From 05647b55512d0fec74fe6d69553a773e852674df Mon Sep 17 00:00:00 2001 From: Denys Vlasenko Date: Tue, 20 Oct 2009 00:06:03 +0200 Subject: [PATCH] mkfs_ext2: compat fixes Signed-off-by: Denys Vlasenko --- util-linux/mkfs_ext2.c | 387 ++++++++++++++++++----------------- util-linux/mkfs_ext2_test.sh | 79 ++++--- 2 files changed, 245 insertions(+), 221 deletions(-) diff --git a/util-linux/mkfs_ext2.c b/util-linux/mkfs_ext2.c index 38654ef04..d0ebc1dd9 100644 --- a/util-linux/mkfs_ext2.c +++ b/util-linux/mkfs_ext2.c @@ -10,6 +10,10 @@ #include "libbb.h" #include #include +#include /* PAGE_SIZE */ +#ifndef PAGE_SIZE +# define PAGE_SIZE 4096 +#endif #include "volume_id/volume_id_internal.h" #define ENABLE_FEATURE_MKFS_EXT2_RESERVED_GDT 0 @@ -26,8 +30,8 @@ //#define LINUX_S_IFWHT 0160000 //#define EXT2_FEATURE_INCOMPAT_WHITEOUT 0x0020 -// storage helper -void BUG_unsupported_field_size(void); +// storage helpers +char BUG_wrong_field_size(void); #define STORE_LE(field, value) \ do { \ if (sizeof(field) == 4) \ @@ -37,9 +41,12 @@ do { \ else if (sizeof(field) == 1) \ field = (value); \ else \ - BUG_unsupported_field_size(); \ + BUG_wrong_field_size(); \ } while (0) +#define FETCH_LE32(field) \ + (sizeof(field) == 4 ? cpu_to_le32(field) : BUG_wrong_field_size()) + // All fields are little-endian struct ext2_dir { uint32_t inode1; @@ -68,10 +75,10 @@ static unsigned int_log2(unsigned arg) } // taken from mkfs_minix.c. libbb candidate? -static unsigned div_roundup(uint32_t size, uint32_t n) +static unsigned div_roundup(uint64_t size, uint32_t n) { // Overflow-resistant - uint32_t res = size / n; + uint64_t res = size / n; if (res * n != size) res++; return res; @@ -83,8 +90,8 @@ static void allocate(uint8_t *bitmap, uint32_t blocksize, uint32_t start, uint32 memset(bitmap, 0, blocksize); i = start / 8; memset(bitmap, 0xFF, i); - bitmap[i] = 0xFF >> (8-(start&7)); -//bb_info_msg("ALLOC: [%u][%u][%u]: [%u]:=[%x]", blocksize, start, end, blocksize - end/8 - 1, (uint8_t)(0xFF << (8-(end&7)))); + bitmap[i] = 0xFF >> (8 - (start & 7)); +//bb_info_msg("ALLOC: [%u][%u][%u]: [%u-%u]:=[%x],[%x]", blocksize, start, end, start/8, blocksize - end/8 - 1, 0xFF >> (8 - (start & 7)), (uint8_t)(0xFF << (8-(end&7)))); i = end / 8; bitmap[blocksize - i - 1] = 0xFF << (8 - (end & 7)); memset(bitmap + blocksize - i, 0xFF, i); // N.B. no overflow here! @@ -167,15 +174,15 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) unsigned blocksize, blocksize_log2; unsigned nreserved = 5; unsigned long long kilobytes; - uint32_t nblocks; + uint32_t nblocks, nblocks_full; uint32_t ngroups; uint32_t bytes_per_inode; uint32_t first_data_block; - uint32_t ninodes_per_group; + uint32_t inodes_per_group; uint32_t gdtsz, itsz; time_t timestamp; unsigned opts; - const char *label; + const char *label = ""; struct stat st; struct ext2_super_block *sb; // superblock struct ext2_group_desc *gd; // group descriptors @@ -245,145 +252,85 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) nblocks = kilobytes; if (nblocks != kilobytes) bb_error_msg_and_die("block count doesn't fit in 32 bits"); +#define kilobytes kilobytes_unused_after_this + + if (blocksize < PAGE_SIZE) + nblocks &= ~((PAGE_SIZE / blocksize)-1); + + // N.B. killing e2fsprogs feature! Unused blocks don't account in calculations + nblocks_full = nblocks; + retry: if (nblocks < 8) bb_error_msg_and_die("need >= 8 blocks"); -#define kilobytes kilobytes_unused_after_this // number of bits in one block, i.e. 8*blocksize #define blocks_per_group (8 * blocksize) -/* e2fsprogs-1.41.9 - overhead = 2 + fs->inode_blocks_per_group; - if (has_super(fs->group_desc_count - 1)) - overhead += 1 + fs->desc_blocks + super->s_reserved_gdt_blocks; - rem = (nblocks - first_data_block) % blocks_per_group); - if ((fs->group_desc_count == 1) && rem && (rem < overhead)) { - retval = EXT2_ET_TOOSMALL; - goto cleanup; - } - if (rem && (rem < overhead+50)) { - nblocks -= rem; - goto retry; - } -*/ - // N.B. a block group can have no more than blocks_per_group blocks first_data_block = (EXT2_MIN_BLOCK_SIZE == blocksize); ngroups = div_roundup(nblocks - first_data_block, blocks_per_group); if (0 == ngroups) bb_error_msg_and_die("ngroups"); - { - // ninodes is the total number of inodes (files) in the file system - uint32_t ninodes = nblocks / (bytes_per_inode >> blocksize_log2); - if (ninodes < EXT2_GOOD_OLD_FIRST_INO+1) - ninodes = EXT2_GOOD_OLD_FIRST_INO+1; - ninodes_per_group = div_roundup(ninodes, ngroups); - // minimum number because the first EXT2_GOOD_OLD_FIRST_INO-1 are reserved - if (ninodes_per_group < 16) - ninodes_per_group = 16; - } - - // TODO: 5?/5 WE MUST NOT DEPEND ON WHETHER DEVICE IS /dev/zero 'ed OR NOT - // TODO: 3/5 refuse if mounted - // TODO: 4/5 compat options - // TODO: 1/5 sanity checks - // TODO: 0/5 more verbose error messages - // TODO: 0/5 info printing - // TODO: 2/5 bigendianness! Spot where it comes to play! sb->, gd-> - // TODO: 2/5 reserved GDT: how to mark but not allocate? - // TODO: 3/5 dir_index? - - // fill the superblock - sb = xzalloc(blocksize); - sb->s_rev_level = 1; // revision 1 filesystem - sb->s_magic = EXT2_SUPER_MAGIC; - sb->s_inode_size = sizeof(*inode); - sb->s_first_ino = EXT2_GOOD_OLD_FIRST_INO; - sb->s_log_block_size = sb->s_log_frag_size = blocksize_log2 - EXT2_MIN_BLOCK_LOG_SIZE; - // first 1024 bytes of the device are for boot record. If block size is 1024 bytes, then - // the first block available for data is 1, otherwise 0 - sb->s_first_data_block = first_data_block; // 0 or 1 - // block and inode bitmaps occupy no more than one block, so maximum number of blocks is - sb->s_blocks_per_group = sb->s_frags_per_group = blocks_per_group; - timestamp = time(NULL); - sb->s_mkfs_time = sb->s_wtime = sb->s_lastcheck = timestamp; - sb->s_state = 1; - sb->s_creator_os = EXT2_OS_LINUX; - sb->s_checkinterval = 24*60*60 * 180; // 180 days - sb->s_errors = EXT2_ERRORS_DEFAULT; - sb->s_feature_compat = EXT2_FEATURE_COMPAT_SUPP - | (EXT2_FEATURE_COMPAT_RESIZE_INO * ENABLE_FEATURE_MKFS_EXT2_RESERVED_GDT) - | (EXT2_FEATURE_COMPAT_DIR_INDEX * ENABLE_FEATURE_MKFS_EXT2_DIR_INDEX) - ; - // e2fsprogs-1.41.9 doesn't like EXT2_FEATURE_INCOMPAT_WHITEOUT - sb->s_feature_incompat = EXT2_FEATURE_INCOMPAT_FILETYPE;// | EXT2_FEATURE_INCOMPAT_WHITEOUT; - sb->s_feature_ro_compat = EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER; - sb->s_flags = EXT2_FLAGS_SIGNED_HASH * ENABLE_FEATURE_MKFS_EXT2_DIR_INDEX; - generate_uuid(sb->s_uuid); - if (ENABLE_FEATURE_MKFS_EXT2_DIR_INDEX) { - sb->s_def_hash_version = EXT2_HASH_HALF_MD4; - generate_uuid((uint8_t *)sb->s_hash_seed); - } - /* - * From e2fsprogs: add "jitter" to the superblock's check interval so that we - * don't check all the filesystems at the same time. We use a - * kludgy hack of using the UUID to derive a random jitter value. - */ - sb->s_max_mnt_count = EXT2_DFL_MAX_MNT_COUNT - + (sb->s_uuid[ARRAY_SIZE(sb->s_uuid)-1] % EXT2_DFL_MAX_MNT_COUNT); - - sb->s_blocks_count = nblocks; - - // reserve blocks for superuser - sb->s_r_blocks_count = ((uint64_t) nblocks * nreserved) / 100; - - gdtsz = div_roundup(ngroups, EXT2_DESC_PER_BLOCK(sb)); - /* - * From e2fsprogs: Calculate the number of GDT blocks to reserve for online - * filesystem growth. - * The absolute maximum number of GDT blocks we can reserve is determined by - * the number of block pointers that can fit into a single block. - */ - /* We set it at 1024x the current filesystem size, or - * the upper block count limit (2^32), whichever is lower. - */ + gdtsz = div_roundup(ngroups, blocksize / sizeof(*gd)); + // TODO: reserved blocks must be marked as such in the bitmaps, + // or resulting filesystem is corrupt if (ENABLE_FEATURE_MKFS_EXT2_RESERVED_GDT) { + /* + * From e2fsprogs: Calculate the number of GDT blocks to reserve for online + * filesystem growth. + * The absolute maximum number of GDT blocks we can reserve is determined by + * the number of block pointers that can fit into a single block. + * We set it at 1024x the current filesystem size, or + * the upper block count limit (2^32), whichever is lower. + */ uint32_t rgdtsz = 0xFFFFFFFF; // maximum block number if (nblocks < rgdtsz / 1024) rgdtsz = nblocks * 1024; rgdtsz = div_roundup(rgdtsz - first_data_block, blocks_per_group); - rgdtsz = div_roundup(rgdtsz, EXT2_DESC_PER_BLOCK(sb)) - gdtsz; - if (rgdtsz > EXT2_ADDR_PER_BLOCK(sb)) - rgdtsz = EXT2_ADDR_PER_BLOCK(sb); - STORE_LE(sb->s_reserved_gdt_blocks, rgdtsz); + rgdtsz = div_roundup(rgdtsz, blocksize / sizeof(*gd)) - gdtsz; + if (rgdtsz > blocksize / sizeof(uint32_t)) + rgdtsz = blocksize / sizeof(uint32_t); + //TODO: STORE_LE(sb->s_reserved_gdt_blocks, rgdtsz); gdtsz += rgdtsz; } - // N.B. a block group can have no more than 8*blocksize inodes - if (ninodes_per_group > blocks_per_group) - ninodes_per_group = blocks_per_group; - // adjust inodes per group so they completely fill the inode table blocks in the descriptor - ninodes_per_group = (div_roundup(ninodes_per_group * EXT2_INODE_SIZE(sb), blocksize) << blocksize_log2) / EXT2_INODE_SIZE(sb); - // make sure the number of inodes per group is a multiple of 8 - ninodes_per_group &= ~7; - sb->s_inodes_per_group = ninodes_per_group;// = div_roundup(ninodes_per_group * sb->s_inode_size, blocksize); - // total ninodes - sb->s_inodes_count = ninodes_per_group * ngroups; - - itsz = ninodes_per_group * sb->s_inode_size / blocksize; - sb->s_free_inodes_count = sb->s_inodes_count - EXT2_GOOD_OLD_FIRST_INO; - - // write the label, if any - if (opts & OPT_L) - safe_strncpy((char *)sb->s_volume_name, label, sizeof(sb->s_volume_name)); - -#if 1 -/* if (fs_param.s_blocks_count != s->s_blocks_count) - fprintf(stderr, _("warning: %u blocks unused.