Skip to contentSkip to conversation

Revert "drm/i915: Add eDP intermediate frequencies for CHV"

PATCH In review Sivakumar Thulasimani 4 participants 10 messages Jul 31, 2015 – Aug 14, 2015

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

PATCH 1/2 Revert "drm/i915: Add eDP intermediate frequencies for CHV"

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

---
 drivers/gpu/drm/i915/intel_dp.c |    6 ------
 1 file changed, 6 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -6
index 44f8a32..d9fb7a8 100644
@@ -95,9 +95,6 @@ static const int bxt_rates[] = { 162000, 216000, 243000, 270000,
 				  324000, 432000, 540000 };
 static const int skl_rates[] = { 162000, 216000, 270000,
 				  324000, 432000, 540000 };
-static const int chv_rates[] = { 162000, 202500, 210000, 216000,
-				 243000, 270000, 324000, 405000,
-				 420000, 432000, 540000 };
 static const int default_rates[] = { 162000, 270000, 540000 };
 
 /**
@@ -1185,9 +1182,6 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
 	} else if (IS_SKYLAKE(dev)) {
 		*source_rates = skl_rates;
 		return ARRAY_SIZE(skl_rates);
-	} else if (IS_CHERRYVIEW(dev)) {
-		*source_rates = chv_rates;
-		return ARRAY_SIZE(chv_rates);
 	}
 
 	*source_rates = default_rates;
Signature · 6 lines
1.7.9.5

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

PATCH 2/2 drm/i915: remove HBR2 from chv supported list

From: "Thulasimani,Sivakumar" <[email protected]>

This patch removes 5.4Gbps from supported link rate for CHV since it is not supported in it.

---
 drivers/gpu/drm/i915/intel_dp.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +3 -2
index d9fb7a8..4e53433 100644
@@ -1186,8 +1186,9 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
 
 	*source_rates = default_rates;
 
-	if (IS_SKYLAKE(dev) && INTEL_REVID(dev) <= SKL_REVID_B0)
-		/* WaDisableHBR2:skl */
+	/* WaDisableHBR2:skl */
+	if ((IS_SKYLAKE(dev) && INTEL_REVID(dev) <= SKL_REVID_B0) ||
+		IS_CHERRYVIEW(dev))
 		return (DP_LINK_BW_2_7 >> 3) + 1;
 	else if (INTEL_INFO(dev)->gen >= 8 ||
 	    (IS_HASWELL(dev) && !IS_HSW_ULX(dev)))
Signature · 6 lines
1.7.9.5

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

Re: [PATCH 2/2] drm/i915: remove HBR2 from chv supported list

Quoted 16 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This patch removes 5.4Gbps from supported link rate for CHV since it is not supported in it.

---
 drivers/gpu/drm/i915/intel_dp.c |    5 +++--
 1 file changed, 3 insertions(+), 2 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -0
index d9fb7a8..4e53433 100644
@@ -1186,8 +1186,9 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
 

drivers/gpu/drm/i915/intel_dp.c intel_dp_source_rates(struct drm_device *dev, const int **source_rates) · line 1194

Quoted 10 lines
drivers/gpu/drm/i915/intel_dp.c +3 -2
 	*source_rates = default_rates;
 
-	if (IS_SKYLAKE(dev) && INTEL_REVID(dev) <= SKL_REVID_B0)
-		/* WaDisableHBR2:skl */
+	/* WaDisableHBR2:skl */
+	if ((IS_SKYLAKE(dev) && INTEL_REVID(dev) <= SKL_REVID_B0) ||
+		IS_CHERRYVIEW(dev))
 		return (DP_LINK_BW_2_7 >> 3) + 1;
 	else if (INTEL_INFO(dev)->gen >= 8 ||
 	    (IS_HASWELL(dev) && !IS_HSW_ULX(dev)))

I would suggest reordering this a bit. Something likle this perhaps:

/* WaDisableHBR2:skl */
if (IS_SKYLAKE && ...)

	return (DP_LINK_BW_2_7 >> 3) + 1;

if ((IS_HASWELL && !IS_HSW_ULX) || IS_BROADWELL || gen >= 9)

	return (DP_LINK_BW_5_4 >> 3) + 1;
else
	return (DP_LINK_BW_2_7 >> 3) + 1;

IMO that makes it easier to see exactly which platforms support HBR2, and it keeps the SKL w/a separated neatly for easy removal in the near future (I'm assuming we want to kill the w/a based on the fact that it's for an early stepping).

Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#3 lore raw
In reply to Sivakumar Thulasimani

hi Ville,
can you review these patches ?

regards,
Sivakumar

Quoted 30 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

---
  drivers/gpu/drm/i915/intel_dp.c |    6 ------
  1 file changed, 6 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -3
index 44f8a32..d9fb7a8 100644
@@ -95,9 +95,6 @@ static const int bxt_rates[] = { 162000, 216000, 243000, 270000,
  				  324000, 432000, 540000 };
  static const int skl_rates[] = { 162000, 216000, 270000,
  				  324000, 432000, 540000 };
-static const int chv_rates[] = { 162000, 202500, 210000, 216000,
-				 243000, 270000, 324000, 405000,
-				 420000, 432000, 540000 };
  static const int default_rates[] = { 162000, 270000, 540000 };
 
  /**

drivers/gpu/drm/i915/intel_dp.c intel_dp_source_rates(struct drm_device *dev, const int **source_rates) · line 1187

Quoted 9 lines
drivers/gpu/drm/i915/intel_dp.c +0 -3
@@ -1185,9 +1182,6 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
  	} else if (IS_SKYLAKE(dev)) {
  		*source_rates = skl_rates;
  		return ARRAY_SIZE(skl_rates);
-	} else if (IS_CHERRYVIEW(dev)) {
-		*source_rates = chv_rates;
-		return ARRAY_SIZE(chv_rates);
  	}
 
  	*source_rates = default_rates;

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

#4 lore raw
In reply to Sivakumar Thulasimani
Quoted 9 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

My docs still say it does. Is there some undocumented problem with the PLL or is this just a marketing decision?

Quoted 32 lines
---
 drivers/gpu/drm/i915/intel_dp.c |    6 ------
 1 file changed, 6 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -6
index 44f8a32..d9fb7a8 100644
@@ -95,9 +95,6 @@ static const int bxt_rates[] = { 162000, 216000, 243000, 270000,
 				  324000, 432000, 540000 };
 static const int skl_rates[] = { 162000, 216000, 270000,
 				  324000, 432000, 540000 };
-static const int chv_rates[] = { 162000, 202500, 210000, 216000,
-				 243000, 270000, 324000, 405000,
-				 420000, 432000, 540000 };
 static const int default_rates[] = { 162000, 270000, 540000 };
 
 /**
@@ -1185,9 +1182,6 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
 	} else if (IS_SKYLAKE(dev)) {
 		*source_rates = skl_rates;
 		return ARRAY_SIZE(skl_rates);
-	} else if (IS_CHERRYVIEW(dev)) {
-		*source_rates = chv_rates;
-		return ARRAY_SIZE(chv_rates);
 	}
 
 	*source_rates = default_rates;
Signature · 1 line
1.7.9.5
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#5 lore raw
Quoted 8 lines · Ville Syrjälä
Quoted 7 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200
Quoted 4 lines
Quoted 2 lines

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

My docs still say it does. Is there some undocumented problem with the PLL or is this just a marketing decision?

I don't think so, i too have just one document that shows the phy values for each of the link rates but have not found any where else to say it is supported . Also this is not tested by anyone including us from product team so i highly doubt
it will even work.

regarding HBR2, it was supposed to land on a future stepping that never happened so it is not supported at all.

Quoted 31 lines
Quoted 31 lines
---
  drivers/gpu/drm/i915/intel_dp.c |    6 ------
  1 file changed, 6 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -6
index 44f8a32..d9fb7a8 100644
@@ -95,9 +95,6 @@ static const int bxt_rates[] = { 162000, 216000, 243000, 270000,
  				  324000, 432000, 540000 };
  static const int skl_rates[] = { 162000, 216000, 270000,
  				  324000, 432000, 540000 };
-static const int chv_rates[] = { 162000, 202500, 210000, 216000,
-				 243000, 270000, 324000, 405000,
-				 420000, 432000, 540000 };
  static const int default_rates[] = { 162000, 270000, 540000 };
 
  /**
@@ -1185,9 +1182,6 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
  	} else if (IS_SKYLAKE(dev)) {
  		*source_rates = skl_rates;
  		return ARRAY_SIZE(skl_rates);
-	} else if (IS_CHERRYVIEW(dev)) {
-		*source_rates = chv_rates;
-		return ARRAY_SIZE(chv_rates);
  	}
 
  	*source_rates = default_rates;
Signature · 1 line
1.7.9.5
Signature · 7 lines
regards,
Sivakumar

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#6 lore raw
Quoted 11 lines · Sivakumar Thulasimani
Quoted 8 lines · Ville Syrjälä
Quoted 7 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

My docs still say it does. Is there some undocumented problem with the PLL or is this just a marketing decision?

