[PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[Thomas Zimmermann]

The atomic modesetting code tried to distinguish format changes from
full modesetting operations. But the implementation was buggy and the
format registers were often updated even for simple pageflips.

Fix this problem by distinguishing between format and mode changes; and
handle each in it's own branch.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: <stable@vger.kernel.org> # v5.6+
---
 drivers/gpu/drm/ast/ast_mode.c | 52 ++++++++++++++++------------------
 1 file changed, 25 insertions(+), 27 deletions(-)

diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index 62fe682a7de6..b129833c0821 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -768,34 +768,32 @@ static void ast_crtc_helper_atomic_flush(struct drm_crtc *crtc,
 {
 	struct drm_device *dev = crtc->dev;
 	struct ast_private *ast = to_ast_private(dev);
-	struct ast_crtc_state *ast_state;
-	const struct drm_format_info *format;
-	struct ast_vbios_mode_info *vbios_mode_info;
-	struct drm_display_mode *adjusted_mode;
-
-	ast_state = to_ast_crtc_state(crtc->state);
-
-	format = ast_state->format;
-	if (!format)
-		return;
-
-	vbios_mode_info = &ast_state->vbios_mode_info;
-
-	ast_set_color_reg(ast, format);
-	ast_set_vbios_color_reg(ast, format, vbios_mode_info);
-
-	if (!crtc->state->mode_changed)
-		return;
-
-	adjusted_mode = &crtc->state->adjusted_mode;
+	struct drm_crtc_state *crtc_state = crtc->state;
+	struct ast_crtc_state *ast_crtc_state = to_ast_crtc_state(crtc_state);
+	struct ast_crtc_state *old_ast_crtc_state =
+		to_ast_crtc_state(old_crtc_state);
+	const struct drm_format_info *format = ast_crtc_state->format;
+	struct ast_vbios_mode_info *vbios_mode_info =
+		&ast_crtc_state->vbios_mode_info;
+	struct drm_display_mode *adjusted_mode = &crtc_state->adjusted_mode;
+
+	if (drm_WARN_ON_ONCE(dev, !format))
+		return; /* BUG: We didn't set format in primary check(). */
+
+	if (format != old_ast_crtc_state->format) {
+		ast_set_color_reg(ast, format);
+		ast_set_vbios_color_reg(ast, format, vbios_mode_info);
+	}
 
-	ast_set_vbios_mode_reg(ast, adjusted_mode, vbios_mode_info);
-	ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa1, 0x06);
-	ast_set_std_reg(ast, adjusted_mode, vbios_mode_info);
-	ast_set_crtc_reg(ast, adjusted_mode, vbios_mode_info);
-	ast_set_dclk_reg(ast, adjusted_mode, vbios_mode_info);
-	ast_set_crtthd_reg(ast);
-	ast_set_sync_reg(ast, adjusted_mode, vbios_mode_info);
+	if (drm_atomic_crtc_needs_modeset(crtc_state)) {
+		ast_set_vbios_mode_reg(ast, adjusted_mode, vbios_mode_info);
+		ast_set_index_reg(ast, AST_IO_CRTC_PORT, 0xa1, 0x06);
+		ast_set_std_reg(ast, adjusted_mode, vbios_mode_info);
+		ast_set_crtc_reg(ast, adjusted_mode, vbios_mode_info);
+		ast_set_dclk_reg(ast, adjusted_mode, vbios_mode_info);
+		ast_set_crtthd_reg(ast);
+		ast_set_sync_reg(ast, adjusted_mode, vbios_mode_info);
+	}
 }
 
 static void
-- 
2.28.0

Re: [PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.13.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.13: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.14.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.14: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.1, v5.7.15.

v5.8.1: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.15: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.2, v5.7.16.

v5.8.2: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.16: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 1/4] drm/ast: Only set format registers if primary plane's format changes

[Thomas Zimmermann]

[-- Attachment #1.1: Type: text/plain, Size: 1214 bytes --]

Hi

Am 26.08.20 um 15:54 schrieb Sasha Levin:
> Hi
> 
> [This is an automated email]
> 
> This commit has been processed because it contains a "Fixes:" tag
> fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").
> 
> The bot has tested the following trees: v5.8.2, v5.7.16.
> 
> v5.8.2: Failed to apply! Possible dependencies:
>     05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
>     fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")
> 
> v5.7.16: Failed to apply! Possible dependencies:
>     05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
>     3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
>     fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")
> 
> 
> NOTE: The patch will not be queued to stable trees until it is upstream.
> 
> How should we proceed with this patch?

Please drop this patch and the rest of the series.

> 

-- 
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Felix Imendörffer


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 516 bytes --]

[PATCH v1 2/4] drm/ast: Set display mode in atomic_begin()

[Thomas Zimmermann]

Move the mode-setting code from atomic_flush() to atomic_begin(), and
thus before the plane update. With the CRTC update before the plane
updates, the cursor can be disabled while the mode is being changed.

The patch removes the call ast_open_key() from atomic_begin(), The
function unlocks ast's extended display registers; something that has
been done at this point already.

