mm/list_lru.c

   1 // SPDX-License-Identifier: GPL-2.0-only
   2 /*
   3  * Copyright (c) 2013 Red Hat, Inc. and Parallels Inc. All rights reserved.
   4  * Authors: David Chinner and Glauber Costa
   5  *
   6  * Generic LRU infrastructure
   7  */
   8 #include <linux/kernel.h>
   9 #include <linux/module.h>
  10 #include <linux/mm.h>
  11 #include <linux/list_lru.h>
  12 #include <linux/slab.h>
  13 #include <linux/mutex.h>
  14 #include <linux/memcontrol.h>
  15 #include "slab.h"
  16
  17 #ifdef CONFIG_MEMCG_KMEM
  18 static LIST_HEAD(list_lrus);
  19 static DEFINE_MUTEX(list_lrus_mutex);
  20
  21 static void list_lru_register(struct list_lru *lru)
  22 {
  23         mutex_lock(&list_lrus_mutex);
  24         list_add(&lru->list, &list_lrus);
  25         mutex_unlock(&list_lrus_mutex);
  26 }
  27
  28 static void list_lru_unregister(struct list_lru *lru)
  29 {
  30         mutex_lock(&list_lrus_mutex);
  31         list_del(&lru->list);
  32         mutex_unlock(&list_lrus_mutex);
  33 }
  34
  35 static int lru_shrinker_id(struct list_lru *lru)
  36 {
  37         return lru->shrinker_id;
  38 }
  39
  40 static inline bool list_lru_memcg_aware(struct list_lru *lru)
  41 {
  42         return lru->memcg_aware;
  43 }
  44
  45 static inline struct list_lru_one *
  46 list_lru_from_memcg_idx(struct list_lru_node *nlru, int idx)
  47 {
  48         struct list_lru_memcg *memcg_lrus;
  49         /*
  50          * Either lock or RCU protects the array of per cgroup lists
  51          * from relocation (see memcg_update_list_lru_node).
  52          */
  53         memcg_lrus = rcu_dereference_check(nlru->memcg_lrus,
  54                                            lockdep_is_held(&nlru->lock));
  55         if (memcg_lrus && idx >= 0)
  56                 return memcg_lrus->lru[idx];
  57         return &nlru->lru;
  58 }
  59
  60 static __always_inline struct mem_cgroup *mem_cgroup_from_kmem(void *ptr)
  61 {
  62         struct page *page;
  63
  64         if (!memcg_kmem_enabled())
  65                 return NULL;
  66         page = virt_to_head_page(ptr);
  67         return memcg_from_slab_page(page);
  68 }
  69
  70 static inline struct list_lru_one *
  71 list_lru_from_kmem(struct list_lru_node *nlru, void *ptr,
  72                    struct mem_cgroup **memcg_ptr)
  73 {
  74         struct list_lru_one *l = &nlru->lru;
  75         struct mem_cgroup *memcg = NULL;
  76
  77         if (!nlru->memcg_lrus)
  78                 goto out;
  79
  80         memcg = mem_cgroup_from_kmem(ptr);
  81         if (!memcg)
  82                 goto out;
  83
  84         l = list_lru_from_memcg_idx(nlru, memcg_cache_id(memcg));
  85 out:
  86         if (memcg_ptr)
  87                 *memcg_ptr = memcg;
  88         return l;
  89 }
  90 #else
  91 static void list_lru_register(struct list_lru *lru)
  92 {
  93 }
  94
  95 static void list_lru_unregister(struct list_lru *lru)
  96 {
  97 }
  98
  99 static int lru_shrinker_id(struct list_lru *lru)
 100 {
 101         return -1;
 102 }
 103
 104 static inline bool list_lru_memcg_aware(struct list_lru *lru)
 105 {
 106         return false;
 107 }
 108
 109 static inline struct list_lru_one *
 110 list_lru_from_memcg_idx(struct list_lru_node *nlru, int idx)
 111 {
 112         return &nlru->lru;
 113 }
 114
 115 static inline struct list_lru_one *
 116 list_lru_from_kmem(struct list_lru_node *nlru, void *ptr,
 117                    struct mem_cgroup **memcg_ptr)
 118 {
 119         if (memcg_ptr)
 120                 *memcg_ptr = NULL;
 121         return &nlru->lru;
 122 }
 123 #endif /* CONFIG_MEMCG_KMEM */
 124
 125 bool list_lru_add(struct list_lru *lru, struct list_head *item)
 126 {
 127         int nid = page_to_nid(virt_to_page(item));
 128         struct list_lru_node *nlru = &lru->node[nid];
 129         struct mem_cgroup *memcg;
 130         struct list_lru_one *l;
 131
 132         spin_lock(&nlru->lock);
 133         if (list_empty(item)) {
