Blame view
mm/zsmalloc.c
48.4 KB
61989a80f staging: zsmalloc... |
1 2 3 4 |
/* * zsmalloc memory allocator * * Copyright (C) 2011 Nitin Gupta |
31fc00bb7 zsmalloc: add cop... |
5 |
* Copyright (C) 2012, 2013 Minchan Kim |
61989a80f staging: zsmalloc... |
6 7 8 9 10 11 12 |
* * This code is released using a dual license strategy: BSD/GPL * You can choose the license that better fits your requirements. * * Released under the terms of 3-clause BSD License * Released under the terms of GNU General Public License Version 2.0 */ |
2db51dae5 staging: zsmalloc... |
13 |
/* |
2db51dae5 staging: zsmalloc... |
14 15 16 17 |
* Following is how we use various fields and flags of underlying * struct page(s) to form a zspage. * * Usage of struct page fields: |
32e7ba1ea zsmalloc: use pag... |
18 |
* page->private: points to the first component (0-order) page |
2db51dae5 staging: zsmalloc... |
19 20 21 22 23 24 25 26 27 |
* page->index (union with page->freelist): offset of the first object * starting in this page. For the first page, this is * always 0, so we use this field (aka freelist) to point * to the first free object in zspage. * page->lru: links together all component pages (except the first page) * of a zspage * * For _first_ page only: * |
32e7ba1ea zsmalloc: use pag... |
28 |
* page->private: refers to the component page after the first page |
7b60a6852 zsmalloc: record ... |
29 30 |
* If the page is first_page for huge object, it stores handle. * Look at size_class->huge. |
2db51dae5 staging: zsmalloc... |
31 32 33 34 35 36 37 38 |
* page->freelist: points to the first free object in zspage. * Free objects are linked together using in-place * metadata. * page->objects: maximum number of objects we can store in this * zspage (class->zspage_order * PAGE_SIZE / class->size) * page->lru: links together first pages of various zspages. * Basically forming list of zspages in a fullness group. * page->mapping: class index and fullness group of the zspage |
8f958c98f zsmalloc: add com... |
39 |
* page->inuse: the number of objects that are used in this zspage |
2db51dae5 staging: zsmalloc... |
40 41 42 43 44 45 |
* * Usage of struct page flags: * PG_private: identifies the first component page * PG_private2: identifies the last component page * */ |
61989a80f staging: zsmalloc... |
46 47 |
#include <linux/module.h> #include <linux/kernel.h> |
312fcae22 zsmalloc: support... |
48 |
#include <linux/sched.h> |
61989a80f staging: zsmalloc... |
49 50 51 |
#include <linux/bitops.h> #include <linux/errno.h> #include <linux/highmem.h> |
61989a80f staging: zsmalloc... |
52 53 54 55 56 57 |
#include <linux/string.h> #include <linux/slab.h> #include <asm/tlbflush.h> #include <asm/pgtable.h> #include <linux/cpumask.h> #include <linux/cpu.h> |
0cbb613fa staging: fix powe... |
58 |
#include <linux/vmalloc.h> |
759b26b29 zsmalloc: use pre... |
59 |
#include <linux/preempt.h> |
0959c63f1 zsmalloc: collaps... |
60 61 |
#include <linux/spinlock.h> #include <linux/types.h> |
0f050d997 mm/zsmalloc: add ... |
62 |
#include <linux/debugfs.h> |
bcf1647d0 zsmalloc: move it... |
63 |
#include <linux/zsmalloc.h> |
c795779df mm/zpool: zbud/zs... |
64 |
#include <linux/zpool.h> |
0959c63f1 zsmalloc: collaps... |
65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 |
/* * This must be power of 2 and greater than of equal to sizeof(link_free). * These two conditions ensure that any 'struct link_free' itself doesn't * span more than 1 page which avoids complex case of mapping 2 pages simply * to restore link_free pointer values. */ #define ZS_ALIGN 8 /* * A single 'zspage' is composed of up to 2^N discontiguous 0-order (single) * pages. ZS_MAX_ZSPAGE_ORDER defines upper limit on N. */ #define ZS_MAX_ZSPAGE_ORDER 2 #define ZS_MAX_PAGES_PER_ZSPAGE (_AC(1, UL) << ZS_MAX_ZSPAGE_ORDER) |
2e40e163a zsmalloc: decoupl... |
80 |
#define ZS_HANDLE_SIZE (sizeof(unsigned long)) |
0959c63f1 zsmalloc: collaps... |
81 82 |
/* * Object location (<PFN>, <obj_idx>) is encoded as |
c3e3e88ad zsmalloc: add mor... |
83 |
* as single (unsigned long) handle value. |
0959c63f1 zsmalloc: collaps... |
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 |
* * Note that object index <obj_idx> is relative to system * page <PFN> it is stored in, so for each sub-page belonging * to a zspage, obj_idx starts with 0. * * This is made more complicated by various memory models and PAE. */ #ifndef MAX_PHYSMEM_BITS #ifdef CONFIG_HIGHMEM64G #define MAX_PHYSMEM_BITS 36 #else /* !CONFIG_HIGHMEM64G */ /* * If this definition of MAX_PHYSMEM_BITS is used, OBJ_INDEX_BITS will just * be PAGE_SHIFT */ #define MAX_PHYSMEM_BITS BITS_PER_LONG #endif #endif #define _PFN_BITS (MAX_PHYSMEM_BITS - PAGE_SHIFT) |
312fcae22 zsmalloc: support... |
104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 |
/* * Memory for allocating for handle keeps object position by * encoding <page, obj_idx> and the encoded value has a room * in least bit(ie, look at obj_to_location). * We use the bit to synchronize between object access by * user and migration. */ #define HANDLE_PIN_BIT 0 /* * Head in allocated object should have OBJ_ALLOCATED_TAG * to identify the object was allocated or not. * It's okay to add the status bit in the least bit because * header keeps handle which is 4byte-aligned address so we * have room for two bit at least. */ #define OBJ_ALLOCATED_TAG 1 #define OBJ_TAG_BITS 1 #define OBJ_INDEX_BITS (BITS_PER_LONG - _PFN_BITS - OBJ_TAG_BITS) |
0959c63f1 zsmalloc: collaps... |
124 125 126 127 128 129 |
#define OBJ_INDEX_MASK ((_AC(1, UL) << OBJ_INDEX_BITS) - 1) #define MAX(a, b) ((a) >= (b) ? (a) : (b)) /* ZS_MIN_ALLOC_SIZE must be multiple of ZS_ALIGN */ #define ZS_MIN_ALLOC_SIZE \ MAX(32, (ZS_MAX_PAGES_PER_ZSPAGE << PAGE_SHIFT >> OBJ_INDEX_BITS)) |
2e40e163a zsmalloc: decoupl... |
130 |
/* each chunk includes extra space to keep handle */ |
7b60a6852 zsmalloc: record ... |
131 |
#define ZS_MAX_ALLOC_SIZE PAGE_SIZE |
0959c63f1 zsmalloc: collaps... |
132 133 |
/* |
7eb52512a zsmalloc: fixup t... |
134 |
* On systems with 4K page size, this gives 255 size classes! There is a |
0959c63f1 zsmalloc: collaps... |
135 136 137 138 139 140 141 142 143 144 145 |
* trader-off here: * - Large number of size classes is potentially wasteful as free page are * spread across these classes * - Small number of size classes causes large internal fragmentation * - Probably its better to use specific size classes (empirically * determined). NOTE: all those class sizes must be set as multiple of * ZS_ALIGN to make sure link_free itself never has to span 2 pages. * * ZS_MIN_ALLOC_SIZE and ZS_SIZE_CLASS_DELTA must be multiple of ZS_ALIGN * (reason above) */ |
d662b8eba staging: zsmalloc... |
146 |
#define ZS_SIZE_CLASS_DELTA (PAGE_SIZE >> 8) |
0959c63f1 zsmalloc: collaps... |
147 148 149 150 151 152 153 154 155 156 157 158 |
/* * We do not maintain any list for completely empty or full pages */ enum fullness_group { ZS_ALMOST_FULL, ZS_ALMOST_EMPTY, _ZS_NR_FULLNESS_GROUPS, ZS_EMPTY, ZS_FULL }; |
0f050d997 mm/zsmalloc: add ... |
159 160 161 |
enum zs_stat_type { OBJ_ALLOCATED, OBJ_USED, |
248ca1b05 zsmalloc: add ful... |
162 163 |
CLASS_ALMOST_FULL, CLASS_ALMOST_EMPTY, |
0f050d997 mm/zsmalloc: add ... |
164 |
}; |
6fe5186f0 zsmalloc: reduce ... |
165 166 167 168 169 |
#ifdef CONFIG_ZSMALLOC_STAT #define NR_ZS_STAT_TYPE (CLASS_ALMOST_EMPTY + 1) #else #define NR_ZS_STAT_TYPE (OBJ_USED + 1) #endif |
0f050d997 mm/zsmalloc: add ... |
170 171 172 |
struct zs_size_stat { unsigned long objs[NR_ZS_STAT_TYPE]; }; |
572445941 zsmalloc: always ... |
173 174 |
#ifdef CONFIG_ZSMALLOC_STAT static struct dentry *zs_stat_root; |
0f050d997 mm/zsmalloc: add ... |
175 |
#endif |
0959c63f1 zsmalloc: collaps... |
176 |
/* |
40f9fb8cf mm/zsmalloc: supp... |
177 178 179 180 181 |
* number of size_classes */ static int zs_size_classes; /* |
0959c63f1 zsmalloc: collaps... |
182 183 184 185 |
* We assign a page to ZS_ALMOST_EMPTY fullness group when: * n <= N / f, where * n = number of allocated objects * N = total number of objects zspage can store |
6dd9737e3 mm/zsmalloc.c: co... |
186 |
* f = fullness_threshold_frac |
0959c63f1 zsmalloc: collaps... |
187 188 189 190 191 192 193 194 195 196 197 |
* * Similarly, we assign zspage to: * ZS_ALMOST_FULL when n > N / f * ZS_EMPTY when n == 0 * ZS_FULL when n == N * * (see: fix_fullness_group()) */ static const int fullness_threshold_frac = 4; struct size_class { |
572445941 zsmalloc: always ... |
198 199 |
spinlock_t lock; struct page *fullness_list[_ZS_NR_FULLNESS_GROUPS]; |
0959c63f1 zsmalloc: collaps... |
200 201 202 203 204 205 |
/* * Size of objects stored in this class. Must be multiple * of ZS_ALIGN. */ int size; unsigned int index; |
0f050d997 mm/zsmalloc: add ... |
206 |
struct zs_size_stat stats; |
0959c63f1 zsmalloc: collaps... |
207 |
|
7dfa46122 zsmalloc: reorgan... |
208 209 |
/* Number of PAGE_SIZE sized pages to combine to form a 'zspage' */ int pages_per_zspage; |
572445941 zsmalloc: always ... |
210 211 |
/* huge object: pages_per_zspage == 1 && maxobj_per_zspage == 1 */ bool huge; |
0959c63f1 zsmalloc: collaps... |
212 213 214 215 216 217 218 219 220 |
}; /* * Placed within free objects to form a singly linked list. * For every zspage, first_page->freelist gives head of this list. * * This must be power of 2 and less than or equal to ZS_ALIGN */ struct link_free { |
2e40e163a zsmalloc: decoupl... |
221 222 223 224 225 226 227 228 229 230 231 |
union { /* * Position of next free chunk (encodes <PFN, obj_idx>) * It's valid for non-allocated object */ void *next; /* * Handle of allocated object. */ unsigned long handle; }; |
