Blame view
mm/slub.c
106 KB
81819f0fc
|
1 2 3 4 5 6 7 |
/* * SLUB: A slab allocator that limits cache line use instead of queuing * objects in per cpu and per node lists. * * The allocator synchronizes using per slab locks and only * uses a centralized lock to manage a pool of partial slabs. * |
cde535359
|
8 |
* (C) 2007 SGI, Christoph Lameter |
81819f0fc
|
9 10 11 12 13 14 15 16 |
*/ #include <linux/mm.h> #include <linux/module.h> #include <linux/bit_spinlock.h> #include <linux/interrupt.h> #include <linux/bitops.h> #include <linux/slab.h> |
7b3c3a50a
|
17 |
#include <linux/proc_fs.h> |
81819f0fc
|
18 19 20 21 22 |
#include <linux/seq_file.h> #include <linux/cpu.h> #include <linux/cpuset.h> #include <linux/mempolicy.h> #include <linux/ctype.h> |
3ac7fe5a4
|
23 |
#include <linux/debugobjects.h> |
81819f0fc
|
24 |
#include <linux/kallsyms.h> |
b9049e234
|
25 |
#include <linux/memory.h> |
f8bd2258e
|
26 |
#include <linux/math64.h> |
773ff60e8
|
27 |
#include <linux/fault-inject.h> |
81819f0fc
|
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 |
/* * Lock order: * 1. slab_lock(page) * 2. slab->list_lock * * The slab_lock protects operations on the object of a particular * slab and its metadata in the page struct. If the slab lock * has been taken then no allocations nor frees can be performed * on the objects in the slab nor can the slab be added or removed * from the partial or full lists since this would mean modifying * the page_struct of the slab. * * The list_lock protects the partial and full list on each node and * the partial slab counter. If taken then no new slabs may be added or * removed from the lists nor make the number of partial slabs be modified. * (Note that the total number of slabs is an atomic value that may be * modified without taking the list lock). * * The list_lock is a centralized lock and thus we avoid taking it as * much as possible. As long as SLUB does not have to handle partial * slabs, operations can continue without any centralized lock. F.e. * allocating a long series of objects that fill up slabs does not require * the list lock. * * The lock order is sometimes inverted when we are trying to get a slab * off a list. We take the list_lock and then look for a page on the list * to use. While we do that objects in the slabs may be freed. We can * only operate on the slab if we have also taken the slab_lock. So we use * a slab_trylock() on the slab. If trylock was successful then no frees * can occur anymore and we can use the slab for allocations etc. If the * slab_trylock() does not succeed then frees are in progress in the slab and * we must stay away from it for a while since we may cause a bouncing * cacheline if we try to acquire the lock. So go onto the next slab. * If all pages are busy then we may allocate a new slab instead of reusing * a partial slab. A new slab has noone operating on it and thus there is * no danger of cacheline contention. * * Interrupts are disabled during allocation and deallocation in order to * make the slab allocator safe to use in the context of an irq. In addition * interrupts are disabled to ensure that the processor does not change * while handling per_cpu slabs, due to kernel preemption. * * SLUB assigns one slab for allocation to each processor. * Allocations only occur from these slabs called cpu slabs. * |
672bba3a4
|
74 75 |
* Slabs with free elements are kept on a partial list and during regular * operations no list for full slabs is used. If an object in a full slab is |
81819f0fc
|
76 |
* freed then the slab will show up again on the partial lists. |
672bba3a4
|
77 78 |
* We track full slabs for debugging purposes though because otherwise we * cannot scan all objects. |
81819f0fc
|
79 80 81 82 83 84 85 |
* * Slabs are freed when they become empty. Teardown and setup is * minimal so we rely on the page allocators per cpu caches for * fast frees and allocs. * * Overloading of page flags that are otherwise used for LRU management. * |
4b6f07504
|
86 87 88 89 90 91 92 93 94 95 96 97 |
* PageActive The slab is frozen and exempt from list processing. * This means that the slab is dedicated to a purpose * such as satisfying allocations for a specific * processor. Objects may be freed in the slab while * it is frozen but slab_free will then skip the usual * list operations. It is up to the processor holding * the slab to integrate the slab into the slab lists * when the slab is no longer needed. * * One use of this flag is to mark slabs that are * used for allocations. Then such a slab becomes a cpu * slab. The cpu slab may be equipped with an additional |
dfb4f0960
|
98 |
* freelist that allows lockless access to |
894b8788d
|
99 100 |
* free objects in addition to the regular freelist * that requires the slab lock. |
81819f0fc
|
101 102 103 |
* * PageError Slab requires special handling due to debug * options set. This moves slab handling out of |
894b8788d
|
104 |
* the fast path and disables lockless freelists. |
81819f0fc
|
105 |
*/ |
5577bd8a8
|
106 |
#ifdef CONFIG_SLUB_DEBUG |
8a38082d2
|
107 |
#define SLABDEBUG 1 |
5577bd8a8
|
108 109 110 |
#else #define SLABDEBUG 0 #endif |
81819f0fc
|
111 112 113 |
/* * Issues still to be resolved: * |
81819f0fc
|
114 115 |
* - Support PAGE_ALLOC_DEBUG. Should be easy to do. * |
81819f0fc
|
116 117 118 119 120 |
* - Variable sizing of the per node arrays */ /* Enable to test recovery from slab corruption on boot */ #undef SLUB_RESILIENCY_TEST |
81819f0fc
|
121 |
/* |
2086d26a0
|
122 123 124 |
* Mininum number of partial slabs. These will be left on the partial * lists even if they are empty. kmem_cache_shrink may reclaim them. */ |
76be89500
|
125 |
#define MIN_PARTIAL 5 |
e95eed571
|
126 |
|
2086d26a0
|
127 128 129 130 131 132 |
/* * Maximum number of desirable partial slabs. * The existence of more partial slabs makes kmem_cache_shrink * sort the partial list by the number of objects in the. */ #define MAX_PARTIAL 10 |
81819f0fc
|
133 134 |
#define DEBUG_DEFAULT_FLAGS (SLAB_DEBUG_FREE | SLAB_RED_ZONE | \ SLAB_POISON | SLAB_STORE_USER) |
672bba3a4
|
135 |
|
81819f0fc
|
136 137 138 139 140 141 142 143 144 145 |
/* * Set of flags that will prevent slab merging */ #define SLUB_NEVER_MERGE (SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | \ SLAB_TRACE | SLAB_DESTROY_BY_RCU) #define SLUB_MERGE_SAME (SLAB_DEBUG_FREE | SLAB_RECLAIM_ACCOUNT | \ SLAB_CACHE_DMA) #ifndef ARCH_KMALLOC_MINALIGN |
47bfdc0d5
|
146 |
#define ARCH_KMALLOC_MINALIGN __alignof__(unsigned long long) |
81819f0fc
|
147 148 149 |
#endif #ifndef ARCH_SLAB_MINALIGN |
47bfdc0d5
|
150 |
#define ARCH_SLAB_MINALIGN __alignof__(unsigned long long) |
81819f0fc
|
151 |
#endif |
210b5c061
|
152 153 154 |
#define OO_SHIFT 16 #define OO_MASK ((1 << OO_SHIFT) - 1) #define MAX_OBJS_PER_PAGE 65535 /* since page.objects is u16 */ |
81819f0fc
|
155 |
/* Internal SLUB flags */ |
1ceef4024
|
156 157 |
#define __OBJECT_POISON 0x80000000 /* Poison object */ #define __SYSFS_ADD_DEFERRED 0x40000000 /* Not yet visible via sysfs */ |
81819f0fc
|
158 159 160 161 162 163 164 165 166 167 |
static int kmem_size = sizeof(struct kmem_cache); #ifdef CONFIG_SMP static struct notifier_block slab_notifier; #endif static enum { DOWN, /* No slab functionality available */ PARTIAL, /* kmem_cache_open() works but kmalloc does not */ |
672bba3a4
|
168 |
UP, /* Everything works but does not show up in sysfs */ |
81819f0fc
|
169 170 171 172 173 |
SYSFS /* Sysfs up */ } slab_state = DOWN; /* A list of all slab caches on the system */ static DECLARE_RWSEM(slub_lock); |
5af328a51
|
174 |
static LIST_HEAD(slab_caches); |
81819f0fc
|
175 |
|
02cbc8744
|
176 177 178 179 |
/* * Tracking user of a slab. */ struct track { |
ce71e27c6
|
180 |
unsigned long addr; /* Called from address */ |
02cbc8744
|
181 182 183 184 185 186 |
int cpu; /* Was running on cpu */ int pid; /* Pid context */ unsigned long when; /* When did the operation occur */ }; enum track_item { TRACK_ALLOC, TRACK_FREE }; |
f6acb6350
|
187 |
#ifdef CONFIG_SLUB_DEBUG |
81819f0fc
|
188 189 190 |
static int sysfs_slab_add(struct kmem_cache *); static int sysfs_slab_alias(struct kmem_cache *, const char *); static void sysfs_slab_remove(struct kmem_cache *); |
8ff12cfc0
|
191 |
|
81819f0fc
|
192 |
#else |
0c7100132
|
193 194 195 |
static inline int sysfs_slab_add(struct kmem_cache *s) { return 0; } static inline int sysfs_slab_alias(struct kmem_cache *s, const char *p) { return 0; } |
151c602f7
|
196 197 198 199 |
static inline void sysfs_slab_remove(struct kmem_cache *s) { kfree(s); } |
8ff12cfc0
|
200 |
|
81819f0fc
|
201 |
#endif |
8ff12cfc0
|
202 203 204 205 206 207 |
static inline void stat(struct kmem_cache_cpu *c, enum stat_item si) { #ifdef CONFIG_SLUB_STATS c->stat[si]++; #endif } |
81819f0fc
|
208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 |
/******************************************************************** * Core slab cache functions *******************************************************************/ int slab_is_available(void) { return slab_state >= UP; } static inline struct kmem_cache_node *get_node(struct kmem_cache *s, int node) { #ifdef CONFIG_NUMA return s->node[node]; #else return &s->local_node; #endif } |
dfb4f0960
|
225 226 |
static inline struct kmem_cache_cpu *get_cpu_slab(struct kmem_cache *s, int cpu) { |
4c93c355d
|
227 228 229 230 231 |
#ifdef CONFIG_SMP return s->cpu_slab[cpu]; #else return &s->cpu_slab; #endif |
dfb4f0960
|
232 |
} |
6446faa2f
|
233 |
/* Verify that a pointer has an address that is valid within a slab page */ |
02cbc8744
|
234 235 236 237 |
static inline int check_valid_pointer(struct kmem_cache *s, struct page *page, const void *object) { void *base; |
a973e9dd1
|
238 |
if (!object) |
02cbc8744
|
239 |
return 1; |
a973e9dd1
|
240 |
base = page_address(page); |
39b264641
|
241 |
if (object < base || object >= base + page->objects * s->size || |
02cbc8744
|
242 243 244 245 246 247 |
(object - base) % s->size) { return 0; } return 1; } |
81819f0fc
|
248 |
/* |
7656c72b5
|
249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 |
* Slow version of get and set free pointer. * * This version requires touching the cache lines of kmem_cache which * we avoid to do in the fast alloc free paths. There we obtain the offset * from the page struct. */ static inline void *get_freepointer(struct kmem_cache *s, void *object) { return *(void **)(object + s->offset); } static inline void set_freepointer(struct kmem_cache *s, void *object, void *fp) { *(void **)(object + s->offset) = fp; } /* Loop over all objects in a slab */ |
224a88be4
|
266 267 |
#define for_each_object(__p, __s, __addr, __objects) \ for (__p = (__addr); __p < (__addr) + (__objects) * (__s)->size;\ |
7656c72b5
|
268 269 270 271 |
__p += (__s)->size) /* Scan freelist */ #define for_each_free_object(__p, __s, __free) \ |
a973e9dd1
|
272 |
for (__p = (__free); __p; __p = get_freepointer((__s), __p)) |
7656c72b5
|
273 274 275 276 277 278 |
/* Determine object index from a given position */ static inline int slab_index(void *p, struct kmem_cache *s, void *addr) { return (p - addr) / s->size; } |
834f3d119
|
279 280 281 282 |
static inline struct kmem_cache_order_objects oo_make(int order, unsigned long size) { struct kmem_cache_order_objects x = { |
210b5c061
|
283 |
(order << OO_SHIFT) + (PAGE_SIZE << order) / size |
834f3d119
|
284 285 286 287 288 289 290 |
}; return x; } static inline int oo_order(struct kmem_cache_order_objects x) { |
210b5c061
|
291 |
return x.x >> OO_SHIFT; |
834f3d119
|
292 293 294 295 |
} static inline int oo_objects(struct kmem_cache_order_objects x) { |
210b5c061
|
296 |
return x.x & OO_MASK; |
834f3d119
|
297 |
} |
41ecc55b8
|
298 299 300 301 |
#ifdef CONFIG_SLUB_DEBUG /* * Debug settings: */ |
f0630fff5
|
302 303 304 |
#ifdef CONFIG_SLUB_DEBUG_ON static int slub_debug = DEBUG_DEFAULT_FLAGS; #else |
41ecc55b8
|
305 |
static int slub_debug; |
f0630fff5
|
306 |
#endif |
41ecc55b8
|
307 308 |
static char *slub_debug_slabs; |
7656c72b5
|
309 |
/* |
81819f0fc
|
310 311 312 313 314 315 316 317 318 319 320 321 |
* Object debugging */ static void print_section(char *text, u8 *addr, unsigned int length) { int i, offset; int newline = 1; char ascii[17]; ascii[16] = 0; for (i = 0; i < length; i++) { if (newline) { |
249226847
|
322 |
printk(KERN_ERR "%8s 0x%p: ", text, addr + i); |
81819f0fc
|
323 324 |
newline = 0; } |
064287807
|
325 |
printk(KERN_CONT " %02x", addr[i]); |
81819f0fc
|
326 327 328 |
offset = i % 16; ascii[offset] = isgraph(addr[i]) ? addr[i] : '.'; if (offset == 15) { |
064287807
|
329 330 |
printk(KERN_CONT " %s ", ascii); |
81819f0fc
|
331 332 333 334 335 336 |
newline = 1; } } if (!newline) { i %= 16; while (i < 16) { |
064287807
|
337 |
printk(KERN_CONT " "); |
81819f0fc
|
338 339 340 |
ascii[i] = ' '; i++; } |
064287807
|
341 342 |
printk(KERN_CONT " %s ", ascii); |
81819f0fc
|
343 344 |
} } |
81819f0fc
|
345 346 347 348 349 350 351 352 353 354 355 356 357 358 |
static struct track *get_track(struct kmem_cache *s, void *object, enum track_item alloc) { struct track *p; if (s->offset) p = object + s->offset + sizeof(void *); else p = object + s->inuse; return p + alloc; } static void set_track(struct kmem_cache *s, void *object, |
ce71e27c6
|
359 |
enum track_item alloc, unsigned long addr) |
81819f0fc
|
360 361 362 363 364 365 366 367 368 369 370 371 |
{ struct track *p; if (s->offset) p = object + s->offset + sizeof(void *); else p = object + s->inuse; p += alloc; if (addr) { p->addr = addr; p->cpu = smp_processor_id(); |
88e4ccf29
|
372 |
p->pid = current->pid; |
81819f0fc
|
373 374 375 376 |
p->when = jiffies; } else memset(p, 0, sizeof(struct track)); } |
81819f0fc
|
377 378 |
static void init_tracking(struct kmem_cache *s, void *object) { |
249226847
|
379 380 |
if (!(s->flags & SLAB_STORE_USER)) return; |
ce71e27c6
|
381 382 |
set_track(s, object, TRACK_FREE, 0UL); set_track(s, object, TRACK_ALLOC, 0UL); |
81819f0fc
|
383 384 385 386 387 388 |
} static void print_track(const char *s, struct track *t) { if (!t->addr) return; |
7daf705f3
|
389 390 |
printk(KERN_ERR "INFO: %s in %pS age=%lu cpu=%u pid=%d ", |
ce71e27c6
|
391 |
s, (void *)t->addr, jiffies - t->when, t->cpu, t->pid); |
249226847
|
392 393 394 395 396 397 398 399 400 401 402 403 404 |
} static void print_tracking(struct kmem_cache *s, void *object) { if (!(s->flags & SLAB_STORE_USER)) return; print_track("Allocated", get_track(s, object, TRACK_ALLOC)); print_track("Freed", get_track(s, object, TRACK_FREE)); } static void print_page_info(struct page *page) { |
39b264641
|
405 406 407 |
printk(KERN_ERR "INFO: Slab 0x%p objects=%u used=%u fp=0x%p flags=0x%04lx ", page, page->objects, page->inuse, page->freelist, page->flags); |
249226847
|
408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 |
} static void slab_bug(struct kmem_cache *s, char *fmt, ...) { va_list args; char buf[100]; va_start(args, fmt); vsnprintf(buf, sizeof(buf), fmt, args); va_end(args); printk(KERN_ERR "========================================" "===================================== "); printk(KERN_ERR "BUG %s: %s ", s->name, buf); printk(KERN_ERR "----------------------------------------" "------------------------------------- "); |
81819f0fc
|
428 |
} |
249226847
|
429 430 431 432 433 434 435 436 437 438 439 440 441 |
static void slab_fix(struct kmem_cache *s, char *fmt, ...) { va_list args; char buf[100]; va_start(args, fmt); vsnprintf(buf, sizeof(buf), fmt, args); va_end(args); printk(KERN_ERR "FIX %s: %s ", s->name, buf); } static void print_trailer(struct kmem_cache *s, struct page *page, u8 *p) |
81819f0fc
|
442 443 |
{ unsigned int off; /* Offset of last byte */ |
a973e9dd1
|
444 |
u8 *addr = page_address(page); |
249226847
|
445 446 447 448 449 450 451 452 453 454 455 456 |
print_tracking(s, p); print_page_info(page); printk(KERN_ERR "INFO: Object 0x%p @offset=%tu fp=0x%p ", p, p - addr, get_freepointer(s, p)); if (p > addr + 16) print_section("Bytes b4", p - 16, 16); |
0ebd652b3
|
457 |
print_section("Object", p, min_t(unsigned long, s->objsize, PAGE_SIZE)); |
81819f0fc
|
458 459 460 461 |
if (s->flags & SLAB_RED_ZONE) print_section("Redzone", p + s->objsize, s->inuse - s->objsize); |
81819f0fc
|
462 463 464 465 |
if (s->offset) off = s->offset + sizeof(void *); else off = s->inuse; |
249226847
|
466 |
if (s->flags & SLAB_STORE_USER) |
81819f0fc
|
467 |
off += 2 * sizeof(struct track); |
81819f0fc
|
468 469 470 |
if (off != s->size) /* Beginning of the filler is the free pointer */ |
249226847
|
471 472 473 |
print_section("Padding", p + off, s->size - off); dump_stack(); |
81819f0fc
|
474 475 476 477 478 |
} static void object_err(struct kmem_cache *s, struct page *page, u8 *object, char *reason) { |
3dc506378
|
479 |
slab_bug(s, "%s", reason); |
249226847
|
480 |
print_trailer(s, page, object); |
81819f0fc
|
481 |
} |
249226847
|
482 |
static void slab_err(struct kmem_cache *s, struct page *page, char *fmt, ...) |
81819f0fc
|
483 484 485 |
{ va_list args; char buf[100]; |
249226847
|
486 487 |
va_start(args, fmt); vsnprintf(buf, sizeof(buf), fmt, args); |
81819f0fc
|
488 |
va_end(args); |
3dc506378
|
489 |
slab_bug(s, "%s", buf); |
249226847
|
490 |
print_page_info(page); |
81819f0fc
|
491 492 493 494 495 496 497 498 499 |
dump_stack(); } static void init_object(struct kmem_cache *s, void *object, int active) { u8 *p = object; if (s->flags & __OBJECT_POISON) { memset(p, POISON_FREE, s->objsize - 1); |
064287807
|
500 |
p[s->objsize - 1] = POISON_END; |
81819f0fc
|
501 502 503 504 505 506 507 |
} if (s->flags & SLAB_RED_ZONE) memset(p + s->objsize, active ? SLUB_RED_ACTIVE : SLUB_RED_INACTIVE, s->inuse - s->objsize); } |
249226847
|
508 |
static u8 *check_bytes(u8 *start, unsigned int value, unsigned int bytes) |
81819f0fc
|
509 510 511 |
{ while (bytes) { if (*start != (u8)value) |
249226847
|
512 |
return start; |
81819f0fc
|
513 514 515 |
start++; bytes--; } |
249226847
|
516 517 518 519 520 521 522 523 524 525 526 527 528 |
return NULL; } static void restore_bytes(struct kmem_cache *s, char *message, u8 data, void *from, void *to) { slab_fix(s, "Restoring 0x%p-0x%p=0x%x ", from, to - 1, data); memset(from, data, to - from); } static int check_bytes_and_report(struct kmem_cache *s, struct page *page, u8 *object, char *what, |
