Whamcloud - gitweb
tune2fs: Don't allow the -I option if the flex_bg feature is enabled
[tools/e2fsprogs.git] / misc / tune2fs.c
index 85046e2..14529e3 100644 (file)
@@ -25,7 +25,7 @@
  * 94/03/06    - Added the checks interval from Uwe Ohse (uwe@tirka.gun.de)
  */
 
-#define _XOPEN_SOURCE 500 /* for inclusion of strptime() */
+#define _XOPEN_SOURCE 600 /* for inclusion of strptime() */
 #define _BSD_SOURCE /* for inclusion of strcasecmp() */
 #include <fcntl.h>
 #include <grp.h>
@@ -37,11 +37,15 @@ extern int optind;
 #endif
 #include <pwd.h>
 #include <stdio.h>
+#ifdef HAVE_STDLIB_H
 #include <stdlib.h>
+#endif
 #include <string.h>
 #include <time.h>
 #include <unistd.h>
 #include <sys/types.h>
+#include <libgen.h>
+#include <limits.h>
 
 #include "ext2fs/ext2_fs.h"
 #include "ext2fs/ext2fs.h"
@@ -55,22 +59,13 @@ extern int optind;
 #include "../version.h"
 #include "nls-enable.h"
 
-/* 
- * Tune2fs supports these features in addition to the standard features.
- */
-#define EXT2_TUNE2FS_INCOMPAT  (EXT3_FEATURE_INCOMPAT_EXTENTS)
-#define EXT2_TUNE2FS_RO_COMPAT (EXT4_FEATURE_RO_COMPAT_HUGE_FILE|\
-                                EXT4_FEATURE_RO_COMPAT_GDT_CSUM|       \
-                                EXT4_FEATURE_RO_COMPAT_DIR_NLINK|      \
-                                EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE)
-
-
 const char * program_name = "tune2fs";
 char * device_name;
 char * new_label, *new_last_mounted, *new_UUID;
 char * io_options;
 static int c_flag, C_flag, e_flag, f_flag, g_flag, i_flag, l_flag, L_flag;
 static int m_flag, M_flag, r_flag, s_flag = -1, u_flag, U_flag, T_flag;
+static int I_flag;
 static time_t last_check_time;
 static int print_label;
 static int max_mount_count, mount_count, mount_flags;
@@ -84,10 +79,20 @@ static char *mntopts_cmd;
 static int stride, stripe_width;
 static int stride_set, stripe_width_set;
 static char *extended_cmd;
+static unsigned long new_inode_size;
 
 int journal_size, journal_flags;
 char *journal_device;
 
+static struct list_head blk_move_list;
+
+struct blk_move {
+       struct list_head list;
+       blk_t old_loc;
+       blk_t new_loc;
+};
+
+
 static const char *please_fsck = N_("Please run e2fsck on the filesystem.\n");
 
 void do_findfs(int argc, char **argv);
@@ -104,7 +109,7 @@ static void usage(void)
                  "[-L volume_label]\n"
                  "\t[-M last_mounted_dir] [-O [^]feature[,...]]\n"
                  "\t[-E extended-option[,...]] [-T last_check_time] "
-                 "[-U UUID] device\n"), program_name);
+                 "[-U UUID]\n\t[ -I new_inode_size ] device\n"), program_name);
        exit (1);
 }
 
@@ -118,6 +123,9 @@ static __u32 ok_features[3] = {
                EXT4_FEATURE_INCOMPAT_FLEX_BG,
        /* R/O compat */
        EXT2_FEATURE_RO_COMPAT_LARGE_FILE |
+               EXT4_FEATURE_RO_COMPAT_HUGE_FILE|
+               EXT4_FEATURE_RO_COMPAT_DIR_NLINK|
+               EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE|
                EXT4_FEATURE_RO_COMPAT_GDT_CSUM |
                EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER
 };
@@ -132,6 +140,9 @@ static __u32 clear_ok_features[3] = {
                EXT4_FEATURE_INCOMPAT_FLEX_BG,
        /* R/O compat */
        EXT2_FEATURE_RO_COMPAT_LARGE_FILE |
+               EXT4_FEATURE_RO_COMPAT_HUGE_FILE|
+               EXT4_FEATURE_RO_COMPAT_DIR_NLINK|
+               EXT4_FEATURE_RO_COMPAT_EXTRA_ISIZE|
                EXT4_FEATURE_RO_COMPAT_GDT_CSUM
 };
 
@@ -163,11 +174,12 @@ static void remove_journal_device(ext2_filsys fs)
        }
 
