (force_flush) { |
569f48b85 mm: hugetlb: fix ... |
2773 |
address += sz; |
cb900f412 mm, hugetlb: conv... |
2774 |
spin_unlock(ptl); |
24669e584 hugetlb: use mmu_... |
2775 |
break; |
cb900f412 mm, hugetlb: conv... |
2776 |
} |
9e81130b7 mm: hugetlb: bail... |
2777 |
/* Bail out after unmapping reference page if supplied */ |
cb900f412 mm, hugetlb: conv... |
2778 2779 |
if (ref_page) { spin_unlock(ptl); |
9e81130b7 mm: hugetlb: bail... |
2780 |
break; |
cb900f412 mm, hugetlb: conv... |
2781 2782 2783 |
} unlock: spin_unlock(ptl); |
63551ae0f [PATCH] Hugepage ... |
2784 |
} |
24669e584 hugetlb: use mmu_... |
2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 |
/* * mmu_gather ran out of room to batch pages, we break out of * the PTE lock to avoid doing the potential expensive TLB invalidate * and page-free while holding it. */ if (force_flush) { force_flush = 0; tlb_flush_mmu(tlb); if (address < end && !ref_page) goto again; |
fe1668ae5 [PATCH] enforce p... |
2795 |
} |
2ec74c3ef mm: move all mmu ... |
2796 |
mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); |
24669e584 hugetlb: use mmu_... |
2797 |
tlb_end_vma(tlb, vma); |
1da177e4c Linux-2.6.12-rc2 |
2798 |
} |
63551ae0f [PATCH] Hugepage ... |
2799 |
|
d833352a4 mm: hugetlbfs: cl... |
2800 2801 2802 2803 2804 2805 2806 2807 2808 2809 2810 2811 |
void __unmap_hugepage_range_final(struct mmu_gather *tlb, struct vm_area_struct *vma, unsigned long start, unsigned long end, struct page *ref_page) { __unmap_hugepage_range(tlb, vma, start, end, ref_page); /* * Clear this flag so that x86's huge_pmd_share page_table_shareable * test will fail on a vma being torn down, and not grab a page table * on its way out. We're lucky that the flag has such an appropriate * name, and can in fact be safely cleared here. We could clear it * before the __unmap_hugepage_range above, but all that's necessary |
c8c06efa8 mm: convert i_mma... |
2812 |
* is to clear it before releasing the i_mmap_rwsem. This works |
d833352a4 mm: hugetlbfs: cl... |
2813 |
* because in the context this is called, the VMA is about to be |
c8c06efa8 mm: convert i_mma... |
2814 |
* destroyed and the i_mmap_rwsem is held. |
d833352a4 mm: hugetlbfs: cl... |
2815 2816 2817 |
*/ vma->vm_flags &= ~VM_MAYSHARE; } |
502717f4e [PATCH] hugetlb: ... |
2818 |
void unmap_hugepage_range(struct vm_area_struct *vma, unsigned long start, |
04f2cbe35 hugetlb: guarante... |
2819 |
unsigned long end, struct page *ref_page) |
502717f4e [PATCH] hugetlb: ... |
2820 |
{ |
24669e584 hugetlb: use mmu_... |
2821 2822 2823 2824 |
struct mm_struct *mm; struct mmu_gather tlb; mm = vma->vm_mm; |
2b047252d Fix TLB gather vi... |
2825 |
tlb_gather_mmu(&tlb, mm, start, end); |
24669e584 hugetlb: use mmu_... |
2826 2827 |
__unmap_hugepage_range(&tlb, vma, start, end, ref_page); tlb_finish_mmu(&tlb, start, end); |
502717f4e [PATCH] hugetlb: ... |
2828 |
} |
04f2cbe35 hugetlb: guarante... |
2829 2830 2831 2832 2833 2834 |
/* * This is called when the original mapper is failing to COW a MAP_PRIVATE * mappping it owns the reserve page for. The intention is to unmap the page * from other VMAs and let the children be SIGKILLed if they are faulting the * same region. */ |
2f4612af4 mm,hugetlb: make ... |
2835 2836 |
static void unmap_ref_private(struct mm_struct *mm, struct vm_area_struct *vma, struct page *page, unsigned long address) |
04f2cbe35 hugetlb: guarante... |
2837 |
{ |
7526674de hugetlb: make unm... |
2838 |
struct hstate *h = hstate_vma(vma); |
04f2cbe35 hugetlb: guarante... |
2839 2840 |
struct vm_area_struct *iter_vma; struct address_space *mapping; |
04f2cbe35 hugetlb: guarante... |
2841 2842 2843 2844 2845 2846 |
pgoff_t pgoff; /* * vm_pgoff is in PAGE_SIZE units, hence the different calculation * from page cache lookup which is in HPAGE_SIZE units. */ |
7526674de hugetlb: make unm... |
2847 |
address = address & huge_page_mask(h); |
36e4f20af hugetlb: do not u... |
2848 2849 |
pgoff = ((address - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff; |
496ad9aa8 new helper: file_... |
2850 |
mapping = file_inode(vma->vm_file)->i_mapping; |
04f2cbe35 hugetlb: guarante... |
2851 |
|
4eb2b1dcd hugetlb: acquire ... |
2852 2853 2854 2855 2856 |
/* * Take the mapping lock for the duration of the table walk. As * this mapping should be shared between all the VMAs, * __unmap_hugepage_range() is called as the lock is already held */ |
83cde9e8b mm: use new helpe... |
2857 |
i_mmap_lock_write(mapping); |
6b2dbba8b mm: replace vma p... |
2858 |
vma_interval_tree_foreach(iter_vma, &mapping->i_mmap, pgoff, pgoff) { |
04f2cbe35 hugetlb: guarante... |
2859 2860 2861 2862 2863 2864 2865 2866 2867 2868 2869 2870 |
/* Do not unmap the current VMA */ if (iter_vma == vma) continue; /* * Unmap the page from other VMAs without their own reserves. * They get marked to be SIGKILLed if they fault in these * areas. This is because a future no-page fault on this VMA * could insert a zeroed page instead of the data existing * from the time of fork. This would look like data corruption */ if (!is_vma_resv_set(iter_vma, HPAGE_RESV_OWNER)) |
24669e584 hugetlb: use mmu_... |
2871 2872 |
unmap_hugepage_range(iter_vma, address, address + huge_page_size(h), page); |
04f2cbe35 hugetlb: guarante... |
2873 |
} |
83cde9e8b mm: use new helpe... |
2874 |
i_mmap_unlock_write(mapping); |
04f2cbe35 hugetlb: guarante... |
2875 |
} |
0fe6e20b9 hugetlb, rmap: ad... |
2876 2877 |
/* * Hugetlb_cow() should be called with page lock of the original hugepage held. |
ef009b25f hugetlb: clarify ... |
2878 2879 2880 |
* Called with hugetlb_instantiation_mutex held and pte_page locked so we * cannot race with other handlers or page migration. * Keep the pte_same checks anyway to make transition from the mutex easier. |
0fe6e20b9 hugetlb, rmap: ad... |
2881 |
*/ |
1e8f889b1 [PATCH] Hugetlb: ... |
2882 |
static int hugetlb_cow(struct mm_struct *mm, struct vm_area_struct *vma, |
04f2cbe35 hugetlb: guarante... |
2883 |
unsigned long address, pte_t *ptep, pte_t pte, |
cb900f412 mm, hugetlb: conv... |
2884 |
struct page *pagecache_page, spinlock_t *ptl) |
1e8f889b1 [PATCH] Hugetlb: ... |
2885 |
{ |
a55164389 hugetlb: modular ... |
2886 |
struct hstate *h = hstate_vma(vma); |
1e8f889b1 [PATCH] Hugetlb: ... |
2887 |
struct page *old_page, *new_page; |
ad4404a22 mm,hugetlb: simpl... |
2888 |
int ret = 0, outside_reserve = 0; |
2ec74c3ef mm: move all mmu ... |
2889 2890 |
unsigned long mmun_start; /* For mmu_notifiers */ unsigned long mmun_end; /* For mmu_notifiers */ |
1e8f889b1 [PATCH] Hugetlb: ... |
2891 2892 |
old_page = pte_page(pte); |
04f2cbe35 hugetlb: guarante... |
2893 |
retry_avoidcopy: |
1e8f889b1 [PATCH] Hugetlb: ... |
2894 2895 |
/* If no-one else is actually using this page, avoid the copy * and just make the page writable */ |