064287807
|
529 |
u8 *start, unsigned int value, unsigned int bytes) |
249226847
|
530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 |
{ u8 *fault; u8 *end; fault = check_bytes(start, value, bytes); if (!fault) return 1; end = start + bytes; while (end > fault && end[-1] == value) end--; slab_bug(s, "%s overwritten", what); printk(KERN_ERR "INFO: 0x%p-0x%p. First byte 0x%x instead of 0x%x ", fault, end - 1, fault[0], value); print_trailer(s, page, object); restore_bytes(s, what, value, fault, end); return 0; |
81819f0fc
|
550 |
} |
81819f0fc
|
551 552 553 554 555 556 557 |
/* * Object layout: * * object address * Bytes of the object to be managed. * If the freepointer may overlay the object then the free * pointer is the first word of the object. |
672bba3a4
|
558 |
* |
81819f0fc
|
559 560 561 562 563 |
* Poisoning uses 0x6b (POISON_FREE) and the last byte is * 0xa5 (POISON_END) * * object + s->objsize * Padding to reach word boundary. This is also used for Redzoning. |
672bba3a4
|
564 565 566 |
* Padding is extended by another word if Redzoning is enabled and * objsize == inuse. * |
81819f0fc
|
567 568 569 570 |
* We fill with 0xbb (RED_INACTIVE) for inactive objects and with * 0xcc (RED_ACTIVE) for objects in use. * * object + s->inuse |
672bba3a4
|
571 572 |
* Meta data starts here. * |
81819f0fc
|
573 574 |
* A. Free pointer (if we cannot overwrite object on free) * B. Tracking data for SLAB_STORE_USER |
672bba3a4
|
575 |
* C. Padding to reach required alignment boundary or at mininum |
6446faa2f
|
576 |
* one word if debugging is on to be able to detect writes |
672bba3a4
|
577 578 579 |
* before the word boundary. * * Padding is done using 0x5a (POISON_INUSE) |
81819f0fc
|
580 581 |
* * object + s->size |
672bba3a4
|
582 |
* Nothing is used beyond s->size. |
81819f0fc
|
583 |
* |
672bba3a4
|
584 585 |
* If slabcaches are merged then the objsize and inuse boundaries are mostly * ignored. And therefore no slab options that rely on these boundaries |
81819f0fc
|
586 587 |
* may be used with merged slabcaches. */ |
81819f0fc
|
588 589 590 591 592 593 594 595 596 597 598 599 600 601 |
static int check_pad_bytes(struct kmem_cache *s, struct page *page, u8 *p) { unsigned long off = s->inuse; /* The end of info */ if (s->offset) /* Freepointer is placed after the object. */ off += sizeof(void *); if (s->flags & SLAB_STORE_USER) /* We also have user information there */ off += 2 * sizeof(struct track); if (s->size == off) return 1; |
249226847
|
602 603 |
return check_bytes_and_report(s, page, p, "Object padding", p + off, POISON_INUSE, s->size - off); |
81819f0fc
|
604 |
} |
39b264641
|
605 |
/* Check the pad bytes at the end of a slab page */ |
81819f0fc
|
606 607 |
static int slab_pad_check(struct kmem_cache *s, struct page *page) { |
249226847
|
608 609 610 611 612 |
u8 *start; u8 *fault; u8 *end; int length; int remainder; |
81819f0fc
|
613 614 615 |
if (!(s->flags & SLAB_POISON)) return 1; |
a973e9dd1
|
616 |
start = page_address(page); |
834f3d119
|
617 |
length = (PAGE_SIZE << compound_order(page)); |
39b264641
|
618 619 |
end = start + length; remainder = length % s->size; |
81819f0fc
|
620 621 |
if (!remainder) return 1; |
39b264641
|
622 |
fault = check_bytes(end - remainder, POISON_INUSE, remainder); |
249226847
|
623 624 625 626 627 628 |
if (!fault) return 1; while (end > fault && end[-1] == POISON_INUSE) end--; slab_err(s, page, "Padding overwritten. 0x%p-0x%p", fault, end - 1); |
39b264641
|
629 |
print_section("Padding", end - remainder, remainder); |
249226847
|
630 631 632 |
restore_bytes(s, "slab padding", POISON_INUSE, start, end); return 0; |
81819f0fc
|
633 634 635 636 637 638 639 640 641 642 643 |
} static int check_object(struct kmem_cache *s, struct page *page, void *object, int active) { u8 *p = object; u8 *endobject = object + s->objsize; if (s->flags & SLAB_RED_ZONE) { unsigned int red = active ? SLUB_RED_ACTIVE : SLUB_RED_INACTIVE; |
249226847
|
644 645 |
if (!check_bytes_and_report(s, page, object, "Redzone", endobject, red, s->inuse - s->objsize)) |
81819f0fc
|
646 |
return 0; |
81819f0fc
|
647 |
} else { |
3adbefee6
|
648 649 650 651 |
if ((s->flags & SLAB_POISON) && s->objsize < s->inuse) { check_bytes_and_report(s, page, p, "Alignment padding", endobject, POISON_INUSE, s->inuse - s->objsize); } |
81819f0fc
|
652 653 654 655 |
} if (s->flags & SLAB_POISON) { if (!active && (s->flags & __OBJECT_POISON) && |
249226847
|
656 657 658 |
(!check_bytes_and_report(s, page, p, "Poison", p, POISON_FREE, s->objsize - 1) || !check_bytes_and_report(s, page, p, "Poison", |
064287807
|
659 |
p + s->objsize - 1, POISON_END, 1))) |
81819f0fc
|
660 |
return 0; |
81819f0fc
|
661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 |
/* * check_pad_bytes cleans up on its own. */ check_pad_bytes(s, page, p); } if (!s->offset && active) /* * Object and freepointer overlap. Cannot check * freepointer while object is allocated. */ return 1; /* Check free pointer validity */ if (!check_valid_pointer(s, page, get_freepointer(s, p))) { object_err(s, page, p, "Freepointer corrupt"); /* |
9f6c708e5
|
678 |
* No choice but to zap it and thus lose the remainder |
81819f0fc
|
679 |
* of the free objects in this slab. May cause |
672bba3a4
|
680 |
* another error because the object count is now wrong. |
81819f0fc
|
681 |
*/ |
a973e9dd1
|
682 |
set_freepointer(s, p, NULL); |
81819f0fc
|
683 684 685 686 687 688 689 |
return 0; } return 1; } static int check_slab(struct kmem_cache *s, struct page *page) { |
39b264641
|
690 |
int maxobj; |
81819f0fc
|
691 692 693 |
VM_BUG_ON(!irqs_disabled()); if (!PageSlab(page)) { |
249226847
|
694 |
slab_err(s, page, "Not a valid slab page"); |
81819f0fc
|
695 696 |
return 0; } |
39b264641
|
697 698 699 700 701 702 703 704 |
maxobj = (PAGE_SIZE << compound_order(page)) / s->size; if (page->objects > maxobj) { slab_err(s, page, "objects %u > max %u", s->name, page->objects, maxobj); return 0; } if (page->inuse > page->objects) { |
249226847
|
705 |
slab_err(s, page, "inuse %u > max %u", |
39b264641
|
706 |
s->name, page->inuse, page->objects); |
81819f0fc
|
707 708 709 710 711 712 713 714 |
return 0; } /* Slab_pad_check fixes things up after itself */ slab_pad_check(s, page); return 1; } /* |
672bba3a4
|
715 716 |
* Determine if a certain object on a page is on the freelist. Must hold the * slab lock to guarantee that the chains are in a consistent state. |
81819f0fc
|
717 718 719 720 721 722 |
*/ static int on_freelist(struct kmem_cache *s, struct page *page, void *search) { int nr = 0; void *fp = page->freelist; void *object = NULL; |
224a88be4
|
723 |
unsigned long max_objects; |
81819f0fc
|
724 |
|
39b264641
|
725 |
while (fp && nr <= page->objects) { |
81819f0fc
|
726 727 728 729 730 731 |
if (fp == search) return 1; if (!check_valid_pointer(s, page, fp)) { if (object) { object_err(s, page, object, "Freechain corrupt"); |
a973e9dd1
|
732 |
set_freepointer(s, object, NULL); |
81819f0fc
|
733 734 |
break; } else { |
249226847
|
735 |
slab_err(s, page, "Freepointer corrupt"); |
a973e9dd1
|
736 |
page->freelist = NULL; |
39b264641
|
737 |
page->inuse = page->objects; |
249226847
|
738 |
slab_fix(s, "Freelist cleared"); |
81819f0fc
|
739 740 741 742 743 744 745 746 |
return 0; } break; } object = fp; fp = get_freepointer(s, object); nr++; } |
224a88be4
|
747 |
max_objects = (PAGE_SIZE << compound_order(page)) / s->size; |
210b5c061
|
748 749 |
if (max_objects > MAX_OBJS_PER_PAGE) max_objects = MAX_OBJS_PER_PAGE; |
224a88be4
|
750 751 752 753 754 755 756 |
if (page->objects != max_objects) { slab_err(s, page, "Wrong number of objects. Found %d but " "should be %d", page->objects, max_objects); page->objects = max_objects; slab_fix(s, "Number of objects adjusted."); } |
39b264641
|
757 |
if (page->inuse != page->objects - nr) { |
70d71228a
|
758 |
slab_err(s, page, "Wrong object count. Counter is %d but " |
39b264641
|
759 760 |
"counted were %d", page->inuse, page->objects - nr); page->inuse = page->objects - nr; |
249226847
|
761 |
slab_fix(s, "Object count adjusted."); |
81819f0fc
|
762 763 764 |
} return search == NULL; } |
0121c619d
|
765 766 |
static void trace(struct kmem_cache *s, struct page *page, void *object, int alloc) |
3ec097421
|
767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 |
{ if (s->flags & SLAB_TRACE) { printk(KERN_INFO "TRACE %s %s 0x%p inuse=%d fp=0x%p ", s->name, alloc ? "alloc" : "free", object, page->inuse, page->freelist); if (!alloc) print_section("Object", (void *)object, s->objsize); dump_stack(); } } |
643b11384
|
782 |
/* |
672bba3a4
|
783 |
* Tracking of fully allocated slabs for debugging purposes. |
643b11384
|
784 |
*/ |
e95eed571
|
785 |
static void add_full(struct kmem_cache_node *n, struct page *page) |
643b11384
|
786 |
{ |
643b11384
|
787 788 789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 |
spin_lock(&n->list_lock); list_add(&page->lru, &n->full); spin_unlock(&n->list_lock); } static void remove_full(struct kmem_cache *s, struct page *page) { struct kmem_cache_node *n; if (!(s->flags & SLAB_STORE_USER)) return; n = get_node(s, page_to_nid(page)); spin_lock(&n->list_lock); list_del(&page->lru); spin_unlock(&n->list_lock); } |
0f389ec63
|
805 806 807 808 809 810 811 |
/* Tracking of the number of slabs for debugging purposes */ static inline unsigned long slabs_node(struct kmem_cache *s, int node) { struct kmem_cache_node *n = get_node(s, node); return atomic_long_read(&n->nr_slabs); } |
205ab99dd
|
812 |
static inline void inc_slabs_node(struct kmem_cache *s, int node, int objects) |
0f389ec63
|
813 814 815 816 817 818 819 820 821 |
{ struct kmem_cache_node *n = get_node(s, node); /* * May be called early in order to allocate a slab for the * kmem_cache_node structure. Solve the chicken-egg * dilemma by deferring the increment of the count during * bootstrap (see early_kmem_cache_node_alloc). */ |
205ab99dd
|
822 |
if (!NUMA_BUILD || n) { |
0f389ec63
|
823 |
atomic_long_inc(&n->nr_slabs); |
205ab99dd
|
824 825 |
atomic_long_add(objects, &n->total_objects); } |
0f389ec63
|
826 |
} |
205ab99dd
|
827 |
static inline void dec_slabs_node(struct kmem_cache *s, int node, int objects) |
0f389ec63
|
828 829 830 831 |
{ struct kmem_cache_node *n = get_node(s, node); atomic_long_dec(&n->nr_slabs); |
205ab99dd
|
832 |
atomic_long_sub(objects, &n->total_objects); |
0f389ec63
|
833 834 835 |
} /* Object debug checks for alloc/free paths */ |
3ec097421
|
836 837 838 839 840 841 842 843 844 845 846 |
static void setup_object_debug(struct kmem_cache *s, struct page *page, void *object) { if (!(s->flags & (SLAB_STORE_USER|SLAB_RED_ZONE|__OBJECT_POISON))) return; init_object(s, object, 0); init_tracking(s, object); } static int alloc_debug_processing(struct kmem_cache *s, struct page *page, |
ce71e27c6
|
847 |
void *object, unsigned long addr) |
81819f0fc
|
848 849 850 |
{ if (!check_slab(s, page)) goto bad; |
d692ef6dc
|
851 |
if (!on_freelist(s, page, object)) { |
249226847
|
852 |
object_err(s, page, object, "Object already allocated"); |
70d71228a
|
853 |
goto bad; |
81819f0fc
|
854 855 856 857 |
} if (!check_valid_pointer(s, page, object)) { object_err(s, page, object, "Freelist Pointer check fails"); |
70d71228a
|
858 |
goto bad; |
81819f0fc
|
859 |
} |
d692ef6dc
|
860 |
if (!check_object(s, page, object, 0)) |
81819f0fc
|
861 |
goto bad; |
81819f0fc
|
862 |
|
3ec097421
|
863 864 865 866 867 |
/* Success perform special debug activities for allocs */ if (s->flags & SLAB_STORE_USER) set_track(s, object, TRACK_ALLOC, addr); trace(s, page, object, 1); init_object(s, object, 1); |
81819f0fc
|
868 |
return 1; |
3ec097421
|
869 |
|
81819f0fc
|
870 871 872 873 874 |
bad: if (PageSlab(page)) { /* * If this is a slab page then lets do the best we can * to avoid issues in the future. Marking all objects |
672bba3a4
|
875 |
* as used avoids touching the remaining objects. |
81819f0fc
|
876 |
*/ |
249226847
|
877 |
slab_fix(s, "Marking all objects used"); |
39b264641
|
878 |
page->inuse = page->objects; |
a973e9dd1
|
879 |
page->freelist = NULL; |
81819f0fc
|
880 881 882 |
} return 0; } |
3ec097421
|
883 |
static int free_debug_processing(struct kmem_cache *s, struct page *page, |
ce71e27c6
|
884 |
void *object, unsigned long addr) |
81819f0fc
|
885 886 887 888 889 |
{ if (!check_slab(s, page)) goto fail; if (!check_valid_pointer(s, page, object)) { |
70d71228a
|
890 |
slab_err(s, page, "Invalid object pointer 0x%p", object); |
81819f0fc
|
891 892 893 894 |
goto fail; } if (on_freelist(s, page, object)) { |
249226847
|
895 |
object_err(s, page, object, "Object already free"); |
81819f0fc
|
896 897 898 899 900 901 902 |
goto fail; } if (!check_object(s, page, object, 1)) return 0; if (unlikely(s != page->slab)) { |
3adbefee6
|
903 |
if (!PageSlab(page)) { |
70d71228a
|
904 905 |
slab_err(s, page, "Attempt to free object(0x%p) " "outside of slab", object); |
3adbefee6
|
906 |
} else if (!page->slab) { |
81819f0fc
|
907 |
printk(KERN_ERR |
70d71228a
|
908 909 |
"SLUB <none>: no slab for object 0x%p. ", |
81819f0fc
|
910 |
object); |
70d71228a
|
911 |
dump_stack(); |
064287807
|
912 |
} else |
249226847
|
913 914 |
object_err(s, page, object, "page slab pointer corrupt."); |
81819f0fc
|
915 916 |
goto fail; } |
3ec097421
|
917 918 |
/* Special debug activities for freeing objects */ |
8a38082d2
|
919 |
if (!PageSlubFrozen(page) && !page->freelist) |
3ec097421
|
920 921 922 923 924 |
remove_full(s, page); if (s->flags & SLAB_STORE_USER) set_track(s, object, TRACK_FREE, addr); trace(s, page, object, 0); init_object(s, object, 0); |
81819f0fc
|
925 |
return 1; |
3ec097421
|
926 |
|
81819f0fc
|
927 |
fail: |
249226847
|
928 |
slab_fix(s, "Object at 0x%p not freed", object); |
81819f0fc
|
929 930 |
return 0; } |
41ecc55b8
|
931 932 |
static int __init setup_slub_debug(char *str) { |
f0630fff5
|
933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 |
slub_debug = DEBUG_DEFAULT_FLAGS; if (*str++ != '=' || !*str) /* * No options specified. Switch on full debugging. */ goto out; if (*str == ',') /* * No options but restriction on slabs. This means full * debugging for slabs matching a pattern. */ goto check_slabs; slub_debug = 0; if (*str == '-') /* * Switch off all debugging measures. */ goto out; /* * Determine which debug features should be switched on */ |
064287807
|
957 |
for (; *str && *str != ','; str++) { |
f0630fff5
|
958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 |
switch (tolower(*str)) { case 'f': slub_debug |= SLAB_DEBUG_FREE; break; case 'z': slub_debug |= SLAB_RED_ZONE; break; case 'p': slub_debug |= SLAB_POISON; break; case 'u': slub_debug |= SLAB_STORE_USER; break; case 't': slub_debug |= SLAB_TRACE; break; default: printk(KERN_ERR "slub_debug option '%c' " |
064287807
|
976 977 |
"unknown. skipped ", *str); |
f0630fff5
|
978 |
} |
41ecc55b8
|
979 |
} |
f0630fff5
|
980 |
check_slabs: |
41ecc55b8
|
981 982 |
if (*str == ',') slub_debug_slabs = str + 1; |
f0630fff5
|
983 |
out: |
41ecc55b8
|
984 985 986 987 |
return 1; } __setup("slub_debug", setup_slub_debug); |
ba0268a8b
|
988 989 |
static unsigned long kmem_cache_flags(unsigned long objsize, unsigned long flags, const char *name, |
51cc50685
|
990 |
void (*ctor)(void *)) |
41ecc55b8
|
991 992 |
{ /* |
e153362a5
|
993 |
* Enable debugging if selected on the kernel commandline. |
41ecc55b8
|
994 |
*/ |
e153362a5
|
995 996 997 |
if (slub_debug && (!slub_debug_slabs || strncmp(slub_debug_slabs, name, strlen(slub_debug_slabs)) == 0)) flags |= slub_debug; |
ba0268a8b
|
998 999 |
return flags; |
41ecc55b8
|
1000 1001 |
} #else |
3ec097421
|
1002 1003 |
static inline void setup_object_debug(struct kmem_cache *s, struct page *page, void *object) {} |
41ecc55b8
|
1004 |
|
3ec097421
|
1005 |
static inline int alloc_debug_processing(struct kmem_cache *s, |
ce71e27c6
|
1006 |
struct page *page, void *object, unsigned long addr) { return 0; } |
41ecc55b8
|
1007 |
|
3ec097421
|
1008 |
static inline int free_debug_processing(struct kmem_cache *s, |
ce71e27c6
|
1009 |
struct page *page, void *object, unsigned long addr) { return 0; } |
41ecc55b8
|
1010 |
|
41ecc55b8
|
1011 1012 1013 1014 |
static inline int slab_pad_check(struct kmem_cache *s, struct page *page) { return 1; } static inline int check_object(struct kmem_cache *s, struct page *page, void *object, int active) { return 1; } |
3ec097421
|
1015 |
static inline void add_full(struct kmem_cache_node *n, struct page *page) {} |
ba0268a8b
|
1016 1017 |
static inline unsigned long kmem_cache_flags(unsigned long objsize, unsigned long flags, const char *name, |
51cc50685
|
1018 |
void (*ctor)(void *)) |
ba0268a8b
|
1019 1020 1021 |
{ return flags; } |
41ecc55b8
|
1022 |
#define slub_debug 0 |
0f389ec63
|
1023 1024 1025 |
static inline unsigned long slabs_node(struct kmem_cache *s, int node) { return 0; } |
205ab99dd
|
1026 1027 1028 1029 |
static inline void inc_slabs_node(struct kmem_cache *s, int node, int objects) {} static inline void dec_slabs_node(struct kmem_cache *s, int node, int objects) {} |
41ecc55b8
|
1030 |
#endif |
205ab99dd
|
1031 |
|
81819f0fc
|
1032 1033 1034 |
/* * Slab allocation and freeing */ |
65c3376aa
|
1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 |
static inline struct page *alloc_slab_page(gfp_t flags, int node, struct kmem_cache_order_objects oo) { int order = oo_order(oo); if (node == -1) return alloc_pages(flags, order); else return alloc_pages_node(node, flags, order); } |
81819f0fc
|
1045 1046 |
static struct page *allocate_slab(struct kmem_cache *s, gfp_t flags, int node) { |
064287807
|
1047 |
struct page *page; |
834f3d119
|
1048 |
struct kmem_cache_order_objects oo = s->oo; |
81819f0fc
|
1049 |
|
b7a49f0d4
|
1050 |
flags |= s->allocflags; |
e12ba74d8
|
1051 |
|
65c3376aa
|
1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 |
page = alloc_slab_page(flags | __GFP_NOWARN | __GFP_NORETRY, node, oo); if (unlikely(!page)) { oo = s->min; /* * Allocation may have failed due to fragmentation. * Try a lower order alloc if possible */ page = alloc_slab_page(flags, node, oo); if (!page) return NULL; |