 134                 l = list_lru_from_kmem(nlru, item, &memcg);
 135                 list_add_tail(item, &l->list);
 136                 /* Set shrinker bit if the first element was added */
 137                 if (!l->nr_items++)
 138                         memcg_set_shrinker_bit(memcg, nid,
 139                                                lru_shrinker_id(lru));
 140                 nlru->nr_items++;
 141                 spin_unlock(&nlru->lock);
 142                 return true;
 143         }
 144         spin_unlock(&nlru->lock);
 145         return false;
 146 }
 147 EXPORT_SYMBOL_GPL(list_lru_add);
 148
 149 bool list_lru_del(struct list_lru *lru, struct list_head *item)
 150 {
 151         int nid = page_to_nid(virt_to_page(item));
 152         struct list_lru_node *nlru = &lru->node[nid];
 153         struct list_lru_one *l;
 154
 155         spin_lock(&nlru->lock);
 156         if (!list_empty(item)) {
 157                 l = list_lru_from_kmem(nlru, item, NULL);
 158                 list_del_init(item);
 159                 l->nr_items--;
 160                 nlru->nr_items--;
 161                 spin_unlock(&nlru->lock);
 162                 return true;
 163         }
 164         spin_unlock(&nlru->lock);
 165         return false;
 166 }
 167 EXPORT_SYMBOL_GPL(list_lru_del);
 168
 169 void list_lru_isolate(struct list_lru_one *list, struct list_head *item)
 170 {
 171         list_del_init(item);
 172         list->nr_items--;
 173 }
 174 EXPORT_SYMBOL_GPL(list_lru_isolate);
 175
 176 void list_lru_isolate_move(struct list_lru_one *list, struct list_head *item,
 177                            struct list_head *head)
 178 {
 179         list_move(item, head);
 180         list->nr_items--;
 181 }
 182 EXPORT_SYMBOL_GPL(list_lru_isolate_move);
 183
 184 unsigned long list_lru_count_one(struct list_lru *lru,
 185                                  int nid, struct mem_cgroup *memcg)
 186 {
 187         struct list_lru_node *nlru = &lru->node[nid];
 188         struct list_lru_one *l;
 189         unsigned long count;
 190
 191         rcu_read_lock();
 192         l = list_lru_from_memcg_idx(nlru, memcg_cache_id(memcg));
 193         count = l->nr_items;
 194         rcu_read_unlock();
 195
 196         return count;
 197 }
 198 EXPORT_SYMBOL_GPL(list_lru_count_one);
 199
 200 unsigned long list_lru_count_node(struct list_lru *lru, int nid)
 201 {
 202         struct list_lru_node *nlru;
 203
 204         nlru = &lru->node[nid];
 205         return nlru->nr_items;
 206 }
 207 EXPORT_SYMBOL_GPL(list_lru_count_node);
 208
 209 static unsigned long
 210 __list_lru_walk_one(struct list_lru_node *nlru, int memcg_idx,
 211                     list_lru_walk_cb isolate, void *cb_arg,
 212                     unsigned long *nr_to_walk)
 213 {
 214
 215         struct list_lru_one *l;
 216         struct list_head *item, *n;
 217         unsigned long isolated = 0;
 218
 219         l = list_lru_from_memcg_idx(nlru, memcg_idx);
 220 restart:
 221         list_for_each_safe(item, n, &l->list) {
 222                 enum lru_status ret;
 223
 224                 /*
 225                  * decrement nr_to_walk first so that we don't livelock if we
 226                  * get stuck on large numbesr of LRU_RETRY items
 227                  */
 228                 if (!*nr_to_walk)
 229                         break;
 230                 --*nr_to_walk;
 231
 232                 ret = isolate(item, l, &nlru->lock, cb_arg);
 233                 switch (ret) {
 234                 case LRU_REMOVED_RETRY:
 235                         assert_spin_locked(&nlru->lock);
 236                         /* fall through */
 237                 case LRU_REMOVED:
 238                         isolated++;
 239                         nlru->nr_items--;
 240                         /*
 241                          * If the lru lock has been dropped, our list
 242                          * traversal is now invalid and so we have to
 243                          * restart from scratch.
