sys/vm/vm_radix.c

   1 /*
   2  * Copyright (c) 2013 EMC Corp.
   3  * Copyright (c) 2011 Jeffrey Roberson <jeff@freebsd.org>
   4  * Copyright (c) 2008 Mayur Shardul <mayur.shardul@gmail.com>
   5  * All rights reserved.
   6  *
   7  * Redistribution and use in source and binary forms, with or without
   8  * modification, are permitted provided that the following conditions
   9  * are met:
  10  * 1. Redistributions of source code must retain the above copyright
  11  *    notice, this list of conditions and the following disclaimer.
  12  * 2. Redistributions in binary form must reproduce the above copyright
  13  *    notice, this list of conditions and the following disclaimer in the
  14  *    documentation and/or other materials provided with the distribution.
  15  *
  16  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
  17  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
  18  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
  19  * ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
  20  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
  21  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
  22  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
  23  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
  24  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
  25  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
  26  * SUCH DAMAGE.
  27  *
  28  */
  29
  30 /*
  31  * Path-compressed radix trie implementation.
  32  * The following code is not generalized into a general purpose library
  33  * because there are way too many parameters embedded that should really
  34  * be decided by the library consumers.  At the same time, consumers
  35  * of this code must achieve highest possible performance.
  36  *
  37  * The implementation takes into account the following rationale:
  38  * - Size of the nodes should be as small as possible but still big enough
  39  *   to avoid a large maximum depth for the trie.  This is a balance
  40  *   between the necessity to not wire too much physical memory for the nodes
  41  *   and the necessity to avoid too much cache pollution during the trie
  42  *   operations.
  43  * - There is not a huge bias toward the number of lookup operations over
  44  *   the number of insert and remove operations.  This basically implies
  45  *   that optimizations supposedly helping one operation but hurting the
  46  *   other might be carefully evaluated.
  47  * - On average not many nodes are expected to be fully populated, hence
  48  *   level compression may just complicate things.
  49  */
  50
  51 #include <sys/cdefs.h>
  52 __FBSDID("$FreeBSD$");
  53
  54 #include "opt_ddb.h"
  55
  56 #include <sys/param.h>
  57 #include <sys/systm.h>
  58 #include <sys/kernel.h>
  59 #include <sys/vmmeter.h>
  60
  61 #include <vm/uma.h>
  62 #include <vm/vm.h>
  63 #include <vm/vm_param.h>
  64 #include <vm/vm_page.h>
  65 #include <vm/vm_radix.h>
  66
  67 #ifdef DDB
  68 #include <ddb/ddb.h>
  69 #endif
  70
  71 /*
  72  * These widths should allow the pointers to a node's children to fit within
  73  * a single cache line.  The extra levels from a narrow width should not be
  74  * a problem thanks to path compression.
  75  */
  76 #ifdef __LP64__
  77 #define VM_RADIX_WIDTH  4
  78 #else
  79 #define VM_RADIX_WIDTH  3
  80 #endif
  81
  82 #define VM_RADIX_COUNT  (1 << VM_RADIX_WIDTH)
  83 #define VM_RADIX_MASK   (VM_RADIX_COUNT - 1)
  84 #define VM_RADIX_LIMIT                                                  \
  85         (howmany(sizeof(vm_pindex_t) * NBBY, VM_RADIX_WIDTH) - 1)
  86
  87 /* Flag bits stored in node pointers. */
  88 #define VM_RADIX_ISLEAF 0x1
  89 #define VM_RADIX_FLAGS  0x1
  90 #define VM_RADIX_PAD    VM_RADIX_FLAGS
  91
  92 /* Returns one unit associated with specified level. */
  93 #define VM_RADIX_UNITLEVEL(lev)                                         \
  94         ((vm_pindex_t)1 << ((lev) * VM_RADIX_WIDTH))
  95
  96 struct vm_radix_node {
  97         vm_pindex_t      rn_owner;                      /* Owner of record. */
  98         uint16_t         rn_count;                      /* Valid children. */
  99         uint16_t         rn_clev;                       /* Current level. */
 100         void            *rn_child[VM_RADIX_COUNT];      /* Child nodes. */
 101 };
 102
 103 static uma_zone_t vm_radix_node_zone;
 104
 105 /*
 106  * Allocate a radix node.
