fs/jffs2/build.c

   1 /*
   2  * JFFS2 -- Journalling Flash File System, Version 2.
   3  *
   4  * Copyright © 2001-2007 Red Hat, Inc.
   5  * Copyright © 2004-2010 David Woodhouse <dwmw2@infradead.org>
   6  *
   7  * Created by David Woodhouse <dwmw2@infradead.org>
   8  *
   9  * For licensing information, see the file 'LICENCE' in this directory.
  10  *
  11  */
  12
  13 #include <linux/kernel.h>
  14 #include <linux/sched.h>
  15 #include <linux/slab.h>
  16 #include <linux/vmalloc.h>
  17 #include <linux/mtd/mtd.h>
  18 #include "nodelist.h"
  19
  20 static void jffs2_build_remove_unlinked_inode(struct jffs2_sb_info *,
  21                 struct jffs2_inode_cache *, struct jffs2_full_dirent **);
  22
  23 static inline struct jffs2_inode_cache *
  24 first_inode_chain(int *i, struct jffs2_sb_info *c)
  25 {
  26         for (; *i < c->inocache_hashsize; (*i)++) {
  27                 if (c->inocache_list[*i])
  28                         return c->inocache_list[*i];
  29         }
  30         return NULL;
  31 }
  32
  33 static inline struct jffs2_inode_cache *
  34 next_inode(int *i, struct jffs2_inode_cache *ic, struct jffs2_sb_info *c)
  35 {
  36         /* More in this chain? */
  37         if (ic->next)
  38                 return ic->next;
  39         (*i)++;
  40         return first_inode_chain(i, c);
  41 }
  42
  43 #define for_each_inode(i, c, ic)                        \
  44         for (i = 0, ic = first_inode_chain(&i, (c));    \
  45              ic;                                        \
  46              ic = next_inode(&i, ic, (c)))
  47
  48
  49 static void jffs2_build_inode_pass1(struct jffs2_sb_info *c,
  50                                     struct jffs2_inode_cache *ic)
  51 {
  52         struct jffs2_full_dirent *fd;
  53
  54         dbg_fsbuild("building directory inode #%u\n", ic->ino);
  55
  56         /* For each child, increase nlink */
  57         for(fd = ic->scan_dents; fd; fd = fd->next) {
  58                 struct jffs2_inode_cache *child_ic;
  59                 if (!fd->ino)
  60                         continue;
  61
  62                 /* we can get high latency here with huge directories */
  63
  64                 child_ic = jffs2_get_ino_cache(c, fd->ino);
  65                 if (!child_ic) {
  66                         dbg_fsbuild("child \"%s\" (ino #%u) of dir ino #%u doesn't exist!\n",
  67                                   fd->name, fd->ino, ic->ino);
  68                         jffs2_mark_node_obsolete(c, fd->raw);
  69                         continue;
  70                 }
  71
  72                 if (fd->type == DT_DIR) {
  73                         if (child_ic->pino_nlink) {
  74                                 JFFS2_ERROR("child dir \"%s\" (ino #%u) of dir ino #%u appears to be a hard link\n",
  75                                             fd->name, fd->ino, ic->ino);
  76                                 /* TODO: What do we do about it? */
  77                         } else {
  78                                 child_ic->pino_nlink = ic->ino;
  79                         }
  80                 } else
  81                         child_ic->pino_nlink++;
  82
  83                 dbg_fsbuild("increased nlink for child \"%s\" (ino #%u)\n", fd->name, fd->ino);
  84                 /* Can't free scan_dents so far. We might need them in pass 2 */
  85         }
  86 }
  87
  88 /* Scan plan:
  89  - Scan physical nodes. Build map of inodes/dirents. Allocate inocaches as we go
  90  - Scan directory tree from top down, setting nlink in inocaches
  91  - Scan inocaches for inodes with nlink==0
  92 */
  93 static int jffs2_build_filesystem(struct jffs2_sb_info *c)
  94 {
  95         int ret;
  96         int i;
  97         struct jffs2_inode_cache *ic;
  98         struct jffs2_full_dirent *fd;
  99         struct jffs2_full_dirent *dead_fds = NULL;
 100
 101         dbg_fsbuild("build FS data structures\n");
 102
 103         /* First, scan the medium and build all the inode caches with
 104            lists of physical nodes */
 105
 106         c->flags |= JFFS2_SB_FLAG_SCANNING;
