Skip to contentSkip to conversation

drm/i915/skl: Bypass debug message if scalers are not requested

PATCH In review Tvrtko Ursulin 4 participants 12 messages Apr 24, 2015 – Aug 5, 2015

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

PATCH 1/1 drm/i915/skl: Bypass debug message if scalers are not requested

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be quiet.

---
 drivers/gpu/drm/i915/intel_atomic.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +3 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct drm_device *dev,
 	scaler_state = &crtc_state->scaler_state;
 	drm_state = crtc_state->base.state;
 
+	if (!scaler_state->scaler_users)
+		return 0;
+
 	num_scalers_need = hweight32(scaler_state->scaler_users);
 	DRM_DEBUG_KMS("crtc_state = %p need = %d avail = %d scaler_users = 0x%x\n",
 		crtc_state, num_scalers_need, intel_crtc->num_scalers,
Signature · 6 lines
2.3.5

_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx
#1 lore raw
Quoted 27 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be quiet.

---
 drivers/gpu/drm/i915/intel_atomic.c | 3 +++
 1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 5 lines
drivers/gpu/drm/i915/intel_atomic.c +2 -0
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct drm_device *dev,
 	scaler_state = &crtc_state->scaler_state;
 	drm_state = crtc_state->base.state;
 
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use before and no more required now. I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Quoted 7 lines
drivers/gpu/drm/i915/intel_atomic.c +1 -0
+
 	num_scalers_need = hweight32(scaler_state->scaler_users);
 	DRM_DEBUG_KMS("crtc_state = %p need = %d avail = %d scaler_users =
 0x%x\n",

crtc_state, num_scalers_need, intel_crtc->num_scalers,
--
2.3.5

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

#2 lore raw
Quoted 30 lines · Chandra
Quoted 30 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be quiet.

---
  drivers/gpu/drm/i915/intel_atomic.c | 3 +++
  1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct drm_device *dev,
  	scaler_state = &crtc_state->scaler_state;
  	drm_state = crtc_state->base.state;
 
Quoted 6 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 2 lines
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use before and no more required now. I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Regards,

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

#3 lore raw
Quoted 48 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 10 lines · Chandra
Quoted 10 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 7 lines
Quoted 7 lines

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 10 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device

Quoted 10 lines
---
  drivers/gpu/drm/i915/intel_atomic.c | 3 +++
  1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct drm_device

*dev,

Quoted 7 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 5 lines
drivers/gpu/drm/i915/intel_atomic.c +2 -0
  	scaler_state = &crtc_state->scaler_state;
  	drm_state = crtc_state->base.state;
 
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 6 lines

before and no more required now.

Quoted 2 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Oh, that's right, ignore my prev comment (with updated scaler design above shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

Quoted 4 lines

Regards,

Tvrtko

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

#4 lore raw
In reply to Tvrtko Ursulin
Quoted 67 lines

-----Original Message-----
From: Konduru, Chandra
Sent: Friday, April 24, 2015 10:53 AM
To: 'Tvrtko Ursulin'; [email protected]
Cc: Ursulin, Tvrtko
Subject: RE: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 49 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 10 lines · Chandra
Quoted 10 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 8 lines
Quoted 8 lines

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 11 lines
Quoted 11 lines
---
  drivers/gpu/drm/i915/intel_atomic.c | 3 +++
  1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct

drm_device

*dev,

Quoted 8 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 5 lines
  	scaler_state = &crtc_state->scaler_state;
  	drm_state = crtc_state->base.state;
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 6 lines

before and no more required now.

Quoted 2 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Quoted 8 lines

Oh, that's right, ignore my prev comment (with updated scaler design above shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

One more: This change address logs when no scaler is required,
but might comeback when a scaler is active (panel fitting or plane scaling is enabled). I don't know general policy in these kind of situations, but perhaps debug print can be deleted.

Quoted 4 lines
Quoted 4 lines

Regards,

Tvrtko

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

#5 lore raw
Quoted 75 lines · Chandra
Quoted 67 lines

-----Original Message-----
From: Konduru, Chandra
Sent: Friday, April 24, 2015 10:53 AM
To: 'Tvrtko Ursulin'; [email protected]
Cc: Ursulin, Tvrtko
Subject: RE: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 49 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 10 lines · Chandra
Quoted 10 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 8 lines
Quoted 8 lines

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 11 lines
Quoted 11 lines
---
  drivers/gpu/drm/i915/intel_atomic.c | 3 +++
  1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct

drm_device

*dev,

Quoted 8 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 5 lines
  	scaler_state = &crtc_state->scaler_state;
  	drm_state = crtc_state->base.state;
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 6 lines

before and no more required now.