 107  */
 108 static __inline struct vm_radix_node *
 109 vm_radix_node_get(vm_pindex_t owner, uint16_t count, uint16_t clevel)
 110 {
 111         struct vm_radix_node *rnode;
 112
 113         rnode = uma_zalloc(vm_radix_node_zone, M_NOWAIT | M_ZERO);
 114         if (rnode == NULL)
 115                 return (NULL);
 116         rnode->rn_owner = owner;
 117         rnode->rn_count = count;
 118         rnode->rn_clev = clevel;
 119         return (rnode);
 120 }
 121
 122 /*
 123  * Free radix node.
 124  */
 125 static __inline void
 126 vm_radix_node_put(struct vm_radix_node *rnode)
 127 {
 128
 129         uma_zfree(vm_radix_node_zone, rnode);
 130 }
 131
 132 /*
 133  * Return the position in the array for a given level.
 134  */
 135 static __inline int
 136 vm_radix_slot(vm_pindex_t index, uint16_t level)
 137 {
 138
 139         return ((index >> (level * VM_RADIX_WIDTH)) & VM_RADIX_MASK);
 140 }
 141
 142 /* Trims the key after the specified level. */
 143 static __inline vm_pindex_t
 144 vm_radix_trimkey(vm_pindex_t index, uint16_t level)
 145 {
 146         vm_pindex_t ret;
 147
 148         ret = index;
 149         if (level > 0) {
 150                 ret >>= level * VM_RADIX_WIDTH;
 151                 ret <<= level * VM_RADIX_WIDTH;
 152         }
 153         return (ret);
 154 }
 155
 156 /*
 157  * Get the root node for a radix tree.
 158  */
 159 static __inline struct vm_radix_node *
 160 vm_radix_getroot(struct vm_radix *rtree)
 161 {
 162
 163         return ((struct vm_radix_node *)rtree->rt_root);
 164 }
 165
 166 /*
 167  * Set the root node for a radix tree.
 168  */
 169 static __inline void
 170 vm_radix_setroot(struct vm_radix *rtree, struct vm_radix_node *rnode)
 171 {
 172
 173         rtree->rt_root = (uintptr_t)rnode;
 174 }
 175
 176 /*
 177  * Returns TRUE if the specified radix node is a leaf and FALSE otherwise.
 178  */
 179 static __inline boolean_t
 180 vm_radix_isleaf(struct vm_radix_node *rnode)
 181 {
 182
 183         return (((uintptr_t)rnode & VM_RADIX_ISLEAF) != 0);
 184 }
 185
 186 /*
 187  * Returns the associated page extracted from rnode.
 188  */
 189 static __inline vm_page_t
 190 vm_radix_topage(struct vm_radix_node *rnode)
 191 {
 192
 193         return ((vm_page_t)((uintptr_t)rnode & ~VM_RADIX_FLAGS));
 194 }
 195
 196 /*
 197  * Adds the page as a child of the provided node.
 198  */
 199 static __inline void
 200 vm_radix_addpage(struct vm_radix_node *rnode, vm_pindex_t index, uint16_t clev,
 201     vm_page_t page)
 202 {
 203         int slot;
 204
 205         slot = vm_radix_slot(index, clev);
 206         rnode->rn_child[slot] = (void *)((uintptr_t)page | VM_RADIX_ISLEAF);
 207 }
 208
 209 /*
 210  * Returns the slot where two keys differ.
 211  * It cannot accept 2 equal keys.