 #ifdef CONFIG_TESTIO_DEBUG
-       io_ptr = test_io_manager;
-       test_io_backing_manager = unix_io_manager;
-#else
-       io_ptr = unix_io_manager;
+       if (getenv("TEST_IO_FLAGS") || getenv("TEST_IO_BLOCK")) {
+               io_ptr = test_io_manager;
+               test_io_backing_manager = unix_io_manager;
+       } else
 #endif
+               io_ptr = unix_io_manager;
        retval = ext2fs_open(journal_path, EXT2_FLAG_RW|
                             EXT2_FLAG_JOURNAL_DEV_OK, 0,
                             fs->blocksize, io_ptr, &jfs);
@@ -204,7 +216,7 @@ static void remove_journal_device(ext2_filsys fs)
                        break;
        }
        if (i >= nr_users) {
-               fputs(_("Filesystem's UUID not found on journal device.\n"), 
+               fputs(_("Filesystem's UUID not found on journal device.\n"),
                      stderr);
                commit_remove_journal = 1;
                goto no_valid_journal;
@@ -237,7 +249,7 @@ no_valid_journal:
 
 /* Helper function for remove_journal_inode */
 static int release_blocks_proc(ext2_filsys fs, blk_t *blocknr,
-                              int blockcnt EXT2FS_ATTR((unused)), 
+                              int blockcnt EXT2FS_ATTR((unused)),
                               void *private EXT2FS_ATTR((unused)))
 {
        blk_t   block;
@@ -260,7 +272,7 @@ static void remove_journal_inode(ext2_filsys fs)
        struct ext2_inode       inode;
        errcode_t               retval;
        ino_t                   ino = fs->super->s_journal_inum;
-       
+
        retval = ext2fs_read_inode(fs, ino,  &inode);
        if (retval) {
                com_err(program_name, retval,
@@ -356,7 +368,7 @@ static void update_feature_set(ext2_filsys fs, char *features)
        if (FEATURE_OFF(E2P_FEATURE_COMPAT, EXT3_FEATURE_COMPAT_HAS_JOURNAL)) {
                if ((mount_flags & EXT2_MF_MOUNTED) &&
                    !(mount_flags & EXT2_MF_READONLY)) {
-                       fputs(_("The has_journal flag may only be "
+                       fputs(_("The has_journal feature may only be "
                                "cleared when the filesystem is\n"
                                "unmounted or mounted "
                                "read-only.\n"), stderr);
@@ -380,9 +392,8 @@ static void update_feature_set(ext2_filsys fs, char *features)
        if (FEATURE_ON(E2P_FEATURE_COMPAT, EXT3_FEATURE_COMPAT_HAS_JOURNAL)) {
                /*
                 * If adding a journal flag, let the create journal
-                * code below handle creating setting the flag and
-                * creating the journal.  We supply a default size if
-                * necessary.
+                * code below handle setting the flag and creating the
+                * journal.  We supply a default size if necessary.
                 */
                if (!journal_size)
                        journal_size = -1;
@@ -391,7 +402,7 @@ static void update_feature_set(ext2_filsys fs, char *features)
 
        if (FEATURE_ON(E2P_FEATURE_COMPAT, EXT2_FEATURE_COMPAT_DIR_INDEX)) {
                if (!sb->s_def_hash_version)
-                       sb->s_def_hash_version = EXT2_HASH_TEA;
+                       sb->s_def_hash_version = EXT2_HASH_HALF_MD4;
                if (uuid_is_null((unsigned char *) sb->s_hash_seed))
                        uuid_generate((unsigned char *) sb->s_hash_seed);
        }
@@ -405,6 +416,18 @@ static void update_feature_set(ext2_filsys fs, char *features)
                }
        }
 
+       if (FEATURE_OFF(E2P_FEATURE_RO_INCOMPAT,
+                           EXT4_FEATURE_RO_COMPAT_HUGE_FILE)) {
+               if ((mount_flags & EXT2_MF_MOUNTED) &&
+                   !(mount_flags & EXT2_MF_READONLY)) {
+                       fputs(_("The huge_file feature may only be "
+                               "cleared when the filesystem is\n"
+                               "unmounted or mounted "
+                               "read-only.\n"), stderr);
+                       exit(1);
+               }
+       }
+
        if (sb->s_rev_level == EXT2_GOOD_OLD_REV &&
            (sb->s_feature_compat || sb->s_feature_ro_compat ||
             sb->s_feature_incompat))
@@ -414,6 +437,8 @@ static void update_feature_set(ext2_filsys fs, char *features)
                            EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER) ||
            FEATURE_CHANGED(E2P_FEATURE_RO_INCOMPAT,
                            EXT4_FEATURE_RO_COMPAT_GDT_CSUM) ||
+           FEATURE_OFF(E2P_FEATURE_RO_INCOMPAT,
+                       EXT4_FEATURE_RO_COMPAT_HUGE_FILE) ||
            FEATURE_CHANGED(E2P_FEATURE_INCOMPAT,
                            EXT2_FEATURE_INCOMPAT_FILETYPE) ||
            FEATURE_CHANGED(E2P_FEATURE_COMPAT,
@@ -422,6 +447,8 @@ static void update_feature_set(ext2_filsys fs, char *features)
                        EXT2_FEATURE_RO_COMPAT_LARGE_FILE)) {
                sb->s_state &= ~EXT2_VALID_FS;
                printf("\n%s\n", _(please_fsck));
+               if (mount_flags & EXT2_MF_READONLY)
+                       printf(_("(and reboot afterwards!)\n"));
        }
 
