From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1757244AbZBJWmS (ORCPT ); Tue, 10 Feb 2009 17:42:18 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1756180AbZBJWmI (ORCPT ); Tue, 10 Feb 2009 17:42:08 -0500 Received: from 69-30-77-85.dq1sn.easystreet.com ([69.30.77.85]:48254 "EHLO kingsolver.anholt.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756119AbZBJWmG (ORCPT ); Tue, 10 Feb 2009 17:42:06 -0500 Subject: Re: [Bug #12574] possible circular locking dependency detected From: Eric Anholt To: "Michael S. Tsirkin" Cc: Dave Airlie , dri-devel@lists.sourceforge.net, Linux Kernel Mailing List , Kernel Testers List , "Rafael J. Wysocki" In-Reply-To: <20090210223711.GA6809@google.com> References: <20090210223711.GA6809@google.com> Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-1Z7gR/hHA0tX+ZN1mhZM" Date: Tue, 10 Feb 2009 14:41:56 -0800 Message-Id: <1234305716.32506.10.camel@gaiman> Mime-Version: 1.0 X-Mailer: Evolution 2.22.3.1 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org --=-1Z7gR/hHA0tX+ZN1mhZM Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2009-02-11 at 00:37 +0200, Michael S. Tsirkin wrote: > Dave, dri guys, > Could you take a look at this circular dependency please (below)? I > observe it when suspending laptop with radeon drm loaded and with > lockdep enabled. It seems that the root of the problem is that > various vm ops such as drm_vm_open, drm_mmap) are called with mm > semaphore taken, and take dev->struct_mutex. On the other hand, > drm_rmmap_locked is called with dev->struct_mutex, and calls mtrr_del > which depends on mm semaphore indirectly. >=20 > What do you think? Yes, there are real lock inversions now due to the GTT mmap code. It's going to be a pain to fix (I tried getting the mmap_sem -> struct_mutex path to go away, but the fact that mmap_sem is held over the fault handler pretty much kills that). It's high on the list, though. --=20 Eric Anholt eric@anholt.net eric.anholt@intel.com --=-1Z7gR/hHA0tX+ZN1mhZM Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEABECAAYFAkmSArQACgkQHUdvYGzw6ved3ACff4huObnqoU+6sC0vjVh1NYy4 CFAAoIKSScvTf88cLX+SVQh+PDgQbU8B =GKtc -----END PGP SIGNATURE----- --=-1Z7gR/hHA0tX+ZN1mhZM-- From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Anholt Subject: Re: [Bug #12574] possible circular locking dependency detected Date: Tue, 10 Feb 2009 14:41:56 -0800 Message-ID: <1234305716.32506.10.camel@gaiman> References: <20090210223711.GA6809@google.com> Mime-Version: 1.0 Content-Type: multipart/signed; micalg=pgp-sha1; protocol="application/pgp-signature"; boundary="=-1Z7gR/hHA0tX+ZN1mhZM" Return-path: In-Reply-To: <20090210223711.GA6809-hpIqsD4AKlfQT0dZR+AlfA@public.gmane.org> Sender: kernel-testers-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org List-ID: To: "Michael S. Tsirkin" Cc: Dave Airlie , dri-devel-5NWGOfrQmneRv+LV9MX5uipxlwaOVQ5f@public.gmane.org, Linux Kernel Mailing List , Kernel Testers List , "Rafael J. Wysocki" --=-1Z7gR/hHA0tX+ZN1mhZM Content-Type: text/plain Content-Transfer-Encoding: quoted-printable On Wed, 2009-02-11 at 00:37 +0200, Michael S. Tsirkin wrote: > Dave, dri guys, > Could you take a look at this circular dependency please (below)? I > observe it when suspending laptop with radeon drm loaded and with > lockdep enabled. It seems that the root of the problem is that > various vm ops such as drm_vm_open, drm_mmap) are called with mm > semaphore taken, and take dev->struct_mutex. On the other hand, > drm_rmmap_locked is called with dev->struct_mutex, and calls mtrr_del > which depends on mm semaphore indirectly. >=20 > What do you think? Yes, there are real lock inversions now due to the GTT mmap code. It's going to be a pain to fix (I tried getting the mmap_sem -> struct_mutex path to go away, but the fact that mmap_sem is held over the fault handler pretty much kills that). It's high on the list, though. --=20 Eric Anholt eric-WhKQ6XTQaPysTnJN9+BGXg@public.gmane.org eric.anholt-ral2JQCrhuEAvxtiuMwx3w@public.gmane.org --=-1Z7gR/hHA0tX+ZN1mhZM Content-Type: application/pgp-signature; name=signature.asc Content-Description: This is a digitally signed message part -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.9 (GNU/Linux) iEYEABECAAYFAkmSArQACgkQHUdvYGzw6ved3ACff4huObnqoU+6sC0vjVh1NYy4 CFAAoIKSScvTf88cLX+SVQh+PDgQbU8B =GKtc -----END PGP SIGNATURE----- --=-1Z7gR/hHA0tX+ZN1mhZM--