From mboxrd@z Thu Jan  1 00:00:00 1970
From: bjorn@mork.no (=?utf-8?Q?Bj=C3=B8rn_Mork?=)
Date: Sun, 10 Jul 2016 18:02:23 +0200
Subject: Proper way to get device file minor number from struct file
In-Reply-To: <nltp7i$5oi$1@ger.gmane.org> (Matwey V. Kornilov's message of
	"Sun, 10 Jul 2016 18:22:25 +0300")
References: <nltp7i$5oi$1@ger.gmane.org>
Message-ID: <877fcta3cg.fsf@miraculix.mork.no>
To: kernelnewbies@lists.kernelnewbies.org
List-Id: kernelnewbies.lists.kernelnewbies.org

"Matwey V. Kornilov" <matwey.kornilov@gmail.com> writes:

> Hello,
>
> I am looking through some legacy out of tree custom linux kernel device
> driver.
>
> It is full of lines like the following:
>
> iminor(fp->f_dentry->d_inode)
>
> here struct file *fp is an argument of callback of struct file_operations.
>
> Documentation/filesystems/porting says:
>
> "f_dentry is gone; use f_path.dentry, or, better yet, see if you can
> avoid it entirely"
>
> So, is there a proper compatible way to do the same?

Don't know how proper it is, but a "git grep iminor drivers/" shows that

   minor = iminor(file_inode(file));

is a very common construct.

Or alternatively, do the iminor(inode) lookup once in open() to lookup
up your device struct or whatever, and then save that reference in
file->private_data, thereby avoiding the need to know the minor anywhere
else.



Bj?rn