# This is a BitKeeper generated patch for the following project: # Project Name: Linux kernel tree # This patch format is intended for GNU patch command version 2.5 or higher. # This patch includes the following deltas: # ChangeSet 1.775 -> 1.783 # fs/ext3/namei.c 1.2 -> 1.6 # include/linux/ext3_fs.h 1.4 -> 1.9 # fs/ext2/inode.c 1.15 -> 1.16 # lib/rbtree.c 1.2 -> 1.3 # fs/ext3/Makefile 1.2 -> 1.3 # include/linux/ext3_jbd.h 1.2 -> 1.3 # include/linux/ext2_fs.h 1.6 -> 1.9 # include/linux/ext2_fs_sb.h 1.1 -> 1.2 # include/linux/rbtree.h 1.1 -> 1.2 # include/linux/ext2_fs_i.h 1.4 -> 1.5 # fs/ext3/ialloc.c 1.4 -> 1.5 # fs/ext2/ialloc.c 1.8 -> 1.10 # fs/ext3/dir.c 1.1 -> 1.3 # fs/ext2/super.c 1.9 -> 1.13 # fs/ext3/inode.c 1.10 -> 1.11 # fs/ext3/super.c 1.7 -> 1.12 # include/linux/ext3_fs_sb.h 1.2 -> 1.4 # fs/ext3/file.c 1.3 -> 1.4 # (new) -> 1.1 fs/ext3/hash.c # # The following is the BitKeeper ChangeSet Log # -------------------------------------------- # 02/11/06 tytso@snap.thunk.org 1.776 # Port patch-ext3-dxdir-2.4.19-4 to 2.4.20-rc1 # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.777 # Add '.' and '..' entries to be returned by readdir of htree directories # # This patch from Chris Li adds '.' and '..' to the rbtree so that they # are properly returned by readdir. # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.778 # Check for failed kmalloc() in ext3_htree_store_dirent() # # This patch checks for a failed kmalloc() in ext3_htree_store_dirent(), # and passes the error up to its caller, ext3_htree_fill_tree(). # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.779 # Fix ext3 htree rename bug. # # This fixes an ext3 htree bug pointed out by Christopher Li; if # adding the new name to the directory causes a split, this can cause # the directory entry containing the old name to move to another # block, and then the removal of the old name will fail. # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.780 # Default mount options from superblock for ext2/3 filesystems # # This patch adds support for default mount options to be stored in the # superblock, so they don't have to be specified on the mount command line # (or in /etc/fstab). While I was in the code, I also cleaned up the # handling of how mount options are processed in the ext2 and ext3 # filesystems. # # Most mount options are now processed *after* the superblock has been # read in. This allows for a much cleaner handling of those default mount # option parameters that were already stored in the superblock: the # resuid, resgid, and s_errors fields were handled using some fairly gross # special cases. Now the only mount option which is processed first is # the sb option, which specifies the location of the superblock. This # allows the handling of all of the default mount parameters to be much # more cleanly and more generally handled. # # This does change the behaviour from earlier kernels, in that if the sb # mount option is specified, it must be specified *first*. However, this # option is rarely used, and if it is, it generally is specified first, so # this seems to be a reasonable restriction. # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.781 # Ext2/3 forward compatibility: on-line resizing # # This patch allows forward compatibility with future filesystems which # are dynamically grown by using an alternate algorithm for storing the # block group descriptors. It's also a bit more efficient, in that it # uses just a little bit less disk space. Currently, the ext2 filesystem # format requires either relocating the inode table, or reserving space in # before doing the on-line resize. The new scheme, which is documented in # "Planned Extensions to the Ext2/3 Filesystem", by Stephen Tweedie and I # (see: http://e2fsprogs.sourceforge.net/extensions-ext23) # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.782 # Ext2/3 forward compatibility: inode size # # This patch allows filesystems with expanded inodes to be mounted. # (compatibility feature flags will be used to control whether or not the # filesystem should be mounted in case the new inode fields will result in # compatibility issues). This allows for future compatibility with newer # versions of ext2fs. # -------------------------------------------- # 02/11/07 tytso@snap.thunk.org 1.783 # Orlov block allocator for ext2/3 # # This is Al's implementation of the Orlov block allocator for ext2/3. # # At least doubles the throughput for the traverse-a-kernel-tree # test and is well tested. # -------------------------------------------- # diff -Nru a/fs/ext2/ialloc.c b/fs/ext2/ialloc.c --- a/fs/ext2/ialloc.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext2/ialloc.c Thu Nov 7 11:58:05 2002 @@ -17,6 +17,7 @@ #include #include #include +#include /* @@ -228,8 +229,7 @@ * For other inodes, search forward from the parent directory\'s block * group to find a free inode. */ - -static int find_group_dir(struct super_block *sb, int parent_group) +static int find_group_dir(struct super_block *sb, const struct inode *parent) { struct ext2_super_block * es = sb->u.ext2_sb.s_es; int ngroups = sb->u.ext2_sb.s_groups_count; @@ -262,8 +262,140 @@ return best_group; } -static int find_group_other(struct super_block *sb, int parent_group) +/* + * Orlov's allocator for directories. + * + * We always try to spread first-level directories. + * + * If there are blockgroups with both free inodes and free blocks counts + * not worse than average we return one with smallest directory count. + * Otherwise we simply return a random group. + * + * For the rest rules look so: + * + * It's OK to put directory into a group unless + * it has too many directories already (max_dirs) or + * it has too few free inodes left (min_inodes) or + * it has too few free blocks left (min_blocks) or + * it's already running too large debt (max_debt). + * Parent's group is prefered, if it doesn't satisfy these + * conditions we search cyclically through the rest. If none + * of the groups look good we just look for a group with more + * free inodes than average (starting at parent's group). + * + * Debt is incremented each time we allocate a directory and decremented + * when we allocate an inode, within 0--255. + */ + +#define INODE_COST 64 +#define BLOCK_COST 256 + +static int find_group_orlov(struct super_block *sb, const struct inode *parent) { + int parent_group = parent->u.ext2_i.i_block_group; + struct ext2_sb_info *sbi = EXT2_SB(sb); + struct ext2_super_block *es = sbi->s_es; + int ngroups = sbi->s_groups_count; + int inodes_per_group = EXT2_INODES_PER_GROUP(sb); + int avefreei = le32_to_cpu(es->s_free_inodes_count) / ngroups; + int avefreeb = le32_to_cpu(es->s_free_blocks_count) / ngroups; + int blocks_per_dir; + int ndirs = sbi->s_dir_count; + int max_debt, max_dirs, min_blocks, min_inodes; + int group = -1, i; + struct ext2_group_desc *desc; + struct buffer_head *bh; + + if ((parent == sb->s_root->d_inode) || + (parent->i_flags & EXT2_TOPDIR_FL)) { + struct ext2_group_desc *best_desc = NULL; + struct buffer_head *best_bh = NULL; + int best_ndir = inodes_per_group; + int best_group = -1; + + get_random_bytes(&group, sizeof(group)); + parent_group = (unsigned)group % ngroups; + for (i = 0; i < ngroups; i++) { + group = (parent_group + i) % ngroups; + desc = ext2_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (le16_to_cpu(desc->bg_used_dirs_count) >= best_ndir) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) < avefreei) + continue; + if (le16_to_cpu(desc->bg_free_blocks_count) < avefreeb) + continue; + best_group = group; + best_ndir = le16_to_cpu(desc->bg_used_dirs_count); + best_desc = desc; + best_bh = bh; + } + if (best_group >= 0) { + desc = best_desc; + bh = best_bh; + group = best_group; + goto found; + } + goto fallback; + } + + blocks_per_dir = (le32_to_cpu(es->s_blocks_count) - + le32_to_cpu(es->s_free_blocks_count)) / ndirs; + + max_dirs = ndirs / ngroups + inodes_per_group / 16; + min_inodes = avefreei - inodes_per_group / 4; + min_blocks = avefreeb - EXT2_BLOCKS_PER_GROUP(sb) / 4; + + max_debt = EXT2_BLOCKS_PER_GROUP(sb) / max(blocks_per_dir, BLOCK_COST); + if (max_debt * INODE_COST > inodes_per_group) + max_debt = inodes_per_group / INODE_COST; + if (max_debt > 255) + max_debt = 255; + if (max_debt == 0) + max_debt = 1; + + for (i = 0; i < ngroups; i++) { + group = (parent_group + i) % ngroups; + desc = ext2_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (sbi->s_debts[group] >= max_debt) + continue; + if (le16_to_cpu(desc->bg_used_dirs_count) >= max_dirs) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) < min_inodes) + continue; + if (le16_to_cpu(desc->bg_free_blocks_count) < min_blocks) + continue; + goto found; + } + +fallback: + for (i = 0; i < ngroups; i++) { + group = (parent_group + i) % ngroups; + desc = ext2_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) >= avefreei) + goto found; + } + + return -1; + +found: + desc->bg_free_inodes_count = + cpu_to_le16(le16_to_cpu(desc->bg_free_inodes_count) - 1); + desc->bg_used_dirs_count = + cpu_to_le16(le16_to_cpu(desc->bg_used_dirs_count) + 1); + sbi->s_dir_count++; + mark_buffer_dirty(bh); + return group; +} + +static int find_group_other(struct super_block *sb, const struct inode *parent) +{ + int parent_group = parent->u.ext2_i.i_block_group; int ngroups = sb->u.ext2_sb.s_groups_count; struct ext2_group_desc *desc; struct buffer_head *bh; @@ -331,10 +463,13 @@ lock_super (sb); es = sb->u.ext2_sb.s_es; repeat: - if (S_ISDIR(mode)) - group = find_group_dir(sb, dir->u.ext2_i.i_block_group); - else - group = find_group_other(sb, dir->u.ext2_i.i_block_group); + if (S_ISDIR(mode)) { + if (test_opt (sb, OLDALLOC)) + group = find_group_dir(sb, dir); + else + group = find_group_orlov(sb, dir); + } else + group = find_group_other(sb, dir); err = -ENOSPC; if (group == -1) @@ -368,6 +503,15 @@ es->s_free_inodes_count = cpu_to_le32(le32_to_cpu(es->s_free_inodes_count) - 1); + + if (S_ISDIR(mode)) { + if (EXT2_SB(sb)->s_debts[group] < 255) + EXT2_SB(sb)->s_debts[group]++; + } else { + if (EXT2_SB(sb)->s_debts[group]) + EXT2_SB(sb)->s_debts[group]--; + } + mark_buffer_dirty(sb->u.ext2_sb.s_sbh); sb->s_dirt = 1; inode->i_uid = current->fsuid; @@ -385,7 +529,7 @@ inode->i_blksize = PAGE_SIZE; /* This is the optimal IO size (for stat), not the fs block size */ inode->i_blocks = 0; inode->i_mtime = inode->i_atime = inode->i_ctime = CURRENT_TIME; - inode->u.ext2_i.i_new_inode = 1; + inode->u.ext2_i.i_state = EXT2_STATE_NEW; inode->u.ext2_i.i_flags = dir->u.ext2_i.i_flags; if (S_ISLNK(mode)) inode->u.ext2_i.i_flags &= ~(EXT2_IMMUTABLE_FL|EXT2_APPEND_FL); @@ -469,6 +613,21 @@ #else return le32_to_cpu(sb->u.ext2_sb.s_es->s_free_inodes_count); #endif +} + +/* Called at mount-time, super-block is locked */ +unsigned long ext2_count_dirs (struct super_block * sb) +{ + unsigned long count = 0; + int i; + + for (i = 0; i < EXT2_SB(sb)->s_groups_count; i++) { + struct ext2_group_desc *gdp = ext2_get_group_desc (sb, i, NULL); + if (!gdp) + continue; + count += le16_to_cpu(gdp->bg_used_dirs_count); + } + return count; } #ifdef CONFIG_EXT2_CHECK diff -Nru a/fs/ext2/inode.c b/fs/ext2/inode.c --- a/fs/ext2/inode.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext2/inode.c Thu Nov 7 11:58:05 2002 @@ -955,6 +955,7 @@ else inode->u.ext2_i.i_dir_acl = le32_to_cpu(raw_inode->i_dir_acl); inode->i_generation = le32_to_cpu(raw_inode->i_generation); + inode->u.ext2_i.i_state = 0; inode->u.ext2_i.i_prealloc_count = 0; inode->u.ext2_i.i_block_group = block_group; @@ -1061,6 +1062,11 @@ offset &= EXT2_BLOCK_SIZE(inode->i_sb) - 1; raw_inode = (struct ext2_inode *) (bh->b_data + offset); + /* For fields not tracked in the in-memory inode, + * initialise them to zero for new inodes. */ + if (inode->u.ext2_i.i_state & EXT2_STATE_NEW) + memset(raw_inode, 0, EXT2_SB(inode->i_sb)->s_inode_size); + raw_inode->i_mode = cpu_to_le16(inode->i_mode); if(!(test_opt(inode->i_sb, NO_UID32))) { raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid)); @@ -1133,6 +1139,7 @@ err = -EIO; } } + inode->u.ext2_i.i_state &= ~EXT2_STATE_NEW; brelse (bh); return err; } diff -Nru a/fs/ext2/super.c b/fs/ext2/super.c --- a/fs/ext2/super.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext2/super.c Thu Nov 7 11:58:05 2002 @@ -48,16 +48,12 @@ va_start (args, fmt); vsprintf (error_buf, fmt, args); va_end (args); - if (test_opt (sb, ERRORS_PANIC) || - (le16_to_cpu(sb->u.ext2_sb.s_es->s_errors) == EXT2_ERRORS_PANIC && - !test_opt (sb, ERRORS_CONT) && !test_opt (sb, ERRORS_RO))) + if (test_opt (sb, ERRORS_PANIC)) panic ("EXT2-fs panic (device %s): %s: %s\n", bdevname(sb->s_dev), function, error_buf); printk (KERN_CRIT "EXT2-fs error (device %s): %s: %s\n", bdevname(sb->s_dev), function, error_buf); - if (test_opt (sb, ERRORS_RO) || - (le16_to_cpu(sb->u.ext2_sb.s_es->s_errors) == EXT2_ERRORS_RO && - !test_opt (sb, ERRORS_CONT) && !test_opt (sb, ERRORS_PANIC))) { + if (test_opt (sb, ERRORS_RO)) { printk ("Remounting filesystem read-only\n"); sb->s_flags |= MS_RDONLY; } @@ -136,6 +132,7 @@ if (sb->u.ext2_sb.s_group_desc[i]) brelse (sb->u.ext2_sb.s_group_desc[i]); kfree(sb->u.ext2_sb.s_group_desc); + kfree(sb->u.ext2_sb.s_debts); for (i = 0; i < EXT2_MAX_GROUP_LOADED; i++) if (sb->u.ext2_sb.s_inode_bitmap[i]) brelse (sb->u.ext2_sb.s_inode_bitmap[i]); @@ -158,12 +155,61 @@ remount_fs: ext2_remount, }; +static unsigned long get_sb_block(void **data) +{ + unsigned long sb_block; + char *options = (char *) *data; + + if (!options || strncmp(options, "sb=", 3) != 0) + return 1; /* Default location */ + options += 3; + sb_block = simple_strtoul(options, &options, 0); + if (*options && *options != ',') { + printk("EXT2-fs: Invalid sb specification: %s\n", + (char *) *data); + return 1; + } + if (*options == ',') + options++; + *data = (void *) options; + return sb_block; +} + +static int want_value(char *value, char *option) +{ + if (!value || !*value) { + printk(KERN_NOTICE "EXT2-fs: the %s option needs an argument\n", + option); + return -1; + } + return 0; +} + +static int want_null_value(char *value, char *option) +{ + if (*value) { + printk(KERN_NOTICE "EXT2-fs: Invalid %s argument: %s\n", + option, value); + return -1; + } + return 0; +} + +static int want_numeric(char *value, char *option, unsigned long *number) +{ + if (want_value(value, option)) + return -1; + *number = simple_strtoul(value, &value, 0); + if (want_null_value(value, option)) + return -1; + return 0; +} + /* * This function has been shamelessly adapted from the msdos fs */ -static int parse_options (char * options, unsigned long * sb_block, - unsigned short *resuid, unsigned short * resgid, - unsigned long * mount_options) +static int parse_options (char * options, + struct ext2_sb_info *sbi) { char * this_char; char * value; @@ -176,22 +222,22 @@ if ((value = strchr (this_char, '=')) != NULL) *value++ = 0; if (!strcmp (this_char, "bsddf")) - clear_opt (*mount_options, MINIX_DF); + clear_opt (sbi->s_mount_opt, MINIX_DF); else if (!strcmp (this_char, "nouid32")) { - set_opt (*mount_options, NO_UID32); + set_opt (sbi->s_mount_opt, NO_UID32); } else if (!strcmp (this_char, "check")) { if (!value || !*value || !strcmp (value, "none")) - clear_opt (*mount_options, CHECK); + clear_opt (sbi->s_mount_opt, CHECK); else #ifdef CONFIG_EXT2_CHECK - set_opt (*mount_options, CHECK); + set_opt (sbi->s_mount_opt, CHECK); #else printk("EXT2 Check option not supported\n"); #endif } else if (!strcmp (this_char, "debug")) - set_opt (*mount_options, DEBUG); + set_opt (sbi->s_mount_opt, DEBUG); else if (!strcmp (this_char, "errors")) { if (!value || !