The now-empty implementation of atomic_flush() is also being removed.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: <stable@vger.kernel.org> # v5.6+
---
 drivers/gpu/drm/ast/ast_mode.c | 9 ---------
 1 file changed, 9 deletions(-)

diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index b129833c0821..0ea8a68ac2d9 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -757,14 +757,6 @@ static int ast_crtc_helper_atomic_check(struct drm_crtc *crtc,
 
 static void ast_crtc_helper_atomic_begin(struct drm_crtc *crtc,
 					 struct drm_crtc_state *old_crtc_state)
-{
-	struct ast_private *ast = to_ast_private(crtc->dev);
-
-	ast_open_key(ast);
-}
-
-static void ast_crtc_helper_atomic_flush(struct drm_crtc *crtc,
-					 struct drm_crtc_state *old_crtc_state)
 {
 	struct drm_device *dev = crtc->dev;
 	struct ast_private *ast = to_ast_private(dev);
@@ -813,7 +805,6 @@ ast_crtc_helper_atomic_disable(struct drm_crtc *crtc,
 static const struct drm_crtc_helper_funcs ast_crtc_helper_funcs = {
 	.atomic_check = ast_crtc_helper_atomic_check,
 	.atomic_begin = ast_crtc_helper_atomic_begin,
-	.atomic_flush = ast_crtc_helper_atomic_flush,
 	.atomic_enable = ast_crtc_helper_atomic_enable,
 	.atomic_disable = ast_crtc_helper_atomic_disable,
 };
-- 
2.28.0

Re: [PATCH v1 2/4] drm/ast: Set display mode in atomic_begin()

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.13.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.13: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 2/4] drm/ast: Set display mode in atomic_begin()

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.14.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.14: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 2/4] drm/ast: Set display mode in atomic_begin()

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.1, v5.7.15.

v5.8.1: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.15: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 2/4] drm/ast: Set display mode in atomic_begin()

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.2, v5.7.16.

v5.8.2: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.16: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

[PATCH v1 3/4] drm/ast: Add commit-tail function

[Thomas Zimmermann]

Duplicates drm_atomic_helper_commit_tail(), so that planes can be
disabled on full mode-setting changes.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: <stable@vger.kernel.org> # v5.6+
---
 drivers/gpu/drm/ast/ast_mode.c | 27 +++++++++++++++++++++++++++
 1 file changed, 27 insertions(+)

diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index 0ea8a68ac2d9..ae5cb0a333f7 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -1043,6 +1043,31 @@ static int ast_connector_init(struct drm_device *dev)
  * Mode config
  */
 
+static void
+ast_mode_config_helper_commit_tail(struct drm_atomic_state *old_state)
+{
+	struct drm_device *dev = old_state->dev;
+
+	drm_atomic_helper_commit_modeset_disables(dev, old_state);
+
+	drm_atomic_helper_commit_planes(dev, old_state, 0);
+
+	drm_atomic_helper_commit_modeset_enables(dev, old_state);
+
+	drm_atomic_helper_fake_vblank(old_state);
+
+	drm_atomic_helper_commit_hw_done(old_state);
+
+	drm_atomic_helper_wait_for_vblanks(dev, old_state);
+
+	drm_atomic_helper_cleanup_planes(dev, old_state);
+}
+
+static const struct drm_mode_config_helper_funcs
+ast_mode_config_helper_funcs = {
+	.atomic_commit_tail = ast_mode_config_helper_commit_tail,
+};
+
 static const struct drm_mode_config_funcs ast_mode_config_funcs = {
 	.fb_create = drm_gem_fb_create,
 	.mode_valid = drm_vram_helper_mode_valid,
@@ -1082,6 +1107,8 @@ int ast_mode_config_init(struct ast_private *ast)
 		dev->mode_config.max_height = 1200;
 	}
 
+	dev->mode_config.helper_private = &ast_mode_config_helper_funcs;
+
 	memset(&ast->primary_plane, 0, sizeof(ast->primary_plane));
 	ret = drm_universal_plane_init(dev, &ast->primary_plane, 0x01,
 				       &ast_primary_plane_funcs,
-- 
2.28.0

Re: [PATCH v1 3/4] drm/ast: Add commit-tail function

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.13.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.13: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 3/4] drm/ast: Add commit-tail function

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.14.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.14: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 3/4] drm/ast: Add commit-tail function

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.1, v5.7.15.

v5.8.1: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.15: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 3/4] drm/ast: Add commit-tail function

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.2, v5.7.16.

v5.8.2: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.16: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

[PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Thomas Zimmermann]

The ast HW cursor requires the primary plane and CRTC to display at
a valid mode and format. This is not the case while switching
display modes, which can lead to the screen turing permanently dark.