37a2140dc mm, hugetlb: do n... |
2896 2897 |
if (page_mapcount(old_page) == 1 && PageAnon(old_page)) { page_move_anon_rmap(old_page, vma, address); |
1e8f889b1 [PATCH] Hugetlb: ... |
2898 |
set_huge_ptep_writable(vma, address, ptep); |
83c54070e mm: fault feedbac... |
2899 |
return 0; |
1e8f889b1 [PATCH] Hugetlb: ... |
2900 |
} |
04f2cbe35 hugetlb: guarante... |
2901 2902 2903 2904 2905 2906 2907 2908 2909 |
/* * If the process that created a MAP_PRIVATE mapping is about to * perform a COW due to a shared page count, attempt to satisfy * the allocation without using the existing reserves. The pagecache * page is used to determine if the reserve at this address was * consumed or not. If reserves were used, a partial faulted mapping * at the time of fork() could consume its reserves on COW instead * of the full address range. */ |
5944d0116 mm, hugetlb: remo... |
2910 |
if (is_vma_resv_set(vma, HPAGE_RESV_OWNER) && |
04f2cbe35 hugetlb: guarante... |
2911 2912 |
old_page != pagecache_page) outside_reserve = 1; |
1e8f889b1 [PATCH] Hugetlb: ... |
2913 |
page_cache_get(old_page); |
b76c8cfbf hugetlb: prevent ... |
2914 |
|
ad4404a22 mm,hugetlb: simpl... |
2915 2916 2917 2918 |
/* * Drop page table lock as buddy allocator may be called. It will * be acquired again before returning to the caller, as expected. */ |
cb900f412 mm, hugetlb: conv... |
2919 |
spin_unlock(ptl); |
04f2cbe35 hugetlb: guarante... |
2920 |
new_page = alloc_huge_page(vma, address, outside_reserve); |
1e8f889b1 [PATCH] Hugetlb: ... |
2921 |
|
2fc39cec6 hugetlb: debit qu... |
2922 |
if (IS_ERR(new_page)) { |
04f2cbe35 hugetlb: guarante... |
2923 2924 2925 2926 2927 2928 2929 2930 |
/* * If a process owning a MAP_PRIVATE mapping fails to COW, * it is due to references held by a child and an insufficient * huge page pool. To guarantee the original mappers * reliability, unmap the page from child processes. The child * may get SIGKILLed if it later faults. */ if (outside_reserve) { |
ad4404a22 mm,hugetlb: simpl... |
2931 |
page_cache_release(old_page); |
04f2cbe35 hugetlb: guarante... |
2932 |
BUG_ON(huge_pte_none(pte)); |
2f4612af4 mm,hugetlb: make ... |
2933 2934 2935 2936 2937 2938 2939 2940 2941 2942 2943 2944 |
unmap_ref_private(mm, vma, old_page, address); BUG_ON(huge_pte_none(pte)); spin_lock(ptl); ptep = huge_pte_offset(mm, address & huge_page_mask(h)); if (likely(ptep && pte_same(huge_ptep_get(ptep), pte))) goto retry_avoidcopy; /* * race occurs while re-acquiring page table * lock, and our job is done. */ return 0; |
04f2cbe35 hugetlb: guarante... |
2945 |
} |
ad4404a22 mm,hugetlb: simpl... |
2946 2947 2948 |
ret = (PTR_ERR(new_page) == -ENOMEM) ? VM_FAULT_OOM : VM_FAULT_SIGBUS; goto out_release_old; |
1e8f889b1 [PATCH] Hugetlb: ... |
2949 |
} |
0fe6e20b9 hugetlb, rmap: ad... |
2950 2951 2952 2953 |
/* * When the original hugepage is shared one, it does not have * anon_vma prepared. */ |
44e2aa937 mm/hugetlb.c: add... |
2954 |
if (unlikely(anon_vma_prepare(vma))) { |
ad4404a22 mm,hugetlb: simpl... |
2955 2956 |
ret = VM_FAULT_OOM; goto out_release_all; |
44e2aa937 mm/hugetlb.c: add... |
2957 |
} |
0fe6e20b9 hugetlb, rmap: ad... |
2958 |
|
47ad8475c thp: clear_copy_h... |
2959 2960 |
copy_user_huge_page(new_page, old_page, address, vma, pages_per_huge_page(h)); |
0ed361dec mm: fix PageUptod... |
2961 |
__SetPageUptodate(new_page); |
bcc542223 mm: hugetlb: intr... |
2962 |
set_page_huge_active(new_page); |
1e8f889b1 [PATCH] Hugetlb: ... |
2963 |
|
2ec74c3ef mm: move all mmu ... |
2964 2965 2966 |
mmun_start = address & huge_page_mask(h); mmun_end = mmun_start + huge_page_size(h); mmu_notifier_invalidate_range_start(mm, mmun_start, mmun_end); |
ad4404a22 mm,hugetlb: simpl... |
2967 |
|
b76c8cfbf hugetlb: prevent ... |
2968 |
/* |
cb900f412 mm, hugetlb: conv... |
2969 |
* Retake the page table lock to check for racing updates |
b76c8cfbf hugetlb: prevent ... |
2970 2971 |
* before the page tables are altered */ |
cb900f412 mm, hugetlb: conv... |
2972 |
spin_lock(ptl); |
a55164389 hugetlb: modular ... |
2973 |
ptep = huge_pte_offset(mm, address & huge_page_mask(h)); |
a9af0c5df mm/hugetlb.c: add... |
2974 |
if (likely(ptep && pte_same(huge_ptep_get(ptep), pte))) { |
07443a85a mm, hugetlb: retu... |
2975 |
ClearPagePrivate(new_page); |
1e8f889b1 [PATCH] Hugetlb: ... |
2976 |
/* Break COW */ |
8fe627ec5 hugetlbfs: add mi... |
2977 |
huge_ptep_clear_flush(vma, address, ptep); |
34ee645e8 mmu_notifier: cal... |
2978 |
mmu_notifier_invalidate_range(mm, mmun_start, mmun_end); |
1e8f889b1 [PATCH] Hugetlb: ... |
2979 2980 |
set_huge_pte_at(mm, address, ptep, make_huge_pte(vma, new_page, 1)); |
0fe6e20b9 hugetlb, rmap: ad... |
2981 |
page_remove_rmap(old_page); |
cd67f0d2a hugetlb, rmap: us... |
2982 |
hugepage_add_new_anon_rmap(new_page, vma, address); |
1e8f889b1 [PATCH] Hugetlb: ... |
2983 2984 2985 |
/* Make the old page be freed below */ new_page = old_page; } |
cb900f412 mm, hugetlb: conv... |
2986 |
spin_unlock(ptl); |
2ec74c3ef mm: move all mmu ... |
2987 |
mmu_notifier_invalidate_range_end(mm, mmun_start, mmun_end); |
ad4404a22 mm,hugetlb: simpl... |
2988 |
out_release_all: |
1e8f889b1 [PATCH] Hugetlb: ... |
2989 |
page_cache_release(new_page); |
ad4404a22 mm,hugetlb: simpl... |
2990 |
out_release_old: |
1e8f889b1 [PATCH] Hugetlb: ... |
2991 |
page_cache_release(old_page); |
8312034f3 mm, hugetlb: grab... |
2992 |
|
ad4404a22 mm,hugetlb: simpl... |
2993 2994 |
spin_lock(ptl); /* Caller expects lock to be held */ return ret; |
1e8f889b1 [PATCH] Hugetlb: ... |
2995 |
} |
04f2cbe35 hugetlb: guarante... |
2996 |
/* Return the pagecache page at a given address within a VMA */ |
a55164389 hugetlb: modular ... |
2997 2998 |
static struct page *hugetlbfs_pagecache_page(struct hstate *h, struct vm_area_struct *vma, unsigned long address) |
04f2cbe35 hugetlb: guarante... |
2999 3000 |
{ struct address_space *mapping; |
e7c4b0bfd huge page private... |
3001 |
pgoff_t idx; |
04f2cbe35 hugetlb: guarante... |
3002 3003 |
mapping = vma->vm_file->f_mapping; |
a55164389 hugetlb: modular ... |
3004 |
idx = vma_hugecache_offset(h, vma, address); |
04f2cbe35 hugetlb: guarante... |
3005 3006 3007 |
return find_lock_page(mapping, idx); } |
3ae77f43b mm: hugetlbfs_pag... |
3008 3009 3010 3011 3012 |
/* * Return whether there is a pagecache page to back given address within VMA. * Caller follow_hugetlb_page() holds page_table_lock so we cannot lock_page. */ static bool hugetlbfs_pagecache_present(struct hstate *h, |
2a15efc95 mm: follow_hugetl... |
3013 3014 3015 3016 3017 3018 3019 3020 3021 3022 3023 3024 3025 3026 |
struct vm_area_struct *vma, unsigned long address) { struct address_space *mapping; pgoff_t idx; struct page *page; mapping = vma->vm_file->f_mapping; idx = vma_hugecache_offset(h, vma, address); page = find_get_page(mapping, idx); if (page) put_page(page); return page != NULL; } |
