Blame view
mm/page_isolation.c
8.74 KB
a5d76b54a memory unplug: pa... |
1 2 3 |
/* * linux/mm/page_isolation.c */ |
a5d76b54a memory unplug: pa... |
4 5 6 |
#include <linux/mm.h> #include <linux/page-isolation.h> #include <linux/pageblock-flags.h> |
ee6f509c3 mm: factor out me... |
7 |
#include <linux/memory.h> |
c8721bbbd mm: memory-hotplu... |
8 |
#include <linux/hugetlb.h> |
a5d76b54a memory unplug: pa... |
9 |
#include "internal.h" |
b023f4681 memory-hotplug: s... |
10 |
int set_migratetype_isolate(struct page *page, bool skip_hwpoisoned_pages) |
ee6f509c3 mm: factor out me... |
11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 |
{ struct zone *zone; unsigned long flags, pfn; struct memory_isolate_notify arg; int notifier_ret; int ret = -EBUSY; zone = page_zone(page); spin_lock_irqsave(&zone->lock, flags); pfn = page_to_pfn(page); arg.start_pfn = pfn; arg.nr_pages = pageblock_nr_pages; arg.pages_found = 0; /* * It may be possible to isolate a pageblock even if the * migratetype is not MIGRATE_MOVABLE. The memory isolation * notifier chain is used by balloon drivers to return the * number of pages in a range that are held by the balloon * driver to shrink memory. If all the pages are accounted for * by balloons, are free, or on the LRU, isolation can continue. * Later, for example, when memory hotplug notifier runs, these * pages reported as "can be isolated" should be isolated(freed) * by the balloon driver through the memory notifier chain. */ notifier_ret = memory_isolate_notify(MEM_ISOLATE_COUNT, &arg); notifier_ret = notifier_to_errno(notifier_ret); if (notifier_ret) goto out; /* * FIXME: Now, memory hotplug doesn't call shrink_slab() by itself. * We just check MOVABLE pages. */ |
b023f4681 memory-hotplug: s... |
46 47 |
if (!has_unmovable_pages(zone, page, arg.pages_found, skip_hwpoisoned_pages)) |
ee6f509c3 mm: factor out me... |
48 49 50 51 52 53 54 55 56 |
ret = 0; /* * immobile means "not-on-lru" paes. If immobile is larger than * removable-by-driver pages reported by notifier, we'll fail. */ out: if (!ret) { |
2139cbe62 cma: fix counting... |
57 |
unsigned long nr_pages; |
d1ce749a0 cma: count free C... |
58 |
int migratetype = get_pageblock_migratetype(page); |
2139cbe62 cma: fix counting... |
59 |
|
a458431e1 mm: fix zone_wate... |
60 |
set_pageblock_migratetype(page, MIGRATE_ISOLATE); |
ad53f92eb mm/page_alloc: fi... |
61 |
zone->nr_isolate_pageblock++; |
2139cbe62 cma: fix counting... |
62 |
nr_pages = move_freepages_block(zone, page, MIGRATE_ISOLATE); |
d1ce749a0 cma: count free C... |
63 |
__mod_zone_freepage_state(zone, -nr_pages, migratetype); |
ee6f509c3 mm: factor out me... |
64 65 66 67 |
} spin_unlock_irqrestore(&zone->lock, flags); if (!ret) |
ec25af84b mm, page_isolatio... |
68 |
drain_all_pages(zone); |
ee6f509c3 mm: factor out me... |
69 70 71 72 73 74 |
return ret; } void unset_migratetype_isolate(struct page *page, unsigned migratetype) { struct zone *zone; |
2139cbe62 cma: fix counting... |
75 |
unsigned long flags, nr_pages; |
3c605096d mm/page_alloc: re... |
76 77 78 79 |
struct page *isolated_page = NULL; unsigned int order; unsigned long page_idx, buddy_idx; struct page *buddy; |
2139cbe62 cma: fix counting... |
80 |
|
ee6f509c3 mm: factor out me... |
81 82 83 84 |
zone = page_zone(page); spin_lock_irqsave(&zone->lock, flags); if (get_pageblock_migratetype(page) != MIGRATE_ISOLATE) goto out; |
3c605096d mm/page_alloc: re... |
85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 |
/* * Because freepage with more than pageblock_order on isolated * pageblock is restricted to merge due to freepage counting problem, * it is possible that there is free buddy page. * move_freepages_block() doesn't care of merge so we need other * approach in order to merge them. Isolation and free will make * these pages to be merged. */ if (PageBuddy(page)) { order = page_order(page); if (order >= pageblock_order) { page_idx = page_to_pfn(page) & ((1 << MAX_ORDER) - 1); buddy_idx = __find_buddy_index(page_idx, order); buddy = page + (buddy_idx - page_idx); |
1ae7013df CMA: page_isolati... |
100 101 |
if (pfn_valid_within(page_to_pfn(buddy)) && !is_migrate_isolate_page(buddy)) { |
3c605096d mm/page_alloc: re... |
102 |