 244                          */
 245                         if (ret == LRU_REMOVED_RETRY)
 246                                 goto restart;
 247                         break;
 248                 case LRU_ROTATE:
 249                         list_move_tail(item, &l->list);
 250                         break;
 251                 case LRU_SKIP:
 252                         break;
 253                 case LRU_RETRY:
 254                         /*
 255                          * The lru lock has been dropped, our list traversal is
 256                          * now invalid and so we have to restart from scratch.
 257                          */
 258                         assert_spin_locked(&nlru->lock);
 259                         goto restart;
 260                 default:
 261                         BUG();
 262                 }
 263         }
 264         return isolated;
 265 }
 266
 267 unsigned long
 268 list_lru_walk_one(struct list_lru *lru, int nid, struct mem_cgroup *memcg,
 269                   list_lru_walk_cb isolate, void *cb_arg,
 270                   unsigned long *nr_to_walk)
 271 {
 272         struct list_lru_node *nlru = &lru->node[nid];
 273         unsigned long ret;
 274
 275         spin_lock(&nlru->lock);
 276         ret = __list_lru_walk_one(nlru, memcg_cache_id(memcg), isolate, cb_arg,
 277                                   nr_to_walk);
 278         spin_unlock(&nlru->lock);
 279         return ret;
 280 }
 281 EXPORT_SYMBOL_GPL(list_lru_walk_one);
 282
 283 unsigned long
 284 list_lru_walk_one_irq(struct list_lru *lru, int nid, struct mem_cgroup *memcg,
 285                       list_lru_walk_cb isolate, void *cb_arg,
 286                       unsigned long *nr_to_walk)
 287 {
 288         struct list_lru_node *nlru = &lru->node[nid];
 289         unsigned long ret;
 290
 291         spin_lock_irq(&nlru->lock);
 292         ret = __list_lru_walk_one(nlru, memcg_cache_id(memcg), isolate, cb_arg,
 293                                   nr_to_walk);
 294         spin_unlock_irq(&nlru->lock);
 295         return ret;
 296 }
 297
 298 unsigned long list_lru_walk_node(struct list_lru *lru, int nid,
 299                                  list_lru_walk_cb isolate, void *cb_arg,
 300                                  unsigned long *nr_to_walk)
 301 {
 302         long isolated = 0;
 303         int memcg_idx;
 304
 305         isolated += list_lru_walk_one(lru, nid, NULL, isolate, cb_arg,
 306                                       nr_to_walk);
 307         if (*nr_to_walk > 0 && list_lru_memcg_aware(lru)) {
 308                 for_each_memcg_cache_index(memcg_idx) {
 309                         struct list_lru_node *nlru = &lru->node[nid];
 310
 311                         spin_lock(&nlru->lock);
 312                         isolated += __list_lru_walk_one(nlru, memcg_idx,
 313                                                         isolate, cb_arg,
 314                                                         nr_to_walk);
 315                         spin_unlock(&nlru->lock);
 316
 317                         if (*nr_to_walk <= 0)
 318                                 break;
 319                 }
 320         }
 321         return isolated;
 322 }
 323 EXPORT_SYMBOL_GPL(list_lru_walk_node);
 324
 325 static void init_one_lru(struct list_lru_one *l)
 326 {
 327         INIT_LIST_HEAD(&l->list);
 328         l->nr_items = 0;
 329 }
 330