a1ed3dda0 MM: Make needless... |
3027 |
static int hugetlb_no_page(struct mm_struct *mm, struct vm_area_struct *vma, |
8382d914e mm, hugetlb: impr... |
3028 3029 |
struct address_space *mapping, pgoff_t idx, unsigned long address, pte_t *ptep, unsigned int flags) |
ac9b9c667 [PATCH] Fix handl... |
3030 |
{ |
a55164389 hugetlb: modular ... |
3031 |
struct hstate *h = hstate_vma(vma); |
ac9b9c667 [PATCH] Fix handl... |
3032 |
int ret = VM_FAULT_SIGBUS; |
409eb8c26 mm/hugetlb.c: und... |
3033 |
int anon_rmap = 0; |
4c8872659 [PATCH] hugetlb: ... |
3034 |
unsigned long size; |
4c8872659 [PATCH] hugetlb: ... |
3035 |
struct page *page; |
1e8f889b1 [PATCH] Hugetlb: ... |
3036 |
pte_t new_pte; |
cb900f412 mm, hugetlb: conv... |
3037 |
spinlock_t *ptl; |
4c8872659 [PATCH] hugetlb: ... |
3038 |
|
04f2cbe35 hugetlb: guarante... |
3039 3040 3041 |
/* * Currently, we are forced to kill the process in the event the * original mapper has unmapped pages from the child due to a failed |
25985edce Fix common misspe... |
3042 |
* COW. Warn that such a situation has occurred as it may not be obvious |
04f2cbe35 hugetlb: guarante... |
3043 3044 |
*/ if (is_vma_resv_set(vma, HPAGE_RESV_UNMAPPED)) { |
ffb22af5b mm/hugetlb.c: con... |
3045 3046 3047 |
pr_warning("PID %d killed due to inadequate hugepage pool ", current->pid); |
04f2cbe35 hugetlb: guarante... |
3048 3049 |
return ret; } |
4c8872659 [PATCH] hugetlb: ... |
3050 3051 3052 3053 |
/* * Use page lock to guard against racing truncation * before we get page_table_lock. */ |
6bda666a0 [PATCH] hugepages... |
3054 3055 3056 |
retry: page = find_lock_page(mapping, idx); if (!page) { |
a55164389 hugetlb: modular ... |
3057 |
size = i_size_read(mapping->host) >> huge_page_shift(h); |
ebed4bfc8 [PATCH] hugetlb: ... |
3058 3059 |
if (idx >= size) goto out; |
04f2cbe35 hugetlb: guarante... |
3060 |
page = alloc_huge_page(vma, address, 0); |
2fc39cec6 hugetlb: debit qu... |
3061 |
if (IS_ERR(page)) { |
76dcee75c hugetlb: don't us... |
3062 3063 3064 3065 3066 |
ret = PTR_ERR(page); if (ret == -ENOMEM) ret = VM_FAULT_OOM; else ret = VM_FAULT_SIGBUS; |
6bda666a0 [PATCH] hugepages... |
3067 3068 |
goto out; } |
47ad8475c thp: clear_copy_h... |
3069 |
clear_huge_page(page, address, pages_per_huge_page(h)); |
0ed361dec mm: fix PageUptod... |
3070 |
__SetPageUptodate(page); |
bcc542223 mm: hugetlb: intr... |
3071 |
set_page_huge_active(page); |
ac9b9c667 [PATCH] Fix handl... |
3072 |
|
f83a275db mm: account for M... |
3073 |
if (vma->vm_flags & VM_MAYSHARE) { |
6bda666a0 [PATCH] hugepages... |
3074 |
int err; |
45c682a68 hugetlb: fix i_bl... |
3075 |
struct inode *inode = mapping->host; |
6bda666a0 [PATCH] hugepages... |
3076 3077 3078 3079 |
err = add_to_page_cache(page, mapping, idx, GFP_KERNEL); if (err) { put_page(page); |
6bda666a0 [PATCH] hugepages... |
3080 3081 3082 3083 |
if (err == -EEXIST) goto retry; goto out; } |
07443a85a mm, hugetlb: retu... |
3084 |
ClearPagePrivate(page); |
45c682a68 hugetlb: fix i_bl... |
3085 3086 |
spin_lock(&inode->i_lock); |
a55164389 hugetlb: modular ... |
3087 |
inode->i_blocks += blocks_per_huge_page(h); |
45c682a68 hugetlb: fix i_bl... |
3088 |
spin_unlock(&inode->i_lock); |
23be7468e hugetlb: fix infi... |
3089 |
} else { |
6bda666a0 [PATCH] hugepages... |
3090 |
lock_page(page); |
0fe6e20b9 hugetlb, rmap: ad... |
3091 3092 3093 3094 |
if (unlikely(anon_vma_prepare(vma))) { ret = VM_FAULT_OOM; goto backout_unlocked; } |
409eb8c26 mm/hugetlb.c: und... |
3095 |
anon_rmap = 1; |
23be7468e hugetlb: fix infi... |
3096 |
} |
0fe6e20b9 hugetlb, rmap: ad... |
3097 |
} else { |
998b4382c hugetlb: fix meta... |
3098 3099 3100 3101 3102 3103 |
/* * If memory error occurs between mmap() and fault, some process * don't have hwpoisoned swap entry for errored virtual address. * So we need to block hugepage fault by PG_hwpoison bit check. */ if (unlikely(PageHWPoison(page))) { |
32f84528f mm: hugetlb: fix ... |
3104 |
ret = VM_FAULT_HWPOISON | |
972dc4de1 hugetlb: add an i... |
3105 |
VM_FAULT_SET_HINDEX(hstate_index(h)); |
998b4382c hugetlb: fix meta... |
3106 3107 |
goto backout_unlocked; } |
6bda666a0 [PATCH] hugepages... |
3108 |
} |
1e8f889b1 [PATCH] Hugetlb: ... |
3109 |
|
57303d801 hugetlbfs: alloca... |
3110 3111 3112 3113 3114 3115 |
/* * If we are going to COW a private mapping later, we examine the * pending reservations for this page now. This will ensure that * any allocations necessary to record that reservation occur outside * the spinlock. */ |
788c7df45 hugetlb: fault fl... |
3116 |
if ((flags & FAULT_FLAG_WRITE) && !(vma->vm_flags & VM_SHARED)) |
2b26736c8 allocate structur... |
3117 3118 3119 3120 |
if (vma_needs_reservation(h, vma, address) < 0) { ret = VM_FAULT_OOM; goto backout_unlocked; } |
57303d801 hugetlbfs: alloca... |
3121 |
|
cb900f412 mm, hugetlb: conv... |
3122 3123 |
ptl = huge_pte_lockptr(h, mm, ptep); spin_lock(ptl); |
a55164389 hugetlb: modular ... |
3124 |
size = i_size_read(mapping->host) >> huge_page_shift(h); |
4c8872659 [PATCH] hugetlb: ... |
3125 3126 |
if (idx >= size) goto backout; |
83c54070e mm: fault feedbac... |
3127 |
ret = 0; |
7f2e9525b hugetlbfs: common... |
3128 |
if (!huge_pte_none(huge_ptep_get(ptep))) |
4c8872659 [PATCH] hugetlb: ... |
3129 |
goto backout; |
07443a85a mm, hugetlb: retu... |
3130 3131 |
if (anon_rmap) { ClearPagePrivate(page); |
409eb8c26 mm/hugetlb.c: und... |
3132 |
hugepage_add_new_anon_rmap(page, vma, address); |
ac7149045 mm: fix 'ERROR: d... |
3133 |
} else |
409eb8c26 mm/hugetlb.c: und... |
3134 |
page_dup_rmap(page); |
1e8f889b1 [PATCH] Hugetlb: ... |
3135 3136 3137 |
new_pte = make_huge_pte(vma, page, ((vma->vm_flags & VM_WRITE) && (vma->vm_flags & VM_SHARED))); set_huge_pte_at(mm, address, ptep, new_pte); |
788c7df45 hugetlb: fault fl... |
3138 |
if ((flags & FAULT_FLAG_WRITE) && !(vma->vm_flags & VM_SHARED)) { |
1e8f889b1 [PATCH] Hugetlb: ... |
3139 |
/* Optimization, do the COW without a second fault */ |
cb900f412 mm, hugetlb: conv... |
3140 |
ret = hugetlb_cow(mm, vma, address, ptep, new_pte, page, ptl); |
1e8f889b1 [PATCH] Hugetlb: ... |
3141 |
} |
cb900f412 mm, hugetlb: conv... |
3142 |
spin_unlock(ptl); |
4c8872659 [PATCH] hugetlb: ... |
3143 3144 |
unlock_page(page); out: |
ac9b9c667 [PATCH] Fix handl... |
3145 |
return ret; |
4c8872659 [PATCH] hugetlb: ... |
3146 3147 |
backout: |
cb900f412 mm, hugetlb: conv... |
3148 |
spin_unlock(ptl); |
2b26736c8 allocate structur... |
3149 |
backout_unlocked: |
4c8872659 [PATCH] hugetlb: ... |
3150 3151 3152 |
unlock_page(page); put_page(page); goto out; |
ac9b9c667 [PATCH] Fix handl... |
3153 |
} |
8382d914e mm, hugetlb: impr... |
3154 3155 3156 3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 3172 3173 3174 3175 3176 3177 3178 3179 3180 3181 3182 3183 3184 3185 3186 3187 |