        if ((old_features[E2P_FEATURE_COMPAT] != sb->s_feature_compat) ||
@@ -449,11 +476,12 @@ static void add_journal(ext2_filsys fs)
                check_plausibility(journal_device);
                check_mount(journal_device, 0, _("journal"));
 #ifdef CONFIG_TESTIO_DEBUG
-               io_ptr = test_io_manager;
-               test_io_backing_manager = unix_io_manager;
-#else
-               io_ptr = unix_io_manager;
+               if (getenv("TEST_IO_FLAGS") || getenv("TEST_IO_BLOCK")) {
+                       io_ptr = test_io_manager;
+                       test_io_backing_manager = unix_io_manager;
+               } else
 #endif
+                       io_ptr = unix_io_manager;
                retval = ext2fs_open(journal_device, EXT2_FLAG_RW|
                                     EXT2_FLAG_JOURNAL_DEV_OK, 0,
                                     fs->blocksize, io_ptr, &jfs);
@@ -518,16 +546,16 @@ static void parse_e2label_options(int argc, char ** argv)
                *io_options++ = 0;
        device_name = blkid_get_devname(NULL, argv[1], NULL);
        if (!device_name) {
-               com_err("e2label", 0, _("Unable to resolve '%s'"), 
+               com_err("e2label", 0, _("Unable to resolve '%s'"),
                        argv[1]);
                exit(1);
        }
-       open_flag = EXT2_FLAG_SOFTSUPP_FEATURES | EXT2_FLAG_JOURNAL_DEV_OK;
+       open_flag = EXT2_FLAG_JOURNAL_DEV_OK;
        if (argc == 3) {
                open_flag |= EXT2_FLAG_RW;
                L_flag = 1;
                new_label = argv[2];
-       } else 
+       } else
                print_label++;
 }
 
@@ -568,10 +596,10 @@ static void parse_tune2fs_options(int argc, char **argv)
        struct group * gr;
        struct passwd * pw;
 
-       open_flag = EXT2_FLAG_SOFTSUPP_FEATURES;
+       open_flag = 0;
 
        printf("tune2fs %s (%s)\n", E2FSPROGS_VERSION, E2FSPROGS_DATE);
-       while ((c = getopt(argc, argv, "c:e:fg:i:jlm:o:r:s:u:C:E:J:L:M:O:T:U:")) != EOF)
+       while ((c = getopt(argc, argv, "c:e:fg:i:jlm:o:r:s:u:C:E:I:J:L:M:O:T:U:")) != EOF)
                switch (c)
                {
                        case 'c':
@@ -716,7 +744,7 @@ static void parse_tune2fs_options(int argc, char **argv)
                                mntopts_cmd = optarg;
                                open_flag = EXT2_FLAG_RW;
                                break;
-                               
+
                        case 'O':
                                if (features_cmd) {
                                        com_err (program_name, 0,
@@ -772,6 +800,25 @@ static void parse_tune2fs_options(int argc, char **argv)
                                open_flag = EXT2_FLAG_RW |
                                        EXT2_FLAG_JOURNAL_DEV_OK;
                                break;
+                       case 'I':
+                               new_inode_size = strtoul (optarg, &tmp, 0);
+                               if (*tmp) {
+                                       com_err (program_name, 0,
+                                               _("bad inode size - %s"),
+                                                       optarg);
+                                       usage();
+                               }
+                               if (!((new_inode_size &
+                                               (new_inode_size - 1)) == 0)) {
+                                       com_err (program_name, 0,
+                                               _("Inode size must be a "
+                                               "power of two- %s"),
+                                                       optarg);
+                                       usage();
+                               }
+                               open_flag = EXT2_FLAG_RW;
+                               I_flag = 1;
+                               break;
                        default:
                                usage();
                }
@@ -784,7 +831,7 @@ static void parse_tune2fs_options(int argc, char **argv)
                *io_options++ = 0;
        device_name = blkid_get_devname(NULL, argv[optind], NULL);
        if (!device_name) {
-               com_err("tune2fs", 0, _("Unable to resolve '%s'"), 
+               com_err("tune2fs", 0, _("Unable to resolve '%s'"),
                        argv[optind]);
                exit(1);
        }
@@ -801,7 +848,7 @@ void do_findfs(int argc, char **argv)
        }
        dev = blkid_get_devname(NULL, argv[1], NULL);
        if (!dev) {
-               com_err("findfs", 0, _("Unable to resolve '%s'"), 
+               com_err("findfs", 0, _("Unable to resolve '%s'"),
                        argv[1]);
                exit(1);
        }
@@ -809,16 +856,10 @@ void do_findfs(int argc, char **argv)
        exit(0);
 }
 
