Skip to contentSkip to conversation

drm/atomic-helpers: Fix CRTC primary-plane test

PATCH Superseded Thomas Zimmermann 3 participants 11 messages Oct 5, 2022 – Oct 7, 2022

Subsystems (5)

v4 of 5 Superseded Latest is v5 →

lore
Share & tools
lore

Thomas Zimmermann

author maintainer
#0 lore raw

v4 cover drm/atomic-helpers: Fix CRTC primary-plane test

(was: drm/atomic-helper: Don't allocated plane state in CRTC check)

Fix the test for a CRTC's primary plane and clean up the test function to only do the test. Up to v3, this patchset was a single patch, but the cleanup turns it into a series.

v4:
* clean up the helper function (Ville)

Thomas Zimmermann (2):
drm/atomic-helper: Don't allocated plane state in CRTC check drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

 drivers/gpu/drm/ast/ast_mode.c          |  8 ++--
 drivers/gpu/drm/drm_atomic_helper.c     | 60 ++++++++-----------------
 drivers/gpu/drm/drm_simple_kms_helper.c |  6 ++-
 drivers/gpu/drm/mgag200/mgag200_mode.c  |  8 ++--
 drivers/gpu/drm/solomon/ssd130x.c       |  6 ++-
 drivers/gpu/drm/tiny/simpledrm.c        |  6 ++-
 include/drm/drm_atomic_helper.h         |  3 +-
 7 files changed, 42 insertions(+), 55 deletions(-)
Signature · 1 line
2.37.3
author maintainer
#4 lore raw

PATCH v4 2/2 drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

Rename the atomic helper function drm_atomic_helper_check_crtc_state() to drm_atomic_helper_check_crtc_primary_plane() and only check for an attached primary plane. Adapt callers.

Instead of having one big function to check for various CRTC state conditions, we rather want smaller functions that drivers can pick individually.

---
 drivers/gpu/drm/ast/ast_mode.c          |  8 ++--
 drivers/gpu/drm/drm_atomic_helper.c     | 52 +++++++++----------------
 drivers/gpu/drm/drm_simple_kms_helper.c |  6 ++-
 drivers/gpu/drm/mgag200/mgag200_mode.c  |  8 ++--
 drivers/gpu/drm/solomon/ssd130x.c       |  6 ++-
 drivers/gpu/drm/tiny/simpledrm.c        |  6 ++-
 include/drm/drm_atomic_helper.h         |  3 +-
 7 files changed, 42 insertions(+), 47 deletions(-)
drivers/gpu/drm/ast/ast_mode.c +4 -4
index 1bc0220e6783..d598b1e1447f 100644
@@ -1129,13 +1129,13 @@ static int ast_crtc_helper_atomic_check(struct drm_crtc *crtc,
 	bool succ;
 	int ret;
 
-	ret = drm_atomic_helper_check_crtc_state(crtc_state, false);
-	if (ret)
-		return ret;
-
 	if (!crtc_state->enable)
 		goto out;
 
+	ret = drm_atomic_helper_check_crtc_primary_plane(crtc_state);
+	if (ret)
+		return ret;
+
 	ast_state = to_ast_crtc_state(crtc_state);
 
 	format = ast_state->format;
drivers/gpu/drm/drm_atomic_helper.c +18 -34

69 lines not on this page · raw message

drivers/gpu/drm/drm_simple_kms_helper.c +5 -1
index e9f782119d3d..31233c6ae3c4 100644
@@ -102,10 +102,14 @@ static int drm_simple_kms_crtc_check(struct drm_crtc *crtc,
 	struct drm_crtc_state *crtc_state = drm_atomic_get_new_crtc_state(state, crtc);
 	int ret;
 
-	ret = drm_atomic_helper_check_crtc_state(crtc_state, false);
+	if (!crtc_state->enable)
+		goto out;
+
+	ret = drm_atomic_helper_check_crtc_primary_plane(crtc_state);
 	if (ret)
 		return ret;
 