81819f0fc
|
1063 |
|
65c3376aa
|
1064 1065 |
stat(get_cpu_slab(s, raw_smp_processor_id()), ORDER_FALLBACK); } |
834f3d119
|
1066 |
page->objects = oo_objects(oo); |
81819f0fc
|
1067 1068 1069 |
mod_zone_page_state(page_zone(page), (s->flags & SLAB_RECLAIM_ACCOUNT) ? NR_SLAB_RECLAIMABLE : NR_SLAB_UNRECLAIMABLE, |
65c3376aa
|
1070 |
1 << oo_order(oo)); |
81819f0fc
|
1071 1072 1073 1074 1075 1076 1077 |
return page; } static void setup_object(struct kmem_cache *s, struct page *page, void *object) { |
3ec097421
|
1078 |
setup_object_debug(s, page, object); |
4f1049345
|
1079 |
if (unlikely(s->ctor)) |
51cc50685
|
1080 |
s->ctor(object); |
81819f0fc
|
1081 1082 1083 1084 1085 |
} static struct page *new_slab(struct kmem_cache *s, gfp_t flags, int node) { struct page *page; |
81819f0fc
|
1086 |
void *start; |
81819f0fc
|
1087 1088 |
void *last; void *p; |
6cb062296
|
1089 |
BUG_ON(flags & GFP_SLAB_BUG_MASK); |
81819f0fc
|
1090 |
|
6cb062296
|
1091 1092 |
page = allocate_slab(s, flags & (GFP_RECLAIM_MASK | GFP_CONSTRAINT_MASK), node); |
81819f0fc
|
1093 1094 |
if (!page) goto out; |
205ab99dd
|
1095 |
inc_slabs_node(s, page_to_nid(page), page->objects); |
81819f0fc
|
1096 1097 1098 1099 |
page->slab = s; page->flags |= 1 << PG_slab; if (s->flags & (SLAB_DEBUG_FREE | SLAB_RED_ZONE | SLAB_POISON | SLAB_STORE_USER | SLAB_TRACE)) |
8a38082d2
|
1100 |
__SetPageSlubDebug(page); |
81819f0fc
|
1101 1102 |
start = page_address(page); |
81819f0fc
|
1103 1104 |
if (unlikely(s->flags & SLAB_POISON)) |
834f3d119
|
1105 |
memset(start, POISON_INUSE, PAGE_SIZE << compound_order(page)); |
81819f0fc
|
1106 1107 |
last = start; |
224a88be4
|
1108 |
for_each_object(p, s, start, page->objects) { |
81819f0fc
|
1109 1110 1111 1112 1113 |
setup_object(s, page, last); set_freepointer(s, last, p); last = p; } setup_object(s, page, last); |
a973e9dd1
|
1114 |
set_freepointer(s, last, NULL); |
81819f0fc
|
1115 1116 1117 1118 |
page->freelist = start; page->inuse = 0; out: |
81819f0fc
|
1119 1120 1121 1122 1123 |
return page; } static void __free_slab(struct kmem_cache *s, struct page *page) { |
834f3d119
|
1124 1125 |
int order = compound_order(page); int pages = 1 << order; |
81819f0fc
|
1126 |
|
8a38082d2
|
1127 |
if (unlikely(SLABDEBUG && PageSlubDebug(page))) { |
81819f0fc
|
1128 1129 1130 |
void *p; slab_pad_check(s, page); |
224a88be4
|
1131 1132 |
for_each_object(p, s, page_address(page), page->objects) |
81819f0fc
|
1133 |
check_object(s, page, p, 0); |
8a38082d2
|
1134 |
__ClearPageSlubDebug(page); |
81819f0fc
|
1135 1136 1137 1138 1139 |
} mod_zone_page_state(page_zone(page), (s->flags & SLAB_RECLAIM_ACCOUNT) ? NR_SLAB_RECLAIMABLE : NR_SLAB_UNRECLAIMABLE, |
064287807
|
1140 |
-pages); |
81819f0fc
|
1141 |
|
49bd5221c
|
1142 1143 |
__ClearPageSlab(page); reset_page_mapcount(page); |
834f3d119
|
1144 |
__free_pages(page, order); |
81819f0fc
|
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 |
} static void rcu_free_slab(struct rcu_head *h) { struct page *page; page = container_of((struct list_head *)h, struct page, lru); __free_slab(page->slab, page); } static void free_slab(struct kmem_cache *s, struct page *page) { if (unlikely(s->flags & SLAB_DESTROY_BY_RCU)) { /* * RCU free overloads the RCU head over the LRU */ struct rcu_head *head = (void *)&page->lru; call_rcu(head, rcu_free_slab); } else __free_slab(s, page); } static void discard_slab(struct kmem_cache *s, struct page *page) { |
205ab99dd
|
1170 |
dec_slabs_node(s, page_to_nid(page), page->objects); |
81819f0fc
|
1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 |
free_slab(s, page); } /* * Per slab locking using the pagelock */ static __always_inline void slab_lock(struct page *page) { bit_spin_lock(PG_locked, &page->flags); } static __always_inline void slab_unlock(struct page *page) { |
a76d35462
|
1184 |
__bit_spin_unlock(PG_locked, &page->flags); |
81819f0fc
|
1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 |
} static __always_inline int slab_trylock(struct page *page) { int rc = 1; rc = bit_spin_trylock(PG_locked, &page->flags); return rc; } /* * Management of partially allocated slabs */ |
7c2e132c5
|
1198 1199 |
static void add_partial(struct kmem_cache_node *n, struct page *page, int tail) |
81819f0fc
|
1200 |
{ |
e95eed571
|
1201 1202 |
spin_lock(&n->list_lock); n->nr_partial++; |
7c2e132c5
|
1203 1204 1205 1206 |
if (tail) list_add_tail(&page->lru, &n->partial); else list_add(&page->lru, &n->partial); |
81819f0fc
|
1207 1208 |
spin_unlock(&n->list_lock); } |
0121c619d
|
1209 |
static void remove_partial(struct kmem_cache *s, struct page *page) |
81819f0fc
|
1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 |
{ struct kmem_cache_node *n = get_node(s, page_to_nid(page)); spin_lock(&n->list_lock); list_del(&page->lru); n->nr_partial--; spin_unlock(&n->list_lock); } /* |
672bba3a4
|
1220 |
* Lock slab and remove from the partial list. |
81819f0fc
|
1221 |
* |
672bba3a4
|
1222 |
* Must hold list_lock. |
81819f0fc
|
1223 |
*/ |
0121c619d
|
1224 1225 |
static inline int lock_and_freeze_slab(struct kmem_cache_node *n, struct page *page) |
81819f0fc
|
1226 1227 1228 1229 |
{ if (slab_trylock(page)) { list_del(&page->lru); n->nr_partial--; |
8a38082d2
|
1230 |
__SetPageSlubFrozen(page); |
81819f0fc
|
1231 1232 1233 1234 1235 1236 |
return 1; } return 0; } /* |
672bba3a4
|
1237 |
* Try to allocate a partial slab from a specific node. |
81819f0fc
|
1238 1239 1240 1241 1242 1243 1244 1245 |
*/ static struct page *get_partial_node(struct kmem_cache_node *n) { struct page *page; /* * Racy check. If we mistakenly see no partial slabs then we * just allocate an empty slab. If we mistakenly try to get a |
672bba3a4
|
1246 1247 |
* partial slab and there is none available then get_partials() * will return NULL. |
81819f0fc
|
1248 1249 1250 1251 1252 1253 |
*/ if (!n || !n->nr_partial) return NULL; spin_lock(&n->list_lock); list_for_each_entry(page, &n->partial, lru) |
4b6f07504
|
1254 |
if (lock_and_freeze_slab(n, page)) |
81819f0fc
|
1255 1256 1257 1258 1259 1260 1261 1262 |
goto out; page = NULL; out: spin_unlock(&n->list_lock); return page; } /* |
672bba3a4
|
1263 |
* Get a page from somewhere. Search in increasing NUMA distances. |
81819f0fc
|
1264 1265 1266 1267 1268 |
*/ static struct page *get_any_partial(struct kmem_cache *s, gfp_t flags) { #ifdef CONFIG_NUMA struct zonelist *zonelist; |
dd1a239f6
|
1269 |
struct zoneref *z; |
54a6eb5c4
|
1270 1271 |
struct zone *zone; enum zone_type high_zoneidx = gfp_zone(flags); |
81819f0fc
|
1272 1273 1274 |
struct page *page; /* |
672bba3a4
|
1275 1276 1277 1278 |
* The defrag ratio allows a configuration of the tradeoffs between * inter node defragmentation and node local allocations. A lower * defrag_ratio increases the tendency to do local allocations * instead of attempting to obtain partial slabs from other nodes. |
81819f0fc
|
1279 |
* |
672bba3a4
|
1280 1281 1282 1283 |
* If the defrag_ratio is set to 0 then kmalloc() always * returns node local objects. If the ratio is higher then kmalloc() * may return off node objects because partial slabs are obtained * from other nodes and filled up. |
81819f0fc
|
1284 |
* |
6446faa2f
|
1285 |
* If /sys/kernel/slab/xx/defrag_ratio is set to 100 (which makes |
672bba3a4
|
1286 1287 1288 1289 1290 |
* defrag_ratio = 1000) then every (well almost) allocation will * first attempt to defrag slab caches on other nodes. This means * scanning over all nodes to look for partial slabs which may be * expensive if we do it every time we are trying to find a slab * with available objects. |
81819f0fc
|
1291 |
*/ |
9824601ea
|
1292 1293 |
if (!s->remote_node_defrag_ratio || get_cycles() % 1024 > s->remote_node_defrag_ratio) |
81819f0fc
|
1294 |
return NULL; |
0e88460da
|
1295 |
zonelist = node_zonelist(slab_node(current->mempolicy), flags); |
54a6eb5c4
|
1296 |
for_each_zone_zonelist(zone, z, zonelist, high_zoneidx) { |
81819f0fc
|
1297 |
struct kmem_cache_node *n; |
54a6eb5c4
|
1298 |
n = get_node(s, zone_to_nid(zone)); |
81819f0fc
|
1299 |
|
54a6eb5c4
|
1300 |
if (n && cpuset_zone_allowed_hardwall(zone, flags) && |
5595cffc8
|
1301 |
n->nr_partial > n->min_partial) { |
81819f0fc
|
1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 |
page = get_partial_node(n); if (page) return page; } } #endif return NULL; } /* * Get a partial page, lock it and return it. */ static struct page *get_partial(struct kmem_cache *s, gfp_t flags, int node) { struct page *page; int searchnode = (node == -1) ? numa_node_id() : node; page = get_partial_node(get_node(s, searchnode)); if (page || (flags & __GFP_THISNODE)) return page; return get_any_partial(s, flags); } /* * Move a page back to the lists. * * Must be called with the slab lock held. * * On exit the slab lock will have been dropped. */ |
7c2e132c5
|
1333 |
static void unfreeze_slab(struct kmem_cache *s, struct page *page, int tail) |
81819f0fc
|
1334 |
{ |
e95eed571
|
1335 |
struct kmem_cache_node *n = get_node(s, page_to_nid(page)); |
8ff12cfc0
|
1336 |
struct kmem_cache_cpu *c = get_cpu_slab(s, smp_processor_id()); |
e95eed571
|
1337 |
|
8a38082d2
|
1338 |
__ClearPageSlubFrozen(page); |
81819f0fc
|
1339 |
if (page->inuse) { |
e95eed571
|
1340 |
|
a973e9dd1
|
1341 |
if (page->freelist) { |
7c2e132c5
|
1342 |
add_partial(n, page, tail); |
8ff12cfc0
|
1343 1344 1345 |
stat(c, tail ? DEACTIVATE_TO_TAIL : DEACTIVATE_TO_HEAD); } else { stat(c, DEACTIVATE_FULL); |
8a38082d2
|
1346 1347 |
if (SLABDEBUG && PageSlubDebug(page) && (s->flags & SLAB_STORE_USER)) |
8ff12cfc0
|
1348 1349 |
add_full(n, page); } |
81819f0fc
|
1350 1351 |
slab_unlock(page); } else { |
8ff12cfc0
|
1352 |
stat(c, DEACTIVATE_EMPTY); |
5595cffc8
|
1353 |
if (n->nr_partial < n->min_partial) { |
e95eed571
|
1354 |
/* |
672bba3a4
|
1355 1356 1357 |
* Adding an empty slab to the partial slabs in order * to avoid page allocator overhead. This slab needs * to come after the other slabs with objects in |
6446faa2f
|
1358 1359 1360 |
* so that the others get filled first. That way the * size of the partial list stays small. * |
0121c619d
|
1361 1362 |
* kmem_cache_shrink can reclaim any empty slabs from * the partial list. |
e95eed571
|
1363 |
*/ |
7c2e132c5
|
1364 |
add_partial(n, page, 1); |
e95eed571
|
1365 1366 1367 |
slab_unlock(page); } else { slab_unlock(page); |
8ff12cfc0
|
1368 |
stat(get_cpu_slab(s, raw_smp_processor_id()), FREE_SLAB); |
e95eed571
|
1369 1370 |
discard_slab(s, page); } |
81819f0fc
|
1371 1372 1373 1374 1375 1376 |
} } /* * Remove the cpu slab */ |
dfb4f0960
|
1377 |
static void deactivate_slab(struct kmem_cache *s, struct kmem_cache_cpu *c) |
81819f0fc
|
1378 |
{ |
dfb4f0960
|
1379 |
struct page *page = c->page; |
7c2e132c5
|
1380 |
int tail = 1; |
8ff12cfc0
|
1381 |
|
b773ad736
|
1382 |
if (page->freelist) |
8ff12cfc0
|
1383 |
stat(c, DEACTIVATE_REMOTE_FREES); |
894b8788d
|
1384 |
/* |
6446faa2f
|
1385 |
* Merge cpu freelist into slab freelist. Typically we get here |
894b8788d
|
1386 1387 1388 |
* because both freelists are empty. So this is unlikely * to occur. */ |
a973e9dd1
|
1389 |
while (unlikely(c->freelist)) { |
894b8788d
|
1390 |
void **object; |
7c2e132c5
|
1391 |
tail = 0; /* Hot objects. Put the slab first */ |
894b8788d
|
1392 |
/* Retrieve object from cpu_freelist */ |
dfb4f0960
|
1393 |
object = c->freelist; |
b3fba8da6
|
1394 |
c->freelist = c->freelist[c->offset]; |
894b8788d
|
1395 1396 |
/* And put onto the regular freelist */ |
b3fba8da6
|
1397 |
object[c->offset] = page->freelist; |
894b8788d
|
1398 1399 1400 |
page->freelist = object; page->inuse--; } |
dfb4f0960
|
1401 |
c->page = NULL; |
7c2e132c5
|
1402 |
unfreeze_slab(s, page, tail); |
81819f0fc
|
1403 |
} |
dfb4f0960
|
1404 |
static inline void flush_slab(struct kmem_cache *s, struct kmem_cache_cpu *c) |
81819f0fc
|
1405 |
{ |
8ff12cfc0
|
1406 |
stat(c, CPUSLAB_FLUSH); |
dfb4f0960
|
1407 1408 |
slab_lock(c->page); deactivate_slab(s, c); |
81819f0fc
|
1409 1410 1411 1412 |
} /* * Flush cpu slab. |
6446faa2f
|
1413 |
* |
81819f0fc
|
1414 1415 |
* Called from IPI handler with interrupts disabled. */ |
0c7100132
|
1416 |
static inline void __flush_cpu_slab(struct kmem_cache *s, int cpu) |
81819f0fc
|
1417 |
{ |
dfb4f0960
|
1418 |
struct kmem_cache_cpu *c = get_cpu_slab(s, cpu); |
81819f0fc
|
1419 |
|
dfb4f0960
|
1420 1421 |
if (likely(c && c->page)) flush_slab(s, c); |
81819f0fc
|
1422 1423 1424 1425 1426 |
} static void flush_cpu_slab(void *d) { struct kmem_cache *s = d; |
81819f0fc
|
1427 |
|
dfb4f0960
|
1428 |
__flush_cpu_slab(s, smp_processor_id()); |
81819f0fc
|
1429 1430 1431 1432 |
} static void flush_all(struct kmem_cache *s) { |
15c8b6c1a
|
1433 |
on_each_cpu(flush_cpu_slab, s, 1); |
81819f0fc
|
1434 1435 1436 |
} /* |
dfb4f0960
|
1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 |
* Check if the objects in a per cpu structure fit numa * locality expectations. */ static inline int node_match(struct kmem_cache_cpu *c, int node) { #ifdef CONFIG_NUMA if (node != -1 && c->node != node) return 0; #endif return 1; } /* |
894b8788d
|
1450 1451 1452 1453 |
* Slow path. The lockless freelist is empty or we need to perform * debugging duties. * * Interrupts are disabled. |
81819f0fc
|
1454 |
* |
894b8788d
|
1455 1456 1457 |
* Processing is still very fast if new objects have been freed to the * regular freelist. In that case we simply take over the regular freelist * as the lockless freelist and zap the regular freelist. |
81819f0fc
|
1458 |
* |
894b8788d
|
1459 1460 1461 |
* If that is not working then we fall back to the partial lists. We take the * first element of the freelist as the object to allocate now and move the * rest of the freelist to the lockless freelist. |
81819f0fc
|
1462 |
* |
894b8788d
|
1463 |
* And if we were unable to get a new slab from the partial slab lists then |
6446faa2f
|
1464 1465 |
* we need to allocate a new slab. This is the slowest path since it involves * a call to the page allocator and the setup of a new slab. |
81819f0fc
|
1466 |
*/ |
ce71e27c6
|
1467 1468 |
static void *__slab_alloc(struct kmem_cache *s, gfp_t gfpflags, int node, unsigned long addr, struct kmem_cache_cpu *c) |
81819f0fc
|
1469 |
{ |
81819f0fc
|
1470 |
void **object; |
dfb4f0960
|
1471 |
struct page *new; |
81819f0fc
|
1472 |
|
e72e9c23e
|
1473 1474 |
/* We handle __GFP_ZERO in the caller */ gfpflags &= ~__GFP_ZERO; |
dfb4f0960
|
1475 |
if (!c->page) |
81819f0fc
|
1476 |
goto new_slab; |
dfb4f0960
|
1477 1478 |
slab_lock(c->page); if (unlikely(!node_match(c, node))) |
81819f0fc
|
1479 |
goto another_slab; |
6446faa2f
|
1480 |
|
8ff12cfc0
|
1481 |
stat(c, ALLOC_REFILL); |
6446faa2f
|
1482 |
|
894b8788d
|
1483 |
load_freelist: |
dfb4f0960
|
1484 |
object = c->page->freelist; |
a973e9dd1
|
1485 |
if (unlikely(!object)) |
81819f0fc
|
1486 |
goto another_slab; |
8a38082d2
|
1487 |
if (unlikely(SLABDEBUG && PageSlubDebug(c->page))) |
81819f0fc
|
1488 |
goto debug; |
b3fba8da6
|
1489 |
c->freelist = object[c->offset]; |
39b264641
|
1490 |
c->page->inuse = c->page->objects; |
a973e9dd1
|
1491 |
c->page->freelist = NULL; |
dfb4f0960
|
1492 |
c->node = page_to_nid(c->page); |
1f84260c8
|
1493 |
unlock_out: |
dfb4f0960
|
1494 |
slab_unlock(c->page); |
8ff12cfc0
|
1495 |
stat(c, ALLOC_SLOWPATH); |
81819f0fc
|
1496 1497 1498 |
return object; another_slab: |
dfb4f0960
|
1499 |
deactivate_slab(s, c); |
81819f0fc
|
1500 1501 |
new_slab: |
dfb4f0960
|
1502 1503 1504 |
new = get_partial(s, gfpflags, node); if (new) { c->page = new; |
8ff12cfc0
|
1505 |
stat(c, ALLOC_FROM_PARTIAL); |
894b8788d
|
1506 |
goto load_freelist; |
81819f0fc
|
1507 |
} |
b811c202a
|
1508 1509 |
if (gfpflags & __GFP_WAIT) local_irq_enable(); |
dfb4f0960
|
1510 |
new = new_slab(s, gfpflags, node); |
b811c202a
|
1511 1512 1513 |
if (gfpflags & __GFP_WAIT) local_irq_disable(); |
dfb4f0960
|
1514 1515 |
if (new) { c = get_cpu_slab(s, smp_processor_id()); |
8ff12cfc0
|
1516 |
stat(c, ALLOC_SLAB); |
05aa34503
|
1517 |
if (c->page) |
dfb4f0960
|
1518 |
flush_slab(s, c); |
dfb4f0960
|
1519 |
slab_lock(new); |
8a38082d2
|
1520 |
__SetPageSlubFrozen(new); |
dfb4f0960
|
1521 |
c->page = new; |
4b6f07504
|
1522 |
goto load_freelist; |
81819f0fc
|
1523 |
} |
71c7a06ff
|
1524 |
return NULL; |
81819f0fc
|
1525 |
debug: |
dfb4f0960
|
1526 |
if (!alloc_debug_processing(s, c->page, object, addr)) |
81819f0fc
|
1527 |
goto another_slab; |
894b8788d
|
1528 |
|
dfb4f0960
|
1529 |
c->page->inuse++; |
b3fba8da6
|
1530 |
c->page->freelist = object[c->offset]; |
ee3c72a14
|
1531 |
c->node = -1; |
1f84260c8
|
1532 |
goto unlock_out; |
894b8788d
|
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 |
} /* * Inlined fastpath so that allocation functions (kmalloc, kmem_cache_alloc) * have the fastpath folded into their functions. So no function call * overhead for requests that can be satisfied on the fastpath. * * The fastpath works by first checking if the lockless freelist can be used. * If not then __slab_alloc is called for slow processing. * * Otherwise we can simply pick the next object from the lockless free list. */ |
064287807
|
1545 |
static __always_inline void *slab_alloc(struct kmem_cache *s, |
ce71e27c6
|
1546 |
gfp_t gfpflags, int node, unsigned long addr) |
894b8788d
|
1547 |
{ |
894b8788d
|
1548 |
void **object; |
dfb4f0960
|
1549 |
struct kmem_cache_cpu *c; |
1f84260c8
|
1550 |
unsigned long flags; |
bdb219285
|
1551 |
unsigned int objsize; |
1f84260c8
|
1552 |
|
89124d706
|
1553 |
might_sleep_if(gfpflags & __GFP_WAIT); |
3c506efd7
|
1554 |
|
773ff60e8
|
1555 1556 |
if (should_failslab(s->objsize, gfpflags)) return NULL; |
1f84260c8
|
1557 |
|
894b8788d
|
1558 |
local_irq_save(flags); |
dfb4f0960
|
1559 |
c = get_cpu_slab(s, smp_processor_id()); |
bdb219285
|
1560 |
objsize = c->objsize; |
a973e9dd1
|
1561 |
if (unlikely(!c->freelist || !node_match(c, node))) |
894b8788d
|
1562 |
|
dfb4f0960
|
1563 |
object = __slab_alloc(s, gfpflags, node, addr, c); |
894b8788d
|
1564 1565 |
else { |
dfb4f0960
|
1566 |
object = c->freelist; |
b3fba8da6
|
1567 |
c->freelist = object[c->offset]; |
8ff12cfc0
|
1568 |
stat(c, ALLOC_FASTPATH); |
894b8788d
|
1569 1570 |
} local_irq_restore(flags); |