 212  */
 213 static __inline uint16_t
 214 vm_radix_keydiff(vm_pindex_t index1, vm_pindex_t index2)
 215 {
 216         uint16_t clev;
 217
 218         KASSERT(index1 != index2, ("%s: passing the same key value %jx",
 219             __func__, (uintmax_t)index1));
 220
 221         index1 ^= index2;
 222         for (clev = VM_RADIX_LIMIT;; clev--)
 223                 if (vm_radix_slot(index1, clev) != 0)
 224                         return (clev);
 225 }
 226
 227 /*
 228  * Returns TRUE if it can be determined that key does not belong to the
 229  * specified rnode.  Otherwise, returns FALSE.
 230  */
 231 static __inline boolean_t
 232 vm_radix_keybarr(struct vm_radix_node *rnode, vm_pindex_t idx)
 233 {
 234
 235         if (rnode->rn_clev < VM_RADIX_LIMIT) {
 236                 idx = vm_radix_trimkey(idx, rnode->rn_clev + 1);
 237                 return (idx != rnode->rn_owner);
 238         }
 239         return (FALSE);
 240 }
 241
 242 /*
 243  * Internal helper for vm_radix_reclaim_allnodes().
 244  * This function is recursive.
 245  */
 246 static void
 247 vm_radix_reclaim_allnodes_int(struct vm_radix_node *rnode)
 248 {
 249         int slot;
 250
 251         KASSERT(rnode->rn_count <= VM_RADIX_COUNT,
 252             ("vm_radix_reclaim_allnodes_int: bad count in rnode %p", rnode));
 253         for (slot = 0; rnode->rn_count != 0; slot++) {
 254                 if (rnode->rn_child[slot] == NULL)
 255                         continue;
 256                 if (!vm_radix_isleaf(rnode->rn_child[slot]))
 257                         vm_radix_reclaim_allnodes_int(rnode->rn_child[slot]);
 258                 rnode->rn_child[slot] = NULL;
 259                 rnode->rn_count--;
 260         }
 261         vm_radix_node_put(rnode);
 262 }
 263
 264 #ifdef INVARIANTS
 265 /*
 266  * Radix node zone destructor.
 267  */
 268 static void
 269 vm_radix_node_zone_dtor(void *mem, int size __unused, void *arg __unused)
 270 {
 271         struct vm_radix_node *rnode;
 272         int slot;
 273
 274         rnode = mem;
 275         KASSERT(rnode->rn_count == 0,
 276             ("vm_radix_node_put: rnode %p has %d children", rnode,
 277             rnode->rn_count));
 278         for (slot = 0; slot < VM_RADIX_COUNT; slot++)
 279                 KASSERT(rnode->rn_child[slot] == NULL,
 280                     ("vm_radix_node_put: rnode %p has a child", rnode));
 281 }
 282 #endif
 283
 284 #ifndef UMA_MD_SMALL_ALLOC
 285 /*
 286  * Reserve the KVA necessary to satisfy the node allocation.
 287  * This is mandatory in architectures not supporting direct
 288  * mapping as they will need otherwise to carve into the kernel maps for
 289  * every node allocation, resulting into deadlocks for consumers already
 290  * working with kernel maps.
 291  */
 292 static void
 293 vm_radix_reserve_kva(void *arg __unused)
 294 {
 295
 296         /*
 297          * Calculate the number of reserved nodes, discounting the pages that
 298          * are needed to store them.
 299          */
 300         if (!uma_zone_reserve_kva(vm_radix_node_zone,
 301             ((vm_paddr_t)vm_cnt.v_page_count * PAGE_SIZE) / (PAGE_SIZE +
 302             sizeof(struct vm_radix_node))))
 303                 panic("%s: unable to reserve KVA", __func__);
 304 }
 305 SYSINIT(vm_radix_reserve_kva, SI_SUB_KMEM, SI_ORDER_THIRD,
 306     vm_radix_reserve_kva, NULL);
 307 #endif
 308
 309 /*
 310  * Initialize the UMA slab zone.
