From e3f4bdaf2c5bfebcfb483a8caa1fb989ec042e5b Mon Sep 17 00:00:00 2001
From: Lyude Paul <lyude@redhat.com>
Date: Thu, 11 Sep 2025 18:57:38 -0400
Subject: drm/gem/shmem: Extract drm_gem_shmem_init() from
 drm_gem_shmem_create()

With gem objects in rust, the most ideal way for us to be able to handle
gem shmem object creation is to be able to handle the memory allocation of
a gem object ourselves - and then have the DRM gem shmem helpers initialize
the object we've allocated afterwards. So, let's split out
drm_gem_shmem_init() from drm_gem_shmem_create() to allow for doing this.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Link: https://lore.kernel.org/r/20250911230147.650077-2-lyude@redhat.com
---
 include/drm/drm_gem_shmem_helper.h | 1 +
 1 file changed, 1 insertion(+)

(limited to 'include')

diff --git a/include/drm/drm_gem_shmem_helper.h b/include/drm/drm_gem_shmem_helper.h
index 92f5db84b9c2..235dc33127b9 100644
--- a/include/drm/drm_gem_shmem_helper.h
+++ b/include/drm/drm_gem_shmem_helper.h
@@ -107,6 +107,7 @@ struct drm_gem_shmem_object {
 #define to_drm_gem_shmem_obj(obj) \
 	container_of(obj, struct drm_gem_shmem_object, base)
 
+int drm_gem_shmem_init(struct drm_device *dev, struct drm_gem_shmem_object *shmem, size_t size);
 struct drm_gem_shmem_object *drm_gem_shmem_create(struct drm_device *dev, size_t size);
 struct drm_gem_shmem_object *drm_gem_shmem_create_with_mnt(struct drm_device *dev,
 							   size_t size,
-- 
cgit v1.2.3


From c08c931060c7e44452e635e115913dd88214848c Mon Sep 17 00:00:00 2001
From: Lyude Paul <lyude@redhat.com>
Date: Thu, 11 Sep 2025 18:57:39 -0400
Subject: drm/gem/shmem: Extract drm_gem_shmem_release() from
 drm_gem_shmem_free()

At the moment, the way that we currently free gem shmem objects is not
ideal for rust bindings. drm_gem_shmem_free() releases all of the
associated memory with a gem shmem object with kfree(), which means that
for us to correctly release a gem shmem object in rust we have to manually
drop all of the contents of our gem object structure in-place by hand
before finally calling drm_gem_shmem_free() to release the shmem resources
and the allocation for the gem object.

Since the only reason this is an issue is because of drm_gem_shmem_free()
calling kfree(), we can fix this by splitting drm_gem_shmem_free() out into
itself and drm_gem_shmem_release(), where drm_gem_shmem_release() releases
the various gem shmem resources without freeing the structure itself. With
this, we can safely re-acquire the KBox for the gem object's memory
allocation and let rust handle cleaning up all of the other struct members
automatically.

Signed-off-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Daniel Almeida <daniel.almeida@collabora.com>
Link: https://lore.kernel.org/r/20250911230147.650077-3-lyude@redhat.com
---
 drivers/gpu/drm/drm_gem_shmem_helper.c | 23 ++++++++++++++++++-----
 include/drm/drm_gem_shmem_helper.h     |  1 +
 2 files changed, 19 insertions(+), 5 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_gem_shmem_helper.c b/drivers/gpu/drm/drm_gem_shmem_helper.c
index b20a7b75c722..50594cf8e17c 100644
--- a/drivers/gpu/drm/drm_gem_shmem_helper.c
+++ b/drivers/gpu/drm/drm_gem_shmem_helper.c
@@ -175,13 +175,13 @@ struct drm_gem_shmem_object *drm_gem_shmem_create_with_mnt(struct drm_device *de
 EXPORT_SYMBOL_GPL(drm_gem_shmem_create_with_mnt);
 
 /**
- * drm_gem_shmem_free - Free resources associated with a shmem GEM object
- * @shmem: shmem GEM object to free
+ * drm_gem_shmem_release - Release resources associated with a shmem GEM object.
+ * @shmem: shmem GEM object
  *
- * This function cleans up the GEM object state and frees the memory used to
- * store the object itself.
+ * This function cleans up the GEM object state, but does not free the memory used to store the
+ * object itself. This function is meant to be a dedicated helper for the Rust GEM bindings.
  */
-void drm_gem_shmem_free(struct drm_gem_shmem_object *shmem)
+void drm_gem_shmem_release(struct drm_gem_shmem_object *shmem)
 {
 	struct drm_gem_object *obj = &shmem->base;
 
@@ -208,6 +208,19 @@ void drm_gem_shmem_free(struct drm_gem_shmem_object *shmem)
 	}
 
 	drm_gem_object_release(obj);
+}
+EXPORT_SYMBOL_GPL(drm_gem_shmem_release);
+
+/**
+ * drm_gem_shmem_free - Free resources associated with a shmem GEM object
+ * @shmem: shmem GEM object to free
+ *
+ * This function cleans up the GEM object state and frees the memory used to
+ * store the object itself.
+ */
+void drm_gem_shmem_free(struct drm_gem_shmem_object *shmem)
+{
+	drm_gem_shmem_release(shmem);
 	kfree(shmem);
 }
 EXPORT_SYMBOL_GPL(drm_gem_shmem_free);
diff --git a/include/drm/drm_gem_shmem_helper.h b/include/drm/drm_gem_shmem_helper.h
index 235dc33127b9..589f7bfe7506 100644
--- a/include/drm/drm_gem_shmem_helper.h
+++ b/include/drm/drm_gem_shmem_helper.h
@@ -112,6 +112,7 @@ struct drm_gem_shmem_object *drm_gem_shmem_create(struct drm_device *dev, size_t
 struct drm_gem_shmem_object *drm_gem_shmem_create_with_mnt(struct drm_device *dev,
 							   size_t size,
 							   struct vfsmount *gemfs);
+void drm_gem_shmem_release(struct drm_gem_shmem_object *shmem);
 void drm_gem_shmem_free(struct drm_gem_shmem_object *shmem);
 
 void drm_gem_shmem_put_pages_locked(struct drm_gem_shmem_object *shmem);
-- 
cgit v1.2.3


From d8c4bddcd8bcb41885d3db2ba18c840c411564c2 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Fri, 29 Aug 2025 11:13:45 +0200
Subject: drm/fb-helper: Synchronize dirty worker with vblank

Before updating the display from the console's shadow buffer, the dirty
worker now waits for a vblank. This allows several screen updates to pile
up and acts as a rate limiter. If a DRM master is present, it could
interfere with the vblank. Don't wait in this case.

v4:
	* share code with WAITFORVSYNC ioctl (Emil)
	* use lock guard
v3:
	* add back helper->lock
	* acquire DRM master status while waiting for vblank
v2:
	* don't hold helper->lock while waiting for vblank

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Sam Ravnborg <sam@ravnborg.org>
Link: https://lore.kernel.org/r/20250829091447.46719-1-tzimmermann@suse.de
---
 drivers/gpu/drm/drm_client_modeset.c | 44 ++++++++++++++++++++++++++++++++++++
 drivers/gpu/drm/drm_fb_helper.c      | 30 +++++-------------------
 include/drm/drm_client.h             |  1 +
 3 files changed, 51 insertions(+), 24 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_client_modeset.c b/drivers/gpu/drm/drm_client_modeset.c
index 9c2c3b0c8c47..fc4caf7da5fc 100644
--- a/drivers/gpu/drm/drm_client_modeset.c
+++ b/drivers/gpu/drm/drm_client_modeset.c
@@ -1293,6 +1293,50 @@ int drm_client_modeset_dpms(struct drm_client_dev *client, int mode)
 }
 EXPORT_SYMBOL(drm_client_modeset_dpms);
 
+/**
+ * drm_client_modeset_wait_for_vblank() - Wait for the next VBLANK to occur
+ * @client: DRM client
+ * @crtc_index: The ndex of the CRTC to wait on
+ *
+ * Block the caller until the given CRTC has seen a VBLANK. Do nothing
+ * if the CRTC is disabled. If there's another DRM master present, fail
+ * with -EBUSY.
+ *
+ * Returns:
+ * 0 on success, or negative error code otherwise.
+ */
+int drm_client_modeset_wait_for_vblank(struct drm_client_dev *client, unsigned int crtc_index)
+{
+	struct drm_device *dev = client->dev;
+	struct drm_crtc *crtc;
+	int ret;
+
+	/*
+	 * Rate-limit update frequency to vblank. If there's a DRM master
+	 * present, it could interfere while we're waiting for the vblank
+	 * event. Don't wait in this case.
+	 */
+	if (!drm_master_internal_acquire(dev))
+		return -EBUSY;
+
+	crtc = client->modesets[crtc_index].crtc;
+
+	/*
+	 * Only wait for a vblank event if the CRTC is enabled, otherwise
+	 * just don't do anything, not even report an error.
+	 */
+	ret = drm_crtc_vblank_get(crtc);
+	if (!ret) {
+		drm_crtc_wait_one_vblank(crtc);
+		drm_crtc_vblank_put(crtc);
+	}
+
+	drm_master_internal_release(dev);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_client_modeset_wait_for_vblank);
+
 #ifdef CONFIG_DRM_KUNIT_TEST
 #include "tests/drm_client_modeset_test.c"
 #endif
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 11a5b60cb9ce..53e9dc0543de 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -368,6 +368,10 @@ static void drm_fb_helper_fb_dirty(struct drm_fb_helper *helper)
 	unsigned long flags;
 	int ret;
 
+	mutex_lock(&helper->lock);
+	drm_client_modeset_wait_for_vblank(&helper->client, 0);
+	mutex_unlock(&helper->lock);
+
 	if (drm_WARN_ON_ONCE(dev, !helper->funcs->fb_dirty))
 		return;
 
@@ -1068,15 +1072,9 @@ int drm_fb_helper_ioctl(struct fb_info *info, unsigned int cmd,
 			unsigned long arg)
 {
 	struct drm_fb_helper *fb_helper = info->par;
-	struct drm_device *dev = fb_helper->dev;
-	struct drm_crtc *crtc;
 	int ret = 0;
 
-	mutex_lock(&fb_helper->lock);
-	if (!drm_master_internal_acquire(dev)) {
-		ret = -EBUSY;
-		goto unlock;
-	}
+	guard(mutex)(&fb_helper->lock);
 
 	switch (cmd) {
 	case FBIO_WAITFORVSYNC:
@@ -1096,28 +1094,12 @@ int drm_fb_helper_ioctl(struct fb_info *info, unsigned int cmd,
 		 * make. If we're not smart enough here, one should
 		 * just consider switch the userspace to KMS.
 		 */
-		crtc = fb_helper->client.modesets[0].crtc;
-
-		/*
-		 * Only wait for a vblank event if the CRTC is
-		 * enabled, otherwise just don't do anythintg,
-		 * not even report an error.
-		 */
-		ret = drm_crtc_vblank_get(crtc);
-		if (!ret) {
-			drm_crtc_wait_one_vblank(crtc);
-			drm_crtc_vblank_put(crtc);
-		}
-
-		ret = 0;
+		ret = drm_client_modeset_wait_for_vblank(&fb_helper->client, 0);
 		break;
 	default:
 		ret = -ENOTTY;
 	}
 
-	drm_master_internal_release(dev);
-unlock:
-	mutex_unlock(&fb_helper->lock);
 	return ret;
 }
 EXPORT_SYMBOL(drm_fb_helper_ioctl);
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index 146ca80e35db..bdd845e383ef 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -220,6 +220,7 @@ int drm_client_modeset_check(struct drm_client_dev *client);
 int drm_client_modeset_commit_locked(struct drm_client_dev *client);
 int drm_client_modeset_commit(struct drm_client_dev *client);
 int drm_client_modeset_dpms(struct drm_client_dev *client, int mode);
+int drm_client_modeset_wait_for_vblank(struct drm_client_dev *client, unsigned int crtc_index);
 
 /**
  * drm_client_for_each_modeset() - Iterate over client modesets
-- 
cgit v1.2.3


From e46efc6a7d288830c4aeaf3c65c7e913a8ca35d7 Mon Sep 17 00:00:00 2001
From: Luca Ceresoli <luca.ceresoli@bootlin.com>
Date: Fri, 8 Aug 2025 16:49:10 +0200
Subject: drm/bridge: add drm_for_each_bridge_in_chain_scoped()

drm_for_each_bridge_in_chain() iterates ofer the bridges in an encoder
chain without protecting the lifetime of the bridges using
drm_bridge_get/put(). This creates a risk window where the bridge could be
freed while iterating on it. Users of drm_for_each_bridge_in_chain() cannot
solve this reliably.

Add variant of drm_for_each_bridge_in_chain() that gets/puts the bridge
reference at the beginning/end of each iteration, and puts it if breaking
ot of the loop.

Note that this requires adding a new drm_bridge_get_next_bridge_and_put()
function because, unlike similar functions as __of_get_next_child(),
drm_bridge_get_next_bridge() gets the "next" pointer but does not put the
"prev" pointer. Unfortunately drm_bridge_get_next_bridge() cannot be
modified to put the "prev" pointer because some of its users rely on
this, such as drm_atomic_bridge_propagate_bus_flags().

Also deprecate drm_for_each_bridge_in_chain(), in preparation for removing
it after converting all users to the scoped version.

Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250808-drm-bridge-alloc-getput-for_each_bridge-v2-3-edb6ee81edf1@bootlin.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
---
 .clang-format            |  1 +
 include/drm/drm_bridge.h | 41 +++++++++++++++++++++++++++++++++++++++++
 2 files changed, 42 insertions(+)

(limited to 'include')

diff --git a/.clang-format b/.clang-format
index 48405c54ef27..1cac7d497664 100644
--- a/.clang-format
+++ b/.clang-format
@@ -168,6 +168,7 @@ ForEachMacros:
   - 'drm_exec_for_each_locked_object'
   - 'drm_exec_for_each_locked_object_reverse'
   - 'drm_for_each_bridge_in_chain'
+  - 'drm_for_each_bridge_in_chain_scoped'
   - 'drm_for_each_connector_iter'
   - 'drm_for_each_crtc'
   - 'drm_for_each_crtc_reverse'
diff --git a/include/drm/drm_bridge.h b/include/drm/drm_bridge.h
index 76e05930f50e..57a3d3cd08e7 100644
--- a/include/drm/drm_bridge.h
+++ b/include/drm/drm_bridge.h
@@ -1440,10 +1440,51 @@ drm_bridge_chain_get_last_bridge(struct drm_encoder *encoder)
  *	    iteration
  *
  * Iterate over all bridges present in the bridge chain attached to @encoder.
+ *
+ * This is deprecated, do not use!
+ * New drivers shall use drm_for_each_bridge_in_chain_scoped().
  */
 #define drm_for_each_bridge_in_chain(encoder, bridge)			\
 	list_for_each_entry(bridge, &(encoder)->bridge_chain, chain_node)
 
+/**
+ * drm_bridge_get_next_bridge_and_put - Get the next bridge in the chain
+ *                                      and put the previous
+ * @bridge: bridge object
+ *
+ * Same as drm_bridge_get_next_bridge() but additionally puts the @bridge.
+ *
+ * RETURNS:
+ * the next bridge in the chain after @bridge, or NULL if @bridge is the last.
+ */
+static inline struct drm_bridge *
+drm_bridge_get_next_bridge_and_put(struct drm_bridge *bridge)
+{
+	struct drm_bridge *next = drm_bridge_get_next_bridge(bridge);
+
+	drm_bridge_put(bridge);
+
+	return next;
+}
+
+/**
+ * drm_for_each_bridge_in_chain_scoped - iterate over all bridges attached
+ *                                       to an encoder
+ * @encoder: the encoder to iterate bridges on
+ * @bridge: a bridge pointer updated to point to the current bridge at each
+ *	    iteration
+ *
+ * Iterate over all bridges present in the bridge chain attached to @encoder.
+ *
+ * Automatically gets/puts the bridge reference while iterating, and puts
+ * the reference even if returning or breaking in the middle of the loop.
+ */
+#define drm_for_each_bridge_in_chain_scoped(encoder, bridge)		\
+	for (struct drm_bridge *bridge __free(drm_bridge_put) =		\
+	     drm_bridge_chain_get_first_bridge(encoder);		\
+	     bridge;							\
+	     bridge = drm_bridge_get_next_bridge_and_put(bridge))
+
 enum drm_mode_status
 drm_bridge_chain_mode_valid(struct drm_bridge *bridge,
 			    const struct drm_display_info *info,
-- 
cgit v1.2.3


From 2f08387a444c4fb2e983305356f030d50978d21d Mon Sep 17 00:00:00 2001
From: Luca Ceresoli <luca.ceresoli@bootlin.com>
Date: Fri, 8 Aug 2025 16:49:14 +0200
Subject: drm/bridge: remove drm_for_each_bridge_in_chain()

All users have been replaced by drm_for_each_bridge_in_chain_scoped().

Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250808-drm-bridge-alloc-getput-for_each_bridge-v2-7-edb6ee81edf1@bootlin.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
---
 .clang-format            |  1 -
 include/drm/drm_bridge.h | 14 --------------
 2 files changed, 15 deletions(-)

(limited to 'include')

diff --git a/.clang-format b/.clang-format
index 1cac7d497664..d5c05db1a0d9 100644
--- a/.clang-format
+++ b/.clang-format
@@ -167,7 +167,6 @@ ForEachMacros:
   - 'drm_connector_for_each_possible_encoder'
   - 'drm_exec_for_each_locked_object'
   - 'drm_exec_for_each_locked_object_reverse'
-  - 'drm_for_each_bridge_in_chain'
   - 'drm_for_each_bridge_in_chain_scoped'
   - 'drm_for_each_connector_iter'
   - 'drm_for_each_crtc'
diff --git a/include/drm/drm_bridge.h b/include/drm/drm_bridge.h
index 57a3d3cd08e7..c45dece368ad 100644
--- a/include/drm/drm_bridge.h
+++ b/include/drm/drm_bridge.h
@@ -1433,20 +1433,6 @@ drm_bridge_chain_get_last_bridge(struct drm_encoder *encoder)
 						      struct drm_bridge, chain_node));
 }
 
-/**
- * drm_for_each_bridge_in_chain() - Iterate over all bridges present in a chain
- * @encoder: the encoder to iterate bridges on
- * @bridge: a bridge pointer updated to point to the current bridge at each
- *	    iteration
- *
- * Iterate over all bridges present in the bridge chain attached to @encoder.
- *
- * This is deprecated, do not use!
- * New drivers shall use drm_for_each_bridge_in_chain_scoped().
- */
-#define drm_for_each_bridge_in_chain(encoder, bridge)			\
-	list_for_each_entry(bridge, &(encoder)->bridge_chain, chain_node)
-
 /**
  * drm_bridge_get_next_bridge_and_put - Get the next bridge in the chain
  *                                      and put the previous
-- 
cgit v1.2.3


From 78f4eec62097754fed1b910544ff60ef10a76554 Mon Sep 17 00:00:00 2001
From: Luca Ceresoli <luca.ceresoli@bootlin.com>
Date: Fri, 8 Aug 2025 16:49:15 +0200
Subject: drm/bridge: add drm_for_each_bridge_in_chain_from()

Add variant of drm_for_each_bridge_in_chain_scoped() that iterates on the
encoder bridge from a given bridge until the end of the chain.

Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250808-drm-bridge-alloc-getput-for_each_bridge-v2-8-edb6ee81edf1@bootlin.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
---
 include/drm/drm_bridge.h | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

(limited to 'include')

diff --git a/include/drm/drm_bridge.h b/include/drm/drm_bridge.h
index c45dece368ad..d1aba4e25d35 100644
--- a/include/drm/drm_bridge.h
+++ b/include/drm/drm_bridge.h
@@ -1471,6 +1471,25 @@ drm_bridge_get_next_bridge_and_put(struct drm_bridge *bridge)
 	     bridge;							\
 	     bridge = drm_bridge_get_next_bridge_and_put(bridge))
 
+/**
+ * drm_for_each_bridge_in_chain_from - iterate over all bridges starting
+ *                                     from the given bridge
+ * @first_bridge: the bridge to start from
+ * @bridge: a bridge pointer updated to point to the current bridge at each
+ *	    iteration
+ *
+ * Iterate over all bridges in the encoder chain starting from
+ * @first_bridge, included.
+ *
+ * Automatically gets/puts the bridge reference while iterating, and puts
+ * the reference even if returning or breaking in the middle of the loop.
+ */
+#define drm_for_each_bridge_in_chain_from(first_bridge, bridge)		\
+	for (struct drm_bridge *bridge __free(drm_bridge_put) =		\
+		     drm_bridge_get(first_bridge);			\
+	     bridge;							\
+	     bridge = drm_bridge_get_next_bridge_and_put(bridge))
+
 enum drm_mode_status
 drm_bridge_chain_mode_valid(struct drm_bridge *bridge,
 			    const struct drm_display_info *info,
-- 
cgit v1.2.3


From 61aa4f7a600871f9ad96a0ad355b483b0ac4b67d Mon Sep 17 00:00:00 2001
From: Luca Ceresoli <luca.ceresoli@bootlin.com>
Date: Fri, 1 Aug 2025 19:05:29 +0200
Subject: drm/bridge: get the bridge returned by drm_bridge_get_next_bridge()

drm_bridge_get_next_bridge() returns a bridge pointer that the
caller could hold for a long time. Increment the refcount of the returned
bridge and document it must be put by the caller.

Reviewed-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250801-drm-bridge-alloc-getput-drm_bridge_get_next_bridge-v2-7-888912b0be13@bootlin.com
Signed-off-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
---
 include/drm/drm_bridge.h | 9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_bridge.h b/include/drm/drm_bridge.h
index d1aba4e25d35..0ff7ab4aa868 100644
--- a/include/drm/drm_bridge.h
+++ b/include/drm/drm_bridge.h
@@ -1362,6 +1362,13 @@ drm_bridge_get_current_state(struct drm_bridge *bridge)
  * drm_bridge_get_next_bridge() - Get the next bridge in the chain
  * @bridge: bridge object
  *
+ * The caller is responsible of having a reference to @bridge via
+ * drm_bridge_get() or equivalent. This function leaves the refcount of
+ * @bridge unmodified.
+ *
+ * The refcount of the returned bridge is incremented. Use drm_bridge_put()
+ * when done with it.
+ *
  * RETURNS:
  * the next bridge in the chain after @bridge, or NULL if @bridge is the last.
  */
@@ -1371,7 +1378,7 @@ drm_bridge_get_next_bridge(struct drm_bridge *bridge)
 	if (list_is_last(&bridge->chain_node, &bridge->encoder->bridge_chain))
 		return NULL;
 
-	return list_next_entry(bridge, chain_node);
+	return drm_bridge_get(list_next_entry(bridge, chain_node));
 }
 
 /**
-- 
cgit v1.2.3


From ed7a4397f55bfacf2c4c3e466940ed4961707094 Mon Sep 17 00:00:00 2001
From: Christian König <christian.koenig@amd.com>
Date: Fri, 13 Jun 2025 14:09:08 +0200
Subject: drm/ttm: rename ttm_bo_put to _fini v3
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Give TTM BOs a separate cleanup function.

No funktional change, but the next step in removing the TTM BO reference
counting and replacing it with the GEM object reference counting.

v2: move the code around a bit to make it clearer what's happening
v3: fix nouveau_bo_fini as well

Signed-off-by: Christian König <christian.koenig@amd.com>
Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Acked-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Link: https://lore.kernel.org/r/20250909144311.1927-1-christian.koenig@amd.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c          |  2 +-
 drivers/gpu/drm/drm_gem_vram_helper.c            |  6 +--
 drivers/gpu/drm/i915/gem/i915_gem_ttm.c          |  4 +-
 drivers/gpu/drm/loongson/lsdc_gem.c              |  2 +-
 drivers/gpu/drm/nouveau/nouveau_bo.h             |  2 +-
 drivers/gpu/drm/nouveau/nouveau_gem.c            |  2 +-
 drivers/gpu/drm/qxl/qxl_gem.c                    |  2 +-
 drivers/gpu/drm/radeon/radeon_gem.c              |  2 +-
 drivers/gpu/drm/ttm/tests/ttm_bo_test.c          | 12 ++---
 drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c | 60 ++++++++++++------------
 drivers/gpu/drm/ttm/ttm_bo.c                     | 15 +++---
 drivers/gpu/drm/ttm/ttm_bo_internal.h            |  2 +
 drivers/gpu/drm/vmwgfx/vmwgfx_gem.c              |  2 +-
 drivers/gpu/drm/xe/xe_bo.c                       |  2 +-
 include/drm/ttm/ttm_bo.h                         |  2 +-
 15 files changed, 59 insertions(+), 58 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
index c049848a56b2..1c9b5009304e 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_gem.c
@@ -198,7 +198,7 @@ static void amdgpu_gem_object_free(struct drm_gem_object *gobj)
 	struct amdgpu_bo *aobj = gem_to_amdgpu_bo(gobj);
 
 	amdgpu_hmm_unregister(aobj);
-	ttm_bo_put(&aobj->tbo);
+	ttm_bo_fini(&aobj->tbo);
 }
 
 int amdgpu_gem_object_create(struct amdgpu_device *adev, unsigned long size,
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index b04cde4a60e7..90760d0ca071 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -107,7 +107,7 @@ static const struct drm_gem_object_funcs drm_gem_vram_object_funcs;
 
 static void drm_gem_vram_cleanup(struct drm_gem_vram_object *gbo)
 {
-	/* We got here via ttm_bo_put(), which means that the
+	/* We got here via ttm_bo_fini(), which means that the
 	 * TTM buffer object in 'bo' has already been cleaned
 	 * up; only release the GEM object.
 	 */
@@ -234,11 +234,11 @@ EXPORT_SYMBOL(drm_gem_vram_create);
  * drm_gem_vram_put() - Releases a reference to a VRAM-backed GEM object
  * @gbo:	the GEM VRAM object
  *
- * See ttm_bo_put() for more information.
+ * See ttm_bo_fini() for more information.
  */
 void drm_gem_vram_put(struct drm_gem_vram_object *gbo)
 {
-	ttm_bo_put(&gbo->bo);
+	ttm_bo_fini(&gbo->bo);
 }
 EXPORT_SYMBOL(drm_gem_vram_put);
 
diff --git a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
index 1f4814968868..57bb111d65da 100644
--- a/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
+++ b/drivers/gpu/drm/i915/gem/i915_gem_ttm.c
@@ -1029,7 +1029,7 @@ static void i915_ttm_delayed_free(struct drm_i915_gem_object *obj)
 {
 	GEM_BUG_ON(!obj->ttm.created);
 
-	ttm_bo_put(i915_gem_to_ttm(obj));
+	ttm_bo_fini(i915_gem_to_ttm(obj));
 }
 
 static vm_fault_t vm_fault_ttm(struct vm_fault *vmf)
@@ -1325,7 +1325,7 @@ int __i915_gem_ttm_object_init(struct intel_memory_region *mem,
 	 * If this function fails, it will call the destructor, but
 	 * our caller still owns the object. So no freeing in the
 	 * destructor until obj->ttm.created is true.
-	 * Similarly, in delayed_destroy, we can't call ttm_bo_put()
+	 * Similarly, in delayed_destroy, we can't call ttm_bo_fini()
 	 * until successful initialization.
 	 */
 	ret = ttm_bo_init_reserved(&i915->bdev, i915_gem_to_ttm(obj), bo_type,
diff --git a/drivers/gpu/drm/loongson/lsdc_gem.c b/drivers/gpu/drm/loongson/lsdc_gem.c
index a720d8f53209..22d0eced95da 100644
--- a/drivers/gpu/drm/loongson/lsdc_gem.c
+++ b/drivers/gpu/drm/loongson/lsdc_gem.c
@@ -57,7 +57,7 @@ static void lsdc_gem_object_free(struct drm_gem_object *obj)
 	struct ttm_buffer_object *tbo = to_ttm_bo(obj);
 
 	if (tbo)
-		ttm_bo_put(tbo);
+		ttm_bo_fini(tbo);
 }
 
 static int lsdc_gem_object_vmap(struct drm_gem_object *obj, struct iosys_map *map)
diff --git a/drivers/gpu/drm/nouveau/nouveau_bo.h b/drivers/gpu/drm/nouveau/nouveau_bo.h
index d59fd12268b9..6c26beeb427f 100644
--- a/drivers/gpu/drm/nouveau/nouveau_bo.h
+++ b/drivers/gpu/drm/nouveau/nouveau_bo.h
@@ -57,7 +57,7 @@ nouveau_bo(struct ttm_buffer_object *bo)
 static inline void
 nouveau_bo_fini(struct nouveau_bo *bo)
 {
-	ttm_bo_put(&bo->bo);
+	ttm_bo_fini(&bo->bo);
 }
 
 extern struct ttm_device_funcs nouveau_bo_driver;
diff --git a/drivers/gpu/drm/nouveau/nouveau_gem.c b/drivers/gpu/drm/nouveau/nouveau_gem.c
index 690e10fbf0bd..395d92ab6271 100644
--- a/drivers/gpu/drm/nouveau/nouveau_gem.c
+++ b/drivers/gpu/drm/nouveau/nouveau_gem.c
@@ -87,7 +87,7 @@ nouveau_gem_object_del(struct drm_gem_object *gem)
 		return;
 	}
 
-	ttm_bo_put(&nvbo->bo);
+	ttm_bo_fini(&nvbo->bo);
 
 	pm_runtime_mark_last_busy(dev);
 	pm_runtime_put_autosuspend(dev);
diff --git a/drivers/gpu/drm/qxl/qxl_gem.c b/drivers/gpu/drm/qxl/qxl_gem.c
index fc5e3763c359..d26043424e95 100644
--- a/drivers/gpu/drm/qxl/qxl_gem.c
+++ b/drivers/gpu/drm/qxl/qxl_gem.c
@@ -39,7 +39,7 @@ void qxl_gem_object_free(struct drm_gem_object *gobj)
 	qxl_surface_evict(qdev, qobj, false);
 
 	tbo = &qobj->tbo;
-	ttm_bo_put(tbo);
+	ttm_bo_fini(tbo);
 }
 
 int qxl_gem_object_create(struct qxl_device *qdev, int size,
diff --git a/drivers/gpu/drm/radeon/radeon_gem.c b/drivers/gpu/drm/radeon/radeon_gem.c
index f86773f3db20..18ca1bcfd2f9 100644
--- a/drivers/gpu/drm/radeon/radeon_gem.c
+++ b/drivers/gpu/drm/radeon/radeon_gem.c
@@ -86,7 +86,7 @@ static void radeon_gem_object_free(struct drm_gem_object *gobj)
 
 	if (robj) {
 		radeon_mn_unregister(robj);
-		ttm_bo_put(&robj->tbo);
+		ttm_bo_fini(&robj->tbo);
 	}
 }
 
diff --git a/drivers/gpu/drm/ttm/tests/ttm_bo_test.c b/drivers/gpu/drm/ttm/tests/ttm_bo_test.c
index 6c77550c51af..5426b435f702 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_bo_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_bo_test.c
@@ -379,7 +379,7 @@ static void ttm_bo_unreserve_bulk(struct kunit *test)
 	dma_resv_fini(resv);
 }
 
-static void ttm_bo_put_basic(struct kunit *test)
+static void ttm_bo_fini_basic(struct kunit *test)
 {
 	struct ttm_test_devices *priv = test->priv;
 	struct ttm_buffer_object *bo;
@@ -410,7 +410,7 @@ static void ttm_bo_put_basic(struct kunit *test)
 	dma_resv_unlock(bo->base.resv);
 	KUNIT_EXPECT_EQ(test, err, 0);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static const char *mock_name(struct dma_fence *f)
@@ -423,7 +423,7 @@ static const struct dma_fence_ops mock_fence_ops = {
 	.get_timeline_name = mock_name,
 };
 
-static void ttm_bo_put_shared_resv(struct kunit *test)
+static void ttm_bo_fini_shared_resv(struct kunit *test)
 {
 	struct ttm_test_devices *priv = test->priv;
 	struct ttm_buffer_object *bo;
@@ -463,7 +463,7 @@ static void ttm_bo_put_shared_resv(struct kunit *test)
 	bo->type = ttm_bo_type_device;
 	bo->base.resv = external_resv;
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static void ttm_bo_pin_basic(struct kunit *test)
@@ -616,8 +616,8 @@ static struct kunit_case ttm_bo_test_cases[] = {
 	KUNIT_CASE(ttm_bo_unreserve_basic),
 	KUNIT_CASE(ttm_bo_unreserve_pinned),
 	KUNIT_CASE(ttm_bo_unreserve_bulk),
-	KUNIT_CASE(ttm_bo_put_basic),
-	KUNIT_CASE(ttm_bo_put_shared_resv),
+	KUNIT_CASE(ttm_bo_fini_basic),
+	KUNIT_CASE(ttm_bo_fini_shared_resv),
 	KUNIT_CASE(ttm_bo_pin_basic),
 	KUNIT_CASE(ttm_bo_pin_unpin_resource),
 	KUNIT_CASE(ttm_bo_multiple_pin_one_unpin),
diff --git a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
index 1bcc67977f48..3a1eef83190c 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
@@ -144,7 +144,7 @@ static void ttm_bo_init_reserved_sys_man(struct kunit *test)
 				  drm_mm_node_allocated(&bo->base.vma_node.vm_node));
 
 	ttm_resource_free(bo, &bo->resource);
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static void ttm_bo_init_reserved_mock_man(struct kunit *test)
@@ -186,7 +186,7 @@ static void ttm_bo_init_reserved_mock_man(struct kunit *test)
 				  drm_mm_node_allocated(&bo->base.vma_node.vm_node));
 
 	ttm_resource_free(bo, &bo->resource);
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 }
 
@@ -221,7 +221,7 @@ static void ttm_bo_init_reserved_resv(struct kunit *test)
 	KUNIT_EXPECT_PTR_EQ(test, bo->base.resv, &resv);
 
 	ttm_resource_free(bo, &bo->resource);
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static void ttm_bo_validate_basic(struct kunit *test)
@@ -265,7 +265,7 @@ static void ttm_bo_validate_basic(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, bo->resource->placement,
 			DRM_BUDDY_TOPDOWN_ALLOCATION);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 	ttm_mock_manager_fini(priv->ttm_dev, snd_mem);
 }
 
@@ -292,7 +292,7 @@ static void ttm_bo_validate_invalid_placement(struct kunit *test)
 
 	KUNIT_EXPECT_EQ(test, err, -ENOMEM);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static void ttm_bo_validate_failed_alloc(struct kunit *test)
@@ -321,7 +321,7 @@ static void ttm_bo_validate_failed_alloc(struct kunit *test)
 
 	KUNIT_EXPECT_EQ(test, err, -ENOMEM);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 	ttm_bad_manager_fini(priv->ttm_dev, mem_type);
 }
 
@@ -353,7 +353,7 @@ static void ttm_bo_validate_pinned(struct kunit *test)
 	ttm_bo_unpin(bo);
 	dma_resv_unlock(bo->base.resv);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static const struct ttm_bo_validate_test_case ttm_mem_type_cases[] = {
@@ -403,7 +403,7 @@ static void ttm_bo_validate_same_placement(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, err, 0);
 	KUNIT_EXPECT_EQ(test, ctx_val.bytes_moved, 0);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 
 	if (params->mem_type != TTM_PL_SYSTEM)
 		ttm_mock_manager_fini(priv->ttm_dev, params->mem_type);
@@ -452,7 +452,7 @@ static void ttm_bo_validate_busy_placement(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, bo->resource->mem_type, snd_mem);
 	KUNIT_ASSERT_TRUE(test, list_is_singular(&man->lru[bo->priority]));
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 	ttm_bad_manager_fini(priv->ttm_dev, fst_mem);
 	ttm_mock_manager_fini(priv->ttm_dev, snd_mem);
 }
@@ -495,7 +495,7 @@ static void ttm_bo_validate_multihop(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, ctx_val.bytes_moved, size * 2);
 	KUNIT_EXPECT_EQ(test, bo->resource->mem_type, final_mem);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 
 	ttm_mock_manager_fini(priv->ttm_dev, fst_mem);
 	ttm_mock_manager_fini(priv->ttm_dev, tmp_mem);
@@ -567,7 +567,7 @@ static void ttm_bo_validate_no_placement_signaled(struct kunit *test)
 		KUNIT_ASSERT_TRUE(test, flags & TTM_TT_FLAG_ZERO_ALLOC);
 	}
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static int threaded_dma_resv_signal(void *arg)
@@ -635,7 +635,7 @@ static void ttm_bo_validate_no_placement_not_signaled(struct kunit *test)
 	/* Make sure we have an idle object at this point */
 	dma_resv_wait_timeout(bo->base.resv, usage, false, MAX_SCHEDULE_TIMEOUT);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 }
 
 static void ttm_bo_validate_move_fence_signaled(struct kunit *test)
@@ -668,7 +668,7 @@ static void ttm_bo_validate_move_fence_signaled(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, bo->resource->mem_type, mem_type);
 	KUNIT_EXPECT_EQ(test, ctx.bytes_moved, size);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 	dma_fence_put(man->move);
 }
 
@@ -753,7 +753,7 @@ static void ttm_bo_validate_move_fence_not_signaled(struct kunit *test)
 	else
 		KUNIT_EXPECT_EQ(test, bo->resource->mem_type, fst_mem);
 
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo);
 	ttm_mock_manager_fini(priv->ttm_dev, fst_mem);
 	ttm_mock_manager_fini(priv->ttm_dev, snd_mem);
 }
@@ -807,8 +807,8 @@ static void ttm_bo_validate_happy_evict(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, bos[1].resource->mem_type, mem_type);
 
 	for (i = 0; i < bo_no; i++)
-		ttm_bo_put(&bos[i]);
-	ttm_bo_put(bo_val);
+		ttm_bo_fini(&bos[i]);
+	ttm_bo_fini(bo_val);
 
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 	ttm_mock_manager_fini(priv->ttm_dev, mem_multihop);
@@ -852,12 +852,12 @@ static void ttm_bo_validate_all_pinned_evict(struct kunit *test)
 
 	KUNIT_EXPECT_EQ(test, err, -ENOMEM);
 
-	ttm_bo_put(bo_small);
+	ttm_bo_fini(bo_small);
 
 	ttm_bo_reserve(bo_big, false, false, NULL);
 	ttm_bo_unpin(bo_big);
 	dma_resv_unlock(bo_big->base.resv);
-	ttm_bo_put(bo_big);
+	ttm_bo_fini(bo_big);
 
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 	ttm_mock_manager_fini(priv->ttm_dev, mem_multihop);
@@ -916,13 +916,13 @@ static void ttm_bo_validate_allowed_only_evict(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, bo_evictable->resource->mem_type, mem_type_evict);
 	KUNIT_EXPECT_EQ(test, ctx_val.bytes_moved, size * 2 + BO_SIZE);
 
-	ttm_bo_put(bo);
-	ttm_bo_put(bo_evictable);
+	ttm_bo_fini(bo);
+	ttm_bo_fini(bo_evictable);
 
 	ttm_bo_reserve(bo_pinned, false, false, NULL);
 	ttm_bo_unpin(bo_pinned);
 	dma_resv_unlock(bo_pinned->base.resv);
-	ttm_bo_put(bo_pinned);
+	ttm_bo_fini(bo_pinned);
 
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 	ttm_mock_manager_fini(priv->ttm_dev, mem_multihop);
@@ -973,8 +973,8 @@ static void ttm_bo_validate_deleted_evict(struct kunit *test)
 	KUNIT_EXPECT_NULL(test, bo_big->ttm);
 	KUNIT_EXPECT_NULL(test, bo_big->resource);
 
-	ttm_bo_put(bo_small);
-	ttm_bo_put(bo_big);
+	ttm_bo_fini(bo_small);
+	ttm_bo_fini(bo_big);
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 }
 
@@ -1025,8 +1025,8 @@ static void ttm_bo_validate_busy_domain_evict(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, bo_init->resource->mem_type, mem_type);
 	KUNIT_EXPECT_NULL(test, bo_val->resource);
 
-	ttm_bo_put(bo_init);
-	ttm_bo_put(bo_val);
+	ttm_bo_fini(bo_init);
+	ttm_bo_fini(bo_val);
 
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 	ttm_bad_manager_fini(priv->ttm_dev, mem_type_evict);
@@ -1070,8 +1070,8 @@ static void ttm_bo_validate_evict_gutting(struct kunit *test)
 	KUNIT_ASSERT_NULL(test, bo_evict->resource);
 	KUNIT_ASSERT_TRUE(test, bo_evict->ttm->page_flags & TTM_TT_FLAG_ZERO_ALLOC);
 
-	ttm_bo_put(bo_evict);
-	ttm_bo_put(bo);
+	ttm_bo_fini(bo_evict);
+	ttm_bo_fini(bo);
 
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 }
@@ -1128,9 +1128,9 @@ static void ttm_bo_validate_recrusive_evict(struct kunit *test)
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type);
 	ttm_mock_manager_fini(priv->ttm_dev, mem_type_evict);
 
-	ttm_bo_put(bo_val);
-	ttm_bo_put(bo_tt);
-	ttm_bo_put(bo_mock);
+	ttm_bo_fini(bo_val);
+	ttm_bo_fini(bo_tt);
+	ttm_bo_fini(bo_mock);
 }
 
 static struct kunit_case ttm_bo_validate_test_cases[] = {
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 29423ceeec5c..fba2a68a556e 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -318,18 +318,17 @@ static void ttm_bo_release(struct kref *kref)
 	bo->destroy(bo);
 }
 
-/**
- * ttm_bo_put
- *
- * @bo: The buffer object.
- *
- * Unreference a buffer object.
- */
+/* TODO: remove! */
 void ttm_bo_put(struct ttm_buffer_object *bo)
 {
 	kref_put(&bo->kref, ttm_bo_release);
 }
-EXPORT_SYMBOL(ttm_bo_put);
+
+void ttm_bo_fini(struct ttm_buffer_object *bo)
+{
+	ttm_bo_put(bo);
+}
+EXPORT_SYMBOL(ttm_bo_fini);
 
 static int ttm_bo_bounce_temp_buffer(struct ttm_buffer_object *bo,
 				     struct ttm_operation_ctx *ctx,
diff --git a/drivers/gpu/drm/ttm/ttm_bo_internal.h b/drivers/gpu/drm/ttm/ttm_bo_internal.h
index 9d8b747a34db..e0d48eac74b0 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_internal.h
+++ b/drivers/gpu/drm/ttm/ttm_bo_internal.h
@@ -55,4 +55,6 @@ ttm_bo_get_unless_zero(struct ttm_buffer_object *bo)
 	return bo;
 }
 
+void ttm_bo_put(struct ttm_buffer_object *bo);
+
 #endif
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c b/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c
index eedf1fe60be7..39f8c46550c2 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_gem.c
@@ -37,7 +37,7 @@ static void vmw_gem_object_free(struct drm_gem_object *gobj)
 {
 	struct ttm_buffer_object *bo = drm_gem_ttm_of_gem(gobj);
 	if (bo)
-		ttm_bo_put(bo);
+		ttm_bo_fini(bo);
 }
 
 static int vmw_gem_object_open(struct drm_gem_object *obj,
diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index 870f43347281..ebcd191034df 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -1696,7 +1696,7 @@ static void xe_gem_object_free(struct drm_gem_object *obj)
 	 * refcount directly if needed.
 	 */
 	__xe_bo_vunmap(gem_to_xe_bo(obj));
-	ttm_bo_put(container_of(obj, struct ttm_buffer_object, base));
+	ttm_bo_fini(container_of(obj, struct ttm_buffer_object, base));
 }
 
 static void xe_gem_object_close(struct drm_gem_object *obj,
diff --git a/include/drm/ttm/ttm_bo.h b/include/drm/ttm/ttm_bo.h
index e664a96540eb..bca3a8849d47 100644
--- a/include/drm/ttm/ttm_bo.h
+++ b/include/drm/ttm/ttm_bo.h
@@ -391,7 +391,7 @@ int ttm_bo_wait_ctx(struct ttm_buffer_object *bo,
 int ttm_bo_validate(struct ttm_buffer_object *bo,
 		    struct ttm_placement *placement,
 		    struct ttm_operation_ctx *ctx);
-void ttm_bo_put(struct ttm_buffer_object *bo);
+void ttm_bo_fini(struct ttm_buffer_object *bo);
 void ttm_bo_set_bulk_move(struct ttm_buffer_object *bo,
 			  struct ttm_lru_bulk_move *bulk);
 bool ttm_bo_eviction_valuable(struct ttm_buffer_object *bo,
-- 
cgit v1.2.3


From 091767ee7510765886b8475c44fd81b2e6e9da87 Mon Sep 17 00:00:00 2001
From: Luc Ma <onion0709@gmail.com>
Date: Mon, 15 Sep 2025 21:23:26 +0800
Subject: drm/sched: backend_ops doc fix

Function drm_sched_entity_do_release() has been renamed in
commit 180fc134d712 ("drm/scheduler: Rename cleanup functions v2.").

Refer to the correct function in the documentation.

Signed-off-by: Luc Ma <onion0709@gmail.com>
[phasta: commit message]
Signed-off-by: Philipp Stanner <phasta@kernel.org>
Link: https://lore.kernel.org/r/20250915132327.6293-1-onion0709@gmail.com
---
 include/drm/gpu_scheduler.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/gpu_scheduler.h b/include/drm/gpu_scheduler.h
index 323a505e6e6a..fb88301b3c45 100644
--- a/include/drm/gpu_scheduler.h
+++ b/include/drm/gpu_scheduler.h
@@ -546,7 +546,7 @@ struct drm_sched_backend_ops {
  * @num_rqs: Number of run-queues. This is at most DRM_SCHED_PRIORITY_COUNT,
  *           as there's usually one run-queue per priority, but could be less.
  * @sched_rq: An allocated array of run-queues of size @num_rqs;
- * @job_scheduled: once @drm_sched_entity_do_release is called the scheduler
+ * @job_scheduled: once drm_sched_entity_flush() is called the scheduler
  *                 waits on this wait queue until all the scheduled jobs are
  *                 finished.
  * @job_id_count: used to assign unique id to the each job.
-- 
cgit v1.2.3


From 0bf37f45d5c472aebdf32da64775cac1110c085c Mon Sep 17 00:00:00 2001
From: Andrzej Kacprowski <Andrzej.Kacprowski@intel.com>
Date: Mon, 15 Sep 2025 12:34:37 +0200
Subject: accel/ivpu: Add support for user-managed preemption buffer

Allow user mode drivers to manage preemption buffers, enabling
memory savings by sharing a single buffer across multiple
command queues within the same memory context.

Introduce DRM_IVPU_PARAM_PREEMPT_BUFFER_SIZE to report the required
preemption buffer size as specified by the firmware.

The preemption buffer is now passed from user space as an entry
in the BO list of DRM_IVPU_CMDQ_SUBMIT. The buffer must be
non-mappable and large enough to hold preemption data.

For backward compatibility, the kernel will allocate an internal
preemption buffer if user space does not provide one.

User space can only provide a single preemption buffer,
simplifying the ioctl interface and parameter validation.
A separate secondary preemption buffer is only needed
to save below 4GB address space on 37xx and only if preemption
buffers are not shared.

Signed-off-by: Andrzej Kacprowski <Andrzej.Kacprowski@intel.com>
Reviewed-by: Lizhi Hou <lizhi.hou@amd.com>
Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com>
Link: https://lore.kernel.org/r/20250915103437.830086-1-karol.wachowski@linux.intel.com
---
 drivers/accel/ivpu/ivpu_drv.c |  3 ++
 drivers/accel/ivpu/ivpu_fw.c  | 57 ++++++++++++++++++++-----
 drivers/accel/ivpu/ivpu_fw.h  |  7 +++-
 drivers/accel/ivpu/ivpu_gem.h |  7 +++-
 drivers/accel/ivpu/ivpu_job.c | 96 +++++++++++++++++++++++++++++--------------
 drivers/accel/ivpu/ivpu_job.h |  4 +-
 include/uapi/drm/ivpu_accel.h | 11 +++++
 7 files changed, 141 insertions(+), 44 deletions(-)

(limited to 'include')

diff --git a/drivers/accel/ivpu/ivpu_drv.c b/drivers/accel/ivpu/ivpu_drv.c
index 3289751b4757..a08f99c3ba4a 100644
--- a/drivers/accel/ivpu/ivpu_drv.c
+++ b/drivers/accel/ivpu/ivpu_drv.c
@@ -200,6 +200,9 @@ static int ivpu_get_param_ioctl(struct drm_device *dev, void *data, struct drm_f
 	case DRM_IVPU_PARAM_CAPABILITIES:
 		args->value = ivpu_is_capable(vdev, args->index);
 		break;
+	case DRM_IVPU_PARAM_PREEMPT_BUFFER_SIZE:
+		args->value = ivpu_fw_preempt_buf_size(vdev);
+		break;
 	default:
 		ret = -EINVAL;
 		break;
diff --git a/drivers/accel/ivpu/ivpu_fw.c b/drivers/accel/ivpu/ivpu_fw.c
index df9631fef058..32f513499829 100644
--- a/drivers/accel/ivpu/ivpu_fw.c
+++ b/drivers/accel/ivpu/ivpu_fw.c
@@ -26,6 +26,8 @@
 #define FW_RUNTIME_MIN_ADDR	(FW_GLOBAL_MEM_START)
 #define FW_RUNTIME_MAX_ADDR	(FW_GLOBAL_MEM_END - FW_SHARED_MEM_SIZE)
 #define FW_FILE_IMAGE_OFFSET	(VPU_FW_HEADER_SIZE + FW_VERSION_HEADER_SIZE)
+#define FW_PREEMPT_BUF_MIN_SIZE SZ_4K
+#define FW_PREEMPT_BUF_MAX_SIZE SZ_32M
 
 #define WATCHDOG_MSS_REDIRECT	32
 #define WATCHDOG_NCE_REDIRECT	33
@@ -151,6 +153,47 @@ ivpu_fw_sched_mode_select(struct ivpu_device *vdev, const struct vpu_firmware_he
 	return VPU_SCHEDULING_MODE_HW;
 }
 
+static void
+ivpu_preemption_config_parse(struct ivpu_device *vdev, const struct vpu_firmware_header *fw_hdr)
+{
+	struct ivpu_fw_info *fw = vdev->fw;
+	u32 primary_preempt_buf_size, secondary_preempt_buf_size;
+
+	if (fw_hdr->preemption_buffer_1_max_size)
+		primary_preempt_buf_size = fw_hdr->preemption_buffer_1_max_size;
+	else
+		primary_preempt_buf_size = fw_hdr->preemption_buffer_1_size;
+
+	if (fw_hdr->preemption_buffer_2_max_size)
+		secondary_preempt_buf_size = fw_hdr->preemption_buffer_2_max_size;
+	else
+		secondary_preempt_buf_size = fw_hdr->preemption_buffer_2_size;
+
+	ivpu_dbg(vdev, FW_BOOT, "Preemption buffer size, primary: %u, secondary: %u\n",
+		 primary_preempt_buf_size, secondary_preempt_buf_size);
+
+	if (primary_preempt_buf_size < FW_PREEMPT_BUF_MIN_SIZE ||
+	    secondary_preempt_buf_size < FW_PREEMPT_BUF_MIN_SIZE) {
+		ivpu_warn(vdev, "Preemption buffers size too small\n");
+		return;
+	}
+
+	if (primary_preempt_buf_size > FW_PREEMPT_BUF_MAX_SIZE ||
+	    secondary_preempt_buf_size > FW_PREEMPT_BUF_MAX_SIZE) {
+		ivpu_warn(vdev, "Preemption buffers size too big\n");
+		return;
+	}
+
+	if (fw->sched_mode != VPU_SCHEDULING_MODE_HW)
+		return;
+
+	if (ivpu_test_mode & IVPU_TEST_MODE_MIP_DISABLE)
+		return;
+
+	vdev->fw->primary_preempt_buf_size = ALIGN(primary_preempt_buf_size, PAGE_SIZE);
+	vdev->fw->secondary_preempt_buf_size = ALIGN(secondary_preempt_buf_size, PAGE_SIZE);
+}
+
 static int ivpu_fw_parse(struct ivpu_device *vdev)
 {
 	struct ivpu_fw_info *fw = vdev->fw;
@@ -235,17 +278,9 @@ static int ivpu_fw_parse(struct ivpu_device *vdev)
 	fw->sched_mode = ivpu_fw_sched_mode_select(vdev, fw_hdr);
 	ivpu_info(vdev, "Scheduler mode: %s\n", fw->sched_mode ? "HW" : "OS");
 
-	if (fw_hdr->preemption_buffer_1_max_size)
-		fw->primary_preempt_buf_size = fw_hdr->preemption_buffer_1_max_size;
-	else
-		fw->primary_preempt_buf_size = fw_hdr->preemption_buffer_1_size;
-
-	if (fw_hdr->preemption_buffer_2_max_size)
-		fw->secondary_preempt_buf_size = fw_hdr->preemption_buffer_2_max_size;
-	else
-		fw->secondary_preempt_buf_size = fw_hdr->preemption_buffer_2_size;
-	ivpu_dbg(vdev, FW_BOOT, "Preemption buffer sizes: primary %u, secondary %u\n",
-		 fw->primary_preempt_buf_size, fw->secondary_preempt_buf_size);
+	ivpu_preemption_config_parse(vdev, fw_hdr);
+	ivpu_dbg(vdev, FW_BOOT, "Mid-inference preemption %s supported\n",
+		 ivpu_fw_preempt_buf_size(vdev) ? "is" : "is not");
 
 	if (fw_hdr->ro_section_start_address && !is_within_range(fw_hdr->ro_section_start_address,
 								 fw_hdr->ro_section_size,
diff --git a/drivers/accel/ivpu/ivpu_fw.h b/drivers/accel/ivpu/ivpu_fw.h
index 7081913fb0dd..6fe2917abda6 100644
--- a/drivers/accel/ivpu/ivpu_fw.h
+++ b/drivers/accel/ivpu/ivpu_fw.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (C) 2020-2024 Intel Corporation
+ * Copyright (C) 2020-2025 Intel Corporation
  */
 
 #ifndef __IVPU_FW_H__
@@ -52,4 +52,9 @@ static inline bool ivpu_fw_is_cold_boot(struct ivpu_device *vdev)
 	return vdev->fw->entry_point == vdev->fw->cold_boot_entry_point;
 }
 
+static inline u32 ivpu_fw_preempt_buf_size(struct ivpu_device *vdev)
+{
+	return vdev->fw->primary_preempt_buf_size + vdev->fw->secondary_preempt_buf_size;
+}
+
 #endif /* __IVPU_FW_H__ */
diff --git a/drivers/accel/ivpu/ivpu_gem.h b/drivers/accel/ivpu/ivpu_gem.h
index aa8ff14f7aae..3ee996d503b2 100644
--- a/drivers/accel/ivpu/ivpu_gem.h
+++ b/drivers/accel/ivpu/ivpu_gem.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (C) 2020-2023 Intel Corporation
+ * Copyright (C) 2020-2025 Intel Corporation
  */
 #ifndef __IVPU_GEM_H__
 #define __IVPU_GEM_H__
@@ -96,4 +96,9 @@ static inline u32 cpu_to_vpu_addr(struct ivpu_bo *bo, void *cpu_addr)
 	return bo->vpu_addr + (cpu_addr - ivpu_bo_vaddr(bo));
 }
 
+static inline bool ivpu_bo_is_mappable(struct ivpu_bo *bo)
+{
+	return bo->flags & DRM_IVPU_BO_MAPPABLE;
+}
+
 #endif /* __IVPU_GEM_H__ */
diff --git a/drivers/accel/ivpu/ivpu_job.c b/drivers/accel/ivpu/ivpu_job.c
index b9902dd0cad5..044268d0fc87 100644
--- a/drivers/accel/ivpu/ivpu_job.c
+++ b/drivers/accel/ivpu/ivpu_job.c
@@ -34,22 +34,20 @@ static void ivpu_cmdq_ring_db(struct ivpu_device *vdev, struct ivpu_cmdq *cmdq)
 static int ivpu_preemption_buffers_create(struct ivpu_device *vdev,
 					  struct ivpu_file_priv *file_priv, struct ivpu_cmdq *cmdq)
 {
-	u64 primary_size = ALIGN(vdev->fw->primary_preempt_buf_size, PAGE_SIZE);
-	u64 secondary_size = ALIGN(vdev->fw->secondary_preempt_buf_size, PAGE_SIZE);
-
-	if (vdev->fw->sched_mode != VPU_SCHEDULING_MODE_HW ||
-	    ivpu_test_mode & IVPU_TEST_MODE_MIP_DISABLE)
+	if (ivpu_fw_preempt_buf_size(vdev) == 0)
 		return 0;
 
 	cmdq->primary_preempt_buf = ivpu_bo_create(vdev, &file_priv->ctx, &vdev->hw->ranges.user,
-						   primary_size, DRM_IVPU_BO_WC);
+						   vdev->fw->primary_preempt_buf_size,
+						   DRM_IVPU_BO_WC);
 	if (!cmdq->primary_preempt_buf) {
 		ivpu_err(vdev, "Failed to create primary preemption buffer\n");
 		return -ENOMEM;
 	}
 
 	cmdq->secondary_preempt_buf = ivpu_bo_create(vdev, &file_priv->ctx, &vdev->hw->ranges.dma,
-						     secondary_size, DRM_IVPU_BO_WC);
+						     vdev->fw->secondary_preempt_buf_size,
+						     DRM_IVPU_BO_WC);
 	if (!cmdq->secondary_preempt_buf) {
 		ivpu_err(vdev, "Failed to create secondary preemption buffer\n");
 		goto err_free_primary;
@@ -66,20 +64,39 @@ err_free_primary:
 static void ivpu_preemption_buffers_free(struct ivpu_device *vdev,
 					 struct ivpu_file_priv *file_priv, struct ivpu_cmdq *cmdq)
 {
-	if (vdev->fw->sched_mode != VPU_SCHEDULING_MODE_HW)
-		return;
-
 	if (cmdq->primary_preempt_buf)
 		ivpu_bo_free(cmdq->primary_preempt_buf);
 	if (cmdq->secondary_preempt_buf)
 		ivpu_bo_free(cmdq->secondary_preempt_buf);
 }
 
+static int ivpu_preemption_job_init(struct ivpu_device *vdev, struct ivpu_file_priv *file_priv,
+				    struct ivpu_cmdq *cmdq, struct ivpu_job *job)
+{
+	int ret;
+
+	/* Use preemption buffer provided by the user space */
+	if (job->primary_preempt_buf)
+		return 0;
+
+	if (!cmdq->primary_preempt_buf) {
+		/* Allocate per command queue preemption buffers */
+		ret = ivpu_preemption_buffers_create(vdev, file_priv, cmdq);
+		if (ret)
+			return ret;
+	}
+
+	/* Use preemption buffers allocated by the kernel */
+	job->primary_preempt_buf = cmdq->primary_preempt_buf;
+	job->secondary_preempt_buf = cmdq->secondary_preempt_buf;
+
+	return 0;
+}
+
 static struct ivpu_cmdq *ivpu_cmdq_alloc(struct ivpu_file_priv *file_priv)
 {
 	struct ivpu_device *vdev = file_priv->vdev;
 	struct ivpu_cmdq *cmdq;
-	int ret;
 
 	cmdq = kzalloc(sizeof(*cmdq), GFP_KERNEL);
 	if (!cmdq)
@@ -89,10 +106,6 @@ static struct ivpu_cmdq *ivpu_cmdq_alloc(struct ivpu_file_priv *file_priv)
 	if (!cmdq->mem)
 		goto err_free_cmdq;
 
-	ret = ivpu_preemption_buffers_create(vdev, file_priv, cmdq);
-	if (ret)
-		ivpu_warn(vdev, "Failed to allocate preemption buffers, preemption limited\n");
-
 	return cmdq;
 
 err_free_cmdq:
@@ -429,17 +442,14 @@ static int ivpu_cmdq_push_job(struct ivpu_cmdq *cmdq, struct ivpu_job *job)
 	if (unlikely(ivpu_test_mode & IVPU_TEST_MODE_NULL_SUBMISSION))
 		entry->flags = VPU_JOB_FLAGS_NULL_SUBMISSION_MASK;
 
-	if (vdev->fw->sched_mode == VPU_SCHEDULING_MODE_HW) {
-		if (cmdq->primary_preempt_buf) {
-			entry->primary_preempt_buf_addr = cmdq->primary_preempt_buf->vpu_addr;
-			entry->primary_preempt_buf_size = ivpu_bo_size(cmdq->primary_preempt_buf);
-		}
+	if (job->primary_preempt_buf) {
+		entry->primary_preempt_buf_addr = job->primary_preempt_buf->vpu_addr;
+		entry->primary_preempt_buf_size = ivpu_bo_size(job->primary_preempt_buf);
+	}
 
-		if (cmdq->secondary_preempt_buf) {
-			entry->secondary_preempt_buf_addr = cmdq->secondary_preempt_buf->vpu_addr;
-			entry->secondary_preempt_buf_size =
-				ivpu_bo_size(cmdq->secondary_preempt_buf);
-		}
+	if (job->secondary_preempt_buf) {
+		entry->secondary_preempt_buf_addr = job->secondary_preempt_buf->vpu_addr;
+		entry->secondary_preempt_buf_size = ivpu_bo_size(job->secondary_preempt_buf);
 	}
 
 	wmb(); /* Ensure that tail is updated after filling entry */
@@ -663,6 +673,13 @@ static int ivpu_job_submit(struct ivpu_job *job, u8 priority, u32 cmdq_id)
 		goto err_unlock;
 	}
 
+	ret = ivpu_preemption_job_init(vdev, file_priv, cmdq, job);
+	if (ret) {
+		ivpu_err(vdev, "Failed to initialize preemption buffers for job %d: %d\n",
+			 job->job_id, ret);
+		goto err_unlock;
+	}
+
 	job->cmdq_id = cmdq->id;
 
 	is_first_job = xa_empty(&vdev->submitted_jobs_xa);
@@ -716,7 +733,7 @@ err_unlock:
 
 static int
 ivpu_job_prepare_bos_for_submit(struct drm_file *file, struct ivpu_job *job, u32 *buf_handles,
-				u32 buf_count, u32 commands_offset)
+				u32 buf_count, u32 commands_offset, u32 preempt_buffer_index)
 {
 	struct ivpu_file_priv *file_priv = job->file_priv;
 	struct ivpu_device *vdev = file_priv->vdev;
@@ -752,6 +769,20 @@ ivpu_job_prepare_bos_for_submit(struct drm_file *file, struct ivpu_job *job, u32
 
 	job->cmd_buf_vpu_addr = bo->vpu_addr + commands_offset;
 
+	if (preempt_buffer_index) {
+		struct ivpu_bo *preempt_bo = job->bos[preempt_buffer_index];
+
+		if (ivpu_bo_size(preempt_bo) < ivpu_fw_preempt_buf_size(vdev)) {
+			ivpu_warn(vdev, "Preemption buffer is too small\n");
+			return -EINVAL;
+		}
+		if (ivpu_bo_is_mappable(preempt_bo)) {
+			ivpu_warn(vdev, "Preemption buffer cannot be mappable\n");
+			return -EINVAL;
+		}
+		job->primary_preempt_buf = preempt_bo;
+	}
+
 	ret = drm_gem_lock_reservations((struct drm_gem_object **)job->bos, buf_count,
 					&acquire_ctx);
 	if (ret) {
@@ -782,7 +813,7 @@ unlock_reservations:
 
 static int ivpu_submit(struct drm_file *file, struct ivpu_file_priv *file_priv, u32 cmdq_id,
 		       u32 buffer_count, u32 engine, void __user *buffers_ptr, u32 cmds_offset,
-		       u8 priority)
+		       u32 preempt_buffer_index, u8 priority)
 {
 	struct ivpu_device *vdev = file_priv->vdev;
 	struct ivpu_job *job;
@@ -814,7 +845,8 @@ static int ivpu_submit(struct drm_file *file, struct ivpu_file_priv *file_priv,
 		goto err_exit_dev;
 	}
 
-	ret = ivpu_job_prepare_bos_for_submit(file, job, buf_handles, buffer_count, cmds_offset);
+	ret = ivpu_job_prepare_bos_for_submit(file, job, buf_handles, buffer_count, cmds_offset,
+					      preempt_buffer_index);
 	if (ret) {
 		ivpu_err(vdev, "Failed to prepare job: %d\n", ret);
 		goto err_destroy_job;
@@ -868,7 +900,7 @@ int ivpu_submit_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	priority = ivpu_job_to_jsm_priority(args->priority);
 
 	return ivpu_submit(file, file_priv, 0, args->buffer_count, args->engine,
-			   (void __user *)args->buffers_ptr, args->commands_offset, priority);
+			   (void __user *)args->buffers_ptr, args->commands_offset, 0, priority);
 }
 
 int ivpu_cmdq_submit_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
@@ -885,6 +917,9 @@ int ivpu_cmdq_submit_ioctl(struct drm_device *dev, void *data, struct drm_file *
 	if (args->buffer_count == 0 || args->buffer_count > JOB_MAX_BUFFER_COUNT)
 		return -EINVAL;
 
+	if (args->preempt_buffer_index >= args->buffer_count)
+		return -EINVAL;
+
 	if (!IS_ALIGNED(args->commands_offset, 8))
 		return -EINVAL;
 
@@ -895,7 +930,8 @@ int ivpu_cmdq_submit_ioctl(struct drm_device *dev, void *data, struct drm_file *
 		return -EBADFD;
 
 	return ivpu_submit(file, file_priv, args->cmdq_id, args->buffer_count, VPU_ENGINE_COMPUTE,
-			   (void __user *)args->buffers_ptr, args->commands_offset, 0);
+			   (void __user *)args->buffers_ptr, args->commands_offset,
+			   args->preempt_buffer_index, 0);
 }
 
 int ivpu_cmdq_create_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
diff --git a/drivers/accel/ivpu/ivpu_job.h b/drivers/accel/ivpu/ivpu_job.h
index 2e301c2eea7b..6c8b9c739b51 100644
--- a/drivers/accel/ivpu/ivpu_job.h
+++ b/drivers/accel/ivpu/ivpu_job.h
@@ -1,6 +1,6 @@
 /* SPDX-License-Identifier: GPL-2.0-only */
 /*
- * Copyright (C) 2020-2024 Intel Corporation
+ * Copyright (C) 2020-2025 Intel Corporation
  */
 
 #ifndef __IVPU_JOB_H__
@@ -55,6 +55,8 @@ struct ivpu_job {
 	u32 job_id;
 	u32 engine_idx;
 	size_t bo_count;
+	struct ivpu_bo *primary_preempt_buf;
+	struct ivpu_bo *secondary_preempt_buf;
 	struct ivpu_bo *bos[] __counted_by(bo_count);
 };
 
diff --git a/include/uapi/drm/ivpu_accel.h b/include/uapi/drm/ivpu_accel.h
index 160ee1411d4a..e470b0221e02 100644
--- a/include/uapi/drm/ivpu_accel.h
+++ b/include/uapi/drm/ivpu_accel.h
@@ -90,6 +90,7 @@ extern "C" {
 #define DRM_IVPU_PARAM_TILE_CONFIG	    11
 #define DRM_IVPU_PARAM_SKU		    12
 #define DRM_IVPU_PARAM_CAPABILITIES	    13
+#define DRM_IVPU_PARAM_PREEMPT_BUFFER_SIZE  14
 
 #define DRM_IVPU_PLATFORM_TYPE_SILICON	    0
 
@@ -176,6 +177,9 @@ struct drm_ivpu_param {
 	 *
 	 * %DRM_IVPU_PARAM_CAPABILITIES:
 	 * Supported capabilities (read-only)
+	 *
+	 * %DRM_IVPU_PARAM_PREEMPT_BUFFER_SIZE:
+	 * Size of the preemption buffer (read-only)
 	 */
 	__u32 param;
 
@@ -371,6 +375,13 @@ struct drm_ivpu_cmdq_submit {
 	 * to be executed. The offset has to be 8-byte aligned.
 	 */
 	__u32 commands_offset;
+	/**
+	 * @preempt_buffer_index:
+	 *
+	 * Index of the preemption buffer in the buffers_ptr array.
+	 */
+	__u32 preempt_buffer_index;
+	__u32 reserved;
 };
 
 /* drm_ivpu_bo_wait job status codes */
-- 
cgit v1.2.3


From b060004f06ae0a3064bddb87a3f8ad13f859fcf3 Mon Sep 17 00:00:00 2001
From: Boris Brezillon <boris.brezillon@collabora.com>
Date: Wed, 17 Sep 2025 20:18:37 +0100
Subject: drm/panfrost: Introduce uAPI for JM context creation
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The new uAPI lets user space query the KM driver for the available
priorities a job can be given at submit time. These are managed through
the notion of a context, for which we also provide new creation and
destruction ioctls.

Reviewed-by: Steven Price <steven.price@arm.com>
Signed-off-by: Boris Brezillon <boris.brezillon@collabora.com>
Signed-off-by: Adrián Larumbe <adrian.larumbe@collabora.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Link: https://lore.kernel.org/r/20250917191859.500279-2-adrian.larumbe@collabora.com
---
 include/uapi/drm/panfrost_drm.h | 50 +++++++++++++++++++++++++++++++++++++++++
 1 file changed, 50 insertions(+)

(limited to 'include')

diff --git a/include/uapi/drm/panfrost_drm.h b/include/uapi/drm/panfrost_drm.h
index ed67510395bd..e8b47c9f6976 100644
--- a/include/uapi/drm/panfrost_drm.h
+++ b/include/uapi/drm/panfrost_drm.h
@@ -22,6 +22,8 @@ extern "C" {
 #define DRM_PANFROST_PERFCNT_DUMP		0x07
 #define DRM_PANFROST_MADVISE			0x08
 #define DRM_PANFROST_SET_LABEL_BO		0x09
+#define DRM_PANFROST_JM_CTX_CREATE		0x0a
+#define DRM_PANFROST_JM_CTX_DESTROY		0x0b
 
 #define DRM_IOCTL_PANFROST_SUBMIT		DRM_IOW(DRM_COMMAND_BASE + DRM_PANFROST_SUBMIT, struct drm_panfrost_submit)
 #define DRM_IOCTL_PANFROST_WAIT_BO		DRM_IOW(DRM_COMMAND_BASE + DRM_PANFROST_WAIT_BO, struct drm_panfrost_wait_bo)
@@ -31,6 +33,8 @@ extern "C" {
 #define DRM_IOCTL_PANFROST_GET_BO_OFFSET	DRM_IOWR(DRM_COMMAND_BASE + DRM_PANFROST_GET_BO_OFFSET, struct drm_panfrost_get_bo_offset)
 #define DRM_IOCTL_PANFROST_MADVISE		DRM_IOWR(DRM_COMMAND_BASE + DRM_PANFROST_MADVISE, struct drm_panfrost_madvise)
 #define DRM_IOCTL_PANFROST_SET_LABEL_BO		DRM_IOWR(DRM_COMMAND_BASE + DRM_PANFROST_SET_LABEL_BO, struct drm_panfrost_set_label_bo)
+#define DRM_IOCTL_PANFROST_JM_CTX_CREATE	DRM_IOWR(DRM_COMMAND_BASE + DRM_PANFROST_JM_CTX_CREATE, struct drm_panfrost_jm_ctx_create)
+#define DRM_IOCTL_PANFROST_JM_CTX_DESTROY	DRM_IOWR(DRM_COMMAND_BASE + DRM_PANFROST_JM_CTX_DESTROY, struct drm_panfrost_jm_ctx_destroy)
 
 /*
  * Unstable ioctl(s): only exposed when the unsafe unstable_ioctls module
@@ -71,6 +75,12 @@ struct drm_panfrost_submit {
 
 	/** A combination of PANFROST_JD_REQ_* */
 	__u32 requirements;
+
+	/** JM context handle. Zero if you want to use the default context. */
+	__u32 jm_ctx_handle;
+
+	/** Padding field. MBZ. */
+	__u32 pad;
 };
 
 /**
@@ -177,6 +187,7 @@ enum drm_panfrost_param {
 	DRM_PANFROST_PARAM_AFBC_FEATURES,
 	DRM_PANFROST_PARAM_SYSTEM_TIMESTAMP,
 	DRM_PANFROST_PARAM_SYSTEM_TIMESTAMP_FREQUENCY,
+	DRM_PANFROST_PARAM_ALLOWED_JM_CTX_PRIORITIES,
 };
 
 struct drm_panfrost_get_param {
@@ -299,6 +310,45 @@ struct panfrost_dump_registers {
 	__u32 value;
 };
 
+enum drm_panfrost_jm_ctx_priority {
+	/**
+	 * @PANFROST_JM_CTX_PRIORITY_LOW: Low priority context.
+	 */
+	PANFROST_JM_CTX_PRIORITY_LOW = 0,
+
+	/**
+	 * @PANFROST_JM_CTX_PRIORITY_MEDIUM: Medium priority context.
+	 */
+	PANFROST_JM_CTX_PRIORITY_MEDIUM,
+
+	/**
+	 * @PANFROST_JM_CTX_PRIORITY_HIGH: High priority context.
+	 *
+	 * Requires CAP_SYS_NICE or DRM_MASTER.
+	 */
+	PANFROST_JM_CTX_PRIORITY_HIGH,
+};
+
+struct drm_panfrost_jm_ctx_create {
+	/** @handle: Handle of the created JM context */
+	__u32 handle;
+
+	/** @priority: Context priority (see enum drm_panfrost_jm_ctx_priority). */
+	__u32 priority;
+};
+
+struct drm_panfrost_jm_ctx_destroy {
+	/**
+	 * @handle: Handle of the JM context to destroy.
+	 *
+	 * Must be a valid context handle returned by DRM_IOCTL_PANTHOR_JM_CTX_CREATE.
+	 */
+	__u32 handle;
+
+	/** @pad: Padding field, MBZ. */
+	__u32 pad;
+};
+
 #if defined(__cplusplus)
 }
 #endif
-- 
cgit v1.2.3


From 2b8e4b94c1b6cea1b4d3f989b3dfe4301d0d3a4a Mon Sep 17 00:00:00 2001
From: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date: Fri, 18 Jul 2025 15:01:51 +0300
Subject: drm/dp: Add definitions for POST_LT_ADJ training sequence
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add the bit definitions needed for POST_LT_ADJ sequence.

v2: DP_POST_LT_ADJ_REQ_IN_PROGRESS is bit 1 not 5 (Jani)

Tested-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250718120154.15492-2-ville.syrjala@linux.intel.com
---
 include/drm/display/drm_dp.h | 3 +++
 1 file changed, 3 insertions(+)

(limited to 'include')

diff --git a/include/drm/display/drm_dp.h b/include/drm/display/drm_dp.h
index 811e9238a77c..cf318e3ddb5c 100644
--- a/include/drm/display/drm_dp.h
+++ b/include/drm/display/drm_dp.h
@@ -115,6 +115,7 @@
 
 #define DP_MAX_LANE_COUNT                   0x002
 # define DP_MAX_LANE_COUNT_MASK		    0x1f
+# define DP_POST_LT_ADJ_REQ_SUPPORTED	    (1 << 5) /* 1.3 */
 # define DP_TPS3_SUPPORTED		    (1 << 6) /* 1.2 */
 # define DP_ENHANCED_FRAME_CAP		    (1 << 7)
 
@@ -583,6 +584,7 @@
 
 #define DP_LANE_COUNT_SET	            0x101
 # define DP_LANE_COUNT_MASK		    0x0f
+# define DP_POST_LT_ADJ_REQ_GRANTED         (1 << 5) /* 1.3 */
 # define DP_LANE_COUNT_ENHANCED_FRAME_EN    (1 << 7)
 
 #define DP_TRAINING_PATTERN_SET	            0x102
@@ -800,6 +802,7 @@
 
 #define DP_LANE_ALIGN_STATUS_UPDATED                    0x204
 #define  DP_INTERLANE_ALIGN_DONE                        (1 << 0)
+#define  DP_POST_LT_ADJ_REQ_IN_PROGRESS                 (1 << 1) /* 1.3 */
 #define  DP_128B132B_DPRX_EQ_INTERLANE_ALIGN_DONE       (1 << 2) /* 2.0 E11 */
 #define  DP_128B132B_DPRX_CDS_INTERLANE_ALIGN_DONE      (1 << 3) /* 2.0 E11 */
 #define  DP_128B132B_LT_FAILED                          (1 << 4) /* 2.0 E11 */
-- 
cgit v1.2.3


From 3a9cf301794c1a49d95eeb13119ff490fb5cfe88 Mon Sep 17 00:00:00 2001
From: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date: Fri, 18 Jul 2025 15:01:52 +0300
Subject: drm/dp: Add POST_LT_ADJ_REQ helpers
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add small helpers (drm_dp_post_lt_adj_req_supported() and
drm_dp_post_lt_adj_req_in_progress()) to help with implementing
the POST_LT_ADJ_REQ sequence.

Tested-by: Imre Deak <imre.deak@intel.com>
Reviewed-by: Imre Deak <imre.deak@intel.com>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20250718120154.15492-3-ville.syrjala@linux.intel.com
---
 drivers/gpu/drm/display/drm_dp_helper.c | 8 ++++++++
 include/drm/display/drm_dp_helper.h     | 8 ++++++++
 2 files changed, 16 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/display/drm_dp_helper.c b/drivers/gpu/drm/display/drm_dp_helper.c
index 4aaeae4fa03c..5426db21e53f 100644
--- a/drivers/gpu/drm/display/drm_dp_helper.c
+++ b/drivers/gpu/drm/display/drm_dp_helper.c
@@ -123,6 +123,14 @@ bool drm_dp_clock_recovery_ok(const u8 link_status[DP_LINK_STATUS_SIZE],
 }
 EXPORT_SYMBOL(drm_dp_clock_recovery_ok);
 
+bool drm_dp_post_lt_adj_req_in_progress(const u8 link_status[DP_LINK_STATUS_SIZE])
+{
+	u8 lane_align = dp_link_status(link_status, DP_LANE_ALIGN_STATUS_UPDATED);
+
+	return lane_align & DP_POST_LT_ADJ_REQ_IN_PROGRESS;
+}
+EXPORT_SYMBOL(drm_dp_post_lt_adj_req_in_progress);
+
 u8 drm_dp_get_adjust_request_voltage(const u8 link_status[DP_LINK_STATUS_SIZE],
 				     int lane)
 {
diff --git a/include/drm/display/drm_dp_helper.h b/include/drm/display/drm_dp_helper.h
index 87caa4f1fdb8..52ce28097015 100644
--- a/include/drm/display/drm_dp_helper.h
+++ b/include/drm/display/drm_dp_helper.h
@@ -37,6 +37,7 @@ bool drm_dp_channel_eq_ok(const u8 link_status[DP_LINK_STATUS_SIZE],
 			  int lane_count);
 bool drm_dp_clock_recovery_ok(const u8 link_status[DP_LINK_STATUS_SIZE],
 			      int lane_count);
+bool drm_dp_post_lt_adj_req_in_progress(const u8 link_status[DP_LINK_STATUS_SIZE]);
 u8 drm_dp_get_adjust_request_voltage(const u8 link_status[DP_LINK_STATUS_SIZE],
 				     int lane);
 u8 drm_dp_get_adjust_request_pre_emphasis(const u8 link_status[DP_LINK_STATUS_SIZE],
@@ -155,6 +156,13 @@ drm_dp_enhanced_frame_cap(const u8 dpcd[DP_RECEIVER_CAP_SIZE])
 		(dpcd[DP_MAX_LANE_COUNT] & DP_ENHANCED_FRAME_CAP);
 }
 
+static inline bool
+drm_dp_post_lt_adj_req_supported(const u8 dpcd[DP_RECEIVER_CAP_SIZE])
+{
+	return dpcd[DP_DPCD_REV] >= 0x13 &&
+		(dpcd[DP_MAX_LANE_COUNT] & DP_POST_LT_ADJ_REQ_SUPPORTED);
+}
+
 static inline bool
 drm_dp_fast_training_cap(const u8 dpcd[DP_RECEIVER_CAP_SIZE])
 {
-- 
cgit v1.2.3


From 3a33c48876bc437d149cc15e0c4b96efb14912a1 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Thu, 18 Sep 2025 17:39:51 +0200
Subject: drm/format-helper: Remove drm_fb_blit()

The function is unused; remove it.

Instead of relying on a general blit helper, drivers should pick a blit
function by themselves from their list of supported color formats.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://lore.kernel.org/r/20250918154207.84714-4-tzimmermann@suse.de
---
 drivers/gpu/drm/drm_format_helper.c | 91 -------------------------------------
 include/drm/drm_format_helper.h     |  4 --
 2 files changed, 95 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_format_helper.c b/drivers/gpu/drm/drm_format_helper.c
index 006836554cc2..6cddf05c493b 100644
--- a/drivers/gpu/drm/drm_format_helper.c
+++ b/drivers/gpu/drm/drm_format_helper.c
@@ -1165,97 +1165,6 @@ void drm_fb_argb8888_to_argb4444(struct iosys_map *dst, const unsigned int *dst_
 }
 EXPORT_SYMBOL(drm_fb_argb8888_to_argb4444);
 
-/**
- * drm_fb_blit - Copy parts of a framebuffer to display memory
- * @dst:	Array of display-memory addresses to copy to
- * @dst_pitch: Array of numbers of bytes between the start of two consecutive scanlines
- *             within @dst; can be NULL if scanlines are stored next to each other.
- * @dst_format:	FOURCC code of the display's color format
- * @src:	The framebuffer memory to copy from
- * @fb:		The framebuffer to copy from
- * @clip:	Clip rectangle area to copy
- * @state: Transform and conversion state
- *
- * This function copies parts of a framebuffer to display memory. If the
- * formats of the display and the framebuffer mismatch, the blit function
- * will attempt to convert between them during the process. The parameters @dst,
- * @dst_pitch and @src refer to arrays. Each array must have at least as many
- * entries as there are planes in @dst_format's format. Each entry stores the
- * value for the format's respective color plane at the same index.
- *
- * This function does not apply clipping on @dst (i.e. the destination is at the
- * top-left corner).
- *
- * Returns:
- * 0 on success, or
- * -EINVAL if the color-format conversion failed, or
- * a negative error code otherwise.
- */
-int drm_fb_blit(struct iosys_map *dst, const unsigned int *dst_pitch, uint32_t dst_format,
-		const struct iosys_map *src, const struct drm_framebuffer *fb,
-		const struct drm_rect *clip, struct drm_format_conv_state *state)
-{
-	uint32_t fb_format = fb->format->format;
-
-	if (fb_format == dst_format) {
-		drm_fb_memcpy(dst, dst_pitch, src, fb, clip);
-		return 0;
-	} else if (fb_format == (dst_format | DRM_FORMAT_BIG_ENDIAN)) {
-		drm_fb_swab(dst, dst_pitch, src, fb, clip, false, state);
-		return 0;
-	} else if (fb_format == (dst_format & ~DRM_FORMAT_BIG_ENDIAN)) {
-		drm_fb_swab(dst, dst_pitch, src, fb, clip, false, state);
-		return 0;
-	} else if (fb_format == DRM_FORMAT_XRGB8888) {
-		if (dst_format == DRM_FORMAT_RGB565) {
-			drm_fb_xrgb8888_to_rgb565(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_XRGB1555) {
-			drm_fb_xrgb8888_to_xrgb1555(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_ARGB1555) {
-			drm_fb_xrgb8888_to_argb1555(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_RGBA5551) {
-			drm_fb_xrgb8888_to_rgba5551(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_RGB888) {
-			drm_fb_xrgb8888_to_rgb888(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_BGR888) {
-			drm_fb_xrgb8888_to_bgr888(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_ARGB8888) {
-			drm_fb_xrgb8888_to_argb8888(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_XBGR8888) {
-			drm_fb_xrgb8888_to_xbgr8888(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_ABGR8888) {
-			drm_fb_xrgb8888_to_abgr8888(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_XRGB2101010) {
-			drm_fb_xrgb8888_to_xrgb2101010(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_ARGB2101010) {
-			drm_fb_xrgb8888_to_argb2101010(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_BGRX8888) {
-			drm_fb_swab(dst, dst_pitch, src, fb, clip, false, state);
-			return 0;
-		} else if (dst_format == DRM_FORMAT_RGB332) {
-			drm_fb_xrgb8888_to_rgb332(dst, dst_pitch, src, fb, clip, state);
-			return 0;
-		}
-	}
-
-	drm_warn_once(fb->dev, "No conversion helper from %p4cc to %p4cc found.\n",
-		      &fb_format, &dst_format);
-
-	return -EINVAL;
-}
-EXPORT_SYMBOL(drm_fb_blit);
-
 static void drm_fb_gray8_to_gray2_line(void *dbuf, const void *sbuf, unsigned int pixels)
 {
 	u8 *dbuf8 = dbuf;
diff --git a/include/drm/drm_format_helper.h b/include/drm/drm_format_helper.h
index 32d57d6c5327..2b5c1aef80b0 100644
--- a/include/drm/drm_format_helper.h
+++ b/include/drm/drm_format_helper.h
@@ -128,10 +128,6 @@ void drm_fb_argb8888_to_argb4444(struct iosys_map *dst, const unsigned int *dst_
 				 const struct iosys_map *src, const struct drm_framebuffer *fb,
 				 const struct drm_rect *clip, struct drm_format_conv_state *state);
 
-int drm_fb_blit(struct iosys_map *dst, const unsigned int *dst_pitch, uint32_t dst_format,
-		const struct iosys_map *src, const struct drm_framebuffer *fb,
-		const struct drm_rect *clip, struct drm_format_conv_state *state);
-
 void drm_fb_xrgb8888_to_mono(struct iosys_map *dst, const unsigned int *dst_pitch,
 			     const struct iosys_map *src, const struct drm_framebuffer *fb,
 			     const struct drm_rect *clip, struct drm_format_conv_state *state);
-- 
cgit v1.2.3


From 32620e176443bf23ec81bfe8f177c6721a904864 Mon Sep 17 00:00:00 2001
From: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Date: Mon, 22 Sep 2025 20:33:15 +0530
Subject: drm/pcids: Split PTL pciids group to make wcl subplatform

To form the WCL platform as a subplatform of PTL in definition,
WCL pci ids are splited into saparate group from PTL.
So update the pciidlist struct to cover all the pci ids.

v2:
- Squash wcl description in single patch for display and xe.(jani,gustavo)

Signed-off-by: Dnyaneshwar Bhadane <dnyaneshwar.bhadane@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Signed-off-by: Suraj Kandpal <suraj.kandpal@intel.com>
Link: https://lore.kernel.org/r/20250922150317.2334680-2-dnyaneshwar.bhadane@intel.com
---
 drivers/gpu/drm/i915/display/intel_display_device.c | 1 +
 drivers/gpu/drm/xe/xe_pci.c                         | 1 +
 include/drm/intel/pciids.h                          | 5 ++++-
 3 files changed, 6 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/i915/display/intel_display_device.c b/drivers/gpu/drm/i915/display/intel_display_device.c
index a002bc6ce7b0..a9a36176096f 100644
--- a/drivers/gpu/drm/i915/display/intel_display_device.c
+++ b/drivers/gpu/drm/i915/display/intel_display_device.c
@@ -1482,6 +1482,7 @@ static const struct {
 	INTEL_LNL_IDS(INTEL_DISPLAY_DEVICE, &lnl_desc),
 	INTEL_BMG_IDS(INTEL_DISPLAY_DEVICE, &bmg_desc),
 	INTEL_PTL_IDS(INTEL_DISPLAY_DEVICE, &ptl_desc),
+	INTEL_WCL_IDS(INTEL_DISPLAY_DEVICE, &ptl_desc),
 };
 
 static const struct {
diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
index 046d330bad34..7ae316534eb6 100644
--- a/drivers/gpu/drm/xe/xe_pci.c
+++ b/drivers/gpu/drm/xe/xe_pci.c
@@ -374,6 +374,7 @@ static const struct pci_device_id pciidlist[] = {
 	INTEL_LNL_IDS(INTEL_VGA_DEVICE, &lnl_desc),
 	INTEL_BMG_IDS(INTEL_VGA_DEVICE, &bmg_desc),
 	INTEL_PTL_IDS(INTEL_VGA_DEVICE, &ptl_desc),
+	INTEL_WCL_IDS(INTEL_VGA_DEVICE, &ptl_desc),
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, pciidlist);
diff --git a/include/drm/intel/pciids.h b/include/drm/intel/pciids.h
index da6301a6fcea..69d4ae92d822 100644
--- a/include/drm/intel/pciids.h
+++ b/include/drm/intel/pciids.h
@@ -877,7 +877,10 @@
 	MACRO__(0xB08F, ## __VA_ARGS__), \
 	MACRO__(0xB090, ## __VA_ARGS__), \
 	MACRO__(0xB0A0, ## __VA_ARGS__), \
-	MACRO__(0xB0B0, ## __VA_ARGS__), \
+	MACRO__(0xB0B0, ## __VA_ARGS__)
+
+/* WCL */
+#define INTEL_WCL_IDS(MACRO__, ...) \
 	MACRO__(0xFD80, ## __VA_ARGS__), \
 	MACRO__(0xFD81, ## __VA_ARGS__)
 
-- 
cgit v1.2.3


From be0bd958cedd4adaaaa4c5aa1e0361de01e208e3 Mon Sep 17 00:00:00 2001
From: Shengjiu Wang <shengjiu.wang@nxp.com>
Date: Tue, 23 Sep 2025 13:29:56 +0800
Subject: ALSA: Add definitions for the bits in IEC958 subframe

The IEC958 subframe format SNDRV_PCM_FMTBIT_IEC958_SUBFRAME_LE are used
in HDMI and DisplayPort to describe the audio stream, but hardware device
may need to reorder the IEC958 bits for internal transmission, so need
these standard bits definitions for IEC958 subframe format.

Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Reviewed-by: Takashi Iwai <tiwai@suse.de>
Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Signed-off-by: Liu Ying <victor.liu@nxp.com>
Link: https://lore.kernel.org/r/20250923053001.2678596-3-shengjiu.wang@nxp.com
---
 include/sound/asoundef.h | 9 +++++++++
 1 file changed, 9 insertions(+)

(limited to 'include')

diff --git a/include/sound/asoundef.h b/include/sound/asoundef.h
index 09b2c3dffb30..c4a929d4fd51 100644
--- a/include/sound/asoundef.h
+++ b/include/sound/asoundef.h
@@ -12,6 +12,15 @@
  *        Digital audio interface					    *
  *                                                                          *
  ****************************************************************************/
+/* IEC958 subframe format */
+#define IEC958_SUBFRAME_PREAMBLE_MASK	(0xfU)
+#define IEC958_SUBFRAME_AUXILIARY_MASK	(0xfU << 4)
+#define IEC958_SUBFRAME_SAMPLE_24_MASK	(0xffffffU << 4)
+#define IEC958_SUBFRAME_SAMPLE_20_MASK	(0xfffffU << 8)
+#define IEC958_SUBFRAME_VALIDITY	(0x1U << 28)
+#define IEC958_SUBFRAME_USER_DATA	(0x1U << 29)
+#define IEC958_SUBFRAME_CHANNEL_STATUS	(0x1U << 30)
+#define IEC958_SUBFRAME_PARITY		(0x1U << 31)
 
 /* AES/IEC958 channel status bits */
 #define IEC958_AES0_PROFESSIONAL	(1<<0)	/* 0 = consumer, 1 = professional */
-- 
cgit v1.2.3


From 21d4c95e4b06e6f2d981f476bbbd934451a2edcd Mon Sep 17 00:00:00 2001
From: Shengjiu Wang <shengjiu.wang@nxp.com>
Date: Tue, 23 Sep 2025 13:29:57 +0800
Subject: drm/bridge: dw-hdmi: Add API dw_hdmi_to_plat_data() to get plat_data

Add API dw_hdmi_to_plat_data() to fetch plat_data because audio device
driver needs it to enable(disable)_audio().

Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Acked-by: Liu Ying <victor.liu@nxp.com>
Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Signed-off-by: Liu Ying <victor.liu@nxp.com>
Link: https://lore.kernel.org/r/20250923053001.2678596-4-shengjiu.wang@nxp.com
---
 drivers/gpu/drm/bridge/synopsys/dw-hdmi.c | 6 ++++++
 include/drm/bridge/dw_hdmi.h              | 2 ++
 2 files changed, 8 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
index 206b099a35e9..8d096b569cf1 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
@@ -198,6 +198,12 @@ struct dw_hdmi {
 	enum drm_connector_status last_connector_result;
 };
 
+const struct dw_hdmi_plat_data *dw_hdmi_to_plat_data(struct dw_hdmi *hdmi)
+{
+	return hdmi->plat_data;
+}
+EXPORT_SYMBOL_GPL(dw_hdmi_to_plat_data);
+
 #define HDMI_IH_PHY_STAT0_RX_SENSE \
 	(HDMI_IH_PHY_STAT0_RX_SENSE0 | HDMI_IH_PHY_STAT0_RX_SENSE1 | \
 	 HDMI_IH_PHY_STAT0_RX_SENSE2 | HDMI_IH_PHY_STAT0_RX_SENSE3)
diff --git a/include/drm/bridge/dw_hdmi.h b/include/drm/bridge/dw_hdmi.h
index 6a46baa0737c..b8fc4fdf5a21 100644
--- a/include/drm/bridge/dw_hdmi.h
+++ b/include/drm/bridge/dw_hdmi.h
@@ -208,4 +208,6 @@ void dw_hdmi_phy_setup_hpd(struct dw_hdmi *hdmi, void *data);
 
 bool dw_hdmi_bus_fmt_is_420(struct dw_hdmi *hdmi);
 
+const struct dw_hdmi_plat_data *dw_hdmi_to_plat_data(struct dw_hdmi *hdmi);
+
 #endif /* __IMX_HDMI_H__ */
-- 
cgit v1.2.3


From 80c5d14434c94074fcf89016f89c65c209475268 Mon Sep 17 00:00:00 2001
From: Shengjiu Wang <shengjiu.wang@nxp.com>
Date: Tue, 23 Sep 2025 13:29:58 +0800
Subject: drm/bridge: dw-hdmi: Add API dw_hdmi_set_sample_iec958() for iec958
 format

Add API dw_hdmi_set_sample_iec958() for IEC958 format because audio device
driver needs IEC958 information to configure this specific setting.

Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Acked-by: Liu Ying <victor.liu@nxp.com>
Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Signed-off-by: Liu Ying <victor.liu@nxp.com>
Link: https://lore.kernel.org/r/20250923053001.2678596-5-shengjiu.wang@nxp.com
---
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-gp-audio.c |  5 +++++
 drivers/gpu/drm/bridge/synopsys/dw-hdmi.c          | 12 +++++++++++-
 include/drm/bridge/dw_hdmi.h                       |  3 ++-
 3 files changed, 18 insertions(+), 2 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-gp-audio.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-gp-audio.c
index ab18f9a3bf23..df7a37eb47f4 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-gp-audio.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-gp-audio.c
@@ -90,6 +90,11 @@ static int audio_hw_params(struct device *dev,  void *data,
 				   params->iec.status[0] & IEC958_AES0_NONAUDIO);
 	dw_hdmi_set_sample_width(dw->data.hdmi, params->sample_width);
 
+	if (daifmt->bit_fmt == SNDRV_PCM_FORMAT_IEC958_SUBFRAME_LE)
+		dw_hdmi_set_sample_iec958(dw->data.hdmi, 1);
+	else
+		dw_hdmi_set_sample_iec958(dw->data.hdmi, 0);
+
 	return 0;
 }
 
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
index 8d096b569cf1..3b77e73ac0ea 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi.c
@@ -177,6 +177,7 @@ struct dw_hdmi {
 
 	spinlock_t audio_lock;
 	struct mutex audio_mutex;
+	unsigned int sample_iec958;
 	unsigned int sample_non_pcm;
 	unsigned int sample_width;
 	unsigned int sample_rate;
@@ -718,6 +719,14 @@ void dw_hdmi_set_sample_non_pcm(struct dw_hdmi *hdmi, unsigned int non_pcm)
 }
 EXPORT_SYMBOL_GPL(dw_hdmi_set_sample_non_pcm);
 
+void dw_hdmi_set_sample_iec958(struct dw_hdmi *hdmi, unsigned int iec958)
+{
+	mutex_lock(&hdmi->audio_mutex);
+	hdmi->sample_iec958 = iec958;
+	mutex_unlock(&hdmi->audio_mutex);
+}
+EXPORT_SYMBOL_GPL(dw_hdmi_set_sample_iec958);
+
 void dw_hdmi_set_sample_rate(struct dw_hdmi *hdmi, unsigned int rate)
 {
 	mutex_lock(&hdmi->audio_mutex);
@@ -849,7 +858,8 @@ static void dw_hdmi_gp_audio_enable(struct dw_hdmi *hdmi)
 				    hdmi->channels,
 				    hdmi->sample_width,
 				    hdmi->sample_rate,
-				    hdmi->sample_non_pcm);
+				    hdmi->sample_non_pcm,
+				    hdmi->sample_iec958);
 }
 
 static void dw_hdmi_gp_audio_disable(struct dw_hdmi *hdmi)
diff --git a/include/drm/bridge/dw_hdmi.h b/include/drm/bridge/dw_hdmi.h
index b8fc4fdf5a21..095cdd9b7424 100644
--- a/include/drm/bridge/dw_hdmi.h
+++ b/include/drm/bridge/dw_hdmi.h
@@ -145,7 +145,7 @@ struct dw_hdmi_plat_data {
 
 	/* Platform-specific audio enable/disable (optional) */
 	void (*enable_audio)(struct dw_hdmi *hdmi, int channel,
-			     int width, int rate, int non_pcm);
+			     int width, int rate, int non_pcm, int iec958);
 	void (*disable_audio)(struct dw_hdmi *hdmi);
 
 	/* Vendor PHY support */
@@ -179,6 +179,7 @@ void dw_hdmi_setup_rx_sense(struct dw_hdmi *hdmi, bool hpd, bool rx_sense);
 int dw_hdmi_set_plugged_cb(struct dw_hdmi *hdmi, hdmi_codec_plugged_cb fn,
 			   struct device *codec_dev);
 void dw_hdmi_set_sample_non_pcm(struct dw_hdmi *hdmi, unsigned int non_pcm);
+void dw_hdmi_set_sample_iec958(struct dw_hdmi *hdmi, unsigned int iec958);
 void dw_hdmi_set_sample_width(struct dw_hdmi *hdmi, unsigned int width);
 void dw_hdmi_set_sample_rate(struct dw_hdmi *hdmi, unsigned int rate);
 void dw_hdmi_set_channel_count(struct dw_hdmi *hdmi, unsigned int cnt);
-- 
cgit v1.2.3


From 0205fae6327a4ef6bdb9b6dd9722c17613c422cb Mon Sep 17 00:00:00 2001
From: Shengjiu Wang <shengjiu.wang@nxp.com>
Date: Tue, 23 Sep 2025 13:29:59 +0800
Subject: drm/bridge: imx: add driver for HDMI TX Parallel Audio Interface

The HDMI TX Parallel Audio Interface (HTX_PAI) is a digital module that
acts as the bridge between the Audio Subsystem to the HDMI TX Controller.
This IP block is found in the HDMI subsystem of the i.MX8MP SoC.

Data received from the audio subsystem can have an arbitrary component
ordering. The HTX_PAI block has integrated muxing options to select which
sections of the 32-bit input data word will be mapped to each IEC60958
field. The HTX_PAI_FIELD_CTRL register contains mux selects to
individually select P,C,U,V,Data, and Preamble.

Use component helper so that imx8mp-hdmi-tx will be aggregate driver,
imx8mp-hdmi-pai will be component driver, then imx8mp-hdmi-pai can use
bind() ops to get the plat_data from imx8mp-hdmi-tx device.

Signed-off-by: Shengjiu Wang <shengjiu.wang@nxp.com>
Reviewed-by: Liu Ying <victor.liu@nxp.com>
Tested-by: Alexander Stein <alexander.stein@ew.tq-group.com>
Signed-off-by: Liu Ying <victor.liu@nxp.com>
Link: https://lore.kernel.org/r/20250923053001.2678596-6-shengjiu.wang@nxp.com
---
 drivers/gpu/drm/bridge/imx/Kconfig           |  11 ++
 drivers/gpu/drm/bridge/imx/Makefile          |   1 +
 drivers/gpu/drm/bridge/imx/imx8mp-hdmi-pai.c | 158 +++++++++++++++++++++++++++
 drivers/gpu/drm/bridge/imx/imx8mp-hdmi-tx.c  |  65 ++++++++++-
 include/drm/bridge/dw_hdmi.h                 |   6 +
 5 files changed, 236 insertions(+), 5 deletions(-)
 create mode 100644 drivers/gpu/drm/bridge/imx/imx8mp-hdmi-pai.c

(limited to 'include')

diff --git a/drivers/gpu/drm/bridge/imx/Kconfig b/drivers/gpu/drm/bridge/imx/Kconfig
index 9a480c6abb85..b9028a5e5a06 100644
--- a/drivers/gpu/drm/bridge/imx/Kconfig
+++ b/drivers/gpu/drm/bridge/imx/Kconfig
@@ -18,12 +18,23 @@ config DRM_IMX8MP_DW_HDMI_BRIDGE
 	depends on OF
 	depends on COMMON_CLK
 	select DRM_DW_HDMI
+	imply DRM_IMX8MP_HDMI_PAI
 	imply DRM_IMX8MP_HDMI_PVI
 	imply PHY_FSL_SAMSUNG_HDMI_PHY
 	help
 	  Choose this to enable support for the internal HDMI encoder found
 	  on the i.MX8MP SoC.
 