d07dbea46
|
1571 1572 |
if (unlikely((gfpflags & __GFP_ZERO) && object)) |
bdb219285
|
1573 |
memset(object, 0, objsize); |
d07dbea46
|
1574 |
|
894b8788d
|
1575 |
return object; |
81819f0fc
|
1576 1577 1578 1579 |
} void *kmem_cache_alloc(struct kmem_cache *s, gfp_t gfpflags) { |
ce71e27c6
|
1580 |
return slab_alloc(s, gfpflags, -1, _RET_IP_); |
81819f0fc
|
1581 1582 1583 1584 1585 1586 |
} EXPORT_SYMBOL(kmem_cache_alloc); #ifdef CONFIG_NUMA void *kmem_cache_alloc_node(struct kmem_cache *s, gfp_t gfpflags, int node) { |
ce71e27c6
|
1587 |
return slab_alloc(s, gfpflags, node, _RET_IP_); |
81819f0fc
|
1588 1589 1590 1591 1592 |
} EXPORT_SYMBOL(kmem_cache_alloc_node); #endif /* |
894b8788d
|
1593 1594 |
* Slow patch handling. This may still be called frequently since objects * have a longer lifetime than the cpu slabs in most processing loads. |
81819f0fc
|
1595 |
* |
894b8788d
|
1596 1597 1598 |
* So we still attempt to reduce cache line usage. Just take the slab * lock and free the item. If there is no additional partial page * handling required then we can return immediately. |
81819f0fc
|
1599 |
*/ |
894b8788d
|
1600 |
static void __slab_free(struct kmem_cache *s, struct page *page, |
ce71e27c6
|
1601 |
void *x, unsigned long addr, unsigned int offset) |
81819f0fc
|
1602 1603 1604 |
{ void *prior; void **object = (void *)x; |
8ff12cfc0
|
1605 |
struct kmem_cache_cpu *c; |
81819f0fc
|
1606 |
|
8ff12cfc0
|
1607 1608 |
c = get_cpu_slab(s, raw_smp_processor_id()); stat(c, FREE_SLOWPATH); |
81819f0fc
|
1609 |
slab_lock(page); |
8a38082d2
|
1610 |
if (unlikely(SLABDEBUG && PageSlubDebug(page))) |
81819f0fc
|
1611 |
goto debug; |
6446faa2f
|
1612 |
|
81819f0fc
|
1613 |
checks_ok: |
b3fba8da6
|
1614 |
prior = object[offset] = page->freelist; |
81819f0fc
|
1615 1616 |
page->freelist = object; page->inuse--; |
8a38082d2
|
1617 |
if (unlikely(PageSlubFrozen(page))) { |
8ff12cfc0
|
1618 |
stat(c, FREE_FROZEN); |
81819f0fc
|
1619 |
goto out_unlock; |
8ff12cfc0
|
1620 |
} |
81819f0fc
|
1621 1622 1623 1624 1625 |
if (unlikely(!page->inuse)) goto slab_empty; /* |
6446faa2f
|
1626 |
* Objects left in the slab. If it was not on the partial list before |
81819f0fc
|
1627 1628 |
* then add it. */ |
a973e9dd1
|
1629 |
if (unlikely(!prior)) { |
7c2e132c5
|
1630 |
add_partial(get_node(s, page_to_nid(page)), page, 1); |
8ff12cfc0
|
1631 1632 |
stat(c, FREE_ADD_PARTIAL); } |
81819f0fc
|
1633 1634 1635 |
out_unlock: slab_unlock(page); |
81819f0fc
|
1636 1637 1638 |
return; slab_empty: |
a973e9dd1
|
1639 |
if (prior) { |
81819f0fc
|
1640 |
/* |
672bba3a4
|
1641 |
* Slab still on the partial list. |
81819f0fc
|
1642 1643 |
*/ remove_partial(s, page); |
8ff12cfc0
|
1644 1645 |
stat(c, FREE_REMOVE_PARTIAL); } |
81819f0fc
|
1646 |
slab_unlock(page); |
8ff12cfc0
|
1647 |
stat(c, FREE_SLAB); |
81819f0fc
|
1648 |
discard_slab(s, page); |
81819f0fc
|
1649 1650 1651 |
return; debug: |
3ec097421
|
1652 |
if (!free_debug_processing(s, page, x, addr)) |
77c5e2d01
|
1653 |
goto out_unlock; |
77c5e2d01
|
1654 |
goto checks_ok; |
81819f0fc
|
1655 |
} |
894b8788d
|
1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 |
/* * Fastpath with forced inlining to produce a kfree and kmem_cache_free that * can perform fastpath freeing without additional function calls. * * The fastpath is only possible if we are freeing to the current cpu slab * of this processor. This typically the case if we have just allocated * the item before. * * If fastpath is not possible then fall back to __slab_free where we deal * with all sorts of special processing. */ |
064287807
|
1667 |
static __always_inline void slab_free(struct kmem_cache *s, |
ce71e27c6
|
1668 |
struct page *page, void *x, unsigned long addr) |
894b8788d
|
1669 1670 |
{ void **object = (void *)x; |
dfb4f0960
|
1671 |
struct kmem_cache_cpu *c; |
1f84260c8
|
1672 |
unsigned long flags; |
894b8788d
|
1673 |
local_irq_save(flags); |
dfb4f0960
|
1674 |
c = get_cpu_slab(s, smp_processor_id()); |
27d9e4e94
|
1675 |
debug_check_no_locks_freed(object, c->objsize); |
3ac7fe5a4
|
1676 1677 |
if (!(s->flags & SLAB_DEBUG_OBJECTS)) debug_check_no_obj_freed(object, s->objsize); |
ee3c72a14
|
1678 |
if (likely(page == c->page && c->node >= 0)) { |
b3fba8da6
|
1679 |
object[c->offset] = c->freelist; |
dfb4f0960
|
1680 |
c->freelist = object; |
8ff12cfc0
|
1681 |
stat(c, FREE_FASTPATH); |
894b8788d
|
1682 |
} else |
b3fba8da6
|
1683 |
__slab_free(s, page, x, addr, c->offset); |
894b8788d
|
1684 1685 1686 |
local_irq_restore(flags); } |
81819f0fc
|
1687 1688 |
void kmem_cache_free(struct kmem_cache *s, void *x) { |
77c5e2d01
|
1689 |
struct page *page; |
81819f0fc
|
1690 |
|
b49af68ff
|
1691 |
page = virt_to_head_page(x); |
81819f0fc
|
1692 |
|
ce71e27c6
|
1693 |
slab_free(s, page, x, _RET_IP_); |
81819f0fc
|
1694 1695 |
} EXPORT_SYMBOL(kmem_cache_free); |
e9beef181
|
1696 |
/* Figure out on which slab page the object resides */ |
81819f0fc
|
1697 1698 |
static struct page *get_object_page(const void *x) { |
b49af68ff
|
1699 |
struct page *page = virt_to_head_page(x); |
81819f0fc
|
1700 1701 1702 1703 1704 1705 1706 1707 |
if (!PageSlab(page)) return NULL; return page; } /* |
672bba3a4
|
1708 1709 1710 1711 |
* Object placement in a slab is made very easy because we always start at * offset 0. If we tune the size of the object to the alignment then we can * get the required alignment by putting one properly sized object after * another. |
81819f0fc
|
1712 1713 1714 1715 |
* * Notice that the allocation order determines the sizes of the per cpu * caches. Each processor has always one slab available for allocations. * Increasing the allocation order reduces the number of times that slabs |
672bba3a4
|
1716 |
* must be moved on and off the partial lists and is therefore a factor in |
81819f0fc
|
1717 |
* locking overhead. |
81819f0fc
|
1718 1719 1720 1721 1722 1723 1724 1725 1726 |
*/ /* * Mininum / Maximum order of slab pages. This influences locking overhead * and slab fragmentation. A higher order reduces the number of partial slabs * and increases the number of allocations possible without having to * take the list_lock. */ static int slub_min_order; |
114e9e89e
|
1727 |
static int slub_max_order = PAGE_ALLOC_COSTLY_ORDER; |
9b2cd506e
|
1728 |
static int slub_min_objects; |
81819f0fc
|
1729 1730 1731 |
/* * Merge control. If this is set then no merging of slab caches will occur. |
672bba3a4
|
1732 |
* (Could be removed. This was introduced to pacify the merge skeptics.) |
81819f0fc
|
1733 1734 1735 1736 |
*/ static int slub_nomerge; /* |
81819f0fc
|
1737 1738 |
* Calculate the order of allocation given an slab object size. * |
672bba3a4
|
1739 1740 1741 1742 |
* The order of allocation has significant impact on performance and other * system components. Generally order 0 allocations should be preferred since * order 0 does not cause fragmentation in the page allocator. Larger objects * be problematic to put into order 0 slabs because there may be too much |
c124f5b54
|
1743 |
* unused space left. We go to a higher order if more than 1/16th of the slab |
672bba3a4
|
1744 1745 1746 1747 1748 1749 |
* would be wasted. * * In order to reach satisfactory performance we must ensure that a minimum * number of objects is in one slab. Otherwise we may generate too much * activity on the partial lists which requires taking the list_lock. This is * less a concern for large slabs though which are rarely used. |
81819f0fc
|
1750 |
* |
672bba3a4
|
1751 1752 1753 1754 |
* slub_max_order specifies the order where we begin to stop considering the * number of objects in a slab as critical. If we reach slub_max_order then * we try to keep the page order as low as possible. So we accept more waste * of space in favor of a small page order. |
81819f0fc
|
1755 |
* |
672bba3a4
|
1756 1757 1758 1759 |
* Higher order allocations also allow the placement of more objects in a * slab and thereby reduce object handling overhead. If the user has * requested a higher mininum order then we start with that one instead of * the smallest order which will fit the object. |
81819f0fc
|
1760 |
*/ |
5e6d444ea
|
1761 1762 |
static inline int slab_order(int size, int min_objects, int max_order, int fract_leftover) |
81819f0fc
|
1763 1764 1765 |
{ int order; int rem; |
6300ea750
|
1766 |
int min_order = slub_min_order; |
81819f0fc
|
1767 |
|
210b5c061
|
1768 1769 |
if ((PAGE_SIZE << min_order) / size > MAX_OBJS_PER_PAGE) return get_order(size * MAX_OBJS_PER_PAGE) - 1; |
39b264641
|
1770 |
|
6300ea750
|
1771 |
for (order = max(min_order, |
5e6d444ea
|
1772 1773 |
fls(min_objects * size - 1) - PAGE_SHIFT); order <= max_order; order++) { |
81819f0fc
|
1774 |
|
5e6d444ea
|
1775 |
unsigned long slab_size = PAGE_SIZE << order; |
81819f0fc
|
1776 |
|
5e6d444ea
|
1777 |
if (slab_size < min_objects * size) |
81819f0fc
|
1778 1779 1780 |
continue; rem = slab_size % size; |
5e6d444ea
|
1781 |
if (rem <= slab_size / fract_leftover) |
81819f0fc
|
1782 1783 1784 |
break; } |
672bba3a4
|
1785 |
|
81819f0fc
|
1786 1787 |
return order; } |
5e6d444ea
|
1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 |
static inline int calculate_order(int size) { int order; int min_objects; int fraction; /* * Attempt to find best configuration for a slab. This * works by first attempting to generate a layout with * the best configuration and backing off gradually. * * First we reduce the acceptable waste in a slab. Then * we reduce the minimum objects required in a slab. */ min_objects = slub_min_objects; |
9b2cd506e
|
1803 1804 |
if (!min_objects) min_objects = 4 * (fls(nr_cpu_ids) + 1); |
5e6d444ea
|
1805 |
while (min_objects > 1) { |
c124f5b54
|
1806 |
fraction = 16; |
5e6d444ea
|
1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823 1824 1825 1826 1827 1828 1829 1830 1831 1832 |
while (fraction >= 4) { order = slab_order(size, min_objects, slub_max_order, fraction); if (order <= slub_max_order) return order; fraction /= 2; } min_objects /= 2; } /* * We were unable to place multiple objects in a slab. Now * lets see if we can place a single object there. */ order = slab_order(size, 1, slub_max_order, 1); if (order <= slub_max_order) return order; /* * Doh this slab cannot be placed using slub_max_order. */ order = slab_order(size, 1, MAX_ORDER, 1); if (order <= MAX_ORDER) return order; return -ENOSYS; } |
81819f0fc
|
1833 |
/* |
672bba3a4
|
1834 |
* Figure out what the alignment of the objects will be. |
81819f0fc
|
1835 1836 1837 1838 1839 |
*/ static unsigned long calculate_alignment(unsigned long flags, unsigned long align, unsigned long size) { /* |
6446faa2f
|
1840 1841 |
* If the user wants hardware cache aligned objects then follow that * suggestion if the object is sufficiently large. |
81819f0fc
|
1842 |
* |
6446faa2f
|
1843 1844 |
* The hardware cache alignment cannot override the specified * alignment though. If that is greater then use it. |
81819f0fc
|
1845 |
*/ |
b62103867
|
1846 1847 1848 1849 1850 1851 |
if (flags & SLAB_HWCACHE_ALIGN) { unsigned long ralign = cache_line_size(); while (size <= ralign / 2) ralign /= 2; align = max(align, ralign); } |
81819f0fc
|
1852 1853 |
if (align < ARCH_SLAB_MINALIGN) |
b62103867
|
1854 |
align = ARCH_SLAB_MINALIGN; |
81819f0fc
|
1855 1856 1857 |
return ALIGN(align, sizeof(void *)); } |
dfb4f0960
|
1858 1859 1860 1861 |
static void init_kmem_cache_cpu(struct kmem_cache *s, struct kmem_cache_cpu *c) { c->page = NULL; |
a973e9dd1
|
1862 |
c->freelist = NULL; |
dfb4f0960
|
1863 |
c->node = 0; |
42a9fdbb1
|
1864 1865 |
c->offset = s->offset / sizeof(void *); c->objsize = s->objsize; |
62f75532b
|
1866 1867 1868 |
#ifdef CONFIG_SLUB_STATS memset(c->stat, 0, NR_SLUB_STAT_ITEMS * sizeof(unsigned)); #endif |
dfb4f0960
|
1869 |
} |
5595cffc8
|
1870 1871 |
static void init_kmem_cache_node(struct kmem_cache_node *n, struct kmem_cache *s) |
81819f0fc
|
1872 1873 |
{ n->nr_partial = 0; |
5595cffc8
|
1874 1875 1876 1877 1878 1879 1880 1881 1882 1883 |
/* * The larger the object size is, the more pages we want on the partial * list to avoid pounding the page allocator excessively. */ n->min_partial = ilog2(s->size); if (n->min_partial < MIN_PARTIAL) n->min_partial = MIN_PARTIAL; else if (n->min_partial > MAX_PARTIAL) n->min_partial = MAX_PARTIAL; |
81819f0fc
|
1884 1885 |
spin_lock_init(&n->list_lock); INIT_LIST_HEAD(&n->partial); |
8ab1372fa
|
1886 |
#ifdef CONFIG_SLUB_DEBUG |
0f389ec63
|
1887 |
atomic_long_set(&n->nr_slabs, 0); |
02b71b701
|
1888 |
atomic_long_set(&n->total_objects, 0); |
643b11384
|
1889 |
INIT_LIST_HEAD(&n->full); |
8ab1372fa
|
1890 |
#endif |
81819f0fc
|
1891 |
} |
4c93c355d
|
1892 1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 |
#ifdef CONFIG_SMP /* * Per cpu array for per cpu structures. * * The per cpu array places all kmem_cache_cpu structures from one processor * close together meaning that it becomes possible that multiple per cpu * structures are contained in one cacheline. This may be particularly * beneficial for the kmalloc caches. * * A desktop system typically has around 60-80 slabs. With 100 here we are * likely able to get per cpu structures for all caches from the array defined * here. We must be able to cover all kmalloc caches during bootstrap. * * If the per cpu array is exhausted then fall back to kmalloc * of individual cachelines. No sharing is possible then. */ #define NR_KMEM_CACHE_CPU 100 static DEFINE_PER_CPU(struct kmem_cache_cpu, kmem_cache_cpu)[NR_KMEM_CACHE_CPU]; static DEFINE_PER_CPU(struct kmem_cache_cpu *, kmem_cache_cpu_free); |
174596a0b
|
1914 |
static DECLARE_BITMAP(kmem_cach_cpu_free_init_once, CONFIG_NR_CPUS); |
4c93c355d
|
1915 1916 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958 1959 1960 1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 |
static struct kmem_cache_cpu *alloc_kmem_cache_cpu(struct kmem_cache *s, int cpu, gfp_t flags) { struct kmem_cache_cpu *c = per_cpu(kmem_cache_cpu_free, cpu); if (c) per_cpu(kmem_cache_cpu_free, cpu) = (void *)c->freelist; else { /* Table overflow: So allocate ourselves */ c = kmalloc_node( ALIGN(sizeof(struct kmem_cache_cpu), cache_line_size()), flags, cpu_to_node(cpu)); if (!c) return NULL; } init_kmem_cache_cpu(s, c); return c; } static void free_kmem_cache_cpu(struct kmem_cache_cpu *c, int cpu) { if (c < per_cpu(kmem_cache_cpu, cpu) || c > per_cpu(kmem_cache_cpu, cpu) + NR_KMEM_CACHE_CPU) { kfree(c); return; } c->freelist = (void *)per_cpu(kmem_cache_cpu_free, cpu); per_cpu(kmem_cache_cpu_free, cpu) = c; } static void free_kmem_cache_cpus(struct kmem_cache *s) { int cpu; for_each_online_cpu(cpu) { struct kmem_cache_cpu *c = get_cpu_slab(s, cpu); if (c) { s->cpu_slab[cpu] = NULL; free_kmem_cache_cpu(c, cpu); } } } static int alloc_kmem_cache_cpus(struct kmem_cache *s, gfp_t flags) { int cpu; for_each_online_cpu(cpu) { struct kmem_cache_cpu *c = get_cpu_slab(s, cpu); if (c) continue; c = alloc_kmem_cache_cpu(s, cpu, flags); if (!c) { free_kmem_cache_cpus(s); return 0; } s->cpu_slab[cpu] = c; } return 1; } /* * Initialize the per cpu array. */ static void init_alloc_cpu_cpu(int cpu) { int i; |
174596a0b
|
1988 |
if (cpumask_test_cpu(cpu, to_cpumask(kmem_cach_cpu_free_init_once))) |
4c93c355d
|
1989 1990 1991 1992 |
return; for (i = NR_KMEM_CACHE_CPU - 1; i >= 0; i--) free_kmem_cache_cpu(&per_cpu(kmem_cache_cpu, cpu)[i], cpu); |
174596a0b
|
1993 |
cpumask_set_cpu(cpu, to_cpumask(kmem_cach_cpu_free_init_once)); |
4c93c355d
|
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 |
} static void __init init_alloc_cpu(void) { int cpu; for_each_online_cpu(cpu) init_alloc_cpu_cpu(cpu); } #else static inline void free_kmem_cache_cpus(struct kmem_cache *s) {} static inline void init_alloc_cpu(void) {} static inline int alloc_kmem_cache_cpus(struct kmem_cache *s, gfp_t flags) { init_kmem_cache_cpu(s, &s->cpu_slab); return 1; } #endif |
81819f0fc
|
2014 2015 2016 2017 2018 2019 2020 |
#ifdef CONFIG_NUMA /* * No kmalloc_node yet so do it by hand. We know that this is the first * slab on the node for this slabcache. There are no concurrent accesses * possible. * * Note that this function only works on the kmalloc_node_cache |
4c93c355d
|
2021 2022 |
* when allocating for the kmalloc_node_cache. This is used for bootstrapping * memory on a fresh node that has no slab structures yet. |
81819f0fc
|
2023 |
*/ |
0094de92a
|
2024 |
static void early_kmem_cache_node_alloc(gfp_t gfpflags, int node) |
81819f0fc
|
2025 2026 2027 |
{ struct page *page; struct kmem_cache_node *n; |
ba84c73c7
|
2028 |
unsigned long flags; |
81819f0fc
|
2029 2030 |
BUG_ON(kmalloc_caches->size < sizeof(struct kmem_cache_node)); |
a2f92ee7e
|
2031 |
page = new_slab(kmalloc_caches, gfpflags, node); |
81819f0fc
|
2032 2033 |
BUG_ON(!page); |
a2f92ee7e
|
2034 2035 2036 2037 2038 2039 2040 2041 |
if (page_to_nid(page) != node) { printk(KERN_ERR "SLUB: Unable to allocate memory from " "node %d ", node); printk(KERN_ERR "SLUB: Allocating a useless per node structure " "in order to be able to continue "); } |
81819f0fc
|
2042 2043 2044 2045 2046 |
n = page->freelist; BUG_ON(!n); page->freelist = get_freepointer(kmalloc_caches, n); page->inuse++; kmalloc_caches->node[node] = n; |
8ab1372fa
|
2047 |
#ifdef CONFIG_SLUB_DEBUG |
d45f39cb0
|
2048 2049 |
init_object(kmalloc_caches, n, 1); init_tracking(kmalloc_caches, n); |
8ab1372fa
|
2050 |
#endif |
5595cffc8
|
2051 |
init_kmem_cache_node(n, kmalloc_caches); |
205ab99dd
|
2052 |
inc_slabs_node(kmalloc_caches, node, page->objects); |
6446faa2f
|
2053 |
|
ba84c73c7
|
2054 2055 2056 2057 2058 2059 |
/* * lockdep requires consistent irq usage for each lock * so even though there cannot be a race this early in * the boot sequence, we still disable irqs. */ local_irq_save(flags); |
7c2e132c5
|
2060 |
add_partial(n, page, 0); |
ba84c73c7
|
2061 |
local_irq_restore(flags); |
81819f0fc
|
2062 2063 2064 2065 2066 |
} static void free_kmem_cache_nodes(struct kmem_cache *s) { int node; |
f64dc58c5
|
2067 |
for_each_node_state(node, N_NORMAL_MEMORY) { |
81819f0fc
|
2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 |