\n\n"), - fs_param.s_blocks_count - s->s_blocks_count); -*/ + { + // N.B. e2fsprogs does as follows! + // ninodes is the total number of inodes (files) in the file system + uint32_t ninodes = nblocks_full / (blocksize >= 4096 ? 1 : 4096 / blocksize); + uint32_t overhead, remainder; + if (ninodes < EXT2_GOOD_OLD_FIRST_INO+1) + ninodes = EXT2_GOOD_OLD_FIRST_INO+1; + inodes_per_group = div_roundup(ninodes, ngroups); + // minimum number because the first EXT2_GOOD_OLD_FIRST_INO-1 are reserved + if (inodes_per_group < 16) + inodes_per_group = 16; + + // N.B. a block group can have no more than 8*blocksize inodes + if (inodes_per_group > blocks_per_group) + inodes_per_group = blocks_per_group; + // adjust inodes per group so they completely fill the inode table blocks in the descriptor + inodes_per_group = (div_roundup(inodes_per_group * sizeof(*inode), blocksize) << blocksize_log2) / sizeof(*inode); + // make sure the number of inodes per group is a multiple of 8 + inodes_per_group &= ~7; + itsz = div_roundup(inodes_per_group * sizeof(*inode), blocksize); + + // the last block needs more attention: doesn't it too small for possible overhead? + overhead = (has_super(ngroups - 1) ? (1/*sb*/ + gdtsz) : 0) + 1/*bbmp*/ + 1/*ibmp*/ + itsz; + remainder = (nblocks - first_data_block) % blocks_per_group; + if ((1 == ngroups) && remainder && (remainder < overhead)) + bb_error_msg_and_die("way small device"); + if (remainder && (remainder < overhead + 50)) { +//bb_info_msg("CHOP[%u]", remainder); + nblocks -= remainder; + goto retry; + } + } + // print info + if (nblocks_full - nblocks) + printf("warning: %u blocks unused\n\n", nblocks_full - nblocks); printf( "Filesystem label=%s\n" "OS type: Linux\n" @@ -396,16 +343,17 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) "%u block groups\n" "%u blocks per group, %u fragments per group\n" "%u inodes per group" - , (char *)sb->s_volume_name - , blocksize, sb->s_log_block_size - , blocksize, sb->s_log_block_size - , sb->s_inodes_count, sb->s_blocks_count - , sb->s_r_blocks_count, nreserved + , label + , blocksize, blocksize_log2 - EXT2_MIN_BLOCK_LOG_SIZE + , blocksize, blocksize_log2 - EXT2_MIN_BLOCK_LOG_SIZE + , inodes_per_group * ngroups, nblocks + //, div_roundup((uint64_t) nblocks * nreserved, 100), nreserved + , (unsigned)((uint64_t) nblocks_full * nreserved / 100), nreserved , first_data_block - , gdtsz * EXT2_DESC_PER_BLOCK(sb) * blocks_per_group + , gdtsz * (blocksize / sizeof(*gd)) * blocks_per_group , ngroups , blocks_per_group, blocks_per_group - , ninodes_per_group + , inodes_per_group ); { const char *fmt = "\nSuperblock backups stored on blocks:\n" @@ -420,84 +368,149 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) } } bb_putchar('\n'); -#endif + // dry run? -> we are done if (opts & OPT_n) goto done; + // TODO: 3/5 refuse if mounted + // TODO: 4/5 compat options + // TODO: 1/5 sanity checks + // TODO: 0/5 more verbose error messages + // TODO: 4/5 bigendianness: recheck, wait for ARM reporters + // TODO: 2/5 reserved