-/*
- * Note!  If any extended options are incompatible with the
- * intersection of the SOFTSUPP features and those features explicitly
- * enabled for tune2fs, there needs to be an explicit test for them
- * here.
- */
 static void parse_extended_opts(ext2_filsys fs, const char *opts)
 {
        char    *buf, *token, *next, *p, *arg;
-       int     len;
+       int     len, hash_alg;
        int     r_usage = 0;
 
        len = strlen(opts);
@@ -878,7 +919,26 @@ static void parse_extended_opts(ext2_filsys fs, const char *opts)
                                continue;
                        }
                        stripe_width_set = 1;
-               } else 
+               } else if (strcmp(token, "hash_alg") == 0 ||
+                          strcmp(token, "hash-alg") == 0) {
+                       if (!arg) {
+                               r_usage++;
+                               continue;
+                       }
+                       hash_alg = e2p_string2hash(arg);
+                       if (hash_alg < 0) {
+                               fprintf(stderr, 
+                                       _("Invalid hash algorithm: %s\n"),
+                                       arg);
+                               r_usage++;
+                               continue;
+                       }
+                       fs->super->s_def_hash_version = hash_alg;
+                       printf(_("Setting default hash algorithm "
+                                "to %s (%d)\n"), 
+                              arg, hash_alg);
+                       ext2fs_mark_super_dirty(fs);
+               } else
                        r_usage++;
        }
        if (r_usage) {
@@ -888,22 +948,480 @@ static void parse_extended_opts(ext2_filsys fs, const char *opts)
                        "\tis set off by an equals ('=') sign.\n\n"
                        "Valid extended options are:\n"
                        "\tstride=<RAID per-disk chunk size in blocks>\n"
-                       "\tstripe-width=<RAID stride*data disks in blocks>\n"
+                       "\tstripe_width=<RAID stride*data disks in blocks>\n"
+                       "\thash_alg=<hash algorithm>\n"
                        "\ttest_fs\n"
                        "\t^test_fs\n"));
                free(buf);
                exit(1);
        }
        free(buf);
-}      
+}
+
+static int get_move_bitmap(ext2_filsys fs, int new_ino_blks_per_grp,
+                                       ext2fs_block_bitmap bmap)
+{
+       dgrp_t i;
+       blk_t j, needed_blocks = 0;
+       blk_t start_blk, end_blk;
+
+       for (i = 0; i < fs->group_desc_count; i++) {
+
+               start_blk = fs->group_desc[i].bg_inode_table +
+                                       fs->inode_blocks_per_group;
+
+               end_blk = fs->group_desc[i].bg_inode_table +
+                                       new_ino_blks_per_grp;
+
+               for (j = start_blk; j < end_blk; j++) {
+
+                       if (ext2fs_test_block_bitmap(fs->block_map, j)) {
+                               /* FIXME!!
+                                * What happens if the block is marked
+                                * as a bad block
+                                */
+                               ext2fs_mark_block_bitmap(bmap, j);
+                               needed_blocks++;
+                       } else {
+                               /*
+                                * We are going to use this block for
+                                * inode table. So mark them used.
+                                */
+                               ext2fs_mark_block_bitmap(fs->block_map, j);
+                       }
+               }
+       }
+
+       if (needed_blocks > fs->super->s_free_blocks_count ) {
+               return ENOSPC;
+       }
+
+       return 0;
+}
+
+static int move_block(ext2_filsys fs, ext2fs_block_bitmap bmap)
+{
+       char *buf;
+       errcode_t retval;
+       blk_t blk, new_blk;
+       struct blk_move *bmv;
+
+
+       retval = ext2fs_get_mem(fs->blocksize, &buf);
+       if (retval)
+               return retval;
+
+       for (new_blk = blk = fs->super->s_first_data_block;
+            blk < fs->super->s_blocks_count; blk++) {
+