+config DRM_IMX8MP_HDMI_PAI
+	tristate "Freescale i.MX8MP HDMI PAI bridge support"
+	depends on OF
+	select DRM_DW_HDMI
+	select REGMAP
+	select REGMAP_MMIO
+	help
+	  Choose this to enable support for the internal HDMI TX Parallel
+	  Audio Interface found on the Freescale i.MX8MP SoC.
+
 config DRM_IMX8MP_HDMI_PVI
 	tristate "Freescale i.MX8MP HDMI PVI bridge support"
 	depends on OF
diff --git a/drivers/gpu/drm/bridge/imx/Makefile b/drivers/gpu/drm/bridge/imx/Makefile
index dd5d48584806..8d01fda25451 100644
--- a/drivers/gpu/drm/bridge/imx/Makefile
+++ b/drivers/gpu/drm/bridge/imx/Makefile
@@ -1,6 +1,7 @@
 obj-$(CONFIG_DRM_IMX_LDB_HELPER) += imx-ldb-helper.o
 obj-$(CONFIG_DRM_IMX_LEGACY_BRIDGE) += imx-legacy-bridge.o
 obj-$(CONFIG_DRM_IMX8MP_DW_HDMI_BRIDGE) += imx8mp-hdmi-tx.o
+obj-$(CONFIG_DRM_IMX8MP_HDMI_PAI) += imx8mp-hdmi-pai.o
 obj-$(CONFIG_DRM_IMX8MP_HDMI_PVI) += imx8mp-hdmi-pvi.o
 obj-$(CONFIG_DRM_IMX8QM_LDB) += imx8qm-ldb.o
 obj-$(CONFIG_DRM_IMX8QXP_LDB) += imx8qxp-ldb.o
diff --git a/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-pai.c b/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-pai.c
new file mode 100644
index 000000000000..8d13a35b206a
--- /dev/null
+++ b/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-pai.c
@@ -0,0 +1,158 @@
+// SPDX-License-Identifier: GPL-2.0+
+/*
+ * Copyright 2025 NXP
+ */
+
+#include <linux/bitfield.h>
+#include <linux/component.h>
+#include <linux/module.h>
+#include <linux/of_platform.h>
+#include <linux/platform_device.h>
+#include <linux/regmap.h>
+#include <drm/bridge/dw_hdmi.h>
+#include <sound/asoundef.h>
+
+#define HTX_PAI_CTRL			0x00
+#define   ENABLE			BIT(0)
+
+#define HTX_PAI_CTRL_EXT		0x04
+#define   WTMK_HIGH_MASK		GENMASK(31, 24)
+#define   WTMK_LOW_MASK			GENMASK(23, 16)
+#define   NUM_CH_MASK			GENMASK(10, 8)
+#define   WTMK_HIGH(n)			FIELD_PREP(WTMK_HIGH_MASK, (n))
+#define   WTMK_LOW(n)			FIELD_PREP(WTMK_LOW_MASK, (n))
+#define   NUM_CH(n)			FIELD_PREP(NUM_CH_MASK, (n) - 1)
+
+#define HTX_PAI_FIELD_CTRL		0x08
+#define   PRE_SEL			GENMASK(28, 24)
+#define   D_SEL				GENMASK(23, 20)
+#define   V_SEL				GENMASK(19, 15)
+#define   U_SEL				GENMASK(14, 10)
+#define   C_SEL				GENMASK(9, 5)
+#define   P_SEL				GENMASK(4, 0)
+
+struct imx8mp_hdmi_pai {
+	struct regmap	*regmap;
+};
+
+static void imx8mp_hdmi_pai_enable(struct dw_hdmi *dw_hdmi, int channel,
+				   int width, int rate, int non_pcm,
+				   int iec958)
+{
+	const struct dw_hdmi_plat_data *pdata = dw_hdmi_to_plat_data(dw_hdmi);
+	struct imx8mp_hdmi_pai *hdmi_pai = pdata->priv_audio;
+	int val;
+
+	/* PAI set control extended */
+	val =  WTMK_HIGH(3) | WTMK_LOW(3);
+	val |= NUM_CH(channel);
+	regmap_write(hdmi_pai->regmap, HTX_PAI_CTRL_EXT, val);
+
+	/* IEC60958 format */
+	if (iec958) {
+		val = FIELD_PREP_CONST(P_SEL,
+				       __bf_shf(IEC958_SUBFRAME_PARITY));
+		val |= FIELD_PREP_CONST(C_SEL,
+					__bf_shf(IEC958_SUBFRAME_CHANNEL_STATUS));
+		val |= FIELD_PREP_CONST(U_SEL,
+					__bf_shf(IEC958_SUBFRAME_USER_DATA));
+		val |= FIELD_PREP_CONST(V_SEL,
+					__bf_shf(IEC958_SUBFRAME_VALIDITY));
+		val |= FIELD_PREP_CONST(D_SEL,
+					__bf_shf(IEC958_SUBFRAME_SAMPLE_24_MASK));
+		val |= FIELD_PREP_CONST(PRE_SEL,
+					__bf_shf(IEC958_SUBFRAME_PREAMBLE_MASK));
+	} else {
+		/*
+		 * The allowed PCM widths are 24bit and 32bit, as they are supported
+		 * by aud2htx module.
+		 * for 24bit, D_SEL = 0, select all the bits.
+		 * for 32bit, D_SEL = 8, select 24bit in MSB.
+		 */
+		val = FIELD_PREP(D_SEL, width - 24);
+	}
+
+	regmap_write(hdmi_pai->regmap, HTX_PAI_FIELD_CTRL, val);
+
+	/* PAI start running */
+	regmap_write(hdmi_pai->regmap, HTX_PAI_CTRL, ENABLE);
+}
+
+static void imx8mp_hdmi_pai_disable(struct dw_hdmi *dw_hdmi)
+{
+	const struct dw_hdmi_plat_data *pdata = dw_hdmi_to_plat_data(dw_hdmi);
+	struct imx8mp_hdmi_pai *hdmi_pai = pdata->priv_audio;
+
+	/* Stop PAI */
+	regmap_write(hdmi_pai->regmap, HTX_PAI_CTRL, 0);
+}
+
+static const struct regmap_config imx8mp_hdmi_pai_regmap_config = {
+	.reg_bits = 32,
+	.reg_stride = 4,
+	.val_bits = 32,
+	.max_register = HTX_PAI_FIELD_CTRL,
+};
+
+static int imx8mp_hdmi_pai_bind(struct device *dev, struct device *master, void *data)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct dw_hdmi_plat_data *plat_data = data;
+	struct imx8mp_hdmi_pai *hdmi_pai;
+	struct resource *res;
+	void __iomem *base;
+
+	hdmi_pai = devm_kzalloc(dev, sizeof(*hdmi_pai), GFP_KERNEL);
+	if (!hdmi_pai)
+		return -ENOMEM;
+
+	base = devm_platform_get_and_ioremap_resource(pdev, 0, &res);
+	if (IS_ERR(base))
+		return PTR_ERR(base);
+
+	hdmi_pai->regmap = devm_regmap_init_mmio_clk(dev, "apb", base,
+						     &imx8mp_hdmi_pai_regmap_config);
+	if (IS_ERR(hdmi_pai->regmap)) {
+		dev_err(dev, "regmap init failed\n");
+		return PTR_ERR(hdmi_pai->regmap);
+	}
+
+	plat_data->enable_audio = imx8mp_hdmi_pai_enable;
+	plat_data->disable_audio = imx8mp_hdmi_pai_disable;
+	plat_data->priv_audio = hdmi_pai;
+
+	return 0;
+}
+
+static const struct component_ops imx8mp_hdmi_pai_ops = {
+	.bind   = imx8mp_hdmi_pai_bind,
+};
+
+static int imx8mp_hdmi_pai_probe(struct platform_device *pdev)
+{
+	return component_add(&pdev->dev, &imx8mp_hdmi_pai_ops);
+}
+
+static void imx8mp_hdmi_pai_remove(struct platform_device *pdev)
+{
+	component_del(&pdev->dev, &imx8mp_hdmi_pai_ops);
+}
+
+static const struct of_device_id imx8mp_hdmi_pai_of_table[] = {
+	{ .compatible = "fsl,imx8mp-hdmi-pai" },
+	{ /* Sentinel */ }
+};
+MODULE_DEVICE_TABLE(of, imx8mp_hdmi_pai_of_table);
+
+static struct platform_driver imx8mp_hdmi_pai_platform_driver = {
+	.probe		= imx8mp_hdmi_pai_probe,
+	.remove		= imx8mp_hdmi_pai_remove,
+	.driver		= {
+		.name	= "imx8mp-hdmi-pai",
+		.of_match_table = imx8mp_hdmi_pai_of_table,
+	},
+};
+module_platform_driver(imx8mp_hdmi_pai_platform_driver);
+
+MODULE_DESCRIPTION("i.MX8MP HDMI PAI driver");
+MODULE_LICENSE("GPL");
diff --git a/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-tx.c b/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-tx.c
index 1e7a789ec289..32fd3554e267 100644
--- a/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-tx.c
+++ b/drivers/gpu/drm/bridge/imx/imx8mp-hdmi-tx.c
@@ -5,11 +5,13 @@
  */
 
 #include <linux/clk.h>
+#include <linux/component.h>
 #include <linux/mod_devicetable.h>
 #include <linux/module.h>
 #include <linux/platform_device.h>
 #include <drm/bridge/dw_hdmi.h>
 #include <drm/drm_modes.h>
+#include <drm/drm_of.h>
 
 struct imx8mp_hdmi {
 	struct dw_hdmi_plat_data plat_data;
@@ -79,10 +81,45 @@ static const struct dw_hdmi_phy_ops imx8mp_hdmi_phy_ops = {
 	.update_hpd	= dw_hdmi_phy_update_hpd,
 };
 
+static int imx8mp_dw_hdmi_bind(struct device *dev)
+{
+	struct platform_device *pdev = to_platform_device(dev);
+	struct imx8mp_hdmi *hdmi = dev_get_drvdata(dev);
+	int ret;
+
+	ret = component_bind_all(dev, &hdmi->plat_data);
+	if (ret)
+		return dev_err_probe(dev, ret, "component_bind_all failed!\n");
+
+	hdmi->dw_hdmi = dw_hdmi_probe(pdev, &hdmi->plat_data);
+	if (IS_ERR(hdmi->dw_hdmi)) {
+		component_unbind_all(dev, &hdmi->plat_data);
+		return PTR_ERR(hdmi->dw_hdmi);
+	}
+
+	return 0;
+}
+
+static void imx8mp_dw_hdmi_unbind(struct device *dev)
+{
+	struct imx8mp_hdmi *hdmi = dev_get_drvdata(dev);
+
+	dw_hdmi_remove(hdmi->dw_hdmi);
+
+	component_unbind_all(dev, &hdmi->plat_data);
+}
+
+static const struct component_master_ops imx8mp_dw_hdmi_ops = {
+	.bind   = imx8mp_dw_hdmi_bind,
+	.unbind = imx8mp_dw_hdmi_unbind,
+};
+
 static int imx8mp_dw_hdmi_probe(struct platform_device *pdev)
 {
 	struct device *dev = &pdev->dev;
 	struct dw_hdmi_plat_data *plat_data;
+	struct component_match *match = NULL;
+	struct device_node *remote;
 	struct imx8mp_hdmi *hdmi;
 
 	hdmi = devm_kzalloc(dev, sizeof(*hdmi), GFP_KERNEL);
@@ -102,20 +139,38 @@ static int imx8mp_dw_hdmi_probe(struct platform_device *pdev)
 	plat_data->priv_data = hdmi;
 	plat_data->phy_force_vendor = true;
 
-	hdmi->dw_hdmi = dw_hdmi_probe(pdev, plat_data);
-	if (IS_ERR(hdmi->dw_hdmi))
-		return PTR_ERR(hdmi->dw_hdmi);
-
 	platform_set_drvdata(pdev, hdmi);
 
+	/* port@2 is for hdmi_pai device */
+	remote = of_graph_get_remote_node(pdev->dev.of_node, 2, 0);
+	if (!remote) {
+		hdmi->dw_hdmi = dw_hdmi_probe(pdev, plat_data);
+		if (IS_ERR(hdmi->dw_hdmi))
+			return PTR_ERR(hdmi->dw_hdmi);
+	} else {
+		drm_of_component_match_add(dev, &match, component_compare_of, remote);
+
+		of_node_put(remote);
+
+		return component_master_add_with_match(dev, &imx8mp_dw_hdmi_ops, match);
+	}
+
 	return 0;
 }
 
 static void imx8mp_dw_hdmi_remove(struct platform_device *pdev)
 {
 	struct imx8mp_hdmi *hdmi = platform_get_drvdata(pdev);
+	struct device_node *remote;
 
-	dw_hdmi_remove(hdmi->dw_hdmi);
+	remote = of_graph_get_remote_node(pdev->dev.of_node, 2, 0);
+	if (remote) {
+		of_node_put(remote);
+
+		component_master_del(&pdev->dev, &imx8mp_dw_hdmi_ops);
+	} else {
+		dw_hdmi_remove(hdmi->dw_hdmi);
+	}
 }
 
 static int imx8mp_dw_hdmi_pm_suspend(struct device *dev)
diff --git a/include/drm/bridge/dw_hdmi.h b/include/drm/bridge/dw_hdmi.h
index 095cdd9b7424..336f062e1f9d 100644
--- a/include/drm/bridge/dw_hdmi.h
+++ b/include/drm/bridge/dw_hdmi.h
@@ -143,6 +143,12 @@ struct dw_hdmi_plat_data {
 					   const struct drm_display_info *info,
 					   const struct drm_display_mode *mode);
 
+	/*
+	 * priv_audio is specially used for additional audio device to get
+	 * driver data through this dw_hdmi_plat_data.
+	 */
+	void *priv_audio;
+
 	/* Platform-specific audio enable/disable (optional) */
 	void (*enable_audio)(struct dw_hdmi *hdmi, int channel,
 			     int width, int rate, int non_pcm, int iec958);
-- 
cgit v1.2.3


From 97825e1c6de7315cba9acb6c1371f1a87dedd904 Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Fri, 26 Sep 2025 14:10:32 +0300
Subject: drm/{i915,xe}: driver agnostic drm to display pointer chase

The display driver needs to get from the struct drm_device pointer to
the struct intel_display pointer. Currently, this depends on knowledge
of the struct drm_i915_private and struct xe_device definitions, but
we'd like to hide those definitions from display.

Require the struct drm_device and struct intel_display * members within
struct drm_i915_private and struct xe_device to be placed next to each
other, to be able to figure out the display pointer without knowledge of
the structures.

Use a generic dummy device structure to define the relative offsets of
the drm and display members, and add static assertions to ensure this
holds for both i915 and xe. Use the dummy structure to do the pointer
chase from struct drm_device * to struct intel_display *.

This requires moving the display member in struct xe_device after the
drm member.

Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjala <ville.syrjala@linux.intel.com>
Suggested-by: Simona Vetter <simona.vetter@ffwll.ch>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://lore.kernel.org/r/20250926111032.1188876-1-jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
---
 .../drm/i915/display/intel_display_conversion.c    | 20 +++++++----
 drivers/gpu/drm/i915/i915_driver.c                 |  4 +++
 drivers/gpu/drm/i915/i915_drv.h                    |  1 +
 drivers/gpu/drm/xe/display/xe_display.c            |  4 +++
 drivers/gpu/drm/xe/xe_device_types.h               |  7 ++--
 include/drm/intel/display_member.h                 | 42 ++++++++++++++++++++++
 6 files changed, 69 insertions(+), 9 deletions(-)
 create mode 100644 include/drm/intel/display_member.h

(limited to 'include')

diff --git a/drivers/gpu/drm/i915/display/intel_display_conversion.c b/drivers/gpu/drm/i915/display/intel_display_conversion.c
index d56065f22655..9a47aa38cf82 100644
--- a/drivers/gpu/drm/i915/display/intel_display_conversion.c
+++ b/drivers/gpu/drm/i915/display/intel_display_conversion.c
@@ -1,15 +1,21 @@
 // SPDX-License-Identifier: MIT
 /* Copyright © 2024 Intel Corporation */
 
-#include "i915_drv.h"
-#include "intel_display_conversion.h"
+#include <drm/intel/display_member.h>
 
-static struct intel_display *__i915_to_display(struct drm_i915_private *i915)
-{
-	return i915->display;
-}
+#include "intel_display_conversion.h"
 
 struct intel_display *__drm_to_display(struct drm_device *drm)
 {
-	return __i915_to_display(to_i915(drm));
+	/*
+	 * Note: This relies on both struct drm_i915_private and struct
+	 * xe_device having the struct drm_device and struct intel_display *
+	 * members at the same relative offsets, as defined by struct
+	 * __intel_generic_device.
+	 *
+	 * See also INTEL_DISPLAY_MEMBER_STATIC_ASSERT().
+	 */
+	struct __intel_generic_device *d = container_of(drm, struct __intel_generic_device, drm);
+
+	return d->display;
 }
diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c
index 95165e45de74..b46cb54ef5dc 100644
--- a/drivers/gpu/drm/i915/i915_driver.c
+++ b/drivers/gpu/drm/i915/i915_driver.c
@@ -46,6 +46,7 @@
 #include <drm/drm_ioctl.h>
 #include <drm/drm_managed.h>
 #include <drm/drm_probe_helper.h>
+#include <drm/intel/display_member.h>
 
 #include "display/i9xx_display_sr.h"
 #include "display/intel_bw.h"
@@ -737,6 +738,9 @@ static void i915_welcome_messages(struct drm_i915_private *dev_priv)
 			 "DRM_I915_DEBUG_RUNTIME_PM enabled\n");
 }
 
+/* Ensure drm and display members are placed properly. */
+INTEL_DISPLAY_MEMBER_STATIC_ASSERT(struct drm_i915_private, drm, display);
+
 static struct drm_i915_private *
 i915_driver_create(struct pci_dev *pdev, const struct pci_device_id *ent)
 {
diff --git a/drivers/gpu/drm/i915/i915_drv.h b/drivers/gpu/drm/i915/i915_drv.h
index 03e497d2081e..6e159bb8ad2f 100644
--- a/drivers/gpu/drm/i915/i915_drv.h
+++ b/drivers/gpu/drm/i915/i915_drv.h
@@ -174,6 +174,7 @@ struct i915_selftest_stash {
 struct drm_i915_private {
 	struct drm_device drm;
 
+	/* display device data, must be placed after drm device member */
 	struct intel_display *display;
 
 	/* FIXME: Device release actions should all be moved to drmm_ */
diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c
index 19e691fccf8c..5f4044e63185 100644
--- a/drivers/gpu/drm/xe/display/xe_display.c
+++ b/drivers/gpu/drm/xe/display/xe_display.c
@@ -13,6 +13,7 @@
 #include <drm/drm_drv.h>
 #include <drm/drm_managed.h>
 #include <drm/drm_probe_helper.h>
+#include <drm/intel/display_member.h>
 #include <uapi/drm/xe_drm.h>
 
 #include "soc/intel_dram.h"
@@ -35,6 +36,9 @@
 #include "skl_watermark.h"
 #include "xe_module.h"
 
+/* Ensure drm and display members are placed properly. */
+INTEL_DISPLAY_MEMBER_STATIC_ASSERT(struct xe_device, drm, display);
+
 /* Xe device functions */
 
 /**
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index a6c361db11d9..53264b2bb832 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -217,6 +217,11 @@ struct xe_device {
 	/** @drm: drm device */
 	struct drm_device drm;
 
+#if IS_ENABLED(CONFIG_DRM_XE_DISPLAY)
+	/** @display: display device data, must be placed after drm device member */
+	struct intel_display *display;
+#endif
+
 	/** @devcoredump: device coredump */
 	struct xe_devcoredump devcoredump;
 
@@ -617,8 +622,6 @@ struct xe_device {
 	 * drm_i915_private during build. After cleanup these should go away,
 	 * migrating to the right sub-structs
 	 */
-	struct intel_display *display;
-
 	const struct dram_info *dram_info;
 
 	/*
diff --git a/include/drm/intel/display_member.h b/include/drm/intel/display_member.h
new file mode 100644
index 000000000000..0319ea560b60
--- /dev/null
+++ b/include/drm/intel/display_member.h
@@ -0,0 +1,42 @@
+/* SPDX-License-Identifier: MIT */
+/* Copyright © 2025 Intel Corporation */
+
+#ifndef __DRM_INTEL_DISPLAY_H__
+#define __DRM_INTEL_DISPLAY_H__
+
+#include <linux/build_bug.h>
+#include <linux/stddef.h>
+#include <linux/stringify.h>
+
+#include <drm/drm_device.h>
+
+struct intel_display;
+
+/*
+ * A dummy device struct to define the relative offsets of drm and display
+ * members. With the members identically placed in struct drm_i915_private and
+ * struct xe_device, this allows figuring out the struct intel_display pointer
+ * without the definition of either driver specific structure.
+ */
+struct __intel_generic_device {
+	struct drm_device drm;
+	struct intel_display *display;
+};
+
+/**
+ * INTEL_DISPLAY_MEMBER_STATIC_ASSERT() - ensure correct placing of drm and display members
+ * @type: The struct to check
+ * @drm_member: Name of the struct drm_device member
+ * @display_member: Name of the struct intel_display * member.
+ *
+ * Use this static assert macro to ensure the struct drm_i915_private and struct
+ * xe_device struct drm_device and struct intel_display * members are at the
+ * same relative offsets.
+ */
+#define INTEL_DISPLAY_MEMBER_STATIC_ASSERT(type, drm_member, display_member) \
+	static_assert( \
+		offsetof(struct __intel_generic_device, display) - offsetof(struct __intel_generic_device, drm) == \
+		offsetof(type, display_member) - offsetof(type, drm_member), \
+		__stringify(type) " " __stringify(drm_member) " and " __stringify(display_member) " members at invalid offsets")
+
+#endif
-- 
cgit v1.2.3


From fb24aaf5415cc686fb0473eb782a7c8a7bab0469 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Thu, 21 Aug 2025 10:17:09 +0200
Subject: drm/dumb-buffers: Provide helper to set pitch and size

Add drm_modes_size_dumb(), a helper to calculate the dumb-buffer
scanline pitch and allocation size. Implementations of struct
drm_driver.dumb_create can call the new helper for their size
computations.

There is currently quite a bit of code duplication among DRM's
memory managers. Each calculates scanline pitch and buffer size
from the given arguments, but the implementations are inconsistent
in how they treat alignment and format support. Later patches will
unify this code on top of drm_mode_size_dumb() as much as possible.

drm_mode_size_dumb() uses existing 4CC format helpers to interpret
the given color mode. This makes the dumb-buffer interface behave
similar the kernel's video= parameter. Current per-driver implementations
again likely have subtle differences or bugs in how they support color
modes.

The dumb-buffer UAPI is only specified for known color modes. These
values describe linear, single-plane RGB color formats or legacy index
formats. Other values should not be specified. But some user space
still does. So for unknown color modes, there are a number of known
exceptions for which drm_mode_size_dumb() calculates the pitch from
the bpp value, as before. All other values work the same but print
an error.

v6:
- document additional use cases for DUMB_CREATE2 in TODO list (Tomi)
- fix typos in documentation (Tomi)
v5:
- check for overflows with check_mul_overflow() (Tomi)
v4:
- use %u conversion specifier (Geert)
- list DRM_FORMAT_Dn in UAPI docs (Geert)
- avoid dmesg spamming with drm_warn_once() (Sima)
- add more information about bpp special case (Sima)
- clarify parameters for hardware alignment
- add a TODO item for DUMB_CREATE2
v3:
- document the UAPI semantics
- compute scanline pitch from for unknown color modes (Andy, Tomi)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Reviewed-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com>
Link: https://lore.kernel.org/r/20250821081918.79786-3-tzimmermann@suse.de
---
 Documentation/gpu/todo.rst         |  37 +++++++++++
 drivers/gpu/drm/drm_dumb_buffers.c | 130 +++++++++++++++++++++++++++++++++++++
 include/drm/drm_dumb_buffers.h     |  14 ++++
 include/uapi/drm/drm_mode.h        |  50 +++++++++++++-
 4 files changed, 230 insertions(+), 1 deletion(-)
 create mode 100644 include/drm/drm_dumb_buffers.h

(limited to 'include')

diff --git a/Documentation/gpu/todo.rst b/Documentation/gpu/todo.rst
index 92db80793bba..98ed38241dc6 100644
--- a/Documentation/gpu/todo.rst
+++ b/Documentation/gpu/todo.rst
@@ -648,6 +648,43 @@ Contact: Thomas Zimmermann <tzimmermann@suse.de>, Simona Vetter
 
 Level: Advanced
 
+Implement a new DUMB_CREATE2 ioctl
+----------------------------------
+
+The current DUMB_CREATE ioctl is not well defined. Instead of a pixel and
+framebuffer format, it only accepts a color mode of vague semantics. Assuming
+a linear framebuffer, the color mode gives an idea of the supported pixel
+format. But userspace effectively has to guess the correct values. It really
+only works reliably with framebuffers in XRGB8888. Userspace has begun to
+workaround these limitations by computing arbitrary format's buffer sizes and
+calculating their sizes in terms of XRGB8888 pixels.
+
+One possible solution is a new ioctl DUMB_CREATE2. It should accept a DRM
+format and a format modifier to resolve the color mode's ambiguity. As
+framebuffers can be multi-planar, the new ioctl has to return the buffer size,
+pitch and GEM handle for each individual color plane.
+
+In the first step, the new ioctl can be limited to the current features of
+the existing DUMB_CREATE. Individual drivers can then be extended to support
+multi-planar formats. Rockchip might require this and would be a good candidate.
+
+It might also be helpful to userspace to query information about the size of
+a potential buffer, if allocated. Userspace would supply geometry and format;
+the kernel would return minimal allocation sizes and scanline pitch. There is
+interest to allocate that memory from another device and provide it to the
+DRM driver (say via dma-buf).
+
+Another requested feature is the ability to allocate a buffer by size, without
+format. Accelators use this for their buffer allocation and it could likely be
+generalized.
+
+In addition to the kernel implementation, there must be user-space support
+for the new ioctl. There's code in Mesa that might be able to use the new
+call.
+
+Contact: Thomas Zimmermann <tzimmermann@suse.de>
+
+Level: Advanced
 
 Better Testing
 ==============
diff --git a/drivers/gpu/drm/drm_dumb_buffers.c b/drivers/gpu/drm/drm_dumb_buffers.c
index 9916aaf5b3f2..e9eed9a5b760 100644
--- a/drivers/gpu/drm/drm_dumb_buffers.c
+++ b/drivers/gpu/drm/drm_dumb_buffers.c
@@ -25,6 +25,8 @@
 
 #include <drm/drm_device.h>
 #include <drm/drm_drv.h>
+#include <drm/drm_dumb_buffers.h>
+#include <drm/drm_fourcc.h>
 #include <drm/drm_gem.h>
 #include <drm/drm_mode.h>
 
@@ -57,6 +59,134 @@
  * a hardware-specific ioctl to allocate suitable buffer objects.
  */
 
+static int drm_mode_align_dumb(struct drm_mode_create_dumb *args,
+			       unsigned long hw_pitch_align,
+			       unsigned long hw_size_align)
+{
+	u32 pitch = args->pitch;
+	u32 size;
+
+	if (!pitch)
+		return -EINVAL;
+
+	if (hw_pitch_align)
+		pitch = roundup(pitch, hw_pitch_align);
+
+	if (!hw_size_align)
+		hw_size_align = PAGE_SIZE;
+	else if (!IS_ALIGNED(hw_size_align, PAGE_SIZE))
+		return -EINVAL; /* TODO: handle this if necessary */
+
+	if (check_mul_overflow(args->height, pitch, &size))
+		return -EINVAL;
+	size = ALIGN(size, hw_size_align);
+	if (!size)
+		return -EINVAL;
+
+	args->pitch = pitch;
+	args->size = size;
+
+	return 0;
+}
+
+/**
+ * drm_mode_size_dumb - Calculates the scanline and buffer sizes for dumb buffers
+ * @dev: DRM device
+ * @args: Parameters for the dumb buffer
+ * @hw_pitch_align: Hardware scanline alignment in bytes
+ * @hw_size_align: Hardware buffer-size alignment in bytes
+ *
+ * The helper drm_mode_size_dumb() calculates the size of the buffer
+ * allocation and the scanline size for a dumb buffer. Callers have to
+ * set the buffers width, height and color mode in the argument @arg.
+ * The helper validates the correctness of the input and tests for
+ * possible overflows. If successful, it returns the dumb buffer's
+ * required scanline pitch and size in &args.
+ *
+ * The parameter @hw_pitch_align allows the driver to specifies an
+ * alignment for the scanline pitch, if the hardware requires any. The
+ * calculated pitch will be a multiple of the alignment. The parameter
+ * @hw_size_align allows to specify an alignment for buffer sizes. The
+ * provided alignment should represent requirements of the graphics
+ * hardware. drm_mode_size_dumb() handles GEM-related constraints
+ * automatically across all drivers and hardware. For example, the
+ * returned buffer size is always a multiple of PAGE_SIZE, which is
+ * required by mmap().
+ *
+ * Returns:
+ * Zero on success, or a negative error code otherwise.
+ */
+int drm_mode_size_dumb(struct drm_device *dev,
+		       struct drm_mode_create_dumb *args,
+		       unsigned long hw_pitch_align,
+		       unsigned long hw_size_align)
+{
+	u64 pitch = 0;
+	u32 fourcc;
+
+	/*
+	 * The scanline pitch depends on the buffer width and the color
+	 * format. The latter is specified as a color-mode constant for
+	 * which we first have to find the corresponding color format.
+	 *
+	 * Different color formats can have the same color-mode constant.
+	 * For example XRGB8888 and BGRX8888 both have a color mode of 32.
+	 * It is possible to use different formats for dumb-buffer allocation
+	 * and rendering as long as all involved formats share the same
+	 * color-mode constant.
+	 */
+	fourcc = drm_driver_color_mode_format(dev, args->bpp);
+	if (fourcc != DRM_FORMAT_INVALID) {
+		const struct drm_format_info *info = drm_format_info(fourcc);
+
+		if (!info)
+			return -EINVAL;
+		pitch = drm_format_info_min_pitch(info, 0, args->width);
+	} else if (args->bpp) {
+		/*
+		 * Some userspace throws in arbitrary values for bpp and
+		 * relies on the kernel to figure it out. In this case we
+		 * fall back to the old method of using bpp directly. The
+		 * over-commitment of memory from the rounding is acceptable
+		 * for compatibility with legacy userspace. We have a number
+		 * of deprecated legacy values that are explicitly supported.
+		 */
+		switch (args->bpp) {
+		default:
+			drm_warn_once(dev,
+				      "Unknown color mode %u; guessing buffer size.\n",
+				      args->bpp);
+			fallthrough;
+		/*
+		 * These constants represent various YUV formats supported by
+		 * drm_gem_afbc_get_bpp().
+		 */
+		case 12: // DRM_FORMAT_YUV420_8BIT
+		case 15: // DRM_FORMAT_YUV420_10BIT
+		case 30: // DRM_FORMAT_VUY101010
+			fallthrough;
+		/*
+		 * Used by Mesa and Gstreamer to allocate NV formats and others
+		 * as RGB buffers. Technically, XRGB16161616F formats are RGB,
+		 * but the dumb buffers are not supposed to be used for anything
+		 * beyond 32 bits per pixels.
+		 */
+		case 10: // DRM_FORMAT_NV{15,20,30}, DRM_FORMAT_P010
+		case 64: // DRM_FORMAT_{XRGB,XBGR,ARGB,ABGR}16161616F
+			pitch = args->width * DIV_ROUND_UP(args->bpp, SZ_8);
+			break;
+		}
+	}
+
+	if (!pitch || pitch > U32_MAX)
+		return -EINVAL;
+
+	args->pitch = pitch;
+
+	return drm_mode_align_dumb(args, hw_pitch_align, hw_size_align);
+}
+EXPORT_SYMBOL(drm_mode_size_dumb);
+
 int drm_mode_create_dumb(struct drm_device *dev,
 			 struct drm_mode_create_dumb *args,
 			 struct drm_file *file_priv)
diff --git a/include/drm/drm_dumb_buffers.h b/include/drm/drm_dumb_buffers.h
new file mode 100644
index 000000000000..1f3a8236fb3d
--- /dev/null
+++ b/include/drm/drm_dumb_buffers.h
@@ -0,0 +1,14 @@
+/* SPDX-License-Identifier: MIT */
+
+#ifndef __DRM_DUMB_BUFFERS_H__
+#define __DRM_DUMB_BUFFERS_H__
+
+struct drm_device;
+struct drm_mode_create_dumb;
+
+int drm_mode_size_dumb(struct drm_device *dev,
+		       struct drm_mode_create_dumb *args,
+		       unsigned long hw_pitch_align,
+		       unsigned long hw_size_align);
+
+#endif
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index a122bea25593..1e0e02a79b5c 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -1066,7 +1066,7 @@ struct drm_mode_crtc_page_flip_target {
  * struct drm_mode_create_dumb - Create a KMS dumb buffer for scanout.
  * @height: buffer height in pixels
  * @width: buffer width in pixels
- * @bpp: bits per pixel
+ * @bpp: color mode
  * @flags: must be zero
  * @handle: buffer object handle
  * @pitch: number of bytes between two consecutive lines
@@ -1074,6 +1074,54 @@ struct drm_mode_crtc_page_flip_target {
  *
  * User-space fills @height, @width, @bpp and @flags. If the IOCTL succeeds,
  * the kernel fills @handle, @pitch and @size.
+ *
+ * The value of @bpp is a color-mode number describing a specific format
+ * or a variant thereof. The value often corresponds to the number of bits
+ * per pixel for most modes, although there are exceptions. Each color mode
+ * maps to a DRM format plus a number of modes with similar pixel layout.
+ * Framebuffer layout is always linear.
+ *
+ * Support for all modes and formats is optional. Even if dumb-buffer
+ * creation with a certain color mode succeeds, it is not guaranteed that
+ * the DRM driver supports any of the related formats. Most drivers support
+ * a color mode of 32 with a format of DRM_FORMAT_XRGB8888 on their primary
+ * plane.
+ *
+ * +------------+------------------------+------------------------+
+ * | Color mode | Framebuffer format     | Compatible formats     |
+ * +============+========================+========================+
+ * |     32     |  * DRM_FORMAT_XRGB8888 |  * DRM_FORMAT_BGRX8888 |
+ * |            |                        |  * DRM_FORMAT_RGBX8888 |
+ * |            |                        |  * DRM_FORMAT_XBGR8888 |
+ * +------------+------------------------+------------------------+
+ * |     24     |  * DRM_FORMAT_RGB888   |  * DRM_FORMAT_BGR888   |
+ * +------------+------------------------+------------------------+
+ * |     16     |  * DRM_FORMAT_RGB565   |  * DRM_FORMAT_BGR565   |
+ * +------------+------------------------+------------------------+
+ * |     15     |  * DRM_FORMAT_XRGB1555 |  * DRM_FORMAT_BGRX1555 |
+ * |            |                        |  * DRM_FORMAT_RGBX1555 |
+ * |            |                        |  * DRM_FORMAT_XBGR1555 |
+ * +------------+------------------------+------------------------+
+ * |      8     |  * DRM_FORMAT_C8       |  * DRM_FORMAT_D8       |
+ * |            |                        |  * DRM_FORMAT_R8       |
+ * +------------+------------------------+------------------------+
+ * |      4     |  * DRM_FORMAT_C4       |  * DRM_FORMAT_D4       |
+ * |            |                        |  * DRM_FORMAT_R4       |
+ * +------------+------------------------+------------------------+
+ * |      2     |  * DRM_FORMAT_C2       |  * DRM_FORMAT_D2       |
+ * |            |                        |  * DRM_FORMAT_R2       |
+ * +------------+------------------------+------------------------+
+ * |      1     |  * DRM_FORMAT_C1       |  * DRM_FORMAT_D1       |
+ * |            |                        |  * DRM_FORMAT_R1       |
+ * +------------+------------------------+------------------------+
+ *
+ * Color modes of 10, 12, 15, 30 and 64 are only supported for use by
+ * legacy user space. Please don't use them in new code. Other modes
+ * are not support.
+ *
+ * Do not attempt to allocate anything but linear framebuffer memory
+ * with single-plane RGB data. Allocation of other framebuffer
+ * layouts requires dedicated ioctls in the respective DRM driver.
  */
 struct drm_mode_create_dumb {
 	__u32 height;
-- 
cgit v1.2.3


From 0359a849f4e7d6f41249e0741ab8cf7babdfaa41 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 2 Sep 2025 10:32:39 +0200
Subject: drm/crtc: Drop no_vblank bit field

The no_vblank field in drm_crtc_state is defined as a bit-field with a
single bit.

This will create a syntax issue with the macros we'll introduce next,
and most other booleans but the *_changed ones in drm_crtc_state do not
use a bit field anyway.

Let's drop it.

Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://lore.kernel.org/r/20250902-drm-state-readout-v1-11-14ad5315da3f@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20250902-drm-state-readout-v1-11-14ad5315da3f@kernel.org
---
 include/drm/drm_crtc.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index caa56e039da2..2d2a0bd526cf 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -186,7 +186,7 @@ struct drm_crtc_state {
 	 * this case the driver will send the VBLANK event on its own when the
 	 * writeback job is complete.
 	 */
-	bool no_vblank : 1;
+	bool no_vblank;
 
 	/**
 	 * @plane_mask: Bitmask of drm_plane_mask(plane) of planes attached to
-- 
cgit v1.2.3


From 74afeb8128502a529041a2566febd26053a7be11 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Tue, 16 Sep 2025 10:36:19 +0200
Subject: drm/vblank: Add vblank timer

The vblank timer simulates a vblank interrupt for hardware without
support. Rate-limits the display update frequency.

DRM drivers for hardware without vblank support apply display updates
ASAP. A vblank event informs DRM clients of the completed update.
Userspace compositors immediately schedule the next update, which
creates significant load on virtualization outputs. Display updates
are usually fast on virtualization outputs, as their framebuffers are
in regular system memory and there's no hardware vblank interrupt to
throttle the update rate.

The vblank timer is a HR timer that signals the vblank in software.
It limits the update frequency of a DRM driver similar to a hardware
vblank interrupt. The timer is not synchronized to the actual vblank
interval of the display.

The code has been adopted from vkms, which added the funtionality
in commit 3a0709928b17 ("drm/vkms: Add vblank events simulated by
hrtimers").

The new implementation is part of the existing vblank support,
which sets up the timer automatically. Drivers only have to start
and cancel the vblank timer as part of enabling and disabling the
CRTC. The new vblank helper library provides callbacks for struct
drm_crtc_funcs.

The standard way for handling vblank is to call drm_crtc_handle_vblank().
Drivers that require additional processing, such as vkms, can init
handle_vblank_timeout in struct drm_crtc_helper_funcs to refer to
their timeout handler.

There's a possible deadlock between drm_crtc_handle_vblank() and
hrtimer_cancel(). [1] The implementation avoids to call hrtimer_cancel()
directly and instead signals to the timer function to not restart
itself.

v4:
- fix possible race condition between timeout and atomic commit (Michael)
v3:
- avoid deadlock when cancelling timer (Ville, Lyude)
v2:
- implement vblank timer entirely in vblank helpers
- downgrade overrun warning to debug
- fix docs

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Tested-by: Louis Chauvet <louis.chauvet@bootlin.com>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>
Link: https://lore.kernel.org/all/20250510094757.4174662-1-zengheng4@huawei.com/ # [1]
Link: https://lore.kernel.org/r/20250916083816.30275-2-tzimmermann@suse.de
---
 Documentation/gpu/drm-kms-helpers.rst    |  12 +++
 drivers/gpu/drm/Makefile                 |   3 +-
 drivers/gpu/drm/drm_vblank.c             | 172 ++++++++++++++++++++++++++++++-
 drivers/gpu/drm/drm_vblank_helper.c      |  96 +++++++++++++++++
 include/drm/drm_modeset_helper_vtables.h |  12 +++
 include/drm/drm_vblank.h                 |  32 ++++++
 include/drm/drm_vblank_helper.h          |  33 ++++++
 7 files changed, 357 insertions(+), 3 deletions(-)
 create mode 100644 drivers/gpu/drm/drm_vblank_helper.c
 create mode 100644 include/drm/drm_vblank_helper.h

(limited to 'include')

diff --git a/Documentation/gpu/drm-kms-helpers.rst b/Documentation/gpu/drm-kms-helpers.rst
index 5139705089f2..781129f78b06 100644
--- a/Documentation/gpu/drm-kms-helpers.rst
+++ b/Documentation/gpu/drm-kms-helpers.rst
@@ -92,6 +92,18 @@ GEM Atomic Helper Reference
 .. kernel-doc:: drivers/gpu/drm/drm_gem_atomic_helper.c
    :export:
 
+VBLANK Helper Reference
+-----------------------
+
+.. kernel-doc:: drivers/gpu/drm/drm_vblank_helper.c
+   :doc: overview
+
+.. kernel-doc:: include/drm/drm_vblank_helper.h
+   :internal:
+
+.. kernel-doc:: drivers/gpu/drm/drm_vblank_helper.c
+   :export:
+
 Simple KMS Helper Reference
 ===========================
 
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 4dafbdc8f86a..5ba4ffdb8055 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -150,7 +150,8 @@ drm_kms_helper-y := \
 	drm_plane_helper.o \
 	drm_probe_helper.o \
 	drm_self_refresh_helper.o \
-	drm_simple_kms_helper.o
+	drm_simple_kms_helper.o \
+	drm_vblank_helper.o
 drm_kms_helper-$(CONFIG_DRM_PANEL_BRIDGE) += bridge/panel.o
 drm_kms_helper-$(CONFIG_DRM_FBDEV_EMULATION) += drm_fb_helper.o
 obj-$(CONFIG_DRM_KMS_HELPER) += drm_kms_helper.o
diff --git a/drivers/gpu/drm/drm_vblank.c b/drivers/gpu/drm/drm_vblank.c
index 46f59883183d..61e211fd3c9c 100644
--- a/drivers/gpu/drm/drm_vblank.c
+++ b/drivers/gpu/drm/drm_vblank.c
@@ -136,8 +136,17 @@
  * vblanks after a timer has expired, which can be configured through the
  * ``vblankoffdelay`` module parameter.
  *
- * Drivers for hardware without support for vertical-blanking interrupts
- * must not call drm_vblank_init(). For such drivers, atomic helpers will
+ * Drivers for hardware without support for vertical-blanking interrupts can
+ * use DRM vblank timers to send vblank events at the rate of the current
+ * display mode's refresh. While not synchronized to the hardware's
+ * vertical-blanking regions, the timer helps DRM clients and compositors to
+ * adapt their update cycle to the display output. Drivers should set up
+ * vblanking as usual, but call drm_crtc_vblank_start_timer() and
+ * drm_crtc_vblank_cancel_timer() as part of their atomic mode setting.
+ * See also DRM vblank helpers for more information.
+ *
+ * Drivers without support for vertical-blanking interrupts nor timers must
+ * not call drm_vblank_init(). For these drivers, atomic helpers will
  * automatically generate fake vblank events as part of the display update.
  * This functionality also can be controlled by the driver by enabling and
  * disabling struct drm_crtc_state.no_vblank.
@@ -508,6 +517,9 @@ static void drm_vblank_init_release(struct drm_device *dev, void *ptr)
 	drm_WARN_ON(dev, READ_ONCE(vblank->enabled) &&
 		    drm_core_check_feature(dev, DRIVER_MODESET));
 
+	if (vblank->vblank_timer.crtc)
+		hrtimer_cancel(&vblank->vblank_timer.timer);
+
 	drm_vblank_destroy_worker(vblank);
 	timer_delete_sync(&vblank->disable_timer);
 }
@@ -2162,3 +2174,159 @@ err_free:
 	return ret;
 }
 
+/*
+ * VBLANK timer
+ */
+
+static enum hrtimer_restart drm_vblank_timer_function(struct hrtimer *timer)
+{
+	struct drm_vblank_crtc_timer *vtimer =
+		container_of(timer, struct drm_vblank_crtc_timer, timer);
+	struct drm_crtc *crtc = vtimer->crtc;
+	const struct drm_crtc_helper_funcs *crtc_funcs = crtc->helper_private;
+	struct drm_device *dev = crtc->dev;
+	unsigned long flags;
+	ktime_t interval;
+	u64 ret_overrun;
+	bool succ;
+
+	spin_lock_irqsave(&vtimer->interval_lock, flags);
+	interval = vtimer->interval;
+	spin_unlock_irqrestore(&vtimer->interval_lock, flags);
+
+	if (!interval)
+		return HRTIMER_NORESTART;
+
+	ret_overrun = hrtimer_forward_now(&vtimer->timer, interval);
+	if (ret_overrun != 1)
+		drm_dbg_vbl(dev, "vblank timer overrun\n");
+
+	if (crtc_funcs->handle_vblank_timeout)
+		succ = crtc_funcs->handle_vblank_timeout(crtc);
+	else
+		succ = drm_crtc_handle_vblank(crtc);
+	if (!succ)
+		return HRTIMER_NORESTART;
+
+	return HRTIMER_RESTART;
+}
+
+/**
+ * drm_crtc_vblank_start_timer - Starts the vblank timer on the given CRTC
+ * @crtc: the CRTC
+ *
+ * Drivers should call this function from their CRTC's enable_vblank
+ * function to start a vblank timer. The timer will fire after the duration
+ * of a full frame. drm_crtc_vblank_cancel_timer() disables a running timer.
+ *
+ * Returns:
+ * 0 on success, or a negative errno code otherwise.
+ */
+int drm_crtc_vblank_start_timer(struct drm_crtc *crtc)
+{
+	struct drm_vblank_crtc *vblank = drm_crtc_vblank_crtc(crtc);
+	struct drm_vblank_crtc_timer *vtimer = &vblank->vblank_timer;
+	unsigned long flags;
+
+	if (!vtimer->crtc) {
+		/*
+		 * Set up the data structures on the first invocation.
+		 */
+		vtimer->crtc = crtc;
+		spin_lock_init(&vtimer->interval_lock);
+		hrtimer_setup(&vtimer->timer, drm_vblank_timer_function,
+			      CLOCK_MONOTONIC, HRTIMER_MODE_REL);
+	} else {
+		/*
+		 * Timer should not be active. If it is, wait for the
+		 * previous cancel operations to finish.
+		 */
+		while (hrtimer_active(&vtimer->timer))
+			hrtimer_try_to_cancel(&vtimer->timer);
+	}
+
+	drm_calc_timestamping_constants(crtc, &crtc->mode);
+
+	spin_lock_irqsave(&vtimer->interval_lock, flags);
+	vtimer->interval = ns_to_ktime(vblank->framedur_ns);
+	spin_unlock_irqrestore(&vtimer->interval_lock, flags);
+
+	hrtimer_start(&vtimer->timer, vtimer->interval, HRTIMER_MODE_REL);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_crtc_vblank_start_timer);
+
+/**
+ * drm_crtc_vblank_start_timer - Cancels the given CRTC's vblank timer
+ * @crtc: the CRTC
+ *
+ * Drivers should call this function from their CRTC's disable_vblank
+ * function to stop a vblank timer.
+ */
+void drm_crtc_vblank_cancel_timer(struct drm_crtc *crtc)
+{
+	struct drm_vblank_crtc *vblank = drm_crtc_vblank_crtc(crtc);
+	struct drm_vblank_crtc_timer *vtimer = &vblank->vblank_timer;
+	unsigned long flags;
+
+	/*
+	 * Calling hrtimer_cancel() can result in a deadlock with DRM's
+	 * vblank_time_lime_lock and hrtimers' softirq_expiry_lock. So
+	 * clear interval and indicate cancellation. The timer function
+	 * will cancel itself on the next invocation.
+	 */
+
+	spin_lock_irqsave(&vtimer->interval_lock, flags);
+	vtimer->interval = 0;
+	spin_unlock_irqrestore(&vtimer->interval_lock, flags);
+
+	hrtimer_try_to_cancel(&vtimer->timer);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_cancel_timer);
+
+/**
+ * drm_crtc_vblank_get_vblank_timeout - Returns the vblank timeout
+ * @crtc: The CRTC
+ * @vblank_time: Returns the next vblank timestamp
+ *
+ * The helper drm_crtc_vblank_get_vblank_timeout() returns the next vblank
+ * timestamp of the CRTC's vblank timer according to the timer's expiry
+ * time.
+ */
+void drm_crtc_vblank_get_vblank_timeout(struct drm_crtc *crtc, ktime_t *vblank_time)
+{
+	struct drm_vblank_crtc *vblank = drm_crtc_vblank_crtc(crtc);
+	struct drm_vblank_crtc_timer *vtimer = &vblank->vblank_timer;
+	u64 cur_count;
+	ktime_t cur_time;
+
+	if (!READ_ONCE(vblank->enabled)) {
+		*vblank_time = ktime_get();
+		return;
+	}
+
+	/*
+	 * A concurrent vblank timeout could update the expires field before
+	 * we compare it with the vblank time. Hence we'd compare the old
+	 * expiry time to the new vblank time; deducing the timer had already
+	 * expired. Reread until we get consistent values from both fields.
+	 */
+	do {
+		cur_count = drm_crtc_vblank_count_and_time(crtc, &cur_time);
+		*vblank_time = READ_ONCE(vtimer->timer.node.expires);
+	} while (cur_count != drm_crtc_vblank_count_and_time(crtc, &cur_time));
+
+	if (drm_WARN_ON(crtc->dev, !ktime_compare(*vblank_time, cur_time)))
+		return; /* Already expired */
+
+	/*
+	 * To prevent races we roll the hrtimer forward before we do any
+	 * interrupt processing - this is how real hw works (the interrupt
+	 * is only generated after all the vblank registers are updated)
+	 * and what the vblank core expects. Therefore we need to always
+	 * correct the timestamp by one frame.
+	 */
+	*vblank_time = ktime_sub(*vblank_time, vtimer->interval);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_get_vblank_timeout);
diff --git a/drivers/gpu/drm/drm_vblank_helper.c b/drivers/gpu/drm/drm_vblank_helper.c
new file mode 100644
index 000000000000..f94d1e706191
--- /dev/null
+++ b/drivers/gpu/drm/drm_vblank_helper.c
@@ -0,0 +1,96 @@
+// SPDX-License-Identifier: MIT
+
+#include <drm/drm_crtc.h>
+#include <drm/drm_managed.h>
+#include <drm/drm_modeset_helper_vtables.h>
+#include <drm/drm_print.h>
+#include <drm/drm_vblank.h>
+#include <drm/drm_vblank_helper.h>
+
+/**
+ * DOC: overview
+ *
+ * The vblank helper library provides functions for supporting vertical
+ * blanking in DRM drivers.
+ *
+ * For vblank timers, several callback implementations are available.
+ * Drivers enable support for vblank timers by setting the vblank callbacks
+ * in struct &drm_crtc_funcs to the helpers provided by this library. The
+ * initializer macro DRM_CRTC_VBLANK_TIMER_FUNCS does this conveniently.
+ *
+ * Once the driver enables vblank support with drm_vblank_init(), each
+ * CRTC's vblank timer fires according to the programmed display mode. By
+ * default, the vblank timer invokes drm_crtc_handle_vblank(). Drivers with
+ * more specific requirements can set their own handler function in
+ * struct &drm_crtc_helper_funcs.handle_vblank_timeout.
+ */
+
+/*
+ * VBLANK timer
+ */
+
+/**
+ * drm_crtc_vblank_helper_enable_vblank_timer - Implements struct &drm_crtc_funcs.enable_vblank
+ * @crtc: The CRTC
+ *
+ * The helper drm_crtc_vblank_helper_enable_vblank_timer() implements
+ * enable_vblank of struct drm_crtc_helper_funcs for CRTCs that require
+ * a VBLANK timer. It sets up the timer on the first invocation. The
+ * started timer expires after the current frame duration. See struct
+ * &drm_vblank_crtc.framedur_ns.
+ *
+ * See also struct &drm_crtc_helper_funcs.enable_vblank.
+ *
+ * Returns:
+ * 0 on success, or a negative errno code otherwise.
+ */
+int drm_crtc_vblank_helper_enable_vblank_timer(struct drm_crtc *crtc)
+{
+	return drm_crtc_vblank_start_timer(crtc);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_helper_enable_vblank_timer);
+
+/**
+ * drm_crtc_vblank_helper_disable_vblank_timer - Implements struct &drm_crtc_funcs.disable_vblank
+ * @crtc: The CRTC
+ *
+ * The helper drm_crtc_vblank_helper_disable_vblank_timer() implements
+ * disable_vblank of struct drm_crtc_funcs for CRTCs that require a
+ * VBLANK timer.
+ *
+ * See also struct &drm_crtc_helper_funcs.disable_vblank.
+ */
+void drm_crtc_vblank_helper_disable_vblank_timer(struct drm_crtc *crtc)
+{
+	drm_crtc_vblank_cancel_timer(crtc);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_helper_disable_vblank_timer);
+
+/**
+ * drm_crtc_vblank_helper_get_vblank_timestamp_from_timer -
+ *	Implements struct &drm_crtc_funcs.get_vblank_timestamp
+ * @crtc: The CRTC
+ * @max_error: Maximum acceptable error
+ * @vblank_time: Returns the next vblank timestamp
+ * @in_vblank_irq: True is called from drm_crtc_handle_vblank()
+ *
+ * The helper drm_crtc_helper_get_vblank_timestamp_from_timer() implements
+ * get_vblank_timestamp of struct drm_crtc_funcs for CRTCs that require a
+ * VBLANK timer. It returns the timestamp according to the timer's expiry
+ * time.
+ *
+ * See also struct &drm_crtc_funcs.get_vblank_timestamp.
+ *
+ * Returns:
+ * True on success, or false otherwise.
+ */
+bool drm_crtc_vblank_helper_get_vblank_timestamp_from_timer(struct drm_crtc *crtc,
+							    int *max_error,
+							    ktime_t *vblank_time,
+							    bool in_vblank_irq)
+{
+	drm_crtc_vblank_get_vblank_timeout(crtc, vblank_time);
+
+	return true;
+}
+EXPORT_SYMBOL(drm_crtc_vblank_helper_get_vblank_timestamp_from_timer);
diff --git a/include/drm/drm_modeset_helper_vtables.h b/include/drm/drm_modeset_helper_vtables.h
index ce7c7aeac887..fe32854b7ffe 100644
--- a/include/drm/drm_modeset_helper_vtables.h
+++ b/include/drm/drm_modeset_helper_vtables.h
@@ -490,6 +490,18 @@ struct drm_crtc_helper_funcs {
 				     bool in_vblank_irq, int *vpos, int *hpos,
 				     ktime_t *stime, ktime_t *etime,
 				     const struct drm_display_mode *mode);
+
+	/**
+	 * @handle_vblank_timeout: Handles timeouts of the vblank timer.
+	 *
+	 * Called by CRTC's the vblank timer on each timeout. Semantics is
+	 * equivalient to drm_crtc_handle_vblank(). Implementations should
+	 * invoke drm_crtc_handle_vblank() as part of processing the timeout.
+	 *
+	 * This callback is optional. If unset, the vblank timer invokes
+	 * drm_crtc_handle_vblank() directly.
+	 */
+	bool (*handle_vblank_timeout)(struct drm_crtc *crtc);
 };
 
 /**
diff --git a/include/drm/drm_vblank.h b/include/drm/drm_vblank.h
index 151ab1e85b1b..ffa564d79638 100644
--- a/include/drm/drm_vblank.h
+++ b/include/drm/drm_vblank.h
@@ -25,6 +25,7 @@
 #define _DRM_VBLANK_H_
 
 #include <linux/seqlock.h>
+#include <linux/hrtimer.h>
 #include <linux/idr.h>
 #include <linux/poll.h>
 #include <linux/kthread.h>
@@ -103,6 +104,28 @@ struct drm_vblank_crtc_config {
 	bool disable_immediate;
 };
 
+/**
+ * struct drm_vblank_crtc_timer - vblank timer for a CRTC
+ */
+struct drm_vblank_crtc_timer {
+	/**
+	 * @timer: The vblank's high-resolution timer
+	 */
+	struct hrtimer timer;
+	/**
+	 * @interval_lock: Protects @interval
+	 */
+	spinlock_t interval_lock;
+	/**
+	 * @interval: Duration between two vblanks
+	 */
+	ktime_t interval;
+	/**
+	 * @crtc: The timer's CRTC
+	 */
+	struct drm_crtc *crtc;
+};
+
 /**
  * struct drm_vblank_crtc - vblank tracking for a CRTC
  *
@@ -254,6 +277,11 @@ struct drm_vblank_crtc {
 	 * cancelled.
 	 */
 	wait_queue_head_t work_wait_queue;
+
+	/**
+	 * @vblank_timer: Holds the state of the vblank timer
+	 */
+	struct drm_vblank_crtc_timer vblank_timer;
 };
 
 struct drm_vblank_crtc *drm_crtc_vblank_crtc(struct drm_crtc *crtc);
@@ -290,6 +318,10 @@ wait_queue_head_t *drm_crtc_vblank_waitqueue(struct drm_crtc *crtc);
 void drm_crtc_set_max_vblank_count(struct drm_crtc *crtc,
 				   u32 max_vblank_count);
 
+int drm_crtc_vblank_start_timer(struct drm_crtc *crtc);
+void drm_crtc_vblank_cancel_timer(struct drm_crtc *crtc);
+void drm_crtc_vblank_get_vblank_timeout(struct drm_crtc *crtc, ktime_t *vblank_time);
+
 /*
  * Helpers for struct drm_crtc_funcs
  */
diff --git a/include/drm/drm_vblank_helper.h b/include/drm/drm_vblank_helper.h
new file mode 100644
index 000000000000..74a971d0cfba
--- /dev/null
+++ b/include/drm/drm_vblank_helper.h
@@ -0,0 +1,33 @@
+/* SPDX-License-Identifier: GPL-2.0+ */
+
+#ifndef _DRM_VBLANK_HELPER_H_
+#define _DRM_VBLANK_HELPER_H_
+
+#include <linux/hrtimer_types.h>
+#include <linux/types.h>
+
+struct drm_crtc;
+
+/*
+ * VBLANK timer
+ */
+
+int drm_crtc_vblank_helper_enable_vblank_timer(struct drm_crtc *crtc);
+void drm_crtc_vblank_helper_disable_vblank_timer(struct drm_crtc *crtc);
+bool drm_crtc_vblank_helper_get_vblank_timestamp_from_timer(struct drm_crtc *crtc,
+							    int *max_error,
+							    ktime_t *vblank_time,
+							    bool in_vblank_irq);
+
+/**
+ * DRM_CRTC_VBLANK_TIMER_FUNCS - Default implementation for VBLANK timers
+ *
+ * This macro initializes struct &drm_crtc_funcs to default helpers for
+ * VBLANK timers.
+ */
+#define DRM_CRTC_VBLANK_TIMER_FUNCS \
+	.enable_vblank = drm_crtc_vblank_helper_enable_vblank_timer, \
+	.disable_vblank = drm_crtc_vblank_helper_disable_vblank_timer, \
+	.get_vblank_timestamp = drm_crtc_vblank_helper_get_vblank_timestamp_from_timer
+
+#endif
-- 
cgit v1.2.3


From d54dbb5963bdbdf8559903fe2b2343e871adcb30 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Tue, 16 Sep 2025 10:36:20 +0200
Subject: drm/vblank: Add CRTC helpers for simple use cases

Implement atomic_flush, atomic_enable and atomic_disable of struct
drm_crtc_helper_funcs for vblank handling. Driver with no further
requirements can use these functions instead of adding their own.
Also simplifies the use of vblank timers.

The code has been adopted from vkms, which added the funtionality
in commit 3a0709928b17 ("drm/vkms: Add vblank events simulated by
hrtimers").

v3:
- mention vkms (Javier)
v2:
- fix docs

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Tested-by: Michael Kelley <mhklinux@outlook.com>
Link: https://lore.kernel.org/r/20250916083816.30275-3-tzimmermann@suse.de
---
 drivers/gpu/drm/drm_vblank_helper.c | 80 +++++++++++++++++++++++++++++++++++++
 include/drm/drm_vblank_helper.h     | 23 +++++++++++
 2 files changed, 103 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_vblank_helper.c b/drivers/gpu/drm/drm_vblank_helper.c
index f94d1e706191..a04a6ba1b0ca 100644
--- a/drivers/gpu/drm/drm_vblank_helper.c
+++ b/drivers/gpu/drm/drm_vblank_helper.c
@@ -1,5 +1,6 @@
 // SPDX-License-Identifier: MIT
 
+#include <drm/drm_atomic.h>
 #include <drm/drm_crtc.h>
 #include <drm/drm_managed.h>
 #include <drm/drm_modeset_helper_vtables.h>
@@ -17,6 +18,12 @@
  * Drivers enable support for vblank timers by setting the vblank callbacks
  * in struct &drm_crtc_funcs to the helpers provided by this library. The
  * initializer macro DRM_CRTC_VBLANK_TIMER_FUNCS does this conveniently.
+ * The driver further has to send the VBLANK event from its atomic_flush
+ * callback and control vblank from the CRTC's atomic_enable and atomic_disable
+ * callbacks. The callbacks are located in struct &drm_crtc_helper_funcs.
+ * The vblank helper library provides implementations of these callbacks
+ * for drivers without further requirements. The initializer macro
+ * DRM_CRTC_HELPER_VBLANK_FUNCS sets them coveniently.
  *
  * Once the driver enables vblank support with drm_vblank_init(), each
  * CRTC's vblank timer fires according to the programmed display mode. By
@@ -25,6 +32,79 @@
  * struct &drm_crtc_helper_funcs.handle_vblank_timeout.
  */
 
+/*
+ * VBLANK helpers
+ */
+
+/**
+ * drm_crtc_vblank_atomic_flush -
+ *	Implements struct &drm_crtc_helper_funcs.atomic_flush
+ * @crtc: The CRTC
+ * @state: The atomic state to apply
+ *
+ * The helper drm_crtc_vblank_atomic_flush() implements atomic_flush of
+ * struct drm_crtc_helper_funcs for CRTCs that only need to send out a
+ * VBLANK event.
+ *
+ * See also struct &drm_crtc_helper_funcs.atomic_flush.
+ */
+void drm_crtc_vblank_atomic_flush(struct drm_crtc *crtc,
+				  struct drm_atomic_state *state)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_crtc_state *crtc_state = drm_atomic_get_new_crtc_state(state, crtc);
+	struct drm_pending_vblank_event *event;
+
+	spin_lock_irq(&dev->event_lock);
+
+	event = crtc_state->event;
+	crtc_state->event = NULL;
+
+	if (event) {
+		if (drm_crtc_vblank_get(crtc) == 0)
+			drm_crtc_arm_vblank_event(crtc, event);
+		else
+			drm_crtc_send_vblank_event(crtc, event);
+	}
+
+	spin_unlock_irq(&dev->event_lock);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_atomic_flush);
+
+/**
+ * drm_crtc_vblank_atomic_enable - Implements struct &drm_crtc_helper_funcs.atomic_enable
+ * @crtc: The CRTC
+ * @state: The atomic state
+ *
+ * The helper drm_crtc_vblank_atomic_enable() implements atomic_enable
+ * of struct drm_crtc_helper_funcs for CRTCs the only need to enable VBLANKs.
+ *
+ * See also struct &drm_crtc_helper_funcs.atomic_enable.
+ */
+void drm_crtc_vblank_atomic_enable(struct drm_crtc *crtc,
+				   struct drm_atomic_state *state)
+{
+	drm_crtc_vblank_on(crtc);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_atomic_enable);
+
+/**
+ * drm_crtc_vblank_atomic_disable - Implements struct &drm_crtc_helper_funcs.atomic_disable
+ * @crtc: The CRTC
+ * @state: The atomic state
+ *
+ * The helper drm_crtc_vblank_atomic_disable() implements atomic_disable
+ * of struct drm_crtc_helper_funcs for CRTCs the only need to disable VBLANKs.
+ *
+ * See also struct &drm_crtc_funcs.atomic_disable.
+ */
+void drm_crtc_vblank_atomic_disable(struct drm_crtc *crtc,
+				    struct drm_atomic_state *state)
+{
+	drm_crtc_vblank_off(crtc);
+}
+EXPORT_SYMBOL(drm_crtc_vblank_atomic_disable);
+
 /*
  * VBLANK timer
  */
diff --git a/include/drm/drm_vblank_helper.h b/include/drm/drm_vblank_helper.h
index 74a971d0cfba..fcd8a9b35846 100644
--- a/include/drm/drm_vblank_helper.h
+++ b/include/drm/drm_vblank_helper.h
@@ -6,8 +6,31 @@
 #include <linux/hrtimer_types.h>
 #include <linux/types.h>
 
+struct drm_atomic_state;
 struct drm_crtc;
 
+/*
+ * VBLANK helpers
+ */
+
+void drm_crtc_vblank_atomic_flush(struct drm_crtc *crtc,
+				  struct drm_atomic_state *state);
+void drm_crtc_vblank_atomic_enable(struct drm_crtc *crtc,
+				   struct drm_atomic_state *state);
+void drm_crtc_vblank_atomic_disable(struct drm_crtc *crtc,
+				    struct drm_atomic_state *crtc_state);
+
+/**
+ * DRM_CRTC_HELPER_VBLANK_FUNCS - Default implementation for VBLANK helpers
+ *
+ * This macro initializes struct &drm_crtc_helper_funcs to default helpers
+ * for VBLANK handling.
+ */
+#define DRM_CRTC_HELPER_VBLANK_FUNCS \
+	.atomic_flush = drm_crtc_vblank_atomic_flush, \
+	.atomic_enable = drm_crtc_vblank_atomic_enable, \
+	.atomic_disable = drm_crtc_vblank_atomic_disable
+
 /*
  * VBLANK timer
  */
-- 
cgit v1.2.3


From ad298d9ec957414dbf3d51f3c8bca4b6d2416c0c Mon Sep 17 00:00:00 2001
From: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Date: Tue, 30 Sep 2025 14:27:52 +0200
Subject: drm/gpusvm, drm/xe: Fix userptr to not allow device private pages
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When userptr is used on SVM-enabled VMs, a non-NULL
hmm_range::dev_private_owner value might mean that
hmm_range_fault() attempts to return device private pages.
Either that will fail, or the userptr code will not know
how to handle those.

Use NULL for hmm_range::dev_private_owner to migrate
such pages to system. In order to do that, move the
struct drm_gpusvm::device_private_page_owner field to
struct drm_gpusvm_ctx::device_private_page_owner so that
it doesn't remain immutable over the drm_gpusvm lifetime.

v2:
- Don't conditionally compile xe_svm_devm_owner().
- Kerneldoc xe_svm_devm_owner().

Fixes: 9e9787414882 ("drm/xe/userptr: replace xe_hmm with gpusvm")
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Link: https://lore.kernel.org/r/20250930122752.96034-1-thomas.hellstrom@linux.intel.com
---
 drivers/gpu/drm/drm_gpusvm.c    | 24 +++++++++++++-----------
 drivers/gpu/drm/xe/xe_svm.c     | 11 +++--------
 drivers/gpu/drm/xe/xe_svm.h     | 14 ++++++++++++++
 drivers/gpu/drm/xe/xe_userptr.c |  1 +
 drivers/gpu/drm/xe/xe_vm.c      |  1 +
 include/drm/drm_gpusvm.h        |  7 ++++---
 6 files changed, 36 insertions(+), 22 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
index eeeeb99cfdf6..cb906765897e 100644
--- a/drivers/gpu/drm/drm_gpusvm.c
+++ b/drivers/gpu/drm/drm_gpusvm.c
@@ -361,7 +361,6 @@ static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
  * @name: Name of the GPU SVM.
  * @drm: Pointer to the DRM device structure.
  * @mm: Pointer to the mm_struct for the address space.
- * @device_private_page_owner: Device private pages owner.
  * @mm_start: Start address of GPU SVM.
  * @mm_range: Range of the GPU SVM.
  * @notifier_size: Size of individual notifiers.
@@ -383,7 +382,7 @@ static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
  */
 int drm_gpusvm_init(struct drm_gpusvm *gpusvm,
 		    const char *name, struct drm_device *drm,
-		    struct mm_struct *mm, void *device_private_page_owner,
+		    struct mm_struct *mm,
 		    unsigned long mm_start, unsigned long mm_range,
 		    unsigned long notifier_size,
 		    const struct drm_gpusvm_ops *ops,
@@ -395,15 +394,13 @@ int drm_gpusvm_init(struct drm_gpusvm *gpusvm,
 		mmgrab(mm);
 	} else {
 		/* No full SVM mode, only core drm_gpusvm_pages API. */
-		if (ops || num_chunks || mm_range || notifier_size ||
-		    device_private_page_owner)
+		if (ops || num_chunks || mm_range || notifier_size)
 			return -EINVAL;
 	}
 
 	gpusvm->name = name;
 	gpusvm->drm = drm;
 	gpusvm->mm = mm;
-	gpusvm->device_private_page_owner = device_private_page_owner;
 	gpusvm->mm_start = mm_start;
 	gpusvm->mm_range = mm_range;
 	gpusvm->notifier_size = notifier_size;
@@ -684,6 +681,7 @@ static unsigned int drm_gpusvm_hmm_pfn_to_order(unsigned long hmm_pfn,
  * @notifier: Pointer to the GPU SVM notifier structure
  * @start: Start address
  * @end: End address
+ * @dev_private_owner: The device private page owner
  *
  * Check if pages between start and end have been faulted in on the CPU. Use to
  * prevent migration of pages without CPU backing store.
@@ -692,14 +690,15 @@ static unsigned int drm_gpusvm_hmm_pfn_to_order(unsigned long hmm_pfn,
  */
 static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
 				   struct drm_gpusvm_notifier *notifier,
-				   unsigned long start, unsigned long end)
+				   unsigned long start, unsigned long end,
+				   void *dev_private_owner)
 {
 	struct hmm_range hmm_range = {
 		.default_flags = 0,
 		.notifier = &notifier->notifier,
 		.start = start,
 		.end = end,
-		.dev_private_owner = gpusvm->device_private_page_owner,
+		.dev_private_owner = dev_private_owner,
 	};
 	unsigned long timeout =
 		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
@@ -753,6 +752,7 @@ err_free:
  * @gpuva_start: Start address of GPUVA which mirrors CPU
  * @gpuva_end: End address of GPUVA which mirrors CPU
  * @check_pages_threshold: Check CPU pages for present threshold
+ * @dev_private_owner: The device private page owner
  *
  * This function determines the chunk size for the GPU SVM range based on the
  * fault address, GPU SVM chunk sizes, existing GPU SVM ranges, and the virtual
@@ -767,7 +767,8 @@ drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
 			    unsigned long fault_addr,
 			    unsigned long gpuva_start,
 			    unsigned long gpuva_end,
-			    unsigned long check_pages_threshold)
+			    unsigned long check_pages_threshold,
+			    void *dev_private_owner)
 {
 	unsigned long start, end;
 	int i = 0;
@@ -814,7 +815,7 @@ retry:
 		 * process-many-malloc' mallocs at least 64k at a time.
 		 */
 		if (end - start <= check_pages_threshold &&
-		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
+		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end, dev_private_owner)) {
 			++i;
 			goto retry;
 		}
@@ -957,7 +958,8 @@ drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
 	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
 						 fault_addr, gpuva_start,
 						 gpuva_end,
-						 ctx->check_pages_threshold);
+						 ctx->check_pages_threshold,
+						 ctx->device_private_page_owner);
 	if (chunk_size == LONG_MAX) {
 		err = -EINVAL;
 		goto err_notifier_remove;
@@ -1268,7 +1270,7 @@ int drm_gpusvm_get_pages(struct drm_gpusvm *gpusvm,
 		.notifier = notifier,
 		.start = pages_start,
 		.end = pages_end,
-		.dev_private_owner = gpusvm->device_private_page_owner,
+		.dev_private_owner = ctx->device_private_page_owner,
 	};
 	void *zdd;
 	unsigned long timeout =
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 7f2f1f041f1d..7e2db71ff34e 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -67,11 +67,6 @@ void xe_svm_range_debug(struct xe_svm_range *range, const char *operation)
 	range_debug(range, operation);
 }
 
-static void *xe_svm_devm_owner(struct xe_device *xe)
-{
-	return xe;
-}
-
 static struct drm_gpusvm_range *
 xe_svm_range_alloc(struct drm_gpusvm *gpusvm)
 {
@@ -744,15 +739,14 @@ int xe_svm_init(struct xe_vm *vm)
 			  xe_svm_garbage_collector_work_func);
 
 		err = drm_gpusvm_init(&vm->svm.gpusvm, "Xe SVM", &vm->xe->drm,
-				      current->mm, xe_svm_devm_owner(vm->xe), 0,
-				      vm->size,
+				      current->mm, 0, vm->size,
 				      xe_modparam.svm_notifier_size * SZ_1M,
 				      &gpusvm_ops, fault_chunk_sizes,
 				      ARRAY_SIZE(fault_chunk_sizes));
 		drm_gpusvm_driver_set_lock(&vm->svm.gpusvm, &vm->lock);
 	} else {
 		err = drm_gpusvm_init(&vm->svm.gpusvm, "Xe SVM (simple)",
-				      &vm->xe->drm, NULL, NULL, 0, 0, 0, NULL,
+				      &vm->xe->drm, NULL, 0, 0, 0, NULL,
 				      NULL, 0);
 	}
 
@@ -1017,6 +1011,7 @@ static int __xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 		.devmem_only = need_vram && devmem_possible,
 		.timeslice_ms = need_vram && devmem_possible ?
 			vm->xe->atomic_svm_timeslice_ms : 0,
+		.device_private_page_owner = xe_svm_devm_owner(vm->xe),
 	};
 	struct xe_validation_ctx vctx;
 	struct drm_exec exec;
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index cef6ee7d6fe3..0955d2ac8d74 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -6,6 +6,20 @@
 #ifndef _XE_SVM_H_
 #define _XE_SVM_H_
 
+struct xe_device;
+
+/**
+ * xe_svm_devm_owner() - Return the owner of device private memory
+ * @xe: The xe device.
+ *
+ * Return: The owner of this device's device private memory to use in
+ * hmm_range_fault()-
+ */
+static inline void *xe_svm_devm_owner(struct xe_device *xe)
+{
+	return xe;
+}
+
 #if IS_ENABLED(CONFIG_DRM_XE_GPUSVM)
 
 #include <drm/drm_pagemap.h>
diff --git a/drivers/gpu/drm/xe/xe_userptr.c b/drivers/gpu/drm/xe/xe_userptr.c
index 91d09af71ced..f16e92cd8090 100644
--- a/drivers/gpu/drm/xe/xe_userptr.c
+++ b/drivers/gpu/drm/xe/xe_userptr.c
@@ -54,6 +54,7 @@ int xe_vma_userptr_pin_pages(struct xe_userptr_vma *uvma)
 	struct xe_device *xe = vm->xe;
 	struct drm_gpusvm_ctx ctx = {
 		.read_only = xe_vma_read_only(vma),
+		.device_private_page_owner = NULL,
 	};
 
 	lockdep_assert_held(&vm->lock);
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 80b7f13ecd80..4e914928e0a9 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -2883,6 +2883,7 @@ static int prefetch_ranges(struct xe_vm *vm, struct xe_vma_op *op)
 	ctx.read_only = xe_vma_read_only(vma);
 	ctx.devmem_possible = devmem_possible;
 	ctx.check_pages_threshold = devmem_possible ? SZ_64K : 0;
+	ctx.device_private_page_owner = xe_svm_devm_owner(vm->xe);
 
 	/* TODO: Threading the migration */
 	xa_for_each(&op->prefetch_range.range, i, svm_range) {
diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
index 5434048a2ca4..b92faa9a26b2 100644
--- a/include/drm/drm_gpusvm.h
+++ b/include/drm/drm_gpusvm.h
@@ -179,7 +179,6 @@ struct drm_gpusvm_range {
  * @name: Name of the GPU SVM
  * @drm: Pointer to the DRM device structure
  * @mm: Pointer to the mm_struct for the address space
- * @device_private_page_owner: Device private pages owner
  * @mm_start: Start address of GPU SVM
  * @mm_range: Range of the GPU SVM
  * @notifier_size: Size of individual notifiers
@@ -204,7 +203,6 @@ struct drm_gpusvm {
 	const char *name;
 	struct drm_device *drm;
 	struct mm_struct *mm;
-	void *device_private_page_owner;
 	unsigned long mm_start;
 	unsigned long mm_range;
 	unsigned long notifier_size;
@@ -226,6 +224,8 @@ struct drm_gpusvm {
 /**
  * struct drm_gpusvm_ctx - DRM GPU SVM context
  *
+ * @device_private_page_owner: The device-private page owner to use for
+ * this operation
  * @check_pages_threshold: Check CPU pages for present if chunk is less than or
  *                         equal to threshold. If not present, reduce chunk
  *                         size.
@@ -239,6 +239,7 @@ struct drm_gpusvm {
  * Context that is DRM GPUSVM is operating in (i.e. user arguments).
  */
 struct drm_gpusvm_ctx {
+	void *device_private_page_owner;
 	unsigned long check_pages_threshold;
 	unsigned long timeslice_ms;
 	unsigned int in_notifier :1;
@@ -249,7 +250,7 @@ struct drm_gpusvm_ctx {
 
 int drm_gpusvm_init(struct drm_gpusvm *gpusvm,
 		    const char *name, struct drm_device *drm,
-		    struct mm_struct *mm, void *device_private_page_owner,
+		    struct mm_struct *mm,
 		    unsigned long mm_start, unsigned long mm_range,
 		    unsigned long notifier_size,
 		    const struct drm_gpusvm_ops *ops,
-- 
cgit v1.2.3


From fcf6bddebfe03574041d7bb5f7e3f18e6b46a426 Mon Sep 17 00:00:00 2001
From: Imre Deak <imre.deak@intel.com>
Date: Tue, 30 Sep 2025 21:24:45 +0300
Subject: drm/dp: Add quirk for Synaptics DSC throughput link-bpp limit
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Some Synaptics MST branch devices have a problem decompressing a stream
with a compressed link-bpp higher than 12, if the pixel clock is higher
than ~50 % of the maximum throughput capability reported by the branch
device. The screen remains blank, or for some - mostly black content -
gets enabled, but may stil have jitter artifacts.

At least the following docking stations are affected, based on testing
both with any Intel devices or the UCD-500 reference device as a source:

- DELL WD19DCS, DELL WD19TB3, DELL WD22TB4
- ThinkPad 40AN
- HP G2

At least the following docking stations are free from this problem,
based on tests with a source/sink/mode etc. configuration matching the
test cases used above:

- DELL Dual Charge HD22Q, DELL WD25TB5
- ThinkPad 40B0
- Anker 565

All the affected devices have an older version of the Synaptics MST
branch device (Panamera), whereas all the non-affected docking stations
have a newer branch device (at least Synaptics Panamera with a higher HW
revision number and Synaptics Cayenne models). Add the required quirk
entries accordingly. The quirk will be handled by the i915/xe drivers in
a follow-up change.

The latest firmware version of the Synaptics branch device for all the
affected devices tested above is 5.7 (as reported at DPCD address
0x50a/0x50b). For the DELL devices this corresponds to the latest
01.00.14.01.A03 firmware package version of the docking station.

v2:
- Document the DP_DPCD_QUIRK_DSC_THROUGHPUT_BPP_LIMIT enum.
- Describe the quirk in more detail in the dpcd_quirk_list.
v3:
- s/Panarema/Panamera in the commit log.

Cc: dri-devel@lists.freedesktop.org
Reported-by: Vidya Srinivas <vidya.srinivas@intel.com>
Reported-and-tested-by: Swati Sharma <swati2.sharma@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://lore.kernel.org/r/20250930182450.563016-2-imre.deak@intel.com
---
 drivers/gpu/drm/display/drm_dp_helper.c | 4 ++++
 include/drm/display/drm_dp_helper.h     | 9 +++++++++
 2 files changed, 13 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/display/drm_dp_helper.c b/drivers/gpu/drm/display/drm_dp_helper.c
index 4aaeae4fa03c..4cbcb685f562 100644
--- a/drivers/gpu/drm/display/drm_dp_helper.c
+++ b/drivers/gpu/drm/display/drm_dp_helper.c
@@ -2543,6 +2543,10 @@ static const struct dpcd_quirk dpcd_quirk_list[] = {
 	{ OUI(0x00, 0x0C, 0xE7), DEVICE_ID_ANY, false, BIT(DP_DPCD_QUIRK_HBLANK_EXPANSION_REQUIRES_DSC) },
 	/* Apple MacBookPro 2017 15 inch eDP Retina panel reports too low DP_MAX_LINK_RATE */
 	{ OUI(0x00, 0x10, 0xfa), DEVICE_ID(101, 68, 21, 101, 98, 97), false, BIT(DP_DPCD_QUIRK_CAN_DO_MAX_LINK_RATE_3_24_GBPS) },
+	/* Synaptics Panamera supports only a compressed bpp of 12 above 50% of its max DSC pixel throughput */
+	{ OUI(0x90, 0xCC, 0x24), DEVICE_ID('S', 'Y', 'N', 'A', 0x53, 0x22), true, BIT(DP_DPCD_QUIRK_DSC_THROUGHPUT_BPP_LIMIT) },
+	{ OUI(0x90, 0xCC, 0x24), DEVICE_ID('S', 'Y', 'N', 'A', 0x53, 0x31), true, BIT(DP_DPCD_QUIRK_DSC_THROUGHPUT_BPP_LIMIT) },
+	{ OUI(0x90, 0xCC, 0x24), DEVICE_ID('S', 'Y', 'N', 'A', 0x53, 0x33), true, BIT(DP_DPCD_QUIRK_DSC_THROUGHPUT_BPP_LIMIT) },
 };
 
 #undef OUI
diff --git a/include/drm/display/drm_dp_helper.h b/include/drm/display/drm_dp_helper.h
index 87caa4f1fdb8..558f5ce924ac 100644
--- a/include/drm/display/drm_dp_helper.h
+++ b/include/drm/display/drm_dp_helper.h
@@ -820,6 +820,15 @@ enum drm_dp_quirk {
 	 * requires enabling DSC.
 	 */
 	DP_DPCD_QUIRK_HBLANK_EXPANSION_REQUIRES_DSC,
+	/**
+	 * @DP_DPCD_QUIRK_DSC_THROUGHPUT_BPP_LIMIT:
+	 *
+	 * The device doesn't support DSC decompression at the maximum DSC
+	 * pixel throughput and compressed bpp it indicates via its DPCD DSC
+	 * capabilities. The compressed bpp must be limited above a device
+	 * specific DSC pixel throughput.
+	 */
+	DP_DPCD_QUIRK_DSC_THROUGHPUT_BPP_LIMIT,
 };
 
 /**
-- 
cgit v1.2.3


From 1f95871207db4439a3116e9a86f5b5658a5157c4 Mon Sep 17 00:00:00 2001
From: Imre Deak <imre.deak@intel.com>
Date: Tue, 30 Sep 2025 21:24:46 +0300
Subject: drm/dp: Add helpers to query the branch DSC max throughput/line-width
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add helpers to query the DP DSC sink device's per-slice throughput as
well as a DSC branch device's overall throughput and line-width
capabilities.

v2 (Ville):
- Rename pixel_clock to peak_pixel_rate, document what the value means
  in case of MST tiled displays.
- Fix name of drm_dp_dsc_branch_max_slice_throughput() to
  drm_dp_dsc_sink_max_slice_throughput().
v3:
- Fix the DSC branch device minimum valid line width value from 2560
  to 5120 pixels.
- Fix drm_dp_dsc_sink_max_slice_throughput()'s pixel_clock parameter
  name to peak_pixel_rate in header file.
- Add handling for throughput mode 0 granular delta, defined by DP
  Standard v2.1a.
v4:
- Remove the default switch case in
  drm_dp_dsc_sink_max_slice_throughput(), which is unreachable in the
  current code. (Ville)

Cc: dri-devel@lists.freedesktop.org
Suggested-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reported-and-tested-by: Swati Sharma <swati2.sharma@intel.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Imre Deak <imre.deak@intel.com>
Link: https://lore.kernel.org/r/20250930182450.563016-3-imre.deak@intel.com
---
 drivers/gpu/drm/display/drm_dp_helper.c | 152 ++++++++++++++++++++++++++++++++
 include/drm/display/drm_dp.h            |   3 +
 include/drm/display/drm_dp_helper.h     |   5 ++
 3 files changed, 160 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/display/drm_dp_helper.c b/drivers/gpu/drm/display/drm_dp_helper.c
index 4cbcb685f562..1ebed7c41b0d 100644
--- a/drivers/gpu/drm/display/drm_dp_helper.c
+++ b/drivers/gpu/drm/display/drm_dp_helper.c
@@ -2836,6 +2836,158 @@ int drm_dp_dsc_sink_supported_input_bpcs(const u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_S
 }
 EXPORT_SYMBOL(drm_dp_dsc_sink_supported_input_bpcs);
 
+/*
+ * See DP Standard v2.1a 2.8.4 Minimum Slices/Display, Table 2-159 and
+ * Appendix L.1 Derivation of Slice Count Requirements.
+ */
+static int dsc_sink_min_slice_throughput(int peak_pixel_rate)
+{
+	if (peak_pixel_rate >= 4800000)
+		return 600000;
+	else if (peak_pixel_rate >= 2700000)
+		return 400000;
+	else
+		return 340000;
+}
+
+/**
+ * drm_dp_dsc_sink_max_slice_throughput() - Get a DSC sink's maximum pixel throughput per slice
+ * @dsc_dpcd: DSC sink's capabilities from DPCD
+ * @peak_pixel_rate: Cumulative peak pixel rate in kHz
+ * @is_rgb_yuv444: The mode is either RGB or YUV444
+ *
+ * Return the DSC sink device's maximum pixel throughput per slice, based on
+ * the device's @dsc_dpcd capabilities, the @peak_pixel_rate of the transferred
+ * stream(s) and whether the output format @is_rgb_yuv444 or yuv422/yuv420.
+ *
+ * Note that @peak_pixel_rate is the total pixel rate transferred to the same
+ * DSC/display sink. For instance to calculate a tile's slice count of an MST
+ * multi-tiled display sink (not considering here the required
+ * rounding/alignment of slice count)::
+ *
+ *   @peak_pixel_rate = tile_pixel_rate * tile_count
+ *   total_slice_count = @peak_pixel_rate / drm_dp_dsc_sink_max_slice_throughput(@peak_pixel_rate)
+ *   tile_slice_count = total_slice_count / tile_count
+ *
+ * Returns:
+ * The maximum pixel throughput per slice supported by the DSC sink device
+ * in kPixels/sec.
+ */
+int drm_dp_dsc_sink_max_slice_throughput(const u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE],
+					 int peak_pixel_rate, bool is_rgb_yuv444)
+{
+	int throughput;
+	int delta = 0;
+	int base;
+
+	throughput = dsc_dpcd[DP_DSC_PEAK_THROUGHPUT - DP_DSC_SUPPORT];
+
+	if (is_rgb_yuv444) {
+		throughput = (throughput & DP_DSC_THROUGHPUT_MODE_0_MASK) >>
+			     DP_DSC_THROUGHPUT_MODE_0_SHIFT;
+
+		delta = ((dsc_dpcd[DP_DSC_RC_BUF_BLK_SIZE - DP_DSC_SUPPORT]) &
+			 DP_DSC_THROUGHPUT_MODE_0_DELTA_MASK) >>
+			DP_DSC_THROUGHPUT_MODE_0_DELTA_SHIFT;	/* in units of 2 MPixels/sec */
+		delta *= 2000;
+	} else {
+		throughput = (throughput & DP_DSC_THROUGHPUT_MODE_1_MASK) >>
+			     DP_DSC_THROUGHPUT_MODE_1_SHIFT;
+	}
+
+	switch (throughput) {
+	case 0:
+		return dsc_sink_min_slice_throughput(peak_pixel_rate);
+	case 1:
+		base = 340000;
+		break;
+	case 2 ... 14:
+		base = 400000 + 50000 * (throughput - 2);
+		break;
+	case 15:
+		base = 170000;
+		break;
+	}
+
+	return base + delta;
+}
+EXPORT_SYMBOL(drm_dp_dsc_sink_max_slice_throughput);
+
+static u8 dsc_branch_dpcd_cap(const u8 dpcd[DP_DSC_BRANCH_CAP_SIZE], int reg)
+{
+	return dpcd[reg - DP_DSC_BRANCH_OVERALL_THROUGHPUT_0];
+}
+
+/**
+ * drm_dp_dsc_branch_max_overall_throughput() - Branch device's max overall DSC pixel throughput
+ * @dsc_branch_dpcd: DSC branch capabilities from DPCD
+ * @is_rgb_yuv444: The mode is either RGB or YUV444
+ *
+ * Return the branch device's maximum overall DSC pixel throughput, based on
+ * the device's DPCD DSC branch capabilities, and whether the output
+ * format @is_rgb_yuv444 or yuv422/yuv420.
+ *
+ * Returns:
+ * - 0:   The maximum overall throughput capability is not indicated by
+ *        the device separately and it must be determined from the per-slice
+ *        max throughput (see @drm_dp_dsc_branch_slice_max_throughput())
+ *        and the maximum slice count supported by the device.
+ * - > 0: The maximum overall DSC pixel throughput supported by the branch
+ *        device in kPixels/sec.
+ */
+int drm_dp_dsc_branch_max_overall_throughput(const u8 dsc_branch_dpcd[DP_DSC_BRANCH_CAP_SIZE],
+					     bool is_rgb_yuv444)
+{
+	int throughput;
+
+	if (is_rgb_yuv444)
+		throughput = dsc_branch_dpcd_cap(dsc_branch_dpcd,
+						 DP_DSC_BRANCH_OVERALL_THROUGHPUT_0);
+	else
+		throughput = dsc_branch_dpcd_cap(dsc_branch_dpcd,
+						 DP_DSC_BRANCH_OVERALL_THROUGHPUT_1);
+
+	switch (throughput) {
+	case 0:
+		return 0;
+	case 1:
+		return 680000;
+	default:
+		return 600000 + 50000 * throughput;
+	}
+}
+EXPORT_SYMBOL(drm_dp_dsc_branch_max_overall_throughput);
+
+/**
+ * drm_dp_dsc_branch_max_line_width() - Branch device's max DSC line width
+ * @dsc_branch_dpcd: DSC branch capabilities from DPCD
+ *
+ * Return the branch device's maximum overall DSC line width, based on
+ * the device's @dsc_branch_dpcd capabilities.
+ *
+ * Returns:
+ * - 0:        The maximum line width is not indicated by the device
+ *             separately and it must be determined from the maximum
+ *             slice count and slice-width supported by the device.
+ * - %-EINVAL: The device indicates an invalid maximum line width
+ *             (< 5120 pixels).
+ * - >= 5120:  The maximum line width in pixels.
+ */
+int drm_dp_dsc_branch_max_line_width(const u8 dsc_branch_dpcd[DP_DSC_BRANCH_CAP_SIZE])
+{
+	int line_width = dsc_branch_dpcd_cap(dsc_branch_dpcd, DP_DSC_BRANCH_MAX_LINE_WIDTH);
+
+	switch (line_width) {
+	case 0:
+		return 0;
+	case 1 ... 15:
+		return -EINVAL;
+	default:
+		return line_width * 320;
+	}
+}
+EXPORT_SYMBOL(drm_dp_dsc_branch_max_line_width);
+
 static int drm_dp_read_lttpr_regs(struct drm_dp_aux *aux,
 				  const u8 dpcd[DP_RECEIVER_CAP_SIZE], int address,
 				  u8 *buf, int buf_size)
diff --git a/include/drm/display/drm_dp.h b/include/drm/display/drm_dp.h
index 811e9238a77c..600bca40249b 100644
--- a/include/drm/display/drm_dp.h
+++ b/include/drm/display/drm_dp.h
@@ -257,6 +257,8 @@
 # define DP_DSC_RC_BUF_BLK_SIZE_4           0x1
 # define DP_DSC_RC_BUF_BLK_SIZE_16          0x2
 # define DP_DSC_RC_BUF_BLK_SIZE_64          0x3
+# define DP_DSC_THROUGHPUT_MODE_0_DELTA_SHIFT 3 /* DP 2.1a, in units of 2 MPixels/sec */
+# define DP_DSC_THROUGHPUT_MODE_0_DELTA_MASK  (0x1f << DP_DSC_THROUGHPUT_MODE_0_DELTA_SHIFT)
 
 #define DP_DSC_RC_BUF_SIZE                  0x063
 
@@ -1683,6 +1685,7 @@ enum drm_dp_phy {
 #define DP_BRANCH_OUI_HEADER_SIZE	0xc
 #define DP_RECEIVER_CAP_SIZE		0xf
 #define DP_DSC_RECEIVER_CAP_SIZE        0x10 /* DSC Capabilities 0x60 through 0x6F */
+#define DP_DSC_BRANCH_CAP_SIZE		3
 #define EDP_PSR_RECEIVER_CAP_SIZE	2
 #define EDP_DISPLAY_CTL_CAP_SIZE	5
 #define DP_LTTPR_COMMON_CAP_SIZE	8
diff --git a/include/drm/display/drm_dp_helper.h b/include/drm/display/drm_dp_helper.h
index 558f5ce924ac..fc02eb888be8 100644
--- a/include/drm/display/drm_dp_helper.h
+++ b/include/drm/display/drm_dp_helper.h
@@ -203,6 +203,11 @@ u8 drm_dp_dsc_sink_max_slice_count(const u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE],
 u8 drm_dp_dsc_sink_line_buf_depth(const u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE]);
 int drm_dp_dsc_sink_supported_input_bpcs(const u8 dsc_dpc[DP_DSC_RECEIVER_CAP_SIZE],
 					 u8 dsc_bpc[3]);
+int drm_dp_dsc_sink_max_slice_throughput(const u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE],
+					 int peak_pixel_rate, bool is_rgb_yuv444);
+int drm_dp_dsc_branch_max_overall_throughput(const u8 dsc_branch_dpcd[DP_DSC_BRANCH_CAP_SIZE],
+					     bool is_rgb_yuv444);
+int drm_dp_dsc_branch_max_line_width(const u8 dsc_branch_dpcd[DP_DSC_BRANCH_CAP_SIZE]);
 
 static inline bool
 drm_dp_sink_supports_dsc(const u8 dsc_dpcd[DP_DSC_RECEIVER_CAP_SIZE])
-- 
cgit v1.2.3


From 54cea7e655fe3d031c9e7eaaf127bffe0496e7cd Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:17 +0200
Subject: drm/atomic: Remove unused drm_atomic_get_existing_connector_state()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The drm_atomic_get_existing_connector_state() function is deprecated and
isn't used anymore, so let's remove it.

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Tested-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-2-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 23 -----------------------
 1 file changed, 23 deletions(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index 38636a593c9d..321c866d5b0a 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -731,29 +731,6 @@ drm_atomic_get_new_plane_state(const struct drm_atomic_state *state,
 	return state->planes[drm_plane_index(plane)].new_state;
 }
 
-/**
- * drm_atomic_get_existing_connector_state - get connector state, if it exists
- * @state: global atomic state object
- * @connector: connector to grab
- *
- * This function returns the connector state for the given connector,
- * or NULL if the connector is not part of the global atomic state.
- *
- * This function is deprecated, @drm_atomic_get_old_connector_state or
- * @drm_atomic_get_new_connector_state should be used instead.
- */
-static inline struct drm_connector_state *
-drm_atomic_get_existing_connector_state(const struct drm_atomic_state *state,
-					struct drm_connector *connector)
-{
-	int index = drm_connector_index(connector);
-
-	if (index >= state->num_connector)
-		return NULL;
-
-	return state->connectors[index].state;
-}
-
 /**
  * drm_atomic_get_old_connector_state - get connector state, if it exists
  * @state: global atomic state object
-- 
cgit v1.2.3


From 9525f537ce00143f77e4208812cce3aa1d1d1f56 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:18 +0200
Subject: drm/atomic: Document __drm_connectors_state state pointer
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

While the old and new state pointers are somewhat self-explanatory, the
state pointer and its relation to the other two really isn't.

Now that we've cleaned up everything and it isn't used in any
modesetting path, we can document what it's still useful for: to free
the right state when we free the global state.

Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-3-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 19 ++++++++++++++++++-
 1 file changed, 18 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index 321c866d5b0a..fe47715c31e4 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -182,7 +182,24 @@ struct __drm_crtcs_state {
 
 struct __drm_connnectors_state {
 	struct drm_connector *ptr;
-	struct drm_connector_state *state, *old_state, *new_state;
+
+	/**
+	 * @state:
+	 *
+	 * Used to track the @drm_connector_state we will need to free
+	 * when tearing down the associated &drm_atomic_state in
+	 * $drm_mode_config_funcs.atomic_state_clear or
+	 * drm_atomic_state_default_clear().
+	 *
+	 * Before a commit, and the call to
+	 * drm_atomic_helper_swap_state() in particular, it points to
+	 * the same state than @new_state. After a commit, it points to
+	 * the same state than @old_state.
+	 */
+	struct drm_connector_state *state;
+
+	struct drm_connector_state *old_state, *new_state;
+
 	/**
 	 * @out_fence_ptr:
 	 *
-- 
cgit v1.2.3


From 6b592431c5b07f4fd2fa3ea123eda3ae92d8a2a0 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:19 +0200
Subject: drm/atomic: Convert __drm_atomic_get_current_plane_state() to modern
 accessor
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The __drm_atomic_get_current_plane_state() function tries to get and
return the existing plane state, and if it doesn't exist returns the one
stored in the drm_plane->state field.

Using the current nomenclature, it tries to get the existing plane state
with an ad-hoc implementation of drm_atomic_get_existing_plane_state(),
and falls back to either the old or new plane state, depending on
whether it is called before or after drm_atomic_helper_swap_state().

The existing plane state itself is deprecated, because it also changes
when swapping states from the new state to the old state.

Fortunately for us, we can simplify things. Indeed,
__drm_atomic_get_current_plane_state() is only used in two macros:
intel_atomic_crtc_state_for_each_plane_state and
drm_atomic_crtc_state_for_each_plane_state().

The intel variant is only used through the intel_wm_compute() function
that is only ever called in intel_crtc_atomic_check().

The generic variant is more widely used, and can be found in the malidp,
msm, tegra and vc4 drivers. All of these call sites though are during
atomic_check(), so we end up in the same situation than Intel's.

Thus, we only ever use the existing state as the new state, and
plane->state is always going to be the old state. Any plane isn't
guaranteed to be part of the state though, so we can't rely on
drm_atomic_get_old_plane_state() and we still need to use plane->state.

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-4-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 20 +++++++++++++-------
 1 file changed, 13 insertions(+), 7 deletions(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index fe47715c31e4..c8b88990506e 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -793,11 +793,11 @@ drm_atomic_get_new_connector_state(const struct drm_atomic_state *state,
  * @state: global atomic state object
  * @plane: plane to grab
  *
- * This function returns the plane state for the given plane, either from
- * @state, or if the plane isn't part of the atomic state update, from @plane.
- * This is useful in atomic check callbacks, when drivers need to peek at, but
- * not change, state of other planes, since it avoids threading an error code
- * back up the call chain.
+ * This function returns the plane state for the given plane, either the
+ * new plane state from @state, or if the plane isn't part of the atomic
+ * state update, from @plane. This is useful in atomic check callbacks,
+ * when drivers need to peek at, but not change, state of other planes,
+ * since it avoids threading an error code back up the call chain.
  *
  * WARNING:
  *
@@ -818,9 +818,15 @@ static inline const struct drm_plane_state *
 __drm_atomic_get_current_plane_state(const struct drm_atomic_state *state,
 				     struct drm_plane *plane)
 {
-	if (state->planes[drm_plane_index(plane)].state)
-		return state->planes[drm_plane_index(plane)].state;
+	struct drm_plane_state *plane_state;
 
+	plane_state = drm_atomic_get_new_plane_state(state, plane);
+	if (plane_state)
+		return plane_state;
+
+	/*
+	 * If the plane isn't part of the state, fallback to the currently active one.
+	 */
 	return plane->state;
 }
 
-- 
cgit v1.2.3


From 93d851281755f0a279402e1b88d29584ed8244e9 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:23 +0200
Subject: drm/atomic: Remove unused drm_atomic_get_existing_plane_state()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The drm_atomic_get_existing_plane_state() function is deprecated and
isn't used anymore, so let's remove it.