 331 #ifdef CONFIG_MEMCG_KMEM
 332 static void __memcg_destroy_list_lru_node(struct list_lru_memcg *memcg_lrus,
 333                                           int begin, int end)
 334 {
 335         int i;
 336
 337         for (i = begin; i < end; i++)
 338                 kfree(memcg_lrus->lru[i]);
 339 }
 340
 341 static int __memcg_init_list_lru_node(struct list_lru_memcg *memcg_lrus,
 342                                       int begin, int end)
 343 {
 344         int i;
 345
 346         for (i = begin; i < end; i++) {
 347                 struct list_lru_one *l;
 348
 349                 l = kmalloc(sizeof(struct list_lru_one), GFP_KERNEL);
 350                 if (!l)
 351                         goto fail;
 352
 353                 init_one_lru(l);
 354                 memcg_lrus->lru[i] = l;
 355         }
 356         return 0;
 357 fail:
 358         __memcg_destroy_list_lru_node(memcg_lrus, begin, i);
 359         return -ENOMEM;
 360 }
 361
 362 static int memcg_init_list_lru_node(struct list_lru_node *nlru)
 363 {
 364         struct list_lru_memcg *memcg_lrus;
 365         int size = memcg_nr_cache_ids;
 366
 367         memcg_lrus = kvmalloc(sizeof(*memcg_lrus) +
 368                               size * sizeof(void *), GFP_KERNEL);
 369         if (!memcg_lrus)
 370                 return -ENOMEM;
 371
 372         if (__memcg_init_list_lru_node(memcg_lrus, 0, size)) {
 373                 kvfree(memcg_lrus);
 374                 return -ENOMEM;
 375         }
 376         RCU_INIT_POINTER(nlru->memcg_lrus, memcg_lrus);
 377
 378         return 0;
 379 }
 380
 381 static void memcg_destroy_list_lru_node(struct list_lru_node *nlru)
 382 {
 383         struct list_lru_memcg *memcg_lrus;
 384         /*
 385          * This is called when shrinker has already been unregistered,
 386          * and nobody can use it. So, there is no need to use kvfree_rcu().
 387          */
 388         memcg_lrus = rcu_dereference_protected(nlru->memcg_lrus, true);
 389         __memcg_destroy_list_lru_node(memcg_lrus, 0, memcg_nr_cache_ids);
 390         kvfree(memcg_lrus);
 391 }
 392
 393 static void kvfree_rcu(struct rcu_head *head)
 394 {
 395         struct list_lru_memcg *mlru;
 396
 397         mlru = container_of(head, struct list_lru_memcg, rcu);
 398         kvfree(mlru);
 399 }
 400
 401 static int memcg_update_list_lru_node(struct list_lru_node *nlru,
 402                                       int old_size, int new_size)
 403 {
 404         struct list_lru_memcg *old, *new;
 405
 406         BUG_ON(old_size > new_size);
 407
 408         old = rcu_dereference_protected(nlru->memcg_lrus,
 409                                         lockdep_is_held(&list_lrus_mutex));
 410         new = kvmalloc(sizeof(*new) + new_size * sizeof(void *), GFP_KERNEL);
 411         if (!new)
 412                 return -ENOMEM;
 413
 414         if (__memcg_init_list_lru_node(new, old_size, new_size)) {
 415                 kvfree(new);
 416                 return -ENOMEM;
 417         }
 418
 419         memcpy(&new->lru, &old->lru, old_size * sizeof(void *));
 420
 421         /*
 422          * The locking below allows readers that hold nlru->lock avoid taking
 423          * rcu_read_lock (see list_lru_from_memcg_idx).
 424          *
 425          * Since list_lru_{add,del} may be called under an IRQ-safe lock,
 426          * we have to use IRQ-safe primitives here to avoid deadlock.