+out:
 	return drm_atomic_add_affected_planes(state, crtc);
 }
 
drivers/gpu/drm/mgag200/mgag200_mode.c +4 -4
index bbab2549243a..5f7eb642f0c6 100644
@@ -579,13 +579,13 @@ int mgag200_crtc_helper_atomic_check(struct drm_crtc *crtc, struct drm_atomic_st
 	struct drm_property_blob *new_gamma_lut = new_crtc_state->gamma_lut;
 	int ret;
 
-	ret = drm_atomic_helper_check_crtc_state(new_crtc_state, false);
-	if (ret)
-		return ret;
-
 	if (!new_crtc_state->enable)
 		return 0;
 
+	ret = drm_atomic_helper_check_crtc_primary_plane(new_crtc_state);
+	if (ret)
+		return ret;
+
 	if (new_crtc_state->mode_changed) {
 		if (funcs->pixpllc_atomic_check) {
 			ret = funcs->pixpllc_atomic_check(crtc, new_state);
drivers/gpu/drm/solomon/ssd130x.c +5 -1
index bc41a5ae810a..473db3dc0d55 100644
@@ -648,10 +648,14 @@ static int ssd130x_crtc_helper_atomic_check(struct drm_crtc *crtc,
 	struct drm_crtc_state *new_crtc_state = drm_atomic_get_new_crtc_state(new_state, crtc);
 	int ret;
 
-	ret = drm_atomic_helper_check_crtc_state(new_crtc_state, false);
+	if (!new_crtc_state->enable)
+		goto out;
+
+	ret = drm_atomic_helper_check_crtc_primary_plane(new_crtc_state);
 	if (ret)
 		return ret;
 
+out:
 	return drm_atomic_add_affected_planes(new_state, crtc);
 }
 
drivers/gpu/drm/tiny/simpledrm.c +5 -1
index 18489779fb8a..ecd49a8f3334 100644
@@ -551,10 +551,14 @@ static int simpledrm_crtc_helper_atomic_check(struct drm_crtc *crtc,
 	struct drm_crtc_state *new_crtc_state = drm_atomic_get_new_crtc_state(new_state, crtc);
 	int ret;
 
-	ret = drm_atomic_helper_check_crtc_state(new_crtc_state, false);
+	if (!new_crtc_state->enable)
+		goto out;
+
+	ret = drm_atomic_helper_check_crtc_primary_plane(new_crtc_state);
 	if (ret)
 		return ret;
 
+out:
 	return drm_atomic_add_affected_planes(new_state, crtc);
 }
 
include/drm/drm_atomic_helper.h +1 -2
index 06d8902a8097..33f982cd1a27 100644
@@ -58,10 +58,9 @@ int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state,
 					int max_scale,
 					bool can_position,
 					bool can_update_disabled);
-int drm_atomic_helper_check_crtc_state(struct drm_crtc_state *crtc_state,
-				       bool can_disable_primary_plane);
 int drm_atomic_helper_check_planes(struct drm_device *dev,
 			       struct drm_atomic_state *state);
+int drm_atomic_helper_check_crtc_primary_plane(struct drm_crtc_state *crtc_state);
 int drm_atomic_helper_check(struct drm_device *dev,
 			    struct drm_atomic_state *state);
 void drm_atomic_helper_commit_tail(struct drm_atomic_state *state);
Signature · 1 line
2.37.3
author maintainer
#1 lore raw

PATCH v4 1/2 drm/atomic-helper: Don't allocated plane state in CRTC check

In reply to Thomas Zimmermann

In drm_atomic_helper_check_crtc_state(), do not add a new plane state to the global state if it does not exist already. Adding a new plane state will result in overhead for the plane during the atomic-commit step.

For the test in drm_atomic_helper_check_crtc_state() to succeed, it is important that the CRTC has an enabled primary plane after the commit. Simply testing the CRTC state's plane_mask for a primary plane is sufficient.

Note that the helper still only tests for an attached primary plane. Drivers have to ensure that the plane contains valid pixel information.

v3:
* test for a primary plane in plane_mask (Ville)
v2:
* remove unnecessary test for plane->crtc (Ville)
* inline drm_atomic_get_next_plane_state() (Ville)
* acquire plane lock before accessing plane->state (Ville)

Fixes d6b9af1097fe ("drm/atomic-helper: Add helper drm_atomic_helper_check_crtc_state()")
Cc Thomas Zimmermann <[email protected]>
Cc Jocelyn Falempe <[email protected]>
Cc Maarten Lankhorst <[email protected]>
Cc Maxime Ripard <[email protected]>
Cc David Airlie <[email protected]>
Cc Daniel Vetter <[email protected]>
---
 drivers/gpu/drm/drm_atomic_helper.c | 10 +---------
 1 file changed, 1 insertion(+), 9 deletions(-)
drivers/gpu/drm/drm_atomic_helper.c +1 -9
index 98cc3137c062..02b4a7dc92f5 100644
@@ -945,7 +945,6 @@ int drm_atomic_helper_check_crtc_state(struct drm_crtc_state *crtc_state,
 				       bool can_disable_primary_planes)
 {
 	struct drm_device *dev = crtc_state->crtc->dev;
-	struct drm_atomic_state *state = crtc_state->state;
 
 	if (!crtc_state->enable)
 		return 0;
@@ -956,14 +955,7 @@ int drm_atomic_helper_check_crtc_state(struct drm_crtc_state *crtc_state,
 		struct drm_plane *plane;
 
 		drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
-			struct drm_plane_state *plane_state;
-
-			if (plane->type != DRM_PLANE_TYPE_PRIMARY)
-				continue;
-			plane_state = drm_atomic_get_plane_state(state, plane);
-			if (IS_ERR(plane_state))
-				return PTR_ERR(plane_state);
-			if (plane_state->fb && plane_state->crtc) {
+			if (plane->type == DRM_PLANE_TYPE_PRIMARY) {
 				has_primary_plane = true;
 				break;
 			}
Signature · 1 line
2.37.3
maintainer Reviewed-by on v4 1/2
#2 lore raw

Re: [PATCH v4 1/2] drm/atomic-helper: Don't allocated plane state in CRTC check

Hello Thomas,

Quoted 20 lines · Thomas Zimmermann

In drm_atomic_helper_check_crtc_state(), do not add a new plane state to the global state if it does not exist already. Adding a new plane state will result in overhead for the plane during the atomic-commit step.

For the test in drm_atomic_helper_check_crtc_state() to succeed, it is important that the CRTC has an enabled primary plane after the commit. Simply testing the CRTC state's plane_mask for a primary plane is sufficient.

Note that the helper still only tests for an attached primary plane. Drivers have to ensure that the plane contains valid pixel information.

v3:
* test for a primary plane in plane_mask (Ville)
v2:
* remove unnecessary test for plane->crtc (Ville)
* inline drm_atomic_get_next_plane_state() (Ville)
* acquire plane lock before accessing plane->state (Ville)

Quoted 2 lines
Fixes d6b9af1097fe ("drm/atomic-helper: Add helper drm_atomic_helper_check_crtc_state()")

This patch makes sense to me.

but I've a hard time parsing the subject line. Did you mean instead:

"drm/atomic-helper: Don't allocate new plane state in CRTC check" ?

or "drm/atomic-helper: Don't add a new plane state in CRTC check" ?

In any case you can fix that while applying so no need to resend IMO.

Signature · 5 lines
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat
maintainer Reviewed-by on v4 2/2
#5 lore raw

Re: [PATCH v4 2/2] drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

In reply to Thomas Zimmermann
Quoted 10 lines · Thomas Zimmermann

Rename the atomic helper function drm_atomic_helper_check_crtc_state() to drm_atomic_helper_check_crtc_primary_plane() and only check for an attached primary plane. Adapt callers.