+               if (!ext2fs_test_block_bitmap(bmap, blk))
+                       continue;
+
+               retval = ext2fs_new_block(fs, new_blk, NULL, &new_blk);
+               if (retval)
+                       goto err_out;
+
+               /* Mark this block as allocated */
+               ext2fs_mark_block_bitmap(fs->block_map, new_blk);
+
+               /* Add it to block move list */
+               retval = ext2fs_get_mem(sizeof(struct blk_move), &bmv);
+               if (retval)
+                       goto err_out;
+
+               bmv->old_loc = blk;
+               bmv->new_loc = new_blk;
+
+               list_add(&(bmv->list), &blk_move_list);
+
+               retval = io_channel_read_blk(fs->io, blk, 1, buf);
+               if (retval)
+                       goto err_out;
+
+               retval = io_channel_write_blk(fs->io, new_blk, 1, buf);
+               if (retval)
+                       goto err_out;
+       }
+
+err_out:
+       ext2fs_free_mem(&buf);
+       return retval;
+}
+
+static blk_t transalate_block(blk_t blk)
+{
+       struct list_head *entry;
+       struct blk_move *bmv;
+
+       list_for_each(entry, &blk_move_list) {
+
+               bmv = list_entry(entry, struct blk_move, list);
+               if (bmv->old_loc == blk)
+                       return bmv->new_loc;
+       }
+
+       return 0;
+}
+
+static int process_block(ext2_filsys fs EXT2FS_ATTR((unused)),
+                        blk_t *block_nr,
+                        e2_blkcnt_t blockcnt EXT2FS_ATTR((unused)),
+                        blk_t ref_block EXT2FS_ATTR((unused)),
+                        int ref_offset EXT2FS_ATTR((unused)),
+                        void *priv_data)
+{
+       int ret = 0;
+       blk_t new_blk;
+       ext2fs_block_bitmap bmap = (ext2fs_block_bitmap) priv_data;
+
+       if (!ext2fs_test_block_bitmap(bmap, *block_nr))
+               return 0;
+       new_blk = transalate_block(*block_nr);
+       if (new_blk) {
+               *block_nr = new_blk;
+               /*
+                * This will force the ext2fs_write_inode in the iterator
+                */
+               ret |= BLOCK_CHANGED;
+       }
+
+       return ret;
+}
+
+static int inode_scan_and_fix(ext2_filsys fs, ext2fs_block_bitmap bmap)
+{
+       errcode_t retval = 0;
+       ext2_ino_t ino;
+       blk_t blk;
+       char *block_buf = 0;
+       struct ext2_inode inode;
+       ext2_inode_scan scan = NULL;
+
+       retval = ext2fs_get_mem(fs->blocksize * 3, &block_buf);
+       if (retval)
+               return retval;
+
+       retval = ext2fs_open_inode_scan(fs, 0, &scan);
+       if (retval)
+               goto err_out;
+
+       while (1) {
+
+               retval = ext2fs_get_next_inode(scan, &ino, &inode);
+               if (retval)
+                       goto err_out;
+
+               if (!ino)
+                       break;
+
+               if (inode.i_links_count == 0)
+                       continue; /* inode not in use */
+
+               /* FIXME!!
+                * If we end up modifying the journal inode
+                * the sb->s_jnl_blocks will differ. But a
+                * subsequent e2fsck fixes that.
+                * Do we need to fix this ??
+                */
+
+               if (inode.i_file_acl &&
+                   ext2fs_test_block_bitmap(bmap, inode.i_file_acl)) {
+                       blk = transalate_block(inode.i_file_acl);
+                       if (!blk)
+                               continue;
+
+                       inode.i_file_acl = blk;
+
+                       /*
+                        * Write the inode to disk so that inode table
+                        * resizing can work
+                        */
+                       retval = ext2fs_write_inode(fs, ino, &inode);
+                       if (retval)
+                               goto err_out;
+               }
+
+               if (!ext2fs_inode_has_valid_blocks(&inode))
+                       continue;
+
+               retval = ext2fs_block_iterate2(fs, ino, 0, block_buf,
+                                              process_block, bmap);
+               if (retval)
+                       goto err_out;
+
+       }
+
+err_out:
+       ext2fs_free_mem(&block_buf);
 