0959c63f1 zsmalloc: collaps... |
232 233 234 |
}; struct zs_pool { |
6f3526d6d mm: zsmalloc: con... |
235 |
const char *name; |
0f050d997 mm/zsmalloc: add ... |
236 |
|
40f9fb8cf mm/zsmalloc: supp... |
237 |
struct size_class **size_class; |
2e40e163a zsmalloc: decoupl... |
238 |
struct kmem_cache *handle_cachep; |
0959c63f1 zsmalloc: collaps... |
239 240 |
gfp_t flags; /* allocation flags used when growing pool */ |
13de8933c zsmalloc: move pa... |
241 |
atomic_long_t pages_allocated; |
0f050d997 mm/zsmalloc: add ... |
242 |
|
7d3f39382 zsmalloc/zram: in... |
243 |
struct zs_pool_stats stats; |
ab9d306d9 zsmalloc: use shr... |
244 245 246 247 248 249 250 251 |
/* Compact classes */ struct shrinker shrinker; /* * To signify that register_shrinker() was successful * and unregister_shrinker() will not Oops. */ bool shrinker_enabled; |
0f050d997 mm/zsmalloc: add ... |
252 253 254 |
#ifdef CONFIG_ZSMALLOC_STAT struct dentry *stat_dentry; #endif |
0959c63f1 zsmalloc: collaps... |
255 |
}; |
61989a80f staging: zsmalloc... |
256 257 258 259 260 261 262 263 264 |
/* * A zspage's class index and fullness group * are encoded in its (first)page->mapping */ #define CLASS_IDX_BITS 28 #define FULLNESS_BITS 4 #define CLASS_IDX_MASK ((1 << CLASS_IDX_BITS) - 1) #define FULLNESS_MASK ((1 << FULLNESS_BITS) - 1) |
f553646a6 staging: zsmalloc... |
265 |
struct mapping_area { |
1b945aeef zsmalloc: add Kco... |
266 |
#ifdef CONFIG_PGTABLE_MAPPING |
f553646a6 staging: zsmalloc... |
267 268 269 270 271 272 273 |
struct vm_struct *vm; /* vm area for mapping object that span pages */ #else char *vm_buf; /* copy buffer for objects that span pages */ #endif char *vm_addr; /* address of kmap_atomic()'ed pages */ enum zs_mapmode vm_mm; /* mapping mode */ }; |
2e40e163a zsmalloc: decoupl... |
274 275 276 277 278 279 280 281 282 |
static int create_handle_cache(struct zs_pool *pool) { pool->handle_cachep = kmem_cache_create("zs_handle", ZS_HANDLE_SIZE, 0, 0, NULL); return pool->handle_cachep ? 0 : 1; } static void destroy_handle_cache(struct zs_pool *pool) { |
cd10add00 zsmalloc: remove ... |
283 |
kmem_cache_destroy(pool->handle_cachep); |
2e40e163a zsmalloc: decoupl... |
284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 |
} static unsigned long alloc_handle(struct zs_pool *pool) { return (unsigned long)kmem_cache_alloc(pool->handle_cachep, pool->flags & ~__GFP_HIGHMEM); } static void free_handle(struct zs_pool *pool, unsigned long handle) { kmem_cache_free(pool->handle_cachep, (void *)handle); } static void record_obj(unsigned long handle, unsigned long obj) { |
c102f07ca zsmalloc: fix mig... |
299 300 301 302 303 304 |
/* * lsb of @obj represents handle lock while other bits * represent object value the handle is pointing so * updating shouldn't do store tearing. */ WRITE_ONCE(*(unsigned long *)handle, obj); |
2e40e163a zsmalloc: decoupl... |
305 |
} |
c795779df mm/zpool: zbud/zs... |
306 307 308 |
/* zpool driver */ #ifdef CONFIG_ZPOOL |
6f3526d6d mm: zsmalloc: con... |
309 |
static void *zs_zpool_create(const char *name, gfp_t gfp, |
786727799 mm: zpool: consti... |
310 |
const struct zpool_ops *zpool_ops, |
479305fd7 zpool: remove zpo... |
311 |
struct zpool *zpool) |
c795779df mm/zpool: zbud/zs... |
312 |
{ |
3eba0c6a5 mm/zpool: add nam... |
313 |
return zs_create_pool(name, gfp); |
c795779df mm/zpool: zbud/zs... |
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 |
} static void zs_zpool_destroy(void *pool) { zs_destroy_pool(pool); } static int zs_zpool_malloc(void *pool, size_t size, gfp_t gfp, unsigned long *handle) { *handle = zs_malloc(pool, size); return *handle ? 0 : -1; } static void zs_zpool_free(void *pool, unsigned long handle) { zs_free(pool, handle); } static int zs_zpool_shrink(void *pool, unsigned int pages, unsigned int *reclaimed) { return -EINVAL; } static void *zs_zpool_map(void *pool, unsigned long handle, enum zpool_mapmode mm) { enum zs_mapmode zs_mm; switch (mm) { case ZPOOL_MM_RO: zs_mm = ZS_MM_RO; break; case ZPOOL_MM_WO: zs_mm = ZS_MM_WO; break; case ZPOOL_MM_RW: /* fallthru */ default: zs_mm = ZS_MM_RW; break; } return zs_map_object(pool, handle, zs_mm); } static void zs_zpool_unmap(void *pool, unsigned long handle) { zs_unmap_object(pool, handle); } static u64 zs_zpool_total_size(void *pool) { |
722cdc172 zsmalloc: change ... |
365 |
return zs_get_total_pages(pool) << PAGE_SHIFT; |
c795779df mm/zpool: zbud/zs... |
366 367 368 369 370 371 372 373 374 375 376 377 378 379 |
} static struct zpool_driver zs_zpool_driver = { .type = "zsmalloc", .owner = THIS_MODULE, .create = zs_zpool_create, .destroy = zs_zpool_destroy, .malloc = zs_zpool_malloc, .free = zs_zpool_free, .shrink = zs_zpool_shrink, .map = zs_zpool_map, .unmap = zs_zpool_unmap, .total_size = zs_zpool_total_size, }; |
137f8cff5 mm/zpool: use pre... |
380 |
MODULE_ALIAS("zpool-zsmalloc"); |
c795779df mm/zpool: zbud/zs... |
381 |
#endif /* CONFIG_ZPOOL */ |
248ca1b05 zsmalloc: add ful... |
382 383 384 385 |
static unsigned int get_maxobj_per_zspage(int size, int pages_per_zspage) { return pages_per_zspage * PAGE_SIZE / size; } |
61989a80f staging: zsmalloc... |
386 387 388 389 390 |
/* per-cpu VM mapping areas for zspage accesses that cross page boundaries */ static DEFINE_PER_CPU(struct mapping_area, zs_map_area); static int is_first_page(struct page *page) { |
a27545bf0 zsmalloc: use Pag... |
391 |
return PagePrivate(page); |
61989a80f staging: zsmalloc... |
392 393 394 395 |
} static int is_last_page(struct page *page) { |
a27545bf0 zsmalloc: use Pag... |
396 |
return PagePrivate2(page); |
61989a80f staging: zsmalloc... |
397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 |
} static void get_zspage_mapping(struct page *page, unsigned int *class_idx, enum fullness_group *fullness) { unsigned long m; BUG_ON(!is_first_page(page)); m = (unsigned long)page->mapping; *fullness = m & FULLNESS_MASK; *class_idx = (m >> FULLNESS_BITS) & CLASS_IDX_MASK; } static void set_zspage_mapping(struct page *page, unsigned int class_idx, enum fullness_group fullness) { unsigned long m; BUG_ON(!is_first_page(page)); m = ((class_idx & CLASS_IDX_MASK) << FULLNESS_BITS) | (fullness & FULLNESS_MASK); page->mapping = (struct address_space *)m; } |
c3e3e88ad zsmalloc: add mor... |
420 421 422 423 424 425 426 |
/* * zsmalloc divides the pool into various size classes where each * class maintains a list of zspages where each zspage is divided * into equal sized chunks. Each allocation falls into one of these * classes depending on its size. This function returns index of the * size class which has chunk size big enough to hold the give size. */ |
61989a80f staging: zsmalloc... |
427 428 429 430 431 432 433 |
static int get_size_class_index(int size) { int idx = 0; if (likely(size > ZS_MIN_ALLOC_SIZE)) idx = DIV_ROUND_UP(size - ZS_MIN_ALLOC_SIZE, ZS_SIZE_CLASS_DELTA); |
7b60a6852 zsmalloc: record ... |
434 |
return min(zs_size_classes - 1, idx); |
61989a80f staging: zsmalloc... |
435 |
} |
248ca1b05 zsmalloc: add ful... |
436 437 438 |
static inline void zs_stat_inc(struct size_class *class, enum zs_stat_type type, unsigned long cnt) { |
6fe5186f0 zsmalloc: reduce ... |
439 440 |
if (type < NR_ZS_STAT_TYPE) class->stats.objs[type] += cnt; |
248ca1b05 zsmalloc: add ful... |
441 442 443 444 445 |
} static inline void zs_stat_dec(struct size_class *class, enum zs_stat_type type, unsigned long cnt) { |
6fe5186f0 zsmalloc: reduce ... |
446 447 |
if (type < NR_ZS_STAT_TYPE) class->stats.objs[type] -= cnt; |
248ca1b05 zsmalloc: add ful... |
448 449 450 451 452 |
} static inline unsigned long zs_stat_get(struct size_class *class, enum zs_stat_type type) { |
6fe5186f0 zsmalloc: reduce ... |
453 454 455 |
if (type < NR_ZS_STAT_TYPE) return class->stats.objs[type]; return 0; |
248ca1b05 zsmalloc: add ful... |
456 |
} |
572445941 zsmalloc: always ... |
457 |
#ifdef CONFIG_ZSMALLOC_STAT |
248ca1b05 zsmalloc: add ful... |
458 459 460 461 462 463 464 465 466 467 468 469 470 471 472 473 |
static int __init zs_stat_init(void) { if (!debugfs_initialized()) return -ENODEV; zs_stat_root = debugfs_create_dir("zsmalloc", NULL); if (!zs_stat_root) return -ENOMEM; return 0; } static void __exit zs_stat_exit(void) { debugfs_remove_recursive(zs_stat_root); } |
1120ed548 mm/zsmalloc: add ... |
474 |
static unsigned long zs_can_compact(struct size_class *class); |
248ca1b05 zsmalloc: add ful... |
475 476 477 478 479 480 481 |
static int zs_stats_size_show(struct seq_file *s, void *v) { int i; struct zs_pool *pool = s->private; struct size_class *class; int objs_per_zspage; unsigned long class_almost_full, class_almost_empty; |
1120ed548 mm/zsmalloc: add ... |
482 |
unsigned long obj_allocated, obj_used, pages_used, freeable; |
248ca1b05 zsmalloc: add ful... |
483 484 |
unsigned long total_class_almost_full = 0, total_class_almost_empty = 0; unsigned long total_objs = 0, total_used_objs = 0, total_pages = 0; |
1120ed548 mm/zsmalloc: add ... |
485 |
unsigned long total_freeable = 0; |
248ca1b05 zsmalloc: add ful... |
486 |
|
1120ed548 mm/zsmalloc: add ... |
487 488 |
seq_printf(s, " %5s %5s %11s %12s %13s %10s %10s %16s %8s ", |
248ca1b05 zsmalloc: add ful... |
489 490 |
"class", "size", "almost_full", "almost_empty", "obj_allocated", "obj_used", "pages_used", |
1120ed548 mm/zsmalloc: add ... |
491 |
"pages_per_zspage", "freeable"); |
248ca1b05 zsmalloc: add ful... |
492 493 494 495 496 497 498 499 500 501 502 503 |