#ifdef CONFIG_SMP static u32 fault_mutex_hash(struct hstate *h, struct mm_struct *mm, struct vm_area_struct *vma, struct address_space *mapping, pgoff_t idx, unsigned long address) { unsigned long key[2]; u32 hash; if (vma->vm_flags & VM_SHARED) { key[0] = (unsigned long) mapping; key[1] = idx; } else { key[0] = (unsigned long) mm; key[1] = address >> huge_page_shift(h); } hash = jhash2((u32 *)&key, sizeof(key)/sizeof(u32), 0); return hash & (num_fault_mutexes - 1); } #else /* * For uniprocesor systems we always use a single mutex, so just * return 0 and avoid the hashing overhead. */ static u32 fault_mutex_hash(struct hstate *h, struct mm_struct *mm, struct vm_area_struct *vma, struct address_space *mapping, pgoff_t idx, unsigned long address) { return 0; } #endif |
86e5216f8 [PATCH] Hugetlb: ... |
3188 |
int hugetlb_fault(struct mm_struct *mm, struct vm_area_struct *vma, |
788c7df45 hugetlb: fault fl... |
3189 |
unsigned long address, unsigned int flags) |
86e5216f8 [PATCH] Hugetlb: ... |
3190 |
{ |
8382d914e mm, hugetlb: impr... |
3191 |
pte_t *ptep, entry; |
cb900f412 mm, hugetlb: conv... |
3192 |
spinlock_t *ptl; |
1e8f889b1 [PATCH] Hugetlb: ... |
3193 |
int ret; |
8382d914e mm, hugetlb: impr... |
3194 3195 |
u32 hash; pgoff_t idx; |
0fe6e20b9 hugetlb, rmap: ad... |
3196 |
struct page *page = NULL; |
57303d801 hugetlbfs: alloca... |
3197 |
struct page *pagecache_page = NULL; |
a55164389 hugetlb: modular ... |
3198 |
struct hstate *h = hstate_vma(vma); |
8382d914e mm, hugetlb: impr... |
3199 |
struct address_space *mapping; |
0f792cf94 mm/hugetlb: fix g... |
3200 |
int need_wait_lock = 0; |
86e5216f8 [PATCH] Hugetlb: ... |
3201 |
|
1e16a539a mm/hugetlb.c: fix... |
3202 |
address &= huge_page_mask(h); |
fd6a03edd HWPOISON, hugetlb... |
3203 3204 3205 |
ptep = huge_pte_offset(mm, address); if (ptep) { entry = huge_ptep_get(ptep); |
290408d4a hugetlb: hugepage... |
3206 |
if (unlikely(is_hugetlb_entry_migration(entry))) { |
cb900f412 mm, hugetlb: conv... |
3207 |
migration_entry_wait_huge(vma, mm, ptep); |
290408d4a hugetlb: hugepage... |
3208 3209 |
return 0; } else if (unlikely(is_hugetlb_entry_hwpoisoned(entry))) |
32f84528f mm: hugetlb: fix ... |
3210 |
return VM_FAULT_HWPOISON_LARGE | |
972dc4de1 hugetlb: add an i... |
3211 |
VM_FAULT_SET_HINDEX(hstate_index(h)); |
fd6a03edd HWPOISON, hugetlb... |
3212 |
} |
a55164389 hugetlb: modular ... |
3213 |
ptep = huge_pte_alloc(mm, address, huge_page_size(h)); |
86e5216f8 [PATCH] Hugetlb: ... |
3214 3215 |
if (!ptep) return VM_FAULT_OOM; |
8382d914e mm, hugetlb: impr... |
3216 3217 |
mapping = vma->vm_file->f_mapping; idx = vma_hugecache_offset(h, vma, address); |
3935baa9b [PATCH] hugepage:... |
3218 3219 3220 3221 3222 |
/* * Serialize hugepage allocation and instantiation, so that we don't * get spurious allocation failures if two CPUs race to instantiate * the same page in the page cache. */ |
8382d914e mm, hugetlb: impr... |
3223 3224 |
hash = fault_mutex_hash(h, mm, vma, mapping, idx, address); mutex_lock(&htlb_fault_mutex_table[hash]); |
7f2e9525b hugetlbfs: common... |
3225 3226 |
entry = huge_ptep_get(ptep); if (huge_pte_none(entry)) { |
8382d914e mm, hugetlb: impr... |
3227 |
ret = hugetlb_no_page(mm, vma, mapping, idx, address, ptep, flags); |
b4d1d99fd hugetlb: handle u... |
3228 |
goto out_mutex; |
3935baa9b [PATCH] hugepage:... |
3229 |
} |
86e5216f8 [PATCH] Hugetlb: ... |
3230 |
|
83c54070e mm: fault feedbac... |
3231 |
ret = 0; |
1e8f889b1 [PATCH] Hugetlb: ... |
3232 |
|
57303d801 hugetlbfs: alloca... |
3233 |
/* |
0f792cf94 mm/hugetlb: fix g... |
3234 3235 3236 3237 3238 3239 3240 3241 3242 3243 |
* entry could be a migration/hwpoison entry at this point, so this * check prevents the kernel from going below assuming that we have * a active hugepage in pagecache. This goto expects the 2nd page fault, * and is_hugetlb_entry_(migration|hwpoisoned) check will properly * handle it. */ if (!pte_present(entry)) goto out_mutex; /* |
57303d801 hugetlbfs: alloca... |
3244 3245 3246 3247 3248 3249 3250 |
* If we are going to COW the mapping later, we examine the pending * reservations for this page now. This will ensure that any * allocations necessary to record that reservation occur outside the * spinlock. For private mappings, we also lookup the pagecache * page now as it is used to determine if a reservation has been * consumed. */ |
106c992a5 mm/hugetlb: add m... |
3251 |
if ((flags & FAULT_FLAG_WRITE) && !huge_pte_write(entry)) { |
2b26736c8 allocate structur... |
3252 3253 |
if (vma_needs_reservation(h, vma, address) < 0) { ret = VM_FAULT_OOM; |
b4d1d99fd hugetlb: handle u... |
3254 |
goto out_mutex; |
2b26736c8 allocate structur... |
3255 |
} |
57303d801 hugetlbfs: alloca... |
3256 |
|
f83a275db mm: account for M... |
3257 |
if (!(vma->vm_flags & VM_MAYSHARE)) |
57303d801 hugetlbfs: alloca... |
3258 3259 3260 |
pagecache_page = hugetlbfs_pagecache_page(h, vma, address); } |
0f792cf94 mm/hugetlb: fix g... |
3261 3262 3263 3264 3265 |
ptl = huge_pte_lock(h, mm, ptep); /* Check for a racing update before calling hugetlb_cow */ if (unlikely(!pte_same(entry, huge_ptep_get(ptep)))) goto out_ptl; |
56c9cfb13 hugetlb, rmap: fi... |
3266 3267 3268 3269 |
/* * hugetlb_cow() requires page locks of pte_page(entry) and * pagecache_page, so here we need take the former one * when page != pagecache_page or !pagecache_page. |
56c9cfb13 hugetlb, rmap: fi... |
3270 3271 3272 |
*/ page = pte_page(entry); if (page != pagecache_page) |
0f792cf94 mm/hugetlb: fix g... |
3273 3274 3275 3276 |
if (!trylock_page(page)) { need_wait_lock = 1; goto out_ptl; } |
b4d1d99fd hugetlb: handle u... |
3277 |
|
0f792cf94 mm/hugetlb: fix g... |
3278 |
get_page(page); |
b4d1d99fd hugetlb: handle u... |
3279 |
|
788c7df45 hugetlb: fault fl... |
3280 |
if (flags & FAULT_FLAG_WRITE) { |
106c992a5 mm/hugetlb: add m... |
3281 |
if (!huge_pte_write(entry)) { |
57303d801 hugetlbfs: alloca... |
3282 |
ret = hugetlb_cow(mm, vma, address, ptep, entry, |
cb900f412 mm, hugetlb: conv... |
3283 |
pagecache_page, ptl); |
0f792cf94 mm/hugetlb: fix g... |
3284 |
goto out_put_page; |
b4d1d99fd hugetlb: handle u... |
3285 |
} |
106c992a5 mm/hugetlb: add m... |
3286 |
entry = huge_pte_mkdirty(entry); |
b4d1d99fd hugetlb: handle u... |
3287 3288 |
} entry = pte_mkyoung(entry); |
788c7df45 hugetlb: fault fl... |
3289 3290 |
if (huge_ptep_set_access_flags(vma, address, ptep, entry, flags & FAULT_FLAG_WRITE)) |
4b3073e1c MM: Pass a PTE po... |
3291 |
update_mmu_cache(vma, address, ptep); |
0f792cf94 mm/hugetlb: fix g... |
3292 3293 3294 3295 |
out_put_page: if (page != pagecache_page) unlock_page(page); put_page(page); |
cb900f412 mm, hugetlb: conv... |
3296 3297 |
out_ptl: spin_unlock(ptl); |
57303d801 hugetlbfs: alloca... |
3298 3299 3300 3301 3302 |