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-8-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 18 ------------------
 1 file changed, 18 deletions(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index c8b88990506e..b2cc9c4a9beb 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -700,24 +700,6 @@ drm_atomic_get_new_crtc_state(const struct drm_atomic_state *state,
 	return state->crtcs[drm_crtc_index(crtc)].new_state;
 }
 
-/**
- * drm_atomic_get_existing_plane_state - get plane state, if it exists
- * @state: global atomic state object
- * @plane: plane to grab
- *
- * This function returns the plane state for the given plane, or NULL
- * if the plane is not part of the global atomic state.
- *
- * This function is deprecated, @drm_atomic_get_old_plane_state or
- * @drm_atomic_get_new_plane_state should be used instead.
- */
-static inline struct drm_plane_state *
-drm_atomic_get_existing_plane_state(const struct drm_atomic_state *state,
-				    struct drm_plane *plane)
-{
-	return state->planes[drm_plane_index(plane)].state;
-}
-
 /**
  * drm_atomic_get_old_plane_state - get plane state, if it exists
  * @state: global atomic state object
-- 
cgit v1.2.3


From 7ca4e7fa4d57ce6eac4842e925ea1f98ea54a365 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:24 +0200
Subject: drm/atomic: Document __drm_planes_state state pointer
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

While the old and new state pointers are somewhat self-explanatory, the
state pointer and its relation to the other two really isn't.

Now that we've cleaned up everything and it isn't used in any
modesetting path, we can document what it's still useful for: to free
the right state when we free the global state.

Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-9-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index b2cc9c4a9beb..a05012adcf94 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -159,7 +159,23 @@ struct drm_crtc_commit {
 
 struct __drm_planes_state {
 	struct drm_plane *ptr;
-	struct drm_plane_state *state, *old_state, *new_state;
+
+	/**
+	 * @state:
+	 *
+	 * Used to track the @drm_plane_state we will need to free when
+	 * tearing down the associated &drm_atomic_state in
+	 * $drm_mode_config_funcs.atomic_state_clear or
+	 * drm_atomic_state_default_clear().
+	 *
+	 * Before a commit, and the call to
+	 * drm_atomic_helper_swap_state() in particular, it points to
+	 * the same state than @new_state. After a commit, it points to
+	 * the same state than @old_state.
+	 */
+	struct drm_plane_state *state;
+
+	struct drm_plane_state *old_state, *new_state;
 };
 
 struct __drm_crtcs_state {
-- 
cgit v1.2.3


From 3c2b2136debfc200b8ebcd8e5302348b6c86158d Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:50 +0200
Subject: drm/atomic: Remove unused drm_atomic_get_existing_crtc_state()
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The drm_atomic_get_existing_crtc_state() function is deprecated and
isn't used anymore, so let's remove it.

Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-35-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 18 ------------------
 1 file changed, 18 deletions(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index a05012adcf94..f53a32688b33 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -669,24 +669,6 @@ struct drm_crtc *
 drm_atomic_get_new_crtc_for_encoder(struct drm_atomic_state *state,
 					 struct drm_encoder *encoder);
 
-/**
- * drm_atomic_get_existing_crtc_state - get CRTC state, if it exists
- * @state: global atomic state object
- * @crtc: CRTC to grab
- *
- * This function returns the CRTC state for the given CRTC, or NULL
- * if the CRTC is not part of the global atomic state.
- *
- * This function is deprecated, @drm_atomic_get_old_crtc_state or
- * @drm_atomic_get_new_crtc_state should be used instead.
- */
-static inline struct drm_crtc_state *
-drm_atomic_get_existing_crtc_state(const struct drm_atomic_state *state,
-				   struct drm_crtc *crtc)
-{
-	return state->crtcs[drm_crtc_index(crtc)].state;
-}
-
 /**
  * drm_atomic_get_old_crtc_state - get old CRTC state, if it exists
  * @state: global atomic state object
-- 
cgit v1.2.3


From 4ae41729a658aee3e68a609cc0275658b724ebc4 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:51 +0200
Subject: drm/atomic: Document __drm_crtcs_state state pointer
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

While the old and new state pointers are somewhat self-explanatory, the
state pointer and its relation to the other two really isn't.

Now that we've cleaned up everything and it isn't used in any
modesetting path, we can document what it's still useful for: to free
the right state when we free the global state.

Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-36-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index f53a32688b33..6b634b747490 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -180,7 +180,23 @@ struct __drm_planes_state {
 
 struct __drm_crtcs_state {
 	struct drm_crtc *ptr;
-	struct drm_crtc_state *state, *old_state, *new_state;
+
+	/**
+	 * @state:
+	 *
+	 * Used to track the @drm_crtc_state we will need to free when
+	 * tearing down the associated &drm_atomic_state in
+	 * $drm_mode_config_funcs.atomic_state_clear or
+	 * drm_atomic_state_default_clear().
+	 *
+	 * Before a commit, and the call to
+	 * drm_atomic_helper_swap_state() in particular, it points to
+	 * the same state than @new_state. After a commit, it points to
+	 * the same state than @old_state.
+	 */
+	struct drm_crtc_state *state;
+
+	struct drm_crtc_state *old_state, *new_state;
 
 	/**
 	 * @commit:
-- 
cgit v1.2.3


From 7a031e8d3528ba0860d282ffd3c88fbda4bf8c4c Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Tue, 30 Sep 2025 12:59:54 +0200
Subject: drm/atomic: Document __drm_private_objs_state state pointer
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

While the old and new state pointers are somewhat self-explanatory, the
state pointer and its relation to the other two really isn't.

Now that we've cleaned up everything and it isn't used in any
modesetting path, we can document what it's still useful for: to free
the right state when we free the global state.

Reviewed-by: Luca Ceresoli <luca.ceresoli@bootlin.com>
Reviewed-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://lore.kernel.org/r/20250930-drm-no-more-existing-state-v5-39-eeb9e1287907@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 include/drm/drm_atomic.h | 18 +++++++++++++++++-
 1 file changed, 17 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index 6b634b747490..c8ab2163bf65 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -391,7 +391,23 @@ struct drm_private_state {
 
 struct __drm_private_objs_state {
 	struct drm_private_obj *ptr;
-	struct drm_private_state *state, *old_state, *new_state;
+
+	/**
+	 * @state:
+	 *
+	 * Used to track the @drm_private_state we will need to free
+	 * when tearing down the associated &drm_atomic_state in
+	 * $drm_mode_config_funcs.atomic_state_clear or
+	 * drm_atomic_state_default_clear().
+	 *
+	 * Before a commit, and the call to
+	 * drm_atomic_helper_swap_state() in particular, it points to
+	 * the same state than @new_state. After a commit, it points to
+	 * the same state than @old_state.
+	 */
+	struct drm_private_state *state;
+
+	struct drm_private_state *old_state, *new_state;
 };
 
 /**
-- 
cgit v1.2.3


From c178e534fff1d5a74da80ea03b20e2b948a00113 Mon Sep 17 00:00:00 2001
From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Date: Mon, 6 Oct 2025 15:21:22 +0530
Subject: drm/buddy: Optimize free block management with RB tree

Replace the freelist (O(n)) used for free block management with a
red-black tree, providing more efficient O(log n) search, insert,
and delete operations. This improves scalability and performance
when managing large numbers of free blocks per order (e.g., hundreds
or thousands).

In the VK-CTS memory stress subtest, the buddy manager merges
fragmented memory and inserts freed blocks into the freelist. Since
freelist insertion is O(n), this becomes a bottleneck as fragmentation
increases. Benchmarking shows list_insert_sorted() consumes ~52.69% CPU
with the freelist, compared to just 0.03% with the RB tree
(rbtree_insert.isra.0), despite performing the same sorted insert.

This also improves performance in heavily fragmented workloads,
such as games or graphics tests that stress memory.

As the buddy allocator evolves with new features such as clear-page
tracking, the resulting fragmentation and complexity have grown.
These RB-tree based design changes are introduced to address that
growth and ensure the allocator continues to perform efficiently
under fragmented conditions.

The RB tree implementation with separate clear/dirty trees provides:
- O(n log n) aggregate complexity for all operations instead of O(n^2)
- Elimination of soft lockups and system instability
- Improved code maintainability and clarity
- Better scalability for large memory systems
- Predictable performance under fragmentation

v3(Matthew):
  - Remove RB_EMPTY_NODE check in force_merge function.
  - Rename rb for loop macros to have less generic names and move to
    .c file.
  - Make the rb node rb and link field as union.

v4(Jani Nikula):
  - The kernel-doc comment should be "/**"
  - Move all the rbtree macros to rbtree.h and add parens to ensure
    correct precedence.

v5:
  - Remove the inline in a .c file (Jani Nikula).

v6(Peter Zijlstra):
  - Add rb_add() function replacing the existing rbtree_insert() code.

v7:
  - A full walk iteration in rbtree is slower than the list (Peter Zijlstra).
  - The existing rbtree_postorder_for_each_entry_safe macro should be used
    in scenarios where traversal order is not a critical factor (Christian).

v8(Matthew):
  - Remove the rbtree_is_empty() check in this patch as well.

Cc: stable@vger.kernel.org
Fixes: a68c7eaa7a8f ("drm/amdgpu: Enable clear page functionality")
Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://lore.kernel.org/r/20251006095124.1663-1-Arunpravin.PaneerSelvam@amd.com
---
 drivers/gpu/drm/drm_buddy.c | 195 +++++++++++++++++++++++++++-----------------
 include/drm/drm_buddy.h     |  11 ++-
 2 files changed, 126 insertions(+), 80 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_buddy.c b/drivers/gpu/drm/drm_buddy.c
index a94061f373de..c87210a06c31 100644
--- a/drivers/gpu/drm/drm_buddy.c
+++ b/drivers/gpu/drm/drm_buddy.c
@@ -14,6 +14,8 @@
 
 static struct kmem_cache *slab_blocks;
 
+#define rbtree_get_free_block(node) rb_entry((node), struct drm_buddy_block, rb)
+
 static struct drm_buddy_block *drm_block_alloc(struct drm_buddy *mm,
 					       struct drm_buddy_block *parent,
 					       unsigned int order,
@@ -31,6 +33,8 @@ static struct drm_buddy_block *drm_block_alloc(struct drm_buddy *mm,
 	block->header |= order;
 	block->parent = parent;
 
+	RB_CLEAR_NODE(&block->rb);
+
 	BUG_ON(block->header & DRM_BUDDY_HEADER_UNUSED);
 	return block;
 }
@@ -41,23 +45,49 @@ static void drm_block_free(struct drm_buddy *mm,
 	kmem_cache_free(slab_blocks, block);
 }
 
-static void list_insert_sorted(struct drm_buddy *mm,
-			       struct drm_buddy_block *block)
+static bool drm_buddy_block_offset_less(const struct drm_buddy_block *block,
+					const struct drm_buddy_block *node)
 {
-	struct drm_buddy_block *node;
-	struct list_head *head;
+	return drm_buddy_block_offset(block) < drm_buddy_block_offset(node);
+}
 
-	head = &mm->free_list[drm_buddy_block_order(block)];
-	if (list_empty(head)) {
-		list_add(&block->link, head);
-		return;
-	}
+static bool rbtree_block_offset_less(struct rb_node *block,
+				     const struct rb_node *node)
+{
+	return drm_buddy_block_offset_less(rbtree_get_free_block(block),
+					   rbtree_get_free_block(node));
+}
 
-	list_for_each_entry(node, head, link)
-		if (drm_buddy_block_offset(block) < drm_buddy_block_offset(node))
-			break;
+static void rbtree_insert(struct drm_buddy *mm,
+			  struct drm_buddy_block *block)
+{
+	rb_add(&block->rb,
+	       &mm->free_tree[drm_buddy_block_order(block)],
+	       rbtree_block_offset_less);
+}
+
+static void rbtree_remove(struct drm_buddy *mm,
+			  struct drm_buddy_block *block)
+{
+	struct rb_root *root;
+
+	root = &mm->free_tree[drm_buddy_block_order(block)];
+	rb_erase(&block->rb, root);
+
+	RB_CLEAR_NODE(&block->rb);
+}
+
+static struct drm_buddy_block *
+rbtree_last_entry(struct drm_buddy *mm, unsigned int order)
+{
+	struct rb_node *node = rb_last(&mm->free_tree[order]);
+
+	return node ? rb_entry(node, struct drm_buddy_block, rb) : NULL;
+}
 
-	__list_add(&block->link, node->link.prev, &node->link);
+static bool rbtree_is_empty(struct drm_buddy *mm, unsigned int order)
+{
+	return RB_EMPTY_ROOT(&mm->free_tree[order]);
 }
 
 static void clear_reset(struct drm_buddy_block *block)
@@ -70,12 +100,13 @@ static void mark_cleared(struct drm_buddy_block *block)
 	block->header |= DRM_BUDDY_HEADER_CLEAR;
 }
 
-static void mark_allocated(struct drm_buddy_block *block)
+static void mark_allocated(struct drm_buddy *mm,
+			   struct drm_buddy_block *block)
 {
 	block->header &= ~DRM_BUDDY_HEADER_STATE;
 	block->header |= DRM_BUDDY_ALLOCATED;
 
-	list_del(&block->link);
+	rbtree_remove(mm, block);
 }
 
 static void mark_free(struct drm_buddy *mm,
@@ -84,15 +115,16 @@ static void mark_free(struct drm_buddy *mm,
 	block->header &= ~DRM_BUDDY_HEADER_STATE;
 	block->header |= DRM_BUDDY_FREE;
 
-	list_insert_sorted(mm, block);
+	rbtree_insert(mm, block);
 }
 
-static void mark_split(struct drm_buddy_block *block)
+static void mark_split(struct drm_buddy *mm,
+		       struct drm_buddy_block *block)
 {
 	block->header &= ~DRM_BUDDY_HEADER_STATE;
 	block->header |= DRM_BUDDY_SPLIT;
 
-	list_del(&block->link);
+	rbtree_remove(mm, block);
 }
 
 static inline bool overlaps(u64 s1, u64 e1, u64 s2, u64 e2)
@@ -148,7 +180,7 @@ static unsigned int __drm_buddy_free(struct drm_buddy *mm,
 				mark_cleared(parent);
 		}
 
-		list_del(&buddy->link);
+		rbtree_remove(mm, buddy);
 		if (force_merge && drm_buddy_block_is_clear(buddy))
 			mm->clear_avail -= drm_buddy_block_size(mm, buddy);
 
@@ -179,13 +211,19 @@ static int __force_merge(struct drm_buddy *mm,
 		return -EINVAL;
 
 	for (i = min_order - 1; i >= 0; i--) {
-		struct drm_buddy_block *block, *prev;
+		struct rb_root *root = &mm->free_tree[i];
+		struct rb_node *iter;
+
+		iter = rb_last(root);
 
-		list_for_each_entry_safe_reverse(block, prev, &mm->free_list[i], link) {
-			struct drm_buddy_block *buddy;
+		while (iter) {
+			struct drm_buddy_block *block, *buddy;
 			u64 block_start, block_end;
 
-			if (!block->parent)
+			block = rbtree_get_free_block(iter);
+			iter = rb_prev(iter);
+
+			if (!block || !block->parent)
 				continue;
 
 			block_start = drm_buddy_block_offset(block);
@@ -201,15 +239,10 @@ static int __force_merge(struct drm_buddy *mm,
 			WARN_ON(drm_buddy_block_is_clear(block) ==
 				drm_buddy_block_is_clear(buddy));
 
-			/*
-			 * If the prev block is same as buddy, don't access the
-			 * block in the next iteration as we would free the
-			 * buddy block as part of the free function.
-			 */
-			if (prev == buddy)
-				prev = list_prev_entry(prev, link);
+			if (iter == &buddy->rb)
+				iter = rb_prev(iter);
 
-			list_del(&block->link);
+			rbtree_remove(mm, block);
 			if (drm_buddy_block_is_clear(block))
 				mm->clear_avail -= drm_buddy_block_size(mm, block);
 
@@ -237,7 +270,7 @@ static int __force_merge(struct drm_buddy *mm,
 int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 {
 	unsigned int i;
-	u64 offset;
+	u64 offset = 0;
 
 	if (size < chunk_size)
 		return -EINVAL;
@@ -258,14 +291,14 @@ int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 
 	BUG_ON(mm->max_order > DRM_BUDDY_MAX_ORDER);
 
-	mm->free_list = kmalloc_array(mm->max_order + 1,
-				      sizeof(struct list_head),
+	mm->free_tree = kmalloc_array(mm->max_order + 1,
+				      sizeof(struct rb_root),
 				      GFP_KERNEL);
-	if (!mm->free_list)
+	if (!mm->free_tree)
 		return -ENOMEM;
 
 	for (i = 0; i <= mm->max_order; ++i)
-		INIT_LIST_HEAD(&mm->free_list[i]);
+		mm->free_tree[i] = RB_ROOT;
 
 	mm->n_roots = hweight64(size);
 
@@ -273,9 +306,8 @@ int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 				  sizeof(struct drm_buddy_block *),
 				  GFP_KERNEL);
 	if (!mm->roots)
-		goto out_free_list;
+		goto out_free_tree;
 
-	offset = 0;
 	i = 0;
 
 	/*
@@ -312,8 +344,8 @@ out_free_roots:
 	while (i--)
 		drm_block_free(mm, mm->roots[i]);
 	kfree(mm->roots);
-out_free_list:
-	kfree(mm->free_list);
+out_free_tree:
+	kfree(mm->free_tree);
 	return -ENOMEM;
 }
 EXPORT_SYMBOL(drm_buddy_init);
@@ -323,7 +355,7 @@ EXPORT_SYMBOL(drm_buddy_init);
  *
  * @mm: DRM buddy manager to free
  *
- * Cleanup memory manager resources and the freelist
+ * Cleanup memory manager resources and the freetree
  */
 void drm_buddy_fini(struct drm_buddy *mm)
 {
@@ -350,7 +382,7 @@ void drm_buddy_fini(struct drm_buddy *mm)
 	WARN_ON(mm->avail != mm->size);
 
 	kfree(mm->roots);
-	kfree(mm->free_list);
+	kfree(mm->free_tree);
 }
 EXPORT_SYMBOL(drm_buddy_fini);
 
@@ -383,7 +415,7 @@ static int split_block(struct drm_buddy *mm,
 		clear_reset(block);
 	}
 
-	mark_split(block);
+	mark_split(mm, block);
 
 	return 0;
 }
@@ -412,7 +444,7 @@ EXPORT_SYMBOL(drm_get_buddy);
  * @is_clear: blocks clear state
  *
  * Reset the clear state based on @is_clear value for each block
- * in the freelist.
+ * in the freetree.
  */
 void drm_buddy_reset_clear(struct drm_buddy *mm, bool is_clear)
 {
@@ -431,9 +463,9 @@ void drm_buddy_reset_clear(struct drm_buddy *mm, bool is_clear)
 	}
 
 	for (i = 0; i <= mm->max_order; ++i) {
-		struct drm_buddy_block *block;
+		struct drm_buddy_block *block, *tmp;
 
-		list_for_each_entry_reverse(block, &mm->free_list[i], link) {
+		rbtree_postorder_for_each_entry_safe(block, tmp, &mm->free_tree[i], rb) {
 			if (is_clear != drm_buddy_block_is_clear(block)) {
 				if (is_clear) {
 					mark_cleared(block);
@@ -639,14 +671,18 @@ get_maxblock(struct drm_buddy *mm, unsigned int order,
 	unsigned int i;
 
 	for (i = order; i <= mm->max_order; ++i) {
+		struct rb_node *iter = rb_last(&mm->free_tree[i]);
 		struct drm_buddy_block *tmp_block;
 
-		list_for_each_entry_reverse(tmp_block, &mm->free_list[i], link) {
-			if (block_incompatible(tmp_block, flags))
-				continue;
+		while (iter) {
+			tmp_block = rbtree_get_free_block(iter);
 
-			block = tmp_block;
-			break;
+			if (!block_incompatible(tmp_block, flags)) {
+				block = tmp_block;
+				break;
+			}
+
+			iter = rb_prev(iter);
 		}
 
 		if (!block)
@@ -667,7 +703,7 @@ get_maxblock(struct drm_buddy *mm, unsigned int order,
 }
 
 static struct drm_buddy_block *
-alloc_from_freelist(struct drm_buddy *mm,
+alloc_from_freetree(struct drm_buddy *mm,
 		    unsigned int order,
 		    unsigned long flags)
 {
@@ -682,14 +718,18 @@ alloc_from_freelist(struct drm_buddy *mm,
 			tmp = drm_buddy_block_order(block);
 	} else {
 		for (tmp = order; tmp <= mm->max_order; ++tmp) {
+			struct rb_node *iter = rb_last(&mm->free_tree[tmp]);
 			struct drm_buddy_block *tmp_block;
 
-			list_for_each_entry_reverse(tmp_block, &mm->free_list[tmp], link) {
-				if (block_incompatible(tmp_block, flags))
-					continue;
+			while (iter) {
+				tmp_block = rbtree_get_free_block(iter);
 
-				block = tmp_block;
-				break;
+				if (!block_incompatible(tmp_block, flags)) {
+					block = tmp_block;
+					break;
+				}
+
+				iter = rb_prev(iter);
 			}
 
 			if (block)
@@ -700,13 +740,9 @@ alloc_from_freelist(struct drm_buddy *mm,
 	if (!block) {
 		/* Fallback method */
 		for (tmp = order; tmp <= mm->max_order; ++tmp) {
-			if (!list_empty(&mm->free_list[tmp])) {
-				block = list_last_entry(&mm->free_list[tmp],
-							struct drm_buddy_block,
-							link);
-				if (block)
-					break;
-			}
+			block = rbtree_last_entry(mm, tmp);
+			if (block)
+				break;
 		}
 
 		if (!block)
@@ -771,7 +807,7 @@ static int __alloc_range(struct drm_buddy *mm,
 
 		if (contains(start, end, block_start, block_end)) {
 			if (drm_buddy_block_is_free(block)) {
-				mark_allocated(block);
+				mark_allocated(mm, block);
 				total_allocated += drm_buddy_block_size(mm, block);
 				mm->avail -= drm_buddy_block_size(mm, block);
 				if (drm_buddy_block_is_clear(block))
@@ -849,8 +885,8 @@ static int __alloc_contig_try_harder(struct drm_buddy *mm,
 {
 	u64 rhs_offset, lhs_offset, lhs_size, filled;
 	struct drm_buddy_block *block;
-	struct list_head *list;
 	LIST_HEAD(blocks_lhs);
+	struct rb_node *iter;
 	unsigned long pages;
 	unsigned int order;
 	u64 modify_size;
@@ -862,11 +898,14 @@ static int __alloc_contig_try_harder(struct drm_buddy *mm,
 	if (order == 0)
 		return -ENOSPC;
 
-	list = &mm->free_list[order];
-	if (list_empty(list))
+	if (rbtree_is_empty(mm, order))
 		return -ENOSPC;
 
-	list_for_each_entry_reverse(block, list, link) {
+	iter = rb_last(&mm->free_tree[order]);
+
+	while (iter) {
+		block = rbtree_get_free_block(iter);
+
 		/* Allocate blocks traversing RHS */
 		rhs_offset = drm_buddy_block_offset(block);
 		err =  __drm_buddy_alloc_range(mm, rhs_offset, size,
@@ -891,6 +930,8 @@ static int __alloc_contig_try_harder(struct drm_buddy *mm,
 		}
 		/* Free blocks for the next iteration */
 		drm_buddy_free_list_internal(mm, blocks);
+
+		iter = rb_prev(iter);
 	}
 
 	return -ENOSPC;
@@ -976,7 +1017,7 @@ int drm_buddy_block_trim(struct drm_buddy *mm,
 	list_add(&block->tmp_link, &dfs);
 	err =  __alloc_range(mm, &dfs, new_start, new_size, blocks, NULL);
 	if (err) {
-		mark_allocated(block);
+		mark_allocated(mm, block);
 		mm->avail -= drm_buddy_block_size(mm, block);
 		if (drm_buddy_block_is_clear(block))
 			mm->clear_avail -= drm_buddy_block_size(mm, block);
@@ -999,8 +1040,8 @@ __drm_buddy_alloc_blocks(struct drm_buddy *mm,
 		return  __drm_buddy_alloc_range_bias(mm, start, end,
 						     order, flags);
 	else
-		/* Allocate from freelist */
-		return alloc_from_freelist(mm, order, flags);
+		/* Allocate from freetree */
+		return alloc_from_freetree(mm, order, flags);
 }
 
 /**
@@ -1017,8 +1058,8 @@ __drm_buddy_alloc_blocks(struct drm_buddy *mm,
  * alloc_range_bias() called on range limitations, which traverses
  * the tree and returns the desired block.
  *
- * alloc_from_freelist() called when *no* range restrictions
- * are enforced, which picks the block from the freelist.
+ * alloc_from_freetree() called when *no* range restrictions
+ * are enforced, which picks the block from the freetree.
  *
  * Returns:
  * 0 on success, error code on failure.
@@ -1120,7 +1161,7 @@ int drm_buddy_alloc_blocks(struct drm_buddy *mm,
 			}
 		} while (1);
 
-		mark_allocated(block);
+		mark_allocated(mm, block);
 		mm->avail -= drm_buddy_block_size(mm, block);
 		if (drm_buddy_block_is_clear(block))
 			mm->clear_avail -= drm_buddy_block_size(mm, block);
@@ -1201,10 +1242,10 @@ void drm_buddy_print(struct drm_buddy *mm, struct drm_printer *p)
 		   mm->chunk_size >> 10, mm->size >> 20, mm->avail >> 20, mm->clear_avail >> 20);
 
 	for (order = mm->max_order; order >= 0; order--) {
-		struct drm_buddy_block *block;
+		struct drm_buddy_block *block, *tmp;
 		u64 count = 0, free;
 
-		list_for_each_entry(block, &mm->free_list[order], link) {
+		rbtree_postorder_for_each_entry_safe(block, tmp, &mm->free_tree[order], rb) {
 			BUG_ON(!drm_buddy_block_is_free(block));
 			count++;
 		}
diff --git a/include/drm/drm_buddy.h b/include/drm/drm_buddy.h
index 513837632b7d..9ee105d4309f 100644
--- a/include/drm/drm_buddy.h
+++ b/include/drm/drm_buddy.h
@@ -10,6 +10,7 @@
 #include <linux/list.h>
 #include <linux/slab.h>
 #include <linux/sched.h>
+#include <linux/rbtree.h>
 
 #include <drm/drm_print.h>
 
@@ -53,7 +54,11 @@ struct drm_buddy_block {
 	 * a list, if so desired. As soon as the block is freed with
 	 * drm_buddy_free* ownership is given back to the mm.
 	 */
-	struct list_head link;
+	union {
+		struct rb_node rb;
+		struct list_head link;
+	};
+
 	struct list_head tmp_link;
 };
 
@@ -68,7 +73,7 @@ struct drm_buddy_block {
  */
 struct drm_buddy {
 	/* Maintain a free list for each order. */
-	struct list_head *free_list;
+	struct rb_root *free_tree;
 
 	/*
 	 * Maintain explicit binary tree(s) to track the allocation of the
@@ -94,7 +99,7 @@ struct drm_buddy {
 };
 
 static inline u64
-drm_buddy_block_offset(struct drm_buddy_block *block)
+drm_buddy_block_offset(const struct drm_buddy_block *block)
 {
 	return block->header & DRM_BUDDY_HEADER_OFFSET;
 }
-- 
cgit v1.2.3


From d4cd665c98c144dd6ad5d66d30396e13d23118c9 Mon Sep 17 00:00:00 2001
From: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Date: Mon, 6 Oct 2025 15:21:23 +0530
Subject: drm/buddy: Separate clear and dirty free block trees

Maintain two separate RB trees per order - one for clear (zeroed) blocks
and another for dirty (uncleared) blocks. This separation improves
code clarity and makes it more obvious which tree is being searched
during allocation. It also improves scalability and efficiency when
searching for a specific type of block, avoiding unnecessary checks
and making the allocator more predictable under fragmentation.

The changes have been validated using the existing drm_buddy_test
KUnit test cases, along with selected graphics workloads,
to ensure correctness and avoid regressions.

v2: Missed adding the suggested-by tag. Added it in v2.

v3(Matthew):
  - Remove the double underscores from the internal functions.
  - Rename the internal functions to have less generic names.
  - Fix the error handling code.
  - Pass tree argument for the tree macro.
  - Use the existing dirty/free bit instead of new tree field.
  - Make free_trees[] instead of clear_tree and dirty_tree for
    more cleaner approach.

v4:
  - A bug was reported by Intel CI and it is fixed by
    Matthew Auld.
  - Replace the get_root function with
    &mm->free_trees[tree][order] (Matthew)
  - Remove the unnecessary rbtree_is_empty() check (Matthew)
  - Remove the unnecessary get_tree_for_flags() function.
  - Rename get_tree_for_block() name with get_block_tree() for more
    clarity.

v5(Jani Nikula):
  - Don't use static inline in .c files.
  - enum free_tree and enumerator names are quite generic for a header
    and usage and the whole enum should be an implementation detail.

v6:
  - Rewrite the __force_merge() function using the rb_last() and rb_prev().

v7(Matthew):
  - Replace the open-coded tree iteration for loops with the
    for_each_free_tree() macro throughout the code.
  - Fixed out_free_roots to prevent double decrement of i,
    addressing potential crash.
  - Replaced enum drm_buddy_free_tree with unsigned int
    in for_each_free_tree loops.

Cc: stable@vger.kernel.org
Fixes: a68c7eaa7a8f ("drm/amdgpu: Enable clear page functionality")
Signed-off-by: Arunpravin Paneer Selvam <Arunpravin.PaneerSelvam@amd.com>
Suggested-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Closes: https://gitlab.freedesktop.org/drm/amd/-/issues/4260
Link: https://lore.kernel.org/r/20251006095124.1663-2-Arunpravin.PaneerSelvam@amd.com
---
 drivers/gpu/drm/drm_buddy.c | 333 +++++++++++++++++++++++++-------------------
 include/drm/drm_buddy.h     |   2 +-
 2 files changed, 188 insertions(+), 147 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_buddy.c b/drivers/gpu/drm/drm_buddy.c
index c87210a06c31..f2c92902e4a3 100644
--- a/drivers/gpu/drm/drm_buddy.c
+++ b/drivers/gpu/drm/drm_buddy.c
@@ -12,9 +12,16 @@
 
 #include <drm/drm_buddy.h>
 
+enum drm_buddy_free_tree {
+	DRM_BUDDY_CLEAR_TREE = 0,
+	DRM_BUDDY_DIRTY_TREE,
+	DRM_BUDDY_MAX_FREE_TREES,
+};
+
 static struct kmem_cache *slab_blocks;
 
-#define rbtree_get_free_block(node) rb_entry((node), struct drm_buddy_block, rb)
+#define for_each_free_tree(tree) \
+	for ((tree) = 0; (tree) < DRM_BUDDY_MAX_FREE_TREES; (tree)++)
 
 static struct drm_buddy_block *drm_block_alloc(struct drm_buddy *mm,
 					       struct drm_buddy_block *parent,
@@ -45,6 +52,30 @@ static void drm_block_free(struct drm_buddy *mm,
 	kmem_cache_free(slab_blocks, block);
 }
 
+static enum drm_buddy_free_tree
+get_block_tree(struct drm_buddy_block *block)
+{
+	return drm_buddy_block_is_clear(block) ?
+	       DRM_BUDDY_CLEAR_TREE : DRM_BUDDY_DIRTY_TREE;
+}
+
+static struct drm_buddy_block *
+rbtree_get_free_block(const struct rb_node *node)
+{
+	return node ? rb_entry(node, struct drm_buddy_block, rb) : NULL;
+}
+
+static struct drm_buddy_block *
+rbtree_last_free_block(struct rb_root *root)
+{
+	return rbtree_get_free_block(rb_last(root));
+}
+
+static bool rbtree_is_empty(struct rb_root *root)
+{
+	return RB_EMPTY_ROOT(root);
+}
+
 static bool drm_buddy_block_offset_less(const struct drm_buddy_block *block,
 					const struct drm_buddy_block *node)
 {
@@ -59,37 +90,28 @@ static bool rbtree_block_offset_less(struct rb_node *block,
 }
 
 static void rbtree_insert(struct drm_buddy *mm,
-			  struct drm_buddy_block *block)
+			  struct drm_buddy_block *block,
+			  enum drm_buddy_free_tree tree)
 {
 	rb_add(&block->rb,
-	       &mm->free_tree[drm_buddy_block_order(block)],
+	       &mm->free_trees[tree][drm_buddy_block_order(block)],
 	       rbtree_block_offset_less);
 }
 
 static void rbtree_remove(struct drm_buddy *mm,
 			  struct drm_buddy_block *block)
 {
+	unsigned int order = drm_buddy_block_order(block);
+	enum drm_buddy_free_tree tree;
 	struct rb_root *root;
 
-	root = &mm->free_tree[drm_buddy_block_order(block)];
-	rb_erase(&block->rb, root);
+	tree = get_block_tree(block);
+	root = &mm->free_trees[tree][order];
 
+	rb_erase(&block->rb, root);
 	RB_CLEAR_NODE(&block->rb);
 }
 
-static struct drm_buddy_block *
-rbtree_last_entry(struct drm_buddy *mm, unsigned int order)
-{
-	struct rb_node *node = rb_last(&mm->free_tree[order]);
-
-	return node ? rb_entry(node, struct drm_buddy_block, rb) : NULL;
-}
-
-static bool rbtree_is_empty(struct drm_buddy *mm, unsigned int order)
-{
-	return RB_EMPTY_ROOT(&mm->free_tree[order]);
-}
-
 static void clear_reset(struct drm_buddy_block *block)
 {
 	block->header &= ~DRM_BUDDY_HEADER_CLEAR;
@@ -112,10 +134,13 @@ static void mark_allocated(struct drm_buddy *mm,
 static void mark_free(struct drm_buddy *mm,
 		      struct drm_buddy_block *block)
 {
+	enum drm_buddy_free_tree tree;
+
 	block->header &= ~DRM_BUDDY_HEADER_STATE;
 	block->header |= DRM_BUDDY_FREE;
 
-	rbtree_insert(mm, block);
+	tree = get_block_tree(block);
+	rbtree_insert(mm, block, tree);
 }
 
 static void mark_split(struct drm_buddy *mm,
@@ -201,7 +226,7 @@ static int __force_merge(struct drm_buddy *mm,
 			 u64 end,
 			 unsigned int min_order)
 {
-	unsigned int order;
+	unsigned int tree, order;
 	int i;
 
 	if (!min_order)
@@ -210,45 +235,48 @@ static int __force_merge(struct drm_buddy *mm,
 	if (min_order > mm->max_order)
 		return -EINVAL;
 
-	for (i = min_order - 1; i >= 0; i--) {
-		struct rb_root *root = &mm->free_tree[i];
-		struct rb_node *iter;
+	for_each_free_tree(tree) {
+		for (i = min_order - 1; i >= 0; i--) {
+			struct rb_node *iter = rb_last(&mm->free_trees[tree][i]);
 
-		iter = rb_last(root);
-
-		while (iter) {
-			struct drm_buddy_block *block, *buddy;
-			u64 block_start, block_end;
+			while (iter) {
+				struct drm_buddy_block *block, *buddy;
+				u64 block_start, block_end;
 
-			block = rbtree_get_free_block(iter);
-			iter = rb_prev(iter);
+				block = rbtree_get_free_block(iter);
+				iter = rb_prev(iter);
 
-			if (!block || !block->parent)
-				continue;
+				if (!block || !block->parent)
+					continue;
 
-			block_start = drm_buddy_block_offset(block);
-			block_end = block_start + drm_buddy_block_size(mm, block) - 1;
+				block_start = drm_buddy_block_offset(block);
+				block_end = block_start + drm_buddy_block_size(mm, block) - 1;
 
-			if (!contains(start, end, block_start, block_end))
-				continue;
+				if (!contains(start, end, block_start, block_end))
+					continue;
 
-			buddy = __get_buddy(block);
-			if (!drm_buddy_block_is_free(buddy))
-				continue;
+				buddy = __get_buddy(block);
+				if (!drm_buddy_block_is_free(buddy))
+					continue;
 
-			WARN_ON(drm_buddy_block_is_clear(block) ==
-				drm_buddy_block_is_clear(buddy));
+				WARN_ON(drm_buddy_block_is_clear(block) ==
+					drm_buddy_block_is_clear(buddy));
 
-			if (iter == &buddy->rb)
-				iter = rb_prev(iter);
+				/*
+				 * Advance to the next node when the current node is the buddy,
+				 * as freeing the block will also remove its buddy from the tree.
+				 */
+				if (iter == &buddy->rb)
+					iter = rb_prev(iter);
 
-			rbtree_remove(mm, block);
-			if (drm_buddy_block_is_clear(block))
-				mm->clear_avail -= drm_buddy_block_size(mm, block);
+				rbtree_remove(mm, block);
+				if (drm_buddy_block_is_clear(block))
+					mm->clear_avail -= drm_buddy_block_size(mm, block);
 
-			order = __drm_buddy_free(mm, block, true);
-			if (order >= min_order)
-				return 0;
+				order = __drm_buddy_free(mm, block, true);
+				if (order >= min_order)
+					return 0;
+			}
 		}
 	}
 
@@ -269,7 +297,7 @@ static int __force_merge(struct drm_buddy *mm,
  */
 int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 {
-	unsigned int i;
+	unsigned int i, j, root_count = 0;
 	u64 offset = 0;
 
 	if (size < chunk_size)
@@ -291,14 +319,22 @@ int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 
 	BUG_ON(mm->max_order > DRM_BUDDY_MAX_ORDER);
 
-	mm->free_tree = kmalloc_array(mm->max_order + 1,
-				      sizeof(struct rb_root),
-				      GFP_KERNEL);
-	if (!mm->free_tree)
+	mm->free_trees = kmalloc_array(DRM_BUDDY_MAX_FREE_TREES,
+				       sizeof(*mm->free_trees),
+				       GFP_KERNEL);
+	if (!mm->free_trees)
 		return -ENOMEM;
 
-	for (i = 0; i <= mm->max_order; ++i)
-		mm->free_tree[i] = RB_ROOT;
+	for_each_free_tree(i) {
+		mm->free_trees[i] = kmalloc_array(mm->max_order + 1,
+						  sizeof(struct rb_root),
+						  GFP_KERNEL);
+		if (!mm->free_trees[i])
+			goto out_free_tree;
+
+		for (j = 0; j <= mm->max_order; ++j)
+			mm->free_trees[i][j] = RB_ROOT;
+	}
 
 	mm->n_roots = hweight64(size);
 
@@ -308,8 +344,6 @@ int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 	if (!mm->roots)
 		goto out_free_tree;
 
-	i = 0;
-
 	/*
 	 * Split into power-of-two blocks, in case we are given a size that is
 	 * not itself a power-of-two.
@@ -328,24 +362,26 @@ int drm_buddy_init(struct drm_buddy *mm, u64 size, u64 chunk_size)
 
 		mark_free(mm, root);
 
-		BUG_ON(i > mm->max_order);
+		BUG_ON(root_count > mm->max_order);
 		BUG_ON(drm_buddy_block_size(mm, root) < chunk_size);
 
-		mm->roots[i] = root;
+		mm->roots[root_count] = root;
 
 		offset += root_size;
 		size -= root_size;
-		i++;
+		root_count++;
 	} while (size);
 
 	return 0;
 
 out_free_roots:
-	while (i--)
-		drm_block_free(mm, mm->roots[i]);
+	while (root_count--)
+		drm_block_free(mm, mm->roots[root_count]);
 	kfree(mm->roots);
 out_free_tree:
-	kfree(mm->free_tree);
+	while (i--)
+		kfree(mm->free_trees[i]);
+	kfree(mm->free_trees);
 	return -ENOMEM;
 }
 EXPORT_SYMBOL(drm_buddy_init);
@@ -381,8 +417,9 @@ void drm_buddy_fini(struct drm_buddy *mm)
 
 	WARN_ON(mm->avail != mm->size);
 
+	for_each_free_tree(i)
+		kfree(mm->free_trees[i]);
 	kfree(mm->roots);
-	kfree(mm->free_tree);
 }
 EXPORT_SYMBOL(drm_buddy_fini);
 
@@ -406,8 +443,7 @@ static int split_block(struct drm_buddy *mm,
 		return -ENOMEM;
 	}
 
-	mark_free(mm, block->left);
-	mark_free(mm, block->right);
+	mark_split(mm, block);
 
 	if (drm_buddy_block_is_clear(block)) {
 		mark_cleared(block->left);
@@ -415,7 +451,8 @@ static int split_block(struct drm_buddy *mm,
 		clear_reset(block);
 	}
 
-	mark_split(mm, block);
+	mark_free(mm, block->left);
+	mark_free(mm, block->right);
 
 	return 0;
 }
@@ -448,6 +485,7 @@ EXPORT_SYMBOL(drm_get_buddy);
  */
 void drm_buddy_reset_clear(struct drm_buddy *mm, bool is_clear)
 {
+	enum drm_buddy_free_tree src_tree, dst_tree;
 	u64 root_size, size, start;
 	unsigned int order;
 	int i;
@@ -462,19 +500,24 @@ void drm_buddy_reset_clear(struct drm_buddy *mm, bool is_clear)
 		size -= root_size;
 	}
 
+	src_tree = is_clear ? DRM_BUDDY_DIRTY_TREE : DRM_BUDDY_CLEAR_TREE;
+	dst_tree = is_clear ? DRM_BUDDY_CLEAR_TREE : DRM_BUDDY_DIRTY_TREE;
+
 	for (i = 0; i <= mm->max_order; ++i) {
+		struct rb_root *root = &mm->free_trees[src_tree][i];
 		struct drm_buddy_block *block, *tmp;
 
-		rbtree_postorder_for_each_entry_safe(block, tmp, &mm->free_tree[i], rb) {
-			if (is_clear != drm_buddy_block_is_clear(block)) {
-				if (is_clear) {
-					mark_cleared(block);
-					mm->clear_avail += drm_buddy_block_size(mm, block);
-				} else {
-					clear_reset(block);
-					mm->clear_avail -= drm_buddy_block_size(mm, block);
-				}
+		rbtree_postorder_for_each_entry_safe(block, tmp, root, rb) {
+			rbtree_remove(mm, block);
+			if (is_clear) {
+				mark_cleared(block);
+				mm->clear_avail += drm_buddy_block_size(mm, block);
+			} else {
+				clear_reset(block);
+				mm->clear_avail -= drm_buddy_block_size(mm, block);
 			}
+
+			rbtree_insert(mm, block, dst_tree);
 		}
 	}
 }
@@ -664,27 +707,17 @@ __drm_buddy_alloc_range_bias(struct drm_buddy *mm,
 }
 
 static struct drm_buddy_block *
-get_maxblock(struct drm_buddy *mm, unsigned int order,
-	     unsigned long flags)
+get_maxblock(struct drm_buddy *mm,
+	     unsigned int order,
+	     enum drm_buddy_free_tree tree)
 {
 	struct drm_buddy_block *max_block = NULL, *block = NULL;
+	struct rb_root *root;
 	unsigned int i;
 
 	for (i = order; i <= mm->max_order; ++i) {
-		struct rb_node *iter = rb_last(&mm->free_tree[i]);
-		struct drm_buddy_block *tmp_block;
-
-		while (iter) {
-			tmp_block = rbtree_get_free_block(iter);
-
-			if (!block_incompatible(tmp_block, flags)) {
-				block = tmp_block;
-				break;
-			}
-
-			iter = rb_prev(iter);
-		}
-
+		root = &mm->free_trees[tree][i];
+		block = rbtree_last_free_block(root);
 		if (!block)
 			continue;
 
@@ -708,39 +741,37 @@ alloc_from_freetree(struct drm_buddy *mm,
 		    unsigned long flags)
 {
 	struct drm_buddy_block *block = NULL;
+	struct rb_root *root;
+	enum drm_buddy_free_tree tree;
 	unsigned int tmp;
 	int err;
 
+	tree = (flags & DRM_BUDDY_CLEAR_ALLOCATION) ?
+		DRM_BUDDY_CLEAR_TREE : DRM_BUDDY_DIRTY_TREE;
+
 	if (flags & DRM_BUDDY_TOPDOWN_ALLOCATION) {
-		block = get_maxblock(mm, order, flags);
+		block = get_maxblock(mm, order, tree);
 		if (block)
 			/* Store the obtained block order */
 			tmp = drm_buddy_block_order(block);
 	} else {
 		for (tmp = order; tmp <= mm->max_order; ++tmp) {
-			struct rb_node *iter = rb_last(&mm->free_tree[tmp]);
-			struct drm_buddy_block *tmp_block;
-
-			while (iter) {
-				tmp_block = rbtree_get_free_block(iter);
-
-				if (!block_incompatible(tmp_block, flags)) {
-					block = tmp_block;
-					break;
-				}
-
-				iter = rb_prev(iter);
-			}
-
+			/* Get RB tree root for this order and tree */
+			root = &mm->free_trees[tree][tmp];
+			block = rbtree_last_free_block(root);
 			if (block)
 				break;
 		}
 	}
 
 	if (!block) {
-		/* Fallback method */
+		/* Try allocating from the other tree */
+		tree = (tree == DRM_BUDDY_CLEAR_TREE) ?
+			DRM_BUDDY_DIRTY_TREE : DRM_BUDDY_CLEAR_TREE;
+
 		for (tmp = order; tmp <= mm->max_order; ++tmp) {
-			block = rbtree_last_entry(mm, tmp);
+			root = &mm->free_trees[tree][tmp];
+			block = rbtree_last_free_block(root);
 			if (block)
 				break;
 		}
@@ -885,10 +916,9 @@ static int __alloc_contig_try_harder(struct drm_buddy *mm,
 {
 	u64 rhs_offset, lhs_offset, lhs_size, filled;
 	struct drm_buddy_block *block;
+	unsigned int tree, order;
 	LIST_HEAD(blocks_lhs);
-	struct rb_node *iter;
 	unsigned long pages;
-	unsigned int order;
 	u64 modify_size;
 	int err;
 
@@ -898,40 +928,45 @@ static int __alloc_contig_try_harder(struct drm_buddy *mm,
 	if (order == 0)
 		return -ENOSPC;
 
-	if (rbtree_is_empty(mm, order))
-		return -ENOSPC;
+	for_each_free_tree(tree) {
+		struct rb_root *root;
+		struct rb_node *iter;
 
-	iter = rb_last(&mm->free_tree[order]);
-
-	while (iter) {
-		block = rbtree_get_free_block(iter);
-
-		/* Allocate blocks traversing RHS */
-		rhs_offset = drm_buddy_block_offset(block);
-		err =  __drm_buddy_alloc_range(mm, rhs_offset, size,
-					       &filled, blocks);
-		if (!err || err != -ENOSPC)
-			return err;
-
-		lhs_size = max((size - filled), min_block_size);
-		if (!IS_ALIGNED(lhs_size, min_block_size))
-			lhs_size = round_up(lhs_size, min_block_size);
-
-		/* Allocate blocks traversing LHS */
-		lhs_offset = drm_buddy_block_offset(block) - lhs_size;
-		err =  __drm_buddy_alloc_range(mm, lhs_offset, lhs_size,
-					       NULL, &blocks_lhs);
-		if (!err) {
-			list_splice(&blocks_lhs, blocks);
-			return 0;
-		} else if (err != -ENOSPC) {
+		root = &mm->free_trees[tree][order];
+		if (rbtree_is_empty(root))
+			continue;
+
+		iter = rb_last(root);
+		while (iter) {
+			block = rbtree_get_free_block(iter);
+
+			/* Allocate blocks traversing RHS */
+			rhs_offset = drm_buddy_block_offset(block);
+			err =  __drm_buddy_alloc_range(mm, rhs_offset, size,
+						       &filled, blocks);
+			if (!err || err != -ENOSPC)
+				return err;
+
+			lhs_size = max((size - filled), min_block_size);
+			if (!IS_ALIGNED(lhs_size, min_block_size))
+				lhs_size = round_up(lhs_size, min_block_size);
+
+			/* Allocate blocks traversing LHS */
+			lhs_offset = drm_buddy_block_offset(block) - lhs_size;
+			err =  __drm_buddy_alloc_range(mm, lhs_offset, lhs_size,
+						       NULL, &blocks_lhs);
+			if (!err) {
+				list_splice(&blocks_lhs, blocks);
+				return 0;
+			} else if (err != -ENOSPC) {
+				drm_buddy_free_list_internal(mm, blocks);
+				return err;
+			}
+			/* Free blocks for the next iteration */
 			drm_buddy_free_list_internal(mm, blocks);
-			return err;
-		}
-		/* Free blocks for the next iteration */
-		drm_buddy_free_list_internal(mm, blocks);
 
-		iter = rb_prev(iter);
+			iter = rb_prev(iter);
+		}
 	}
 
 	return -ENOSPC;
@@ -1243,11 +1278,17 @@ void drm_buddy_print(struct drm_buddy *mm, struct drm_printer *p)
 
 	for (order = mm->max_order; order >= 0; order--) {
 		struct drm_buddy_block *block, *tmp;
+		struct rb_root *root;
 		u64 count = 0, free;
+		unsigned int tree;
+
+		for_each_free_tree(tree) {
+			root = &mm->free_trees[tree][order];
 
-		rbtree_postorder_for_each_entry_safe(block, tmp, &mm->free_tree[order], rb) {
-			BUG_ON(!drm_buddy_block_is_free(block));
-			count++;
+			rbtree_postorder_for_each_entry_safe(block, tmp, root, rb) {
+				BUG_ON(!drm_buddy_block_is_free(block));
+				count++;
+			}
 		}
 
 		drm_printf(p, "order-%2d ", order);
diff --git a/include/drm/drm_buddy.h b/include/drm/drm_buddy.h
index 9ee105d4309f..d7891d08f67a 100644
--- a/include/drm/drm_buddy.h
+++ b/include/drm/drm_buddy.h
@@ -73,7 +73,7 @@ struct drm_buddy_block {
  */
 struct drm_buddy {
 	/* Maintain a free list for each order. */
-	struct rb_root *free_tree;
+	struct rb_root **free_trees;
 
 	/*
 	 * Maintain explicit binary tree(s) to track the allocation of the
-- 
cgit v1.2.3


From 1226cd7c7686aee9363ec39a8a80292e27216c6b Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Wed, 8 Oct 2025 15:11:42 +0200
Subject: drm/atomic: Change state pointers to a more meaningful name

The state pointer found in the struct drm_atomic_state internals for
most object is a bit ambiguous, and confusing when those internals also
have old state and new state.

After the recent cleanups, the state pointer only use is to point to the
state we need to free when destroying the atomic state.

We can thus rename it something less ambiguous, and hopefully more
meaningful.

Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://lore.kernel.org/r/20251008-drm-rename-state-v2-1-49b490b2676a@kernel.org
Signed-off-by: Maxime Ripard <mripard@kernel.org>
---
 drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c |  4 ++--
 drivers/gpu/drm/drm_atomic.c                      | 24 +++++++++++------------
 drivers/gpu/drm/drm_atomic_helper.c               |  8 ++++----
 include/drm/drm_atomic.h                          | 16 +++++++--------
 4 files changed, 26 insertions(+), 26 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
index 0d03e324d5b9..fe41494635c5 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm.c
@@ -12450,7 +12450,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 				int j = state->num_private_objs-1;
 
 				dm_atomic_destroy_state(obj,
-						state->private_objs[i].state);
+						state->private_objs[i].state_to_destroy);
 
 				/* If i is not at the end of the array then the
 				 * last element needs to be moved to where i was
@@ -12461,7 +12461,7 @@ static int amdgpu_dm_atomic_check(struct drm_device *dev,
 						state->private_objs[j];
 
 				state->private_objs[j].ptr = NULL;
-				state->private_objs[j].state = NULL;
+				state->private_objs[j].state_to_destroy = NULL;
 				state->private_objs[j].old_state = NULL;
 				state->private_objs[j].new_state = NULL;
 
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 0fda567c3900..be2cb6e43cb0 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -207,9 +207,9 @@ void drm_atomic_state_default_clear(struct drm_atomic_state *state)
 			continue;
 
 		connector->funcs->atomic_destroy_state(connector,
-						       state->connectors[i].state);
+						       state->connectors[i].state_to_destroy);
 		state->connectors[i].ptr = NULL;
-		state->connectors[i].state = NULL;
+		state->connectors[i].state_to_destroy = NULL;
 		state->connectors[i].old_state = NULL;
 		state->connectors[i].new_state = NULL;
 		drm_connector_put(connector);
@@ -222,10 +222,10 @@ void drm_atomic_state_default_clear(struct drm_atomic_state *state)
 			continue;
 
 		crtc->funcs->atomic_destroy_state(crtc,
-						  state->crtcs[i].state);
+						  state->crtcs[i].state_to_destroy);
 
 		state->crtcs[i].ptr = NULL;
-		state->crtcs[i].state = NULL;
+		state->crtcs[i].state_to_destroy = NULL;
 		state->crtcs[i].old_state = NULL;
 		state->crtcs[i].new_state = NULL;
 
@@ -242,9 +242,9 @@ void drm_atomic_state_default_clear(struct drm_atomic_state *state)
 			continue;
 
 		plane->funcs->atomic_destroy_state(plane,
-						   state->planes[i].state);
+						   state->planes[i].state_to_destroy);
 		state->planes[i].ptr = NULL;
-		state->planes[i].state = NULL;
+		state->planes[i].state_to_destroy = NULL;
 		state->planes[i].old_state = NULL;
 		state->planes[i].new_state = NULL;
 	}
@@ -253,9 +253,9 @@ void drm_atomic_state_default_clear(struct drm_atomic_state *state)
 		struct drm_private_obj *obj = state->private_objs[i].ptr;
 
 		obj->funcs->atomic_destroy_state(obj,
-						 state->private_objs[i].state);
+						 state->private_objs[i].state_to_destroy);
 		state->private_objs[i].ptr = NULL;
-		state->private_objs[i].state = NULL;
+		state->private_objs[i].state_to_destroy = NULL;
 		state->private_objs[i].old_state = NULL;
 		state->private_objs[i].new_state = NULL;
 	}
@@ -361,7 +361,7 @@ drm_atomic_get_crtc_state(struct drm_atomic_state *state,
 	if (!crtc_state)
 		return ERR_PTR(-ENOMEM);
 
-	state->crtcs[index].state = crtc_state;
+	state->crtcs[index].state_to_destroy = crtc_state;
 	state->crtcs[index].old_state = crtc->state;
 	state->crtcs[index].new_state = crtc_state;
 	state->crtcs[index].ptr = crtc;
@@ -546,7 +546,7 @@ drm_atomic_get_plane_state(struct drm_atomic_state *state,
 	if (!plane_state)
 		return ERR_PTR(-ENOMEM);
 
-	state->planes[index].state = plane_state;
+	state->planes[index].state_to_destroy = plane_state;
 	state->planes[index].ptr = plane;
 	state->planes[index].old_state = plane->state;
 	state->planes[index].new_state = plane_state;
@@ -858,7 +858,7 @@ drm_atomic_get_private_obj_state(struct drm_atomic_state *state,
 	if (!obj_state)
 		return ERR_PTR(-ENOMEM);
 
-	state->private_objs[index].state = obj_state;
+	state->private_objs[index].state_to_destroy = obj_state;
 	state->private_objs[index].old_state = obj->state;
 	state->private_objs[index].new_state = obj_state;
 	state->private_objs[index].ptr = obj;
@@ -1161,7 +1161,7 @@ drm_atomic_get_connector_state(struct drm_atomic_state *state,
 		return ERR_PTR(-ENOMEM);
 
 	drm_connector_get(connector);
-	state->connectors[index].state = connector_state;
+	state->connectors[index].state_to_destroy = connector_state;
 	state->connectors[index].old_state = connector->state;
 	state->connectors[index].new_state = connector_state;
 	state->connectors[index].ptr = connector;
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index d5ebe6ea0acb..5a473a274ff0 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -3236,7 +3236,7 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state,
 		old_conn_state->state = state;
 		new_conn_state->state = NULL;
 
-		state->connectors[i].state = old_conn_state;
+		state->connectors[i].state_to_destroy = old_conn_state;
 		connector->state = new_conn_state;
 	}
 
@@ -3246,7 +3246,7 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state,
 		old_crtc_state->state = state;
 		new_crtc_state->state = NULL;
 
-		state->crtcs[i].state = old_crtc_state;
+		state->crtcs[i].state_to_destroy = old_crtc_state;
 		crtc->state = new_crtc_state;
 
 		if (new_crtc_state->commit) {
@@ -3266,7 +3266,7 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state,
 		old_plane_state->state = state;
 		new_plane_state->state = NULL;
 
-		state->planes[i].state = old_plane_state;
+		state->planes[i].state_to_destroy = old_plane_state;
 		plane->state = new_plane_state;
 	}
 	drm_panic_unlock(state->dev, flags);
@@ -3277,7 +3277,7 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state,
 		old_obj_state->state = state;
 		new_obj_state->state = NULL;
 
-		state->private_objs[i].state = old_obj_state;
+		state->private_objs[i].state_to_destroy = old_obj_state;
 		obj->state = new_obj_state;
 	}
 
diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index c8ab2163bf65..155e82f87e4d 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -161,7 +161,7 @@ struct __drm_planes_state {
 	struct drm_plane *ptr;
 
 	/**
-	 * @state:
+	 * @state_to_destroy:
 	 *
 	 * Used to track the @drm_plane_state we will need to free when
 	 * tearing down the associated &drm_atomic_state in
@@ -173,7 +173,7 @@ struct __drm_planes_state {
 	 * the same state than @new_state. After a commit, it points to
 	 * the same state than @old_state.
 	 */
-	struct drm_plane_state *state;
+	struct drm_plane_state *state_to_destroy;
 
 	struct drm_plane_state *old_state, *new_state;
 };
@@ -182,7 +182,7 @@ struct __drm_crtcs_state {
 	struct drm_crtc *ptr;
 
 	/**
-	 * @state:
+	 * @state_to_destroy:
 	 *
 	 * Used to track the @drm_crtc_state we will need to free when
 	 * tearing down the associated &drm_atomic_state in
@@ -194,7 +194,7 @@ struct __drm_crtcs_state {
 	 * the same state than @new_state. After a commit, it points to
 	 * the same state than @old_state.
 	 */
-	struct drm_crtc_state *state;
+	struct drm_crtc_state *state_to_destroy;
 
 	struct drm_crtc_state *old_state, *new_state;
 
@@ -216,7 +216,7 @@ struct __drm_connnectors_state {
 	struct drm_connector *ptr;
 
 	/**
-	 * @state:
+	 * @state_to_destroy:
 	 *
 	 * Used to track the @drm_connector_state we will need to free
 	 * when tearing down the associated &drm_atomic_state in
@@ -228,7 +228,7 @@ struct __drm_connnectors_state {
 	 * the same state than @new_state. After a commit, it points to
 	 * the same state than @old_state.
 	 */
-	struct drm_connector_state *state;
+	struct drm_connector_state *state_to_destroy;
 
 	struct drm_connector_state *old_state, *new_state;
 
@@ -393,7 +393,7 @@ struct __drm_private_objs_state {
 	struct drm_private_obj *ptr;
 
 	/**
-	 * @state:
+	 * @state_to_destroy:
 	 *
 	 * Used to track the @drm_private_state we will need to free
 	 * when tearing down the associated &drm_atomic_state in
@@ -405,7 +405,7 @@ struct __drm_private_objs_state {
 	 * the same state than @new_state. After a commit, it points to
 	 * the same state than @old_state.
 	 */
-	struct drm_private_state *state;
+	struct drm_private_state *state_to_destroy;
 
 	struct drm_private_state *old_state, *new_state;
 };
-- 
cgit v1.2.3


From 78de8f87668334a3bfadad52e5142fc19dad1807 Mon Sep 17 00:00:00 2001
From: Matt Roper <matthew.d.roper@intel.com>
Date: Mon, 13 Oct 2025 13:09:58 -0700
Subject: drm/xe: Handle Wa_22010954014 and Wa_14022085890 as device
 workarounds

When Wa_22010954014 and Wa_14022085890 were first implemented, we didn't
have a device workaround infrastructure so we hacked them into the GT
workaround list.  Now that we have proper device workaround support,
move them to the proper place.  Note that Wa_14022085890 specifically
applies to BMG-G21 platforms, so this requires defining a BMG
subplatform to capture the correct subset of device IDs.

Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Link: https://lore.kernel.org/r/20251013200944.2499947-40-matthew.d.roper@intel.com
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
---
 drivers/gpu/drm/xe/xe_device_wa_oob.rules | 2 ++
 drivers/gpu/drm/xe/xe_guc_pc.c            | 3 ++-
 drivers/gpu/drm/xe/xe_pci.c               | 6 ++++++
 drivers/gpu/drm/xe/xe_platform_types.h    | 1 +
 drivers/gpu/drm/xe/xe_wa.c                | 2 +-
 drivers/gpu/drm/xe/xe_wa_oob.rules        | 5 -----
 include/drm/intel/pciids.h                | 7 +++++--
 7 files changed, 17 insertions(+), 9 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/xe/xe_device_wa_oob.rules b/drivers/gpu/drm/xe/xe_device_wa_oob.rules
index 3cc93f0e77f8..55ba01bc8f38 100644
--- a/drivers/gpu/drm/xe/xe_device_wa_oob.rules
+++ b/drivers/gpu/drm/xe/xe_device_wa_oob.rules
@@ -1,3 +1,5 @@
+22010954014	PLATFORM(DG2)
 15015404425     PLATFORM(LUNARLAKE)
 		PLATFORM(PANTHERLAKE)
 22019338487_display	PLATFORM(LUNARLAKE)
+14022085890	SUBPLATFORM(BATTLEMAGE, G21)
diff --git a/drivers/gpu/drm/xe/xe_guc_pc.c b/drivers/gpu/drm/xe/xe_guc_pc.c
index 3c0feb50a1e2..ff22235857f8 100644
--- a/drivers/gpu/drm/xe/xe_guc_pc.c
+++ b/drivers/gpu/drm/xe/xe_guc_pc.c
@@ -14,6 +14,7 @@
 
 #include <drm/drm_managed.h>
 #include <drm/drm_print.h>
+#include <generated/xe_device_wa_oob.h>
 #include <generated/xe_wa_oob.h>
 
 #include "abi/guc_actions_slpc_abi.h"
@@ -886,7 +887,7 @@ static int pc_adjust_freq_bounds(struct xe_guc_pc *pc)
 	if (pc_get_min_freq(pc) > pc->rp0_freq)
 		ret = pc_set_min_freq(pc, pc->rp0_freq);
 
-	if (XE_GT_WA(tile->primary_gt, 14022085890))
+	if (XE_DEVICE_WA(tile_to_xe(tile), 14022085890))
 		ret = pc_set_min_freq(pc, max(BMG_MIN_FREQ, pc_get_min_freq(pc)));
 
 out:
diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
index 0fea2ddb3c99..368181b4d0b3 100644
--- a/drivers/gpu/drm/xe/xe_pci.c
+++ b/drivers/gpu/drm/xe/xe_pci.c
@@ -335,6 +335,8 @@ static const struct xe_device_desc lnl_desc = {
 	.vm_max_level = 4,
 };
 
+static const u16 bmg_g21_ids[] = { INTEL_BMG_G21_IDS(NOP), 0 };
+
 static const struct xe_device_desc bmg_desc = {
 	DGFX_FEATURES,
 	PLATFORM(BATTLEMAGE),
@@ -349,6 +351,10 @@ static const struct xe_device_desc bmg_desc = {
 	.has_sriov = true,
 	.max_gt_per_tile = 2,
 	.needs_scratch = true,
+	.subplatforms = (const struct xe_subplatform_desc[]) {
+		{ XE_SUBPLATFORM_BATTLEMAGE_G21, "G21", bmg_g21_ids },
+		{ }
+	},
 	.va_bits = 48,
 	.vm_max_level = 4,
 };
diff --git a/drivers/gpu/drm/xe/xe_platform_types.h b/drivers/gpu/drm/xe/xe_platform_types.h
index d08574c4cdb8..3e332214c7bb 100644
--- a/drivers/gpu/drm/xe/xe_platform_types.h
+++ b/drivers/gpu/drm/xe/xe_platform_types.h
@@ -34,6 +34,7 @@ enum xe_subplatform {
 	XE_SUBPLATFORM_DG2_G10,
 	XE_SUBPLATFORM_DG2_G11,
 	XE_SUBPLATFORM_DG2_G12,
+	XE_SUBPLATFORM_BATTLEMAGE_G21,
 };
 
 #endif
diff --git a/drivers/gpu/drm/xe/xe_wa.c b/drivers/gpu/drm/xe/xe_wa.c
index c60159a13001..aa1b69f48f6f 100644
--- a/drivers/gpu/drm/xe/xe_wa.c
+++ b/drivers/gpu/drm/xe/xe_wa.c
@@ -1138,6 +1138,6 @@ void xe_wa_apply_tile_workarounds(struct xe_tile *tile)
 	if (IS_SRIOV_VF(tile->xe))
 		return;
 
-	if (XE_GT_WA(tile->primary_gt, 22010954014))
+	if (XE_DEVICE_WA(tile->xe, 22010954014))
 		xe_mmio_rmw32(mmio, XEHP_CLOCK_GATE_DIS, 0, SGSI_SIDECLK_DIS);
 }
diff --git a/drivers/gpu/drm/xe/xe_wa_oob.rules b/drivers/gpu/drm/xe/xe_wa_oob.rules
index eb761d30e066..113a62f1b541 100644
--- a/drivers/gpu/drm/xe/xe_wa_oob.rules
+++ b/drivers/gpu/drm/xe/xe_wa_oob.rules
@@ -14,7 +14,6 @@
 14016763929	SUBPLATFORM(DG2, G10)
 		SUBPLATFORM(DG2, G12)
 16017236439	PLATFORM(PVC)
-22010954014	PLATFORM(DG2)
 14019821291	MEDIA_VERSION_RANGE(1300, 2000)
 14015076503	MEDIA_VERSION(1300)
 16020292621	GRAPHICS_VERSION(2004), GRAPHICS_STEP(A0, B0)
@@ -74,9 +73,5 @@
 16023683509	MEDIA_VERSION(2000), FUNC(xe_rtp_match_psmi_enabled)
 		MEDIA_VERSION(3000), MEDIA_STEP(A0, B0), FUNC(xe_rtp_match_psmi_enabled)
 
-# SoC workaround - currently applies to all platforms with the following
-# primary GT GMDID
-14022085890	GRAPHICS_VERSION(2001)
-
 15015404425_disable	PLATFORM(PANTHERLAKE), MEDIA_STEP(B0, FOREVER)
 16026007364    MEDIA_VERSION(3000)
diff --git a/include/drm/intel/pciids.h b/include/drm/intel/pciids.h
index da6301a6fcea..0cd12616d621 100644
--- a/include/drm/intel/pciids.h
+++ b/include/drm/intel/pciids.h
@@ -849,7 +849,7 @@
 	MACRO__(0x64B0, ## __VA_ARGS__)
 
 /* BMG */
-#define INTEL_BMG_IDS(MACRO__, ...) \
+#define INTEL_BMG_G21_IDS(MACRO__, ...) \
 	MACRO__(0xE202, ## __VA_ARGS__), \
 	MACRO__(0xE209, ## __VA_ARGS__), \
 	MACRO__(0xE20B, ## __VA_ARGS__), \
@@ -858,7 +858,10 @@
 	MACRO__(0xE210, ## __VA_ARGS__), \
 	MACRO__(0xE211, ## __VA_ARGS__), \
 	MACRO__(0xE212, ## __VA_ARGS__), \
-	MACRO__(0xE216, ## __VA_ARGS__), \
+	MACRO__(0xE216, ## __VA_ARGS__)
+
+#define INTEL_BMG_IDS(MACRO__, ...) \
+	INTEL_BMG_G21_IDS(MACRO__, __VA_ARGS__), \
 	MACRO__(0xE220, ## __VA_ARGS__), \
 	MACRO__(0xE221, ## __VA_ARGS__), \
 	MACRO__(0xE222, ## __VA_ARGS__), \
-- 
cgit v1.2.3


From 409b9499099b4ad14ca60b59c6edfebfaf74f907 Mon Sep 17 00:00:00 2001
From: Sanjay Yadav <sanjay.kumar.yadav@intel.com>
Date: Tue, 14 Oct 2025 19:58:24 +0530
Subject: drm/xe/uapi: Add documentation for DRM_XE_GEM_CREATE_FLAG_SCANOUT

Add documentation for drm_xe_gem_create structure flag
DRM_XE_GEM_CREATE_FLAG_SCANOUT.

Signed-off-by: Sanjay Yadav <sanjay.kumar.yadav@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Link: https://lore.kernel.org/r/20251014142823.3701228-2-sanjay.kumar.yadav@intel.com
---
 include/uapi/drm/xe_drm.h | 6 +++++-
 1 file changed, 5 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
index 40ff19f52a8d..2d7945cda739 100644
--- a/include/uapi/drm/xe_drm.h
+++ b/include/uapi/drm/xe_drm.h
@@ -771,7 +771,11 @@ struct drm_xe_device_query {
  *    until the object is either bound to a virtual memory region via
  *    VM_BIND or accessed by the CPU. As a result, no backing memory is
  *    reserved at the time of GEM object creation.
- *  - %DRM_XE_GEM_CREATE_FLAG_SCANOUT
+ *  - %DRM_XE_GEM_CREATE_FLAG_SCANOUT - Indicates that the GEM object is
+ *    intended for scanout via the display engine. When set, kernel ensures
+ *    that the allocation is placed in a memory region compatible with the
+ *    display engine requirements. This may impose restrictions on tiling,
+ *    alignment, and memory placement to guarantee proper display functionality.
  *  - %DRM_XE_GEM_CREATE_FLAG_NEEDS_VISIBLE_VRAM - When using VRAM as a
  *    possible placement, ensure that the corresponding VRAM allocation
  *    will always use the CPU accessible part of VRAM. This is important
-- 
cgit v1.2.3


From e4a2d54a2f1a9c9a1971651832c8f0ad9d3782c4 Mon Sep 17 00:00:00 2001
From: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Date: Wed, 3 Sep 2025 21:50:59 +0300
Subject: drm/bridge: dw-hdmi-qp: Add CEC support

Add support for the CEC interface of the Synopsys DesignWare HDMI QP TX
controller.

This is based on the downstream implementation, but rewritten on top of
the CEC helpers added recently to the DRM HDMI connector framework.

Also note struct dw_hdmi_qp_plat_data has been extended to include the
CEC IRQ number to be provided by the platform driver.

Co-developed-by: Algea Cao <algea.cao@rock-chips.com>
Signed-off-by: Algea Cao <algea.cao@rock-chips.com>
Co-developed-by: Derek Foreman <derek.foreman@collabora.com>
Signed-off-by: Derek Foreman <derek.foreman@collabora.com>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://lore.kernel.org/r/20250903-rk3588-hdmi-cec-v4-1-fa25163c4b08@collabora.com
---
 drivers/gpu/drm/bridge/synopsys/Kconfig      |   8 +
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c | 212 +++++++++++++++++++++++++++
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.h |  14 ++
 include/drm/bridge/dw_hdmi_qp.h              |   1 +
 4 files changed, 235 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/bridge/synopsys/Kconfig b/drivers/gpu/drm/bridge/synopsys/Kconfig
index 2c5e532410de..a46df7583bcf 100644
--- a/drivers/gpu/drm/bridge/synopsys/Kconfig
+++ b/drivers/gpu/drm/bridge/synopsys/Kconfig
@@ -61,6 +61,14 @@ config DRM_DW_HDMI_QP
 	select DRM_KMS_HELPER
 	select REGMAP_MMIO
 
+config DRM_DW_HDMI_QP_CEC
+	bool "Synopsis Designware QP CEC interface"
+	depends on DRM_DW_HDMI_QP
+	select DRM_DISPLAY_HDMI_CEC_HELPER
+	help
+	  Support the CEC interface which is part of the Synopsys
+	  Designware HDMI QP block.
+
 config DRM_DW_MIPI_DSI
 	tristate
 	select DRM_KMS_HELPER
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
index 39332c57f2c5..fc98953672b6 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
@@ -18,6 +18,7 @@
 
 #include <drm/bridge/dw_hdmi_qp.h>
 #include <drm/display/drm_hdmi_helper.h>
+#include <drm/display/drm_hdmi_cec_helper.h>
 #include <drm/display/drm_hdmi_state_helper.h>
 #include <drm/drm_atomic.h>
 #include <drm/drm_atomic_helper.h>
@@ -26,6 +27,8 @@
 #include <drm/drm_edid.h>
 #include <drm/drm_modes.h>
 
+#include <media/cec.h>
+
 #include <sound/hdmi-codec.h>
 
 #include "dw-hdmi-qp.h"
@@ -131,12 +134,28 @@ struct dw_hdmi_qp_i2c {
 	bool			is_segment;
 };
 
+#ifdef CONFIG_DRM_DW_HDMI_QP_CEC
+struct dw_hdmi_qp_cec {
+	struct drm_connector *connector;
+	int irq;
+	u32 addresses;
+	struct cec_msg rx_msg;
+	u8 tx_status;
+	bool tx_done;
+	bool rx_done;
+};
+#endif
+
 struct dw_hdmi_qp {
 	struct drm_bridge bridge;
 
 	struct device *dev;
 	struct dw_hdmi_qp_i2c *i2c;
 
+#ifdef CONFIG_DRM_DW_HDMI_QP_CEC
+	struct dw_hdmi_qp_cec *cec;
+#endif
+
 	struct {
 		const struct dw_hdmi_qp_phy_ops *ops;
 		void *data;
@@ -965,6 +984,179 @@ static int dw_hdmi_qp_bridge_write_infoframe(struct drm_bridge *bridge,
 	}
 }
 
+#ifdef CONFIG_DRM_DW_HDMI_QP_CEC
+static irqreturn_t dw_hdmi_qp_cec_hardirq(int irq, void *dev_id)
+{
+	struct dw_hdmi_qp *hdmi = dev_id;
+	struct dw_hdmi_qp_cec *cec = hdmi->cec;
+	irqreturn_t ret = IRQ_HANDLED;
+	u32 stat;
+
+	stat = dw_hdmi_qp_read(hdmi, CEC_INT_STATUS);
+	if (stat == 0)
+		return IRQ_NONE;
+
+	dw_hdmi_qp_write(hdmi, stat, CEC_INT_CLEAR);
+
+	if (stat & CEC_STAT_LINE_ERR) {
+		cec->tx_status = CEC_TX_STATUS_ERROR;
+		cec->tx_done = true;
+		ret = IRQ_WAKE_THREAD;
+	} else if (stat & CEC_STAT_DONE) {
+		cec->tx_status = CEC_TX_STATUS_OK;
+		cec->tx_done = true;
+		ret = IRQ_WAKE_THREAD;
+	} else if (stat & CEC_STAT_NACK) {
+		cec->tx_status = CEC_TX_STATUS_NACK;
+		cec->tx_done = true;
+		ret = IRQ_WAKE_THREAD;
+	}
+
+	if (stat & CEC_STAT_EOM) {
+		unsigned int len, i, val;
+
+		val = dw_hdmi_qp_read(hdmi, CEC_RX_COUNT_STATUS);
+		len = (val & 0xf) + 1;
+
+		if (len > sizeof(cec->rx_msg.msg))
+			len = sizeof(cec->rx_msg.msg);
+
+		for (i = 0; i < 4; i++) {
+			val = dw_hdmi_qp_read(hdmi, CEC_RX_DATA3_0 + i * 4);
+			cec->rx_msg.msg[i * 4] = val & 0xff;
+			cec->rx_msg.msg[i * 4 + 1] = (val >> 8) & 0xff;
+			cec->rx_msg.msg[i * 4 + 2] = (val >> 16) & 0xff;
+			cec->rx_msg.msg[i * 4 + 3] = (val >> 24) & 0xff;
+		}
+
+		dw_hdmi_qp_write(hdmi, 1, CEC_LOCK_CONTROL);
+
+		cec->rx_msg.len = len;
+		cec->rx_done = true;
+
+		ret = IRQ_WAKE_THREAD;
+	}
+
+	return ret;
+}
+
+static irqreturn_t dw_hdmi_qp_cec_thread(int irq, void *dev_id)
+{
+	struct dw_hdmi_qp *hdmi = dev_id;
+	struct dw_hdmi_qp_cec *cec = hdmi->cec;
+
+	if (cec->tx_done) {
+		cec->tx_done = false;
+		drm_connector_hdmi_cec_transmit_attempt_done(cec->connector,
+							     cec->tx_status);
+	}
+
+	if (cec->rx_done) {
+		cec->rx_done = false;
+		drm_connector_hdmi_cec_received_msg(cec->connector, &cec->rx_msg);
+	}
+
+	return IRQ_HANDLED;
+}
+
+static int dw_hdmi_qp_cec_init(struct drm_bridge *bridge,
+			       struct drm_connector *connector)
+{
+	struct dw_hdmi_qp *hdmi = dw_hdmi_qp_from_bridge(bridge);
+	struct dw_hdmi_qp_cec *cec = hdmi->cec;
+
+	cec->connector = connector;
+
+	dw_hdmi_qp_write(hdmi, 0, CEC_TX_COUNT);
+	dw_hdmi_qp_write(hdmi, ~0, CEC_INT_CLEAR);
+	dw_hdmi_qp_write(hdmi, 0, CEC_INT_MASK_N);
+
+	return devm_request_threaded_irq(hdmi->dev, cec->irq,
+					 dw_hdmi_qp_cec_hardirq,
+					 dw_hdmi_qp_cec_thread, IRQF_SHARED,
+					 dev_name(hdmi->dev), hdmi);
+}
+
+static int dw_hdmi_qp_cec_log_addr(struct drm_bridge *bridge, u8 logical_addr)
+{
+	struct dw_hdmi_qp *hdmi = dw_hdmi_qp_from_bridge(bridge);
+	struct dw_hdmi_qp_cec *cec = hdmi->cec;
+
+	if (logical_addr == CEC_LOG_ADDR_INVALID)
+		cec->addresses = 0;
+	else
+		cec->addresses |= BIT(logical_addr) | CEC_ADDR_BROADCAST;
+
+	dw_hdmi_qp_write(hdmi, cec->addresses, CEC_ADDR);
+
+	return 0;
+}
+
+static int dw_hdmi_qp_cec_enable(struct drm_bridge *bridge, bool enable)
+{
+	struct dw_hdmi_qp *hdmi = dw_hdmi_qp_from_bridge(bridge);
+	unsigned int irqs;
+	u32 swdisable;
+
+	if (!enable) {
+		dw_hdmi_qp_write(hdmi, 0, CEC_INT_MASK_N);
+		dw_hdmi_qp_write(hdmi, ~0, CEC_INT_CLEAR);
+
+		swdisable = dw_hdmi_qp_read(hdmi, GLOBAL_SWDISABLE);
+		swdisable = swdisable | CEC_SWDISABLE;
+		dw_hdmi_qp_write(hdmi, swdisable, GLOBAL_SWDISABLE);
+	} else {
+		swdisable = dw_hdmi_qp_read(hdmi, GLOBAL_SWDISABLE);
+		swdisable = swdisable & ~CEC_SWDISABLE;
+		dw_hdmi_qp_write(hdmi, swdisable, GLOBAL_SWDISABLE);
+
+		dw_hdmi_qp_write(hdmi, ~0, CEC_INT_CLEAR);
+		dw_hdmi_qp_write(hdmi, 1, CEC_LOCK_CONTROL);
+
+		dw_hdmi_qp_cec_log_addr(bridge, CEC_LOG_ADDR_INVALID);
+
+		irqs = CEC_STAT_LINE_ERR | CEC_STAT_NACK | CEC_STAT_EOM |
+		       CEC_STAT_DONE;
+		dw_hdmi_qp_write(hdmi, ~0, CEC_INT_CLEAR);
+		dw_hdmi_qp_write(hdmi, irqs, CEC_INT_MASK_N);
+	}
+
+	return 0;
+}
+
+static int dw_hdmi_qp_cec_transmit(struct drm_bridge *bridge, u8 attempts,
+				   u32 signal_free_time, struct cec_msg *msg)
+{
+	struct dw_hdmi_qp *hdmi = dw_hdmi_qp_from_bridge(bridge);
+	unsigned int i;
+	u32 val;
+
+	for (i = 0; i < msg->len; i++) {
+		if (!(i % 4))
+			val = msg->msg[i];
+		if ((i % 4) == 1)
+			val |= msg->msg[i] << 8;
+		if ((i % 4) == 2)
+			val |= msg->msg[i] << 16;
+		if ((i % 4) == 3)
+			val |= msg->msg[i] << 24;
+
+		if (i == (msg->len - 1) || (i % 4) == 3)
+			dw_hdmi_qp_write(hdmi, val, CEC_TX_DATA3_0 + (i / 4) * 4);
+	}
+
+	dw_hdmi_qp_write(hdmi, msg->len - 1, CEC_TX_COUNT);
+	dw_hdmi_qp_write(hdmi, CEC_CTRL_START, CEC_TX_CONTROL);
+
+	return 0;
+}
+#else
+#define dw_hdmi_qp_cec_init NULL
+#define dw_hdmi_qp_cec_enable NULL
+#define dw_hdmi_qp_cec_log_addr NULL
+#define dw_hdmi_qp_cec_transmit NULL
+#endif /* CONFIG_DRM_DW_HDMI_QP_CEC */
+
 static const struct drm_bridge_funcs dw_hdmi_qp_bridge_funcs = {
 	.atomic_duplicate_state = drm_atomic_helper_bridge_duplicate_state,
 	.atomic_destroy_state = drm_atomic_helper_bridge_destroy_state,
@@ -979,6 +1171,10 @@ static const struct drm_bridge_funcs dw_hdmi_qp_bridge_funcs = {
 	.hdmi_audio_startup = dw_hdmi_qp_audio_enable,
 	.hdmi_audio_shutdown = dw_hdmi_qp_audio_disable,
 	.hdmi_audio_prepare = dw_hdmi_qp_audio_prepare,
+	.hdmi_cec_init = dw_hdmi_qp_cec_init,
+	.hdmi_cec_enable = dw_hdmi_qp_cec_enable,
+	.hdmi_cec_log_addr = dw_hdmi_qp_cec_log_addr,
+	.hdmi_cec_transmit = dw_hdmi_qp_cec_transmit,
 };
 
 static irqreturn_t dw_hdmi_qp_main_hardirq(int irq, void *dev_id)
@@ -1093,6 +1289,22 @@ struct dw_hdmi_qp *dw_hdmi_qp_bind(struct platform_device *pdev,
 	hdmi->bridge.hdmi_audio_dev = dev;
 	hdmi->bridge.hdmi_audio_dai_port = 1;
 
+#ifdef CONFIG_DRM_DW_HDMI_QP_CEC
+	if (plat_data->cec_irq) {
+		hdmi->bridge.ops |= DRM_BRIDGE_OP_HDMI_CEC_ADAPTER;
+		hdmi->bridge.hdmi_cec_dev = dev;
+		hdmi->bridge.hdmi_cec_adapter_name = dev_name(dev);
+
+		hdmi->cec = devm_kzalloc(hdmi->dev, sizeof(*hdmi->cec), GFP_KERNEL);
+		if (!hdmi->cec)
+			return ERR_PTR(-ENOMEM);
+
+		hdmi->cec->irq = plat_data->cec_irq;
+	} else {
+		dev_warn(dev, "Disabled CEC support due to missing IRQ\n");
+	}
+#endif
+
 	ret = devm_drm_bridge_add(dev, &hdmi->bridge);
 	if (ret)
 		return ERR_PTR(ret);
diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.h b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.h
index 72987e6c4689..91a15f82e32a 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.h
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.h
@@ -488,9 +488,23 @@
 #define AUDPKT_VBIT_OVR0				0xf24
 /* CEC Registers */
 #define CEC_TX_CONTROL					0x1000
+#define CEC_CTRL_CLEAR					BIT(0)
+#define CEC_CTRL_START					BIT(0)
 #define CEC_STATUS					0x1004
+#define CEC_STAT_DONE					BIT(0)
+#define CEC_STAT_NACK					BIT(1)
+#define CEC_STAT_ARBLOST				BIT(2)
+#define CEC_STAT_LINE_ERR				BIT(3)
+#define CEC_STAT_RETRANS_FAIL				BIT(4)
+#define CEC_STAT_DISCARD				BIT(5)
+#define CEC_STAT_TX_BUSY				BIT(8)
+#define CEC_STAT_RX_BUSY				BIT(9)
+#define CEC_STAT_DRIVE_ERR				BIT(10)
+#define CEC_STAT_EOM					BIT(11)
+#define CEC_STAT_NOTIFY_ERR				BIT(12)
 #define CEC_CONFIG					0x1008
 #define CEC_ADDR					0x100c
+#define CEC_ADDR_BROADCAST				BIT(15)
 #define CEC_TX_COUNT					0x1020
 #define CEC_TX_DATA3_0					0x1024
 #define CEC_TX_DATA7_4					0x1028
diff --git a/include/drm/bridge/dw_hdmi_qp.h b/include/drm/bridge/dw_hdmi_qp.h
index e9be6d507ad9..b4a9b739734e 100644
--- a/include/drm/bridge/dw_hdmi_qp.h
+++ b/include/drm/bridge/dw_hdmi_qp.h
@@ -23,6 +23,7 @@ struct dw_hdmi_qp_plat_data {
 	const struct dw_hdmi_qp_phy_ops *phy_ops;
 	void *phy_data;
 	int main_irq;
+	int cec_irq;
 };
 
 struct dw_hdmi_qp *dw_hdmi_qp_bind(struct platform_device *pdev,
-- 
cgit v1.2.3


From f7a1de0d86221000dc0699a8b48ad3a848e766d9 Mon Sep 17 00:00:00 2001
From: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Date: Wed, 3 Sep 2025 21:51:00 +0300
Subject: drm/bridge: dw-hdmi-qp: Fixup timer base setup

Currently the TIMER_BASE_CONFIG0 register gets initialized to a fixed
value as initially found in vendor driver code supporting the RK3588
SoC.  As a matter of fact the value matches the rate of the HDMI TX
reference clock, which is roughly 428.57 MHz.

However, on RK3576 SoC that rate is slightly lower, i.e. 396.00 MHz, and
the incorrect register configuration breaks CEC functionality.

Set the timer base according to the actual reference clock rate that
shall be provided by the platform driver.  Otherwise fallback to the
vendor default.

While at it, also drop the unnecessary empty lines in
dw_hdmi_qp_init_hw().

Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://lore.kernel.org/r/20250903-rk3588-hdmi-cec-v4-2-fa25163c4b08@collabora.com
---
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c | 12 +++++++++---
 include/drm/bridge/dw_hdmi_qp.h              |  1 +
 2 files changed, 10 insertions(+), 3 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
index fc98953672b6..4ba7b339eff6 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
@@ -161,6 +161,7 @@ struct dw_hdmi_qp {
 		void *data;
 	} phy;
 
+	unsigned long ref_clk_rate;
 	struct regmap *regm;
 
 	unsigned long tmds_char_rate;
@@ -1210,13 +1211,11 @@ static void dw_hdmi_qp_init_hw(struct dw_hdmi_qp *hdmi)
 {
 	dw_hdmi_qp_write(hdmi, 0, MAINUNIT_0_INT_MASK_N);
 	dw_hdmi_qp_write(hdmi, 0, MAINUNIT_1_INT_MASK_N);
-	dw_hdmi_qp_write(hdmi, 428571429, TIMER_BASE_CONFIG0);
+	dw_hdmi_qp_write(hdmi, hdmi->ref_clk_rate, TIMER_BASE_CONFIG0);
 
 	/* Software reset */
 	dw_hdmi_qp_write(hdmi, 0x01, I2CM_CONTROL0);
-
 	dw_hdmi_qp_write(hdmi, 0x085c085c, I2CM_FM_SCL_CONFIG0);
-
 	dw_hdmi_qp_mod(hdmi, 0, I2CM_FM_EN, I2CM_INTERFACE_CONTROL0);
 
 	/* Clear DONE and ERROR interrupts */
@@ -1262,6 +1261,13 @@ struct dw_hdmi_qp *dw_hdmi_qp_bind(struct platform_device *pdev,
 	hdmi->phy.ops = plat_data->phy_ops;
 	hdmi->phy.data = plat_data->phy_data;
 
+	if (plat_data->ref_clk_rate) {
+		hdmi->ref_clk_rate = plat_data->ref_clk_rate;
+	} else {
+		hdmi->ref_clk_rate = 428571429;
+		dev_warn(dev, "Set ref_clk_rate to vendor default\n");
+	}
+
 	dw_hdmi_qp_init_hw(hdmi);
 
 	ret = devm_request_threaded_irq(dev, plat_data->main_irq,
diff --git a/include/drm/bridge/dw_hdmi_qp.h b/include/drm/bridge/dw_hdmi_qp.h
index b4a9b739734e..76ecf3130199 100644
--- a/include/drm/bridge/dw_hdmi_qp.h
+++ b/include/drm/bridge/dw_hdmi_qp.h
@@ -24,6 +24,7 @@ struct dw_hdmi_qp_plat_data {
 	void *phy_data;
 	int main_irq;
 	int cec_irq;
+	unsigned long ref_clk_rate;
 };
 
 struct dw_hdmi_qp *dw_hdmi_qp_bind(struct platform_device *pdev,
-- 
cgit v1.2.3


From b291e4f1a4951204ce858cd01801291d34962a33 Mon Sep 17 00:00:00 2001
From: Lizhi Hou <lizhi.hou@amd.com>
Date: Tue, 14 Oct 2025 16:41:19 -0700
Subject: accel/amdxdna: Support getting last hardware error

Add new parameter DRM_AMDXDNA_HW_LAST_ASYNC_ERR to get array IOCTL. When
hardware reports an error, the driver save the error information and
timestamp. This new get array parameter retrieves the last error.

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Link: https://lore.kernel.org/r/20251014234119.628453-1-lizhi.hou@amd.com
---
 drivers/accel/amdxdna/aie2_error.c      | 95 +++++++++++++++++++++++++++------
 drivers/accel/amdxdna/aie2_pci.c        |  3 ++
 drivers/accel/amdxdna/aie2_pci.h        |  5 +-
 drivers/accel/amdxdna/amdxdna_error.h   | 59 ++++++++++++++++++++
 drivers/accel/amdxdna/amdxdna_pci_drv.c |  3 +-
 include/uapi/drm/amdxdna_accel.h        | 13 +++++
 6 files changed, 159 insertions(+), 19 deletions(-)
 create mode 100644 drivers/accel/amdxdna/amdxdna_error.h

(limited to 'include')

diff --git a/drivers/accel/amdxdna/aie2_error.c b/drivers/accel/amdxdna/aie2_error.c
index 5ee905632a39..d452008ec4f4 100644
--- a/drivers/accel/amdxdna/aie2_error.c
+++ b/drivers/accel/amdxdna/aie2_error.c
@@ -13,6 +13,7 @@
 
 #include "aie2_msg_priv.h"
 #include "aie2_pci.h"
+#include "amdxdna_error.h"
 #include "amdxdna_mailbox.h"
 #include "amdxdna_pci_drv.h"
 
@@ -46,6 +47,7 @@ enum aie_module_type {
 	AIE_MEM_MOD = 0,
 	AIE_CORE_MOD,
 	AIE_PL_MOD,
+	AIE_UNKNOWN_MOD,
 };
 
 enum aie_error_category {
@@ -143,6 +145,31 @@ static const struct aie_event_category aie_ml_shim_tile_event_cat[] = {
 	EVENT_CATEGORY(74U, AIE_ERROR_LOCK),
 };
 
+static const enum amdxdna_error_num aie_cat_err_num_map[] = {
+	[AIE_ERROR_SATURATION] = AMDXDNA_ERROR_NUM_AIE_SATURATION,
+	[AIE_ERROR_FP] = AMDXDNA_ERROR_NUM_AIE_FP,
+	[AIE_ERROR_STREAM] = AMDXDNA_ERROR_NUM_AIE_STREAM,
+	[AIE_ERROR_ACCESS] = AMDXDNA_ERROR_NUM_AIE_ACCESS,
+	[AIE_ERROR_BUS] = AMDXDNA_ERROR_NUM_AIE_BUS,
+	[AIE_ERROR_INSTRUCTION] = AMDXDNA_ERROR_NUM_AIE_INSTRUCTION,
+	[AIE_ERROR_ECC] = AMDXDNA_ERROR_NUM_AIE_ECC,
+	[AIE_ERROR_LOCK] = AMDXDNA_ERROR_NUM_AIE_LOCK,
+	[AIE_ERROR_DMA] = AMDXDNA_ERROR_NUM_AIE_DMA,
+	[AIE_ERROR_MEM_PARITY] = AMDXDNA_ERROR_NUM_AIE_MEM_PARITY,
+	[AIE_ERROR_UNKNOWN] = AMDXDNA_ERROR_NUM_UNKNOWN,
+};
+
+static_assert(ARRAY_SIZE(aie_cat_err_num_map) == AIE_ERROR_UNKNOWN + 1);
+
+static const enum amdxdna_error_module aie_err_mod_map[] = {
+	[AIE_MEM_MOD] = AMDXDNA_ERROR_MODULE_AIE_MEMORY,
+	[AIE_CORE_MOD] = AMDXDNA_ERROR_MODULE_AIE_CORE,
+	[AIE_PL_MOD] = AMDXDNA_ERROR_MODULE_AIE_PL,
+	[AIE_UNKNOWN_MOD] = AMDXDNA_ERROR_MODULE_UNKNOWN,
+};
+
+static_assert(ARRAY_SIZE(aie_err_mod_map) == AIE_UNKNOWN_MOD + 1);
+
 static enum aie_error_category
 aie_get_error_category(u8 row, u8 event_id, enum aie_module_type mod_type)
 {
@@ -176,12 +203,40 @@ aie_get_error_category(u8 row, u8 event_id, enum aie_module_type mod_type)
 		if (event_id != lut[i].event_id)
 			continue;
 
+		if (lut[i].category > AIE_ERROR_UNKNOWN)
+			return AIE_ERROR_UNKNOWN;
+
 		return lut[i].category;
 	}
 
 	return AIE_ERROR_UNKNOWN;
 }
 
+static void aie2_update_last_async_error(struct amdxdna_dev_hdl *ndev, void *err_info, u32 num_err)
+{
+	struct aie_error *errs = err_info;
+	enum amdxdna_error_module err_mod;
+	enum aie_error_category aie_err;
+	enum amdxdna_error_num err_num;
+	struct aie_error *last_err;
+
+	last_err = &errs[num_err - 1];
+	if (last_err->mod_type >= AIE_UNKNOWN_MOD) {
+		err_num = aie_cat_err_num_map[AIE_ERROR_UNKNOWN];
+		err_mod = aie_err_mod_map[AIE_UNKNOWN_MOD];
+	} else {
+		aie_err = aie_get_error_category(last_err->row,
+						 last_err->event_id,
+						 last_err->mod_type);
+		err_num = aie_cat_err_num_map[aie_err];
+		err_mod = aie_err_mod_map[last_err->mod_type];
+	}
+
+	ndev->last_async_err.err_code = AMDXDNA_ERROR_ENCODE(err_num, err_mod);
+	ndev->last_async_err.ts_us = ktime_to_us(ktime_get_real());
+	ndev->last_async_err.ex_err_code = AMDXDNA_EXTRA_ERR_ENCODE(last_err->row, last_err->col);
+}
+
 static u32 aie2_error_backtrack(struct amdxdna_dev_hdl *ndev, void *err_info, u32 num_err)
 {
 	struct aie_error *errs = err_info;
@@ -264,29 +319,14 @@ static void aie2_error_worker(struct work_struct *err_work)
 	}
 
 	mutex_lock(&xdna->dev_lock);
+	aie2_update_last_async_error(e->ndev, info->payload, info->err_cnt);
+
 	/* Re-sent this event to firmware */
 	if (aie2_error_event_send(e))
 		XDNA_WARN(xdna, "Unable to register async event");
 	mutex_unlock(&xdna->dev_lock);
 }
 
-int aie2_error_async_events_send(struct amdxdna_dev_hdl *ndev)
-{
-	struct amdxdna_dev *xdna = ndev->xdna;
-	struct async_event *e;
-	int i, ret;
-
-	drm_WARN_ON(&xdna->ddev, !mutex_is_locked(&xdna->dev_lock));
-	for (i = 0; i < ndev->async_events->event_cnt; i++) {
-		e = &ndev->async_events->event[i];
-		ret = aie2_error_event_send(e);
-		if (ret)
-			return ret;
-	}
-
-	return 0;
-}
-
 void aie2_error_async_events_free(struct amdxdna_dev_hdl *ndev)
 {
 	struct amdxdna_dev *xdna = ndev->xdna;
@@ -341,6 +381,10 @@ int aie2_error_async_events_alloc(struct amdxdna_dev_hdl *ndev)
 		e->size = ASYNC_BUF_SIZE;
 		e->resp.status = MAX_AIE2_STATUS_CODE;
 		INIT_WORK(&e->work, aie2_error_worker);
+
+		ret = aie2_error_event_send(e);
+		if (ret)
+			goto free_wq;
 	}
 
 	ndev->async_events = events;
@@ -349,6 +393,8 @@ int aie2_error_async_events_alloc(struct amdxdna_dev_hdl *ndev)
 		 events->event_cnt, events->size);
 	return 0;
 
+free_wq:
+	destroy_workqueue(events->wq);
 free_buf:
 	dma_free_noncoherent(xdna->ddev.dev, events->size, events->buf,
 			     events->addr, DMA_FROM_DEVICE);
@@ -356,3 +402,18 @@ free_events:
 	kfree(events);
 	return ret;
 }
+
+int aie2_get_array_async_error(struct amdxdna_dev_hdl *ndev, struct amdxdna_drm_get_array *args)
+{
+	struct amdxdna_dev *xdna = ndev->xdna;
+
+	drm_WARN_ON(&xdna->ddev, !mutex_is_locked(&xdna->dev_lock));
+
+	args->num_element = 1;
+	args->element_size = sizeof(ndev->last_async_err);
+	if (copy_to_user(u64_to_user_ptr(args->buffer),
+			 &ndev->last_async_err, args->element_size))
+		return -EFAULT;
+
+	return 0;
+}
diff --git a/drivers/accel/amdxdna/aie2_pci.c b/drivers/accel/amdxdna/aie2_pci.c
index 8a66f276100e..cfca4e456b61 100644
--- a/drivers/accel/amdxdna/aie2_pci.c
+++ b/drivers/accel/amdxdna/aie2_pci.c
@@ -924,6 +924,9 @@ static int aie2_get_array(struct amdxdna_client *client,
 	case DRM_AMDXDNA_HW_CONTEXT_ALL:
 		ret = aie2_query_ctx_status_array(client, args);
 		break;
+	case DRM_AMDXDNA_HW_LAST_ASYNC_ERR:
+		ret = aie2_get_array_async_error(xdna->dev_handle, args);
+		break;
 	default:
 		XDNA_ERR(xdna, "Not supported request parameter %u", args->param);
 		ret = -EOPNOTSUPP;
diff --git a/drivers/accel/amdxdna/aie2_pci.h b/drivers/accel/amdxdna/aie2_pci.h
index 289a23ecd5f1..34bc35479f42 100644
--- a/drivers/accel/amdxdna/aie2_pci.h
+++ b/drivers/accel/amdxdna/aie2_pci.h
@@ -190,6 +190,8 @@ struct amdxdna_dev_hdl {
 
 	enum aie2_dev_status		dev_status;
 	u32				hwctx_num;
+
+	struct amdxdna_async_error	last_async_err;
 };
 
 #define DEFINE_BAR_OFFSET(reg_name, bar, reg_addr) \
@@ -253,8 +255,9 @@ void aie2_psp_stop(struct psp_device *psp);
 /* aie2_error.c */
 int aie2_error_async_events_alloc(struct amdxdna_dev_hdl *ndev);
 void aie2_error_async_events_free(struct amdxdna_dev_hdl *ndev);
-int aie2_error_async_events_send(struct amdxdna_dev_hdl *ndev);
 int aie2_error_async_msg_thread(void *data);
+int aie2_get_array_async_error(struct amdxdna_dev_hdl *ndev,
+			       struct amdxdna_drm_get_array *args);
 
 /* aie2_message.c */
 int aie2_suspend_fw(struct amdxdna_dev_hdl *ndev);
diff --git a/drivers/accel/amdxdna/amdxdna_error.h b/drivers/accel/amdxdna/amdxdna_error.h
new file mode 100644
index 000000000000..c51de86ec12b
--- /dev/null
+++ b/drivers/accel/amdxdna/amdxdna_error.h
@@ -0,0 +1,59 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * Copyright (C) 2025, Advanced Micro Devices, Inc.
+ */
+
+#ifndef _AMDXDNA_ERROR_H_
+#define _AMDXDNA_ERROR_H_
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+
+#define AMDXDNA_ERR_DRV_AIE		4
+#define AMDXDNA_ERR_SEV_CRITICAL	3
+#define AMDXDNA_ERR_CLASS_AIE		2
+
+#define AMDXDNA_ERR_NUM_MASK		GENMASK_U64(15, 0)
+#define AMDXDNA_ERR_DRV_MASK		GENMASK_U64(23, 16)
+#define AMDXDNA_ERR_SEV_MASK		GENMASK_U64(31, 24)
+#define AMDXDNA_ERR_MOD_MASK		GENMASK_U64(39, 32)
+#define AMDXDNA_ERR_CLASS_MASK		GENMASK_U64(47, 40)
+
+enum amdxdna_error_num {
+	AMDXDNA_ERROR_NUM_AIE_SATURATION = 3,
+	AMDXDNA_ERROR_NUM_AIE_FP,
+	AMDXDNA_ERROR_NUM_AIE_STREAM,
+	AMDXDNA_ERROR_NUM_AIE_ACCESS,
+	AMDXDNA_ERROR_NUM_AIE_BUS,
+	AMDXDNA_ERROR_NUM_AIE_INSTRUCTION,
+	AMDXDNA_ERROR_NUM_AIE_ECC,
+	AMDXDNA_ERROR_NUM_AIE_LOCK,
+	AMDXDNA_ERROR_NUM_AIE_DMA,
+	AMDXDNA_ERROR_NUM_AIE_MEM_PARITY,
+	AMDXDNA_ERROR_NUM_UNKNOWN = 15,
+};
+
+enum amdxdna_error_module {
+	AMDXDNA_ERROR_MODULE_AIE_CORE = 3,
+	AMDXDNA_ERROR_MODULE_AIE_MEMORY,
+	AMDXDNA_ERROR_MODULE_AIE_SHIM,
+	AMDXDNA_ERROR_MODULE_AIE_NOC,
+	AMDXDNA_ERROR_MODULE_AIE_PL,
+	AMDXDNA_ERROR_MODULE_UNKNOWN = 8,
+};
+
+#define AMDXDNA_ERROR_ENCODE(err_num, err_mod)				\
+	(FIELD_PREP(AMDXDNA_ERR_NUM_MASK, err_num) |			\
+	 FIELD_PREP_CONST(AMDXDNA_ERR_DRV_MASK, AMDXDNA_ERR_DRV_AIE) |	\
+	 FIELD_PREP_CONST(AMDXDNA_ERR_SEV_MASK, AMDXDNA_ERR_SEV_CRITICAL) | \
+	 FIELD_PREP(AMDXDNA_ERR_MOD_MASK, err_mod) |			\
+	 FIELD_PREP_CONST(AMDXDNA_ERR_CLASS_MASK, AMDXDNA_ERR_CLASS_AIE))
+
+#define AMDXDNA_EXTRA_ERR_COL_MASK	GENMASK_U64(7, 0)
+#define AMDXDNA_EXTRA_ERR_ROW_MASK	GENMASK_U64(15, 8)
+
+#define AMDXDNA_EXTRA_ERR_ENCODE(row, col)				\
+	(FIELD_PREP(AMDXDNA_EXTRA_ERR_COL_MASK, col) |			\
+	 FIELD_PREP(AMDXDNA_EXTRA_ERR_ROW_MASK, row))
+
+#endif /* _AMDXDNA_ERROR_H_ */
diff --git a/drivers/accel/amdxdna/amdxdna_pci_drv.c b/drivers/accel/amdxdna/amdxdna_pci_drv.c
index aa04452310e5..696fdac8ad3c 100644
--- a/drivers/accel/amdxdna/amdxdna_pci_drv.c
+++ b/drivers/accel/amdxdna/amdxdna_pci_drv.c
@@ -27,9 +27,10 @@ MODULE_FIRMWARE("amdnpu/17f0_20/npu.sbin");
 /*
  * 0.0: Initial version
  * 0.1: Support getting all hardware contexts by DRM_IOCTL_AMDXDNA_GET_ARRAY
+ * 0.2: Support getting last error hardware error
  */
 #define AMDXDNA_DRIVER_MAJOR		0
-#define AMDXDNA_DRIVER_MINOR		1
+#define AMDXDNA_DRIVER_MINOR		2
 
 /*
  * Bind the driver base on (vendor_id, device_id) pair and later use the
diff --git a/include/uapi/drm/amdxdna_accel.h b/include/uapi/drm/amdxdna_accel.h
index a1fb9785db77..c7eec9ceb2ae 100644
--- a/include/uapi/drm/amdxdna_accel.h
+++ b/include/uapi/drm/amdxdna_accel.h
@@ -523,7 +523,20 @@ struct amdxdna_drm_hwctx_entry {
 	__u32 pad;
 };
 
+/**
+ * struct amdxdna_async_error - XDNA async error structure
+ */
+struct amdxdna_async_error {
+	/** @err_code: Error code. */
+	__u64 err_code;
+	/** @ts_us: Timestamp. */
+	__u64 ts_us;
+	/** @ex_err_code: Extra error code */
+	__u64 ex_err_code;
+};
+
 #define DRM_AMDXDNA_HW_CONTEXT_ALL	0
+#define DRM_AMDXDNA_HW_LAST_ASYNC_ERR	2
 
 /**
  * struct amdxdna_drm_get_array - Get information array.
-- 
cgit v1.2.3


From bce13d6ecd6c0e6c3376ac53617b30d19ce2a5c6 Mon Sep 17 00:00:00 2001
From: Matthew Brost <matthew.brost@intel.com>
Date: Wed, 15 Oct 2025 14:03:20 +0200
Subject: drm/gpusvm, drm/xe: Allow mixed mappings for userptr
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Compute kernels often issue memory copies immediately after completion.
If the memory being copied is an SVM pointer that was faulted into the
device and then bound via userptr, it is undesirable to move that
memory. Worse, if userptr is mixed between system and device memory, the
bind operation may be rejected.