 311  */
 312 void
 313 vm_radix_zinit(void)
 314 {
 315
 316         vm_radix_node_zone = uma_zcreate("RADIX NODE",
 317             sizeof(struct vm_radix_node), NULL,
 318 #ifdef INVARIANTS
 319             vm_radix_node_zone_dtor,
 320 #else
 321             NULL,
 322 #endif
 323             NULL, NULL, VM_RADIX_PAD, UMA_ZONE_VM);
 324 }
 325
 326 /*
 327  * Inserts the key-value pair into the trie.
 328  * Panics if the key already exists.
 329  */
 330 int
 331 vm_radix_insert(struct vm_radix *rtree, vm_page_t page)
 332 {
 333         vm_pindex_t index, newind;
 334         void **parentp;
 335         struct vm_radix_node *rnode, *tmp;
 336         vm_page_t m;
 337         int slot;
 338         uint16_t clev;
 339
 340         index = page->pindex;
 341
 342         /*
 343          * The owner of record for root is not really important because it
 344          * will never be used.
 345          */
 346         rnode = vm_radix_getroot(rtree);
 347         if (rnode == NULL) {
 348                 rtree->rt_root = (uintptr_t)page | VM_RADIX_ISLEAF;
 349                 return (0);
 350         }
 351         parentp = (void **)&rtree->rt_root;
 352         for (;;) {
 353                 if (vm_radix_isleaf(rnode)) {
 354                         m = vm_radix_topage(rnode);
 355                         if (m->pindex == index)
 356                                 panic("%s: key %jx is already present",
 357                                     __func__, (uintmax_t)index);
 358                         clev = vm_radix_keydiff(m->pindex, index);
 359                         tmp = vm_radix_node_get(vm_radix_trimkey(index,
 360                             clev + 1), 2, clev);
 361                         if (tmp == NULL)
 362                                 return (ENOMEM);
 363                         *parentp = tmp;
 364                         vm_radix_addpage(tmp, index, clev, page);
 365                         vm_radix_addpage(tmp, m->pindex, clev, m);
 366                         return (0);
 367                 } else if (vm_radix_keybarr(rnode, index))
 368                         break;
 369                 slot = vm_radix_slot(index, rnode->rn_clev);
 370                 if (rnode->rn_child[slot] == NULL) {
 371                         rnode->rn_count++;
 372                         vm_radix_addpage(rnode, index, rnode->rn_clev, page);
 373                         return (0);
 374                 }
 375                 parentp = &rnode->rn_child[slot];
 376                 rnode = rnode->rn_child[slot];
 377         }
 378
 379         /*
 380          * A new node is needed because the right insertion level is reached.
 381          * Setup the new intermediate node and add the 2 children: the
 382          * new object and the older edge.
 383          */
 384         newind = rnode->rn_owner;
 385         clev = vm_radix_keydiff(newind, index);
 386         tmp = vm_radix_node_get(vm_radix_trimkey(index, clev + 1), 2, clev);
 387         if (tmp == NULL)
 388                 return (ENOMEM);
 389         *parentp = tmp;
 390         vm_radix_addpage(tmp, index, clev, page);
 391         slot = vm_radix_slot(newind, clev);
 392         tmp->rn_child[slot] = rnode;
 393         return (0);
 394 }
 395
 396 /*
 397  * Returns TRUE if the specified radix tree contains a single leaf and FALSE
 398  * otherwise.
 399  */
 400 boolean_t
 401 vm_radix_is_singleton(struct vm_radix *rtree)
 402 {
 403         struct vm_radix_node *rnode;
 404
 405         rnode = vm_radix_getroot(rtree);
 406         if (rnode == NULL)
 407                 return (FALSE);
 408         return (vm_radix_isleaf(rnode));
 409 }
 410
 411 /*
 412  * Returns the value stored at the index.  If the index is not present,
 413  * NULL is returned.