struct kmem_cache_node *n = s->node[node]; if (n && n != &s->local_node) kmem_cache_free(kmalloc_caches, n); s->node[node] = NULL; } } static int init_kmem_cache_nodes(struct kmem_cache *s, gfp_t gfpflags) { int node; int local_node; if (slab_state >= UP) local_node = page_to_nid(virt_to_page(s)); else local_node = 0; |
f64dc58c5
|
2084 |
for_each_node_state(node, N_NORMAL_MEMORY) { |
81819f0fc
|
2085 2086 2087 2088 2089 2090 |
struct kmem_cache_node *n; if (local_node == node) n = &s->local_node; else { if (slab_state == DOWN) { |
0094de92a
|
2091 |
early_kmem_cache_node_alloc(gfpflags, node); |
81819f0fc
|
2092 2093 2094 2095 2096 2097 2098 2099 2100 2101 2102 2103 |
continue; } n = kmem_cache_alloc_node(kmalloc_caches, gfpflags, node); if (!n) { free_kmem_cache_nodes(s); return 0; } } s->node[node] = n; |
5595cffc8
|
2104 |
init_kmem_cache_node(n, s); |
81819f0fc
|
2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 |
} return 1; } #else static void free_kmem_cache_nodes(struct kmem_cache *s) { } static int init_kmem_cache_nodes(struct kmem_cache *s, gfp_t gfpflags) { |
5595cffc8
|
2115 |
init_kmem_cache_node(&s->local_node, s); |
81819f0fc
|
2116 2117 2118 2119 2120 2121 2122 2123 |
return 1; } #endif /* * calculate_sizes() determines the order and the distribution of data within * a slab object. */ |
06b285dc3
|
2124 |
static int calculate_sizes(struct kmem_cache *s, int forced_order) |
81819f0fc
|
2125 2126 2127 2128 |
{ unsigned long flags = s->flags; unsigned long size = s->objsize; unsigned long align = s->align; |
834f3d119
|
2129 |
int order; |
81819f0fc
|
2130 2131 |
/* |
d8b42bf54
|
2132 2133 2134 2135 2136 2137 2138 2139 |
* Round up object size to the next word boundary. We can only * place the free pointer at word boundaries and this determines * the possible location of the free pointer. */ size = ALIGN(size, sizeof(void *)); #ifdef CONFIG_SLUB_DEBUG /* |
81819f0fc
|
2140 2141 2142 2143 2144 |
* Determine if we can poison the object itself. If the user of * the slab may touch the object after free or before allocation * then we should never poison the object itself. */ if ((flags & SLAB_POISON) && !(flags & SLAB_DESTROY_BY_RCU) && |
c59def9f2
|
2145 |
!s->ctor) |
81819f0fc
|
2146 2147 2148 |
s->flags |= __OBJECT_POISON; else s->flags &= ~__OBJECT_POISON; |
81819f0fc
|
2149 2150 |
/* |
672bba3a4
|
2151 |
* If we are Redzoning then check if there is some space between the |
81819f0fc
|
2152 |
* end of the object and the free pointer. If not then add an |
672bba3a4
|
2153 |
* additional word to have some bytes to store Redzone information. |
81819f0fc
|
2154 2155 2156 |
*/ if ((flags & SLAB_RED_ZONE) && size == s->objsize) size += sizeof(void *); |
41ecc55b8
|
2157 |
#endif |
81819f0fc
|
2158 2159 |
/* |
672bba3a4
|
2160 2161 |
* With that we have determined the number of bytes in actual use * by the object. This is the potential offset to the free pointer. |
81819f0fc
|
2162 2163 2164 2165 |
*/ s->inuse = size; if (((flags & (SLAB_DESTROY_BY_RCU | SLAB_POISON)) || |
c59def9f2
|
2166 |
s->ctor)) { |
81819f0fc
|
2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 |
/* * Relocate free pointer after the object if it is not * permitted to overwrite the first word of the object on * kmem_cache_free. * * This is the case if we do RCU, have a constructor or * destructor or are poisoning the objects. */ s->offset = size; size += sizeof(void *); } |
c12b3c625
|
2178 |
#ifdef CONFIG_SLUB_DEBUG |
81819f0fc
|
2179 2180 2181 2182 2183 2184 |
if (flags & SLAB_STORE_USER) /* * Need to store information about allocs and frees after * the object. */ size += 2 * sizeof(struct track); |
be7b3fbce
|
2185 |
if (flags & SLAB_RED_ZONE) |
81819f0fc
|
2186 2187 2188 2189 |
/* * Add some empty padding so that we can catch * overwrites from earlier objects rather than let * tracking information or the free pointer be |
0211a9c85
|
2190 |
* corrupted if a user writes before the start |
81819f0fc
|
2191 2192 2193 |
* of the object. */ size += sizeof(void *); |
41ecc55b8
|
2194 |
#endif |
672bba3a4
|
2195 |
|
81819f0fc
|
2196 2197 |
/* * Determine the alignment based on various parameters that the |
65c02d4cf
|
2198 2199 |
* user specified and the dynamic determination of cache line size * on bootup. |
81819f0fc
|
2200 2201 2202 2203 2204 2205 2206 2207 2208 2209 |
*/ align = calculate_alignment(flags, align, s->objsize); /* * SLUB stores one object immediately after another beginning from * offset 0. In order to align the objects we have to simply size * each object to conform to the alignment. */ size = ALIGN(size, align); s->size = size; |
06b285dc3
|
2210 2211 2212 2213 |
if (forced_order >= 0) order = forced_order; else order = calculate_order(size); |
81819f0fc
|
2214 |
|
834f3d119
|
2215 |
if (order < 0) |
81819f0fc
|
2216 |
return 0; |
b7a49f0d4
|
2217 |
s->allocflags = 0; |
834f3d119
|
2218 |
if (order) |
b7a49f0d4
|
2219 2220 2221 2222 2223 2224 2225 |
s->allocflags |= __GFP_COMP; if (s->flags & SLAB_CACHE_DMA) s->allocflags |= SLUB_DMA; if (s->flags & SLAB_RECLAIM_ACCOUNT) s->allocflags |= __GFP_RECLAIMABLE; |
81819f0fc
|
2226 2227 2228 |
/* * Determine the number of objects per slab */ |
834f3d119
|
2229 |
s->oo = oo_make(order, size); |
65c3376aa
|
2230 |
s->min = oo_make(get_order(size), size); |
205ab99dd
|
2231 2232 |
if (oo_objects(s->oo) > oo_objects(s->max)) s->max = s->oo; |
81819f0fc
|
2233 |
|
834f3d119
|
2234 |
return !!oo_objects(s->oo); |
81819f0fc
|
2235 2236 |
} |
81819f0fc
|
2237 2238 2239 |
static int kmem_cache_open(struct kmem_cache *s, gfp_t gfpflags, const char *name, size_t size, size_t align, unsigned long flags, |
51cc50685
|
2240 |
void (*ctor)(void *)) |
81819f0fc
|
2241 2242 2243 2244 |
{ memset(s, 0, kmem_size); s->name = name; s->ctor = ctor; |
81819f0fc
|
2245 |
s->objsize = size; |
81819f0fc
|
2246 |
s->align = align; |
ba0268a8b
|
2247 |
s->flags = kmem_cache_flags(size, flags, name, ctor); |
81819f0fc
|
2248 |
|
06b285dc3
|
2249 |
if (!calculate_sizes(s, -1)) |
81819f0fc
|
2250 2251 2252 2253 |
goto error; s->refcount = 1; #ifdef CONFIG_NUMA |
e2cb96b7e
|
2254 |
s->remote_node_defrag_ratio = 1000; |
81819f0fc
|
2255 |
#endif |
dfb4f0960
|
2256 2257 |
if (!init_kmem_cache_nodes(s, gfpflags & ~SLUB_DMA)) goto error; |
81819f0fc
|
2258 |
|
dfb4f0960
|
2259 |
if (alloc_kmem_cache_cpus(s, gfpflags & ~SLUB_DMA)) |
81819f0fc
|
2260 |
return 1; |
4c93c355d
|
2261 |
free_kmem_cache_nodes(s); |
81819f0fc
|
2262 2263 2264 2265 2266 |
error: if (flags & SLAB_PANIC) panic("Cannot create slab %s size=%lu realsize=%u " "order=%u offset=%u flags=%lx ", |
834f3d119
|
2267 |
s->name, (unsigned long)size, s->size, oo_order(s->oo), |
81819f0fc
|
2268 2269 2270 |
s->offset, flags); return 0; } |
81819f0fc
|
2271 2272 2273 2274 2275 2276 |
/* * Check if a given pointer is valid */ int kmem_ptr_validate(struct kmem_cache *s, const void *object) { |
064287807
|
2277 |
struct page *page; |
81819f0fc
|
2278 2279 2280 2281 2282 2283 |
page = get_object_page(object); if (!page || s != page->slab) /* No slab or wrong slab */ return 0; |
abcd08a6f
|
2284 |
if (!check_valid_pointer(s, page, object)) |
81819f0fc
|
2285 2286 2287 2288 2289 |
return 0; /* * We could also check if the object is on the slabs freelist. * But this would be too expensive and it seems that the main |
6446faa2f
|
2290 |
* purpose of kmem_ptr_valid() is to check if the object belongs |
81819f0fc
|
2291 2292 2293 2294 2295 2296 2297 2298 2299 2300 2301 2302 2303 2304 2305 2306 2307 2308 2309 2310 |
* to a certain slab. */ return 1; } EXPORT_SYMBOL(kmem_ptr_validate); /* * Determine the size of a slab object */ unsigned int kmem_cache_size(struct kmem_cache *s) { return s->objsize; } EXPORT_SYMBOL(kmem_cache_size); const char *kmem_cache_name(struct kmem_cache *s) { return s->name; } EXPORT_SYMBOL(kmem_cache_name); |
33b12c381
|
2311 2312 2313 2314 2315 2316 2317 2318 2319 2320 2321 2322 2323 2324 2325 2326 2327 2328 2329 2330 2331 2332 2333 2334 2335 2336 |
static void list_slab_objects(struct kmem_cache *s, struct page *page, const char *text) { #ifdef CONFIG_SLUB_DEBUG void *addr = page_address(page); void *p; DECLARE_BITMAP(map, page->objects); bitmap_zero(map, page->objects); slab_err(s, page, "%s", text); slab_lock(page); for_each_free_object(p, s, page->freelist) set_bit(slab_index(p, s, addr), map); for_each_object(p, s, addr, page->objects) { if (!test_bit(slab_index(p, s, addr), map)) { printk(KERN_ERR "INFO: Object 0x%p @offset=%tu ", p, p - addr); print_tracking(s, p); } } slab_unlock(page); #endif } |
81819f0fc
|
2337 |
/* |
599870b17
|
2338 |
* Attempt to free all partial slabs on a node. |
81819f0fc
|
2339 |
*/ |
599870b17
|
2340 |
static void free_partial(struct kmem_cache *s, struct kmem_cache_node *n) |
81819f0fc
|
2341 |
{ |
81819f0fc
|
2342 2343 2344 2345 |
unsigned long flags; struct page *page, *h; spin_lock_irqsave(&n->list_lock, flags); |
33b12c381
|
2346 |
list_for_each_entry_safe(page, h, &n->partial, lru) { |
81819f0fc
|
2347 2348 2349 |
if (!page->inuse) { list_del(&page->lru); discard_slab(s, page); |
599870b17
|
2350 |
n->nr_partial--; |
33b12c381
|
2351 2352 2353 |
} else { list_slab_objects(s, page, "Objects remaining on kmem_cache_close()"); |
599870b17
|
2354 |
} |
33b12c381
|
2355 |
} |
81819f0fc
|
2356 |
spin_unlock_irqrestore(&n->list_lock, flags); |
81819f0fc
|
2357 2358 2359 |
} /* |
672bba3a4
|
2360 |
* Release all resources used by a slab cache. |
81819f0fc
|
2361 |
*/ |
0c7100132
|
2362 |
static inline int kmem_cache_close(struct kmem_cache *s) |
81819f0fc
|
2363 2364 2365 2366 2367 2368 |
{ int node; flush_all(s); /* Attempt to free all objects */ |
4c93c355d
|
2369 |
free_kmem_cache_cpus(s); |
f64dc58c5
|
2370 |
for_each_node_state(node, N_NORMAL_MEMORY) { |
81819f0fc
|
2371 |
struct kmem_cache_node *n = get_node(s, node); |
599870b17
|
2372 2373 |
free_partial(s, n); if (n->nr_partial || slabs_node(s, node)) |
81819f0fc
|
2374 2375 2376 2377 2378 2379 2380 2381 2382 2383 2384 2385 2386 2387 2388 2389 |
return 1; } free_kmem_cache_nodes(s); return 0; } /* * Close a cache and release the kmem_cache structure * (must be used for caches created using kmem_cache_create) */ void kmem_cache_destroy(struct kmem_cache *s) { down_write(&slub_lock); s->refcount--; if (!s->refcount) { list_del(&s->list); |
a0e1d1be2
|
2390 |
up_write(&slub_lock); |
d629d8195
|
2391 2392 2393 2394 2395 2396 |
if (kmem_cache_close(s)) { printk(KERN_ERR "SLUB %s: %s called for cache that " "still has objects. ", s->name, __func__); dump_stack(); } |
81819f0fc
|
2397 |
sysfs_slab_remove(s); |
a0e1d1be2
|
2398 2399 |
} else up_write(&slub_lock); |
81819f0fc
|
2400 2401 2402 2403 2404 2405 |
} EXPORT_SYMBOL(kmem_cache_destroy); /******************************************************************** * Kmalloc subsystem *******************************************************************/ |
331dc558f
|
2406 |
struct kmem_cache kmalloc_caches[PAGE_SHIFT + 1] __cacheline_aligned; |
81819f0fc
|
2407 |
EXPORT_SYMBOL(kmalloc_caches); |
81819f0fc
|
2408 2409 |
static int __init setup_slub_min_order(char *str) { |
064287807
|
2410 |
get_option(&str, &slub_min_order); |
81819f0fc
|
2411 2412 2413 2414 2415 2416 2417 2418 |
return 1; } __setup("slub_min_order=", setup_slub_min_order); static int __init setup_slub_max_order(char *str) { |
064287807
|
2419 |
get_option(&str, &slub_max_order); |
81819f0fc
|
2420 2421 2422 2423 2424 2425 2426 2427 |
return 1; } __setup("slub_max_order=", setup_slub_max_order); static int __init setup_slub_min_objects(char *str) { |
064287807
|
2428 |
get_option(&str, &slub_min_objects); |
81819f0fc
|
2429 2430 2431 2432 2433 2434 2435 2436 2437 2438 2439 2440 2441 |
return 1; } __setup("slub_min_objects=", setup_slub_min_objects); static int __init setup_slub_nomerge(char *str) { slub_nomerge = 1; return 1; } __setup("slub_nomerge", setup_slub_nomerge); |
81819f0fc
|
2442 2443 2444 2445 2446 2447 2448 2449 2450 2451 |
static struct kmem_cache *create_kmalloc_cache(struct kmem_cache *s, const char *name, int size, gfp_t gfp_flags) { unsigned int flags = 0; if (gfp_flags & SLUB_DMA) flags = SLAB_CACHE_DMA; down_write(&slub_lock); if (!kmem_cache_open(s, gfp_flags, name, size, ARCH_KMALLOC_MINALIGN, |
319d1e240
|
2452 |
flags, NULL)) |
81819f0fc
|
2453 2454 2455 2456 2457 2458 2459 2460 2461 2462 2463 2464 |
goto panic; list_add(&s->list, &slab_caches); up_write(&slub_lock); if (sysfs_slab_add(s)) goto panic; return s; panic: panic("Creation of kmalloc slab %s size=%d failed. ", name, size); } |
2e443fd00
|
2465 |
#ifdef CONFIG_ZONE_DMA |
4097d6017
|
2466 |
static struct kmem_cache *kmalloc_caches_dma[PAGE_SHIFT + 1]; |
1ceef4024
|
2467 2468 2469 2470 2471 2472 2473 2474 2475 2476 2477 2478 2479 2480 2481 2482 |
static void sysfs_add_func(struct work_struct *w) { struct kmem_cache *s; down_write(&slub_lock); list_for_each_entry(s, &slab_caches, list) { if (s->flags & __SYSFS_ADD_DEFERRED) { s->flags &= ~__SYSFS_ADD_DEFERRED; sysfs_slab_add(s); } } up_write(&slub_lock); } static DECLARE_WORK(sysfs_add_work, sysfs_add_func); |
2e443fd00
|
2483 2484 2485 |
static noinline struct kmem_cache *dma_kmalloc_cache(int index, gfp_t flags) { struct kmem_cache *s; |
2e443fd00
|
2486 2487 2488 2489 2490 2491 2492 2493 |
char *text; size_t realsize; s = kmalloc_caches_dma[index]; if (s) return s; /* Dynamically create dma cache */ |
1ceef4024
|
2494 2495 2496 2497 2498 2499 2500 2501 2502 |
if (flags & __GFP_WAIT) down_write(&slub_lock); else { if (!down_write_trylock(&slub_lock)) goto out; } if (kmalloc_caches_dma[index]) goto unlock_out; |
2e443fd00
|
2503 |
|
7b55f620e
|
2504 |
realsize = kmalloc_caches[index].objsize; |
3adbefee6
|
2505 2506 |
text = kasprintf(flags & ~SLUB_DMA, "kmalloc_dma-%d", (unsigned int)realsize); |
1ceef4024
|
2507 2508 2509 2510 2511 2512 2513 2514 |
s = kmalloc(kmem_size, flags & ~SLUB_DMA); if (!s || !text || !kmem_cache_open(s, flags, text, realsize, ARCH_KMALLOC_MINALIGN, SLAB_CACHE_DMA|__SYSFS_ADD_DEFERRED, NULL)) { kfree(s); kfree(text); goto unlock_out; |
dfce8648d
|
2515 |
} |
1ceef4024
|
2516 2517 2518 2519 2520 2521 2522 |
list_add(&s->list, &slab_caches); kmalloc_caches_dma[index] = s; schedule_work(&sysfs_add_work); unlock_out: |
dfce8648d
|
2523 |
up_write(&slub_lock); |
1ceef4024
|
2524 |
out: |
dfce8648d
|
2525 |
return kmalloc_caches_dma[index]; |
2e443fd00
|
2526 2527 |
} #endif |
f1b263393
|
2528 2529 2530 2531 2532 2533 2534 2535 2536 2537 2538 2539 2540 2541 2542 2543 2544 2545 2546 2547 2548 2549 2550 2551 2552 2553 2554 2555 2556 2557 2558 2559 |
/* * Conversion table for small slabs sizes / 8 to the index in the * kmalloc array. This is necessary for slabs < 192 since we have non power * of two cache sizes there. The size of larger slabs can be determined using * fls. */ static s8 size_index[24] = { 3, /* 8 */ 4, /* 16 */ 5, /* 24 */ 5, /* 32 */ 6, /* 40 */ 6, /* 48 */ 6, /* 56 */ 6, /* 64 */ 1, /* 72 */ 1, /* 80 */ 1, /* 88 */ 1, /* 96 */ 7, /* 104 */ 7, /* 112 */ 7, /* 120 */ 7, /* 128 */ 2, /* 136 */ 2, /* 144 */ 2, /* 152 */ 2, /* 160 */ 2, /* 168 */ 2, /* 176 */ 2, /* 184 */ 2 /* 192 */ }; |
81819f0fc
|
2560 2561 |
static struct kmem_cache *get_slab(size_t size, gfp_t flags) { |
f1b263393
|
2562 |
int index; |
81819f0fc
|
2563 |
|
f1b263393
|
2564 2565 2566 |
if (size <= 192) { if (!size) return ZERO_SIZE_PTR; |
81819f0fc
|
2567 |
|
f1b263393
|
2568 |
index = size_index[(size - 1) / 8]; |
aadb4bc4a
|
2569 |
} else |
f1b263393
|
2570 |
index = fls(size - 1); |
81819f0fc
|
2571 2572 |
#ifdef CONFIG_ZONE_DMA |
f1b263393
|
2573 |
if (unlikely((flags & SLUB_DMA))) |
2e443fd00
|
2574 |
return dma_kmalloc_cache(index, flags); |
f1b263393
|
2575 |
|
81819f0fc
|
2576 2577 2578 2579 2580 2581 |
#endif return &kmalloc_caches[index]; } void *__kmalloc(size_t size, gfp_t flags) { |
aadb4bc4a
|
2582 |
struct kmem_cache *s; |
81819f0fc
|
2583 |
|
331dc558f
|
2584 |
if (unlikely(size > PAGE_SIZE)) |
eada35efc
|
2585 |
return kmalloc_large(size, flags); |
aadb4bc4a
|
2586 2587 2588 2589 |
s = get_slab(size, flags); if (unlikely(ZERO_OR_NULL_PTR(s))) |
6cb8f9132
|
2590 |
return s; |
ce71e27c6
|
2591 |
return slab_alloc(s, flags, -1, _RET_IP_); |
81819f0fc
|
2592 2593 |
} EXPORT_SYMBOL(__kmalloc); |
f619cfe1b
|
2594 2595 2596 2597 2598 2599 2600 2601 2602 2603 |
static void *kmalloc_large_node(size_t size, gfp_t flags, int node) { struct page *page = alloc_pages_node(node, flags | __GFP_COMP, get_order(size)); if (page) return page_address(page); else return NULL; } |
81819f0fc
|
2604 2605 2606 |
#ifdef CONFIG_NUMA void *__kmalloc_node(size_t size, gfp_t flags, int node) { |
aadb4bc4a
|
2607 |
struct kmem_cache *s; |
81819f0fc
|
2608 |
|
331dc558f
|
2609 |
if (unlikely(size > PAGE_SIZE)) |
f619cfe1b
|
2610 |
return kmalloc_large_node(size, flags, node); |
aadb4bc4a
|
2611 2612 2613 2614 |
s = get_slab(size, flags); if (unlikely(ZERO_OR_NULL_PTR(s))) |
6cb8f9132
|
2615 |
return s; |
ce71e27c6
|
2616 |
return slab_alloc(s, flags, node, _RET_IP_); |
81819f0fc
|
2617 2618 2619 2620 2621 2622 |
} EXPORT_SYMBOL(__kmalloc_node); #endif size_t ksize(const void *object) { |
272c1d21d
|
2623 |
struct page *page; |
81819f0fc
|
2624 |
struct kmem_cache *s; |
ef8b4520b
|
2625 |
if (unlikely(object == ZERO_SIZE_PTR)) |
272c1d21d
|
2626 |
return 0; |
294a80a8e
|
2627 |
page = virt_to_head_page(object); |
294a80a8e
|
2628 |
|
76994412f
|
2629 2630 |
if (unlikely(!PageSlab(page))) { WARN_ON(!PageCompound(page)); |
294a80a8e
|
2631 |
return PAGE_SIZE << compound_order(page); |
76994412f
|
2632 |
} |
81819f0fc
|
2633 |
s = page->slab; |
81819f0fc
|
2634 |
|
ae20bfda6
|
2635 |
#ifdef CONFIG_SLUB_DEBUG |
81819f0fc
|
2636 2637 2638 2639 2640 2641 |
/* * Debugging requires use of the padding between object * and whatever may come after it. */ if (s->flags & (SLAB_RED_ZONE | SLAB_POISON)) return s->objsize; |
ae20bfda6
|
2642 |
#endif |
81819f0fc
|
2643 2644 2645 2646 2647 2648 2649 |
/* * If we have the need to store the freelist pointer * back there or track user information then we can * only use the space before that information. */ if (s->flags & (SLAB_DESTROY_BY_RCU | SLAB_STORE_USER)) return s->inuse; |
81819f0fc
|
2650 2651 2652 2653 2654 |
/* * Else we can use all the padding etc for the allocation */ return s->size; } |
