executing vfs_write on remote machine, which other calls should be hooked?

saumitra · 04-04-2007, 01:49 AM

hi,

Im working on a project, where i'm having two same copies of same files on two remote machines.

i'm tracking the changes in files from vfs_write in read_write.c

i hooked vfs_write and took its all parameters like buf, count, position...

opened _related file structure_ on remote machine and called vfs_write ON REMOTE MACHINE WITH THESE TRANSFERED PARAMETES...

so now, if i APPEND something to one file on machine A, same append is done on machine B!!

e.g.
if on machine A,
a.txt is "abcd"
then

if i do #echo "e" >> a.txt

vfs_write calls are called on both machies (for their file structure of a.txt ) with same parameters (position buffer etc)

its working fine for append..

NOW, i want to keep file on machine B exactly as on machine A at any point of time.. by carrying out all required calls.. WHICH ALL CALLS SHOULD I HOOK?????

PLEASE HELP!!

details of problem faced:

a.txt is "abcdefg"

i call #echo "XYZ" > a.txt NOTE: NO APPENT.. truncate is done..

RESULT :

machine A : a.txt = "XYZ"
machine B : a.txt = "XYZdefg"

please tell me which all other call i should hook and execute remotely ASAP.. thanks

theNbomr · 04-05-2007, 04:36 PM

I've only done a little bit of VFS stuff, using fuse, so maybe I'm way off base. But no one else has jumped in yet, so...
Isn't there actually a whole series of file_operations and/or inode_operations performed to create a new file? And don't you have to implement the functionality of all of those operations on both peers to faithfully track all changes? As I understand it, you would have to implement pretty well all of the functions in

Code:

   struct file_operations {
           int (*lseek) (struct inode *, struct file *, off_t, int);
           int (*read) (struct inode *, struct file *, char *, int);
           int (*write) (struct inode *, struct file *, const char *, int);
           int (*readdir) (struct inode *, struct file *, void *, filldir_t);
           int (*select) (struct inode *, struct file *, int, select_table *);
           int (*ioctl) (struct inode *, struct file *, unsigned int, unsigned long);
           int (*mmap) (struct inode *, struct file *, struct vm_area_struct *);
           int (*open) (struct inode *, struct file *);
           void (*release) (struct inode *, struct file *);
           int (*fsync) (struct inode *, struct file *);
           int (*fasync) (struct inode *, struct file *, int);
           int (*check_media_change) (kdev_t dev);
           int (*revalidate) (kdev_t dev);
   };

   ... and ...

   struct inode_operations {
           struct file_operations * default_file_ops;
           int (*create) (struct inode *,const char *,int,int,struct inode **);
           int (*lookup) (struct inode *,const char *,int,struct inode **);
           int (*link) (struct inode *,struct inode *,const char *,int);
           int (*unlink) (struct inode *,const char *,int);
           int (*symlink) (struct inode *,const char *,int,const char *);
           int (*mkdir) (struct inode *,const char *,int,int);
           int (*rmdir) (struct inode *,const char *,int);
           int (*mknod) (struct inode *,const char *,int,int,int);
           int (*rename) (struct inode *,const char *,int,struct inode *,const char *,int);
           int (*readlink) (struct inode *,char *,int);
           int (*follow_link) (struct inode *,struct inode *,int,int,struct inode **);
           int (*readpage) (struct inode *, struct page *);
           int (*writepage) (struct inode *, struct page *);
           int (*bmap) (struct inode *,int);
           void (*truncate) (struct inode *);
           int (*permission) (struct inode *, int);
           int (*smap) (struct inode *,int);
   };

--- rod.

saumitra · 04-07-2007, 08:43 AM

yes i'm planning the same now..

just have to look where all these are IMPLEMENTED... (defined!)

any idea where are they??

i think they are specifically implemented in kernel/fs/<file system>/....

they are related to a file system.

so consider.. file->f_ops->write calls VFS_WRITE in any case..

so are there some functions for all file->f_ops->....

i can hook those to get in generic file system tracking??

theNbomr · 04-07-2007, 01:26 PM

How did you implement the write function? The rest should be done the same way. Are you using fuse?

--- rod.

saumitra · 04-07-2007, 11:44 PM

Quote:

Originally Posted by theNbomr

How did you implement the write function? The rest should be done the same way. Are you using fuse?

--- rod.

i implemented vfs_write function.

it has 4 parameters. ## MACHINE A##
1> file struct
2> buf
3> count (buf)
4> file pos (to be written at)

so i took
1>__ FILE NAME __ (not struct)
2> buf
3> count (buf)
4> file pos (to be written at)

so on ## MACHINE B ##

opened respective file structure with __FILE NAME __ i had...

so on machine B i have again,,.
1> file struct
2> buf
3> count (buf)
4> file pos (to be written at)

((btw initially files on both machines are same... exactly identicall in all aspects))