 107         ret = jffs2_scan_medium(c);
 108         c->flags &= ~JFFS2_SB_FLAG_SCANNING;
 109         if (ret)
 110                 goto exit;
 111
 112         dbg_fsbuild("scanned flash completely\n");
 113         jffs2_dbg_dump_block_lists_nolock(c);
 114
 115         dbg_fsbuild("pass 1 starting\n");
 116         c->flags |= JFFS2_SB_FLAG_BUILDING;
 117         /* Now scan the directory tree, increasing nlink according to every dirent found. */
 118         for_each_inode(i, c, ic) {
 119                 if (ic->scan_dents) {
 120                         jffs2_build_inode_pass1(c, ic);
 121                         cond_resched();
 122                 }
 123         }
 124
 125         dbg_fsbuild("pass 1 complete\n");
 126
 127         /* Next, scan for inodes with nlink == 0 and remove them. If
 128            they were directories, then decrement the nlink of their
 129            children too, and repeat the scan. As that's going to be
 130            a fairly uncommon occurrence, it's not so evil to do it this
 131            way. Recursion bad. */
 132         dbg_fsbuild("pass 2 starting\n");
 133
 134         for_each_inode(i, c, ic) {
 135                 if (ic->pino_nlink)
 136                         continue;
 137
 138                 jffs2_build_remove_unlinked_inode(c, ic, &dead_fds);
 139                 cond_resched();
 140         }
 141
 142         dbg_fsbuild("pass 2a starting\n");
 143
 144         while (dead_fds) {
 145                 fd = dead_fds;
 146                 dead_fds = fd->next;
 147
 148                 ic = jffs2_get_ino_cache(c, fd->ino);
 149
 150                 if (ic)
 151                         jffs2_build_remove_unlinked_inode(c, ic, &dead_fds);
 152                 jffs2_free_full_dirent(fd);
 153         }
 154
 155         dbg_fsbuild("pass 2a complete\n");
 156         dbg_fsbuild("freeing temporary data structures\n");
 157
 158         /* Finally, we can scan again and free the dirent structs */
 159         for_each_inode(i, c, ic) {
 160                 while(ic->scan_dents) {
 161                         fd = ic->scan_dents;
 162                         ic->scan_dents = fd->next;
 163                         jffs2_free_full_dirent(fd);
 164                 }
 165                 ic->scan_dents = NULL;
 166                 cond_resched();
 167         }
 168         jffs2_build_xattr_subsystem(c);
 169         c->flags &= ~JFFS2_SB_FLAG_BUILDING;
 170
 171         dbg_fsbuild("FS build complete\n");
 172
 173         /* Rotate the lists by some number to ensure wear levelling */
 174         jffs2_rotate_lists(c);
 175
 176         ret = 0;
 177
 178 exit:
 179         if (ret) {
 180                 for_each_inode(i, c, ic) {
 181                         while(ic->scan_dents) {
 182                                 fd = ic->scan_dents;
 183                                 ic->scan_dents = fd->next;
 184                                 jffs2_free_full_dirent(fd);
 185                         }
 186                 }
 187                 jffs2_clear_xattr_subsystem(c);
 188         }
 189
 190         return ret;
 191 }
 192
 193 static void jffs2_build_remove_unlinked_inode(struct jffs2_sb_info *c,
 194                                         struct jffs2_inode_cache *ic,
 195                                         struct jffs2_full_dirent **dead_fds)
 196 {
 197         struct jffs2_raw_node_ref *raw;
 198         struct jffs2_full_dirent *fd;
 199
 200         dbg_fsbuild("removing ino #%u with nlink == zero.\n", ic->ino);
 201
 202         raw = ic->nodes;
 203         while (raw != (void *)ic) {
 204                 struct jffs2_raw_node_ref *next = raw->next_in_ino;
 205                 dbg_fsbuild("obsoleting node at 0x%08x\n", ref_offset(raw));
 206                 jffs2_mark_node_obsolete(c, raw);
 207                 raw = next;
 208         }
 209
 210         if (ic->scan_dents) {
 211                 int whinged = 0;
 212                 dbg_fsbuild("inode #%u was a directory which may have children...\n", ic->ino);
 213
 214                 while(ic->scan_dents) {