__isolate_free_page(page, order); |
cfa869438 mm/page_alloc.c: ... |
103 |
kernel_map_pages(page, (1 << order), 1); |
3c605096d mm/page_alloc: re... |
104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 |
set_page_refcounted(page); isolated_page = page; } } } /* * If we isolate freepage with more than pageblock_order, there * should be no freepage in the range, so we could avoid costly * pageblock scanning for freepage moving. */ if (!isolated_page) { nr_pages = move_freepages_block(zone, page, migratetype); __mod_zone_freepage_state(zone, nr_pages, migratetype); } |
a458431e1 mm: fix zone_wate... |
119 |
set_pageblock_migratetype(page, migratetype); |
ad53f92eb mm/page_alloc: fi... |
120 |
zone->nr_isolate_pageblock--; |
ee6f509c3 mm: factor out me... |
121 122 |
out: spin_unlock_irqrestore(&zone->lock, flags); |
3c605096d mm/page_alloc: re... |
123 124 |
if (isolated_page) __free_pages(isolated_page, order); |
ee6f509c3 mm: factor out me... |
125 |
} |
a5d76b54a memory unplug: pa... |
126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 |
static inline struct page * __first_valid_page(unsigned long pfn, unsigned long nr_pages) { int i; for (i = 0; i < nr_pages; i++) if (pfn_valid_within(pfn + i)) break; if (unlikely(i == nr_pages)) return NULL; return pfn_to_page(pfn + i); } /* * start_isolate_page_range() -- make page-allocation-type of range of pages * to be MIGRATE_ISOLATE. * @start_pfn: The lower PFN of the range to be isolated. * @end_pfn: The upper PFN of the range to be isolated. |
0815f3d81 mm: page_isolatio... |
143 |
* @migratetype: migrate type to set in error recovery. |
a5d76b54a memory unplug: pa... |
144 145 146 147 148 149 150 151 |
* * Making page-allocation-type to be MIGRATE_ISOLATE means free pages in * the range will never be allocated. Any free pages and pages freed in the * future will not be allocated again. * * start_pfn/end_pfn must be aligned to pageblock_order. * Returns 0 on success and -EBUSY if any part of range cannot be isolated. */ |
0815f3d81 mm: page_isolatio... |
152 |
int start_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn, |
b023f4681 memory-hotplug: s... |
153 |
unsigned migratetype, bool skip_hwpoisoned_pages) |
a5d76b54a memory unplug: pa... |
154 155 156 157 158 159 160 161 162 163 164 165 |
{ unsigned long pfn; unsigned long undo_pfn; struct page *page; BUG_ON((start_pfn) & (pageblock_nr_pages - 1)); BUG_ON((end_pfn) & (pageblock_nr_pages - 1)); for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) { page = __first_valid_page(pfn, pageblock_nr_pages); |
b023f4681 memory-hotplug: s... |
166 167 |
if (page && set_migratetype_isolate(page, skip_hwpoisoned_pages)) { |
a5d76b54a memory unplug: pa... |
168 169 170 171 172 173 174 |
undo_pfn = pfn; goto undo; } } return 0; undo: for (pfn = start_pfn; |
dbc0e4cef memory hotremove:... |
175 |
pfn < undo_pfn; |
a5d76b54a memory unplug: pa... |
176 |
pfn += pageblock_nr_pages) |
0815f3d81 mm: page_isolatio... |
177 |
unset_migratetype_isolate(pfn_to_page(pfn), migratetype); |
a5d76b54a memory unplug: pa... |
178 179 180 181 182 183 184 |
return -EBUSY; } /* * Make isolated pages available again. */ |
0815f3d81 mm: page_isolatio... |
185 186 |
int undo_isolate_page_range(unsigned long start_pfn, unsigned long end_pfn, unsigned migratetype) |
a5d76b54a memory unplug: pa... |
187 188 189 190 191 192 193 194 195 |
{ unsigned long pfn; struct page *page; BUG_ON((start_pfn) & (pageblock_nr_pages - 1)); BUG_ON((end_pfn) & (pageblock_nr_pages - 1)); for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) { page = __first_valid_page(pfn, pageblock_nr_pages); |
dbc0e4cef memory hotremove:... |
196 |
if (!page || get_pageblock_migratetype(page) != MIGRATE_ISOLATE) |
a5d76b54a memory unplug: pa... |
197 |
continue; |
0815f3d81 mm: page_isolatio... |
198 |
unset_migratetype_isolate(page, migratetype); |
a5d76b54a memory unplug: pa... |
199 200 201 202 203 204 205 206 |
} return 0; } /* * Test all pages in the range is free(means isolated) or not. * all pages in [start_pfn...end_pfn) must be in the same zone. * zone->lock must be held before call this. * |
0815f3d81 mm: page_isolatio... |
207 |
* Returns 1 if all pages in the range are isolated. |