Quoted 2 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Quoted 8 lines

Oh, that's right, ignore my prev comment (with updated scaler design above shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

Quoted 4 lines

One more: This change address logs when no scaler is required,
but might comeback when a scaler is active (panel fitting or plane scaling is enabled). I don't know general policy in these kind of situations, but perhaps debug print can be deleted.

Atomic is really complicated, but doing fully diagnostics for each frame is also way too noisy. For that reason we've add a DRM_DEBUG_ATOMIC which can be used for all these state tracking debug lines.
-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
#6 lore raw
Quoted 76 lines · Daniel Vetter
Quoted 75 lines · Chandra
Quoted 75 lines

-----Original Message-----
From: Konduru, Chandra
Sent: Friday, April 24, 2015 10:53 AM
To: 'Tvrtko Ursulin'; [email protected]
Cc: Ursulin, Tvrtko
Subject: RE: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 49 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 10 lines · Chandra
Quoted 10 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 8 lines
Quoted 8 lines

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 11 lines
Quoted 11 lines
---
   drivers/gpu/drm/i915/intel_atomic.c | 3 +++
   1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct

drm_device

*dev,

Quoted 8 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 5 lines
   	scaler_state = &crtc_state->scaler_state;
   	drm_state = crtc_state->base.state;
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 6 lines

before and no more required now.

Quoted 2 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Oh, that's right, ignore my prev comment (with updated scaler design above shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

Quoted 8 lines
Quoted 4 lines

One more: This change address logs when no scaler is required,
but might comeback when a scaler is active (panel fitting or plane scaling is enabled). I don't know general policy in these kind of situations, but perhaps debug print can be deleted.

Atomic is really complicated, but doing fully diagnostics for each frame is also way too noisy. For that reason we've add a DRM_DEBUG_ATOMIC which can be used for all these state tracking debug lines.

We didn't do anything here and I just noticed kernel is still too spammy with regards to this issue.

Should we just merge my patch? Still looks completely safe to me...

Regards,

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

#7 lore raw
Quoted 83 lines · Tvrtko Ursulin
Quoted 81 lines · Daniel Vetter
Quoted 79 lines · Chandra
Quoted 67 lines

-----Original Message-----
From: Konduru, Chandra
Sent: Friday, April 24, 2015 10:53 AM
To: 'Tvrtko Ursulin'; [email protected]
Cc: Ursulin, Tvrtko
Subject: RE: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 49 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 10 lines · Chandra
Quoted 10 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 8 lines
Quoted 8 lines

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 11 lines
Quoted 11 lines
---
  drivers/gpu/drm/i915/intel_atomic.c | 3 +++
  1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct

drm_device

*dev,

Quoted 8 lines

drivers/gpu/drm/i915/intel_atomic.c int intel_atomic_setup_scalers(struct drm_device *dev, · line 306

Quoted 5 lines
  	scaler_state = &crtc_state->scaler_state;
  	drm_state = crtc_state->base.state;
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 6 lines

before and no more required now.

Quoted 2 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Quoted 8 lines

Oh, that's right, ignore my prev comment (with updated scaler design above shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

One more: This change address logs when no scaler is required,
but might comeback when a scaler is active (panel fitting or plane scaling is enabled). I don't know general policy in these kind of situations, but perhaps debug print can be deleted.

Quoted 8 lines
Quoted 3 lines

Atomic is really complicated, but doing fully diagnostics for each frame is also way too noisy. For that reason we've add a DRM_DEBUG_ATOMIC which can be used for all these state tracking debug lines.

We didn't do anything here and I just noticed kernel is still too spammy with regards to this issue.

Should we just merge my patch? Still looks completely safe to me...

doesn't seem to apply any more:(
-Daniel

Quoted 4 lines

Regards,

Tvrtko

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
#8 lore raw
Quoted 88 lines · Daniel Vetter
Quoted 87 lines · Tvrtko Ursulin
Quoted 84 lines · Daniel Vetter
Quoted 75 lines · Chandra
Quoted 67 lines

-----Original Message-----
From: Konduru, Chandra
Sent: Friday, April 24, 2015 10:53 AM
To: 'Tvrtko Ursulin'; [email protected]
Cc: Ursulin, Tvrtko
Subject: RE: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 49 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 10 lines · Chandra
Quoted 10 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

Quoted 8 lines
Quoted 8 lines

From: Tvrtko Ursulin <[email protected]>

---
Up for discussion I suppose, but like it is, with typical drm.debug = 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 11 lines
Quoted 11 lines
---
   drivers/gpu/drm/i915/intel_atomic.c | 3 +++
   1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct

drm_device

*dev,

Quoted 8 lines

drivers/gpu/drm/i915/intel_atomic.c

Quoted 5 lines
   	scaler_state = &crtc_state->scaler_state;
   	drm_state = crtc_state->base.state;
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+	if (!scaler_state->scaler_users)
+		return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 6 lines

before and no more required now.