Xe already has the necessary plumbing to support userptr with mixed
mappings. This update modifies GPUSVM's get_pages to correctly locate
pages in such mixed mapping scenarios.

v2:
- Rebase (Thomas Hellström)
v3:
- Remove Fixes tag.
v4:
- Break out from series since the other patch was merged.
- Update patch subject, ensure dri-devel and Maarten are CC'd.

Cc: Maarten Lankhorst <maarten.lankhorst@intel.com>
Cc: dri-devel@lists.freedesktop.org
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Link: https://lore.kernel.org/r/20251015120320.176338-1-thomas.hellstrom@linux.intel.com
Acked-by: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/drm_gpusvm.c    | 6 ++++--
 drivers/gpu/drm/xe/xe_userptr.c | 4 +++-
 include/drm/drm_gpusvm.h        | 4 ++++
 3 files changed, 11 insertions(+), 3 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
index cb906765897e..73e550c8ff8c 100644
--- a/drivers/gpu/drm/drm_gpusvm.c
+++ b/drivers/gpu/drm/drm_gpusvm.c
@@ -1363,7 +1363,8 @@ map_pages:
 		order = drm_gpusvm_hmm_pfn_to_order(pfns[i], i, npages);
 		if (is_device_private_page(page) ||
 		    is_device_coherent_page(page)) {
-			if (zdd != page->zone_device_data && i > 0) {
+			if (!ctx->allow_mixed &&
+			    zdd != page->zone_device_data && i > 0) {
 				err = -EOPNOTSUPP;
 				goto err_unmap;
 			}
@@ -1399,7 +1400,8 @@ map_pages:
 		} else {
 			dma_addr_t addr;
 
-			if (is_zone_device_page(page) || pagemap) {
+			if (is_zone_device_page(page) ||
+			    (pagemap && !ctx->allow_mixed)) {
 				err = -EOPNOTSUPP;
 				goto err_unmap;
 			}
diff --git a/drivers/gpu/drm/xe/xe_userptr.c b/drivers/gpu/drm/xe/xe_userptr.c
index f16e92cd8090..0d9130b1958a 100644
--- a/drivers/gpu/drm/xe/xe_userptr.c
+++ b/drivers/gpu/drm/xe/xe_userptr.c
@@ -3,6 +3,7 @@
  * Copyright © 2025 Intel Corporation
  */
 
+#include "xe_svm.h"
 #include "xe_userptr.h"
 
 #include <linux/mm.h>
@@ -54,7 +55,8 @@ int xe_vma_userptr_pin_pages(struct xe_userptr_vma *uvma)
 	struct xe_device *xe = vm->xe;
 	struct drm_gpusvm_ctx ctx = {
 		.read_only = xe_vma_read_only(vma),
-		.device_private_page_owner = NULL,
+		.device_private_page_owner = xe_svm_devm_owner(xe),
+		.allow_mixed = true,
 	};
 
 	lockdep_assert_held(&vm->lock);
diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
index b92faa9a26b2..632e100e6efb 100644
--- a/include/drm/drm_gpusvm.h
+++ b/include/drm/drm_gpusvm.h
@@ -235,6 +235,9 @@ struct drm_gpusvm {
  * @read_only: operating on read-only memory
  * @devmem_possible: possible to use device memory
  * @devmem_only: use only device memory
+ * @allow_mixed: Allow mixed mappings in get pages. Mixing between system and
+ *               single dpagemap is supported, mixing between multiple dpagemap
+ *               is unsupported.
  *
  * Context that is DRM GPUSVM is operating in (i.e. user arguments).
  */
@@ -246,6 +249,7 @@ struct drm_gpusvm_ctx {
 	unsigned int read_only :1;
 	unsigned int devmem_possible :1;
 	unsigned int devmem_only :1;
+	unsigned int allow_mixed :1;
 };
 
 int drm_gpusvm_init(struct drm_gpusvm *gpusvm,
-- 
cgit v1.2.3


From 59a2d3f38ab23cce4cd9f0c4a5e08fdfe9e67ae7 Mon Sep 17 00:00:00 2001
From: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Date: Wed, 15 Oct 2025 19:07:26 +0200
Subject: drm/xe/uapi: Hide the madvise autoreset behind a VM_BIND flag
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

The madvise implementation currently resets the SVM madvise if the
underlying CPU map is unmapped. This is in an attempt to mimic the
CPU madvise behaviour. However, it's not clear that this is a desired
behaviour since if the end app user relies on it for malloc()ed
objects or stack objects, it may not work as intended.

Instead of having the autoreset functionality being a direct
application-facing implicit UAPI, make the UMD explicitly choose
this behaviour if it wants to expose it by introducing
DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET, and add a semantics
description.

v2:
- Kerneldoc fixes. Fix a commit log message.

Fixes: a2eb8aec3ebe ("drm/xe: Reset VMA attributes to default in SVM garbage collector")
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Cc: "Falkowski, John" <john.falkowski@intel.com>
Cc: "Mrozek, Michal" <michal.mrozek@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Himal Prasad Ghimiray <himal.prasad.ghimiray@intel.com>
Link: https://lore.kernel.org/r/20251015170726.178685-2-thomas.hellstrom@linux.intel.com
---
 drivers/gpu/drm/xe/xe_svm.c      |  5 +++++
 drivers/gpu/drm/xe/xe_vm.c       | 12 +++++++++---
 drivers/gpu/drm/xe/xe_vm_types.h |  1 +
 include/uapi/drm/xe_drm.h        | 15 +++++++++++++++
 4 files changed, 30 insertions(+), 3 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index da2a412f80c0..129e7818565c 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -302,6 +302,11 @@ static int xe_svm_range_set_default_attr(struct xe_vm *vm, u64 range_start, u64
 	if (!vma)
 		return -EINVAL;
 
+	if (!(vma->gpuva.flags & XE_VMA_MADV_AUTORESET)) {
+		drm_dbg(&vm->xe->drm, "Skipping madvise reset for vma.\n");
+		return 0;
+	}
+
 	if (xe_vma_has_default_mem_attrs(vma))
 		return 0;
 
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index c3230d3f9e6f..10d77666a425 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -646,7 +646,8 @@ static void xe_vma_ops_incr_pt_update_ops(struct xe_vma_ops *vops, u8 tile_mask,
 	XE_VMA_READ_ONLY |		    \
 	XE_VMA_DUMPABLE |		    \
 	XE_VMA_SYSTEM_ALLOCATOR |           \
-	DRM_GPUVA_SPARSE)
+	DRM_GPUVA_SPARSE |		    \
+	XE_VMA_MADV_AUTORESET)
 
 static void xe_vm_populate_rebind(struct xe_vma_op *op, struct xe_vma *vma,
 				  u8 tile_mask)
@@ -2297,6 +2298,8 @@ vm_bind_ioctl_ops_create(struct xe_vm *vm, struct xe_vma_ops *vops,
 				op->map.vma_flags |= XE_VMA_SYSTEM_ALLOCATOR;
 			if (flags & DRM_XE_VM_BIND_FLAG_DUMPABLE)
 				op->map.vma_flags |= XE_VMA_DUMPABLE;
+			if (flags & DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET)
+				op->map.vma_flags |= XE_VMA_MADV_AUTORESET;
 			op->map.pat_index = pat_index;
 			op->map.invalidate_on_bind =
 				__xe_vm_needs_clear_scratch_pages(vm, flags);
@@ -3280,7 +3283,8 @@ ALLOW_ERROR_INJECTION(vm_bind_ioctl_ops_execute, ERRNO);
 	 DRM_XE_VM_BIND_FLAG_NULL | \
 	 DRM_XE_VM_BIND_FLAG_DUMPABLE | \
 	 DRM_XE_VM_BIND_FLAG_CHECK_PXP | \
-	 DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR)
+	 DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR | \
+	 DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET)
 
 #ifdef TEST_VM_OPS_ERROR
 #define SUPPORTED_FLAGS	(SUPPORTED_FLAGS_STUB | FORCE_OP_ERROR)
@@ -3395,7 +3399,9 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe, struct xe_vm *vm,
 		    XE_IOCTL_DBG(xe,  (prefetch_region != DRM_XE_CONSULT_MEM_ADVISE_PREF_LOC &&
 				       !(BIT(prefetch_region) & xe->info.mem_region_mask))) ||
 		    XE_IOCTL_DBG(xe, obj &&
-				 op == DRM_XE_VM_BIND_OP_UNMAP)) {
+				 op == DRM_XE_VM_BIND_OP_UNMAP) ||
+		    XE_IOCTL_DBG(xe, (flags & DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET) &&
+				 (!is_cpu_addr_mirror || op != DRM_XE_VM_BIND_OP_MAP))) {
 			err = -EINVAL;
 			goto free_bind_ops;
 		}
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index a3b422b27ae8..d6e2a0fdd4b3 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -46,6 +46,7 @@ struct xe_vm_pgtable_update_op;
 #define XE_VMA_PTE_COMPACT	(DRM_GPUVA_USERBITS << 7)
 #define XE_VMA_DUMPABLE		(DRM_GPUVA_USERBITS << 8)
 #define XE_VMA_SYSTEM_ALLOCATOR	(DRM_GPUVA_USERBITS << 9)
+#define XE_VMA_MADV_AUTORESET	(DRM_GPUVA_USERBITS << 10)
 
 /**
  * struct xe_vma_mem_attr - memory attributes associated with vma
diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
index 2d7945cda739..47853659a705 100644
--- a/include/uapi/drm/xe_drm.h
+++ b/include/uapi/drm/xe_drm.h
@@ -1017,6 +1017,20 @@ struct drm_xe_vm_destroy {
  *    valid on VMs with DRM_XE_VM_CREATE_FLAG_FAULT_MODE set. The CPU address
  *    mirror flag are only valid for DRM_XE_VM_BIND_OP_MAP operations, the BO
  *    handle MBZ, and the BO offset MBZ.
+ *  - %DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET - Can be used in combination with
+ *    %DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR to reset madvises when the underlying
+ *    CPU address space range is unmapped (typically with munmap(2) or brk(2)).
+ *    The madvise values set with &DRM_IOCTL_XE_MADVISE are reset to the values
+ *    that were present immediately after the &DRM_IOCTL_XE_VM_BIND.
+ *    The reset GPU virtual address range is the intersection of the range bound
+ *    using &DRM_IOCTL_XE_VM_BIND and the virtual CPU address space range
+ *    unmapped.
+ *    This functionality is present to mimic the behaviour of CPU address space
+ *    madvises set using madvise(2), which are typically reset on unmap.
+ *    Note: free(3) may or may not call munmap(2) and/or brk(2), and may thus
+ *    not invoke autoreset. Neither will stack variables going out of scope.
+ *    Therefore it's recommended to always explicitly reset the madvises when
+ *    freeing the memory backing a region used in a &DRM_IOCTL_XE_MADVISE call.
  *
  * The @prefetch_mem_region_instance for %DRM_XE_VM_BIND_OP_PREFETCH can also be:
  *  - %DRM_XE_CONSULT_MEM_ADVISE_PREF_LOC, which ensures prefetching occurs in
@@ -1123,6 +1137,7 @@ struct drm_xe_vm_bind_op {
 #define DRM_XE_VM_BIND_FLAG_DUMPABLE	(1 << 3)
 #define DRM_XE_VM_BIND_FLAG_CHECK_PXP	(1 << 4)
 #define DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR	(1 << 5)
+#define DRM_XE_VM_BIND_FLAG_MADVISE_AUTORESET	(1 << 6)
 	/** @flags: Bind flags */
 	__u32 flags;
 
-- 
cgit v1.2.3


From c002b1764e7b0ffd181f32d5103d9d0ac283beeb Mon Sep 17 00:00:00 2001
From: Matt Roper <matthew.d.roper@intel.com>
Date: Thu, 16 Oct 2025 19:26:30 -0700
Subject: drm/xe/nvl: Define NVL-S platform

Provide the basic platform definitions and PCI IDs for NVL-S.

Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Shekhar Chauhan <shekhar.chauhan@intel.com>
Reviewed-by: Gustavo Sousa <gustavo.sousa@intel.com>
Link: https://lore.kernel.org/r/20251016-xe3p-v3-11-3dd173a3097a@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
---
 drivers/gpu/drm/xe/xe_pci.c            | 12 ++++++++++++
 drivers/gpu/drm/xe/xe_platform_types.h |  1 +
 include/drm/intel/pciids.h             |  9 +++++++++
 3 files changed, 22 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
index f3ab19d3ee91..6dc5124967a1 100644
--- a/drivers/gpu/drm/xe/xe_pci.c
+++ b/drivers/gpu/drm/xe/xe_pci.c
@@ -377,6 +377,17 @@ static const struct xe_device_desc ptl_desc = {
 	.vm_max_level = 4,
 };
 
+static const struct xe_device_desc nvls_desc = {
+	PLATFORM(NOVALAKE_S),
+	.dma_mask_size = 46,
+	.has_display = true,
+	.has_flat_ccs = 1,
+	.max_gt_per_tile = 2,
+	.require_force_probe = true,
+	.va_bits = 48,
+	.vm_max_level = 4,
+};
+
 #undef PLATFORM
 __diag_pop();
 
@@ -403,6 +414,7 @@ static const struct pci_device_id pciidlist[] = {
 	INTEL_LNL_IDS(INTEL_VGA_DEVICE, &lnl_desc),
 	INTEL_BMG_IDS(INTEL_VGA_DEVICE, &bmg_desc),
 	INTEL_PTL_IDS(INTEL_VGA_DEVICE, &ptl_desc),
+	INTEL_NVLS_IDS(INTEL_VGA_DEVICE, &nvls_desc),
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, pciidlist);
diff --git a/drivers/gpu/drm/xe/xe_platform_types.h b/drivers/gpu/drm/xe/xe_platform_types.h
index 3e332214c7bb..78286285c249 100644
--- a/drivers/gpu/drm/xe/xe_platform_types.h
+++ b/drivers/gpu/drm/xe/xe_platform_types.h
@@ -24,6 +24,7 @@ enum xe_platform {
 	XE_LUNARLAKE,
 	XE_BATTLEMAGE,
 	XE_PANTHERLAKE,
+	XE_NOVALAKE_S,
 };
 
 enum xe_subplatform {
diff --git a/include/drm/intel/pciids.h b/include/drm/intel/pciids.h
index 0cd12616d621..9f095a99d6c9 100644
--- a/include/drm/intel/pciids.h
+++ b/include/drm/intel/pciids.h
@@ -884,4 +884,13 @@
 	MACRO__(0xFD80, ## __VA_ARGS__), \
 	MACRO__(0xFD81, ## __VA_ARGS__)
 
+/* NVL-S */
+#define INTEL_NVLS_IDS(MACRO__, ...) \
+	MACRO__(0xD740, ## __VA_ARGS__), \
+	MACRO__(0xD741, ## __VA_ARGS__), \
+	MACRO__(0xD742, ## __VA_ARGS__), \
+	MACRO__(0xD743, ## __VA_ARGS__), \
+	MACRO__(0xD744, ## __VA_ARGS__), \
+	MACRO__(0xD745, ## __VA_ARGS__)
+
 #endif /* __PCIIDS_H__ */
-- 
cgit v1.2.3


From 7910d69376cde30e5871970d97d1a2e360568474 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Wed, 1 Oct 2025 16:37:03 +0200
Subject: drm/client: Remove holds_console_lock parameter from suspend/resume

No caller of the client resume/suspend helpers holds the console
lock. The last such cases were removed from radeon in the patch
series at [1]. Now remove the related parameter and the TODO items.

v2:
- update placeholders for CONFIG_DRM_CLIENT=n

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patchwork.freedesktop.org/series/151624/ # [1]
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Andi Shyti <andi.shyti@linux.intel.com>
Acked-by: Danilo Krummrich <dakr@kernel.org>
Reviewed-by: Lyude Paul <lyude@redhat.com>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://lore.kernel.org/r/20251001143709.419736-1-tzimmermann@suse.de
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_device.c |  8 ++++----
 drivers/gpu/drm/clients/drm_fbdev_client.c | 14 ++++----------
 drivers/gpu/drm/clients/drm_log.c          |  4 ++--
 drivers/gpu/drm/drm_client_event.c         | 16 ++++++++--------
 drivers/gpu/drm/drm_modeset_helper.c       |  6 +++---
 drivers/gpu/drm/i915/i915_driver.c         |  6 +++---
 drivers/gpu/drm/nouveau/nouveau_display.c  |  4 ++--
 drivers/gpu/drm/radeon/radeon_device.c     |  4 ++--
 drivers/gpu/drm/xe/display/xe_display.c    |  6 +++---
 include/drm/drm_client.h                   | 14 ++------------
 include/drm/drm_client_event.h             |  8 ++++----
 11 files changed, 37 insertions(+), 53 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
index 7a899fb4de29..9c472f896498 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_device.c
@@ -5212,7 +5212,7 @@ int amdgpu_device_suspend(struct drm_device *dev, bool notify_clients)
 		dev_warn(adev->dev, "smart shift update failed\n");
 
 	if (notify_clients)
-		drm_client_dev_suspend(adev_to_drm(adev), false);
+		drm_client_dev_suspend(adev_to_drm(adev));
 
 	cancel_delayed_work_sync(&adev->delayed_init_work);
 
@@ -5346,7 +5346,7 @@ exit:
 	flush_delayed_work(&adev->delayed_init_work);
 
 	if (notify_clients)
-		drm_client_dev_resume(adev_to_drm(adev), false);
+		drm_client_dev_resume(adev_to_drm(adev));
 
 	amdgpu_ras_resume(adev);
 
@@ -5951,7 +5951,7 @@ int amdgpu_device_reinit_after_reset(struct amdgpu_reset_context *reset_context)
 				if (r)
 					goto out;
 
-				drm_client_dev_resume(adev_to_drm(tmp_adev), false);
+				drm_client_dev_resume(adev_to_drm(tmp_adev));
 
 				/*
 				 * The GPU enters bad state once faulty pages
@@ -6286,7 +6286,7 @@ static void amdgpu_device_halt_activities(struct amdgpu_device *adev,
 		 */
 		amdgpu_unregister_gpu_instance(tmp_adev);
 
-		drm_client_dev_suspend(adev_to_drm(tmp_adev), false);
+		drm_client_dev_suspend(adev_to_drm(tmp_adev));
 
 		/* disable ras on ALL IPs */
 		if (!need_emergency_restart && !amdgpu_reset_in_dpc(adev) &&
diff --git a/drivers/gpu/drm/clients/drm_fbdev_client.c b/drivers/gpu/drm/clients/drm_fbdev_client.c
index f894ba52bdb5..ec5ab9f30547 100644
--- a/drivers/gpu/drm/clients/drm_fbdev_client.c
+++ b/drivers/gpu/drm/clients/drm_fbdev_client.c
@@ -62,26 +62,20 @@ err_drm_err:
 	return ret;
 }
 
-static int drm_fbdev_client_suspend(struct drm_client_dev *client, bool holds_console_lock)
+static int drm_fbdev_client_suspend(struct drm_client_dev *client)
 {
 	struct drm_fb_helper *fb_helper = drm_fb_helper_from_client(client);
 
-	if (holds_console_lock)
-		drm_fb_helper_set_suspend(fb_helper, true);
-	else
-		drm_fb_helper_set_suspend_unlocked(fb_helper, true);
+	drm_fb_helper_set_suspend_unlocked(fb_helper, true);
 
 	return 0;
 }
 
-static int drm_fbdev_client_resume(struct drm_client_dev *client, bool holds_console_lock)
+static int drm_fbdev_client_resume(struct drm_client_dev *client)
 {
 	struct drm_fb_helper *fb_helper = drm_fb_helper_from_client(client);
 
-	if (holds_console_lock)
-		drm_fb_helper_set_suspend(fb_helper, false);
-	else
-		drm_fb_helper_set_suspend_unlocked(fb_helper, false);
+	drm_fb_helper_set_suspend_unlocked(fb_helper, false);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/clients/drm_log.c b/drivers/gpu/drm/clients/drm_log.c
index d239f1e3c456..fd8556dd58ed 100644
--- a/drivers/gpu/drm/clients/drm_log.c
+++ b/drivers/gpu/drm/clients/drm_log.c
@@ -319,7 +319,7 @@ static int drm_log_client_hotplug(struct drm_client_dev *client)
 	return 0;
 }
 
-static int drm_log_client_suspend(struct drm_client_dev *client, bool _console_lock)
+static int drm_log_client_suspend(struct drm_client_dev *client)
 {
 	struct drm_log *dlog = client_to_drm_log(client);
 
@@ -328,7 +328,7 @@ static int drm_log_client_suspend(struct drm_client_dev *client, bool _console_l
 	return 0;
 }
 
-static int drm_log_client_resume(struct drm_client_dev *client, bool _console_lock)
+static int drm_log_client_resume(struct drm_client_dev *client)
 {
 	struct drm_log *dlog = client_to_drm_log(client);
 
diff --git a/drivers/gpu/drm/drm_client_event.c b/drivers/gpu/drm/drm_client_event.c
index c83196ad8b59..c3baeb4d4e6b 100644
--- a/drivers/gpu/drm/drm_client_event.c
+++ b/drivers/gpu/drm/drm_client_event.c
@@ -122,7 +122,7 @@ void drm_client_dev_restore(struct drm_device *dev)
 	mutex_unlock(&dev->clientlist_mutex);
 }
 
-static int drm_client_suspend(struct drm_client_dev *client, bool holds_console_lock)
+static int drm_client_suspend(struct drm_client_dev *client)
 {
 	struct drm_device *dev = client->dev;
 	int ret = 0;
@@ -131,7 +131,7 @@ static int drm_client_suspend(struct drm_client_dev *client, bool holds_console_
 		return 0;
 
 	if (client->funcs && client->funcs->suspend)
-		ret = client->funcs->suspend(client, holds_console_lock);
+		ret = client->funcs->suspend(client);
 	drm_dbg_kms(dev, "%s: ret=%d\n", client->name, ret);
 
 	client->suspended = true;
@@ -139,20 +139,20 @@ static int drm_client_suspend(struct drm_client_dev *client, bool holds_console_
 	return ret;
 }
 
-void drm_client_dev_suspend(struct drm_device *dev, bool holds_console_lock)
+void drm_client_dev_suspend(struct drm_device *dev)
 {
 	struct drm_client_dev *client;
 
 	mutex_lock(&dev->clientlist_mutex);
 	list_for_each_entry(client, &dev->clientlist, list) {
 		if (!client->suspended)
-			drm_client_suspend(client, holds_console_lock);
+			drm_client_suspend(client);
 	}
 	mutex_unlock(&dev->clientlist_mutex);
 }
 EXPORT_SYMBOL(drm_client_dev_suspend);
 
-static int drm_client_resume(struct drm_client_dev *client, bool holds_console_lock)
+static int drm_client_resume(struct drm_client_dev *client)
 {
 	struct drm_device *dev = client->dev;
 	int ret = 0;
@@ -161,7 +161,7 @@ static int drm_client_resume(struct drm_client_dev *client, bool holds_console_l
 		return 0;
 
 	if (client->funcs && client->funcs->resume)
-		ret = client->funcs->resume(client, holds_console_lock);
+		ret = client->funcs->resume(client);
 	drm_dbg_kms(dev, "%s: ret=%d\n", client->name, ret);
 
 	client->suspended = false;
@@ -172,14 +172,14 @@ static int drm_client_resume(struct drm_client_dev *client, bool holds_console_l
 	return ret;
 }
 
-void drm_client_dev_resume(struct drm_device *dev, bool holds_console_lock)
+void drm_client_dev_resume(struct drm_device *dev)
 {
 	struct drm_client_dev *client;
 
 	mutex_lock(&dev->clientlist_mutex);
 	list_for_each_entry(client, &dev->clientlist, list) {
 		if  (client->suspended)
-			drm_client_resume(client, holds_console_lock);
+			drm_client_resume(client);
 	}
 	mutex_unlock(&dev->clientlist_mutex);
 }
diff --git a/drivers/gpu/drm/drm_modeset_helper.c b/drivers/gpu/drm/drm_modeset_helper.c
index 988735560570..a57f6a10ada4 100644
--- a/drivers/gpu/drm/drm_modeset_helper.c
+++ b/drivers/gpu/drm/drm_modeset_helper.c
@@ -203,10 +203,10 @@ int drm_mode_config_helper_suspend(struct drm_device *dev)
 	if (dev->mode_config.poll_enabled)
 		drm_kms_helper_poll_disable(dev);
 
-	drm_client_dev_suspend(dev, false);
+	drm_client_dev_suspend(dev);
 	state = drm_atomic_helper_suspend(dev);
 	if (IS_ERR(state)) {
-		drm_client_dev_resume(dev, false);
+		drm_client_dev_resume(dev);
 
 		/*
 		 * Don't enable polling if it was never initialized
@@ -252,7 +252,7 @@ int drm_mode_config_helper_resume(struct drm_device *dev)
 		DRM_ERROR("Failed to resume (%d)\n", ret);
 	dev->mode_config.suspend_state = NULL;
 
-	drm_client_dev_resume(dev, false);
+	drm_client_dev_resume(dev);
 
 	/*
 	 * Don't enable polling if it is not initialized
diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c
index a28c3710c4d5..89be8da79d3b 100644
--- a/drivers/gpu/drm/i915/i915_driver.c
+++ b/drivers/gpu/drm/i915/i915_driver.c
@@ -978,7 +978,7 @@ void i915_driver_shutdown(struct drm_i915_private *i915)
 	intel_runtime_pm_disable(&i915->runtime_pm);
 	intel_power_domains_disable(display);
 
-	drm_client_dev_suspend(&i915->drm, false);
+	drm_client_dev_suspend(&i915->drm);
 	if (intel_display_device_present(display)) {
 		drm_kms_helper_poll_disable(&i915->drm);
 		intel_display_driver_disable_user_access(display);
@@ -1061,7 +1061,7 @@ static int i915_drm_suspend(struct drm_device *dev)
 	/* We do a lot of poking in a lot of registers, make sure they work
 	 * properly. */
 	intel_power_domains_disable(display);
-	drm_client_dev_suspend(dev, false);
+	drm_client_dev_suspend(dev);
 	if (intel_display_device_present(display)) {
 		drm_kms_helper_poll_disable(dev);
 		intel_display_driver_disable_user_access(display);
@@ -1245,7 +1245,7 @@ static int i915_drm_resume(struct drm_device *dev)
 
 	intel_opregion_resume(display);
 
-	drm_client_dev_resume(dev, false);
+	drm_client_dev_resume(dev);
 
 	intel_power_domains_enable(display);
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_display.c b/drivers/gpu/drm/nouveau/nouveau_display.c
index 54aed3656a4c..00515623a2cc 100644
--- a/drivers/gpu/drm/nouveau/nouveau_display.c
+++ b/drivers/gpu/drm/nouveau/nouveau_display.c
@@ -765,7 +765,7 @@ nouveau_display_suspend(struct drm_device *dev, bool runtime)
 {
 	struct nouveau_display *disp = nouveau_display(dev);
 
-	drm_client_dev_suspend(dev, false);
+	drm_client_dev_suspend(dev);
 
 	if (drm_drv_uses_atomic_modeset(dev)) {
 		if (!runtime) {
@@ -796,7 +796,7 @@ nouveau_display_resume(struct drm_device *dev, bool runtime)
 		}
 	}
 
-	drm_client_dev_resume(dev, false);
+	drm_client_dev_resume(dev);
 }
 
 int
diff --git a/drivers/gpu/drm/radeon/radeon_device.c b/drivers/gpu/drm/radeon/radeon_device.c
index 9e35b14e2bf0..60afaa8e56b4 100644
--- a/drivers/gpu/drm/radeon/radeon_device.c
+++ b/drivers/gpu/drm/radeon/radeon_device.c
@@ -1635,7 +1635,7 @@ int radeon_suspend_kms(struct drm_device *dev, bool suspend,
 	}
 
 	if (notify_clients)
-		drm_client_dev_suspend(dev, false);
+		drm_client_dev_suspend(dev);
 
 	return 0;
 }
@@ -1739,7 +1739,7 @@ int radeon_resume_kms(struct drm_device *dev, bool resume, bool notify_clients)
 		radeon_pm_compute_clocks(rdev);
 
 	if (notify_clients)
-		drm_client_dev_resume(dev, false);
+		drm_client_dev_resume(dev);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c
index 19e691fccf8c..d3cc6181842c 100644
--- a/drivers/gpu/drm/xe/display/xe_display.c
+++ b/drivers/gpu/drm/xe/display/xe_display.c
@@ -324,7 +324,7 @@ void xe_display_pm_suspend(struct xe_device *xe)
 	 * properly.
 	 */
 	intel_power_domains_disable(display);
-	drm_client_dev_suspend(&xe->drm, false);
+	drm_client_dev_suspend(&xe->drm);
 
 	if (intel_display_device_present(display)) {
 		drm_kms_helper_poll_disable(&xe->drm);
@@ -356,7 +356,7 @@ void xe_display_pm_shutdown(struct xe_device *xe)
 		return;
 
 	intel_power_domains_disable(display);
-	drm_client_dev_suspend(&xe->drm, false);
+	drm_client_dev_suspend(&xe->drm);
 
 	if (intel_display_device_present(display)) {
 		drm_kms_helper_poll_disable(&xe->drm);
@@ -481,7 +481,7 @@ void xe_display_pm_resume(struct xe_device *xe)
 
 	intel_opregion_resume(display);
 
-	drm_client_dev_resume(&xe->drm, false);
+	drm_client_dev_resume(&xe->drm);
 
 	intel_power_domains_enable(display);
 }
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index bdd845e383ef..3556928d3938 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -70,13 +70,8 @@ struct drm_client_funcs {
 	 * Called when suspending the device.
 	 *
 	 * This callback is optional.
-	 *
-	 * FIXME: Some callers hold the console lock when invoking this
-	 *        function. This interferes with fbdev emulation, which
-	 *        also tries to acquire the lock. Push the console lock
-	 *        into the callback and remove 'holds_console_lock'.
 	 */
-	int (*suspend)(struct drm_client_dev *client, bool holds_console_lock);
+	int (*suspend)(struct drm_client_dev *client);
 
 	/**
 	 * @resume:
@@ -84,13 +79,8 @@ struct drm_client_funcs {
 	 * Called when resuming the device from suspend.
 	 *
 	 * This callback is optional.
-	 *
-	 * FIXME: Some callers hold the console lock when invoking this
-	 *        function. This interferes with fbdev emulation, which
-	 *        also tries to acquire the lock. Push the console lock
-	 *        into the callback and remove 'holds_console_lock'.
 	 */
-	int (*resume)(struct drm_client_dev *client, bool holds_console_lock);
+	int (*resume)(struct drm_client_dev *client);
 };
 
 /**
diff --git a/include/drm/drm_client_event.h b/include/drm/drm_client_event.h
index 1d544d3a3228..985d6f02a4c4 100644
--- a/include/drm/drm_client_event.h
+++ b/include/drm/drm_client_event.h
@@ -11,8 +11,8 @@ struct drm_device;
 void drm_client_dev_unregister(struct drm_device *dev);
 void drm_client_dev_hotplug(struct drm_device *dev);
 void drm_client_dev_restore(struct drm_device *dev);
-void drm_client_dev_suspend(struct drm_device *dev, bool holds_console_lock);
-void drm_client_dev_resume(struct drm_device *dev, bool holds_console_lock);
+void drm_client_dev_suspend(struct drm_device *dev);
+void drm_client_dev_resume(struct drm_device *dev);
 #else
 static inline void drm_client_dev_unregister(struct drm_device *dev)
 { }
@@ -20,9 +20,9 @@ static inline void drm_client_dev_hotplug(struct drm_device *dev)
 { }
 static inline void drm_client_dev_restore(struct drm_device *dev)
 { }
-static inline void drm_client_dev_suspend(struct drm_device *dev, bool holds_console_lock)
+static inline void drm_client_dev_suspend(struct drm_device *dev)
 { }
-static inline void drm_client_dev_resume(struct drm_device *dev, bool holds_console_lock)
+static inline void drm_client_dev_resume(struct drm_device *dev)
 { }
 #endif
 
-- 
cgit v1.2.3


From 8b5690d5c942a80535f095c8965028262d4ab0f7 Mon Sep 17 00:00:00 2001
From: Maxime Ripard <mripard@kernel.org>
Date: Mon, 13 Oct 2025 10:35:17 +0200
Subject: dma-buf: heaps: cma: Register list of CMA regions at boot

In order to create a CMA heap instance for each CMA region found in the
system, we need to register each of these instances.

While it would appear trivial, the CMA regions are created super early
in the kernel boot process, before most of the subsystems are
initialized. Thus, we can't just create an exported function to create a
heap from the CMA region being initialized.

What we can do however is create a two-step process, where we collect
all the CMA regions into an array early on, and then when we initialize
the heaps we iterate over that array and create the heaps from the CMA
regions we collected.

Reviewed-by: T.J. Mercier <tjmercier@google.com>
Signed-off-by: Maxime Ripard <mripard@kernel.org>
Signed-off-by: Sumit Semwal <sumit.semwal@linaro.org>
Link: https://lore.kernel.org/r/20251013-dma-buf-ecc-heap-v8-2-04ce150ea3d9@kernel.org
---
 MAINTAINERS                       |  1 +
 drivers/dma-buf/heaps/cma_heap.c  | 14 ++++++++++++++
 include/linux/dma-buf/heaps/cma.h | 16 ++++++++++++++++
 3 files changed, 31 insertions(+)
 create mode 100644 include/linux/dma-buf/heaps/cma.h

(limited to 'include')

diff --git a/MAINTAINERS b/MAINTAINERS
index f9f985c7d747..cbfbdf722dc8 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -7308,6 +7308,7 @@ F:	Documentation/userspace-api/dma-buf-alloc-exchange.rst
 F:	drivers/dma-buf/
 F:	include/linux/*fence.h
 F:	include/linux/dma-buf.h
+F:	include/linux/dma-buf/
 F:	include/linux/dma-resv.h
 K:	\bdma_(?:buf|fence|resv)\b
 
diff --git a/drivers/dma-buf/heaps/cma_heap.c b/drivers/dma-buf/heaps/cma_heap.c
index 0df007111975..2a901af635ed 100644
--- a/drivers/dma-buf/heaps/cma_heap.c
+++ b/drivers/dma-buf/heaps/cma_heap.c
@@ -14,6 +14,7 @@
 
 #include <linux/cma.h>
 #include <linux/dma-buf.h>
+#include <linux/dma-buf/heaps/cma.h>
 #include <linux/dma-heap.h>
 #include <linux/dma-map-ops.h>
 #include <linux/err.h>
@@ -27,6 +28,19 @@
 
 #define DEFAULT_CMA_NAME "default_cma_region"
 
+static struct cma *dma_areas[MAX_CMA_AREAS] __initdata;
+static unsigned int dma_areas_num __initdata;
+
+int __init dma_heap_cma_register_heap(struct cma *cma)
+{
+	if (dma_areas_num >= ARRAY_SIZE(dma_areas))
+		return -EINVAL;
+
+	dma_areas[dma_areas_num++] = cma;
+
+	return 0;
+}
+
 struct cma_heap {
 	struct dma_heap *heap;
 	struct cma *cma;
diff --git a/include/linux/dma-buf/heaps/cma.h b/include/linux/dma-buf/heaps/cma.h
new file mode 100644
index 000000000000..e751479e21e7
--- /dev/null
+++ b/include/linux/dma-buf/heaps/cma.h
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+#ifndef DMA_BUF_HEAP_CMA_H_
+#define DMA_BUF_HEAP_CMA_H_
+
+struct cma;
+
+#ifdef CONFIG_DMABUF_HEAPS_CMA
+int dma_heap_cma_register_heap(struct cma *cma);
+#else
+static inline int dma_heap_cma_register_heap(struct cma *cma)
+{
+	return 0;
+}
+#endif // CONFIG_DMABUF_HEAPS_CMA
+
+#endif // DMA_BUF_HEAP_CMA_H_
-- 
cgit v1.2.3


From 8e4865faf7a97de2a0fd797556a62b31528b42bc Mon Sep 17 00:00:00 2001
From: Alice Ryhl <aliceryhl@google.com>
Date: Mon, 6 Oct 2025 12:05:55 +0000
Subject: drm/gpuvm: add deferred vm_bo cleanup

When using GPUVM in immediate mode, it is necessary to call
drm_gpuvm_unlink() from the fence signalling critical path. However,
unlink may call drm_gpuvm_bo_put(), which causes some challenges:

1. drm_gpuvm_bo_put() often requires you to take resv locks, which you
   can't do from the fence signalling critical path.
2. drm_gpuvm_bo_put() calls drm_gem_object_put(), which is often going
   to be unsafe to call from the fence signalling critical path.

To solve these issues, add a deferred version of drm_gpuvm_unlink() that
adds the vm_bo to a deferred cleanup list, and then clean it up later.

The new methods take the GEMs GPUVA lock internally rather than letting
the caller do it because it also needs to perform an operation after
releasing the mutex again. This is to prevent freeing the GEM while
holding the mutex (more info as comments in the patch). This means that
the new methods can only be used with DRM_GPUVM_IMMEDIATE_MODE.

Reviewed-by: Boris Brezillon <boris.brezillon@collabora.com>
Acked-by: Danilo Krummrich <dakr@kernel.org>
Link: https://lore.kernel.org/r/20251006-vmbo-defer-v4-1-30cbd2c05adb@google.com
[aliceryhl: fix formatting of vm_bo = llist_entry(...) line]
Signed-off-by: Alice Ryhl <aliceryhl@google.com>
---
 drivers/gpu/drm/drm_gpuvm.c | 190 ++++++++++++++++++++++++++++++++++++++++++++
 include/drm/drm_gpuvm.h     |  16 ++++
 2 files changed, 206 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_gpuvm.c b/drivers/gpu/drm/drm_gpuvm.c
index af63f4d00315..936e6c1a60c1 100644
--- a/drivers/gpu/drm/drm_gpuvm.c
+++ b/drivers/gpu/drm/drm_gpuvm.c
@@ -876,6 +876,31 @@ __drm_gpuvm_bo_list_add(struct drm_gpuvm *gpuvm, spinlock_t *lock,
 	cond_spin_unlock(lock, !!lock);
 }
 
+/**
+ * drm_gpuvm_bo_is_zombie() - check whether this vm_bo is scheduled for cleanup
+ * @vm_bo: the &drm_gpuvm_bo
+ *
+ * When a vm_bo is scheduled for cleanup using the bo_defer list, it is not
+ * immediately removed from the evict and extobj lists. Therefore, anyone
+ * iterating these lists should skip entries that are being destroyed.
+ *
+ * Checking the refcount without incrementing it is okay as long as the lock
+ * protecting the evict/extobj list is held for as long as you are using the
+ * vm_bo, because even if the refcount hits zero while you are using it, freeing
+ * the vm_bo requires taking the list's lock.
+ *
+ * Zombie entries can be observed on the evict and extobj lists regardless of
+ * whether DRM_GPUVM_RESV_PROTECTED is used, but they remain on the lists for a
+ * longer time when the resv lock is used because we can't take the resv lock
+ * during run_job() in immediate mode, meaning that they need to remain on the
+ * lists until drm_gpuvm_bo_deferred_cleanup() is called.
+ */
+static bool
+drm_gpuvm_bo_is_zombie(struct drm_gpuvm_bo *vm_bo)
+{
+	return !kref_read(&vm_bo->kref);
+}
+
 /**
  * drm_gpuvm_bo_list_add() - insert a vm_bo into the given list
  * @__vm_bo: the &drm_gpuvm_bo
@@ -1081,6 +1106,8 @@ drm_gpuvm_init(struct drm_gpuvm *gpuvm, const char *name,
 	INIT_LIST_HEAD(&gpuvm->evict.list);
 	spin_lock_init(&gpuvm->evict.lock);
 
+	init_llist_head(&gpuvm->bo_defer);
+
 	kref_init(&gpuvm->kref);
 
 	gpuvm->name = name ? name : "unknown";
@@ -1122,6 +1149,8 @@ drm_gpuvm_fini(struct drm_gpuvm *gpuvm)
 		 "Extobj list should be empty.\n");
 	drm_WARN(gpuvm->drm, !list_empty(&gpuvm->evict.list),
 		 "Evict list should be empty.\n");
+	drm_WARN(gpuvm->drm, !llist_empty(&gpuvm->bo_defer),
+		 "VM BO cleanup list should be empty.\n");
 
 	drm_gem_object_put(gpuvm->r_obj);
 }
@@ -1217,6 +1246,9 @@ drm_gpuvm_prepare_objects_locked(struct drm_gpuvm *gpuvm,
 
 	drm_gpuvm_resv_assert_held(gpuvm);
 	list_for_each_entry(vm_bo, &gpuvm->extobj.list, list.entry.extobj) {
+		if (drm_gpuvm_bo_is_zombie(vm_bo))
+			continue;
+
 		ret = exec_prepare_obj(exec, vm_bo->obj, num_fences);
 		if (ret)
 			break;
@@ -1460,6 +1492,9 @@ drm_gpuvm_validate_locked(struct drm_gpuvm *gpuvm, struct drm_exec *exec)
 
 	list_for_each_entry_safe(vm_bo, next, &gpuvm->evict.list,
 				 list.entry.evict) {
+		if (drm_gpuvm_bo_is_zombie(vm_bo))
+			continue;
+
 		ret = ops->vm_bo_validate(vm_bo, exec);
 		if (ret)
 			break;
@@ -1560,6 +1595,7 @@ drm_gpuvm_bo_create(struct drm_gpuvm *gpuvm,
 
 	INIT_LIST_HEAD(&vm_bo->list.entry.extobj);
 	INIT_LIST_HEAD(&vm_bo->list.entry.evict);
+	init_llist_node(&vm_bo->list.entry.bo_defer);
 
 	return vm_bo;
 }
@@ -1621,6 +1657,126 @@ drm_gpuvm_bo_put(struct drm_gpuvm_bo *vm_bo)
 }
 EXPORT_SYMBOL_GPL(drm_gpuvm_bo_put);
 
+/*
+ * drm_gpuvm_bo_into_zombie() - called when the vm_bo becomes a zombie due to
+ * deferred cleanup
+ *
+ * If deferred cleanup is used, then this must be called right after the vm_bo
+ * refcount drops to zero. Must be called with GEM mutex held. After releasing
+ * the GEM mutex, drm_gpuvm_bo_defer_zombie_cleanup() must be called.
+ */
+static void
+drm_gpuvm_bo_into_zombie(struct kref *kref)
+{
+	struct drm_gpuvm_bo *vm_bo = container_of(kref, struct drm_gpuvm_bo,
+						  kref);
+
+	if (!drm_gpuvm_resv_protected(vm_bo->vm)) {
+		drm_gpuvm_bo_list_del(vm_bo, extobj, true);
+		drm_gpuvm_bo_list_del(vm_bo, evict, true);
+	}
+
+	list_del(&vm_bo->list.entry.gem);
+}
+
+/*
+ * drm_gpuvm_bo_defer_zombie_cleanup() - adds a new zombie vm_bo to the
+ * bo_defer list
+ *
+ * Called after drm_gpuvm_bo_into_zombie(). GEM mutex must not be held.
+ *
+ * It's important that the GEM stays alive for the duration in which we hold
+ * the mutex, but the instant we add the vm_bo to bo_defer, another thread
+ * might call drm_gpuvm_bo_deferred_cleanup() and put the GEM. Therefore, to
+ * avoid kfreeing a mutex we are holding, the GEM mutex must be released
+ * *before* calling this function.
+ */
+static void
+drm_gpuvm_bo_defer_zombie_cleanup(struct drm_gpuvm_bo *vm_bo)
+{
+	llist_add(&vm_bo->list.entry.bo_defer, &vm_bo->vm->bo_defer);
+}
+
+static void
+drm_gpuvm_bo_defer_free(struct kref *kref)
+{
+	struct drm_gpuvm_bo *vm_bo = container_of(kref, struct drm_gpuvm_bo,
+						  kref);
+
+	drm_gpuvm_bo_into_zombie(kref);
+	mutex_unlock(&vm_bo->obj->gpuva.lock);
+	drm_gpuvm_bo_defer_zombie_cleanup(vm_bo);
+}
+
+/**
+ * drm_gpuvm_bo_put_deferred() - drop a struct drm_gpuvm_bo reference with
+ * deferred cleanup
+ * @vm_bo: the &drm_gpuvm_bo to release the reference of
+ *
+ * This releases a reference to @vm_bo.
+ *
+ * This might take and release the GEMs GPUVA lock. You should call
+ * drm_gpuvm_bo_deferred_cleanup() later to complete the cleanup process.
+ *
+ * Returns: true if vm_bo is being destroyed, false otherwise.
+ */
+bool
+drm_gpuvm_bo_put_deferred(struct drm_gpuvm_bo *vm_bo)
+{
+	if (!vm_bo)
+		return false;
+
+	drm_WARN_ON(vm_bo->vm->drm, !drm_gpuvm_immediate_mode(vm_bo->vm));
+
+	return !!kref_put_mutex(&vm_bo->kref,
+				drm_gpuvm_bo_defer_free,
+				&vm_bo->obj->gpuva.lock);
+}
+EXPORT_SYMBOL_GPL(drm_gpuvm_bo_put_deferred);
+
+/**
+ * drm_gpuvm_bo_deferred_cleanup() - clean up BOs in the deferred list
+ * deferred cleanup
+ * @gpuvm: the VM to clean up
+ *
+ * Cleans up &drm_gpuvm_bo instances in the deferred cleanup list.
+ */
+void
+drm_gpuvm_bo_deferred_cleanup(struct drm_gpuvm *gpuvm)
+{
+	const struct drm_gpuvm_ops *ops = gpuvm->ops;
+	struct drm_gpuvm_bo *vm_bo;
+	struct drm_gem_object *obj;
+	struct llist_node *bo_defer;
+
+	bo_defer = llist_del_all(&gpuvm->bo_defer);
+	if (!bo_defer)
+		return;
+
+	if (drm_gpuvm_resv_protected(gpuvm)) {
+		dma_resv_lock(drm_gpuvm_resv(gpuvm), NULL);
+		llist_for_each_entry(vm_bo, bo_defer, list.entry.bo_defer) {
+			drm_gpuvm_bo_list_del(vm_bo, extobj, false);
+			drm_gpuvm_bo_list_del(vm_bo, evict, false);
+		}
+		dma_resv_unlock(drm_gpuvm_resv(gpuvm));
+	}
+
+	while (bo_defer) {
+		vm_bo = llist_entry(bo_defer, struct drm_gpuvm_bo, list.entry.bo_defer);
+		bo_defer = bo_defer->next;
+		obj = vm_bo->obj;
+		if (ops && ops->vm_bo_free)
+			ops->vm_bo_free(vm_bo);
+		else
+			kfree(vm_bo);
+
+		drm_gpuvm_put(gpuvm);
+		drm_gem_object_put(obj);
+	}
+}
+EXPORT_SYMBOL_GPL(drm_gpuvm_bo_deferred_cleanup);
+
 static struct drm_gpuvm_bo *
 __drm_gpuvm_bo_find(struct drm_gpuvm *gpuvm,
 		    struct drm_gem_object *obj)
@@ -1948,6 +2104,40 @@ drm_gpuva_unlink(struct drm_gpuva *va)
 }
 EXPORT_SYMBOL_GPL(drm_gpuva_unlink);
 
+/**
+ * drm_gpuva_unlink_defer() - unlink a &drm_gpuva with deferred vm_bo cleanup
+ * @va: the &drm_gpuva to unlink
+ *
+ * Similar to drm_gpuva_unlink(), but uses drm_gpuvm_bo_put_deferred() and takes
+ * the lock for the caller.
+ */
+void
+drm_gpuva_unlink_defer(struct drm_gpuva *va)
+{
+	struct drm_gem_object *obj = va->gem.obj;
+	struct drm_gpuvm_bo *vm_bo = va->vm_bo;
+	bool should_defer_bo;
+
+	if (unlikely(!obj))
+		return;
+
+	drm_WARN_ON(vm_bo->vm->drm, !drm_gpuvm_immediate_mode(vm_bo->vm));
+
+	mutex_lock(&obj->gpuva.lock);
+	list_del_init(&va->gem.entry);
+
+	/*
+	 * This is drm_gpuvm_bo_put_deferred() except we already hold the mutex.
+	 */
+	should_defer_bo = kref_put(&vm_bo->kref, drm_gpuvm_bo_into_zombie);
+	mutex_unlock(&obj->gpuva.lock);
+	if (should_defer_bo)
+		drm_gpuvm_bo_defer_zombie_cleanup(vm_bo);
+
+	va->vm_bo = NULL;
+}
+EXPORT_SYMBOL_GPL(drm_gpuva_unlink_defer);
+
 /**
  * drm_gpuva_find_first() - find the first &drm_gpuva in the given range
  * @gpuvm: the &drm_gpuvm to search in
diff --git a/include/drm/drm_gpuvm.h b/include/drm/drm_gpuvm.h
index 8890ded1d907..81cc7672cf2d 100644
--- a/include/drm/drm_gpuvm.h
+++ b/include/drm/drm_gpuvm.h
@@ -27,6 +27,7 @@
 
 #include <linux/dma-resv.h>
 #include <linux/list.h>
+#include <linux/llist.h>
 #include <linux/rbtree.h>
 #include <linux/types.h>
 
@@ -152,6 +153,7 @@ void drm_gpuva_remove(struct drm_gpuva *va);
 
 void drm_gpuva_link(struct drm_gpuva *va, struct drm_gpuvm_bo *vm_bo);
 void drm_gpuva_unlink(struct drm_gpuva *va);
+void drm_gpuva_unlink_defer(struct drm_gpuva *va);
 
 struct drm_gpuva *drm_gpuva_find(struct drm_gpuvm *gpuvm,
 				 u64 addr, u64 range);
@@ -331,6 +333,11 @@ struct drm_gpuvm {
 		 */
 		spinlock_t lock;
 	} evict;
+
+	/**
+	 * @bo_defer: structure holding vm_bos that need to be destroyed
+	 */
+	struct llist_head bo_defer;
 };
 
 void drm_gpuvm_init(struct drm_gpuvm *gpuvm, const char *name,
@@ -714,6 +721,12 @@ struct drm_gpuvm_bo {
 			 * &drm_gpuvms evict list.
 			 */
 			struct list_head evict;
+
+			/**
+			 * @list.entry.bo_defer: List entry to attach to
+			 * the &drm_gpuvms bo_defer list.
+			 */
+			struct llist_node bo_defer;
 		} entry;
 	} list;
 };
@@ -746,6 +759,9 @@ drm_gpuvm_bo_get(struct drm_gpuvm_bo *vm_bo)
 
 bool drm_gpuvm_bo_put(struct drm_gpuvm_bo *vm_bo);
 
+bool drm_gpuvm_bo_put_deferred(struct drm_gpuvm_bo *vm_bo);
+void drm_gpuvm_bo_deferred_cleanup(struct drm_gpuvm *gpuvm);
+
 struct drm_gpuvm_bo *
 drm_gpuvm_bo_find(struct drm_gpuvm *gpuvm,
 		  struct drm_gem_object *obj);
-- 
cgit v1.2.3


From 5e0de2dfbc1bddbd47fbcc9dabb4917000b0ca27 Mon Sep 17 00:00:00 2001
From: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Date: Tue, 21 Oct 2025 22:17:33 -0700
Subject: drm/xe/cri: Add CRI platform definition

Add platform definition and PCI IDs for Crescent Island.

Other platforms use INTEL_VGA_DEVICE since they have a
PCI_BASE_CLASS_DISPLAY class.  This is not the case for CRI, so just
match on devid, which should be sufficient.

Signed-off-by: Balasubramani Vivekanandan <balasubramani.vivekanandan@intel.com>
Reviewed-by: Shekhar Chauhan <shekhar.chauhan@intel.com>
Link: https://lore.kernel.org/r/20251021-cri-v1-1-bf11e61d9f49@intel.com
Signed-off-by: Lucas De Marchi <lucas.demarchi@intel.com>
---
 drivers/gpu/drm/xe/xe_pci.c            | 15 +++++++++++++++
 drivers/gpu/drm/xe/xe_platform_types.h |  1 +
 include/drm/intel/pciids.h             |  4 ++++
 3 files changed, 20 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/xe/xe_pci.c b/drivers/gpu/drm/xe/xe_pci.c
index ece45157fd31..6e59642e7820 100644
--- a/drivers/gpu/drm/xe/xe_pci.c
+++ b/drivers/gpu/drm/xe/xe_pci.c
@@ -400,6 +400,20 @@ static const struct xe_device_desc nvls_desc = {
 	.vm_max_level = 4,
 };
 
+static const struct xe_device_desc cri_desc = {
+	DGFX_FEATURES,
+	PLATFORM(CRESCENTISLAND),
+	.dma_mask_size = 52,
+	.has_display = false,
+	.has_flat_ccs = false,
+	.has_mbx_power_limits = true,
+	.has_sriov = true,
+	.max_gt_per_tile = 2,
+	.require_force_probe = true,
+	.va_bits = 57,
+	.vm_max_level = 4,
+};
+
 #undef PLATFORM
 __diag_pop();
 
@@ -427,6 +441,7 @@ static const struct pci_device_id pciidlist[] = {
 	INTEL_BMG_IDS(INTEL_VGA_DEVICE, &bmg_desc),
 	INTEL_PTL_IDS(INTEL_VGA_DEVICE, &ptl_desc),
 	INTEL_NVLS_IDS(INTEL_VGA_DEVICE, &nvls_desc),
+	INTEL_CRI_IDS(INTEL_PCI_DEVICE, &cri_desc),
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, pciidlist);
diff --git a/drivers/gpu/drm/xe/xe_platform_types.h b/drivers/gpu/drm/xe/xe_platform_types.h
index 78286285c249..f516dbddfd88 100644
--- a/drivers/gpu/drm/xe/xe_platform_types.h
+++ b/drivers/gpu/drm/xe/xe_platform_types.h
@@ -25,6 +25,7 @@ enum xe_platform {
 	XE_BATTLEMAGE,
 	XE_PANTHERLAKE,
 	XE_NOVALAKE_S,
+	XE_CRESCENTISLAND,
 };
 
 enum xe_subplatform {
diff --git a/include/drm/intel/pciids.h b/include/drm/intel/pciids.h
index 9f095a99d6c9..6e53fb4cdd37 100644
--- a/include/drm/intel/pciids.h
+++ b/include/drm/intel/pciids.h
@@ -893,4 +893,8 @@
 	MACRO__(0xD744, ## __VA_ARGS__), \
 	MACRO__(0xD745, ## __VA_ARGS__)
 
+/* CRI */
+#define INTEL_CRI_IDS(MACRO__, ...) \
+	MACRO__(0x674C, ## __VA_ARGS__)
+
 #endif /* __PCIIDS_H__ */
-- 
cgit v1.2.3


From a16f6ba43d9d19996ace3aa08218fa399009f4b7 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Thu, 9 Oct 2025 15:16:28 +0200
Subject: drm/client: Add client free callback to unprepare fb_helper

Add free callback to struct drm_client_funcs. Invoke function to
free the client memory as part of the release process. Implement
free for fbdev emulation.

Fbdev emulation allocates and prepares client memory in
drm_fbdev_client_setup(). The release happens in fb_destroy from
struct fb_ops. Multiple implementations of this callback exist in
the various drivers that provide an fbdev implementation. Each of
them needs to follow the implementation details of the fbdev setup
code.

Adding a free callback for the client puts the unprepare and release
of the fbdev client in a single place.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com> # core, msm
Acked-by: Tomi Valkeinen <tomi.valkeinen@ideasonboard.com> # omapdrm
Acked-by: Patrik Jakobsson <patrik.r.jakobsson@gmail.com> # gma500
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Link: https://lore.kernel.org/r/20251009132006.45834-2-tzimmermann@suse.de
---
 drivers/gpu/drm/armada/armada_fbdev.c      |  2 --
 drivers/gpu/drm/clients/drm_fbdev_client.c | 17 +++++++++++++++--
 drivers/gpu/drm/drm_client.c               |  4 ++++
 drivers/gpu/drm/drm_fbdev_dma.c            |  4 ----
 drivers/gpu/drm/drm_fbdev_shmem.c          |  2 --
 drivers/gpu/drm/drm_fbdev_ttm.c            |  2 --
 drivers/gpu/drm/exynos/exynos_drm_fbdev.c  |  2 --
 drivers/gpu/drm/gma500/fbdev.c             |  3 ---
 drivers/gpu/drm/i915/display/intel_fbdev.c |  2 --
 drivers/gpu/drm/msm/msm_fbdev.c            |  2 --
 drivers/gpu/drm/omapdrm/omap_fbdev.c       |  2 --
 drivers/gpu/drm/radeon/radeon_fbdev.c      |  2 --
 drivers/gpu/drm/tegra/fbdev.c              |  2 --
 include/drm/drm_client.h                   | 10 ++++++++++
 14 files changed, 29 insertions(+), 27 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c
index cb53cc91bafb..22e2081bfa04 100644
--- a/drivers/gpu/drm/armada/armada_fbdev.c
+++ b/drivers/gpu/drm/armada/armada_fbdev.c
@@ -28,8 +28,6 @@ static void armada_fbdev_fb_destroy(struct fb_info *info)
 	fbh->fb->funcs->destroy(fbh->fb);
 
 	drm_client_release(&fbh->client);
-	drm_fb_helper_unprepare(fbh);
-	kfree(fbh);
 }
 
 static const struct fb_ops armada_fb_ops = {
diff --git a/drivers/gpu/drm/clients/drm_fbdev_client.c b/drivers/gpu/drm/clients/drm_fbdev_client.c
index ec5ab9f30547..47e5f27eee58 100644
--- a/drivers/gpu/drm/clients/drm_fbdev_client.c
+++ b/drivers/gpu/drm/clients/drm_fbdev_client.c
@@ -13,16 +13,28 @@
  * struct drm_client_funcs
  */
 
+static void drm_fbdev_client_free(struct drm_client_dev *client)
+{
+	struct drm_fb_helper *fb_helper = drm_fb_helper_from_client(client);
+
+	drm_fb_helper_unprepare(fb_helper);
+	kfree(fb_helper);
+}
+
 static void drm_fbdev_client_unregister(struct drm_client_dev *client)
 {
 	struct drm_fb_helper *fb_helper = drm_fb_helper_from_client(client);
 
 	if (fb_helper->info) {
+		/*
+		 * Fully probed framebuffer device
+		 */
 		drm_fb_helper_unregister_info(fb_helper);
 	} else {
+		/*
+		 * Partially initialized client, no framebuffer device yet
+		 */
 		drm_client_release(&fb_helper->client);
-		drm_fb_helper_unprepare(fb_helper);
-		kfree(fb_helper);
 	}
 }
 
@@ -82,6 +94,7 @@ static int drm_fbdev_client_resume(struct drm_client_dev *client)
 
 static const struct drm_client_funcs drm_fbdev_client_funcs = {
 	.owner		= THIS_MODULE,
+	.free		= drm_fbdev_client_free,
 	.unregister	= drm_fbdev_client_unregister,
 	.restore	= drm_fbdev_client_restore,
 	.hotplug	= drm_fbdev_client_hotplug,
diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index 3fa38d4ac70b..fe9c6d7083ea 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -168,6 +168,10 @@ void drm_client_release(struct drm_client_dev *client)
 
 	drm_client_modeset_free(client);
 	drm_client_close(client);
+
+	if (client->funcs && client->funcs->free)
+		client->funcs->free(client);
+
 	drm_dev_put(dev);
 }
 EXPORT_SYMBOL(drm_client_release);
diff --git a/drivers/gpu/drm/drm_fbdev_dma.c b/drivers/gpu/drm/drm_fbdev_dma.c
index 8bd626ef16c7..c6196293e424 100644
--- a/drivers/gpu/drm/drm_fbdev_dma.c
+++ b/drivers/gpu/drm/drm_fbdev_dma.c
@@ -57,8 +57,6 @@ static void drm_fbdev_dma_fb_destroy(struct fb_info *info)
 	drm_client_buffer_vunmap(fb_helper->buffer);
 	drm_client_framebuffer_delete(fb_helper->buffer);
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops drm_fbdev_dma_fb_ops = {
@@ -92,8 +90,6 @@ static void drm_fbdev_dma_shadowed_fb_destroy(struct fb_info *info)
 	drm_client_buffer_vunmap(fb_helper->buffer);
 	drm_client_framebuffer_delete(fb_helper->buffer);
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops drm_fbdev_dma_shadowed_fb_ops = {
diff --git a/drivers/gpu/drm/drm_fbdev_shmem.c b/drivers/gpu/drm/drm_fbdev_shmem.c
index 1e827bf8b815..51573058df6f 100644
--- a/drivers/gpu/drm/drm_fbdev_shmem.c
+++ b/drivers/gpu/drm/drm_fbdev_shmem.c
@@ -65,8 +65,6 @@ static void drm_fbdev_shmem_fb_destroy(struct fb_info *info)
 	drm_client_buffer_vunmap(fb_helper->buffer);
 	drm_client_framebuffer_delete(fb_helper->buffer);
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops drm_fbdev_shmem_fb_ops = {
diff --git a/drivers/gpu/drm/drm_fbdev_ttm.c b/drivers/gpu/drm/drm_fbdev_ttm.c
index 85feb55bba11..ccf460fbc1f0 100644
--- a/drivers/gpu/drm/drm_fbdev_ttm.c
+++ b/drivers/gpu/drm/drm_fbdev_ttm.c
@@ -53,8 +53,6 @@ static void drm_fbdev_ttm_fb_destroy(struct fb_info *info)
 	drm_client_framebuffer_delete(fb_helper->buffer);
 
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops drm_fbdev_ttm_fb_ops = {
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index 93de25b77e68..a3bd21a827ad 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -42,8 +42,6 @@ static void exynos_drm_fb_destroy(struct fb_info *info)
 	drm_framebuffer_remove(fb);
 
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops exynos_drm_fb_ops = {
diff --git a/drivers/gpu/drm/gma500/fbdev.c b/drivers/gpu/drm/gma500/fbdev.c
index a6af21514cff..bc92fa24a1e2 100644
--- a/drivers/gpu/drm/gma500/fbdev.c
+++ b/drivers/gpu/drm/gma500/fbdev.c
@@ -84,9 +84,6 @@ static void psb_fbdev_fb_destroy(struct fb_info *info)
 	drm_gem_object_put(obj);
 
 	drm_client_release(&fb_helper->client);
-
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops psb_fbdev_fb_ops = {
diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.c b/drivers/gpu/drm/i915/display/intel_fbdev.c
index 7c4709d58aa3..bf5721856f3c 100644
--- a/drivers/gpu/drm/i915/display/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/display/intel_fbdev.c
@@ -146,8 +146,6 @@ static void intel_fbdev_fb_destroy(struct fb_info *info)
 	drm_framebuffer_remove(fb_helper->fb);
 
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 __diag_push();
diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c
index b5969374d53f..aad6fb77f0de 100644
--- a/drivers/gpu/drm/msm/msm_fbdev.c
+++ b/drivers/gpu/drm/msm/msm_fbdev.c
@@ -52,8 +52,6 @@ static void msm_fbdev_fb_destroy(struct fb_info *info)
 	drm_framebuffer_remove(fb);
 
 	drm_client_release(&helper->client);
-	drm_fb_helper_unprepare(helper);
-	kfree(helper);
 }
 
 static const struct fb_ops msm_fb_ops = {
diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
index 948af7ec1130..b5df2923d2a6 100644
--- a/drivers/gpu/drm/omapdrm/omap_fbdev.c
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -103,8 +103,6 @@ static void omap_fbdev_fb_destroy(struct fb_info *info)
 	drm_framebuffer_remove(fb);
 
 	drm_client_release(&helper->client);
-	drm_fb_helper_unprepare(helper);
-	kfree(helper);
 }
 
 /*
diff --git a/drivers/gpu/drm/radeon/radeon_fbdev.c b/drivers/gpu/drm/radeon/radeon_fbdev.c
index dc81b0c2dbff..4df6c9167bf0 100644
--- a/drivers/gpu/drm/radeon/radeon_fbdev.c
+++ b/drivers/gpu/drm/radeon/radeon_fbdev.c
@@ -184,8 +184,6 @@ static void radeon_fbdev_fb_destroy(struct fb_info *info)
 	radeon_fbdev_destroy_pinned_object(gobj);
 
 	drm_client_release(&fb_helper->client);
-	drm_fb_helper_unprepare(fb_helper);
-	kfree(fb_helper);
 }
 
 static const struct fb_ops radeon_fbdev_fb_ops = {
diff --git a/drivers/gpu/drm/tegra/fbdev.c b/drivers/gpu/drm/tegra/fbdev.c
index 1b70f5e164af..91aece6f34e0 100644
--- a/drivers/gpu/drm/tegra/fbdev.c
+++ b/drivers/gpu/drm/tegra/fbdev.c
@@ -53,8 +53,6 @@ static void tegra_fbdev_fb_destroy(struct fb_info *info)
 	drm_framebuffer_remove(fb);
 
 	drm_client_release(&helper->client);
-	drm_fb_helper_unprepare(helper);
-	kfree(helper);
 }
 
 static const struct fb_ops tegra_fb_ops = {
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index 3556928d3938..715b422952ee 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -28,6 +28,16 @@ struct drm_client_funcs {
 	 */
 	struct module *owner;
 
+	/**
+	 * @free:
+	 *
+	 * Called when the client gets unregistered. Implementations should
+	 * release all client-specific data and free the memory.
+	 *
+	 * This callback is optional.
+	 */
+	void (*free)(struct drm_client_dev *client);
+
 	/**
 	 * @unregister:
 	 *
-- 
cgit v1.2.3


From 5a5e9c0228e613f0ef2a58b9782d7c0ea8f1e58b Mon Sep 17 00:00:00 2001
From: "Rob Herring (Arm)" <robh@kernel.org>
Date: Mon, 20 Oct 2025 14:33:28 -0500
Subject: accel: Add Arm Ethos-U NPU driver

Add a driver for Arm Ethos-U65/U85 NPUs. The Ethos-U NPU has a
relatively simple interface with single command stream to describe
buffers, operation settings, and network operations. It supports up to 8
memory regions (though no h/w bounds on a region). The Ethos NPUs
are designed to use an SRAM for scratch memory. Region 2 is reserved
for SRAM (like the downstream driver stack and compiler). Userspace
doesn't need access to the SRAM.

The h/w has no MMU nor external IOMMU and is a DMA engine which can
read and write anywhere in memory without h/w bounds checks. The user
submitted command streams must be validated against the bounds of the
GEM BOs. This is similar to the VC4 design which validates shaders.

The job submit is based on the rocket driver for the Rockchip NPU
utilizing the GPU scheduler. It is simpler as there's only 1 core rather
than 3.

Tested on i.MX93 platform (U65) and FVP (U85) with Mesa Teflon
support.

Acked-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Tomeu Vizoso <tomeu@tomeuvizoso.net>
Reviewed-by: Frank Li <Frank.Li@nxp.com>
Link: https://patch.msgid.link/20251020-ethos-v6-2-ecebc383c4b7@kernel.org
Signed-off-by: Rob Herring (Arm) <robh@kernel.org>
---
 MAINTAINERS                          |   9 +
 drivers/accel/Kconfig                |   1 +
 drivers/accel/Makefile               |   1 +
 drivers/accel/ethosu/Kconfig         |  11 +
 drivers/accel/ethosu/Makefile        |   4 +
 drivers/accel/ethosu/ethosu_device.h | 197 ++++++++++
 drivers/accel/ethosu/ethosu_drv.c    | 403 ++++++++++++++++++++
 drivers/accel/ethosu/ethosu_drv.h    |  15 +
 drivers/accel/ethosu/ethosu_gem.c    | 704 +++++++++++++++++++++++++++++++++++
 drivers/accel/ethosu/ethosu_gem.h    |  46 +++
 drivers/accel/ethosu/ethosu_job.c    | 496 ++++++++++++++++++++++++
 drivers/accel/ethosu/ethosu_job.h    |  40 ++
 include/uapi/drm/ethosu_accel.h      | 261 +++++++++++++
 13 files changed, 2188 insertions(+)
 create mode 100644 drivers/accel/ethosu/Kconfig
 create mode 100644 drivers/accel/ethosu/Makefile
 create mode 100644 drivers/accel/ethosu/ethosu_device.h
 create mode 100644 drivers/accel/ethosu/ethosu_drv.c
 create mode 100644 drivers/accel/ethosu/ethosu_drv.h
 create mode 100644 drivers/accel/ethosu/ethosu_gem.c
 create mode 100644 drivers/accel/ethosu/ethosu_gem.h
 create mode 100644 drivers/accel/ethosu/ethosu_job.c
 create mode 100644 drivers/accel/ethosu/ethosu_job.h
 create mode 100644 include/uapi/drm/ethosu_accel.h

(limited to 'include')

diff --git a/MAINTAINERS b/MAINTAINERS
index 5f53e385a261..874fcbe59990 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -2017,6 +2017,15 @@ F:	arch/arm64/include/asm/arch_timer.h
 F:	drivers/clocksource/arm_arch_timer.c
 F:	drivers/clocksource/arm_arch_timer_mmio.c
 
+ARM ETHOS-U NPU DRIVER
+M:	Rob Herring (Arm) <robh@kernel.org>
+M:	Tomeu Vizoso <tomeu@tomeuvizoso.net>
+L:	dri-devel@lists.freedesktop.org
+S:	Supported
+T:	git https://gitlab.freedesktop.org/drm/misc/kernel.git
+F:	drivers/accel/ethosu/
+F:	include/uapi/drm/ethosu_accel.h
+
 ARM GENERIC INTERRUPT CONTROLLER DRIVERS
 M:	Marc Zyngier <maz@kernel.org>
 L:	linux-arm-kernel@lists.infradead.org (moderated for non-subscribers)
diff --git a/drivers/accel/Kconfig b/drivers/accel/Kconfig
index bb01cebc42bf..bdf48ccafcf2 100644
--- a/drivers/accel/Kconfig
+++ b/drivers/accel/Kconfig
@@ -25,6 +25,7 @@ menuconfig DRM_ACCEL
 	  and debugfs).
 
 source "drivers/accel/amdxdna/Kconfig"
+source "drivers/accel/ethosu/Kconfig"
 source "drivers/accel/habanalabs/Kconfig"
 source "drivers/accel/ivpu/Kconfig"
 source "drivers/accel/qaic/Kconfig"
diff --git a/drivers/accel/Makefile b/drivers/accel/Makefile
index ffc3fa588666..1d3a7251b950 100644
--- a/drivers/accel/Makefile
+++ b/drivers/accel/Makefile
@@ -1,6 +1,7 @@
 # SPDX-License-Identifier: GPL-2.0-only
 
