diff mbox

drm/i915/bdw: Don't use forcewake needlessly

Message ID 1386681892-31586-1-git-send-email-ville.syrjala@linux.intel.com (mailing list archive)
State New, archived
Headers show

Commit Message

Ville Syrjälä Dec. 10, 2013, 1:24 p.m. UTC 
From: Ville Syrjälä <ville.syrjala@linux.intel.com>

Not all registers need forcewake even if they're not shadowed.
Add the missing NEEDS_FORCE_WAKE() check to gen8_writeX() to
avoid needless forcewake usage when writing eg. display
registers.

Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
---
I didn't test this at all, so be warned.

 drivers/gpu/drm/i915/intel_uncore.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Comments

Ben Widawsky Dec. 10, 2013, 7:47 p.m. UTC | #1 
On Tue, Dec 10, 2013 at 03:24:52PM +0200, ville.syrjala@linux.intel.com wrote:
> From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> 
> Not all registers need forcewake even if they're not shadowed.
> Add the missing NEEDS_FORCE_WAKE() check to gen8_writeX() to
> avoid needless forcewake usage when writing eg. display
> registers.
> 
> Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> ---
> I didn't test this at all, so be warned.
> 
>  drivers/gpu/drm/i915/intel_uncore.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
> index e63658e..78fe8d6 100644
> --- a/drivers/gpu/drm/i915/intel_uncore.c
> +++ b/drivers/gpu/drm/i915/intel_uncore.c
> @@ -616,7 +616,8 @@ static bool is_gen8_shadowed(struct drm_i915_private *dev_priv, u32 reg)
>  #define __gen8_write(x) \
>  static void \
>  gen8_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace) { \
> -	bool __needs_put = !is_gen8_shadowed(dev_priv, reg); \
> +	bool __needs_put = NEEDS_FORCE_WAKE(dev_priv, reg) && \
> +		!is_gen8_shadowed(dev_priv, reg); \
>  	REG_WRITE_HEADER; \
>  	if (__needs_put) { \
>  		dev_priv->uncore.funcs.force_wake_get(dev_priv, \

I would prefer if this was simply checking the display offset, as
opposed to NEEDS_FORCE_WAKE - since if someone adds to the macro we run
the risk of a potentially hard to track down bug.

In any case, the patch is functionally correct, and I'd prefer the
change, but up to you:
Reviewed-by: Ben Widawsky <ben@bwidawsk.net>

> -- 
> 1.8.3.2
> 
> _______________________________________________
> Intel-gfx mailing list
> Intel-gfx@lists.freedesktop.org
> http://lists.freedesktop.org/mailman/listinfo/intel-gfx
Daniel Vetter Dec. 10, 2013, 10:30 p.m. UTC | #2 
On Tue, Dec 10, 2013 at 11:47:17AM -0800, Ben Widawsky wrote:
> On Tue, Dec 10, 2013 at 03:24:52PM +0200, ville.syrjala@linux.intel.com wrote:
> > From: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > 
> > Not all registers need forcewake even if they're not shadowed.
> > Add the missing NEEDS_FORCE_WAKE() check to gen8_writeX() to
> > avoid needless forcewake usage when writing eg. display
> > registers.
> > 
> > Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
> > ---
> > I didn't test this at all, so be warned.
> > 
> >  drivers/gpu/drm/i915/intel_uncore.c | 3 ++-
> >  1 file changed, 2 insertions(+), 1 deletion(-)
> > 
> > diff --git a/drivers/gpu/drm/i915/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
> > index e63658e..78fe8d6 100644
> > --- a/drivers/gpu/drm/i915/intel_uncore.c
> > +++ b/drivers/gpu/drm/i915/intel_uncore.c
> > @@ -616,7 +616,8 @@ static bool is_gen8_shadowed(struct drm_i915_private *dev_priv, u32 reg)
> >  #define __gen8_write(x) \
> >  static void \
> >  gen8_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace) { \
> > -	bool __needs_put = !is_gen8_shadowed(dev_priv, reg); \
> > +	bool __needs_put = NEEDS_FORCE_WAKE(dev_priv, reg) && \
> > +		!is_gen8_shadowed(dev_priv, reg); \
> >  	REG_WRITE_HEADER; \
> >  	if (__needs_put) { \
> >  		dev_priv->uncore.funcs.force_wake_get(dev_priv, \
> 
> I would prefer if this was simply checking the display offset, as
> opposed to NEEDS_FORCE_WAKE - since if someone adds to the macro we run
> the risk of a potentially hard to track down bug.

Yeah, I agree that an explicit check would be better since the forcewake
stuff works a bit different on gen8+
-Daniel

> 
> In any case, the patch is functionally correct, and I'd prefer the
> change, but up to you:
> Reviewed-by: Ben Widawsky <ben@bwidawsk.net>
> 
> > -- 
> > 1.8.3.2
> > 
> > _______________________________________________
> > Intel-gfx mailing list
> > Intel-gfx@lists.freedesktop.org
> > http://lists.freedesktop.org/mailman/listinfo/intel-gfx
> 
> -- 
> Ben Widawsky, Intel Open Source Technology Center
> _______________________________________________
> 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/intel_uncore.c b/drivers/gpu/drm/i915/intel_uncore.c
index e63658e..78fe8d6 100644
--- a/drivers/gpu/drm/i915/intel_uncore.c
+++ b/drivers/gpu/drm/i915/intel_uncore.c
@@ -616,7 +616,8 @@  static bool is_gen8_shadowed(struct drm_i915_private *dev_priv, u32 reg)
 #define __gen8_write(x) \
 static void \
 gen8_write##x(struct drm_i915_private *dev_priv, off_t reg, u##x val, bool trace) { \
-	bool __needs_put = !is_gen8_shadowed(dev_priv, reg); \
+	bool __needs_put = NEEDS_FORCE_WAKE(dev_priv, reg) && \
+		!is_gen8_shadowed(dev_priv, reg); \
 	REG_WRITE_HEADER; \
 	if (__needs_put) { \
 		dev_priv->uncore.funcs.force_wake_get(dev_priv, \