As a workaround, the ast driver now disables active planes while the
mode or format switch takes place. It also synchronizes with the vertical
refresh to give CRTC and planes some time to catch up on each other.
The active planes planes (primary or cursor) will be re-enabled by
each plane's atomic_update() function.

v2:
	* move the logic into the commit-tail function

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Sam Ravnborg <sam@ravnborg.org>
Cc: Emil Velikov <emil.l.velikov@gmail.com>
Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
Cc: <stable@vger.kernel.org> # v5.6+
---
 drivers/gpu/drm/ast/ast_drv.h  |  2 +
 drivers/gpu/drm/ast/ast_mode.c | 68 ++++++++++++++++++++++++++++++++--
 2 files changed, 66 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
index c1af6b725933..467049ca8430 100644
--- a/drivers/gpu/drm/ast/ast_drv.h
+++ b/drivers/gpu/drm/ast/ast_drv.h
@@ -177,6 +177,8 @@ struct ast_private *ast_device_create(struct drm_driver *drv,
 
 #define AST_IO_MM_OFFSET		(0x380)
 
+#define AST_IO_VGAIR1_VREFRESH		BIT(3)
+
 #define __ast_read(x) \
 static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
 u##x val = 0;\
diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
index ae5cb0a333f7..a379d51f3543 100644
--- a/drivers/gpu/drm/ast/ast_mode.c
+++ b/drivers/gpu/drm/ast/ast_mode.c
@@ -514,6 +514,17 @@ static void ast_set_start_address_crt1(struct ast_private *ast,
 
 }
 
+static void ast_wait_for_vretrace(struct ast_private *ast)
+{
+	unsigned long timeout = jiffies + HZ;
+	u8 vgair1;
+
+	do {
+		vgair1 = ast_io_read8(ast, AST_IO_INPUT_STATUS1_READ);
+	} while (!(vgair1 & AST_IO_VGAIR1_VREFRESH) &&
+		 time_before(jiffies, timeout));
+}
+
 /*
  * Primary plane
  */
@@ -1043,23 +1054,72 @@ static int ast_connector_init(struct drm_device *dev)
  * Mode config
  */
 
+static bool
+ast_crtc_needs_planes_disabled(struct drm_crtc_state *old_crtc_state,
+			       struct drm_crtc_state *new_crtc_state)
+{
+	struct ast_crtc_state *old_ast_crtc_state, *new_ast_crtc_state;
+
+	if (drm_atomic_crtc_needs_modeset(new_crtc_state))
+		return true;
+
+	old_ast_crtc_state = to_ast_crtc_state(old_crtc_state);
+	new_ast_crtc_state = to_ast_crtc_state(new_crtc_state);
+
+	if (old_ast_crtc_state->format != new_ast_crtc_state->format)
+		return true;
+
+	return false;
+}
+
 static void
 ast_mode_config_helper_commit_tail(struct drm_atomic_state *old_state)
 {
 	struct drm_device *dev = old_state->dev;
+	struct ast_private *ast = to_ast_private(dev);
+	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
+	struct drm_crtc *crtc;
+	int i;
+	bool wait_for_vretrace = false;
 
 	drm_atomic_helper_commit_modeset_disables(dev, old_state);
 
-	drm_atomic_helper_commit_planes(dev, old_state, 0);
+	/*
+	 * HW cursors require the underlying primary plane and CRTC to
+	 * display a valid mode and image. This is not the case during
+	 * full modeset operations. So we temporarily disable any active
+	 * plane, including the HW cursor. Each plane's atomic_update()
+	 * helper will re-enable it if necessary.
+	 *
+	 * We only do this during *full* modesets. It does not affect
+	 * simple pageflips on the planes.
+	 */
+	for_each_oldnew_crtc_in_state(old_state, crtc,
+				      old_crtc_state,
+				      new_crtc_state, i) {
+		if (!ast_crtc_needs_planes_disabled(old_crtc_state,
+						    new_crtc_state))
+			continue;
+		drm_atomic_helper_disable_planes_on_crtc(old_crtc_state,
+							 false);
+		wait_for_vretrace = true;
+	}
+
+	/*
+	 * Ensure that no scanout takes place before reprogramming mode
+	 * and format registers.
+	 */
+	if (wait_for_vretrace)
+		ast_wait_for_vretrace(ast);
+
+	drm_atomic_helper_commit_planes(dev, old_state,
+					DRM_PLANE_COMMIT_ACTIVE_ONLY);
 
 	drm_atomic_helper_commit_modeset_enables(dev, old_state);
 
 	drm_atomic_helper_fake_vblank(old_state);
-
 	drm_atomic_helper_commit_hw_done(old_state);
-
 	drm_atomic_helper_wait_for_vblanks(dev, old_state);
-
 	drm_atomic_helper_cleanup_planes(dev, old_state);
 }
 
-- 
2.28.0

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[daniel]