 obj-$(CONFIG_DRM_ACCEL_AMDXDNA)		+= amdxdna/
+obj-$(CONFIG_DRM_ACCEL_ARM_ETHOSU)	+= ethosu/
 obj-$(CONFIG_DRM_ACCEL_HABANALABS)	+= habanalabs/
 obj-$(CONFIG_DRM_ACCEL_IVPU)		+= ivpu/
 obj-$(CONFIG_DRM_ACCEL_QAIC)		+= qaic/
diff --git a/drivers/accel/ethosu/Kconfig b/drivers/accel/ethosu/Kconfig
new file mode 100644
index 000000000000..d25f9b3eb317
--- /dev/null
+++ b/drivers/accel/ethosu/Kconfig
@@ -0,0 +1,11 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+config DRM_ACCEL_ARM_ETHOSU
+	tristate "Arm Ethos-U65/U85 NPU"
+	depends on HAS_IOMEM
+	depends on DRM_ACCEL
+	select DRM_GEM_DMA_HELPER
+	select DRM_SCHED
+	select GENERIC_ALLOCATOR
+	help
+	  Enables driver for Arm Ethos-U65/U85 NPUs
diff --git a/drivers/accel/ethosu/Makefile b/drivers/accel/ethosu/Makefile
new file mode 100644
index 000000000000..17db5a600416
--- /dev/null
+++ b/drivers/accel/ethosu/Makefile
@@ -0,0 +1,4 @@
+# SPDX-License-Identifier: GPL-2.0-only
+
+obj-$(CONFIG_DRM_ACCEL_ARM_ETHOSU) := ethosu.o
+ethosu-y += ethosu_drv.o ethosu_gem.o ethosu_job.o
diff --git a/drivers/accel/ethosu/ethosu_device.h b/drivers/accel/ethosu/ethosu_device.h
new file mode 100644
index 000000000000..b189fa783d6a
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_device.h
@@ -0,0 +1,197 @@
+/* SPDX-License-Identifier: GPL-2.0-only or MIT */
+/* Copyright 2025 Arm, Ltd. */
+
+#ifndef __ETHOSU_DEVICE_H__
+#define __ETHOSU_DEVICE_H__
+
+#include <linux/bitfield.h>
+#include <linux/bits.h>
+#include <linux/types.h>
+
+#include <drm/drm_device.h>
+#include <drm/gpu_scheduler.h>
+
+#include <drm/ethosu_accel.h>
+
+struct clk;
+struct gen_pool;
+
+#define NPU_REG_ID		0x0000
+#define NPU_REG_STATUS		0x0004
+#define NPU_REG_CMD		0x0008
+#define NPU_REG_RESET		0x000c
+#define NPU_REG_QBASE		0x0010
+#define NPU_REG_QBASE_HI	0x0014
+#define NPU_REG_QREAD		0x0018
+#define NPU_REG_QCONFIG		0x001c
+#define NPU_REG_QSIZE		0x0020
+#define NPU_REG_PROT		0x0024
+#define NPU_REG_CONFIG		0x0028
+#define NPU_REG_REGIONCFG	0x003c
+#define NPU_REG_AXILIMIT0	0x0040		// U65
+#define NPU_REG_AXILIMIT1	0x0044		// U65
+#define NPU_REG_AXILIMIT2	0x0048		// U65
+#define NPU_REG_AXILIMIT3	0x004c		// U65
+#define NPU_REG_MEM_ATTR0	0x0040		// U85
+#define NPU_REG_MEM_ATTR1	0x0044		// U85
+#define NPU_REG_MEM_ATTR2	0x0048		// U85
+#define NPU_REG_MEM_ATTR3	0x004c		// U85
+#define NPU_REG_AXI_SRAM	0x0050		// U85
+#define NPU_REG_AXI_EXT		0x0054		// U85
+
+#define NPU_REG_BASEP(x)	(0x0080 + (x) * 8)
+#define NPU_REG_BASEP_HI(x)	(0x0084 + (x) * 8)
+#define NPU_BASEP_REGION_MAX	8
+
+#define ID_ARCH_MAJOR_MASK	GENMASK(31, 28)
+#define ID_ARCH_MINOR_MASK	GENMASK(27, 20)
+#define ID_ARCH_PATCH_MASK	GENMASK(19, 16)
+#define ID_VER_MAJOR_MASK	GENMASK(11, 8)
+#define ID_VER_MINOR_MASK	GENMASK(7, 4)
+
+#define CONFIG_MACS_PER_CC_MASK	GENMASK(3, 0)
+#define CONFIG_CMD_STREAM_VER_MASK	GENMASK(7, 4)
+
+#define STATUS_STATE_RUNNING	BIT(0)
+#define STATUS_IRQ_RAISED	BIT(1)
+#define STATUS_BUS_STATUS	BIT(2)
+#define STATUS_RESET_STATUS	BIT(3)
+#define STATUS_CMD_PARSE_ERR	BIT(4)
+#define STATUS_CMD_END_REACHED	BIT(5)
+
+#define CMD_CLEAR_IRQ		BIT(1)
+#define CMD_TRANSITION_TO_RUN	BIT(0)
+
+#define RESET_PENDING_CSL	BIT(1)
+#define RESET_PENDING_CPL	BIT(0)
+
+#define PROT_ACTIVE_CSL		BIT(1)
+
+enum ethosu_cmds {
+	NPU_OP_CONV = 0x2,
+	NPU_OP_DEPTHWISE = 0x3,
+	NPU_OP_POOL = 0x5,
+	NPU_OP_ELEMENTWISE = 0x6,
+	NPU_OP_RESIZE = 0x7,	// U85 only
+	NPU_OP_DMA_START = 0x10,
+	NPU_SET_IFM_PAD_TOP = 0x100,
+	NPU_SET_IFM_PAD_LEFT = 0x101,
+	NPU_SET_IFM_PAD_RIGHT = 0x102,
+	NPU_SET_IFM_PAD_BOTTOM = 0x103,
+	NPU_SET_IFM_DEPTH_M1 = 0x104,
+	NPU_SET_IFM_PRECISION = 0x105,
+	NPU_SET_IFM_BROADCAST = 0x108,
+	NPU_SET_IFM_WIDTH0_M1 = 0x10a,
+	NPU_SET_IFM_HEIGHT0_M1 = 0x10b,
+	NPU_SET_IFM_HEIGHT1_M1 = 0x10c,
+	NPU_SET_IFM_REGION = 0x10f,
+	NPU_SET_OFM_WIDTH_M1 = 0x111,
+	NPU_SET_OFM_HEIGHT_M1 = 0x112,
+	NPU_SET_OFM_DEPTH_M1 = 0x113,
+	NPU_SET_OFM_PRECISION = 0x114,
+	NPU_SET_OFM_WIDTH0_M1 = 0x11a,
+	NPU_SET_OFM_HEIGHT0_M1 = 0x11b,
+	NPU_SET_OFM_HEIGHT1_M1 = 0x11c,
+	NPU_SET_OFM_REGION = 0x11f,
+	NPU_SET_KERNEL_WIDTH_M1 = 0x120,
+	NPU_SET_KERNEL_HEIGHT_M1 = 0x121,
+	NPU_SET_KERNEL_STRIDE = 0x122,
+	NPU_SET_WEIGHT_REGION = 0x128,
+	NPU_SET_SCALE_REGION = 0x129,
+	NPU_SET_DMA0_SRC_REGION = 0x130,
+	NPU_SET_DMA0_DST_REGION = 0x131,
+	NPU_SET_DMA0_SIZE0 = 0x132,
+	NPU_SET_DMA0_SIZE1 = 0x133,
+	NPU_SET_IFM2_BROADCAST = 0x180,
+	NPU_SET_IFM2_PRECISION = 0x185,
+	NPU_SET_IFM2_WIDTH0_M1 = 0x18a,
+	NPU_SET_IFM2_HEIGHT0_M1 = 0x18b,
+	NPU_SET_IFM2_HEIGHT1_M1 = 0x18c,
+	NPU_SET_IFM2_REGION = 0x18f,
+	NPU_SET_IFM_BASE0 = 0x4000,
+	NPU_SET_IFM_BASE1 = 0x4001,
+	NPU_SET_IFM_BASE2 = 0x4002,
+	NPU_SET_IFM_BASE3 = 0x4003,
+	NPU_SET_IFM_STRIDE_X = 0x4004,
+	NPU_SET_IFM_STRIDE_Y = 0x4005,
+	NPU_SET_IFM_STRIDE_C = 0x4006,
+	NPU_SET_OFM_BASE0 = 0x4010,
+	NPU_SET_OFM_BASE1 = 0x4011,
+	NPU_SET_OFM_BASE2 = 0x4012,
+	NPU_SET_OFM_BASE3 = 0x4013,
+	NPU_SET_OFM_STRIDE_X = 0x4014,
+	NPU_SET_OFM_STRIDE_Y = 0x4015,
+	NPU_SET_OFM_STRIDE_C = 0x4016,
+	NPU_SET_WEIGHT_BASE = 0x4020,
+	NPU_SET_WEIGHT_LENGTH = 0x4021,
+	NPU_SET_SCALE_BASE = 0x4022,
+	NPU_SET_SCALE_LENGTH = 0x4023,
+	NPU_SET_DMA0_SRC = 0x4030,
+	NPU_SET_DMA0_DST = 0x4031,
+	NPU_SET_DMA0_LEN = 0x4032,
+	NPU_SET_DMA0_SRC_STRIDE0 = 0x4033,
+	NPU_SET_DMA0_SRC_STRIDE1 = 0x4034,
+	NPU_SET_DMA0_DST_STRIDE0 = 0x4035,
+	NPU_SET_DMA0_DST_STRIDE1 = 0x4036,
+	NPU_SET_IFM2_BASE0 = 0x4080,
+	NPU_SET_IFM2_BASE1 = 0x4081,
+	NPU_SET_IFM2_BASE2 = 0x4082,
+	NPU_SET_IFM2_BASE3 = 0x4083,
+	NPU_SET_IFM2_STRIDE_X = 0x4084,
+	NPU_SET_IFM2_STRIDE_Y = 0x4085,
+	NPU_SET_IFM2_STRIDE_C = 0x4086,
+	NPU_SET_WEIGHT1_BASE = 0x4090,
+	NPU_SET_WEIGHT1_LENGTH = 0x4091,
+	NPU_SET_SCALE1_BASE = 0x4092,
+	NPU_SET_WEIGHT2_BASE = 0x4092,
+	NPU_SET_SCALE1_LENGTH = 0x4093,
+	NPU_SET_WEIGHT2_LENGTH = 0x4093,
+	NPU_SET_WEIGHT3_BASE = 0x4094,
+	NPU_SET_WEIGHT3_LENGTH = 0x4095,
+};
+
+#define ETHOSU_SRAM_REGION	2	/* Matching Vela compiler */
+
+/**
+ * struct ethosu_device - Ethosu device
+ */
+struct ethosu_device {
+	/** @base: Base drm_device. */
+	struct drm_device base;
+
+	/** @iomem: CPU mapping of the registers. */
+	void __iomem *regs;
+
+	void __iomem *sram;
+	struct gen_pool *srampool;
+	dma_addr_t sramphys;
+
+	struct clk_bulk_data *clks;
+	int num_clks;
+	int irq;
+
+	struct drm_ethosu_npu_info npu_info;
+
+	struct ethosu_job *in_flight_job;
+	/* For in_flight_job and ethosu_job_hw_submit() */
+	struct mutex job_lock;
+
+	/* For dma_fence */
+	spinlock_t fence_lock;
+
+	struct drm_gpu_scheduler sched;
+	/* For ethosu_job_do_push() */
+	struct mutex sched_lock;
+	u64 fence_context;
+	u64 emit_seqno;
+};
+
+#define to_ethosu_device(drm_dev) \
+	((struct ethosu_device *)container_of(drm_dev, struct ethosu_device, base))
+
+static inline bool ethosu_is_u65(const struct ethosu_device *ethosudev)
+{
+	return FIELD_GET(ID_ARCH_MAJOR_MASK, ethosudev->npu_info.id) == 1;
+}
+
+#endif
diff --git a/drivers/accel/ethosu/ethosu_drv.c b/drivers/accel/ethosu/ethosu_drv.c
new file mode 100644
index 000000000000..e05a69bf5574
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_drv.c
@@ -0,0 +1,403 @@
+// SPDX-License-Identifier: GPL-2.0-only or MIT
+// Copyright (C) 2025 Arm, Ltd.
+
+#include <linux/bitfield.h>
+#include <linux/clk.h>
+#include <linux/genalloc.h>
+#include <linux/io.h>
+#include <linux/iopoll.h>
+#include <linux/module.h>
+#include <linux/mod_devicetable.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include <drm/drm_drv.h>
+#include <drm/drm_ioctl.h>
+#include <drm/drm_utils.h>
+#include <drm/drm_gem.h>
+#include <drm/drm_accel.h>
+#include <drm/ethosu_accel.h>
+
+#include "ethosu_drv.h"
+#include "ethosu_device.h"
+#include "ethosu_gem.h"
+#include "ethosu_job.h"
+
+static int ethosu_ioctl_dev_query(struct drm_device *ddev, void *data,
+				  struct drm_file *file)
+{
+	struct ethosu_device *ethosudev = to_ethosu_device(ddev);
+	struct drm_ethosu_dev_query *args = data;
+
+	if (!args->pointer) {
+		switch (args->type) {
+		case DRM_ETHOSU_DEV_QUERY_NPU_INFO:
+			args->size = sizeof(ethosudev->npu_info);
+			return 0;
+		default:
+			return -EINVAL;
+		}
+	}
+
+	switch (args->type) {
+	case DRM_ETHOSU_DEV_QUERY_NPU_INFO:
+		if (args->size < offsetofend(struct drm_ethosu_npu_info, sram_size))
+			return -EINVAL;
+		return copy_struct_to_user(u64_to_user_ptr(args->pointer),
+					   args->size,
+					   &ethosudev->npu_info,
+					   sizeof(ethosudev->npu_info), NULL);
+	default:
+		return -EINVAL;
+	}
+}
+
+#define ETHOSU_BO_FLAGS		DRM_ETHOSU_BO_NO_MMAP
+
+static int ethosu_ioctl_bo_create(struct drm_device *ddev, void *data,
+				  struct drm_file *file)
+{
+	struct drm_ethosu_bo_create *args = data;
+	int cookie, ret;
+
+	if (!drm_dev_enter(ddev, &cookie))
+		return -ENODEV;
+
+	if (!args->size || (args->flags & ~ETHOSU_BO_FLAGS)) {
+		ret = -EINVAL;
+		goto out_dev_exit;
+	}
+
+	ret = ethosu_gem_create_with_handle(file, ddev, &args->size,
+					    args->flags, &args->handle);
+
+out_dev_exit:
+	drm_dev_exit(cookie);
+	return ret;
+}
+
+static int ethosu_ioctl_bo_wait(struct drm_device *ddev, void *data,
+				struct drm_file *file)
+{
+	struct drm_ethosu_bo_wait *args = data;
+	int cookie, ret;
+	unsigned long timeout = drm_timeout_abs_to_jiffies(args->timeout_ns);
+
+	if (args->pad)
+		return -EINVAL;
+
+	if (!drm_dev_enter(ddev, &cookie))
+		return -ENODEV;
+
+	ret = drm_gem_dma_resv_wait(file, args->handle, true, timeout);
+
+	drm_dev_exit(cookie);
+	return ret;
+}
+
+static int ethosu_ioctl_bo_mmap_offset(struct drm_device *ddev, void *data,
+				       struct drm_file *file)
+{
+	struct drm_ethosu_bo_mmap_offset *args = data;
+	struct drm_gem_object *obj;
+
+	if (args->pad)
+		return -EINVAL;
+
+	obj = drm_gem_object_lookup(file, args->handle);
+	if (!obj)
+		return -ENOENT;
+
+	args->offset = drm_vma_node_offset_addr(&obj->vma_node);
+	drm_gem_object_put(obj);
+	return 0;
+}
+
+static int ethosu_ioctl_cmdstream_bo_create(struct drm_device *ddev, void *data,
+					    struct drm_file *file)
+{
+	struct drm_ethosu_cmdstream_bo_create *args = data;
+	int cookie, ret;
+
+	if (!drm_dev_enter(ddev, &cookie))
+		return -ENODEV;
+
+	if (!args->size || !args->data || args->pad || args->flags) {
+		ret = -EINVAL;
+		goto out_dev_exit;
+	}
+
+	args->flags |= DRM_ETHOSU_BO_NO_MMAP;
+
+	ret = ethosu_gem_cmdstream_create(file, ddev, args->size, args->data,
+					  args->flags, &args->handle);
+
+out_dev_exit:
+	drm_dev_exit(cookie);
+	return ret;
+}
+
+static int ethosu_open(struct drm_device *ddev, struct drm_file *file)
+{
+	int ret = 0;
+
+	if (!try_module_get(THIS_MODULE))
+		return -EINVAL;
+
+	struct ethosu_file_priv __free(kfree) *priv = kzalloc(sizeof(*priv), GFP_KERNEL);
+	if (!priv) {
+		ret = -ENOMEM;
+		goto err_put_mod;
+	}
+	priv->edev = to_ethosu_device(ddev);
+
+	ret = ethosu_job_open(priv);
+	if (ret)
+		goto err_put_mod;
+
+	file->driver_priv = no_free_ptr(priv);
+	return 0;
+
+err_put_mod:
+	module_put(THIS_MODULE);
+	return ret;
+}
+
+static void ethosu_postclose(struct drm_device *ddev, struct drm_file *file)
+{
+	ethosu_job_close(file->driver_priv);
+	kfree(file->driver_priv);
+	module_put(THIS_MODULE);
+}
+
+static const struct drm_ioctl_desc ethosu_drm_driver_ioctls[] = {
+#define ETHOSU_IOCTL(n, func, flags) \
+	DRM_IOCTL_DEF_DRV(ETHOSU_##n, ethosu_ioctl_##func, flags)
+
+	ETHOSU_IOCTL(DEV_QUERY, dev_query, 0),
+	ETHOSU_IOCTL(BO_CREATE, bo_create, 0),
+	ETHOSU_IOCTL(BO_WAIT, bo_wait, 0),
+	ETHOSU_IOCTL(BO_MMAP_OFFSET, bo_mmap_offset, 0),
+	ETHOSU_IOCTL(CMDSTREAM_BO_CREATE, cmdstream_bo_create, 0),
+	ETHOSU_IOCTL(SUBMIT, submit, 0),
+};
+
+DEFINE_DRM_ACCEL_FOPS(ethosu_drm_driver_fops);
+
+/*
+ * Ethosu driver version:
+ * - 1.0 - initial interface
+ */
+static const struct drm_driver ethosu_drm_driver = {
+	.driver_features = DRIVER_COMPUTE_ACCEL | DRIVER_GEM,
+	.open = ethosu_open,
+	.postclose = ethosu_postclose,
+	.ioctls = ethosu_drm_driver_ioctls,
+	.num_ioctls = ARRAY_SIZE(ethosu_drm_driver_ioctls),
+	.fops = &ethosu_drm_driver_fops,
+	.name = "ethosu",
+	.desc = "Arm Ethos-U Accel driver",
+	.major = 1,
+	.minor = 0,
+
+	.gem_create_object = ethosu_gem_create_object,
+};
+
+#define U65_DRAM_AXI_LIMIT_CFG	0x1f3f0002
+#define U65_SRAM_AXI_LIMIT_CFG	0x1f3f00b0
+#define U85_AXI_EXT_CFG		0x00021f3f
+#define U85_AXI_SRAM_CFG	0x00021f3f
+#define U85_MEM_ATTR0_CFG	0x00000000
+#define U85_MEM_ATTR2_CFG	0x000000b7
+
+static int ethosu_reset(struct ethosu_device *ethosudev)
+{
+	int ret;
+	u32 reg;
+
+	writel_relaxed(RESET_PENDING_CSL, ethosudev->regs + NPU_REG_RESET);
+	ret = readl_poll_timeout(ethosudev->regs + NPU_REG_STATUS, reg,
+				 !FIELD_GET(STATUS_RESET_STATUS, reg),
+				 USEC_PER_MSEC, USEC_PER_SEC);
+	if (ret)
+		return ret;
+
+	if (!FIELD_GET(PROT_ACTIVE_CSL, readl_relaxed(ethosudev->regs + NPU_REG_PROT))) {
+		dev_warn(ethosudev->base.dev, "Could not reset to non-secure mode (PROT = %x)\n",
+			 readl_relaxed(ethosudev->regs + NPU_REG_PROT));
+	}
+
+	/*
+	 * Assign region 2 (SRAM) to AXI M0 (AXILIMIT0),
+	 * everything else to AXI M1 (AXILIMIT2)
+	 */
+	writel_relaxed(0x0000aa8a, ethosudev->regs + NPU_REG_REGIONCFG);
+	if (ethosu_is_u65(ethosudev)) {
+		writel_relaxed(U65_SRAM_AXI_LIMIT_CFG, ethosudev->regs + NPU_REG_AXILIMIT0);
+		writel_relaxed(U65_DRAM_AXI_LIMIT_CFG, ethosudev->regs + NPU_REG_AXILIMIT2);
+	} else {
+		writel_relaxed(U85_AXI_SRAM_CFG, ethosudev->regs + NPU_REG_AXI_SRAM);
+		writel_relaxed(U85_AXI_EXT_CFG, ethosudev->regs + NPU_REG_AXI_EXT);
+		writel_relaxed(U85_MEM_ATTR0_CFG, ethosudev->regs + NPU_REG_MEM_ATTR0);	// SRAM
+		writel_relaxed(U85_MEM_ATTR2_CFG, ethosudev->regs + NPU_REG_MEM_ATTR2);	// DRAM
+	}
+
+	if (ethosudev->sram)
+		memset_io(ethosudev->sram, 0, ethosudev->npu_info.sram_size);
+
+	return 0;
+}
+
+static int ethosu_device_resume(struct device *dev)
+{
+	struct ethosu_device *ethosudev = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_bulk_prepare_enable(ethosudev->num_clks, ethosudev->clks);
+	if (ret)
+		return ret;
+
+	ret = ethosu_reset(ethosudev);
+	if (!ret)
+		return 0;
+
+	clk_bulk_disable_unprepare(ethosudev->num_clks, ethosudev->clks);
+	return ret;
+}
+
+static int ethosu_device_suspend(struct device *dev)
+{
+	struct ethosu_device *ethosudev = dev_get_drvdata(dev);
+
+	clk_bulk_disable_unprepare(ethosudev->num_clks, ethosudev->clks);
+	return 0;
+}
+
+static int ethosu_sram_init(struct ethosu_device *ethosudev)
+{
+	ethosudev->npu_info.sram_size = 0;
+
+	ethosudev->srampool = of_gen_pool_get(ethosudev->base.dev->of_node, "sram", 0);
+	if (!ethosudev->srampool)
+		return 0;
+
+	ethosudev->npu_info.sram_size = gen_pool_size(ethosudev->srampool);
+
+	ethosudev->sram = (void __iomem *)gen_pool_dma_alloc(ethosudev->srampool,
+							     ethosudev->npu_info.sram_size,
+							     &ethosudev->sramphys);
+	if (!ethosudev->sram) {
+		dev_err(ethosudev->base.dev, "failed to allocate from SRAM pool\n");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+static int ethosu_init(struct ethosu_device *ethosudev)
+{
+	int ret;
+	u32 id, config;
+
+	ret = ethosu_device_resume(ethosudev->base.dev);
+	if (ret)
+		return ret;
+
+	pm_runtime_set_autosuspend_delay(ethosudev->base.dev, 50);
+	pm_runtime_use_autosuspend(ethosudev->base.dev);
+	ret = devm_pm_runtime_set_active_enabled(ethosudev->base.dev);
+	if (ret)
+		return ret;
+	pm_runtime_get_noresume(ethosudev->base.dev);
+
+	ethosudev->npu_info.id = id = readl_relaxed(ethosudev->regs + NPU_REG_ID);
+	ethosudev->npu_info.config = config = readl_relaxed(ethosudev->regs + NPU_REG_CONFIG);
+
+	ethosu_sram_init(ethosudev);
+
+	dev_info(ethosudev->base.dev,
+		 "Ethos-U NPU, arch v%ld.%ld.%ld, rev r%ldp%ld, cmd stream ver%ld, %d MACs, %dKB SRAM\n",
+		 FIELD_GET(ID_ARCH_MAJOR_MASK, id),
+		 FIELD_GET(ID_ARCH_MINOR_MASK, id),
+		 FIELD_GET(ID_ARCH_PATCH_MASK, id),
+		 FIELD_GET(ID_VER_MAJOR_MASK, id),
+		 FIELD_GET(ID_VER_MINOR_MASK, id),
+		 FIELD_GET(CONFIG_CMD_STREAM_VER_MASK, config),
+		 1 << FIELD_GET(CONFIG_MACS_PER_CC_MASK, config),
+		 ethosudev->npu_info.sram_size / 1024);
+
+	return 0;
+}
+
+static int ethosu_probe(struct platform_device *pdev)
+{
+	int ret;
+	struct ethosu_device *ethosudev;
+
+	ethosudev = devm_drm_dev_alloc(&pdev->dev, &ethosu_drm_driver,
+				       struct ethosu_device, base);
+	if (IS_ERR(ethosudev))
+		return -ENOMEM;
+	platform_set_drvdata(pdev, ethosudev);
+
+	dma_set_mask_and_coherent(&pdev->dev, DMA_BIT_MASK(40));
+
+	ethosudev->regs = devm_platform_ioremap_resource(pdev, 0);
+
+	ethosudev->num_clks = devm_clk_bulk_get_all(&pdev->dev, &ethosudev->clks);
+	if (ethosudev->num_clks < 0)
+		return ethosudev->num_clks;
+
+	ret = ethosu_job_init(ethosudev);
+	if (ret)
+		return ret;
+
+	ret = ethosu_init(ethosudev);
+	if (ret)
+		return ret;
+
+	ret = drm_dev_register(&ethosudev->base, 0);
+	if (ret)
+		pm_runtime_dont_use_autosuspend(ethosudev->base.dev);
+
+	pm_runtime_put_autosuspend(ethosudev->base.dev);
+	return ret;
+}
+
+static void ethosu_remove(struct platform_device *pdev)
+{
+	struct ethosu_device *ethosudev = dev_get_drvdata(&pdev->dev);
+
+	drm_dev_unregister(&ethosudev->base);
+	ethosu_job_fini(ethosudev);
+	if (ethosudev->sram)
+		gen_pool_free(ethosudev->srampool, (unsigned long)ethosudev->sram,
+			      ethosudev->npu_info.sram_size);
+}
+
+static const struct of_device_id dt_match[] = {
+	{ .compatible = "arm,ethos-u65" },
+	{ .compatible = "arm,ethos-u85" },
+	{}
+};
+MODULE_DEVICE_TABLE(of, dt_match);
+
+static DEFINE_RUNTIME_DEV_PM_OPS(ethosu_pm_ops,
+				 ethosu_device_suspend,
+				 ethosu_device_resume,
+				 NULL);
+
+static struct platform_driver ethosu_driver = {
+	.probe = ethosu_probe,
+	.remove = ethosu_remove,
+	.driver = {
+		.name = "ethosu",
+		.pm = pm_ptr(&ethosu_pm_ops),
+		.of_match_table = dt_match,
+	},
+};
+module_platform_driver(ethosu_driver);
+
+MODULE_AUTHOR("Rob Herring <robh@kernel.org>");
+MODULE_DESCRIPTION("Arm Ethos-U Accel Driver");
+MODULE_LICENSE("Dual MIT/GPL");
diff --git a/drivers/accel/ethosu/ethosu_drv.h b/drivers/accel/ethosu/ethosu_drv.h
new file mode 100644
index 000000000000..9e21dfe94184
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_drv.h
@@ -0,0 +1,15 @@
+/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
+/* Copyright 2025 Arm, Ltd. */
+#ifndef __ETHOSU_DRV_H__
+#define __ETHOSU_DRV_H__
+
+#include <drm/gpu_scheduler.h>
+
+struct ethosu_device;
+
+struct ethosu_file_priv {
+	struct ethosu_device *edev;
+	struct drm_sched_entity sched_entity;
+};
+
+#endif
diff --git a/drivers/accel/ethosu/ethosu_gem.c b/drivers/accel/ethosu/ethosu_gem.c
new file mode 100644
index 000000000000..473b5f5d7514
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_gem.c
@@ -0,0 +1,704 @@
+// SPDX-License-Identifier: GPL-2.0-only or MIT
+/* Copyright 2025 Arm, Ltd. */
+
+#include <linux/err.h>
+#include <linux/slab.h>
+
+#include <drm/ethosu_accel.h>
+
+#include "ethosu_device.h"
+#include "ethosu_gem.h"
+
+static void ethosu_gem_free_object(struct drm_gem_object *obj)
+{
+	struct ethosu_gem_object *bo = to_ethosu_bo(obj);
+
+	kfree(bo->info);
+	drm_gem_free_mmap_offset(&bo->base.base);
+	drm_gem_dma_free(&bo->base);
+}
+
+static int ethosu_gem_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
+{
+	struct ethosu_gem_object *bo = to_ethosu_bo(obj);
+
+	/* Don't allow mmap on objects that have the NO_MMAP flag set. */
+	if (bo->flags & DRM_ETHOSU_BO_NO_MMAP)
+		return -EINVAL;
+
+	return drm_gem_dma_object_mmap(obj, vma);
+}
+
+static const struct drm_gem_object_funcs ethosu_gem_funcs = {
+	.free = ethosu_gem_free_object,
+	.print_info = drm_gem_dma_object_print_info,
+	.get_sg_table = drm_gem_dma_object_get_sg_table,
+	.vmap = drm_gem_dma_object_vmap,
+	.mmap = ethosu_gem_mmap,
+	.vm_ops = &drm_gem_dma_vm_ops,
+};
+
+/**
+ * ethosu_gem_create_object - Implementation of driver->gem_create_object.
+ * @ddev: DRM device
+ * @size: Size in bytes of the memory the object will reference
+ *
+ * This lets the GEM helpers allocate object structs for us, and keep
+ * our BO stats correct.
+ */
+struct drm_gem_object *ethosu_gem_create_object(struct drm_device *ddev, size_t size)
+{
+	struct ethosu_gem_object *obj;
+
+	obj = kzalloc(sizeof(*obj), GFP_KERNEL);
+	if (!obj)
+		return ERR_PTR(-ENOMEM);
+
+	obj->base.base.funcs = &ethosu_gem_funcs;
+	return &obj->base.base;
+}
+
+/**
+ * ethosu_gem_create_with_handle() - Create a GEM object and attach it to a handle.
+ * @file: DRM file.
+ * @ddev: DRM device.
+ * @size: Size of the GEM object to allocate.
+ * @flags: Combination of drm_ethosu_bo_flags flags.
+ * @handle: Pointer holding the handle pointing to the new GEM object.
+ *
+ * Return: Zero on success
+ */
+int ethosu_gem_create_with_handle(struct drm_file *file,
+				  struct drm_device *ddev,
+				  u64 *size, u32 flags, u32 *handle)
+{
+	struct drm_gem_dma_object *mem;
+	struct ethosu_gem_object *bo;
+	int ret;
+
+	mem = drm_gem_dma_create(ddev, *size);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+
+	bo = to_ethosu_bo(&mem->base);
+	bo->flags = flags;
+
+	/*
+	 * Allocate an id of idr table where the obj is registered
+	 * and handle has the id what user can see.
+	 */
+	ret = drm_gem_handle_create(file, &mem->base, handle);
+	if (!ret)
+		*size = bo->base.base.size;
+
+	/* drop reference from allocate - handle holds it now. */
+	drm_gem_object_put(&mem->base);
+
+	return ret;
+}
+
+struct dma {
+	s8 region;
+	u64 len;
+	u64 offset;
+	s64 stride[2];
+};
+
+struct dma_state {
+	u16 size0;
+	u16 size1;
+	s8 mode;
+	struct dma src;
+	struct dma dst;
+};
+
+struct buffer {
+	u64 base;
+	u32 length;
+	s8 region;
+};
+
+struct feat_matrix {
+	u64 base[4];
+	s64 stride_x;
+	s64 stride_y;
+	s64 stride_c;
+	s8 region;
+	u8 broadcast;
+	u16 stride_kernel;
+	u16 precision;
+	u16 depth;
+	u16 width;
+	u16 width0;
+	u16 height[3];
+	u8 pad_top;
+	u8 pad_left;
+	u8 pad_bottom;
+	u8 pad_right;
+};
+
+struct cmd_state {
+	struct dma_state dma;
+	struct buffer scale[2];
+	struct buffer weight[4];
+	struct feat_matrix ofm;
+	struct feat_matrix ifm;
+	struct feat_matrix ifm2;
+};
+
+static void cmd_state_init(struct cmd_state *st)
+{
+	/* Initialize to all 1s to detect missing setup */
+	memset(st, 0xff, sizeof(*st));
+}
+
+static u64 cmd_to_addr(u32 *cmd)
+{
+	return ((u64)((cmd[0] & 0xff0000) << 16)) | cmd[1];
+}
+
+static u64 dma_length(struct ethosu_validated_cmdstream_info *info,
+		      struct dma_state *dma_st, struct dma *dma)
+{
+	s8 mode = dma_st->mode;
+	u64 len = dma->len;
+
+	if (mode >= 1) {
+		len += dma->stride[0];
+		len *= dma_st->size0;
+	}
+	if (mode == 2) {
+		len += dma->stride[1];
+		len *= dma_st->size1;
+	}
+	if (dma->region >= 0)
+		info->region_size[dma->region] = max(info->region_size[dma->region],
+						     len + dma->offset);
+
+	return len;
+}
+
+static u64 feat_matrix_length(struct ethosu_validated_cmdstream_info *info,
+			      struct feat_matrix *fm,
+			      u32 x, u32 y, u32 c)
+{
+	u32 element_size, storage = fm->precision >> 14;
+	int tile = 0;
+	u64 addr;
+
+	if (fm->region < 0)
+		return U64_MAX;
+
+	switch (storage) {
+	case 0:
+		if (x >= fm->width0 + 1) {
+			x -= fm->width0 + 1;
+			tile += 1;
+		}
+		if (y >= fm->height[tile] + 1) {
+			y -= fm->height[tile] + 1;
+			tile += 2;
+		}
+		break;
+	case 1:
+		if (y >= fm->height[1] + 1) {
+			y -= fm->height[1] + 1;
+			tile = 2;
+		} else if (y >= fm->height[0] + 1) {
+			y -= fm->height[0] + 1;
+			tile = 1;
+		}
+		break;
+	}
+	if (fm->base[tile] == U64_MAX)
+		return U64_MAX;
+
+	addr = fm->base[tile] + y * fm->stride_y;
+
+	switch ((fm->precision >> 6) & 0x3) { // format
+	case 0: //nhwc:
+		addr += x * fm->stride_x + c;
+		break;
+	case 1: //nhcwb16:
+		element_size = BIT((fm->precision >> 1) & 0x3);
+
+		addr += (c / 16) * fm->stride_c + (16 * x + (c & 0xf)) * element_size;
+		break;
+	}
+
+	info->region_size[fm->region] = max(info->region_size[fm->region], addr + 1);
+
+	return addr;
+}
+
+static int calc_sizes(struct drm_device *ddev,
+		      struct ethosu_validated_cmdstream_info *info,
+		      u16 op, struct cmd_state *st,
+		      bool ifm, bool ifm2, bool weight, bool scale)
+{
+	u64 len;
+
+	if (ifm) {
+		if (st->ifm.stride_kernel == U16_MAX)
+			return -EINVAL;
+		u32 stride_y = ((st->ifm.stride_kernel >> 8) & 0x2) +
+			((st->ifm.stride_kernel >> 1) & 0x1) + 1;
+		u32 stride_x = ((st->ifm.stride_kernel >> 5) & 0x2) +
+			(st->ifm.stride_kernel & 0x1) + 1;
+		u32 ifm_height = st->ofm.height[2] * stride_y +
+			st->ifm.height[2] - (st->ifm.pad_top + st->ifm.pad_bottom);
+		u32 ifm_width  = st->ofm.width * stride_x +
+			st->ifm.width - (st->ifm.pad_left + st->ifm.pad_right);
+
+		len = feat_matrix_length(info, &st->ifm, ifm_width,
+					 ifm_height, st->ifm.depth);
+		dev_dbg(ddev->dev, "op %d: IFM:%d:0x%llx-0x%llx\n",
+			op, st->ifm.region, st->ifm.base[0], len);
+		if (len == U64_MAX)
+			return -EINVAL;
+	}
+
+	if (ifm2) {
+		len = feat_matrix_length(info, &st->ifm2, st->ifm.depth,
+					 0, st->ofm.depth);
+		dev_dbg(ddev->dev, "op %d: IFM2:%d:0x%llx-0x%llx\n",
+			op, st->ifm2.region, st->ifm2.base[0], len);
+		if (len == U64_MAX)
+			return -EINVAL;
+	}
+
+	if (weight) {
+		dev_dbg(ddev->dev, "op %d: W:%d:0x%llx-0x%llx\n",
+			op, st->weight[0].region, st->weight[0].base,
+			st->weight[0].base + st->weight[0].length - 1);
+		if (st->weight[0].region < 0 || st->weight[0].base == U64_MAX ||
+		    st->weight[0].length == U32_MAX)
+			return -EINVAL;
+		info->region_size[st->weight[0].region] =
+			max(info->region_size[st->weight[0].region],
+			    st->weight[0].base + st->weight[0].length);
+	}
+
+	if (scale) {
+		dev_dbg(ddev->dev, "op %d: S:%d:0x%llx-0x%llx\n",
+			op, st->scale[0].region, st->scale[0].base,
+			st->scale[0].base + st->scale[0].length - 1);
+		if (st->scale[0].region < 0 || st->scale[0].base == U64_MAX ||
+		    st->scale[0].length == U32_MAX)
+			return -EINVAL;
+		info->region_size[st->scale[0].region] =
+			max(info->region_size[st->scale[0].region],
+			    st->scale[0].base + st->scale[0].length);
+	}
+
+	len = feat_matrix_length(info, &st->ofm, st->ofm.width,
+				 st->ofm.height[2], st->ofm.depth);
+	dev_dbg(ddev->dev, "op %d: OFM:%d:0x%llx-0x%llx\n",
+		op, st->ofm.region, st->ofm.base[0], len);
+	if (len == U64_MAX)
+		return -EINVAL;
+	info->output_region[st->ofm.region] = true;
+
+	return 0;
+}
+
+static int calc_sizes_elemwise(struct drm_device *ddev,
+			       struct ethosu_validated_cmdstream_info *info,
+			       u16 op, struct cmd_state *st,
+			       bool ifm, bool ifm2)
+{
+	u32 height, width, depth;
+	u64 len;
+
+	if (ifm) {
+		height = st->ifm.broadcast & 0x1 ? 0 : st->ofm.height[2];
+		width = st->ifm.broadcast & 0x2 ? 0 : st->ofm.width;
+		depth = st->ifm.broadcast & 0x4 ? 0 : st->ofm.depth;
+
+		len = feat_matrix_length(info, &st->ifm, width,
+					 height, depth);
+		dev_dbg(ddev->dev, "op %d: IFM:%d:0x%llx-0x%llx\n",
+			op, st->ifm.region, st->ifm.base[0], len);
+		if (len == U64_MAX)
+			return -EINVAL;
+	}
+
+	if (ifm2) {
+		height = st->ifm2.broadcast & 0x1 ? 0 : st->ofm.height[2];
+		width = st->ifm2.broadcast & 0x2 ? 0 : st->ofm.width;
+		depth = st->ifm2.broadcast & 0x4 ? 0 : st->ofm.depth;
+
+		len = feat_matrix_length(info, &st->ifm2, width,
+					 height, depth);
+		dev_dbg(ddev->dev, "op %d: IFM2:%d:0x%llx-0x%llx\n",
+			op, st->ifm2.region, st->ifm2.base[0], len);
+		if (len == U64_MAX)
+			return -EINVAL;
+	}
+
+	len = feat_matrix_length(info, &st->ofm, st->ofm.width,
+				 st->ofm.height[2], st->ofm.depth);
+	dev_dbg(ddev->dev, "op %d: OFM:%d:0x%llx-0x%llx\n",
+		op, st->ofm.region, st->ofm.base[0], len);
+	if (len == U64_MAX)
+		return -EINVAL;
+	info->output_region[st->ofm.region] = true;
+
+	return 0;
+}
+
+static int ethosu_gem_cmdstream_copy_and_validate(struct drm_device *ddev,
+						  u32 __user *ucmds,
+						  struct ethosu_gem_object *bo,
+						  u32 size)
+{
+	struct ethosu_validated_cmdstream_info __free(kfree) *info = kzalloc(sizeof(*info), GFP_KERNEL);
+	struct ethosu_device *edev = to_ethosu_device(ddev);
+	u32 *bocmds = bo->base.vaddr;
+	struct cmd_state st;
+	int i, ret;
+
+	if (!info)
+		return -ENOMEM;
+	info->cmd_size = size;
+
+	cmd_state_init(&st);
+
+	for (i = 0; i < size / 4; i++) {
+		bool use_ifm, use_ifm2, use_scale;
+		u64 dstlen, srclen;
+		u16 cmd, param;
+		u32 cmds[2];
+		u64 addr;
+
+		if (get_user(cmds[0], ucmds++))
+			return -EFAULT;
+
+		bocmds[i] = cmds[0];
+
+		cmd = cmds[0];
+		param = cmds[0] >> 16;
+
+		if (cmd & 0x4000) {
+			if (get_user(cmds[1], ucmds++))
+				return -EFAULT;
+
+			i++;
+			bocmds[i] = cmds[1];
+			addr = cmd_to_addr(cmds);
+		}
+
+		switch (cmd) {
+		case NPU_OP_DMA_START:
+			srclen = dma_length(info, &st.dma, &st.dma.src);
+			dstlen = dma_length(info, &st.dma, &st.dma.dst);
+
+			if (st.dma.dst.region >= 0)
+				info->output_region[st.dma.dst.region] = true;
+			dev_dbg(ddev->dev, "cmd: DMA SRC:%d:0x%llx+0x%llx DST:%d:0x%llx+0x%llx\n",
+				st.dma.src.region, st.dma.src.offset, srclen,
+				st.dma.dst.region, st.dma.dst.offset, dstlen);
+			break;
+		case NPU_OP_CONV:
+		case NPU_OP_DEPTHWISE:
+			use_ifm2 = param & 0x1;  // weights_ifm2
+			use_scale = !(st.ofm.precision & 0x100);
+			ret = calc_sizes(ddev, info, cmd, &st, true, use_ifm2,
+					 !use_ifm2, use_scale);
+			if (ret)
+				return ret;
+			break;
+		case NPU_OP_POOL:
+			use_ifm = param != 0x4;  // pooling mode
+			use_scale = !(st.ofm.precision & 0x100);
+			ret = calc_sizes(ddev, info, cmd, &st, use_ifm, false,
+					 false, use_scale);
+			if (ret)
+				return ret;
+			break;
+		case NPU_OP_ELEMENTWISE:
+			use_ifm2 = !((st.ifm2.broadcast == 8) || (param == 5) ||
+				(param == 6) || (param == 7) || (param == 0x24));
+			use_ifm = st.ifm.broadcast != 8;
+			ret = calc_sizes_elemwise(ddev, info, cmd, &st, use_ifm, use_ifm2);
+			if (ret)
+				return ret;
+			break;
+		case NPU_OP_RESIZE: // U85 only
+			WARN_ON(1); // TODO
+			break;
+		case NPU_SET_KERNEL_WIDTH_M1:
+			st.ifm.width = param;
+			break;
+		case NPU_SET_KERNEL_HEIGHT_M1:
+			st.ifm.height[2] = param;
+			break;
+		case NPU_SET_KERNEL_STRIDE:
+			st.ifm.stride_kernel = param;
+			break;
+		case NPU_SET_IFM_PAD_TOP:
+			st.ifm.pad_top = param & 0x7f;
+			break;
+		case NPU_SET_IFM_PAD_LEFT:
+			st.ifm.pad_left = param & 0x7f;
+			break;
+		case NPU_SET_IFM_PAD_RIGHT:
+			st.ifm.pad_right = param & 0xff;
+			break;
+		case NPU_SET_IFM_PAD_BOTTOM:
+			st.ifm.pad_bottom = param & 0xff;
+			break;
+		case NPU_SET_IFM_DEPTH_M1:
+			st.ifm.depth = param;
+			break;
+		case NPU_SET_IFM_PRECISION:
+			st.ifm.precision = param;
+			break;
+		case NPU_SET_IFM_BROADCAST:
+			st.ifm.broadcast = param;
+			break;
+		case NPU_SET_IFM_REGION:
+			st.ifm.region = param & 0x7f;
+			break;
+		case NPU_SET_IFM_WIDTH0_M1:
+			st.ifm.width0 = param;
+			break;
+		case NPU_SET_IFM_HEIGHT0_M1:
+			st.ifm.height[0] = param;
+			break;
+		case NPU_SET_IFM_HEIGHT1_M1:
+			st.ifm.height[1] = param;
+			break;
+		case NPU_SET_IFM_BASE0:
+		case NPU_SET_IFM_BASE1:
+		case NPU_SET_IFM_BASE2:
+		case NPU_SET_IFM_BASE3:
+			st.ifm.base[cmd & 0x3] = addr;
+			break;
+		case NPU_SET_IFM_STRIDE_X:
+			st.ifm.stride_x = addr;
+			break;
+		case NPU_SET_IFM_STRIDE_Y:
+			st.ifm.stride_y = addr;
+			break;
+		case NPU_SET_IFM_STRIDE_C:
+			st.ifm.stride_c = addr;
+			break;
+
+		case NPU_SET_OFM_WIDTH_M1:
+			st.ofm.width = param;
+			break;
+		case NPU_SET_OFM_HEIGHT_M1:
+			st.ofm.height[2] = param;
+			break;
+		case NPU_SET_OFM_DEPTH_M1:
+			st.ofm.depth = param;
+			break;
+		case NPU_SET_OFM_PRECISION:
+			st.ofm.precision = param;
+			break;
+		case NPU_SET_OFM_REGION:
+			st.ofm.region = param & 0x7;
+			break;
+		case NPU_SET_OFM_WIDTH0_M1:
+			st.ofm.width0 = param;
+			break;
+		case NPU_SET_OFM_HEIGHT0_M1:
+			st.ofm.height[0] = param;
+			break;
+		case NPU_SET_OFM_HEIGHT1_M1:
+			st.ofm.height[1] = param;
+			break;
+		case NPU_SET_OFM_BASE0:
+		case NPU_SET_OFM_BASE1:
+		case NPU_SET_OFM_BASE2:
+		case NPU_SET_OFM_BASE3:
+			st.ofm.base[cmd & 0x3] = addr;
+			break;
+		case NPU_SET_OFM_STRIDE_X:
+			st.ofm.stride_x = addr;
+			break;
+		case NPU_SET_OFM_STRIDE_Y:
+			st.ofm.stride_y = addr;
+			break;
+		case NPU_SET_OFM_STRIDE_C:
+			st.ofm.stride_c = addr;
+			break;
+
+		case NPU_SET_IFM2_BROADCAST:
+			st.ifm2.broadcast = param;
+			break;
+		case NPU_SET_IFM2_PRECISION:
+			st.ifm2.precision = param;
+			break;
+		case NPU_SET_IFM2_REGION:
+			st.ifm2.region = param & 0x7;
+			break;
+		case NPU_SET_IFM2_WIDTH0_M1:
+			st.ifm2.width0 = param;
+			break;
+		case NPU_SET_IFM2_HEIGHT0_M1:
+			st.ifm2.height[0] = param;
+			break;
+		case NPU_SET_IFM2_HEIGHT1_M1:
+			st.ifm2.height[1] = param;
+			break;
+		case NPU_SET_IFM2_BASE0:
+		case NPU_SET_IFM2_BASE1:
+		case NPU_SET_IFM2_BASE2:
+		case NPU_SET_IFM2_BASE3:
+			st.ifm2.base[cmd & 0x3] = addr;
+			break;
+		case NPU_SET_IFM2_STRIDE_X:
+			st.ifm2.stride_x = addr;
+			break;
+		case NPU_SET_IFM2_STRIDE_Y:
+			st.ifm2.stride_y = addr;
+			break;
+		case NPU_SET_IFM2_STRIDE_C:
+			st.ifm2.stride_c = addr;
+			break;
+
+		case NPU_SET_WEIGHT_REGION:
+			st.weight[0].region = param & 0x7;
+			break;
+		case NPU_SET_SCALE_REGION:
+			st.scale[0].region = param & 0x7;
+			break;
+		case NPU_SET_WEIGHT_BASE:
+			st.weight[0].base = addr;
+			break;
+		case NPU_SET_WEIGHT_LENGTH:
+			st.weight[0].length = cmds[1];
+			break;
+		case NPU_SET_SCALE_BASE:
+			st.scale[0].base = addr;
+			break;
+		case NPU_SET_SCALE_LENGTH:
+			st.scale[0].length = cmds[1];
+			break;
+		case NPU_SET_WEIGHT1_BASE:
+			st.weight[1].base = addr;
+			break;
+		case NPU_SET_WEIGHT1_LENGTH:
+			st.weight[1].length = cmds[1];
+			break;
+		case NPU_SET_SCALE1_BASE: // NPU_SET_WEIGHT2_BASE (U85)
+			if (ethosu_is_u65(edev))
+				st.scale[1].base = addr;
+			else
+				st.weight[2].base = addr;
+			break;
+		case NPU_SET_SCALE1_LENGTH: // NPU_SET_WEIGHT2_LENGTH (U85)
+			if (ethosu_is_u65(edev))
+				st.scale[1].length = cmds[1];
+			else
+				st.weight[1].length = cmds[1];
+			break;
+		case NPU_SET_WEIGHT3_BASE:
+			st.weight[3].base = addr;
+			break;
+		case NPU_SET_WEIGHT3_LENGTH:
+			st.weight[3].length = cmds[1];
+			break;
+
+		case NPU_SET_DMA0_SRC_REGION:
+			if (param & 0x100)
+				st.dma.src.region = -1;
+			else
+				st.dma.src.region = param & 0x7;
+			st.dma.mode = (param >> 9) & 0x3;
+			break;
+		case NPU_SET_DMA0_DST_REGION:
+			if (param & 0x100)
+				st.dma.dst.region = -1;
+			else
+				st.dma.dst.region = param & 0x7;
+			break;
+		case NPU_SET_DMA0_SIZE0:
+			st.dma.size0 = param;
+			break;
+		case NPU_SET_DMA0_SIZE1:
+			st.dma.size1 = param;
+			break;
+		case NPU_SET_DMA0_SRC_STRIDE0:
+			st.dma.src.stride[0] = ((s64)addr << 24) >> 24;
+			break;
+		case NPU_SET_DMA0_SRC_STRIDE1:
+			st.dma.src.stride[1] = ((s64)addr << 24) >> 24;
+			break;
+		case NPU_SET_DMA0_DST_STRIDE0:
+			st.dma.dst.stride[0] = ((s64)addr << 24) >> 24;
+			break;
+		case NPU_SET_DMA0_DST_STRIDE1:
+			st.dma.dst.stride[1] = ((s64)addr << 24) >> 24;
+			break;
+		case NPU_SET_DMA0_SRC:
+			st.dma.src.offset = addr;
+			break;
+		case NPU_SET_DMA0_DST:
+			st.dma.dst.offset = addr;
+			break;
+		case NPU_SET_DMA0_LEN:
+			st.dma.src.len = st.dma.dst.len = addr;
+			break;
+		default:
+			break;
+		}
+	}
+
+	for (i = 0; i < NPU_BASEP_REGION_MAX; i++) {
+		if (!info->region_size[i])
+			continue;
+		dev_dbg(ddev->dev, "region %d max size: 0x%llx\n",
+			i, info->region_size[i]);
+	}
+
+	bo->info = no_free_ptr(info);
+	return 0;
+}
+
+/**
+ * ethosu_gem_cmdstream_create() - Create a GEM object and attach it to a handle.
+ * @file: DRM file.
+ * @ddev: DRM device.
+ * @exclusive_vm: Exclusive VM. Not NULL if the GEM object can't be shared.
+ * @size: Size of the GEM object to allocate.
+ * @flags: Combination of drm_ethosu_bo_flags flags.
+ * @handle: Pointer holding the handle pointing to the new GEM object.
+ *
+ * Return: Zero on success
+ */
+int ethosu_gem_cmdstream_create(struct drm_file *file,
+				struct drm_device *ddev,
+				u32 size, u64 data, u32 flags, u32 *handle)
+{
+	int ret;
+	struct drm_gem_dma_object *mem;
+	struct ethosu_gem_object *bo;
+
+	mem = drm_gem_dma_create(ddev, size);
+	if (IS_ERR(mem))
+		return PTR_ERR(mem);
+
+	bo = to_ethosu_bo(&mem->base);
+	bo->flags = flags;
+
+	ret = ethosu_gem_cmdstream_copy_and_validate(ddev,
+						     (void __user *)(uintptr_t)data,
+						     bo, size);
+	if (ret)
+		goto fail;
+
+	/*
+	 * Allocate an id of idr table where the obj is registered
+	 * and handle has the id what user can see.
+	 */
+	ret = drm_gem_handle_create(file, &mem->base, handle);
+
+fail:
+	/* drop reference from allocate - handle holds it now. */
+	drm_gem_object_put(&mem->base);
+
+	return ret;
+}
diff --git a/drivers/accel/ethosu/ethosu_gem.h b/drivers/accel/ethosu/ethosu_gem.h
new file mode 100644
index 000000000000..3922895a60fb
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_gem.h
@@ -0,0 +1,46 @@
+/* SPDX-License-Identifier: GPL-2.0 or MIT */
+/* Copyright 2025 Arm, Ltd. */
+
+#ifndef __ETHOSU_GEM_H__
+#define __ETHOSU_GEM_H__
+
+#include "ethosu_device.h"
+#include <drm/drm_gem_dma_helper.h>
+
+struct ethosu_validated_cmdstream_info {
+	u32 cmd_size;
+	u64 region_size[NPU_BASEP_REGION_MAX];
+	bool output_region[NPU_BASEP_REGION_MAX];
+};
+
+/**
+ * struct ethosu_gem_object - Driver specific GEM object.
+ */
+struct ethosu_gem_object {
+	/** @base: Inherit from drm_gem_shmem_object. */
+	struct drm_gem_dma_object base;
+
+	struct ethosu_validated_cmdstream_info *info;
+
+	/** @flags: Combination of drm_ethosu_bo_flags flags. */
+	u32 flags;
+};
+
+static inline
+struct ethosu_gem_object *to_ethosu_bo(struct drm_gem_object *obj)
+{
+	return container_of(to_drm_gem_dma_obj(obj), struct ethosu_gem_object, base);
+}
+
+struct drm_gem_object *ethosu_gem_create_object(struct drm_device *ddev,
+						size_t size);
+
+int ethosu_gem_create_with_handle(struct drm_file *file,
+				  struct drm_device *ddev,
+				  u64 *size, u32 flags, uint32_t *handle);
+
+int ethosu_gem_cmdstream_create(struct drm_file *file,
+				struct drm_device *ddev,
+				u32 size, u64 data, u32 flags, u32 *handle);
+
+#endif /* __ETHOSU_GEM_H__ */
diff --git a/drivers/accel/ethosu/ethosu_job.c b/drivers/accel/ethosu/ethosu_job.c
new file mode 100644
index 000000000000..32b89cbfbaad
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_job.c
@@ -0,0 +1,496 @@
+// SPDX-License-Identifier: GPL-2.0-only OR MIT
+/* Copyright 2024-2025 Tomeu Vizoso <tomeu@tomeuvizoso.net> */
+/* Copyright 2025 Arm, Ltd. */
+
+#include <linux/bitfield.h>
+#include <linux/genalloc.h>
+#include <linux/interrupt.h>
+#include <linux/iopoll.h>
+#include <linux/platform_device.h>
+#include <linux/pm_runtime.h>
+
+#include <drm/drm_file.h>
+#include <drm/drm_gem.h>
+#include <drm/drm_gem_dma_helper.h>
+#include <drm/ethosu_accel.h>
+
+#include "ethosu_device.h"
+#include "ethosu_drv.h"
+#include "ethosu_gem.h"
+#include "ethosu_job.h"
+
+#define JOB_TIMEOUT_MS 500
+
+static struct ethosu_job *to_ethosu_job(struct drm_sched_job *sched_job)
+{
+	return container_of(sched_job, struct ethosu_job, base);
+}
+
+static const char *ethosu_fence_get_driver_name(struct dma_fence *fence)
+{
+	return "ethosu";
+}
+
+static const char *ethosu_fence_get_timeline_name(struct dma_fence *fence)
+{
+	return "ethosu-npu";
+}
+
+static const struct dma_fence_ops ethosu_fence_ops = {
+	.get_driver_name = ethosu_fence_get_driver_name,
+	.get_timeline_name = ethosu_fence_get_timeline_name,
+};
+
+static void ethosu_job_hw_submit(struct ethosu_device *dev, struct ethosu_job *job)
+{
+	struct drm_gem_dma_object *cmd_bo = to_drm_gem_dma_obj(job->cmd_bo);
+	struct ethosu_validated_cmdstream_info *cmd_info = to_ethosu_bo(job->cmd_bo)->info;
+
+	for (int i = 0; i < job->region_cnt; i++) {
+		struct drm_gem_dma_object *bo;
+		int region = job->region_bo_num[i];
+
+		bo = to_drm_gem_dma_obj(job->region_bo[i]);
+		writel_relaxed(lower_32_bits(bo->dma_addr), dev->regs + NPU_REG_BASEP(region));
+		writel_relaxed(upper_32_bits(bo->dma_addr), dev->regs + NPU_REG_BASEP_HI(region));
+		dev_dbg(dev->base.dev, "Region %d base addr = %pad\n", region, &bo->dma_addr);
+	}
+
+	if (job->sram_size) {
+		writel_relaxed(lower_32_bits(dev->sramphys),
+			       dev->regs + NPU_REG_BASEP(ETHOSU_SRAM_REGION));
+		writel_relaxed(upper_32_bits(dev->sramphys),
+			       dev->regs + NPU_REG_BASEP_HI(ETHOSU_SRAM_REGION));
+		dev_dbg(dev->base.dev, "Region %d base addr = %pad (SRAM)\n",
+			ETHOSU_SRAM_REGION, &dev->sramphys);
+	}
+
+	writel_relaxed(lower_32_bits(cmd_bo->dma_addr), dev->regs + NPU_REG_QBASE);
+	writel_relaxed(upper_32_bits(cmd_bo->dma_addr), dev->regs + NPU_REG_QBASE_HI);
+	writel_relaxed(cmd_info->cmd_size, dev->regs + NPU_REG_QSIZE);
+
+	writel(CMD_TRANSITION_TO_RUN, dev->regs + NPU_REG_CMD);
+
+	dev_dbg(dev->base.dev,
+		"Submitted cmd at %pad to core\n", &cmd_bo->dma_addr);
+}
+
+static int ethosu_acquire_object_fences(struct ethosu_job *job)
+{
+	int i, ret;
+	struct drm_gem_object **bos = job->region_bo;
+	struct ethosu_validated_cmdstream_info *info = to_ethosu_bo(job->cmd_bo)->info;
+
+	for (i = 0; i < job->region_cnt; i++) {
+		bool is_write;
+
+		if (!bos[i])
+			break;
+
+		ret = dma_resv_reserve_fences(bos[i]->resv, 1);
+		if (ret)
+			return ret;
+
+		is_write = info->output_region[job->region_bo_num[i]];
+		ret = drm_sched_job_add_implicit_dependencies(&job->base, bos[i],
+							      is_write);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
+
+static void ethosu_attach_object_fences(struct ethosu_job *job)
+{
+	int i;
+	struct dma_fence *fence = job->inference_done_fence;
+	struct drm_gem_object **bos = job->region_bo;
+	struct ethosu_validated_cmdstream_info *info = to_ethosu_bo(job->cmd_bo)->info;
+
+	for (i = 0; i < job->region_cnt; i++)
+		if (info->output_region[job->region_bo_num[i]])
+			dma_resv_add_fence(bos[i]->resv, fence, DMA_RESV_USAGE_WRITE);
+}
+
+static int ethosu_job_push(struct ethosu_job *job)
+{
+	struct ww_acquire_ctx acquire_ctx;
+	int ret;
+
+	ret = drm_gem_lock_reservations(job->region_bo, job->region_cnt, &acquire_ctx);
+	if (ret)
+		return ret;
+
+	ret = ethosu_acquire_object_fences(job);
+	if (ret)
+		goto out;
+
+	ret = pm_runtime_resume_and_get(job->dev->base.dev);
+	if (!ret) {
+		guard(mutex)(&job->dev->sched_lock);
+
+		drm_sched_job_arm(&job->base);
+		job->inference_done_fence = dma_fence_get(&job->base.s_fence->finished);
+		kref_get(&job->refcount); /* put by scheduler job completion */
+		drm_sched_entity_push_job(&job->base);
+		ethosu_attach_object_fences(job);
+	}
+
+out:
+	drm_gem_unlock_reservations(job->region_bo, job->region_cnt, &acquire_ctx);
+	return ret;
+}
+
+static void ethosu_job_cleanup(struct kref *ref)
+{
+	struct ethosu_job *job = container_of(ref, struct ethosu_job,
+						refcount);
+	unsigned int i;
+
+	pm_runtime_put_autosuspend(job->dev->base.dev);
+
+	dma_fence_put(job->done_fence);
+	dma_fence_put(job->inference_done_fence);
+
+	for (i = 0; i < job->region_cnt; i++)
+		drm_gem_object_put(job->region_bo[i]);
+
+	drm_gem_object_put(job->cmd_bo);
+
+	kfree(job);
+}
+
+static void ethosu_job_put(struct ethosu_job *job)
+{
+	kref_put(&job->refcount, ethosu_job_cleanup);
+}
+
+static void ethosu_job_free(struct drm_sched_job *sched_job)
+{
+	struct ethosu_job *job = to_ethosu_job(sched_job);
+
+	drm_sched_job_cleanup(sched_job);
+	ethosu_job_put(job);
+}
+
+static struct dma_fence *ethosu_job_run(struct drm_sched_job *sched_job)
+{
+	struct ethosu_job *job = to_ethosu_job(sched_job);
+	struct ethosu_device *dev = job->dev;
+	struct dma_fence *fence = job->done_fence;
+
+	if (unlikely(job->base.s_fence->finished.error))
+		return NULL;
+
+	dma_fence_init(fence, &ethosu_fence_ops, &dev->fence_lock,
+		       dev->fence_context, ++dev->emit_seqno);
+	dma_fence_get(fence);
+
+	scoped_guard(mutex, &dev->job_lock) {
+		dev->in_flight_job = job;
+		ethosu_job_hw_submit(dev, job);
+	}
+
+	return fence;
+}
+
+static void ethosu_job_handle_irq(struct ethosu_device *dev)
+{
+	u32 status = readl_relaxed(dev->regs + NPU_REG_STATUS);
+
+	if (status & (STATUS_BUS_STATUS | STATUS_CMD_PARSE_ERR)) {
+		dev_err(dev->base.dev, "Error IRQ - %x\n", status);
+		drm_sched_fault(&dev->sched);
+		return;
+	}
+
+	scoped_guard(mutex, &dev->job_lock) {
+		if (dev->in_flight_job) {
+			dma_fence_signal(dev->in_flight_job->done_fence);
+			dev->in_flight_job = NULL;
+		}
+	}
+}
+
+static irqreturn_t ethosu_job_irq_handler_thread(int irq, void *data)
+{
+	struct ethosu_device *dev = data;
+
+	ethosu_job_handle_irq(dev);
+
+	return IRQ_HANDLED;
+}
+
+static irqreturn_t ethosu_job_irq_handler(int irq, void *data)
+{
+	struct ethosu_device *dev = data;
+	u32 status = readl_relaxed(dev->regs + NPU_REG_STATUS);
+
+	if (!(status & STATUS_IRQ_RAISED))
+		return IRQ_NONE;
+
+	writel_relaxed(CMD_CLEAR_IRQ, dev->regs + NPU_REG_CMD);
+	return IRQ_WAKE_THREAD;
+}
+
+static enum drm_gpu_sched_stat ethosu_job_timedout(struct drm_sched_job *bad)
+{
+	struct ethosu_job *job = to_ethosu_job(bad);
+	struct ethosu_device *dev = job->dev;
+	bool running;
+	u32 *bocmds = to_drm_gem_dma_obj(job->cmd_bo)->vaddr;
+	u32 cmdaddr;
+
+	cmdaddr = readl_relaxed(dev->regs + NPU_REG_QREAD);
+	running = FIELD_GET(STATUS_STATE_RUNNING, readl_relaxed(dev->regs + NPU_REG_STATUS));
+
+	if (running) {
+		int ret;
+		u32 reg;
+
+		ret = readl_relaxed_poll_timeout(dev->regs + NPU_REG_QREAD,
+						 reg,
+						 reg != cmdaddr,
+						 USEC_PER_MSEC, 100 * USEC_PER_MSEC);
+
+		/* If still running and progress is being made, just return */
+		if (!ret)
+			return DRM_GPU_SCHED_STAT_NO_HANG;
+	}
+
+	dev_err(dev->base.dev, "NPU sched timed out: NPU %s, cmdstream offset 0x%x: 0x%x\n",
+		running ? "running" : "stopped",
+		cmdaddr, bocmds[cmdaddr / 4]);
+
+	drm_sched_stop(&dev->sched, bad);
+
+	scoped_guard(mutex, &dev->job_lock)
+		dev->in_flight_job = NULL;
+
+	/* Proceed with reset now. */
+	pm_runtime_force_suspend(dev->base.dev);
+	pm_runtime_force_resume(dev->base.dev);
+
+	/* Restart the scheduler */
+	drm_sched_start(&dev->sched, 0);
+
+	return DRM_GPU_SCHED_STAT_RESET;
+}
+
+static const struct drm_sched_backend_ops ethosu_sched_ops = {
+	.run_job = ethosu_job_run,
+	.timedout_job = ethosu_job_timedout,
+	.free_job = ethosu_job_free
+};
+
+int ethosu_job_init(struct ethosu_device *edev)
+{
+	struct device *dev = edev->base.dev;
+	struct drm_sched_init_args args = {
+		.ops = &ethosu_sched_ops,
+		.num_rqs = DRM_SCHED_PRIORITY_COUNT,
+		.credit_limit = 1,
+		.timeout = msecs_to_jiffies(JOB_TIMEOUT_MS),
+		.name = dev_name(dev),
+		.dev = dev,
+	};
+	int ret;
+
+	spin_lock_init(&edev->fence_lock);
+	ret = devm_mutex_init(dev, &edev->job_lock);
+	if (ret)
+		return ret;
+	ret = devm_mutex_init(dev, &edev->sched_lock);
+	if (ret)
+		return ret;
+
+	edev->irq = platform_get_irq(to_platform_device(dev), 0);
+	if (edev->irq < 0)
+		return edev->irq;
+
+	ret = devm_request_threaded_irq(dev, edev->irq,
+					ethosu_job_irq_handler,
+					ethosu_job_irq_handler_thread,
+					IRQF_SHARED, KBUILD_MODNAME,
+					edev);
+	if (ret) {
+		dev_err(dev, "failed to request irq\n");
+		return ret;
+	}
+
+	edev->fence_context = dma_fence_context_alloc(1);
+
+	ret = drm_sched_init(&edev->sched, &args);
+	if (ret) {
+		dev_err(dev, "Failed to create scheduler: %d\n", ret);
+		goto err_sched;
+	}
+
+	return 0;
+
+err_sched:
+	drm_sched_fini(&edev->sched);
+	return ret;
+}
+
+void ethosu_job_fini(struct ethosu_device *dev)
+{
+	drm_sched_fini(&dev->sched);
+}
+
+int ethosu_job_open(struct ethosu_file_priv *ethosu_priv)
+{
+	struct ethosu_device *dev = ethosu_priv->edev;
+	struct drm_gpu_scheduler *sched = &dev->sched;
+	int ret;
+
+	ret = drm_sched_entity_init(&ethosu_priv->sched_entity,
+				    DRM_SCHED_PRIORITY_NORMAL,
+				    &sched, 1, NULL);
+	return WARN_ON(ret);
+}
+
+void ethosu_job_close(struct ethosu_file_priv *ethosu_priv)
+{
+	struct drm_sched_entity *entity = &ethosu_priv->sched_entity;
+
+	drm_sched_entity_destroy(entity);
+}
+
+static int ethosu_ioctl_submit_job(struct drm_device *dev, struct drm_file *file,
+				   struct drm_ethosu_job *job)
+{
+	struct ethosu_device *edev = to_ethosu_device(dev);
+	struct ethosu_file_priv *file_priv = file->driver_priv;
+	struct ethosu_job *ejob = NULL;
+	struct ethosu_validated_cmdstream_info *cmd_info;
+	int ret = 0;
+
+	/* BO region 2 is reserved if SRAM is used */
+	if (job->region_bo_handles[ETHOSU_SRAM_REGION] && job->sram_size)
+		return -EINVAL;
+
+	if (edev->npu_info.sram_size < job->sram_size)
+		return -EINVAL;
+
+	ejob = kzalloc(sizeof(*ejob), GFP_KERNEL);
+	if (!ejob)
+		return -ENOMEM;
+
+	kref_init(&ejob->refcount);
+
+	ejob->dev = edev;
+	ejob->sram_size = job->sram_size;
+
+	ejob->done_fence = kzalloc(sizeof(*ejob->done_fence), GFP_KERNEL);
+	if (!ejob->done_fence) {
+		ret = -ENOMEM;
+		goto out_cleanup_job;
+	}
+
+	ret = drm_sched_job_init(&ejob->base,
+				 &file_priv->sched_entity,
+				 1, NULL, file->client_id);
+	if (ret)
+		goto out_put_job;
+
+	ejob->cmd_bo = drm_gem_object_lookup(file, job->cmd_bo);
+	if (!ejob->cmd_bo) {
+		ret = -ENOENT;
+		goto out_cleanup_job;
+	}
+	cmd_info = to_ethosu_bo(ejob->cmd_bo)->info;
+	if (!cmd_info) {
+		ret = -EINVAL;
+		goto out_cleanup_job;
+	}
+
+	for (int i = 0; i < NPU_BASEP_REGION_MAX; i++) {
+		struct drm_gem_object *gem;
+
+		/* Can only omit a BO handle if the region is not used or used for SRAM */
+		if (!job->region_bo_handles[i] &&
+		    (!cmd_info->region_size[i] || (i == ETHOSU_SRAM_REGION && job->sram_size)))
+			continue;
+
+		if (job->region_bo_handles[i] && !cmd_info->region_size[i]) {
+			dev_err(dev->dev,
+				"Cmdstream BO handle %d set for unused region %d\n",
+				job->region_bo_handles[i], i);
+			ret = -EINVAL;
+			goto out_cleanup_job;
+		}
+
+		gem = drm_gem_object_lookup(file, job->region_bo_handles[i]);
+		if (!gem) {
+			dev_err(dev->dev,
+				"Invalid BO handle %d for region %d\n",
+				job->region_bo_handles[i], i);
+			ret = -ENOENT;
+			goto out_cleanup_job;
+		}
+
+		ejob->region_bo[ejob->region_cnt] = gem;
+		ejob->region_bo_num[ejob->region_cnt] = i;
+		ejob->region_cnt++;
+
+		if (to_ethosu_bo(gem)->info) {
+			dev_err(dev->dev,
+				"Cmdstream BO handle %d used for region %d\n",
+				job->region_bo_handles[i], i);
+			ret = -EINVAL;
+			goto out_cleanup_job;
+		}
+
+		/* Verify the command stream doesn't have accesses outside the BO */
+		if (cmd_info->region_size[i] > gem->size) {
+			dev_err(dev->dev,
+				"cmd stream region %d size greater than BO size (%llu > %zu)\n",
+				i, cmd_info->region_size[i], gem->size);
+			ret = -EOVERFLOW;
+			goto out_cleanup_job;
+		}
+	}
+	ret = ethosu_job_push(ejob);
+
+out_cleanup_job:
+	if (ret)
+		drm_sched_job_cleanup(&ejob->base);
+out_put_job:
+	ethosu_job_put(ejob);
+
+	return ret;
+}
+
+int ethosu_ioctl_submit(struct drm_device *dev, void *data, struct drm_file *file)
+{
+	struct drm_ethosu_submit *args = data;
+	int ret = 0;
+	unsigned int i = 0;
+
+	if (args->pad) {
+		drm_dbg(dev, "Reserved field in drm_ethosu_submit struct should be 0.\n");
+		return -EINVAL;
+	}
+
+	struct drm_ethosu_job __free(kvfree) *jobs =
+		kvmalloc_array(args->job_count, sizeof(*jobs), GFP_KERNEL);
+	if (!jobs)
+		return -ENOMEM;
+
+	if (copy_from_user(jobs,
+			   (void __user *)(uintptr_t)args->jobs,
+			   args->job_count * sizeof(*jobs))) {
+		drm_dbg(dev, "Failed to copy incoming job array\n");
+		return -EFAULT;
+	}
+
+	for (i = 0; i < args->job_count; i++) {
+		ret = ethosu_ioctl_submit_job(dev, file, &jobs[i]);
+		if (ret)
+			return ret;
+	}
+
+	return 0;
+}
diff --git a/drivers/accel/ethosu/ethosu_job.h b/drivers/accel/ethosu/ethosu_job.h
new file mode 100644
index 000000000000..ff1cf448d094
--- /dev/null
+++ b/drivers/accel/ethosu/ethosu_job.h
@@ -0,0 +1,40 @@
+/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
+/* Copyright 2024-2025 Tomeu Vizoso <tomeu@tomeuvizoso.net> */
+/* Copyright 2025 Arm, Ltd. */
+
+#ifndef __ETHOSU_JOB_H__
+#define __ETHOSU_JOB_H__
+
+#include <linux/kref.h>
+#include <drm/gpu_scheduler.h>
+
+struct ethosu_device;
+struct ethosu_file_priv;
+
+struct ethosu_job {
+	struct drm_sched_job base;
+	struct ethosu_device *dev;
+
+	struct drm_gem_object *cmd_bo;
+	struct drm_gem_object *region_bo[NPU_BASEP_REGION_MAX];
+	u8 region_bo_num[NPU_BASEP_REGION_MAX];
+	u8 region_cnt;
+	u32 sram_size;
+
+	/* Fence to be signaled by drm-sched once its done with the job */
+	struct dma_fence *inference_done_fence;
+
+	/* Fence to be signaled by IRQ handler when the job is complete. */
+	struct dma_fence *done_fence;
+
+	struct kref refcount;
+};
+
+int ethosu_ioctl_submit(struct drm_device *dev, void *data, struct drm_file *file);
+
+int ethosu_job_init(struct ethosu_device *dev);
+void ethosu_job_fini(struct ethosu_device *dev);
+int ethosu_job_open(struct ethosu_file_priv *ethosu_priv);
+void ethosu_job_close(struct ethosu_file_priv *ethosu_priv);
+
+#endif
diff --git a/include/uapi/drm/ethosu_accel.h b/include/uapi/drm/ethosu_accel.h
new file mode 100644
index 000000000000..af78bb4686d7
--- /dev/null
+++ b/include/uapi/drm/ethosu_accel.h
@@ -0,0 +1,261 @@
+/* SPDX-License-Identifier: MIT */
+/* Copyright (C) 2025 Arm, Ltd. */
+#ifndef _ETHOSU_DRM_H_
+#define _ETHOSU_DRM_H_
+
+#include "drm.h"
+
+#if defined(__cplusplus)
+extern "C" {
+#endif
+
+/**
+ * DOC: IOCTL IDs
+ *
+ * enum drm_ethosu_ioctl_id - IOCTL IDs
+ *
+ * Place new ioctls at the end, don't re-order, don't replace or remove entries.
+ *
+ * These IDs are not meant to be used directly. Use the DRM_IOCTL_ETHOSU_xxx
+ * definitions instead.
+ */
+enum drm_ethosu_ioctl_id {
+	/** @DRM_ETHOSU_DEV_QUERY: Query device information. */
+	DRM_ETHOSU_DEV_QUERY = 0,
+
+	/** @DRM_ETHOSU_BO_CREATE: Create a buffer object. */
+	DRM_ETHOSU_BO_CREATE,
+
+	/** @DRM_ETHOSU_BO_WAIT: Wait on a buffer object's fence. */
+	DRM_ETHOSU_BO_WAIT,
+
+	/**
+	 * @DRM_ETHOSU_BO_MMAP_OFFSET: Get the file offset to pass to
+	 * mmap to map a GEM object.
+	 */
+	DRM_ETHOSU_BO_MMAP_OFFSET,
+
+	/**
+	 * @DRM_ETHOSU_CMDSTREAM_BO_CREATE: Create a command stream buffer
+	 * object.
+	 */
+	DRM_ETHOSU_CMDSTREAM_BO_CREATE,
+
+	/** @DRM_ETHOSU_SUBMIT: Submit a job and BOs to run. */
+	DRM_ETHOSU_SUBMIT,
+};
+
+/**
+ * DOC: IOCTL arguments
+ */
+
+/**
+ * enum drm_ethosu_dev_query_type - Query type
+ *
+ * Place new types at the end, don't re-order, don't remove or replace.
+ */
+enum drm_ethosu_dev_query_type {
+	/** @DRM_ETHOSU_DEV_QUERY_NPU_INFO: Query NPU information. */
+	DRM_ETHOSU_DEV_QUERY_NPU_INFO = 0,
+};
+
+/**
+ * struct drm_ethosu_gpu_info - NPU information
+ *
+ * Structure grouping all queryable information relating to the NPU.
+ */
+struct drm_ethosu_npu_info {
+	/** @id : NPU ID. */
+	__u32 id;
+#define DRM_ETHOSU_ARCH_MAJOR(x)			((x) >> 28)
+#define DRM_ETHOSU_ARCH_MINOR(x)			(((x) >> 20) & 0xff)
+#define DRM_ETHOSU_ARCH_PATCH(x)			(((x) >> 16) & 0xf)
+#define DRM_ETHOSU_PRODUCT_MAJOR(x)		(((x) >> 12) & 0xf)
+#define DRM_ETHOSU_VERSION_MAJOR(x)		(((x) >> 8) & 0xf)
+#define DRM_ETHOSU_VERSION_MINOR(x)		(((x) >> 4) & 0xff)
+#define DRM_ETHOSU_VERSION_STATUS(x)		((x) & 0xf)
+
+	/** @gpu_rev: GPU revision. */
+	__u32 config;
+
+	__u32 sram_size;
+};
+
+/**
+ * struct drm_ethosu_dev_query - Arguments passed to DRM_ETHOSU_IOCTL_DEV_QUERY
+ */
+struct drm_ethosu_dev_query {
+	/** @type: the query type (see drm_ethosu_dev_query_type). */
+	__u32 type;
+
+	/**
+	 * @size: size of the type being queried.
+	 *
+	 * If pointer is NULL, size is updated by the driver to provide the
+	 * output structure size. If pointer is not NULL, the driver will
+	 * only copy min(size, actual_structure_size) bytes to the pointer,
+	 * and update the size accordingly. This allows us to extend query
+	 * types without breaking userspace.
+	 */
+	__u32 size;
+
+	/**
+	 * @pointer: user pointer to a query type struct.
+	 *
+	 * Pointer can be NULL, in which case, nothing is copied, but the
+	 * actual structure size is returned. If not NULL, it must point to
+	 * a location that's large enough to hold size bytes.
+	 */
+	__u64 pointer;
+};
+
+/**
+ * enum drm_ethosu_bo_flags - Buffer object flags, passed at creation time.
+ */
+enum drm_ethosu_bo_flags {
+	/**
+	 * @DRM_ETHOSU_BO_NO_MMAP: The buffer object will never be CPU-mapped
+	 * in userspace.
+	 */
+	DRM_ETHOSU_BO_NO_MMAP = (1 << 0),
+};
+
+/**
+ * struct drm_ethosu_bo_create - Arguments passed to DRM_IOCTL_ETHOSU_BO_CREATE.
+ */
+struct drm_ethosu_bo_create {
+	/**
+	 * @size: Requested size for the object
+	 *
+	 * The (page-aligned) allocated size for the object will be returned.
+	 */
+	__u64 size;
+
+	/**
+	 * @flags: Flags. Must be a combination of drm_ethosu_bo_flags flags.
+	 */
+	__u32 flags;
+
+	/**
+	 * @handle: Returned handle for the object.
+	 *
+	 * Object handles are nonzero.
+	 */
+	__u32 handle;
+};
+
+/**
+ * struct drm_ethosu_bo_mmap_offset - Arguments passed to DRM_IOCTL_ETHOSU_BO_MMAP_OFFSET.
+ */
+struct drm_ethosu_bo_mmap_offset {
+	/** @handle: Handle of the object we want an mmap offset for. */
+	__u32 handle;
+
+	/** @pad: MBZ. */
+	__u32 pad;
+
+	/** @offset: The fake offset to use for subsequent mmap calls. */
+	__u64 offset;
+};
+
+/**
+ * struct drm_ethosu_wait_bo - ioctl argument for waiting for
+ * completion of the last DRM_ETHOSU_SUBMIT on a BO.
+ *
+ * This is useful for cases where multiple processes might be
+ * rendering to a BO and you want to wait for all rendering to be
+ * completed.
+ */
+struct drm_ethosu_bo_wait {
+	__u32 handle;
+	__u32 pad;
+	__s64 timeout_ns;	/* absolute */
+};
+
+struct drm_ethosu_cmdstream_bo_create {
+	/* Size of the data argument. */
+	__u32 size;
+
+	/* Flags, currently must be 0. */
+	__u32 flags;
+
+	/* Pointer to the data. */
+	__u64 data;
+
+	/** Returned GEM handle for the BO. */
+	__u32 handle;
+
+	/* Pad, must be 0. */
+	__u32 pad;
+};
+
+/**
+ * struct drm_ethosu_job - A job to be run on the NPU
+ *
+ * The kernel will schedule the execution of this job taking into account its
+ * dependencies with other jobs. All tasks in the same job will be executed
+ * sequentially on the same core, to benefit from memory residency in SRAM.
+ */
+struct drm_ethosu_job {
+	/** Input: BO handle for cmdstream. */
+	__u32 cmd_bo;
+
+	/** Input: Amount of SRAM to use. */
+	__u32 sram_size;
+
+#define ETHOSU_MAX_REGIONS	8
+	/** Input: Array of BO handles for each region. */
+	__u32 region_bo_handles[ETHOSU_MAX_REGIONS];
+};
+
+/**
+ * struct drm_ethosu_submit - ioctl argument for submitting commands to the NPU.
+ *
+ * The kernel will schedule the execution of these jobs in dependency order.
+ */
+struct drm_ethosu_submit {
+	/** Input: Pointer to an array of struct drm_ethosu_job. */
+	__u64 jobs;
+
+	/** Input: Number of jobs passed in. */
+	__u32 job_count;
+
+	/** Reserved, must be zero. */
+	__u32 pad;
+};
+
+/**
+ * DRM_IOCTL_ETHOSU() - Build a ethosu IOCTL number
+ * @__access: Access type. Must be R, W or RW.
+ * @__id: One of the DRM_ETHOSU_xxx id.
+ * @__type: Suffix of the type being passed to the IOCTL.
+ *
+ * Don't use this macro directly, use the DRM_IOCTL_ETHOSU_xxx
+ * values instead.
+ *
+ * Return: An IOCTL number to be passed to ioctl() from userspace.
+ */
+#define DRM_IOCTL_ETHOSU(__access, __id, __type) \
+	DRM_IO ## __access(DRM_COMMAND_BASE + DRM_ETHOSU_ ## __id, \
+			   struct drm_ethosu_ ## __type)
+
+enum {
+	DRM_IOCTL_ETHOSU_DEV_QUERY =
+		DRM_IOCTL_ETHOSU(WR, DEV_QUERY, dev_query),
+	DRM_IOCTL_ETHOSU_BO_CREATE =
+		DRM_IOCTL_ETHOSU(WR, BO_CREATE, bo_create),
+	DRM_IOCTL_ETHOSU_BO_WAIT =
+		DRM_IOCTL_ETHOSU(WR, BO_WAIT, bo_wait),
+	DRM_IOCTL_ETHOSU_BO_MMAP_OFFSET =
+		DRM_IOCTL_ETHOSU(WR, BO_MMAP_OFFSET, bo_mmap_offset),
+	DRM_IOCTL_ETHOSU_CMDSTREAM_BO_CREATE =
+		DRM_IOCTL_ETHOSU(WR, CMDSTREAM_BO_CREATE, cmdstream_bo_create),
+	DRM_IOCTL_ETHOSU_SUBMIT =
+		DRM_IOCTL_ETHOSU(WR, SUBMIT, submit),
+};
+
+#if defined(__cplusplus)
+}
+#endif
+
+#endif /* _ETHOSU_DRM_H_ */
-- 
cgit v1.2.3


From 0a0e79a2d9ed846fc3c3f5ef92b691e81bc9721a Mon Sep 17 00:00:00 2001
From: Ville Syrjälä <ville.syrjala@linux.intel.com>
Date: Fri, 17 Oct 2025 19:33:27 +0300
Subject: drm/atomic: WARN about invalid drm_foo_get_state() usage
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

drm_{crtc,plane,connector,private_obj}_get_state() must not
be called after the atomic check phase. At that point the commit
has been carved in stone and no new objects must be introduced
into it. WARN if anyone attempts to violate this rule.

Cc: Maxime Ripard <mripard@kernel.org>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: https://patch.msgid.link/20251017163327.9074-2-ville.syrjala@linux.intel.com
Reviewed-by: Maxime Ripard <mripard@kernel.org>
---
 drivers/gpu/drm/drm_atomic.c | 8 ++++++++
 include/drm/drm_atomic.h     | 8 ++++++++
 2 files changed, 16 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 5d31b20e67ab..e05820b18832 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -200,6 +200,8 @@ void drm_atomic_state_default_clear(struct drm_atomic_state *state)
 
 	drm_dbg_atomic(dev, "Clearing atomic state %p\n", state);
 
+	state->checked = false;
+
 	for (i = 0; i < state->num_connector; i++) {
 		struct drm_connector *connector = state->connectors[i].ptr;
 
@@ -348,6 +350,7 @@ drm_atomic_get_crtc_state(struct drm_atomic_state *state,
 	struct drm_crtc_state *crtc_state;
 
 	WARN_ON(!state->acquire_ctx);
+	drm_WARN_ON(state->dev, state->checked);
 
 	crtc_state = drm_atomic_get_new_crtc_state(state, crtc);
 	if (crtc_state)
@@ -528,6 +531,7 @@ drm_atomic_get_plane_state(struct drm_atomic_state *state,
 	struct drm_plane_state *plane_state;
 
 	WARN_ON(!state->acquire_ctx);
+	drm_WARN_ON(state->dev, state->checked);
 
 	/* the legacy pointers should never be set */
 	WARN_ON(plane->fb);
@@ -837,6 +841,7 @@ drm_atomic_get_private_obj_state(struct drm_atomic_state *state,
 	struct drm_private_state *obj_state;
 
 	WARN_ON(!state->acquire_ctx);
+	drm_WARN_ON(state->dev, state->checked);
 
 	obj_state = drm_atomic_get_new_private_obj_state(state, obj);
 	if (obj_state)
@@ -1131,6 +1136,7 @@ drm_atomic_get_connector_state(struct drm_atomic_state *state,
 	struct drm_connector_state *connector_state;
 
 	WARN_ON(!state->acquire_ctx);
+	drm_WARN_ON(state->dev, state->checked);
 
 	ret = drm_modeset_lock(&config->connection_mutex, state->acquire_ctx);
 	if (ret)
@@ -1543,6 +1549,8 @@ int drm_atomic_check_only(struct drm_atomic_state *state)
 		     requested_crtc, affected_crtc);
 	}
 
+	state->checked = true;
+
 	return 0;
 }
 EXPORT_SYMBOL(drm_atomic_check_only);
diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index 155e82f87e4d..2e433d44658d 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -523,6 +523,14 @@ struct drm_atomic_state {
 	 */
 	bool duplicated : 1;
 
+	/**
+	 * @checked:
+	 *
+	 * Indicates the state has been checked and thus must no longer
+	 * be mutated. For internal use only, do not consult from drivers.
+	 */
+	bool checked : 1;
+
 	/**
 	 * @planes:
 	 *
-- 
cgit v1.2.3


From 57557964b582238d5ee4b8538d1c4694f91c2186 Mon Sep 17 00:00:00 2001
From: Jacek Lawrynowicz <jacek.lawrynowicz@linux.intel.com>
Date: Wed, 29 Oct 2025 10:17:52 +0100
Subject: accel/ivpu: Add support for userptr buffer objects

Introduce a new ioctl `drm_ivpu_bo_create_from_userptr` that allows
users to create GEM buffer objects from user pointers to memory regions.
The user pointer must be page-aligned and the memory region must remain
valid for the buffer object's lifetime.

Userptr buffers enable direct use of mmapped files (e.g. inference
weights) in NPU workloads without copying data to NPU buffer objects.
This reduces memory usage and provides better flexibility for NPU
applications.

Signed-off-by: Jacek Lawrynowicz <jacek.lawrynowicz@linux.intel.com>
Reviewed-by: Jeff Hugo <jeff.hugo@oss.qualcomm.com>
Signed-off-by: Karol Wachowski <karol.wachowski@linux.intel.com>
Link: https://patch.msgid.link/20251029091752.203198-1-karol.wachowski@linux.intel.com
---
 drivers/accel/ivpu/Makefile           |   1 +
 drivers/accel/ivpu/ivpu_drv.c         |   3 +
 drivers/accel/ivpu/ivpu_gem.c         |   2 +-
 drivers/accel/ivpu/ivpu_gem.h         |   7 ++
 drivers/accel/ivpu/ivpu_gem_userptr.c | 202 ++++++++++++++++++++++++++++++++++
 drivers/accel/ivpu/ivpu_mmu_context.c |   4 +-
 drivers/accel/ivpu/ivpu_mmu_context.h |   2 +-
 include/uapi/drm/ivpu_accel.h         |  52 +++++++++
 8 files changed, 270 insertions(+), 3 deletions(-)
 create mode 100644 drivers/accel/ivpu/ivpu_gem_userptr.c

(limited to 'include')

diff --git a/drivers/accel/ivpu/Makefile b/drivers/accel/ivpu/Makefile
index 1029e0bab061..dbf76b8a5b4c 100644
--- a/drivers/accel/ivpu/Makefile
+++ b/drivers/accel/ivpu/Makefile
@@ -6,6 +6,7 @@ intel_vpu-y := \
 	ivpu_fw.o \
 	ivpu_fw_log.o \
 	ivpu_gem.o \
+	ivpu_gem_userptr.o \
 	ivpu_hw.o \
 	ivpu_hw_btrs.o \
 	ivpu_hw_ip.o \
diff --git a/drivers/accel/ivpu/ivpu_drv.c b/drivers/accel/ivpu/ivpu_drv.c
index c6fe7a408912..ca68730dee88 100644
--- a/drivers/accel/ivpu/ivpu_drv.c
+++ b/drivers/accel/ivpu/ivpu_drv.c
@@ -134,6 +134,8 @@ bool ivpu_is_capable(struct ivpu_device *vdev, u32 capability)
 		return true;
 	case DRM_IVPU_CAP_DMA_MEMORY_RANGE:
 		return true;
+	case DRM_IVPU_CAP_BO_CREATE_FROM_USERPTR:
+		return true;
 	case DRM_IVPU_CAP_MANAGE_CMDQ:
 		return vdev->fw->sched_mode == VPU_SCHEDULING_MODE_HW;
 	default:
@@ -313,6 +315,7 @@ static const struct drm_ioctl_desc ivpu_drm_ioctls[] = {
 	DRM_IOCTL_DEF_DRV(IVPU_CMDQ_CREATE, ivpu_cmdq_create_ioctl, 0),
 	DRM_IOCTL_DEF_DRV(IVPU_CMDQ_DESTROY, ivpu_cmdq_destroy_ioctl, 0),
 	DRM_IOCTL_DEF_DRV(IVPU_CMDQ_SUBMIT, ivpu_cmdq_submit_ioctl, 0),
+	DRM_IOCTL_DEF_DRV(IVPU_BO_CREATE_FROM_USERPTR, ivpu_bo_create_from_userptr_ioctl, 0),
 };
 
 static int ivpu_wait_for_ready(struct ivpu_device *vdev)
diff --git a/drivers/accel/ivpu/ivpu_gem.c b/drivers/accel/ivpu/ivpu_gem.c
index 7353cfb73bcb..03d39615ad37 100644
--- a/drivers/accel/ivpu/ivpu_gem.c
+++ b/drivers/accel/ivpu/ivpu_gem.c
@@ -96,7 +96,7 @@ int __must_check ivpu_bo_bind(struct ivpu_bo *bo)
 	if (!bo->mmu_mapped) {
 		drm_WARN_ON(&vdev->drm, !bo->ctx);
 		ret = ivpu_mmu_context_map_sgt(vdev, bo->ctx, bo->vpu_addr, sgt,
-					       ivpu_bo_is_snooped(bo));
+					       ivpu_bo_is_snooped(bo), ivpu_bo_is_read_only(bo));
 		if (ret) {
 			ivpu_err(vdev, "Failed to map BO in MMU: %d\n", ret);
 			goto unlock;
diff --git a/drivers/accel/ivpu/ivpu_gem.h b/drivers/accel/ivpu/ivpu_gem.h
index 54452eb8a41f..2dcd7eba9cb7 100644
--- a/drivers/accel/ivpu/ivpu_gem.h
+++ b/drivers/accel/ivpu/ivpu_gem.h
@@ -38,6 +38,8 @@ void ivpu_bo_free(struct ivpu_bo *bo);
 int ivpu_bo_create_ioctl(struct drm_device *dev, void *data, struct drm_file *file);
 int ivpu_bo_info_ioctl(struct drm_device *dev, void *data, struct drm_file *file);
 int ivpu_bo_wait_ioctl(struct drm_device *dev, void *data, struct drm_file *file);
+int ivpu_bo_create_from_userptr_ioctl(struct drm_device *dev, void *data,
+				      struct drm_file *file);
 
 void ivpu_bo_list(struct drm_device *dev, struct drm_printer *p);
 void ivpu_bo_list_print(struct drm_device *dev);
@@ -75,6 +77,11 @@ static inline bool ivpu_bo_is_snooped(struct ivpu_bo *bo)
 	return ivpu_bo_cache_mode(bo) == DRM_IVPU_BO_CACHED;
 }
 
+static inline bool ivpu_bo_is_read_only(struct ivpu_bo *bo)
+{
+	return bo->flags & DRM_IVPU_BO_READ_ONLY;
+}
+
 static inline void *ivpu_to_cpu_addr(struct ivpu_bo *bo, u32 vpu_addr)
 {
 	if (vpu_addr < bo->vpu_addr)
diff --git a/drivers/accel/ivpu/ivpu_gem_userptr.c b/drivers/accel/ivpu/ivpu_gem_userptr.c
new file mode 100644
index 000000000000..235c67959453
--- /dev/null
+++ b/drivers/accel/ivpu/ivpu_gem_userptr.c
@@ -0,0 +1,202 @@
+// SPDX-License-Identifier: GPL-2.0-only
+/*
+ * Copyright (C) 2020-2025 Intel Corporation
+ */
+
+#include <linux/dma-buf.h>
+#include <linux/err.h>
+#include <linux/highmem.h>
+#include <linux/mm.h>
+#include <linux/mman.h>
+#include <linux/scatterlist.h>
+#include <linux/slab.h>
+#include <linux/capability.h>
+
+#include <drm/drm_device.h>
+#include <drm/drm_file.h>
+#include <drm/drm_gem.h>
+
+#include "ivpu_drv.h"
+#include "ivpu_gem.h"
+
+static struct sg_table *
+ivpu_gem_userptr_dmabuf_map(struct dma_buf_attachment *attachment,
+			    enum dma_data_direction direction)
+{
+	struct sg_table *sgt = attachment->dmabuf->priv;
+	int ret;
+
+	ret = dma_map_sgtable(attachment->dev, sgt, direction, DMA_ATTR_SKIP_CPU_SYNC);
+	if (ret)
+		return ERR_PTR(ret);
+
+	return sgt;
+}
+
+static void ivpu_gem_userptr_dmabuf_unmap(struct dma_buf_attachment *attachment,
+					  struct sg_table *sgt,
+					  enum dma_data_direction direction)
+{
+	dma_unmap_sgtable(attachment->dev, sgt, direction, DMA_ATTR_SKIP_CPU_SYNC);
+}
+
+static void ivpu_gem_userptr_dmabuf_release(struct dma_buf *dma_buf)
+{
+	struct sg_table *sgt = dma_buf->priv;
+	struct sg_page_iter page_iter;
+	struct page *page;
+
+	for_each_sgtable_page(sgt, &page_iter, 0) {
+		page = sg_page_iter_page(&page_iter);
+		unpin_user_page(page);
+	}
+
+	sg_free_table(sgt);
+	kfree(sgt);
+}
+
+static const struct dma_buf_ops ivpu_gem_userptr_dmabuf_ops = {
+	.map_dma_buf = ivpu_gem_userptr_dmabuf_map,
+	.unmap_dma_buf = ivpu_gem_userptr_dmabuf_unmap,
+	.release = ivpu_gem_userptr_dmabuf_release,
+};
+
+static struct dma_buf *
+ivpu_create_userptr_dmabuf(struct ivpu_device *vdev, void __user *user_ptr,
+			   size_t size, uint32_t flags)
+{
+	struct dma_buf_export_info exp_info = {};
+	struct dma_buf *dma_buf;
+	struct sg_table *sgt;
+	struct page **pages;
+	unsigned long nr_pages = size >> PAGE_SHIFT;
+	unsigned int gup_flags = FOLL_LONGTERM;
+	int ret, i, pinned;
+
+	/* Add FOLL_WRITE only if the BO is not read-only */
+	if (!(flags & DRM_IVPU_BO_READ_ONLY))
+		gup_flags |= FOLL_WRITE;
+
+	pages = kvmalloc_array(nr_pages, sizeof(*pages), GFP_KERNEL);
+	if (!pages)
+		return ERR_PTR(-ENOMEM);
+
+	pinned = pin_user_pages_fast((unsigned long)user_ptr, nr_pages, gup_flags, pages);
+	if (pinned < 0) {
+		ret = pinned;
+		ivpu_warn(vdev, "Failed to pin user pages: %d\n", ret);
+		goto free_pages_array;
+	}
+
+	if (pinned != nr_pages) {
+		ivpu_warn(vdev, "Pinned %d pages, expected %lu\n", pinned, nr_pages);
+		ret = -EFAULT;
+		goto unpin_pages;
+	}
+
+	sgt = kmalloc(sizeof(*sgt), GFP_KERNEL);
+	if (!sgt) {
+		ret = -ENOMEM;
+		goto unpin_pages;
+	}
+
+	ret = sg_alloc_table_from_pages(sgt, pages, nr_pages, 0, size, GFP_KERNEL);
+	if (ret) {
+		ivpu_warn(vdev, "Failed to create sg table: %d\n", ret);
+		goto free_sgt;
+	}
+
+	exp_info.exp_name = "ivpu_userptr_dmabuf";
+	exp_info.owner = THIS_MODULE;
+	exp_info.ops = &ivpu_gem_userptr_dmabuf_ops;
+	exp_info.size = size;
+	exp_info.flags = O_RDWR | O_CLOEXEC;
+	exp_info.priv = sgt;
+
+	dma_buf = dma_buf_export(&exp_info);
+	if (IS_ERR(dma_buf)) {
+		ret = PTR_ERR(dma_buf);
+		ivpu_warn(vdev, "Failed to export userptr dma-buf: %d\n", ret);
+		goto free_sg_table;
+	}
+
+	kvfree(pages);
+	return dma_buf;
+
+free_sg_table:
+	sg_free_table(sgt);
+free_sgt:
+	kfree(sgt);
+unpin_pages:
+	for (i = 0; i < pinned; i++)
+		unpin_user_page(pages[i]);
+free_pages_array:
+	kvfree(pages);
+	return ERR_PTR(ret);
+}
+
+static struct ivpu_bo *
+ivpu_bo_create_from_userptr(struct ivpu_device *vdev, void __user *user_ptr,
+			    size_t size, uint32_t flags)
+{
+	struct dma_buf *dma_buf;
+	struct drm_gem_object *obj;
+	struct ivpu_bo *bo;
+
+	dma_buf = ivpu_create_userptr_dmabuf(vdev, user_ptr, size, flags);
+	if (IS_ERR(dma_buf))
+		return ERR_CAST(dma_buf);
+
+	obj = ivpu_gem_prime_import(&vdev->drm, dma_buf);
+	if (IS_ERR(obj)) {
+		dma_buf_put(dma_buf);
+		return ERR_CAST(obj);
+	}
+
+	dma_buf_put(dma_buf);
+
+	bo = to_ivpu_bo(obj);
+	bo->flags = flags;
+
+	return bo;
+}
+
+int ivpu_bo_create_from_userptr_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
+{
+	struct drm_ivpu_bo_create_from_userptr *args = data;
+	struct ivpu_file_priv *file_priv = file->driver_priv;
+	struct ivpu_device *vdev = to_ivpu_device(dev);
+	void __user *user_ptr = u64_to_user_ptr(args->user_ptr);
+	struct ivpu_bo *bo;
+	int ret;
+
+	if (args->flags & ~(DRM_IVPU_BO_HIGH_MEM | DRM_IVPU_BO_DMA_MEM | DRM_IVPU_BO_READ_ONLY))
+		return -EINVAL;
+
+	if (!args->user_ptr || !args->size)
+		return -EINVAL;
+
+	if (!PAGE_ALIGNED(args->user_ptr) || !PAGE_ALIGNED(args->size))
+		return -EINVAL;
+
+	if (!access_ok(user_ptr, args->size))
+		return -EFAULT;
+
+	bo = ivpu_bo_create_from_userptr(vdev, user_ptr, args->size, args->flags);
+	if (IS_ERR(bo))
+		return PTR_ERR(bo);
+
+	ret = drm_gem_handle_create(file, &bo->base.base, &args->handle);
+	if (ret) {
+		ivpu_err(vdev, "Failed to create handle for BO: %pe (ctx %u size %llu flags 0x%x)",
+			 bo, file_priv->ctx.id, args->size, args->flags);
+	} else {
+		ivpu_dbg(vdev, BO, "Created userptr BO: handle=%u vpu_addr=0x%llx size=%llu flags=0x%x\n",
+			 args->handle, bo->vpu_addr, args->size, bo->flags);
+		args->vpu_addr = bo->vpu_addr;
+	}
+
+	drm_gem_object_put(&bo->base.base);
+
+	return ret;
+}
diff --git a/drivers/accel/ivpu/ivpu_mmu_context.c b/drivers/accel/ivpu/ivpu_mmu_context.c
index 4ffc783426be..d128e8961688 100644
--- a/drivers/accel/ivpu/ivpu_mmu_context.c
+++ b/drivers/accel/ivpu/ivpu_mmu_context.c
@@ -430,7 +430,7 @@ static void ivpu_mmu_context_unmap_pages(struct ivpu_mmu_context *ctx, u64 vpu_a
 
 int
 ivpu_mmu_context_map_sgt(struct ivpu_device *vdev, struct ivpu_mmu_context *ctx,
-			 u64 vpu_addr, struct sg_table *sgt,  bool llc_coherent)
+			 u64 vpu_addr, struct sg_table *sgt, bool llc_coherent, bool read_only)
 {
 	size_t start_vpu_addr = vpu_addr;
 	struct scatterlist *sg;
@@ -450,6 +450,8 @@ ivpu_mmu_context_map_sgt(struct ivpu_device *vdev, struct ivpu_mmu_context *ctx,
 	prot = IVPU_MMU_ENTRY_MAPPED;
 	if (llc_coherent)
 		prot |= IVPU_MMU_ENTRY_FLAG_LLC_COHERENT;
+	if (read_only)
+		prot |= IVPU_MMU_ENTRY_FLAG_RO;
 
 	mutex_lock(&ctx->lock);
 
diff --git a/drivers/accel/ivpu/ivpu_mmu_context.h b/drivers/accel/ivpu/ivpu_mmu_context.h
index f255310968cf..663a11a9db11 100644
--- a/drivers/accel/ivpu/ivpu_mmu_context.h
+++ b/drivers/accel/ivpu/ivpu_mmu_context.h
@@ -42,7 +42,7 @@ int ivpu_mmu_context_insert_node(struct ivpu_mmu_context *ctx, const struct ivpu
 void ivpu_mmu_context_remove_node(struct ivpu_mmu_context *ctx, struct drm_mm_node *node);
 
 int ivpu_mmu_context_map_sgt(struct ivpu_device *vdev, struct ivpu_mmu_context *ctx,
-			     u64 vpu_addr, struct sg_table *sgt, bool llc_coherent);
+			     u64 vpu_addr, struct sg_table *sgt, bool llc_coherent, bool read_only);
 void ivpu_mmu_context_unmap_sgt(struct ivpu_device *vdev, struct ivpu_mmu_context *ctx,
 				u64 vpu_addr, struct sg_table *sgt);
 int ivpu_mmu_context_set_pages_ro(struct ivpu_device *vdev, struct ivpu_mmu_context *ctx,
diff --git a/include/uapi/drm/ivpu_accel.h b/include/uapi/drm/ivpu_accel.h
index e470b0221e02..264505d54f93 100644
--- a/include/uapi/drm/ivpu_accel.h
+++ b/include/uapi/drm/ivpu_accel.h
@@ -25,6 +25,7 @@ extern "C" {
 #define DRM_IVPU_CMDQ_CREATE              0x0b
 #define DRM_IVPU_CMDQ_DESTROY             0x0c
 #define DRM_IVPU_CMDQ_SUBMIT              0x0d
+#define DRM_IVPU_BO_CREATE_FROM_USERPTR	  0x0e
 
 #define DRM_IOCTL_IVPU_GET_PARAM                                               \
 	DRM_IOWR(DRM_COMMAND_BASE + DRM_IVPU_GET_PARAM, struct drm_ivpu_param)
@@ -69,6 +70,10 @@ extern "C" {
 #define DRM_IOCTL_IVPU_CMDQ_SUBMIT                                             \
 	DRM_IOW(DRM_COMMAND_BASE + DRM_IVPU_CMDQ_SUBMIT, struct drm_ivpu_cmdq_submit)
 
+#define DRM_IOCTL_IVPU_BO_CREATE_FROM_USERPTR                        \
+	DRM_IOWR(DRM_COMMAND_BASE + DRM_IVPU_BO_CREATE_FROM_USERPTR, \
+		 struct drm_ivpu_bo_create_from_userptr)
+
 /**
  * DOC: contexts
  *
@@ -127,6 +132,13 @@ extern "C" {
  * command queue destroy and submit job on specific command queue.
  */
 #define DRM_IVPU_CAP_MANAGE_CMDQ       3
+/**
+ * DRM_IVPU_CAP_BO_CREATE_FROM_USERPTR
+ *
+ * Driver supports creating buffer objects from user space memory pointers.
+ * This allows creating GEM buffers from existing user memory regions.
+ */
+#define DRM_IVPU_CAP_BO_CREATE_FROM_USERPTR	4
 
 /**
  * struct drm_ivpu_param - Get/Set VPU parameters
@@ -194,6 +206,7 @@ struct drm_ivpu_param {
 #define DRM_IVPU_BO_HIGH_MEM   DRM_IVPU_BO_SHAVE_MEM
 #define DRM_IVPU_BO_MAPPABLE   0x00000002
 #define DRM_IVPU_BO_DMA_MEM    0x00000004
+#define DRM_IVPU_BO_READ_ONLY  0x00000008
 
 #define DRM_IVPU_BO_CACHED     0x00000000
 #define DRM_IVPU_BO_UNCACHED   0x00010000
@@ -204,6 +217,7 @@ struct drm_ivpu_param {
 	(DRM_IVPU_BO_HIGH_MEM | \
 	 DRM_IVPU_BO_MAPPABLE | \
 	 DRM_IVPU_BO_DMA_MEM | \
+	 DRM_IVPU_BO_READ_ONLY | \
 	 DRM_IVPU_BO_CACHE_MASK)
 
 /**
@@ -255,6 +269,44 @@ struct drm_ivpu_bo_create {
 	__u64 vpu_addr;
 };
 
+/**
+ * struct drm_ivpu_bo_create_from_userptr - Create dma-buf from user pointer
+ *
+ * Create a GEM buffer object from a user pointer to a memory region.
+ */
+struct drm_ivpu_bo_create_from_userptr {
+	/** @user_ptr: User pointer to memory region (must be page aligned) */
+	__u64 user_ptr;
+
+	/** @size: Size of the memory region in bytes (must be page aligned) */
+	__u64 size;
+
+	/**
+	 * @flags:
+	 *
+	 * Supported flags:
+	 *
+	 * %DRM_IVPU_BO_HIGH_MEM:
+	 *
+	 * Allocate VPU address from >4GB range.
+	 *
+	 * %DRM_IVPU_BO_DMA_MEM:
+	 *
+	 * Allocate from DMA memory range accessible by hardware DMA.
+	 *
+	 * %DRM_IVPU_BO_READ_ONLY:
+	 *
+	 * Allocate as a read-only buffer object.
+	 */
+	__u32 flags;
+
+	/** @handle: Returned GEM object handle */
+	__u32 handle;
+
+	/** @vpu_addr: Returned VPU virtual address */
+	__u64 vpu_addr;
+};
+
 /**
  * struct drm_ivpu_bo_info - Query buffer object info
  */
-- 
cgit v1.2.3


From d1ac2573f0085a5aab6f5d4cfa7b2818d703e5fe Mon Sep 17 00:00:00 2001
From: Nemesa Garg <nemesa.garg@intel.com>
Date: Tue, 28 Oct 2025 17:37:37 +0530
Subject: drm/drm_crtc: Introduce sharpness strength property

Introduce a new crtc property "SHARPNESS_STRENGTH" that allows
the user to set the intensity so as to get the sharpness effect.
The value of this property can be set from 0-255.
It is useful in scenario when the output is blurry and user
want to sharpen the pixels. User can increase/decrease the
sharpness level depending on the content displayed.

v2: Rename crtc property variable [Arun]
    Add modeset detail in uapi doc[Uma]
v3: Fix build issue
v4: Modify the subject line[Ankit]

Signed-off-by: Nemesa Garg <nemesa.garg@intel.com>
Reviewed-by: Ankit Nautiyal <ankit.k.nautiyal@intel.com>
Tested-by: Adarsh G M <Adarsh.g.m@intel.com>
Acked-by: Simona Vetter <simona.vetter@ffwll.ch>
Link: https://invent.kde.org/plasma/kwin/-/merge_requests/7689
Link: https://patch.msgid.link/20251028120747.3027332-2-ankit.k.nautiyal@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
---
 drivers/gpu/drm/drm_atomic_uapi.c |  4 ++++
 drivers/gpu/drm/drm_crtc.c        | 35 +++++++++++++++++++++++++++++++++++
 include/drm/drm_crtc.h            | 18 ++++++++++++++++++
 3 files changed, 57 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 85dbdaa4a2e2..b2cb5ae5a139 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -419,6 +419,8 @@ static int drm_atomic_crtc_set_property(struct drm_crtc *crtc,
 		set_out_fence_for_crtc(state->state, crtc, fence_ptr);
 	} else if (property == crtc->scaling_filter_property) {
 		state->scaling_filter = val;
+	} else if (property == crtc->sharpness_strength_property) {
+		state->sharpness_strength = val;
 	} else if (crtc->funcs->atomic_set_property) {
 		return crtc->funcs->atomic_set_property(crtc, state, property, val);
 	} else {
@@ -456,6 +458,8 @@ drm_atomic_crtc_get_property(struct drm_crtc *crtc,
 		*val = 0;
 	else if (property == crtc->scaling_filter_property)
 		*val = state->scaling_filter;
+	else if (property == crtc->sharpness_strength_property)
+		*val = state->sharpness_strength;
 	else if (crtc->funcs->atomic_get_property)
 		return crtc->funcs->atomic_get_property(crtc, state, property, val);
 	else {
diff --git a/drivers/gpu/drm/drm_crtc.c b/drivers/gpu/drm/drm_crtc.c
index 46655339003d..a7797d260f1e 100644
--- a/drivers/gpu/drm/drm_crtc.c
+++ b/drivers/gpu/drm/drm_crtc.c
@@ -229,6 +229,25 @@ struct dma_fence *drm_crtc_create_fence(struct drm_crtc *crtc)
  * 		Driver's default scaling filter
  * 	Nearest Neighbor:
  * 		Nearest Neighbor scaling filter
+ * SHARPNESS_STRENGTH:
+ *	Atomic property for setting the sharpness strength/intensity by userspace.
+ *
+ *	The value of this property is set as an integer value ranging
+ *	from 0 - 255 where:
+ *
+ *	0: Sharpness feature is disabled(default value).
+ *
+ *	1: Minimum sharpness.
+ *
+ *	255: Maximum sharpness.
+ *
+ *	User can gradually increase or decrease the sharpness level and can
+ *	set the optimum value depending on content.
+ *	This value will be passed to kernel through the UAPI.
+ *	The setting of this property does not require modeset.
+ *	The sharpness effect takes place post blending on the final composed output.
+ *	If the feature is disabled, the content remains same without any sharpening effect
+ *	and when this feature is applied, it enhances the clarity of the content.
  */
 
 __printf(6, 0)
@@ -940,6 +959,22 @@ int drm_crtc_create_scaling_filter_property(struct drm_crtc *crtc,
 }
 EXPORT_SYMBOL(drm_crtc_create_scaling_filter_property);
 
+int drm_crtc_create_sharpness_strength_property(struct drm_crtc *crtc)
+{
+	struct drm_device *dev = crtc->dev;
+	struct drm_property *prop =
+		drm_property_create_range(dev, 0, "SHARPNESS_STRENGTH", 0, 255);
+
+	if (!prop)
+		return -ENOMEM;
+
+	crtc->sharpness_strength_property = prop;
+	drm_object_attach_property(&crtc->base, prop, 0);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_crtc_create_sharpness_strength_property);
+
 /**
  * drm_crtc_in_clone_mode - check if the given CRTC state is in clone mode
  *
diff --git a/include/drm/drm_crtc.h b/include/drm/drm_crtc.h
index caa56e039da2..bcdbde681986 100644
--- a/include/drm/drm_crtc.h
+++ b/include/drm/drm_crtc.h
@@ -317,6 +317,17 @@ struct drm_crtc_state {
 	 */
 	enum drm_scaling_filter scaling_filter;
 
+	/**
+	 * @sharpness_strength:
+	 *
+	 * Used by the user to set the sharpness intensity.
+	 * The value ranges from 0-255.
+	 * Default value is 0 which disable the sharpness feature.
+	 * Any value greater than 0 enables sharpening with the
+	 * specified strength.
+	 */
+	u8 sharpness_strength;
+
 	/**
 	 * @event:
 	 *
@@ -1088,6 +1099,12 @@ struct drm_crtc {
 	 */
 	struct drm_property *scaling_filter_property;
 
+	/**
+	 * @sharpness_strength_property: property to apply
+	 * the intensity of the sharpness requested.
+	 */
+	struct drm_property *sharpness_strength_property;
+
 	/**
 	 * @state:
 	 *
@@ -1324,4 +1341,5 @@ static inline struct drm_crtc *drm_crtc_find(struct drm_device *dev,
 int drm_crtc_create_scaling_filter_property(struct drm_crtc *crtc,
 					    unsigned int supported_filters);
 bool drm_crtc_in_clone_mode(struct drm_crtc_state *crtc_state);
+int drm_crtc_create_sharpness_strength_property(struct drm_crtc *crtc);
 #endif /* __DRM_CRTC_H__ */
-- 
cgit v1.2.3


From dce4657ff526b65007fe8d5c92968a933cc7c9da Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 27 Oct 2025 13:09:12 +0100
Subject: drm/client: Remove pitch from struct drm_client_buffer

Only the client-buffer setup uses the pitch field from struct
drm_client_buffer. Remove the field and pass the value among setup
helpers.

Clients that need the pitch should rather look at the framebuffer's
pitches[0] directly.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Francesco Valla <francesco@valla.it>
Link: https://patch.msgid.link/20251027121042.143588-2-tzimmermann@suse.de
---
 drivers/gpu/drm/drm_client.c | 14 +++++++-------
 include/drm/drm_client.h     |  5 -----
 2 files changed, 7 insertions(+), 12 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index fe9c6d7083ea..82b871d62313 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -188,7 +188,7 @@ static void drm_client_buffer_delete(struct drm_client_buffer *buffer)
 
 static struct drm_client_buffer *
 drm_client_buffer_create(struct drm_client_dev *client, u32 width, u32 height,
-			 u32 format, u32 *handle)
+			 u32 format, u32 *handle, u32 *pitch)
 {
 	const struct drm_format_info *info = drm_format_info(format);
 	struct drm_mode_create_dumb dumb_args = { };
@@ -216,9 +216,9 @@ drm_client_buffer_create(struct drm_client_dev *client, u32 width, u32 height,
 		goto err_delete;
 	}
 
-	buffer->pitch = dumb_args.pitch;
 	buffer->gem = obj;
 	*handle = dumb_args.handle;
+	*pitch = dumb_args.pitch;
 
 	return buffer;
 
@@ -353,7 +353,7 @@ static void drm_client_buffer_rmfb(struct drm_client_buffer *buffer)
 
 static int drm_client_buffer_addfb(struct drm_client_buffer *buffer,
 				   u32 width, u32 height, u32 format,
-				   u32 handle)
+				   u32 handle, u32 pitch)
 {
 	struct drm_client_dev *client = buffer->client;
 	struct drm_mode_fb_cmd2 fb_req = { };
@@ -363,7 +363,7 @@ static int drm_client_buffer_addfb(struct drm_client_buffer *buffer,
 	fb_req.height = height;
 	fb_req.pixel_format = format;
 	fb_req.handles[0] = handle;
-	fb_req.pitches[0] = buffer->pitch;
+	fb_req.pitches[0] = pitch;
 
 	ret = drm_mode_addfb2(client->dev, &fb_req, client->file);
 	if (ret)
@@ -399,15 +399,15 @@ struct drm_client_buffer *
 drm_client_framebuffer_create(struct drm_client_dev *client, u32 width, u32 height, u32 format)
 {
 	struct drm_client_buffer *buffer;
-	u32 handle;
+	u32 handle, pitch;
 	int ret;
 
 	buffer = drm_client_buffer_create(client, width, height, format,
-					  &handle);
+					  &handle, &pitch);
 	if (IS_ERR(buffer))
 		return buffer;
 
-	ret = drm_client_buffer_addfb(buffer, width, height, format, handle);
+	ret = drm_client_buffer_addfb(buffer, width, height, format, handle, pitch);
 
 	/*
 	 * The handle is only needed for creating the framebuffer, destroy it
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index 715b422952ee..c674464f7e74 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -173,11 +173,6 @@ struct drm_client_buffer {
 	 */
 	struct drm_client_dev *client;
 
-	/**
-	 * @pitch: Buffer pitch
-	 */
-	u32 pitch;
-
 	/**
 	 * @gem: GEM object backing this buffer
 	 *
-- 
cgit v1.2.3


From ea39f2e66e61035e203530977a3df428345d03e2 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 27 Oct 2025 13:09:15 +0100
Subject: drm/client: Deprecate struct drm_client_buffer.gem

The client buffer's framebuffer holds a reference and pointer on
each of its GEM buffer objects. Thus the field gem in the client-
buffer struct is not necessary. Deprecated the field and convert
the client-buffer helpers to use the framebuffer's objects.

In drm_client_buffer_delete(), do a possible vunmap before releasing
the framebuffer. Otherwise we'd eventually release the framebuffer
before unmaping its buffer objects.

v2:
- avoid dependency on CONFIG_DRM_KMS_HELPER

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Francesco Valla <francesco@valla.it>
Link: https://patch.msgid.link/20251027121042.143588-5-tzimmermann@suse.de
---
 drivers/gpu/drm/drm_client.c | 20 ++++++++++++--------
 include/drm/drm_client.h     |  9 +++------
 2 files changed, 15 insertions(+), 14 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index c4db4fc7ba69..0aa56c4b912b 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -17,6 +17,7 @@
 #include <drm/drm_fourcc.h>
 #include <drm/drm_framebuffer.h>
 #include <drm/drm_gem.h>
+#include <drm/drm_gem_framebuffer_helper.h>
 #include <drm/drm_mode.h>
 #include <drm/drm_print.h>
 
@@ -178,17 +179,17 @@ EXPORT_SYMBOL(drm_client_release);
 
 static void drm_client_buffer_delete(struct drm_client_buffer *buffer)
 {
+	struct drm_gem_object *gem = buffer->fb->obj[0];
 	int ret;
 
+	drm_gem_vunmap(gem, &buffer->map);
+
 	ret = drm_mode_rmfb(buffer->client->dev, buffer->fb->base.id, buffer->client->file);
 	if (ret)
 		drm_err(buffer->client->dev,
 			"Error removing FB:%u (%d)\n", buffer->fb->base.id, ret);
 
-	if (buffer->gem) {
-		drm_gem_vunmap(buffer->gem, &buffer->map);
-		drm_gem_object_put(buffer->gem);
-	}
+	drm_gem_object_put(buffer->gem);
 
 	kfree(buffer);
 }
@@ -278,7 +279,7 @@ err_delete:
 int drm_client_buffer_vmap_local(struct drm_client_buffer *buffer,
 				 struct iosys_map *map_copy)
 {
-	struct drm_gem_object *gem = buffer->gem;
+	struct drm_gem_object *gem = buffer->fb->obj[0];
 	struct iosys_map *map = &buffer->map;
 	int ret;
 
@@ -307,7 +308,7 @@ EXPORT_SYMBOL(drm_client_buffer_vmap_local);
  */
 void drm_client_buffer_vunmap_local(struct drm_client_buffer *buffer)
 {
-	struct drm_gem_object *gem = buffer->gem;
+	struct drm_gem_object *gem = buffer->fb->obj[0];
 	struct iosys_map *map = &buffer->map;
 
 	drm_gem_vunmap_locked(gem, map);
@@ -338,9 +339,10 @@ EXPORT_SYMBOL(drm_client_buffer_vunmap_local);
 int drm_client_buffer_vmap(struct drm_client_buffer *buffer,
 			   struct iosys_map *map_copy)
 {
+	struct drm_gem_object *gem = buffer->fb->obj[0];
 	int ret;
 
-	ret = drm_gem_vmap(buffer->gem, &buffer->map);
+	ret = drm_gem_vmap(gem, &buffer->map);
 	if (ret)
 		return ret;
 	*map_copy = buffer->map;
@@ -359,7 +361,9 @@ EXPORT_SYMBOL(drm_client_buffer_vmap);
  */
 void drm_client_buffer_vunmap(struct drm_client_buffer *buffer)
 {
-	drm_gem_vunmap(buffer->gem, &buffer->map);
+	struct drm_gem_object *gem = buffer->fb->obj[0];
+
+	drm_gem_vunmap(gem, &buffer->map);
 }
 EXPORT_SYMBOL(drm_client_buffer_vunmap);
 
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index c674464f7e74..b01fc2a21f09 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -176,12 +176,9 @@ struct drm_client_buffer {
 	/**
 	 * @gem: GEM object backing this buffer
 	 *
-	 * FIXME: The dependency on GEM here isn't required, we could
-	 * convert the driver handle to a dma-buf instead and use the
-	 * backend-agnostic dma-buf vmap support instead. This would
-	 * require that the handle2fd prime ioctl is reworked to pull the
-	 * fd_install step out of the driver backend hooks, to make that
-	 * final step optional for internal users.
+	 * FIXME: The DRM framebuffer holds a reference on its GEM
+	 * buffer objects. Do not use this field in new code and
+	 * update existing users.
 	 */
 	struct drm_gem_object *gem;
 
-- 
cgit v1.2.3


From 3e3153325fd3693d0f9fe235c4afbcd68ef102e1 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 27 Oct 2025 13:09:16 +0100
Subject: drm/client: Remove drm_client_framebuffer_delete()

Release client buffers with drm_client_buffer_delete() instead of
drm_client_framebuffer_delete(). The latter is just a tiny wrapper
around the former.