+       return retval;
+
+}
+
+static int expand_inode_table(ext2_filsys fs, unsigned long new_ino_size)
+{
+       dgrp_t i;
+       blk_t blk;
+       errcode_t retval;
+       int new_ino_blks_per_grp;
+       unsigned int j;
+       char *old_itable = NULL, *new_itable = NULL;
+       char *tmp_old_itable = NULL, *tmp_new_itable = NULL;
+       unsigned long old_ino_size;
+       int old_itable_size, new_itable_size;
+
+       old_itable_size = fs->inode_blocks_per_group * fs->blocksize;
+       old_ino_size = EXT2_INODE_SIZE(fs->super);
+
+       new_ino_blks_per_grp = ext2fs_div_ceil(
+                                       EXT2_INODES_PER_GROUP(fs->super) *
+                                       new_ino_size,
+                                       fs->blocksize);
+
+       new_itable_size = new_ino_blks_per_grp * fs->blocksize;
+
+       retval = ext2fs_get_mem(old_itable_size, &old_itable);
+       if (retval)
+               return retval;
+
+       retval = ext2fs_get_mem(new_itable_size, &new_itable);
+       if (retval)
+               goto err_out;
+
+       tmp_old_itable = old_itable;
+       tmp_new_itable = new_itable;
+
+       for (i = 0; i < fs->group_desc_count; i++) {
+
+               blk = fs->group_desc[i].bg_inode_table;
+               retval = io_channel_read_blk(fs->io, blk,
+                               fs->inode_blocks_per_group, old_itable);
+               if (retval)
+                       goto err_out;
+
+               for (j = 0; j < EXT2_INODES_PER_GROUP(fs->super); j++) {
+
+                       memcpy(new_itable, old_itable, old_ino_size);
+
+                       memset(new_itable+old_ino_size, 0,
+                                       new_ino_size - old_ino_size);
+
+                       new_itable += new_ino_size;
+                       old_itable += old_ino_size;
+               }
+
+               /* reset the pointer */
+               old_itable = tmp_old_itable;
+               new_itable = tmp_new_itable;
+
+               retval = io_channel_write_blk(fs->io, blk,
+                                       new_ino_blks_per_grp, new_itable);
+               if (retval)
+                       goto err_out;
+       }
+
+       /* Update the meta data */
+       fs->inode_blocks_per_group = new_ino_blks_per_grp;
+       fs->super->s_inode_size = new_ino_size;
+
+err_out:
+       if (old_itable)
+               ext2fs_free_mem(&old_itable);
+
+       if (new_itable)
+               ext2fs_free_mem(&new_itable);
+
+       return retval;
+}
+
+static errcode_t ext2fs_calculate_summary_stats(ext2_filsys fs)
+{
+       blk_t           blk;
+       ext2_ino_t      ino;
+       unsigned int    group = 0;
+       unsigned int    count = 0;
+       int             total_free = 0;
+       int             group_free = 0;
+
+       /*
+        * First calculate the block statistics
+        */
+       for (blk = fs->super->s_first_data_block;
+            blk < fs->super->s_blocks_count; blk++) {
+               if (!ext2fs_fast_test_block_bitmap(fs->block_map, blk)) {
+                       group_free++;
+                       total_free++;
+               }
+               count++;
+               if ((count == fs->super->s_blocks_per_group) ||
+                   (blk == fs->super->s_blocks_count-1)) {
+                       fs->group_desc[group++].bg_free_blocks_count =
+                               group_free;
+                       count = 0;
+                       group_free = 0;
+               }
+       }
+       fs->super->s_free_blocks_count = total_free;
+
+       /*
+        * Next, calculate the inode statistics
+        */
+       group_free = 0;
+       total_free = 0;
+       count = 0;
+       group = 0;
+
+       /* Protect loop from wrap-around if s_inodes_count maxed */
+       for (ino = 1; ino <= fs->super->s_inodes_count && ino > 0; ino++) {
+               if (!ext2fs_fast_test_inode_bitmap(fs->inode_map, ino)) {
+                       group_free++;
+                       total_free++;
+               }
+               count++;
