Irregular Expressions » super block

The EXT file system – Part 3

Dan O'Connor — Wed, 24 Aug 2011 02:12:19 +0000

We also have to deal with the block pointers in the inodes. Each inode can store the address of 12 blocks to store the data. So what happens when you need more then 12 blocks to store the data?

You use indirect block pointers, instead of pointing at blocks of data. You point at blocks that point to the blocks. Basically lists of lists.

The first layer of this is called single indirect block pointers, the lists of lists. Then we go to double indirect block pointers. Lists of lists of lists. Then down to triple indirect block pointers, lists of lists of lists of lists.

The EXT file system – Part 2

Dan O'Connor — Sat, 20 Aug 2011 03:01:48 +0000

The still short version of what happens when you need locate a file on the system is the following;

The SuperBlock is read, to collect information about the file system.
Next to the Group Descriptor Table (After the SuperBlock on the disk), collection of information about the block groups.
Once the block group that the file resides in is determined, the group descriptor table is used to locate where the Inode Table is.
The Inode Table will point us to the block root directory entries.
That is opened to locate the folder that contains the file we are looking for.
Now we have the inode of the folder, we then determine the block group it belongs to (we use information from the SuperBlock for that).
Then we go to that block group and get the location of the folder.
Now we can open the block the contains the folder.

And this will keep going until the location of the file is found depending on the number of sub folders, then the information will be pulled out of the blocks that contain the information.

The EXT file system – Part 1

Dan O'Connor — Wed, 17 Aug 2011 04:34:44 +0000

Partitions are just fine, but how does the system get to a file on the disk?

On a EXT based system, the short answer is the information is stored in the Inode Table.

That really does not mean much if you don’t have understanding of what the filesystem looks like.

A little explanation is needed.

The first section of the filesystem is going to contain the Super Block (block 1), there could be boot code in front of that (Unused). The Super block contains information about the file system, such as block sizes, total number of inodes and the volume name to list a few.

Following the Super Block is the Group Descriptor Tables, this is what we need. This will have a backup of the Super Block, and contains the Block Bitmap, Inode Bitmap and group descriptor data structure for every block in group in the filesystem. These are the basic structures we need to locate information on the disk.