for (i = 0; i < zs_size_classes; i++) { class = pool->size_class[i]; if (class->index != i) continue; spin_lock(&class->lock); class_almost_full = zs_stat_get(class, CLASS_ALMOST_FULL); class_almost_empty = zs_stat_get(class, CLASS_ALMOST_EMPTY); obj_allocated = zs_stat_get(class, OBJ_ALLOCATED); obj_used = zs_stat_get(class, OBJ_USED); |
1120ed548 mm/zsmalloc: add ... |
504 |
freeable = zs_can_compact(class); |
248ca1b05 zsmalloc: add ful... |
505 506 507 508 509 510 |
spin_unlock(&class->lock); objs_per_zspage = get_maxobj_per_zspage(class->size, class->pages_per_zspage); pages_used = obj_allocated / objs_per_zspage * class->pages_per_zspage; |
1120ed548 mm/zsmalloc: add ... |
511 512 513 |
seq_printf(s, " %5u %5u %11lu %12lu %13lu" " %10lu %10lu %16d %8lu ", |
248ca1b05 zsmalloc: add ful... |
514 515 |
i, class->size, class_almost_full, class_almost_empty, obj_allocated, obj_used, pages_used, |
1120ed548 mm/zsmalloc: add ... |
516 |
class->pages_per_zspage, freeable); |
248ca1b05 zsmalloc: add ful... |
517 518 519 520 521 522 |
total_class_almost_full += class_almost_full; total_class_almost_empty += class_almost_empty; total_objs += obj_allocated; total_used_objs += obj_used; total_pages += pages_used; |
1120ed548 mm/zsmalloc: add ... |
523 |
total_freeable += freeable; |
248ca1b05 zsmalloc: add ful... |
524 525 526 527 |
} seq_puts(s, " "); |
1120ed548 mm/zsmalloc: add ... |
528 529 |
seq_printf(s, " %5s %5s %11lu %12lu %13lu %10lu %10lu %16s %8lu ", |
248ca1b05 zsmalloc: add ful... |
530 531 |
"Total", "", total_class_almost_full, total_class_almost_empty, total_objs, |
1120ed548 mm/zsmalloc: add ... |
532 |
total_used_objs, total_pages, "", total_freeable); |
248ca1b05 zsmalloc: add ful... |
533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 |
return 0; } static int zs_stats_size_open(struct inode *inode, struct file *file) { return single_open(file, zs_stats_size_show, inode->i_private); } static const struct file_operations zs_stat_size_ops = { .open = zs_stats_size_open, .read = seq_read, .llseek = seq_lseek, .release = single_release, }; |
6f3526d6d mm: zsmalloc: con... |
548 |
static int zs_pool_stat_create(const char *name, struct zs_pool *pool) |
248ca1b05 zsmalloc: add ful... |
549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 |
{ struct dentry *entry; if (!zs_stat_root) return -ENODEV; entry = debugfs_create_dir(name, zs_stat_root); if (!entry) { pr_warn("debugfs dir <%s> creation failed ", name); return -ENOMEM; } pool->stat_dentry = entry; entry = debugfs_create_file("classes", S_IFREG | S_IRUGO, pool->stat_dentry, pool, &zs_stat_size_ops); if (!entry) { pr_warn("%s: debugfs file entry <%s> creation failed ", name, "classes"); return -ENOMEM; } return 0; } static void zs_pool_stat_destroy(struct zs_pool *pool) { debugfs_remove_recursive(pool->stat_dentry); } #else /* CONFIG_ZSMALLOC_STAT */ |
248ca1b05 zsmalloc: add ful... |
581 582 583 584 585 586 587 588 |
static int __init zs_stat_init(void) { return 0; } static void __exit zs_stat_exit(void) { } |
6f3526d6d mm: zsmalloc: con... |
589 |
static inline int zs_pool_stat_create(const char *name, struct zs_pool *pool) |
248ca1b05 zsmalloc: add ful... |
590 591 592 593 594 595 596 |
{ return 0; } static inline void zs_pool_stat_destroy(struct zs_pool *pool) { } |
248ca1b05 zsmalloc: add ful... |
597 |
#endif |
c3e3e88ad zsmalloc: add mor... |
598 599 600 601 602 603 604 |
/* * For each size class, zspages are divided into different groups * depending on how "full" they are. This was done so that we could * easily find empty or nearly empty zspages when we try to shrink * the pool (not yet implemented). This function returns fullness * status of the given page. */ |
61989a80f staging: zsmalloc... |
605 606 607 608 609 610 611 612 613 614 615 616 617 |
static enum fullness_group get_fullness_group(struct page *page) { int inuse, max_objects; enum fullness_group fg; BUG_ON(!is_first_page(page)); inuse = page->inuse; max_objects = page->objects; if (inuse == 0) fg = ZS_EMPTY; else if (inuse == max_objects) fg = ZS_FULL; |
d3d07c92f zsmalloc: adjust ... |
618 |
else if (inuse <= 3 * max_objects / fullness_threshold_frac) |
61989a80f staging: zsmalloc... |
619 620 621 622 623 624 |
fg = ZS_ALMOST_EMPTY; else fg = ZS_ALMOST_FULL; return fg; } |
c3e3e88ad zsmalloc: add mor... |
625 626 627 628 629 630 |
/* * Each size class maintains various freelists and zspages are assigned * to one of these freelists based on the number of live objects they * have. This functions inserts the given zspage into the freelist * identified by <class, fullness_group>. */ |
61989a80f staging: zsmalloc... |
631 632 633 634 635 636 637 638 639 |
static void insert_zspage(struct page *page, struct size_class *class, enum fullness_group fullness) { struct page **head; BUG_ON(!is_first_page(page)); if (fullness >= _ZS_NR_FULLNESS_GROUPS) return; |
248ca1b05 zsmalloc: add ful... |
640 641 |
zs_stat_inc(class, fullness == ZS_ALMOST_EMPTY ? CLASS_ALMOST_EMPTY : CLASS_ALMOST_FULL, 1); |
58f171174 zsmalloc: partial... |
642 643 644 645 646 647 648 649 650 651 652 653 654 655 |
head = &class->fullness_list[fullness]; if (!*head) { *head = page; return; } /* * We want to see more ZS_FULL pages and less almost * empty/full. Put pages with higher ->inuse first. */ list_add_tail(&page->lru, &(*head)->lru); if (page->inuse >= (*head)->inuse) *head = page; |
61989a80f staging: zsmalloc... |
656 |
} |
c3e3e88ad zsmalloc: add mor... |
657 658 659 660 |
/* * This function removes the given zspage from the freelist identified * by <class, fullness_group>. */ |
61989a80f staging: zsmalloc... |
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 |
static void remove_zspage(struct page *page, struct size_class *class, enum fullness_group fullness) { struct page **head; BUG_ON(!is_first_page(page)); if (fullness >= _ZS_NR_FULLNESS_GROUPS) return; head = &class->fullness_list[fullness]; BUG_ON(!*head); if (list_empty(&(*head)->lru)) *head = NULL; else if (*head == page) *head = (struct page *)list_entry((*head)->lru.next, struct page, lru); list_del_init(&page->lru); |
248ca1b05 zsmalloc: add ful... |
680 681 |
zs_stat_dec(class, fullness == ZS_ALMOST_EMPTY ? CLASS_ALMOST_EMPTY : CLASS_ALMOST_FULL, 1); |
61989a80f staging: zsmalloc... |
682 |
} |
c3e3e88ad zsmalloc: add mor... |
683 684 685 686 687 688 689 690 691 |
/* * Each size class maintains zspages in different fullness groups depending * on the number of live objects they contain. When allocating or freeing * objects, the fullness status of the page can change, say, from ALMOST_FULL * to ALMOST_EMPTY when freeing an object. This function checks if such * a status change has occurred for the given page and accordingly moves the * page from the freelist of the old fullness group to that of the new * fullness group. */ |
c78062612 zsmalloc: factor ... |
692 |
static enum fullness_group fix_fullness_group(struct size_class *class, |
61989a80f staging: zsmalloc... |
693 694 695 |
struct page *page) { int class_idx; |
61989a80f staging: zsmalloc... |
696 697 698 699 700 701 702 703 |
enum fullness_group currfg, newfg; BUG_ON(!is_first_page(page)); get_zspage_mapping(page, &class_idx, &currfg); newfg = get_fullness_group(page); if (newfg == currfg) goto out; |
61989a80f staging: zsmalloc... |
704 705 706 707 708 709 710 711 712 713 714 715 716 |
remove_zspage(page, class, currfg); insert_zspage(page, class, newfg); set_zspage_mapping(page, class_idx, newfg); out: return newfg; } /* * We have to decide on how many pages to link together * to form a zspage for each size class. This is important * to reduce wastage due to unusable space left at end of * each zspage which is given as: |
888fa374e mm/zsmalloc.c: fi... |
717 718 |
* wastage = Zp % class_size * usage = Zp - wastage |
61989a80f staging: zsmalloc... |
719 720 721 722 723 724 |
* where Zp = zspage size = k * PAGE_SIZE where k = 1, 2, ... * * For example, for size class of 3/8 * PAGE_SIZE, we should * link together 3 PAGE_SIZE sized pages to form a zspage * since then we can perfectly fit in 8 such objects. */ |
2e3b61547 staging: zsmalloc... |
725 |
static int get_pages_per_zspage(int class_size) |
61989a80f staging: zsmalloc... |
726 727 728 729 |
{ int i, max_usedpc = 0; /* zspage order which gives maximum used size per KB */ int max_usedpc_order = 1; |
84d4faaba staging: zsmalloc... |
730 |
for (i = 1; i <= ZS_MAX_PAGES_PER_ZSPAGE; i++) { |
61989a80f staging: zsmalloc... |
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 |
int zspage_size; int waste, usedpc; zspage_size = i * PAGE_SIZE; waste = zspage_size % class_size; usedpc = (zspage_size - waste) * 100 / zspage_size; if (usedpc > max_usedpc) { max_usedpc = usedpc; max_usedpc_order = i; } } return max_usedpc_order; } /* * A single 'zspage' is composed of many system pages which are * linked together using fields in struct page. This function finds * the first/head page, given any component page of a zspage. */ static struct page *get_first_page(struct page *page) { if (is_first_page(page)) return page; else |
32e7ba1ea zsmalloc: use pag... |
757 |
return (struct page *)page_private(page); |
61989a80f staging: zsmalloc... |
758 759 760 761 762 763 764 765 766 |
} static struct page *get_next_page(struct page *page) { struct page *next; if (is_last_page(page)) next = NULL; else if (is_first_page(page)) |
e842b976a staging: zsmalloc... |
767 |
next = (struct page *)page_private(page); |
61989a80f staging: zsmalloc... |
768 769 770 771 772 |
else next = list_entry(page->lru.next, struct page, lru); return next; } |
67296874e staging: zsmalloc... |
773 774 |
/* * Encode <page, obj_idx> as a single handle value. |
