diff mbox

[v1,2/2] drm/i915: Deferring uncore early sanitize, sanitize, disable pc8 to resume

Message ID 1417417085-32419-2-git-send-email-sagar.a.kamble@intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

sagar.a.kamble@intel.com Dec. 1, 2014, 6:58 a.m. UTC 
From: Sagar Kamble <sagar.a.kamble@intel.com>

Due to disabling of RC6 in uncore_sanitize in early resume, power is drained
till it RC6 is re-enabled post resume.
With this change RC6 disabling will be done at beginning of resume only.
This helps yield additional power benefits.

Signed-off-by: Akash Goel <akash.goel@intel.com>
Signed-off-by: Sagar Kamble <sagar.a.kamble@intel.com>
---
 drivers/gpu/drm/i915/i915_drv.c | 13 +++++++------
 1 file changed, 7 insertions(+), 6 deletions(-)

Comments

Daniel Vetter Dec. 1, 2014, 9:16 a.m. UTC | #1 
On Mon, Dec 01, 2014 at 12:28:05PM +0530, sagar.a.kamble@intel.com wrote:
> From: Sagar Kamble <sagar.a.kamble@intel.com>
> 
> Due to disabling of RC6 in uncore_sanitize in early resume, power is drained
> till it RC6 is re-enabled post resume.
> With this change RC6 disabling will be done at beginning of resume only.
> This helps yield additional power benefits.
> 
> Signed-off-by: Akash Goel <akash.goel@intel.com>
> Signed-off-by: Sagar Kamble <sagar.a.kamble@intel.com>

A bit of git blame turns up

commit 76c4b250080fff6e4befaa3619942422fd0ea380
Author: Imre Deak <imre.deak@intel.com>
Date:   Tue Apr 1 19:55:22 2014 +0300

    drm/i915: move power domain init earlier during system resume
    
    During resume the intel hda audio driver depends on the i915 driver
    reinitializing the audio power domain. Since the order of calling the
    i915 resume handler wrt. that of the audio driver is not guaranteed,
    move the power domain reinitialization step to the resume_early
    handler. This is guaranteed to run before the resume handler of any
    other driver.
    
    The power domain initialization in turn requires us to enable the i915
    pci device first, so move that part earlier too.
    
    Accordingly disabling of the i915 pci device should happen after the
    audio suspend handler ran. So move the disabling later from the i915
    resume handler to the resume_late handler.
    
    v2:
    - move intel_uncore_sanitize/early_sanitize earlier too, so they don't
      get reordered wrt. intel_power_domains_init_hw()
    
    Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=76152
    Signed-off-by: Imre Deak <imre.deak@intel.com>
    Reviewed-by: Takashi Iwai <tiwai@suse.de>
    Cc: stable@vger.kernel.org
    [danvet: Add cc: stable and loud comments that this is just a hack.]
    [danvet: Fix "Should it be static?" sparse warning reported by Wu
    Fengguang's kbuilder.]
    Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>


How does this patch not break stuff?

And a general rule of thumb: If you change anything in the driver load
sequence please digg around in the git historly (with git log --pickaxe
and git blame) to gather evidence for your changes and make sure you don't
break anything.

Thanks, Daniel

> ---
>  drivers/gpu/drm/i915/i915_drv.c | 13 +++++++------
>  1 file changed, 7 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> index 71be3c9..0e08ec0 100644
> --- a/drivers/gpu/drm/i915/i915_drv.c
> +++ b/drivers/gpu/drm/i915/i915_drv.c
> @@ -675,6 +675,13 @@ static int i915_drm_resume(struct drm_device *dev)
>  {
>  	struct drm_i915_private *dev_priv = dev->dev_private;
>  
> +	intel_uncore_early_sanitize(dev, true);
> +
> +	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
> +		hsw_disable_pc8(dev_priv);
> +
> +	intel_uncore_sanitize(dev);
> +
>  	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
>  		mutex_lock(&dev->struct_mutex);
>  		i915_gem_restore_gtt_mappings(dev);
> @@ -761,12 +768,6 @@ static int i915_drm_resume_early(struct drm_device *dev)
>  	if (ret)
>  		DRM_ERROR("Resume prepare failed: %d,Continuing resume\n", ret);
>  
> -	intel_uncore_early_sanitize(dev, true);
> -
> -	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
> -		hsw_disable_pc8(dev_priv);
> -
> -	intel_uncore_sanitize(dev);
>  	intel_power_domains_init_hw(dev_priv);
>  
>  	return ret;
> -- 
> 1.8.5
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
sagar.a.kamble@intel.com Dec. 1, 2014, 10:29 a.m. UTC | #2 
Thanks Daniel.
This particular commit is moving power_domain_init into resume early.
Does not have details about ordering with uncore early sanitize and
uncore sanitize.

Imre,
Can you please clarify why this ordering with power domain init was
done?

