Skip to contentSkip to conversation

drm/i915: do not enable RC6p on Sandy Bridge

PATCH In review Eugeni Dodonov 3 participants 3 messages Feb 14, 2012 – Feb 16, 2012

Subsystems (2)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 drm/i915: do not enable RC6p on Sandy Bridge

With base on latest findings, RC6p seems to be respondible for RC6-related issues on Sandy Bridge platform. To work-around those issues, the previous solution was to completely disable RC6 on Sandy Bridge for the past few releases, even if plain RC6 was not giving any issues.

What this patch does is preventing RC6p from being enabled on Sandy Bridge even if users enable RC6 via a kernel parameter. So it won't change the defaults in any way, but will ensure that if users do enable RC6 manually it won't break their machines by enabling this extra state.

Proper fix for this (enabling specific RC6 states according to the GPU generation) were proposed for the -next kernel, but we are too late in the release process now to pick such changes.

CC Keith Packard <[email protected]>
CC Jesse Barnes <[email protected]>
---
 drivers/gpu/drm/i915/intel_display.c |    4 ++--
 1 files changed, 2 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_display.c +2 -2
index b3b51c4..834d6ad 100644
@@ -8182,8 +8182,8 @@ void gen6_enable_rps(struct drm_i915_private *dev_priv)
 	I915_WRITE(GEN6_RC6pp_THRESHOLD, 64000); /* unused */
 
 	if (intel_enable_rc6(dev_priv->dev))
-		rc6_mask = GEN6_RC_CTL_RC6p_ENABLE |
-			GEN6_RC_CTL_RC6_ENABLE;
+		rc6_mask = GEN6_RC_CTL_RC6_ENABLE |
+			(IS_GEN7(dev_priv->dev)) ? GEN6_RC_CTL_RC6p_ENABLE : 0;
 
 	I915_WRITE(GEN6_RC_CONTROL,
 		   rc6_mask |
Signature · 1 line
1.7.9
Acked-by on v1 1/1
#1 lore raw

<#part sign=pgpmime>

drivers/gpu/drm/i915/intel_display.c void gen6_enable_rps(struct drm_i915_private *dev_priv) · line 8186

Quoted 4 lines · Eugeni Dodonov
drivers/gpu/drm/i915/intel_display.c +2 -2
-		rc6_mask = GEN6_RC_CTL_RC6p_ENABLE |
-			GEN6_RC_CTL_RC6_ENABLE;
+		rc6_mask = GEN6_RC_CTL_RC6_ENABLE |
+			(IS_GEN7(dev_priv->dev)) ? GEN6_RC_CTL_RC6p_ENABLE : 0;

Jesse -- I asked for this to go into -fixes to see if we can't get some additional data from our existing RC6 failure cases. If you agree that this is reasonable, please feel free to pull this into -fixes and send it along to airlied.

Signature · 1 line
[email protected]
#2 lore raw
Quoted 13 lines · On Wed, 15 Feb 2012 15:39:14 +1300 Keith Packard <[email protected]> wrote:

<#part sign=pgpmime>

drivers/gpu/drm/i915/intel_display.c void gen6_enable_rps(struct drm_i915_private *dev_priv) · line 8186

Quoted 4 lines · Eugeni Dodonov
drivers/gpu/drm/i915/intel_display.c +2 -2
-		rc6_mask = GEN6_RC_CTL_RC6p_ENABLE |
-			GEN6_RC_CTL_RC6_ENABLE;
+		rc6_mask = GEN6_RC_CTL_RC6_ENABLE |
+			(IS_GEN7(dev_priv->dev)) ? GEN6_RC_CTL_RC6p_ENABLE : 0;

Jesse -- I asked for this to go into -fixes to see if we can't get some additional data from our existing RC6 failure cases. If you agree that this is reasonable, please feel free to pull this into -fixes and send it along to airlied.

Quoted 1 line

Yeah it's fine, I've applied it to -fixes. Hopefully the RC6p thing will work out and we can queue a change of default in the stable series on top of this.

Signature · 1 line
Jesse Barnes, Intel Open Source Technology Center

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx