I read about the structure of FAT12.. It has 2 FAT's in it which contain the entry for each cluster.. It also has a Root Directory where it keeps information about each file..

My question is, when we create a new file in that file system, then how do the Root Directory entry and the FAT entries corresponding to the new file get updated.. I mean who updates them..??

The Disk Operating System (DOS) does, but if you are writing the OS, then you would. Pseudo code:

1. Find and empty directory entry to use for the filename and associated attributes
2. Find open locations on the disc via the FAT
3. Update the FATs with corresponding information to where the file is being store

There are some details within that but that's the crux of it.

Ok.. So where can i learn more about it..??
I Googled it, but had no luck..

What does "Open Locations" mean in point 2..??

One more question..
We define a BPB table in the bootloader.. It merely tells about the disk attributes, such as number of sectors per clusters and stuffs.. So how can the system know that we are using a FAT12 file system..?

Here is a good link for a lot of OS developmental items online (there are far more): http://wiki.osdev.org/Main_Page

Try FAT12/FAT16/FAT32/VFAT and you will get tons of hits.

The File Allocation Table(s) (FAT) store file location information and point to a particular location on the disc that stores that part of the file in a sector identified by the address associated with the FAT location (it is a linked list, if I recall correctly). The directory entry points to the first FAT entry. The details you can find in a document I will provide the link for.

This will be explained better within the document here: http://download.microsoft.com/download/1/6/1/161ba512-40e2-4cc9-843a-923143f3456c/fatgen103.doc

Here's a OS Wiki link: http://wiki.osdev.org/FAT

I hope this helps, though be warned, you have a lot of reading to do.

This is one of the best documents that I have found about FAT12:
http://bos.asmhackers.net/docs/filesystems/fat12-16-32/docs/FAT12Description.pdf

I also have many other documents and code examples if you go up the directory structure on the server, like:
http://bos.asmhackers.net/docs/filesystems/fat12-16-32/

The FAT table contains cluster/fat pointers to the next part of any file or directory entries, and you can read in the cluster according to a formula based on the FAT value, but it also serves as a pointer to where in the FAt table the next part is stored.
For any sub-directory you have to look up which clusters to load from the FAT table and then the clusters loaded will have the same entry structure as the root directory itself. So the root is a special case becasue it has a fixed location and size so it's easier to get something from it then to first look up the directory entries for the folder, and then search for files within that.

I'm not sure if i made any sense, but the document that I linked to has it all explained in a good way.

@Smiddy:
OSDev WIki is one of my favorites..
I will read the document you've given..
So what do i need to do to define my own File System..??

@Bubach:
I have read the document..
It was really useful..
I will check the asmhackers..
Thanks..

Knowledge, pen and paper (or ability to write out a plan on a computer, etcetera). Check the same site for file systems. There are so many out there. There was a guy on here who wrote a very simple one, I think, Bogdan? The problem using your own is you run into no one else can read/write to it. BTW, you can think of a File System as a big file with pointers etcetera to locations on a disc. Your canvas, so to speak, is the disc itself, you can do it however you wish.

@Smiddy:
Hehe.. Nice one with the canvas thing..
I will check it out.. And if i get any further problems.. I will post it here..