From mboxrd@z Thu Jan 1 00:00:00 1970 From: Ville =?iso-8859-1?Q?Syrj=E4l=E4?= Subject: Re: [PATCH 2/4] drm/i915: Fix system resume if PCI device remained enabled Date: Mon, 18 Apr 2016 11:44:51 +0300 Message-ID: <20160418084451.GB4329@intel.com> References: <1460963062-13211-1-git-send-email-imre.deak@intel.com> <1460963062-13211-3-git-send-email-imre.deak@intel.com> <20160418082822.GY4329@intel.com> <1460968358.3172.17.camel@intel.com> Mime-Version: 1.0 Content-Type: text/plain; charset=iso-8859-1 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Content-Disposition: inline In-Reply-To: <1460968358.3172.17.camel@intel.com> Sender: stable-owner@vger.kernel.org To: Imre Deak Cc: intel-gfx@lists.freedesktop.org, stable@vger.kernel.org List-Id: intel-gfx@lists.freedesktop.org On Mon, Apr 18, 2016 at 11:32:38AM +0300, Imre Deak wrote: > On ma, 2016-04-18 at 11:28 +0300, Ville Syrj=E4l=E4 wrote: > > On Mon, Apr 18, 2016 at 10:04:20AM +0300, Imre Deak wrote: > > > During system resume we depended on pci_enable_device() also > > > putting the > > > device into PCI D0 state. This won't work if the PCI device was > > > already > > > enabled but still in D3 state. This is because pci_enable_device(= ) > > > is > > > refcounted and will not change the HW state if called with a non- > > > zero > > > refcount. Leaving the device in D3 will make all subsequent devic= e > > > accesses fail. > > >=20 > > > This didn't cause a problem most of the time, since we resumed wi= th > > > an > > > enable refcount of 0. But it fails at least after module reload > > > because > > > after that we also happen to leak a PCI device enable reference: > > > During > > > probing we call drm_get_pci_dev() which will enable the PCI devic= e, > > > but > > > during device removal drm_put_dev() won't disable it. This is a b= ug > > > of > > > its own in DRM core, but without much harm as it only leaves the > > > PCI > > > device enabled. Fixing it is also a bit more involved, due to DRM > > > mid-layering and because it affects non-i915 drivers too. The fix > > > in > > > this patch is valid regardless of the problem in DRM core. > > >=20 > > > CC: Ville Syrj=E4l=E4 > > > CC: stable@vger.kernel.org > > > Signed-off-by: Imre Deak > > > --- > > > =A0drivers/gpu/drm/i915/i915_drv.c | 9 ++++++++- > > > =A01 file changed, 8 insertions(+), 1 deletion(-) > > >=20 > > > diff --git a/drivers/gpu/drm/i915/i915_drv.c > > > b/drivers/gpu/drm/i915/i915_drv.c > > > index d550ae2..7eaa93e 100644 > > > --- a/drivers/gpu/drm/i915/i915_drv.c > > > +++ b/drivers/gpu/drm/i915/i915_drv.c > > > @@ -803,7 +803,7 @@ static int i915_drm_resume(struct drm_device > > > *dev) > > > =A0static int i915_drm_resume_early(struct drm_device *dev) > > > =A0{ > > > =A0 struct drm_i915_private *dev_priv =3D dev->dev_private; > > > - int ret =3D 0; > > > + int ret; > > > =A0 > > > =A0 /* > > > =A0 =A0* We have a resume ordering issue with the snd-hda driver > > > also > > > @@ -814,6 +814,13 @@ static int i915_drm_resume_early(struct > > > drm_device *dev) > > > =A0 =A0* FIXME: This should be solved with a special hdmi sink > > > device or > > > =A0 =A0* similar so that power domains can be employed. > > > =A0 =A0*/ > > > + > > > + ret =3D pci_set_power_state(dev->pdev, PCI_D0); > > > + if (ret) { > > > + DRM_ERROR("failed to set PCI D0 power state > > > (%d)\n", ret); > > > + goto out; > > > + } > >=20 > > Hmm. Doesn't this already happen from pci bus resume_noirq hook? >=20 > It does, but not during thaw_noirq. Maybe put that into a comment? If we ever get to dropping the device state frobbery from freeze/thaw, then we should also be able to throw out the pci_set_power_state() call as well. Perhaps we should have some asserts about the state of the PCI device? >=20 > >=20 > > > + > > > =A0 if (pci_enable_device(dev->pdev)) { > > > =A0 ret =3D -EIO; > > > =A0 goto out; > > > --=20 > > > 2.5.0 > >=20 --=20 Ville Syrj=E4l=E4 Intel OTC