*value) { printk ("EXT2-fs: the errors option requires " @@ -199,19 +245,19 @@ return 0; } if (!strcmp (value, "continue")) { - clear_opt (*mount_options, ERRORS_RO); - clear_opt (*mount_options, ERRORS_PANIC); - set_opt (*mount_options, ERRORS_CONT); + clear_opt (sbi->s_mount_opt, ERRORS_RO); + clear_opt (sbi->s_mount_opt, ERRORS_PANIC); + set_opt (sbi->s_mount_opt, ERRORS_CONT); } else if (!strcmp (value, "remount-ro")) { - clear_opt (*mount_options, ERRORS_CONT); - clear_opt (*mount_options, ERRORS_PANIC); - set_opt (*mount_options, ERRORS_RO); + clear_opt (sbi->s_mount_opt, ERRORS_CONT); + clear_opt (sbi->s_mount_opt, ERRORS_PANIC); + set_opt (sbi->s_mount_opt, ERRORS_RO); } else if (!strcmp (value, "panic")) { - clear_opt (*mount_options, ERRORS_CONT); - clear_opt (*mount_options, ERRORS_RO); - set_opt (*mount_options, ERRORS_PANIC); + clear_opt (sbi->s_mount_opt, ERRORS_CONT); + clear_opt (sbi->s_mount_opt, ERRORS_RO); + set_opt (sbi->s_mount_opt, ERRORS_PANIC); } else { printk ("EXT2-fs: Invalid errors option: %s\n", @@ -221,53 +267,30 @@ } else if (!strcmp (this_char, "grpid") || !strcmp (this_char, "bsdgroups")) - set_opt (*mount_options, GRPID); + set_opt (sbi->s_mount_opt, GRPID); else if (!strcmp (this_char, "minixdf")) - set_opt (*mount_options, MINIX_DF); + set_opt (sbi->s_mount_opt, MINIX_DF); else if (!strcmp (this_char, "nocheck")) - clear_opt (*mount_options, CHECK); + clear_opt (sbi->s_mount_opt, CHECK); else if (!strcmp (this_char, "nogrpid") || !strcmp (this_char, "sysvgroups")) - clear_opt (*mount_options, GRPID); + clear_opt (sbi->s_mount_opt, GRPID); else if (!strcmp (this_char, "resgid")) { - if (!value || !*value) { - printk ("EXT2-fs: the resgid option requires " - "an argument\n"); + unsigned long v; + if (want_numeric(value, "resgid", &v)) return 0; - } - *resgid = simple_strtoul (value, &value, 0); - if (*value) { - printk ("EXT2-fs: Invalid resgid option: %s\n", - value); - return 0; - } + sbi->s_resgid = v; } else if (!strcmp (this_char, "resuid")) { - if (!value || !*value) { - printk ("EXT2-fs: the resuid option requires " - "an argument"); + unsigned long v; + if (want_numeric(value, "resuid", &v)) return 0; - } - *resuid = simple_strtoul (value, &value, 0); - if (*value) { - printk ("EXT2-fs: Invalid resuid option: %s\n", - value); - return 0; - } - } - else if (!strcmp (this_char, "sb")) { - if (!value || !*value) { - printk ("EXT2-fs: the sb option requires " - "an argument"); - return 0; - } - *sb_block = simple_strtoul (value, &value, 0); - if (*value) { - printk ("EXT2-fs: Invalid sb option: %s\n", - value); - return 0; - } + sbi->s_resuid = v; } + else if (!strcmp (this_char, "oldalloc")) + set_opt (sbi->s_mount_opt, OLDALLOC); + else if (!strcmp (this_char, "orlov")) + clear_opt (sbi->s_mount_opt, OLDALLOC); /* Silently ignore the quota options */ else if (!strcmp (this_char, "grpquota") || !strcmp (this_char, "noquota") @@ -397,17 +420,37 @@ return res; } +static unsigned long descriptor_loc(struct super_block *sb, + unsigned long logic_sb_block, + int nr) +{ + struct ext2_sb_info *sbi = EXT2_SB(sb); + unsigned long bg, first_data_block, first_meta_bg; + int has_super = 0; + + first_data_block = le32_to_cpu(sbi->s_es->s_first_data_block); + first_meta_bg = le32_to_cpu(sbi->s_es->s_first_meta_bg); + + if (!EXT2_HAS_INCOMPAT_FEATURE(sb, EXT2_FEATURE_INCOMPAT_META_BG) || + nr < first_meta_bg) + return (logic_sb_block + nr + 1); + bg = sbi->s_desc_per_block * nr; + if (ext2_bg_has_super(sb, bg)) + has_super = 1; + return (first_data_block + has_super + (bg * sbi->s_blocks_per_group)); +} + struct super_block * ext2_read_super (struct super_block * sb, void * data, int silent) { struct buffer_head * bh; + struct ext2_sb_info * sbi = EXT2_SB(sb); struct ext2_super_block * es; - unsigned long sb_block = 1; - unsigned short resuid = EXT2_DEF_RESUID; - unsigned short resgid = EXT2_DEF_RESGID; - unsigned long logic_sb_block = 1; - unsigned long offset = 0; + unsigned long sb_block = get_sb_block(&data); + unsigned long block, logic_sb_block; + unsigned long offset; kdev_t dev = sb->s_dev; + unsigned long def_mount_opts; int blocksize = BLOCK_SIZE; int db_count; int i, j; @@ -423,12 +466,6 @@ if(blocksize < BLOCK_SIZE ) blocksize = BLOCK_SIZE; - sb->u.ext2_sb.s_mount_opt = 0; - if (!parse_options ((char *) data, &sb_block, &resuid, &resgid, - &sb->u.ext2_sb.s_mount_opt)) { - return NULL; - } - if (set_blocksize(dev, blocksize) < 0) { printk ("EXT2-fs: unable to set blocksize %d\n", blocksize); return NULL; @@ -436,14 +473,11 @@ sb->s_blocksize = blocksize; /* - * If the superblock doesn't start on a sector boundary, - * calculate the offset. FIXME(eric) this doesn't make sense - * that we would have to do this. + * If the superblock doesn't start on a hardware sector boundary, + * calculate the offset. */ - if (blocksize != BLOCK_SIZE) { - logic_sb_block = (sb_block*BLOCK_SIZE) / blocksize; - offset = (sb_block*BLOCK_SIZE) % blocksize; - } + logic_sb_block = (sb_block*BLOCK_SIZE) / blocksize; + offset = (sb_block*BLOCK_SIZE) % blocksize; if (!(bh = sb_bread(sb, logic_sb_block))) { printk ("EXT2-fs: unable to read superblock\n"); @@ -462,6 +496,27 @@ bdevname(dev)); goto failed_mount; } + + /* Set defaults before we parse the mount options */ + def_mount_opts = le32_to_cpu(es->s_default_mount_opts); + if (def_mount_opts & EXT2_DEFM_DEBUG) + set_opt(sbi->s_mount_opt, DEBUG); + if (def_mount_opts & EXT2_DEFM_BSDGROUPS) + set_opt(sbi->s_mount_opt, GRPID); + if (def_mount_opts & EXT2_DEFM_UID16) + set_opt(sbi->s_mount_opt, NO_UID32); + + if (le16_to_cpu(sbi->s_es->s_errors) == EXT2_ERRORS_PANIC) + set_opt(sbi->s_mount_opt, ERRORS_PANIC); + else if (le16_to_cpu(sbi->s_es->s_errors) == EXT2_ERRORS_RO) + set_opt(sbi->s_mount_opt, ERRORS_RO); + + sbi->s_resuid = le16_to_cpu(es->s_def_resuid); + sbi->s_resgid = le16_to_cpu(es->s_def_resgid); + + if (!parse_options ((char *) data, sbi)) + goto failed_mount; + if (le32_to_cpu(es->s_rev_level) == EXT2_GOOD_OLD_REV && (EXT2_HAS_COMPAT_FEATURE(sb, ~0U) || EXT2_HAS_RO_COMPAT_FEATURE(sb, ~0U) || @@ -522,14 +577,16 @@ } if (le32_to_cpu(es->s_rev_level) == EXT2_GOOD_OLD_REV) { - sb->u.ext2_sb.s_inode_size = EXT2_GOOD_OLD_INODE_SIZE; - sb->u.ext2_sb.s_first_ino = EXT2_GOOD_OLD_FIRST_INO; + sbi->s_inode_size = EXT2_GOOD_OLD_INODE_SIZE; + sbi->s_first_ino = EXT2_GOOD_OLD_FIRST_INO; } else { - sb->u.ext2_sb.s_inode_size = le16_to_cpu(es->s_inode_size); - sb->u.ext2_sb.s_first_ino = le32_to_cpu(es->s_first_ino); - if (sb->u.ext2_sb.s_inode_size != EXT2_GOOD_OLD_INODE_SIZE) { + sbi->s_inode_size = le16_to_cpu(es->s_inode_size); + sbi->s_first_ino = le32_to_cpu(es->s_first_ino); + if ((sbi->s_inode_size < EXT2_GOOD_OLD_INODE_SIZE) || + (sbi->s_inode_size & (sbi->s_inode_size - 1)) || + (sbi->s_inode_size > blocksize)) { printk ("EXT2-fs: unsupported inode size: %d\n", - sb->u.ext2_sb.s_inode_size); + sbi->s_inode_size); goto failed_mount; } } @@ -550,14 +607,6 @@ sb->u.ext2_sb.s_desc_per_block = sb->s_blocksize / sizeof (struct ext2_group_desc); sb->u.ext2_sb.s_sbh = bh; - if (resuid != EXT2_DEF_RESUID) - sb->u.ext2_sb.s_resuid = resuid; - else - sb->u.ext2_sb.s_resuid = le16_to_cpu(es->s_def_resuid); - if (resgid != EXT2_DEF_RESGID) - sb->u.ext2_sb.s_resgid = resgid; - else - sb->u.ext2_sb.s_resgid = le16_to_cpu(es->s_def_resgid); sb->u.ext2_sb.s_mount_state = le16_to_cpu(es->s_state); sb->u.ext2_sb.s_addr_per_block_bits = log2 (EXT2_ADDR_PER_BLOCK(sb)); @@ -610,14 +659,22 @@ printk ("EXT2-fs: not enough memory\n"); goto failed_mount; } + sbi->s_debts = kmalloc(sbi->s_groups_count * sizeof(*sbi->s_debts), + GFP_KERNEL); + if (!sbi->s_debts) { + printk ("EXT2-fs: not enough memory\n"); + goto failed_mount_group_desc; + } + memset(sbi->s_debts, 0, sbi->s_groups_count * sizeof(*sbi->s_debts)); for (i = 0; i < db_count; i++) { - sb->u.ext2_sb.s_group_desc[i] = sb_bread(sb, logic_sb_block + i + 1); - if (!sb->u.ext2_sb.s_group_desc[i]) { + block = descriptor_loc(sb, logic_sb_block, i); + sbi->s_group_desc[i] = sb_bread(sb, block); + if (!sbi->s_group_desc[i]) { for (j = 0; j < i; j++) - brelse (sb->u.ext2_sb.s_group_desc[j]); - kfree(sb->u.ext2_sb.s_group_desc); + brelse (sbi->s_group_desc[j]); + kfree(sbi->s_group_desc); printk ("EXT2-fs: unable to read group descriptors\n"); - goto failed_mount; + goto failed_mount_group_desc; } } if (!ext2_check_descriptors (sb)) { @@ -634,6 +691,7 @@ sb->u.ext2_sb.s_loaded_inode_bitmaps = 0; sb->u.ext2_sb.s_loaded_block_bitmaps = 0; sb->u.ext2_sb.s_gdb_count = db_count; + sb->u.ext2_sb.s_dir_count = ext2_count_dirs(sb); /* * set up enough so that it can read an inode */ @@ -654,7 +712,10 @@ failed_mount2: for (i = 0; i < db_count; i++) brelse(sb->u.ext2_sb.s_group_desc[i]); +failed_mount_group_desc: kfree(sb->u.ext2_sb.s_group_desc); + if (sb->u.ext2_sb.s_debts) + kfree(sb->u.ext2_sb.s_debts); failed_mount: brelse(bh); return NULL; @@ -709,24 +770,16 @@ int ext2_remount (struct super_block * sb, int * flags, char * data) { + struct ext2_sb_info * sbi = EXT2_SB(sb); struct ext2_super_block * es; - unsigned short resuid = sb->u.ext2_sb.s_resuid; - unsigned short resgid = sb->u.ext2_sb.s_resgid; - unsigned long new_mount_opt; - unsigned long tmp; /* * Allow the "check" option to be passed as a remount option. */ - new_mount_opt = sb->u.ext2_sb.s_mount_opt; - if (!parse_options (data, &tmp, &resuid, &resgid, - &new_mount_opt)) + if (!parse_options (data, sbi)) return -EINVAL; - sb->u.ext2_sb.s_mount_opt = new_mount_opt; - sb->u.ext2_sb.s_resuid = resuid; - sb->u.ext2_sb.s_resgid = resgid; - es = sb->u.ext2_sb.s_es; + es = sbi->s_es; if ((*flags & MS_RDONLY) == (sb->s_flags & MS_RDONLY)) return 0; if (*flags & MS_RDONLY) { diff -Nru a/fs/ext3/Makefile b/fs/ext3/Makefile --- a/fs/ext3/Makefile Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/Makefile Thu Nov 7 11:58:05 2002 @@ -10,7 +10,7 @@ O_TARGET := ext3.o obj-y := balloc.o bitmap.o dir.o file.o fsync.o ialloc.o inode.o \ - ioctl.o namei.o super.o symlink.o + ioctl.o namei.o super.o symlink.o hash.o obj-m := $(O_TARGET) include $(TOPDIR)/Rules.make diff -Nru a/fs/ext3/dir.c b/fs/ext3/dir.c --- a/fs/ext3/dir.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/dir.c Thu Nov 7 11:58:05 2002 @@ -21,12 +21,16 @@ #include #include #include +#include +#include static unsigned char ext3_filetype_table[] = { DT_UNKNOWN, DT_REG, DT_DIR, DT_CHR, DT_BLK, DT_FIFO, DT_SOCK, DT_LNK }; static int ext3_readdir(struct file *, void *, filldir_t); +static int ext3_dx_readdir(struct file * filp, + void * dirent, filldir_t filldir); struct file_operations ext3_dir_operations = { read: generic_read_dir, @@ -35,6 +39,17 @@ fsync: ext3_sync_file, /* BKL held */ }; + +static unsigned char get_dtype(struct super_block *sb, int filetype) +{ + if (!EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_FILETYPE) || + (filetype >= EXT3_FT_MAX)) + return DT_UNKNOWN; + + return (ext3_filetype_table[filetype]); +} + + int ext3_check_dir_entry (const char * function, struct inode * dir, struct ext3_dir_entry_2 * de, struct buffer_head * bh, @@ -79,6 +94,16 @@ sb = inode->i_sb; + if (is_dx(inode)) { + err = ext3_dx_readdir(filp, dirent, filldir); + if (err != ERR_BAD_DX_DIR) + return err; + /* + * We don't set the inode dirty flag since it's not + * critical that it get flushed back to the disk. + */ + EXT3_I(filp->f_dentry->d_inode)->i_flags &= ~EXT3_INDEX_FL; + } stored = 0; bh = NULL; offset = filp->f_pos & (sb->s_blocksize - 1); @@ -162,18 +187,12 @@ * during the copy operation. */ unsigned long version = filp->f_version; - unsigned char d_type = DT_UNKNOWN; - if (EXT3_HAS_INCOMPAT_FEATURE(sb, - EXT3_FEATURE_INCOMPAT_FILETYPE) - && de->file_type < EXT3_FT_MAX) - d_type = - ext3_filetype_table[de->file_type]; error = filldir(dirent, de->name, de->name_len, filp->f_pos, le32_to_cpu(de->inode), - d_type); + get_dtype(sb, de->file_type)); if (error) break; if (version != filp->f_version) @@ -188,3 +207,272 @@ UPDATE_ATIME(inode); return 0; } + +#ifdef CONFIG_EXT3_INDEX +/* + * These functions convert from the major/minor hash to an f_pos + * value. + * + * Currently we only use major hash numer. This is unfortunate, but + * on 32-bit machines, the same VFS interface is used for lseek and + * llseek, so if we use the 64 bit offset, then the 32-bit versions of + * lseek/telldir/seekdir will blow out spectacularly, and from within + * the ext2 low-level routine, we don't know if we're being called by + * a 64-bit version of the system call or the 32-bit version of the + * system call. Worse yet, NFSv2 only allows for a 32-bit readdir + * cookie. Sigh. + */ +#define hash2pos(major, minor) (major >> 1) +#define pos2maj_hash(pos) ((pos << 1) & 0xffffffff) +#define pos2min_hash(pos) (0) + +/* + * This structure holds the nodes of the red-black tree used to store + * the directory entry in hash order. + */ +struct fname { + __u32 hash; + __u32 minor_hash; + rb_node_t rb_hash; + struct fname *next; + __u32 inode; + __u8 name_len; + __u8 file_type; + char name[0]; +}; + +/* + * This functoin implements a non-recursive way of freeing all of the + * nodes in the red-black tree. + */ +static void free_rb_tree_fname(rb_root_t *root) +{ + rb_node_t *n = root->rb_node; + rb_node_t *parent; + struct fname *fname; + + while (n) { + /* Do the node's children first */ + if ((n)->rb_left) { + n = n->rb_left; + continue; + } + if (n->rb_right) { + n = n->rb_right; + continue; + } + /* + * The node has no children; free it, and then zero + * out parent's link to it. Finally go to the + * beginning of the loop and try to free the parent + * node. + */ + parent = n->rb_parent; + fname = rb_entry(n, struct fname, rb_hash); + kfree(fname); + if (!parent) + root->rb_node = 0; + else if (parent->rb_left == n) + parent->rb_left = 0; + else if (parent->rb_right == n) + parent->rb_right = 0; + n = parent; + } + root->rb_node = 0; +} + + +struct dir_private_info *create_dir_info(loff_t pos) +{ + struct dir_private_info *p; + + p = kmalloc(sizeof(struct dir_private_info), GFP_KERNEL); + if (!p) + return NULL; + p->root.rb_node = 0; + p->curr_node = 0; + p->extra_fname = 0; + p->last_pos = 0; + p->curr_hash = pos2maj_hash(pos); + p->curr_minor_hash = pos2min_hash(pos); + p->next_hash = 0; + return p; +} + +void ext3_htree_free_dir_info(struct dir_private_info *p) +{ + free_rb_tree_fname(&p->root); + kfree(p); +} + +/* + * Given a directory entry, enter it into the fname rb tree. + */ +int ext3_htree_store_dirent(struct file *dir_file, __u32 hash, + __u32 minor_hash, + struct ext3_dir_entry_2 *dirent) +{ + rb_node_t **p, *parent = NULL; + struct fname * fname, *new_fn; + struct dir_private_info *info; + int len; + + info = (struct dir_private_info *) dir_file->private_data; + p = &info->root.rb_node; + + /* Create and allocate the fname structure */ + len = sizeof(struct fname) + dirent->name_len + 1; + new_fn = kmalloc(len, GFP_KERNEL); + if (!new_fn) + return -ENOMEM; + memset(new_fn, 0, len); + new_fn->hash = hash; + new_fn->minor_hash = minor_hash; + new_fn->inode = le32_to_cpu(dirent->inode); + new_fn->name_len = dirent->name_len; + new_fn->file_type = dirent->file_type; + memcpy(new_fn->name, dirent->name, dirent->name_len); + new_fn->name[dirent->name_len] = 0; + + while (*p) { + parent = *p; + fname = rb_entry(parent, struct fname, rb_hash); + + /* + * If the hash and minor hash match up, then we put + * them on a linked list. This rarely happens... + */ + if ((new_fn->hash == fname->hash) && + (new_fn->minor_hash == fname->minor_hash)) { + new_fn->next = fname->next; + fname->next = new_fn; + return 0; + } + + if (new_fn->hash < fname->hash) + p = &(*p)->rb_left; + else if (new_fn->hash > fname->hash) + p = &(*p)->rb_right; + else if (new_fn->minor_hash < fname->minor_hash) + p = &(*p)->rb_left; + else /* if (new_fn->minor_hash > fname->minor_hash) */ + p = &(*p)->rb_right; + } + + rb_link_node(&new_fn->rb_hash, parent, p); + rb_insert_color(&new_fn->rb_hash, &info->root); + return 0; +} + + + +/* + * This is a helper function for ext3_dx_readdir. It calls filldir + * for all entres on the fname linked list. (Normally there is only + * one entry on the linked list, unless there are 62 bit hash collisions.) + */ +static int call_filldir(struct file * filp, void * dirent, + filldir_t filldir, struct fname *fname) +{ + struct dir_private_info *info = filp->private_data; + loff_t curr_pos; + struct inode *inode = filp->f_dentry->d_inode; + struct super_block * sb; + int error; + + sb = inode->i_sb; + + if (!fname) { + printk("call_filldir: called with null fname?!?