+               if ((count == fs->super->s_inodes_per_group) ||
+                   (ino == fs->super->s_inodes_count)) {
+                       fs->group_desc[group++].bg_free_inodes_count =
+                               group_free;
+                       count = 0;
+                       group_free = 0;
+               }
+       }
+       fs->super->s_free_inodes_count = total_free;
+       ext2fs_mark_super_dirty(fs);
+       return 0;
+}
+
+#define list_for_each_safe(pos, pnext, head) \
+       for (pos = (head)->next, pnext = pos->next; pos != (head); \
+            pos = pnext, pnext = pos->next)
+
+static void free_blk_move_list(void)
+{
+       struct list_head *entry, *tmp;
+       struct blk_move *bmv;
+
+       list_for_each_safe(entry, tmp, &blk_move_list) {
+
+               bmv = list_entry(entry, struct blk_move, list);
+               list_del(entry);
+               ext2fs_free_mem(&bmv);
+       }
+
+       return ;
+}
+
+static int resize_inode(ext2_filsys fs, unsigned long new_size)
+{
+       errcode_t retval;
+       int new_ino_blks_per_grp;
+       ext2fs_block_bitmap bmap;
+
+       ext2fs_read_inode_bitmap(fs);
+       ext2fs_read_block_bitmap(fs);
+       INIT_LIST_HEAD(&blk_move_list);
+
+
+       new_ino_blks_per_grp = ext2fs_div_ceil(
+                                       EXT2_INODES_PER_GROUP(fs->super)*
+                                       new_size,
+                                       fs->blocksize);
+
+       /* We may change the file system.
+        * Mark the file system as invalid so that
+        * the user is prompted to run fsck.
+        */
+       fs->super->s_state &= ~EXT2_VALID_FS;
+
+       retval = ext2fs_allocate_block_bitmap(fs, _("blocks to be moved"),
+                                               &bmap);
+       if (retval)
+               return retval;
+
+       retval = get_move_bitmap(fs, new_ino_blks_per_grp, bmap);
+       if (retval)
+               goto err_out;
+
+       retval = move_block(fs, bmap);
+       if (retval)
+               goto err_out;
+
+       retval = inode_scan_and_fix(fs, bmap);
+       if (retval)
+               goto err_out;
+
+       retval = expand_inode_table(fs, new_size);
+       if (retval)
+               goto err_out;
+
+       ext2fs_calculate_summary_stats(fs);
+
+       fs->super->s_state |= EXT2_VALID_FS;
+       /* mark super block and block bitmap as dirty */
+       ext2fs_mark_super_dirty(fs);
+       ext2fs_mark_bb_dirty(fs);
+
+err_out:
+       free_blk_move_list();
+       ext2fs_free_block_bitmap(bmap);
+
+       return retval;
+}
+
+static int tune2fs_setup_tdb(const char *name, io_manager *io_ptr)
+{
+       errcode_t retval = 0;
+       const char *tdb_dir;
+       char tdb_file[PATH_MAX];
+       char *dev_name, *tmp_name;
+
+#if 0 /* FIXME!! */
+       /*
+        * Configuration via a conf file would be
+        * nice
+        */
+       profile_get_string(profile, "scratch_files",
+                                       "directory", 0, 0,
+                                       &tdb_dir);
+#endif
+       tmp_name = strdup(name);
+       dev_name = basename(tmp_name);
+
+       tdb_dir = getenv("E2FSPROGS_UNDO_DIR");
+       if (!tdb_dir)
+               tdb_dir="/var/lib/e2fsprogs";
+
+       if (!strcmp(tdb_dir, "none") || (tdb_dir[0] == 0) ||
+           access(tdb_dir, W_OK))
+               return 0;
+
+       sprintf(tdb_file, "%s/tune2fs-%s.e2undo", tdb_dir, dev_name);
+
+       if (!access(tdb_file, F_OK)) {
+               if (unlink(tdb_file) < 0) {
+                       retval = errno;
+                       com_err(program_name, retval,
+                               _("while trying to delete %s"),
+                               tdb_file);
+                       return retval;
+               }
+       }
+
+       set_undo_io_backing_manager(*io_ptr);
+       *io_ptr = undo_io_manager;
+       set_undo_io_backup_file(tdb_file);
+       printf(_("To undo the tune2fs operation please run "
+                "the command\n    e2undo %s %s\n\n"),
+                tdb_file, name);
+       free(tmp_name);
+       return retval;
+}
 