Quoted 2 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in scaler_state->scaler_users ?

Quoted 8 lines

Oh, that's right, ignore my prev comment (with updated scaler design above shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

Quoted 4 lines

One more: This change address logs when no scaler is required,
but might comeback when a scaler is active (panel fitting or plane scaling is enabled). I don't know general policy in these kind of situations, but perhaps debug print can be deleted.

Atomic is really complicated, but doing fully diagnostics for each frame is also way too noisy. For that reason we've add a DRM_DEBUG_ATOMIC which can be used for all these state tracking debug lines.

Quoted 6 lines
Quoted 4 lines

We didn't do anything here and I just noticed kernel is still too spammy with regards to this issue.

Should we just merge my patch? Still looks completely safe to me...

doesn't seem to apply any more:(

Yeah only some months have passed, who would have thought. :)

But I realized that would be only one of the three log lines per cursor update - there is a code path calling skl_detach_scaler two times as well. Looks like this overall, per update:

[drm:intel_atomic_setup_scalers] crtc_state = ffff880074b55c00 need = 0 avail = 2 scaler_users = 0x0 [drm:skl_detach_scaler] CRTC:21 Disabled scaler id 0.0
[drm:skl_detach_scaler] CRTC:21 Disabled scaler id 0.1

I'll rebase this patch for a start.

Regards,

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

#9 lore raw
Quoted 117 lines · Tvrtko Ursulin
Quoted 102 lines · Daniel Vetter
Quoted 101 lines · Tvrtko Ursulin
Quoted 98 lines · Daniel Vetter
Quoted 85 lines · Chandra
Quoted 72 lines

-----Original Message-----
From: Konduru, Chandra
Sent: Friday, April 24, 2015 10:53 AM
To: 'Tvrtko Ursulin'; [email protected]
Cc: Ursulin, Tvrtko
Subject: RE: [PATCH] drm/i915/skl: Bypass debug message if scalers are not
requested

Quoted 51 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:34 AM
To: Konduru, Chandra; [email protected]
Cc: Ursulin, Tvrtko
Subject: Re: [PATCH] drm/i915/skl: Bypass debug message if scalers are
not requested

Quoted 12 lines · Chandra
Quoted 12 lines

-----Original Message-----
From: Tvrtko Ursulin [mailto:[email protected]]
Sent: Friday, April 24, 2015 9:08 AM
To: [email protected]
Cc: Ursulin, Tvrtko; Konduru, Chandra
Subject: [PATCH] drm/i915/skl: Bypass debug message if scalers are not requested

From: Tvrtko Ursulin <[email protected]>

Quoted 7 lines
Quoted 7 lines

---
Up for discussion I suppose, but like it is, with typical drm.debug
= 0xe, it logs one line per cursor movement while the log would otherwise be

quiet.

Quoted 11 lines
Quoted 11 lines
---
   drivers/gpu/drm/i915/intel_atomic.c | 3 +++
   1 file changed, 3 insertions(+)
drivers/gpu/drm/i915/intel_atomic.c +0 -0
index 3c4b7cd..7284c6d 100644
@@ -302,6 +302,9 @@ int intel_atomic_setup_scalers(struct

drm_device

*dev,

Quoted 8 lines

drivers/gpu/drm/i915/intel_atomic.c

Quoted 5 lines
       scaler_state = &crtc_state->scaler_state;
       drm_state = crtc_state->base.state;
drivers/gpu/drm/i915/intel_atomic.c +2 -0
+    if (!scaler_state->scaler_users)
+        return 0;

This will cause issue because scalers will never get freed if they are in use

Quoted 8 lines

before and no more required now.

Quoted 3 lines

I put the debug print to help debug variety of state related issues while we are in development, but perhaps debug print can be deleted.

Doesn't the loop below skip everything anyway when no bits are set in
scaler_state->scaler_users ?

Quoted 13 lines

Oh, that's right, ignore my prev comment (with updated scaler design above
shouldn't cause issues).
By the way, can you pls run kms_panel_fit and kms_plane_scaling to make sure
they pass and there are no related warnings or errors in kernel log? For kms_plane_scaling, pls apply two pending patches that aren't merged:
v5 [PATCH 13/14] drm/i915: skylake primary plane scaling using shared scalers
v5 [PATCH 14/14] drm/i915: skylake sprite plane scaling using shared scalers

Quoted 6 lines

One more: This change address logs when no scaler is required, but might comeback when a scaler is active (panel fitting or plane scaling is enabled).
I don't know general policy in these kind of situations, but perhaps debug print
can be deleted.

