Re: Help on automatic population of directories

[Shailabh Nagar <nagar@watson.ibm.com>]

Maneesh Soni wrote:

>For creating new files, dentry and inode for the new files are needed. So
>that can be done as follows where dentry is the dentry for the directory
>under which you want to create new files.
>
>	down(&dentry->d_inode->i_sem);
>	new_dentry = rcfs_get_dentry(dentry->d_inode, "New_file");
>	if (!IS_ERR(new_dentry)) {
>		retval = rcfs_mknod(dentry->d_inode, new_dentry, mode, 0);
>	}
>	up(&dentry->d_inode->i_sem);
>
>Now for rcfs_get_dentry(), have a look at sysfs_get_dentry(). This should
>allocate new dentry if it is not already in dcache. Also you have to 
>define ->lookup() dir_inode operation as simple_lookup().
>
>  
>
Thanks for that simple code snippet which works well (with the mode 
value set properly).
This is exactly what I was looking for !

Now onto autodelete.....

>>Q2: How can one find the vfsmnt given the dentry of a directory that has
>>just been created ?
>>
>>I was thinking of using d_path() to lookup the complete path of the 
>>just-created directory and then calling a modified version of sys_mknod
>>on the resulting string. Probably not the best way of doing this but it 
>>was simpler to understand :-) But I'm having trouble determining the 
>>vfsmnt from the dentry returned from rcfs_mknod - I tried using
>>current->namespace->root and also current->fs->rootmnt with no success -
>>dpath returns blank or garbage as the resulting string and not 
>>"/rcfs/hello/" which is what I expect when directory hello is being created.
>>
>>    
>>
>
>Why do you need this?..  It seems you want to know the absolute path for the
>new created dentry. 
>  
>
Yes, I was thinking of getting the absolute path and then using a 
modified sys_mknod (without the copy/from user
for the pathname)  to create the entry. Obviously pretty inefficient  as 
Jan Hudec  pointed out :-(

 -- Shailabh