 215                         struct jffs2_inode_cache *child_ic;
 216
 217                         fd = ic->scan_dents;
 218                         ic->scan_dents = fd->next;
 219
 220                         if (!fd->ino) {
 221                                 /* It's a deletion dirent. Ignore it */
 222                                 dbg_fsbuild("child \"%s\" is a deletion dirent, skipping...\n", fd->name);
 223                                 jffs2_free_full_dirent(fd);
 224                                 continue;
 225                         }
 226                         if (!whinged)
 227                                 whinged = 1;
 228
 229                         dbg_fsbuild("removing child \"%s\", ino #%u\n", fd->name, fd->ino);
 230
 231                         child_ic = jffs2_get_ino_cache(c, fd->ino);
 232                         if (!child_ic) {
 233                                 dbg_fsbuild("cannot remove child \"%s\", ino #%u, because it doesn't exist\n",
 234                                                 fd->name, fd->ino);
 235                                 jffs2_free_full_dirent(fd);
 236                                 continue;
 237                         }
 238
 239                         /* Reduce nlink of the child. If it's now zero, stick it on the
 240                            dead_fds list to be cleaned up later. Else just free the fd */
 241
 242                         if (fd->type == DT_DIR)
 243                                 child_ic->pino_nlink = 0;
 244                         else
 245                                 child_ic->pino_nlink--;
 246
 247                         if (!child_ic->pino_nlink) {
 248                                 dbg_fsbuild("inode #%u (\"%s\") now has no links; adding to dead_fds list.\n",
 249                                           fd->ino, fd->name);
 250                                 fd->next = *dead_fds;
 251                                 *dead_fds = fd;
 252                         } else {
 253                                 dbg_fsbuild("inode #%u (\"%s\") has now got nlink %d. Ignoring.\n",
 254                                           fd->ino, fd->name, child_ic->pino_nlink);
 255                                 jffs2_free_full_dirent(fd);
 256                         }
 257                 }
 258         }
 259
 260         /*
 261            We don't delete the inocache from the hash list and free it yet.
 262            The erase code will do that, when all the nodes are completely gone.
 263         */
 264 }
 265
 266 static void jffs2_calc_trigger_levels(struct jffs2_sb_info *c)
 267 {
 268         uint32_t size;
 269
 270         /* Deletion should almost _always_ be allowed. We're fairly
 271            buggered once we stop allowing people to delete stuff
 272            because there's not enough free space... */
 273         c->resv_blocks_deletion = 2;
 274
 275         /* Be conservative about how much space we need before we allow writes.
 276            On top of that which is required for deletia, require an extra 2%
 277            of the medium to be available, for overhead caused by nodes being
 278            split across blocks, etc. */
 279
 280         size = c->flash_size / 50; /* 2% of flash size */
 281         size += c->nr_blocks * 100; /* And 100 bytes per eraseblock */
 282         size += c->sector_size - 1; /* ... and round up */
 283
 284         c->resv_blocks_write = c->resv_blocks_deletion + (size / c->sector_size);
 285
 286         /* When do we let the GC thread run in the background */
 287
 288         c->resv_blocks_gctrigger = c->resv_blocks_write + 1;
 289
 290         /* When do we allow garbage collection to merge nodes to make
 291            long-term progress at the expense of short-term space exhaustion? */
 292         c->resv_blocks_gcmerge = c->resv_blocks_deletion + 1;
 293
 294         /* When do we allow garbage collection to eat from bad blocks rather
 295            than actually making progress? */
 296         c->resv_blocks_gcbad = 0;//c->resv_blocks_deletion + 2;
 297
 298         /* What number of 'very dirty' eraseblocks do we allow before we
