To answer the first half of your questions, this is a compile time option only. Or that is my understanding of it. To enable it, you would need to modify the .config file of the kernel source that you are either recompiling or obtaining from the kernel.org website.

When you navigate into the kernel's directory (usually distributions place their kernel source in /usr/src) you can change the .config file manually (if you know what you are doing) or by using menuconfig:

$ make menuconfig

Actually, one of the best books I have come across dealing with the Linux kernel is Robert Love's "Linux Kernel Development" which I do not know if it is the same Linux Kernel Development in your listing. Love presents the material in a very organized manner to which it becomes easier to follow.

It would seem that for the most part your questions have been answered, but to add a little bit more to your second question:

>>another point , how do i identify partitions and internal devices to mount and umount them?
>>.All i know is they are inside /dev

While the fdisk utility will show you the partition table to a specific device, sometimes it is simpler to type:

$ cat /proc/partitions

In this file you will get a real-time display of all currently connected block devices and the partitions available to each. In the rarest of cases a device may not necessarily automount. You can monitor the device's connection in this file and when listed use the "sudo mount" command to mount the device to a desired path.

Note that the listings in /proc/partitions should be preceded with the /dev/ string when invoking the mount command. That is, if you see a listing of sdb1 showing up, then you would need to type:

$ sudo mount /dev/sdb1 <destination path>