 414  */
 415 vm_page_t
 416 vm_radix_lookup(struct vm_radix *rtree, vm_pindex_t index)
 417 {
 418         struct vm_radix_node *rnode;
 419         vm_page_t m;
 420         int slot;
 421
 422         rnode = vm_radix_getroot(rtree);
 423         while (rnode != NULL) {
 424                 if (vm_radix_isleaf(rnode)) {
 425                         m = vm_radix_topage(rnode);
 426                         if (m->pindex == index)
 427                                 return (m);
 428                         else
 429                                 break;
 430                 } else if (vm_radix_keybarr(rnode, index))
 431                         break;
 432                 slot = vm_radix_slot(index, rnode->rn_clev);
 433                 rnode = rnode->rn_child[slot];
 434         }
 435         return (NULL);
 436 }
 437
 438 /*
 439  * Look up the nearest entry at a position bigger than or equal to index.
 440  */
 441 vm_page_t
 442 vm_radix_lookup_ge(struct vm_radix *rtree, vm_pindex_t index)
 443 {
 444         struct vm_radix_node *stack[VM_RADIX_LIMIT];
 445         vm_pindex_t inc;
 446         vm_page_t m;
 447         struct vm_radix_node *child, *rnode;
 448 #ifdef INVARIANTS
 449         int loops = 0;
 450 #endif
 451         int slot, tos;
 452
 453         rnode = vm_radix_getroot(rtree);
 454         if (rnode == NULL)
 455                 return (NULL);
 456         else if (vm_radix_isleaf(rnode)) {
 457                 m = vm_radix_topage(rnode);
 458                 if (m->pindex >= index)
 459                         return (m);
 460                 else
 461                         return (NULL);
 462         }
 463         tos = 0;
 464         for (;;) {
 465                 /*
 466                  * If the keys differ before the current bisection node,
 467                  * then the search key might rollback to the earliest
 468                  * available bisection node or to the smallest key
 469                  * in the current node (if the owner is bigger than the
 470                  * search key).
 471                  */
 472                 if (vm_radix_keybarr(rnode, index)) {
 473                         if (index > rnode->rn_owner) {
 474 ascend:
 475                                 KASSERT(++loops < 1000,
 476                                     ("vm_radix_lookup_ge: too many loops"));
 477
 478                                 /*
 479                                  * Pop nodes from the stack until either the
 480                                  * stack is empty or a node that could have a
 481                                  * matching descendant is found.
 482                                  */
 483                                 do {
 484                                         if (tos == 0)
 485                                                 return (NULL);
 486                                         rnode = stack[--tos];
 487                                 } while (vm_radix_slot(index,
 488                                     rnode->rn_clev) == (VM_RADIX_COUNT - 1));
 489
 490                                 /*
 491                                  * The following computation cannot overflow
 492                                  * because index's slot at the current level
 493                                  * is less than VM_RADIX_COUNT - 1.
 494                                  */
 495                                 index = vm_radix_trimkey(index,
 496                                     rnode->rn_clev);
 497                                 index += VM_RADIX_UNITLEVEL(rnode->rn_clev);
 498                         } else
 499                                 index = rnode->rn_owner;
 500                         KASSERT(!vm_radix_keybarr(rnode, index),
 501                             ("vm_radix_lookup_ge: keybarr failed"));
 502                 }
 503                 slot = vm_radix_slot(index, rnode->rn_clev);
 504                 child = rnode->rn_child[slot];
 505                 if (vm_radix_isleaf(child)) {
 506                         m = vm_radix_topage(child);
 507                         if (m->pindex >= index)
 508                                 return (m);
 509                 } else if (child != NULL)
 510                         goto descend;
 511
 512                 /*
 513                  * Look for an available edge or page within the current
 514                  * bisection node.