I don't think so, i too have just one document that shows the phy values for each of
the link rates but have not found any where else to say it is supported .

Display cluster HAS says they're supported. And since the spreadsheet has them all in green I assume someone actually figured that they ought to work.

Quoted 3 lines

Also this is not tested by anyone including us from product team so i highly doubt
it will even work.

Well if no one has tested them I guess we shouldn't try to use them. Not a big loss in any case I suppose.

So based on that reasoning I can give an ack for this patch:

Quoted 4 lines

regarding HBR2, it was supposed to land on a future stepping that never happened
so it is not supported at all.

Hmm. Display Cluster HAS listed it as a stretch goal for early steppings. Apart from that there isn't much else to go by. Excepth the training pattern 3 support, but now that I look again the new bit for that has disappeared from the DP register in the spec. Or maybe I was looking at the k0 spec when I added it. It's still in the k0 spec but as you say that's been nuked.

In light of this, I think dropping HBR2 is reasonable. HBR is anyway enough for 4k@30 with 4 lanes, so it's not like we really need HBR2.

And could you also cook up a patch to kill the training pattern 3 support for CHV? Should be mostly a revert of my original patch that added it, but you probably need to adjust the use_tps3 condition a bit too.

Quoted 35 lines
Quoted 31 lines
Quoted 31 lines
---
  drivers/gpu/drm/i915/intel_dp.c |    6 ------
  1 file changed, 6 deletions(-)
drivers/gpu/drm/i915/intel_dp.c +0 -6
index 44f8a32..d9fb7a8 100644
@@ -95,9 +95,6 @@ static const int bxt_rates[] = { 162000, 216000, 243000, 270000,
  				  324000, 432000, 540000 };
  static const int skl_rates[] = { 162000, 216000, 270000,
  				  324000, 432000, 540000 };
-static const int chv_rates[] = { 162000, 202500, 210000, 216000,
-				 243000, 270000, 324000, 405000,
-				 420000, 432000, 540000 };
  static const int default_rates[] = { 162000, 270000, 540000 };
 
  /**
@@ -1185,9 +1182,6 @@ intel_dp_source_rates(struct drm_device *dev, const int **source_rates)
  	} else if (IS_SKYLAKE(dev)) {
  		*source_rates = skl_rates;
  		return ARRAY_SIZE(skl_rates);
-	} else if (IS_CHERRYVIEW(dev)) {
-		*source_rates = chv_rates;
-		return ARRAY_SIZE(chv_rates);
  	}
 
  	*source_rates = default_rates;
Signature · 1 line
1.7.9.5
Signature · 2 lines
regards,
Sivakumar
Signature · 6 lines
Ville Syrjälä
Intel OTC
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#7 lore raw
Quoted 46 lines · Ville Syrjälä
Quoted 11 lines · Sivakumar Thulasimani
Quoted 8 lines · Ville Syrjälä
Quoted 7 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

My docs still say it does. Is there some undocumented problem with the PLL or is this just a marketing decision?

I don't think so, i too have just one document that shows the phy values for each of
the link rates but have not found any where else to say it is supported .

Display cluster HAS says they're supported. And since the spreadsheet has them all in green I assume someone actually figured that they ought to work.

Quoted 3 lines

Also this is not tested by anyone including us from product team so i highly doubt
it will even work.

Well if no one has tested them I guess we shouldn't try to use them. Not a big loss in any case I suppose.

So based on that reasoning I can give an ack for this patch:

Quoted 4 lines

regarding HBR2, it was supposed to land on a future stepping that never happened
so it is not supported at all.

Hmm. Display Cluster HAS listed it as a stretch goal for early steppings. Apart from that there isn't much else to go by. Excepth the training pattern 3 support, but now that I look again the new bit for that has disappeared from the DP register in the spec. Or maybe I was looking at the k0 spec when I added it. It's still in the k0 spec but as you say that's been nuked.

Quoted 7 lines

In light of this, I think dropping HBR2 is reasonable. HBR is anyway enough for 4k@30 with 4 lanes, so it's not like we really need HBR2.

And could you also cook up a patch to kill the training pattern 3 support for CHV? Should be mostly a revert of my original patch that added it, but you probably need to adjust the use_tps3 condition a bit too.

Can we please grill the people responsible for this docs mess some more?