312fcae22 zsmalloc: support... |
775 |
* We use the least bit of handle for tagging. |
67296874e staging: zsmalloc... |
776 |
*/ |
312fcae22 zsmalloc: support... |
777 |
static void *location_to_obj(struct page *page, unsigned long obj_idx) |
61989a80f staging: zsmalloc... |
778 |
{ |
312fcae22 zsmalloc: support... |
779 |
unsigned long obj; |
61989a80f staging: zsmalloc... |
780 781 782 783 784 |
if (!page) { BUG_ON(obj_idx); return NULL; } |
312fcae22 zsmalloc: support... |
785 786 787 |
obj = page_to_pfn(page) << OBJ_INDEX_BITS; obj |= ((obj_idx) & OBJ_INDEX_MASK); obj <<= OBJ_TAG_BITS; |
61989a80f staging: zsmalloc... |
788 |
|
312fcae22 zsmalloc: support... |
789 |
return (void *)obj; |
61989a80f staging: zsmalloc... |
790 |
} |
67296874e staging: zsmalloc... |
791 792 793 |
/* * Decode <page, obj_idx> pair from the given object handle. We adjust the * decoded obj_idx back to its original value since it was adjusted in |
312fcae22 zsmalloc: support... |
794 |
* location_to_obj(). |
67296874e staging: zsmalloc... |
795 |
*/ |
312fcae22 zsmalloc: support... |
796 |
static void obj_to_location(unsigned long obj, struct page **page, |
61989a80f staging: zsmalloc... |
797 798 |
unsigned long *obj_idx) { |
312fcae22 zsmalloc: support... |
799 800 801 |
obj >>= OBJ_TAG_BITS; *page = pfn_to_page(obj >> OBJ_INDEX_BITS); *obj_idx = (obj & OBJ_INDEX_MASK); |
61989a80f staging: zsmalloc... |
802 |
} |
2e40e163a zsmalloc: decoupl... |
803 804 805 806 |
static unsigned long handle_to_obj(unsigned long handle) { return *(unsigned long *)handle; } |
7b60a6852 zsmalloc: record ... |
807 808 |
static unsigned long obj_to_head(struct size_class *class, struct page *page, void *obj) |
312fcae22 zsmalloc: support... |
809 |
{ |
7b60a6852 zsmalloc: record ... |
810 811 |
if (class->huge) { VM_BUG_ON(!is_first_page(page)); |
12a7bfad5 zsmalloc: fix obj... |
812 |
return page_private(page); |
7b60a6852 zsmalloc: record ... |
813 814 |
} else return *(unsigned long *)obj; |
312fcae22 zsmalloc: support... |
815 |
} |
61989a80f staging: zsmalloc... |
816 817 818 819 820 821 822 823 824 825 |
static unsigned long obj_idx_to_offset(struct page *page, unsigned long obj_idx, int class_size) { unsigned long off = 0; if (!is_first_page(page)) off = page->index; return off + obj_idx * class_size; } |
312fcae22 zsmalloc: support... |
826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 |
static inline int trypin_tag(unsigned long handle) { unsigned long *ptr = (unsigned long *)handle; return !test_and_set_bit_lock(HANDLE_PIN_BIT, ptr); } static void pin_tag(unsigned long handle) { while (!trypin_tag(handle)); } static void unpin_tag(unsigned long handle) { unsigned long *ptr = (unsigned long *)handle; clear_bit_unlock(HANDLE_PIN_BIT, ptr); } |
f4477e90b staging: zsmalloc... |
844 845 846 847 848 849 850 |
static void reset_page(struct page *page) { clear_bit(PG_private, &page->flags); clear_bit(PG_private_2, &page->flags); set_page_private(page, 0); page->mapping = NULL; page->freelist = NULL; |
22b751c3d mm: rename page s... |
851 |
page_mapcount_reset(page); |
f4477e90b staging: zsmalloc... |
852 |
} |
61989a80f staging: zsmalloc... |
853 854 |
static void free_zspage(struct page *first_page) { |
f4477e90b staging: zsmalloc... |
855 |
struct page *nextp, *tmp, *head_extra; |
61989a80f staging: zsmalloc... |
856 857 858 |
BUG_ON(!is_first_page(first_page)); BUG_ON(first_page->inuse); |
f4477e90b staging: zsmalloc... |
859 |
head_extra = (struct page *)page_private(first_page); |
61989a80f staging: zsmalloc... |
860 |
|
f4477e90b staging: zsmalloc... |
861 |
reset_page(first_page); |
61989a80f staging: zsmalloc... |
862 863 864 |
__free_page(first_page); /* zspage with only 1 system page */ |
f4477e90b staging: zsmalloc... |
865 |
if (!head_extra) |
61989a80f staging: zsmalloc... |
866 |
return; |
f4477e90b staging: zsmalloc... |
867 |
list_for_each_entry_safe(nextp, tmp, &head_extra->lru, lru) { |
61989a80f staging: zsmalloc... |
868 |
list_del(&nextp->lru); |
f4477e90b staging: zsmalloc... |
869 |
reset_page(nextp); |
61989a80f staging: zsmalloc... |
870 871 |
__free_page(nextp); } |
f4477e90b staging: zsmalloc... |
872 873 |
reset_page(head_extra); __free_page(head_extra); |
61989a80f staging: zsmalloc... |
874 875 876 877 878 879 880 881 882 883 884 885 |
} /* Initialize a newly allocated zspage */ static void init_zspage(struct page *first_page, struct size_class *class) { unsigned long off = 0; struct page *page = first_page; BUG_ON(!is_first_page(first_page)); while (page) { struct page *next_page; struct link_free *link; |
5538c5623 zsmalloc: simplif... |
886 |
unsigned int i = 1; |
af4ee5e97 zsmalloc: correct... |
887 |
void *vaddr; |
61989a80f staging: zsmalloc... |
888 889 890 891 892 893 894 895 896 |
/* * page->index stores offset of first object starting * in the page. For the first page, this is always 0, * so we use first_page->index (aka ->freelist) to store * head of corresponding zspage's freelist. */ if (page != first_page) page->index = off; |
af4ee5e97 zsmalloc: correct... |
897 898 |
vaddr = kmap_atomic(page); link = (struct link_free *)vaddr + off / sizeof(*link); |
5538c5623 zsmalloc: simplif... |
899 900 |
while ((off += class->size) < PAGE_SIZE) { |
312fcae22 zsmalloc: support... |
901 |
link->next = location_to_obj(page, i++); |
5538c5623 zsmalloc: simplif... |
902 |
link += class->size / sizeof(*link); |
61989a80f staging: zsmalloc... |
903 904 905 906 907 908 909 910 |
} /* * We now come to the last (full or partial) object on this * page, which must point to the first object on the next * page (if present) */ next_page = get_next_page(page); |
312fcae22 zsmalloc: support... |
911 |
link->next = location_to_obj(next_page, 0); |
af4ee5e97 zsmalloc: correct... |
912 |
kunmap_atomic(vaddr); |
61989a80f staging: zsmalloc... |
913 |
page = next_page; |
5538c5623 zsmalloc: simplif... |
914 |
off %= PAGE_SIZE; |
61989a80f staging: zsmalloc... |
915 916 917 918 919 920 921 922 923 |
} } /* * Allocate a zspage for the given size class */ static struct page *alloc_zspage(struct size_class *class, gfp_t flags) { int i, error; |
b4b700c5a staging: zsmalloc... |
924 |
struct page *first_page = NULL, *uninitialized_var(prev_page); |
61989a80f staging: zsmalloc... |
925 926 927 928 929 |
/* * Allocate individual pages and link them together as: * 1. first page->private = first sub-page * 2. all sub-pages are linked together using page->lru |
32e7ba1ea zsmalloc: use pag... |
930 |
* 3. each sub-page is linked to the first page using page->private |
61989a80f staging: zsmalloc... |
931 932 933 934 935 936 937 |
* * For each size class, First/Head pages are linked together using * page->lru. Also, we set PG_private to identify the first page * (i.e. no other sub-page has this flag set) and PG_private_2 to * identify the last page. */ error = -ENOMEM; |
2e3b61547 staging: zsmalloc... |
938 |
for (i = 0; i < class->pages_per_zspage; i++) { |
b4b700c5a staging: zsmalloc... |
939 |
struct page *page; |
61989a80f staging: zsmalloc... |
940 941 942 943 944 945 946 |
page = alloc_page(flags); if (!page) goto cleanup; INIT_LIST_HEAD(&page->lru); if (i == 0) { /* first page */ |
a27545bf0 zsmalloc: use Pag... |
947 |
SetPagePrivate(page); |
61989a80f staging: zsmalloc... |
948 949 950 951 952 |
set_page_private(page, 0); first_page = page; first_page->inuse = 0; } if (i == 1) |
e842b976a staging: zsmalloc... |
953 |
set_page_private(first_page, (unsigned long)page); |
61989a80f staging: zsmalloc... |
954 |
if (i >= 1) |
32e7ba1ea zsmalloc: use pag... |
955 |
set_page_private(page, (unsigned long)first_page); |
61989a80f staging: zsmalloc... |
956 957 |
if (i >= 2) list_add(&page->lru, &prev_page->lru); |
2e3b61547 staging: zsmalloc... |
958 |
if (i == class->pages_per_zspage - 1) /* last page */ |
a27545bf0 zsmalloc: use Pag... |
959 |
SetPagePrivate2(page); |
61989a80f staging: zsmalloc... |
960 961 962 963 |
prev_page = page; } init_zspage(first_page, class); |
312fcae22 zsmalloc: support... |
964 |
first_page->freelist = location_to_obj(first_page, 0); |
61989a80f staging: zsmalloc... |
965 |
/* Maximum number of objects we can store in this zspage */ |
2e3b61547 staging: zsmalloc... |
966 |
first_page->objects = class->pages_per_zspage * PAGE_SIZE / class->size; |
61989a80f staging: zsmalloc... |
967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 |
error = 0; /* Success */ cleanup: if (unlikely(error) && first_page) { free_zspage(first_page); first_page = NULL; } return first_page; } static struct page *find_get_zspage(struct size_class *class) { int i; struct page *page; for (i = 0; i < _ZS_NR_FULLNESS_GROUPS; i++) { page = class->fullness_list[i]; if (page) break; } return page; } |
1b945aeef zsmalloc: add Kco... |
992 |
#ifdef CONFIG_PGTABLE_MAPPING |
f553646a6 staging: zsmalloc... |
993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 |
static inline int __zs_cpu_up(struct mapping_area *area) { /* * Make sure we don't leak memory if a cpu UP notification * and zs_init() race and both call zs_cpu_up() on the same cpu */ if (area->vm) return 0; area->vm = alloc_vm_area(PAGE_SIZE * 2, NULL); if (!area->vm) return -ENOMEM; return 0; } static inline void __zs_cpu_down(struct mapping_area *area) { if (area->vm) free_vm_area(area->vm); area->vm = NULL; } static inline void *__zs_map_object(struct mapping_area *area, struct page *pages[2], int off, int size) { |
f6f8ed473 mm/vmalloc.c: cle... |
1017 |
BUG_ON(map_vm_area(area->vm, PAGE_KERNEL, pages)); |
f553646a6 staging: zsmalloc... |
1018 1019 1020 1021 1022 1023 1024 1025 |