On Wed, Aug 05, 2020 at 12:54:28PM +0200, Thomas Zimmermann wrote:
> The ast HW cursor requires the primary plane and CRTC to display at
> a valid mode and format. This is not the case while switching
> display modes, which can lead to the screen turing permanently dark.
> 
> As a workaround, the ast driver now disables active planes while the
> mode or format switch takes place. It also synchronizes with the vertical
> refresh to give CRTC and planes some time to catch up on each other.
> The active planes planes (primary or cursor) will be re-enabled by
> each plane's atomic_update() function.
> 
> v2:
> 	* move the logic into the commit-tail function
> 
> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
> Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
> Cc: Thomas Zimmermann <tzimmermann@suse.de>
> Cc: Gerd Hoffmann <kraxel@redhat.com>
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> Cc: Sam Ravnborg <sam@ravnborg.org>
> Cc: Emil Velikov <emil.l.velikov@gmail.com>
> Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
> Cc: <stable@vger.kernel.org> # v5.6+
> ---
>  drivers/gpu/drm/ast/ast_drv.h  |  2 +
>  drivers/gpu/drm/ast/ast_mode.c | 68 ++++++++++++++++++++++++++++++++--
>  2 files changed, 66 insertions(+), 4 deletions(-)
> 
> diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
> index c1af6b725933..467049ca8430 100644
> --- a/drivers/gpu/drm/ast/ast_drv.h
> +++ b/drivers/gpu/drm/ast/ast_drv.h
> @@ -177,6 +177,8 @@ struct ast_private *ast_device_create(struct drm_driver *drv,
>  
>  #define AST_IO_MM_OFFSET		(0x380)
>  
> +#define AST_IO_VGAIR1_VREFRESH		BIT(3)
> +
>  #define __ast_read(x) \
>  static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
>  u##x val = 0;\
> diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
> index ae5cb0a333f7..a379d51f3543 100644
> --- a/drivers/gpu/drm/ast/ast_mode.c
> +++ b/drivers/gpu/drm/ast/ast_mode.c
> @@ -514,6 +514,17 @@ static void ast_set_start_address_crt1(struct ast_private *ast,
>  
>  }
>  
> +static void ast_wait_for_vretrace(struct ast_private *ast)
> +{
> +	unsigned long timeout = jiffies + HZ;
> +	u8 vgair1;
> +
> +	do {
> +		vgair1 = ast_io_read8(ast, AST_IO_INPUT_STATUS1_READ);
> +	} while (!(vgair1 & AST_IO_VGAIR1_VREFRESH) &&
> +		 time_before(jiffies, timeout));
> +}
> +
>  /*
>   * Primary plane
>   */
> @@ -1043,23 +1054,72 @@ static int ast_connector_init(struct drm_device *dev)
>   * Mode config
>   */
>  
> +static bool
> +ast_crtc_needs_planes_disabled(struct drm_crtc_state *old_crtc_state,
> +			       struct drm_crtc_state *new_crtc_state)
> +{
> +	struct ast_crtc_state *old_ast_crtc_state, *new_ast_crtc_state;
> +
> +	if (drm_atomic_crtc_needs_modeset(new_crtc_state))
> +		return true;
> +
> +	old_ast_crtc_state = to_ast_crtc_state(old_crtc_state);
> +	new_ast_crtc_state = to_ast_crtc_state(new_crtc_state);
> +
> +	if (old_ast_crtc_state->format != new_ast_crtc_state->format)
> +		return true;
> +
> +	return false;
> +}
> +
>  static void
>  ast_mode_config_helper_commit_tail(struct drm_atomic_state *old_state)
>  {
>  	struct drm_device *dev = old_state->dev;
> +	struct ast_private *ast = to_ast_private(dev);
> +	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
> +	struct drm_crtc *crtc;
> +	int i;
> +	bool wait_for_vretrace = false;
>  
>  	drm_atomic_helper_commit_modeset_disables(dev, old_state);
>  
> -	drm_atomic_helper_commit_planes(dev, old_state, 0);
> +	/*
> +	 * HW cursors require the underlying primary plane and CRTC to
> +	 * display a valid mode and image. This is not the case during
> +	 * full modeset operations. So we temporarily disable any active
> +	 * plane, including the HW cursor. Each plane's atomic_update()
> +	 * helper will re-enable it if necessary.
> +	 *
> +	 * We only do this during *full* modesets. It does not affect
> +	 * simple pageflips on the planes.
> +	 */
> +	for_each_oldnew_crtc_in_state(old_state, crtc,
> +				      old_crtc_state,
> +				      new_crtc_state, i) {
> +		if (!ast_crtc_needs_planes_disabled(old_crtc_state,
> +						    new_crtc_state))
> +			continue;
> +		drm_atomic_helper_disable_planes_on_crtc(old_crtc_state,
> +							 false);
> +		wait_for_vretrace = true;
> +	}

Hm this still feels like you're fighting the framework more than using it.
Comment here, but it's kinda review comments on the entire series.

- ast_crtc_needs_planes_disabled feels a bit strange, the usual way to
  handle this kind of stuff is to set crtc_state->needs_modeset from your
  plane's atomic_check function. You might need your own atomic_check
  implementation for that, so that after the plane checks you run the
  modeset checks again.

- with that you can put your call here to disable all planes into the crtc
  ->atomic_disable callback. You can then also put the
  ast_wait_for_retrace in there, at the end.

> +
> +	/*
> +	 * Ensure that no scanout takes place before reprogramming mode
> +	 * and format registers.
> +	 */
> +	if (wait_for_vretrace)
> +		ast_wait_for_vretrace(ast);
> +
> +	drm_atomic_helper_commit_planes(dev, old_state,
> +					DRM_PLANE_COMMIT_ACTIVE_ONLY);

This order also feels a bit strange, especially with the first 2 patches
where you put the crtc modeset code into atomic_begin. It feels a bit like
if you do the plane commit _after_ modeset enables, then you could move
the crtc code into the crtc ->atomic_enable hook, and then let the plane
update stuff roll through all in commit_planes. Moving the modset code
into atomic_begin at least suggests you want modeset enables before plane
commit, and lots of drivers have that sequence in their commit_tail. It's
even a default implementation with drm_atomic_helper_commit_tail_rpm.