Move the test for !buffer into drm_client_buffer_delete(), although
all callers appear to always have a valid pointer.

v2:
- test for !buffer before deref-ing pointer (Jocelyn, Dan)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Francesco Valla <francesco@valla.it>
Link: https://patch.msgid.link/20251027121042.143588-6-tzimmermann@suse.de
---
 drivers/gpu/drm/clients/drm_log.c |  4 ++--
 drivers/gpu/drm/drm_client.c      | 28 ++++++++++++----------------
 drivers/gpu/drm/drm_fbdev_dma.c   |  6 +++---
 drivers/gpu/drm/drm_fbdev_shmem.c |  4 ++--
 drivers/gpu/drm/drm_fbdev_ttm.c   |  8 ++++----
 include/drm/drm_client.h          |  2 +-
 6 files changed, 24 insertions(+), 28 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/clients/drm_log.c b/drivers/gpu/drm/clients/drm_log.c
index 24b08fdcb57a..c2ddc57b538e 100644
--- a/drivers/gpu/drm/clients/drm_log.c
+++ b/drivers/gpu/drm/clients/drm_log.c
@@ -272,7 +272,7 @@ static void drm_log_init_client(struct drm_log *dlog)
 
 err_failed_commit:
 	for (i = 0; i < n_modeset; i++)
-		drm_client_framebuffer_delete(dlog->scanout[i].buffer);
+		drm_client_buffer_delete(dlog->scanout[i].buffer);
 
 err_nomodeset:
 	kfree(dlog->scanout);
@@ -286,7 +286,7 @@ static void drm_log_free_scanout(struct drm_client_dev *client)
 
 	if (dlog->n_scanout) {
 		for (i = 0; i < dlog->n_scanout; i++)
-			drm_client_framebuffer_delete(dlog->scanout[i].buffer);
+			drm_client_buffer_delete(dlog->scanout[i].buffer);
 		dlog->n_scanout = 0;
 		kfree(dlog->scanout);
 		dlog->scanout = NULL;
diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index 0aa56c4b912b..b4e37bb2041b 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -177,11 +177,19 @@ void drm_client_release(struct drm_client_dev *client)
 }
 EXPORT_SYMBOL(drm_client_release);
 
-static void drm_client_buffer_delete(struct drm_client_buffer *buffer)
+/**
+ * drm_client_buffer_delete - Delete a client buffer
+ * @buffer: DRM client buffer
+ */
+void drm_client_buffer_delete(struct drm_client_buffer *buffer)
 {
-	struct drm_gem_object *gem = buffer->fb->obj[0];
+	struct drm_gem_object *gem;
 	int ret;
 
+	if (!buffer)
+		return;
+
+	gem = buffer->fb->obj[0];
 	drm_gem_vunmap(gem, &buffer->map);
 
 	ret = drm_mode_rmfb(buffer->client->dev, buffer->fb->base.id, buffer->client->file);
@@ -193,6 +201,7 @@ static void drm_client_buffer_delete(struct drm_client_buffer *buffer)
 
 	kfree(buffer);
 }
+EXPORT_SYMBOL(drm_client_buffer_delete);
 
 static struct drm_client_buffer *
 drm_client_buffer_create(struct drm_client_dev *client, u32 width, u32 height,
@@ -376,7 +385,7 @@ EXPORT_SYMBOL(drm_client_buffer_vunmap);
  *
  * This function creates a &drm_client_buffer which consists of a
  * &drm_framebuffer backed by a dumb buffer.
- * Call drm_client_framebuffer_delete() to free the buffer.
+ * Call drm_client_buffer_delete() to free the buffer.
  *
  * Returns:
  * Pointer to a client buffer or an error pointer on failure.
@@ -420,19 +429,6 @@ err_drm_mode_destroy_dumb:
 }
 EXPORT_SYMBOL(drm_client_framebuffer_create);
 
-/**
- * drm_client_framebuffer_delete - Delete a client framebuffer
- * @buffer: DRM client buffer (can be NULL)
- */
-void drm_client_framebuffer_delete(struct drm_client_buffer *buffer)
-{
-	if (!buffer)
-		return;
-
-	drm_client_buffer_delete(buffer);
-}
-EXPORT_SYMBOL(drm_client_framebuffer_delete);
-
 /**
  * drm_client_framebuffer_flush - Manually flush client framebuffer
  * @buffer: DRM client buffer (can be NULL)
diff --git a/drivers/gpu/drm/drm_fbdev_dma.c b/drivers/gpu/drm/drm_fbdev_dma.c
index c6196293e424..6216de1446c1 100644
--- a/drivers/gpu/drm/drm_fbdev_dma.c
+++ b/drivers/gpu/drm/drm_fbdev_dma.c
@@ -55,7 +55,7 @@ static void drm_fbdev_dma_fb_destroy(struct fb_info *info)
 	drm_fb_helper_fini(fb_helper);
 
 	drm_client_buffer_vunmap(fb_helper->buffer);
-	drm_client_framebuffer_delete(fb_helper->buffer);
+	drm_client_buffer_delete(fb_helper->buffer);
 	drm_client_release(&fb_helper->client);
 }
 
@@ -88,7 +88,7 @@ static void drm_fbdev_dma_shadowed_fb_destroy(struct fb_info *info)
 	vfree(shadow);
 
 	drm_client_buffer_vunmap(fb_helper->buffer);
-	drm_client_framebuffer_delete(fb_helper->buffer);
+	drm_client_buffer_delete(fb_helper->buffer);
 	drm_client_release(&fb_helper->client);
 }
 
@@ -324,7 +324,7 @@ err_drm_client_buffer_vunmap:
 	fb_helper->buffer = NULL;
 	drm_client_buffer_vunmap(buffer);
 err_drm_client_buffer_delete:
-	drm_client_framebuffer_delete(buffer);
+	drm_client_buffer_delete(buffer);
 	return ret;
 }
 EXPORT_SYMBOL(drm_fbdev_dma_driver_fbdev_probe);
diff --git a/drivers/gpu/drm/drm_fbdev_shmem.c b/drivers/gpu/drm/drm_fbdev_shmem.c
index 51573058df6f..520c2218e5dc 100644
--- a/drivers/gpu/drm/drm_fbdev_shmem.c
+++ b/drivers/gpu/drm/drm_fbdev_shmem.c
@@ -63,7 +63,7 @@ static void drm_fbdev_shmem_fb_destroy(struct fb_info *info)
 	drm_fb_helper_fini(fb_helper);
 
 	drm_client_buffer_vunmap(fb_helper->buffer);
-	drm_client_framebuffer_delete(fb_helper->buffer);
+	drm_client_buffer_delete(fb_helper->buffer);
 	drm_client_release(&fb_helper->client);
 }
 
@@ -204,7 +204,7 @@ err_drm_client_buffer_vunmap:
 	fb_helper->buffer = NULL;
 	drm_client_buffer_vunmap(buffer);
 err_drm_client_buffer_delete:
-	drm_client_framebuffer_delete(buffer);
+	drm_client_buffer_delete(buffer);
 	return ret;
 }
 EXPORT_SYMBOL(drm_fbdev_shmem_driver_fbdev_probe);
diff --git a/drivers/gpu/drm/drm_fbdev_ttm.c b/drivers/gpu/drm/drm_fbdev_ttm.c
index ccf460fbc1f0..7f7c88461228 100644
--- a/drivers/gpu/drm/drm_fbdev_ttm.c
+++ b/drivers/gpu/drm/drm_fbdev_ttm.c
@@ -50,7 +50,7 @@ static void drm_fbdev_ttm_fb_destroy(struct fb_info *info)
 	fb_deferred_io_cleanup(info);
 	drm_fb_helper_fini(fb_helper);
 	vfree(shadow);
-	drm_client_framebuffer_delete(fb_helper->buffer);
+	drm_client_buffer_delete(fb_helper->buffer);
 
 	drm_client_release(&fb_helper->client);
 }
@@ -200,7 +200,7 @@ int drm_fbdev_ttm_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	screen_buffer = vzalloc(screen_size);
 	if (!screen_buffer) {
 		ret = -ENOMEM;
-		goto err_drm_client_framebuffer_delete;
+		goto err_drm_client_buffer_delete;
 	}
 
 	info = drm_fb_helper_alloc_info(fb_helper);
@@ -233,10 +233,10 @@ err_drm_fb_helper_release_info:
 	drm_fb_helper_release_info(fb_helper);
 err_vfree:
 	vfree(screen_buffer);
-err_drm_client_framebuffer_delete:
+err_drm_client_buffer_delete:
 	fb_helper->fb = NULL;
 	fb_helper->buffer = NULL;
-	drm_client_framebuffer_delete(buffer);
+	drm_client_buffer_delete(buffer);
 	return ret;
 }
 EXPORT_SYMBOL(drm_fbdev_ttm_driver_fbdev_probe);
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index b01fc2a21f09..ffc4013b2e18 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -195,7 +195,7 @@ struct drm_client_buffer {
 
 struct drm_client_buffer *
 drm_client_framebuffer_create(struct drm_client_dev *client, u32 width, u32 height, u32 format);
-void drm_client_framebuffer_delete(struct drm_client_buffer *buffer);
+void drm_client_buffer_delete(struct drm_client_buffer *buffer);
 int drm_client_framebuffer_flush(struct drm_client_buffer *buffer, struct drm_rect *rect);
 int drm_client_buffer_vmap_local(struct drm_client_buffer *buffer,
 				 struct iosys_map *map_copy);
-- 
cgit v1.2.3


From c2707e0f8322607b65e5eb8362ba94a2aeb299b9 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 27 Oct 2025 13:09:17 +0100
Subject: drm/client: Create client buffers with
 drm_client_buffer_create_dumb()

Rename drm_client_framebuffer_create() to drm_client_buffer_create_dump()
and adapt callers. The new name reflects the function's purpose. Using
dumb buffers is the easiest way for creating a GEM buffer in a drivers-
independent way.

There's also drm_client_buffer_create(), which creates the client buffer
from a preexisting buffer object. This helper can be exported for drivers
that create their own GEM buffer object.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Tested-by: Francesco Valla <francesco@valla.it>
Link: https://patch.msgid.link/20251027121042.143588-7-tzimmermann@suse.de
---
 drivers/gpu/drm/clients/drm_log.c | 2 +-
 drivers/gpu/drm/drm_client.c      | 6 +++---
 drivers/gpu/drm/drm_fbdev_dma.c   | 2 +-
 drivers/gpu/drm/drm_fbdev_shmem.c | 2 +-
 drivers/gpu/drm/drm_fbdev_ttm.c   | 2 +-
 include/drm/drm_client.h          | 2 +-
 6 files changed, 8 insertions(+), 8 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/clients/drm_log.c b/drivers/gpu/drm/clients/drm_log.c
index c2ddc57b538e..48636bb1a21e 100644
--- a/drivers/gpu/drm/clients/drm_log.c
+++ b/drivers/gpu/drm/clients/drm_log.c
@@ -204,7 +204,7 @@ static int drm_log_setup_modeset(struct drm_client_dev *client,
 	if (format == DRM_FORMAT_INVALID)
 		return -EINVAL;
 
-	scanout->buffer = drm_client_framebuffer_create(client, width, height, format);
+	scanout->buffer = drm_client_buffer_create_dumb(client, width, height, format);
 	if (IS_ERR(scanout->buffer)) {
 		drm_warn(client->dev, "drm_log can't create framebuffer %d %d %p4cc\n",
 			 width, height, &format);
diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index b4e37bb2041b..e7dfbdeca45a 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -377,7 +377,7 @@ void drm_client_buffer_vunmap(struct drm_client_buffer *buffer)
 EXPORT_SYMBOL(drm_client_buffer_vunmap);
 
 /**
- * drm_client_framebuffer_create - Create a client framebuffer
+ * drm_client_buffer_create_dumb - Create a client buffer backed by a dumb buffer
  * @client: DRM client
  * @width: Framebuffer width
  * @height: Framebuffer height
@@ -391,7 +391,7 @@ EXPORT_SYMBOL(drm_client_buffer_vunmap);
  * Pointer to a client buffer or an error pointer on failure.
  */
 struct drm_client_buffer *
-drm_client_framebuffer_create(struct drm_client_dev *client, u32 width, u32 height, u32 format)
+drm_client_buffer_create_dumb(struct drm_client_dev *client, u32 width, u32 height, u32 format)
 {
 	const struct drm_format_info *info = drm_format_info(format);
 	struct drm_device *dev = client->dev;
@@ -427,7 +427,7 @@ err_drm_mode_destroy_dumb:
 	drm_mode_destroy_dumb(client->dev, dumb_args.handle, client->file);
 	return ERR_PTR(ret);
 }
-EXPORT_SYMBOL(drm_client_framebuffer_create);
+EXPORT_SYMBOL(drm_client_buffer_create_dumb);
 
 /**
  * drm_client_framebuffer_flush - Manually flush client framebuffer
diff --git a/drivers/gpu/drm/drm_fbdev_dma.c b/drivers/gpu/drm/drm_fbdev_dma.c
index 6216de1446c1..876bd8cfc5ea 100644
--- a/drivers/gpu/drm/drm_fbdev_dma.c
+++ b/drivers/gpu/drm/drm_fbdev_dma.c
@@ -281,7 +281,7 @@ int drm_fbdev_dma_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 
 	format = drm_driver_legacy_fb_format(dev, sizes->surface_bpp,
 					     sizes->surface_depth);
-	buffer = drm_client_framebuffer_create(client, sizes->surface_width,
+	buffer = drm_client_buffer_create_dumb(client, sizes->surface_width,
 					       sizes->surface_height, format);
 	if (IS_ERR(buffer))
 		return PTR_ERR(buffer);
diff --git a/drivers/gpu/drm/drm_fbdev_shmem.c b/drivers/gpu/drm/drm_fbdev_shmem.c
index 520c2218e5dc..46e43b60b3f9 100644
--- a/drivers/gpu/drm/drm_fbdev_shmem.c
+++ b/drivers/gpu/drm/drm_fbdev_shmem.c
@@ -147,7 +147,7 @@ int drm_fbdev_shmem_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 		    sizes->surface_bpp);
 
 	format = drm_driver_legacy_fb_format(dev, sizes->surface_bpp, sizes->surface_depth);
-	buffer = drm_client_framebuffer_create(client, sizes->surface_width,
+	buffer = drm_client_buffer_create_dumb(client, sizes->surface_width,
 					       sizes->surface_height, format);
 	if (IS_ERR(buffer))
 		return PTR_ERR(buffer);
diff --git a/drivers/gpu/drm/drm_fbdev_ttm.c b/drivers/gpu/drm/drm_fbdev_ttm.c
index 7f7c88461228..c7ad779ba590 100644
--- a/drivers/gpu/drm/drm_fbdev_ttm.c
+++ b/drivers/gpu/drm/drm_fbdev_ttm.c
@@ -187,7 +187,7 @@ int drm_fbdev_ttm_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 
 	format = drm_driver_legacy_fb_format(dev, sizes->surface_bpp,
 					     sizes->surface_depth);
-	buffer = drm_client_framebuffer_create(client, sizes->surface_width,
+	buffer = drm_client_buffer_create_dumb(client, sizes->surface_width,
 					       sizes->surface_height, format);
 	if (IS_ERR(buffer))
 		return PTR_ERR(buffer);
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index ffc4013b2e18..690ef04fccce 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -194,7 +194,7 @@ struct drm_client_buffer {
 };
 
 struct drm_client_buffer *
-drm_client_framebuffer_create(struct drm_client_dev *client, u32 width, u32 height, u32 format);
+drm_client_buffer_create_dumb(struct drm_client_dev *client, u32 width, u32 height, u32 format);
 void drm_client_buffer_delete(struct drm_client_buffer *buffer);
 int drm_client_framebuffer_flush(struct drm_client_buffer *buffer, struct drm_rect *rect);
 int drm_client_buffer_vmap_local(struct drm_client_buffer *buffer,
-- 
cgit v1.2.3


From 231668043d4ffebda28630b120cddcba384a3318 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 27 Oct 2025 13:09:18 +0100
Subject: drm/client: Flush client buffers with drm_client_buffer_sync()

Rename drm_client_framebuffer_flush() to drm_cient_buffer_flush() and
adapt its callers. The old name was left over from previous naming
conventions.

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>>
Tested-by: Francesco Valla <francesco@valla.it>
Link: https://patch.msgid.link/20251027121042.143588-8-tzimmermann@suse.de
---
 drivers/gpu/drm/clients/drm_log.c | 4 ++--
 drivers/gpu/drm/drm_client.c      | 8 ++++----
 include/drm/drm_client.h          | 2 +-
 3 files changed, 7 insertions(+), 7 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/clients/drm_log.c b/drivers/gpu/drm/clients/drm_log.c
index 48636bb1a21e..19e55aa0ed74 100644
--- a/drivers/gpu/drm/clients/drm_log.c
+++ b/drivers/gpu/drm/clients/drm_log.c
@@ -100,7 +100,7 @@ static void drm_log_clear_line(struct drm_log_scanout *scanout, u32 line)
 		return;
 	iosys_map_memset(&map, r.y1 * fb->pitches[0], 0, height * fb->pitches[0]);
 	drm_client_buffer_vunmap_local(scanout->buffer);
-	drm_client_framebuffer_flush(scanout->buffer, &r);
+	drm_client_buffer_flush(scanout->buffer, &r);
 }
 
 static void drm_log_draw_line(struct drm_log_scanout *scanout, const char *s,
@@ -133,7 +133,7 @@ static void drm_log_draw_line(struct drm_log_scanout *scanout, const char *s,
 	if (scanout->line >= scanout->rows)
 		scanout->line = 0;
 	drm_client_buffer_vunmap_local(scanout->buffer);
-	drm_client_framebuffer_flush(scanout->buffer, &r);
+	drm_client_buffer_flush(scanout->buffer, &r);
 }
 
 static void drm_log_draw_new_line(struct drm_log_scanout *scanout,
diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index e7dfbdeca45a..504ec5bdfa2c 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -430,8 +430,8 @@ err_drm_mode_destroy_dumb:
 EXPORT_SYMBOL(drm_client_buffer_create_dumb);
 
 /**
- * drm_client_framebuffer_flush - Manually flush client framebuffer
- * @buffer: DRM client buffer (can be NULL)
+ * drm_client_buffer_flush - Manually flush client buffer
+ * @buffer: DRM client buffer
  * @rect: Damage rectangle (if NULL flushes all)
  *
  * This calls &drm_framebuffer_funcs->dirty (if present) to flush buffer changes
@@ -440,7 +440,7 @@ EXPORT_SYMBOL(drm_client_buffer_create_dumb);
  * Returns:
  * Zero on success or negative error code on failure.
  */
-int drm_client_framebuffer_flush(struct drm_client_buffer *buffer, struct drm_rect *rect)
+int drm_client_buffer_flush(struct drm_client_buffer *buffer, struct drm_rect *rect)
 {
 	if (!buffer || !buffer->fb || !buffer->fb->funcs->dirty)
 		return 0;
@@ -460,4 +460,4 @@ int drm_client_framebuffer_flush(struct drm_client_buffer *buffer, struct drm_re
 	return buffer->fb->funcs->dirty(buffer->fb, buffer->client->file,
 					0, 0, NULL, 0);
 }
-EXPORT_SYMBOL(drm_client_framebuffer_flush);
+EXPORT_SYMBOL(drm_client_buffer_flush);
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index 690ef04fccce..5ecde0f6f591 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -196,7 +196,7 @@ struct drm_client_buffer {
 struct drm_client_buffer *
 drm_client_buffer_create_dumb(struct drm_client_dev *client, u32 width, u32 height, u32 format);
 void drm_client_buffer_delete(struct drm_client_buffer *buffer);
-int drm_client_framebuffer_flush(struct drm_client_buffer *buffer, struct drm_rect *rect);
+int drm_client_buffer_flush(struct drm_client_buffer *buffer, struct drm_rect *rect);
 int drm_client_buffer_vmap_local(struct drm_client_buffer *buffer,
 				 struct iosys_map *map_copy);
 void drm_client_buffer_vunmap_local(struct drm_client_buffer *buffer);
-- 
cgit v1.2.3


From 9695c143b72a7faa2dbbb2a5881269f82e6f9783 Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Wed, 29 Oct 2025 12:39:46 +0200
Subject: drm/buddy: replace drm_print.h include with a forward declaration

The drm_buddy.h header does not really need anything from drm_print.h. A
simple forward declaration for struct drm_printer is sufficient.

An explicit drm_print.h include has previously been added to all the
files that indirectly depended on this include.

v3: Only remove the include here (Thomas)

Cc: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://lore.kernel.org/r/b303996b407fcbe2c7357bea036f79c45d6dae49.1761734313.git.jani.nikula@intel.com
---
 include/drm/drm_buddy.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_buddy.h b/include/drm/drm_buddy.h
index c2e05a281252..b909fa8f810a 100644
--- a/include/drm/drm_buddy.h
+++ b/include/drm/drm_buddy.h
@@ -12,7 +12,7 @@
 #include <linux/sched.h>
 #include <linux/rbtree.h>
 
-#include <drm/drm_print.h>
+struct drm_printer;
 
 #define DRM_BUDDY_RANGE_ALLOCATION		BIT(0)
 #define DRM_BUDDY_TOPDOWN_ALLOCATION		BIT(1)
-- 
cgit v1.2.3


From ea722522d505fa8fb3533a386adeb400607e5072 Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Wed, 29 Oct 2025 12:39:47 +0200
Subject: drm/mm: replace drm_print.h include with a forward declaration

The drm_mm.h header does not really need anything from drm_print.h. A
simple forward declaration for struct drm_printer is sufficient.

An explicit drm_print.h include has previously been added to all the
files that indirectly depended on this include.

v3: Only remove	the include here (Thomas)

Cc: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://lore.kernel.org/r/7d570ed1f0f0f14cac346bea50bce9ef02ddd166.1761734313.git.jani.nikula@intel.com
---
 include/drm/drm_mm.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/drm_mm.h b/include/drm/drm_mm.h
index f654874c4ce6..16ce0e8f36a6 100644
--- a/include/drm/drm_mm.h
+++ b/include/drm/drm_mm.h
@@ -48,7 +48,7 @@
 #endif
 #include <linux/types.h>
 
-#include <drm/drm_print.h>
+struct drm_printer;
 
 #ifdef CONFIG_DRM_DEBUG_MM
 #define DRM_MM_BUG_ON(expr) BUG_ON(expr)
-- 
cgit v1.2.3


From d7a849d126d0a75b2c5101f82d0c9693e04a43fd Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Wed, 29 Oct 2025 12:39:48 +0200
Subject: drm/ttm: replace drm_print.h include with a forward declaration

The ttm/ttm_resource.h header does not really need anything from
drm_print.h. A simple forward declaration for struct drm_printer is
sufficient.

An explicit drm_print.h include has previously been added to all the
files that indirectly depended on this include.

v3: Only remove the include here (Thomas)

Cc: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Thomas Zimmermann <tzimmermann@suse.de>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://lore.kernel.org/r/cfdb1095033112c2a7e58767481c98929984a33c.1761734313.git.jani.nikula@intel.com
---
 include/drm/ttm/ttm_resource.h | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/ttm/ttm_resource.h b/include/drm/ttm/ttm_resource.h
index f49daa504c36..68bf010d8b40 100644
--- a/include/drm/ttm/ttm_resource.h
+++ b/include/drm/ttm/ttm_resource.h
@@ -31,14 +31,15 @@
 #include <linux/iosys-map.h>
 #include <linux/dma-fence.h>
 
-#include <drm/drm_print.h>
 #include <drm/ttm/ttm_caching.h>
 #include <drm/ttm/ttm_kmap_iter.h>
 
 #define TTM_MAX_BO_PRIORITY	4U
 #define TTM_NUM_MEM_TYPES 9
 
+struct dentry;
 struct dmem_cgroup_device;
+struct drm_printer;
 struct ttm_device;
 struct ttm_resource_manager;
 struct ttm_resource;
-- 
cgit v1.2.3


From 0af5b6a8f8dd41fd801bb0f2af3295d69ba8b7fe Mon Sep 17 00:00:00 2001
From: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Date: Mon, 20 Oct 2025 12:54:07 +0100
Subject: drm/ttm: Replace multiple booleans with flags in pool init
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Multiple consecutive boolean function arguments are usually not very
readable.

Replace the ones in ttm_pool_init() with flags with the additional
benefit of soon being able to pass in more data with just this one
code base churning cost.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-3-tvrtko.ursulin@igalia.com
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c     |  2 +-
 drivers/gpu/drm/ttm/tests/ttm_device_test.c | 25 ++++++++++---------------
 drivers/gpu/drm/ttm/tests/ttm_pool_test.c   | 24 +++++++++++-------------
 drivers/gpu/drm/ttm/ttm_device.c            |  5 ++++-
 drivers/gpu/drm/ttm/ttm_pool.c              |  8 +++-----
 drivers/gpu/drm/ttm/ttm_pool_internal.h     |  5 +++--
 include/drm/ttm/ttm_allocation.h            | 10 ++++++++++
 include/drm/ttm/ttm_pool.h                  |  8 +++-----
 8 files changed, 45 insertions(+), 42 deletions(-)
 create mode 100644 include/drm/ttm/ttm_allocation.h

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index aa9ee5dffa45..8f6d331e1ea2 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1837,7 +1837,7 @@ static int amdgpu_ttm_pools_init(struct amdgpu_device *adev)
 	for (i = 0; i < adev->gmc.num_mem_partitions; i++) {
 		ttm_pool_init(&adev->mman.ttm_pools[i], adev->dev,
 			      adev->gmc.mem_partitions[i].numa.node,
-			      false, false);
+			      0);
 	}
 	return 0;
 }
diff --git a/drivers/gpu/drm/ttm/tests/ttm_device_test.c b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
index 1621903818e5..98648d5f20e7 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_device_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
@@ -7,11 +7,11 @@
 #include <drm/ttm/ttm_placement.h>
 
 #include "ttm_kunit_helpers.h"
+#include "../ttm_pool_internal.h"
 
 struct ttm_device_test_case {
 	const char *description;
-	bool use_dma_alloc;
-	bool use_dma32;
+	unsigned int alloc_flags;
 	bool pools_init_expected;
 };
 
@@ -119,26 +119,22 @@ static void ttm_device_init_no_vma_man(struct kunit *test)
 static const struct ttm_device_test_case ttm_device_cases[] = {
 	{
 		.description = "No DMA allocations, no DMA32 required",
-		.use_dma_alloc = false,
-		.use_dma32 = false,
 		.pools_init_expected = false,
 	},
 	{
 		.description = "DMA allocations, DMA32 required",
-		.use_dma_alloc = true,
-		.use_dma32 = true,
+		.alloc_flags = TTM_ALLOCATION_POOL_USE_DMA_ALLOC |
+			       TTM_ALLOCATION_POOL_USE_DMA32,
 		.pools_init_expected = true,
 	},
 	{
 		.description = "No DMA allocations, DMA32 required",
-		.use_dma_alloc = false,
-		.use_dma32 = true,
+		.alloc_flags = TTM_ALLOCATION_POOL_USE_DMA32,
 		.pools_init_expected = false,
 	},
 	{
 		.description = "DMA allocations, no DMA32 required",
-		.use_dma_alloc = true,
-		.use_dma32 = false,
+		.alloc_flags = TTM_ALLOCATION_POOL_USE_DMA_ALLOC,
 		.pools_init_expected = true,
 	},
 };
@@ -163,15 +159,14 @@ static void ttm_device_init_pools(struct kunit *test)
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
 	err = ttm_device_kunit_init(priv, ttm_dev,
-				    params->use_dma_alloc,
-				    params->use_dma32);
+				    params->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA_ALLOC,
+				    params->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA32);
 	KUNIT_ASSERT_EQ(test, err, 0);
 
 	pool = &ttm_dev->pool;
 	KUNIT_ASSERT_NOT_NULL(test, pool);
 	KUNIT_EXPECT_PTR_EQ(test, pool->dev, priv->dev);
-	KUNIT_EXPECT_EQ(test, pool->use_dma_alloc, params->use_dma_alloc);
-	KUNIT_EXPECT_EQ(test, pool->use_dma32, params->use_dma32);
+	KUNIT_EXPECT_EQ(test, pool->alloc_flags, params->alloc_flags);
 
 	if (params->pools_init_expected) {
 		for (int i = 0; i < TTM_NUM_CACHING_TYPES; ++i) {
@@ -181,7 +176,7 @@ static void ttm_device_init_pools(struct kunit *test)
 				KUNIT_EXPECT_EQ(test, pt.caching, i);
 				KUNIT_EXPECT_EQ(test, pt.order, j);
 
-				if (params->use_dma_alloc)
+				if (ttm_pool_uses_dma_alloc(pool))
 					KUNIT_ASSERT_FALSE(test,
 							   list_empty(&pt.pages));
 			}
diff --git a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
index 17ebb9fbd688..11c92bd75779 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_pool_test.c
@@ -13,7 +13,7 @@
 struct ttm_pool_test_case {
 	const char *description;
 	unsigned int order;
-	bool use_dma_alloc;
+	unsigned int alloc_flags;
 };
 
 struct ttm_pool_test_priv {
@@ -87,7 +87,7 @@ static struct ttm_pool *ttm_pool_pre_populated(struct kunit *test,
 	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, pool);
 
-	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, true, false);
+	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, TTM_ALLOCATION_POOL_USE_DMA_ALLOC);
 
 	err = ttm_pool_alloc(pool, tt, &simple_ctx);
 	KUNIT_ASSERT_EQ(test, err, 0);
@@ -114,12 +114,12 @@ static const struct ttm_pool_test_case ttm_pool_basic_cases[] = {
 	{
 		.description = "One page, with coherent DMA mappings enabled",
 		.order = 0,
-		.use_dma_alloc = true,
+		.alloc_flags = TTM_ALLOCATION_POOL_USE_DMA_ALLOC,
 	},
 	{
 		.description = "Above the allocation limit, with coherent DMA mappings enabled",
 		.order = MAX_PAGE_ORDER + 1,
-		.use_dma_alloc = true,
+		.alloc_flags = TTM_ALLOCATION_POOL_USE_DMA_ALLOC,
 	},
 };
 
@@ -151,13 +151,11 @@ static void ttm_pool_alloc_basic(struct kunit *test)
 	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, pool);
 
-	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, params->use_dma_alloc,
-		      false);
+	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, params->alloc_flags);
 
 	KUNIT_ASSERT_PTR_EQ(test, pool->dev, devs->dev);
 	KUNIT_ASSERT_EQ(test, pool->nid, NUMA_NO_NODE);
-	KUNIT_ASSERT_EQ(test, ttm_pool_uses_dma_alloc(pool),
-			params->use_dma_alloc);
+	KUNIT_ASSERT_EQ(test, pool->alloc_flags, params->alloc_flags);
 
 	err = ttm_pool_alloc(pool, tt, &simple_ctx);
 	KUNIT_ASSERT_EQ(test, err, 0);
@@ -167,14 +165,14 @@ static void ttm_pool_alloc_basic(struct kunit *test)
 	last_page = tt->pages[tt->num_pages - 1];
 
 	if (params->order <= MAX_PAGE_ORDER) {
-		if (params->use_dma_alloc) {
+		if (ttm_pool_uses_dma_alloc(pool)) {
 			KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
 			KUNIT_ASSERT_NOT_NULL(test, (void *)last_page->private);
 		} else {
 			KUNIT_ASSERT_EQ(test, fst_page->private, params->order);
 		}
 	} else {
-		if (params->use_dma_alloc) {
+		if (ttm_pool_uses_dma_alloc(pool)) {
 			KUNIT_ASSERT_NOT_NULL(test, (void *)fst_page->private);
 			KUNIT_ASSERT_NULL(test, (void *)last_page->private);
 		} else {
@@ -220,7 +218,7 @@ static void ttm_pool_alloc_basic_dma_addr(struct kunit *test)
 	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, pool);
 
-	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, true, false);
+	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, TTM_ALLOCATION_POOL_USE_DMA_ALLOC);
 
 	err = ttm_pool_alloc(pool, tt, &simple_ctx);
 	KUNIT_ASSERT_EQ(test, err, 0);
@@ -350,7 +348,7 @@ static void ttm_pool_free_dma_alloc(struct kunit *test)
 	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, pool);
 
-	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, true, false);
+	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, TTM_ALLOCATION_POOL_USE_DMA_ALLOC);
 	ttm_pool_alloc(pool, tt, &simple_ctx);
 
 	pt = &pool->caching[caching].orders[order];
@@ -381,7 +379,7 @@ static void ttm_pool_free_no_dma_alloc(struct kunit *test)
 	pool = kunit_kzalloc(test, sizeof(*pool), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, pool);
 
-	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, false, false);
+	ttm_pool_init(pool, devs->dev, NUMA_NO_NODE, 0);
 	ttm_pool_alloc(pool, tt, &simple_ctx);
 
 	pt = &pool->caching[caching].orders[order];
diff --git a/drivers/gpu/drm/ttm/ttm_device.c b/drivers/gpu/drm/ttm/ttm_device.c
index c3e2fcbdd2cc..a97b1444536c 100644
--- a/drivers/gpu/drm/ttm/ttm_device.c
+++ b/drivers/gpu/drm/ttm/ttm_device.c
@@ -31,6 +31,7 @@
 #include <linux/export.h>
 #include <linux/mm.h>
 
+#include <drm/ttm/ttm_allocation.h>
 #include <drm/ttm/ttm_bo.h>
 #include <drm/ttm/ttm_device.h>
 #include <drm/ttm/ttm_tt.h>
@@ -236,7 +237,9 @@ int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *func
 	else
 		nid = NUMA_NO_NODE;
 
-	ttm_pool_init(&bdev->pool, dev, nid, use_dma_alloc, use_dma32);
+	ttm_pool_init(&bdev->pool, dev, nid,
+		      (use_dma_alloc ? TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0) |
+		      (use_dma32 ? TTM_ALLOCATION_POOL_USE_DMA32 : 0));
 
 	bdev->vma_manager = vma_manager;
 	spin_lock_init(&bdev->lru_lock);
diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c
index ff6fab4122bb..4fc69447060c 100644
--- a/drivers/gpu/drm/ttm/ttm_pool.c
+++ b/drivers/gpu/drm/ttm/ttm_pool.c
@@ -1059,13 +1059,12 @@ long ttm_pool_backup(struct ttm_pool *pool, struct ttm_tt *tt,
  * @pool: the pool to initialize
  * @dev: device for DMA allocations and mappings
  * @nid: NUMA node to use for allocations
- * @use_dma_alloc: true if coherent DMA alloc should be used
- * @use_dma32: true if GFP_DMA32 should be used
+ * @alloc_flags: TTM_ALLOCATION_POOL_ flags
  *
  * Initialize the pool and its pool types.
  */
 void ttm_pool_init(struct ttm_pool *pool, struct device *dev,
-		   int nid, bool use_dma_alloc, bool use_dma32)
+		   int nid, unsigned int alloc_flags)
 {
 	unsigned int i, j;
 
@@ -1073,8 +1072,7 @@ void ttm_pool_init(struct ttm_pool *pool, struct device *dev,
 
 	pool->dev = dev;
 	pool->nid = nid;
-	pool->use_dma_alloc = use_dma_alloc;
-	pool->use_dma32 = use_dma32;
+	pool->alloc_flags = alloc_flags;
 
 	for (i = 0; i < TTM_NUM_CACHING_TYPES; ++i) {
 		for (j = 0; j < NR_PAGE_ORDERS; ++j) {
diff --git a/drivers/gpu/drm/ttm/ttm_pool_internal.h b/drivers/gpu/drm/ttm/ttm_pool_internal.h
index 3e50d30bd95a..96b7f21514fb 100644
--- a/drivers/gpu/drm/ttm/ttm_pool_internal.h
+++ b/drivers/gpu/drm/ttm/ttm_pool_internal.h
@@ -4,16 +4,17 @@
 #ifndef _TTM_POOL_INTERNAL_H_
 #define _TTM_POOL_INTERNAL_H_
 
+#include <drm/ttm/ttm_allocation.h>
 #include <drm/ttm/ttm_pool.h>
 
 static inline bool ttm_pool_uses_dma_alloc(struct ttm_pool *pool)
 {
-	return pool->use_dma_alloc;
+	return pool->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA_ALLOC;
 }
 
 static inline bool ttm_pool_uses_dma32(struct ttm_pool *pool)
 {
-	return pool->use_dma32;
+	return pool->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA32;
 }
 
 #endif
diff --git a/include/drm/ttm/ttm_allocation.h b/include/drm/ttm/ttm_allocation.h
new file mode 100644
index 000000000000..7869dc32bd91
--- /dev/null
+++ b/include/drm/ttm/ttm_allocation.h
@@ -0,0 +1,10 @@
+/* SPDX-License-Identifier: GPL-2.0 OR MIT */
+/* Copyright (c) 2025 Valve Corporation */
+
+#ifndef _TTM_ALLOCATION_H_
+#define _TTM_ALLOCATION_H_
+
+#define TTM_ALLOCATION_POOL_USE_DMA_ALLOC	BIT(0) /* Use coherent DMA allocations. */
+#define TTM_ALLOCATION_POOL_USE_DMA32		BIT(1) /* Use GFP_DMA32 allocations. */
+
+#endif
diff --git a/include/drm/ttm/ttm_pool.h b/include/drm/ttm/ttm_pool.h
index 54cd34a6e4c0..67c72de913bb 100644
--- a/include/drm/ttm/ttm_pool.h
+++ b/include/drm/ttm/ttm_pool.h
@@ -64,16 +64,14 @@ struct ttm_pool_type {
  *
  * @dev: the device we allocate pages for
  * @nid: which numa node to use
- * @use_dma_alloc: if coherent DMA allocations should be used
- * @use_dma32: if GFP_DMA32 should be used
+ * @alloc_flags: TTM_ALLOCATION_POOL_ flags
  * @caching: pools for each caching/order
  */
 struct ttm_pool {
 	struct device *dev;
 	int nid;
 
-	bool use_dma_alloc;
-	bool use_dma32;
+	unsigned int alloc_flags;
 
 	struct {
 		struct ttm_pool_type orders[NR_PAGE_ORDERS];
@@ -85,7 +83,7 @@ int ttm_pool_alloc(struct ttm_pool *pool, struct ttm_tt *tt,
 void ttm_pool_free(struct ttm_pool *pool, struct ttm_tt *tt);
 
 void ttm_pool_init(struct ttm_pool *pool, struct device *dev,
-		   int nid, bool use_dma_alloc, bool use_dma32);
+		   int nid, unsigned int alloc_flags);
 void ttm_pool_fini(struct ttm_pool *pool);
 
 int ttm_pool_debugfs(struct ttm_pool *pool, struct seq_file *m);
-- 
cgit v1.2.3


From 77e19f8d32979f00b7c2cbcb35dbbf6f2116518e Mon Sep 17 00:00:00 2001
From: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Date: Mon, 20 Oct 2025 12:54:08 +0100
Subject: drm/ttm: Replace multiple booleans with flags in device init
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Multiple consecutive boolean function arguments are usually not very
readable.

Replace the ones in ttm_device_init() with flags with the additional
benefit of soon being able to pass in more data with just a one off
code base churning cost.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Alex Deucher <alexander.deucher@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Danilo Krummrich <dakr@kernel.org>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Gerd Hoffmann <kraxel@redhat.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Lyude Paul <lyude@redhat.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Cc: Maxime Ripard <mripard@kernel.org>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Sui Jingfeng <suijingfeng@loongson.cn>
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Thomas Zimmermann <tzimmermann@suse.de>
Cc: Zack Rusin <zack.rusin@broadcom.com>
Acked-by: Christian König <christian.koenig@amd.com>
Acked-by: Zack Rusin <zack.rusin@broadcom.com>
Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> # For xe
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-4-tvrtko.ursulin@igalia.com
[tursulin: fixup checkpatch while applying]
---
 drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c          |  6 ++++--
 drivers/gpu/drm/drm_gem_vram_helper.c            |  2 +-
 drivers/gpu/drm/i915/intel_region_ttm.c          |  2 +-
 drivers/gpu/drm/loongson/lsdc_ttm.c              |  3 ++-
 drivers/gpu/drm/nouveau/nouveau_ttm.c            |  6 ++++--
 drivers/gpu/drm/qxl/qxl_ttm.c                    |  2 +-
 drivers/gpu/drm/radeon/radeon_ttm.c              |  6 ++++--
 drivers/gpu/drm/ttm/tests/ttm_bo_test.c          | 16 ++++++++--------
 drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c |  2 +-
 drivers/gpu/drm/ttm/tests/ttm_device_test.c      | 12 +++++-------
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c    | 22 +++++++++-------------
 drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h    |  7 ++-----
 drivers/gpu/drm/ttm/ttm_device.c                 |  9 +++------
 drivers/gpu/drm/vmwgfx/vmwgfx_drv.c              |  4 ++--
 drivers/gpu/drm/xe/xe_device.c                   |  2 +-
 include/drm/ttm/ttm_device.h                     |  3 ++-
 16 files changed, 50 insertions(+), 54 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
index 8f6d331e1ea2..7b144ddea268 100644
--- a/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
+++ b/drivers/gpu/drm/amd/amdgpu/amdgpu_ttm.c
@@ -1930,8 +1930,10 @@ int amdgpu_ttm_init(struct amdgpu_device *adev)
 	r = ttm_device_init(&adev->mman.bdev, &amdgpu_bo_driver, adev->dev,
 			       adev_to_drm(adev)->anon_inode->i_mapping,
 			       adev_to_drm(adev)->vma_offset_manager,
-			       adev->need_swiotlb,
-			       dma_addressing_limited(adev->dev));
+			       (adev->need_swiotlb ?
+				TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0) |
+			       (dma_addressing_limited(adev->dev) ?
+				TTM_ALLOCATION_POOL_USE_DMA32 : 0));
 	if (r) {
 		dev_err(adev->dev,
 			"failed initializing buffer object driver(%d).\n", r);
diff --git a/drivers/gpu/drm/drm_gem_vram_helper.c b/drivers/gpu/drm/drm_gem_vram_helper.c
index 9996d42ddc97..5e5b70518dbe 100644
--- a/drivers/gpu/drm/drm_gem_vram_helper.c
+++ b/drivers/gpu/drm/drm_gem_vram_helper.c
@@ -860,7 +860,7 @@ static int drm_vram_mm_init(struct drm_vram_mm *vmm, struct drm_device *dev,
 	ret = ttm_device_init(&vmm->bdev, &bo_driver, dev->dev,
 				 dev->anon_inode->i_mapping,
 				 dev->vma_offset_manager,
-				 false, true);
+				 TTM_ALLOCATION_POOL_USE_DMA32);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/i915/intel_region_ttm.c b/drivers/gpu/drm/i915/intel_region_ttm.c
index 04525d92bec5..47a69aad5c3f 100644
--- a/drivers/gpu/drm/i915/intel_region_ttm.c
+++ b/drivers/gpu/drm/i915/intel_region_ttm.c
@@ -34,7 +34,7 @@ int intel_region_ttm_device_init(struct drm_i915_private *dev_priv)
 
 	return ttm_device_init(&dev_priv->bdev, i915_ttm_driver(),
 			       drm->dev, drm->anon_inode->i_mapping,
-			       drm->vma_offset_manager, false, false);
+			       drm->vma_offset_manager, 0);
 }
 
 /**
diff --git a/drivers/gpu/drm/loongson/lsdc_ttm.c b/drivers/gpu/drm/loongson/lsdc_ttm.c
index 383a758bbd7e..5d9075634bf8 100644
--- a/drivers/gpu/drm/loongson/lsdc_ttm.c
+++ b/drivers/gpu/drm/loongson/lsdc_ttm.c
@@ -545,7 +545,8 @@ int lsdc_ttm_init(struct lsdc_device *ldev)
 
 	ret = ttm_device_init(&ldev->bdev, &lsdc_bo_driver, ddev->dev,
 			      ddev->anon_inode->i_mapping,
-			      ddev->vma_offset_manager, false, true);
+			      ddev->vma_offset_manager,
+			      TTM_ALLOCATION_POOL_USE_DMA32);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/nouveau/nouveau_ttm.c b/drivers/gpu/drm/nouveau/nouveau_ttm.c
index 7d2436e5d50d..0a55babdf667 100644
--- a/drivers/gpu/drm/nouveau/nouveau_ttm.c
+++ b/drivers/gpu/drm/nouveau/nouveau_ttm.c
@@ -302,8 +302,10 @@ nouveau_ttm_init(struct nouveau_drm *drm)
 	ret = ttm_device_init(&drm->ttm.bdev, &nouveau_bo_driver, drm->dev->dev,
 				  dev->anon_inode->i_mapping,
 				  dev->vma_offset_manager,
-				  drm_need_swiotlb(drm->client.mmu.dmabits),
-				  drm->client.mmu.dmabits <= 32);
+				  (drm_need_swiotlb(drm->client.mmu.dmabits) ?
+				   TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0) |
+				  (drm->client.mmu.dmabits <= 32 ?
+				   TTM_ALLOCATION_POOL_USE_DMA32 : 0));
 	if (ret) {
 		NV_ERROR(drm, "error initialising bo driver, %d\n", ret);
 		return ret;
diff --git a/drivers/gpu/drm/qxl/qxl_ttm.c b/drivers/gpu/drm/qxl/qxl_ttm.c
index a6dd2d1f1004..1a40590077dd 100644
--- a/drivers/gpu/drm/qxl/qxl_ttm.c
+++ b/drivers/gpu/drm/qxl/qxl_ttm.c
@@ -197,7 +197,7 @@ int qxl_ttm_init(struct qxl_device *qdev)
 	r = ttm_device_init(&qdev->mman.bdev, &qxl_bo_driver, NULL,
 			    qdev->ddev.anon_inode->i_mapping,
 			    qdev->ddev.vma_offset_manager,
-			    false, false);
+			    0);
 	if (r) {
 		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
 		return r;
diff --git a/drivers/gpu/drm/radeon/radeon_ttm.c b/drivers/gpu/drm/radeon/radeon_ttm.c
index 616d25c8c2de..695ac32f7535 100644
--- a/drivers/gpu/drm/radeon/radeon_ttm.c
+++ b/drivers/gpu/drm/radeon/radeon_ttm.c
@@ -683,8 +683,10 @@ int radeon_ttm_init(struct radeon_device *rdev)
 	r = ttm_device_init(&rdev->mman.bdev, &radeon_bo_driver, rdev->dev,
 			       rdev_to_drm(rdev)->anon_inode->i_mapping,
 			       rdev_to_drm(rdev)->vma_offset_manager,
-			       rdev->need_swiotlb,
-			       dma_addressing_limited(&rdev->pdev->dev));
+			       (rdev->need_swiotlb ?
+				TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0) |
+			       (dma_addressing_limited(&rdev->pdev->dev) ?
+				TTM_ALLOCATION_POOL_USE_DMA32 : 0));
 	if (r) {
 		DRM_ERROR("failed initializing buffer object driver(%d).\n", r);
 		return r;
diff --git a/drivers/gpu/drm/ttm/tests/ttm_bo_test.c b/drivers/gpu/drm/ttm/tests/ttm_bo_test.c
index 5426b435f702..d468f8322072 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_bo_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_bo_test.c
@@ -251,7 +251,7 @@ static void ttm_bo_unreserve_basic(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -290,7 +290,7 @@ static void ttm_bo_unreserve_pinned(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -342,7 +342,7 @@ static void ttm_bo_unreserve_bulk(struct kunit *test)
 	resv = kunit_kzalloc(test, sizeof(*resv), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, resv);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -394,7 +394,7 @@ static void ttm_bo_fini_basic(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -437,7 +437,7 @@ static void ttm_bo_fini_shared_resv(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -477,7 +477,7 @@ static void ttm_bo_pin_basic(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -512,7 +512,7 @@ static void ttm_bo_pin_unpin_resource(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
@@ -563,7 +563,7 @@ static void ttm_bo_multiple_pin_one_unpin(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 	priv->ttm_dev = ttm_dev;
 
diff --git a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
index 3a1eef83190c..17a570af296c 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
@@ -995,7 +995,7 @@ static void ttm_bo_validate_busy_domain_evict(struct kunit *test)
 	 */
 	ttm_device_fini(priv->ttm_dev);
 
-	err = ttm_device_kunit_init_bad_evict(test->priv, priv->ttm_dev, false, false);
+	err = ttm_device_kunit_init_bad_evict(test->priv, priv->ttm_dev);
 	KUNIT_ASSERT_EQ(test, err, 0);
 
 	ttm_mock_manager_init(priv->ttm_dev, mem_type, MANAGER_SIZE);
diff --git a/drivers/gpu/drm/ttm/tests/ttm_device_test.c b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
index 98648d5f20e7..2d55ad34fe48 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_device_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_device_test.c
@@ -25,7 +25,7 @@ static void ttm_device_init_basic(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 
 	KUNIT_EXPECT_PTR_EQ(test, ttm_dev->funcs, &ttm_dev_funcs);
@@ -55,7 +55,7 @@ static void ttm_device_init_multiple(struct kunit *test)
 	KUNIT_ASSERT_NOT_NULL(test, ttm_devs);
 
 	for (i = 0; i < num_dev; i++) {
-		err = ttm_device_kunit_init(priv, &ttm_devs[i], false, false);
+		err = ttm_device_kunit_init(priv, &ttm_devs[i], 0);
 		KUNIT_ASSERT_EQ(test, err, 0);
 
 		KUNIT_EXPECT_PTR_EQ(test, ttm_devs[i].dev_mapping,
@@ -81,7 +81,7 @@ static void ttm_device_fini_basic(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 
 	man = ttm_manager_type(ttm_dev, TTM_PL_SYSTEM);
@@ -109,7 +109,7 @@ static void ttm_device_init_no_vma_man(struct kunit *test)
 	vma_man = drm->vma_offset_manager;
 	drm->vma_offset_manager = NULL;
 
-	err = ttm_device_kunit_init(priv, ttm_dev, false, false);
+	err = ttm_device_kunit_init(priv, ttm_dev, 0);
 	KUNIT_EXPECT_EQ(test, err, -EINVAL);
 
 	/* Bring the manager back for a graceful cleanup */
@@ -158,9 +158,7 @@ static void ttm_device_init_pools(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(priv, ttm_dev,
-				    params->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA_ALLOC,
-				    params->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA32);
+	err = ttm_device_kunit_init(priv, ttm_dev, params->alloc_flags);
 	KUNIT_ASSERT_EQ(test, err, 0);
 
 	pool = &ttm_dev->pool;
diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
index 7aaf0d1395ff..7b533e4e1e04 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.c
@@ -117,8 +117,7 @@ static void bad_evict_flags(struct ttm_buffer_object *bo,
 
 static int ttm_device_kunit_init_with_funcs(struct ttm_test_devices *priv,
 					    struct ttm_device *ttm,
-					    bool use_dma_alloc,
-					    bool use_dma32,
+					    unsigned int alloc_flags,
 					    struct ttm_device_funcs *funcs)
 {
 	struct drm_device *drm = priv->drm;
@@ -127,7 +126,7 @@ static int ttm_device_kunit_init_with_funcs(struct ttm_test_devices *priv,
 	err = ttm_device_init(ttm, funcs, drm->dev,
 			      drm->anon_inode->i_mapping,
 			      drm->vma_offset_manager,
-			      use_dma_alloc, use_dma32);
+			      alloc_flags);
 
 	return err;
 }
@@ -143,11 +142,10 @@ EXPORT_SYMBOL_GPL(ttm_dev_funcs);
 
 int ttm_device_kunit_init(struct ttm_test_devices *priv,
 			  struct ttm_device *ttm,
-			  bool use_dma_alloc,
-			  bool use_dma32)
+			  unsigned int alloc_flags)
 {
-	return ttm_device_kunit_init_with_funcs(priv, ttm, use_dma_alloc,
-						use_dma32, &ttm_dev_funcs);
+	return ttm_device_kunit_init_with_funcs(priv, ttm, alloc_flags,
+						&ttm_dev_funcs);
 }
 EXPORT_SYMBOL_GPL(ttm_device_kunit_init);
 
@@ -161,12 +159,10 @@ struct ttm_device_funcs ttm_dev_funcs_bad_evict = {
 EXPORT_SYMBOL_GPL(ttm_dev_funcs_bad_evict);
 
 int ttm_device_kunit_init_bad_evict(struct ttm_test_devices *priv,
-				    struct ttm_device *ttm,
-				    bool use_dma_alloc,
-				    bool use_dma32)
+				    struct ttm_device *ttm)
 {
-	return ttm_device_kunit_init_with_funcs(priv, ttm, use_dma_alloc,
-						use_dma32, &ttm_dev_funcs_bad_evict);
+	return ttm_device_kunit_init_with_funcs(priv, ttm, 0,
+						&ttm_dev_funcs_bad_evict);
 }
 EXPORT_SYMBOL_GPL(ttm_device_kunit_init_bad_evict);
 
@@ -252,7 +248,7 @@ struct ttm_test_devices *ttm_test_devices_all(struct kunit *test)
 	ttm_dev = kunit_kzalloc(test, sizeof(*ttm_dev), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, ttm_dev);
 
-	err = ttm_device_kunit_init(devs, ttm_dev, false, false);
+	err = ttm_device_kunit_init(devs, ttm_dev, 0);
 	KUNIT_ASSERT_EQ(test, err, 0);
 
 	devs->ttm_dev = ttm_dev;
diff --git a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
index c7da23232ffa..f8402b979d05 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
+++ b/drivers/gpu/drm/ttm/tests/ttm_kunit_helpers.h
@@ -28,12 +28,9 @@ struct ttm_test_devices {
 /* Building blocks for test-specific init functions */
 int ttm_device_kunit_init(struct ttm_test_devices *priv,
 			  struct ttm_device *ttm,
-			  bool use_dma_alloc,
-			  bool use_dma32);
+			  unsigned int alloc_flags);
 int ttm_device_kunit_init_bad_evict(struct ttm_test_devices *priv,
-				    struct ttm_device *ttm,
-				    bool use_dma_alloc,
-				    bool use_dma32);
+				    struct ttm_device *ttm);
 struct ttm_buffer_object *ttm_bo_kunit_init(struct kunit *test,
 					    struct ttm_test_devices *devs,
 					    size_t size,
diff --git a/drivers/gpu/drm/ttm/ttm_device.c b/drivers/gpu/drm/ttm/ttm_device.c
index a97b1444536c..87c85ccb21ac 100644
--- a/drivers/gpu/drm/ttm/ttm_device.c
+++ b/drivers/gpu/drm/ttm/ttm_device.c
@@ -199,8 +199,7 @@ EXPORT_SYMBOL(ttm_device_swapout);
  * @dev: The core kernel device pointer for DMA mappings and allocations.
  * @mapping: The address space to use for this bo.
  * @vma_manager: A pointer to a vma manager.
- * @use_dma_alloc: If coherent DMA allocation API should be used.
- * @use_dma32: If we should use GFP_DMA32 for device memory allocations.
+ * @alloc_flags: TTM_ALLOCATION_ flags.
  *
  * Initializes a struct ttm_device:
  * Returns:
@@ -209,7 +208,7 @@ EXPORT_SYMBOL(ttm_device_swapout);
 int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *funcs,
 		    struct device *dev, struct address_space *mapping,
 		    struct drm_vma_offset_manager *vma_manager,
-		    bool use_dma_alloc, bool use_dma32)
+		    unsigned int alloc_flags)
 {
 	struct ttm_global *glob = &ttm_glob;
 	int ret, nid;
@@ -237,9 +236,7 @@ int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *func
 	else
 		nid = NUMA_NO_NODE;
 
-	ttm_pool_init(&bdev->pool, dev, nid,
-		      (use_dma_alloc ? TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0) |
-		      (use_dma32 ? TTM_ALLOCATION_POOL_USE_DMA32 : 0));
+	ttm_pool_init(&bdev->pool, dev, nid, alloc_flags);
 
 	bdev->vma_manager = vma_manager;
 	spin_lock_init(&bdev->lru_lock);
diff --git a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
index 8ff958d119be..599052d07ae8 100644
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_drv.c
@@ -1023,8 +1023,8 @@ static int vmw_driver_load(struct vmw_private *dev_priv, u32 pci_id)
 			      dev_priv->drm.dev,
 			      dev_priv->drm.anon_inode->i_mapping,
 			      dev_priv->drm.vma_offset_manager,
-			      dev_priv->map_mode == vmw_dma_alloc_coherent,
-			      false);
+			      (dev_priv->map_mode == vmw_dma_alloc_coherent) ?
+			      TTM_ALLOCATION_POOL_USE_DMA_ALLOC : 0);
 	if (unlikely(ret != 0)) {
 		drm_err(&dev_priv->drm,
 			"Failed initializing TTM buffer object driver.\n");
diff --git a/drivers/gpu/drm/xe/xe_device.c b/drivers/gpu/drm/xe/xe_device.c
index 5f6a412b571c..58e7996160a0 100644
--- a/drivers/gpu/drm/xe/xe_device.c
+++ b/drivers/gpu/drm/xe/xe_device.c
@@ -437,7 +437,7 @@ struct xe_device *xe_device_create(struct pci_dev *pdev,
 
 	err = ttm_device_init(&xe->ttm, &xe_ttm_funcs, xe->drm.dev,
 			      xe->drm.anon_inode->i_mapping,
-			      xe->drm.vma_offset_manager, false, false);
+			      xe->drm.vma_offset_manager, 0);
 	if (WARN_ON(err))
 		goto err;
 
diff --git a/include/drm/ttm/ttm_device.h b/include/drm/ttm/ttm_device.h
index 592b5f802859..074b98572275 100644
--- a/include/drm/ttm/ttm_device.h
+++ b/include/drm/ttm/ttm_device.h
@@ -27,6 +27,7 @@
 
 #include <linux/types.h>
 #include <linux/workqueue.h>
+#include <drm/ttm/ttm_allocation.h>
 #include <drm/ttm/ttm_resource.h>
 #include <drm/ttm/ttm_pool.h>
 
@@ -292,7 +293,7 @@ static inline void ttm_set_driver_manager(struct ttm_device *bdev, int type,
 int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *funcs,
 		    struct device *dev, struct address_space *mapping,
 		    struct drm_vma_offset_manager *vma_manager,
-		    bool use_dma_alloc, bool use_dma32);
+		    unsigned int alloc_flags);
 void ttm_device_fini(struct ttm_device *bdev);
 void ttm_device_clear_dma_mappings(struct ttm_device *bdev);
 
-- 
cgit v1.2.3


From 7e9c548d3709c76601c953834bed9c888f3e17b2 Mon Sep 17 00:00:00 2001
From: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Date: Mon, 20 Oct 2025 12:54:09 +0100
Subject: drm/ttm: Allow drivers to specify maximum beneficial TTM pool size
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

GPUs typically benefit from contiguous memory via reduced TLB pressure and
improved caching performance, where the maximum size of contiguous block
which adds a performance benefit is related to hardware design.

TTM pool allocator by default tries (hard) to allocate up to the system
MAX_PAGE_ORDER blocks. This varies by the CPU platform and can also be
configured via Kconfig.

If that limit was set to be higher than the GPU can make an extra use of,
lets allow the individual drivers to let TTM know over which allocation
order can the pool allocator afford to make a little bit less effort with.

We implement this by disabling direct reclaim for those allocations, which
reduces the allocation latency and lowers the demands on the page
allocator, in cases where expending this effort is not critical for the
GPU in question.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Thadeu Lima de Souza Cascardo <cascardo@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-5-tvrtko.ursulin@igalia.com
---
 drivers/gpu/drm/ttm/ttm_pool.c          | 8 ++++++++
 drivers/gpu/drm/ttm/ttm_pool_internal.h | 5 +++++
 include/drm/ttm/ttm_allocation.h        | 5 +++--
 3 files changed, 16 insertions(+), 2 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c
index 4fc69447060c..97e9ce505cf6 100644
--- a/drivers/gpu/drm/ttm/ttm_pool.c
+++ b/drivers/gpu/drm/ttm/ttm_pool.c
@@ -136,6 +136,7 @@ static DECLARE_RWSEM(pool_shrink_rwsem);
 static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags,
 					unsigned int order)
 {
+	const unsigned int beneficial_order = ttm_pool_beneficial_order(pool);
 	unsigned long attr = DMA_ATTR_FORCE_CONTIGUOUS;
 	struct ttm_pool_dma *dma;
 	struct page *p;
@@ -149,6 +150,13 @@ static struct page *ttm_pool_alloc_page(struct ttm_pool *pool, gfp_t gfp_flags,
 		gfp_flags |= __GFP_NOMEMALLOC | __GFP_NORETRY | __GFP_NOWARN |
 			__GFP_THISNODE;
 
+	/*
+	 * Do not add latency to the allocation path for allocations orders
+	 * device tolds us do not bring them additional performance gains.
+	 */
+	if (beneficial_order && order > beneficial_order)
+		gfp_flags &= ~__GFP_DIRECT_RECLAIM;
+
 	if (!ttm_pool_uses_dma_alloc(pool)) {
 		p = alloc_pages_node(pool->nid, gfp_flags, order);
 		if (p)
diff --git a/drivers/gpu/drm/ttm/ttm_pool_internal.h b/drivers/gpu/drm/ttm/ttm_pool_internal.h
index 96b7f21514fb..82c4b7e56a99 100644
--- a/drivers/gpu/drm/ttm/ttm_pool_internal.h
+++ b/drivers/gpu/drm/ttm/ttm_pool_internal.h
@@ -17,4 +17,9 @@ static inline bool ttm_pool_uses_dma32(struct ttm_pool *pool)
 	return pool->alloc_flags & TTM_ALLOCATION_POOL_USE_DMA32;
 }
 
+static inline bool ttm_pool_beneficial_order(struct ttm_pool *pool)
+{
+	return pool->alloc_flags & 0xff;
+}
+
 #endif
diff --git a/include/drm/ttm/ttm_allocation.h b/include/drm/ttm/ttm_allocation.h
index 7869dc32bd91..8f8544760306 100644
--- a/include/drm/ttm/ttm_allocation.h
+++ b/include/drm/ttm/ttm_allocation.h
@@ -4,7 +4,8 @@
 #ifndef _TTM_ALLOCATION_H_
 #define _TTM_ALLOCATION_H_
 
-#define TTM_ALLOCATION_POOL_USE_DMA_ALLOC	BIT(0) /* Use coherent DMA allocations. */
-#define TTM_ALLOCATION_POOL_USE_DMA32		BIT(1) /* Use GFP_DMA32 allocations. */
+#define TTM_ALLOCATION_POOL_BENEFICIAL_ORDER(n)	((n) & 0xff) /* Max order which caller can benefit from */
+#define TTM_ALLOCATION_POOL_USE_DMA_ALLOC 	BIT(8) /* Use coherent DMA allocations. */
+#define TTM_ALLOCATION_POOL_USE_DMA32		BIT(9) /* Use GFP_DMA32 allocations. */
 
 #endif
-- 
cgit v1.2.3


From 402b3a865090578f9245115e17ee230e01daf60e Mon Sep 17 00:00:00 2001
From: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Date: Mon, 20 Oct 2025 12:54:11 +0100
Subject: drm/ttm: Add an allocation flag to propagate -ENOSPC on OOM
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Some graphics APIs differentiate between out-of-graphics-memory and
out-of-host-memory (system memory). Add a device init flag to have -ENOSPC
propagated from the resource managers instead of being converted to
-ENOMEM, to aid driver stacks in determining what error code to return or
whether corrective action can be taken at the driver level.

Co-developed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251020115411.36818-7-tvrtko.ursulin@igalia.com
---
 drivers/gpu/drm/ttm/ttm_bo.c     | 4 +++-
 drivers/gpu/drm/ttm/ttm_device.c | 1 +
 include/drm/ttm/ttm_allocation.h | 1 +
 include/drm/ttm/ttm_device.h     | 5 +++++
 4 files changed, 10 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index 44559dbf62f9..c4e669686fd6 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -32,6 +32,7 @@
 #define pr_fmt(fmt) "[TTM] " fmt
 
 #include <drm/drm_print.h>
+#include <drm/ttm/ttm_allocation.h>
 #include <drm/ttm/ttm_bo.h>
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_tt.h>
@@ -878,7 +879,8 @@ bounce:
 
 	/* For backward compatibility with userspace */
 	if (ret == -ENOSPC)
-		return -ENOMEM;
+		return bo->bdev->alloc_flags & TTM_ALLOCATION_PROPAGATE_ENOSPC ?
+		       ret : -ENOMEM;
 
 	/*
 	 * We might need to add a TTM.
diff --git a/drivers/gpu/drm/ttm/ttm_device.c b/drivers/gpu/drm/ttm/ttm_device.c
index 87c85ccb21ac..5c10e5fbf43b 100644
--- a/drivers/gpu/drm/ttm/ttm_device.c
+++ b/drivers/gpu/drm/ttm/ttm_device.c
@@ -227,6 +227,7 @@ int ttm_device_init(struct ttm_device *bdev, const struct ttm_device_funcs *func
 		return -ENOMEM;
 	}
 
+	bdev->alloc_flags = alloc_flags;
 	bdev->funcs = funcs;
 
 	ttm_sys_man_init(bdev);
diff --git a/include/drm/ttm/ttm_allocation.h b/include/drm/ttm/ttm_allocation.h
index 8f8544760306..655d1e44aba7 100644
--- a/include/drm/ttm/ttm_allocation.h
+++ b/include/drm/ttm/ttm_allocation.h
@@ -7,5 +7,6 @@
 #define TTM_ALLOCATION_POOL_BENEFICIAL_ORDER(n)	((n) & 0xff) /* Max order which caller can benefit from */
 #define TTM_ALLOCATION_POOL_USE_DMA_ALLOC 	BIT(8) /* Use coherent DMA allocations. */
 #define TTM_ALLOCATION_POOL_USE_DMA32		BIT(9) /* Use GFP_DMA32 allocations. */
+#define TTM_ALLOCATION_PROPAGATE_ENOSPC		BIT(10) /* Do not convert ENOSPC from resource managers to ENOMEM. */
 
 #endif
diff --git a/include/drm/ttm/ttm_device.h b/include/drm/ttm/ttm_device.h
index 074b98572275..d016360e5ceb 100644
--- a/include/drm/ttm/ttm_device.h
+++ b/include/drm/ttm/ttm_device.h
@@ -220,6 +220,11 @@ struct ttm_device {
 	 */
 	struct list_head device_list;
 
+	/**
+	 * @alloc_flags: TTM_ALLOCATION_ flags.
+	 */
+	unsigned int alloc_flags;
+
 	/**
 	 * @funcs: Function table for the device.
 	 * Constant after bo device init
-- 
cgit v1.2.3


From b3c8fa0d9c2650ab9bdc7ccc980a6826c4f9021d Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Thu, 30 Oct 2025 22:28:31 +0200
Subject: drm/{i915, xe}/display: pass parent interface to display probe
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Let's gradually start calling i915 and xe parent, or core, drivers from
display via function pointers passed at display probe.

Going forward, the struct intel_display_parent_interface is expected to
include const pointers to sub-structs by functionality, for example:

struct intel_display_rpm {
	struct ref_tracker *(*get)(struct drm_device *drm);
	/* ... */
};

struct intel_display_parent_interface {
	/* ... */
	const struct intel_display_rpm *rpm;
};

This is a baby step towards not building display as part of both i915
and xe drivers, but rather making it an independent driver interfacing
with the two.

v3: useless include additions dropped
v2: unrelated include removal dropped

Cc: Jouni Högander <jouni.hogander@intel.com>
Cc: Lucas De Marchi <lucas.demarchi@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patch.msgid.link/20251030202836.1815680-2-jouni.hogander@intel.com
---
 drivers/gpu/drm/i915/display/intel_display_core.h  |  4 ++++
 .../gpu/drm/i915/display/intel_display_device.c    |  5 ++++-
 .../gpu/drm/i915/display/intel_display_device.h    |  4 +++-
 drivers/gpu/drm/i915/i915_driver.c                 | 11 ++++++++-
 drivers/gpu/drm/i915/i915_driver.h                 |  2 ++
 drivers/gpu/drm/i915/selftests/mock_gem_device.c   |  4 +++-
 drivers/gpu/drm/xe/display/xe_display.c            |  6 ++++-
 include/drm/intel/display_parent_interface.h       | 26 ++++++++++++++++++++++
 8 files changed, 57 insertions(+), 5 deletions(-)
 create mode 100644 include/drm/intel/display_parent_interface.h

(limited to 'include')

diff --git a/drivers/gpu/drm/i915/display/intel_display_core.h b/drivers/gpu/drm/i915/display/intel_display_core.h
index 32664098b407..893279be8409 100644
--- a/drivers/gpu/drm/i915/display/intel_display_core.h
+++ b/drivers/gpu/drm/i915/display/intel_display_core.h
@@ -41,6 +41,7 @@ struct intel_cdclk_vals;
 struct intel_color_funcs;
 struct intel_crtc;
 struct intel_crtc_state;
+struct intel_display_parent_interface;
 struct intel_dmc;
 struct intel_dpll_global_funcs;
 struct intel_dpll_mgr;
@@ -291,6 +292,9 @@ struct intel_display {
 	/* Intel PCH: where the south display engine lives */
 	enum intel_pch pch_type;
 
+	/* Parent, or core, driver functions exposed to display */
+	const struct intel_display_parent_interface *parent;
+
 	/* Display functions */
 	struct {
 		/* Top level crtc-ish functions */
diff --git a/drivers/gpu/drm/i915/display/intel_display_device.c b/drivers/gpu/drm/i915/display/intel_display_device.c
index f3f1f25b0f38..328447a5e5e8 100644
--- a/drivers/gpu/drm/i915/display/intel_display_device.c
+++ b/drivers/gpu/drm/i915/display/intel_display_device.c
@@ -1647,7 +1647,8 @@ static void display_platforms_or(struct intel_display_platforms *dst,
 	bitmap_or(dst->bitmap, dst->bitmap, src->bitmap, display_platforms_num_bits());
 }
 
-struct intel_display *intel_display_device_probe(struct pci_dev *pdev)
+struct intel_display *intel_display_device_probe(struct pci_dev *pdev,
+						 const struct intel_display_parent_interface *parent)
 {
 	struct intel_display *display;
 	const struct intel_display_device_info *info;
@@ -1663,6 +1664,8 @@ struct intel_display *intel_display_device_probe(struct pci_dev *pdev)
 	/* Add drm device backpointer as early as possible. */
 	display->drm = pci_get_drvdata(pdev);
 
+	display->parent = parent;
+
 	intel_display_params_copy(&display->params);
 
 	if (has_no_display(pdev)) {
diff --git a/drivers/gpu/drm/i915/display/intel_display_device.h b/drivers/gpu/drm/i915/display/intel_display_device.h
index ece66c8c4ce6..b559ef43d547 100644
--- a/drivers/gpu/drm/i915/display/intel_display_device.h
+++ b/drivers/gpu/drm/i915/display/intel_display_device.h
@@ -13,6 +13,7 @@
 
 struct drm_printer;
 struct intel_display;
+struct intel_display_parent_interface;
 struct pci_dev;
 
 /*
@@ -313,7 +314,8 @@ struct intel_display_device_info {
 
 bool intel_display_device_present(struct intel_display *display);
 bool intel_display_device_enabled(struct intel_display *display);
-struct intel_display *intel_display_device_probe(struct pci_dev *pdev);
+struct intel_display *intel_display_device_probe(struct pci_dev *pdev,
+						 const struct intel_display_parent_interface *parent);
 void intel_display_device_remove(struct intel_display *display);
 void intel_display_device_info_runtime_init(struct intel_display *display);
 
diff --git a/drivers/gpu/drm/i915/i915_driver.c b/drivers/gpu/drm/i915/i915_driver.c
index b46cb54ef5dc..18ff10c63e98 100644
--- a/drivers/gpu/drm/i915/i915_driver.c
+++ b/drivers/gpu/drm/i915/i915_driver.c
@@ -47,6 +47,7 @@
 #include <drm/drm_managed.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/intel/display_member.h>
+#include <drm/intel/display_parent_interface.h>
 
 #include "display/i9xx_display_sr.h"
 #include "display/intel_bw.h"
@@ -738,6 +739,14 @@ static void i915_welcome_messages(struct drm_i915_private *dev_priv)
 			 "DRM_I915_DEBUG_RUNTIME_PM enabled\n");
 }
 
+static const struct intel_display_parent_interface parent = {
+};
+
+const struct intel_display_parent_interface *i915_driver_parent_interface(void)
+{
+	return &parent;
+}
+
 /* Ensure drm and display members are placed properly. */
 INTEL_DISPLAY_MEMBER_STATIC_ASSERT(struct drm_i915_private, drm, display);
 
@@ -762,7 +771,7 @@ i915_driver_create(struct pci_dev *pdev, const struct pci_device_id *ent)
 	/* Set up device info and initial runtime info. */
 	intel_device_info_driver_create(i915, pdev->device, match_info);
 
-	display = intel_display_device_probe(pdev);
+	display = intel_display_device_probe(pdev, &parent);
 	if (IS_ERR(display))
 		return ERR_CAST(display);
 
diff --git a/drivers/gpu/drm/i915/i915_driver.h b/drivers/gpu/drm/i915/i915_driver.h
index 1e95ecb2a163..9551519ab429 100644
--- a/drivers/gpu/drm/i915/i915_driver.h
+++ b/drivers/gpu/drm/i915/i915_driver.h
@@ -12,6 +12,7 @@ struct pci_dev;
 struct pci_device_id;
 struct drm_i915_private;
 struct drm_printer;
+struct intel_display_parent_interface;
 
 #define DRIVER_NAME		"i915"
 #define DRIVER_DESC		"Intel Graphics"
@@ -24,6 +25,7 @@ void i915_driver_shutdown(struct drm_i915_private *i915);
 
 int i915_driver_resume_switcheroo(struct drm_i915_private *i915);
 int i915_driver_suspend_switcheroo(struct drm_i915_private *i915, pm_message_t state);
+const struct intel_display_parent_interface *i915_driver_parent_interface(void);
 
 void
 i915_print_iommu_status(struct drm_i915_private *i915, struct drm_printer *p);
diff --git a/drivers/gpu/drm/i915/selftests/mock_gem_device.c b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
index fb8751bd5df0..b59626c4994c 100644
--- a/drivers/gpu/drm/i915/selftests/mock_gem_device.c
+++ b/drivers/gpu/drm/i915/selftests/mock_gem_device.c
@@ -33,6 +33,7 @@
 #include "gt/intel_gt.h"
 #include "gt/intel_gt_requests.h"
 #include "gt/mock_engine.h"
+#include "i915_driver.h"
 #include "intel_memory_region.h"
 #include "intel_region_ttm.h"
 
@@ -183,7 +184,8 @@ struct drm_i915_private *mock_gem_device(void)
 	/* Set up device info and initial runtime info. */
 	intel_device_info_driver_create(i915, pdev->device, &mock_info);
 
-	display = intel_display_device_probe(pdev);
+	/* FIXME: Can we run selftests using a mock device without display? */
+	display = intel_display_device_probe(pdev, i915_driver_parent_interface());
 	if (IS_ERR(display))
 		goto err_device;
 
diff --git a/drivers/gpu/drm/xe/display/xe_display.c b/drivers/gpu/drm/xe/display/xe_display.c
index 5f4044e63185..074341645960 100644
--- a/drivers/gpu/drm/xe/display/xe_display.c
+++ b/drivers/gpu/drm/xe/display/xe_display.c
@@ -14,6 +14,7 @@
 #include <drm/drm_managed.h>
 #include <drm/drm_probe_helper.h>
 #include <drm/intel/display_member.h>
+#include <drm/intel/display_parent_interface.h>
 #include <uapi/drm/xe_drm.h>
 
 #include "soc/intel_dram.h"
@@ -515,6 +516,9 @@ static void display_device_remove(struct drm_device *dev, void *arg)
 	intel_display_device_remove(display);
 }
 
+static const struct intel_display_parent_interface parent = {
+};
+
 /**
  * xe_display_probe - probe display and create display struct
  * @xe: XE device instance
@@ -535,7 +539,7 @@ int xe_display_probe(struct xe_device *xe)
 	if (!xe->info.probe_display)
 		goto no_display;
 
-	display = intel_display_device_probe(pdev);
+	display = intel_display_device_probe(pdev, &parent);
 	if (IS_ERR(display))
 		return PTR_ERR(display);
 
diff --git a/include/drm/intel/display_parent_interface.h b/include/drm/intel/display_parent_interface.h
new file mode 100644
index 000000000000..28c976815327
--- /dev/null
+++ b/include/drm/intel/display_parent_interface.h
@@ -0,0 +1,26 @@
+/* SPDX-License-Identifier: MIT */
+/* Copyright © 2025 Intel Corporation x*/
+
+#ifndef __DISPLAY_PARENT_INTERFACE_H__
+#define __DISPLAY_PARENT_INTERFACE_H__
+
+#include <linux/types.h>
+
+struct drm_device;
+
+/**
+ * struct intel_display_parent_interface - services parent driver provides to display
+ *
+ * The parent, or core, driver provides a pointer to this structure to display
+ * driver when calling intel_display_device_probe(). The display driver uses it
+ * to access services provided by the parent driver. The structure may contain
+ * sub-struct pointers to group function pointers by functionality.
+ *
+ * All function and sub-struct pointers must be initialized and callable unless
+ * explicitly marked as "optional" below. The display driver will only NULL
+ * check the optional pointers.
+ */
+struct intel_display_parent_interface {
+};
+
+#endif
-- 
cgit v1.2.3


From 1914d6861b594dde7cbb359ab1ba43f2e3a82f91 Mon Sep 17 00:00:00 2001
From: Jouni Högander <jouni.hogander@intel.com>
Date: Thu, 30 Oct 2025 22:28:32 +0200
Subject: drm/{i915, xe}/display: Add display runtime pm parent interface
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

We have differing implementations for display runtime pm in i915 and xe
drivers. Add struct of function pointers into display_parent_interface
which will contain used implementation of runtime pm.

v2:
  - add _interface suffix to rpm function pointer struct
  - add struct ref_tracker forward declaration
  - use kernel-doc comments

Reviewed-by: Jani Nikula <jani.nikula@intel.com>
Signed-off-by: Jouni Högander <jouni.hogander@intel.com>
Link: https://patch.msgid.link/20251030202836.1815680-3-jouni.hogander@intel.com
---
 include/drm/intel/display_parent_interface.h | 19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

(limited to 'include')

diff --git a/include/drm/intel/display_parent_interface.h b/include/drm/intel/display_parent_interface.h
index 28c976815327..26bedc360044 100644
--- a/include/drm/intel/display_parent_interface.h
+++ b/include/drm/intel/display_parent_interface.h
@@ -7,6 +7,23 @@
 #include <linux/types.h>
 
 struct drm_device;
+struct ref_tracker;
+
+struct intel_display_rpm_interface {
+	struct ref_tracker *(*get)(const struct drm_device *drm);
+	struct ref_tracker *(*get_raw)(const struct drm_device *drm);
+	struct ref_tracker *(*get_if_in_use)(const struct drm_device *drm);
+	struct ref_tracker *(*get_noresume)(const struct drm_device *drm);
+
+	void (*put)(const struct drm_device *drm, struct ref_tracker *wakeref);
+	void (*put_raw)(const struct drm_device *drm, struct ref_tracker *wakeref);
+	void (*put_unchecked)(const struct drm_device *drm);
+
+	bool (*suspended)(const struct drm_device *drm);
+	void (*assert_held)(const struct drm_device *drm);
+	void (*assert_block)(const struct drm_device *drm);
+	void (*assert_unblock)(const struct drm_device *drm);
+};
 
 /**
  * struct intel_display_parent_interface - services parent driver provides to display
@@ -21,6 +38,8 @@ struct drm_device;
  * check the optional pointers.
  */
 struct intel_display_parent_interface {
+	/** @rpm: Runtime PM functions */
+	const struct intel_display_rpm_interface *rpm;
 };
 
 #endif
-- 
cgit v1.2.3


From 8b61583f993589a64c061aa91b44f5bd350d90a5 Mon Sep 17 00:00:00 2001
From: Jani Nikula <jani.nikula@intel.com>
Date: Tue, 28 Oct 2025 22:07:26 +0200
Subject: drm/edid: add DRM_EDID_IDENT_INIT() to initialize struct
 drm_edid_ident
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Add a convenience helper for initializing struct drm_edid_ident.

Cc: Tiago Martins Araújo <tiago.martins.araujo@gmail.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Tested-by: Tiago Martins Araújo <tiago.martins.araujo@gmail.com>
Cc: stable@vger.kernel.org
Link: https://patch.msgid.link/710b2ac6a211606ec1f90afa57b79e8c7375a27e.1761681968.git.jani.nikula@intel.com
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
---
 include/drm/drm_edid.h | 6 ++++++
 1 file changed, 6 insertions(+)

(limited to 'include')

diff --git a/include/drm/drm_edid.h b/include/drm/drm_edid.h
index 3d1aecfec9b2..04f7a7f1f108 100644
--- a/include/drm/drm_edid.h
+++ b/include/drm/drm_edid.h
@@ -340,6 +340,12 @@ struct drm_edid_ident {
 	const char *name;
 };
 
+#define DRM_EDID_IDENT_INIT(_vend_chr_0, _vend_chr_1, _vend_chr_2, _product_id, _name) \
+{ \
+	.panel_id = drm_edid_encode_panel_id(_vend_chr_0, _vend_chr_1, _vend_chr_2, _product_id), \
+	.name = _name, \
+}
+
 #define EDID_PRODUCT_ID(e) ((e)->prod_code[0] | ((e)->prod_code[1] << 8))
 
 /* Short Audio Descriptor */
-- 
cgit v1.2.3


From 1556c170d2f78344a9eee567fbfcee4651689813 Mon Sep 17 00:00:00 2001
From: Lizhi Hou <lizhi.hou@amd.com>
Date: Mon, 3 Nov 2025 22:25:44 -0800
Subject: accel/amdxdna: Add IOCTL parameter for resource data

Extend DRM_IOCTL_AMDXDNA_GET_INFO to include additional parameters
that allow collection of resource data.

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Link: https://patch.msgid.link/20251104062546.833771-2-lizhi.hou@amd.com
---
 drivers/accel/amdxdna/aie2_ctx.c        |  6 ------
 drivers/accel/amdxdna/aie2_message.c    |  2 ++
 drivers/accel/amdxdna/aie2_pci.c        | 27 +++++++++++++++++++++++++++
 drivers/accel/amdxdna/amdxdna_pci_drv.c |  3 ++-
 include/uapi/drm/amdxdna_accel.h        | 17 +++++++++++++++++
 5 files changed, 48 insertions(+), 7 deletions(-)

(limited to 'include')

diff --git a/drivers/accel/amdxdna/aie2_ctx.c b/drivers/accel/amdxdna/aie2_ctx.c
index 289a2aaf4cae..b78c47ed0d34 100644
--- a/drivers/accel/amdxdna/aie2_ctx.c
+++ b/drivers/accel/amdxdna/aie2_ctx.c
@@ -556,7 +556,6 @@ int aie2_hwctx_init(struct amdxdna_hwctx *hwctx)
 	struct drm_gpu_scheduler *sched;
 	struct amdxdna_hwctx_priv *priv;
 	struct amdxdna_gem_obj *heap;
-	struct amdxdna_dev_hdl *ndev;
 	int i, ret;
 
 	priv = kzalloc(sizeof(*hwctx->priv), GFP_KERNEL);
@@ -654,8 +653,6 @@ int aie2_hwctx_init(struct amdxdna_hwctx *hwctx)
 	amdxdna_pm_suspend_put(xdna);
 
 	hwctx->status = HWCTX_STAT_INIT;
-	ndev = xdna->dev_handle;
-	ndev->hwctx_num++;
 	init_waitqueue_head(&priv->job_free_wq);
 
 	XDNA_DBG(xdna, "hwctx %s init completed", hwctx->name);
@@ -688,13 +685,10 @@ free_priv:
 
 void aie2_hwctx_fini(struct amdxdna_hwctx *hwctx)
 {
-	struct amdxdna_dev_hdl *ndev;
 	struct amdxdna_dev *xdna;
 	int idx;
 
 	xdna = hwctx->client->xdna;
-	ndev = xdna->dev_handle;
-	ndev->hwctx_num--;
 
 	XDNA_DBG(xdna, "%s sequence number %lld", hwctx->name, hwctx->priv->seq);
 	drm_sched_entity_destroy(&hwctx->priv->entity);
diff --git a/drivers/accel/amdxdna/aie2_message.c b/drivers/accel/amdxdna/aie2_message.c
index 339dec998247..39214253d804 100644
--- a/drivers/accel/amdxdna/aie2_message.c
+++ b/drivers/accel/amdxdna/aie2_message.c
@@ -235,6 +235,7 @@ int aie2_create_context(struct amdxdna_dev_hdl *ndev, struct amdxdna_hwctx *hwct
 		ret = -EINVAL;
 		goto out_destroy_context;
 	}
+	ndev->hwctx_num++;
 
 	XDNA_DBG(xdna, "%s mailbox channel irq: %d, msix_id: %d",
 		 hwctx->name, ret, resp.msix_id);
@@ -269,6 +270,7 @@ int aie2_destroy_context(struct amdxdna_dev_hdl *ndev, struct amdxdna_hwctx *hwc
 		 hwctx->fw_ctx_id);
 	hwctx->priv->mbox_chann = NULL;
 	hwctx->fw_ctx_id = -1;
+	ndev->hwctx_num--;
 
 	return ret;
 }
diff --git a/drivers/accel/amdxdna/aie2_pci.c b/drivers/accel/amdxdna/aie2_pci.c
index ce57b915004e..396dc6e06007 100644
--- a/drivers/accel/amdxdna/aie2_pci.c
+++ b/drivers/accel/amdxdna/aie2_pci.c
@@ -838,6 +838,30 @@ static int aie2_get_hwctx_status(struct amdxdna_client *client,
 	return 0;
 }
 
+static int aie2_query_resource_info(struct amdxdna_client *client,
+				    struct amdxdna_drm_get_info *args)
+{
+	struct amdxdna_drm_get_resource_info res_info;
+	const struct amdxdna_dev_priv *priv;
+	struct amdxdna_dev_hdl *ndev;
+	struct amdxdna_dev *xdna;
+
+	xdna = client->xdna;
+	ndev = xdna->dev_handle;
+	priv = ndev->priv;
+
+	res_info.npu_clk_max = priv->dpm_clk_tbl[ndev->max_dpm_level].hclk;
+	res_info.npu_tops_max = ndev->max_tops;
+	res_info.npu_task_max = priv->hwctx_limit;
+	res_info.npu_tops_curr = ndev->curr_tops;
+	res_info.npu_task_curr = ndev->hwctx_num;
+
+	if (copy_to_user(u64_to_user_ptr(args->buffer), &res_info, sizeof(res_info)))
+		return -EFAULT;
+
+	return 0;
+}
+
 static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_info *args)
 {
 	struct amdxdna_dev *xdna = client->xdna;
@@ -872,6 +896,9 @@ static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_i
 	case DRM_AMDXDNA_GET_POWER_MODE:
 		ret = aie2_get_power_mode(client, args);
 		break;
+	case DRM_AMDXDNA_QUERY_RESOURCE_INFO:
+		ret = aie2_query_resource_info(client, args);
+		break;
 	default:
 		XDNA_ERR(xdna, "Not supported request parameter %u", args->param);
 		ret = -EOPNOTSUPP;
diff --git a/drivers/accel/amdxdna/amdxdna_pci_drv.c b/drivers/accel/amdxdna/amdxdna_pci_drv.c
index 3599e713bfcb..af943a603ad1 100644
--- a/drivers/accel/amdxdna/amdxdna_pci_drv.c
+++ b/drivers/accel/amdxdna/amdxdna_pci_drv.c
@@ -29,9 +29,10 @@ MODULE_FIRMWARE("amdnpu/17f0_20/npu.sbin");
  * 0.1: Support getting all hardware contexts by DRM_IOCTL_AMDXDNA_GET_ARRAY
  * 0.2: Support getting last error hardware error
  * 0.3: Support firmware debug buffer
+ * 0.4: Support getting resource information
  */
 #define AMDXDNA_DRIVER_MAJOR		0
-#define AMDXDNA_DRIVER_MINOR		3
+#define AMDXDNA_DRIVER_MINOR		4
 
 /*
  * Bind the driver base on (vendor_id, device_id) pair and later use the
diff --git a/include/uapi/drm/amdxdna_accel.h b/include/uapi/drm/amdxdna_accel.h
index c7eec9ceb2ae..8b679c38d308 100644
--- a/include/uapi/drm/amdxdna_accel.h
+++ b/include/uapi/drm/amdxdna_accel.h
@@ -442,6 +442,23 @@ enum amdxdna_drm_get_param {
 	DRM_AMDXDNA_QUERY_HW_CONTEXTS,
 	DRM_AMDXDNA_QUERY_FIRMWARE_VERSION = 8,
 	DRM_AMDXDNA_GET_POWER_MODE,
+	DRM_AMDXDNA_QUERY_RESOURCE_INFO = 12,
+};
+
+/**
+ * struct amdxdna_drm_get_resource_info - Get resource information
+ */
+struct amdxdna_drm_get_resource_info {
+	/** @npu_clk_max: max H-Clocks */
+	__u64 npu_clk_max;
+	/** @npu_tops_max: max TOPs */
+	__u64 npu_tops_max;
+	/** @npu_task_max: max number of tasks */
+	__u64 npu_task_max;
+	/** @npu_tops_curr: current TOPs */
+	__u64 npu_tops_curr;
+	/** @npu_task_curr: current number of tasks */
+	__u64 npu_task_curr;
 };
 
 /**
-- 
cgit v1.2.3


From e568dc3e625d818f199bd085005213cce3271453 Mon Sep 17 00:00:00 2001
From: Lizhi Hou <lizhi.hou@amd.com>
Date: Mon, 3 Nov 2025 22:25:45 -0800
Subject: accel/amdxdna: Add IOCTL parameter for telemetry data

Extend DRM_IOCTL_AMDXDNA_GET_INFO to include additional parameters
that allow collection of telemetry data.

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Link: https://patch.msgid.link/20251104062546.833771-3-lizhi.hou@amd.com
---
 drivers/accel/amdxdna/aie2_message.c           | 56 +++++++++++++++++---
 drivers/accel/amdxdna/aie2_msg_priv.h          | 25 ++++++++-
 drivers/accel/amdxdna/aie2_pci.c               | 73 ++++++++++++++++++++++++++
 drivers/accel/amdxdna/aie2_pci.h               |  3 ++
 drivers/accel/amdxdna/amdxdna_mailbox_helper.h |  6 ++-
 drivers/accel/amdxdna/amdxdna_pci_drv.c        |  3 +-
 include/uapi/drm/amdxdna_accel.h               | 17 ++++++
 7 files changed, 173 insertions(+), 10 deletions(-)

(limited to 'include')

diff --git a/drivers/accel/amdxdna/aie2_message.c b/drivers/accel/amdxdna/aie2_message.c
index 39214253d804..69cdce9ff208 100644
--- a/drivers/accel/amdxdna/aie2_message.c
+++ b/drivers/accel/amdxdna/aie2_message.c
@@ -47,7 +47,7 @@ static int aie2_send_mgmt_msg_wait(struct amdxdna_dev_hdl *ndev,
 		ndev->mgmt_chann = NULL;
 	}
 
-	if (!ret && *hdl->data != AIE2_STATUS_SUCCESS) {
+	if (!ret && *hdl->status != AIE2_STATUS_SUCCESS) {
 		XDNA_ERR(xdna, "command opcode 0x%x failed, status 0x%x",
 			 msg->opcode, *hdl->data);
 		ret = -EINVAL;
@@ -336,11 +336,6 @@ int aie2_query_status(struct amdxdna_dev_hdl *ndev, char __user *buf,
 		goto fail;
 	}
 
-	if (resp.status != AIE2_STATUS_SUCCESS) {
-		XDNA_ERR(xdna, "Query NPU status failed, status 0x%x", resp.status);
-		ret = -EINVAL;
-		goto fail;
-	}
 	XDNA_DBG(xdna, "Query NPU status completed");
 
 	if (size < resp.size) {
@@ -362,6 +357,55 @@ fail:
 	return ret;
 }
 
+int aie2_query_telemetry(struct amdxdna_dev_hdl *ndev,
+			 char __user *buf, u32 size,
+			 struct amdxdna_drm_query_telemetry_header *header)
+{
+	DECLARE_AIE2_MSG(get_telemetry, MSG_OP_GET_TELEMETRY);
+	struct amdxdna_dev *xdna = ndev->xdna;
+	dma_addr_t dma_addr;
+	u8 *addr;
+	int ret;
+
+	if (header->type >= MAX_TELEMETRY_TYPE)
+		return -EINVAL;
+
+	addr = dma_alloc_noncoherent(xdna->ddev.dev, size, &dma_addr,
+				     DMA_FROM_DEVICE, GFP_KERNEL);
+	if (!addr)
+		return -ENOMEM;
+
+	req.buf_addr = dma_addr;
+	req.buf_size = size;
+	req.type = header->type;
+
+	drm_clflush_virt_range(addr, size); /* device can access */
+	ret = aie2_send_mgmt_msg_wait(ndev, &msg);
+	if (ret) {
+		XDNA_ERR(xdna, "Query telemetry failed, status %d", ret);
+		goto free_buf;
+	}
+
+	if (size < resp.size) {
+		ret = -EINVAL;
+		XDNA_ERR(xdna, "Bad buffer size. Available: %u. Needs: %u", size, resp.size);
+		goto free_buf;
+	}
+
+	if (copy_to_user(buf, addr, resp.size)) {
+		ret = -EFAULT;
+		XDNA_ERR(xdna, "Failed to copy telemetry to user space");
+		goto free_buf;
+	}
+
+	header->major = resp.major;
+	header->minor = resp.minor;
+
+free_buf:
+	dma_free_noncoherent(xdna->ddev.dev, size, addr, dma_addr, DMA_FROM_DEVICE);
+	return ret;
+}
+
 int aie2_register_asyn_event_msg(struct amdxdna_dev_hdl *ndev, dma_addr_t addr, u32 size,
 				 void *handle, int (*cb)(void*, void __iomem *, size_t))
 {
diff --git a/drivers/accel/amdxdna/aie2_msg_priv.h b/drivers/accel/amdxdna/aie2_msg_priv.h
index 945140011763..947daa63f064 100644
--- a/drivers/accel/amdxdna/aie2_msg_priv.h
+++ b/drivers/accel/amdxdna/aie2_msg_priv.h
@@ -9,7 +9,8 @@
 enum aie2_msg_opcode {
 	MSG_OP_CREATE_CONTEXT              = 0x2,
 	MSG_OP_DESTROY_CONTEXT             = 0x3,
-	MSG_OP_SYNC_BO			   = 0x7,
+	MSG_OP_GET_TELEMETRY               = 0x4,
+	MSG_OP_SYNC_BO                     = 0x7,
 	MSG_OP_EXECUTE_BUFFER_CF           = 0xC,
 	MSG_OP_QUERY_COL_STATUS            = 0xD,
 	MSG_OP_QUERY_AIE_TILE_INFO         = 0xE,
@@ -137,6 +138,28 @@ struct destroy_ctx_resp {
 	enum aie2_msg_status	status;
 } __packed;
 
+enum telemetry_type {
+	TELEMETRY_TYPE_DISABLED,
+	TELEMETRY_TYPE_HEALTH,
+	TELEMETRY_TYPE_ERROR_INFO,
+	TELEMETRY_TYPE_PROFILING,
+	TELEMETRY_TYPE_DEBUG,
+	MAX_TELEMETRY_TYPE
+};
+
+struct get_telemetry_req {
+	enum telemetry_type	type;
+	__u64	buf_addr;
+	__u32	buf_size;
+} __packed;
+
+struct get_telemetry_resp {
+	__u32	major;
+	__u32	minor;
+	__u32	size;
+	enum aie2_msg_status	status;
+} __packed;
+
 struct execute_buffer_req {
 	__u32	cu_idx;
 	__u32	payload[19];
diff --git a/drivers/accel/amdxdna/aie2_pci.c b/drivers/accel/amdxdna/aie2_pci.c
index 396dc6e06007..d7ccbdaf47f5 100644
--- a/drivers/accel/amdxdna/aie2_pci.c
+++ b/drivers/accel/amdxdna/aie2_pci.c
@@ -862,6 +862,76 @@ static int aie2_query_resource_info(struct amdxdna_client *client,
 	return 0;
 }
 
+static int aie2_fill_hwctx_map(struct amdxdna_hwctx *hwctx, void *arg)
+{
+	struct amdxdna_dev *xdna = hwctx->client->xdna;
+	u32 *map = arg;
+
+	if (hwctx->fw_ctx_id >= xdna->dev_handle->priv->hwctx_limit) {
+		XDNA_ERR(xdna, "Invalid fw ctx id %d/%d ", hwctx->fw_ctx_id,
+			 xdna->dev_handle->priv->hwctx_limit);
+		return -EINVAL;
+	}
+
+	map[hwctx->fw_ctx_id] = hwctx->id;
+	return 0;
+}
+
+static int aie2_get_telemetry(struct amdxdna_client *client,
+			      struct amdxdna_drm_get_info *args)
+{
+	struct amdxdna_drm_query_telemetry_header *header __free(kfree) = NULL;
+	u32 telemetry_data_sz, header_sz, elem_num;
+	struct amdxdna_dev *xdna = client->xdna;
+	struct amdxdna_client *tmp_client;
+	int ret;
+
+	elem_num = xdna->dev_handle->priv->hwctx_limit;
+	header_sz = struct_size(header, map, elem_num);
+	if (args->buffer_size <= header_sz) {
+		XDNA_ERR(xdna, "Invalid buffer size");
+		return -EINVAL;
+	}
+
+	telemetry_data_sz = args->buffer_size - header_sz;
+	if (telemetry_data_sz > SZ_4M) {
+		XDNA_ERR(xdna, "Buffer size is too big, %d", telemetry_data_sz);
+		return -EINVAL;
+	}
+
+	header = kzalloc(header_sz, GFP_KERNEL);
+	if (!header)
+		return -ENOMEM;
+
+	if (copy_from_user(header, u64_to_user_ptr(args->buffer), sizeof(*header))) {
+		XDNA_ERR(xdna, "Failed to copy telemetry header from user");
+		return -EFAULT;
+	}
+
+	header->map_num_elements = elem_num;
+	list_for_each_entry(tmp_client, &xdna->client_list, node) {
+		ret = amdxdna_hwctx_walk(tmp_client, &header->map,
+					 aie2_fill_hwctx_map);
+		if (ret)
+			return ret;
+	}
+
+	ret = aie2_query_telemetry(xdna->dev_handle,
+				   u64_to_user_ptr(args->buffer + header_sz),
+				   telemetry_data_sz, header);
+	if (ret) {
+		XDNA_ERR(xdna, "Query telemetry failed ret %d", ret);
+		return ret;
+	}
+
+	if (copy_to_user(u64_to_user_ptr(args->buffer), header, header_sz)) {
+		XDNA_ERR(xdna, "Copy header failed");
+		return -EFAULT;
+	}
+
+	return 0;
+}
+
 static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_info *args)
 {
 	struct amdxdna_dev *xdna = client->xdna;
@@ -896,6 +966,9 @@ static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_i
 	case DRM_AMDXDNA_GET_POWER_MODE:
 		ret = aie2_get_power_mode(client, args);
 		break;
+	case DRM_AMDXDNA_QUERY_TELEMETRY:
+		ret = aie2_get_telemetry(client, args);
+		break;
 	case DRM_AMDXDNA_QUERY_RESOURCE_INFO:
 		ret = aie2_query_resource_info(client, args);
 		break;
diff --git a/drivers/accel/amdxdna/aie2_pci.h b/drivers/accel/amdxdna/aie2_pci.h
index a79f4f71ff6b..9793cd1e0c55 100644
--- a/drivers/accel/amdxdna/aie2_pci.h
+++ b/drivers/accel/amdxdna/aie2_pci.h
@@ -305,6 +305,9 @@ int aie2_create_context(struct amdxdna_dev_hdl *ndev, struct amdxdna_hwctx *hwct
 int aie2_destroy_context(struct amdxdna_dev_hdl *ndev, struct amdxdna_hwctx *hwctx);
 int aie2_map_host_buf(struct amdxdna_dev_hdl *ndev, u32 context_id, u64 addr, u64 size);
 int aie2_query_status(struct amdxdna_dev_hdl *ndev, char __user *buf, u32 size, u32 *cols_filled);
+int aie2_query_telemetry(struct amdxdna_dev_hdl *ndev,
+			 char __user *buf, u32 size,
+			 struct amdxdna_drm_query_telemetry_header *header);
 int aie2_register_asyn_event_msg(struct amdxdna_dev_hdl *ndev, dma_addr_t addr, u32 size,
 				 void *handle, int (*cb)(void*, void __iomem *, size_t));
 int aie2_config_cu(struct amdxdna_hwctx *hwctx,
diff --git a/drivers/accel/amdxdna/amdxdna_mailbox_helper.h b/drivers/accel/amdxdna/amdxdna_mailbox_helper.h
index 710ff8873d61..556c712cad0a 100644
--- a/drivers/accel/amdxdna/amdxdna_mailbox_helper.h
+++ b/drivers/accel/amdxdna/amdxdna_mailbox_helper.h
@@ -16,16 +16,18 @@ struct xdna_notify {
 	u32			*data;
 	size_t			size;
 	int			error;
+	u32			*status;
 };
 
-#define DECLARE_XDNA_MSG_COMMON(name, op, status)			\
+#define DECLARE_XDNA_MSG_COMMON(name, op, s)				\
 	struct name##_req	req = { 0 };				\
-	struct name##_resp	resp = { status	};			\
+	struct name##_resp	resp = { .status = s };			\
 	struct xdna_notify	hdl = {					\
 		.error = 0,						\
 		.data = (u32 *)&resp,					\
 		.size = sizeof(resp),					\
 		.comp = COMPLETION_INITIALIZER_ONSTACK(hdl.comp),	\
+		.status = (u32 *)&resp.status,				\
 	};								\
 	struct xdna_mailbox_msg msg = {					\
 		.send_data = (u8 *)&req,				\
diff --git a/drivers/accel/amdxdna/amdxdna_pci_drv.c b/drivers/accel/amdxdna/amdxdna_pci_drv.c
index af943a603ad1..7590265d4485 100644
--- a/drivers/accel/amdxdna/amdxdna_pci_drv.c
+++ b/drivers/accel/amdxdna/amdxdna_pci_drv.c
@@ -30,9 +30,10 @@ MODULE_FIRMWARE("amdnpu/17f0_20/npu.sbin");
  * 0.2: Support getting last error hardware error
  * 0.3: Support firmware debug buffer
  * 0.4: Support getting resource information
+ * 0.5: Support getting telemetry data
  */
 #define AMDXDNA_DRIVER_MAJOR		0
-#define AMDXDNA_DRIVER_MINOR		4
+#define AMDXDNA_DRIVER_MINOR		5
 
 /*
  * Bind the driver base on (vendor_id, device_id) pair and later use the
diff --git a/include/uapi/drm/amdxdna_accel.h b/include/uapi/drm/amdxdna_accel.h
index 8b679c38d308..8ad254bc35a5 100644
--- a/include/uapi/drm/amdxdna_accel.h
+++ b/include/uapi/drm/amdxdna_accel.h
@@ -442,6 +442,7 @@ enum amdxdna_drm_get_param {
 	DRM_AMDXDNA_QUERY_HW_CONTEXTS,
 	DRM_AMDXDNA_QUERY_FIRMWARE_VERSION = 8,
 	DRM_AMDXDNA_GET_POWER_MODE,
+	DRM_AMDXDNA_QUERY_TELEMETRY,
 	DRM_AMDXDNA_QUERY_RESOURCE_INFO = 12,
 };
 
@@ -461,6 +462,22 @@ struct amdxdna_drm_get_resource_info {
 	__u64 npu_task_curr;
 };
 
+/**
+ * struct amdxdna_drm_query_telemetry_header - Telemetry data header
+ */
+struct amdxdna_drm_query_telemetry_header {
+	/** @major: Firmware telemetry interface major version number */
+	__u32 major;
+	/** @minor: Firmware telemetry interface minor version number */
+	__u32 minor;
+	/** @type: Telemetry query type */
+	__u32 type;
+	/** @map_num_elements: Total number of elements in the map table */
+	__u32 map_num_elements;
+	/** @map: Element map */
+	__u32 map[];
+};
+
 /**
  * struct amdxdna_drm_get_info - Get some information from the AIE hardware.
  * @param: Value in enum amdxdna_drm_get_param. Specifies the structure passed in the buffer.
-- 
cgit v1.2.3


From 3a0ff7b98af4a5de1b995dfb57e65843f9b7b628 Mon Sep 17 00:00:00 2001
From: Lizhi Hou <lizhi.hou@amd.com>
Date: Tue, 4 Nov 2025 10:53:39 -0800
Subject: accel/amdxdna: Support preemption requests

The driver checks the firmware version during initialization.If preemption
is supported, the driver configures preemption accordingly and handles
userspace preemption requests. Otherwise, the driver returns an error for
userspace preemption requests.