GDT: how to mark but not allocate? + // TODO: 3/5 dir_index? + + // fill the superblock + sb = xzalloc(blocksize); + STORE_LE(sb->s_rev_level, 1); // revision 1 filesystem + STORE_LE(sb->s_magic, EXT2_SUPER_MAGIC); + STORE_LE(sb->s_inode_size, sizeof(*inode)); + STORE_LE(sb->s_first_ino, EXT2_GOOD_OLD_FIRST_INO); + STORE_LE(sb->s_log_block_size, blocksize_log2 - EXT2_MIN_BLOCK_LOG_SIZE); + STORE_LE(sb->s_log_frag_size, blocksize_log2 - EXT2_MIN_BLOCK_LOG_SIZE); + // first 1024 bytes of the device are for boot record. If block size is 1024 bytes, then + // the first block available for data is 1, otherwise 0 + STORE_LE(sb->s_first_data_block, first_data_block); // 0 or 1 + // block and inode bitmaps occupy no more than one block, so maximum number of blocks is + STORE_LE(sb->s_blocks_per_group, blocks_per_group); + STORE_LE(sb->s_frags_per_group, blocks_per_group); + // blocks + STORE_LE(sb->s_blocks_count, nblocks); + // reserve blocks for superuser + STORE_LE(sb->s_r_blocks_count, (uint32_t)((uint64_t) nblocks_full * nreserved / 100)); + // ninodes + STORE_LE(sb->s_inodes_per_group, inodes_per_group); + STORE_LE(sb->s_inodes_count, inodes_per_group * ngroups); + STORE_LE(sb->s_free_inodes_count, inodes_per_group * ngroups - EXT2_GOOD_OLD_FIRST_INO); + // timestamps + timestamp = time(NULL); + STORE_LE(sb->s_mkfs_time, timestamp); + STORE_LE(sb->s_wtime, timestamp); + STORE_LE(sb->s_lastcheck, timestamp); + // misc + STORE_LE(sb->s_state, 1); // TODO: what's 1? + STORE_LE(sb->s_creator_os, EXT2_OS_LINUX); + STORE_LE(sb->s_checkinterval, 24*60*60 * 180); // 180 days + STORE_LE(sb->s_errors, EXT2_ERRORS_DEFAULT); + STORE_LE(sb->s_feature_compat, EXT2_FEATURE_COMPAT_SUPP + | (EXT2_FEATURE_COMPAT_RESIZE_INO * ENABLE_FEATURE_MKFS_EXT2_RESERVED_GDT) + | (EXT2_FEATURE_COMPAT_DIR_INDEX * ENABLE_FEATURE_MKFS_EXT2_DIR_INDEX) + ); + // e2fsprogs-1.41.9 doesn't like EXT2_FEATURE_INCOMPAT_WHITEOUT + STORE_LE(sb->s_feature_incompat, EXT2_FEATURE_INCOMPAT_FILETYPE);// | EXT2_FEATURE_INCOMPAT_WHITEOUT; + STORE_LE(sb->s_feature_ro_compat, EXT2_FEATURE_RO_COMPAT_SPARSE_SUPER); + STORE_LE(sb->s_flags, EXT2_FLAGS_SIGNED_HASH * ENABLE_FEATURE_MKFS_EXT2_DIR_INDEX); + generate_uuid(sb->s_uuid); + if (ENABLE_FEATURE_MKFS_EXT2_DIR_INDEX) { + STORE_LE(sb->s_def_hash_version, EXT2_HASH_HALF_MD4); + generate_uuid((uint8_t *)sb->s_hash_seed); + } + /* + * From e2fsprogs: add "jitter" to the superblock's check interval so that we + * don't check all the filesystems at the same time. We use a + * kludgy hack of using the UUID to derive a random jitter value. + */ + STORE_LE(sb->s_max_mnt_count, EXT2_DFL_MAX_MNT_COUNT + + (sb->s_uuid[ARRAY_SIZE(sb->s_uuid)-1] % EXT2_DFL_MAX_MNT_COUNT)); + + // write the label, if any + if (label) //opts & OPT_L) + safe_strncpy((char *)sb->s_volume_name, label, sizeof(sb->s_volume_name)); + // fill group descriptors gd = xzalloc(gdtsz * blocksize); + buf = xmalloc(blocksize); sb->s_free_blocks_count = 0; - for (i = 0, pos = first_data_block, n = nblocks; + for (i = 0, pos = first_data_block, n = nblocks - first_data_block; i < ngroups; i++, pos += blocks_per_group, n -= blocks_per_group ) { uint32_t overhead = pos + (has_super(i) ? (1/*sb*/ + gdtsz) : 0); - gd[i].bg_block_bitmap = overhead + 0; - gd[i].bg_inode_bitmap = overhead + 1; - gd[i].bg_inode_table = overhead + 2; + uint32_t fb; + STORE_LE(gd[i].bg_block_bitmap, overhead + 0); + STORE_LE(gd[i].bg_inode_bitmap, overhead + 1); + STORE_LE(gd[i].bg_inode_table, overhead + 2); overhead = overhead - pos + 1/*bbmp*/ + 1/*ibmp*/ + itsz; - gd[i].bg_free_inodes_count = ninodes_per_group; - //gd[i].bg_used_dirs_count = 0; + gd[i].bg_free_inodes_count = inodes_per_group; + //STORE_LE(gd[i].bg_used_dirs_count, 0); // N.B. both root and lost+found dirs are within the first block group, thus +2 if (0 == i) { overhead += 2; - gd[i].bg_used_dirs_count = 2; + STORE_LE(gd[i].bg_used_dirs_count, 2); gd[i].bg_free_inodes_count -= EXT2_GOOD_OLD_FIRST_INO; } - // N.B. the following is pure heuristics! - // Likely to cope with 1024-byte blocks, when first block is for boot sectors - if (ngroups-1 == i) { - overhead += first_data_block; - } - gd[i].bg_free_blocks_count = (n < blocks_per_group ? n : blocks_per_group) - overhead; - sb->s_free_blocks_count += gd[i].bg_free_blocks_count; +// // N.B. the following is pure heuristics! +// // Likely to cope with 1024-byte blocks, when first block is for boot sectors +// if (ngroups-1 == i) { +// n -= first_data_block; +// } + + // mark preallocated blocks as allocated + fb = (n < blocks_per_group ? n : blocks_per_group) - overhead; +//bb_info_msg("ALLOC: [%u][%u][%u]", blocksize, overhead, blocks_per_group - (fb + overhead)); + allocate(buf, blocksize, + overhead, + blocks_per_group - (fb + overhead) + ); + // dump block bitmap + PUT((uint64_t)(FETCH_LE32(gd[i].bg_block_bitmap)) * blocksize, buf, blocksize); + STORE_LE(gd[i].bg_free_blocks_count, fb); + + // mark preallocated inodes as allocated + allocate(buf, blocksize, + inodes_per_group - gd[i].bg_free_inodes_count, + blocks_per_group - inodes_per_group + ); + // dump inode bitmap + //PUT((uint64_t)(FETCH_LE32(gd[i].bg_block_bitmap)) * blocksize, buf, blocksize); + //but it's right after block bitmap, so we can just: + xwrite(fd, buf, blocksize); + STORE_LE(gd[i].bg_free_inodes_count, gd[i].bg_free_inodes_count); + + // count overall free blocks + sb->s_free_blocks_count += fb; } STORE_LE(sb->s_free_blocks_count, sb->s_free_blocks_count); // dump filesystem skeleton structures // printf("Writing superblocks and filesystem accounting information: "); - buf = xmalloc(blocksize); for (i = 0, pos = first_data_block; i < ngroups; i++, pos += blocks_per_group) { - uint32_t overhead = has_super(i) ? (1/*sb*/ + gdtsz) : 0; - uint32_t start; - uint32_t end; - // dump superblock and group descriptors and their backups - if (overhead) { // N.B. in fact, we want (has_super(i)) condition, but it is equal to (overhead != 0) and is cheaper + if (has_super(i)) { // N.B. 1024 byte blocks are special PUT(((uint64_t)pos << blocksize_log2) + ((0 == i && 0 == first_data_block) ? 