area->vm_addr = area->vm->addr; return area->vm_addr + off; } static inline void __zs_unmap_object(struct mapping_area *area, struct page *pages[2], int off, int size) { unsigned long addr = (unsigned long)area->vm_addr; |
f553646a6 staging: zsmalloc... |
1026 |
|
d95abbbb2 staging: zsmalloc... |
1027 |
unmap_kernel_range(addr, PAGE_SIZE * 2); |
f553646a6 staging: zsmalloc... |
1028 |
} |
1b945aeef zsmalloc: add Kco... |
1029 |
#else /* CONFIG_PGTABLE_MAPPING */ |
f553646a6 staging: zsmalloc... |
1030 1031 1032 1033 1034 1035 1036 1037 1038 |
static inline int __zs_cpu_up(struct mapping_area *area) { /* * Make sure we don't leak memory if a cpu UP notification * and zs_init() race and both call zs_cpu_up() on the same cpu */ if (area->vm_buf) return 0; |
40f9fb8cf mm/zsmalloc: supp... |
1039 |
area->vm_buf = kmalloc(ZS_MAX_ALLOC_SIZE, GFP_KERNEL); |
f553646a6 staging: zsmalloc... |
1040 1041 1042 1043 1044 1045 1046 |
if (!area->vm_buf) return -ENOMEM; return 0; } static inline void __zs_cpu_down(struct mapping_area *area) { |
40f9fb8cf mm/zsmalloc: supp... |
1047 |
kfree(area->vm_buf); |
f553646a6 staging: zsmalloc... |
1048 1049 1050 1051 1052 |
area->vm_buf = NULL; } static void *__zs_map_object(struct mapping_area *area, struct page *pages[2], int off, int size) |
5f601902c staging: zsmalloc... |
1053 |
{ |
5f601902c staging: zsmalloc... |
1054 1055 |
int sizes[2]; void *addr; |
f553646a6 staging: zsmalloc... |
1056 |
char *buf = area->vm_buf; |
5f601902c staging: zsmalloc... |
1057 |
|
f553646a6 staging: zsmalloc... |
1058 1059 1060 1061 1062 1063 |
/* disable page faults to match kmap_atomic() return conditions */ pagefault_disable(); /* no read fastpath */ if (area->vm_mm == ZS_MM_WO) goto out; |
5f601902c staging: zsmalloc... |
1064 1065 1066 |
sizes[0] = PAGE_SIZE - off; sizes[1] = size - sizes[0]; |
5f601902c staging: zsmalloc... |
1067 1068 1069 1070 1071 1072 1073 |
/* copy object to per-cpu buffer */ addr = kmap_atomic(pages[0]); memcpy(buf, addr + off, sizes[0]); kunmap_atomic(addr); addr = kmap_atomic(pages[1]); memcpy(buf + sizes[0], addr, sizes[1]); kunmap_atomic(addr); |
f553646a6 staging: zsmalloc... |
1074 1075 |
out: return area->vm_buf; |
5f601902c staging: zsmalloc... |
1076 |
} |
f553646a6 staging: zsmalloc... |
1077 1078 |
static void __zs_unmap_object(struct mapping_area *area, struct page *pages[2], int off, int size) |
5f601902c staging: zsmalloc... |
1079 |
{ |
5f601902c staging: zsmalloc... |
1080 1081 |
int sizes[2]; void *addr; |
2e40e163a zsmalloc: decoupl... |
1082 |
char *buf; |
5f601902c staging: zsmalloc... |
1083 |
|
f553646a6 staging: zsmalloc... |
1084 1085 1086 |
/* no write fastpath */ if (area->vm_mm == ZS_MM_RO) goto out; |
5f601902c staging: zsmalloc... |
1087 |
|
7b60a6852 zsmalloc: record ... |
1088 |
buf = area->vm_buf; |
a82cbf071 zsmalloc: drop un... |
1089 1090 1091 |
buf = buf + ZS_HANDLE_SIZE; size -= ZS_HANDLE_SIZE; off += ZS_HANDLE_SIZE; |
2e40e163a zsmalloc: decoupl... |
1092 |
|
5f601902c staging: zsmalloc... |
1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 |
sizes[0] = PAGE_SIZE - off; sizes[1] = size - sizes[0]; /* copy per-cpu buffer to object */ addr = kmap_atomic(pages[0]); memcpy(addr + off, buf, sizes[0]); kunmap_atomic(addr); addr = kmap_atomic(pages[1]); memcpy(addr, buf + sizes[0], sizes[1]); kunmap_atomic(addr); |
f553646a6 staging: zsmalloc... |
1103 1104 1105 1106 |
out: /* enable page faults to match kunmap_atomic() return conditions */ pagefault_enable(); |
5f601902c staging: zsmalloc... |
1107 |
} |
61989a80f staging: zsmalloc... |
1108 |
|
1b945aeef zsmalloc: add Kco... |
1109 |
#endif /* CONFIG_PGTABLE_MAPPING */ |
f553646a6 staging: zsmalloc... |
1110 |
|
61989a80f staging: zsmalloc... |
1111 1112 1113 |
static int zs_cpu_notifier(struct notifier_block *nb, unsigned long action, void *pcpu) { |
f553646a6 staging: zsmalloc... |
1114 |
int ret, cpu = (long)pcpu; |
61989a80f staging: zsmalloc... |
1115 1116 1117 1118 1119 |
struct mapping_area *area; switch (action) { case CPU_UP_PREPARE: area = &per_cpu(zs_map_area, cpu); |
f553646a6 staging: zsmalloc... |
1120 1121 1122 |
ret = __zs_cpu_up(area); if (ret) return notifier_from_errno(ret); |
61989a80f staging: zsmalloc... |
1123 1124 1125 1126 |
break; case CPU_DEAD: case CPU_UP_CANCELED: area = &per_cpu(zs_map_area, cpu); |
f553646a6 staging: zsmalloc... |
1127 |
__zs_cpu_down(area); |
61989a80f staging: zsmalloc... |
1128 1129 1130 1131 1132 1133 1134 1135 1136 |
break; } return NOTIFY_OK; } static struct notifier_block zs_cpu_nb = { .notifier_call = zs_cpu_notifier }; |
b1b00a5b8 zsmalloc: fix zs_... |
1137 |
static int zs_register_cpu_notifier(void) |
61989a80f staging: zsmalloc... |
1138 |
{ |
b1b00a5b8 zsmalloc: fix zs_... |
1139 |
int cpu, uninitialized_var(ret); |
61989a80f staging: zsmalloc... |
1140 |
|
f0e71fcd0 zsmalloc: Fix CPU... |
1141 1142 1143 |
cpu_notifier_register_begin(); __register_cpu_notifier(&zs_cpu_nb); |
61989a80f staging: zsmalloc... |
1144 1145 |
for_each_online_cpu(cpu) { ret = zs_cpu_notifier(NULL, CPU_UP_PREPARE, (void *)(long)cpu); |
b1b00a5b8 zsmalloc: fix zs_... |
1146 1147 |
if (notifier_to_errno(ret)) break; |
61989a80f staging: zsmalloc... |
1148 |
} |
f0e71fcd0 zsmalloc: Fix CPU... |
1149 1150 |
cpu_notifier_register_done(); |
b1b00a5b8 zsmalloc: fix zs_... |
1151 1152 |
return notifier_to_errno(ret); } |
f0e71fcd0 zsmalloc: Fix CPU... |
1153 |
|
66cdef663 mm/zsmalloc: adju... |
1154 |
static void zs_unregister_cpu_notifier(void) |
40f9fb8cf mm/zsmalloc: supp... |
1155 |
{ |
66cdef663 mm/zsmalloc: adju... |
1156 |
int cpu; |
40f9fb8cf mm/zsmalloc: supp... |
1157 |
|
66cdef663 mm/zsmalloc: adju... |
1158 |
cpu_notifier_register_begin(); |
40f9fb8cf mm/zsmalloc: supp... |
1159 |
|
66cdef663 mm/zsmalloc: adju... |
1160 1161 1162 |
for_each_online_cpu(cpu) zs_cpu_notifier(NULL, CPU_DEAD, (void *)(long)cpu); __unregister_cpu_notifier(&zs_cpu_nb); |
40f9fb8cf mm/zsmalloc: supp... |
1163 |
|
66cdef663 mm/zsmalloc: adju... |
1164 |
cpu_notifier_register_done(); |
b1b00a5b8 zsmalloc: fix zs_... |
1165 |
} |
66cdef663 mm/zsmalloc: adju... |
1166 |
static void init_zs_size_classes(void) |
b1b00a5b8 zsmalloc: fix zs_... |
1167 |
{ |
66cdef663 mm/zsmalloc: adju... |
1168 |
int nr; |
c795779df mm/zpool: zbud/zs... |
1169 |
|
66cdef663 mm/zsmalloc: adju... |
1170 1171 1172 |
nr = (ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) / ZS_SIZE_CLASS_DELTA + 1; if ((ZS_MAX_ALLOC_SIZE - ZS_MIN_ALLOC_SIZE) % ZS_SIZE_CLASS_DELTA) nr += 1; |
40f9fb8cf mm/zsmalloc: supp... |
1173 |
|
66cdef663 mm/zsmalloc: adju... |
1174 |
zs_size_classes = nr; |
61989a80f staging: zsmalloc... |
1175 |
} |
9eec4cd53 zsmalloc: merge s... |
1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 |
static bool can_merge(struct size_class *prev, int size, int pages_per_zspage) { if (prev->pages_per_zspage != pages_per_zspage) return false; if (get_maxobj_per_zspage(prev->size, prev->pages_per_zspage) != get_maxobj_per_zspage(size, pages_per_zspage)) return false; return true; } |
312fcae22 zsmalloc: support... |
1187 1188 1189 1190 1191 1192 |
static bool zspage_full(struct page *page) { BUG_ON(!is_first_page(page)); return page->inuse == page->objects; } |
66cdef663 mm/zsmalloc: adju... |
1193 1194 1195 1196 1197 |
unsigned long zs_get_total_pages(struct zs_pool *pool) { return atomic_long_read(&pool->pages_allocated); } EXPORT_SYMBOL_GPL(zs_get_total_pages); |
4bbc0bc06 staging: zsmalloc... |
1198 |
/** |
66cdef663 mm/zsmalloc: adju... |
1199 1200 1201 |
* zs_map_object - get address of allocated object from handle. * @pool: pool from which the object was allocated * @handle: handle returned from zs_malloc |
4bbc0bc06 staging: zsmalloc... |
1202 |
* |
66cdef663 mm/zsmalloc: adju... |
1203 1204 1205 |
* Before using an object allocated from zs_malloc, it must be mapped using * this function. When done with the object, it must be unmapped using * zs_unmap_object. |
4bbc0bc06 staging: zsmalloc... |
1206 |
* |
66cdef663 mm/zsmalloc: adju... |
1207 1208 1209 1210 |
* Only one object can be mapped per cpu at a time. There is no protection * against nested mappings. * * This function returns with preemption and page faults disabled. |
4bbc0bc06 staging: zsmalloc... |
1211 |
*/ |
66cdef663 mm/zsmalloc: adju... |
1212 1213 |
void *zs_map_object(struct zs_pool *pool, unsigned long handle, enum zs_mapmode mm) |
61989a80f staging: zsmalloc... |
1214 |
{ |
66cdef663 mm/zsmalloc: adju... |
1215 |
struct page *page; |
2e40e163a zsmalloc: decoupl... |
1216 |
unsigned long obj, obj_idx, off; |
61989a80f staging: zsmalloc... |
1217 |
|
66cdef663 mm/zsmalloc: adju... |
1218 1219 1220 1221 1222 |
unsigned int class_idx; enum fullness_group fg; struct size_class *class; struct mapping_area *area; struct page *pages[2]; |
2e40e163a zsmalloc: decoupl... |
1223 |
void *ret; |
61989a80f staging: zsmalloc... |
1224 |
|
66cdef663 mm/zsmalloc: adju... |
1225 |
BUG_ON(!handle); |
40f9fb8cf mm/zsmalloc: supp... |
1226 |
|
9eec4cd53 zsmalloc: merge s... |
1227 |
/* |
66cdef663 mm/zsmalloc: adju... |
1228 1229 1230 |
* Because we use per-cpu mapping areas shared among the * pools/users, we can't allow mapping in interrupt context * because it can corrupt another users mappings. |
9eec4cd53 zsmalloc: merge s... |
1231 |
*/ |
66cdef663 mm/zsmalloc: adju... |
1232 |
BUG_ON(in_interrupt()); |
61989a80f staging: zsmalloc... |
1233 |
|
312fcae22 zsmalloc: support... |
1234 1235 |
/* From now on, migration cannot move the object */ pin_tag(handle); |
2e40e163a zsmalloc: decoupl... |
1236 1237 |