 515                  */
 516                 if (slot < (VM_RADIX_COUNT - 1)) {
 517                         inc = VM_RADIX_UNITLEVEL(rnode->rn_clev);
 518                         index = vm_radix_trimkey(index, rnode->rn_clev);
 519                         do {
 520                                 index += inc;
 521                                 slot++;
 522                                 child = rnode->rn_child[slot];
 523                                 if (vm_radix_isleaf(child)) {
 524                                         m = vm_radix_topage(child);
 525                                         if (m->pindex >= index)
 526                                                 return (m);
 527                                 } else if (child != NULL)
 528                                         goto descend;
 529                         } while (slot < (VM_RADIX_COUNT - 1));
 530                 }
 531                 KASSERT(child == NULL || vm_radix_isleaf(child),
 532                     ("vm_radix_lookup_ge: child is radix node"));
 533
 534                 /*
 535                  * If a page or edge bigger than the search slot is not found
 536                  * in the current node, ascend to the next higher-level node.
 537                  */
 538                 goto ascend;
 539 descend:
 540                 KASSERT(rnode->rn_clev > 0,
 541                     ("vm_radix_lookup_ge: pushing leaf's parent"));
 542                 KASSERT(tos < VM_RADIX_LIMIT,
 543                     ("vm_radix_lookup_ge: stack overflow"));
 544                 stack[tos++] = rnode;
 545                 rnode = child;
 546         }
 547 }
 548
 549 /*
 550  * Look up the nearest entry at a position less than or equal to index.
 551  */
 552 vm_page_t
 553 vm_radix_lookup_le(struct vm_radix *rtree, vm_pindex_t index)
 554 {
 555         struct vm_radix_node *stack[VM_RADIX_LIMIT];
 556         vm_pindex_t inc;
 557         vm_page_t m;
 558         struct vm_radix_node *child, *rnode;
 559 #ifdef INVARIANTS
 560         int loops = 0;
 561 #endif
 562         int slot, tos;
 563
 564         rnode = vm_radix_getroot(rtree);
 565         if (rnode == NULL)
 566                 return (NULL);
 567         else if (vm_radix_isleaf(rnode)) {
 568                 m = vm_radix_topage(rnode);
 569                 if (m->pindex <= index)
 570                         return (m);
 571                 else
 572                         return (NULL);
 573         }
 574         tos = 0;
 575         for (;;) {
 576                 /*
 577                  * If the keys differ before the current bisection node,
 578                  * then the search key might rollback to the earliest
 579                  * available bisection node or to the largest key
 580                  * in the current node (if the owner is smaller than the
 581                  * search key).
 582                  */
 583                 if (vm_radix_keybarr(rnode, index)) {
 584                         if (index > rnode->rn_owner) {
 585                                 index = rnode->rn_owner + VM_RADIX_COUNT *
 586                                     VM_RADIX_UNITLEVEL(rnode->rn_clev);
 587                         } else {
 588 ascend:
 589                                 KASSERT(++loops < 1000,
 590                                     ("vm_radix_lookup_le: too many loops"));
 591
 592                                 /*
 593                                  * Pop nodes from the stack until either the
 594                                  * stack is empty or a node that could have a
 595                                  * matching descendant is found.
 596                                  */
 597                                 do {
 598                                         if (tos == 0)
 599                                                 return (NULL);
 600                                         rnode = stack[--tos];
 601                                 } while (vm_radix_slot(index,
 602                                     rnode->rn_clev) == 0);
 603
 604                                 /*
 605                                  * The following computation cannot overflow
 606                                  * because index's slot at the current level
 607                                  * is greater than 0.
