File System linux

impulse

Hi, i've task to make linux system file simulation, i want to ask 2 question:

1. It's said that in linux have inode size and another file system attribute(superblock size, free space mgmt,etc) depend on partition size, i want to ask that in linux,

given the size of linux os partition, say 90Gb,

how's calculation of inode table size, block in inode and inode level(level 1 or 2 or 3) is implemented??

and i want a reasonable answer why that calculation is implemented according to time and space performance.

2. Same as question 1, but this is about size of block in system file, it's said that 4kb block size is the best of all block size(512,etc). i don't get it why?? anyone can help me??

Thanks....

marc

1. It's said that in linux have inode size and another file system attribute(superblock size, free space mgmt,etc) depend on partition size, i want to ask that in linux,
given the size of linux os partition, say 90Gb,
how's calculation of inode table size, block in inode and inode level(level 1 or 2 or 3) is implemented??
and i want a reasonable answer why that calculation is implemented according to time and space performance.

I'm sorry but I don't understand you there :S

2. Same as question 1, but this is about size of block in system file, it's said that 4kb block size is the best of all block size(512,etc). i don't get it why?? anyone can help me??

IIRC the 4kb block size is usually better because that's what the most common physical size in commercial hard disks is. That doesn't mean it's always the best

Regards

CORRECTION: I was wrong

Generation one Advanced Format, 4K sector technology, utilizes the storage surface media more efficiently by combining eight 512-byte sectors into one single sector that is 4096-bytes in length. Key design elements of the traditional 512-byte sector architecture are maintained, specifically, the identification and synchronization marks at the beginning and the error correction coding (ECC) area at the end of the sector. Between the sector header and ECC areas, eight 512 byte sectors are combined, eliminating the need for redundant header areas between each individual chunk of 512-byte data. The Long Data Sector Committee selected the 4K block length for the first generation AF standard for several reasons, including its correspondence to the paging size used by processors and some operating systems as well as its correlation to the size of standard transactions in relational data base systems.[4]

Extracted from http://en.wikipedia.org/wiki/Advanced_Format