Okay a FAT File Entry is same as a Directory Entry (FAT16/12 btw)
except that the Attribute Byte (0xB I guess) has bit 4 set
The directory names are different as they don't need an extension,
but are simply padded, some like 'DIR ' instead of a file like
'DIR TXT'. Cool so finding directories is same as finding files,
I have already written a dir_convert which pads the directory names.
The other difference is that the "cluster" which a file points to contains the
contents of file, while a directory's cluster points to the contents of that directory.

;; Change to a Directory.
;; IN : ESI - Name of Directory
;; OUT: CF on Ok, or current_directory.offset set to the new
;; directory
chdir:
        pushad  
        clc
        mov eax, esi
        ;; Convert it to FAT16 Directory filename
        call dir_convert
        mov esi, eax
        xor eax, eax
        ;; Query for existence of directory
        ;; and it's starting cluster
        call entry_exists
        ;; Good it exists
        jc .ok
        ;; Maybe not, clear carry and bailout
        clc
        popad
        ret
.ok:
        ;; AX contains starting cluster, now we need to calculate
        ;; the first sector of the cluster
        ;; Subtract 2 from
        ;; ax
        sub ax, 2
        ;; Save ECX
        push ecx
        xor ecx, ecx
        mov cl, byte [BPB.fat_16_sectors_per_cluster]
        ;; EAX now holds starting sector of cluster
        imul cx
        ;; Add Data Start region to the 
        ;; Formula:
        ;; DataRegion + ((N - 2) * SectorsPerCluster)
        add eax, [BPB.data_start]
        ;; Cool now, EAX contains the starting sector
        ;; of the sub-directory 
        ;; Update Values.
        mov dword [current_directory.offset], eax
        ;; Reload ECX
        pop ecx
        ;; Reload All registers
        popad
        ;; Set Carry Sucessful
        stc
        ;; Return
        ret