Instead of having one big function to check for various CRTC state conditions, we rather want smaller functions that drivers can pick individually.

---

[...]

drivers/gpu/drm/drm_atomic_helper.c int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state, · line 949

Quoted 4 lines
drivers/gpu/drm/drm_atomic_helper.c +3 -0
+	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
+		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+			return 0;
 	}

I believe the code convention is to drop the curly braces when you have a single statement inside the a loop ?

Feel free to ignore it though. I particularly don't agree with that convention anyways, because I think that makes the code more error prone. But still thought that was worth to point that out.

Signature · 5 lines
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat
author maintainer
#3 lore raw

Re: [PATCH v4 1/2] drm/atomic-helper: Don't allocated plane state in CRTC check

In reply to Javier Martinez Canillas

Hi

Am 06.10.22 um 22:18 schrieb Javier Martinez Canillas:

Quoted 26 lines

Hello Thomas,

Quoted 20 lines · Thomas Zimmermann

In drm_atomic_helper_check_crtc_state(), do not add a new plane state to the global state if it does not exist already. Adding a new plane state will result in overhead for the plane during the atomic-commit step.

For the test in drm_atomic_helper_check_crtc_state() to succeed, it is important that the CRTC has an enabled primary plane after the commit. Simply testing the CRTC state's plane_mask for a primary plane is sufficient.

Note that the helper still only tests for an attached primary plane. Drivers have to ensure that the plane contains valid pixel information.

v3:
* test for a primary plane in plane_mask (Ville)
v2:
* remove unnecessary test for plane->crtc (Ville)
* inline drm_atomic_get_next_plane_state() (Ville)
* acquire plane lock before accessing plane->state (Ville)

Quoted 2 lines
Fixes d6b9af1097fe ("drm/atomic-helper: Add helper drm_atomic_helper_check_crtc_state()")
Quoted 7 lines

This patch makes sense to me.

but I've a hard time parsing the subject line. Did you mean instead:

"drm/atomic-helper: Don't allocate new plane state in CRTC check" ?

Ok, I'll do that.

Best regard
Thomas

Quoted 5 lines

or "drm/atomic-helper: Don't add a new plane state in CRTC check" ?

In any case you can fix that while applying so no need to resend IMO.

Signature · 6 lines
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Ivo Totev
#6 lore raw

Re: [PATCH v4 2/2] drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

In reply to Javier Martinez Canillas
Quoted 16 lines · Javier Martinez Canillas
Quoted 10 lines · Thomas Zimmermann

Rename the atomic helper function drm_atomic_helper_check_crtc_state() to drm_atomic_helper_check_crtc_primary_plane() and only check for an attached primary plane. Adapt callers.

Instead of having one big function to check for various CRTC state conditions, we rather want smaller functions that drivers can pick individually.

---

[...]

Quoted 7 lines

drivers/gpu/drm/drm_atomic_helper.c int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state, · line 949

Quoted 4 lines
drivers/gpu/drm/drm_atomic_helper.c +3 -0
+	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
+		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+			return 0;
 	}

I believe the code convention is to drop the curly braces when you have a single statement inside the a loop ?

This has two.

Quoted 11 lines

Feel free to ignore it though. I particularly don't agree with that convention anyways, because I think that makes the code more error prone. But still thought that was worth to point that out.

Signature · 5 lines
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat
Signature · 2 lines
Ville Syrjälä
Intel
#7 lore raw

Re: [PATCH v4 2/2] drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

Quoted 25 lines · Ville Syrjälä
Quoted 16 lines · Javier Martinez Canillas
Quoted 10 lines · Thomas Zimmermann

Rename the atomic helper function drm_atomic_helper_check_crtc_state() to drm_atomic_helper_check_crtc_primary_plane() and only check for an attached primary plane. Adapt callers.

Instead of having one big function to check for various CRTC state conditions, we rather want smaller functions that drivers can pick individually.

---

[...]

