Hi, On Thu, 2006-04-27 at 19:46 +0200, Niki Hammler wrote: > I have got a question concerning directory entries. I have the following > block containing exactly the filenames I had in one specified folder on > the same file system: > > http://www.sbox.tugraz.at/home/n/nobaq/ext2.dat > > I really hoped that this is an directory block which could point me to > the inode of the files. Yes, it is. > But when I try to extract the data, I only get garbage. I'm reading the > block this way: First 4 bytes are pointer to inode, second 4 bytes are > length of the name and the the rest is the name itself and so on. Not quite. It's an ext2_dir_entry_2 struct from linux/include/linux/ext2_fs.h : struct ext2_dir_entry_2 { __le32 inode; /* Inode number */ __le16 rec_len; /* Directory entry length */ __u8 name_len; /* Name length */ __u8 file_type; char name[EXT2_NAME_LEN]; /* File name */ }; so yes, the first 4 bytes are the inode number; but then you've got a 2- byte record length, which includes the 8 byte directory entry struct plus the name length rounded up to the next 4 bytes (to keep the entries 4-byte aligned on disk); then the name length itself, and the inode type, both of them just 1 byte long. > The first two entries should be '.' and '..', so the name lengths should > be only 1 and 2, shouldn't they? They are: looking at the "hexdump -C" of the data, I see 00000000 01 40 01 00 0c 00 01 02 2e 00 00 00 b6 c1 08 00 |.@..............| 00000010 0c 00 02 02 2e 2e 00 00 02 40 01 00 14 00 09 01 |.........@......| so you've got inode number 0x00014001, then 0x000c = 12 bytes record length, then a 1-byte name and file_type 2, EXT2_FT_DIR; then "." for the name. That completes the first record. Then you have inode 0x0008c1b6, record length 12, name length 2 and file_type 2, for the name "..". And so on. Cheers, Stephen _______________________________________________ Ext3-users@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/ext3-users