1024 : 0), sb, 1024);//blocksize); PUT(((uint64_t)pos << blocksize_log2) + blocksize, gd, gdtsz * blocksize); } - - start = overhead + 1/*bbmp*/ + 1/*ibmp*/ + itsz; - if (i == 0) - start += 2; // for "/" and "/lost+found" - end = blocks_per_group - (start + gd[i].bg_free_blocks_count); - - // mark preallocated blocks as allocated - allocate(buf, blocksize, start, end); - // dump block bitmap - PUT((uint64_t)(pos + overhead) * blocksize, buf, blocksize); - - // mark preallocated inodes as allocated - allocate(buf, blocksize, - ninodes_per_group - gd[i].bg_free_inodes_count, - blocks_per_group - ninodes_per_group - ); - // dump inode bitmap - //PUT((uint64_t)(pos + overhead + 1) * blocksize, buf, blocksize); - //but it's right after block bitmap, so we can just: - xwrite(fd, buf, blocksize); } // zero boot sectors memset(buf, 0, blocksize); - PUT(0, buf, 1024); // N.B. 1024 <= blocksize + PUT(0, buf, 1024); // N.B. 1024 <= blocksize, so buf[0..1023] contains zeros // zero inode tables for (i = 0; i < ngroups; ++i) for (n = 0; n < itsz; ++n) - PUT((uint64_t)(gd[i].bg_inode_table + n) * blocksize, buf, blocksize); + PUT((uint64_t)(FETCH_LE32(gd[i].bg_inode_table) + n) * blocksize, buf, blocksize); // prepare directory inode inode = (struct ext2_inode *)buf; @@ -511,13 +524,13 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) // dump root dir inode STORE_LE(inode->i_links_count, 3); // "/.", "/..", "/lost+found/.." point to this inode - STORE_LE(inode->i_block[0], gd[0].bg_inode_table + itsz); - PUT(((uint64_t)gd[0].bg_inode_table << blocksize_log2) + (EXT2_ROOT_INO-1) * sizeof(*inode), buf, sizeof(*inode)); + STORE_LE(inode->i_block[0], FETCH_LE32(gd[0].bg_inode_table) + itsz); + PUT(((uint64_t)FETCH_LE32(gd[0].bg_inode_table) << blocksize_log2) + (EXT2_ROOT_INO-1) * sizeof(*inode), buf, sizeof(*inode)); // dump lost+found dir inode STORE_LE(inode->i_links_count, 2); // both "/lost+found" and "/lost+found/." point to this inode - STORE_LE(inode->i_block[0], inode->i_block[0]+1); // use next block //= gd[0].bg_inode_table + itsz + 1; - PUT(((uint64_t)gd[0].bg_inode_table << blocksize_log2) + (EXT2_GOOD_OLD_FIRST_INO-1) * sizeof(*inode), buf, sizeof(*inode)); + STORE_LE(inode->i_block[0], inode->i_block[0]+1); // use next block + PUT(((uint64_t)FETCH_LE32(gd[0].bg_inode_table) << blocksize_log2) + (EXT2_GOOD_OLD_FIRST_INO-1) * sizeof(*inode), buf, sizeof(*inode)); // dump directories memset(buf, 0, blocksize); @@ -534,7 +547,7 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) STORE_LE(dir->name_len2, 2); STORE_LE(dir->file_type2, EXT2_FT_DIR); dir->name2[0] = '.'; dir->name2[1] = '.'; - PUT((uint64_t)(gd[0].bg_inode_table + itsz + 1) * blocksize, buf, blocksize); + PUT((uint64_t)(FETCH_LE32(gd[0].bg_inode_table) + itsz + 1) * blocksize, buf, blocksize); // dump root dir block STORE_LE(dir->inode1, EXT2_ROOT_INO); @@ -544,13 +557,7 @@ int mkfs_ext2_main(int argc UNUSED_PARAM, char **argv) STORE_LE(dir->name_len3, 10); STORE_LE(dir->file_type3, EXT2_FT_DIR); strcpy(dir->name3, "lost+found"); - PUT((uint64_t)(gd[0].bg_inode_table + itsz + 0) * blocksize, buf, blocksize); - -// bb_info_msg("done\n" -// "This filesystem will be automatically checked every %u mounts or\n" -// "180 days, whichever comes first. Use tune2fs -c or -i to override.", -// sb->s_max_mnt_count -// ); + PUT((uint64_t)(FETCH_LE32(gd[0].bg_inode_table) + itsz + 0) * blocksize, buf, blocksize); done: // cleanup diff --git a/util-linux/mkfs_ext2_test.sh b/util-linux/mkfs_ext2_test.sh index fe6291fc5..9c20b3b89 100755 --- a/util-linux/mkfs_ext2_test.sh +++ b/util-linux/mkfs_ext2_test.sh @@ -1,54 +1,71 @@ #!