81819f0fc
|
2655 2656 2657 |
void kfree(const void *x) { |
81819f0fc
|
2658 |
struct page *page; |
5bb983b0c
|
2659 |
void *object = (void *)x; |
81819f0fc
|
2660 |
|
2408c5503
|
2661 |
if (unlikely(ZERO_OR_NULL_PTR(x))) |
81819f0fc
|
2662 |
return; |
b49af68ff
|
2663 |
page = virt_to_head_page(x); |
aadb4bc4a
|
2664 |
if (unlikely(!PageSlab(page))) { |
0937502af
|
2665 |
BUG_ON(!PageCompound(page)); |
aadb4bc4a
|
2666 2667 2668 |
put_page(page); return; } |
ce71e27c6
|
2669 |
slab_free(page->slab, page, object, _RET_IP_); |
81819f0fc
|
2670 2671 |
} EXPORT_SYMBOL(kfree); |
2086d26a0
|
2672 |
/* |
672bba3a4
|
2673 2674 2675 2676 2677 2678 2679 2680 |
* kmem_cache_shrink removes empty slabs from the partial lists and sorts * the remaining slabs by the number of items in use. The slabs with the * most items in use come first. New allocations will then fill those up * and thus they can be removed from the partial lists. * * The slabs with the least items are placed last. This results in them * being allocated from last increasing the chance that the last objects * are freed in them. |
2086d26a0
|
2681 2682 2683 2684 2685 2686 2687 2688 |
*/ int kmem_cache_shrink(struct kmem_cache *s) { int node; int i; struct kmem_cache_node *n; struct page *page; struct page *t; |
205ab99dd
|
2689 |
int objects = oo_objects(s->max); |
2086d26a0
|
2690 |
struct list_head *slabs_by_inuse = |
834f3d119
|
2691 |
kmalloc(sizeof(struct list_head) * objects, GFP_KERNEL); |
2086d26a0
|
2692 2693 2694 2695 2696 2697 |
unsigned long flags; if (!slabs_by_inuse) return -ENOMEM; flush_all(s); |
f64dc58c5
|
2698 |
for_each_node_state(node, N_NORMAL_MEMORY) { |
2086d26a0
|
2699 2700 2701 2702 |
n = get_node(s, node); if (!n->nr_partial) continue; |
834f3d119
|
2703 |
for (i = 0; i < objects; i++) |
2086d26a0
|
2704 2705 2706 2707 2708 |
INIT_LIST_HEAD(slabs_by_inuse + i); spin_lock_irqsave(&n->list_lock, flags); /* |
672bba3a4
|
2709 |
* Build lists indexed by the items in use in each slab. |
2086d26a0
|
2710 |
* |
672bba3a4
|
2711 2712 |
* Note that concurrent frees may occur while we hold the * list_lock. page->inuse here is the upper limit. |
2086d26a0
|
2713 2714 2715 2716 2717 2718 2719 2720 2721 2722 2723 2724 2725 |
*/ list_for_each_entry_safe(page, t, &n->partial, lru) { if (!page->inuse && slab_trylock(page)) { /* * Must hold slab lock here because slab_free * may have freed the last object and be * waiting to release the slab. */ list_del(&page->lru); n->nr_partial--; slab_unlock(page); discard_slab(s, page); } else { |
fcda3d89b
|
2726 2727 |
list_move(&page->lru, slabs_by_inuse + page->inuse); |
2086d26a0
|
2728 2729 |
} } |
2086d26a0
|
2730 |
/* |
672bba3a4
|
2731 2732 |
* Rebuild the partial list with the slabs filled up most * first and the least used slabs at the end. |
2086d26a0
|
2733 |
*/ |
834f3d119
|
2734 |
for (i = objects - 1; i >= 0; i--) |
2086d26a0
|
2735 |
list_splice(slabs_by_inuse + i, n->partial.prev); |
2086d26a0
|
2736 2737 2738 2739 2740 2741 2742 |
spin_unlock_irqrestore(&n->list_lock, flags); } kfree(slabs_by_inuse); return 0; } EXPORT_SYMBOL(kmem_cache_shrink); |
b9049e234
|
2743 2744 2745 2746 2747 2748 2749 2750 2751 2752 2753 2754 2755 2756 2757 2758 2759 2760 2761 2762 2763 2764 2765 2766 2767 2768 2769 2770 2771 2772 2773 2774 2775 2776 2777 2778 2779 2780 2781 |
#if defined(CONFIG_NUMA) && defined(CONFIG_MEMORY_HOTPLUG) static int slab_mem_going_offline_callback(void *arg) { struct kmem_cache *s; down_read(&slub_lock); list_for_each_entry(s, &slab_caches, list) kmem_cache_shrink(s); up_read(&slub_lock); return 0; } static void slab_mem_offline_callback(void *arg) { struct kmem_cache_node *n; struct kmem_cache *s; struct memory_notify *marg = arg; int offline_node; offline_node = marg->status_change_nid; /* * If the node still has available memory. we need kmem_cache_node * for it yet. */ if (offline_node < 0) return; down_read(&slub_lock); list_for_each_entry(s, &slab_caches, list) { n = get_node(s, offline_node); if (n) { /* * if n->nr_slabs > 0, slabs still exist on the node * that is going down. We were unable to free them, * and offline_pages() function shoudn't call this * callback. So, we must fail. */ |
0f389ec63
|
2782 |
BUG_ON(slabs_node(s, offline_node)); |
b9049e234
|
2783 2784 2785 2786 2787 2788 2789 2790 2791 2792 2793 2794 2795 2796 2797 2798 2799 2800 2801 2802 2803 2804 2805 2806 |
s->node[offline_node] = NULL; kmem_cache_free(kmalloc_caches, n); } } up_read(&slub_lock); } static int slab_mem_going_online_callback(void *arg) { struct kmem_cache_node *n; struct kmem_cache *s; struct memory_notify *marg = arg; int nid = marg->status_change_nid; int ret = 0; /* * If the node's memory is already available, then kmem_cache_node is * already created. Nothing to do. */ if (nid < 0) return 0; /* |
0121c619d
|
2807 |
* We are bringing a node online. No memory is available yet. We must |
b9049e234
|
2808 2809 2810 2811 2812 2813 2814 2815 2816 2817 2818 2819 2820 2821 2822 |
* allocate a kmem_cache_node structure in order to bring the node * online. */ down_read(&slub_lock); list_for_each_entry(s, &slab_caches, list) { /* * XXX: kmem_cache_alloc_node will fallback to other nodes * since memory is not yet available from the node that * is brought up. */ n = kmem_cache_alloc(kmalloc_caches, GFP_KERNEL); if (!n) { ret = -ENOMEM; goto out; } |
5595cffc8
|
2823 |
init_kmem_cache_node(n, s); |
b9049e234
|
2824 2825 2826 2827 2828 2829 2830 2831 2832 2833 2834 2835 2836 2837 2838 2839 2840 2841 2842 2843 2844 2845 2846 2847 2848 2849 2850 |
s->node[nid] = n; } out: up_read(&slub_lock); return ret; } static int slab_memory_callback(struct notifier_block *self, unsigned long action, void *arg) { int ret = 0; switch (action) { case MEM_GOING_ONLINE: ret = slab_mem_going_online_callback(arg); break; case MEM_GOING_OFFLINE: ret = slab_mem_going_offline_callback(arg); break; case MEM_OFFLINE: case MEM_CANCEL_ONLINE: slab_mem_offline_callback(arg); break; case MEM_ONLINE: case MEM_CANCEL_OFFLINE: break; } |
dc19f9db3
|
2851 2852 2853 2854 |
if (ret) ret = notifier_from_errno(ret); else ret = NOTIFY_OK; |
b9049e234
|
2855 2856 2857 2858 |
return ret; } #endif /* CONFIG_MEMORY_HOTPLUG */ |
81819f0fc
|
2859 2860 2861 2862 2863 2864 2865 |
/******************************************************************** * Basic setup of slabs *******************************************************************/ void __init kmem_cache_init(void) { int i; |
4b356be01
|
2866 |
int caches = 0; |
81819f0fc
|
2867 |
|
4c93c355d
|
2868 |
init_alloc_cpu(); |
81819f0fc
|
2869 2870 2871 |
#ifdef CONFIG_NUMA /* * Must first have the slab cache available for the allocations of the |
672bba3a4
|
2872 |
* struct kmem_cache_node's. There is special bootstrap code in |
81819f0fc
|
2873 2874 2875 2876 |
* kmem_cache_open for slab_state == DOWN. */ create_kmalloc_cache(&kmalloc_caches[0], "kmem_cache_node", sizeof(struct kmem_cache_node), GFP_KERNEL); |
8ffa68755
|
2877 |
kmalloc_caches[0].refcount = -1; |
4b356be01
|
2878 |
caches++; |
b9049e234
|
2879 |
|
0c40ba4fd
|
2880 |
hotplug_memory_notifier(slab_memory_callback, SLAB_CALLBACK_PRI); |
81819f0fc
|
2881 2882 2883 2884 2885 2886 |
#endif /* Able to allocate the per node structures */ slab_state = PARTIAL; /* Caches that are not of the two-to-the-power-of size */ |
4b356be01
|
2887 2888 |
if (KMALLOC_MIN_SIZE <= 64) { create_kmalloc_cache(&kmalloc_caches[1], |
81819f0fc
|
2889 |
"kmalloc-96", 96, GFP_KERNEL); |
4b356be01
|
2890 |
caches++; |
4b356be01
|
2891 |
create_kmalloc_cache(&kmalloc_caches[2], |
81819f0fc
|
2892 |
"kmalloc-192", 192, GFP_KERNEL); |
4b356be01
|
2893 2894 |
caches++; } |
81819f0fc
|
2895 |
|
331dc558f
|
2896 |
for (i = KMALLOC_SHIFT_LOW; i <= PAGE_SHIFT; i++) { |
81819f0fc
|
2897 2898 |
create_kmalloc_cache(&kmalloc_caches[i], "kmalloc", 1 << i, GFP_KERNEL); |
4b356be01
|
2899 2900 |
caches++; } |
81819f0fc
|
2901 |
|
f1b263393
|
2902 2903 2904 2905 |
/* * Patch up the size_index table if we have strange large alignment * requirements for the kmalloc array. This is only the case for |
6446faa2f
|
2906 |
* MIPS it seems. The standard arches will not generate any code here. |
f1b263393
|
2907 2908 2909 2910 2911 2912 2913 2914 2915 |
* * Largest permitted alignment is 256 bytes due to the way we * handle the index determination for the smaller caches. * * Make sure that nothing crazy happens if someone starts tinkering * around with ARCH_KMALLOC_MINALIGN */ BUILD_BUG_ON(KMALLOC_MIN_SIZE > 256 || (KMALLOC_MIN_SIZE & (KMALLOC_MIN_SIZE - 1))); |
12ad6843d
|
2916 |
for (i = 8; i < KMALLOC_MIN_SIZE; i += 8) |
f1b263393
|
2917 |
size_index[(i - 1) / 8] = KMALLOC_SHIFT_LOW; |
41d54d3bf
|
2918 2919 2920 2921 2922 2923 2924 2925 2926 |
if (KMALLOC_MIN_SIZE == 128) { /* * The 192 byte sized cache is not used if the alignment * is 128 byte. Redirect kmalloc to use the 256 byte cache * instead. */ for (i = 128 + 8; i <= 192; i += 8) size_index[(i - 1) / 8] = 8; } |
81819f0fc
|
2927 2928 2929 |
slab_state = UP; /* Provide the correct kmalloc names now that the caches are up */ |
331dc558f
|
2930 |
for (i = KMALLOC_SHIFT_LOW; i <= PAGE_SHIFT; i++) |
81819f0fc
|
2931 2932 2933 2934 2935 |
kmalloc_caches[i]. name = kasprintf(GFP_KERNEL, "kmalloc-%d", 1 << i); #ifdef CONFIG_SMP register_cpu_notifier(&slab_notifier); |
4c93c355d
|
2936 2937 2938 2939 |
kmem_size = offsetof(struct kmem_cache, cpu_slab) + nr_cpu_ids * sizeof(struct kmem_cache_cpu *); #else kmem_size = sizeof(struct kmem_cache); |
81819f0fc
|
2940 |
#endif |
3adbefee6
|
2941 2942 |
printk(KERN_INFO "SLUB: Genslabs=%d, HWalign=%d, Order=%d-%d, MinObjects=%d," |
4b356be01
|
2943 2944 2945 |
" CPUs=%d, Nodes=%d ", caches, cache_line_size(), |
81819f0fc
|
2946 2947 2948 2949 2950 2951 2952 2953 2954 2955 2956 |
slub_min_order, slub_max_order, slub_min_objects, nr_cpu_ids, nr_node_ids); } /* * Find a mergeable slab cache */ static int slab_unmergeable(struct kmem_cache *s) { if (slub_nomerge || (s->flags & SLUB_NEVER_MERGE)) return 1; |
c59def9f2
|
2957 |
if (s->ctor) |
81819f0fc
|
2958 |
return 1; |
8ffa68755
|
2959 2960 2961 2962 2963 |
/* * We may have set a slab to be unmergeable during bootstrap. */ if (s->refcount < 0) return 1; |
81819f0fc
|
2964 2965 2966 2967 |
return 0; } static struct kmem_cache *find_mergeable(size_t size, |
ba0268a8b
|
2968 |
size_t align, unsigned long flags, const char *name, |
51cc50685
|
2969 |
void (*ctor)(void *)) |
81819f0fc
|
2970 |
{ |
5b95a4acf
|
2971 |
struct kmem_cache *s; |
81819f0fc
|
2972 2973 2974 |
if (slub_nomerge || (flags & SLUB_NEVER_MERGE)) return NULL; |
c59def9f2
|
2975 |
if (ctor) |
81819f0fc
|
2976 2977 2978 2979 2980 |
return NULL; size = ALIGN(size, sizeof(void *)); align = calculate_alignment(flags, align, size); size = ALIGN(size, align); |
ba0268a8b
|
2981 |
flags = kmem_cache_flags(size, flags, name, NULL); |
81819f0fc
|
2982 |
|
5b95a4acf
|
2983 |
list_for_each_entry(s, &slab_caches, list) { |
81819f0fc
|
2984 2985 2986 2987 2988 |
if (slab_unmergeable(s)) continue; if (size > s->size) continue; |
ba0268a8b
|
2989 |
if ((flags & SLUB_MERGE_SAME) != (s->flags & SLUB_MERGE_SAME)) |
81819f0fc
|
2990 2991 2992 2993 2994 |
continue; /* * Check if alignment is compatible. * Courtesy of Adrian Drzewiecki */ |
064287807
|
2995 |
if ((s->size & ~(align - 1)) != s->size) |
81819f0fc
|
2996 2997 2998 2999 3000 3001 3002 3003 3004 3005 3006 |
continue; if (s->size - size >= sizeof(void *)) continue; return s; } return NULL; } struct kmem_cache *kmem_cache_create(const char *name, size_t size, |
51cc50685
|
3007 |
size_t align, unsigned long flags, void (*ctor)(void *)) |
81819f0fc
|
3008 3009 3010 3011 |
{ struct kmem_cache *s; down_write(&slub_lock); |
ba0268a8b
|
3012 |
s = find_mergeable(size, align, flags, name, ctor); |
81819f0fc
|
3013 |
if (s) { |
42a9fdbb1
|
3014 |
int cpu; |
81819f0fc
|
3015 3016 3017 3018 3019 3020 |
s->refcount++; /* * Adjust the object sizes so that we clear * the complete object on kzalloc. */ s->objsize = max(s->objsize, (int)size); |
42a9fdbb1
|
3021 3022 3023 3024 3025 3026 3027 |
/* * And then we need to update the object size in the * per cpu structures */ for_each_online_cpu(cpu) get_cpu_slab(s, cpu)->objsize = s->objsize; |
6446faa2f
|
3028 |
|
81819f0fc
|
3029 |
s->inuse = max_t(int, s->inuse, ALIGN(size, sizeof(void *))); |
a0e1d1be2
|
3030 |
up_write(&slub_lock); |
6446faa2f
|
3031 |
|
7b8f3b66d
|
3032 3033 3034 3035 |
if (sysfs_slab_alias(s, name)) { down_write(&slub_lock); s->refcount--; up_write(&slub_lock); |
81819f0fc
|
3036 |
goto err; |
7b8f3b66d
|
3037 |
} |
a0e1d1be2
|
3038 3039 |
return s; } |
6446faa2f
|
3040 |
|
a0e1d1be2
|
3041 3042 3043 |
s = kmalloc(kmem_size, GFP_KERNEL); if (s) { if (kmem_cache_open(s, GFP_KERNEL, name, |
c59def9f2
|
3044 |
size, align, flags, ctor)) { |
81819f0fc
|
3045 |
list_add(&s->list, &slab_caches); |
a0e1d1be2
|
3046 |
up_write(&slub_lock); |
7b8f3b66d
|
3047 3048 3049 3050 3051 |
if (sysfs_slab_add(s)) { down_write(&slub_lock); list_del(&s->list); up_write(&slub_lock); kfree(s); |
a0e1d1be2
|
3052 |
goto err; |
7b8f3b66d
|
3053 |
} |
a0e1d1be2
|
3054 3055 3056 |
return s; } kfree(s); |
81819f0fc
|
3057 3058 |
} up_write(&slub_lock); |
81819f0fc
|
3059 3060 |
err: |
81819f0fc
|
3061 3062 3063 3064 3065 3066 3067 3068 |
if (flags & SLAB_PANIC) panic("Cannot create slabcache %s ", name); else s = NULL; return s; } EXPORT_SYMBOL(kmem_cache_create); |
81819f0fc
|
3069 |
#ifdef CONFIG_SMP |
27390bc33
|
3070 |
/* |
672bba3a4
|
3071 3072 |
* Use the cpu notifier to insure that the cpu slabs are flushed when * necessary. |
81819f0fc
|
3073 3074 3075 3076 3077 |
*/ static int __cpuinit slab_cpuup_callback(struct notifier_block *nfb, unsigned long action, void *hcpu) { long cpu = (long)hcpu; |
5b95a4acf
|
3078 3079 |
struct kmem_cache *s; unsigned long flags; |
81819f0fc
|
3080 3081 |
switch (action) { |
4c93c355d
|
3082 3083 3084 3085 3086 3087 3088 3089 3090 |
case CPU_UP_PREPARE: case CPU_UP_PREPARE_FROZEN: init_alloc_cpu_cpu(cpu); down_read(&slub_lock); list_for_each_entry(s, &slab_caches, list) s->cpu_slab[cpu] = alloc_kmem_cache_cpu(s, cpu, GFP_KERNEL); up_read(&slub_lock); break; |
81819f0fc
|
3091 |
case CPU_UP_CANCELED: |
8bb784428
|
3092 |
case CPU_UP_CANCELED_FROZEN: |
81819f0fc
|
3093 |
case CPU_DEAD: |
8bb784428
|
3094 |
case CPU_DEAD_FROZEN: |
5b95a4acf
|
3095 3096 |
down_read(&slub_lock); list_for_each_entry(s, &slab_caches, list) { |
4c93c355d
|
3097 |
struct kmem_cache_cpu *c = get_cpu_slab(s, cpu); |
5b95a4acf
|
3098 3099 3100 |
local_irq_save(flags); __flush_cpu_slab(s, cpu); local_irq_restore(flags); |
4c93c355d
|
3101 3102 |
free_kmem_cache_cpu(c, cpu); s->cpu_slab[cpu] = NULL; |
5b95a4acf
|
3103 3104 |
} up_read(&slub_lock); |
81819f0fc
|
3105 3106 3107 3108 3109 3110 |
break; default: break; } return NOTIFY_OK; } |
064287807
|
3111 |
static struct notifier_block __cpuinitdata slab_notifier = { |
3adbefee6
|
3112 |
.notifier_call = slab_cpuup_callback |
064287807
|
3113 |
}; |
81819f0fc
|
3114 3115 |
#endif |
ce71e27c6
|
3116 |
void *__kmalloc_track_caller(size_t size, gfp_t gfpflags, unsigned long caller) |
81819f0fc
|
3117 |
{ |
aadb4bc4a
|
3118 |
struct kmem_cache *s; |
331dc558f
|
3119 |
if (unlikely(size > PAGE_SIZE)) |
eada35efc
|
3120 |
return kmalloc_large(size, gfpflags); |
aadb4bc4a
|
3121 |
s = get_slab(size, gfpflags); |
81819f0fc
|
3122 |
|
2408c5503
|
3123 |
if (unlikely(ZERO_OR_NULL_PTR(s))) |
6cb8f9132
|
3124 |
return s; |
81819f0fc
|
3125 |
|
ce15fea82
|
3126 |
return slab_alloc(s, gfpflags, -1, caller); |
81819f0fc
|
3127 3128 3129 |
} void *__kmalloc_node_track_caller(size_t size, gfp_t gfpflags, |
ce71e27c6
|
3130 |
int node, unsigned long caller) |
81819f0fc
|
3131 |
{ |
aadb4bc4a
|
3132 |
struct kmem_cache *s; |
331dc558f
|
3133 |
if (unlikely(size > PAGE_SIZE)) |
f619cfe1b
|
3134 |
return kmalloc_large_node(size, gfpflags, node); |
eada35efc
|
3135 |
|
aadb4bc4a
|
3136 |
s = get_slab(size, gfpflags); |
81819f0fc
|
3137 |
|
2408c5503
|
3138 |
if (unlikely(ZERO_OR_NULL_PTR(s))) |
6cb8f9132
|
3139 |
return s; |
81819f0fc
|
3140 |
|
ce15fea82
|
3141 |
return slab_alloc(s, gfpflags, node, caller); |
81819f0fc
|
3142 |
} |
f6acb6350
|
3143 |
#ifdef CONFIG_SLUB_DEBUG |
205ab99dd
|
3144 3145 |
static unsigned long count_partial(struct kmem_cache_node *n, int (*get_count)(struct page *)) |
5b06c853a
|
3146 3147 3148 3149 3150 3151 3152 |
{ unsigned long flags; unsigned long x = 0; struct page *page; spin_lock_irqsave(&n->list_lock, flags); list_for_each_entry(page, &n->partial, lru) |
205ab99dd
|
3153 |
x += get_count(page); |
5b06c853a
|
3154 3155 3156 |
spin_unlock_irqrestore(&n->list_lock, flags); return x; } |
205ab99dd
|
3157 3158 3159 3160 3161 3162 3163 3164 3165 3166 3167 3168 3169 3170 3171 |
static int count_inuse(struct page *page) { return page->inuse; } static int count_total(struct page *page) { return page->objects; } static int count_free(struct page *page) { return page->objects - page->inuse; } |
5b06c853a
|
3172 |
|
434e245dd
|
3173 3174 |
static int validate_slab(struct kmem_cache *s, struct page *page, unsigned long *map) |
53e15af03
|
3175 3176 |
{ void *p; |
a973e9dd1
|
3177 |
void *addr = page_address(page); |
53e15af03
|
3178 3179 3180 3181 3182 3183 |
if (!check_slab(s, page) || !on_freelist(s, page, NULL)) return 0; /* Now we know that a valid freelist exists */ |
39b264641
|
3184 |
bitmap_zero(map, page->objects); |
53e15af03
|
3185 |
|
7656c72b5
|
3186 3187 |
for_each_free_object(p, s, page->freelist) { set_bit(slab_index(p, s, addr), map); |
53e15af03
|
3188 3189 3190 |
if (!check_object(s, page, p, 0)) return 0; } |
224a88be4
|
3191 |
for_each_object(p, s, addr, page->objects) |
7656c72b5
|
3192 |
if (!test_bit(slab_index(p, s, addr), map)) |
53e15af03
|
3193 3194 3195 3196 |
if (!check_object(s, page, p, 1)) return 0; return 1; } |
434e245dd
|
3197 3198 |
static void validate_slab_slab(struct kmem_cache *s, struct page *page, unsigned long *map) |
53e15af03
|
3199 3200 |
{ if (slab_trylock(page)) { |
434e245dd
|
3201 |
validate_slab(s, page, map); |
53e15af03
|
3202 3203 3204 3205 3206 3207 3208 |