\n"); + return 0; + } + curr_pos = hash2pos(fname->hash, fname->minor_hash); + while (fname) { + error = filldir(dirent, fname->name, + fname->name_len, curr_pos, + fname->inode, + get_dtype(sb, fname->file_type)); + if (error) { + filp->f_pos = curr_pos; + info->extra_fname = fname->next; + return error; + } + fname = fname->next; + } + return 0; +} + +static int ext3_dx_readdir(struct file * filp, + void * dirent, filldir_t filldir) +{ + struct dir_private_info *info = filp->private_data; + struct inode *inode = filp->f_dentry->d_inode; + struct fname *fname; + int ret; + + if (!info) { + info = create_dir_info(filp->f_pos); + if (!info) + return -ENOMEM; + filp->private_data = info; + } + + /* Some one has messed with f_pos; reset the world */ + if (info->last_pos != filp->f_pos) { + free_rb_tree_fname(&info->root); + info->curr_node = 0; + info->extra_fname = 0; + info->curr_hash = pos2maj_hash(filp->f_pos); + info->curr_minor_hash = pos2min_hash(filp->f_pos); + } + + /* + * If there are any leftover names on the hash collision + * chain, return them first. + */ + if (info->extra_fname && + call_filldir(filp, dirent, filldir, info->extra_fname)) + goto finished; + + if (!info->curr_node) + info->curr_node = rb_get_first(&info->root); + + while (1) { + /* + * Fill the rbtree if we have no more entries, + * or the inode has changed since we last read in the + * cached entries. + */ + if ((!info->curr_node) || + (filp->f_version != inode->i_version)) { + info->curr_node = 0; + free_rb_tree_fname(&info->root); + filp->f_version = inode->i_version; + ret = ext3_htree_fill_tree(filp, info->curr_hash, + info->curr_minor_hash, + &info->next_hash); + if (ret < 0) + return ret; + if (ret == 0) + break; + info->curr_node = rb_get_first(&info->root); + } + + fname = rb_entry(info->curr_node, struct fname, rb_hash); + info->curr_hash = fname->hash; + info->curr_minor_hash = fname->minor_hash; + if (call_filldir(filp, dirent, filldir, fname)) + break; + + info->curr_node = rb_get_next(info->curr_node); + if (!info->curr_node) { + info->curr_hash = info->next_hash; + info->curr_minor_hash = 0; + } + } +finished: + info->last_pos = filp->f_pos; + UPDATE_ATIME(inode); + return 0; +} +#endif diff -Nru a/fs/ext3/file.c b/fs/ext3/file.c --- a/fs/ext3/file.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/file.c Thu Nov 7 11:58:05 2002 @@ -35,6 +35,9 @@ { if (filp->f_mode & FMODE_WRITE) ext3_discard_prealloc (inode); + if (is_dx(inode) && filp->private_data) + ext3_htree_free_dir_info(filp->private_data); + return 0; } diff -Nru a/fs/ext3/hash.c b/fs/ext3/hash.c --- /dev/null Wed Dec 31 16:00:00 1969 +++ b/fs/ext3/hash.c Thu Nov 7 11:58:05 2002 @@ -0,0 +1,215 @@ +/* + * linux/fs/ext3/hash.c + * + * Copyright (C) 2002 by Theodore Ts'o + * + * This file is released under the GPL v2. + * + * This file may be redistributed under the terms of the GNU Public + * License. + */ + +#include +#include +#include +#include + +#define DELTA 0x9E3779B9 + +static void TEA_transform(__u32 buf[4], __u32 const in[]) +{ + __u32 sum = 0; + __u32 b0 = buf[0], b1 = buf[1]; + __u32 a = in[0], b = in[1], c = in[2], d = in[3]; + int n = 16; + + do { + sum += DELTA; + b0 += ((b1 << 4)+a) ^ (b1+sum) ^ ((b1 >> 5)+b); + b1 += ((b0 << 4)+c) ^ (b0+sum) ^ ((b0 >> 5)+d); + } while(--n); + + buf[0] += b0; + buf[1] += b1; +} + +/* F, G and H are basic MD4 functions: selection, majority, parity */ +#define F(x, y, z) ((z) ^ ((x) & ((y) ^ (z)))) +#define G(x, y, z) (((x) & (y)) + (((x) ^ (y)) & (z))) +#define H(x, y, z) ((x) ^ (y) ^ (z)) + +/* + * The generic round function. The application is so specific that + * we don't bother protecting all the arguments with parens, as is generally + * good macro practice, in favor of extra legibility. + * Rotation is separate from addition to prevent recomputation + */ +#define ROUND(f, a, b, c, d, x, s) \ + (a += f(b, c, d) + x, a = (a << s) | (a >> (32-s))) +#define K1 0 +#define K2 013240474631UL +#define K3 015666365641UL + +/* + * Basic cut-down MD4 transform. Returns only 32 bits of result. + */ +static void halfMD4Transform (__u32 buf[4], __u32 const in[]) +{ + __u32 a = buf[0], b = buf[1], c = buf[2], d = buf[3]; + + /* Round 1 */ + ROUND(F, a, b, c, d, in[0] + K1, 3); + ROUND(F, d, a, b, c, in[1] + K1, 7); + ROUND(F, c, d, a, b, in[2] + K1, 11); + ROUND(F, b, c, d, a, in[3] + K1, 19); + ROUND(F, a, b, c, d, in[4] + K1, 3); + ROUND(F, d, a, b, c, in[5] + K1, 7); + ROUND(F, c, d, a, b, in[6] + K1, 11); + ROUND(F, b, c, d, a, in[7] + K1, 19); + + /* Round 2 */ + ROUND(G, a, b, c, d, in[1] + K2, 3); + ROUND(G, d, a, b, c, in[3] + K2, 5); + ROUND(G, c, d, a, b, in[5] + K2, 9); + ROUND(G, b, c, d, a, in[7] + K2, 13); + ROUND(G, a, b, c, d, in[0] + K2, 3); + ROUND(G, d, a, b, c, in[2] + K2, 5); + ROUND(G, c, d, a, b, in[4] + K2, 9); + ROUND(G, b, c, d, a, in[6] + K2, 13); + + /* Round 3 */ + ROUND(H, a, b, c, d, in[3] + K3, 3); + ROUND(H, d, a, b, c, in[7] + K3, 9); + ROUND(H, c, d, a, b, in[2] + K3, 11); + ROUND(H, b, c, d, a, in[6] + K3, 15); + ROUND(H, a, b, c, d, in[1] + K3, 3); + ROUND(H, d, a, b, c, in[5] + K3, 9); + ROUND(H, c, d, a, b, in[0] + K3, 11); + ROUND(H, b, c, d, a, in[4] + K3, 15); + + buf[0] += a; + buf[1] += b; + buf[2] += c; + buf[3] += d; +} + +#undef ROUND +#undef F +#undef G +#undef H +#undef K1 +#undef K2 +#undef K3 + +/* The old legacy hash */ +static __u32 dx_hack_hash (const char *name, int len) +{ + __u32 hash0 = 0x12a3fe2d, hash1 = 0x37abe8f9; + while (len--) { + __u32 hash = hash1 + (hash0 ^ (*name++ * 7152373)); + + if (hash & 0x80000000) hash -= 0x7fffffff; + hash1 = hash0; + hash0 = hash; + } + return (hash0 << 1); +} + +static void str2hashbuf(const char *msg, int len, __u32 *buf, int num) +{ + __u32 pad, val; + int i; + + pad = (__u32)len | ((__u32)len << 8); + pad |= pad << 16; + + val = pad; + if (len > num*4) + len = num * 4; + for (i=0; i < len; i++) { + if ((i % 4) == 0) + val = pad; + val = msg[i] + (val << 8); + if ((i % 4) == 3) { + *buf++ = val; + val = pad; + num--; + } + } + if (--num >= 0) + *buf++ = val; + while (--num >= 0) + *buf++ = pad; +} + +/* + * Returns the hash of a filename. If len is 0 and name is NULL, then + * this function can be used to test whether or not a hash version is + * supported. + * + * The seed is an 4 longword (32 bits) "secret" which can be used to + * uniquify a hash. If the seed is all zero's, then some default seed + * may be used. + * + * A particular hash version specifies whether or not the seed is + * represented, and whether or not the returned hash is 32 bits or 64 + * bits. 32 bit hashes will return 0 for the minor hash. + */ +int ext3fs_dirhash(const char *name, int len, struct dx_hash_info *hinfo) +{ + __u32 hash; + __u32 minor_hash = 0; + const char *p; + int i; + __u32 in[8], buf[4]; + + /* Initialize the default seed for the hash checksum functions */ + buf[0] = 0x67452301; + buf[1] = 0xefcdab89; + buf[2] = 0x98badcfe; + buf[3] = 0x10325476; + + /* Check to see if the seed is all zero's */ + if (hinfo->seed) { + for (i=0; i < 4; i++) { + if (hinfo->seed[i]) + break; + } + if (i < 4) + memcpy(buf, hinfo->seed, sizeof(buf)); + } + + switch (hinfo->hash_version) { + case DX_HASH_LEGACY: + hash = dx_hack_hash(name, len); + break; + case DX_HASH_HALF_MD4: + p = name; + while (len > 0) { + str2hashbuf(p, len, in, 8); + halfMD4Transform(buf, in); + len -= 32; + p += 32; + } + minor_hash = buf[2]; + hash = buf[1]; + break; + case DX_HASH_TEA: + p = name; + while (len > 0) { + str2hashbuf(p, len, in, 4); + TEA_transform(buf, in); + len -= 16; + p += 16; + } + hash = buf[0]; + minor_hash = buf[1]; + break; + default: + hinfo->hash = 0; + return -1; + } + hinfo->hash = hash & ~1; + hinfo->minor_hash = minor_hash; + return 0; +} diff -Nru a/fs/ext3/ialloc.c b/fs/ext3/ialloc.c --- a/fs/ext3/ialloc.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/ialloc.c Thu Nov 7 11:58:05 2002 @@ -21,6 +21,7 @@ #include #include #include +#include #include #include @@ -293,6 +294,199 @@ * the groups with above-average free space, that group with the fewest * directories already is chosen. * + * For other inodes, search forward from the parent directory\'s block + * group to find a free inode. + */ +static int find_group_dir(struct super_block *sb, const struct inode *parent) +{ + struct ext3_super_block * es = EXT3_SB(sb)->s_es; + int ngroups = EXT3_SB(sb)->s_groups_count; + int avefreei = le32_to_cpu(es->s_free_inodes_count) / ngroups; + struct ext3_group_desc *desc, *best_desc = NULL; + struct buffer_head *bh; + int group, best_group = -1; + + for (group = 0; group < ngroups; group++) { + desc = ext3_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) < avefreei) + continue; + if (!best_desc || + (le16_to_cpu(desc->bg_free_blocks_count) > + le16_to_cpu(best_desc->bg_free_blocks_count))) { + best_group = group; + best_desc = desc; + } + } + return best_group; +} + +/* + * Orlov's allocator for directories. + * + * We always try to spread first-level directories. + * + * If there are blockgroups with both free inodes and free blocks counts + * not worse than average we return one with smallest directory count. + * Otherwise we simply return a random group. + * + * For the rest rules look so: + * + * It's OK to put directory into a group unless + * it has too many directories already (max_dirs) or + * it has too few free inodes left (min_inodes) or + * it has too few free blocks left (min_blocks) or + * it's already running too large debt (max_debt). + * Parent's group is prefered, if it doesn't satisfy these + * conditions we search cyclically through the rest. If none + * of the groups look good we just look for a group with more + * free inodes than average (starting at parent's group). + * + * Debt is incremented each time we allocate a directory and decremented + * when we allocate an inode, within 0--255. + */ + +#define INODE_COST 64 +#define BLOCK_COST 256 + +static int find_group_orlov(struct super_block *sb, const struct inode *parent) +{ + int parent_group = EXT3_I(parent)->i_block_group; + struct ext3_sb_info *sbi = EXT3_SB(sb); + struct ext3_super_block *es = sbi->s_es; + int ngroups = sbi->s_groups_count; + int inodes_per_group = EXT3_INODES_PER_GROUP(sb); + int avefreei = le32_to_cpu(es->s_free_inodes_count) / ngroups; + int avefreeb = le32_to_cpu(es->s_free_blocks_count) / ngroups; + int blocks_per_dir; + int ndirs = sbi->s_dir_count; + int max_debt, max_dirs, min_blocks, min_inodes; + int group = -1, i; + struct ext3_group_desc *desc; + struct buffer_head *bh; + + if ((parent == sb->s_root->d_inode) || + (parent->i_flags & EXT3_TOPDIR_FL)) { + int best_ndir = inodes_per_group; + int best_group = -1; + + get_random_bytes(&group, sizeof(group)); + parent_group = (unsigned)group % ngroups; + for (i = 0; i < ngroups; i++) { + group = (parent_group + i) % ngroups; + desc = ext3_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (le16_to_cpu(desc->bg_used_dirs_count) >= best_ndir) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) < avefreei) + continue; + if (le16_to_cpu(desc->bg_free_blocks_count) < avefreeb) + continue; + best_group = group; + best_ndir = le16_to_cpu(desc->bg_used_dirs_count); + } + if (best_group >= 0) + return best_group; + goto fallback; + } + + blocks_per_dir = (le32_to_cpu(es->s_blocks_count) - + le32_to_cpu(es->s_free_blocks_count)) / ndirs; + + max_dirs = ndirs / ngroups + inodes_per_group / 16; + min_inodes = avefreei - inodes_per_group / 4; + min_blocks = avefreeb - EXT3_BLOCKS_PER_GROUP(sb) / 4; + + max_debt = EXT3_BLOCKS_PER_GROUP(sb) / max(blocks_per_dir, BLOCK_COST); + if (max_debt * INODE_COST > inodes_per_group) + max_debt = inodes_per_group / INODE_COST; + if (max_debt > 255) + max_debt = 255; + if (max_debt == 0) + max_debt = 1; + + for (i = 0; i < ngroups; i++) { + group = (parent_group + i) % ngroups; + desc = ext3_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (sbi->s_debts[group] >= max_debt) + continue; + if (le16_to_cpu(desc->bg_used_dirs_count) >= max_dirs) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) < min_inodes) + continue; + if (le16_to_cpu(desc->bg_free_blocks_count) < min_blocks) + continue; + return group; + } + +fallback: + for (i = 0; i < ngroups; i++) { + group = (parent_group + i) % ngroups; + desc = ext3_get_group_desc (sb, group, &bh); + if (!desc || !desc->bg_free_inodes_count) + continue; + if (le16_to_cpu(desc->bg_free_inodes_count) >= avefreei) + return group; + } + + return -1; +} + +static int find_group_other(struct super_block *sb, const struct inode *parent) +{ + int parent_group = EXT3_I(parent)->i_block_group; + int ngroups = EXT3_SB(sb)->s_groups_count; + struct ext3_group_desc *desc; + struct buffer_head *bh; + int group, i; + + /* + * Try to place the inode in its parent directory + */ + group = parent_group; + desc = ext3_get_group_desc (sb, group, &bh); + if (desc && le16_to_cpu(desc->bg_free_inodes_count)) + return group; + + /* + * Use a quadratic hash to find a group with a + * free inode + */ + for (i = 1; i < ngroups; i <<= 1) { + group += i; + if (group >= ngroups) + group -= ngroups; + desc = ext3_get_group_desc (sb, group, &bh); + if (desc && le16_to_cpu(desc->bg_free_inodes_count)) + return group; + } + + /* + * That failed: try linear search for a free inode + */ + group = parent_group + 1; + for (i = 2; i < ngroups; i++) { + if (++group >= ngroups) + group = 0; + desc = ext3_get_group_desc (sb, group, &bh); + if (desc && le16_to_cpu(desc->bg_free_inodes_count)) + return group; + } + + return -1; +} + +/* + * There are two policies for allocating an inode. If the new inode is + * a directory, then a forward search is made for a block group with both + * free space and a low directory-to-inode ratio; if that fails, then of + * the groups with above-average free space, that group with the fewest + * directories already is chosen. + * * For other inodes, search forward from the parent directory's block * group to find a free inode. */ @@ -302,11 +496,11 @@ struct super_block * sb; struct buffer_head * bh; struct buffer_head * bh2; - int i, j, avefreei; + int group; + ino_t ino; struct inode * inode; int bitmap_nr; struct ext3_group_desc * gdp; - struct ext3_group_desc * tmp; struct ext3_super_block * es; int err = 0; @@ -323,94 +517,36 @@ lock_super (sb); es = sb->u.ext3_sb.s_es; repeat: - gdp = NULL; - i = 0; - if (S_ISDIR(mode)) { - avefreei = le32_to_cpu(es->s_free_inodes_count) / - sb->u.ext3_sb.s_groups_count; - if (!gdp) { - for (j = 0; j < sb->u.ext3_sb.s_groups_count; j++) { - struct buffer_head *temp_buffer; - tmp = ext3_get_group_desc (sb, j, &temp_buffer); - if (tmp && - le16_to_cpu(tmp->bg_free_inodes_count) && - le16_to_cpu(tmp->bg_free_inodes_count) >= - avefreei) { - if (!gdp || (le16_to_cpu(tmp->bg_free_blocks_count) > - le16_to_cpu(gdp->bg_free_blocks_count))) { - i = j; - gdp = tmp; - bh2 = temp_buffer; - } - } - } - } - } else { - /* - * Try to place the inode in its parent directory - */ - i = dir->u.ext3_i.i_block_group; - tmp = ext3_get_group_desc (sb, i, &bh2); - if (tmp && le16_to_cpu(tmp->bg_free_inodes_count)) - gdp = tmp; + if (test_opt (sb, OLDALLOC)) + group = find_group_dir(sb, dir); else - { - /* - * Use a quadratic hash to find a group with a - * free inode - */ - for (j = 1; j < sb->u.ext3_sb.s_groups_count; j <<= 1) { - i += j; - if (i >= sb->u.ext3_sb.s_groups_count) - i -= sb->u.ext3_sb.s_groups_count; - tmp = ext3_get_group_desc (sb, i, &bh2); - if (tmp && - le16_to_cpu(tmp->bg_free_inodes_count)) { - gdp = tmp; - break; - } - } - } - if (!gdp) { - /* - * That failed: try linear search for a free inode - */ - i = dir->u.ext3_i.i_block_group + 1; - for (j = 2; j < sb->u.ext3_sb.s_groups_count; j++) { - if (++i >= sb->u.ext3_sb.s_groups_count) - i = 0; - tmp = ext3_get_group_desc (sb, i, &bh2); - if (tmp && - le16_to_cpu(tmp->bg_free_inodes_count)) { - gdp = tmp; - break; - } - } - } - } - + group = find_group_orlov(sb, dir); + } else + group = find_group_other(sb, dir); + err = -ENOSPC; - if (!gdp) + if (!group == -1) goto out; err = -EIO; - bitmap_nr = load_inode_bitmap (sb, i); + bitmap_nr = load_inode_bitmap (sb, group); if (bitmap_nr < 0) goto fail; bh = sb->u.ext3_sb.s_inode_bitmap[bitmap_nr]; + gdp = ext3_get_group_desc (sb, group, &bh2); - if ((j = ext3_find_first_zero_bit ((unsigned long *) bh->b_data, + if ((ino = ext3_find_first_zero_bit ((unsigned long *) bh->b_data, EXT3_INODES_PER_GROUP(sb))) < EXT3_INODES_PER_GROUP(sb)) { BUFFER_TRACE(bh, "get_write_access"); err = ext3_journal_get_write_access(handle, bh); if (err) goto fail; - if (ext3_set_bit (j, bh->b_data)) { + if (ext3_set_bit (ino, bh->b_data)) { ext3_error (sb, "ext3_new_inode", - "bit already set for inode %d", j); + "bit already set for inode %lu", ino); goto repeat; } BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata"); @@ -420,7 +556,7 @@ if (le16_to_cpu(gdp->bg_free_inodes_count) != 0) { ext3_error (sb, "ext3_new_inode", "Free inodes count corrupted in group %d", - i); + group); /* Is it really ENOSPC? */ err = -ENOSPC; if (sb->s_flags & MS_RDONLY) @@ -436,11 +572,11 @@ } goto repeat; } - j += i * EXT3_INODES_PER_GROUP(sb) + 1; - if (j < EXT3_FIRST_INO(sb) || j > le32_to_cpu(es->s_inodes_count)) { + ino += group * EXT3_INODES_PER_GROUP(sb) + 1; + if (ino < EXT3_FIRST_INO(sb) || ino > le32_to_cpu(es->s_inodes_count)) { ext3_error (sb, "ext3_new_inode", "reserved inode or inode > inodes count - " - "block_group = %d,inode=%d", i, j); + "block_group = %d, inode=%lu", group, ino); err = -EIO; goto fail; } @@ -450,9 +586,11 @@ if (err) goto fail; gdp->bg_free_inodes_count = cpu_to_le16(le16_to_cpu(gdp->bg_free_inodes_count) - 1); - if (S_ISDIR(mode)) + if (S_ISDIR(mode)) { gdp->bg_used_dirs_count = cpu_to_le16(le16_to_cpu(gdp->bg_used_dirs_count) + 1); + EXT3_SB(sb)->s_dir_count++; + } BUFFER_TRACE(bh2, "call ext3_journal_dirty_metadata"); err = ext3_journal_dirty_metadata(handle, bh2); if (err) goto fail; @@ -478,7 +616,7 @@ inode->i_gid = current->fsgid; inode->i_mode = mode; - inode->i_ino = j; + inode->i_ino = ino; /* This is the optimal IO size (for stat), not the fs block size */ inode->i_blksize = PAGE_SIZE; inode->i_blocks = 0; @@ -498,7 +636,7 @@ #ifdef EXT3_PREALLOCATE inode->u.ext3_i.i_prealloc_count = 0; #endif - inode->u.ext3_i.i_block_group = i; + inode->u.ext3_i.i_block_group = group; if (inode->u.ext3_i.i_flags & EXT3_SYNC_FL) inode->i_flags |= S_SYNC; @@ -620,6 +758,21 @@ #else return le32_to_cpu(sb->u.ext3_sb.s_es->s_free_inodes_count); #endif +} + +/* Called at mount-time, super-block is locked */ +unsigned long ext3_count_dirs (struct super_block * sb) +{ + unsigned long count = 0; + int i; + + for (i = 0; i < EXT3_SB(sb)->s_groups_count; i++) { + struct ext3_group_desc *gdp = ext3_get_group_desc (sb, i, NULL); + if (!gdp) + continue; + count += le16_to_cpu(gdp->bg_used_dirs_count); + } + return count; } #ifdef CONFIG_EXT3_CHECK diff -Nru a/fs/ext3/inode.c b/fs/ext3/inode.c --- a/fs/ext3/inode.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/inode.c Thu Nov 7 11:58:05 2002 @@ -2194,6 +2194,11 @@ if (err) goto out_brelse; } + /* For fields not not tracking in the in-memory inode, + * initialise them to zero for new inodes. */ + if (EXT3_I(inode)->i_state & EXT3_STATE_NEW) + memset(raw_inode, 0, EXT3_SB(inode->i_sb)->s_inode_size); + raw_inode->i_mode = cpu_to_le16(inode->i_mode); if(!(test_opt(inode->i_sb, NO_UID32))) { raw_inode->i_uid_low = cpu_to_le16(low_16_bits(inode->i_uid)); @@ -2231,15 +2236,6 @@ raw_inode->i_faddr = cpu_to_le32(inode->u.ext3_i.i_faddr); raw_inode->i_frag = inode->u.ext3_i.i_frag_no; raw_inode->i_fsize = inode->u.ext3_i.i_frag_size; -#else - /* If we are not tracking these fields in the in-memory inode, - * then preserve them on disk, but still initialise them to zero - * for new inodes. */ - if (EXT3_I(inode)->i_state & EXT3_STATE_NEW) { - raw_inode->i_faddr = 0; - raw_inode->i_frag = 0; - raw_inode->i_fsize = 0; - } #endif raw_inode->i_file_acl = cpu_to_le32(inode->u.ext3_i.i_file_acl); if (!S_ISREG(inode->i_mode)) { diff -Nru a/fs/ext3/namei.c b/fs/ext3/namei.c --- a/fs/ext3/namei.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/namei.c Thu Nov 7 11:58:05 2002 @@ -16,6 +16,12 @@ * David S. Miller (davem@caip.rutgers.edu), 1995 * Directory entry file type support and forward compatibility hooks * for B-tree directories by Theodore Ts'o (tytso@mit.edu), 1998 + * Hash Tree Directory indexing (c) + * Daniel Phillips, 2001 + * Hash Tree Directory indexing porting + * Christopher Li, 2002 + * Hash Tree Directory indexing cleanup + * Theodore Ts'o, 2002 */ #include @@ -38,6 +44,642 @@ #define NAMEI_RA_SIZE (NAMEI_RA_CHUNKS * NAMEI_RA_BLOCKS) #define NAMEI_RA_INDEX(c,b) (((c) * NAMEI_RA_BLOCKS) + (b)) +static struct buffer_head *ext3_append(handle_t *handle, + struct inode *inode, + u32 *block, int *err) +{ + struct buffer_head *bh; + + *block = inode->i_size >> inode->i_sb->s_blocksize_bits; + + if ((bh = ext3_bread(handle, inode, *block, 1, err))) { + inode->i_size += inode->i_sb->s_blocksize; + EXT3_I(inode)->i_disksize = inode->i_size; + ext3_journal_get_write_access(handle,bh); + } + return bh; +} + +#ifndef assert +#define assert(test) J_ASSERT(test) +#endif + +#ifndef swap +#define swap(x, y) do { typeof(x) z = x; x = y; y = z; } while (0) +#endif + +typedef struct { u32 v; } le_u32; +typedef struct { u16 v; } le_u16; + +#ifdef DX_DEBUG +#define dxtrace(command) command +#else +#define dxtrace(command) +#endif + +struct fake_dirent +{ + /*le*/u32 inode; + /*le*/u16 rec_len; + u8 name_len; + u8 file_type; +}; + +struct dx_countlimit +{ + le_u16 limit; + le_u16 count; +}; + +struct dx_entry +{ + le_u32 hash; + le_u32 block; +}; + +/* + * dx_root_info is laid out so that if it should somehow get overlaid by a + * dirent the two low bits of the hash version will be zero. Therefore, the + * hash version mod 4 should never be 0. Sincerely, the paranoia department. + */ + +struct dx_root +{ + struct fake_dirent dot; + char dot_name[4]; + struct fake_dirent dotdot; + char dotdot_name[4]; + struct dx_root_info + { + le_u32 reserved_zero; + u8 hash_version; + u8 info_length; /* 8 */ + u8 indirect_levels; + u8 unused_flags; + } + info; + struct dx_entry entries[0]; +}; + +struct dx_node +{ + struct fake_dirent fake; + struct dx_entry entries[0]; +}; + + +struct dx_frame +{ + struct buffer_head *bh; + struct dx_entry *entries; + struct dx_entry *at; +}; + +struct dx_map_entry +{ + u32 hash; + u32 offs; +}; + +#ifdef CONFIG_EXT3_INDEX +static inline unsigned dx_get_block (struct dx_entry *entry); +static void dx_set_block (struct dx_entry *entry, unsigned value); +static inline unsigned dx_get_hash (struct dx_entry *entry); +static void dx_set_hash (struct dx_entry *entry, unsigned value); +static unsigned dx_get_count (struct dx_entry *entries); +static unsigned dx_get_limit (struct dx_entry *entries); +static void dx_set_count (struct dx_entry *entries, unsigned value); +static void dx_set_limit (struct dx_entry *entries, unsigned value); +static unsigned dx_root_limit (struct inode *dir, unsigned infosize); +static unsigned dx_node_limit (struct inode *dir); +static struct dx_frame *dx_probe(struct dentry *dentry, + struct inode *dir, + struct dx_hash_info *hinfo, + struct dx_frame *frame, + int *err); +static void dx_release (struct dx_frame *frames); +static int dx_make_map (struct ext3_dir_entry_2 *de, int size, + struct dx_hash_info *hinfo, struct dx_map_entry map[]); +static void dx_sort_map(struct dx_map_entry *map, unsigned count); +static struct ext3_dir_entry_2 *dx_move_dirents (char *from, char *to, + struct dx_map_entry *offsets, int count); +static struct ext3_dir_entry_2* dx_pack_dirents (char *base, int size); +static void dx_insert_block (struct dx_frame *frame, u32 hash, u32 block); +static int ext3_htree_next_block(struct inode *dir, __u32 hash, + struct dx_frame *frame, + struct dx_frame *frames, int *err, + __u32 *start_hash); +static struct buffer_head * ext3_dx_find_entry(struct dentry *dentry, + struct ext3_dir_entry_2 **res_dir, int *err); +static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry, + struct inode *inode); + +/* + * Future: use high four bits of block for coalesce-on-delete flags + * Mask them off for now. + */ + +static inline unsigned dx_get_block (struct dx_entry *entry) +{ + return le32_to_cpu(entry->block.v) & 0x00ffffff; +} + +static inline void dx_set_block (struct dx_entry *entry, unsigned value) +{ + entry->block.v = cpu_to_le32(value); +} + +static inline unsigned dx_get_hash (struct dx_entry *entry) +{ + return le32_to_cpu(entry->hash.v); +} + +static inline void dx_set_hash (struct dx_entry *entry, unsigned value) +{ + entry->hash.v = cpu_to_le32(value); +} + +static inline unsigned dx_get_count (struct dx_entry *entries) +{ + return le16_to_cpu(((struct dx_countlimit *) entries)->count.v); +} + +static inline unsigned dx_get_limit (struct dx_entry *entries) +{ + return le16_to_cpu(((struct dx_countlimit *) entries)->limit.v); +} + +static inline void dx_set_count (struct dx_entry *entries, unsigned value) +{ + ((struct dx_countlimit *) entries)->count.v = cpu_to_le16(value); +} + +static inline void dx_set_limit (struct dx_entry *entries, unsigned value) +{ + ((struct dx_countlimit *) entries)->limit.v = cpu_to_le16(value); +} + +static inline unsigned dx_root_limit (struct inode *dir, unsigned infosize) +{ + unsigned entry_space = dir->i_sb->s_blocksize - EXT3_DIR_REC_LEN(1) - + EXT3_DIR_REC_LEN(2) - infosize; + return 0? 20: entry_space / sizeof(struct dx_entry); +} + +static inline unsigned dx_node_limit (struct inode *dir) +{ + unsigned entry_space = dir->i_sb->s_blocksize - EXT3_DIR_REC_LEN(0); + return 0? 22: entry_space / sizeof(struct dx_entry); +} + +/* + * Debug + */ +#ifdef DX_DEBUG +struct stats +{ + unsigned names; + unsigned space; + unsigned bcount; +}; + +static struct stats dx_show_leaf(struct dx_hash_info *hinfo, struct ext3_dir_entry_2 *de, + int size, int show_names) +{ + unsigned names = 0, space = 0; + char *base = (char *) de; + struct dx_hash_info h = *hinfo; + + printk("names: "); + while ((char *) de < base + size) + { + if (de->inode) + { + if (show_names) + { + int len = de->name_len; + char *name = de->name; + while (len--) printk("%c", *name++); + ext3fs_dirhash(de->name, de->name_len, &h); + printk(":%x.%u ", h.hash, + ((char *) de - base)); + } + space += EXT3_DIR_REC_LEN(de->name_len); + names++; + } + de = (struct ext3_dir_entry_2 *) ((char *) de + le16_to_cpu(de->rec_len)); + } + printk("(%i)\n", names); + return (struct stats) { names, space, 1 }; +} + +struct stats dx_show_entries(struct dx_hash_info *hinfo, struct inode *dir, + struct dx_entry *entries, int levels) +{ + unsigned blocksize = dir->i_sb->s_blocksize; + unsigned count = dx_get_count (entries), names = 0, space = 0, i; + unsigned bcount = 0; + struct buffer_head *bh; + int err; + printk("%i indexed blocks...\n", count); + for (i = 0; i < count; i++, entries++) + { + u32 block = dx_get_block(entries), hash = i? dx_get_hash(entries): 0; + u32 range = i < count - 1? (dx_get_hash(entries + 1) - hash): ~hash; + struct stats stats; + printk("%s%3u:%03u hash %8x/%8x ",levels?"":" ", i, block, hash, range); + if (!(bh = ext3_bread (NULL,dir, block, 0,&err))) continue; + stats = levels? + dx_show_entries(hinfo, dir, ((struct dx_node *) bh->b_data)->entries, levels - 1): + dx_show_leaf(hinfo, (struct ext3_dir_entry_2 *) bh->b_data, blocksize, 0); + names += stats.names; + space += stats.space; + bcount += stats.bcount; + brelse (bh); + } + if (bcount) + printk("%snames %u, fullness %u (%u%%)\n", levels?"":" ", + names, space/bcount,(space/bcount)*100/blocksize); + return (struct stats) { names, space, bcount}; +} +#endif /* DX_DEBUG */ + +/* + * Probe for a directory leaf block to search. + * + * dx_probe can return ERR_BAD_DX_DIR, which means there was a format + * error in the directory index, and the caller should fall back to + * searching the directory normally. The callers of dx_probe **MUST** + * check for this error code, and make sure it never gets reflected + * back to userspace. + */ +static struct dx_frame * +dx_probe(struct dentry *dentry, struct inode *dir, + struct dx_hash_info *hinfo, struct dx_frame *frame_in, int *err) +{ + unsigned count, indirect; + struct dx_entry *at, *entries, *p, *q, *m; + struct dx_root *root; + struct buffer_head *bh; + struct dx_frame *frame = frame_in; + u32 hash; + + frame->bh = NULL; + if (dentry) + dir = dentry->d_parent->d_inode; + if (!(bh = ext3_bread (NULL,dir, 0, 0, err))) + goto fail; + root = (struct dx_root *) bh->b_data; + if (root->info.hash_version != DX_HASH_TEA && + root->info.hash_version != DX_HASH_HALF_MD4 && + root->info.hash_version != DX_HASH_LEGACY) { + ext3_warning(dir->i_sb, __FUNCTION__, + "Unrecognised inode hash code %d", + root->info.hash_version); + brelse(bh); + *err = ERR_BAD_DX_DIR; + goto fail; + } + hinfo->hash_version = root->info.hash_version; + hinfo->seed = dir->i_sb->u.ext3_sb.s_hash_seed; + if (dentry) + ext3fs_dirhash(dentry->d_name.name, dentry->d_name.len, hinfo); + hash = hinfo->hash; + + if (root->info.unused_flags & 1) { + ext3_warning(dir->i_sb, __FUNCTION__, + "Unimplemented inode hash flags: %#06x", + root->info.unused_flags); + brelse(bh); + *err = ERR_BAD_DX_DIR; + goto fail; + } + + if ((indirect = root->info.indirect_levels) > 1) { + ext3_warning(dir->i_sb, __FUNCTION__, + "Unimplemented inode hash depth: %#06x", + root->info.indirect_levels); + brelse(bh); + *err = ERR_BAD_DX_DIR; + goto fail; + } + + entries = (struct dx_entry *) (((char *)&root->info) + + root->info.info_length); + assert(dx_get_limit(entries) == dx_root_limit(dir, + root->info.info_length)); + dxtrace (printk("Look up %x", hash)); + while (1) + { + count = dx_get_count(entries); + assert (count && count <= dx_get_limit(entries)); + p = entries + 1; + q = entries + count - 1; + while (p <= q) + { + m = p + (q - p)/2; + dxtrace(printk(".")); + if (dx_get_hash(m) > hash) + q = m - 1; + else + p = m + 1; + } + + if (0) // linear search cross check + { + unsigned n = count - 1; + at = entries; + while (n--) + { + dxtrace(printk(",")); + if (dx_get_hash(++at) > hash) + { + at--; + break; + } + } + assert (at == p - 1); + } + + at = p - 1; + dxtrace(printk(" %x->%u\n", at == entries? 