Sorry this is all dragging around so much, figuring out the best atomic
flow is occasionally a bit an endeavour :-/

Cheers, Daniel

>  
>  	drm_atomic_helper_commit_modeset_enables(dev, old_state);
>  
>  	drm_atomic_helper_fake_vblank(old_state);
> -
>  	drm_atomic_helper_commit_hw_done(old_state);
> -
>  	drm_atomic_helper_wait_for_vblanks(dev, old_state);
> -
>  	drm_atomic_helper_cleanup_planes(dev, old_state);
>  }
>  
> -- 
> 2.28.0
> 

-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Thomas Zimmermann]

[-- Attachment #1.1: Type: text/plain, Size: 7261 bytes --]

Hi

Am 07.08.20 um 10:50 schrieb daniel@ffwll.ch:
> On Wed, Aug 05, 2020 at 12:54:28PM +0200, Thomas Zimmermann wrote:
>> The ast HW cursor requires the primary plane and CRTC to display at
>> a valid mode and format. This is not the case while switching
>> display modes, which can lead to the screen turing permanently dark.
>>
>> As a workaround, the ast driver now disables active planes while the
>> mode or format switch takes place. It also synchronizes with the vertical
>> refresh to give CRTC and planes some time to catch up on each other.
>> The active planes planes (primary or cursor) will be re-enabled by
>> each plane's atomic_update() function.
>>
>> v2:
>> 	* move the logic into the commit-tail function
>>
>> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
>> Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
>> Cc: Thomas Zimmermann <tzimmermann@suse.de>
>> Cc: Gerd Hoffmann <kraxel@redhat.com>
>> Cc: Dave Airlie <airlied@redhat.com>
>> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
>> Cc: Sam Ravnborg <sam@ravnborg.org>
>> Cc: Emil Velikov <emil.l.velikov@gmail.com>
>> Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
>> Cc: <stable@vger.kernel.org> # v5.6+
>> ---
>>  drivers/gpu/drm/ast/ast_drv.h  |  2 +
>>  drivers/gpu/drm/ast/ast_mode.c | 68 ++++++++++++++++++++++++++++++++--
>>  2 files changed, 66 insertions(+), 4 deletions(-)
>>
>> diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
>> index c1af6b725933..467049ca8430 100644
>> --- a/drivers/gpu/drm/ast/ast_drv.h
>> +++ b/drivers/gpu/drm/ast/ast_drv.h
>> @@ -177,6 +177,8 @@ struct ast_private *ast_device_create(struct drm_driver *drv,
>>  
>>  #define AST_IO_MM_OFFSET		(0x380)
>>  
>> +#define AST_IO_VGAIR1_VREFRESH		BIT(3)
>> +
>>  #define __ast_read(x) \
>>  static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
>>  u##x val = 0;\
>> diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
>> index ae5cb0a333f7..a379d51f3543 100644
>> --- a/drivers/gpu/drm/ast/ast_mode.c
>> +++ b/drivers/gpu/drm/ast/ast_mode.c
>> @@ -514,6 +514,17 @@ static void ast_set_start_address_crt1(struct ast_private *ast,
>>  
>>  }
>>  
>> +static void ast_wait_for_vretrace(struct ast_private *ast)
>> +{
>> +	unsigned long timeout = jiffies + HZ;
>> +	u8 vgair1;
>> +
>> +	do {
>> +		vgair1 = ast_io_read8(ast, AST_IO_INPUT_STATUS1_READ);
>> +	} while (!(vgair1 & AST_IO_VGAIR1_VREFRESH) &&
>> +		 time_before(jiffies, timeout));
>> +}
>> +
>>  /*
>>   * Primary plane
>>   */
>> @@ -1043,23 +1054,72 @@ static int ast_connector_init(struct drm_device *dev)
>>   * Mode config
>>   */
>>  
>> +static bool
>> +ast_crtc_needs_planes_disabled(struct drm_crtc_state *old_crtc_state,
>> +			       struct drm_crtc_state *new_crtc_state)
>> +{
>> +	struct ast_crtc_state *old_ast_crtc_state, *new_ast_crtc_state;
>> +
>> +	if (drm_atomic_crtc_needs_modeset(new_crtc_state))
>> +		return true;
>> +
>> +	old_ast_crtc_state = to_ast_crtc_state(old_crtc_state);
>> +	new_ast_crtc_state = to_ast_crtc_state(new_crtc_state);
>> +
>> +	if (old_ast_crtc_state->format != new_ast_crtc_state->format)
>> +		return true;
>> +
>> +	return false;
>> +}
>> +
>>  static void
>>  ast_mode_config_helper_commit_tail(struct drm_atomic_state *old_state)
>>  {
>>  	struct drm_device *dev = old_state->dev;
>> +	struct ast_private *ast = to_ast_private(dev);
>> +	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
>> +	struct drm_crtc *crtc;
>> +	int i;
>> +	bool wait_for_vretrace = false;
>>  
>>  	drm_atomic_helper_commit_modeset_disables(dev, old_state);
>>  
>> -	drm_atomic_helper_commit_planes(dev, old_state, 0);
>> +	/*
>> +	 * HW cursors require the underlying primary plane and CRTC to
>> +	 * display a valid mode and image. This is not the case during
>> +	 * full modeset operations. So we temporarily disable any active
>> +	 * plane, including the HW cursor. Each plane's atomic_update()
>> +	 * helper will re-enable it if necessary.
>> +	 *
>> +	 * We only do this during *full* modesets. It does not affect
>> +	 * simple pageflips on the planes.
>> +	 */
>> +	for_each_oldnew_crtc_in_state(old_state, crtc,
>> +				      old_crtc_state,
>> +				      new_crtc_state, i) {
>> +		if (!ast_crtc_needs_planes_disabled(old_crtc_state,
>> +						    new_crtc_state))
>> +			continue;
>> +		drm_atomic_helper_disable_planes_on_crtc(old_crtc_state,
>> +							 false);
>> +		wait_for_vretrace = true;
>> +	}
> 
> Hm this still feels like you're fighting the framework more than using it.
> Comment here, but it's kinda review comments on the entire series.
> 
> - ast_crtc_needs_planes_disabled feels a bit strange, the usual way to
>   handle this kind of stuff is to set crtc_state->needs_modeset from your
>   plane's atomic_check function. You might need your own atomic_check
>   implementation for that, so that after the plane checks you run the
>   modeset checks again.
> 
> - with that you can put your call here to disable all planes into the crtc
>   ->atomic_disable callback. You can then also put the
>   ast_wait_for_retrace in there, at the end.

