Re: [PATCH] drm/nouveau: add lockdep annotations

From: Maarten Lankhorst <maarten.lankhorst@canonical.com>
To: Ben Skeggs <skeggsb@gmail.com>
Cc: Marcin Slusarz <marcin.slusarz@gmail.com>,
	Peter Hurley <peter@hurleysoftware.com>,
	nouveau@lists.freedesktop.org,
	Linux Kernel <linux-kernel@vger.kernel.org>,
	dri-devel@lists.freedesktop.org,
	Daniel J Blueman <daniel@quora.org>,
	Ben Skeggs <bskeggs@redhat.com>,
	Arend van Spriel <arend@broadcom.com>
Subject: Re: [PATCH] drm/nouveau: add lockdep annotations
Date: Thu, 07 Feb 2013 17:19:11 +0100	[thread overview]
Message-ID: <5113D3FF.6060009@canonical.com> (raw)
In-Reply-To: <20130205205255.GA10051@localhost.bne.redhat.com>

Hey,

Op 05-02-13 21:52, Ben Skeggs schreef:
> On Mon, Feb 04, 2013 at 10:59:28PM +0100, Maarten Lankhorst wrote:
>> Op 04-02-13 22:30, Marcin Slusarz schreef:
>>> 1) Lockdep thinks all nouveau subdevs belong to the same class and can be
>>> locked in arbitrary order, which is not true (at least in general case).
>>> Tell it to distinguish subdevs by (o)class type.
>> Apart from this specific case, is there any other reason why we require being able to nest 2 subdev locks?
> I think I tend to prefer Marcin's fix for this actually.  The subdev's
> are completely separate classes of objects and as interaction between
> them increases (PM will be very much like this), we may very well
> require holding multiple subdev mutexes at once.
>
> Ben.
Depends, I think for this specific example I think my cleanup is better.

For the generic case you could use nested mutexes, which will give you a
different lockdep class when you need it. It's probably better to have those
cases where you do need to nest locking annotated:

mutex_lock_nested(&mutex, SINGLE_DEPTH_NESTING);

See also Documentation/lockdep-design.txt

~Maarten