if (pagecache_page) { unlock_page(pagecache_page); put_page(pagecache_page); } |
b4d1d99fd hugetlb: handle u... |
3303 |
out_mutex: |
8382d914e mm, hugetlb: impr... |
3304 |
mutex_unlock(&htlb_fault_mutex_table[hash]); |
0f792cf94 mm/hugetlb: fix g... |
3305 3306 3307 3308 3309 3310 3311 3312 3313 |
/* * Generally it's safe to hold refcount during waiting page lock. But * here we just wait to defer the next page fault to avoid busy loop and * the page is not used after unlocked before returning from the current * page fault. So we are safe from accessing freed page, even if we wait * here without taking refcount. */ if (need_wait_lock) wait_on_page_locked(page); |
1e8f889b1 [PATCH] Hugetlb: ... |
3314 |
return ret; |
86e5216f8 [PATCH] Hugetlb: ... |
3315 |
} |
28a35716d mm: use long type... |
3316 3317 3318 3319 |
long follow_hugetlb_page(struct mm_struct *mm, struct vm_area_struct *vma, struct page **pages, struct vm_area_struct **vmas, unsigned long *position, unsigned long *nr_pages, long i, unsigned int flags) |
63551ae0f [PATCH] Hugepage ... |
3320 |
{ |
d5d4b0aa4 [PATCH] optimize ... |
3321 3322 |
unsigned long pfn_offset; unsigned long vaddr = *position; |
28a35716d mm: use long type... |
3323 |
unsigned long remainder = *nr_pages; |
a55164389 hugetlb: modular ... |
3324 |
struct hstate *h = hstate_vma(vma); |
63551ae0f [PATCH] Hugepage ... |
3325 |
|
63551ae0f [PATCH] Hugepage ... |
3326 |
while (vaddr < vma->vm_end && remainder) { |
4c8872659 [PATCH] hugetlb: ... |
3327 |
pte_t *pte; |
cb900f412 mm, hugetlb: conv... |
3328 |
spinlock_t *ptl = NULL; |
2a15efc95 mm: follow_hugetl... |
3329 |
int absent; |
4c8872659 [PATCH] hugetlb: ... |
3330 |
struct page *page; |
63551ae0f [PATCH] Hugepage ... |
3331 |
|
4c8872659 [PATCH] hugetlb: ... |
3332 |
/* |
02057967b mm, hugetlb: abor... |
3333 3334 3335 3336 3337 3338 3339 3340 3341 |
* If we have a pending SIGKILL, don't keep faulting pages and * potentially allocating memory. */ if (unlikely(fatal_signal_pending(current))) { remainder = 0; break; } /* |
4c8872659 [PATCH] hugetlb: ... |
3342 |
* Some archs (sparc64, sh*) have multiple pte_ts to |
2a15efc95 mm: follow_hugetl... |
3343 |
* each hugepage. We have to make sure we get the |
4c8872659 [PATCH] hugetlb: ... |
3344 |
* first, for the page indexing below to work. |
cb900f412 mm, hugetlb: conv... |
3345 3346 |
* * Note that page table lock is not held when pte is null. |
4c8872659 [PATCH] hugetlb: ... |
3347 |
*/ |
a55164389 hugetlb: modular ... |
3348 |
pte = huge_pte_offset(mm, vaddr & huge_page_mask(h)); |
cb900f412 mm, hugetlb: conv... |
3349 3350 |
if (pte) ptl = huge_pte_lock(h, mm, pte); |
2a15efc95 mm: follow_hugetl... |
3351 3352 3353 3354 |
absent = !pte || huge_pte_none(huge_ptep_get(pte)); /* * When coredumping, it suits get_dump_page if we just return |
3ae77f43b mm: hugetlbfs_pag... |
3355 3356 3357 3358 |
* an error where there's an empty slot with no huge pagecache * to back it. This way, we avoid allocating a hugepage, and * the sparse dumpfile avoids allocating disk blocks, but its * huge holes still show up with zeroes where they need to be. |
2a15efc95 mm: follow_hugetl... |
3359 |
*/ |
3ae77f43b mm: hugetlbfs_pag... |
3360 3361 |
if (absent && (flags & FOLL_DUMP) && !hugetlbfs_pagecache_present(h, vma, vaddr)) { |
cb900f412 mm, hugetlb: conv... |
3362 3363 |
if (pte) spin_unlock(ptl); |
2a15efc95 mm: follow_hugetl... |
3364 3365 3366 |
remainder = 0; break; } |
63551ae0f [PATCH] Hugepage ... |
3367 |
|
9cc3a5bd4 hugetlbfs: add sw... |
3368 3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 |
/* * We need call hugetlb_fault for both hugepages under migration * (in which case hugetlb_fault waits for the migration,) and * hwpoisoned hugepages (in which case we need to prevent the * caller from accessing to them.) In order to do this, we use * here is_swap_pte instead of is_hugetlb_entry_migration and * is_hugetlb_entry_hwpoisoned. This is because it simply covers * both cases, and because we can't follow correct pages * directly from any kind of swap entries. */ if (absent || is_swap_pte(huge_ptep_get(pte)) || |
106c992a5 mm/hugetlb: add m... |
3379 3380 |
((flags & FOLL_WRITE) && !huge_pte_write(huge_ptep_get(pte)))) { |
4c8872659 [PATCH] hugetlb: ... |
3381 |
int ret; |
63551ae0f [PATCH] Hugepage ... |
3382 |
|
cb900f412 mm, hugetlb: conv... |
3383 3384 |
if (pte) spin_unlock(ptl); |
2a15efc95 mm: follow_hugetl... |
3385 3386 |
ret = hugetlb_fault(mm, vma, vaddr, (flags & FOLL_WRITE) ? FAULT_FLAG_WRITE : 0); |
a89182c76 Fix VM_FAULT flag... |
3387 |
if (!(ret & VM_FAULT_ERROR)) |
4c8872659 [PATCH] hugetlb: ... |
3388 |
continue; |
63551ae0f [PATCH] Hugepage ... |
3389 |
|
4c8872659 [PATCH] hugetlb: ... |
3390 |
remainder = 0; |
4c8872659 [PATCH] hugetlb: ... |
3391 3392 |
break; } |
a55164389 hugetlb: modular ... |
3393 |
pfn_offset = (vaddr & ~huge_page_mask(h)) >> PAGE_SHIFT; |
7f2e9525b hugetlbfs: common... |
3394 |
page = pte_page(huge_ptep_get(pte)); |
d5d4b0aa4 [PATCH] optimize ... |
3395 |
same_page: |
d6692183a [PATCH] fix extra... |
3396 |
if (pages) { |
2a15efc95 mm: follow_hugetl... |
3397 |
pages[i] = mem_map_offset(page, pfn_offset); |
a0368d4e4 mm: hugetlb: use ... |
3398 |
get_page_foll(pages[i]); |
d6692183a [PATCH] fix extra... |
3399 |
} |
63551ae0f [PATCH] Hugepage ... |
3400 3401 3402 3403 3404 |
if (vmas) vmas[i] = vma; vaddr += PAGE_SIZE; |
d5d4b0aa4 [PATCH] optimize ... |
3405 |
++pfn_offset; |
63551ae0f [PATCH] Hugepage ... |
3406 3407 |
--remainder; ++i; |
d5d4b0aa4 [PATCH] optimize ... |
3408 |
if (vaddr < vma->vm_end && remainder && |
a55164389 hugetlb: modular ... |
3409 |
pfn_offset < pages_per_huge_page(h)) { |
d5d4b0aa4 [PATCH] optimize ... |
3410 3411 3412 3413 3414 3415 |
/* * We use pfn_offset to avoid touching the pageframes * of this compound page. */ goto same_page; } |
cb900f412 mm, hugetlb: conv... |
3416 |
spin_unlock(ptl); |
63551ae0f [PATCH] Hugepage ... |
3417 |
} |
28a35716d mm: use long type... |
3418 |
*nr_pages = remainder; |
63551ae0f [PATCH] Hugepage ... |
3419 |
*position = vaddr; |
2a15efc95 mm: follow_hugetl... |
3420 |
return i ? i : -EFAULT; |
63551ae0f [PATCH] Hugepage ... |
3421 |
} |
8f860591f [PATCH] Enable mp... |
3422 |
|
7da4d641c mm: Count the num... |
3423 |
unsigned long hugetlb_change_protection(struct vm_area_struct *vma, |
8f860591f [PATCH] Enable mp... |
3424 3425 3426 3427 3428 3429 |
unsigned long address, unsigned long end, pgprot_t newprot) { struct mm_struct *mm = vma->vm_mm; unsigned long start = address; pte_t *ptep; pte_t pte; |
a55164389 hugetlb: modular ... |
3430 |
struct hstate *h = hstate_vma(vma); |
7da4d641c mm: Count the num... |
3431 |
unsigned long pages = 0; |
8f860591f [PATCH] Enable mp... |
3432 3433 3434 |
BUG_ON(address >= end); flush_cache_range(vma, address, end); |