 608                                  */
 609                                 index = vm_radix_trimkey(index,
 610                                     rnode->rn_clev);
 611                         }
 612                         index--;
 613                         KASSERT(!vm_radix_keybarr(rnode, index),
 614                             ("vm_radix_lookup_le: keybarr failed"));
 615                 }
 616                 slot = vm_radix_slot(index, rnode->rn_clev);
 617                 child = rnode->rn_child[slot];
 618                 if (vm_radix_isleaf(child)) {
 619                         m = vm_radix_topage(child);
 620                         if (m->pindex <= index)
 621                                 return (m);
 622                 } else if (child != NULL)
 623                         goto descend;
 624
 625                 /*
 626                  * Look for an available edge or page within the current
 627                  * bisection node.
 628                  */
 629                 if (slot > 0) {
 630                         inc = VM_RADIX_UNITLEVEL(rnode->rn_clev);
 631                         index |= inc - 1;
 632                         do {
 633                                 index -= inc;
 634                                 slot--;
 635                                 child = rnode->rn_child[slot];
 636                                 if (vm_radix_isleaf(child)) {
 637                                         m = vm_radix_topage(child);
 638                                         if (m->pindex <= index)
 639                                                 return (m);
 640                                 } else if (child != NULL)
 641                                         goto descend;
 642                         } while (slot > 0);
 643                 }
 644                 KASSERT(child == NULL || vm_radix_isleaf(child),
 645                     ("vm_radix_lookup_le: child is radix node"));
 646
 647                 /*
 648                  * If a page or edge smaller than the search slot is not found
 649                  * in the current node, ascend to the next higher-level node.
 650                  */
 651                 goto ascend;
 652 descend:
 653                 KASSERT(rnode->rn_clev > 0,
 654                     ("vm_radix_lookup_le: pushing leaf's parent"));
 655                 KASSERT(tos < VM_RADIX_LIMIT,
 656                     ("vm_radix_lookup_le: stack overflow"));
 657                 stack[tos++] = rnode;
 658                 rnode = child;
 659         }
 660 }
 661
 662 /*
 663  * Remove the specified index from the trie, and return the value stored at
 664  * that index.  If the index is not present, return NULL.
 665  */
 666 vm_page_t
 667 vm_radix_remove(struct vm_radix *rtree, vm_pindex_t index)
 668 {
 669         struct vm_radix_node *rnode, *parent;
 670         vm_page_t m;
 671         int i, slot;
 672
 673         rnode = vm_radix_getroot(rtree);
 674         if (vm_radix_isleaf(rnode)) {
 675                 m = vm_radix_topage(rnode);
 676                 if (m->pindex != index)
 677                         return (NULL);
 678                 vm_radix_setroot(rtree, NULL);
 679                 return (m);
 680         }
 681         parent = NULL;
 682         for (;;) {
 683                 if (rnode == NULL)
 684                         return (NULL);
 685                 slot = vm_radix_slot(index, rnode->rn_clev);
 686                 if (vm_radix_isleaf(rnode->rn_child[slot])) {
 687                         m = vm_radix_topage(rnode->rn_child[slot]);
 688                         if (m->pindex != index)
 689                                 return (NULL);
 690                         rnode->rn_child[slot] = NULL;
 691                         rnode->rn_count--;
 692                         if (rnode->rn_count > 1)
 693                                 return (m);
 694                         for (i = 0; i < VM_RADIX_COUNT; i++)
 695                                 if (rnode->rn_child[i] != NULL)
 696                                         break;
 697                         KASSERT(i != VM_RADIX_COUNT,
 698                             ("%s: invalid node configuration", __func__));
 699                         if (parent == NULL)
 700                                 vm_radix_setroot(rtree, rnode->rn_child[i]);
 701                         else {
 702                                 slot = vm_radix_slot(index, parent->rn_clev);
 703                                 KASSERT(parent->rn_child[slot] == rnode,
 704                                     ("%s: invalid child value", __func__));
 705                                 parent->rn_child[slot] = rnode->rn_child[i];
 706                         }
 707                         rnode->rn_count--;
 708                         rnode->rn_child[i] = NULL;
 709                         vm_radix_node_put(rnode);
 710                         return (m);
 711                 }
 712                 parent = rnode;
 713                 rnode = rnode->rn_child[slot];
 714         }
 715 }
 716
 717 /*
 718  * Remove and free all the nodes from the radix tree.