obj = handle_to_obj(handle); obj_to_location(obj, &page, &obj_idx); |
66cdef663 mm/zsmalloc: adju... |
1238 1239 1240 |
get_zspage_mapping(get_first_page(page), &class_idx, &fg); class = pool->size_class[class_idx]; off = obj_idx_to_offset(page, obj_idx, class->size); |
df8b5bb99 mm/zsmalloc: avoi... |
1241 |
|
66cdef663 mm/zsmalloc: adju... |
1242 1243 1244 1245 1246 |
area = &get_cpu_var(zs_map_area); area->vm_mm = mm; if (off + class->size <= PAGE_SIZE) { /* this object is contained entirely within a page */ area->vm_addr = kmap_atomic(page); |
2e40e163a zsmalloc: decoupl... |
1247 1248 |
ret = area->vm_addr + off; goto out; |
61989a80f staging: zsmalloc... |
1249 |
} |
66cdef663 mm/zsmalloc: adju... |
1250 1251 1252 1253 |
/* this object spans two pages */ pages[0] = page; pages[1] = get_next_page(page); BUG_ON(!pages[1]); |
9eec4cd53 zsmalloc: merge s... |
1254 |
|
2e40e163a zsmalloc: decoupl... |
1255 1256 |
ret = __zs_map_object(area, pages, off, class->size); out: |
7b60a6852 zsmalloc: record ... |
1257 1258 1259 1260 |
if (!class->huge) ret += ZS_HANDLE_SIZE; return ret; |
61989a80f staging: zsmalloc... |
1261 |
} |
66cdef663 mm/zsmalloc: adju... |
1262 |
EXPORT_SYMBOL_GPL(zs_map_object); |
61989a80f staging: zsmalloc... |
1263 |
|
66cdef663 mm/zsmalloc: adju... |
1264 |
void zs_unmap_object(struct zs_pool *pool, unsigned long handle) |
61989a80f staging: zsmalloc... |
1265 |
{ |
66cdef663 mm/zsmalloc: adju... |
1266 |
struct page *page; |
2e40e163a zsmalloc: decoupl... |
1267 |
unsigned long obj, obj_idx, off; |
61989a80f staging: zsmalloc... |
1268 |
|
66cdef663 mm/zsmalloc: adju... |
1269 1270 1271 1272 |
unsigned int class_idx; enum fullness_group fg; struct size_class *class; struct mapping_area *area; |
9eec4cd53 zsmalloc: merge s... |
1273 |
|
66cdef663 mm/zsmalloc: adju... |
1274 |
BUG_ON(!handle); |
9eec4cd53 zsmalloc: merge s... |
1275 |
|
2e40e163a zsmalloc: decoupl... |
1276 1277 |
obj = handle_to_obj(handle); obj_to_location(obj, &page, &obj_idx); |
66cdef663 mm/zsmalloc: adju... |
1278 1279 1280 |
get_zspage_mapping(get_first_page(page), &class_idx, &fg); class = pool->size_class[class_idx]; off = obj_idx_to_offset(page, obj_idx, class->size); |
61989a80f staging: zsmalloc... |
1281 |
|
66cdef663 mm/zsmalloc: adju... |
1282 1283 1284 1285 1286 |
area = this_cpu_ptr(&zs_map_area); if (off + class->size <= PAGE_SIZE) kunmap_atomic(area->vm_addr); else { struct page *pages[2]; |
40f9fb8cf mm/zsmalloc: supp... |
1287 |
|
66cdef663 mm/zsmalloc: adju... |
1288 1289 1290 1291 1292 1293 1294 |
pages[0] = page; pages[1] = get_next_page(page); BUG_ON(!pages[1]); __zs_unmap_object(area, pages, off, class->size); } put_cpu_var(zs_map_area); |
312fcae22 zsmalloc: support... |
1295 |
unpin_tag(handle); |
61989a80f staging: zsmalloc... |
1296 |
} |
66cdef663 mm/zsmalloc: adju... |
1297 |
EXPORT_SYMBOL_GPL(zs_unmap_object); |
61989a80f staging: zsmalloc... |
1298 |
|
c78062612 zsmalloc: factor ... |
1299 1300 1301 1302 1303 1304 1305 1306 1307 |
static unsigned long obj_malloc(struct page *first_page, struct size_class *class, unsigned long handle) { unsigned long obj; struct link_free *link; struct page *m_page; unsigned long m_objidx, m_offset; void *vaddr; |
312fcae22 zsmalloc: support... |
1308 |
handle |= OBJ_ALLOCATED_TAG; |
c78062612 zsmalloc: factor ... |
1309 1310 1311 1312 1313 1314 1315 |
obj = (unsigned long)first_page->freelist; obj_to_location(obj, &m_page, &m_objidx); m_offset = obj_idx_to_offset(m_page, m_objidx, class->size); vaddr = kmap_atomic(m_page); link = (struct link_free *)vaddr + m_offset / sizeof(*link); first_page->freelist = link->next; |
7b60a6852 zsmalloc: record ... |
1316 1317 1318 1319 1320 1321 |
if (!class->huge) /* record handle in the header of allocated chunk */ link->handle = handle; else /* record handle in first_page->private */ set_page_private(first_page, handle); |
c78062612 zsmalloc: factor ... |
1322 1323 1324 1325 1326 1327 |
kunmap_atomic(vaddr); first_page->inuse++; zs_stat_inc(class, OBJ_USED, 1); return obj; } |
61989a80f staging: zsmalloc... |
1328 1329 1330 1331 |
/** * zs_malloc - Allocate block of given size from pool. * @pool: pool to allocate from * @size: size of block to allocate |
61989a80f staging: zsmalloc... |
1332 |
* |
00a61d861 staging: zsmalloc... |
1333 |
* On success, handle to the allocated object is returned, |
c23443483 staging: zsmalloc... |
1334 |
* otherwise 0. |
61989a80f staging: zsmalloc... |
1335 1336 |
* Allocation requests with size > ZS_MAX_ALLOC_SIZE will fail. */ |
c23443483 staging: zsmalloc... |
1337 |
unsigned long zs_malloc(struct zs_pool *pool, size_t size) |
61989a80f staging: zsmalloc... |
1338 |
{ |
2e40e163a zsmalloc: decoupl... |
1339 |
unsigned long handle, obj; |
61989a80f staging: zsmalloc... |
1340 |
struct size_class *class; |
c78062612 zsmalloc: factor ... |
1341 |
struct page *first_page; |
61989a80f staging: zsmalloc... |
1342 |
|
7b60a6852 zsmalloc: record ... |
1343 |
if (unlikely(!size || size > ZS_MAX_ALLOC_SIZE)) |
2e40e163a zsmalloc: decoupl... |
1344 1345 1346 1347 |
return 0; handle = alloc_handle(pool); if (!handle) |
c23443483 staging: zsmalloc... |
1348 |
return 0; |
61989a80f staging: zsmalloc... |
1349 |
|
2e40e163a zsmalloc: decoupl... |
1350 1351 |
/* extra space in chunk to keep the handle */ size += ZS_HANDLE_SIZE; |
9eec4cd53 zsmalloc: merge s... |
1352 |
class = pool->size_class[get_size_class_index(size)]; |
61989a80f staging: zsmalloc... |
1353 1354 1355 1356 1357 1358 1359 |
spin_lock(&class->lock); first_page = find_get_zspage(class); if (!first_page) { spin_unlock(&class->lock); first_page = alloc_zspage(class, pool->flags); |
2e40e163a zsmalloc: decoupl... |
1360 1361 |
if (unlikely(!first_page)) { free_handle(pool, handle); |
c23443483 staging: zsmalloc... |
1362 |
return 0; |
2e40e163a zsmalloc: decoupl... |
1363 |
} |
61989a80f staging: zsmalloc... |
1364 1365 |
set_zspage_mapping(first_page, class->index, ZS_EMPTY); |
13de8933c zsmalloc: move pa... |
1366 1367 |
atomic_long_add(class->pages_per_zspage, &pool->pages_allocated); |
0f050d997 mm/zsmalloc: add ... |
1368 |
|
61989a80f staging: zsmalloc... |
1369 |
spin_lock(&class->lock); |
0f050d997 mm/zsmalloc: add ... |
1370 1371 |
zs_stat_inc(class, OBJ_ALLOCATED, get_maxobj_per_zspage( class->size, class->pages_per_zspage)); |
61989a80f staging: zsmalloc... |
1372 |
} |
c78062612 zsmalloc: factor ... |
1373 |
obj = obj_malloc(first_page, class, handle); |
61989a80f staging: zsmalloc... |
1374 |
/* Now move the zspage to another fullness group, if required */ |
c78062612 zsmalloc: factor ... |
1375 |
fix_fullness_group(class, first_page); |
2e40e163a zsmalloc: decoupl... |
1376 |
record_obj(handle, obj); |
61989a80f staging: zsmalloc... |
1377 |
spin_unlock(&class->lock); |
2e40e163a zsmalloc: decoupl... |
1378 |
return handle; |
61989a80f staging: zsmalloc... |
1379 1380 |
} EXPORT_SYMBOL_GPL(zs_malloc); |
c78062612 zsmalloc: factor ... |
1381 1382 |
static void obj_free(struct zs_pool *pool, struct size_class *class, unsigned long obj) |
61989a80f staging: zsmalloc... |
1383 1384 1385 |
{ struct link_free *link; struct page *first_page, *f_page; |
c78062612 zsmalloc: factor ... |
1386 |
unsigned long f_objidx, f_offset; |
af4ee5e97 zsmalloc: correct... |
1387 |
void *vaddr; |
61989a80f staging: zsmalloc... |
1388 |
|
c78062612 zsmalloc: factor ... |
1389 |
BUG_ON(!obj); |
61989a80f staging: zsmalloc... |
1390 |
|
312fcae22 zsmalloc: support... |
1391 |
obj &= ~OBJ_ALLOCATED_TAG; |
2e40e163a zsmalloc: decoupl... |
1392 |
obj_to_location(obj, &f_page, &f_objidx); |
61989a80f staging: zsmalloc... |
1393 |
first_page = get_first_page(f_page); |
61989a80f staging: zsmalloc... |
1394 |
f_offset = obj_idx_to_offset(f_page, f_objidx, class->size); |
c78062612 zsmalloc: factor ... |
1395 |
vaddr = kmap_atomic(f_page); |
61989a80f staging: zsmalloc... |
1396 1397 |
/* Insert this object in containing zspage's freelist */ |
af4ee5e97 zsmalloc: correct... |
1398 |
link = (struct link_free *)(vaddr + f_offset); |
61989a80f staging: zsmalloc... |
1399 |
link->next = first_page->freelist; |
7b60a6852 zsmalloc: record ... |
1400 1401 |
if (class->huge) set_page_private(first_page, 0); |
af4ee5e97 zsmalloc: correct... |
1402 |
kunmap_atomic(vaddr); |
c23443483 staging: zsmalloc... |
1403 |
first_page->freelist = (void *)obj; |
61989a80f staging: zsmalloc... |
1404 |
first_page->inuse--; |
0f050d997 mm/zsmalloc: add ... |
1405 |
zs_stat_dec(class, OBJ_USED, 1); |
c78062612 zsmalloc: factor ... |
1406 1407 1408 1409 1410 1411 1412 1413 1414 1415 1416 1417 |
} void zs_free(struct zs_pool *pool, unsigned long handle) { struct page *first_page, *f_page; unsigned long obj, f_objidx; int class_idx; struct size_class *class; enum fullness_group fullness; if (unlikely(!handle)) return; |
312fcae22 zsmalloc: support... |
1418 |
pin_tag(handle); |
c78062612 zsmalloc: factor ... |
1419 |
obj = handle_to_obj(handle); |
c78062612 zsmalloc: factor ... |
1420 1421 1422 1423 1424 1425 1426 1427 1428 |
obj_to_location(obj, &f_page, &f_objidx); first_page = get_first_page(f_page); get_zspage_mapping(first_page, &class_idx, &fullness); class = pool->size_class[class_idx]; spin_lock(&class->lock); obj_free(pool, class, obj); fullness = fix_fullness_group(class, first_page); |
312fcae22 zsmalloc: support... |
1429 |
if (fullness == ZS_EMPTY) { |
0f050d997 mm/zsmalloc: add ... |
1430 1431 |
zs_stat_dec(class, OBJ_ALLOCATED, get_maxobj_per_zspage( class->size, class->pages_per_zspage)); |
312fcae22 zsmalloc: support... |
1432 1433 1434 1435 |