/bin/sh -test_mke2fs() { - echo Testing $kilobytes - - >image_std - dd seek=$((kilobytes-1)) bs=1K count=1 /dev/null 2>&1 || exit 1 - /usr/bin/mke2fs -F image_std $kilobytes >image_std.raw_out 2>&1 || return 1 - cat image_std.raw_out \ - | grep -v '^mke2fs ' \ +run_test() { # params: mke2fs_invocation image_name + >$2 + dd seek=$((kilobytes-1)) bs=1K count=1 /dev/null 2>&1 || exit 1 + $1 -F $2 $kilobytes >$2.raw_out 2>&1 || return 1 + cat $2.raw_out \ + | grep -v '^mke2fs [0-9]*\.[0-9]*\.[0-9]* ' \ | grep -v '^Maximum filesystem' \ - | grep -v '^warning: .* blocks unused' \ | grep -v '^Writing inode tables' \ | grep -v '^Writing superblocks and filesystem accounting information' \ | grep -v '^This filesystem will be automatically checked every' \ | grep -v '^180 days, whichever comes first' \ - | sed 's/block groups/block group/' \ + | sed 's/inodes, [0-9]* blocks/inodes, N blocks/' \ + | sed 's/blocks* unused./blocks unused/' \ + | sed 's/block groups*/block groups/' \ | sed 's/ *$//' \ | sed 's/blocks (.*%) reserved/blocks reserved/' \ | grep -v '^$' \ - >image_std.out + >$2.out +} - >image_bb - dd seek=$((kilobytes-1)) bs=1K count=1 /dev/null 2>&1 || exit 1 - ./busybox mke2fs -F image_bb $kilobytes >image_bb.raw_out 2>&1 || return 1 - cat image_bb.raw_out \ - | grep -v '^mke2fs ' \ - | grep -v '^Maximum filesystem' \ - | grep -v '^warning: .* blocks unused' \ - | grep -v '^Writing inode tables' \ - | grep -v '^Writing superblocks and filesystem accounting information' \ - | grep -v '^This filesystem will be automatically checked every' \ - | grep -v '^180 days, whichever comes first' \ - | sed 's/block groups/block group/' \ - | sed 's/ *$//' \ - | sed 's/blocks (.*%) reserved/blocks reserved/' \ - | grep -v '^$' \ - >image_bb.out +test_mke2fs() { + echo Testing $kilobytes + + run_test '/usr/bin/mke2fs' image_std || return 1 + run_test './busybox mke2fs' image_bb || return 1 diff -ua image_bb.out image_std.out >image.out.diff || { cat image.out.diff return 1 } + + e2fsck -f -n image_bb >/dev/null 2>&1 || { echo "e2fsck error on image_bb"; exit 1; } } +# -:bbox +:standard + +# -6240 inodes, 24908 blocks +# +6240 inodes, 24577 blocks +# -4 block group +# +3 block group +# -1560 inodes per group +# +2080 inodes per group kilobytes=24908 test_mke2fs -kilobytes=81940 test_mke2fs -kilobytes=98392 test_mke2fs -exit + +# -304 inodes, N blocks +# +152 inodes, N blocks +# -304 inodes per group +# +152 inodes per group +kilobytes=1218 test_mke2fs + +# -14464 inodes, N blocks +# +14448 inodes, N blocks +# -8 block group +# +7 block group +# -1808 inodes per group +# +2064 inodes per group +kilobytes=57696 test_mke2fs + +# This size results in "warning: 75 blocks unused" +kilobytes=98380 test_mke2fs + +# -warning: 239 blocks unused. +# +warning: 242 blocks unused. +kilobytes=49395 test_mke2fs while true; do - kilobytes=$(( (RANDOM*RANDOM) % 100000 + 100)) + kilobytes=$(( (RANDOM*RANDOM) % 1000000 + 2000)) test_mke2fs || exit 1 done -- 2.25.1