The CRTC's atomic_disable/enable only run if needs_modeset() is true.

I brought back support for fast format changes of the primary plane.
Moving that code into atomic_disable/enable would require to set
needs_modeset in atomic_check() for format changes. And later figure out
in atomic_disable/enable if it's really a modeset or just a change of
the format. That's not good either.

Best regards
Thomas

> 
>> +
>> +	/*
>> +	 * Ensure that no scanout takes place before reprogramming mode
>> +	 * and format registers.
>> +	 */
>> +	if (wait_for_vretrace)
>> +		ast_wait_for_vretrace(ast);
>> +
>> +	drm_atomic_helper_commit_planes(dev, old_state,
>> +					DRM_PLANE_COMMIT_ACTIVE_ONLY);
> 
> This order also feels a bit strange, especially with the first 2 patches
> where you put the crtc modeset code into atomic_begin. It feels a bit like
> if you do the plane commit _after_ modeset enables, then you could move
> the crtc code into the crtc ->atomic_enable hook, and then let the plane
> update stuff roll through all in commit_planes. Moving the modset code
> into atomic_begin at least suggests you want modeset enables before plane
> commit, and lots of drivers have that sequence in their commit_tail. It's
> even a default implementation with drm_atomic_helper_commit_tail_rpm.
> 
> Sorry this is all dragging around so much, figuring out the best atomic
> flow is occasionally a bit an endeavour :-/
> 
> Cheers, Daniel
> 
>>  
>>  	drm_atomic_helper_commit_modeset_enables(dev, old_state);
>>  
>>  	drm_atomic_helper_fake_vblank(old_state);
>> -
>>  	drm_atomic_helper_commit_hw_done(old_state);
>> -
>>  	drm_atomic_helper_wait_for_vblanks(dev, old_state);
>> -
>>  	drm_atomic_helper_cleanup_planes(dev, old_state);
>>  }
>>  
>> -- 
>> 2.28.0
>>
> 

-- 
Thomas Zimmermann
Graphics Driver Developer
SUSE Software Solutions Germany GmbH
Maxfeldstr. 5, 90409 Nürnberg, Germany
(HRB 36809, AG Nürnberg)
Geschäftsführer: Felix Imendörffer


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 516 bytes --]

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Daniel Vetter]