atomic_long_sub(class->pages_per_zspage, &pool->pages_allocated); free_zspage(first_page); } |
61989a80f staging: zsmalloc... |
1436 |
spin_unlock(&class->lock); |
312fcae22 zsmalloc: support... |
1437 |
unpin_tag(handle); |
61989a80f staging: zsmalloc... |
1438 |
|
312fcae22 zsmalloc: support... |
1439 1440 1441 |
free_handle(pool, handle); } EXPORT_SYMBOL_GPL(zs_free); |
0dc63d488 zsmalloc: cosmeti... |
1442 |
static void zs_object_copy(unsigned long dst, unsigned long src, |
312fcae22 zsmalloc: support... |
1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 |
struct size_class *class) { struct page *s_page, *d_page; unsigned long s_objidx, d_objidx; unsigned long s_off, d_off; void *s_addr, *d_addr; int s_size, d_size, size; int written = 0; s_size = d_size = class->size; obj_to_location(src, &s_page, &s_objidx); obj_to_location(dst, &d_page, &d_objidx); s_off = obj_idx_to_offset(s_page, s_objidx, class->size); d_off = obj_idx_to_offset(d_page, d_objidx, class->size); if (s_off + class->size > PAGE_SIZE) s_size = PAGE_SIZE - s_off; if (d_off + class->size > PAGE_SIZE) d_size = PAGE_SIZE - d_off; s_addr = kmap_atomic(s_page); d_addr = kmap_atomic(d_page); while (1) { size = min(s_size, d_size); memcpy(d_addr + d_off, s_addr + s_off, size); written += size; if (written == class->size) break; |
495819ead zsmalloc: micro-o... |
1476 1477 1478 1479 1480 1481 |
s_off += size; s_size -= size; d_off += size; d_size -= size; if (s_off >= PAGE_SIZE) { |
312fcae22 zsmalloc: support... |
1482 1483 1484 1485 1486 1487 1488 1489 |
kunmap_atomic(d_addr); kunmap_atomic(s_addr); s_page = get_next_page(s_page); BUG_ON(!s_page); s_addr = kmap_atomic(s_page); d_addr = kmap_atomic(d_page); s_size = class->size - written; s_off = 0; |
312fcae22 zsmalloc: support... |
1490 |
} |
495819ead zsmalloc: micro-o... |
1491 |
if (d_off >= PAGE_SIZE) { |
312fcae22 zsmalloc: support... |
1492 1493 1494 1495 1496 1497 |
kunmap_atomic(d_addr); d_page = get_next_page(d_page); BUG_ON(!d_page); d_addr = kmap_atomic(d_page); d_size = class->size - written; d_off = 0; |
312fcae22 zsmalloc: support... |
1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 |
} } kunmap_atomic(d_addr); kunmap_atomic(s_addr); } /* * Find alloced object in zspage from index object and * return handle. */ static unsigned long find_alloced_obj(struct page *page, int index, struct size_class *class) { unsigned long head; int offset = 0; unsigned long handle = 0; void *addr = kmap_atomic(page); if (!is_first_page(page)) offset = page->index; offset += class->size * index; while (offset < PAGE_SIZE) { |
7b60a6852 zsmalloc: record ... |
1522 |
head = obj_to_head(class, page, addr + offset); |
312fcae22 zsmalloc: support... |
1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 |
if (head & OBJ_ALLOCATED_TAG) { handle = head & ~OBJ_ALLOCATED_TAG; if (trypin_tag(handle)) break; handle = 0; } offset += class->size; index++; } kunmap_atomic(addr); return handle; } struct zs_compact_control { /* Source page for migration which could be a subpage of zspage. */ struct page *s_page; /* Destination page for migration which should be a first page * of zspage. */ struct page *d_page; /* Starting object index within @s_page which used for live object * in the subpage. */ int index; |
312fcae22 zsmalloc: support... |
1547 1548 1549 1550 1551 1552 1553 1554 1555 1556 |
}; static int migrate_zspage(struct zs_pool *pool, struct size_class *class, struct zs_compact_control *cc) { unsigned long used_obj, free_obj; unsigned long handle; struct page *s_page = cc->s_page; struct page *d_page = cc->d_page; unsigned long index = cc->index; |
312fcae22 zsmalloc: support... |
1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 |
int ret = 0; while (1) { handle = find_alloced_obj(s_page, index, class); if (!handle) { s_page = get_next_page(s_page); if (!s_page) break; index = 0; continue; } /* Stop if there is no more space */ if (zspage_full(d_page)) { unpin_tag(handle); ret = -ENOMEM; break; } used_obj = handle_to_obj(handle); free_obj = obj_malloc(d_page, class, handle); |
0dc63d488 zsmalloc: cosmeti... |
1578 |
zs_object_copy(free_obj, used_obj, class); |
312fcae22 zsmalloc: support... |
1579 |
index++; |
c102f07ca zsmalloc: fix mig... |
1580 1581 1582 1583 1584 1585 1586 |
/* * record_obj updates handle's value to free_obj and it will * invalidate lock bit(ie, HANDLE_PIN_BIT) of handle, which * breaks synchronization using pin_tag(e,g, zs_free) so * let's keep the lock bit. */ free_obj |= BIT(HANDLE_PIN_BIT); |
312fcae22 zsmalloc: support... |
1587 1588 1589 |
record_obj(handle, free_obj); unpin_tag(handle); obj_free(pool, class, used_obj); |
312fcae22 zsmalloc: support... |
1590 1591 1592 1593 1594 |
} /* Remember last position in this iteration */ cc->s_page = s_page; cc->index = index; |
312fcae22 zsmalloc: support... |
1595 1596 1597 |
return ret; } |
0dc63d488 zsmalloc: cosmeti... |
1598 |
static struct page *isolate_target_page(struct size_class *class) |
312fcae22 zsmalloc: support... |
1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 |
{ int i; struct page *page; for (i = 0; i < _ZS_NR_FULLNESS_GROUPS; i++) { page = class->fullness_list[i]; if (page) { remove_zspage(page, class, i); break; } } return page; } |
860c707dc zsmalloc: account... |
1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 |
/* * putback_zspage - add @first_page into right class's fullness list * @pool: target pool * @class: destination class * @first_page: target page * * Return @fist_page's fullness_group */ static enum fullness_group putback_zspage(struct zs_pool *pool, struct size_class *class, struct page *first_page) |
312fcae22 zsmalloc: support... |
1624 |
{ |
312fcae22 zsmalloc: support... |
1625 1626 1627 |
enum fullness_group fullness; BUG_ON(!is_first_page(first_page)); |
839373e64 zsmalloc: remove ... |
1628 |
fullness = get_fullness_group(first_page); |
312fcae22 zsmalloc: support... |
1629 |
insert_zspage(first_page, class, fullness); |
839373e64 zsmalloc: remove ... |
1630 |
set_zspage_mapping(first_page, class->index, fullness); |
13de8933c zsmalloc: move pa... |
1631 |
if (fullness == ZS_EMPTY) { |
312fcae22 zsmalloc: support... |
1632 1633 |
zs_stat_dec(class, OBJ_ALLOCATED, get_maxobj_per_zspage( class->size, class->pages_per_zspage)); |
13de8933c zsmalloc: move pa... |
1634 1635 |
atomic_long_sub(class->pages_per_zspage, &pool->pages_allocated); |
312fcae22 zsmalloc: support... |
1636 |
|
61989a80f staging: zsmalloc... |
1637 |
free_zspage(first_page); |
13de8933c zsmalloc: move pa... |
1638 |
} |
860c707dc zsmalloc: account... |
1639 1640 |
return fullness; |
61989a80f staging: zsmalloc... |
1641 |
} |
312fcae22 zsmalloc: support... |
1642 1643 1644 |
static struct page *isolate_source_page(struct size_class *class) { |
ad9d5e175 zsmalloc: conside... |
1645 1646 1647 1648 1649 1650 1651 |
int i; struct page *page = NULL; for (i = ZS_ALMOST_EMPTY; i >= ZS_ALMOST_FULL; i--) { page = class->fullness_list[i]; if (!page) continue; |
312fcae22 zsmalloc: support... |
1652 |
|
ad9d5e175 zsmalloc: conside... |
1653 1654 1655 |
remove_zspage(page, class, i); break; } |
312fcae22 zsmalloc: support... |
1656 1657 1658 |
return page; } |
04f05909e zsmalloc: introdu... |
1659 1660 1661 1662 |
/* * * Based on the number of unused allocated objects calculate * and return the number of pages that we can free. |
04f05909e zsmalloc: introdu... |
1663 1664 1665 1666 |
*/ static unsigned long zs_can_compact(struct size_class *class) { unsigned long obj_wasted; |
04f05909e zsmalloc: introdu... |
1667 1668 1669 1670 1671 |
obj_wasted = zs_stat_get(class, OBJ_ALLOCATED) - zs_stat_get(class, OBJ_USED); obj_wasted /= get_maxobj_per_zspage(class->size, class->pages_per_zspage); |
6cbf16b3b zsmalloc: use cla... |
1672 |
return obj_wasted * class->pages_per_zspage; |
04f05909e zsmalloc: introdu... |
1673 |
} |
7d3f39382 zsmalloc/zram: in... |
1674 |
static void __zs_compact(struct zs_pool *pool, struct size_class *class) |
312fcae22 zsmalloc: support... |
1675 |
{ |
312fcae22 zsmalloc: support... |
1676 1677 1678 |
struct zs_compact_control cc; struct page *src_page; struct page *dst_page = NULL; |
312fcae22 zsmalloc: support... |
1679 |
|
312fcae22 zsmalloc: support... |
1680 1681 1682 1683 |
spin_lock(&class->lock); while ((src_page = isolate_source_page(class))) { BUG_ON(!is_first_page(src_page)); |
04f05909e zsmalloc: introdu... |
1684 1685 |
if (!zs_can_compact(class)) break; |
312fcae22 zsmalloc: support... |
1686 1687 |
cc.index = 0; cc.s_page = src_page; |
0dc63d488 zsmalloc: cosmeti... |
1688 |
while ((dst_page = isolate_target_page(class))) { |
312fcae22 zsmalloc: support... |
1689 1690 |
cc.d_page = dst_page; /* |
0dc63d488 zsmalloc: cosmeti... |
1691 1692 |
* If there is no more space in dst_page, resched * and see if anyone had allocated another zspage. |
312fcae22 zsmalloc: support... |
1693 1694 1695 1696 1697 |
*/ if (!migrate_zspage(pool, class, &cc)) break; putback_zspage(pool, class, dst_page); |
312fcae22 zsmalloc: support... |
1698 1699 1700 1701 1702 1703 1704 |
} /* Stop if we couldn't find slot */ if (dst_page == NULL) break; putback_zspage(pool, class, dst_page); |
860c707dc zsmalloc: account... |
1705 |
if (putback_zspage(pool, class, src_page) == ZS_EMPTY) |
6cbf16b3b zsmalloc: use cla... |
1706 |
pool->stats.pages_compacted += class->pages_per_zspage; |
312fcae22 zsmalloc: support... |
1707 |
spin_unlock(&class->lock); |
312fcae22 zsmalloc: support... |
1708 1709 1710 1711 1712 1713 |
cond_resched(); spin_lock(&class->lock); } if (src_page) putback_zspage(pool, class, src_page); |