slab_unlock(page); } else printk(KERN_INFO "SLUB %s: Skipped busy slab 0x%p ", s->name, page); if (s->flags & DEBUG_DEFAULT_FLAGS) { |
8a38082d2
|
3209 3210 |
if (!PageSlubDebug(page)) printk(KERN_ERR "SLUB %s: SlubDebug not set " |
53e15af03
|
3211 3212 3213 |
"on slab 0x%p ", s->name, page); } else { |
8a38082d2
|
3214 3215 |
if (PageSlubDebug(page)) printk(KERN_ERR "SLUB %s: SlubDebug set on " |
53e15af03
|
3216 3217 3218 3219 |
"slab 0x%p ", s->name, page); } } |
434e245dd
|
3220 3221 |
static int validate_slab_node(struct kmem_cache *s, struct kmem_cache_node *n, unsigned long *map) |
53e15af03
|
3222 3223 3224 3225 3226 3227 3228 3229 |
{ unsigned long count = 0; struct page *page; unsigned long flags; spin_lock_irqsave(&n->list_lock, flags); list_for_each_entry(page, &n->partial, lru) { |
434e245dd
|
3230 |
validate_slab_slab(s, page, map); |
53e15af03
|
3231 3232 3233 3234 3235 3236 3237 3238 3239 3240 3241 |
count++; } if (count != n->nr_partial) printk(KERN_ERR "SLUB %s: %ld partial slabs counted but " "counter=%ld ", s->name, count, n->nr_partial); if (!(s->flags & SLAB_STORE_USER)) goto out; list_for_each_entry(page, &n->full, lru) { |
434e245dd
|
3242 |
validate_slab_slab(s, page, map); |
53e15af03
|
3243 3244 3245 3246 3247 3248 3249 3250 3251 3252 3253 3254 |
count++; } if (count != atomic_long_read(&n->nr_slabs)) printk(KERN_ERR "SLUB: %s %ld slabs counted but " "counter=%ld ", s->name, count, atomic_long_read(&n->nr_slabs)); out: spin_unlock_irqrestore(&n->list_lock, flags); return count; } |
434e245dd
|
3255 |
static long validate_slab_cache(struct kmem_cache *s) |
53e15af03
|
3256 3257 3258 |
{ int node; unsigned long count = 0; |
205ab99dd
|
3259 |
unsigned long *map = kmalloc(BITS_TO_LONGS(oo_objects(s->max)) * |
434e245dd
|
3260 3261 3262 3263 |
sizeof(unsigned long), GFP_KERNEL); if (!map) return -ENOMEM; |
53e15af03
|
3264 3265 |
flush_all(s); |
f64dc58c5
|
3266 |
for_each_node_state(node, N_NORMAL_MEMORY) { |
53e15af03
|
3267 |
struct kmem_cache_node *n = get_node(s, node); |
434e245dd
|
3268 |
count += validate_slab_node(s, n, map); |
53e15af03
|
3269 |
} |
434e245dd
|
3270 |
kfree(map); |
53e15af03
|
3271 3272 |
return count; } |
b34597090
|
3273 3274 3275 3276 3277 3278 3279 3280 3281 3282 3283 3284 3285 3286 3287 3288 3289 3290 3291 3292 3293 3294 3295 3296 3297 3298 3299 |
#ifdef SLUB_RESILIENCY_TEST static void resiliency_test(void) { u8 *p; printk(KERN_ERR "SLUB resiliency testing "); printk(KERN_ERR "----------------------- "); printk(KERN_ERR "A. Corruption after allocation "); p = kzalloc(16, GFP_KERNEL); p[16] = 0x12; printk(KERN_ERR " 1. kmalloc-16: Clobber Redzone/next pointer" " 0x12->0x%p ", p + 16); validate_slab_cache(kmalloc_caches + 4); /* Hmmm... The next two are dangerous */ p = kzalloc(32, GFP_KERNEL); p[32 + sizeof(void *)] = 0x34; printk(KERN_ERR " 2. kmalloc-32: Clobber next pointer/next slab" |
3adbefee6
|
3300 3301 3302 3303 3304 3305 |
" 0x34 -> -0x%p ", p); printk(KERN_ERR "If allocated object is overwritten then not detectable "); |
b34597090
|
3306 3307 3308 3309 3310 3311 3312 3313 3314 |
validate_slab_cache(kmalloc_caches + 5); p = kzalloc(64, GFP_KERNEL); p += 64 + (get_cycles() & 0xff) * sizeof(void *); *p = 0x56; printk(KERN_ERR " 3. kmalloc-64: corrupting random byte 0x56->0x%p ", p); |
3adbefee6
|
3315 3316 3317 3318 |
printk(KERN_ERR "If allocated object is overwritten then not detectable "); |
b34597090
|
3319 3320 3321 3322 3323 3324 3325 3326 3327 3328 3329 3330 3331 3332 3333 3334 |
validate_slab_cache(kmalloc_caches + 6); printk(KERN_ERR " B. Corruption after free "); p = kzalloc(128, GFP_KERNEL); kfree(p); *p = 0x78; printk(KERN_ERR "1. kmalloc-128: Clobber first word 0x78->0x%p ", p); validate_slab_cache(kmalloc_caches + 7); p = kzalloc(256, GFP_KERNEL); kfree(p); p[50] = 0x9a; |
3adbefee6
|
3335 3336 3337 3338 3339 |
printk(KERN_ERR " 2. kmalloc-256: Clobber 50th byte 0x9a->0x%p ", p); |
b34597090
|
3340 3341 3342 3343 3344 3345 3346 3347 3348 3349 3350 3351 3352 3353 |
validate_slab_cache(kmalloc_caches + 8); p = kzalloc(512, GFP_KERNEL); kfree(p); p[512] = 0xab; printk(KERN_ERR " 3. kmalloc-512: Clobber redzone 0xab->0x%p ", p); validate_slab_cache(kmalloc_caches + 9); } #else static void resiliency_test(void) {}; #endif |
88a420e4e
|
3354 |
/* |
672bba3a4
|
3355 |
* Generate lists of code addresses where slabcache objects are allocated |
88a420e4e
|
3356 3357 3358 3359 3360 |
* and freed. */ struct location { unsigned long count; |
ce71e27c6
|
3361 |
unsigned long addr; |
45edfa580
|
3362 3363 3364 3365 3366 |
long long sum_time; long min_time; long max_time; long min_pid; long max_pid; |
174596a0b
|
3367 |
DECLARE_BITMAP(cpus, NR_CPUS); |
45edfa580
|
3368 |
nodemask_t nodes; |
88a420e4e
|
3369 3370 3371 3372 3373 3374 3375 3376 3377 3378 3379 3380 3381 3382 |
}; struct loc_track { unsigned long max; unsigned long count; struct location *loc; }; static void free_loc_track(struct loc_track *t) { if (t->max) free_pages((unsigned long)t->loc, get_order(sizeof(struct location) * t->max)); } |
68dff6a9a
|
3383 |
static int alloc_loc_track(struct loc_track *t, unsigned long max, gfp_t flags) |
88a420e4e
|
3384 3385 3386 |
{ struct location *l; int order; |
88a420e4e
|
3387 |
order = get_order(sizeof(struct location) * max); |
68dff6a9a
|
3388 |
l = (void *)__get_free_pages(flags, order); |
88a420e4e
|
3389 3390 3391 3392 3393 3394 3395 3396 3397 3398 3399 3400 3401 |
if (!l) return 0; if (t->count) { memcpy(l, t->loc, sizeof(struct location) * t->count); free_loc_track(t); } t->max = max; t->loc = l; return 1; } static int add_location(struct loc_track *t, struct kmem_cache *s, |
45edfa580
|
3402 |
const struct track *track) |
88a420e4e
|
3403 3404 3405 |
{ long start, end, pos; struct location *l; |
ce71e27c6
|
3406 |
unsigned long caddr; |
45edfa580
|
3407 |
unsigned long age = jiffies - track->when; |
88a420e4e
|
3408 3409 3410 3411 3412 3413 3414 3415 3416 3417 3418 3419 3420 3421 3422 |
start = -1; end = t->count; for ( ; ; ) { pos = start + (end - start + 1) / 2; /* * There is nothing at "end". If we end up there * we need to add something to before end. */ if (pos == end) break; caddr = t->loc[pos].addr; |
45edfa580
|
3423 3424 3425 3426 3427 3428 3429 3430 3431 3432 3433 3434 3435 3436 3437 |
if (track->addr == caddr) { l = &t->loc[pos]; l->count++; if (track->when) { l->sum_time += age; if (age < l->min_time) l->min_time = age; if (age > l->max_time) l->max_time = age; if (track->pid < l->min_pid) l->min_pid = track->pid; if (track->pid > l->max_pid) l->max_pid = track->pid; |
174596a0b
|
3438 3439 |
cpumask_set_cpu(track->cpu, to_cpumask(l->cpus)); |
45edfa580
|
3440 3441 |
} node_set(page_to_nid(virt_to_page(track)), l->nodes); |
88a420e4e
|
3442 3443 |
return 1; } |
45edfa580
|
3444 |
if (track->addr < caddr) |
88a420e4e
|
3445 3446 3447 3448 3449 3450 |
end = pos; else start = pos; } /* |
672bba3a4
|
3451 |
* Not found. Insert new tracking element. |
88a420e4e
|
3452 |
*/ |
68dff6a9a
|
3453 |
if (t->count >= t->max && !alloc_loc_track(t, 2 * t->max, GFP_ATOMIC)) |
88a420e4e
|
3454 3455 3456 3457 3458 3459 3460 3461 |
return 0; l = t->loc + pos; if (pos < t->count) memmove(l + 1, l, (t->count - pos) * sizeof(struct location)); t->count++; l->count = 1; |
45edfa580
|
3462 3463 3464 3465 3466 3467 |
l->addr = track->addr; l->sum_time = age; l->min_time = age; l->max_time = age; l->min_pid = track->pid; l->max_pid = track->pid; |
174596a0b
|
3468 3469 |
cpumask_clear(to_cpumask(l->cpus)); cpumask_set_cpu(track->cpu, to_cpumask(l->cpus)); |
45edfa580
|
3470 3471 |
nodes_clear(l->nodes); node_set(page_to_nid(virt_to_page(track)), l->nodes); |
88a420e4e
|
3472 3473 3474 3475 3476 3477 |
return 1; } static void process_slab(struct loc_track *t, struct kmem_cache *s, struct page *page, enum track_item alloc) { |
a973e9dd1
|
3478 |
void *addr = page_address(page); |
39b264641
|
3479 |
DECLARE_BITMAP(map, page->objects); |
88a420e4e
|
3480 |
void *p; |
39b264641
|
3481 |
bitmap_zero(map, page->objects); |
7656c72b5
|
3482 3483 |
for_each_free_object(p, s, page->freelist) set_bit(slab_index(p, s, addr), map); |
88a420e4e
|
3484 |
|
224a88be4
|
3485 |
for_each_object(p, s, addr, page->objects) |
45edfa580
|
3486 3487 |
if (!test_bit(slab_index(p, s, addr), map)) add_location(t, s, get_track(s, p, alloc)); |
88a420e4e
|
3488 3489 3490 3491 3492 |
} static int list_locations(struct kmem_cache *s, char *buf, enum track_item alloc) { |
e374d4835
|
3493 |
int len = 0; |
88a420e4e
|
3494 |
unsigned long i; |
68dff6a9a
|
3495 |
struct loc_track t = { 0, 0, NULL }; |
88a420e4e
|
3496 |
int node; |
68dff6a9a
|
3497 |
if (!alloc_loc_track(&t, PAGE_SIZE / sizeof(struct location), |
ea3061d22
|
3498 |
GFP_TEMPORARY)) |
68dff6a9a
|
3499 3500 |
return sprintf(buf, "Out of memory "); |
88a420e4e
|
3501 3502 3503 |
/* Push back cpu slabs */ flush_all(s); |
f64dc58c5
|
3504 |
for_each_node_state(node, N_NORMAL_MEMORY) { |
88a420e4e
|
3505 3506 3507 |
struct kmem_cache_node *n = get_node(s, node); unsigned long flags; struct page *page; |
9e86943b6
|
3508 |
if (!atomic_long_read(&n->nr_slabs)) |
88a420e4e
|
3509 3510 3511 3512 3513 3514 3515 3516 3517 3518 3519 |
continue; spin_lock_irqsave(&n->list_lock, flags); list_for_each_entry(page, &n->partial, lru) process_slab(&t, s, page, alloc); list_for_each_entry(page, &n->full, lru) process_slab(&t, s, page, alloc); spin_unlock_irqrestore(&n->list_lock, flags); } for (i = 0; i < t.count; i++) { |
45edfa580
|
3520 |
struct location *l = &t.loc[i]; |
88a420e4e
|
3521 |
|
9c2462472
|
3522 |
if (len > PAGE_SIZE - KSYM_SYMBOL_LEN - 100) |
88a420e4e
|
3523 |
break; |
e374d4835
|
3524 |
len += sprintf(buf + len, "%7ld ", l->count); |
45edfa580
|
3525 3526 |
if (l->addr) |
e374d4835
|
3527 |
len += sprint_symbol(buf + len, (unsigned long)l->addr); |
88a420e4e
|
3528 |
else |
e374d4835
|
3529 |
len += sprintf(buf + len, "<not-available>"); |
45edfa580
|
3530 3531 |
if (l->sum_time != l->min_time) { |
e374d4835
|
3532 |
len += sprintf(buf + len, " age=%ld/%ld/%ld", |
f8bd2258e
|
3533 3534 3535 |
l->min_time, (long)div_u64(l->sum_time, l->count), l->max_time); |
45edfa580
|
3536 |
} else |
e374d4835
|
3537 |
len += sprintf(buf + len, " age=%ld", |
45edfa580
|
3538 3539 3540 |
l->min_time); if (l->min_pid != l->max_pid) |
e374d4835
|
3541 |
len += sprintf(buf + len, " pid=%ld-%ld", |
45edfa580
|
3542 3543 |
l->min_pid, l->max_pid); else |
e374d4835
|
3544 |
len += sprintf(buf + len, " pid=%ld", |
45edfa580
|
3545 |
l->min_pid); |
174596a0b
|
3546 3547 |
if (num_online_cpus() > 1 && !cpumask_empty(to_cpumask(l->cpus)) && |
e374d4835
|
3548 3549 3550 |
len < PAGE_SIZE - 60) { len += sprintf(buf + len, " cpus="); len += cpulist_scnprintf(buf + len, PAGE_SIZE - len - 50, |
174596a0b
|
3551 |
to_cpumask(l->cpus)); |
45edfa580
|
3552 |
} |
849663430
|
3553 |
if (num_online_nodes() > 1 && !nodes_empty(l->nodes) && |
e374d4835
|
3554 3555 3556 |
len < PAGE_SIZE - 60) { len += sprintf(buf + len, " nodes="); len += nodelist_scnprintf(buf + len, PAGE_SIZE - len - 50, |
45edfa580
|
3557 3558 |
l->nodes); } |
e374d4835
|
3559 3560 |
len += sprintf(buf + len, " "); |
88a420e4e
|
3561 3562 3563 3564 |
} free_loc_track(&t); if (!t.count) |
e374d4835
|
3565 3566 3567 |
len += sprintf(buf, "No data "); return len; |
88a420e4e
|
3568 |
} |
81819f0fc
|
3569 |
enum slab_stat_type { |
205ab99dd
|
3570 3571 3572 3573 3574 |
SL_ALL, /* All slabs */ SL_PARTIAL, /* Only partially allocated slabs */ SL_CPU, /* Only slabs used for cpu caches */ SL_OBJECTS, /* Determine allocated objects not slabs */ SL_TOTAL /* Determine object capacity not slabs */ |
81819f0fc
|
3575 |
}; |
205ab99dd
|
3576 |
#define SO_ALL (1 << SL_ALL) |
81819f0fc
|
3577 3578 3579 |
#define SO_PARTIAL (1 << SL_PARTIAL) #define SO_CPU (1 << SL_CPU) #define SO_OBJECTS (1 << SL_OBJECTS) |
205ab99dd
|
3580 |
#define SO_TOTAL (1 << SL_TOTAL) |
81819f0fc
|
3581 |
|
62e5c4b4d
|
3582 3583 |
static ssize_t show_slab_objects(struct kmem_cache *s, char *buf, unsigned long flags) |
81819f0fc
|
3584 3585 |
{ unsigned long total = 0; |
81819f0fc
|
3586 3587 3588 3589 3590 3591 |
int node; int x; unsigned long *nodes; unsigned long *per_cpu; nodes = kzalloc(2 * sizeof(unsigned long) * nr_node_ids, GFP_KERNEL); |
62e5c4b4d
|
3592 3593 |
if (!nodes) return -ENOMEM; |
81819f0fc
|
3594 |
per_cpu = nodes + nr_node_ids; |
205ab99dd
|
3595 3596 |
if (flags & SO_CPU) { int cpu; |
81819f0fc
|
3597 |
|
205ab99dd
|
3598 3599 |
for_each_possible_cpu(cpu) { struct kmem_cache_cpu *c = get_cpu_slab(s, cpu); |
dfb4f0960
|
3600 |
|
205ab99dd
|
3601 3602 3603 3604 3605 3606 3607 3608 |
if (!c || c->node < 0) continue; if (c->page) { if (flags & SO_TOTAL) x = c->page->objects; else if (flags & SO_OBJECTS) x = c->page->inuse; |
81819f0fc
|
3609 3610 |
else x = 1; |
205ab99dd
|
3611 |
|
81819f0fc
|
3612 |
total += x; |
205ab99dd
|
3613 |
nodes[c->node] += x; |
81819f0fc
|
3614 |
} |
205ab99dd
|
3615 |
per_cpu[c->node]++; |
81819f0fc
|
3616 3617 |
} } |
205ab99dd
|
3618 3619 3620 3621 3622 3623 3624 3625 3626 |
if (flags & SO_ALL) { for_each_node_state(node, N_NORMAL_MEMORY) { struct kmem_cache_node *n = get_node(s, node); if (flags & SO_TOTAL) x = atomic_long_read(&n->total_objects); else if (flags & SO_OBJECTS) x = atomic_long_read(&n->total_objects) - count_partial(n, count_free); |
81819f0fc
|
3627 |
|
81819f0fc
|
3628 |
else |
205ab99dd
|
3629 |
x = atomic_long_read(&n->nr_slabs); |
81819f0fc
|
3630 3631 3632 |
total += x; nodes[node] += x; } |
205ab99dd
|
3633 3634 3635 |
} else if (flags & SO_PARTIAL) { for_each_node_state(node, N_NORMAL_MEMORY) { struct kmem_cache_node *n = get_node(s, node); |
81819f0fc
|
3636 |
|
205ab99dd
|
3637 3638 3639 3640 |
if (flags & SO_TOTAL) x = count_partial(n, count_total); else if (flags & SO_OBJECTS) x = count_partial(n, count_inuse); |
81819f0fc
|
3641 |
else |
205ab99dd
|
3642 |
x = n->nr_partial; |
81819f0fc
|
3643 3644 3645 3646 |
total += x; nodes[node] += x; } } |
81819f0fc
|
3647 3648 |
x = sprintf(buf, "%lu", total); #ifdef CONFIG_NUMA |
f64dc58c5
|
3649 |
for_each_node_state(node, N_NORMAL_MEMORY) |
81819f0fc
|
3650 3651 3652 3653 3654 3655 3656 3657 3658 3659 3660 3661 |
if (nodes[node]) x += sprintf(buf + x, " N%d=%lu", node, nodes[node]); #endif kfree(nodes); return x + sprintf(buf + x, " "); } static int any_slab_objects(struct kmem_cache *s) { int node; |
81819f0fc
|
3662 |
|
dfb4f0960
|
3663 |
for_each_online_node(node) { |
81819f0fc
|
3664 |
struct kmem_cache_node *n = get_node(s, node); |
dfb4f0960
|
3665 3666 |
if (!n) continue; |
4ea33e2dc
|
3667 |
if (atomic_long_read(&n->total_objects)) |
81819f0fc
|
3668 3669 3670 3671 3672 3673 3674 3675 3676 3677 3678 3679 3680 3681 3682 3683 3684 3685 3686 3687 |
return 1; } return 0; } #define to_slab_attr(n) container_of(n, struct slab_attribute, attr) #define to_slab(n) container_of(n, struct kmem_cache, kobj); struct slab_attribute { struct attribute attr; ssize_t (*show)(struct kmem_cache *s, char *buf); ssize_t (*store)(struct kmem_cache *s, const char *x, size_t count); }; #define SLAB_ATTR_RO(_name) \ static struct slab_attribute _name##_attr = __ATTR_RO(_name) #define SLAB_ATTR(_name) \ static struct slab_attribute _name##_attr = \ __ATTR(_name, 0644, _name##_show, _name##_store) |
81819f0fc
|
3688 3689 3690 3691 3692 3693 3694 3695 3696 3697 3698 3699 3700 3701 3702 3703 3704 3705 3706 3707 3708 3709 3710 |
static ssize_t slab_size_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", s->size); } SLAB_ATTR_RO(slab_size); static ssize_t align_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", s->align); } SLAB_ATTR_RO(align); static ssize_t object_size_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", s->objsize); } SLAB_ATTR_RO(object_size); static ssize_t objs_per_slab_show(struct kmem_cache *s, char *buf) { |
834f3d119
|
3711 3712 |
return sprintf(buf, "%d ", oo_objects(s->oo)); |
81819f0fc
|
3713 3714 |
} SLAB_ATTR_RO(objs_per_slab); |
06b285dc3
|
3715 3716 3717 |
static ssize_t order_store(struct kmem_cache *s, const char *buf, size_t length) { |
0121c619d
|
3718 3719 3720 3721 3722 3723 |
unsigned long order; int err; err = strict_strtoul(buf, 10, &order); if (err) return err; |
06b285dc3
|
3724 3725 3726 3727 3728 3729 3730 |
if (order > slub_max_order || order < slub_min_order) return -EINVAL; calculate_sizes(s, order); return length; } |
81819f0fc
|
3731 3732 |
static ssize_t order_show(struct kmem_cache *s, char *buf) { |
834f3d119
|
3733 3734 |
return sprintf(buf, "%d ", oo_order(s->oo)); |
81819f0fc
|
3735 |
} |
06b285dc3
|
3736 |
SLAB_ATTR(order); |
81819f0fc
|
3737 3738 3739 3740 3741 3742 3743 3744 3745 3746 3747 3748 |
static ssize_t ctor_show(struct kmem_cache *s, char *buf) { if (s->ctor) { int n = sprint_symbol(buf, (unsigned long)s->ctor); return n + sprintf(buf + n, " "); } return 0; } SLAB_ATTR_RO(ctor); |
81819f0fc
|
3749 3750 3751 3752 3753 3754 3755 3756 3757 |
static ssize_t aliases_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", s->refcount - 1); } SLAB_ATTR_RO(aliases); static ssize_t slabs_show(struct kmem_cache *s, char *buf) { |
205ab99dd
|
3758 |
return show_slab_objects(s, buf, SO_ALL); |
81819f0fc
|
3759 3760 3761 3762 3763 |
} SLAB_ATTR_RO(slabs); static ssize_t partial_show(struct kmem_cache *s, char *buf) { |
d9acf4b7b
|
3764 |
return show_slab_objects(s, buf, SO_PARTIAL); |
81819f0fc
|
3765 3766 3767 3768 3769 |
} SLAB_ATTR_RO(partial); static ssize_t cpu_slabs_show(struct kmem_cache *s, char *buf) { |
d9acf4b7b
|
3770 |
return show_slab_objects(s, buf, SO_CPU); |
81819f0fc
|
3771 3772 3773 3774 3775 |
} SLAB_ATTR_RO(cpu_slabs); static ssize_t objects_show(struct kmem_cache *s, char *buf) { |
205ab99dd
|
3776 |
return show_slab_objects(s, buf, SO_ALL|SO_OBJECTS); |
81819f0fc
|
3777 3778 |
} SLAB_ATTR_RO(objects); |
205ab99dd
|
3779 3780 3781 3782 3783 3784 3785 3786 3787 3788 3789 |
static ssize_t objects_partial_show(struct kmem_cache *s, char *buf) { return show_slab_objects(s, buf, SO_PARTIAL|SO_OBJECTS); } SLAB_ATTR_RO(objects_partial); static ssize_t total_objects_show(struct kmem_cache *s, char *buf) { return show_slab_objects(s, buf, SO_ALL|SO_TOTAL); } SLAB_ATTR_RO(total_objects); |
