Re: inode->i_op->rename semantics .

[David Chow <davidchow@shaolinmicro.com>]

Nir Tzachar wrote:

>hi AV.
>
>thank you for ur help, it solved my problems,
>and thanks for clarifying some of the vfs code....
>( i rather wondered what d_move actually does. )
>
>say, isnt it in order to publish some kind of a guide (more thorough than
>the one in Documentation and anything i could find on the web + books... )
>
Yes, I do have some notes on all i_ops, f_ops, a_ops for my own use 
(some still in my head). May be this should be further extend to form a 
book. Since the vfs.txt in the kernel doc written by Richard is pretty 
out-of-date and didn't have any examples. Especially for semantics from 
the Linux VFS. I thought of doing this for 2.4 but struggling for 2.5 as 
it worked a bit different in some cases.

>to the vfs? even explaining execution paths will do....
>i think this kind of document, coming from the maintainer of
>the vfs himself, will save us all much headaches......
>
>  
>
Al, I think you've been answering too many questions that I think it can 
be solved with an update-to-date simple doc. Do you have any plan on 
documenting the VFS api properly from i_ops, f_ops, a_ops and some 
detail explanation of those calls? Maybe something about 2.4 is quite 
good enough for a lot of people. If you have something, please do tell 
me so that I can collect them.

regards,
David Chow