7d3f39382 zsmalloc/zram: in... |
1714 |
spin_unlock(&class->lock); |
312fcae22 zsmalloc: support... |
1715 1716 1717 1718 1719 |
} unsigned long zs_compact(struct zs_pool *pool) { int i; |
312fcae22 zsmalloc: support... |
1720 1721 1722 1723 1724 1725 1726 1727 |
struct size_class *class; for (i = zs_size_classes - 1; i >= 0; i--) { class = pool->size_class[i]; if (!class) continue; if (class->index != i) continue; |
7d3f39382 zsmalloc/zram: in... |
1728 |
__zs_compact(pool, class); |
312fcae22 zsmalloc: support... |
1729 |
} |
860c707dc zsmalloc: account... |
1730 |
return pool->stats.pages_compacted; |
312fcae22 zsmalloc: support... |
1731 1732 |
} EXPORT_SYMBOL_GPL(zs_compact); |
61989a80f staging: zsmalloc... |
1733 |
|
7d3f39382 zsmalloc/zram: in... |
1734 1735 1736 1737 1738 |
void zs_pool_stats(struct zs_pool *pool, struct zs_pool_stats *stats) { memcpy(stats, &pool->stats, sizeof(struct zs_pool_stats)); } EXPORT_SYMBOL_GPL(zs_pool_stats); |
ab9d306d9 zsmalloc: use shr... |
1739 1740 1741 1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760 1761 1762 1763 1764 |
static unsigned long zs_shrinker_scan(struct shrinker *shrinker, struct shrink_control *sc) { unsigned long pages_freed; struct zs_pool *pool = container_of(shrinker, struct zs_pool, shrinker); pages_freed = pool->stats.pages_compacted; /* * Compact classes and calculate compaction delta. * Can run concurrently with a manually triggered * (by user) compaction. */ pages_freed = zs_compact(pool) - pages_freed; return pages_freed ? pages_freed : SHRINK_STOP; } static unsigned long zs_shrinker_count(struct shrinker *shrinker, struct shrink_control *sc) { int i; struct size_class *class; unsigned long pages_to_free = 0; struct zs_pool *pool = container_of(shrinker, struct zs_pool, shrinker); |
ab9d306d9 zsmalloc: use shr... |
1765 1766 1767 1768 1769 1770 |
for (i = zs_size_classes - 1; i >= 0; i--) { class = pool->size_class[i]; if (!class) continue; if (class->index != i) continue; |
ab9d306d9 zsmalloc: use shr... |
1771 |
pages_to_free += zs_can_compact(class); |
ab9d306d9 zsmalloc: use shr... |
1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 |
} return pages_to_free; } static void zs_unregister_shrinker(struct zs_pool *pool) { if (pool->shrinker_enabled) { unregister_shrinker(&pool->shrinker); pool->shrinker_enabled = false; } } static int zs_register_shrinker(struct zs_pool *pool) { pool->shrinker.scan_objects = zs_shrinker_scan; pool->shrinker.count_objects = zs_shrinker_count; pool->shrinker.batch = 0; pool->shrinker.seeks = DEFAULT_SEEKS; return register_shrinker(&pool->shrinker); } |
00a61d861 staging: zsmalloc... |
1794 |
/** |
66cdef663 mm/zsmalloc: adju... |
1795 1796 |
* zs_create_pool - Creates an allocation pool to work from. * @flags: allocation flags used to allocate pool metadata |
166cfda75 staging: zsmalloc... |
1797 |
* |
66cdef663 mm/zsmalloc: adju... |
1798 1799 |
* This function must be called before anything when using * the zsmalloc allocator. |
166cfda75 staging: zsmalloc... |
1800 |
* |
66cdef663 mm/zsmalloc: adju... |
1801 1802 |
* On success, a pointer to the newly created pool is returned, * otherwise NULL. |
396b7fd6f staging/zsmalloc:... |
1803 |
*/ |
6f3526d6d mm: zsmalloc: con... |
1804 |
struct zs_pool *zs_create_pool(const char *name, gfp_t flags) |
61989a80f staging: zsmalloc... |
1805 |
{ |
66cdef663 mm/zsmalloc: adju... |
1806 1807 1808 |
int i; struct zs_pool *pool; struct size_class *prev_class = NULL; |
61989a80f staging: zsmalloc... |
1809 |
|
66cdef663 mm/zsmalloc: adju... |
1810 1811 1812 |
pool = kzalloc(sizeof(*pool), GFP_KERNEL); if (!pool) return NULL; |
61989a80f staging: zsmalloc... |
1813 |
|
66cdef663 mm/zsmalloc: adju... |
1814 1815 1816 1817 1818 1819 |
pool->size_class = kcalloc(zs_size_classes, sizeof(struct size_class *), GFP_KERNEL); if (!pool->size_class) { kfree(pool); return NULL; } |
61989a80f staging: zsmalloc... |
1820 |
|
2e40e163a zsmalloc: decoupl... |
1821 1822 1823 1824 1825 1826 |
pool->name = kstrdup(name, GFP_KERNEL); if (!pool->name) goto err; if (create_handle_cache(pool)) goto err; |
c60369f01 staging: zsmalloc... |
1827 |
/* |
66cdef663 mm/zsmalloc: adju... |
1828 1829 |
* Iterate reversly, because, size of size_class that we want to use * for merging should be larger or equal to current size. |
c60369f01 staging: zsmalloc... |
1830 |
*/ |
66cdef663 mm/zsmalloc: adju... |
1831 1832 1833 1834 |
for (i = zs_size_classes - 1; i >= 0; i--) { int size; int pages_per_zspage; struct size_class *class; |
c60369f01 staging: zsmalloc... |
1835 |
|
66cdef663 mm/zsmalloc: adju... |
1836 1837 1838 1839 |
size = ZS_MIN_ALLOC_SIZE + i * ZS_SIZE_CLASS_DELTA; if (size > ZS_MAX_ALLOC_SIZE) size = ZS_MAX_ALLOC_SIZE; pages_per_zspage = get_pages_per_zspage(size); |
61989a80f staging: zsmalloc... |
1840 |
|
66cdef663 mm/zsmalloc: adju... |
1841 1842 1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854 1855 1856 1857 1858 1859 1860 1861 1862 1863 |
/* * size_class is used for normal zsmalloc operation such * as alloc/free for that size. Although it is natural that we * have one size_class for each size, there is a chance that we * can get more memory utilization if we use one size_class for * many different sizes whose size_class have same * characteristics. So, we makes size_class point to * previous size_class if possible. */ if (prev_class) { if (can_merge(prev_class, size, pages_per_zspage)) { pool->size_class[i] = prev_class; continue; } } class = kzalloc(sizeof(struct size_class), GFP_KERNEL); if (!class) goto err; class->size = size; class->index = i; class->pages_per_zspage = pages_per_zspage; |
7b60a6852 zsmalloc: record ... |
1864 1865 1866 |
if (pages_per_zspage == 1 && get_maxobj_per_zspage(size, pages_per_zspage) == 1) class->huge = true; |
66cdef663 mm/zsmalloc: adju... |
1867 1868 1869 1870 |
spin_lock_init(&class->lock); pool->size_class[i] = class; prev_class = class; |
61989a80f staging: zsmalloc... |
1871 |
} |
66cdef663 mm/zsmalloc: adju... |
1872 |
pool->flags = flags; |
b74185108 staging: zsmalloc... |
1873 |
|
0f050d997 mm/zsmalloc: add ... |
1874 1875 |
if (zs_pool_stat_create(name, pool)) goto err; |
ab9d306d9 zsmalloc: use shr... |
1876 1877 1878 1879 1880 1881 |
/* * Not critical, we still can use the pool * and user can trigger compaction manually. */ if (zs_register_shrinker(pool) == 0) pool->shrinker_enabled = true; |
66cdef663 mm/zsmalloc: adju... |
1882 1883 1884 1885 1886 |
return pool; err: zs_destroy_pool(pool); return NULL; |
61989a80f staging: zsmalloc... |
1887 |
} |
66cdef663 mm/zsmalloc: adju... |
1888 |
EXPORT_SYMBOL_GPL(zs_create_pool); |
61989a80f staging: zsmalloc... |
1889 |
|
66cdef663 mm/zsmalloc: adju... |
1890 |
void zs_destroy_pool(struct zs_pool *pool) |
61989a80f staging: zsmalloc... |
1891 |
{ |
66cdef663 mm/zsmalloc: adju... |
1892 |
int i; |
61989a80f staging: zsmalloc... |
1893 |
|
ab9d306d9 zsmalloc: use shr... |
1894 |
zs_unregister_shrinker(pool); |
0f050d997 mm/zsmalloc: add ... |
1895 |
zs_pool_stat_destroy(pool); |
66cdef663 mm/zsmalloc: adju... |
1896 1897 1898 |
for (i = 0; i < zs_size_classes; i++) { int fg; struct size_class *class = pool->size_class[i]; |
61989a80f staging: zsmalloc... |
1899 |
|
66cdef663 mm/zsmalloc: adju... |
1900 1901 |
if (!class) continue; |
61989a80f staging: zsmalloc... |
1902 |
|
66cdef663 mm/zsmalloc: adju... |
1903 1904 |
if (class->index != i) continue; |
61989a80f staging: zsmalloc... |
1905 |
|
66cdef663 mm/zsmalloc: adju... |
1906 1907 1908 1909 1910 1911 1912 1913 1914 |
for (fg = 0; fg < _ZS_NR_FULLNESS_GROUPS; fg++) { if (class->fullness_list[fg]) { pr_info("Freeing non-empty class with size %db, fullness group %d ", class->size, fg); } } kfree(class); } |
f553646a6 staging: zsmalloc... |
1915 |
|
2e40e163a zsmalloc: decoupl... |
1916 |
destroy_handle_cache(pool); |
66cdef663 mm/zsmalloc: adju... |
1917 |
kfree(pool->size_class); |
0f050d997 mm/zsmalloc: add ... |
1918 |
kfree(pool->name); |
66cdef663 mm/zsmalloc: adju... |
1919 1920 1921 |
kfree(pool); } EXPORT_SYMBOL_GPL(zs_destroy_pool); |
b74185108 staging: zsmalloc... |
1922 |
|
66cdef663 mm/zsmalloc: adju... |
1923 1924 1925 |
static int __init zs_init(void) { int ret = zs_register_cpu_notifier(); |
0f050d997 mm/zsmalloc: add ... |
1926 1927 |
if (ret) goto notifier_fail; |
66cdef663 mm/zsmalloc: adju... |
1928 1929 1930 1931 1932 1933 |
init_zs_size_classes(); #ifdef CONFIG_ZPOOL zpool_register_driver(&zs_zpool_driver); #endif |
0f050d997 mm/zsmalloc: add ... |
1934 1935 1936 1937 1938 1939 1940 |
ret = zs_stat_init(); if (ret) { pr_err("zs stat initialization failed "); goto stat_fail; } |
66cdef663 mm/zsmalloc: adju... |
1941 |
return 0; |
0f050d997 mm/zsmalloc: add ... |
1942 1943 1944 1945 1946 1947 1948 1949 1950 |
stat_fail: #ifdef CONFIG_ZPOOL zpool_unregister_driver(&zs_zpool_driver); #endif notifier_fail: zs_unregister_cpu_notifier(); return ret; |
61989a80f staging: zsmalloc... |
1951 |
} |
61989a80f staging: zsmalloc... |
1952 |
|
66cdef663 mm/zsmalloc: adju... |
1953 |
static void __exit zs_exit(void) |
61989a80f staging: zsmalloc... |
1954 |
{ |
66cdef663 mm/zsmalloc: adju... |
1955 1956 1957 1958 |
#ifdef CONFIG_ZPOOL zpool_unregister_driver(&zs_zpool_driver); #endif zs_unregister_cpu_notifier(); |
0f050d997 mm/zsmalloc: add ... |
1959 1960 |
zs_stat_exit(); |
61989a80f staging: zsmalloc... |
1961 |
} |
069f101fa staging: zsmalloc... |
1962 1963 1964 1965 1966 1967 |
module_init(zs_init); module_exit(zs_exit); MODULE_LICENSE("Dual BSD/GPL"); MODULE_AUTHOR("Nitin Gupta <ngupta@vflare.org>"); |