 299            trigger the GC thread even if we don't _need_ the space. When we
 300            can't mark nodes obsolete on the medium, the old dirty nodes cause
 301            performance problems because we have to inspect and discard them. */
 302         c->vdirty_blocks_gctrigger = c->resv_blocks_gctrigger;
 303         if (jffs2_can_mark_obsolete(c))
 304                 c->vdirty_blocks_gctrigger *= 10;
 305
 306         /* If there's less than this amount of dirty space, don't bother
 307            trying to GC to make more space. It'll be a fruitless task */
 308         c->nospc_dirty_size = c->sector_size + (c->flash_size / 100);
 309
 310         dbg_fsbuild("JFFS2 trigger levels (size %d KiB, block size %d KiB, %d blocks)\n",
 311                   c->flash_size / 1024, c->sector_size / 1024, c->nr_blocks);
 312         dbg_fsbuild("Blocks required to allow deletion:    %d (%d KiB)\n",
 313                   c->resv_blocks_deletion, c->resv_blocks_deletion*c->sector_size/1024);
 314         dbg_fsbuild("Blocks required to allow writes:      %d (%d KiB)\n",
 315                   c->resv_blocks_write, c->resv_blocks_write*c->sector_size/1024);
 316         dbg_fsbuild("Blocks required to quiesce GC thread: %d (%d KiB)\n",
 317                   c->resv_blocks_gctrigger, c->resv_blocks_gctrigger*c->sector_size/1024);
 318         dbg_fsbuild("Blocks required to allow GC merges:   %d (%d KiB)\n",
 319                   c->resv_blocks_gcmerge, c->resv_blocks_gcmerge*c->sector_size/1024);
 320         dbg_fsbuild("Blocks required to GC bad blocks:     %d (%d KiB)\n",
 321                   c->resv_blocks_gcbad, c->resv_blocks_gcbad*c->sector_size/1024);
 322         dbg_fsbuild("Amount of dirty space required to GC: %d bytes\n",
 323                   c->nospc_dirty_size);
 324         dbg_fsbuild("Very dirty blocks before GC triggered: %d\n",
 325                   c->vdirty_blocks_gctrigger);
 326 }
 327
 328 int jffs2_do_mount_fs(struct jffs2_sb_info *c)
 329 {
 330         int ret;
 331         int i;
 332         int size;
 333
 334         c->free_size = c->flash_size;
 335         c->nr_blocks = c->flash_size / c->sector_size;
 336         size = sizeof(struct jffs2_eraseblock) * c->nr_blocks;
 337 #ifndef __ECOS
 338         if (jffs2_blocks_use_vmalloc(c))
 339                 c->blocks = vzalloc(size);
 340         else
 341 #endif
 342                 c->blocks = kzalloc(size, GFP_KERNEL);
 343         if (!c->blocks)
 344                 return -ENOMEM;
 345
 346         for (i=0; i<c->nr_blocks; i++) {
 347                 INIT_LIST_HEAD(&c->blocks[i].list);
 348                 c->blocks[i].offset = i * c->sector_size;
 349                 c->blocks[i].free_size = c->sector_size;
 350         }
 351
 352         INIT_LIST_HEAD(&c->clean_list);
 353         INIT_LIST_HEAD(&c->very_dirty_list);
 354         INIT_LIST_HEAD(&c->dirty_list);
 355         INIT_LIST_HEAD(&c->erasable_list);
 356         INIT_LIST_HEAD(&c->erasing_list);
 357         INIT_LIST_HEAD(&c->erase_checking_list);
 358         INIT_LIST_HEAD(&c->erase_pending_list);
 359         INIT_LIST_HEAD(&c->erasable_pending_wbuf_list);
 360         INIT_LIST_HEAD(&c->erase_complete_list);
 361         INIT_LIST_HEAD(&c->free_list);
 362         INIT_LIST_HEAD(&c->bad_list);
 363         INIT_LIST_HEAD(&c->bad_used_list);
 364         c->highest_ino = 1;
 365         c->summary = NULL;
 366
 367         ret = jffs2_sum_init(c);
 368         if (ret)
 369                 goto out_free;
 370
 371         if (jffs2_build_filesystem(c)) {
 372                 dbg_fsbuild("build_fs failed\n");
 373                 jffs2_free_ino_caches(c);
 374                 jffs2_free_raw_node_refs(c);
 375                 ret = -EIO;
 376                 goto out_free;
 377         }
 378
 379         jffs2_calc_trigger_levels(c);
 380
 381         return 0;
 382
 383  out_free:
 384 #ifndef __ECOS
 385         if (jffs2_blocks_use_vmalloc(c))
 386                 vfree(c->blocks);
 387         else
 388 #endif
 389                 kfree(c->blocks);
 390
 391         return ret;
 392 }