a5338093b mm: move mmu noti... |
3435 |
mmu_notifier_invalidate_range_start(mm, start, end); |
83cde9e8b mm: use new helpe... |
3436 |
i_mmap_lock_write(vma->vm_file->f_mapping); |
a55164389 hugetlb: modular ... |
3437 |
for (; address < end; address += huge_page_size(h)) { |
cb900f412 mm, hugetlb: conv... |
3438 |
spinlock_t *ptl; |
8f860591f [PATCH] Enable mp... |
3439 3440 3441 |
ptep = huge_pte_offset(mm, address); if (!ptep) continue; |
cb900f412 mm, hugetlb: conv... |
3442 |
ptl = huge_pte_lock(h, mm, ptep); |
7da4d641c mm: Count the num... |
3443 3444 |
if (huge_pmd_unshare(mm, &address, ptep)) { pages++; |
cb900f412 mm, hugetlb: conv... |
3445 |
spin_unlock(ptl); |
39dde65c9 [PATCH] shared pa... |
3446 |
continue; |
7da4d641c mm: Count the num... |
3447 |
} |
a8bda28d8 mm/hugetlb: add m... |
3448 3449 3450 3451 3452 3453 3454 3455 3456 3457 3458 3459 3460 3461 3462 3463 3464 3465 3466 3467 |
pte = huge_ptep_get(ptep); if (unlikely(is_hugetlb_entry_hwpoisoned(pte))) { spin_unlock(ptl); continue; } if (unlikely(is_hugetlb_entry_migration(pte))) { swp_entry_t entry = pte_to_swp_entry(pte); if (is_write_migration_entry(entry)) { pte_t newpte; make_migration_entry_read(&entry); newpte = swp_entry_to_pte(entry); set_huge_pte_at(mm, address, ptep, newpte); pages++; } spin_unlock(ptl); continue; } if (!huge_pte_none(pte)) { |
8f860591f [PATCH] Enable mp... |
3468 |
pte = huge_ptep_get_and_clear(mm, address, ptep); |
106c992a5 mm/hugetlb: add m... |
3469 |
pte = pte_mkhuge(huge_pte_modify(pte, newprot)); |
be7517d6a mm/hugetlb: set P... |
3470 |
pte = arch_make_huge_pte(pte, vma, NULL, 0); |
8f860591f [PATCH] Enable mp... |
3471 |
set_huge_pte_at(mm, address, ptep, pte); |
7da4d641c mm: Count the num... |
3472 |
pages++; |
8f860591f [PATCH] Enable mp... |
3473 |
} |
cb900f412 mm, hugetlb: conv... |
3474 |
spin_unlock(ptl); |
8f860591f [PATCH] Enable mp... |
3475 |
} |
d833352a4 mm: hugetlbfs: cl... |
3476 |
/* |
c8c06efa8 mm: convert i_mma... |
3477 |
* Must flush TLB before releasing i_mmap_rwsem: x86's huge_pmd_unshare |
d833352a4 mm: hugetlbfs: cl... |
3478 |
* may have cleared our pud entry and done put_page on the page table: |
c8c06efa8 mm: convert i_mma... |
3479 |
* once we release i_mmap_rwsem, another task can do the final put_page |
d833352a4 mm: hugetlbfs: cl... |
3480 3481 |
* and that page table be reused and filled with junk. */ |
8f860591f [PATCH] Enable mp... |
3482 |
flush_tlb_range(vma, start, end); |
34ee645e8 mmu_notifier: cal... |
3483 |
mmu_notifier_invalidate_range(mm, start, end); |
83cde9e8b mm: use new helpe... |
3484 |
i_mmap_unlock_write(vma->vm_file->f_mapping); |
a5338093b mm: move mmu noti... |
3485 |
mmu_notifier_invalidate_range_end(mm, start, end); |
7da4d641c mm: Count the num... |
3486 3487 |
return pages << h->order; |
8f860591f [PATCH] Enable mp... |
3488 |
} |
a1e78772d hugetlb: reserve ... |
3489 3490 |
int hugetlb_reserve_pages(struct inode *inode, long from, long to, |
5a6fe1259 Do not account fo... |
3491 |
struct vm_area_struct *vma, |
ca16d140a mm: don't access ... |
3492 |
vm_flags_t vm_flags) |
e4e574b76 hugetlb: Try to g... |
3493 |
{ |
17c9d12e1 Do not account fo... |
3494 |
long ret, chg; |
a55164389 hugetlb: modular ... |
3495 |
struct hstate *h = hstate_inode(inode); |
90481622d hugepages: fix us... |
3496 |
struct hugepage_subpool *spool = subpool_inode(inode); |
9119a41e9 mm, hugetlb: unif... |
3497 |
struct resv_map *resv_map; |
1c5ecae3a hugetlbfs: add mi... |
3498 |
long gbl_reserve; |
e4e574b76 hugetlb: Try to g... |
3499 |
|
a1e78772d hugetlb: reserve ... |
3500 |
/* |
17c9d12e1 Do not account fo... |
3501 3502 |
* Only apply hugepage reservation if asked. At fault time, an * attempt will be made for VM_NORESERVE to allocate a page |
90481622d hugepages: fix us... |
3503 |
* without using reserves |
17c9d12e1 Do not account fo... |
3504 |
*/ |
ca16d140a mm: don't access ... |
3505 |
if (vm_flags & VM_NORESERVE) |
17c9d12e1 Do not account fo... |
3506 3507 3508 |
return 0; /* |
a1e78772d hugetlb: reserve ... |
3509 3510 3511 3512 3513 |
* Shared mappings base their reservation on the number of pages that * are already allocated on behalf of the file. Private mappings need * to reserve the full area even if read-only as mprotect() may be * called to make the mapping read-write. Assume !vma is a shm mapping */ |
9119a41e9 mm, hugetlb: unif... |
3514 |
if (!vma || vma->vm_flags & VM_MAYSHARE) { |
4e35f4838 mm, hugetlb: use ... |
3515 |
resv_map = inode_resv_map(inode); |
9119a41e9 mm, hugetlb: unif... |
3516 |
|
1406ec9ba mm, hugetlb: impr... |
3517 |
chg = region_chg(resv_map, from, to); |
9119a41e9 mm, hugetlb: unif... |
3518 3519 3520 |
} else { resv_map = resv_map_alloc(); |
17c9d12e1 Do not account fo... |
3521 3522 |
if (!resv_map) return -ENOMEM; |
a1e78772d hugetlb: reserve ... |
3523 |
chg = to - from; |
84afd99b8 hugetlb reservati... |
3524 |
|
17c9d12e1 Do not account fo... |
3525 3526 3527 |
set_vma_resv_map(vma, resv_map); set_vma_resv_flags(vma, HPAGE_RESV_OWNER); } |
c50ac0508 hugetlb: fix resv... |
3528 3529 3530 3531 |
if (chg < 0) { ret = chg; goto out_err; } |
8a6301127 pretend cpuset ha... |
3532 |
|
1c5ecae3a hugetlbfs: add mi... |
3533 3534 3535 3536 3537 3538 3539 |
/* * There must be enough pages in the subpool for the mapping. If * the subpool has a minimum size, there may be some global * reservations already in place (gbl_reserve). */ gbl_reserve = hugepage_subpool_get_pages(spool, chg); if (gbl_reserve < 0) { |
c50ac0508 hugetlb: fix resv... |
3540 3541 3542 |
ret = -ENOSPC; goto out_err; } |
5a6fe1259 Do not account fo... |
3543 3544 |
/* |
17c9d12e1 Do not account fo... |
3545 |
* Check enough hugepages are available for the reservation. |
90481622d hugepages: fix us... |
3546 |
* Hand the pages back to the subpool if there are not |
5a6fe1259 Do not account fo... |
3547 |
*/ |
1c5ecae3a hugetlbfs: add mi... |
3548 |
ret = hugetlb_acct_memory(h, gbl_reserve); |
68842c9b9 hugetlbfs: fix qu... |
3549 |
if (ret < 0) { |
1c5ecae3a hugetlbfs: add mi... |
3550 3551 |
/* put back original number of pages, chg */ (void)hugepage_subpool_put_pages(spool, chg); |
c50ac0508 hugetlb: fix resv... |
3552 |
goto out_err; |
68842c9b9 hugetlbfs: fix qu... |
3553 |
} |
17c9d12e1 Do not account fo... |
3554 3555 3556 3557 3558 3559 3560 3561 3562 3563 3564 3565 |
/* * Account for the reservations made. Shared mappings record regions * that have reservations as they are shared by multiple VMAs. * When the last VMA disappears, the region map says how much * the reservation was and the page cache tells how much of * the reservation was consumed. Private mappings are per-VMA and * only the consumed reservations are tracked. When the VMA * disappears, the original reservation is the VMA size and the * consumed reservations are stored in the map. Hence, nothing * else has to be done for private mappings here */ |