Thanks,
Sagar

On Mon, 2014-12-01 at 10:16 +0100, Daniel Vetter wrote:
> On Mon, Dec 01, 2014 at 12:28:05PM +0530, sagar.a.kamble@intel.com wrote:
> > From: Sagar Kamble <sagar.a.kamble@intel.com>
> > 
> > Due to disabling of RC6 in uncore_sanitize in early resume, power is drained
> > till it RC6 is re-enabled post resume.
> > With this change RC6 disabling will be done at beginning of resume only.
> > This helps yield additional power benefits.
> > 
> > Signed-off-by: Akash Goel <akash.goel@intel.com>
> > Signed-off-by: Sagar Kamble <sagar.a.kamble@intel.com>
> 
> A bit of git blame turns up
> 
> commit 76c4b250080fff6e4befaa3619942422fd0ea380
> Author: Imre Deak <imre.deak@intel.com>
> Date:   Tue Apr 1 19:55:22 2014 +0300
> 
>     drm/i915: move power domain init earlier during system resume
>     
>     During resume the intel hda audio driver depends on the i915 driver
>     reinitializing the audio power domain. Since the order of calling the
>     i915 resume handler wrt. that of the audio driver is not guaranteed,
>     move the power domain reinitialization step to the resume_early
>     handler. This is guaranteed to run before the resume handler of any
>     other driver.
>     
>     The power domain initialization in turn requires us to enable the i915
>     pci device first, so move that part earlier too.
>     
>     Accordingly disabling of the i915 pci device should happen after the
>     audio suspend handler ran. So move the disabling later from the i915
>     resume handler to the resume_late handler.
>     
>     v2:
>     - move intel_uncore_sanitize/early_sanitize earlier too, so they don't
>       get reordered wrt. intel_power_domains_init_hw()
>     
>     Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=76152
>     Signed-off-by: Imre Deak <imre.deak@intel.com>
>     Reviewed-by: Takashi Iwai <tiwai@suse.de>
>     Cc: stable@vger.kernel.org
>     [danvet: Add cc: stable and loud comments that this is just a hack.]
>     [danvet: Fix "Should it be static?" sparse warning reported by Wu
>     Fengguang's kbuilder.]
>     Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> 
> 
> How does this patch not break stuff?
> 
> And a general rule of thumb: If you change anything in the driver load
> sequence please digg around in the git historly (with git log --pickaxe
> and git blame) to gather evidence for your changes and make sure you don't
> break anything.
> 
> Thanks, Daniel
> 
> > ---
> >  drivers/gpu/drm/i915/i915_drv.c | 13 +++++++------
> >  1 file changed, 7 insertions(+), 6 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> > index 71be3c9..0e08ec0 100644
> > --- a/drivers/gpu/drm/i915/i915_drv.c
> > +++ b/drivers/gpu/drm/i915/i915_drv.c
> > @@ -675,6 +675,13 @@ static int i915_drm_resume(struct drm_device *dev)
> >  {
> >  	struct drm_i915_private *dev_priv = dev->dev_private;
> >  
> > +	intel_uncore_early_sanitize(dev, true);
> > +
> > +	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
> > +		hsw_disable_pc8(dev_priv);
> > +
> > +	intel_uncore_sanitize(dev);
> > +
> >  	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
> >  		mutex_lock(&dev->struct_mutex);
> >  		i915_gem_restore_gtt_mappings(dev);
> > @@ -761,12 +768,6 @@ static int i915_drm_resume_early(struct drm_device *dev)
> >  	if (ret)
> >  		DRM_ERROR("Resume prepare failed: %d,Continuing resume\n", ret);
> >  
> > -	intel_uncore_early_sanitize(dev, true);
> > -
> > -	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
> > -		hsw_disable_pc8(dev_priv);
> > -
> > -	intel_uncore_sanitize(dev);
> >  	intel_power_domains_init_hw(dev_priv);
> >  
> >  	return ret;
> > -- 
> > 1.8.5
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
>
Imre Deak Dec. 1, 2014, 1:23 p.m. UTC | #3 
On Mon, 2014-12-01 at 15:59 +0530, Sagar Arun Kamble wrote:
> Thanks Daniel.
> This particular commit is moving power_domain_init into resume early.
> Does not have details about ordering with uncore early sanitize and
> uncore sanitize.
> 
> Imre,
> Can you please clarify why this ordering with power domain init was
> done?

We call intel_uncore_early_sanitize() and intel_uncore_sanitize() before
any other HW access. They are called from i915_drm_resume_early() since
the hda driver's resume handler can run before i915_drm_resume() is
called. The hda resume handler can in turn request the display power
well resulting in an i915 HW access.