0: dx_get_hash(at), dx_get_block(at))); + frame->bh = bh; + frame->entries = entries; + frame->at = at; + if (!indirect--) return frame; + if (!(bh = ext3_bread (NULL,dir, dx_get_block(at), 0, err))) + goto fail2; + at = entries = ((struct dx_node *) bh->b_data)->entries; + assert (dx_get_limit(entries) == dx_node_limit (dir)); + frame++; + } +fail2: + while (frame >= frame_in) { + brelse(frame->bh); + frame--; + } +fail: + return NULL; +} + +static void dx_release (struct dx_frame *frames) +{ + if (frames[0].bh == NULL) + return; + + if (((struct dx_root *) frames[0].bh->b_data)->info.indirect_levels) + brelse(frames[1].bh); + brelse(frames[0].bh); +} + +/* + * This function increments the frame pointer to search the next leaf + * block, and reads in the necessary intervening nodes if the search + * should be necessary. Whether or not the search is necessary is + * controlled by the hash parameter. If the hash value is even, then + * the search is only continued if the next block starts with that + * hash value. This is used if we are searching for a specific file. + * + * If the hash value is HASH_NB_ALWAYS, then always go to the next block. + * + * This function returns 1 if the caller should continue to search, + * or 0 if it should not. If there is an error reading one of the + * index blocks, it will return -1. + * + * If start_hash is non-null, it will be filled in with the starting + * hash of the next page. + */ +static int ext3_htree_next_block(struct inode *dir, __u32 hash, + struct dx_frame *frame, + struct dx_frame *frames, int *err, + __u32 *start_hash) +{ + struct dx_frame *p; + struct buffer_head *bh; + int num_frames = 0; + __u32 bhash; + + *err = ENOENT; + p = frame; + /* + * Find the next leaf page by incrementing the frame pointer. + * If we run out of entries in the interior node, loop around and + * increment pointer in the parent node. When we break out of + * this loop, num_frames indicates the number of interior + * nodes need to be read. + */ + while (1) { + if (++(p->at) < p->entries + dx_get_count(p->entries)) + break; + if (p == frames) + return 0; + num_frames++; + p--; + } + + /* + * If the hash is 1, then continue only if the next page has a + * continuation hash of any value. This is used for readdir + * handling. Otherwise, check to see if the hash matches the + * desired contiuation hash. If it doesn't, return since + * there's no point to read in the successive index pages. + */ + bhash = dx_get_hash(p->at); + if (start_hash) + *start_hash = bhash; + if ((hash & 1) == 0) { + if ((bhash & ~1) != hash) + return 0; + } + /* + * If the hash is HASH_NB_ALWAYS, we always go to the next + * block so no check is necessary + */ + while (num_frames--) { + if (!(bh = ext3_bread(NULL, dir, dx_get_block(p->at), + 0, err))) + return -1; /* Failure */ + p++; + brelse (p->bh); + p->bh = bh; + p->at = p->entries = ((struct dx_node *) bh->b_data)->entries; + } + return 1; +} + + +/* + * p is at least 6 bytes before the end of page + */ +static inline struct ext3_dir_entry_2 *ext3_next_entry(struct ext3_dir_entry_2 *p) +{ + return (struct ext3_dir_entry_2 *)((char*)p + le16_to_cpu(p->rec_len)); +} + +/* + * This function fills a red-black tree with information from a + * directory. We start scanning the directory in hash order, starting + * at start_hash and start_minor_hash. + * + * This function returns the number of entries inserted into the tree, + * or a negative error code. + */ +int ext3_htree_fill_tree(struct file *dir_file, __u32 start_hash, + __u32 start_minor_hash, __u32 *next_hash) +{ + struct dx_hash_info hinfo; + struct buffer_head *bh; + struct ext3_dir_entry_2 *de, *top; + static struct dx_frame frames[2], *frame; + struct inode *dir; + int block, err; + int count = 0; + int ret; + __u32 hashval; + + dxtrace(printk("In htree_fill_tree, start hash: %x:%x\n", start_hash, + start_minor_hash)); + dir = dir_file->f_dentry->d_inode; + hinfo.hash = start_hash; + hinfo.minor_hash = 0; + frame = dx_probe(0, dir_file->f_dentry->d_inode, &hinfo, frames, &err); + if (!frame) + return err; + + /* Add '.' and '..' from the htree header */ + if (!start_hash && !start_minor_hash) { + de = (struct ext3_dir_entry_2 *) frames[0].bh->b_data; + if ((err = ext3_htree_store_dirent(dir_file, 0, 0, de)) != 0) + goto errout; + de = ext3_next_entry(de); + if ((err = ext3_htree_store_dirent(dir_file, 0, 0, de)) != 0) + goto errout; + count += 2; + } + + while (1) { + block = dx_get_block(frame->at); + dxtrace(printk("Reading block %d\n", block)); + if (!(bh = ext3_bread (NULL, dir, block, 0, &err))) + goto errout; + + de = (struct ext3_dir_entry_2 *) bh->b_data; + top = (struct ext3_dir_entry_2 *) ((char *) de + dir->i_sb->s_blocksize - + EXT3_DIR_REC_LEN(0)); + for (; de < top; de = ext3_next_entry(de)) { + ext3fs_dirhash(de->name, de->name_len, &hinfo); + if ((hinfo.hash < start_hash) || + ((hinfo.hash == start_hash) && + (hinfo.minor_hash < start_minor_hash))) + continue; + if ((err = ext3_htree_store_dirent(dir_file, + hinfo.hash, hinfo.minor_hash, de)) != 0) + goto errout; + count++; + } + brelse (bh); + hashval = ~1; + ret = ext3_htree_next_block(dir, HASH_NB_ALWAYS, + frame, frames, &err, &hashval); + if (next_hash) + *next_hash = hashval; + if (ret == -1) + goto errout; + /* + * Stop if: (a) there are no more entries, or + * (b) we have inserted at least one entry and the + * next hash value is not a continuation + */ + if ((ret == 0) || + (count && ((hashval & 1) == 0))) + break; + } + dx_release(frames); + dxtrace(printk("Fill tree: returned %d entries\n", count)); + return count; +errout: + dx_release(frames); + return (err); +} + + +/* + * Directory block splitting, compacting + */ + +static int dx_make_map (struct ext3_dir_entry_2 *de, int size, + struct dx_hash_info *hinfo, struct dx_map_entry *map_tail) +{ + int count = 0; + char *base = (char *) de; + struct dx_hash_info h = *hinfo; + + while ((char *) de < base + size) + { + if (de->name_len && de->inode) { + ext3fs_dirhash(de->name, de->name_len, &h); + map_tail--; + map_tail->hash = h.hash; + map_tail->offs = (u32) ((char *) de - base); + count++; + } + /* XXX: do we need to check rec_len == 0 case? -Chris */ + de = (struct ext3_dir_entry_2 *) ((char *) de + le16_to_cpu(de->rec_len)); + } + return count; +} + +static void dx_sort_map (struct dx_map_entry *map, unsigned count) +{ + struct dx_map_entry *p, *q, *top = map + count - 1; + int more; + /* Combsort until bubble sort doesn't suck */ + while (count > 2) + { + count = count*10/13; + if (count - 9 < 2) /* 9, 10 -> 11 */ + count = 11; + for (p = top, q = p - count; q >= map; p--, q--) + if (p->hash < q->hash) + swap(*p, *q); + } + /* Garden variety bubble sort */ + do { + more = 0; + q = top; + while (q-- > map) + { + if (q[1].hash >= q[0].hash) + continue; + swap(*(q+1), *q); + more = 1; + } + } while(more); +} + +static void dx_insert_block(struct dx_frame *frame, u32 hash, u32 block) +{ + struct dx_entry *entries = frame->entries; + struct dx_entry *old = frame->at, *new = old + 1; + int count = dx_get_count(entries); + + assert(count < dx_get_limit(entries)); + assert(old < entries + count); + memmove(new + 1, new, (char *)(entries + count) - (char *)(new)); + dx_set_hash(new, hash); + dx_set_block(new, block); + dx_set_count(entries, count + 1); +} +#endif + + +static void ext3_update_dx_flag(struct inode *inode) +{ + if (!EXT3_HAS_COMPAT_FEATURE(inode->i_sb, + EXT3_FEATURE_COMPAT_DIR_INDEX)) + EXT3_I(inode)->i_flags &= ~EXT3_INDEX_FL; +} + /* * NOTE! unlike strncmp, ext3_match returns 1 for success, 0 for failure. * @@ -94,6 +736,7 @@ return 0; } + /* * ext3_find_entry() * @@ -105,6 +748,8 @@ * The returned buffer_head has ->b_count elevated. The caller is expected * to brelse() it when appropriate. */ + + static struct buffer_head * ext3_find_entry (struct dentry *dentry, struct ext3_dir_entry_2 ** res_dir) { @@ -119,12 +764,32 @@ int num = 0; int nblocks, i, err; struct inode *dir = dentry->d_parent->d_inode; + int namelen; + const u8 *name; + unsigned blocksize; *res_dir = NULL; sb = dir->i_sb; - + blocksize = sb->s_blocksize; + namelen = dentry->d_name.len; + name = dentry->d_name.name; + if (namelen > EXT3_NAME_LEN) + return NULL; +#ifdef CONFIG_EXT3_INDEX + if (is_dx(dir)) { + bh = ext3_dx_find_entry(dentry, res_dir, &err); + /* + * On success, or if the error was file not found, + * return. Otherwise, fall back to doing a search the + * old fashioned way. + */ + if (bh || (err != ERR_BAD_DX_DIR)) + return bh; + dxtrace(printk("ext3_find_entry: dx failed, falling back\n")); + } +#endif nblocks = dir->i_size >> EXT3_BLOCK_SIZE_BITS(sb); - start = dir->u.ext3_i.i_dir_start_lookup; + start = EXT3_I(dir)->i_dir_start_lookup; if (start >= nblocks) start = 0; block = start; @@ -165,7 +830,7 @@ i = search_dirblock(bh, dir, dentry, block << EXT3_BLOCK_SIZE_BITS(sb), res_dir); if (i == 1) { - dir->u.ext3_i.i_dir_start_lookup = block; + EXT3_I(dir)->i_dir_start_lookup = block; ret = bh; goto cleanup_and_exit; } else { @@ -196,6 +861,66 @@ return ret; } +#ifdef CONFIG_EXT3_INDEX +static struct buffer_head * ext3_dx_find_entry(struct dentry *dentry, + struct ext3_dir_entry_2 **res_dir, int *err) +{ + struct super_block * sb; + struct dx_hash_info hinfo; + u32 hash; + struct dx_frame frames[2], *frame; + struct ext3_dir_entry_2 *de, *top; + struct buffer_head *bh; + unsigned long block; + int retval; + int namelen = dentry->d_name.len; + const u8 *name = dentry->d_name.name; + struct inode *dir = dentry->d_parent->d_inode; + + sb = dir->i_sb; + if (!(frame = dx_probe (dentry, 0, &hinfo, frames, err))) + return NULL; + hash = hinfo.hash; + do { + block = dx_get_block(frame->at); + if (!(bh = ext3_bread (NULL,dir, block, 0, err))) + goto errout; + de = (struct ext3_dir_entry_2 *) bh->b_data; + top = (struct ext3_dir_entry_2 *) ((char *) de + sb->s_blocksize - + EXT3_DIR_REC_LEN(0)); + for (; de < top; de = ext3_next_entry(de)) + if (ext3_match (namelen, name, de)) { + if (!ext3_check_dir_entry("ext3_find_entry", + dir, de, bh, + (block<b_data))) { + brelse (bh); + goto errout; + } + *res_dir = de; + dx_release (frames); + return bh; + } + brelse (bh); + /* Check to see if we should continue to search */ + retval = ext3_htree_next_block(dir, hash, frame, + frames, err, 0); + if (retval == -1) { + ext3_warning(sb, __FUNCTION__, + "error reading index page in directory #%lu", + dir->i_ino); + goto errout; + } + } while (retval == 1); + + *err = -ENOENT; +errout: + dxtrace(printk("%s not found\n", name)); + dx_release (frames); + return NULL; +} +#endif + static struct dentry *ext3_lookup(struct inode * dir, struct dentry *dentry) { struct inode * inode; @@ -212,8 +937,9 @@ brelse (bh); inode = iget(dir->i_sb, ino); - if (!inode) + if (!inode) { return ERR_PTR(-EACCES); + } } d_add(dentry, inode); return NULL; @@ -237,6 +963,300 @@ de->file_type = ext3_type_by_mode[(mode & S_IFMT)>>S_SHIFT]; } +#ifdef CONFIG_EXT3_INDEX +static struct ext3_dir_entry_2 * +dx_move_dirents(char *from, char *to, struct dx_map_entry *map, int count) +{ + unsigned rec_len = 0; + + while (count--) { + struct ext3_dir_entry_2 *de = (struct ext3_dir_entry_2 *) (from + map->offs); + rec_len = EXT3_DIR_REC_LEN(de->name_len); + memcpy (to, de, rec_len); + ((struct ext3_dir_entry_2 *) to)->rec_len = rec_len; + de->inode = 0; + map++; + to += rec_len; + } + return (struct ext3_dir_entry_2 *) (to - rec_len); +} + +static struct ext3_dir_entry_2* dx_pack_dirents(char *base, int size) +{ + struct ext3_dir_entry_2 *next, *to, *prev, *de = (struct ext3_dir_entry_2 *) base; + unsigned rec_len = 0; + + prev = to = de; + while ((char*)de < base + size) { + next = (struct ext3_dir_entry_2 *) ((char *) de + + le16_to_cpu(de->rec_len)); + if (de->inode && de->name_len) { + rec_len = EXT3_DIR_REC_LEN(de->name_len); + if (de > to) + memmove(to, de, rec_len); + to->rec_len = rec_len; + prev = to; + to = (struct ext3_dir_entry_2 *) (((char *) to) + rec_len); + } + de = next; + } + return prev; +} + +static struct ext3_dir_entry_2 *do_split(handle_t *handle, struct inode *dir, + struct buffer_head **bh,struct dx_frame *frame, + struct dx_hash_info *hinfo, int *error) +{ + unsigned blocksize = dir->i_sb->s_blocksize; + unsigned count, continued; + struct buffer_head *bh2; + u32 newblock; + u32 hash2; + struct dx_map_entry *map; + char *data1 = (*bh)->b_data, *data2; + unsigned split; + struct ext3_dir_entry_2 *de = NULL, *de2; + int err; + + bh2 = ext3_append (handle, dir, &newblock, error); + if (!(bh2)) { + brelse(*bh); + *bh = NULL; + goto errout; + } + + BUFFER_TRACE(*bh, "get_write_access"); + err = ext3_journal_get_write_access(handle, *bh); + if (err) { + journal_error: + brelse(*bh); + brelse(bh2); + *bh = NULL; + ext3_std_error(dir->i_sb, err); + goto errout; + } + BUFFER_TRACE(frame->bh, "get_write_access"); + err = ext3_journal_get_write_access(handle, frame->bh); + if (err) + goto journal_error; + + data2 = bh2->b_data; + + /* create map in the end of data2 block */ + map = (struct dx_map_entry *) (data2 + blocksize); + count = dx_make_map ((struct ext3_dir_entry_2 *) data1, + blocksize, hinfo, map); + map -= count; + split = count/2; // need to adjust to actual middle + dx_sort_map (map, count); + hash2 = map[split].hash; + continued = hash2 == map[split - 1].hash; + dxtrace(printk("Split block %i at %x, %i/%i\n", + dx_get_block(frame->at), hash2, split, count-split)); + + /* Fancy dance to stay within two buffers */ + de2 = dx_move_dirents(data1, data2, map + split, count - split); + de = dx_pack_dirents(data1,blocksize); + de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de); + de2->rec_len = cpu_to_le16(data2 + blocksize - (char *) de2); + dxtrace(dx_show_leaf (hinfo, (struct ext3_dir_entry_2 *) data1, blocksize, 1)); + dxtrace(dx_show_leaf (hinfo, (struct ext3_dir_entry_2 *) data2, blocksize, 1)); + + /* Which block gets the new entry? */ + if (hinfo->hash >= hash2) + { + swap(*bh, bh2); + de = de2; + } + dx_insert_block (frame, hash2 + continued, newblock); + err = ext3_journal_dirty_metadata (handle, bh2); + if (err) + goto journal_error; + err = ext3_journal_dirty_metadata (handle, frame->bh); + if (err) + goto journal_error; + brelse (bh2); + dxtrace(dx_show_index ("frame", frame->entries)); +errout: + return de; +} +#endif + + +/* + * Add a new entry into a directory (leaf) block. If de is non-NULL, + * it points to a directory entry which is guaranteed to be large + * enough for new directory entry. If de is NULL, then + * add_dirent_to_buf will attempt search the directory block for + * space. It will return -ENOSPC if no space is available, and -EIO + * and -EEXIST if directory entry already exists. + * + * NOTE! bh is NOT released in the case where ENOSPC is returned. In + * all other cases bh is released. + */ +static int add_dirent_to_buf(handle_t *handle, struct dentry *dentry, + struct inode *inode, struct ext3_dir_entry_2 *de, + struct buffer_head * bh) +{ + struct inode *dir = dentry->d_parent->d_inode; + const char *name = dentry->d_name.name; + int namelen = dentry->d_name.len; + unsigned long offset = 0; + unsigned short reclen; + int