Doubt in Creation of EXT2 filesystem on floppy disk...

[Zishan Shaikh]

Hi all, i am Zishan Shaikh, BE COMP, working on the deduplication project

i have attached three files, the first file shows the default
assumptions made while creating the system..please open them first and
help me out.

My doubt is in d second and third file...
(second file ) firstly i did not understand the calculations made in
the paragraph just above the table 18-7 ind second file, i mean how
did the number 184 inodes in 23 blocks come??? did not understand from
where did 23 blocks come...
i mean it says, "1 inode for each 8192 bytes" => 1 inode for 8 blocks
(since 1 block=1024 bytes assumption in 1st file)...
then how did 184 inodes for 23 blocks appear??? confused over this.

my second doubt is : the table in18-7 in the 2nd file and continued in
the third file...i mean i did not get how were those number of blocks
allocated to specific fields...blocks 5-27 have been allocated for
inode table...(23 blocks again, my basic doubt)...

Please help...

Zishan Shaikh
BE Computer, MAE

[Mahesh Mohan]

Hi Zishan,

Firstly, if you read carefully, it is given that 184 inodes are stored IN 23 blocks and not for 23 blocks. The statistic that there should be 1 inode for 8192 bytes or 8 blocks is just the inode to no., of blocks ratio. 

Secondly, the leftmost column in table 18-7 does not specify the number of blocks allocated to certain fields. They are the block numbers where those objects in the other column reside i.  the boot block structure is stored in block number 0, superblock is stored in block number 1 and so on. 

AFAIK the block numbers where the boot block, superblock etc should reside are pre-decided so I don't see a particular reason there. You can try to find out more if you'd like.....

The inode table contains 23 blocks. So all the 184 inodes (mind you, its 184 in this case and not in all cases) will be stored in these 23 blocks. These include the free inodes. So, basically if the disk size is more, the no. of inodes will also be more and more number of blocks may be required to store those inodes. 

Hope it is clear to you now.....

Regards,

Mahesh  

P.S: By the way, in which filesystem are you planning to implement deduplication ?

[Zishan Shaikh]

thanx....that clears all my doubts related to that problem!!!!

n we r implementing deduplication on ext2 first, thn modify d code for ext4....

but first of all we are going to study what d previous group has done,
and then extend it further....
so first study f previous code, then ext2 first n then modifications
for ext4....

[Mahesh Mohan]

Is that the plan that More sir has suggested ? Would it not suffice to have a detailed study of the design of the previous code and then implement the same for ext 4 directly instead of first implementing it in ext2 and then again in ext4 ? Basically what I feel is that the more you are able to study ext4 in detail, the better it is.....just a suggestion though......think about it....or maybe Kashish n Paaras can guide you more on this.....

Regards,

Mahesh

[Zishan Shaikh]

for got to mention,
Zishan Shaikh
BE Comp.

[Zishan Shaikh]

Hi mahesh,

yes that is what Amar sir had suggested.....n we too had considered d
same...but then what u say is also a good idea, coz that will make us
study n understand ext4 in details jst lyk we studied ext2....will
surely talk to othr members, n sir too abt ths....thanx for tht!

@kashish @paras @gaurav:
guys if ur reading ths please u ppl too let us knw wht u think abt
working on ext4 directly....

Zishan Shaikh
BE Comp, MAE.

[Gaurav Mahajan]

On Tue, Aug 9, 2011 at 9:21 AM, Zishan Shaikh <zishan3...@gmail.com> wrote:
> Hi mahesh,
>
> yes that is what Amar sir had suggested.....n we too had considered d
> same...but then what u say is also a good idea, coz that will make us
> study n understand ext4 in details jst lyk we studied ext2....will
> surely talk to othr members, n sir too abt ths....thanx for tht!
>
> @kashish @paras @gaurav:
> guys if ur reading ths please u ppl too let us knw wht u think abt
> working on ext4 directly....

Hi Zishan

It would be good idea to study the EXT4 instate....of implementing it
in EXT2 and then port to EXT4.
I think Paras and Kashish group have faced the biggest problem with
this concern. Everybody asked them about why they have used EXT2 and
not EXT4. But unless you study EXT2 &3 in detail you will not
understand ext4. Because they cover a lot of basics........

Regards,
Gaurav Mahajan.

[KASHISH BHATIA]

yes, Gaurav is correct .

Rather than working on ext4 directly , you should study ext2 and ext3.

 

And if you understand the previous design and code , it would be very easy for you all

to implement it in ext4 because ext4 also provide similar functions as in ext2.

 

 

 

--
Regards,

Kashish Bhatia

[Zishan Shaikh]

thanx for tht help guys!!!

Zishan Shaikh
BE COMP, MAE