Patch itself is for Jani.
-Daniel

Signature · 7 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
maintainer
#8 lore raw
Quoted 50 lines · Daniel Vetter
Quoted 49 lines · Ville Syrjälä
Quoted 11 lines · Sivakumar Thulasimani
Quoted 8 lines · Ville Syrjälä
Quoted 7 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

My docs still say it does. Is there some undocumented problem with the PLL or is this just a marketing decision?

I don't think so, i too have just one document that shows the phy values for each of
the link rates but have not found any where else to say it is supported .

Display cluster HAS says they're supported. And since the spreadsheet has them all in green I assume someone actually figured that they ought to work.

Quoted 3 lines

Also this is not tested by anyone including us from product team so i highly doubt
it will even work.

Well if no one has tested them I guess we shouldn't try to use them. Not a big loss in any case I suppose.

So based on that reasoning I can give an ack for this patch:

Quoted 4 lines

regarding HBR2, it was supposed to land on a future stepping that never happened
so it is not supported at all.

Hmm. Display Cluster HAS listed it as a stretch goal for early steppings. Apart from that there isn't much else to go by. Excepth the training pattern 3 support, but now that I look again the new bit for that has disappeared from the DP register in the spec. Or maybe I was looking at the k0 spec when I added it. It's still in the k0 spec but as you say that's been nuked.

In light of this, I think dropping HBR2 is reasonable. HBR is anyway enough for 4k@30 with 4 lanes, so it's not like we really need HBR2.

Quoted 8 lines
Quoted 4 lines

And could you also cook up a patch to kill the training pattern 3 support for CHV? Should be mostly a revert of my original patch that added it, but you probably need to adjust the use_tps3 condition a bit too.

Can we please grill the people responsible for this docs mess some more?

Patch itself is for Jani.

There was some suggestions from Ville [1] to patch 1/2 that I haven't seen a reply to.

BR,
Jani.

[1] http://mid.gmane.org/[email protected]

Quoted 5 lines

-Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 5 lines
Jani Nikula, Intel Open Source Technology Center
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#9 lore raw
Quoted 52 lines · Jani Nikula
Quoted 51 lines · Daniel Vetter
Quoted 41 lines · Ville Syrjälä
Quoted 10 lines · Sivakumar Thulasimani
Quoted 8 lines · Ville Syrjälä
Quoted 7 lines · Sivakumar Thulasimani

From: "Thulasimani,Sivakumar" <[email protected]>

This reverts
commit fe51bfb95c996733150c44d21e1c9f4b6322a326.

Author: Ville Syrjälä <[email protected]>
Date:   Thu Mar 12 17:10:38 2015 +0200
Quoted 7 lines
Quoted 4 lines
Quoted 2 lines

CHV does not support intermediate frequencies so reverting the patch that added it in the first place

My docs still say it does. Is there some undocumented problem with the PLL or is this just a marketing decision?

I don't think so, i too have just one document that shows the phy values for each of
the link rates but have not found any where else to say it is supported .

Display cluster HAS says they're supported. And since the spreadsheet has them all in green I assume someone actually figured that they ought to work.

Quoted 3 lines

Also this is not tested by anyone including us from product team so i highly doubt
it will even work.

Well if no one has tested them I guess we shouldn't try to use them. Not a big loss in any case I suppose.

So based on that reasoning I can give an ack for this patch:

Quoted 3 lines

regarding HBR2, it was supposed to land on a future stepping that never happened
so it is not supported at all.

Hmm. Display Cluster HAS listed it as a stretch goal for early steppings. Apart from that there isn't much else to go by. Excepth the training pattern 3 support, but now that I look again the new bit for that has disappeared from the DP register in the spec. Or maybe I was looking at the k0 spec when I added it. It's still in the k0 spec but as you say that's been nuked.

Quoted 7 lines

In light of this, I think dropping HBR2 is reasonable. HBR is anyway enough for 4k@30 with 4 lanes, so it's not like we really need HBR2.

And could you also cook up a patch to kill the training pattern 3 support for CHV? Should be mostly a revert of my original patch that added it, but you probably need to adjust the use_tps3 condition a bit too.

Can we please grill the people responsible for this docs mess some more?

Quoted 10 lines
Quoted 1 line

Patch itself is for Jani.

There was some suggestions from Ville [1] to patch 1/2 that I haven't seen a reply to.

BR,
Jani.

[1] http://mid.gmane.org/[email protected]

yes, i can make that change. i was assuming that since Daniel's reply had message "patch itself is for Jani" that you would pick it up :), will check once before waiting next time.

Quoted 5 lines
Quoted 5 lines

-Daniel

Signature · 3 lines
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch
Signature · 7 lines
regards,
Sivakumar

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