nlen, rlen, err; + char *top; + + reclen = EXT3_DIR_REC_LEN(namelen); + if (!de) { + de = (struct ext3_dir_entry_2 *)bh->b_data; + top = bh->b_data + dir->i_sb->s_blocksize - reclen; + while ((char *) de <= top) { + if (!ext3_check_dir_entry("ext3_add_entry", dir, de, + bh, offset)) { + brelse (bh); + return -EIO; + } + if (ext3_match (namelen, name, de)) { + brelse (bh); + return -EEXIST; + } + nlen = EXT3_DIR_REC_LEN(de->name_len); + rlen = le16_to_cpu(de->rec_len); + if ((de->inode? rlen - nlen: rlen) >= reclen) + break; + de = (struct ext3_dir_entry_2 *)((char *)de + rlen); + offset += rlen; + } + if ((char *) de > top) + return -ENOSPC; + } + BUFFER_TRACE(bh, "get_write_access"); + err = ext3_journal_get_write_access(handle, bh); + if (err) { + ext3_std_error(dir->i_sb, err); + brelse(bh); + return err; + } + + /* By now the buffer is marked for journaling */ + nlen = EXT3_DIR_REC_LEN(de->name_len); + rlen = le16_to_cpu(de->rec_len); + if (de->inode) { + struct ext3_dir_entry_2 *de1 = (struct ext3_dir_entry_2 *)((char *)de + nlen); + de1->rec_len = cpu_to_le16(rlen - nlen); + de->rec_len = cpu_to_le16(nlen); + de = de1; + } + de->file_type = EXT3_FT_UNKNOWN; + if (inode) { + de->inode = cpu_to_le32(inode->i_ino); + ext3_set_de_type(dir->i_sb, de, inode->i_mode); + } else + de->inode = 0; + de->name_len = namelen; + memcpy (de->name, name, namelen); + /* + * XXX shouldn't update any times until successful + * completion of syscall, but too many callers depend + * on this. + * + * XXX similarly, too many callers depend on + * ext3_new_inode() setting the times, but error + * recovery deletes the inode, so the worst that can + * happen is that the times are slightly out of date + * and/or different from the directory change time. + */ + dir->i_mtime = dir->i_ctime = CURRENT_TIME; + ext3_update_dx_flag(dir); + dir->i_version = ++event; + ext3_mark_inode_dirty(handle, dir); + BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata"); + err = ext3_journal_dirty_metadata(handle, bh); + if (err) + ext3_std_error(dir->i_sb, err); + brelse(bh); + return 0; +} + +#ifdef CONFIG_EXT3_INDEX +/* + * This converts a one block unindexed directory to a 3 block indexed + * directory, and adds the dentry to the indexed directory. + */ +static int make_indexed_dir(handle_t *handle, struct dentry *dentry, + struct inode *inode, struct buffer_head *bh) +{ + struct inode *dir = dentry->d_parent->d_inode; + const char *name = dentry->d_name.name; + int namelen = dentry->d_name.len; + struct buffer_head *bh2; + struct dx_root *root; + struct dx_frame frames[2], *frame; + struct dx_entry *entries; + struct ext3_dir_entry_2 *de, *de2; + char *data1, *top; + unsigned len; + int retval; + unsigned blocksize; + struct dx_hash_info hinfo; + u32 block; + + blocksize = dir->i_sb->s_blocksize; + dxtrace(printk("Creating index\n")); + retval = ext3_journal_get_write_access(handle, bh); + if (retval) { + ext3_std_error(dir->i_sb, retval); + brelse(bh); + return retval; + } + root = (struct dx_root *) bh->b_data; + + EXT3_I(dir)->i_flags |= EXT3_INDEX_FL; + bh2 = ext3_append (handle, dir, &block, &retval); + if (!(bh2)) { + brelse(bh); + return retval; + } + data1 = bh2->b_data; + + /* The 0th block becomes the root, move the dirents out */ + de = (struct ext3_dir_entry_2 *) &root->info; + len = ((char *) root) + blocksize - (char *) de; + memcpy (data1, de, len); + de = (struct ext3_dir_entry_2 *) data1; + top = data1 + len; + while (((char *) de2=(char*)de+le16_to_cpu(de->rec_len)) < top) + de = de2; + de->rec_len = cpu_to_le16(data1 + blocksize - (char *) de); + /* Initialize the root; the dot dirents already exist */ + de = (struct ext3_dir_entry_2 *) (&root->dotdot); + de->rec_len = cpu_to_le16(blocksize - EXT3_DIR_REC_LEN(2)); + memset (&root->info, 0, sizeof(root->info)); + root->info.info_length = sizeof(root->info); + root->info.hash_version = dir->i_sb->u.ext3_sb.s_def_hash_version; + entries = root->entries; + dx_set_block (entries, 1); + dx_set_count (entries, 1); + dx_set_limit (entries, dx_root_limit(dir, sizeof(root->info))); + + /* Initialize as for dx_probe */ + hinfo.hash_version = root->info.hash_version; + hinfo.seed = dir->i_sb->u.ext3_sb.s_hash_seed; + ext3fs_dirhash(name, namelen, &hinfo); + frame = frames; + frame->entries = entries; + frame->at = entries; + frame->bh = bh; + bh = bh2; + de = do_split(handle,dir, &bh, frame, &hinfo, &retval); + dx_release (frames); + if (!(de)) + return retval; + + return add_dirent_to_buf(handle, dentry, inode, de, bh); +} +#endif + /* * ext3_add_entry() * @@ -247,127 +1267,198 @@ * may not sleep between calling this and putting something into * the entry, as someone else might have used it while you slept. */ - -/* - * AKPM: the journalling code here looks wrong on the error paths - */ static int ext3_add_entry (handle_t *handle, struct dentry *dentry, struct inode *inode) { struct inode *dir = dentry->d_parent->d_inode; - const char *name = dentry->d_name.name; - int namelen = dentry->d_name.len; unsigned long offset; - unsigned short rec_len; struct buffer_head * bh; - struct ext3_dir_entry_2 * de, * de1; + struct ext3_dir_entry_2 *de; struct super_block * sb; int retval; +#ifdef CONFIG_EXT3_INDEX + int dx_fallback=0; +#endif + unsigned blocksize; + unsigned nlen, rlen; + u32 block, blocks; sb = dir->i_sb; - - if (!namelen) + blocksize = sb->s_blocksize; + if (!dentry->d_name.len) return -EINVAL; - bh = ext3_bread (handle, dir, 0, 0, &retval); +#ifdef CONFIG_EXT3_INDEX + if (is_dx(dir)) { + retval = ext3_dx_add_entry(handle, dentry, inode); + if (!retval || (retval != ERR_BAD_DX_DIR)) + return retval; + EXT3_I(dir)->i_flags &= ~EXT3_INDEX_FL; + dx_fallback++; + ext3_mark_inode_dirty(handle, dir); + } +#endif + blocks = dir->i_size >> sb->s_blocksize_bits; + for (block = 0, offset = 0; block < blocks; block++) { + bh = ext3_bread(handle, dir, block, 0, &retval); + if(!bh) + return retval; + retval = add_dirent_to_buf(handle, dentry, inode, 0, bh); + if (retval != -ENOSPC) + return retval; + +#ifdef CONFIG_EXT3_INDEX + if (blocks == 1 && !dx_fallback && + EXT3_HAS_COMPAT_FEATURE(sb, EXT3_FEATURE_COMPAT_DIR_INDEX)) + return make_indexed_dir(handle, dentry, inode, bh); +#endif + brelse(bh); + } + bh = ext3_append(handle, dir, &block, &retval); if (!bh) return retval; - rec_len = EXT3_DIR_REC_LEN(namelen); - offset = 0; de = (struct ext3_dir_entry_2 *) bh->b_data; - while (1) { - if ((char *)de >= sb->s_blocksize + bh->b_data) { - brelse (bh); - bh = NULL; - bh = ext3_bread (handle, dir, - offset >> EXT3_BLOCK_SIZE_BITS(sb), 1, &retval); - if (!bh) - return retval; - if (dir->i_size <= offset) { - if (dir->i_size == 0) { - brelse(bh); - return -ENOENT; - } + de->inode = 0; + de->rec_len = cpu_to_le16(rlen = blocksize); + nlen = 0; + return add_dirent_to_buf(handle, dentry, inode, de, bh); +} - ext3_debug ("creating next block\n"); +#ifdef CONFIG_EXT3_INDEX +/* + * Returns 0 for success, or a negative error value + */ +static int ext3_dx_add_entry(handle_t *handle, struct dentry *dentry, + struct inode *inode) +{ + struct dx_frame frames[2], *frame; + struct dx_entry *entries, *at; + struct dx_hash_info hinfo; + struct buffer_head * bh; + struct inode *dir = dentry->d_parent->d_inode; + struct super_block * sb = dir->i_sb; + struct ext3_dir_entry_2 *de; + int err; - BUFFER_TRACE(bh, "get_write_access"); - ext3_journal_get_write_access(handle, bh); - de = (struct ext3_dir_entry_2 *) bh->b_data; - de->inode = 0; - de->rec_len = le16_to_cpu(sb->s_blocksize); - dir->u.ext3_i.i_disksize = - dir->i_size = offset + sb->s_blocksize; - dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; - ext3_mark_inode_dirty(handle, dir); - } else { + frame = dx_probe(dentry, 0, &hinfo, frames, &err); + if (!frame) + return err; + entries = frame->entries; + at = frame->at; - ext3_debug ("skipping to next block\n"); + if (!(bh = ext3_bread(handle,dir, dx_get_block(frame->at), 0, &err))) + goto cleanup; - de = (struct ext3_dir_entry_2 *) bh->b_data; - } - } - if (!ext3_check_dir_entry ("ext3_add_entry", dir, de, bh, - offset)) { - brelse (bh); - return -ENOENT; - } - if (ext3_match (namelen, name, de)) { - brelse (bh); - return -EEXIST; + BUFFER_TRACE(bh, "get_write_access"); + err = ext3_journal_get_write_access(handle, bh); + if (err) + goto journal_error; + + err = add_dirent_to_buf(handle, dentry, inode, 0, bh); + if (err != -ENOSPC) { + bh = 0; + goto cleanup; + } + + /* Block full, should compress but for now just split */ + dxtrace(printk("using %u of %u node entries\n", + dx_get_count(entries), dx_get_limit(entries))); + /* Need to split index? */ + if (dx_get_count(entries) == dx_get_limit(entries)) { + u32 newblock; + unsigned icount = dx_get_count(entries); + int levels = frame - frames; + struct dx_entry *entries2; + struct dx_node *node2; + struct buffer_head *bh2; + + if (levels && (dx_get_count(frames->entries) == + dx_get_limit(frames->entries))) { + ext3_warning(sb, __FUNCTION__, + "Directory index full!\n"); + err = -ENOSPC; + goto cleanup; } - if ((le32_to_cpu(de->inode) == 0 && - le16_to_cpu(de->rec_len) >= rec_len) || - (le16_to_cpu(de->rec_len) >= - EXT3_DIR_REC_LEN(de->name_len) + rec_len)) { - BUFFER_TRACE(bh, "get_write_access"); - ext3_journal_get_write_access(handle, bh); - /* By now the buffer is marked for journaling */ - offset += le16_to_cpu(de->rec_len); - if (le32_to_cpu(de->inode)) { - de1 = (struct ext3_dir_entry_2 *) ((char *) de + - EXT3_DIR_REC_LEN(de->name_len)); - de1->rec_len = - cpu_to_le16(le16_to_cpu(de->rec_len) - - EXT3_DIR_REC_LEN(de->name_len)); - de->rec_len = cpu_to_le16( - EXT3_DIR_REC_LEN(de->name_len)); - de = de1; + bh2 = ext3_append (handle, dir, &newblock, &err); + if (!(bh2)) + goto cleanup; + node2 = (struct dx_node *)(bh2->b_data); + entries2 = node2->entries; + node2->fake.rec_len = cpu_to_le16(sb->s_blocksize); + node2->fake.inode = 0; + BUFFER_TRACE(frame->bh, "get_write_access"); + err = ext3_journal_get_write_access(handle, frame->bh); + if (err) + goto journal_error; + if (levels) { + unsigned icount1 = icount/2, icount2 = icount - icount1; + unsigned hash2 = dx_get_hash(entries + icount1); + dxtrace(printk("Split index %i/%i\n", icount1, icount2)); + + BUFFER_TRACE(frame->bh, "get_write_access"); /* index root */ + err = ext3_journal_get_write_access(handle, + frames[0].bh); + if (err) + goto journal_error; + + memcpy ((char *) entries2, (char *) (entries + icount1), + icount2 * sizeof(struct dx_entry)); + dx_set_count (entries, icount1); + dx_set_count (entries2, icount2); + dx_set_limit (entries2, dx_node_limit(dir)); + + /* Which index block gets the new entry? */ + if (at - entries >= icount1) { + frame->at = at = at - entries - icount1 + entries2; + frame->entries = entries = entries2; + swap(frame->bh, bh2); } - de->file_type = EXT3_FT_UNKNOWN; - if (inode) { - de->inode = cpu_to_le32(inode->i_ino); - ext3_set_de_type(dir->i_sb, de, inode->i_mode); - } else - de->inode = 0; - de->name_len = namelen; - memcpy (de->name, name, namelen); - /* - * XXX shouldn't update any times until successful - * completion of syscall, but too many callers depend - * on this. - * - * XXX similarly, too many callers depend on - * ext3_new_inode() setting the times, but error - * recovery deletes the inode, so the worst that can - * happen is that the times are slightly out of date - * and/or different from the directory change time. - */ - dir->i_mtime = dir->i_ctime = CURRENT_TIME; - dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; - dir->i_version = ++event; - ext3_mark_inode_dirty(handle, dir); - BUFFER_TRACE(bh, "call ext3_journal_dirty_metadata"); - ext3_journal_dirty_metadata(handle, bh); - brelse(bh); - return 0; + dx_insert_block (frames + 0, hash2, newblock); + dxtrace(dx_show_index ("node", frames[1].entries)); + dxtrace(dx_show_index ("node", + ((struct dx_node *) bh2->b_data)->entries)); + err = ext3_journal_dirty_metadata(handle, bh2); + if (err) + goto journal_error; + brelse (bh2); + } else { + dxtrace(printk("Creating second level index...\n")); + memcpy((char *) entries2, (char *) entries, + icount * sizeof(struct dx_entry)); + dx_set_limit(entries2, dx_node_limit(dir)); + + /* Set up root */ + dx_set_count(entries, 1); + dx_set_block(entries + 0, newblock); + ((struct dx_root *) frames[0].bh->b_data)->info.indirect_levels = 1; + + /* Add new access path frame */ + frame = frames + 1; + frame->at = at = at - entries + entries2; + frame->entries = entries = entries2; + frame->bh = bh2; + err = ext3_journal_get_write_access(handle, + frame->bh); + if (err) + goto journal_error; } - offset += le16_to_cpu(de->rec_len); - de = (struct ext3_dir_entry_2 *) - ((char *) de + le16_to_cpu(de->rec_len)); + ext3_journal_dirty_metadata(handle, frames[0].bh); } - brelse (bh); - return -ENOSPC; + de = do_split(handle, dir, &bh, frame, &hinfo, &err); + if (!de) + goto cleanup; + err = add_dirent_to_buf(handle, dentry, inode, de, bh); + bh = 0; + goto cleanup; + +journal_error: + ext3_std_error(dir->i_sb, err); +cleanup: + if (bh) + brelse(bh); + dx_release(frames); + return err; } +#endif /* * ext3_delete_entry deletes a directory entry by merging it with the @@ -451,9 +1542,11 @@ struct inode * inode; int err; - handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 3); - if (IS_ERR(handle)) + handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + + EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3); + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(dir)) handle->h_sync = 1; @@ -478,9 +1571,11 @@ struct inode *inode; int err; - handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 3); - if (IS_ERR(handle)) + handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + + EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3); + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(dir)) handle->h_sync = 1; @@ -507,9 +1602,11 @@ if (dir->i_nlink >= EXT3_LINK_MAX) return -EMLINK; - handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 3); - if (IS_ERR(handle)) + handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + + EXT3_INDEX_EXTRA_TRANS_BLOCKS + 3); + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(dir)) handle->h_sync = 1; @@ -521,7 +1618,7 @@ inode->i_op = &ext3_dir_inode_operations; inode->i_fop = &ext3_dir_operations; - inode->i_size = inode->u.ext3_i.i_disksize = inode->i_sb->s_blocksize; + inode->i_size = EXT3_I(inode)->i_disksize = inode->i_sb->s_blocksize; inode->i_blocks = 0; dir_block = ext3_bread (handle, inode, 0, 1, &err); if (!dir_block) { @@ -554,21 +1651,19 @@ inode->i_mode |= S_ISGID; ext3_mark_inode_dirty(handle, inode); err = ext3_add_entry (handle, dentry, inode); - if (err) - goto out_no_entry; + if (err) { + inode->i_nlink = 0; + ext3_mark_inode_dirty(handle, inode); + iput (inode); + goto out_stop; + } dir->i_nlink++; - dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; + ext3_update_dx_flag(dir); ext3_mark_inode_dirty(handle, dir); d_instantiate(dentry, inode); out_stop: ext3_journal_stop(handle, dir); return err; - -out_no_entry: - inode->i_nlink = 0; - ext3_mark_inode_dirty(handle, inode); - iput (inode); - goto out_stop; } /* @@ -655,7 +1750,7 @@ int err = 0, rc; lock_super(sb); - if (!list_empty(&inode->u.ext3_i.i_orphan)) + if (!list_empty(&EXT3_I(inode)->i_orphan)) goto out_unlock; /* Orphan handling is only valid for files with data blocks @@ -696,7 +1791,7 @@ * This is safe: on error we're going to ignore the orphan list * anyway on the next recovery. */ if (!err) - list_add(&inode->u.ext3_i.i_orphan, &EXT3_SB(sb)->s_orphan); + list_add(&EXT3_I(inode)->i_orphan, &EXT3_SB(sb)->s_orphan); jbd_debug(4, "superblock will point to %ld\n", inode->i_ino); jbd_debug(4, "orphan inode %ld will point to %d\n", @@ -714,25 +1809,26 @@ int ext3_orphan_del(handle_t *handle, struct inode *inode) { struct list_head *prev; + struct ext3_inode_info *ei = EXT3_I(inode); struct ext3_sb_info *sbi; ino_t ino_next; struct ext3_iloc iloc; int err = 0; lock_super(inode->i_sb); - if (list_empty(&inode->u.ext3_i.i_orphan)) { + if (list_empty(&ei->i_orphan)) { unlock_super(inode->i_sb); return 0; } ino_next = NEXT_ORPHAN(inode); - prev = inode->u.ext3_i.i_orphan.prev; + prev = ei->i_orphan.prev; sbi = EXT3_SB(inode->i_sb); jbd_debug(4, "remove inode %ld from orphan list\n", inode->i_ino); - list_del(&inode->u.ext3_i.i_orphan); - INIT_LIST_HEAD(&inode->u.ext3_i.i_orphan); + list_del(&ei->i_orphan); + INIT_LIST_HEAD(&ei->i_orphan); /* If we're on an error path, we may not have a valid * transaction handle with which to update the orphan list on @@ -793,8 +1889,9 @@ handle_t *handle; handle = ext3_journal_start(dir, EXT3_DELETE_TRANS_BLOCKS); - if (IS_ERR(handle)) + if (IS_ERR(handle)) { return PTR_ERR(handle); + } retval = -ENOENT; bh = ext3_find_entry (dentry, &de); @@ -832,7 +1929,7 @@ dir->i_nlink--; inode->i_ctime = dir->i_ctime = dir->i_mtime = CURRENT_TIME; ext3_mark_inode_dirty(handle, inode); - dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; + ext3_update_dx_flag(dir); ext3_mark_inode_dirty(handle, dir); end_rmdir: @@ -850,8 +1947,9 @@ handle_t *handle; handle = ext3_journal_start(dir, EXT3_DELETE_TRANS_BLOCKS); - if (IS_ERR(handle)) + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(dir)) handle->h_sync = 1; @@ -878,7 +1976,7 @@ if (retval) goto end_unlink; dir->i_ctime = dir->i_mtime = CURRENT_TIME; - dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; + ext3_update_dx_flag(dir); ext3_mark_inode_dirty(handle, dir); inode->i_nlink--; if (!inode->i_nlink) @@ -904,9 +2002,11 @@ if (l > dir->i_sb->s_blocksize) return -ENAMETOOLONG; - handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + 5); - if (IS_ERR(handle)) + handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + + EXT3_INDEX_EXTRA_TRANS_BLOCKS + 5); + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(dir)) handle->h_sync = 1; @@ -916,7 +2016,7 @@ if (IS_ERR(inode)) goto out_stop; - if (l > sizeof (inode->u.ext3_i.i_data)) { + if (l > sizeof (EXT3_I(inode)->i_data)) { inode->i_op = &page_symlink_inode_operations; inode->i_mapping->a_ops = &ext3_aops; /* @@ -925,25 +2025,23 @@ * i_size in generic_commit_write(). */ err = block_symlink(inode, symname, l); - if (err) - goto out_no_entry; + if (err) { + ext3_dec_count(handle, inode); + ext3_mark_inode_dirty(handle, inode); + iput (inode); + goto out_stop; + } } else { inode->i_op = &ext3_fast_symlink_inode_operations; - memcpy((char*)&inode->u.ext3_i.i_data,symname,l); + memcpy((char*)&EXT3_I(inode)->i_data,symname,l); inode->i_size = l-1; } - inode->u.ext3_i.i_disksize = inode->i_size; + EXT3_I(inode)->i_disksize = inode->i_size; err = ext3_add_nondir(handle, dentry, inode); ext3_mark_inode_dirty(handle, inode); out_stop: ext3_journal_stop(handle, dir); return err; - -out_no_entry: - ext3_dec_count(handle, inode); - ext3_mark_inode_dirty(handle, inode); - iput (inode); - goto out_stop; } static int ext3_link (struct dentry * old_dentry, @@ -956,12 +2054,15 @@ if (S_ISDIR(inode->i_mode)) return -EPERM; - if (inode->i_nlink >= EXT3_LINK_MAX) + if (inode->i_nlink >= EXT3_LINK_MAX) { return -EMLINK; + } - handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS); - if (IS_ERR(handle)) + handle = ext3_journal_start(dir, EXT3_DATA_TRANS_BLOCKS + + EXT3_INDEX_EXTRA_TRANS_BLOCKS); + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(dir)) handle->h_sync = 1; @@ -995,9 +2096,11 @@ old_bh = new_bh = dir_bh = NULL; - handle = ext3_journal_start(old_dir, 2 * EXT3_DATA_TRANS_BLOCKS + 2); - if (IS_ERR(handle)) + handle = ext3_journal_start(old_dir, 2 * EXT3_DATA_TRANS_BLOCKS + + EXT3_INDEX_EXTRA_TRANS_BLOCKS + 2); + if (IS_ERR(handle)) { return PTR_ERR(handle); + } if (IS_SYNC(old_dir) || IS_SYNC(new_dir)) handle->h_sync = 1; @@ -1070,14 +2173,33 @@ /* * ok, that's it */ - ext3_delete_entry(handle, old_dir, old_de, old_bh); + retval = ext3_delete_entry(handle, old_dir, old_de, old_bh); + if (retval == -ENOENT) { + /* + * old_de could have moved out from under us. + */ + struct buffer_head *old_bh2; + struct ext3_dir_entry_2 *old_de2; + + old_bh2 = ext3_find_entry(old_dentry, &old_de2); + if (old_bh2) { + retval = ext3_delete_entry(handle, old_dir, + old_de2, old_bh2); + brelse(old_bh2); + } + } + if (retval) { + ext3_warning(old_dir->i_sb, "ext3_rename", + "Deleting old file (%lu), %d, error=%d", + old_dir->i_ino, old_dir->i_nlink, retval); + } if (new_inode) { new_inode->i_nlink--; new_inode->i_ctime = CURRENT_TIME; } old_dir->i_ctime = old_dir->i_mtime = CURRENT_TIME; - old_dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; + ext3_update_dx_flag(old_dir); if (dir_bh) { BUFFER_TRACE(dir_bh, "get_write_access"); ext3_journal_get_write_access(handle, dir_bh); @@ -1089,7 +2211,7 @@ new_inode->i_nlink--; } else { new_dir->i_nlink++; - new_dir->u.ext3_i.i_flags &= ~EXT3_INDEX_FL; + ext3_update_dx_flag(new_dir); ext3_mark_inode_dirty(handle, new_dir); } } diff -Nru a/fs/ext3/super.c b/fs/ext3/super.c --- a/fs/ext3/super.c Thu Nov 7 11:58:05 2002 +++ b/fs/ext3/super.c Thu Nov 7 11:58:05 2002 @@ -106,32 +106,6 @@ static char error_buf[1024]; -/* Determine the appropriate response to ext3_error on a given filesystem */ - -static int ext3_error_behaviour(struct super_block *sb) -{ - /* First check for mount-time options */ - if (test_opt (sb, ERRORS_PANIC)) - return EXT3_ERRORS_PANIC; - if (test_opt (sb, ERRORS_RO)) - return EXT3_ERRORS_RO; - if (test_opt (sb, ERRORS_CONT)) - return EXT3_ERRORS_CONTINUE; - - /* If no overrides were specified on the mount, then fall back - * to the default behaviour set in the filesystem's superblock - * on disk. */ - switch (le16_to_cpu(sb->u.ext3_sb.s_es->s_errors)) { - case EXT3_ERRORS_PANIC: - return EXT3_ERRORS_PANIC; - case EXT3_ERRORS_RO: - return EXT3_ERRORS_RO; - default: - break; - } - return EXT3_ERRORS_CONTINUE; -} - /* Deal with the reporting of failure conditions on a filesystem such as * inconsistencies detected or read IO failures. * @@ -157,18 +131,15 @@ if (sb->s_flags & MS_RDONLY) return; - if (ext3_error_behaviour(sb) != EXT3_ERRORS_CONTINUE) { - EXT3_SB(sb)->s_mount_opt |= EXT3_MOUNT_ABORT; - journal_abort(EXT3_SB(sb)->s_journal, -EIO); - } - - if (ext3_error_behaviour(sb) == EXT3_ERRORS_PANIC) + if (test_opt (sb, ERRORS_PANIC)) panic ("EXT3-fs (device %s): panic forced after error\n", bdevname(sb->s_dev)); - - if (ext3_error_behaviour(sb) == EXT3_ERRORS_RO) { + if (test_opt (sb, ERRORS_RO)) { printk (KERN_CRIT "Remounting filesystem read-only\n"); sb->s_flags |= MS_RDONLY; + } else { + EXT3_SB(sb)->s_mount_opt |= EXT3_MOUNT_ABORT; + journal_abort(EXT3_SB(sb)->s_journal, -EIO); } ext3_commit_super(sb, es, 1); @@ -258,7 +229,7 @@ vsprintf (error_buf, fmt, args); va_end (args); - if (ext3_error_behaviour(sb) == EXT3_ERRORS_PANIC) + if (test_opt (sb, ERRORS_PANIC)) panic ("EXT3-fs panic (device %s): %s: %s\n", bdevname(sb->s_dev), function, error_buf); @@ -416,6 +387,7 @@ for (i = 0; i < sbi->s_gdb_count; i++) brelse(sbi->s_group_desc[i]); kfree(sbi->s_group_desc); + kfree(sbi->s_debts); for (i = 0; i < EXT3_MAX_GROUP_LOADED; i++) brelse(sbi->s_inode_bitmap[i]); for (i = 0; i < EXT3_MAX_GROUP_LOADED; i++) @@ -490,17 +462,32 @@ return 0; } +static unsigned long get_sb_block(void **data) +{ + unsigned long sb_block; + char *options = (char *) *data; + + if (!options || strncmp(options, "sb=", 3) != 0) + return 1; /* Default location */ + options += 3; + sb_block = simple_strtoul(options, &options, 0); + if (*options && *options != ',') { + printk("EXT3-fs: Invalid sb specification: %s\n", + (char *) *data); + return 1; + } + if (*options == ',') + options++; + *data = (void *) options; + return sb_block; +} + /* * This function has been shamelessly adapted from the msdos fs */ -static int parse_options (char * options, unsigned long * sb_block, - struct ext3_sb_info *sbi, - unsigned long * inum, - int is_remount) -{ - unsigned long *mount_options = &sbi->s_mount_opt; - uid_t *resuid = &sbi->s_resuid; - gid_t *resgid = &sbi->s_resgid; +static int parse_options (char * options, struct ext3_sb_info *sbi, + unsigned long * inum, int is_remount) +{ char * this_char; char * value; @@ -512,42 +499,42 @@ if ((value = strchr (this_char, '=')) != NULL) *value++ = 0; if (!strcmp (this_char, "bsddf")) - clear_opt (*mount_options, MINIX_DF); + clear_opt (sbi->s_mount_opt, MINIX_DF); else if (!strcmp (this_char, "nouid32")) { - set_opt (*mount_options, NO_UID32); + set_opt (sbi->s_mount_opt, NO_UID32); } else if (!strcmp (this_char, "abort")) - set_opt (*mount_options, ABORT); + set_opt (sbi->s_mount_opt, ABORT); else if (!strcmp (this_char, "check")) { if (!value || !*value || !strcmp (value, "none")) - clear_opt (*mount_options, CHECK); + clear_opt (sbi->s_mount_opt, CHECK); else #ifdef CONFIG_EXT3_CHECK - set_opt (*mount_options, CHECK); + set_opt (sbi->s_mount_opt, CHECK); #else printk(KERN_ERR "EXT3 Check option not supported\n"); #endif } else if (!strcmp (this_char, "debug")) - set_opt (*mount_options, DEBUG); + set_opt (sbi->s_mount_opt, DEBUG); else if (!strcmp (this_char, "errors")) { if (want_value(value, "errors")) return 0; if (!strcmp (value, "continue")) { - clear_opt (*mount_options, ERRORS_RO); - clear_opt (*mount_options, ERRORS_PANIC); - set_opt (*mount_options, ERRORS_CONT); + clear_opt (sbi->s_mount_opt, ERRORS_RO); + clear_opt (sbi->s_mount_opt, ERRORS_PANIC); + set_opt (sbi->s_mount_opt, ERRORS_CONT); } else if (!strcmp (value, "remount-ro")) { - clear_opt (*mount_options, ERRORS_CONT); - clear_opt (*mount_options, ERRORS_PANIC); - set_opt (*mount_options, ERRORS_RO); + clear_opt (sbi->s_mount_opt, ERRORS_CONT); + clear_opt (sbi->s_mount_opt, ERRORS_PANIC); + set_opt (sbi->s_mount_opt, ERRORS_RO); } else if (!strcmp (value, "panic")) { - clear_opt (*mount_options, ERRORS_CONT); - clear_opt (*mount_options, ERRORS_RO); - set_opt (*mount_options, ERRORS_PANIC); + clear_opt (sbi->s_mount_opt, ERRORS_CONT); + clear_opt (sbi->s_mount_opt, ERRORS_RO); + set_opt (sbi->s_mount_opt, ERRORS_PANIC); } else { printk (KERN_ERR @@ -558,29 +545,25 @@ } else if (!strcmp (this_char, "grpid") || !strcmp (this_char, "bsdgroups")) - set_opt (*mount_options, GRPID); + set_opt (sbi->s_mount_opt, GRPID); else if (!strcmp (this_char, "minixdf")) - set_opt (*mount_options, MINIX_DF); + set_opt (sbi->s_mount_opt, MINIX_DF); else if (!strcmp (this_char, "nocheck")) - clear_opt (*mount_options, CHECK); + clear_opt (sbi->s_mount_opt, CHECK); else if (!strcmp (this_char, "nogrpid") || !strcmp (this_char, "sysvgroups")) - clear_opt (*mount_options, GRPID); + clear_opt (sbi->s_mount_opt, GRPID); else if (!strcmp (this_char, "resgid")) { unsigned long v; if (want_numeric(value, "resgid", &v)) return 0; - *resgid = v; + sbi->s_resgid = v; } else if (!strcmp (this_char, "resuid")) { unsigned long v; if (want_numeric(value, "resuid", &v)) return 0; - *resuid = v; - } - else if (!strcmp (this_char, "sb")) { - if (want_numeric(value, "sb", sb_block)) - return 0; + sbi->s_resuid = v; } #ifdef CONFIG_JBD_DEBUG else if (!strcmp (this_char, "ro-after")) { @@ -611,12 +594,12 @@ if (want_value(value, "journal")) return 0; if (!strcmp (value, "update")) - set_opt (*mount_options, UPDATE_JOURNAL); + set_opt (sbi->s_mount_opt, UPDATE_JOURNAL); else if (want_numeric(value, "journal", inum)) return 0; } else if (!strcmp (this_char, "noload")) - set_opt (*mount_options, NOLOAD); + set_opt (sbi->s_mount_opt, NOLOAD); else if (!strcmp (this_char, "data")) { int data_opt = 0; @@ -635,7 +618,7 @@ return 0; } if (is_remount) { - if ((*mount_options & EXT3_MOUNT_DATA_FLAGS) != + if ((sbi->s_mount_opt & EXT3_MOUNT_DATA_FLAGS) != data_opt) { printk(KERN_ERR "EXT3-fs: cannot change data " @@ -643,8 +626,8 @@ return 0; } } else { - *mount_options &= ~EXT3_MOUNT_DATA_FLAGS; - *mount_options |= data_opt; + sbi->s_mount_opt &= ~EXT3_MOUNT_DATA_FLAGS; + sbi->s_mount_opt |= data_opt; } } else if (!strcmp (this_char, "commit")) { unsigned long v; @@ -707,6 +690,7 @@ es->s_mtime = cpu_to_le32(CURRENT_TIME); ext3_update_dynamic_rev(sb); EXT3_SET_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_RECOVER); + ext3_commit_super (sb, es, 1); if (test_opt (sb, DEBUG)) printk (KERN_INFO @@ -717,6 +701,7 @@ EXT3_BLOCKS_PER_GROUP(sb), EXT3_INODES_PER_GROUP(sb), sbi->s_mount_opt); + printk(KERN_INFO "EXT3 FS " EXT3FS_VERSION ", " EXT3FS_DATE " on %s, ", bdevname(sb->s_dev)); if (EXT3_SB(sb)->s_journal->j_inode == NULL) { @@ -890,17 +875,39 @@ return res; } +static unsigned long descriptor_loc(struct super_block *sb, + unsigned long logic_sb_block, + int nr) +{ + struct ext3_sb_info *sbi = EXT3_SB(sb); + unsigned long bg, first_data_block, first_meta_bg; + int has_super = 0; + + first_data_block = le32_to_cpu(sbi->s_es->s_first_data_block); + first_meta_bg = le32_to_cpu(sbi->s_es->s_first_meta_bg); + + if (!EXT3_HAS_INCOMPAT_FEATURE(sb, EXT3_FEATURE_INCOMPAT_META_BG) || + nr < first_meta_bg) + return (logic_sb_block + nr + 1); + bg = sbi->s_desc_per_block * nr; + if (ext3_bg_has_super(sb, bg)) + has_super = 1; + return (first_data_block + has_super + (bg * sbi->s_blocks_per_group)); +} + + struct super_block * ext3_read_super (struct super_block * sb, void * data, int silent) { struct buffer_head * bh; struct ext3_super_block *es = 0; struct ext3_sb_info *sbi = EXT3_SB(sb); - unsigned long sb_block = 1; - unsigned long logic_sb_block = 1; - unsigned long offset = 0; + unsigned long sb_block = get_sb_block(&data); + unsigned long block, logic_sb_block; + unsigned long offset; unsigned long journal_inum = 0; kdev_t dev = sb->s_dev; + unsigned long def_mount_opts; int blocksize; int hblock; int db_count; @@ -922,14 +929,6 @@ if (blocksize < hblock) blocksize = hblock; - sbi->s_mount_opt = 0; - sbi->s_resuid = EXT3_DEF_RESUID; - sbi->s_resgid = EXT3_DEF_RESGID; - if (!parse_options ((char *) data, &sb_block, sbi, &journal_inum, 0)) { - sb->s_dev = 0; - goto out_fail; - } - sb->s_blocksize = blocksize; set_blocksize (dev, blocksize); @@ -937,10 +936,8 @@ * The ext3 superblock will not be buffer aligned for other than 1kB * block sizes. We need to calculate the offset from buffer start. */ - if (blocksize != EXT3_MIN_BLOCK_SIZE) { - logic_sb_block = (sb_block * EXT3_MIN_BLOCK_SIZE) / blocksize; - offset = (sb_block * EXT3_MIN_BLOCK_SIZE) % blocksize; - } + logic_sb_block = (sb_block * EXT3_MIN_BLOCK_SIZE) / blocksize; + offset = (sb_block * EXT3_MIN_BLOCK_SIZE) % blocksize; if (!