a5d76b54a memory unplug: pa... |
208 209 |
*/ static int |
b023f4681 memory-hotplug: s... |
210 211 |
__test_page_isolated_in_pageblock(unsigned long pfn, unsigned long end_pfn, bool skip_hwpoisoned_pages) |
a5d76b54a memory unplug: pa... |
212 213 214 215 216 217 218 219 220 |
{ struct page *page; while (pfn < end_pfn) { if (!pfn_valid_within(pfn)) { pfn++; continue; } page = pfn_to_page(pfn); |
41d575ad4 memory-hotplug: b... |
221 |
if (PageBuddy(page)) { |
435b405c0 memory-hotplug: f... |
222 223 224 225 226 227 228 229 230 231 232 233 234 235 |
/* * If race between isolatation and allocation happens, * some free pages could be in MIGRATE_MOVABLE list * although pageblock's migratation type of the page * is MIGRATE_ISOLATE. Catch it and move the page into * MIGRATE_ISOLATE list. */ if (get_freepage_migratetype(page) != MIGRATE_ISOLATE) { struct page *end_page; end_page = page + (1 << page_order(page)) - 1; move_freepages(page_zone(page), page, end_page, MIGRATE_ISOLATE); } |
a5d76b54a memory unplug: pa... |
236 |
pfn += 1 << page_order(page); |
41d575ad4 memory-hotplug: b... |
237 |
} |
a5d76b54a memory unplug: pa... |
238 |
else if (page_count(page) == 0 && |
b12c4ad14 mm: page_alloc: u... |
239 |
get_freepage_migratetype(page) == MIGRATE_ISOLATE) |
a5d76b54a memory unplug: pa... |
240 |
pfn += 1; |
b023f4681 memory-hotplug: s... |
241 242 243 244 245 246 247 248 |
else if (skip_hwpoisoned_pages && PageHWPoison(page)) { /* * The HWPoisoned page may be not in buddy * system, and page_count() is not 0. */ pfn++; continue; } |
a5d76b54a memory unplug: pa... |
249 250 251 252 253 254 255 |
else break; } if (pfn < end_pfn) return 0; return 1; } |
b023f4681 memory-hotplug: s... |
256 257 |
int test_pages_isolated(unsigned long start_pfn, unsigned long end_pfn, bool skip_hwpoisoned_pages) |
a5d76b54a memory unplug: pa... |
258 |
{ |
6c1b7f680 memory hotplug: m... |
259 |
unsigned long pfn, flags; |
a5d76b54a memory unplug: pa... |
260 |
struct page *page; |
6c1b7f680 memory hotplug: m... |
261 262 |
struct zone *zone; int ret; |
a5d76b54a memory unplug: pa... |
263 |
|
a5d76b54a memory unplug: pa... |
264 |
/* |
85dbe7060 page_isolation: F... |
265 266 267 |
* Note: pageblock_nr_pages != MAX_ORDER. Then, chunks of free pages * are not aligned to pageblock_nr_pages. * Then we just check migratetype first. |
a5d76b54a memory unplug: pa... |
268 269 270 |
*/ for (pfn = start_pfn; pfn < end_pfn; pfn += pageblock_nr_pages) { page = __first_valid_page(pfn, pageblock_nr_pages); |
dbc0e4cef memory hotremove:... |
271 |
if (page && get_pageblock_migratetype(page) != MIGRATE_ISOLATE) |
a5d76b54a memory unplug: pa... |
272 273 |
break; } |
a70dcb969 memory hotplug: f... |
274 275 |
page = __first_valid_page(start_pfn, end_pfn - start_pfn); if ((pfn < end_pfn) || !page) |
a5d76b54a memory unplug: pa... |
276 |
return -EBUSY; |
85dbe7060 page_isolation: F... |
277 |
/* Check all pages are free or marked as ISOLATED */ |
a70dcb969 memory hotplug: f... |
278 |
zone = page_zone(page); |
6c1b7f680 memory hotplug: m... |
279 |
spin_lock_irqsave(&zone->lock, flags); |
b023f4681 memory-hotplug: s... |
280 281 |
ret = __test_page_isolated_in_pageblock(start_pfn, end_pfn, skip_hwpoisoned_pages); |
6c1b7f680 memory hotplug: m... |
282 283 |
spin_unlock_irqrestore(&zone->lock, flags); return ret ? 0 : -EBUSY; |
a5d76b54a memory unplug: pa... |
284 |
} |
723a0644a mm/page_alloc: re... |
285 286 287 288 289 |
struct page *alloc_migrate_target(struct page *page, unsigned long private, int **resultp) { gfp_t gfp_mask = GFP_USER | __GFP_MOVABLE; |
c8721bbbd mm: memory-hotplu... |
290 291 292 293 294 295 296 297 298 299 300 301 |
/* * TODO: allocate a destination hugepage from a nearest neighbor node, * accordance with memory policy of the user process if possible. For * now as a simple work-around, we use the next node for destination. */ if (PageHuge(page)) { nodemask_t src = nodemask_of_node(page_to_nid(page)); nodemask_t dst; nodes_complement(dst, src); return alloc_huge_page_node(page_hstate(compound_head(page)), next_node(page_to_nid(page), dst)); } |
723a0644a mm/page_alloc: re... |
302 303 304 305 306 |
if (PageHighMem(page)) gfp_mask |= __GFP_HIGHMEM; return alloc_page(gfp_mask); } |