33039678c mm/hugetlb: handl... |
3566 3567 3568 3569 3570 3571 3572 3573 3574 3575 3576 3577 3578 3579 3580 3581 3582 3583 |
if (!vma || vma->vm_flags & VM_MAYSHARE) { long add = region_add(resv_map, from, to); if (unlikely(chg > add)) { /* * pages in this range were added to the reserve * map between region_chg and region_add. This * indicates a race with alloc_huge_page. Adjust * the subpool and reserve counts modified above * based on the difference. */ long rsv_adjust; rsv_adjust = hugepage_subpool_put_pages(spool, chg - add); hugetlb_acct_memory(h, -rsv_adjust); } } |
a43a8c39b [PATCH] tightenin... |
3584 |
return 0; |
c50ac0508 hugetlb: fix resv... |
3585 |
out_err: |
f031dd274 mm, hugetlb: remo... |
3586 3587 |
if (vma && is_vma_resv_set(vma, HPAGE_RESV_OWNER)) kref_put(&resv_map->refs, resv_map_release); |
c50ac0508 hugetlb: fix resv... |
3588 |
return ret; |
a43a8c39b [PATCH] tightenin... |
3589 3590 3591 3592 |
} void hugetlb_unreserve_pages(struct inode *inode, long offset, long freed) { |
a55164389 hugetlb: modular ... |
3593 |
struct hstate *h = hstate_inode(inode); |
4e35f4838 mm, hugetlb: use ... |
3594 |
struct resv_map *resv_map = inode_resv_map(inode); |
9119a41e9 mm, hugetlb: unif... |
3595 |
long chg = 0; |
90481622d hugepages: fix us... |
3596 |
struct hugepage_subpool *spool = subpool_inode(inode); |
1c5ecae3a hugetlbfs: add mi... |
3597 |
long gbl_reserve; |
45c682a68 hugetlb: fix i_bl... |
3598 |
|
9119a41e9 mm, hugetlb: unif... |
3599 |
if (resv_map) |
1406ec9ba mm, hugetlb: impr... |
3600 |
chg = region_truncate(resv_map, offset); |
45c682a68 hugetlb: fix i_bl... |
3601 |
spin_lock(&inode->i_lock); |
e4c6f8bed hugetlbfs: fix i_... |
3602 |
inode->i_blocks -= (blocks_per_huge_page(h) * freed); |
45c682a68 hugetlb: fix i_bl... |
3603 |
spin_unlock(&inode->i_lock); |
1c5ecae3a hugetlbfs: add mi... |
3604 3605 3606 3607 3608 3609 |
/* * If the subpool has a minimum size, the number of global * reservations to be released may be adjusted. */ gbl_reserve = hugepage_subpool_put_pages(spool, (chg - freed)); hugetlb_acct_memory(h, -gbl_reserve); |
a43a8c39b [PATCH] tightenin... |
3610 |
} |
93f70f900 HWPOISON, hugetlb... |
3611 |
|
3212b535f mm: hugetlb: Copy... |
3612 3613 3614 3615 3616 3617 3618 3619 3620 3621 3622 3623 3624 3625 3626 3627 3628 3629 3630 3631 3632 3633 3634 3635 3636 3637 3638 3639 3640 3641 3642 3643 3644 3645 3646 3647 3648 3649 3650 3651 3652 3653 3654 3655 3656 |
#ifdef CONFIG_ARCH_WANT_HUGE_PMD_SHARE static unsigned long page_table_shareable(struct vm_area_struct *svma, struct vm_area_struct *vma, unsigned long addr, pgoff_t idx) { unsigned long saddr = ((idx - svma->vm_pgoff) << PAGE_SHIFT) + svma->vm_start; unsigned long sbase = saddr & PUD_MASK; unsigned long s_end = sbase + PUD_SIZE; /* Allow segments to share if only one is marked locked */ unsigned long vm_flags = vma->vm_flags & ~VM_LOCKED; unsigned long svm_flags = svma->vm_flags & ~VM_LOCKED; /* * match the virtual addresses, permission and the alignment of the * page table page. */ if (pmd_index(addr) != pmd_index(saddr) || vm_flags != svm_flags || sbase < svma->vm_start || svma->vm_end < s_end) return 0; return saddr; } static int vma_shareable(struct vm_area_struct *vma, unsigned long addr) { unsigned long base = addr & PUD_MASK; unsigned long end = base + PUD_SIZE; /* * check on proper vm_flags and page table alignment */ if (vma->vm_flags & VM_MAYSHARE && vma->vm_start <= base && end <= vma->vm_end) return 1; return 0; } /* * Search for a shareable pmd page for hugetlb. In any case calls pmd_alloc() * and returns the corresponding pte. While this is not necessary for the * !shared pmd case because we can allocate the pmd later as well, it makes the * code much cleaner. pmd allocation is essential for the shared case because |
c8c06efa8 mm: convert i_mma... |
3657 |
* pud has to be populated inside the same i_mmap_rwsem section - otherwise |
3212b535f mm: hugetlb: Copy... |
3658 3659 3660 3661 3662 3663 3664 3665 3666 3667 3668 3669 3670 |
* racing tasks could either miss the sharing (see huge_pte_offset) or select a * bad pmd for sharing. */ pte_t *huge_pmd_share(struct mm_struct *mm, unsigned long addr, pud_t *pud) { struct vm_area_struct *vma = find_vma(mm, addr); struct address_space *mapping = vma->vm_file->f_mapping; pgoff_t idx = ((addr - vma->vm_start) >> PAGE_SHIFT) + vma->vm_pgoff; struct vm_area_struct *svma; unsigned long saddr; pte_t *spte = NULL; pte_t *pte; |
cb900f412 mm, hugetlb: conv... |
3671 |
spinlock_t *ptl; |
3212b535f mm: hugetlb: Copy... |
3672 3673 3674 |
if (!vma_shareable(vma, addr)) return (pte_t *)pmd_alloc(mm, pud, addr); |
83cde9e8b mm: use new helpe... |
3675 |
i_mmap_lock_write(mapping); |
3212b535f mm: hugetlb: Copy... |
3676 3677 3678 3679 3680 3681 3682 3683 |
vma_interval_tree_foreach(svma, &mapping->i_mmap, idx, idx) { if (svma == vma) continue; saddr = page_table_shareable(svma, vma, addr, idx); if (saddr) { spte = huge_pte_offset(svma->vm_mm, saddr); if (spte) { |
dc6c9a35b mm: account pmd p... |
3684 |
mm_inc_nr_pmds(mm); |
3212b535f mm: hugetlb: Copy... |
3685 3686 3687 3688 3689 3690 3691 3692 |
get_page(virt_to_page(spte)); break; } } } if (!spte) goto out; |
cb900f412 mm, hugetlb: conv... |
3693 3694 |
ptl = huge_pte_lockptr(hstate_vma(vma), mm, spte); spin_lock(ptl); |
dc6c9a35b mm: account pmd p... |
3695 |
if (pud_none(*pud)) { |
3212b535f mm: hugetlb: Copy... |
3696 3697 |
pud_populate(mm, pud, (pmd_t *)((unsigned long)spte & PAGE_MASK)); |
dc6c9a35b mm: account pmd p... |
3698 |
} else { |
3212b535f mm: hugetlb: Copy... |
3699 |
put_page(virt_to_page(spte)); |
dc6c9a35b mm: account pmd p... |
3700 3701 |
mm_inc_nr_pmds(mm); } |
cb900f412 mm, hugetlb: conv... |
3702 |
spin_unlock(ptl); |
3212b535f mm: hugetlb: Copy... |
3703 3704 |
out: pte = (pte_t *)pmd_alloc(mm, pud, addr); |
83cde9e8b mm: use new helpe... |
3705 |
i_mmap_unlock_write(mapping); |
3212b535f mm: hugetlb: Copy... |
3706 3707 3708 3709 3710 3711 3712 3713 3714 3715 |
return pte; } /* * unmap huge page backed by shared pte. * * Hugetlb pte page is ref counted at the time of mapping. If pte is shared * indicated by page_count > 1, unmap is achieved by clearing pud and * decrementing the ref count. If count == 1, the pte page is not shared. * |
cb900f412 mm, hugetlb: conv... |
3716 |
* called with page table lock held. |
3212b535f mm: hugetlb: Copy... |
3717 3718 3719 3720 3721 3722 3723 3724 3725 3726 3727 3728 3729 3730 3731 |
* * returns: 1 successfully unmapped a shared pte page * 0 the underlying pte page is not shared, or it is the last user */ int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep) { pgd_t *pgd = pgd_offset(mm, *addr); pud_t *pud = pud_offset(pgd, *addr); BUG_ON(page_count(virt_to_page(ptep)) == 0); if (page_count(virt_to_page(ptep)) == 1) return 0; pud_clear(pud); put_page(virt_to_page(ptep)); |