> 
> Thanks,
> Sagar
> 
> On Mon, 2014-12-01 at 10:16 +0100, Daniel Vetter wrote:
> > On Mon, Dec 01, 2014 at 12:28:05PM +0530, sagar.a.kamble@intel.com wrote:
> > > From: Sagar Kamble <sagar.a.kamble@intel.com>
> > > 
> > > Due to disabling of RC6 in uncore_sanitize in early resume, power is drained
> > > till it RC6 is re-enabled post resume.
> > > With this change RC6 disabling will be done at beginning of resume only.
> > > This helps yield additional power benefits.
> > > 
> > > Signed-off-by: Akash Goel <akash.goel@intel.com>
> > > Signed-off-by: Sagar Kamble <sagar.a.kamble@intel.com>
> > 
> > A bit of git blame turns up
> > 
> > commit 76c4b250080fff6e4befaa3619942422fd0ea380
> > Author: Imre Deak <imre.deak@intel.com>
> > Date:   Tue Apr 1 19:55:22 2014 +0300
> > 
> >     drm/i915: move power domain init earlier during system resume
> >     
> >     During resume the intel hda audio driver depends on the i915 driver
> >     reinitializing the audio power domain. Since the order of calling the
> >     i915 resume handler wrt. that of the audio driver is not guaranteed,
> >     move the power domain reinitialization step to the resume_early
> >     handler. This is guaranteed to run before the resume handler of any
> >     other driver.
> >     
> >     The power domain initialization in turn requires us to enable the i915
> >     pci device first, so move that part earlier too.
> >     
> >     Accordingly disabling of the i915 pci device should happen after the
> >     audio suspend handler ran. So move the disabling later from the i915
> >     resume handler to the resume_late handler.
> >     
> >     v2:
> >     - move intel_uncore_sanitize/early_sanitize earlier too, so they don't
> >       get reordered wrt. intel_power_domains_init_hw()
> >     
> >     Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=76152
> >     Signed-off-by: Imre Deak <imre.deak@intel.com>
> >     Reviewed-by: Takashi Iwai <tiwai@suse.de>
> >     Cc: stable@vger.kernel.org
> >     [danvet: Add cc: stable and loud comments that this is just a hack.]
> >     [danvet: Fix "Should it be static?" sparse warning reported by Wu
> >     Fengguang's kbuilder.]
> >     Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
> > 
> > 
> > How does this patch not break stuff?
> > 
> > And a general rule of thumb: If you change anything in the driver load
> > sequence please digg around in the git historly (with git log --pickaxe
> > and git blame) to gather evidence for your changes and make sure you don't
> > break anything.
> > 
> > Thanks, Daniel
> > 
> > > ---
> > >  drivers/gpu/drm/i915/i915_drv.c | 13 +++++++------
> > >  1 file changed, 7 insertions(+), 6 deletions(-)
> > > 
> > > diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
> > > index 71be3c9..0e08ec0 100644
> > > --- a/drivers/gpu/drm/i915/i915_drv.c
> > > +++ b/drivers/gpu/drm/i915/i915_drv.c
> > > @@ -675,6 +675,13 @@ static int i915_drm_resume(struct drm_device *dev)
> > >  {
> > >  	struct drm_i915_private *dev_priv = dev->dev_private;
> > >  
> > > +	intel_uncore_early_sanitize(dev, true);
> > > +
> > > +	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
> > > +		hsw_disable_pc8(dev_priv);
> > > +
> > > +	intel_uncore_sanitize(dev);
> > > +
> > >  	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
> > >  		mutex_lock(&dev->struct_mutex);
> > >  		i915_gem_restore_gtt_mappings(dev);
> > > @@ -761,12 +768,6 @@ static int i915_drm_resume_early(struct drm_device *dev)
> > >  	if (ret)
> > >  		DRM_ERROR("Resume prepare failed: %d,Continuing resume\n", ret);
> > >  
> > > -	intel_uncore_early_sanitize(dev, true);
> > > -
> > > -	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
> > > -		hsw_disable_pc8(dev_priv);
> > > -
> > > -	intel_uncore_sanitize(dev);
> > >  	intel_power_domains_init_hw(dev_priv);
> > >  
> > >  	return ret;
> > > -- 
> > > 1.8.5
> > > 
> > > _______________________________________________
> > > Intel-gfx mailing list
> > > Intel-gfx@lists.freedesktop.org
> > > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> > 
> 
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
diff mbox

Patch

diff --git a/drivers/gpu/drm/i915/i915_drv.c b/drivers/gpu/drm/i915/i915_drv.c
index 71be3c9..0e08ec0 100644
--- a/drivers/gpu/drm/i915/i915_drv.c
+++ b/drivers/gpu/drm/i915/i915_drv.c
@@ -675,6 +675,13 @@  static int i915_drm_resume(struct drm_device *dev)
 {
 	struct drm_i915_private *dev_priv = dev->dev_private;
 
+	intel_uncore_early_sanitize(dev, true);
+
+	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
+		hsw_disable_pc8(dev_priv);
+
+	intel_uncore_sanitize(dev);
+
 	if (drm_core_check_feature(dev, DRIVER_MODESET)) {
 		mutex_lock(&dev->struct_mutex);
 		i915_gem_restore_gtt_mappings(dev);
@@ -761,12 +768,6 @@  static int i915_drm_resume_early(struct drm_device *dev)
 	if (ret)
 		DRM_ERROR("Resume prepare failed: %d,Continuing resume\n", ret);
 
-	intel_uncore_early_sanitize(dev, true);
-
-	if (IS_HASWELL(dev_priv) || IS_BROADWELL(dev_priv))
-		hsw_disable_pc8(dev_priv);
-
-	intel_uncore_sanitize(dev);
 	intel_power_domains_init_hw(dev_priv);
 
 	return ret;