 719  * This function is recursive but there is a tight control on it as the
 720  * maximum depth of the tree is fixed.
 721  */
 722 void
 723 vm_radix_reclaim_allnodes(struct vm_radix *rtree)
 724 {
 725         struct vm_radix_node *root;
 726
 727         root = vm_radix_getroot(rtree);
 728         if (root == NULL)
 729                 return;
 730         vm_radix_setroot(rtree, NULL);
 731         if (!vm_radix_isleaf(root))
 732                 vm_radix_reclaim_allnodes_int(root);
 733 }
 734
 735 /*
 736  * Replace an existing page in the trie with another one.
 737  * Panics if there is not an old page in the trie at the new page's index.
 738  */
 739 vm_page_t
 740 vm_radix_replace(struct vm_radix *rtree, vm_page_t newpage)
 741 {
 742         struct vm_radix_node *rnode;
 743         vm_page_t m;
 744         vm_pindex_t index;
 745         int slot;
 746
 747         index = newpage->pindex;
 748         rnode = vm_radix_getroot(rtree);
 749         if (rnode == NULL)
 750                 panic("%s: replacing page on an empty trie", __func__);
 751         if (vm_radix_isleaf(rnode)) {
 752                 m = vm_radix_topage(rnode);
 753                 if (m->pindex != index)
 754                         panic("%s: original replacing root key not found",
 755                             __func__);
 756                 rtree->rt_root = (uintptr_t)newpage | VM_RADIX_ISLEAF;
 757                 return (m);
 758         }
 759         for (;;) {
 760                 slot = vm_radix_slot(index, rnode->rn_clev);
 761                 if (vm_radix_isleaf(rnode->rn_child[slot])) {
 762                         m = vm_radix_topage(rnode->rn_child[slot]);
 763                         if (m->pindex == index) {
 764                                 rnode->rn_child[slot] =
 765                                     (void *)((uintptr_t)newpage |
 766                                     VM_RADIX_ISLEAF);
 767                                 return (m);
 768                         } else
 769                                 break;
 770                 } else if (rnode->rn_child[slot] == NULL ||
 771                     vm_radix_keybarr(rnode->rn_child[slot], index))
 772                         break;
 773                 rnode = rnode->rn_child[slot];
 774         }
 775         panic("%s: original replacing page not found", __func__);
 776 }
 777
 778 void
 779 vm_radix_wait(void)
 780 {
 781         uma_zwait(vm_radix_node_zone);
 782 }
 783
 784 #ifdef DDB
 785 /*
 786  * Show details about the given radix node.
 787  */
 788 DB_SHOW_COMMAND(radixnode, db_show_radixnode)
 789 {
 790         struct vm_radix_node *rnode;
 791         int i;
 792
 793         if (!have_addr)
 794                 return;
 795         rnode = (struct vm_radix_node *)addr;
 796         db_printf("radixnode %p, owner %jx, children count %u, level %u:\n",
 797             (void *)rnode, (uintmax_t)rnode->rn_owner, rnode->rn_count,
 798             rnode->rn_clev);
 799         for (i = 0; i < VM_RADIX_COUNT; i++)
 800                 if (rnode->rn_child[i] != NULL)
 801                         db_printf("slot: %d, val: %p, page: %p, clev: %d\n",
 802                             i, (void *)rnode->rn_child[i],
 803                             vm_radix_isleaf(rnode->rn_child[i]) ?
 804                             vm_radix_topage(rnode->rn_child[i]) : NULL,
 805                             rnode->rn_clev);
 806 }
 807 #endif /* DDB */