Atomic is really complicated, but doing fully diagnostics for each frame
is also way too noisy. For that reason we've add a DRM_DEBUG_ATOMIC which
can be used for all these state tracking debug lines.

Quoted 6 lines
Quoted 4 lines

We didn't do anything here and I just noticed kernel is still too spammy with regards to this issue.

Should we just merge my patch? Still looks completely safe to me...

doesn't seem to apply any more:(

Yeah only some months have passed, who would have thought. :)

But I realized that would be only one of the three log lines per cursor update - there is a code path calling skl_detach_scaler two times as well. Looks like this overall, per update:

Quoted 6 lines

[drm:intel_atomic_setup_scalers] crtc_state = ffff880074b55c00 need = 0 avail = 2 scaler_users = 0x0
[drm:skl_detach_scaler] CRTC:21 Disabled scaler id 0.0 [drm:skl_detach_scaler] CRTC:21 Disabled scaler id 0.1

I'll rebase this patch for a start.

Sent it as a new thread but forgot v2 in the subject.

Other source of spam is probably intel_begin_crtc_commit -> skl_detach_scalers.

Don't know what the right fix would be there. Looks like it is not tracking transitions in scaler use so it would be able to act and log when something interesting happens, but rather does it every time. I defer to Chandra. :)

Regards,

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

#10 lore raw

On 08/05/2015 03:19 PM, Tvrtko Ursulin wrote:
[snip]

Quoted 29 lines
Quoted 20 lines
Quoted 6 lines
Quoted 6 lines
Quoted 5 lines

Atomic is really complicated, but doing fully diagnostics for each frame
is also way too noisy. For that reason we've add a DRM_DEBUG_ATOMIC which
can be used for all these state tracking debug lines.

Quoted 7 lines
Quoted 5 lines

We didn't do anything here and I just noticed kernel is still too spammy
with regards to this issue.

Should we just merge my patch? Still looks completely safe to me...

doesn't seem to apply any more:(

Yeah only some months have passed, who would have thought. :)

But I realized that would be only one of the three log lines per cursor update - there is a code path calling skl_detach_scaler two times as well. Looks like this overall, per update:

Quoted 6 lines

[drm:intel_atomic_setup_scalers] crtc_state = ffff880074b55c00 need = 0 avail = 2 scaler_users = 0x0
[drm:skl_detach_scaler] CRTC:21 Disabled scaler id 0.0 [drm:skl_detach_scaler] CRTC:21 Disabled scaler id 0.1

I'll rebase this patch for a start.

Sent it as a new thread but forgot v2 in the subject.

Quoted 7 lines

Other source of spam is probably intel_begin_crtc_commit -> skl_detach_scalers.

Don't know what the right fix would be there. Looks like it is not tracking transitions in scaler use so it would be able to act and log when something interesting happens, but rather does it every time. I defer to Chandra. :)

Ok I've sent something named "[PATCH] drm/i915/skl: Only disable scalers once".

It was only compile tested due my current lack of suitable hardware for testing. And I don't pretend to be confident in the world of atomic states etc. so it may well be completely wrong.

So either scrutinize it a lot please, or let it be a motivator to come up with a proper fix of your own. :)

Regards,

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

Tested-by on v1 1/1
#11 lore raw
In reply to Tvrtko Ursulin

Task id: 6262 -------------------------------------Summary-------------------------------------
Platform Delta drm-intel-nightly Series Applied

PNV                                  276/276              276/276
ILK                                  301/301              301/301
SNB                                  318/318              318/318
IVB                                  345/345              345/345

BYT -2 285/285 283/285
BDW 321/321 321/321
-------------------------------------Detailed-------------------------------------
Platform Test drm-intel-nightly Series Applied
*BYT igt@gem_ctx_basic PASS(3) DMESG_WARN(1)PASS(1)
(dmesg patch applied)drm:check_crtc_state[i915]]*ERROR*mismatch_in_has_infoframe(expected#,found#)@mismatch in has_infoframe .* found WARNING:at_drivers/gpu/drm/i915/intel_display.c:#check_crtc_state[i915]()@WARNING:.* at .* check_crtc_state+0x
*BYT igt@kms_setmode@invalid-clone-exclusive-crtc PASS(6) DMESG_WARN(1)PASS(1)
(dmesg patch applied)drm:check_crtc_state[i915]]*ERROR*mismatch_in_has_infoframe(expected#,found#)@mismatch in has_infoframe .* found WARNING:at_drivers/gpu/drm/i915/intel_display.c:#check_crtc_state[i915]()@WARNING:.* at .* check_crtc_state+0x
Note: You need to pay more attention to line start with '*'
_______________________________________________
Intel-gfx mailing list
[email protected]
http://lists.freedesktop.org/mailman/listinfo/intel-gfx