81819f0fc
|
3790 3791 3792 3793 3794 3795 3796 3797 3798 3799 3800 3801 3802 3803 3804 3805 3806 3807 3808 3809 3810 3811 3812 3813 3814 3815 3816 3817 3818 3819 3820 3821 3822 3823 3824 3825 3826 3827 3828 3829 3830 3831 3832 3833 3834 3835 3836 3837 3838 3839 |
static ssize_t sanity_checks_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_DEBUG_FREE)); } static ssize_t sanity_checks_store(struct kmem_cache *s, const char *buf, size_t length) { s->flags &= ~SLAB_DEBUG_FREE; if (buf[0] == '1') s->flags |= SLAB_DEBUG_FREE; return length; } SLAB_ATTR(sanity_checks); static ssize_t trace_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_TRACE)); } static ssize_t trace_store(struct kmem_cache *s, const char *buf, size_t length) { s->flags &= ~SLAB_TRACE; if (buf[0] == '1') s->flags |= SLAB_TRACE; return length; } SLAB_ATTR(trace); static ssize_t reclaim_account_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_RECLAIM_ACCOUNT)); } static ssize_t reclaim_account_store(struct kmem_cache *s, const char *buf, size_t length) { s->flags &= ~SLAB_RECLAIM_ACCOUNT; if (buf[0] == '1') s->flags |= SLAB_RECLAIM_ACCOUNT; return length; } SLAB_ATTR(reclaim_account); static ssize_t hwcache_align_show(struct kmem_cache *s, char *buf) { |
5af608399
|
3840 3841 |
return sprintf(buf, "%d ", !!(s->flags & SLAB_HWCACHE_ALIGN)); |
81819f0fc
|
3842 3843 3844 3845 3846 3847 3848 3849 3850 3851 3852 3853 3854 3855 3856 3857 3858 3859 3860 3861 3862 3863 3864 3865 3866 3867 3868 3869 3870 3871 3872 3873 3874 3875 |
} SLAB_ATTR_RO(hwcache_align); #ifdef CONFIG_ZONE_DMA static ssize_t cache_dma_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_CACHE_DMA)); } SLAB_ATTR_RO(cache_dma); #endif static ssize_t destroy_by_rcu_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_DESTROY_BY_RCU)); } SLAB_ATTR_RO(destroy_by_rcu); static ssize_t red_zone_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_RED_ZONE)); } static ssize_t red_zone_store(struct kmem_cache *s, const char *buf, size_t length) { if (any_slab_objects(s)) return -EBUSY; s->flags &= ~SLAB_RED_ZONE; if (buf[0] == '1') s->flags |= SLAB_RED_ZONE; |
06b285dc3
|
3876 |
calculate_sizes(s, -1); |
81819f0fc
|
3877 3878 3879 3880 3881 3882 3883 3884 3885 3886 3887 3888 3889 3890 3891 3892 3893 3894 3895 |
return length; } SLAB_ATTR(red_zone); static ssize_t poison_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_POISON)); } static ssize_t poison_store(struct kmem_cache *s, const char *buf, size_t length) { if (any_slab_objects(s)) return -EBUSY; s->flags &= ~SLAB_POISON; if (buf[0] == '1') s->flags |= SLAB_POISON; |
06b285dc3
|
3896 |
calculate_sizes(s, -1); |
81819f0fc
|
3897 3898 3899 3900 3901 3902 3903 3904 3905 3906 3907 3908 3909 3910 3911 3912 3913 3914 3915 |
return length; } SLAB_ATTR(poison); static ssize_t store_user_show(struct kmem_cache *s, char *buf) { return sprintf(buf, "%d ", !!(s->flags & SLAB_STORE_USER)); } static ssize_t store_user_store(struct kmem_cache *s, const char *buf, size_t length) { if (any_slab_objects(s)) return -EBUSY; s->flags &= ~SLAB_STORE_USER; if (buf[0] == '1') s->flags |= SLAB_STORE_USER; |
06b285dc3
|
3916 |
calculate_sizes(s, -1); |
81819f0fc
|
3917 3918 3919 |
return length; } SLAB_ATTR(store_user); |
53e15af03
|
3920 3921 3922 3923 3924 3925 3926 3927 |
static ssize_t validate_show(struct kmem_cache *s, char *buf) { return 0; } static ssize_t validate_store(struct kmem_cache *s, const char *buf, size_t length) { |
434e245dd
|
3928 3929 3930 3931 3932 3933 3934 3935 |
int ret = -EINVAL; if (buf[0] == '1') { ret = validate_slab_cache(s); if (ret >= 0) ret = length; } return ret; |
53e15af03
|
3936 3937 |
} SLAB_ATTR(validate); |
2086d26a0
|
3938 3939 3940 3941 3942 3943 3944 3945 3946 3947 3948 3949 3950 3951 3952 3953 3954 3955 |
static ssize_t shrink_show(struct kmem_cache *s, char *buf) { return 0; } static ssize_t shrink_store(struct kmem_cache *s, const char *buf, size_t length) { if (buf[0] == '1') { int rc = kmem_cache_shrink(s); if (rc) return rc; } else return -EINVAL; return length; } SLAB_ATTR(shrink); |
88a420e4e
|
3956 3957 3958 3959 3960 3961 3962 3963 3964 3965 3966 3967 3968 3969 3970 |
static ssize_t alloc_calls_show(struct kmem_cache *s, char *buf) { if (!(s->flags & SLAB_STORE_USER)) return -ENOSYS; return list_locations(s, buf, TRACK_ALLOC); } SLAB_ATTR_RO(alloc_calls); static ssize_t free_calls_show(struct kmem_cache *s, char *buf) { if (!(s->flags & SLAB_STORE_USER)) return -ENOSYS; return list_locations(s, buf, TRACK_FREE); } SLAB_ATTR_RO(free_calls); |
81819f0fc
|
3971 |
#ifdef CONFIG_NUMA |
9824601ea
|
3972 |
static ssize_t remote_node_defrag_ratio_show(struct kmem_cache *s, char *buf) |
81819f0fc
|
3973 |
{ |
9824601ea
|
3974 3975 |
return sprintf(buf, "%d ", s->remote_node_defrag_ratio / 10); |
81819f0fc
|
3976 |
} |
9824601ea
|
3977 |
static ssize_t remote_node_defrag_ratio_store(struct kmem_cache *s, |
81819f0fc
|
3978 3979 |
const char *buf, size_t length) { |
0121c619d
|
3980 3981 3982 3983 3984 3985 |
unsigned long ratio; int err; err = strict_strtoul(buf, 10, &ratio); if (err) return err; |
e2cb96b7e
|
3986 |
if (ratio <= 100) |
0121c619d
|
3987 |
s->remote_node_defrag_ratio = ratio * 10; |
81819f0fc
|
3988 |
|
81819f0fc
|
3989 3990 |
return length; } |
9824601ea
|
3991 |
SLAB_ATTR(remote_node_defrag_ratio); |
81819f0fc
|
3992 |
#endif |
8ff12cfc0
|
3993 |
#ifdef CONFIG_SLUB_STATS |
8ff12cfc0
|
3994 3995 3996 3997 3998 3999 4000 4001 4002 4003 4004 4005 4006 4007 4008 4009 4010 4011 |
static int show_stat(struct kmem_cache *s, char *buf, enum stat_item si) { unsigned long sum = 0; int cpu; int len; int *data = kmalloc(nr_cpu_ids * sizeof(int), GFP_KERNEL); if (!data) return -ENOMEM; for_each_online_cpu(cpu) { unsigned x = get_cpu_slab(s, cpu)->stat[si]; data[cpu] = x; sum += x; } len = sprintf(buf, "%lu", sum); |
50ef37b96
|
4012 |
#ifdef CONFIG_SMP |
8ff12cfc0
|
4013 4014 |
for_each_online_cpu(cpu) { if (data[cpu] && len < PAGE_SIZE - 20) |
50ef37b96
|
4015 |
len += sprintf(buf + len, " C%d=%u", cpu, data[cpu]); |
8ff12cfc0
|
4016 |
} |
50ef37b96
|
4017 |
#endif |
8ff12cfc0
|
4018 4019 4020 4021 4022 4023 4024 4025 4026 4027 4028 4029 4030 4031 4032 4033 4034 4035 4036 4037 4038 4039 4040 4041 4042 4043 4044 4045 4046 |
kfree(data); return len + sprintf(buf + len, " "); } #define STAT_ATTR(si, text) \ static ssize_t text##_show(struct kmem_cache *s, char *buf) \ { \ return show_stat(s, buf, si); \ } \ SLAB_ATTR_RO(text); \ STAT_ATTR(ALLOC_FASTPATH, alloc_fastpath); STAT_ATTR(ALLOC_SLOWPATH, alloc_slowpath); STAT_ATTR(FREE_FASTPATH, free_fastpath); STAT_ATTR(FREE_SLOWPATH, free_slowpath); STAT_ATTR(FREE_FROZEN, free_frozen); STAT_ATTR(FREE_ADD_PARTIAL, free_add_partial); STAT_ATTR(FREE_REMOVE_PARTIAL, free_remove_partial); STAT_ATTR(ALLOC_FROM_PARTIAL, alloc_from_partial); STAT_ATTR(ALLOC_SLAB, alloc_slab); STAT_ATTR(ALLOC_REFILL, alloc_refill); STAT_ATTR(FREE_SLAB, free_slab); STAT_ATTR(CPUSLAB_FLUSH, cpuslab_flush); STAT_ATTR(DEACTIVATE_FULL, deactivate_full); STAT_ATTR(DEACTIVATE_EMPTY, deactivate_empty); STAT_ATTR(DEACTIVATE_TO_HEAD, deactivate_to_head); STAT_ATTR(DEACTIVATE_TO_TAIL, deactivate_to_tail); STAT_ATTR(DEACTIVATE_REMOTE_FREES, deactivate_remote_frees); |
65c3376aa
|
4047 |
STAT_ATTR(ORDER_FALLBACK, order_fallback); |
8ff12cfc0
|
4048 |
#endif |
064287807
|
4049 |
static struct attribute *slab_attrs[] = { |
81819f0fc
|
4050 4051 4052 4053 4054 |
&slab_size_attr.attr, &object_size_attr.attr, &objs_per_slab_attr.attr, &order_attr.attr, &objects_attr.attr, |
205ab99dd
|
4055 4056 |
&objects_partial_attr.attr, &total_objects_attr.attr, |
81819f0fc
|
4057 4058 4059 4060 |
&slabs_attr.attr, &partial_attr.attr, &cpu_slabs_attr.attr, &ctor_attr.attr, |
81819f0fc
|
4061 4062 4063 4064 4065 4066 4067 4068 4069 4070 |
&aliases_attr.attr, &align_attr.attr, &sanity_checks_attr.attr, &trace_attr.attr, &hwcache_align_attr.attr, &reclaim_account_attr.attr, &destroy_by_rcu_attr.attr, &red_zone_attr.attr, &poison_attr.attr, &store_user_attr.attr, |
53e15af03
|
4071 |
&validate_attr.attr, |
2086d26a0
|
4072 |
&shrink_attr.attr, |
88a420e4e
|
4073 4074 |
&alloc_calls_attr.attr, &free_calls_attr.attr, |
81819f0fc
|
4075 4076 4077 4078 |
#ifdef CONFIG_ZONE_DMA &cache_dma_attr.attr, #endif #ifdef CONFIG_NUMA |
9824601ea
|
4079 |
&remote_node_defrag_ratio_attr.attr, |
81819f0fc
|
4080 |
#endif |
8ff12cfc0
|
4081 4082 4083 4084 4085 4086 4087 4088 4089 4090 4091 4092 4093 4094 4095 4096 4097 4098 |
#ifdef CONFIG_SLUB_STATS &alloc_fastpath_attr.attr, &alloc_slowpath_attr.attr, &free_fastpath_attr.attr, &free_slowpath_attr.attr, &free_frozen_attr.attr, &free_add_partial_attr.attr, &free_remove_partial_attr.attr, &alloc_from_partial_attr.attr, &alloc_slab_attr.attr, &alloc_refill_attr.attr, &free_slab_attr.attr, &cpuslab_flush_attr.attr, &deactivate_full_attr.attr, &deactivate_empty_attr.attr, &deactivate_to_head_attr.attr, &deactivate_to_tail_attr.attr, &deactivate_remote_frees_attr.attr, |
65c3376aa
|
4099 |
&order_fallback_attr.attr, |
8ff12cfc0
|
4100 |
#endif |
81819f0fc
|
4101 4102 4103 4104 4105 4106 4107 4108 4109 4110 4111 4112 4113 4114 4115 4116 4117 4118 4119 4120 4121 4122 4123 4124 4125 4126 4127 4128 4129 4130 4131 4132 4133 4134 4135 4136 4137 4138 4139 4140 4141 4142 4143 4144 |
NULL }; static struct attribute_group slab_attr_group = { .attrs = slab_attrs, }; static ssize_t slab_attr_show(struct kobject *kobj, struct attribute *attr, char *buf) { struct slab_attribute *attribute; struct kmem_cache *s; int err; attribute = to_slab_attr(attr); s = to_slab(kobj); if (!attribute->show) return -EIO; err = attribute->show(s, buf); return err; } static ssize_t slab_attr_store(struct kobject *kobj, struct attribute *attr, const char *buf, size_t len) { struct slab_attribute *attribute; struct kmem_cache *s; int err; attribute = to_slab_attr(attr); s = to_slab(kobj); if (!attribute->store) return -EIO; err = attribute->store(s, buf, len); return err; } |
151c602f7
|
4145 4146 4147 4148 4149 4150 |
static void kmem_cache_release(struct kobject *kobj) { struct kmem_cache *s = to_slab(kobj); kfree(s); } |
81819f0fc
|
4151 4152 4153 4154 4155 4156 4157 |
static struct sysfs_ops slab_sysfs_ops = { .show = slab_attr_show, .store = slab_attr_store, }; static struct kobj_type slab_ktype = { .sysfs_ops = &slab_sysfs_ops, |
151c602f7
|
4158 |
.release = kmem_cache_release |
81819f0fc
|
4159 4160 4161 4162 4163 4164 4165 4166 4167 4168 4169 4170 4171 4172 |
}; static int uevent_filter(struct kset *kset, struct kobject *kobj) { struct kobj_type *ktype = get_ktype(kobj); if (ktype == &slab_ktype) return 1; return 0; } static struct kset_uevent_ops slab_uevent_ops = { .filter = uevent_filter, }; |
27c3a314d
|
4173 |
static struct kset *slab_kset; |
81819f0fc
|
4174 4175 4176 4177 |
#define ID_STR_LENGTH 64 /* Create a unique string id for a slab cache: |
6446faa2f
|
4178 4179 |
* * Format :[flags-]size |
81819f0fc
|
4180 4181 4182 4183 4184 4185 4186 4187 4188 4189 4190 4191 4192 4193 4194 4195 4196 4197 4198 4199 4200 4201 4202 4203 4204 4205 4206 4207 4208 4209 4210 4211 4212 4213 4214 4215 4216 4217 4218 4219 4220 4221 4222 4223 4224 4225 |
*/ static char *create_unique_id(struct kmem_cache *s) { char *name = kmalloc(ID_STR_LENGTH, GFP_KERNEL); char *p = name; BUG_ON(!name); *p++ = ':'; /* * First flags affecting slabcache operations. We will only * get here for aliasable slabs so we do not need to support * too many flags. The flags here must cover all flags that * are matched during merging to guarantee that the id is * unique. */ if (s->flags & SLAB_CACHE_DMA) *p++ = 'd'; if (s->flags & SLAB_RECLAIM_ACCOUNT) *p++ = 'a'; if (s->flags & SLAB_DEBUG_FREE) *p++ = 'F'; if (p != name + 1) *p++ = '-'; p += sprintf(p, "%07d", s->size); BUG_ON(p > name + ID_STR_LENGTH - 1); return name; } static int sysfs_slab_add(struct kmem_cache *s) { int err; const char *name; int unmergeable; if (slab_state < SYSFS) /* Defer until later */ return 0; unmergeable = slab_unmergeable(s); if (unmergeable) { /* * Slabcache can never be merged so we can use the name proper. * This is typically the case for debug situations. In that * case we can catch duplicate names easily. */ |
27c3a314d
|
4226 |
sysfs_remove_link(&slab_kset->kobj, s->name); |
81819f0fc
|
4227 4228 4229 4230 4231 4232 4233 4234 |
name = s->name; } else { /* * Create a unique name for the slab as a target * for the symlinks. */ name = create_unique_id(s); } |
27c3a314d
|
4235 |
s->kobj.kset = slab_kset; |
1eada11c8
|
4236 4237 4238 |
err = kobject_init_and_add(&s->kobj, &slab_ktype, NULL, name); if (err) { kobject_put(&s->kobj); |
81819f0fc
|
4239 |
return err; |
1eada11c8
|
4240 |
} |
81819f0fc
|
4241 4242 4243 4244 4245 4246 4247 4248 4249 4250 4251 4252 4253 4254 4255 4256 4257 |
err = sysfs_create_group(&s->kobj, &slab_attr_group); if (err) return err; kobject_uevent(&s->kobj, KOBJ_ADD); if (!unmergeable) { /* Setup first alias */ sysfs_slab_alias(s, s->name); kfree(name); } return 0; } static void sysfs_slab_remove(struct kmem_cache *s) { kobject_uevent(&s->kobj, KOBJ_REMOVE); kobject_del(&s->kobj); |
151c602f7
|
4258 |
kobject_put(&s->kobj); |
81819f0fc
|
4259 4260 4261 4262 |
} /* * Need to buffer aliases during bootup until sysfs becomes |
9f6c708e5
|
4263 |
* available lest we lose that information. |
81819f0fc
|
4264 4265 4266 4267 4268 4269 |
*/ struct saved_alias { struct kmem_cache *s; const char *name; struct saved_alias *next; }; |
5af328a51
|
4270 |
static struct saved_alias *alias_list; |
81819f0fc
|
4271 4272 4273 4274 4275 4276 4277 4278 4279 |
static int sysfs_slab_alias(struct kmem_cache *s, const char *name) { struct saved_alias *al; if (slab_state == SYSFS) { /* * If we have a leftover link then remove it. */ |
27c3a314d
|
4280 4281 |
sysfs_remove_link(&slab_kset->kobj, name); return sysfs_create_link(&slab_kset->kobj, &s->kobj, name); |
81819f0fc
|
4282 4283 4284 4285 4286 4287 4288 4289 4290 4291 4292 4293 4294 4295 4296 |
} al = kmalloc(sizeof(struct saved_alias), GFP_KERNEL); if (!al) return -ENOMEM; al->s = s; al->name = name; al->next = alias_list; alias_list = al; return 0; } static int __init slab_sysfs_init(void) { |
5b95a4acf
|
4297 |
struct kmem_cache *s; |
81819f0fc
|
4298 |
int err; |
0ff21e466
|
4299 |
slab_kset = kset_create_and_add("slab", &slab_uevent_ops, kernel_kobj); |
27c3a314d
|
4300 |
if (!slab_kset) { |
81819f0fc
|
4301 4302 4303 4304 |
printk(KERN_ERR "Cannot register slab subsystem. "); return -ENOSYS; } |
26a7bd030
|
4305 |
slab_state = SYSFS; |
5b95a4acf
|
4306 |
list_for_each_entry(s, &slab_caches, list) { |
26a7bd030
|
4307 |
err = sysfs_slab_add(s); |
5d540fb71
|
4308 4309 4310 4311 |
if (err) printk(KERN_ERR "SLUB: Unable to add boot slab %s" " to sysfs ", s->name); |
26a7bd030
|
4312 |
} |
81819f0fc
|
4313 4314 4315 4316 4317 4318 |
while (alias_list) { struct saved_alias *al = alias_list; alias_list = alias_list->next; err = sysfs_slab_alias(al->s, al->name); |
5d540fb71
|
4319 4320 4321 4322 |
if (err) printk(KERN_ERR "SLUB: Unable to add boot slab alias" " %s to sysfs ", s->name); |
81819f0fc
|
4323 4324 4325 4326 4327 4328 4329 4330 |
kfree(al); } resiliency_test(); return 0; } __initcall(slab_sysfs_init); |
81819f0fc
|
4331 |
#endif |
57ed3eda9
|
4332 4333 4334 4335 |
/* * The /proc/slabinfo ABI */ |
158a96242
|
4336 |
#ifdef CONFIG_SLABINFO |
57ed3eda9
|
4337 4338 4339 4340 4341 4342 4343 4344 4345 4346 4347 4348 4349 4350 4351 4352 4353 4354 4355 4356 4357 4358 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 4371 4372 4373 4374 |
static void print_slabinfo_header(struct seq_file *m) { seq_puts(m, "slabinfo - version: 2.1 "); seq_puts(m, "# name <active_objs> <num_objs> <objsize> " "<objperslab> <pagesperslab>"); seq_puts(m, " : tunables <limit> <batchcount> <sharedfactor>"); seq_puts(m, " : slabdata <active_slabs> <num_slabs> <sharedavail>"); seq_putc(m, ' '); } static void *s_start(struct seq_file *m, loff_t *pos) { loff_t n = *pos; down_read(&slub_lock); if (!n) print_slabinfo_header(m); return seq_list_start(&slab_caches, *pos); } static void *s_next(struct seq_file *m, void *p, loff_t *pos) { return seq_list_next(p, &slab_caches, pos); } static void s_stop(struct seq_file *m, void *p) { up_read(&slub_lock); } static int s_show(struct seq_file *m, void *p) { unsigned long nr_partials = 0; unsigned long nr_slabs = 0; unsigned long nr_inuse = 0; |
205ab99dd
|
4375 4376 |
unsigned long nr_objs = 0; unsigned long nr_free = 0; |
57ed3eda9
|
4377 4378 4379 4380 4381 4382 4383 4384 4385 4386 4387 4388 4389 |
struct kmem_cache *s; int node; s = list_entry(p, struct kmem_cache, list); for_each_online_node(node) { struct kmem_cache_node *n = get_node(s, node); if (!n) continue; nr_partials += n->nr_partial; nr_slabs += atomic_long_read(&n->nr_slabs); |
205ab99dd
|
4390 4391 |
nr_objs += atomic_long_read(&n->total_objects); nr_free += count_partial(n, count_free); |
57ed3eda9
|
4392 |
} |
205ab99dd
|
4393 |
nr_inuse = nr_objs - nr_free; |
57ed3eda9
|
4394 4395 |
seq_printf(m, "%-17s %6lu %6lu %6u %4u %4d", s->name, nr_inuse, |
834f3d119
|
4396 4397 |
nr_objs, s->size, oo_objects(s->oo), (1 << oo_order(s->oo))); |
57ed3eda9
|
4398 4399 4400 4401 4402 4403 4404 |
seq_printf(m, " : tunables %4u %4u %4u", 0, 0, 0); seq_printf(m, " : slabdata %6lu %6lu %6lu", nr_slabs, nr_slabs, 0UL); seq_putc(m, ' '); return 0; } |
7b3c3a50a
|
4405 |
static const struct seq_operations slabinfo_op = { |
57ed3eda9
|
4406 4407 4408 4409 4410 |
.start = s_start, .next = s_next, .stop = s_stop, .show = s_show, }; |
7b3c3a50a
|
4411 4412 4413 4414 4415 4416 4417 4418 4419 4420 4421 4422 4423 4424 4425 4426 4427 4428 |
static int slabinfo_open(struct inode *inode, struct file *file) { return seq_open(file, &slabinfo_op); } static const struct file_operations proc_slabinfo_operations = { .open = slabinfo_open, .read = seq_read, .llseek = seq_lseek, .release = seq_release, }; static int __init slab_proc_init(void) { proc_create("slabinfo",S_IWUSR|S_IRUGO,NULL,&proc_slabinfo_operations); return 0; } module_init(slab_proc_init); |
158a96242
|
4429 |
#endif /* CONFIG_SLABINFO */ |