 427          */
 428         spin_lock_irq(&nlru->lock);
 429         rcu_assign_pointer(nlru->memcg_lrus, new);
 430         spin_unlock_irq(&nlru->lock);
 431
 432         call_rcu(&old->rcu, kvfree_rcu);
 433         return 0;
 434 }
 435
 436 static void memcg_cancel_update_list_lru_node(struct list_lru_node *nlru,
 437                                               int old_size, int new_size)
 438 {
 439         struct list_lru_memcg *memcg_lrus;
 440
 441         memcg_lrus = rcu_dereference_protected(nlru->memcg_lrus,
 442                                                lockdep_is_held(&list_lrus_mutex));
 443         /* do not bother shrinking the array back to the old size, because we
 444          * cannot handle allocation failures here */
 445         __memcg_destroy_list_lru_node(memcg_lrus, old_size, new_size);
 446 }
 447
 448 static int memcg_init_list_lru(struct list_lru *lru, bool memcg_aware)
 449 {
 450         int i;
 451
 452         lru->memcg_aware = memcg_aware;
 453
 454         if (!memcg_aware)
 455                 return 0;
 456
 457         for_each_node(i) {
 458                 if (memcg_init_list_lru_node(&lru->node[i]))
 459                         goto fail;
 460         }
 461         return 0;
 462 fail:
 463         for (i = i - 1; i >= 0; i--) {
 464                 if (!lru->node[i].memcg_lrus)
 465                         continue;
 466                 memcg_destroy_list_lru_node(&lru->node[i]);
 467         }
 468         return -ENOMEM;
 469 }
 470
 471 static void memcg_destroy_list_lru(struct list_lru *lru)
 472 {
 473         int i;
 474
 475         if (!list_lru_memcg_aware(lru))
 476                 return;
 477
 478         for_each_node(i)
 479                 memcg_destroy_list_lru_node(&lru->node[i]);
 480 }
 481
 482 static int memcg_update_list_lru(struct list_lru *lru,
 483                                  int old_size, int new_size)
 484 {
 485         int i;
 486
 487         if (!list_lru_memcg_aware(lru))
 488                 return 0;
 489
 490         for_each_node(i) {
 491                 if (memcg_update_list_lru_node(&lru->node[i],
 492                                                old_size, new_size))
 493                         goto fail;
 494         }
 495         return 0;
 496 fail:
 497         for (i = i - 1; i >= 0; i--) {
 498                 if (!lru->node[i].memcg_lrus)
 499                         continue;
 500
 501                 memcg_cancel_update_list_lru_node(&lru->node[i],
 502                                                   old_size, new_size);
 503         }
 504         return -ENOMEM;
 505 }
 506
 507 static void memcg_cancel_update_list_lru(struct list_lru *lru,
 508                                          int old_size, int new_size)
 509 {
 510         int i;
 511
 512         if (!list_lru_memcg_aware(lru))
 513                 return;
 514
 515         for_each_node(i)
 516                 memcg_cancel_update_list_lru_node(&lru->node[i],
 517                                                   old_size, new_size);
 518 }
 519
 520 int memcg_update_all_list_lrus(int new_size)
 521 {
 522         int ret = 0;
 523         struct list_lru *lru;
 524         int old_size = memcg_nr_cache_ids;
 525
 526         mutex_lock(&list_lrus_mutex);
 527         list_for_each_entry(lru, &list_lrus, list) {
 528                 ret = memcg_update_list_lru(lru, old_size, new_size);
 529                 if (ret)
 530                         goto fail;
 531         }
 532 out:
 533         mutex_unlock(&list_lrus_mutex);
 534         return ret;
 535 fail:
 536         list_for_each_entry_continue_reverse(lru, &list_lrus, list)
 537                 memcg_cancel_update_list_lru(lru, old_size, new_size);
 538         goto out;
 539 }
 540
 541 static void memcg_drain_list_lru_node(struct list_lru *lru, int nid,
 542                                       int src_idx, struct mem_cgroup *dst_memcg)
 543 {
 544         struct list_lru_node *nlru = &lru->node[nid];
 545         int dst_idx = dst_memcg->kmemcg_id;
 546         struct list_lru_one *src, *dst;
 547         bool set;
 548
 549         /*
 550          * Since list_lru_{add,del} may be called under an IRQ-safe lock,
 551          * we have to use IRQ-safe primitives here to avoid deadlock.