dc6c9a35b mm: account pmd p... |
3732 |
mm_dec_nr_pmds(mm); |
3212b535f mm: hugetlb: Copy... |
3733 3734 3735 |
*addr = ALIGN(*addr, HPAGE_SIZE * PTRS_PER_PTE) - HPAGE_SIZE; return 1; } |
9e5fc74c3 mm: hugetlb: Copy... |
3736 3737 3738 3739 3740 3741 |
#define want_pmd_share() (1) #else /* !CONFIG_ARCH_WANT_HUGE_PMD_SHARE */ pte_t *huge_pmd_share(struct mm_struct *mm, unsigned long addr, pud_t *pud) { return NULL; } |
e81f2d223 mm/hugetlb: reduc... |
3742 3743 3744 3745 3746 |
int huge_pmd_unshare(struct mm_struct *mm, unsigned long *addr, pte_t *ptep) { return 0; } |
9e5fc74c3 mm: hugetlb: Copy... |
3747 |
#define want_pmd_share() (0) |
3212b535f mm: hugetlb: Copy... |
3748 |
#endif /* CONFIG_ARCH_WANT_HUGE_PMD_SHARE */ |
9e5fc74c3 mm: hugetlb: Copy... |
3749 3750 3751 3752 3753 3754 3755 3756 3757 3758 3759 3760 3761 3762 3763 3764 3765 3766 3767 3768 3769 3770 3771 3772 3773 3774 3775 3776 3777 3778 3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 3790 3791 |
#ifdef CONFIG_ARCH_WANT_GENERAL_HUGETLB pte_t *huge_pte_alloc(struct mm_struct *mm, unsigned long addr, unsigned long sz) { pgd_t *pgd; pud_t *pud; pte_t *pte = NULL; pgd = pgd_offset(mm, addr); pud = pud_alloc(mm, pgd, addr); if (pud) { if (sz == PUD_SIZE) { pte = (pte_t *)pud; } else { BUG_ON(sz != PMD_SIZE); if (want_pmd_share() && pud_none(*pud)) pte = huge_pmd_share(mm, addr, pud); else pte = (pte_t *)pmd_alloc(mm, pud, addr); } } BUG_ON(pte && !pte_none(*pte) && !pte_huge(*pte)); return pte; } pte_t *huge_pte_offset(struct mm_struct *mm, unsigned long addr) { pgd_t *pgd; pud_t *pud; pmd_t *pmd = NULL; pgd = pgd_offset(mm, addr); if (pgd_present(*pgd)) { pud = pud_offset(pgd, addr); if (pud_present(*pud)) { if (pud_huge(*pud)) return (pte_t *)pud; pmd = pmd_offset(pud, addr); } } return (pte_t *) pmd; } |
61f77eda9 mm/hugetlb: reduc... |
3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 |
#endif /* CONFIG_ARCH_WANT_GENERAL_HUGETLB */ /* * These functions are overwritable if your architecture needs its own * behavior. */ struct page * __weak follow_huge_addr(struct mm_struct *mm, unsigned long address, int write) { return ERR_PTR(-EINVAL); } struct page * __weak |
9e5fc74c3 mm: hugetlb: Copy... |
3806 |
follow_huge_pmd(struct mm_struct *mm, unsigned long address, |
e66f17ff7 mm/hugetlb: take ... |
3807 |
pmd_t *pmd, int flags) |
9e5fc74c3 mm: hugetlb: Copy... |
3808 |
{ |
e66f17ff7 mm/hugetlb: take ... |
3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 |
struct page *page = NULL; spinlock_t *ptl; retry: ptl = pmd_lockptr(mm, pmd); spin_lock(ptl); /* * make sure that the address range covered by this pmd is not * unmapped from other threads. */ if (!pmd_huge(*pmd)) goto out; if (pmd_present(*pmd)) { |
975341270 mm/hugetlb: use p... |
3821 |
page = pmd_page(*pmd) + ((address & ~PMD_MASK) >> PAGE_SHIFT); |
e66f17ff7 mm/hugetlb: take ... |
3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 |
if (flags & FOLL_GET) get_page(page); } else { if (is_hugetlb_entry_migration(huge_ptep_get((pte_t *)pmd))) { spin_unlock(ptl); __migration_entry_wait(mm, (pte_t *)pmd, ptl); goto retry; } /* * hwpoisoned entry is treated as no_page_table in * follow_page_mask(). */ } out: spin_unlock(ptl); |
9e5fc74c3 mm: hugetlb: Copy... |
3837 3838 |
return page; } |
61f77eda9 mm/hugetlb: reduc... |
3839 |
struct page * __weak |
9e5fc74c3 mm: hugetlb: Copy... |
3840 |
follow_huge_pud(struct mm_struct *mm, unsigned long address, |
e66f17ff7 mm/hugetlb: take ... |
3841 |
pud_t *pud, int flags) |
9e5fc74c3 mm: hugetlb: Copy... |
3842 |
{ |
e66f17ff7 mm/hugetlb: take ... |
3843 3844 |
if (flags & FOLL_GET) return NULL; |
9e5fc74c3 mm: hugetlb: Copy... |
3845 |
|
e66f17ff7 mm/hugetlb: take ... |
3846 |
return pte_page(*(pte_t *)pud) + ((address & ~PUD_MASK) >> PAGE_SHIFT); |
9e5fc74c3 mm: hugetlb: Copy... |
3847 |
} |
d5bd91069 hugepage: move is... |
3848 |
#ifdef CONFIG_MEMORY_FAILURE |
93f70f900 HWPOISON, hugetlb... |
3849 3850 3851 3852 |
/* * This function is called from memory failure code. * Assume the caller holds page lock of the head page. */ |
6de2b1aab HWPOISON, hugetlb... |
3853 |
int dequeue_hwpoisoned_huge_page(struct page *hpage) |
93f70f900 HWPOISON, hugetlb... |
3854 3855 3856 |
{ struct hstate *h = page_hstate(hpage); int nid = page_to_nid(hpage); |
6de2b1aab HWPOISON, hugetlb... |
3857 |
int ret = -EBUSY; |
93f70f900 HWPOISON, hugetlb... |
3858 3859 |
spin_lock(&hugetlb_lock); |
7e1f049ef mm: hugetlb: clea... |
3860 3861 3862 3863 3864 |
/* * Just checking !page_huge_active is not enough, because that could be * an isolated/hwpoisoned hugepage (which have >0 refcount). */ if (!page_huge_active(hpage) && !page_count(hpage)) { |
56f2fb147 mm/hugetlb.c: fix... |
3865 3866 3867 3868 3869 3870 3871 |
/* * Hwpoisoned hugepage isn't linked to activelist or freelist, * but dangling hpage->lru can trigger list-debug warnings * (this happens when we call unpoison_memory() on it), * so let it point to itself with list_del_init(). */ list_del_init(&hpage->lru); |
8c6c2ecb4 HWPOSION, hugetlb... |
3872 |
set_page_refcounted(hpage); |
6de2b1aab HWPOISON, hugetlb... |
3873 3874 3875 3876 |
h->free_huge_pages--; h->free_huge_pages_node[nid]--; ret = 0; } |
93f70f900 HWPOISON, hugetlb... |
3877 |
spin_unlock(&hugetlb_lock); |
6de2b1aab HWPOISON, hugetlb... |
3878 |
return ret; |
93f70f900 HWPOISON, hugetlb... |
3879 |
} |
6de2b1aab HWPOISON, hugetlb... |
3880 |
#endif |
31caf665e mm: migrate: make... |
3881 3882 3883 |
bool isolate_huge_page(struct page *page, struct list_head *list) { |
bcc542223 mm: hugetlb: intr... |
3884 |
bool ret = true; |
309381fea mm: dump page whe... |
3885 |
VM_BUG_ON_PAGE(!PageHead(page), page); |
31caf665e mm: migrate: make... |
3886 |
spin_lock(&hugetlb_lock); |
bcc542223 mm: hugetlb: intr... |
3887 3888 3889 3890 3891 |
if (!page_huge_active(page) || !get_page_unless_zero(page)) { ret = false; goto unlock; } clear_page_huge_active(page); |
31caf665e mm: migrate: make... |
3892 |
list_move_tail(&page->lru, list); |
bcc542223 mm: hugetlb: intr... |
3893 |
unlock: |
31caf665e mm: migrate: make... |
3894 |
spin_unlock(&hugetlb_lock); |
bcc542223 mm: hugetlb: intr... |
3895 |
return ret; |
31caf665e mm: migrate: make... |
3896 3897 3898 3899 |
} void putback_active_hugepage(struct page *page) { |
309381fea mm: dump page whe... |
3900 |
VM_BUG_ON_PAGE(!PageHead(page), page); |
31caf665e mm: migrate: make... |
3901 |
spin_lock(&hugetlb_lock); |
bcc542223 mm: hugetlb: intr... |
3902 |
set_page_huge_active(page); |
31caf665e mm: migrate: make... |
3903 3904 3905 3906 |
list_move_tail(&page->lru, &(page_hstate(page))->hugepage_activelist); spin_unlock(&hugetlb_lock); put_page(page); } |