linux-kernel - Re: [RFC] Hack to use mkdir/rmdir in debugfs

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 22 Jan 2013 20:41:46 -0800
From:	Greg Kroah-Hartman <gregkh@...uxfoundation.org>
To:	Steven Rostedt <rostedt@...dmis.org>
Cc:	LKML <linux-kernel@...r.kernel.org>,
	linux-fsdevel <linux-fsdevel@...r.kernel.org>,
	Al Viro <viro@...iv.linux.org.uk>,
	Andrew Morton <akpm@...ux-foundation.org>
Subject: Re: [RFC] Hack to use mkdir/rmdir in debugfs

On Tue, Jan 22, 2013 at 11:31:35PM -0500, Steven Rostedt wrote:
> On Tue, 2013-01-22 at 20:08 -0800, Greg Kroah-Hartman wrote:
> 
> > > Is doing something as silly as the following fine, or is there a better
> > > way?
> > 
> > Yes, why not create your own fs for ftrace?  :)
> 
> But but but...
> 
> debugfs is soooo convenient!
> 
> Do you think it would be worth doing that though? I only need the mkdir
> and rmdir for this one instance. Nothing more.
> 
> > >	mutex_unlock(&inode->i_mutex);
> > > 
> > > 	ret = new_instance_create(dentry->d_iname);
> > > 
> > > 	mutex_lock(&inode->i_mutex);
> > > 
> 
> > But how can you callback to your code to let it know that something was
> > created in it?
> 
> I pass the dentry->d_iname to create a directory.
> 
> > 
> > Don't you need that for both mkdir and rmdir?
> 
> It's a global list. It's very specific and doesn't need to be robust. It
> all deals with modifying one global parameter. Not that hard. And I've
> already implemented this. It works quite well :-)
> 
> > 
> > But again, I'd really not want to do this in debugfs, how about your own
> > filesystem?
> 
> I will note that this never modifies the debugfs code. But it does
> circumvent it.

Ah, I like circumventing debugfs, it kind of fits right into its mission :)

> That is, all this code lives in kernel/trace/trace.c. I don't modify
> any of the debugfs code. I just replace the debugfs
> dentry->d_inode->i_op with my own ops.

Oh, ok, I thought you were talking about modifying the debugfs core.

> I can create my own fs, but that just seems to be overkill. The only
> difference is that I need mkdir and rmdir for this one instance.
> 
> That said, perhaps it wouldn't be too hard to create a ftracefs. Where
> should it go? fs/ftrace or perhaps kernel/trace/fs ?

Sure, it only takes 300 lines to write a fs so it's not hard to do your
own.

> I notice that I only use:
> 
> debugfs_create_file()
> debugfs_remove();
> debugfs_create_dir();
> debugfs_remove_recursive();
> debugfs_initialized()
> 
> so maybe it wouldn't be such a far fetch thing to implement.
> 
> But then would I be able to still mount it in /debug/tracing ? As this
> is where everything currently uses it? But then we need to teach admins
> to add it there, or someplace else. /sys/kernel/ftrace?
> 
> Tools like trace-cmd and perf already expect it to be in the debugfs
> tracing directory, and will automate mounting it to /sys/kernel/debug/
> if not found. This may break userspace if I make another fs.

Yes, you would need to require distros to mount it in the proper place,
which is a pain, but not impossible.  It's up to you, as I thought you
needed to modify debugfs, I didn't like it, but if you can do it in your
own code, I really don't care anymore :)

greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/