On Wed, Aug 12, 2020 at 10:25:25AM +0200, Thomas Zimmermann wrote:
> Hi
> 
> Am 07.08.20 um 10:50 schrieb daniel@ffwll.ch:
> > On Wed, Aug 05, 2020 at 12:54:28PM +0200, Thomas Zimmermann wrote:
> >> The ast HW cursor requires the primary plane and CRTC to display at
> >> a valid mode and format. This is not the case while switching
> >> display modes, which can lead to the screen turing permanently dark.
> >>
> >> As a workaround, the ast driver now disables active planes while the
> >> mode or format switch takes place. It also synchronizes with the vertical
> >> refresh to give CRTC and planes some time to catch up on each other.
> >> The active planes planes (primary or cursor) will be re-enabled by
> >> each plane's atomic_update() function.
> >>
> >> v2:
> >> 	* move the logic into the commit-tail function
> >>
> >> Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
> >> Fixes: 4961eb60f145 ("drm/ast: Enable atomic modesetting")
> >> Cc: Thomas Zimmermann <tzimmermann@suse.de>
> >> Cc: Gerd Hoffmann <kraxel@redhat.com>
> >> Cc: Dave Airlie <airlied@redhat.com>
> >> Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
> >> Cc: Sam Ravnborg <sam@ravnborg.org>
> >> Cc: Emil Velikov <emil.l.velikov@gmail.com>
> >> Cc: "Y.C. Chen" <yc_chen@aspeedtech.com>
> >> Cc: <stable@vger.kernel.org> # v5.6+
> >> ---
> >>  drivers/gpu/drm/ast/ast_drv.h  |  2 +
> >>  drivers/gpu/drm/ast/ast_mode.c | 68 ++++++++++++++++++++++++++++++++--
> >>  2 files changed, 66 insertions(+), 4 deletions(-)
> >>
> >> diff --git a/drivers/gpu/drm/ast/ast_drv.h b/drivers/gpu/drm/ast/ast_drv.h
> >> index c1af6b725933..467049ca8430 100644
> >> --- a/drivers/gpu/drm/ast/ast_drv.h
> >> +++ b/drivers/gpu/drm/ast/ast_drv.h
> >> @@ -177,6 +177,8 @@ struct ast_private *ast_device_create(struct drm_driver *drv,
> >>  
> >>  #define AST_IO_MM_OFFSET		(0x380)
> >>  
> >> +#define AST_IO_VGAIR1_VREFRESH		BIT(3)
> >> +
> >>  #define __ast_read(x) \
> >>  static inline u##x ast_read##x(struct ast_private *ast, u32 reg) { \
> >>  u##x val = 0;\
> >> diff --git a/drivers/gpu/drm/ast/ast_mode.c b/drivers/gpu/drm/ast/ast_mode.c
> >> index ae5cb0a333f7..a379d51f3543 100644
> >> --- a/drivers/gpu/drm/ast/ast_mode.c
> >> +++ b/drivers/gpu/drm/ast/ast_mode.c
> >> @@ -514,6 +514,17 @@ static void ast_set_start_address_crt1(struct ast_private *ast,
> >>  
> >>  }
> >>  
> >> +static void ast_wait_for_vretrace(struct ast_private *ast)
> >> +{
> >> +	unsigned long timeout = jiffies + HZ;
> >> +	u8 vgair1;
> >> +
> >> +	do {
> >> +		vgair1 = ast_io_read8(ast, AST_IO_INPUT_STATUS1_READ);
> >> +	} while (!(vgair1 & AST_IO_VGAIR1_VREFRESH) &&
> >> +		 time_before(jiffies, timeout));
> >> +}
> >> +
> >>  /*
> >>   * Primary plane
> >>   */
> >> @@ -1043,23 +1054,72 @@ static int ast_connector_init(struct drm_device *dev)
> >>   * Mode config
> >>   */
> >>  
> >> +static bool
> >> +ast_crtc_needs_planes_disabled(struct drm_crtc_state *old_crtc_state,
> >> +			       struct drm_crtc_state *new_crtc_state)
> >> +{
> >> +	struct ast_crtc_state *old_ast_crtc_state, *new_ast_crtc_state;
> >> +
> >> +	if (drm_atomic_crtc_needs_modeset(new_crtc_state))
> >> +		return true;
> >> +
> >> +	old_ast_crtc_state = to_ast_crtc_state(old_crtc_state);
> >> +	new_ast_crtc_state = to_ast_crtc_state(new_crtc_state);
> >> +
> >> +	if (old_ast_crtc_state->format != new_ast_crtc_state->format)
> >> +		return true;
> >> +
> >> +	return false;
> >> +}
> >> +
> >>  static void
> >>  ast_mode_config_helper_commit_tail(struct drm_atomic_state *old_state)
> >>  {
> >>  	struct drm_device *dev = old_state->dev;
> >> +	struct ast_private *ast = to_ast_private(dev);
> >> +	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
> >> +	struct drm_crtc *crtc;
> >> +	int i;
> >> +	bool wait_for_vretrace = false;
> >>  
> >>  	drm_atomic_helper_commit_modeset_disables(dev, old_state);
> >>  
> >> -	drm_atomic_helper_commit_planes(dev, old_state, 0);
> >> +	/*
> >> +	 * HW cursors require the underlying primary plane and CRTC to
> >> +	 * display a valid mode and image. This is not the case during
> >> +	 * full modeset operations. So we temporarily disable any active
> >> +	 * plane, including the HW cursor. Each plane's atomic_update()
> >> +	 * helper will re-enable it if necessary.
> >> +	 *
> >> +	 * We only do this during *full* modesets. It does not affect
> >> +	 * simple pageflips on the planes.
> >> +	 */
> >> +	for_each_oldnew_crtc_in_state(old_state, crtc,
> >> +				      old_crtc_state,
> >> +				      new_crtc_state, i) {
> >> +		if (!ast_crtc_needs_planes_disabled(old_crtc_state,
> >> +						    new_crtc_state))
> >> +			continue;
> >> +		drm_atomic_helper_disable_planes_on_crtc(old_crtc_state,
> >> +							 false);
> >> +		wait_for_vretrace = true;
> >> +	}
> > 
> > Hm this still feels like you're fighting the framework more than using it.
> > Comment here, but it's kinda review comments on the entire series.
> > 
> > - ast_crtc_needs_planes_disabled feels a bit strange, the usual way to
> >   handle this kind of stuff is to set crtc_state->needs_modeset from your
> >   plane's atomic_check function. You might need your own atomic_check
> >   implementation for that, so that after the plane checks you run the
> >   modeset checks again.
> > 
> > - with that you can put your call here to disable all planes into the crtc
> >   ->atomic_disable callback. You can then also put the
> >   ast_wait_for_retrace in there, at the end.
> 
> The CRTC's atomic_disable/enable only run if needs_modeset() is true.
> 
> I brought back support for fast format changes of the primary plane.
> Moving that code into atomic_disable/enable would require to set
> needs_modeset in atomic_check() for format changes. And later figure out
> in atomic_disable/enable if it's really a modeset or just a change of
> the format. That's not good either.