(bh = sb_bread(sb, logic_sb_block))) { printk (KERN_ERR "EXT3-fs: unable to read superblock\n"); @@ -960,6 +957,34 @@ bdevname(dev)); goto failed_mount; } + + /* Set defaults before we parse the mount options */ + def_mount_opts = le32_to_cpu(es->s_default_mount_opts); + sbi->s_mount_opt = 0; + if (def_mount_opts & EXT3_DEFM_DEBUG) + set_opt(sbi->s_mount_opt, DEBUG); + if (def_mount_opts & EXT3_DEFM_BSDGROUPS) + set_opt(sbi->s_mount_opt, GRPID); + if (def_mount_opts & EXT3_DEFM_UID16) + set_opt(sbi->s_mount_opt, NO_UID32); + if ((def_mount_opts & EXT3_DEFM_JMODE) == EXT3_DEFM_JMODE_DATA) + sbi->s_mount_opt |= EXT3_MOUNT_JOURNAL_DATA; + else if ((def_mount_opts & EXT3_DEFM_JMODE) == EXT3_DEFM_JMODE_ORDERED) + sbi->s_mount_opt |= EXT3_MOUNT_ORDERED_DATA; + else if ((def_mount_opts & EXT3_DEFM_JMODE) == EXT3_DEFM_JMODE_WBACK) + sbi->s_mount_opt |= EXT3_MOUNT_WRITEBACK_DATA; + + if (le16_to_cpu(sbi->s_es->s_errors) == EXT3_ERRORS_PANIC) + set_opt(sbi->s_mount_opt, ERRORS_PANIC); + else if (le16_to_cpu(sbi->s_es->s_errors) == EXT3_ERRORS_RO) + set_opt(sbi->s_mount_opt, ERRORS_RO); + + sbi->s_resuid = le16_to_cpu(es->s_def_resuid); + sbi->s_resgid = le16_to_cpu(es->s_def_resgid); + + if (!parse_options ((char *) data, sbi, &journal_inum, 0)) + goto failed_mount; + if (le32_to_cpu(es->s_rev_level) == EXT3_GOOD_OLD_REV && (EXT3_HAS_COMPAT_FEATURE(sb, ~0U) || EXT3_HAS_RO_COMPAT_FEATURE(sb, ~0U) || @@ -1036,7 +1061,9 @@ } else { sbi->s_inode_size = le16_to_cpu(es->s_inode_size); sbi->s_first_ino = le32_to_cpu(es->s_first_ino); - if (sbi->s_inode_size != EXT3_GOOD_OLD_INODE_SIZE) { + if ((sbi->s_inode_size < EXT3_GOOD_OLD_INODE_SIZE) || + (sbi->s_inode_size & (sbi->s_inode_size - 1)) || + (sbi->s_inode_size > blocksize)) { printk (KERN_ERR "EXT3-fs: unsupported inode size: %d\n", sbi->s_inode_size); @@ -1059,13 +1086,12 @@ sbi->s_itb_per_group = sbi->s_inodes_per_group /sbi->s_inodes_per_block; sbi->s_desc_per_block = blocksize / sizeof(struct ext3_group_desc); sbi->s_sbh = bh; - if (sbi->s_resuid == EXT3_DEF_RESUID) - sbi->s_resuid = le16_to_cpu(es->s_def_resuid); - if (sbi->s_resgid == EXT3_DEF_RESGID) - sbi->s_resgid = le16_to_cpu(es->s_def_resgid); sbi->s_mount_state = le16_to_cpu(es->s_state); sbi->s_addr_per_block_bits = log2(EXT3_ADDR_PER_BLOCK(sb)); sbi->s_desc_per_block_bits = log2(EXT3_DESC_PER_BLOCK(sb)); + for (i=0; i < 4; i++) + sbi->s_hash_seed[i] = le32_to_cpu(es->s_hash_seed[i]); + sbi->s_def_hash_version = es->s_def_hash_version; if (sbi->s_blocks_per_group > blocksize * 8) { printk (KERN_ERR @@ -1098,8 +1124,16 @@ printk (KERN_ERR "EXT3-fs: not enough memory\n"); goto failed_mount; } + sbi->s_debts = kmalloc(sbi->s_groups_count * sizeof(*sbi->s_debts), + GFP_KERNEL); + if (!sbi->s_debts) { + printk ("EXT3-fs: not enough memory\n"); + goto failed_mount2; + } + memset(sbi->s_debts, 0, sbi->s_groups_count * sizeof(*sbi->s_debts)); for (i = 0; i < db_count; i++) { - sbi->s_group_desc[i] = sb_bread(sb, logic_sb_block + i + 1); + block = descriptor_loc(sb, logic_sb_block, i); + sbi->s_group_desc[i] = sb_bread(sb, block); if (!sbi->s_group_desc[i]) { printk (KERN_ERR "EXT3-fs: " "can't read group descriptor %d\n", i); @@ -1120,6 +1154,7 @@ sbi->s_loaded_inode_bitmaps = 0; sbi->s_loaded_block_bitmaps = 0; sbi->s_gdb_count = db_count; + sbi->s_dir_count = ext3_count_dirs(sb); get_random_bytes(&sbi->s_next_generation, sizeof(u32)); /* * set up enough so that it can read an inode @@ -1223,6 +1258,8 @@ failed_mount3: journal_destroy(sbi->s_journal); failed_mount2: + if (sbi->s_debts) + kfree(sbi->s_debts); for (i = 0; i < db_count; i++) brelse(sbi->s_group_desc[i]); kfree(sbi->s_group_desc); @@ -1651,7 +1688,7 @@ /* * Allow the "check" option to be passed as a remount option. */ - if (!parse_options(data, &tmp, sbi, &tmp, 1)) + if (!parse_options(data, sbi, &tmp, 1)) return -EINVAL; if (sbi->s_mount_opt & EXT3_MOUNT_ABORT) diff -Nru a/include/linux/ext2_fs.h b/include/linux/ext2_fs.h --- a/include/linux/ext2_fs.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/ext2_fs.h Thu Nov 7 11:58:05 2002 @@ -198,10 +198,11 @@ #define EXT2_ECOMPR_FL 0x00000800 /* Compression error */ /* End compression flags --- maybe not all used */ #define EXT2_BTREE_FL 0x00001000 /* btree format dir */ +#define EXT2_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ #define EXT2_RESERVED_FL 0x80000000 /* reserved for ext2 lib */ -#define EXT2_FL_USER_VISIBLE 0x00001FFF /* User visible flags */ -#define EXT2_FL_USER_MODIFIABLE 0x000000FF /* User modifiable flags */ +#define EXT2_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ +#define EXT2_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ /* * ioctl commands @@ -307,6 +308,7 @@ * Mount flags */ #define EXT2_MOUNT_CHECK 0x0001 /* Do mount-time checks */ +#define EXT2_MOUNT_OLDALLOC 0x0002 /* Don't use the new Orlov allocator */ #define EXT2_MOUNT_GRPID 0x0004 /* Create files with directory's group */ #define EXT2_MOUNT_DEBUG 0x0008 /* Some debugging messages */ #define EXT2_MOUNT_ERRORS_CONT 0x0010 /* Continue on errors */ @@ -392,7 +394,20 @@ __u8 s_prealloc_blocks; /* Nr of blocks to try to preallocate*/ __u8 s_prealloc_dir_blocks; /* Nr to preallocate for dirs */ __u16 s_padding1; - __u32 s_reserved[204]; /* Padding to the end of the block */ + /* + * Journaling support valid if EXT3_FEATURE_COMPAT_HAS_JOURNAL set. + */ + __u8 s_journal_uuid[16]; /* uuid of journal superblock */ + __u32 s_journal_inum; /* inode number of journal file */ + __u32 s_journal_dev; /* device number of journal file */ + __u32 s_last_orphan; /* start of list of inodes to delete */ + __u32 s_hash_seed[4]; /* HTREE hash seed */ + __u8 s_def_hash_version; /* Default hash version to use */ + __u8 s_reserved_char_pad; + __u16 s_reserved_word_pad; + __u32 s_default_mount_opts; + __u32 s_first_meta_bg; /* First metablock block group */ + __u32 s_reserved[190]; /* Padding to the end of the block */ }; #ifdef __KERNEL__ @@ -464,10 +479,12 @@ #define EXT2_FEATURE_INCOMPAT_FILETYPE 0x0002 #define EXT3_FEATURE_INCOMPAT_RECOVER 0x0004 #define EXT3_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 +#define EXT2_FEATURE_INCOMPAT_META_BG 0x0010 #define EXT2_FEATURE_INCOMPAT_ANY 0xffffffff #define EXT2_FEATURE_COMPAT_SUPP 0 -#define EXT2_FEATURE_INCOMPAT_SUPP EXT2_FEATURE_INCOMPAT_FILETYPE +#define EXT2_FEATURE_INCOMPAT_SUPP (EXT2_FEATURE_INCOMPAT_FILETYPE| \ + EXT2_FEATURE_INCOMPAT_META_BG) #define EXT2_FEATURE_RO_COMPAT_SUPP (EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER| \ EXT2_FEATURE_RO_COMPAT_LARGE_FILE| \ EXT2_FEATURE_RO_COMPAT_BTREE_DIR) @@ -475,10 +492,18 @@ #define EXT2_FEATURE_INCOMPAT_UNSUPPORTED ~EXT2_FEATURE_INCOMPAT_SUPP /* - * Default values for user and/or group using reserved blocks + * Default mount options */ -#define EXT2_DEF_RESUID 0 -#define EXT2_DEF_RESGID 0 +#define EXT2_DEFM_DEBUG 0x0001 +#define EXT2_DEFM_BSDGROUPS 0x0002 +#define EXT2_DEFM_XATTR_USER 0x0004 +#define EXT2_DEFM_ACL 0x0008 +#define EXT2_DEFM_UID16 0x0010 + /* Not used by ext2, but reserved for use by ext3 */ +#define EXT3_DEFM_JMODE 0x0060 +#define EXT3_DEFM_JMODE_DATA 0x0020 +#define EXT3_DEFM_JMODE_ORDERED 0x0040 +#define EXT3_DEFM_JMODE_WBACK 0x0060 /* * Structure of a directory entry @@ -576,6 +601,7 @@ extern struct inode * ext2_new_inode (const struct inode *, int); extern void ext2_free_inode (struct inode *); extern unsigned long ext2_count_free_inodes (struct super_block *); +extern unsigned long ext2_count_dirs (struct super_block *); extern void ext2_check_inodes_bitmap (struct super_block *); extern unsigned long ext2_count_free (struct buffer_head *, unsigned); diff -Nru a/include/linux/ext2_fs_i.h b/include/linux/ext2_fs_i.h --- a/include/linux/ext2_fs_i.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/ext2_fs_i.h Thu Nov 7 11:58:05 2002 @@ -25,6 +25,7 @@ __u32 i_faddr; __u8 i_frag_no; __u8 i_frag_size; + __u16 i_state; __u32 i_file_acl; __u32 i_dir_acl; __u32 i_dtime; @@ -34,7 +35,11 @@ __u32 i_prealloc_block; __u32 i_prealloc_count; __u32 i_dir_start_lookup; - int i_new_inode:1; /* Is a freshly allocated inode */ }; + +/* + * Inode dynamic state flags + */ +#define EXT2_STATE_NEW 0x00000001 /* inode is newly created */ #endif /* _LINUX_EXT2_FS_I */ diff -Nru a/include/linux/ext2_fs_sb.h b/include/linux/ext2_fs_sb.h --- a/include/linux/ext2_fs_sb.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/ext2_fs_sb.h Thu Nov 7 11:58:05 2002 @@ -56,6 +56,8 @@ int s_desc_per_block_bits; int s_inode_size; int s_first_ino; + unsigned long s_dir_count; + u8 *s_debts; }; #endif /* _LINUX_EXT2_FS_SB */ diff -Nru a/include/linux/ext3_fs.h b/include/linux/ext3_fs.h --- a/include/linux/ext3_fs.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/ext3_fs.h Thu Nov 7 11:58:05 2002 @@ -40,6 +40,11 @@ #define EXT3FS_VERSION "2.4-0.9.19" /* + * Always enable hashed directories + */ +#define CONFIG_EXT3_INDEX + +/* * Debug code */ #ifdef EXT3FS_DEBUG @@ -203,10 +208,11 @@ #define EXT3_INDEX_FL 0x00001000 /* hash-indexed directory */ #define EXT3_IMAGIC_FL 0x00002000 /* AFS directory */ #define EXT3_JOURNAL_DATA_FL 0x00004000 /* file data should be journaled */ +#define EXT3_TOPDIR_FL 0x00020000 /* Top of directory hierarchies*/ #define EXT3_RESERVED_FL 0x80000000 /* reserved for ext3 lib */ -#define EXT3_FL_USER_VISIBLE 0x00005FFF /* User visible flags */ -#define EXT3_FL_USER_MODIFIABLE 0x000000FF /* User modifiable flags */ +#define EXT3_FL_USER_VISIBLE 0x0003DFFF /* User visible flags */ +#define EXT3_FL_USER_MODIFIABLE 0x000380FF /* User modifiable flags */ /* * Inode dynamic state flags @@ -325,6 +331,7 @@ * Mount flags */ #define EXT3_MOUNT_CHECK 0x0001 /* Do mount-time checks */ +#define EXT3_MOUNT_OLDALLOC 0x0002 /* Don't use the new Orlov allocator */ #define EXT3_MOUNT_GRPID 0x0004 /* Create files with directory's group */ #define EXT3_MOUNT_DEBUG 0x0008 /* Some debugging messages */ #define EXT3_MOUNT_ERRORS_CONT 0x0010 /* Continue on errors */ @@ -437,8 +444,13 @@ /*E0*/ __u32 s_journal_inum; /* inode number of journal file */ __u32 s_journal_dev; /* device number of journal file */ __u32 s_last_orphan; /* start of list of inodes to delete */ - -/*EC*/ __u32 s_reserved[197]; /* Padding to the end of the block */ + __u32 s_hash_seed[4]; /* HTREE hash seed */ + __u8 s_def_hash_version; /* Default hash version to use */ + __u8 s_reserved_char_pad; + __u16 s_reserved_word_pad; + __u32 s_default_mount_opts; + __u32 s_first_meta_bg; /* First metablock block group */ + __u32 s_reserved[190]; /* Padding to the end of the block */ }; #ifdef __KERNEL__ @@ -511,19 +523,28 @@ #define EXT3_FEATURE_INCOMPAT_FILETYPE 0x0002 #define EXT3_FEATURE_INCOMPAT_RECOVER 0x0004 /* Needs recovery */ #define EXT3_FEATURE_INCOMPAT_JOURNAL_DEV 0x0008 /* Journal device */ +#define EXT3_FEATURE_INCOMPAT_META_BG 0x0010 #define EXT3_FEATURE_COMPAT_SUPP 0 #define EXT3_FEATURE_INCOMPAT_SUPP (EXT3_FEATURE_INCOMPAT_FILETYPE| \ - EXT3_FEATURE_INCOMPAT_RECOVER) + EXT3_FEATURE_INCOMPAT_RECOVER| \ + EXT3_FEATURE_INCOMPAT_META_BG) #define EXT3_FEATURE_RO_COMPAT_SUPP (EXT3_FEATURE_RO_COMPAT_SPARSE_SUPER| \ EXT3_FEATURE_RO_COMPAT_LARGE_FILE| \ EXT3_FEATURE_RO_COMPAT_BTREE_DIR) /* - * Default values for user and/or group using reserved blocks + * Default mount options */ -#define EXT3_DEF_RESUID 0 -#define EXT3_DEF_RESGID 0 +#define EXT3_DEFM_DEBUG 0x0001 +#define EXT3_DEFM_BSDGROUPS 0x0002 +#define EXT3_DEFM_XATTR_USER 0x0004 +#define EXT3_DEFM_ACL 0x0008 +#define EXT3_DEFM_UID16 0x0010 +#define EXT3_DEFM_JMODE 0x0060 +#define EXT3_DEFM_JMODE_DATA 0x0020 +#define EXT3_DEFM_JMODE_ORDERED 0x0040 +#define EXT3_DEFM_JMODE_WBACK 0x0060 /* * Structure of a directory entry @@ -575,9 +596,46 @@ #define EXT3_DIR_ROUND (EXT3_DIR_PAD - 1) #define EXT3_DIR_REC_LEN(name_len) (((name_len) + 8 + EXT3_DIR_ROUND) & \ ~EXT3_DIR_ROUND) +/* + * Hash Tree Directory indexing + * (c) Daniel Phillips, 2001 + */ + +#ifdef CONFIG_EXT3_INDEX + #define is_dx(dir) (EXT3_HAS_COMPAT_FEATURE(dir->i_sb, \ + EXT3_FEATURE_COMPAT_DIR_INDEX) && \ + (EXT3_I(dir)->i_flags & EXT3_INDEX_FL)) +#define EXT3_DIR_LINK_MAX(dir) (!is_dx(dir) && (dir)->i_nlink >= EXT3_LINK_MAX) +#define EXT3_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2 || (dir)->i_nlink == 1) +#else + #define is_dx(dir) 0 +#define EXT3_DIR_LINK_MAX(dir) ((dir)->i_nlink >= EXT3_LINK_MAX) +#define EXT3_DIR_LINK_EMPTY(dir) ((dir)->i_nlink == 2) +#endif + +/* Legal values for the dx_root hash_version field: */ + +#define DX_HASH_LEGACY 0 +#define DX_HASH_HALF_MD4 1 +#define DX_HASH_TEA 2 + +/* hash info structure used by the directory hash */ +struct dx_hash_info +{ + u32 hash; + u32 minor_hash; + int hash_version; + u32 *seed; +}; #ifdef __KERNEL__ /* + * Control parameters used by ext3_htree_next_block + */ +#define HASH_NB_ALWAYS 1 + + +/* * Describe an inode's exact location on disk and in memory */ struct ext3_iloc @@ -587,6 +645,27 @@ unsigned long block_group; }; + +/* + * This structure is stuffed into the struct file's private_data field + * for directories. It is where we put information so that we can do + * readdir operations in hash tree order. + */ +struct dir_private_info { + rb_root_t root; + rb_node_t *curr_node; + struct fname *extra_fname; + loff_t last_pos; + __u32 curr_hash; + __u32 curr_minor_hash; + __u32 next_hash; +}; + +/* + * Special error return code only used by dx_probe() and its callers. + */ +#define ERR_BAD_DX_DIR -75000 + /* * Function prototypes */ @@ -614,16 +693,26 @@ /* dir.c */ extern int ext3_check_dir_entry(const char *, struct inode *, - struct ext3_dir_entry_2 *, struct buffer_head *, - unsigned long); + struct ext3_dir_entry_2 *, + struct buffer_head *, unsigned long); +extern int ext3_htree_store_dirent(struct file *dir_file, __u32 hash, + __u32 minor_hash, + struct ext3_dir_entry_2 *dirent); +extern void ext3_htree_free_dir_info(struct dir_private_info *p); + /* fsync.c */ extern int ext3_sync_file (struct file *, struct dentry *, int); +/* hash.c */ +extern int ext3fs_dirhash(const char *name, int len, struct + dx_hash_info *hinfo); + /* ialloc.c */ extern struct inode * ext3_new_inode (handle_t *, const struct inode *, int); extern void ext3_free_inode (handle_t *, struct inode *); extern struct inode * ext3_orphan_get (struct super_block *, ino_t); extern unsigned long ext3_count_free_inodes (struct super_block *); +extern unsigned long ext3_count_dirs (struct super_block *); extern void ext3_check_inodes_bitmap (struct super_block *); extern unsigned long ext3_count_free (struct buffer_head *, unsigned); @@ -650,6 +739,8 @@ /* namei.c */ extern int ext3_orphan_add(handle_t *, struct inode *); extern int ext3_orphan_del(handle_t *, struct inode *); +extern int ext3_htree_fill_tree(struct file *dir_file, __u32 start_hash, + __u32 start_minor_hash, __u32 *next_hash); /* super.c */ extern void ext3_error (struct super_block *, const char *, const char *, ...) diff -Nru a/include/linux/ext3_fs_sb.h b/include/linux/ext3_fs_sb.h --- a/include/linux/ext3_fs_sb.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/ext3_fs_sb.h Thu Nov 7 11:58:05 2002 @@ -62,6 +62,10 @@ int s_inode_size; int s_first_ino; u32 s_next_generation; + u32 s_hash_seed[4]; + int s_def_hash_version; + unsigned long s_dir_count; + u8 *s_debts; /* Journaling */ struct inode * s_journal_inode; diff -Nru a/include/linux/ext3_jbd.h b/include/linux/ext3_jbd.h --- a/include/linux/ext3_jbd.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/ext3_jbd.h Thu Nov 7 11:58:05 2002 @@ -63,6 +63,8 @@ #define EXT3_RESERVE_TRANS_BLOCKS 12 +#define EXT3_INDEX_EXTRA_TRANS_BLOCKS 8 + int ext3_mark_iloc_dirty(handle_t *handle, struct inode *inode, diff -Nru a/include/linux/rbtree.h b/include/linux/rbtree.h --- a/include/linux/rbtree.h Thu Nov 7 11:58:05 2002 +++ b/include/linux/rbtree.h Thu Nov 7 11:58:05 2002 @@ -120,6 +120,8 @@ extern void rb_insert_color(rb_node_t *, rb_root_t *); extern void rb_erase(rb_node_t *, rb_root_t *); +extern rb_node_t *rb_get_first(rb_root_t *root); +extern rb_node_t *rb_get_next(rb_node_t *n); static inline void rb_link_node(rb_node_t * node, rb_node_t * parent, rb_node_t ** rb_link) { diff -Nru a/lib/rbtree.c b/lib/rbtree.c --- a/lib/rbtree.c Thu Nov 7 11:58:05 2002 +++ b/lib/