 552          */
 553         spin_lock_irq(&nlru->lock);
 554
 555         src = list_lru_from_memcg_idx(nlru, src_idx);
 556         dst = list_lru_from_memcg_idx(nlru, dst_idx);
 557
 558         list_splice_init(&src->list, &dst->list);
 559         set = (!dst->nr_items && src->nr_items);
 560         dst->nr_items += src->nr_items;
 561         if (set)
 562                 memcg_set_shrinker_bit(dst_memcg, nid, lru_shrinker_id(lru));
 563         src->nr_items = 0;
 564
 565         spin_unlock_irq(&nlru->lock);
 566 }
 567
 568 static void memcg_drain_list_lru(struct list_lru *lru,
 569                                  int src_idx, struct mem_cgroup *dst_memcg)
 570 {
 571         int i;
 572
 573         if (!list_lru_memcg_aware(lru))
 574                 return;
 575
 576         for_each_node(i)
 577                 memcg_drain_list_lru_node(lru, i, src_idx, dst_memcg);
 578 }
 579
 580 void memcg_drain_all_list_lrus(int src_idx, struct mem_cgroup *dst_memcg)
 581 {
 582         struct list_lru *lru;
 583
 584         mutex_lock(&list_lrus_mutex);
 585         list_for_each_entry(lru, &list_lrus, list)
 586                 memcg_drain_list_lru(lru, src_idx, dst_memcg);
 587         mutex_unlock(&list_lrus_mutex);
 588 }
 589 #else
 590 static int memcg_init_list_lru(struct list_lru *lru, bool memcg_aware)
 591 {
 592         return 0;
 593 }
 594
 595 static void memcg_destroy_list_lru(struct list_lru *lru)
 596 {
 597 }
 598 #endif /* CONFIG_MEMCG_KMEM */
 599
 600 int __list_lru_init(struct list_lru *lru, bool memcg_aware,
 601                     struct lock_class_key *key, struct shrinker *shrinker)
 602 {
 603         int i;
 604         int err = -ENOMEM;
 605
 606 #ifdef CONFIG_MEMCG_KMEM
 607         if (shrinker)
 608                 lru->shrinker_id = shrinker->id;
 609         else
 610                 lru->shrinker_id = -1;
 611 #endif
 612         memcg_get_cache_ids();
 613
 614         lru->node = kcalloc(nr_node_ids, sizeof(*lru->node), GFP_KERNEL);
 615         if (!lru->node)
 616                 goto out;
 617
 618         for_each_node(i) {
 619                 spin_lock_init(&lru->node[i].lock);
 620                 if (key)
 621                         lockdep_set_class(&lru->node[i].lock, key);
 622                 init_one_lru(&lru->node[i].lru);
 623         }
 624
 625         err = memcg_init_list_lru(lru, memcg_aware);
 626         if (err) {
 627                 kfree(lru->node);
 628                 /* Do this so a list_lru_destroy() doesn't crash: */
 629                 lru->node = NULL;
 630                 goto out;
 631         }
 632
 633         list_lru_register(lru);
 634 out:
 635         memcg_put_cache_ids();
 636         return err;
 637 }
 638 EXPORT_SYMBOL_GPL(__list_lru_init);
 639
 640 void list_lru_destroy(struct list_lru *lru)
 641 {
 642         /* Already destroyed or not yet initialized? */
 643         if (!lru->node)
 644                 return;
 645
 646         memcg_get_cache_ids();
 647
 648         list_lru_unregister(lru);
 649
 650         memcg_destroy_list_lru(lru);
 651         kfree(lru->node);
 652         lru->node = NULL;
 653
 654 #ifdef CONFIG_MEMCG_KMEM
 655         lru->shrinker_id = -1;
 656 #endif
 657         memcg_put_cache_ids();
 658 }
 659 EXPORT_SYMBOL_GPL(list_lru_destroy);