Hm so the uapi rules are somewhat quirky, but NEEDS_MODESET generally
means "no flickering". You can do transitions which take longer than 1
vblank (if resources need to be reallocated) without signalling a
NEEDS_MODESET. But going to black and back is not something that should be
done without signalling a modeset. Userspace would like to make that
distinction and avoid such flickering as much as possible.

Maybe not super relevant for ast, but still, needs_modeset should flag
true for this I think.

Not sure how to best model that with helpers tbh so you can keep the fast
format change. Clever drivers with tricks like this all have their own
modeset flow by now. Maybe just have ast state for whether you need a
modeset (before you overwrite that due to plane changes) and forced plane
disable (any time a modeset is indicated) and then call functions from
your commit_tail depending upon that directly.
-Daniel

> 
> Best regards
> Thomas
> 
> > 
> >> +
> >> +	/*
> >> +	 * Ensure that no scanout takes place before reprogramming mode
> >> +	 * and format registers.
> >> +	 */
> >> +	if (wait_for_vretrace)
> >> +		ast_wait_for_vretrace(ast);
> >> +
> >> +	drm_atomic_helper_commit_planes(dev, old_state,
> >> +					DRM_PLANE_COMMIT_ACTIVE_ONLY);
> > 
> > This order also feels a bit strange, especially with the first 2 patches
> > where you put the crtc modeset code into atomic_begin. It feels a bit like
> > if you do the plane commit _after_ modeset enables, then you could move
> > the crtc code into the crtc ->atomic_enable hook, and then let the plane
> > update stuff roll through all in commit_planes. Moving the modset code
> > into atomic_begin at least suggests you want modeset enables before plane
> > commit, and lots of drivers have that sequence in their commit_tail. It's
> > even a default implementation with drm_atomic_helper_commit_tail_rpm.
> > 
> > Sorry this is all dragging around so much, figuring out the best atomic
> > flow is occasionally a bit an endeavour :-/
> > 
> > Cheers, Daniel
> > 
> >>  
> >>  	drm_atomic_helper_commit_modeset_enables(dev, old_state);
> >>  
> >>  	drm_atomic_helper_fake_vblank(old_state);
> >> -
> >>  	drm_atomic_helper_commit_hw_done(old_state);
> >> -
> >>  	drm_atomic_helper_wait_for_vblanks(dev, old_state);
> >> -
> >>  	drm_atomic_helper_cleanup_planes(dev, old_state);
> >>  }
> >>  
> >> -- 
> >> 2.28.0
> >>
> > 
> 
> -- 
> Thomas Zimmermann
> Graphics Driver Developer
> SUSE Software Solutions Germany GmbH
> Maxfeldstr. 5, 90409 Nürnberg, Germany
> (HRB 36809, AG Nürnberg)
> Geschäftsführer: Felix Imendörffer
> 




-- 
Daniel Vetter
Software Engineer, Intel Corporation
http://blog.ffwll.ch

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.13.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.13: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8, v5.7.14.

v5.8: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.14: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.1, v5.7.15.

v5.8.1: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.15: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha

Re: [PATCH v1 4/4] drm/ast: Disable planes while switching display modes

[Sasha Levin]

Hi

[This is an automated email]

This commit has been processed because it contains a "Fixes:" tag
fixing commit: 4961eb60f145 ("drm/ast: Enable atomic modesetting").

The bot has tested the following trees: v5.8.2, v5.7.16.

v5.8.2: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")

v5.7.16: Failed to apply! Possible dependencies:
    05f13f5b5996 ("drm/ast: Remove unused code paths for AST 1180")
    1728bf6402c3 ("drm/ast: Use managed mode-config init")
    2ccebf561e4a ("drm/ast: Move cursor functions to ast_cursor.c")
    3a53230e1c4b ("drm/ast: Make ast_primary_plane_helper_atomic_update static")
    6bb18c9be6d2 ("drm/ast: Init cursors before creating modesetting structures")
    beb2355eecbf ("drm/ast: Pass struct ast_private instance to cursor init/fini functions")
    e6949ff3ca85 ("drm/ast: Initialize mode setting in ast_mode_config_init()")
    fa7dbd768884 ("drm/ast: Upcast from DRM device to ast structure via to_ast_private()")


NOTE: The patch will not be queued to stable trees until it is upstream.

How should we proceed with this patch?

-- 
Thanks
Sasha