Quoted 7 lines

drivers/gpu/drm/drm_atomic_helper.c int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state, · line 949

Quoted 4 lines
drivers/gpu/drm/drm_atomic_helper.c +3 -0
+	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
+		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+			return 0;
 	}

I believe the code convention is to drop the curly braces when you have a single statement inside the a loop ?

Quoted 2 lines

This has two.

No, it has only one that is the if statement. So according to the Linux kernel coding style AFAIU it should be written as:

	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask)
		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
			return 0;
Signature · 5 lines
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat
#8 lore raw

Re: [PATCH v4 2/2] drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

Quoted 29 lines · Javier Martinez Canillas
Quoted 25 lines · Ville Syrjälä
Quoted 16 lines · Javier Martinez Canillas
Quoted 10 lines · Thomas Zimmermann

Rename the atomic helper function drm_atomic_helper_check_crtc_state() to drm_atomic_helper_check_crtc_primary_plane() and only check for an attached primary plane. Adapt callers.

Instead of having one big function to check for various CRTC state conditions, we rather want smaller functions that drivers can pick individually.

---

[...]

Quoted 7 lines

drivers/gpu/drm/drm_atomic_helper.c int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state, · line 949

Quoted 4 lines
drivers/gpu/drm/drm_atomic_helper.c +3 -0
+	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
+		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+			return 0;
 	}

I believe the code convention is to drop the curly braces when you have a single statement inside the a loop ?

Quoted 2 lines

This has two.

Quoted 6 lines

No, it has only one that is the if statement. So according to the Linux kernel coding style AFAIU it should be written as:

	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask)
		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
			return 0;

That is exactly what it says not to do.

Signature · 2 lines
Ville Syrjälä
Intel
#9 lore raw

Re: [PATCH v4 2/2] drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

On 10/7/22 09:29, Ville Syrjälä wrote:

[...]

Quoted 6 lines
Quoted 3 lines
	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask)
		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
			return 0;

That is exactly what it says not to do.

Ah, good to know. I re-read the kernel coding style and see now that it only applies to single simple statements.

Better then, since not using braces even for single statements is something that's error prone IMO, as mentioned before.

Signature · 5 lines
Best regards,

Javier Martinez Canillas
Core Platforms
Red Hat
author maintainer
#10 lore raw

Re: [PATCH v4 2/2] drm/atomic-helper: Replace drm_atomic_helper_check_crtc_state()

In reply to Ville Syrjälä

Hi

Am 07.10.22 um 09:29 schrieb Ville Syrjälä:

Quoted 30 lines · Javier Martinez Canillas
Quoted 29 lines · Ville Syrjälä
Quoted 27 lines · Javier Martinez Canillas
Quoted 16 lines · Thomas Zimmermann
Quoted 10 lines

Rename the atomic helper function drm_atomic_helper_check_crtc_state() to drm_atomic_helper_check_crtc_primary_plane() and only check for an attached primary plane. Adapt callers.

Instead of having one big function to check for various CRTC state conditions, we rather want smaller functions that drivers can pick individually.

---

[...]

Quoted 7 lines

drivers/gpu/drm/drm_atomic_helper.c int drm_atomic_helper_check_plane_state(struct drm_plane_state *plane_state, · line 949

Quoted 4 lines
drivers/gpu/drm/drm_atomic_helper.c +3 -0
+	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask) {
+		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
+			return 0;
  	}

I believe the code convention is to drop the curly braces when you have a single statement inside the a loop ?

This has two.

Quoted 8 lines
Quoted 6 lines

No, it has only one that is the if statement. So according to the Linux kernel coding style AFAIU it should be written as:

	drm_for_each_plane_mask(plane, dev, crtc_state->plane_mask)
		if (plane->type == DRM_PLANE_TYPE_PRIMARY)
			return 0;

That is exactly what it says not to do.

Hey, no need to be so upfront about it. Without the outer braces, I'd find it hard to parse anyway.

Best regard
Thomas

Quoted 1 line
Signature · 6 lines
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Ivo Totev