 int main (int argc, char ** argv)
 {
        errcode_t retval;
        ext2_filsys fs;
        struct ext2_super_block *sb;
-       io_manager io_ptr;
+       io_manager io_ptr, io_ptr_orig = NULL;
 
 #ifdef ENABLE_NLS
        setlocale(LC_MESSAGES, "");
@@ -921,14 +1439,17 @@ int main (int argc, char ** argv)
                parse_e2label_options(argc, argv);
        else
                parse_tune2fs_options(argc, argv);
-       
+
 #ifdef CONFIG_TESTIO_DEBUG
-       io_ptr = test_io_manager;
-       test_io_backing_manager = unix_io_manager;
-#else
-       io_ptr = unix_io_manager;
+       if (getenv("TEST_IO_FLAGS") || getenv("TEST_IO_DEBUG")) {
+               io_ptr = test_io_manager;
+               test_io_backing_manager = unix_io_manager;
+       } else
 #endif
-       retval = ext2fs_open2(device_name, io_options, open_flag, 
+               io_ptr = unix_io_manager;
+
+retry_open:
+       retval = ext2fs_open2(device_name, io_options, open_flag,
                              0, 0, io_ptr, &fs);
         if (retval) {
                com_err (program_name, retval, _("while trying to open %s"),
@@ -937,15 +1458,41 @@ int main (int argc, char ** argv)
                        _("Couldn't find valid filesystem superblock.\n"));
                exit(1);
        }
+
+       if (I_flag && !io_ptr_orig) {
+               /*
+                * Check the inode size is right so we can issue an
+                * error message and bail before setting up the tdb
+                * file.
+                */
+               if (new_inode_size == EXT2_INODE_SIZE(fs->super)) {
+                       fprintf(stderr, _("The inode size is already %d\n"),
+                               new_inode_size);
+                       exit(1);
+               }
+               if (new_inode_size < EXT2_INODE_SIZE(fs->super)) {
+                       fprintf(stderr, _("Shrinking the inode size is "
+                                         "not supported\n"));
+                       exit(1);
+               }
+
+               /*
+                * If inode resize is requested use the
+                * Undo I/O manager
+                */
+               io_ptr_orig = io_ptr;
+               retval = tune2fs_setup_tdb(device_name, &io_ptr);
+               if (retval)
+                       exit(1);
+               if (io_ptr != io_ptr_orig) {
+                       ext2fs_close(fs);
+                       goto retry_open;
+               }
+       }
+
        sb = fs->super;
        fs->flags &= ~EXT2_FLAG_MASTER_SB_ONLY;
-       if ((sb->s_feature_incompat & !EXT2_TUNE2FS_INCOMPAT) ||
-           (sb->s_feature_ro_compat & !EXT2_TUNE2FS_RO_COMPAT)) {
-               fprintf(stderr, 
-                       _("Filesystem %s has unsupported features enabled.\n"),
-                       device_name);
-               exit(1);
-       }
+
        if (print_label) {
                /* For e2label emulation */
                printf("%.*s\n", (int) sizeof(sb->s_volume_name),
@@ -953,6 +1500,7 @@ int main (int argc, char ** argv)
                remove_error_table(&et_ext2_error_table);
                exit(0);
        }
+
        retval = ext2fs_check_if_mounted(device_name, &mount_flags);
        if (retval) {
                com_err("ext2fs_check_if_mount", retval,
@@ -990,8 +1538,8 @@ int main (int argc, char ** argv)
                printf (_("Setting interval between checks to %lu seconds\n"), interval);
        }
        if (m_flag) {
-               sb->s_r_blocks_count = e2p_percent(reserved_ratio,
-                                                  sb->s_blocks_count);
+               sb->s_r_blocks_count = (unsigned int) (reserved_ratio *
+                                       sb->s_blocks_count / 100.0);
                ext2fs_mark_super_dirty(fs);
                printf (_("Setting reserved blocks percentage to %g%% (%u blocks)\n"),
                        reserved_ratio, sb->s_r_blocks_count);
@@ -1040,7 +1588,7 @@ int main (int argc, char ** argv)
        }
        if (L_flag) {
                if (strlen(new_label) > sizeof(sb->s_volume_name))
-                       fputs(_("Warning: label too long, truncating.\n"), 
+                       fputs(_("Warning: label too long, truncating.\n"),
                              stderr);
                memset(sb->s_volume_name, 0, sizeof(sb->s_volume_name));
                strncpy(sb->s_volume_name, new_label,
@@ -1061,8 +1609,24 @@ int main (int argc, char ** argv)
                parse_extended_opts(fs, extended_cmd);
        if (journal_size || journal_device)
                add_journal(fs);
-       
+
        if (U_flag) {
+               int set_csum = 0;
+               dgrp_t i;
+
+               if (sb->s_feature_ro_compat &
+                   EXT4_FEATURE_RO_COMPAT_GDT_CSUM) {
+                       /* 
+                        * Determine if the block group checksums are
+                        * correct so we know whether or not to set
+                        * them later on.
+                        */
+                       for (i = 0; i < fs->group_desc_count; i++)
+                               if (!ext2fs_group_desc_csum_verify(fs, i))
+                                       break;
+                       if (i >= fs->group_desc_count)
+                               set_csum = 1;
+               }
                if ((strcasecmp(new_UUID, "null") == 0) ||
                    (strcasecmp(new_UUID, "clear") == 0)) {
                        uuid_clear(sb->s_uuid);
@@ -1074,8 +1638,41 @@ int main (int argc, char ** argv)
                        com_err(program_name, 0, _("Invalid UUID format\n"));
                        exit(1);
                }
+               if (set_csum) {
+                       for (i = 0; i < fs->group_desc_count; i++)
+                               ext2fs_group_desc_csum_set(fs, i);
+                       fs->flags &= ~EXT2_FLAG_SUPER_ONLY;
+               }
                ext2fs_mark_super_dirty(fs);
        }
+       if (I_flag) {
+               if (mount_flags & EXT2_MF_MOUNTED) {
+                       fputs(_("The inode size may only be "
+                               "changed when the filesystem is "
+                               "unmounted.\n"), stderr);
+                       exit(1);
+               }
+               if (fs->super->s_feature_incompat &
+                   EXT4_FEATURE_INCOMPAT_FLEX_BG) {
+                       fputs(_("Changing the inode size not supported for filesystems "
+                               "with the flex_bg\nfeature enabled.\n"),
+                             stderr);
+                       exit(1);
+               }
+               /*
+                * We want to update group descriptor also
+                * with the new free inode count
+                */
+               fs->flags &= ~EXT2_FLAG_SUPER_ONLY;
+               if (resize_inode(fs, new_inode_size)) {
+                       fputs(_("Error in resizing the inode size.\n"
+                               "Run e2undo to undo the "
+                               "file system changes. \n"), stderr);
+               } else {
+                       printf (_("Setting inode size %lu\n"),
+                                                       new_inode_size);
+               }
+       }
 
        if (l_flag)
                list_super (sb);