Reviewed-by: Mario Limonciello (AMD) <superm1@kernel.org>
Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Link: https://patch.msgid.link/20251104185340.897560-1-lizhi.hou@amd.com
---
 drivers/accel/amdxdna/aie2_message.c    | 95 +++++++++++++++++++++++++++++++++
 drivers/accel/amdxdna/aie2_msg_priv.h   |  3 ++
 drivers/accel/amdxdna/aie2_pci.c        | 63 ++++++++++++++++++++++
 drivers/accel/amdxdna/aie2_pci.h        |  8 +++
 drivers/accel/amdxdna/amdxdna_ctx.h     | 17 ++++++
 drivers/accel/amdxdna/amdxdna_pci_drv.c |  3 +-
 drivers/accel/amdxdna/npu4_regs.c       |  4 ++
 include/uapi/drm/amdxdna_accel.h        | 16 +++++-
 8 files changed, 207 insertions(+), 2 deletions(-)

(limited to 'include')

diff --git a/drivers/accel/amdxdna/aie2_message.c b/drivers/accel/amdxdna/aie2_message.c
index 69cdce9ff208..d493bb1c3360 100644
--- a/drivers/accel/amdxdna/aie2_message.c
+++ b/drivers/accel/amdxdna/aie2_message.c
@@ -210,6 +210,14 @@ int aie2_create_context(struct amdxdna_dev_hdl *ndev, struct amdxdna_hwctx *hwct
 	hwctx->fw_ctx_id = resp.context_id;
 	WARN_ONCE(hwctx->fw_ctx_id == -1, "Unexpected context id");
 
+	if (ndev->force_preempt_enabled) {
+		ret = aie2_runtime_cfg(ndev, AIE2_RT_CFG_FORCE_PREEMPT, &hwctx->fw_ctx_id);
+		if (ret) {
+			XDNA_ERR(xdna, "failed to enable force preempt %d", ret);
+			return ret;
+		}
+	}
+
 	cq_pair = &resp.cq_pair[0];
 	x2i.mb_head_ptr_reg = AIE2_MBOX_OFF(ndev, cq_pair->x2i_q.head_addr);
 	x2i.mb_tail_ptr_reg = AIE2_MBOX_OFF(ndev, cq_pair->x2i_q.tail_addr);
@@ -601,6 +609,11 @@ aie2_cmdlist_fill_dpu(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size)
 	return 0;
 }
 
+static int aie2_cmdlist_unsupp(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size)
+{
+	return -EOPNOTSUPP;
+}
+
 static u32 aie2_get_chain_msg_op(u32 cmd_op)
 {
 	switch (cmd_op) {
@@ -621,6 +634,8 @@ static struct aie2_exec_msg_ops legacy_exec_message_ops = {
 	.init_chain_req = aie2_init_exec_chain_req,
 	.fill_cf_slot = aie2_cmdlist_fill_cf,
 	.fill_dpu_slot = aie2_cmdlist_fill_dpu,
+	.fill_preempt_slot = aie2_cmdlist_unsupp,
+	.fill_elf_slot = aie2_cmdlist_unsupp,
 	.get_chain_msg_op = aie2_get_chain_msg_op,
 };
 
@@ -680,6 +695,74 @@ aie2_cmdlist_fill_npu_dpu(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *si
 	return 0;
 }
 
+static int
+aie2_cmdlist_fill_npu_preempt(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size)
+{
+	struct cmd_chain_slot_npu *npu_slot = slot;
+	struct amdxdna_cmd_preempt_data *pd;
+	u32 cmd_len;
+	u32 arg_sz;
+
+	pd = amdxdna_cmd_get_payload(cmd_bo, &cmd_len);
+	arg_sz = cmd_len - sizeof(*pd);
+	if (cmd_len < sizeof(*pd) || arg_sz > MAX_NPU_ARGS_SIZE)
+		return -EINVAL;
+
+	if (*size < sizeof(*npu_slot) + arg_sz)
+		return -EINVAL;
+
+	npu_slot->cu_idx = amdxdna_cmd_get_cu_idx(cmd_bo);
+	if (npu_slot->cu_idx == INVALID_CU_IDX)
+		return -EINVAL;
+
+	memset(npu_slot, 0, sizeof(*npu_slot));
+	npu_slot->type = EXEC_NPU_TYPE_PREEMPT;
+	npu_slot->inst_buf_addr = pd->inst_buf;
+	npu_slot->save_buf_addr = pd->save_buf;
+	npu_slot->restore_buf_addr = pd->restore_buf;
+	npu_slot->inst_size = pd->inst_size;
+	npu_slot->save_size = pd->save_size;
+	npu_slot->restore_size = pd->restore_size;
+	npu_slot->inst_prop_cnt = pd->inst_prop_cnt;
+	npu_slot->arg_cnt = arg_sz / sizeof(u32);
+	memcpy(npu_slot->args, pd->prop_args, arg_sz);
+
+	*size = sizeof(*npu_slot) + arg_sz;
+	return 0;
+}
+
+static int
+aie2_cmdlist_fill_npu_elf(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size)
+{
+	struct cmd_chain_slot_npu *npu_slot = slot;
+	struct amdxdna_cmd_preempt_data *pd;
+	u32 cmd_len;
+	u32 arg_sz;
+
+	pd = amdxdna_cmd_get_payload(cmd_bo, &cmd_len);
+	arg_sz = cmd_len - sizeof(*pd);
+	if (cmd_len < sizeof(*pd) || arg_sz > MAX_NPU_ARGS_SIZE)
+		return -EINVAL;
+
+	if (*size < sizeof(*npu_slot) + arg_sz)
+		return -EINVAL;
+
+	memset(npu_slot, 0, sizeof(*npu_slot));
+	npu_slot->type = EXEC_NPU_TYPE_ELF;
+	npu_slot->inst_buf_addr = pd->inst_buf;
+	npu_slot->save_buf_addr = pd->save_buf;
+	npu_slot->restore_buf_addr = pd->restore_buf;
+	npu_slot->inst_size = pd->inst_size;
+	npu_slot->save_size = pd->save_size;
+	npu_slot->restore_size = pd->restore_size;
+	npu_slot->inst_prop_cnt = pd->inst_prop_cnt;
+	npu_slot->arg_cnt = 1;
+	npu_slot->args[0] = AIE2_EXEC_BUFFER_KERNEL_OP_TXN;
+
+	*size = struct_size(npu_slot, args, npu_slot->arg_cnt);
+	return 0;
+}
+
 static u32 aie2_get_npu_chain_msg_op(u32 cmd_op)
 {
 	return MSG_OP_CHAIN_EXEC_NPU;
@@ -691,6 +774,8 @@ static struct aie2_exec_msg_ops npu_exec_message_ops = {
 	.init_chain_req = aie2_init_npu_chain_req,
 	.fill_cf_slot = aie2_cmdlist_fill_npu_cf,
 	.fill_dpu_slot = aie2_cmdlist_fill_npu_dpu,
+	.fill_preempt_slot = aie2_cmdlist_fill_npu_preempt,
+	.fill_elf_slot = aie2_cmdlist_fill_npu_elf,
 	.get_chain_msg_op = aie2_get_npu_chain_msg_op,
 };
 
@@ -749,6 +834,16 @@ aie2_cmdlist_fill_slot(void *slot, struct amdxdna_gem_obj *cmd_abo,
 	case ERT_START_NPU:
 		ret = EXEC_MSG_OPS(xdna)->fill_dpu_slot(cmd_abo, slot, size);
 		break;
+	case ERT_START_NPU_PREEMPT:
+		if (!AIE2_FEATURE_ON(xdna->dev_handle, AIE2_PREEMPT))
+			return -EOPNOTSUPP;
+		ret = EXEC_MSG_OPS(xdna)->fill_preempt_slot(cmd_abo, slot, size);
+		break;
+	case ERT_START_NPU_PREEMPT_ELF:
+		if (!AIE2_FEATURE_ON(xdna->dev_handle, AIE2_PREEMPT))
+			return -EOPNOTSUPP;
+		ret = EXEC_MSG_OPS(xdna)->fill_elf_slot(cmd_abo, slot, size);
+		break;
 	default:
 		XDNA_INFO(xdna, "Unsupported op %d", op);
 		ret = -EOPNOTSUPP;
diff --git a/drivers/accel/amdxdna/aie2_msg_priv.h b/drivers/accel/amdxdna/aie2_msg_priv.h
index 947daa63f064..1c957a6298d3 100644
--- a/drivers/accel/amdxdna/aie2_msg_priv.h
+++ b/drivers/accel/amdxdna/aie2_msg_priv.h
@@ -176,6 +176,8 @@ struct exec_dpu_req {
 enum exec_npu_type {
 	EXEC_NPU_TYPE_NON_ELF		= 0x1,
 	EXEC_NPU_TYPE_PARTIAL_ELF	= 0x2,
+	EXEC_NPU_TYPE_PREEMPT		= 0x3,
+	EXEC_NPU_TYPE_ELF		= 0x4,
 };
 
 union exec_req {
@@ -372,6 +374,7 @@ struct cmd_chain_slot_dpu {
 };
 
 #define MAX_NPU_ARGS_SIZE (26 * sizeof(__u32))
+#define AIE2_EXEC_BUFFER_KERNEL_OP_TXN	3
 struct cmd_chain_slot_npu {
 	enum exec_npu_type type;
 	u64 inst_buf_addr;
diff --git a/drivers/accel/amdxdna/aie2_pci.c b/drivers/accel/amdxdna/aie2_pci.c
index d7ccbdaf47f5..ceef1c502e9e 100644
--- a/drivers/accel/amdxdna/aie2_pci.c
+++ b/drivers/accel/amdxdna/aie2_pci.c
@@ -183,6 +183,10 @@ int aie2_runtime_cfg(struct amdxdna_dev_hdl *ndev,
 		if (cfg->category != category)
 			continue;
 
+		if (cfg->feature_mask &&
+		    bitmap_subset(&cfg->feature_mask, &ndev->feature_mask, AIE2_FEATURE_MAX))
+			continue;
+
 		value = val ? *val : cfg->value;
 		ret = aie2_set_runtime_cfg(ndev, cfg->type, value);
 		if (ret) {
@@ -932,6 +936,25 @@ static int aie2_get_telemetry(struct amdxdna_client *client,
 	return 0;
 }
 
+static int aie2_get_preempt_state(struct amdxdna_client *client,
+				  struct amdxdna_drm_get_info *args)
+{
+	struct amdxdna_drm_attribute_state state = {};
+	struct amdxdna_dev *xdna = client->xdna;
+	struct amdxdna_dev_hdl *ndev;
+
+	ndev = xdna->dev_handle;
+	if (args->param == DRM_AMDXDNA_GET_FORCE_PREEMPT_STATE)
+		state.state = ndev->force_preempt_enabled;
+	else if (args->param == DRM_AMDXDNA_GET_FRAME_BOUNDARY_PREEMPT_STATE)
+		state.state = ndev->frame_boundary_preempt;
+
+	if (copy_to_user(u64_to_user_ptr(args->buffer), &state, sizeof(state)))
+		return -EFAULT;
+
+	return 0;
+}
+
 static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_info *args)
 {
 	struct amdxdna_dev *xdna = client->xdna;
@@ -972,6 +995,10 @@ static int aie2_get_info(struct amdxdna_client *client, struct amdxdna_drm_get_i
 	case DRM_AMDXDNA_QUERY_RESOURCE_INFO:
 		ret = aie2_query_resource_info(client, args);
 		break;
+	case DRM_AMDXDNA_GET_FORCE_PREEMPT_STATE:
+	case DRM_AMDXDNA_GET_FRAME_BOUNDARY_PREEMPT_STATE:
+		ret = aie2_get_preempt_state(client, args);
+		break;
 	default:
 		XDNA_ERR(xdna, "Not supported request parameter %u", args->param);
 		ret = -EOPNOTSUPP;
@@ -1078,6 +1105,38 @@ static int aie2_set_power_mode(struct amdxdna_client *client,
 	return aie2_pm_set_mode(xdna->dev_handle, power_mode);
 }
 
+static int aie2_set_preempt_state(struct amdxdna_client *client,
+				  struct amdxdna_drm_set_state *args)
+{
+	struct amdxdna_dev_hdl *ndev = client->xdna->dev_handle;
+	struct amdxdna_drm_attribute_state state;
+	u32 val;
+	int ret;
+
+	if (copy_from_user(&state, u64_to_user_ptr(args->buffer), sizeof(state)))
+		return -EFAULT;
+
+	if (state.state > 1)
+		return -EINVAL;
+
+	if (XDNA_MBZ_DBG(client->xdna, state.pad, sizeof(state.pad)))
+		return -EINVAL;
+
+	if (args->param == DRM_AMDXDNA_SET_FORCE_PREEMPT) {
+		ndev->force_preempt_enabled = state.state;
+	} else if (args->param == DRM_AMDXDNA_SET_FRAME_BOUNDARY_PREEMPT) {
+		val = state.state;
+		ret = aie2_runtime_cfg(ndev, AIE2_RT_CFG_FRAME_BOUNDARY_PREEMPT,
+				       &val);
+		if (ret)
+			return ret;
+
+		ndev->frame_boundary_preempt = state.state;
+	}
+
+	return 0;
+}
+
 static int aie2_set_state(struct amdxdna_client *client,
 			  struct amdxdna_drm_set_state *args)
 {
@@ -1095,6 +1154,10 @@ static int aie2_set_state(struct amdxdna_client *client,
 	case DRM_AMDXDNA_SET_POWER_MODE:
 		ret = aie2_set_power_mode(client, args);
 		break;
+	case DRM_AMDXDNA_SET_FORCE_PREEMPT:
+	case DRM_AMDXDNA_SET_FRAME_BOUNDARY_PREEMPT:
+		ret = aie2_set_preempt_state(client, args);
+		break;
 	default:
 		XDNA_ERR(xdna, "Not supported request parameter %u", args->param);
 		ret = -EOPNOTSUPP;
diff --git a/drivers/accel/amdxdna/aie2_pci.h b/drivers/accel/amdxdna/aie2_pci.h
index 9793cd1e0c55..a5f9c42155d1 100644
--- a/drivers/accel/amdxdna/aie2_pci.h
+++ b/drivers/accel/amdxdna/aie2_pci.h
@@ -110,12 +110,15 @@ struct aie_metadata {
 enum rt_config_category {
 	AIE2_RT_CFG_INIT,
 	AIE2_RT_CFG_CLK_GATING,
+	AIE2_RT_CFG_FORCE_PREEMPT,
+	AIE2_RT_CFG_FRAME_BOUNDARY_PREEMPT,
 };
 
 struct rt_config {
 	u32	type;
 	u32	value;
 	u32	category;
+	unsigned long feature_mask;
 };
 
 struct dpm_clk_freq {
@@ -164,6 +167,8 @@ struct aie2_exec_msg_ops {
 	void (*init_chain_req)(void *req, u64 slot_addr, size_t size, u32 cmd_cnt);
 	int (*fill_cf_slot)(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size);
 	int (*fill_dpu_slot)(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size);
+	int (*fill_preempt_slot)(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size);
+	int (*fill_elf_slot)(struct amdxdna_gem_obj *cmd_bo, void *slot, size_t *size);
 	u32 (*get_chain_msg_op)(u32 cmd_op);
 };
 
@@ -197,6 +202,8 @@ struct amdxdna_dev_hdl {
 	u32				hclk_freq;
 	u32				max_tops;
 	u32				curr_tops;
+	u32				force_preempt_enabled;
+	u32				frame_boundary_preempt;
 
 	/* Mailbox and the management channel */
 	struct mailbox			*mbox;
@@ -223,6 +230,7 @@ struct aie2_hw_ops {
 
 enum aie2_fw_feature {
 	AIE2_NPU_COMMAND,
+	AIE2_PREEMPT,
 	AIE2_FEATURE_MAX
 };
 
diff --git a/drivers/accel/amdxdna/amdxdna_ctx.h b/drivers/accel/amdxdna/amdxdna_ctx.h
index d02fb32499fa..b6151244d64f 100644
--- a/drivers/accel/amdxdna/amdxdna_ctx.h
+++ b/drivers/accel/amdxdna/amdxdna_ctx.h
@@ -16,6 +16,8 @@ enum ert_cmd_opcode {
 	ERT_START_CU = 0,
 	ERT_CMD_CHAIN = 19,
 	ERT_START_NPU = 20,
+	ERT_START_NPU_PREEMPT = 21,
+	ERT_START_NPU_PREEMPT_ELF = 22,
 	ERT_INVALID_CMD	= ~0U,
 };
 
@@ -55,6 +57,21 @@ struct amdxdna_cmd_chain {
 	u64 data[] __counted_by(command_count);
 };
 
+/*
+ * Interpretation of the beginning of data payload for ERT_START_NPU_PREEMPT in
+ * amdxdna_cmd. The rest of the payload in amdxdna_cmd is regular kernel args.
+ */
+struct amdxdna_cmd_preempt_data {
+	u64 inst_buf;	    /* instruction buffer address */
+	u64 save_buf;	    /* save buffer address */
+	u64 restore_buf;    /* restore buffer address */
+	u32 inst_size;	    /* size of instruction buffer in bytes */
+	u32 save_size;	    /* size of save buffer in bytes */
+	u32 restore_size;   /* size of restore buffer in bytes */
+	u32 inst_prop_cnt;  /* properties count */
+	u32 prop_args[];    /* properties and regular kernel arguments */
+};
+
 /* Exec buffer command header format */
 #define AMDXDNA_CMD_STATE		GENMASK(3, 0)
 #define AMDXDNA_CMD_EXTRA_CU_MASK	GENMASK(11, 10)
diff --git a/drivers/accel/amdxdna/amdxdna_pci_drv.c b/drivers/accel/amdxdna/amdxdna_pci_drv.c
index 7590265d4485..1973ab67721b 100644
--- a/drivers/accel/amdxdna/amdxdna_pci_drv.c
+++ b/drivers/accel/amdxdna/amdxdna_pci_drv.c
@@ -31,9 +31,10 @@ MODULE_FIRMWARE("amdnpu/17f0_20/npu.sbin");
  * 0.3: Support firmware debug buffer
  * 0.4: Support getting resource information
  * 0.5: Support getting telemetry data
+ * 0.6: Support preemption
  */
 #define AMDXDNA_DRIVER_MAJOR		0
-#define AMDXDNA_DRIVER_MINOR		5
+#define AMDXDNA_DRIVER_MINOR		6
 
 /*
  * Bind the driver base on (vendor_id, device_id) pair and later use the
diff --git a/drivers/accel/amdxdna/npu4_regs.c b/drivers/accel/amdxdna/npu4_regs.c
index d90777275a9f..986a5f28ba24 100644
--- a/drivers/accel/amdxdna/npu4_regs.c
+++ b/drivers/accel/amdxdna/npu4_regs.c
@@ -64,10 +64,13 @@
 const struct rt_config npu4_default_rt_cfg[] = {
 	{ 5, 1, AIE2_RT_CFG_INIT }, /* PDI APP LOAD MODE */
 	{ 10, 1, AIE2_RT_CFG_INIT }, /* DEBUG BUF */
+	{ 14, 0, AIE2_RT_CFG_INIT, BIT_U64(AIE2_PREEMPT) }, /* Frame boundary preemption */
 	{ 1, 1, AIE2_RT_CFG_CLK_GATING }, /* Clock gating on */
 	{ 2, 1, AIE2_RT_CFG_CLK_GATING }, /* Clock gating on */
 	{ 3, 1, AIE2_RT_CFG_CLK_GATING }, /* Clock gating on */
 	{ 4, 1, AIE2_RT_CFG_CLK_GATING }, /* Clock gating on */
+	{ 13, 0, AIE2_RT_CFG_FORCE_PREEMPT },
+	{ 14, 0, AIE2_RT_CFG_FRAME_BOUNDARY_PREEMPT },
 	{ 0 },
 };
 
@@ -85,6 +88,7 @@ const struct dpm_clk_freq npu4_dpm_clk_table[] = {
 
 const struct aie2_fw_feature_tbl npu4_fw_feature_table[] = {
 	{ .feature = AIE2_NPU_COMMAND, .min_minor = 15 },
+	{ .feature = AIE2_PREEMPT, .min_minor = 12 },
 	{ 0 }
 };
 
diff --git a/include/uapi/drm/amdxdna_accel.h b/include/uapi/drm/amdxdna_accel.h
index 8ad254bc35a5..62c917fd4f7b 100644
--- a/include/uapi/drm/amdxdna_accel.h
+++ b/include/uapi/drm/amdxdna_accel.h
@@ -443,7 +443,9 @@ enum amdxdna_drm_get_param {
 	DRM_AMDXDNA_QUERY_FIRMWARE_VERSION = 8,
 	DRM_AMDXDNA_GET_POWER_MODE,
 	DRM_AMDXDNA_QUERY_TELEMETRY,
-	DRM_AMDXDNA_QUERY_RESOURCE_INFO = 12,
+	DRM_AMDXDNA_GET_FORCE_PREEMPT_STATE,
+	DRM_AMDXDNA_QUERY_RESOURCE_INFO,
+	DRM_AMDXDNA_GET_FRAME_BOUNDARY_PREEMPT_STATE,
 };
 
 /**
@@ -462,6 +464,16 @@ struct amdxdna_drm_get_resource_info {
 	__u64 npu_task_curr;
 };
 
+/**
+ * struct amdxdna_drm_attribute_state - State of an attribute
+ */
+struct amdxdna_drm_attribute_state {
+	/** @state: enabled or disabled */
+	__u8 state;
+	/** @pad: MBZ */
+	__u8 pad[7];
+};
+
 /**
  * struct amdxdna_drm_query_telemetry_header - Telemetry data header
  */
@@ -613,6 +625,8 @@ enum amdxdna_drm_set_param {
 	DRM_AMDXDNA_SET_POWER_MODE,
 	DRM_AMDXDNA_WRITE_AIE_MEM,
 	DRM_AMDXDNA_WRITE_AIE_REG,
+	DRM_AMDXDNA_SET_FORCE_PREEMPT,
+	DRM_AMDXDNA_SET_FRAME_BOUNDARY_PREEMPT,
 };
 
 /**
-- 
cgit v1.2.3


From b6fa6100cec0287856ec1b363b0a962a9be90e6c Mon Sep 17 00:00:00 2001
From: Randy Dunlap <rdunlap@infradead.org>
Date: Thu, 30 Oct 2025 22:41:51 -0700
Subject: drm/panfrost: fix UAPI kernel-doc warnings
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Fix all kernel-doc warnings in include/uapi/drm/panfrost_drm.h.
This mostly means modifying existing comments to conform to
kernel-doc format, but there also some additions of missing
kernel-doc comments and changing non-kernel-doc comments to
use "/*" to begin them.

Warning: panfrost_drm.h:83 struct member 'jc' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'in_syncs' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'in_sync_count' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'out_sync' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'bo_handles' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'bo_handle_count' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'requirements' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'jm_ctx_handle' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:83 struct member 'pad' not described
 in 'drm_panfrost_submit'
Warning: panfrost_drm.h:116 Incorrect use of kernel-doc format:
  * Returned offset for the BO in the GPU address space.  This offset
Warning: panfrost_drm.h:124 struct member 'size' not described
 in 'drm_panfrost_create_bo'
Warning: panfrost_drm.h:124 struct member 'flags' not described
 in 'drm_panfrost_create_bo'
Warning: panfrost_drm.h:124 struct member 'handle' not described
 in 'drm_panfrost_create_bo'
Warning: panfrost_drm.h:124 struct member 'pad' not described
 in 'drm_panfrost_create_bo'
Warning: panfrost_drm.h:124 struct member 'nonzero' not described
 in 'drm_panfrost_create_bo'
Warning: panfrost_drm.h:143 struct member 'handle' not described
 in 'drm_panfrost_mmap_bo'
Warning: panfrost_drm.h:143 struct member 'flags' not described
 in 'drm_panfrost_mmap_bo'
Warning: panfrost_drm.h:143 struct member 'offset' not described
 in 'drm_panfrost_mmap_bo'

Signed-off-by: Randy Dunlap <rdunlap@infradead.org>
Reviewed-by: Steven Price <steven.price@arm.com>
Reviewed-by: Adrián Larumbe <adrian.larumbe@collabora.com>
Signed-off-by: Steven Price <steven.price@arm.com>
Link: https://patch.msgid.link/20251031054152.1406764-1-rdunlap@infradead.org
---
 include/uapi/drm/panfrost_drm.h | 118 ++++++++++++++++++++++++++++------------
 1 file changed, 82 insertions(+), 36 deletions(-)

(limited to 'include')

diff --git a/include/uapi/drm/panfrost_drm.h b/include/uapi/drm/panfrost_drm.h
index e8b47c9f6976..1956431bb391 100644
--- a/include/uapi/drm/panfrost_drm.h
+++ b/include/uapi/drm/panfrost_drm.h
@@ -54,32 +54,46 @@ extern "C" {
  * This asks the kernel to have the GPU execute a render command list.
  */
 struct drm_panfrost_submit {
-
-	/** Address to GPU mapping of job descriptor */
+	/**
+	 * @jc: Address to GPU mapping of job descriptor
+	 */
 	__u64 jc;
-
-	/** An optional array of sync objects to wait on before starting this job. */
+	/**
+	 * @in_syncs: An optional array of sync objects to wait on
+	 * before starting this job.
+	 */
 	__u64 in_syncs;
-
-	/** Number of sync objects to wait on before starting this job. */
+	/**
+	 * @in_sync_count: Number of sync objects to wait on before
+	 * starting this job.
+	 */
 	__u32 in_sync_count;
-
-	/** An optional sync object to place the completion fence in. */
+	/**
+	 * @out_sync: An optional sync object to place the completion fence in.
+	 */
 	__u32 out_sync;
-
-	/** Pointer to a u32 array of the BOs that are referenced by the job. */
+	/**
+	 * @bo_handles: Pointer to a u32 array of the BOs that are
+	 * referenced by the job.
+	 */
 	__u64 bo_handles;
-
-	/** Number of BO handles passed in (size is that times 4). */
+	/**
+	 * @bo_handle_count: Number of BO handles passed in (size is
+	 * that times 4).
+	 */
 	__u32 bo_handle_count;
-
-	/** A combination of PANFROST_JD_REQ_* */
+	/**
+	 * @requirements: A combination of PANFROST_JD_REQ_*
+	 */
 	__u32 requirements;
-
-	/** JM context handle. Zero if you want to use the default context. */
+	/**
+	 * @jm_ctx_handle: JM context handle. Zero if you want to use the
+	 * default context.
+	 */
 	__u32 jm_ctx_handle;
-
-	/** Padding field. MBZ. */
+	/**
+	 * @pad: Padding field. Must be zero.
+	 */
 	__u32 pad;
 };
 
@@ -92,9 +106,18 @@ struct drm_panfrost_submit {
  * completed.
  */
 struct drm_panfrost_wait_bo {
+	/**
+	 * @handle: Handle for the object to wait for.
+	 */
 	__u32 handle;
+	/**
+	 * @pad: Padding, must be zero-filled.
+	 */
 	__u32 pad;
-	__s64 timeout_ns;	/* absolute */
+	/**
+	 * @timeout_ns: absolute number of nanoseconds to wait.
+	 */
+	__s64 timeout_ns;
 };
 
 /* Valid flags to pass to drm_panfrost_create_bo */
@@ -107,16 +130,26 @@ struct drm_panfrost_wait_bo {
  * The flags argument is a bit mask of PANFROST_BO_* flags.
  */
 struct drm_panfrost_create_bo {
+	/**
+	 * @size: size of shmem/BO area to create (bytes)
+	 */
 	__u32 size;
+	/**
+	 * @flags: see PANFROST_BO_* flags
+	 */
 	__u32 flags;
-	/** Returned GEM handle for the BO. */
+	/**
+	 * @handle: Returned GEM handle for the BO.
+	 */
 	__u32 handle;
-	/* Pad, must be zero-filled. */
+	/**
+	 * @pad: Padding, must be zero-filled.
+	 */
 	__u32 pad;
 	/**
-	 * Returned offset for the BO in the GPU address space.  This offset
-	 * is private to the DRM fd and is valid for the lifetime of the GEM
-	 * handle.
+	 * @offset: Returned offset for the BO in the GPU address space.
+	 * This offset is private to the DRM fd and is valid for the
+	 * lifetime of the GEM handle.
 	 *
 	 * This offset value will always be nonzero, since various HW
 	 * units treat 0 specially.
@@ -136,10 +169,17 @@ struct drm_panfrost_create_bo {
  * used in a future extension.
  */
 struct drm_panfrost_mmap_bo {
-	/** Handle for the object being mapped. */
+	/**
+	 * @handle: Handle for the object being mapped.
+	 */
 	__u32 handle;
+	/**
+	 * @flags: currently not used (should be zero)
+	 */
 	__u32 flags;
-	/** offset into the drm node to use for subsequent mmap call. */
+	/**
+	 * @offset: offset into the drm node to use for subsequent mmap call.
+	 */
 	__u64 offset;
 };
 
@@ -196,7 +236,7 @@ struct drm_panfrost_get_param {
 	__u64 value;
 };
 
-/**
+/*
  * Returns the offset for the BO in the GPU address space for this DRM fd.
  * This is the same value returned by drm_panfrost_create_bo, if that was called
  * from this DRM fd.
@@ -244,12 +284,14 @@ struct drm_panfrost_madvise {
  * struct drm_panfrost_set_label_bo - ioctl argument for labelling Panfrost BOs.
  */
 struct drm_panfrost_set_label_bo {
-	/** @handle: Handle of the buffer object to label. */
+	/**
+	 * @handle: Handle of the buffer object to label.
+	 */
 	__u32 handle;
-
-	/**  @pad: MBZ. */
+	/**
+	 * @pad: Must be zero.
+	 */
 	__u32 pad;
-
 	/**
 	 * @label: User pointer to a NUL-terminated string
 	 *
@@ -330,10 +372,13 @@ enum drm_panfrost_jm_ctx_priority {
 };
 
 struct drm_panfrost_jm_ctx_create {
-	/** @handle: Handle of the created JM context */
+	/**
+	 * @handle: Handle of the created JM context
+	 */
 	__u32 handle;
-
-	/** @priority: Context priority (see enum drm_panfrost_jm_ctx_priority). */
+	/**
+	 * @priority: Context priority (see enum drm_panfrost_jm_ctx_priority).
+	 */
 	__u32 priority;
 };
 
@@ -344,8 +389,9 @@ struct drm_panfrost_jm_ctx_destroy {
 	 * Must be a valid context handle returned by DRM_IOCTL_PANTHOR_JM_CTX_CREATE.
 	 */
 	__u32 handle;
-
-	/** @pad: Padding field, MBZ. */
+	/**
+	 * @pad: Padding field, must be zero.
+	 */
 	__u32 pad;
 };
 
-- 
cgit v1.2.3


From 3ca8b2668ca5a410c23a7a7932cc406ebbb89ab2 Mon Sep 17 00:00:00 2001
From: Bagas Sanjaya <bagasdotme@gmail.com>
Date: Thu, 6 Nov 2025 07:52:17 +0700
Subject: drm/ttm: Fix @alloc_flags description
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Stephen Rothwell reports htmldocs warnings when merging drm-misc tree:

Documentation/gpu/drm-mm:40: include/drm/ttm/ttm_device.h:225: ERROR: Unknown target name: "ttm_allocation". [docutils]
Documentation/gpu/drm-mm:43: drivers/gpu/drm/ttm/ttm_device.c:202: ERROR: Unknown target name: "ttm_allocation". [docutils]
Documentation/gpu/drm-mm:73: include/drm/ttm/ttm_pool.h:68: ERROR: Unknown target name: "ttm_allocation_pool". [docutils]
Documentation/gpu/drm-mm:76: drivers/gpu/drm/ttm/ttm_pool.c:1070: ERROR: Unknown target name: "ttm_allocation_pool". [docutils]

Fix these by adding missing wildcard on TTM_ALLOCATION_* and
TTM_ALLOCATION_POOL_* in @alloc_flags description.

Fixes: 0af5b6a8f8dd ("drm/ttm: Replace multiple booleans with flags in pool init")
Fixes: 77e19f8d3297 ("drm/ttm: Replace multiple booleans with flags in device init")
Fixes: 402b3a865090 ("drm/ttm: Add an allocation flag to propagate -ENOSPC on OOM")
Reported-by: Stephen Rothwell <sfr@canb.auug.org.au>
Closes: https://lore.kernel.org/linux-next/20251105161838.55b962a3@canb.auug.org.au/
Signed-off-by: Bagas Sanjaya <bagasdotme@gmail.com>
Acked-by: Tvrtko Ursulin <tvrtko.ursulin@igalia.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Tvrtko Ursulin <tursulin@ursulin.net>
Link: https://lore.kernel.org/r/20251106005217.14026-1-bagasdotme@gmail.com
---
 drivers/gpu/drm/ttm/ttm_device.c | 2 +-
 drivers/gpu/drm/ttm/ttm_pool.c   | 2 +-
 include/drm/ttm/ttm_device.h     | 2 +-
 include/drm/ttm/ttm_pool.h       | 2 +-
 4 files changed, 4 insertions(+), 4 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/ttm/ttm_device.c b/drivers/gpu/drm/ttm/ttm_device.c
index 5c10e5fbf43b..9a51afaf0749 100644
--- a/drivers/gpu/drm/ttm/ttm_device.c
+++ b/drivers/gpu/drm/ttm/ttm_device.c
@@ -199,7 +199,7 @@ EXPORT_SYMBOL(ttm_device_swapout);
  * @dev: The core kernel device pointer for DMA mappings and allocations.
  * @mapping: The address space to use for this bo.
  * @vma_manager: A pointer to a vma manager.
- * @alloc_flags: TTM_ALLOCATION_ flags.
+ * @alloc_flags: TTM_ALLOCATION_* flags.
  *
  * Initializes a struct ttm_device:
  * Returns:
diff --git a/drivers/gpu/drm/ttm/ttm_pool.c b/drivers/gpu/drm/ttm/ttm_pool.c
index 97e9ce505cf6..18b6db015619 100644
--- a/drivers/gpu/drm/ttm/ttm_pool.c
+++ b/drivers/gpu/drm/ttm/ttm_pool.c
@@ -1067,7 +1067,7 @@ long ttm_pool_backup(struct ttm_pool *pool, struct ttm_tt *tt,
  * @pool: the pool to initialize
  * @dev: device for DMA allocations and mappings
  * @nid: NUMA node to use for allocations
- * @alloc_flags: TTM_ALLOCATION_POOL_ flags
+ * @alloc_flags: TTM_ALLOCATION_POOL_* flags
  *
  * Initialize the pool and its pool types.
  */
diff --git a/include/drm/ttm/ttm_device.h b/include/drm/ttm/ttm_device.h
index d016360e5ceb..5618aef462f2 100644
--- a/include/drm/ttm/ttm_device.h
+++ b/include/drm/ttm/ttm_device.h
@@ -221,7 +221,7 @@ struct ttm_device {
 	struct list_head device_list;
 
 	/**
-	 * @alloc_flags: TTM_ALLOCATION_ flags.
+	 * @alloc_flags: TTM_ALLOCATION_* flags.
 	 */
 	unsigned int alloc_flags;
 
diff --git a/include/drm/ttm/ttm_pool.h b/include/drm/ttm/ttm_pool.h
index 67c72de913bb..233581670e78 100644
--- a/include/drm/ttm/ttm_pool.h
+++ b/include/drm/ttm/ttm_pool.h
@@ -64,7 +64,7 @@ struct ttm_pool_type {
  *
  * @dev: the device we allocate pages for
  * @nid: which numa node to use
- * @alloc_flags: TTM_ALLOCATION_POOL_ flags
+ * @alloc_flags: TTM_ALLOCATION_POOL_* flags
  * @caching: pools for each caching/order
  */
 struct ttm_pool {
-- 
cgit v1.2.3


From aaecfadc22cca4c7ad381b1df457e8857c01fe14 Mon Sep 17 00:00:00 2001
From: Michał Winiarski <michal.winiarski@intel.com>
Date: Wed, 12 Nov 2025 14:22:20 +0100
Subject: drm/intel/bmg: Allow device ID usage with single-argument macros
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

When INTEL_BMG_G21_IDS were added as a subplatform, token concatenation
operator usage was omitted, making INTEL_BMG_IDS not usable with
single-argument macros.
Fix that by adding the missing operator.

Fixes: 78de8f876683 ("drm/xe: Handle Wa_22010954014 and Wa_14022085890 as device workarounds")
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patch.msgid.link/20251112132220.516975-25-michal.winiarski@intel.com
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
---
 include/drm/intel/pciids.h | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

(limited to 'include')

diff --git a/include/drm/intel/pciids.h b/include/drm/intel/pciids.h
index 6e53fb4cdd37..286e580b2bd6 100644
--- a/include/drm/intel/pciids.h
+++ b/include/drm/intel/pciids.h
@@ -861,7 +861,7 @@
 	MACRO__(0xE216, ## __VA_ARGS__)
 
 #define INTEL_BMG_IDS(MACRO__, ...) \
-	INTEL_BMG_G21_IDS(MACRO__, __VA_ARGS__), \
+	INTEL_BMG_G21_IDS(MACRO__, ## __VA_ARGS__), \
 	MACRO__(0xE220, ## __VA_ARGS__), \
 	MACRO__(0xE221, ## __VA_ARGS__), \
 	MACRO__(0xE222, ## __VA_ARGS__), \
-- 
cgit v1.2.3


From 1ff27c5929ab0f5e34d5062637369ca542a6d385 Mon Sep 17 00:00:00 2001
From: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Date: Tue, 21 Oct 2025 13:19:15 +0300
Subject: drm/bridge: dw-hdmi-qp: Handle platform supported formats and color
 depth

Extend struct dw_hdmi_qp_plat_data to include the supported display
output formats and maximum bits per color channel.  When provided by the
platform driver, use them to setup the HDMI bridge accordingly.

Additionally, improve debug logging in dw_hdmi_qp_bridge_atomic_enable()
to also show the current HDMI output format and bpc.

Acked-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Cristian Ciocaltea <cristian.ciocaltea@collabora.com>
Signed-off-by: Heiko Stuebner <heiko@sntech.de>
Link: https://lore.kernel.org/r/20251021-rk3588-10bpc-v3-2-3d3eed00a6db@collabora.com
---
 drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c | 11 +++++++++--
 include/drm/bridge/dw_hdmi_qp.h              |  4 ++++
 2 files changed, 13 insertions(+), 2 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
index 4ba7b339eff6..fe4c026280f0 100644
--- a/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
+++ b/drivers/gpu/drm/bridge/synopsys/dw-hdmi-qp.c
@@ -868,8 +868,9 @@ static void dw_hdmi_qp_bridge_atomic_enable(struct drm_bridge *bridge,
 		return;
 
 	if (connector->display_info.is_hdmi) {
-		dev_dbg(hdmi->dev, "%s mode=HDMI rate=%llu\n",
-			__func__, conn_state->hdmi.tmds_char_rate);
+		dev_dbg(hdmi->dev, "%s mode=HDMI %s rate=%llu bpc=%u\n", __func__,
+			drm_hdmi_connector_get_output_format_name(conn_state->hdmi.output_format),
+			conn_state->hdmi.tmds_char_rate, conn_state->hdmi.output_bpc);
 		op_mode = 0;
 		hdmi->tmds_char_rate = conn_state->hdmi.tmds_char_rate;
 	} else {
@@ -1287,6 +1288,12 @@ struct dw_hdmi_qp *dw_hdmi_qp_bind(struct platform_device *pdev,
 	hdmi->bridge.vendor = "Synopsys";
 	hdmi->bridge.product = "DW HDMI QP TX";
 
+	if (plat_data->supported_formats)
+		hdmi->bridge.supported_formats = plat_data->supported_formats;
+
+	if (plat_data->max_bpc)
+		hdmi->bridge.max_bpc = plat_data->max_bpc;
+
 	hdmi->bridge.ddc = dw_hdmi_qp_i2c_adapter(hdmi);
 	if (IS_ERR(hdmi->bridge.ddc))
 		return ERR_CAST(hdmi->bridge.ddc);
diff --git a/include/drm/bridge/dw_hdmi_qp.h b/include/drm/bridge/dw_hdmi_qp.h
index 76ecf3130199..3f461f6b9bbf 100644
--- a/include/drm/bridge/dw_hdmi_qp.h
+++ b/include/drm/bridge/dw_hdmi_qp.h
@@ -25,6 +25,10 @@ struct dw_hdmi_qp_plat_data {
 	int main_irq;
 	int cec_irq;
 	unsigned long ref_clk_rate;
+	/* Supported output formats: bitmask of @hdmi_colorspace */
+	unsigned int supported_formats;
+	/* Maximum bits per color channel: 8, 10 or 12 */
+	unsigned int max_bpc;
 };
 
 struct dw_hdmi_qp *dw_hdmi_qp_bind(struct platform_device *pdev,
-- 
cgit v1.2.3


From 943240d342f148896733eb6c7b223a08aa1f520a Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 10 Nov 2025 16:44:21 +0100
Subject: drm/client: Pass force parameter to client restore

Add force parameter to client restore and pass value through the
layers. The only currently used value is false.

If force is true, the client should restore its display even if it
does not hold the DRM master lock. This is be required for emergency
output, such as sysrq.

While at it, inline drm_fb_helper_lastclose(), which is a trivial
wrapper around drm_fb_helper_restore_fbdev_mode_unlocked().

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patch.msgid.link/20251110154616.539328-2-tzimmermann@suse.de
---
 drivers/gpu/drm/clients/drm_fbdev_client.c |  6 ++++--
 drivers/gpu/drm/drm_client_event.c         |  4 ++--
 drivers/gpu/drm/drm_fb_helper.c            | 24 ++++++------------------
 drivers/gpu/drm/drm_file.c                 |  2 +-
 include/drm/drm_client.h                   |  8 +++++---
 include/drm/drm_client_event.h             |  4 ++--
 include/drm/drm_fb_helper.h                |  8 ++------
 7 files changed, 22 insertions(+), 34 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/clients/drm_fbdev_client.c b/drivers/gpu/drm/clients/drm_fbdev_client.c
index 47e5f27eee58..28951e392482 100644
--- a/drivers/gpu/drm/clients/drm_fbdev_client.c
+++ b/drivers/gpu/drm/clients/drm_fbdev_client.c
@@ -38,9 +38,11 @@ static void drm_fbdev_client_unregister(struct drm_client_dev *client)
 	}
 }
 
-static int drm_fbdev_client_restore(struct drm_client_dev *client)
+static int drm_fbdev_client_restore(struct drm_client_dev *client, bool force)
 {
-	drm_fb_helper_lastclose(client->dev);
+	struct drm_fb_helper *fb_helper = drm_fb_helper_from_client(client);
+
+	drm_fb_helper_restore_fbdev_mode_unlocked(fb_helper, force);
 
 	return 0;
 }
diff --git a/drivers/gpu/drm/drm_client_event.c b/drivers/gpu/drm/drm_client_event.c
index d25dc5250983..7b3e362f7926 100644
--- a/drivers/gpu/drm/drm_client_event.c
+++ b/drivers/gpu/drm/drm_client_event.c
@@ -102,7 +102,7 @@ void drm_client_dev_hotplug(struct drm_device *dev)
 }
 EXPORT_SYMBOL(drm_client_dev_hotplug);
 
-void drm_client_dev_restore(struct drm_device *dev)
+void drm_client_dev_restore(struct drm_device *dev, bool force)
 {
 	struct drm_client_dev *client;
 	int ret;
@@ -115,7 +115,7 @@ void drm_client_dev_restore(struct drm_device *dev)
 		if (!client->funcs || !client->funcs->restore)
 			continue;
 
-		ret = client->funcs->restore(client);
+		ret = client->funcs->restore(client, force);
 		drm_dbg_kms(dev, "%s: ret=%d\n", client->name, ret);
 		if (!ret) /* The first one to return zero gets the privilege to restore */
 			break;
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 53e9dc0543de..1392738ce2fe 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -255,6 +255,7 @@ __drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper,
 /**
  * drm_fb_helper_restore_fbdev_mode_unlocked - restore fbdev configuration
  * @fb_helper: driver-allocated fbdev helper, can be NULL
+ * @force: ignore present DRM master
  *
  * This helper should be called from fbdev emulation's &drm_client_funcs.restore
  * callback. It ensures that the user isn't greeted with a black screen when the
@@ -263,9 +264,9 @@ __drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper,
  * Returns:
  * 0 on success, or a negative errno code otherwise.
  */
-int drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper)
+int drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper, bool force)
 {
-	return __drm_fb_helper_restore_fbdev_mode_unlocked(fb_helper, false);
+	return __drm_fb_helper_restore_fbdev_mode_unlocked(fb_helper, force);
 }
 EXPORT_SYMBOL(drm_fb_helper_restore_fbdev_mode_unlocked);
 
@@ -1328,9 +1329,9 @@ int drm_fb_helper_set_par(struct fb_info *info)
 	 * the KDSET IOCTL with KD_TEXT, and only after that drops the master
 	 * status when exiting.
 	 *
-	 * In the past this was caught by drm_fb_helper_lastclose(), but on
-	 * modern systems where logind always keeps a drm fd open to orchestrate
-	 * the vt switching, this doesn't work.
+	 * In the past this was caught by drm_fb_helper_restore_fbdev_mode_unlocked(),
+	 * but on modern systems where logind always keeps a drm fd open to
+	 * orchestrate the vt switching, this doesn't work.
 	 *
 	 * To not break the userspace ABI we have this special case here, which
 	 * is only used for the above case. Everything else uses the normal
@@ -1955,16 +1956,3 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper)
 	return 0;
 }
 EXPORT_SYMBOL(drm_fb_helper_hotplug_event);
-
-/**
- * drm_fb_helper_lastclose - DRM driver lastclose helper for fbdev emulation
- * @dev: DRM device
- *
- * This function is obsolete. Call drm_fb_helper_restore_fbdev_mode_unlocked()
- * instead.
- */
-void drm_fb_helper_lastclose(struct drm_device *dev)
-{
-	drm_fb_helper_restore_fbdev_mode_unlocked(dev->fb_helper);
-}
-EXPORT_SYMBOL(drm_fb_helper_lastclose);
diff --git a/drivers/gpu/drm/drm_file.c b/drivers/gpu/drm/drm_file.c
index eebd1a05ee97..be5e617ceb9f 100644
--- a/drivers/gpu/drm/drm_file.c
+++ b/drivers/gpu/drm/drm_file.c
@@ -405,7 +405,7 @@ EXPORT_SYMBOL(drm_open);
 
 static void drm_lastclose(struct drm_device *dev)
 {
-	drm_client_dev_restore(dev);
+	drm_client_dev_restore(dev, false);
 
 	if (dev_is_pci(dev->dev))
 		vga_switcheroo_process_delayed_switch();
diff --git a/include/drm/drm_client.h b/include/drm/drm_client.h
index 5ecde0f6f591..c972a8a3385b 100644
--- a/include/drm/drm_client.h
+++ b/include/drm/drm_client.h
@@ -57,12 +57,14 @@ struct drm_client_funcs {
 	 *
 	 * Note that the core does not guarantee exclusion against concurrent
 	 * drm_open(). Clients need to ensure this themselves, for example by
-	 * using drm_master_internal_acquire() and
-	 * drm_master_internal_release().
+	 * using drm_master_internal_acquire() and drm_master_internal_release().
+	 *
+	 * If the caller passes force, the client should ignore any present DRM
+	 * master and restore the display anyway.
 	 *
 	 * This callback is optional.
 	 */
-	int (*restore)(struct drm_client_dev *client);
+	int (*restore)(struct drm_client_dev *client, bool force);
 
 	/**
 	 * @hotplug:
diff --git a/include/drm/drm_client_event.h b/include/drm/drm_client_event.h
index 985d6f02a4c4..79369c755bc9 100644
--- a/include/drm/drm_client_event.h
+++ b/include/drm/drm_client_event.h
@@ -10,7 +10,7 @@ struct drm_device;
 #if defined(CONFIG_DRM_CLIENT)
 void drm_client_dev_unregister(struct drm_device *dev);
 void drm_client_dev_hotplug(struct drm_device *dev);
-void drm_client_dev_restore(struct drm_device *dev);
+void drm_client_dev_restore(struct drm_device *dev, bool force);
 void drm_client_dev_suspend(struct drm_device *dev);
 void drm_client_dev_resume(struct drm_device *dev);
 #else
@@ -18,7 +18,7 @@ static inline void drm_client_dev_unregister(struct drm_device *dev)
 { }
 static inline void drm_client_dev_hotplug(struct drm_device *dev)
 { }
-static inline void drm_client_dev_restore(struct drm_device *dev)
+static inline void drm_client_dev_restore(struct drm_device *dev, bool force)
 { }
 static inline void drm_client_dev_suspend(struct drm_device *dev)
 { }
diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h
index c1d38d54a112..63e3af8dd5ed 100644
--- a/include/drm/drm_fb_helper.h
+++ b/include/drm/drm_fb_helper.h
@@ -254,7 +254,8 @@ int drm_fb_helper_set_par(struct fb_info *info);
 int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
 			    struct fb_info *info);
 
-int drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper);
+int drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper,
+					      bool force);
 
 struct fb_info *drm_fb_helper_alloc_info(struct drm_fb_helper *fb_helper);
 void drm_fb_helper_release_info(struct drm_fb_helper *fb_helper);
@@ -283,7 +284,6 @@ int drm_fb_helper_hotplug_event(struct drm_fb_helper *fb_helper);
 int drm_fb_helper_initial_config(struct drm_fb_helper *fb_helper);
 int drm_fb_helper_debug_enter(struct fb_info *info);
 int drm_fb_helper_debug_leave(struct fb_info *info);
-void drm_fb_helper_lastclose(struct drm_device *dev);
 #else
 static inline void drm_fb_helper_prepare(struct drm_device *dev,
 					 struct drm_fb_helper *helper,
@@ -409,10 +409,6 @@ static inline int drm_fb_helper_debug_leave(struct fb_info *info)
 {
 	return 0;
 }
-
-static inline void drm_fb_helper_lastclose(struct drm_device *dev)
-{
-}
 #endif
 
 #endif
-- 
cgit v1.2.3


From 6915190a50e8f7cf13dcbe534b02845be533b60a Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 10 Nov 2025 16:44:22 +0100
Subject: drm/client: Support emergency restore via sysrq for all clients

Move the sysrq functionality from DRM fbdev helpers to the DRM device
and in-kernel clients, so that it becomes available on all clients.

DRM fbdev helpers support emergency restoration of the console output
via a special key combination. Press SysRq+v to replace the current
compositor with the kernel's output on the framebuffer console. This
allows users to see the log messages during system emergencies.

By moving the functionality from fbdev helpers to the DRM device, any
in-kernel client can serve as emergency output. This can be used to
bring up drm_log, for example.

Each DRM device registers itself to the list of possible sysrq handlers.
On receiving SysRq+v, the DRM core goes over all registered devices and
restores an in-kernel DRM client for each of them.

See Documentation/admin-guide/sysrq.rst on how to invoke SysRq. Switch
VTs to bring back the user-space compositor.

v2:
- declare placeholders as 'static inline' (kernel test robot)
- fix grammar in commit description

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Reviewed-by: Jocelyn Falempe <jfalempe@redhat.com>
Link: https://patch.msgid.link/20251110154616.539328-3-tzimmermann@suse.de
---
 drivers/gpu/drm/Makefile           |  3 +-
 drivers/gpu/drm/drm_client.c       |  1 +
 drivers/gpu/drm/drm_client_sysrq.c | 65 ++++++++++++++++++++++++++++++++++++++
 drivers/gpu/drm/drm_drv.c          |  3 ++
 drivers/gpu/drm/drm_fb_helper.c    | 45 +-------------------------
 drivers/gpu/drm/drm_internal.h     | 11 +++++++
 include/drm/drm_device.h           |  8 +++++
 7 files changed, 91 insertions(+), 45 deletions(-)
 create mode 100644 drivers/gpu/drm/drm_client_sysrq.c

(limited to 'include')

diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index c2672f369aed..9901534948e5 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -76,7 +76,8 @@ drm-y := \
 drm-$(CONFIG_DRM_CLIENT) += \
 	drm_client.o \
 	drm_client_event.o \
-	drm_client_modeset.o
+	drm_client_modeset.o \
+	drm_client_sysrq.o
 drm-$(CONFIG_DRM_LIB_RANDOM) += lib/drm_random.o
 drm-$(CONFIG_COMPAT) += drm_ioc32.o
 drm-$(CONFIG_DRM_PANEL) += drm_panel.o
diff --git a/drivers/gpu/drm/drm_client.c b/drivers/gpu/drm/drm_client.c
index 504ec5bdfa2c..a82d741e6630 100644
--- a/drivers/gpu/drm/drm_client.c
+++ b/drivers/gpu/drm/drm_client.c
@@ -11,6 +11,7 @@
 #include <linux/slab.h>
 
 #include <drm/drm_client.h>
+#include <drm/drm_client_event.h>
 #include <drm/drm_device.h>
 #include <drm/drm_drv.h>
 #include <drm/drm_file.h>
diff --git a/drivers/gpu/drm/drm_client_sysrq.c b/drivers/gpu/drm/drm_client_sysrq.c
new file mode 100644
index 000000000000..eea660096f1b
--- /dev/null
+++ b/drivers/gpu/drm/drm_client_sysrq.c
@@ -0,0 +1,65 @@
+// SPDX-License-Identifier: GPL-2.0 or MIT
+
+#include <linux/sysrq.h>
+
+#include <drm/drm_client_event.h>
+#include <drm/drm_device.h>
+#include <drm/drm_print.h>
+
+#include "drm_internal.h"
+
+#ifdef CONFIG_MAGIC_SYSRQ
+static LIST_HEAD(drm_client_sysrq_dev_list);
+static DEFINE_MUTEX(drm_client_sysrq_dev_lock);
+
+/* emergency restore, don't bother with error reporting */
+static void drm_client_sysrq_restore_work_fn(struct work_struct *ignored)
+{
+	struct drm_device *dev;
+
+	guard(mutex)(&drm_client_sysrq_dev_lock);
+
+	list_for_each_entry(dev, &drm_client_sysrq_dev_list, client_sysrq_list) {
+		if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
+			continue;
+
+		drm_client_dev_restore(dev, true);
+	}
+}
+
+static DECLARE_WORK(drm_client_sysrq_restore_work, drm_client_sysrq_restore_work_fn);
+
+static void drm_client_sysrq_restore_handler(u8 ignored)
+{
+	schedule_work(&drm_client_sysrq_restore_work);
+}
+
+static const struct sysrq_key_op drm_client_sysrq_restore_op = {
+	.handler = drm_client_sysrq_restore_handler,
+	.help_msg = "force-fb(v)",
+	.action_msg = "Restore framebuffer console",
+};
+
+void drm_client_sysrq_register(struct drm_device *dev)
+{
+	guard(mutex)(&drm_client_sysrq_dev_lock);
+
+	if (list_empty(&drm_client_sysrq_dev_list))
+		register_sysrq_key('v', &drm_client_sysrq_restore_op);
+
+	list_add(&dev->client_sysrq_list, &drm_client_sysrq_dev_list);
+}
+
+void drm_client_sysrq_unregister(struct drm_device *dev)
+{
+	guard(mutex)(&drm_client_sysrq_dev_lock);
+
+	/* remove device from global restore list */
+	if (!drm_WARN_ON(dev, list_empty(&dev->client_sysrq_list)))
+		list_del(&dev->client_sysrq_list);
+
+	/* no devices left; unregister key */
+	if (list_empty(&drm_client_sysrq_dev_list))
+		unregister_sysrq_key('v', &drm_client_sysrq_restore_op);
+}
+#endif
diff --git a/drivers/gpu/drm/drm_drv.c b/drivers/gpu/drm/drm_drv.c
index 8e3cb08241c8..2915118436ce 100644
--- a/drivers/gpu/drm/drm_drv.c
+++ b/drivers/gpu/drm/drm_drv.c
@@ -733,6 +733,7 @@ static int drm_dev_init(struct drm_device *dev,
 	INIT_LIST_HEAD(&dev->filelist);
 	INIT_LIST_HEAD(&dev->filelist_internal);
 	INIT_LIST_HEAD(&dev->clientlist);
+	INIT_LIST_HEAD(&dev->client_sysrq_list);
 	INIT_LIST_HEAD(&dev->vblank_event_list);
 
 	spin_lock_init(&dev->event_lock);
@@ -1100,6 +1101,7 @@ int drm_dev_register(struct drm_device *dev, unsigned long flags)
 			goto err_unload;
 	}
 	drm_panic_register(dev);
+	drm_client_sysrq_register(dev);
 
 	DRM_INFO("Initialized %s %d.%d.%d for %s on minor %d\n",
 		 driver->name, driver->major, driver->minor,
@@ -1144,6 +1146,7 @@ void drm_dev_unregister(struct drm_device *dev)
 {
 	dev->registered = false;
 
+	drm_client_sysrq_unregister(dev);
 	drm_panic_unregister(dev);
 
 	drm_client_dev_unregister(dev);
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 1392738ce2fe..9a734017756b 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -32,7 +32,6 @@
 #include <linux/console.h>
 #include <linux/export.h>
 #include <linux/pci.h>
-#include <linux/sysrq.h>
 #include <linux/vga_switcheroo.h>
 
 #include <drm/drm_atomic.h>
@@ -270,42 +269,6 @@ int drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper, b
 }
 EXPORT_SYMBOL(drm_fb_helper_restore_fbdev_mode_unlocked);
 
-#ifdef CONFIG_MAGIC_SYSRQ
-/* emergency restore, don't bother with error reporting */
-static void drm_fb_helper_restore_work_fn(struct work_struct *ignored)
-{
-	struct drm_fb_helper *helper;
-
-	mutex_lock(&kernel_fb_helper_lock);
-	list_for_each_entry(helper, &kernel_fb_helper_list, kernel_fb_list) {
-		struct drm_device *dev = helper->dev;
-
-		if (dev->switch_power_state == DRM_SWITCH_POWER_OFF)
-			continue;
-
-		mutex_lock(&helper->lock);
-		drm_client_modeset_commit_locked(&helper->client);
-		mutex_unlock(&helper->lock);
-	}
-	mutex_unlock(&kernel_fb_helper_lock);
-}
-
-static DECLARE_WORK(drm_fb_helper_restore_work, drm_fb_helper_restore_work_fn);
-
-static void drm_fb_helper_sysrq(u8 dummy1)
-{
-	schedule_work(&drm_fb_helper_restore_work);
-}
-
-static const struct sysrq_key_op sysrq_drm_fb_helper_restore_op = {
-	.handler = drm_fb_helper_sysrq,
-	.help_msg = "force-fb(v)",
-	.action_msg = "Restore framebuffer console",
-};
-#else
-static const struct sysrq_key_op sysrq_drm_fb_helper_restore_op = { };
-#endif
-
 static void drm_fb_helper_dpms(struct fb_info *info, int dpms_mode)
 {
 	struct drm_fb_helper *fb_helper = info->par;
@@ -602,11 +565,8 @@ void drm_fb_helper_fini(struct drm_fb_helper *fb_helper)
 	drm_fb_helper_release_info(fb_helper);
 
 	mutex_lock(&kernel_fb_helper_lock);
-	if (!list_empty(&fb_helper->kernel_fb_list)) {
+	if (!list_empty(&fb_helper->kernel_fb_list))
 		list_del(&fb_helper->kernel_fb_list);
-		if (list_empty(&kernel_fb_helper_list))
-			unregister_sysrq_key('v', &sysrq_drm_fb_helper_restore_op);
-	}
 	mutex_unlock(&kernel_fb_helper_lock);
 
 	if (!fb_helper->client.funcs)
@@ -1840,9 +1800,6 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper)
 		 info->node, info->fix.id);
 
 	mutex_lock(&kernel_fb_helper_lock);
-	if (list_empty(&kernel_fb_helper_list))
-		register_sysrq_key('v', &sysrq_drm_fb_helper_restore_op);
-
 	list_add(&fb_helper->kernel_fb_list, &kernel_fb_helper_list);
 	mutex_unlock(&kernel_fb_helper_lock);
 
diff --git a/drivers/gpu/drm/drm_internal.h b/drivers/gpu/drm/drm_internal.h
index 5a3bed48ab1f..f893b1e3a596 100644
--- a/drivers/gpu/drm/drm_internal.h
+++ b/drivers/gpu/drm/drm_internal.h
@@ -56,6 +56,17 @@ static inline void drm_client_debugfs_init(struct drm_device *dev)
 { }
 #endif
 
+/* drm_client_sysrq.c */
+#if defined(CONFIG_DRM_CLIENT) && defined(CONFIG_MAGIC_SYSRQ)
+void drm_client_sysrq_register(struct drm_device *dev);
+void drm_client_sysrq_unregister(struct drm_device *dev);
+#else
+static inline void drm_client_sysrq_register(struct drm_device *dev)
+{ }
+static inline void drm_client_sysrq_unregister(struct drm_device *dev)
+{ }
+#endif
+
 /* drm_file.c */
 extern struct mutex drm_global_mutex;
 bool drm_dev_needs_global_mutex(struct drm_device *dev);
diff --git a/include/drm/drm_device.h b/include/drm/drm_device.h
index 778b2cca6c49..5af49c5c3778 100644
--- a/include/drm/drm_device.h
+++ b/include/drm/drm_device.h
@@ -238,6 +238,14 @@ struct drm_device {
 	 */
 	struct list_head clientlist;
 
+	/**
+	 * @client_sysrq_list:
+	 *
+	 * Entry into list of devices registered for sysrq. Allows in-kernel
+	 * clients on this device to handle sysrq keys.
+	 */
+	struct list_head client_sysrq_list;
+
 	/**
 	 * @vblank_disable_immediate:
 	 *
-- 
cgit v1.2.3


From 63c971af40365ee706c7e24f6a7900d693518f09 Mon Sep 17 00:00:00 2001
From: Thomas Zimmermann <tzimmermann@suse.de>
Date: Mon, 27 Oct 2025 09:12:17 +0100
Subject: drm/fb-helper: Allocate and release fb_info in single place
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Move the calls to drm_fb_helper_alloc_info() from drivers into a
single place in fbdev helpers. Allocates struct fb_info for a new
framebuffer device. Then call drm_fb_helper_single_fb_probe() to
create an fbdev screen buffer. Also release the instance on errors
by calling drm_fb_helper_release_info().

Simplifies the code and fixes the error cleanup for some of the
drivers.

Regular release of the struct fb_info instance still happens in
drm_fb_helper_fini() as before.

v2:
- remove error rollback in driver implementations (kernel test robot)
- initialize info in TTM implementation (kernel test robot)

Signed-off-by: Thomas Zimmermann <tzimmermann@suse.de>
Acked-by: Christian König <christian.koenig@amd.com> # radeon
Acked-by: Dmitry Baryshkov <dmitry.baryshkov@oss.qualcomm.com> # msm
Acked-by: Javier Martinez Canillas <javierm@redhat.com>
Link: https://patch.msgid.link/20251027081245.80262-1-tzimmermann@suse.de
---
 drivers/gpu/drm/armada/armada_fbdev.c      | 12 +--------
 drivers/gpu/drm/drm_fb_helper.c            | 39 +++++++++---------------------
 drivers/gpu/drm/drm_fbdev_dma.c            | 12 ++-------
 drivers/gpu/drm/drm_fbdev_shmem.c          | 12 ++-------
 drivers/gpu/drm/drm_fbdev_ttm.c            | 12 ++-------
 drivers/gpu/drm/exynos/exynos_drm_fbdev.c  |  9 +------
 drivers/gpu/drm/gma500/fbdev.c             | 12 +--------
 drivers/gpu/drm/i915/display/intel_fbdev.c |  9 +------
 drivers/gpu/drm/msm/msm_fbdev.c            |  9 +------
 drivers/gpu/drm/omapdrm/omap_fbdev.c       |  9 +------
 drivers/gpu/drm/radeon/radeon_fbdev.c      | 13 +---------
 drivers/gpu/drm/tegra/fbdev.c              |  9 +------
 include/drm/drm_fb_helper.h                | 12 ---------
 13 files changed, 26 insertions(+), 143 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/armada/armada_fbdev.c b/drivers/gpu/drm/armada/armada_fbdev.c
index be703d35f6b7..8bbae94804f8 100644
--- a/drivers/gpu/drm/armada/armada_fbdev.c
+++ b/drivers/gpu/drm/armada/armada_fbdev.c
@@ -44,10 +44,10 @@ int armada_fbdev_driver_fbdev_probe(struct drm_fb_helper *fbh,
 				    struct drm_fb_helper_surface_size *sizes)
 {
 	struct drm_device *dev = fbh->dev;
+	struct fb_info *info = fbh->info;
 	struct drm_mode_fb_cmd2 mode;
 	struct armada_framebuffer *dfb;
 	struct armada_gem_object *obj;
-	struct fb_info *info;
 	int size, ret;
 	void *ptr;
 
@@ -91,12 +91,6 @@ int armada_fbdev_driver_fbdev_probe(struct drm_fb_helper *fbh,
 	if (IS_ERR(dfb))
 		return PTR_ERR(dfb);
 
-	info = drm_fb_helper_alloc_info(fbh);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto err_fballoc;
-	}
-
 	info->fbops = &armada_fb_ops;
 	info->fix.smem_start = obj->phys_addr;
 	info->fix.smem_len = obj->obj.size;
@@ -112,8 +106,4 @@ int armada_fbdev_driver_fbdev_probe(struct drm_fb_helper *fbh,
 		(unsigned long long)obj->phys_addr);
 
 	return 0;
-
- err_fballoc:
-	dfb->fb.funcs->destroy(&dfb->fb);
-	return ret;
 }
diff --git a/drivers/gpu/drm/drm_fb_helper.c b/drivers/gpu/drm/drm_fb_helper.c
index 9a734017756b..be790fc68707 100644
--- a/drivers/gpu/drm/drm_fb_helper.c
+++ b/drivers/gpu/drm/drm_fb_helper.c
@@ -459,20 +459,7 @@ int drm_fb_helper_init(struct drm_device *dev,
 }
 EXPORT_SYMBOL(drm_fb_helper_init);
 
-/**
- * drm_fb_helper_alloc_info - allocate fb_info and some of its members
- * @fb_helper: driver-allocated fbdev helper
- *
- * A helper to alloc fb_info and the member cmap. Called by the driver
- * within the struct &drm_driver.fbdev_probe callback function. Drivers do
- * not need to release the allocated fb_info structure themselves, this is
- * automatically done when calling drm_fb_helper_fini().
- *
- * RETURNS:
- * fb_info pointer if things went okay, pointer containing error code
- * otherwise
- */
-struct fb_info *drm_fb_helper_alloc_info(struct drm_fb_helper *fb_helper)
+static struct fb_info *drm_fb_helper_alloc_info(struct drm_fb_helper *fb_helper)
 {
 	struct device *dev = fb_helper->dev->dev;
 	struct fb_info *info;
@@ -499,17 +486,8 @@ err_release:
 	framebuffer_release(info);
 	return ERR_PTR(ret);
 }
-EXPORT_SYMBOL(drm_fb_helper_alloc_info);
 
-/**
- * drm_fb_helper_release_info - release fb_info and its members
- * @fb_helper: driver-allocated fbdev helper
- *
- * A helper to release fb_info and the member cmap.  Drivers do not
- * need to release the allocated fb_info structure themselves, this is
- * automatically done when calling drm_fb_helper_fini().
- */
-void drm_fb_helper_release_info(struct drm_fb_helper *fb_helper)
+static void drm_fb_helper_release_info(struct drm_fb_helper *fb_helper)
 {
 	struct fb_info *info = fb_helper->info;
 
@@ -522,7 +500,6 @@ void drm_fb_helper_release_info(struct drm_fb_helper *fb_helper)
 		fb_dealloc_cmap(&info->cmap);
 	framebuffer_release(info);
 }
-EXPORT_SYMBOL(drm_fb_helper_release_info);
 
 /**
  * drm_fb_helper_unregister_info - unregister fb_info framebuffer device
@@ -1770,6 +1747,11 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper)
 	height = dev->mode_config.max_height;
 
 	drm_client_modeset_probe(&fb_helper->client, width, height);
+
+	info = drm_fb_helper_alloc_info(fb_helper);
+	if (IS_ERR(info))
+		return PTR_ERR(info);
+
 	ret = drm_fb_helper_single_fb_probe(fb_helper);
 	if (ret < 0) {
 		if (ret == -EAGAIN) {
@@ -1778,13 +1760,12 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper)
 		}
 		mutex_unlock(&fb_helper->lock);
 
-		return ret;
+		goto err_drm_fb_helper_release_info;
 	}
 	drm_setup_crtcs_fb(fb_helper);
 
 	fb_helper->deferred_setup = false;
 
-	info = fb_helper->info;
 	info->var.pixclock = 0;
 
 	/* Need to drop locks to avoid recursive deadlock in
@@ -1804,6 +1785,10 @@ __drm_fb_helper_initial_config_and_unlock(struct drm_fb_helper *fb_helper)
 	mutex_unlock(&kernel_fb_helper_lock);
 
 	return 0;
+
+err_drm_fb_helper_release_info:
+	drm_fb_helper_release_info(fb_helper);
+	return ret;
 }
 
 /**
diff --git a/drivers/gpu/drm/drm_fbdev_dma.c b/drivers/gpu/drm/drm_fbdev_dma.c
index 12a8f5a5ada5..9412d9fdd74b 100644
--- a/drivers/gpu/drm/drm_fbdev_dma.c
+++ b/drivers/gpu/drm/drm_fbdev_dma.c
@@ -269,9 +269,9 @@ int drm_fbdev_dma_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 {
 	struct drm_client_dev *client = &fb_helper->client;
 	struct drm_device *dev = fb_helper->dev;
+	struct fb_info *info = fb_helper->info;
 	struct drm_client_buffer *buffer;
 	struct drm_framebuffer *fb;
-	struct fb_info *info;
 	u32 format;
 	struct iosys_map map;
 	int ret;
@@ -301,12 +301,6 @@ int drm_fbdev_dma_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	fb_helper->buffer = buffer;
 	fb_helper->fb = fb;
 
-	info = drm_fb_helper_alloc_info(fb_helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto err_drm_client_buffer_vunmap;
-	}
-
 	drm_fb_helper_fill_info(info, fb_helper, sizes);
 
 	if (fb->funcs->dirty)
@@ -314,12 +308,10 @@ int drm_fbdev_dma_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	else
 		ret = drm_fbdev_dma_driver_fbdev_probe_tail(fb_helper, sizes);
 	if (ret)
-		goto err_drm_fb_helper_release_info;
+		goto err_drm_client_buffer_vunmap;
 
 	return 0;
 
-err_drm_fb_helper_release_info:
-	drm_fb_helper_release_info(fb_helper);
 err_drm_client_buffer_vunmap:
 	fb_helper->fb = NULL;
 	fb_helper->buffer = NULL;
diff --git a/drivers/gpu/drm/drm_fbdev_shmem.c b/drivers/gpu/drm/drm_fbdev_shmem.c
index ac2b22e05cd6..458c899b5d4f 100644
--- a/drivers/gpu/drm/drm_fbdev_shmem.c
+++ b/drivers/gpu/drm/drm_fbdev_shmem.c
@@ -135,10 +135,10 @@ int drm_fbdev_shmem_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 {
 	struct drm_client_dev *client = &fb_helper->client;
 	struct drm_device *dev = fb_helper->dev;
+	struct fb_info *info = fb_helper->info;
 	struct drm_client_buffer *buffer;
 	struct drm_gem_shmem_object *shmem;
 	struct drm_framebuffer *fb;
-	struct fb_info *info;
 	u32 format;
 	struct iosys_map map;
 	int ret;
@@ -168,12 +168,6 @@ int drm_fbdev_shmem_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	fb_helper->buffer = buffer;
 	fb_helper->fb = fb;
 
-	info = drm_fb_helper_alloc_info(fb_helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto err_drm_client_buffer_vunmap;
-	}
-
 	drm_fb_helper_fill_info(info, fb_helper, sizes);
 
 	info->fbops = &drm_fbdev_shmem_fb_ops;
@@ -194,12 +188,10 @@ int drm_fbdev_shmem_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	info->fbdefio = &fb_helper->fbdefio;
 	ret = fb_deferred_io_init(info);
 	if (ret)
-		goto err_drm_fb_helper_release_info;
+		goto err_drm_client_buffer_vunmap;
 
 	return 0;
 
-err_drm_fb_helper_release_info:
-	drm_fb_helper_release_info(fb_helper);
 err_drm_client_buffer_vunmap:
 	fb_helper->fb = NULL;
 	fb_helper->buffer = NULL;
diff --git a/drivers/gpu/drm/drm_fbdev_ttm.c b/drivers/gpu/drm/drm_fbdev_ttm.c
index c7ad779ba590..160bc35d8738 100644
--- a/drivers/gpu/drm/drm_fbdev_ttm.c
+++ b/drivers/gpu/drm/drm_fbdev_ttm.c
@@ -174,8 +174,8 @@ int drm_fbdev_ttm_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 {
 	struct drm_client_dev *client = &fb_helper->client;
 	struct drm_device *dev = fb_helper->dev;
+	struct fb_info *info = fb_helper->info;
 	struct drm_client_buffer *buffer;
-	struct fb_info *info;
 	size_t screen_size;
 	void *screen_buffer;
 	u32 format;
@@ -203,12 +203,6 @@ int drm_fbdev_ttm_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 		goto err_drm_client_buffer_delete;
 	}
 
-	info = drm_fb_helper_alloc_info(fb_helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto err_vfree;
-	}
-
 	drm_fb_helper_fill_info(info, fb_helper, sizes);
 
 	info->fbops = &drm_fbdev_ttm_fb_ops;
@@ -225,12 +219,10 @@ int drm_fbdev_ttm_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	info->fbdefio = &fb_helper->fbdefio;
 	ret = fb_deferred_io_init(info);
 	if (ret)
-		goto err_drm_fb_helper_release_info;
+		goto err_vfree;
 
 	return 0;
 
-err_drm_fb_helper_release_info:
-	drm_fb_helper_release_info(fb_helper);
 err_vfree:
 	vfree(screen_buffer);
 err_drm_client_buffer_delete:
diff --git a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
index a9d35e8fca6a..637927818dfe 100644
--- a/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
+++ b/drivers/gpu/drm/exynos/exynos_drm_fbdev.c
@@ -58,18 +58,11 @@ static int exynos_drm_fbdev_update(struct drm_fb_helper *helper,
 				   struct drm_fb_helper_surface_size *sizes,
 				   struct exynos_drm_gem *exynos_gem)
 {
-	struct fb_info *fbi;
+	struct fb_info *fbi = helper->info;
 	struct drm_framebuffer *fb = helper->fb;
 	unsigned int size = fb->width * fb->height * fb->format->cpp[0];
 	unsigned long offset;
 
-	fbi = drm_fb_helper_alloc_info(helper);
-	if (IS_ERR(fbi)) {
-		DRM_DEV_ERROR(to_dma_dev(helper->dev),
-			      "failed to allocate fb info.\n");
-		return PTR_ERR(fbi);
-	}
-
 	fbi->fbops = &exynos_drm_fb_ops;
 
 	drm_fb_helper_fill_info(fbi, helper, sizes);
diff --git a/drivers/gpu/drm/gma500/fbdev.c b/drivers/gpu/drm/gma500/fbdev.c
index bc92fa24a1e2..c26926babc2a 100644
--- a/drivers/gpu/drm/gma500/fbdev.c
+++ b/drivers/gpu/drm/gma500/fbdev.c
@@ -108,7 +108,7 @@ int psb_fbdev_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	struct drm_device *dev = fb_helper->dev;
 	struct drm_psb_private *dev_priv = to_drm_psb_private(dev);
 	struct pci_dev *pdev = to_pci_dev(dev->dev);
-	struct fb_info *info;
+	struct fb_info *info = fb_helper->info;
 	struct drm_framebuffer *fb;
 	struct drm_mode_fb_cmd2 mode_cmd = { };
 	int size;
@@ -167,12 +167,6 @@ int psb_fbdev_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	fb_helper->funcs = &psb_fbdev_fb_helper_funcs;
 	fb_helper->fb = fb;
 
-	info = drm_fb_helper_alloc_info(fb_helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto err_drm_framebuffer_unregister_private;
-	}
-
 	info->fbops = &psb_fbdev_fb_ops;
 
 	/* Accessed stolen memory directly */
@@ -196,10 +190,6 @@ int psb_fbdev_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 
 	return 0;
 
-err_drm_framebuffer_unregister_private:
-	drm_framebuffer_unregister_private(fb);
-	drm_framebuffer_cleanup(fb);
-	kfree(fb);
 err_drm_gem_object_put:
 	drm_gem_object_put(obj);
 	return ret;
diff --git a/drivers/gpu/drm/i915/display/intel_fbdev.c b/drivers/gpu/drm/i915/display/intel_fbdev.c
index e5449c41cfa1..9cd03e2adeb2 100644
--- a/drivers/gpu/drm/i915/display/intel_fbdev.c
+++ b/drivers/gpu/drm/i915/display/intel_fbdev.c
@@ -267,8 +267,8 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 	struct intel_display *display = to_intel_display(helper->dev);
 	struct intel_fbdev *ifbdev = to_intel_fbdev(helper);
 	struct intel_framebuffer *fb = ifbdev->fb;
+	struct fb_info *info = helper->info;
 	struct ref_tracker *wakeref;
-	struct fb_info *info;
 	struct i915_vma *vma;
 	unsigned long flags = 0;
 	bool prealloc = false;
@@ -318,13 +318,6 @@ int intel_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 		goto out_unlock;
 	}
 
-	info = drm_fb_helper_alloc_info(helper);
-	if (IS_ERR(info)) {
-		drm_err(display->drm, "Failed to allocate fb_info (%pe)\n", info);
-		ret = PTR_ERR(info);
-		goto out_unpin;
-	}
-
 	helper->funcs = &intel_fb_helper_funcs;
 	helper->fb = &fb->base;
 
diff --git a/drivers/gpu/drm/msm/msm_fbdev.c b/drivers/gpu/drm/msm/msm_fbdev.c
index aad6fb77f0de..fd19995b12b5 100644
--- a/drivers/gpu/drm/msm/msm_fbdev.c
+++ b/drivers/gpu/drm/msm/msm_fbdev.c
@@ -91,9 +91,9 @@ int msm_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 {
 	struct drm_device *dev = helper->dev;
 	struct msm_drm_private *priv = dev->dev_private;
+	struct fb_info *fbi = helper->info;
 	struct drm_framebuffer *fb = NULL;
 	struct drm_gem_object *bo;
-	struct fb_info *fbi = NULL;
 	uint64_t paddr;
 	uint32_t format;
 	int ret, pitch;
@@ -126,13 +126,6 @@ int msm_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 		goto fail;
 	}
 
-	fbi = drm_fb_helper_alloc_info(helper);
-	if (IS_ERR(fbi)) {
-		DRM_DEV_ERROR(dev->dev, "failed to allocate fb info\n");
-		ret = PTR_ERR(fbi);
-		goto fail;
-	}
-
 	DBG("fbi=%p, dev=%p", fbi, dev);
 
 	helper->funcs = &msm_fbdev_helper_funcs;
diff --git a/drivers/gpu/drm/omapdrm/omap_fbdev.c b/drivers/gpu/drm/omapdrm/omap_fbdev.c
index d89761f13cd7..ca3fb186bf19 100644
--- a/drivers/gpu/drm/omapdrm/omap_fbdev.c
+++ b/drivers/gpu/drm/omapdrm/omap_fbdev.c
@@ -154,9 +154,9 @@ int omap_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 	struct drm_device *dev = helper->dev;
 	struct omap_drm_private *priv = dev->dev_private;
 	struct omap_fbdev *fbdev = priv->fbdev;
+	struct fb_info *fbi = helper->info;
 	struct drm_framebuffer *fb = NULL;
 	union omap_gem_size gsize;
-	struct fb_info *fbi = NULL;
 	struct drm_mode_fb_cmd2 mode_cmd = {0};
 	struct drm_gem_object *bo;
 	dma_addr_t dma_addr;
@@ -225,13 +225,6 @@ int omap_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 		goto fail;
 	}
 
-	fbi = drm_fb_helper_alloc_info(helper);
-	if (IS_ERR(fbi)) {
-		dev_err(dev->dev, "failed to allocate fb info\n");
-		ret = PTR_ERR(fbi);
-		goto fail;
-	}
-
 	DBG("fbi=%p, dev=%p", fbi, dev);
 
 	helper->funcs = &omap_fbdev_helper_funcs;
diff --git a/drivers/gpu/drm/radeon/radeon_fbdev.c b/drivers/gpu/drm/radeon/radeon_fbdev.c
index c2cfe2d7915f..fd083aaa91bb 100644
--- a/drivers/gpu/drm/radeon/radeon_fbdev.c
+++ b/drivers/gpu/drm/radeon/radeon_fbdev.c
@@ -202,7 +202,7 @@ int radeon_fbdev_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	struct radeon_device *rdev = fb_helper->dev->dev_private;
 	const struct drm_format_info *format_info;
 	struct drm_mode_fb_cmd2 mode_cmd = { };
-	struct fb_info *info;
+	struct fb_info *info = fb_helper->info;
 	struct drm_gem_object *gobj;
 	struct radeon_bo *rbo;
 	struct drm_framebuffer *fb;
@@ -243,13 +243,6 @@ int radeon_fbdev_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 	fb_helper->funcs = &radeon_fbdev_fb_helper_funcs;
 	fb_helper->fb = fb;
 
-	/* okay we have an object now allocate the framebuffer */
-	info = drm_fb_helper_alloc_info(fb_helper);
-	if (IS_ERR(info)) {
-		ret = PTR_ERR(info);
-		goto err_drm_framebuffer_unregister_private;
-	}
-
 	info->fbops = &radeon_fbdev_fb_ops;
 
 	/* radeon resume is fragile and needs a vt switch to help it along */
@@ -275,10 +268,6 @@ int radeon_fbdev_driver_fbdev_probe(struct drm_fb_helper *fb_helper,
 
 	return 0;
 
-err_drm_framebuffer_unregister_private:
-	fb_helper->fb = NULL;
-	drm_framebuffer_unregister_private(fb);
-	drm_framebuffer_cleanup(fb);
 err_kfree:
 	kfree(fb);
 err_radeon_fbdev_destroy_pinned_object:
diff --git a/drivers/gpu/drm/tegra/fbdev.c b/drivers/gpu/drm/tegra/fbdev.c
index 91aece6f34e0..8f40882aa76e 100644
--- a/drivers/gpu/drm/tegra/fbdev.c
+++ b/drivers/gpu/drm/tegra/fbdev.c
@@ -73,10 +73,10 @@ int tegra_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 	struct tegra_drm *tegra = helper->dev->dev_private;
 	struct drm_device *drm = helper->dev;
 	struct drm_mode_fb_cmd2 cmd = { 0 };
+	struct fb_info *info = helper->info;
 	unsigned int bytes_per_pixel;
 	struct drm_framebuffer *fb;
 	unsigned long offset;
-	struct fb_info *info;
 	struct tegra_bo *bo;
 	size_t size;
 	int err;
@@ -97,13 +97,6 @@ int tegra_fbdev_driver_fbdev_probe(struct drm_fb_helper *helper,
 	if (IS_ERR(bo))
 		return PTR_ERR(bo);
 
-	info = drm_fb_helper_alloc_info(helper);
-	if (IS_ERR(info)) {
-		dev_err(drm->dev, "failed to allocate framebuffer info\n");
-		drm_gem_object_put(&bo->gem);
-		return PTR_ERR(info);
-	}
-
 	fb = tegra_fb_alloc(drm,
 			    drm_get_format_info(drm, cmd.pixel_format, cmd.modifier[0]),
 			    &cmd, &bo, 1);
diff --git a/include/drm/drm_fb_helper.h b/include/drm/drm_fb_helper.h
index 63e3af8dd5ed..dd9a18f8de5a 100644
--- a/include/drm/drm_fb_helper.h
+++ b/include/drm/drm_fb_helper.h
@@ -257,8 +257,6 @@ int drm_fb_helper_check_var(struct fb_var_screeninfo *var,
 int drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper,
 					      bool force);
 
-struct fb_info *drm_fb_helper_alloc_info(struct drm_fb_helper *fb_helper);
-void drm_fb_helper_release_info(struct drm_fb_helper *fb_helper);
 void drm_fb_helper_unregister_info(struct drm_fb_helper *fb_helper);
 void drm_fb_helper_fill_info(struct fb_info *info,
 			     struct drm_fb_helper *fb_helper,
@@ -340,16 +338,6 @@ drm_fb_helper_restore_fbdev_mode_unlocked(struct drm_fb_helper *fb_helper)
 	return 0;
 }
 
-static inline struct fb_info *
-drm_fb_helper_alloc_info(struct drm_fb_helper *fb_helper)
-{
-	return NULL;
-}
-
-static inline void drm_fb_helper_release_info(struct drm_fb_helper *fb_helper)
-{
-}
-
 static inline void drm_fb_helper_unregister_info(struct drm_fb_helper *fb_helper)
 {
 }
-- 
cgit v1.2.3


From ddf055b80a544d6f36f77be5f0c6d3c80177d57c Mon Sep 17 00:00:00 2001
From: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Date: Fri, 21 Nov 2025 11:12:29 +0100
Subject: drm/ttm: rework pipelined eviction fence handling
MIME-Version: 1.0
Content-Type: text/plain; charset=UTF-8
Content-Transfer-Encoding: 8bit

Until now ttm stored a single pipelined eviction fence which means
drivers had to use a single entity for these evictions.

To lift this requirement, this commit allows up to 8 entities to
be used.

Ideally a dma_resv object would have been used as a container of
the eviction fences, but the locking rules makes it complex.
dma_resv all have the same ww_class, which means "Attempting to
lock more mutexes after ww_acquire_done." is an error.

One alternative considered was to introduced a 2nd ww_class for
specific resv to hold a single "transient" lock (= the resv lock
would only be held for a short period, without taking any other
locks).

The other option, is to statically reserve a fence array, and
extend the existing code to deal with N fences, instead of 1.

The driver is still responsible to reserve the correct number
of fence slots.

Signed-off-by: Pierre-Eric Pelloux-Prayer <pierre-eric.pelloux-prayer@amd.com>
Link: https://lore.kernel.org/r/20251121101315.3585-20-pierre-eric.pelloux-prayer@amd.com
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Christian König <christian.koenig@amd.com>
---
 drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c | 11 +++---
 drivers/gpu/drm/ttm/tests/ttm_resource_test.c    |  5 ++-
 drivers/gpu/drm/ttm/ttm_bo.c                     | 47 ++++++++++++------------
 drivers/gpu/drm/ttm/ttm_bo_util.c                | 38 +++++++++++++++----
 drivers/gpu/drm/ttm/ttm_resource.c               | 33 ++++++++++-------
 include/drm/ttm/ttm_resource.h                   | 29 +++++++++++----
 6 files changed, 105 insertions(+), 58 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
index 17a570af296c..2eda87882e65 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_bo_validate_test.c
@@ -652,7 +652,7 @@ static void ttm_bo_validate_move_fence_signaled(struct kunit *test)
 	int err;
 
 	man = ttm_manager_type(priv->ttm_dev, mem_type);
-	man->move = dma_fence_get_stub();
+	man->eviction_fences[0] = dma_fence_get_stub();
 
 	bo = ttm_bo_kunit_init(test, test->priv, size, NULL);
 	bo->type = bo_type;
@@ -669,7 +669,7 @@ static void ttm_bo_validate_move_fence_signaled(struct kunit *test)
 	KUNIT_EXPECT_EQ(test, ctx.bytes_moved, size);
 
 	ttm_bo_fini(bo);
-	dma_fence_put(man->move);
+	dma_fence_put(man->eviction_fences[0]);
 }
 
 static const struct ttm_bo_validate_test_case ttm_bo_validate_wait_cases[] = {
@@ -733,9 +733,9 @@ static void ttm_bo_validate_move_fence_not_signaled(struct kunit *test)
 
 	spin_lock_init(&fence_lock);
 	man = ttm_manager_type(priv->ttm_dev, fst_mem);
-	man->move = alloc_mock_fence(test);
+	man->eviction_fences[0] = alloc_mock_fence(test);
 
-	task = kthread_create(threaded_fence_signal, man->move, "move-fence-signal");
+	task = kthread_create(threaded_fence_signal, man->eviction_fences[0], "move-fence-signal");
 	if (IS_ERR(task))
 		KUNIT_FAIL(test, "Couldn't create move fence signal task\n");
 
@@ -743,7 +743,8 @@ static void ttm_bo_validate_move_fence_not_signaled(struct kunit *test)
 	err = ttm_bo_validate(bo, placement_val, &ctx_val);
 	dma_resv_unlock(bo->base.resv);
 
-	dma_fence_wait_timeout(man->move, false, MAX_SCHEDULE_TIMEOUT);
+	dma_fence_wait_timeout(man->eviction_fences[0], false, MAX_SCHEDULE_TIMEOUT);
+	man->eviction_fences[0] = NULL;
 
 	KUNIT_EXPECT_EQ(test, err, 0);
 	KUNIT_EXPECT_EQ(test, ctx_val.bytes_moved, size);
diff --git a/drivers/gpu/drm/ttm/tests/ttm_resource_test.c b/drivers/gpu/drm/ttm/tests/ttm_resource_test.c
index e6ea2bd01f07..c0e4e35e0442 100644
--- a/drivers/gpu/drm/ttm/tests/ttm_resource_test.c
+++ b/drivers/gpu/drm/ttm/tests/ttm_resource_test.c
@@ -207,6 +207,7 @@ static void ttm_resource_manager_init_basic(struct kunit *test)
 	struct ttm_resource_test_priv *priv = test->priv;
 	struct ttm_resource_manager *man;
 	size_t size = SZ_16K;
+	int i;
 
 	man = kunit_kzalloc(test, sizeof(*man), GFP_KERNEL);
 	KUNIT_ASSERT_NOT_NULL(test, man);
@@ -216,8 +217,8 @@ static void ttm_resource_manager_init_basic(struct kunit *test)
 	KUNIT_ASSERT_PTR_EQ(test, man->bdev, priv->devs->ttm_dev);
 	KUNIT_ASSERT_EQ(test, man->size, size);
 	KUNIT_ASSERT_EQ(test, man->usage, 0);
-	KUNIT_ASSERT_NULL(test, man->move);
-	KUNIT_ASSERT_NOT_NULL(test, &man->move_lock);
+	for (i = 0; i < TTM_NUM_MOVE_FENCES; i++)
+		KUNIT_ASSERT_NULL(test, man->eviction_fences[i]);
 
 	for (int i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
 		KUNIT_ASSERT_TRUE(test, list_empty(&man->lru[i]));
diff --git a/drivers/gpu/drm/ttm/ttm_bo.c b/drivers/gpu/drm/ttm/ttm_bo.c
index c4e669686fd6..bd27607f8076 100644
--- a/drivers/gpu/drm/ttm/ttm_bo.c
+++ b/drivers/gpu/drm/ttm/ttm_bo.c
@@ -659,34 +659,35 @@ void ttm_bo_unpin(struct ttm_buffer_object *bo)
 EXPORT_SYMBOL(ttm_bo_unpin);
 
 /*
- * Add the last move fence to the BO as kernel dependency and reserve a new
- * fence slot.
+ * Add the pipelined eviction fencesto the BO as kernel dependency and reserve new
+ * fence slots.
  */
-static int ttm_bo_add_move_fence(struct ttm_buffer_object *bo,
-				 struct ttm_resource_manager *man,
-				 bool no_wait_gpu)
+static int ttm_bo_add_pipelined_eviction_fences(struct ttm_buffer_object *bo,
+						struct ttm_resource_manager *man,
+						bool no_wait_gpu)
 {
 	struct dma_fence *fence;
-	int ret;
-
-	spin_lock(&man->move_lock);
-	fence = dma_fence_get(man->move);
-	spin_unlock(&man->move_lock);
+	int i;
 
-	if (!fence)
-		return 0;
+	spin_lock(&man->eviction_lock);
+	for (i = 0; i < TTM_NUM_MOVE_FENCES; i++) {
+		fence = man->eviction_fences[i];
+		if (!fence)
+			continue;
 
-	if (no_wait_gpu) {
-		ret = dma_fence_is_signaled(fence) ? 0 : -EBUSY;
-		dma_fence_put(fence);
-		return ret;
+		if (no_wait_gpu) {
+			if (!dma_fence_is_signaled(fence)) {
+				spin_unlock(&man->eviction_lock);
+				return -EBUSY;
+			}
+		} else {
+			dma_resv_add_fence(bo->base.resv, fence, DMA_RESV_USAGE_KERNEL);
+		}
 	}
+	spin_unlock(&man->eviction_lock);
 
-	dma_resv_add_fence(bo->base.resv, fence, DMA_RESV_USAGE_KERNEL);
-
-	ret = dma_resv_reserve_fences(bo->base.resv, 1);
-	dma_fence_put(fence);
-	return ret;
+	/* TODO: this call should be removed. */
+	return dma_resv_reserve_fences(bo->base.resv, 1);
 }
 
 /**
@@ -719,7 +720,7 @@ static int ttm_bo_alloc_resource(struct ttm_buffer_object *bo,
 	int i, ret;
 
 	ticket = dma_resv_locking_ctx(bo->base.resv);
-	ret = dma_resv_reserve_fences(bo->base.resv, 1);
+	ret = dma_resv_reserve_fences(bo->base.resv, TTM_NUM_MOVE_FENCES);
 	if (unlikely(ret))
 		return ret;
 
@@ -758,7 +759,7 @@ static int ttm_bo_alloc_resource(struct ttm_buffer_object *bo,
 				return ret;
 		}
 
-		ret = ttm_bo_add_move_fence(bo, man, ctx->no_wait_gpu);
+		ret = ttm_bo_add_pipelined_eviction_fences(bo, man, ctx->no_wait_gpu);
 		if (unlikely(ret)) {
 			ttm_resource_free(bo, res);
 			if (ret == -EBUSY)
diff --git a/drivers/gpu/drm/ttm/ttm_bo_util.c b/drivers/gpu/drm/ttm/ttm_bo_util.c
index acbbca9d5c92..2ff35d55e462 100644
--- a/drivers/gpu/drm/ttm/ttm_bo_util.c
+++ b/drivers/gpu/drm/ttm/ttm_bo_util.c
@@ -258,7 +258,7 @@ static int ttm_buffer_object_transfer(struct ttm_buffer_object *bo,
 	ret = dma_resv_trylock(&fbo->base.base._resv);
 	WARN_ON(!ret);
 
-	ret = dma_resv_reserve_fences(&fbo->base.base._resv, 1);
+	ret = dma_resv_reserve_fences(&fbo->base.base._resv, TTM_NUM_MOVE_FENCES);
 	if (ret) {
 		dma_resv_unlock(&fbo->base.base._resv);
 		kfree(fbo);
@@ -646,20 +646,44 @@ static void ttm_bo_move_pipeline_evict(struct ttm_buffer_object *bo,
 {
 	struct ttm_device *bdev = bo->bdev;
 	struct ttm_resource_manager *from;
+	struct dma_fence *tmp;
+	int i;
 
 	from = ttm_manager_type(bdev, bo->resource->mem_type);
 
 	/**
 	 * BO doesn't have a TTM we need to bind/unbind. Just remember
-	 * this eviction and free up the allocation
+	 * this eviction and free up the allocation.
+	 * The fence will be saved in the first free slot or in the slot
+	 * already used to store a fence from the same context. Since
+	 * drivers can't use more than TTM_NUM_MOVE_FENCES contexts for
+	 * evictions we should always find a slot to use.
 	 */
-	spin_lock(&from->move_lock);
-	if (!from->move || dma_fence_is_later(fence, from->move)) {
-		dma_fence_put(from->move);
-		from->move = dma_fence_get(fence);
+	spin_lock(&from->eviction_lock);
+	for (i = 0; i < TTM_NUM_MOVE_FENCES; i++) {
+		tmp = from->eviction_fences[i];
+		if (!tmp)
+			break;
+		if (fence->context != tmp->context)
+			continue;
+		if (dma_fence_is_later(fence, tmp)) {
+			dma_fence_put(tmp);
+			break;
+		}
+		goto unlock;
+	}
+	if (i < TTM_NUM_MOVE_FENCES) {
+		from->eviction_fences[i] = dma_fence_get(fence);
+	} else {
+		WARN(1, "not enough fence slots for all fence contexts");
+		spin_unlock(&from->eviction_lock);
+		dma_fence_wait(fence, false);
+		goto end;
 	}
-	spin_unlock(&from->move_lock);
 
+unlock:
+	spin_unlock(&from->eviction_lock);
+end:
 	ttm_resource_free(bo, &bo->resource);
 }
 
diff --git a/drivers/gpu/drm/ttm/ttm_resource.c b/drivers/gpu/drm/ttm/ttm_resource.c
index 1a39c30f22fb..f5aa29dc6ec0 100644
--- a/drivers/gpu/drm/ttm/ttm_resource.c
+++ b/drivers/gpu/drm/ttm/ttm_resource.c
@@ -524,14 +524,15 @@ void ttm_resource_manager_init(struct ttm_resource_manager *man,
 {
 	unsigned i;
 
-	spin_lock_init(&man->move_lock);
 	man->bdev = bdev;
 	man->size = size;
 	man->usage = 0;
 
 	for (i = 0; i < TTM_MAX_BO_PRIORITY; ++i)
 		INIT_LIST_HEAD(&man->lru[i]);
-	man->move = NULL;
+	spin_lock_init(&man->eviction_lock);
+	for (i = 0; i < TTM_NUM_MOVE_FENCES; i++)
+		man->eviction_fences[i] = NULL;
 }
 EXPORT_SYMBOL(ttm_resource_manager_init);
 
@@ -552,7 +553,7 @@ int ttm_resource_manager_evict_all(struct ttm_device *bdev,
 		.no_wait_gpu = false,
 	};
 	struct dma_fence *fence;
-	int ret;
+	int ret, i;
 
 	do {
 		ret = ttm_bo_evict_first(bdev, man, &ctx);
@@ -562,18 +563,24 @@ int ttm_resource_manager_evict_all(struct ttm_device *bdev,
 	if (ret && ret != -ENOENT)
 		return ret;
 
-	spin_lock(&man->move_lock);
-	fence = dma_fence_get(man->move);
-	spin_unlock(&man->move_lock);
-
-	if (fence) {
-		ret = dma_fence_wait(fence, false);
-		dma_fence_put(fence);
-		if (ret)
-			return ret;
+	ret = 0;
+
+	spin_lock(&man->eviction_lock);
+	for (i = 0; i < TTM_NUM_MOVE_FENCES; i++) {
+		fence = man->eviction_fences[i];
+		if (fence && !dma_fence_is_signaled(fence)) {
+			dma_fence_get(fence);
+			spin_unlock(&man->eviction_lock);
+			ret = dma_fence_wait(fence, false);
+			dma_fence_put(fence);
+			if (ret)
+				return ret;
+			spin_lock(&man->eviction_lock);
+		}
 	}
+	spin_unlock(&man->eviction_lock);
 
-	return 0;
+	return ret;
 }
 EXPORT_SYMBOL(ttm_resource_manager_evict_all);
 
diff --git a/include/drm/ttm/ttm_resource.h b/include/drm/ttm/ttm_resource.h
index 68bf010d8b40..33e80f30b8b8 100644
--- a/include/drm/ttm/ttm_resource.h
+++ b/include/drm/ttm/ttm_resource.h
@@ -51,6 +51,15 @@ struct io_mapping;
 struct sg_table;
 struct scatterlist;
 
+/**
+ * define TTM_NUM_MOVE_FENCES - How many entities can be used for evictions
+ *
+ * Pipelined evictions can be spread on multiple entities. This
+ * is the max number of entities that can be used by the driver
+ * for that purpose.
+ */
+#define TTM_NUM_MOVE_FENCES 8
+
 /**
  * enum ttm_lru_item_type - enumerate ttm_lru_item subclasses
  */
@@ -181,8 +190,8 @@ struct ttm_resource_manager_func {
  * @size: Size of the managed region.
  * @bdev: ttm device this manager belongs to
  * @func: structure pointer implementing the range manager. See above
- * @move_lock: lock for move fence
- * @move: The fence of the last pipelined move operation.
+ * @eviction_lock: lock for eviction fences
+ * @eviction_fences: The fences of the last pipelined move operation.
  * @lru: The lru list for this memory type.
  *
  * This structure is used to identify and manage memory types for a device.
@@ -196,12 +205,12 @@ struct ttm_resource_manager {
 	struct ttm_device *bdev;
 	uint64_t size;
 	const struct ttm_resource_manager_func *func;
-	spinlock_t move_lock;
 
-	/*
-	 * Protected by @move_lock.
+	/* This is very similar to a dma_resv object, but locking rules make
+	 * it difficult to use one in this context.
 	 */
-	struct dma_fence *move;
+	spinlock_t eviction_lock;
+	struct dma_fence *eviction_fences[TTM_NUM_MOVE_FENCES];
 
 	/*
 	 * Protected by the bdev->lru_lock.
@@ -422,8 +431,12 @@ static inline bool ttm_resource_manager_used(struct ttm_resource_manager *man)
 static inline void
 ttm_resource_manager_cleanup(struct ttm_resource_manager *man)
 {
-	dma_fence_put(man->move);
-	man->move = NULL;
+	int i;
+
+	for (i = 0; i < TTM_NUM_MOVE_FENCES; i++) {
+		dma_fence_put(man->eviction_fences[i]);
+		man->eviction_fences[i] = NULL;
+	}
 }
 
 void ttm_lru_bulk_move_init(struct ttm_lru_bulk_move *bulk);
-- 
cgit v1.2.3


From 4fc183828b050c1d7d94347c876c04a5b680b4fc Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:26 -0700
Subject: drm: Add helper for conversion from signed-magnitude

CTM values are defined as signed-magnitude values. Add
a helper that converts from CTM signed-magnitude fixed
point value to the twos-complement value used by
drm_fixed.

Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-2-alex.hung@amd.com
---
 include/drm/drm_fixed.h | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

(limited to 'include')

diff --git a/include/drm/drm_fixed.h b/include/drm/drm_fixed.h
index 1922188f00e8..33de514a5221 100644
--- a/include/drm/drm_fixed.h
+++ b/include/drm/drm_fixed.h
@@ -78,6 +78,23 @@ static inline u32 dfixed_div(fixed20_12 A, fixed20_12 B)
 #define DRM_FIXED_EPSILON	1LL
 #define DRM_FIXED_ALMOST_ONE	(DRM_FIXED_ONE - DRM_FIXED_EPSILON)
 
+/**
+ * @drm_sm2fixp
+ *
+ * Convert a 1.31.32 signed-magnitude fixed point to 32.32
+ * 2s-complement fixed point
+ *
+ * @return s64 2s-complement fixed point
+ */
+static inline s64 drm_sm2fixp(__u64 a)
+{
+	if ((a & (1LL << 63))) {
+		return -(a & 0x7fffffffffffffffll);
+	} else {
+		return a;
+	}
+}
+
 static inline s64 drm_int2fixp(int a)
 {
 	return ((s64)a) << DRM_FIXED_POINT;
-- 
cgit v1.2.3


From cfc27680ee208cdf7a61cda817b4158c4142595f Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:29 -0700
Subject: drm/colorop: Introduce new drm_colorop mode object

This patches introduces a new drm_colorop mode object. This
object represents color transformations and can be used to
define color pipelines.

We also introduce the drm_colorop_state here, as well as
various helpers and state tracking bits.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-5-alex.hung@amd.com
---
 drivers/gpu/drm/Makefile            |   1 +
 drivers/gpu/drm/drm_atomic.c        |  69 +++++++++++++++
 drivers/gpu/drm/drm_atomic_helper.c |  12 +++
 drivers/gpu/drm/drm_atomic_uapi.c   |  45 ++++++++++
 drivers/gpu/drm/drm_colorop.c       | 103 ++++++++++++++++++++++
 drivers/gpu/drm/drm_mode_config.c   |   7 ++
 include/drm/drm_atomic.h            |  70 +++++++++++++++
 include/drm/drm_atomic_uapi.h       |   1 +
 include/drm/drm_colorop.h           | 169 ++++++++++++++++++++++++++++++++++++
 include/drm/drm_mode_config.h       |  18 ++++
 include/drm/drm_plane.h             |   8 ++
 include/uapi/drm/drm_mode.h         |   1 +
 12 files changed, 504 insertions(+)
 create mode 100644 drivers/gpu/drm/drm_colorop.c
 create mode 100644 include/drm/drm_colorop.h

(limited to 'include')

diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 9901534948e5..3d40442d1854 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -41,6 +41,7 @@ drm-y := \
 	drm_bridge.o \
 	drm_cache.o \
 	drm_color_mgmt.o \
+	drm_colorop.o \
 	drm_connector.o \
 	drm_crtc.o \
 	drm_displayid.o \
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index e05820b18832..6438a3938032 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -42,6 +42,7 @@
 #include <drm/drm_mode.h>
 #include <drm/drm_print.h>
 #include <drm/drm_writeback.h>
+#include <drm/drm_colorop.h>
 
 #include "drm_crtc_internal.h"
 #include "drm_internal.h"
@@ -107,6 +108,7 @@ void drm_atomic_state_default_release(struct drm_atomic_state *state)
 	kfree(state->connectors);
 	kfree(state->crtcs);
 	kfree(state->planes);
+	kfree(state->colorops);
 	kfree(state->private_objs);
 }
 EXPORT_SYMBOL(drm_atomic_state_default_release);
@@ -138,6 +140,10 @@ drm_atomic_state_init(struct drm_device *dev, struct drm_atomic_state *state)
 				sizeof(*state->planes), GFP_KERNEL);
 	if (!state->planes)
 		goto fail;
+	state->colorops = kcalloc(dev->mode_config.num_colorop,
+				  sizeof(*state->colorops), GFP_KERNEL);
+	if (!state->colorops)
+		goto fail;
 
 	/*
 	 * Because drm_atomic_state can be committed asynchronously we need our
@@ -251,6 +257,20 @@ void drm_atomic_state_default_clear(struct drm_atomic_state *state)
 		state->planes[i].new_state = NULL;
 	}
 
+	for (i = 0; i < config->num_colorop; i++) {
+		struct drm_colorop *colorop = state->colorops[i].ptr;
+
+		if (!colorop)
+			continue;
+
+		drm_colorop_atomic_destroy_state(colorop,
+						 state->colorops[i].state);
+		state->colorops[i].ptr = NULL;
+		state->colorops[i].state = NULL;
+		state->colorops[i].old_state = NULL;
+		state->colorops[i].new_state = NULL;
+	}
+
 	for (i = 0; i < state->num_private_objs; i++) {
 		struct drm_private_obj *obj = state->private_objs[i].ptr;
 
@@ -572,6 +592,55 @@ drm_atomic_get_plane_state(struct drm_atomic_state *state,
 }
 EXPORT_SYMBOL(drm_atomic_get_plane_state);
 
+/**
+ * drm_atomic_get_colorop_state - get colorop state
+ * @state: global atomic state object
+ * @colorop: colorop to get state object for
+ *
+ * This function returns the colorop state for the given colorop, allocating it
+ * if needed. It will also grab the relevant plane lock to make sure that the
+ * state is consistent.
+ *
+ * Returns:
+ *
+ * Either the allocated state or the error code encoded into the pointer. When
+ * the error is EDEADLK then the w/w mutex code has detected a deadlock and the
+ * entire atomic sequence must be restarted. All other errors are fatal.
+ */
+struct drm_colorop_state *
+drm_atomic_get_colorop_state(struct drm_atomic_state *state,
+			     struct drm_colorop *colorop)
+{
+	int ret, index = drm_colorop_index(colorop);
+	struct drm_colorop_state *colorop_state;
+
+	WARN_ON(!state->acquire_ctx);
+
+	colorop_state = drm_atomic_get_new_colorop_state(state, colorop);
+	if (colorop_state)
+		return colorop_state;
+
+	ret = drm_modeset_lock(&colorop->plane->mutex, state->acquire_ctx);
+	if (ret)
+		return ERR_PTR(ret);
+
+	colorop_state = drm_atomic_helper_colorop_duplicate_state(colorop);
+	if (!colorop_state)
+		return ERR_PTR(-ENOMEM);
+
+	state->colorops[index].state = colorop_state;
+	state->colorops[index].ptr = colorop;
+	state->colorops[index].old_state = colorop->state;
+	state->colorops[index].new_state = colorop_state;
+	colorop_state->state = state;
+
+	drm_dbg_atomic(colorop->dev, "Added [COLOROP:%d] %p state to %p\n",
+		       colorop->base.id, colorop_state, state);
+
+	return colorop_state;
+}
+EXPORT_SYMBOL(drm_atomic_get_colorop_state);
+
 static bool
 plane_switching_crtc(const struct drm_plane_state *old_plane_state,
 		     const struct drm_plane_state *new_plane_state)
diff --git a/drivers/gpu/drm/drm_atomic_helper.c b/drivers/gpu/drm/drm_atomic_helper.c
index e641fcf8c568..10adac9397cf 100644
--- a/drivers/gpu/drm/drm_atomic_helper.c
+++ b/drivers/gpu/drm/drm_atomic_helper.c
@@ -3184,6 +3184,8 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state,
 	struct drm_crtc_state *old_crtc_state, *new_crtc_state;
 	struct drm_plane *plane;
 	struct drm_plane_state *old_plane_state, *new_plane_state;
+	struct drm_colorop *colorop;
+	struct drm_colorop_state *old_colorop_state, *new_colorop_state;
 	struct drm_crtc_commit *commit;
 	struct drm_private_obj *obj;
 	struct drm_private_state *old_obj_state, *new_obj_state;
@@ -3261,6 +3263,16 @@ int drm_atomic_helper_swap_state(struct drm_atomic_state *state,
 		}
 	}
 
+	for_each_oldnew_colorop_in_state(state, colorop, old_colorop_state, new_colorop_state, i) {
+		WARN_ON(colorop->state != old_colorop_state);
+
+		old_colorop_state->state = state;
+		new_colorop_state->state = NULL;
+
+		state->colorops[i].state = old_colorop_state;
+		colorop->state = new_colorop_state;
+	}
+
 	drm_panic_lock(state->dev, flags);
 	for_each_oldnew_plane_in_state(state, plane, old_plane_state, new_plane_state, i) {
 		WARN_ON(plane->state != old_plane_state);
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index b2cb5ae5a139..148f11895b9e 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -35,6 +35,7 @@
 #include <drm/drm_drv.h>
 #include <drm/drm_writeback.h>
 #include <drm/drm_vblank.h>
+#include <drm/drm_colorop.h>
 
 #include <linux/export.h>
 #include <linux/dma-fence.h>
@@ -648,6 +649,26 @@ drm_atomic_plane_get_property(struct drm_plane *plane,
 	return 0;
 }
 
+static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
+					   struct drm_colorop_state *state,
+					   struct drm_file *file_priv,
+					   struct drm_property *property,
+					   uint64_t val)
+{
+	drm_dbg_atomic(colorop->dev,
+		       "[COLOROP:%d] unknown property [PROP:%d:%s]]\n",
+		       colorop->base.id, property->base.id, property->name);
+	return -EINVAL;
+}
+
+static int
+drm_atomic_colorop_get_property(struct drm_colorop *colorop,
+				const struct drm_colorop_state *state,
+				struct drm_property *property, uint64_t *val)
+{
+	return -EINVAL;
+}
+
 static int drm_atomic_set_writeback_fb_for_connector(
 		struct drm_connector_state *conn_state,
 		struct drm_framebuffer *fb)
@@ -914,6 +935,15 @@ int drm_atomic_get_property(struct drm_mode_object *obj,
 				plane->state, property, val);
 		break;
 	}
+	case DRM_MODE_OBJECT_COLOROP: {
+		struct drm_colorop *colorop = obj_to_colorop(obj);
+
+		if (colorop->plane)
+			WARN_ON(!drm_modeset_is_locked(&colorop->plane->mutex));
+
+		ret = drm_atomic_colorop_get_property(colorop, colorop->state, property, val);
+		break;
+	}
 	default:
 		drm_dbg_atomic(dev, "[OBJECT:%d] has no properties\n", obj->id);
 		ret = -EINVAL;
@@ -1111,6 +1141,21 @@ int drm_atomic_set_property(struct drm_atomic_state *state,
 		ret = drm_atomic_plane_set_property(plane,
 				plane_state, file_priv,
 				prop, prop_value);
+
+		break;
+	}
+	case DRM_MODE_OBJECT_COLOROP: {
+		struct drm_colorop *colorop = obj_to_colorop(obj);
+		struct drm_colorop_state *colorop_state;
+
+		colorop_state = drm_atomic_get_colorop_state(state, colorop);
+		if (IS_ERR(colorop_state)) {
+			ret = PTR_ERR(colorop_state);
+			break;
+		}
+
+		ret = drm_atomic_colorop_set_property(colorop, colorop_state,
+						      file_priv, prop, prop_value);
 		break;
 	}
 	default:
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
new file mode 100644
index 000000000000..59af7ac888d6
--- /dev/null
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -0,0 +1,103 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright (C) 2023 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#include <drm/drm_colorop.h>
+#include <drm/drm_print.h>
+#include <drm/drm_drv.h>
+#include <drm/drm_plane.h>
+
+#include "drm_crtc_internal.h"
+
+static void __drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop,
+							struct drm_colorop_state *state)
+{
+	memcpy(state, colorop->state, sizeof(*state));
+}
+
+struct drm_colorop_state *
+drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop)
+{
+	struct drm_colorop_state *state;
+
+	if (WARN_ON(!colorop->state))
+		return NULL;
+
+	state = kmalloc(sizeof(*state), GFP_KERNEL);
+	if (state)
+		__drm_atomic_helper_colorop_duplicate_state(colorop, state);
+
+	return state;
+}
+
+void drm_colorop_atomic_destroy_state(struct drm_colorop *colorop,
+				      struct drm_colorop_state *state)
+{
+	kfree(state);
+}
+
+/**
+ * __drm_colorop_state_reset - resets colorop state to default values
+ * @colorop_state: atomic colorop state, must not be NULL
+ * @colorop: colorop object, must not be NULL
+ *
+ * Initializes the newly allocated @colorop_state with default
+ * values. This is useful for drivers that subclass the CRTC state.
+ */
+static void __drm_colorop_state_reset(struct drm_colorop_state *colorop_state,
+				      struct drm_colorop *colorop)
+{
+	colorop_state->colorop = colorop;
+}
+
+/**
+ * __drm_colorop_reset - reset state on colorop
+ * @colorop: drm colorop
+ * @colorop_state: colorop state to assign
+ *
+ * Initializes the newly allocated @colorop_state and assigns it to
+ * the &drm_crtc->state pointer of @colorop, usually required when
+ * initializing the drivers or when called from the &drm_colorop_funcs.reset
+ * hook.
+ *
+ * This is useful for drivers that subclass the colorop state.
+ */
+static void __drm_colorop_reset(struct drm_colorop *colorop,
+				struct drm_colorop_state *colorop_state)
+{
+	if (colorop_state)
+		__drm_colorop_state_reset(colorop_state, colorop);
+
+	colorop->state = colorop_state;
+}
+
+void drm_colorop_reset(struct drm_colorop *colorop)
+{
+	kfree(colorop->state);
+	colorop->state = kzalloc(sizeof(*colorop->state), GFP_KERNEL);
+
+	if (colorop->state)
+		__drm_colorop_reset(colorop, colorop->state);
+}
diff --git a/drivers/gpu/drm/drm_mode_config.c b/drivers/gpu/drm/drm_mode_config.c
index 25f376869b3a..d12db9b0bab8 100644
--- a/drivers/gpu/drm/drm_mode_config.c
+++ b/drivers/gpu/drm/drm_mode_config.c
@@ -30,6 +30,7 @@
 #include <drm/drm_managed.h>
 #include <drm/drm_mode_config.h>
 #include <drm/drm_print.h>
+#include <drm/drm_colorop.h>
 #include <linux/dma-resv.h>
 
 #include "drm_crtc_internal.h"
@@ -192,11 +193,15 @@ int drm_mode_getresources(struct drm_device *dev, void *data,
 void drm_mode_config_reset(struct drm_device *dev)
 {
 	struct drm_crtc *crtc;
+	struct drm_colorop *colorop;
 	struct drm_plane *plane;
 	struct drm_encoder *encoder;
 	struct drm_connector *connector;
 	struct drm_connector_list_iter conn_iter;
 
+	drm_for_each_colorop(colorop, dev)
+		drm_colorop_reset(colorop);
+
 	drm_for_each_plane(plane, dev)
 		if (plane->funcs->reset)
 			plane->funcs->reset(plane);
@@ -437,6 +442,7 @@ int drmm_mode_config_init(struct drm_device *dev)
 	INIT_LIST_HEAD(&dev->mode_config.property_list);
 	INIT_LIST_HEAD(&dev->mode_config.property_blob_list);
 	INIT_LIST_HEAD(&dev->mode_config.plane_list);
+	INIT_LIST_HEAD(&dev->mode_config.colorop_list);
 	INIT_LIST_HEAD(&dev->mode_config.privobj_list);
 	idr_init_base(&dev->mode_config.object_idr, 1);
 	idr_init_base(&dev->mode_config.tile_idr, 1);
@@ -458,6 +464,7 @@ int drmm_mode_config_init(struct drm_device *dev)
 	dev->mode_config.num_crtc = 0;
 	dev->mode_config.num_encoder = 0;
 	dev->mode_config.num_total_plane = 0;
+	dev->mode_config.num_colorop = 0;
 
 	if (IS_ENABLED(CONFIG_LOCKDEP)) {
 		struct drm_modeset_acquire_ctx modeset_ctx;
diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index 2e433d44658d..895529337d7e 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -30,6 +30,7 @@
 
 #include <drm/drm_crtc.h>
 #include <drm/drm_util.h>
+#include <drm/drm_colorop.h>
 
 /**
  * struct drm_crtc_commit - track modeset commits on a CRTC
@@ -157,6 +158,11 @@ struct drm_crtc_commit {
 	bool abort_completion;
 };
 
+struct __drm_colorops_state {
+	struct drm_colorop *ptr;
+	struct drm_colorop_state *state, *old_state, *new_state;
+};
+
 struct __drm_planes_state {
 	struct drm_plane *ptr;
 
@@ -531,6 +537,14 @@ struct drm_atomic_state {
 	 */
 	bool checked : 1;
 
+	/**
+	 * @colorops:
+	 *
+	 * Pointer to array of @drm_colorop and @drm_colorop_state part of this
+	 * update.
+	 */
+	struct __drm_colorops_state *colorops;
+
 	/**
 	 * @planes:
 	 *
@@ -672,6 +686,9 @@ drm_atomic_get_crtc_state(struct drm_atomic_state *state,
 struct drm_plane_state * __must_check
 drm_atomic_get_plane_state(struct drm_atomic_state *state,
 			   struct drm_plane *plane);
+struct drm_colorop_state *
+drm_atomic_get_colorop_state(struct drm_atomic_state *state,
+			     struct drm_colorop *colorop);
 struct drm_connector_state * __must_check
 drm_atomic_get_connector_state(struct drm_atomic_state *state,
 			       struct drm_connector *connector);
@@ -768,6 +785,36 @@ drm_atomic_get_new_plane_state(const struct drm_atomic_state *state,
 	return state->planes[drm_plane_index(plane)].new_state;
 }
 
+/**
+ * drm_atomic_get_old_colorop_state - get colorop state, if it exists
+ * @state: global atomic state object
+ * @colorop: colorop to grab
+ *
+ * This function returns the old colorop state for the given colorop, or
+ * NULL if the colorop is not part of the global atomic state.
+ */
+static inline struct drm_colorop_state *
+drm_atomic_get_old_colorop_state(struct drm_atomic_state *state,
+				 struct drm_colorop *colorop)
+{
+	return state->colorops[drm_colorop_index(colorop)].old_state;
+}
+
+/**
+ * drm_atomic_get_new_colorop_state - get colorop state, if it exists
+ * @state: global atomic state object
+ * @colorop: colorop to grab
+ *
+ * This function returns the new colorop state for the given colorop, or
+ * NULL if the colorop is not part of the global atomic state.
+ */
+static inline struct drm_colorop_state *
+drm_atomic_get_new_colorop_state(struct drm_atomic_state *state,
+				 struct drm_colorop *colorop)
+{
+	return state->colorops[drm_colorop_index(colorop)].new_state;
+}
+
 /**
  * drm_atomic_get_old_connector_state - get connector state, if it exists
  * @state: global atomic state object
@@ -998,6 +1045,29 @@ void drm_state_dump(struct drm_device *dev, struct drm_printer *p);
 			     (new_crtc_state) = (__state)->crtcs[__i].new_state, \
 			     (void)(new_crtc_state) /* Only to avoid unused-but-set-variable warning */, 1))
 
+/**
+ * for_each_oldnew_colorop_in_state - iterate over all colorops in an atomic update
+ * @__state: &struct drm_atomic_state pointer
+ * @colorop: &struct drm_colorop iteration cursor
+ * @old_colorop_state: &struct drm_colorop_state iteration cursor for the old state
+ * @new_colorop_state: &struct drm_colorop_state iteration cursor for the new state
+ * @__i: int iteration cursor, for macro-internal use
+ *
+ * This iterates over all colorops in an atomic update, tracking both old and
+ * new state. This is useful in places where the state delta needs to be
+ * considered, for example in atomic check functions.
+ */
+#define for_each_oldnew_colorop_in_state(__state, colorop, old_colorop_state,	\
+					 new_colorop_state, __i)		\
+	for ((__i) = 0;								\
+	     (__i) < (__state)->dev->mode_config.num_colorop;			\
+	     (__i)++)								\
+		for_each_if ((__state)->colorops[__i].ptr &&			\
+			     ((colorop) = (__state)->colorops[__i].ptr,		\
+			      (void)(colorop) /* Only to avoid unused-but-set-variable warning */, \
+			      (old_colorop_state) = (__state)->colorops[__i].old_state,\
+			      (new_colorop_state) = (__state)->colorops[__i].new_state, 1))
+
 /**
  * for_each_oldnew_plane_in_state - iterate over all planes in an atomic update
  * @__state: &struct drm_atomic_state pointer
diff --git a/include/drm/drm_atomic_uapi.h b/include/drm/drm_atomic_uapi.h
index 4c6d39d7bdb2..70a115d523cd 100644
--- a/include/drm/drm_atomic_uapi.h
+++ b/include/drm/drm_atomic_uapi.h
@@ -37,6 +37,7 @@ struct drm_crtc;
 struct drm_connector_state;
 struct dma_fence;
 struct drm_framebuffer;
+struct drm_colorop;
 
 int __must_check
 drm_atomic_set_mode_for_crtc(struct drm_crtc_state *state,
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
new file mode 100644
index 000000000000..28bb7091ef1f
--- /dev/null
+++ b/include/drm/drm_colorop.h
@@ -0,0 +1,169 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright (C) 2023 Advanced Micro Devices, Inc. All rights reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice shall be included in
+ * all copies or substantial portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR
+ * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR
+ * OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * Authors: AMD
+ *
+ */
+
+#ifndef __DRM_COLOROP_H__
+#define __DRM_COLOROP_H__
+
+#include <drm/drm_mode_object.h>
+#include <drm/drm_mode.h>
+#include <drm/drm_property.h>
+
+/**
+ * struct drm_colorop_state - mutable colorop state
+ */
+struct drm_colorop_state {
+	/** @colorop: backpointer to the colorop */
+	struct drm_colorop *colorop;
+
+	/*
+	 * Color properties
+	 *
+	 * The following fields are not always valid, their usage depends
+	 * on the colorop type. See their associated comment for more
+	 * information.
+	 */
+
+	/** @state: backpointer to global drm_atomic_state */
+	struct drm_atomic_state *state;
+};
+
+/**
+ * struct drm_colorop - DRM color operation control structure
+ *
+ * A colorop represents one color operation. They can be chained via
+ * the 'next' pointer to build a color pipeline.
+ *
+ * Since colorops cannot stand-alone and are used to describe colorop
+ * operations on a plane they don't have their own locking mechanism but
+ * are locked and programmed along with their associated &drm_plane.
+ *
+ */
+struct drm_colorop {
+	/** @dev: parent DRM device */
+	struct drm_device *dev;
+
+	/**
+	 * @head:
+	 *
+	 * List of all colorops on @dev, linked from &drm_mode_config.colorop_list.
+	 * Invariant over the lifetime of @dev and therefore does not need
+	 * locking.
+	 */
+	struct list_head head;
+
+	/**
+	 * @index: Position inside the mode_config.list, can be used as an array
+	 * index. It is invariant over the lifetime of the colorop.
+	 */
+	unsigned int index;
+
+	/** @base: base mode object */
+	struct drm_mode_object base;
+
+	/**
+	 * @plane:
+	 *
+	 * The plane on which the colorop sits. A drm_colorop is always unique
+	 * to a plane.
+	 */
+	struct drm_plane *plane;
+
+	/**
+	 * @state:
+	 *
+	 * Current atomic state for this colorop.
+	 *
+	 * This is protected by @mutex. Note that nonblocking atomic commits
+	 * access the current colorop state without taking locks.
+	 */
+	struct drm_colorop_state *state;
+
+	/*
+	 * Color properties
+	 *
+	 * The following fields are not always valid, their usage depends
+	 * on the colorop type. See their associated comment for more
+	 * information.
+	 */
+
+	/** @properties: property tracking for this colorop */
+	struct drm_object_properties properties;
+
+};
+
+#define obj_to_colorop(x) container_of(x, struct drm_colorop, base)
+
+/**
+ * drm_colorop_find - look up a Colorop object from its ID
+ * @dev: DRM device
+ * @file_priv: drm file to check for lease against.
+ * @id: &drm_mode_object ID
+ *
+ * This can be used to look up a Colorop from its userspace ID. Only used by
+ * drivers for legacy IOCTLs and interface, nowadays extensions to the KMS
+ * userspace interface should be done using &drm_property.
+ */
+static inline struct drm_colorop *drm_colorop_find(struct drm_device *dev,
+						   struct drm_file *file_priv,
+						   uint32_t id)
+{
+	struct drm_mode_object *mo;
+
+	mo = drm_mode_object_find(dev, file_priv, id, DRM_MODE_OBJECT_COLOROP);
+	return mo ? obj_to_colorop(mo) : NULL;
+}
+
+struct drm_colorop_state *
+drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop);
+
+void drm_colorop_atomic_destroy_state(struct drm_colorop *colorop,
+				      struct drm_colorop_state *state);
+
+/**
+ * drm_colorop_reset - reset colorop atomic state
+ * @colorop: drm colorop
+ *
+ * Resets the atomic state for @colorop by freeing the state pointer (which might
+ * be NULL, e.g. at driver load time) and allocating a new empty state object.
+ */
+void drm_colorop_reset(struct drm_colorop *colorop);
+
+/**
+ * drm_colorop_index - find the index of a registered colorop
+ * @colorop: colorop to find index for
+ *
+ * Given a registered colorop, return the index of that colorop within a DRM
+ * device's list of colorops.
+ */
+static inline unsigned int drm_colorop_index(const struct drm_colorop *colorop)
+{
+	return colorop->index;
+}
+
+#define drm_for_each_colorop(colorop, dev) \
+	list_for_each_entry(colorop, &(dev)->mode_config.colorop_list, head)
+
+#endif /* __DRM_COLOROP_H__ */
diff --git a/include/drm/drm_mode_config.h b/include/drm/drm_mode_config.h
index 2e848b816218..895fb820dba0 100644
--- a/include/drm/drm_mode_config.h
+++ b/include/drm/drm_mode_config.h
@@ -500,6 +500,24 @@ struct drm_mode_config {
 	 */
 	struct raw_spinlock panic_lock;
 
+	/**
+	 * @num_colorop:
+	 *
+	 * Number of colorop objects on this device.
+	 * This is invariant over the lifetime of a device and hence doesn't
+	 * need any locks.
+	 */
+	int num_colorop;
+
+	/**
+	 * @colorop_list:
+	 *
+	 * List of colorop objects linked with &drm_colorop.head. This is
+	 * invariant over the lifetime of a device and hence doesn't need any
+	 * locks.
+	 */
+	struct list_head colorop_list;
+
 	/**
 	 * @num_crtc:
 	 *
diff --git a/include/drm/drm_plane.h b/include/drm/drm_plane.h
index 01479dd94e76..61fedd4e253c 100644
--- a/include/drm/drm_plane.h
+++ b/include/drm/drm_plane.h
@@ -243,6 +243,14 @@ struct drm_plane_state {
 	 */
 	enum drm_scaling_filter scaling_filter;
 
+	/**
+	 * @color_pipeline:
+	 *
+	 * The first colorop of the active color pipeline, or NULL, if no
+	 * color pipeline is active.
+	 */
+	struct drm_colorop *color_pipeline;
+
 	/**
 	 * @commit: Tracks the pending commit to prevent use-after-free conditions,
 	 * and for async plane updates.
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index 1e0e02a79b5c..ec27125c2928 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -629,6 +629,7 @@ struct drm_mode_connector_set_property {
 #define DRM_MODE_OBJECT_FB 0xfbfbfbfb
 #define DRM_MODE_OBJECT_BLOB 0xbbbbbbbb
 #define DRM_MODE_OBJECT_PLANE 0xeeeeeeee
+#define DRM_MODE_OBJECT_COLOROP 0xfafafafa
 #define DRM_MODE_OBJECT_ANY 0
 
 struct drm_mode_obj_get_properties {
-- 
cgit v1.2.3


From 84423e561208054f872b3ca66e3e99a10d06c0ac Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:30 -0700
Subject: drm/colorop: Add TYPE property

Add a read-only TYPE property. The TYPE specifies the colorop
type, such as enumerated curve, 1D LUT, CTM, 3D LUT, PWL LUT,
etc.

For now we're only introducing an enumerated 1D LUT type to
illustrate the concept.

Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-6-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic.c      |  4 ++--
 drivers/gpu/drm/drm_atomic_uapi.c |  7 ++++++-
 drivers/gpu/drm/drm_colorop.c     | 12 ++++++++++++
 include/drm/drm_colorop.h         | 24 ++++++++++++++++++++++++
 include/uapi/drm/drm_mode.h       | 19 +++++++++++++++++++
 5 files changed, 63 insertions(+), 3 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 6438a3938032..99545184960f 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -634,8 +634,8 @@ drm_atomic_get_colorop_state(struct drm_atomic_state *state,
 	state->colorops[index].new_state = colorop_state;
 	colorop_state->state = state;
 
-	drm_dbg_atomic(colorop->dev, "Added [COLOROP:%d] %p state to %p\n",
-		       colorop->base.id, colorop_state, state);
+	drm_dbg_atomic(colorop->dev, "Added [COLOROP:%d:%d] %p state to %p\n",
+		       colorop->base.id, colorop->type, colorop_state, state);
 
 	return colorop_state;
 }
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 148f11895b9e..55b3046c5f1c 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -666,7 +666,12 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 				const struct drm_colorop_state *state,
 				struct drm_property *property, uint64_t *val)
 {
-	return -EINVAL;
+	if (property == colorop->type_property)
+		*val = colorop->type;
+	else
+		return -EINVAL;
+
+	return 0;
 }
 
 static int drm_atomic_set_writeback_fb_for_connector(
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 59af7ac888d6..d0e839a1df7c 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -101,3 +101,15 @@ void drm_colorop_reset(struct drm_colorop *colorop)
 	if (colorop->state)
 		__drm_colorop_reset(colorop, colorop->state);
 }
+
+static const char * const colorop_type_name[] = {
+	[DRM_COLOROP_1D_CURVE] = "1D Curve",
+};
+
+const char *drm_get_colorop_type_name(enum drm_colorop_type type)
+{
+	if (WARN_ON(type >= ARRAY_SIZE(colorop_type_name)))
+		return "unknown";
+
+	return colorop_type_name[type];
+}
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 28bb7091ef1f..0bda70f4e82a 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -112,6 +112,21 @@ struct drm_colorop {
 	/** @properties: property tracking for this colorop */
 	struct drm_object_properties properties;
 
+	/**
+	 * @type:
+	 *
+	 * Read-only
+	 * Type of color operation
+	 */
+	enum drm_colorop_type type;
+
+	/**
+	 * @type_property:
+	 *
+	 * Read-only "TYPE" property for specifying the type of
+	 * this color operation. The type is enum drm_colorop_type.
+	 */
+	struct drm_property *type_property;
 };
 
 #define obj_to_colorop(x) container_of(x, struct drm_colorop, base)
@@ -166,4 +181,13 @@ static inline unsigned int drm_colorop_index(const struct drm_colorop *colorop)
 #define drm_for_each_colorop(colorop, dev) \
 	list_for_each_entry(colorop, &(dev)->mode_config.colorop_list, head)
 
+/**
+ * drm_get_colorop_type_name - return a string for colorop type
+ * @type: colorop type to compute name of
+ *
+ * In contrast to the other drm_get_*_name functions this one here returns a
+ * const pointer and hence is threadsafe.
+ */
+const char *drm_get_colorop_type_name(enum drm_colorop_type type);
+
 #endif /* __DRM_COLOROP_H__ */
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index ec27125c2928..c419f93eb94d 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -858,6 +858,25 @@ struct drm_color_lut {
 	__u16 reserved;
 };
 
+/**
+ * enum drm_colorop_type - Type of color operation
+ *
+ * drm_colorops can be of many different types. Each type behaves differently
+ * and defines a different set of properties. This enum defines all types and
+ * gives a high-level description.
+ */
+enum drm_colorop_type {
+	/**
+	 * @DRM_COLOROP_1D_CURVE:
+	 *
+	 * enum string "1D Curve"
+	 *
+	 * A 1D curve that is being applied to all color channels. The
+	 * curve is specified via the CURVE_1D_TYPE colorop property.
+	 */
+	DRM_COLOROP_1D_CURVE
+};
+
 /**
  * struct drm_plane_size_hint - Plane size hints
  * @width: The width of the plane in pixel
-- 
cgit v1.2.3


From 41651f9d42eb24186a38a11be0a75dbc148d2991 Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:31 -0700
Subject: drm/colorop: Add 1D Curve subtype

Add a new drm_colorop with DRM_COLOROP_1D_CURVE with two subtypes:
DRM_COLOROP_1D_CURVE_SRGB_EOTF and DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF.

Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Co-developed-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-7-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic_uapi.c |  17 +++--
 drivers/gpu/drm/drm_colorop.c     | 132 ++++++++++++++++++++++++++++++++++++++
 include/drm/drm_colorop.h         |  62 ++++++++++++++++++
 3 files changed, 207 insertions(+), 4 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 55b3046c5f1c..e01eaab9b2f7 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -655,10 +655,17 @@ static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 					   struct drm_property *property,
 					   uint64_t val)
 {
-	drm_dbg_atomic(colorop->dev,
-		       "[COLOROP:%d] unknown property [PROP:%d:%s]]\n",
-		       colorop->base.id, property->base.id, property->name);
-	return -EINVAL;
+	if (property == colorop->curve_1d_type_property) {
+		state->curve_1d_type = val;
+	} else {
+		drm_dbg_atomic(colorop->dev,
+			       "[COLOROP:%d:%d] unknown property [PROP:%d:%s]\n",
+			       colorop->base.id, colorop->type,
+			       property->base.id, property->name);
+		return -EINVAL;
+	}
+
+	return 0;
 }
 
 static int
@@ -668,6 +675,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 {
 	if (property == colorop->type_property)
 		*val = colorop->type;
+	else if (property == colorop->curve_1d_type_property)
+		*val = state->curve_1d_type;
 	else
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index d0e839a1df7c..81c55e0639c3 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -31,6 +31,121 @@
 
 #include "drm_crtc_internal.h"
 
+static const struct drm_prop_enum_list drm_colorop_type_enum_list[] = {
+	{ DRM_COLOROP_1D_CURVE, "1D Curve" },
+};
+
+static const char * const colorop_curve_1d_type_names[] = {
+	[DRM_COLOROP_1D_CURVE_SRGB_EOTF] = "sRGB EOTF",
+	[DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF] = "sRGB Inverse EOTF",
+};
+
+/* Init Helpers */
+
+static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *colorop,
+				  struct drm_plane *plane, enum drm_colorop_type type)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_property *prop;
+	int ret = 0;
+
+	ret = drm_mode_object_add(dev, &colorop->base, DRM_MODE_OBJECT_COLOROP);
+	if (ret)
+		return ret;
+
+	colorop->base.properties = &colorop->properties;
+	colorop->dev = dev;
+	colorop->type = type;
+	colorop->plane = plane;
+
+	list_add_tail(&colorop->head, &config->colorop_list);
+	colorop->index = config->num_colorop++;
+
+	/* add properties */
+
+	/* type */
+	prop = drm_property_create_enum(dev,
+					DRM_MODE_PROP_IMMUTABLE,
+					"TYPE", drm_colorop_type_enum_list,
+					ARRAY_SIZE(drm_colorop_type_enum_list));
+
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->type_property = prop;
+
+	drm_object_attach_property(&colorop->base,
+				   colorop->type_property,
+				   colorop->type);
+
+	return ret;
+}
+
+/**
+ * drm_plane_colorop_curve_1d_init - Initialize a DRM_COLOROP_1D_CURVE
+ *
+ * @dev: DRM device
+ * @colorop: The drm_colorop object to initialize
+ * @plane: The associated drm_plane
+ * @supported_tfs: A bitfield of supported drm_plane_colorop_curve_1d_init enum values,
+ *                 created using BIT(curve_type) and combined with the OR '|'
+ *                 operator.
+ * @return zero on success, -E value on failure
+ */
+int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
+				    struct drm_plane *plane, u64 supported_tfs)
+{
+	struct drm_prop_enum_list enum_list[DRM_COLOROP_1D_CURVE_COUNT];
+	int i, len;
+
+	struct drm_property *prop;
+	int ret;
+
+	if (!supported_tfs) {
+		drm_err(dev,
+			"No supported TFs for new 1D curve colorop on [PLANE:%d:%s]\n",
+			plane->base.id, plane->name);
+		return -EINVAL;
+	}
+
+	if ((supported_tfs & -BIT(DRM_COLOROP_1D_CURVE_COUNT)) != 0) {
+		drm_err(dev, "Unknown TF provided on [PLANE:%d:%s]\n",
+			plane->base.id, plane->name);
+		return -EINVAL;
+	}
+
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_1D_CURVE);
+	if (ret)
+		return ret;
+
+	len = 0;
+	for (i = 0; i < DRM_COLOROP_1D_CURVE_COUNT; i++) {
+		if ((supported_tfs & BIT(i)) == 0)
+			continue;
+
+		enum_list[len].type = i;
+		enum_list[len].name = colorop_curve_1d_type_names[i];
+		len++;
+	}
+
+	if (WARN_ON(len <= 0))
+		return -EINVAL;
+
+	/* initialize 1D curve only attribute */
+	prop = drm_property_create_enum(dev, DRM_MODE_PROP_ATOMIC, "CURVE_1D_TYPE",
+					enum_list, len);
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->curve_1d_type_property = prop;
+	drm_object_attach_property(&colorop->base, colorop->curve_1d_type_property,
+				   enum_list[0].type);
+	drm_colorop_reset(colorop);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_plane_colorop_curve_1d_init);
+
 static void __drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop,
 							struct drm_colorop_state *state)
 {
@@ -69,7 +184,16 @@ void drm_colorop_atomic_destroy_state(struct drm_colorop *colorop,
 static void __drm_colorop_state_reset(struct drm_colorop_state *colorop_state,
 				      struct drm_colorop *colorop)
 {
+	u64 val;
+
 	colorop_state->colorop = colorop;
+
+	if (colorop->curve_1d_type_property) {
+		drm_object_property_get_default_value(&colorop->base,
+						      colorop->curve_1d_type_property,
+						      &val);
+		colorop_state->curve_1d_type = val;
+	}
 }
 
 /**
@@ -113,3 +237,11 @@ const char *drm_get_colorop_type_name(enum drm_colorop_type type)
 
 	return colorop_type_name[type];
 }
+
+const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type type)
+{
+	if (WARN_ON(type >= ARRAY_SIZE(colorop_curve_1d_type_names)))
+		return "unknown";
+
+	return colorop_curve_1d_type_names[type];
+}
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 0bda70f4e82a..3594e50cfd68 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -31,6 +31,41 @@
 #include <drm/drm_mode.h>
 #include <drm/drm_property.h>
 
+/**
+ * enum drm_colorop_curve_1d_type - type of 1D curve
+ *
+ * Describes a 1D curve to be applied by the DRM_COLOROP_1D_CURVE colorop.
+ */
+enum drm_colorop_curve_1d_type {
+	/**
+	 * @DRM_COLOROP_1D_CURVE_SRGB_EOTF:
+	 *
+	 * enum string "sRGB EOTF"
+	 *
+	 * sRGB piece-wise electro-optical transfer function. Transfer
+	 * characteristics as defined by IEC 61966-2-1 sRGB. Equivalent
+	 * to H.273 TransferCharacteristics code point 13 with
+	 * MatrixCoefficients set to 0.
+	 */
+	DRM_COLOROP_1D_CURVE_SRGB_EOTF,
+
+	/**
+	 * @DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF:
+	 *
+	 * enum string "sRGB Inverse EOTF"
+	 *
+	 * The inverse of &DRM_COLOROP_1D_CURVE_SRGB_EOTF
+	 */
+	DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF,
+
+	/**
+	 * @DRM_COLOROP_1D_CURVE_COUNT:
+	 *
+	 * enum value denoting the size of the enum
+	 */
+	DRM_COLOROP_1D_CURVE_COUNT
+};
+
 /**
  * struct drm_colorop_state - mutable colorop state
  */
@@ -46,6 +81,13 @@ struct drm_colorop_state {
 	 * information.
 	 */
 
+	/**
+	 * @curve_1d_type:
+	 *
+	 * Type of 1D curve.
+	 */
+	enum drm_colorop_curve_1d_type curve_1d_type;
+
 	/** @state: backpointer to global drm_atomic_state */
 	struct drm_atomic_state *state;
 };
@@ -127,6 +169,14 @@ struct drm_colorop {
 	 * this color operation. The type is enum drm_colorop_type.
 	 */
 	struct drm_property *type_property;
+
+	/**
+	 * @curve_1d_type_property:
+	 *
+	 * Sub-type for DRM_COLOROP_1D_CURVE type.
+	 */
+	struct drm_property *curve_1d_type_property;
+
 };
 
 #define obj_to_colorop(x) container_of(x, struct drm_colorop, base)
@@ -151,6 +201,9 @@ static inline struct drm_colorop *drm_colorop_find(struct drm_device *dev,
 	return mo ? obj_to_colorop(mo) : NULL;
 }
 
+int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
+				    struct drm_plane *plane, u64 supported_tfs);
+
 struct drm_colorop_state *
 drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop);
 
@@ -190,4 +243,13 @@ static inline unsigned int drm_colorop_index(const struct drm_colorop *colorop)
  */
 const char *drm_get_colorop_type_name(enum drm_colorop_type type);
 
+/**
+ * drm_get_colorop_curve_1d_type_name - return a string for 1D curve type
+ * @type: 1d curve type to compute name of
+ *
+ * In contrast to the other drm_get_*_name functions this one here returns a
+ * const pointer and hence is threadsafe.
+ */
+const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type type);
+
 #endif /* __DRM_COLOROP_H__ */
-- 
cgit v1.2.3


From 8c5ea1745f4c89576bc6d213ea7f9a7068ada4ad Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:32 -0700
Subject: drm/colorop: Add BYPASS property

We want to be able to bypass each colorop at all times.
Introduce a new BYPASS boolean property for this.

Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-8-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic_uapi.c |  6 +++++-
 drivers/gpu/drm/drm_colorop.c     | 15 +++++++++++++++
 include/drm/drm_colorop.h         | 21 +++++++++++++++++++++
 3 files changed, 41 insertions(+), 1 deletion(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index e01eaab9b2f7..25a898090fd9 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -655,7 +655,9 @@ static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 					   struct drm_property *property,
 					   uint64_t val)
 {
-	if (property == colorop->curve_1d_type_property) {
+	if (property == colorop->bypass_property) {
+		state->bypass = val;
+	} else if (property == colorop->curve_1d_type_property) {
 		state->curve_1d_type = val;
 	} else {
 		drm_dbg_atomic(colorop->dev,
@@ -675,6 +677,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 {
 	if (property == colorop->type_property)
 		*val = colorop->type;
+	else if (property == colorop->bypass_property)
+		*val = state->bypass;
 	else if (property == colorop->curve_1d_type_property)
 		*val = state->curve_1d_type;
 	else
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 81c55e0639c3..4c088a5a489c 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -78,6 +78,17 @@ static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *co
 				   colorop->type_property,
 				   colorop->type);
 
+	/* bypass */
+	prop = drm_property_create_bool(dev, DRM_MODE_PROP_ATOMIC,
+					"BYPASS");
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->bypass_property = prop;
+	drm_object_attach_property(&colorop->base,
+				   colorop->bypass_property,
+				   1);
+
 	return ret;
 }
 
@@ -134,6 +145,7 @@ int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *
 	/* initialize 1D curve only attribute */
 	prop = drm_property_create_enum(dev, DRM_MODE_PROP_ATOMIC, "CURVE_1D_TYPE",
 					enum_list, len);
+
 	if (!prop)
 		return -ENOMEM;
 
@@ -150,6 +162,8 @@ static void __drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colo
 							struct drm_colorop_state *state)
 {
 	memcpy(state, colorop->state, sizeof(*state));
+
+	state->bypass = true;
 }
 
 struct drm_colorop_state *
@@ -187,6 +201,7 @@ static void __drm_colorop_state_reset(struct drm_colorop_state *colorop_state,
 	u64 val;
 
 	colorop_state->colorop = colorop;
+	colorop_state->bypass = true;
 
 	if (colorop->curve_1d_type_property) {
 		drm_object_property_get_default_value(&colorop->base,
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 3594e50cfd68..2d24f7248831 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -81,6 +81,15 @@ struct drm_colorop_state {
 	 * information.
 	 */
 
+	/**
+	 * @bypass:
+	 *
+	 * When the property BYPASS exists on this colorop, this stores
+	 * the requested bypass state: true if colorop shall be bypassed,
+	 * false if colorop is enabled.
+	 */
+	bool bypass;
+
 	/**
 	 * @curve_1d_type:
 	 *
@@ -170,6 +179,18 @@ struct drm_colorop {
 	 */
 	struct drm_property *type_property;
 
+	/**
+	 * @bypass_property:
+	 *
+	 * Boolean property to control enablement of the color
+	 * operation. Setting bypass to "true" shall always be supported
+	 * in order to allow compositors to quickly fall back to
+	 * alternate methods of color processing. This is important
+	 * since setting color operations can fail due to unique
+	 * HW constraints.
+	 */
+	struct drm_property *bypass_property;
+
 	/**
 	 * @curve_1d_type_property:
 	 *
-- 
cgit v1.2.3


From 78a5add824186da775be4a10ac0a95cca239718c Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:33 -0700
Subject: drm/colorop: Add NEXT property

We'll construct color pipelines out of drm_colorop by
chaining them via the NEXT pointer. NEXT will point to
the next drm_colorop in the pipeline, or by 0 if we're
at the end of the pipeline.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-9-alex.hung@amd.com
---
 drivers/gpu/drm/drm_colorop.c | 27 +++++++++++++++++++++++++++
 include/drm/drm_colorop.h     | 17 +++++++++++++++++
 2 files changed, 44 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 4c088a5a489c..6dd685b7348b 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -57,6 +57,7 @@ static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *co
 	colorop->dev = dev;
 	colorop->type = type;
 	colorop->plane = plane;
+	colorop->next = NULL;
 
 	list_add_tail(&colorop->head, &config->colorop_list);
 	colorop->index = config->num_colorop++;
@@ -89,6 +90,16 @@ static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *co
 				   colorop->bypass_property,
 				   1);
 
+	/* next */
+	prop = drm_property_create_object(dev, DRM_MODE_PROP_IMMUTABLE | DRM_MODE_PROP_ATOMIC,
+					  "NEXT", DRM_MODE_OBJECT_COLOROP);
+	if (!prop)
+		return -ENOMEM;
+	colorop->next_property = prop;
+	drm_object_attach_property(&colorop->base,
+				   colorop->next_property,
+				   0);
+
 	return ret;
 }
 
@@ -260,3 +271,19 @@ const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type ty
 
 	return colorop_curve_1d_type_names[type];
 }
+
+/**
+ * drm_colorop_set_next_property - sets the next pointer
+ * @colorop: drm colorop
+ * @next: next colorop
+ *
+ * Should be used when constructing the color pipeline
+ */
+void drm_colorop_set_next_property(struct drm_colorop *colorop, struct drm_colorop *next)
+{
+	drm_object_property_set_value(&colorop->base,
+				      colorop->next_property,
+				      next ? next->base.id : 0);
+	colorop->next = next;
+}
+EXPORT_SYMBOL(drm_colorop_set_next_property);
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 2d24f7248831..d8c17c537586 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -171,6 +171,14 @@ struct drm_colorop {
 	 */
 	enum drm_colorop_type type;
 
+	/**
+	 * @next:
+	 *
+	 * Read-only
+	 * Pointer to next drm_colorop in pipeline
+	 */
+	struct drm_colorop *next;
+
 	/**
 	 * @type_property:
 	 *
@@ -198,6 +206,13 @@ struct drm_colorop {
 	 */
 	struct drm_property *curve_1d_type_property;
 
+	/**
+	 * @next_property:
+	 *
+	 * Read-only property to next colorop in the pipeline
+	 */
+	struct drm_property *next_property;
+
 };
 
 #define obj_to_colorop(x) container_of(x, struct drm_colorop, base)
@@ -273,4 +288,6 @@ const char *drm_get_colorop_type_name(enum drm_colorop_type type);
  */
 const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type type);
 
+void drm_colorop_set_next_property(struct drm_colorop *colorop, struct drm_colorop *next);
+
 #endif /* __DRM_COLOROP_H__ */
-- 
cgit v1.2.3


From 2afc3184f3b3f05de05ee1d8fede76d3c20802be Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:35 -0700
Subject: drm/plane: Add COLOR PIPELINE property

We're adding a new enum COLOR PIPELINE property. This
property will have entries for each COLOR PIPELINE by
referencing the DRM object ID of the first drm_colorop
of the pipeline. 0 disables the entire COLOR PIPELINE.

Userspace can use this to discover the available color
pipelines, as well as set the desired one. The color
pipelines are programmed via properties on the actual
drm_colorop objects.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-11-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic.c              | 46 ++++++++++++++++++++++++
 drivers/gpu/drm/drm_atomic_state_helper.c |  5 +++
 drivers/gpu/drm/drm_atomic_uapi.c         | 40 +++++++++++++++++++++
 drivers/gpu/drm/drm_plane.c               | 59 +++++++++++++++++++++++++++++++
 include/drm/drm_atomic.h                  |  3 ++
 include/drm/drm_atomic_uapi.h             |  2 ++
 include/drm/drm_plane.h                   | 11 ++++++
 7 files changed, 166 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index b996fa1c27b8..95111f9a8635 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -1536,6 +1536,52 @@ drm_atomic_add_affected_planes(struct drm_atomic_state *state,
 }
 EXPORT_SYMBOL(drm_atomic_add_affected_planes);
 
+/**
+ * drm_atomic_add_affected_colorops - add colorops for plane
+ * @state: atomic state
+ * @plane: DRM plane
+ *
+ * This function walks the current configuration and adds all colorops
+ * currently used by @plane to the atomic configuration @state. This is useful
+ * when an atomic commit also needs to check all currently enabled colorop on
+ * @plane, e.g. when changing the mode. It's also useful when re-enabling a plane
+ * to avoid special code to force-enable all colorops.
+ *
+ * Since acquiring a colorop state will always also acquire the w/w mutex of the
+ * current plane for that colorop (if there is any) adding all the colorop states for
+ * a plane will not reduce parallelism of atomic updates.
+ *
+ * Returns:
+ * 0 on success or can fail with -EDEADLK or -ENOMEM. When the error is EDEADLK
+ * then the w/w mutex code has detected a deadlock and the entire atomic
+ * sequence must be restarted. All other errors are fatal.
+ */
+int
+drm_atomic_add_affected_colorops(struct drm_atomic_state *state,
+				 struct drm_plane *plane)
+{
+	struct drm_colorop *colorop;
+	struct drm_colorop_state *colorop_state;
+
+	WARN_ON(!drm_atomic_get_new_plane_state(state, plane));
+
+	drm_dbg_atomic(plane->dev,
+		       "Adding all current colorops for [PLANE:%d:%s] to %p\n",
+		       plane->base.id, plane->name, state);
+
+	drm_for_each_colorop(colorop, plane->dev) {
+		if (colorop->plane != plane)
+			continue;
+
+		colorop_state = drm_atomic_get_colorop_state(state, colorop);
+		if (IS_ERR(colorop_state))
+			return PTR_ERR(colorop_state);
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_atomic_add_affected_colorops);
+
 /**
  * drm_atomic_check_only - check whether a given config would work
  * @state: atomic configuration to check
diff --git a/drivers/gpu/drm/drm_atomic_state_helper.c b/drivers/gpu/drm/drm_atomic_state_helper.c
index 7142e163e618..cee6d8fc44ad 100644
--- a/drivers/gpu/drm/drm_atomic_state_helper.c
+++ b/drivers/gpu/drm/drm_atomic_state_helper.c
@@ -268,6 +268,11 @@ void __drm_atomic_helper_plane_state_reset(struct drm_plane_state *plane_state,
 			plane_state->color_range = val;
 	}
 
+	if (plane->color_pipeline_property) {
+		/* default is always NULL, i.e., bypass */
+		plane_state->color_pipeline = NULL;
+	}
+
 	if (plane->zpos_property) {
 		if (!drm_object_property_get_default_value(&plane->base,
 							   plane->zpos_property,
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 25a898090fd9..608b0b571c2e 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -258,6 +258,34 @@ drm_atomic_set_fb_for_plane(struct drm_plane_state *plane_state,
 }
 EXPORT_SYMBOL(drm_atomic_set_fb_for_plane);
 
+/**
+ * drm_atomic_set_colorop_for_plane - set colorop for plane
+ * @plane_state: atomic state object for the plane
+ * @colorop: colorop to use for the plane
+ *
+ * Helper function to select the color pipeline on a plane by setting
+ * it to the first drm_colorop element of the pipeline.
+ */
+void
+drm_atomic_set_colorop_for_plane(struct drm_plane_state *plane_state,
+				 struct drm_colorop *colorop)
+{
+	struct drm_plane *plane = plane_state->plane;
+
+	if (colorop)
+		drm_dbg_atomic(plane->dev,
+			       "Set [COLOROP:%d] for [PLANE:%d:%s] state %p\n",
+			       colorop->base.id, plane->base.id, plane->name,
+			       plane_state);
+	else
+		drm_dbg_atomic(plane->dev,
+			       "Set [NOCOLOROP] for [PLANE:%d:%s] state %p\n",
+			       plane->base.id, plane->name, plane_state);
+
+	plane_state->color_pipeline = colorop;
+}
+EXPORT_SYMBOL(drm_atomic_set_colorop_for_plane);
+
 /**
  * drm_atomic_set_crtc_for_connector - set CRTC for connector
  * @conn_state: atomic state object for the connector
@@ -545,6 +573,16 @@ static int drm_atomic_plane_set_property(struct drm_plane *plane,
 		state->color_encoding = val;
 	} else if (property == plane->color_range_property) {
 		state->color_range = val;
+	} else if (property == plane->color_pipeline_property) {
+		/* find DRM colorop object */
+		struct drm_colorop *colorop = NULL;
+
+		colorop = drm_colorop_find(dev, file_priv, val);
+
+		if (val && !colorop)
+			return -EACCES;
+
+		drm_atomic_set_colorop_for_plane(state, colorop);
 	} else if (property == config->prop_fb_damage_clips) {
 		ret = drm_property_replace_blob_from_id(dev,
 					&state->fb_damage_clips,
@@ -627,6 +665,8 @@ drm_atomic_plane_get_property(struct drm_plane *plane,
 		*val = state->color_encoding;
 	} else if (property == plane->color_range_property) {
 		*val = state->color_range;
+	} else if (property == plane->color_pipeline_property) {
+		*val = (state->color_pipeline) ? state->color_pipeline->base.id : 0;
 	} else if (property == config->prop_fb_damage_clips) {
 		*val = (state->fb_damage_clips) ?
 			state->fb_damage_clips->base.id : 0;
diff --git a/drivers/gpu/drm/drm_plane.c b/drivers/gpu/drm/drm_plane.c
index 38f82391bfda..f6cfa8ac090c 100644
--- a/drivers/gpu/drm/drm_plane.c
+++ b/drivers/gpu/drm/drm_plane.c
@@ -1820,3 +1820,62 @@ int drm_plane_add_size_hints_property(struct drm_plane *plane,
 	return 0;
 }
 EXPORT_SYMBOL(drm_plane_add_size_hints_property);
+
+/**
+ * drm_plane_create_color_pipeline_property - create a new color pipeline
+ * property
+ *
+ * @plane: drm plane
+ * @pipelines: list of pipelines
+ * @num_pipelines: number of pipelines
+ *
+ * Create the COLOR_PIPELINE plane property to specific color pipelines on
+ * the plane.
+ *
+ * RETURNS:
+ * Zero for success or -errno
+ */
+int drm_plane_create_color_pipeline_property(struct drm_plane *plane,
+					     const struct drm_prop_enum_list *pipelines,
+					     int num_pipelines)
+{
+	struct drm_prop_enum_list *all_pipelines;
+	struct drm_property *prop;
+	int len = 0;
+	int i;
+
+	all_pipelines = kcalloc(num_pipelines + 1,
+				sizeof(*all_pipelines),
+				GFP_KERNEL);
+
+	if (!all_pipelines) {
+		drm_err(plane->dev, "failed to allocate color pipeline\n");
+		return -ENOMEM;
+	}
+
+	/* Create default Bypass color pipeline */
+	all_pipelines[len].type = 0;
+	all_pipelines[len].name = "Bypass";
+	len++;
+
+	/* Add all other color pipelines */
+	for (i = 0; i < num_pipelines; i++, len++) {
+		all_pipelines[len].type = pipelines[i].type;
+		all_pipelines[len].name = pipelines[i].name;
+	}
+
+	prop = drm_property_create_enum(plane->dev, DRM_MODE_PROP_ATOMIC,
+					"COLOR_PIPELINE",
+					all_pipelines, len);
+	if (IS_ERR(prop)) {
+		kfree(all_pipelines);
+		return PTR_ERR(prop);
+	}
+
+	drm_object_attach_property(&plane->base, prop, 0);
+	plane->color_pipeline_property = prop;
+
+	kfree(all_pipelines);
+	return 0;
+}
+EXPORT_SYMBOL(drm_plane_create_color_pipeline_property);
diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index 895529337d7e..faeb5de74d50 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -906,6 +906,9 @@ drm_atomic_add_affected_connectors(struct drm_atomic_state *state,
 int __must_check
 drm_atomic_add_affected_planes(struct drm_atomic_state *state,
 			       struct drm_crtc *crtc);
+int __must_check
+drm_atomic_add_affected_colorops(struct drm_atomic_state *state,
+				 struct drm_plane *plane);
 
 int __must_check drm_atomic_check_only(struct drm_atomic_state *state);
 int __must_check drm_atomic_commit(struct drm_atomic_state *state);
diff --git a/include/drm/drm_atomic_uapi.h b/include/drm/drm_atomic_uapi.h
index 70a115d523cd..436315523326 100644
--- a/include/drm/drm_atomic_uapi.h
+++ b/include/drm/drm_atomic_uapi.h
@@ -50,6 +50,8 @@ drm_atomic_set_crtc_for_plane(struct drm_plane_state *plane_state,
 			      struct drm_crtc *crtc);
 void drm_atomic_set_fb_for_plane(struct drm_plane_state *plane_state,
 				 struct drm_framebuffer *fb);
+void drm_atomic_set_colorop_for_plane(struct drm_plane_state *plane_state,
+				      struct drm_colorop *colorop);
 int __must_check
 drm_atomic_set_crtc_for_connector(struct drm_connector_state *conn_state,
 				  struct drm_crtc *crtc);
diff --git a/include/drm/drm_plane.h b/include/drm/drm_plane.h
index 61fedd4e253c..703ef4d1bbbc 100644
--- a/include/drm/drm_plane.h
+++ b/include/drm/drm_plane.h
@@ -791,6 +791,14 @@ struct drm_plane {
 	 */
 	struct drm_property *color_range_property;
 
+	/**
+	 * @color_pipeline_property:
+	 *
+	 * Optional "COLOR_PIPELINE" enum property for specifying
+	 * a color pipeline to use on the plane.
+	 */
+	struct drm_property *color_pipeline_property;
+
 	/**
 	 * @scaling_filter_property: property to apply a particular filter while
 	 * scaling.
@@ -1014,4 +1022,7 @@ int drm_plane_add_size_hints_property(struct drm_plane *plane,
 				      const struct drm_plane_size_hint *hints,
 				      int num_hints);
 
+int drm_plane_create_color_pipeline_property(struct drm_plane *plane,
+					     const struct drm_prop_enum_list *pipelines,
+					     int num_pipelines);
 #endif
-- 
cgit v1.2.3


From 179ab8e7d7b378f1cd3ff10113458133b73dc52e Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:36 -0700
Subject: drm/colorop: Introduce DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE

With the introduction of the pre-blending color pipeline we
can no longer have color operations that don't have a clear
position in the color pipeline. We deprecate all existing
plane properties. For upstream drivers those are:
 - COLOR_ENCODING
 - COLOR_RANGE

Drivers are expected to ignore these properties when
programming the HW. DRM clients that register with
DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE will not be allowed to
set the COLOR_ENCODING and COLOR_RANGE properties.

Setting of the COLOR_PIPELINE plane property or drm_colorop
properties is only allowed for userspace that sets this
client cap.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-12-alex.hung@amd.com
---
 drivers/gpu/drm/drm_connector.c     |  1 +
 drivers/gpu/drm/drm_crtc_internal.h |  1 +
 drivers/gpu/drm/drm_ioctl.c         |  7 +++++++
 drivers/gpu/drm/drm_mode_object.c   | 18 ++++++++++++++++++
 include/drm/drm_file.h              |  7 +++++++
 include/uapi/drm/drm.h              | 15 +++++++++++++++
 6 files changed, 49 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_connector.c b/drivers/gpu/drm/drm_connector.c
index 272d6254ea47..4d6dc9ebfdb5 100644
--- a/drivers/gpu/drm/drm_connector.c
+++ b/drivers/gpu/drm/drm_connector.c
@@ -3439,6 +3439,7 @@ int drm_mode_getconnector(struct drm_device *dev, void *data,
 	 * properties reflect the latest status.
 	 */
 	ret = drm_mode_object_get_properties(&connector->base, file_priv->atomic,
+			file_priv->plane_color_pipeline,
 			(uint32_t __user *)(unsigned long)(out_resp->props_ptr),
 			(uint64_t __user *)(unsigned long)(out_resp->prop_values_ptr),
 			&out_resp->count_props);
diff --git a/drivers/gpu/drm/drm_crtc_internal.h b/drivers/gpu/drm/drm_crtc_internal.h
index 89706aa8232f..c09409229644 100644
--- a/drivers/gpu/drm/drm_crtc_internal.h
+++ b/drivers/gpu/drm/drm_crtc_internal.h
@@ -163,6 +163,7 @@ struct drm_mode_object *__drm_mode_object_find(struct drm_device *dev,
 void drm_mode_object_unregister(struct drm_device *dev,
 				struct drm_mode_object *object);
 int drm_mode_object_get_properties(struct drm_mode_object *obj, bool atomic,
+				   bool plane_color_pipeline,
 				   uint32_t __user *prop_ptr,
 				   uint64_t __user *prop_values,
 				   uint32_t *arg_count_props);
diff --git a/drivers/gpu/drm/drm_ioctl.c b/drivers/gpu/drm/drm_ioctl.c
index d8a24875a7ba..ff193155129e 100644
--- a/drivers/gpu/drm/drm_ioctl.c
+++ b/drivers/gpu/drm/drm_ioctl.c
@@ -373,6 +373,13 @@ drm_setclientcap(struct drm_device *dev, void *data, struct drm_file *file_priv)
 			return -EINVAL;
 		file_priv->supports_virtualized_cursor_plane = req->value;
 		break;
+	case DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE:
+		if (!file_priv->atomic)
+			return -EINVAL;
+		if (req->value > 1)
+			return -EINVAL;
+		file_priv->plane_color_pipeline = req->value;
+		break;
 	default:
 		return -EINVAL;
 	}
diff --git a/drivers/gpu/drm/drm_mode_object.c b/drivers/gpu/drm/drm_mode_object.c
index e943205a2394..b45d501b10c8 100644
--- a/drivers/gpu/drm/drm_mode_object.c
+++ b/drivers/gpu/drm/drm_mode_object.c
@@ -28,6 +28,7 @@
 #include <drm/drm_device.h>
 #include <drm/drm_file.h>
 #include <drm/drm_mode_object.h>
+#include <drm/drm_plane.h>
 #include <drm/drm_print.h>
 
 #include "drm_crtc_internal.h"
@@ -386,6 +387,7 @@ EXPORT_SYMBOL(drm_object_property_get_default_value);
 
 /* helper for getconnector and getproperties ioctls */
 int drm_mode_object_get_properties(struct drm_mode_object *obj, bool atomic,
+				   bool plane_color_pipeline,
 				   uint32_t __user *prop_ptr,
 				   uint64_t __user *prop_values,
 				   uint32_t *arg_count_props)
@@ -399,6 +401,21 @@ int drm_mode_object_get_properties(struct drm_mode_object *obj, bool atomic,
 		if ((prop->flags & DRM_MODE_PROP_ATOMIC) && !atomic)
 			continue;
 
+		if (plane_color_pipeline && obj->type == DRM_MODE_OBJECT_PLANE) {
+			struct drm_plane *plane = obj_to_plane(obj);
+
+			if (prop == plane->color_encoding_property ||
+			    prop == plane->color_range_property)
+				continue;
+		}
+
+		if (!plane_color_pipeline && obj->type == DRM_MODE_OBJECT_PLANE) {
+			struct drm_plane *plane = obj_to_plane(obj);
+
+			if (prop == plane->color_pipeline_property)
+				continue;
+		}
+
 		if (*arg_count_props > count) {
 			ret = __drm_object_property_get_value(obj, prop, &val);
 			if (ret)
@@ -457,6 +474,7 @@ int drm_mode_obj_get_properties_ioctl(struct drm_device *dev, void *data,
 	}
 
 	ret = drm_mode_object_get_properties(obj, file_priv->atomic,
+			file_priv->plane_color_pipeline,
 			(uint32_t __user *)(unsigned long)(arg->props_ptr),
 			(uint64_t __user *)(unsigned long)(arg->prop_values_ptr),
 			&arg->count_props);
diff --git a/include/drm/drm_file.h b/include/drm/drm_file.h
index 115763799625..1a3018e4a537 100644
--- a/include/drm/drm_file.h
+++ b/include/drm/drm_file.h
@@ -206,6 +206,13 @@ struct drm_file {
 	 */
 	bool writeback_connectors;
 
+	/**
+	 * @plane_color_pipeline:
+	 *
+	 * True if client understands plane color pipelines
+	 */
+	bool plane_color_pipeline;
+
 	/**
 	 * @was_master:
 	 *
diff --git a/include/uapi/drm/drm.h b/include/uapi/drm/drm.h
index 3cd5cf15e3c9..27cc159c1d27 100644
--- a/include/uapi/drm/drm.h
+++ b/include/uapi/drm/drm.h
@@ -906,6 +906,21 @@ struct drm_get_cap {
  */
 #define DRM_CLIENT_CAP_CURSOR_PLANE_HOTSPOT	6
 
+/**
+ * DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE
+ *
+ * If set to 1 the DRM core will allow setting the COLOR_PIPELINE
+ * property on a &drm_plane, as well as drm_colorop properties.
+ *
+ * Setting of these plane properties will be rejected when this client
+ * cap is set:
+ * - COLOR_ENCODING
+ * - COLOR_RANGE
+ *
+ * The client must enable &DRM_CLIENT_CAP_ATOMIC first.
+ */
+#define DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE	7
+
 /* DRM_IOCTL_SET_CLIENT_CAP ioctl argument type */
 struct drm_set_client_cap {
 	__u64 capability;
-- 
cgit v1.2.3


From 9cf87f864d8328b06060ce11c4f6d04a12c3047f Mon Sep 17 00:00:00 2001
From: Alex Hung <alex.hung@amd.com>
Date: Fri, 14 Nov 2025 17:01:38 -0700
Subject: drm/colorop: Add destroy functions for color pipeline

The functions are to clean up color pipeline when a device driver
fails to create its color pipeline.

Signed-off-by: Alex Hung <alex.hung@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-14-alex.hung@amd.com
---
 drivers/gpu/drm/drm_colorop.c | 36 ++++++++++++++++++++++++++++++++++++
 include/drm/drm_colorop.h     |  3 +++
 2 files changed, 39 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 6f9c27d216c1..b75ad8544315 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -134,6 +134,42 @@ static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *co
 	return ret;
 }
 
+/**
+ * drm_colorop_cleanup - Cleanup a drm_colorop object in color_pipeline
+ *
+ * @colorop: The drm_colorop object to be cleaned
+ */
+void drm_colorop_cleanup(struct drm_colorop *colorop)
+{
+	struct drm_device *dev = colorop->dev;
+	struct drm_mode_config *config = &dev->mode_config;
+
+	list_del(&colorop->head);
+	config->num_colorop--;
+
+	kfree(colorop->state);
+}
+EXPORT_SYMBOL(drm_colorop_cleanup);
+
+/**
+ * drm_colorop_pipeline_destroy - Helper for color pipeline destruction
+ *
+ * @dev: - The drm_device containing the drm_planes with the color_pipelines
+ *
+ * Provides a default color pipeline destroy handler for drm_device.
+ */
+void drm_colorop_pipeline_destroy(struct drm_device *dev)
+{
+	struct drm_mode_config *config = &dev->mode_config;
+	struct drm_colorop *colorop, *next;
+
+	list_for_each_entry_safe(colorop, next, &config->colorop_list, head) {
+		drm_colorop_cleanup(colorop);
+		kfree(colorop);
+	}
+}
+EXPORT_SYMBOL(drm_colorop_pipeline_destroy);
+
 /**
  * drm_plane_colorop_curve_1d_init - Initialize a DRM_COLOROP_1D_CURVE
  *
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index d8c17c537586..ba03b35454da 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -237,6 +237,9 @@ static inline struct drm_colorop *drm_colorop_find(struct drm_device *dev,
 	return mo ? obj_to_colorop(mo) : NULL;
 }
 
+void drm_colorop_pipeline_destroy(struct drm_device *dev);
+void drm_colorop_cleanup(struct drm_colorop *colorop);
+
 int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
 				    struct drm_plane *plane, u64 supported_tfs);
 
-- 
cgit v1.2.3


From e5719e7f19009d4fbedf685fc22eec9cd8de154f Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:43 -0700
Subject: drm/colorop: Add 3x4 CTM type

This type is used to support a 3x4 matrix in colorops. A 3x4
matrix uses the last column as a "bias" column. Some HW exposes
support for 3x4. The calculation looks like:

 out   matrix    in
 |R|   |0  1  2  3 |   | R |
 |G| = |4  5  6  7 | x | G |
 |B|   |8  9  10 11|   | B |
                       |1.0|

This is also the first colorop where we need a blob property to
program the property. For that we'll introduce a new DATA
property that can be used by all colorop TYPEs requiring a
blob. The way a DATA blob is read depends on the TYPE of
the colorop.

We only create the DATA property for property types that
need it.

Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Louis Chauvet <louis.chauvet@bootlin.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-19-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic.c      |  3 +++
 drivers/gpu/drm/drm_atomic_uapi.c | 31 ++++++++++++++++++++++++++
 drivers/gpu/drm/drm_colorop.c     | 47 +++++++++++++++++++++++++++++++++++++++
 include/drm/drm_colorop.h         | 24 ++++++++++++++++++++
 include/uapi/drm/amdgpu_drm.h     |  9 --------
 include/uapi/drm/drm_mode.h       | 32 +++++++++++++++++++++++++-
 6 files changed, 136 insertions(+), 10 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 95111f9a8635..60b3a069f4dd 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -794,6 +794,9 @@ static void drm_atomic_colorop_print_state(struct drm_printer *p,
 		drm_printf(p, "\tcurve_1d_type=%s\n",
 			   drm_get_colorop_curve_1d_type_name(state->curve_1d_type));
 		break;
+	case DRM_COLOROP_CTM_3X4:
+		drm_printf(p, "\tdata blob id=%d\n", state->data ? state->data->base.id : 0);
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 608b0b571c2e..392198aae072 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -689,6 +689,32 @@ drm_atomic_plane_get_property(struct drm_plane *plane,
 	return 0;
 }
 
+static int drm_atomic_color_set_data_property(struct drm_colorop *colorop,
+					      struct drm_colorop_state *state,
+					      struct drm_property *property,
+					      uint64_t val)
+{
+	ssize_t elem_size = -1;
+	ssize_t size = -1;
+	bool replaced = false;
+
+	switch (colorop->type) {
+	case DRM_COLOROP_CTM_3X4:
+		size = sizeof(struct drm_color_ctm_3x4);
+		break;
+	default:
+		/* should never get here */
+		return -EINVAL;
+	}
+
+	return drm_property_replace_blob_from_id(colorop->dev,
+						 &state->data,
+						 val,
+						 size,
+						 elem_size,
+						 &replaced);
+}
+
 static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 					   struct drm_colorop_state *state,
 					   struct drm_file *file_priv,
@@ -699,6 +725,9 @@ static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 		state->bypass = val;
 	} else if (property == colorop->curve_1d_type_property) {
 		state->curve_1d_type = val;
+	} else if (property == colorop->data_property) {
+		return drm_atomic_color_set_data_property(colorop, state,
+							  property, val);
 	} else {
 		drm_dbg_atomic(colorop->dev,
 			       "[COLOROP:%d:%d] unknown property [PROP:%d:%s]\n",
@@ -721,6 +750,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 		*val = state->bypass;
 	else if (property == colorop->curve_1d_type_property)
 		*val = state->curve_1d_type;
+	else if (property == colorop->data_property)
+		*val = (state->data) ? state->data->base.id : 0;
 	else
 		return -EINVAL;
 
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index b75ad8544315..c68b85a7b261 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -64,6 +64,7 @@
 
 static const struct drm_prop_enum_list drm_colorop_type_enum_list[] = {
 	{ DRM_COLOROP_1D_CURVE, "1D Curve" },
+	{ DRM_COLOROP_CTM_3X4, "3x4 Matrix"},
 };
 
 static const char * const colorop_curve_1d_type_names[] = {
@@ -147,6 +148,11 @@ void drm_colorop_cleanup(struct drm_colorop *colorop)
 	list_del(&colorop->head);
 	config->num_colorop--;
 
+	if (colorop->state && colorop->state->data) {
+		drm_property_blob_put(colorop->state->data);
+		colorop->state->data = NULL;
+	}
+
 	kfree(colorop->state);
 }
 EXPORT_SYMBOL(drm_colorop_cleanup);
@@ -236,11 +242,51 @@ int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *
 }
 EXPORT_SYMBOL(drm_plane_colorop_curve_1d_init);
 
+static int drm_colorop_create_data_prop(struct drm_device *dev, struct drm_colorop *colorop)
+{
+	struct drm_property *prop;
+
+	/* data */
+	prop = drm_property_create(dev, DRM_MODE_PROP_ATOMIC | DRM_MODE_PROP_BLOB,
+				   "DATA", 0);
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->data_property = prop;
+	drm_object_attach_property(&colorop->base,
+				   colorop->data_property,
+				   0);
+
+	return 0;
+}
+
+int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
+				   struct drm_plane *plane)
+{
+	int ret;
+
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_CTM_3X4);
+	if (ret)
+		return ret;
+
+	ret = drm_colorop_create_data_prop(dev, colorop);
+	if (ret)
+		return ret;
+
+	drm_colorop_reset(colorop);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_plane_colorop_ctm_3x4_init);
+
 static void __drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop,
 							struct drm_colorop_state *state)
 {
 	memcpy(state, colorop->state, sizeof(*state));
 
+	if (state->data)
+		drm_property_blob_get(state->data);
+
 	state->bypass = true;
 }
 
@@ -321,6 +367,7 @@ void drm_colorop_reset(struct drm_colorop *colorop)
 
 static const char * const colorop_type_name[] = {
 	[DRM_COLOROP_1D_CURVE] = "1D Curve",
+	[DRM_COLOROP_CTM_3X4] = "3x4 Matrix",
 };
 
 const char *drm_get_colorop_type_name(enum drm_colorop_type type)
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index ba03b35454da..8b5f8aaac2f4 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -97,6 +97,17 @@ struct drm_colorop_state {
 	 */
 	enum drm_colorop_curve_1d_type curve_1d_type;
 
+	/**
+	 * @data:
+	 *
+	 * Data blob for any TYPE that requires such a blob. The
+	 * interpretation of the blob is TYPE-specific.
+	 *
+	 * See the &drm_colorop_type documentation for how blob is laid
+	 * out.
+	 */
+	struct drm_property_blob *data;
+
 	/** @state: backpointer to global drm_atomic_state */
 	struct drm_atomic_state *state;
 };
@@ -206,6 +217,17 @@ struct drm_colorop {
 	 */
 	struct drm_property *curve_1d_type_property;
 
+	/**
+	 * @data_property:
+	 *
+	 * blob property for any TYPE that requires a blob of data,
+	 * such as 1DLUT, CTM, 3DLUT, etc.
+	 *
+	 * The way this blob is interpreted depends on the TYPE of
+	 * this
+	 */
+	struct drm_property *data_property;
+
 	/**
 	 * @next_property:
 	 *
@@ -242,6 +264,8 @@ void drm_colorop_cleanup(struct drm_colorop *colorop);
 
 int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
 				    struct drm_plane *plane, u64 supported_tfs);
+int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
+				   struct drm_plane *plane);
 
 struct drm_colorop_state *
 drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop);
diff --git a/include/uapi/drm/amdgpu_drm.h b/include/uapi/drm/amdgpu_drm.h
index 406a42be429b..f80aa4c9d88f 100644
--- a/include/uapi/drm/amdgpu_drm.h
+++ b/include/uapi/drm/amdgpu_drm.h
@@ -1656,15 +1656,6 @@ struct drm_amdgpu_info_uq_metadata {
 #define AMDGPU_FAMILY_GC_11_5_0			150 /* GC 11.5.0 */
 #define AMDGPU_FAMILY_GC_12_0_0			152 /* GC 12.0.0 */
 
-/* FIXME wrong namespace! */
-struct drm_color_ctm_3x4 {
-	/*
-	 * Conversion matrix with 3x4 dimensions in S31.32 sign-magnitude
-	 * (not two's complement!) format.
-	 */
-	__u64 matrix[12];
-};
-
 #if defined(__cplusplus)
 }
 #endif
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index c419f93eb94d..054561022953 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -847,6 +847,20 @@ struct drm_color_ctm {
 	__u64 matrix[9];
 };
 
+struct drm_color_ctm_3x4 {
+	/*
+	 * Conversion matrix with 3x4 dimensions in S31.32 sign-magnitude
+	 * (not two's complement!) format.
+	 *
+	 * out   matrix          in
+	 * |R|   |0  1  2  3 |   | R |
+	 * |G| = |4  5  6  7 | x | G |
+	 * |B|   |8  9  10 11|   | B |
+	 *                       |1.0|
+	 */
+	__u64 matrix[12];
+};
+
 struct drm_color_lut {
 	/*
 	 * Values are mapped linearly to 0.0 - 1.0 range, with 0x0 == 0.0 and
@@ -874,7 +888,23 @@ enum drm_colorop_type {
 	 * A 1D curve that is being applied to all color channels. The
 	 * curve is specified via the CURVE_1D_TYPE colorop property.
 	 */
-	DRM_COLOROP_1D_CURVE
+	DRM_COLOROP_1D_CURVE,
+
+	/**
+	 * @DRM_COLOROP_CTM_3X4:
+	 *
+	 * enum string "3x4 Matrix"
+	 *
+	 * A 3x4 matrix. Its values are specified via the
+	 * &drm_color_ctm_3x4 struct provided via the DATA property.
+	 *
+	 * The DATA blob is a float[12]:
+	 * out   matrix          in
+	 * | R |   | 0  1  2  3  |   | R |
+	 * | G | = | 4  5  6  7  | x | G |
+	 * | B |   | 8  9  10 12 |   | B |
+	 */
+	DRM_COLOROP_CTM_3X4,
 };
 
 /**
-- 
cgit v1.2.3


From dabeebae4ab7b4372e17310aa7a2168d32268833 Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:48 -0700
Subject: drm/colorop: pass plane_color_pipeline client cap to atomic check

Drivers will need to know whether an atomic check/commit
originated from a client with DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE
so they can ignore deprecated properties, like COLOR_ENCODING
and COLOR_RANGE.

Pass the plane_color_pipeline bit to drm_atomic_state.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-24-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic_uapi.c |  1 +
 include/drm/drm_atomic.h          | 18 ++++++++++++++++++
 2 files changed, 19 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 392198aae072..6f4345b9b2a3 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -1584,6 +1584,7 @@ int drm_mode_atomic_ioctl(struct drm_device *dev,
 	drm_modeset_acquire_init(&ctx, DRM_MODESET_ACQUIRE_INTERRUPTIBLE);
 	state->acquire_ctx = &ctx;
 	state->allow_modeset = !!(arg->flags & DRM_MODE_ATOMIC_ALLOW_MODESET);
+	state->plane_color_pipeline = file_priv->plane_color_pipeline;
 
 retry:
 	copied_objs = 0;
diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index faeb5de74d50..e65ea288cb8b 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -537,6 +537,24 @@ struct drm_atomic_state {
 	 */
 	bool checked : 1;
 
+	/**
+	 * @plane_color_pipeline:
+	 *
+	 * Indicates whether this atomic state originated with a client that
+	 * set the DRM_CLIENT_CAP_PLANE_COLOR_PIPELINE.
+	 *
+	 * Drivers and helper functions should use this to ignore legacy
+	 * properties that are incompatible with the drm_plane COLOR_PIPELINE
+	 * behavior, such as:
+	 *
+	 *  - COLOR_RANGE
+	 *  - COLOR_ENCODING
+	 *
+	 * or any other driver-specific properties that might affect pixel
+	 * values.
+	 */
+	bool plane_color_pipeline : 1;
+
 	/**
 	 * @colorops:
 	 *
-- 
cgit v1.2.3


From e341cc6e85df7ab413471f8c1a3a10f076b2dfcc Mon Sep 17 00:00:00 2001
From: Alex Hung <alex.hung@amd.com>
Date: Fri, 14 Nov 2025 17:01:49 -0700
Subject: drm/colorop: define a new macro for_each_new_colorop_in_state

Create a new macro for_each_new_colorop_in_state to access new
drm_colorop_state updated from uapi.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-25-alex.hung@amd.com
---
 include/drm/drm_atomic.h | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

(limited to 'include')

diff --git a/include/drm/drm_atomic.h b/include/drm/drm_atomic.h
index e65ea288cb8b..43783891d359 100644
--- a/include/drm/drm_atomic.h
+++ b/include/drm/drm_atomic.h
@@ -1089,6 +1089,26 @@ void drm_state_dump(struct drm_device *dev, struct drm_printer *p);
 			      (old_colorop_state) = (__state)->colorops[__i].old_state,\
 			      (new_colorop_state) = (__state)->colorops[__i].new_state, 1))
 
+/**
+ * for_each_new_colorop_in_state - iterate over all colorops in an atomic update
+ * @__state: &struct drm_atomic_state pointer
+ * @colorop: &struct drm_colorop iteration cursor
+ * @new_colorop_state: &struct drm_colorop_state iteration cursor for the new state
+ * @__i: int iteration cursor, for macro-internal use
+ *
+ * This iterates over all colorops in an atomic update, tracking new state. This is
+ * useful in places where the state delta needs to be considered, for example in
+ * atomic check functions.
+ */
+#define for_each_new_colorop_in_state(__state, colorop, new_colorop_state, __i) \
+	for ((__i) = 0;							\
+	     (__i) < (__state)->dev->mode_config.num_colorop;		\
+	     (__i)++)							\
+		for_each_if ((__state)->colorops[__i].ptr &&		\
+			     ((colorop) = (__state)->colorops[__i].ptr,	\
+			      (void)(colorop) /* Only to avoid unused-but-set-variable warning */, \
+			      (new_colorop_state) = (__state)->colorops[__i].new_state, 1))
+
 /**
  * for_each_oldnew_plane_in_state - iterate over all planes in an atomic update
  * @__state: &struct drm_atomic_state pointer
-- 
cgit v1.2.3


From 1b75447412d6e29948c9b687a60352748945c917 Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:56 -0700
Subject: drm/colorop: Add PQ 125 EOTF and its inverse

The PQ function defines a mapping of code values to nits (cd/m^2).
The max code value maps to 10,000 nits.

Windows DWM's canonical composition color space (CCCS)  defaults
to composing SDR contents to 80 nits and uses a float value of
1.0 to represent this. For this reason AMD HW hard-codes a PQ
function that is scaled by 125, yielding 80 nit PQ values for
1.0 and 10,000 nits at 125.0.

This patch introduces this scaled PQ EOTF and its inverse as
1D curve types.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-32-alex.hung@amd.com
---
 drivers/gpu/drm/drm_colorop.c |  2 ++
 include/drm/drm_colorop.h     | 24 ++++++++++++++++++++++++
 2 files changed, 26 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index c68b85a7b261..feaca07b429c 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -70,6 +70,8 @@ static const struct drm_prop_enum_list drm_colorop_type_enum_list[] = {
 static const char * const colorop_curve_1d_type_names[] = {
 	[DRM_COLOROP_1D_CURVE_SRGB_EOTF] = "sRGB EOTF",
 	[DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF] = "sRGB Inverse EOTF",
+	[DRM_COLOROP_1D_CURVE_PQ_125_EOTF] = "PQ 125 EOTF",
+	[DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF] = "PQ 125 Inverse EOTF",
 };
 
 /* Init Helpers */
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 8b5f8aaac2f4..139c9cb38b67 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -58,6 +58,30 @@ enum drm_colorop_curve_1d_type {
 	 */
 	DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF,
 
+	/**
+	 * @DRM_COLOROP_1D_CURVE_PQ_125_EOTF:
+	 *
+	 * enum string "PQ 125 EOTF"
+	 *
+	 * The PQ transfer function, scaled by 125.0f, so that 10,000
+	 * nits correspond to 125.0f.
+	 *
+	 * Transfer characteristics of the PQ function as defined by
+	 * SMPTE ST 2084 (2014) for 10-, 12-, 14-, and 16-bit systems
+	 * and Rec. ITU-R BT.2100-2 perceptual quantization (PQ) system,
+	 * represented by H.273 TransferCharacteristics code point 16.
+	 */
+	DRM_COLOROP_1D_CURVE_PQ_125_EOTF,
+
+	/**
+	 * @DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF:
+	 *
+	 * enum string "PQ 125 Inverse EOTF"
+	 *
+	 * The inverse of DRM_COLOROP_1D_CURVE_PQ_125_EOTF.
+	 */
+	DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF,
+
 	/**
 	 * @DRM_COLOROP_1D_CURVE_COUNT:
 	 *
-- 
cgit v1.2.3


From a355b3d6a46e42b05fafa587cebe76039ed388c6 Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:01:58 -0700
Subject: drm/colorop: add BT2020/BT709 OETF and Inverse OETF

The BT.709 and BT.2020 OETFs are the same, the only difference
being that the BT.2020 variant is defined with more precision
for 10 and 12-bit per color encodings.

Both are used as encoding functions for video content, and are
therefore defined as OETF (opto-electronic transfer function)
instead of as EOTF (electro-optical transfer function).

Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-34-alex.hung@amd.com
---
 drivers/gpu/drm/drm_colorop.c |  2 ++
 include/drm/drm_colorop.h     | 23 +++++++++++++++++++++++
 2 files changed, 25 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index feaca07b429c..46e4942b66e6 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -72,6 +72,8 @@ static const char * const colorop_curve_1d_type_names[] = {
 	[DRM_COLOROP_1D_CURVE_SRGB_INV_EOTF] = "sRGB Inverse EOTF",
 	[DRM_COLOROP_1D_CURVE_PQ_125_EOTF] = "PQ 125 EOTF",
 	[DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF] = "PQ 125 Inverse EOTF",
+	[DRM_COLOROP_1D_CURVE_BT2020_INV_OETF] = "BT.2020 Inverse OETF",
+	[DRM_COLOROP_1D_CURVE_BT2020_OETF] = "BT.2020 OETF",
 };
 
 /* Init Helpers */
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 139c9cb38b67..067805276b15 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -82,6 +82,29 @@ enum drm_colorop_curve_1d_type {
 	 */
 	DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF,
 
+	/**
+	 * @DRM_COLOROP_1D_CURVE_BT2020_INV_OETF:
+	 *
+	 * enum string "BT.2020 Inverse OETF"
+	 *
+	 * The inverse of &DRM_COLOROP_1D_CURVE_BT2020_OETF
+	 */
+	DRM_COLOROP_1D_CURVE_BT2020_INV_OETF,
+
+	/**
+	 * @DRM_COLOROP_1D_CURVE_BT2020_OETF:
+	 *
+	 * enum string "BT.2020 OETF"
+	 *
+	 * The BT.2020/BT.709 transfer function. The BT.709 and BT.2020
+	 * transfer functions are the same, the only difference is that
+	 * BT.2020 is defined with more precision for 10 and 12-bit
+	 * encodings.
+	 *
+	 *
+	 */
+	DRM_COLOROP_1D_CURVE_BT2020_OETF,
+
 	/**
 	 * @DRM_COLOROP_1D_CURVE_COUNT:
 	 *
-- 
cgit v1.2.3


From 621c45ca12ed9bd5a8ef434925fe51c319c6e28d Mon Sep 17 00:00:00 2001
From: Uma Shankar <uma.shankar@intel.com>
Date: Fri, 14 Nov 2025 17:02:00 -0700
Subject: drm: Add Enhanced LUT precision structure

Existing LUT precision structure drm_color_lut has only 16 bit
precision. This is not enough for upcoming enhanced hardwares
and advance usecases like HDR processing. Hence added a new
structure with 32 bit precision values.

Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-36-alex.hung@amd.com
---
 drivers/gpu/drm/drm_color_mgmt.c | 43 ++++++++++++++++++++++++++++++++++++++++
 include/drm/drm_color_mgmt.h     | 13 ++++++++++++
 include/uapi/drm/drm_mode.h      | 12 +++++++++++
 3 files changed, 68 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_color_mgmt.c b/drivers/gpu/drm/drm_color_mgmt.c
index 131c1c9ae92f..c598b99673fc 100644
--- a/drivers/gpu/drm/drm_color_mgmt.c
+++ b/drivers/gpu/drm/drm_color_mgmt.c
@@ -874,3 +874,46 @@ void drm_crtc_fill_palette_8(struct drm_crtc *crtc, drm_crtc_set_lut_func set_pa
 		fill_palette_8(crtc, i, set_palette);
 }
 EXPORT_SYMBOL(drm_crtc_fill_palette_8);
+
+/**
+ * drm_color_lut32_check - check validity of extended lookup table
+ * @lut: property blob containing extended LUT to check
+ * @tests: bitmask of tests to run
+ *
+ * Helper to check whether a userspace-provided extended lookup table is valid and
+ * satisfies hardware requirements.  Drivers pass a bitmask indicating which of
+ * the tests in &drm_color_lut_tests should be performed.
+ *
+ * Returns 0 on success, -EINVAL on failure.
+ */
+int drm_color_lut32_check(const struct drm_property_blob *lut, u32 tests)
+{
+	const struct drm_color_lut32 *entry;
+	int i;
+
+	if (!lut || !tests)
+		return 0;
+
+	entry = lut->data;
+	for (i = 0; i < drm_color_lut32_size(lut); i++) {
+		if (tests & DRM_COLOR_LUT_EQUAL_CHANNELS) {
+			if (entry[i].red != entry[i].blue ||
+			    entry[i].red != entry[i].green) {
+				DRM_DEBUG_KMS("All LUT entries must have equal r/g/b\n");
+				return -EINVAL;
+			}
+		}
+
+		if (i > 0 && tests & DRM_COLOR_LUT_NON_DECREASING) {
+			if (entry[i].red < entry[i - 1].red ||
+			    entry[i].green < entry[i - 1].green ||
+			    entry[i].blue < entry[i - 1].blue) {
+				DRM_DEBUG_KMS("LUT entries must never decrease.\n");
+				return -EINVAL;
+			}
+		}
+	}
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_color_lut32_check);
diff --git a/include/drm/drm_color_mgmt.h b/include/drm/drm_color_mgmt.h
index eccb71ab335a..527582c20885 100644
--- a/include/drm/drm_color_mgmt.h
+++ b/include/drm/drm_color_mgmt.h
@@ -72,6 +72,18 @@ static inline int drm_color_lut_size(const struct drm_property_blob *blob)
 	return blob->length / sizeof(struct drm_color_lut);
 }
 
+/**
+ * drm_color_lut32_size - calculate the number of entries in the extended LUT
+ * @blob: blob containing the LUT
+ *
+ * Returns:
+ * The number of entries in the color LUT stored in @blob.
+ */
+static inline int drm_color_lut32_size(const struct drm_property_blob *blob)
+{
+	return blob->length / sizeof(struct drm_color_lut32);
+}
+
 enum drm_color_encoding {
 	DRM_COLOR_YCBCR_BT601,
 	DRM_COLOR_YCBCR_BT709,
@@ -146,4 +158,5 @@ void drm_crtc_load_palette_8(struct drm_crtc *crtc, const struct drm_color_lut *
 void drm_crtc_fill_palette_332(struct drm_crtc *crtc, drm_crtc_set_lut_func set_palette);
 void drm_crtc_fill_palette_8(struct drm_crtc *crtc, drm_crtc_set_lut_func set_palette);
 
+int drm_color_lut32_check(const struct drm_property_blob *lut, u32 tests);
 #endif
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index 054561022953..5e637ec7b64c 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -872,6 +872,18 @@ struct drm_color_lut {
 	__u16 reserved;
 };
 
+/*
+ * struct drm_color_lut32
+ *
+ * 32-bit per channel color LUT entry, similar to drm_color_lut.
+ */
+struct drm_color_lut32 {
+	__u32 red;
+	__u32 green;
+	__u32 blue;
+	__u32 reserved;
+};
+
 /**
  * enum drm_colorop_type - Type of color operation
  *
-- 
cgit v1.2.3


From 94529775135c301efff78373310102f47f3c8671 Mon Sep 17 00:00:00 2001
From: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Date: Fri, 14 Nov 2025 17:02:01 -0700
Subject: drm: Add helper to extract lut from struct drm_color_lut32

Add helper to extract lut values in 32-bit precision needed by
hardware.

Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: Chaitanya Kumar Borah <chaitanya.kumar.borah@intel.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-37-alex.hung@amd.com
---
 include/drm/drm_color_mgmt.h | 16 ++++++++++++++++
 1 file changed, 16 insertions(+)

(limited to 'include')

diff --git a/include/drm/drm_color_mgmt.h b/include/drm/drm_color_mgmt.h
index 527582c20885..5140691f476a 100644
--- a/include/drm/drm_color_mgmt.h
+++ b/include/drm/drm_color_mgmt.h
@@ -50,6 +50,22 @@ static inline u32 drm_color_lut_extract(u32 user_input, int bit_precision)
 					 (1 << 16) - 1);
 }
 
+/**
+ * drm_color_lut32_extract - clamp and round LUT entries
+ * @user_input: input value
+ * @bit_precision: number of bits the hw LUT supports
+ *
+ * Extract U0.bit_precision from a U0.32 LUT value.
+ *
+ */
+static inline u32 drm_color_lut32_extract(u32 user_input, int bit_precision)
+{
+	u64 max = (bit_precision >= 64) ? ~0ULL : (1ULL << bit_precision) - 1;
+
+	return DIV_ROUND_CLOSEST_ULL((u64)user_input * max,
+				     (1ULL << 32) - 1);
+}
+
 u64 drm_color_ctm_s31_32_to_qm_n(u64 user_input, u32 m, u32 n);
 
 void drm_crtc_enable_color_mgmt(struct drm_crtc *crtc,
-- 
cgit v1.2.3


From 99a4e4f08abe253a7812e4872882a75cecd87703 Mon Sep 17 00:00:00 2001
From: Alex Hung <alex.hung@amd.com>
Date: Fri, 14 Nov 2025 17:02:02 -0700
Subject: drm/colorop: Add 1D Curve Custom LUT type

We've previously introduced DRM_COLOROP_1D_CURVE for
pre-defined 1D curves. But we also have HW that supports
custom curves and userspace needs the ability to pass
custom curves, aka LUTs.

This patch introduces a new colorop type, called
DRM_COLOROP_1D_LUT that provides a SIZE property which
is used by a driver to advertise the supported SIZE
of the LUT, as well as a DATA property which userspace
uses to set the LUT.

DATA and size function in the same way as current drm_crtc
GAMMA and DEGAMMA LUTs.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Co-developed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-38-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic.c      |  4 ++++
 drivers/gpu/drm/drm_atomic_uapi.c |  5 +++++
 drivers/gpu/drm/drm_colorop.c     | 43 +++++++++++++++++++++++++++++++++++++++
 include/drm/drm_colorop.h         | 16 +++++++++++++++
 include/uapi/drm/drm_mode.h       | 14 +++++++++++++
 5 files changed, 82 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 60b3a069f4dd..142fc52bc5b2 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -794,6 +794,10 @@ static void drm_atomic_colorop_print_state(struct drm_printer *p,
 		drm_printf(p, "\tcurve_1d_type=%s\n",
 			   drm_get_colorop_curve_1d_type_name(state->curve_1d_type));
 		break;
+	case DRM_COLOROP_1D_LUT:
+		drm_printf(p, "\tsize=%d\n", colorop->size);
+		drm_printf(p, "\tdata blob id=%d\n", state->data ? state->data->base.id : 0);
+		break;
 	case DRM_COLOROP_CTM_3X4:
 		drm_printf(p, "\tdata blob id=%d\n", state->data ? state->data->base.id : 0);
 		break;
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 6f4345b9b2a3..b9045b17fa62 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -699,6 +699,9 @@ static int drm_atomic_color_set_data_property(struct drm_colorop *colorop,
 	bool replaced = false;
 
 	switch (colorop->type) {
+	case DRM_COLOROP_1D_LUT:
+		size = colorop->size * sizeof(struct drm_color_lut32);
+		break;
 	case DRM_COLOROP_CTM_3X4:
 		size = sizeof(struct drm_color_ctm_3x4);
 		break;
@@ -750,6 +753,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 		*val = state->bypass;
 	else if (property == colorop->curve_1d_type_property)
 		*val = state->curve_1d_type;
+	else if (property == colorop->size_property)
+		*val = colorop->size;
 	else if (property == colorop->data_property)
 		*val = (state->data) ? state->data->base.id : 0;
 	else
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 46e4942b66e6..eaf457d0700c 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -64,6 +64,7 @@
 
 static const struct drm_prop_enum_list drm_colorop_type_enum_list[] = {
 	{ DRM_COLOROP_1D_CURVE, "1D Curve" },
+	{ DRM_COLOROP_1D_LUT, "1D LUT" },
 	{ DRM_COLOROP_CTM_3X4, "3x4 Matrix"},
 };
 
@@ -264,6 +265,47 @@ static int drm_colorop_create_data_prop(struct drm_device *dev, struct drm_color
 	return 0;
 }
 
+/**
+ * drm_plane_colorop_curve_1d_lut_init - Initialize a DRM_COLOROP_1D_LUT
+ *
+ * @dev: DRM device
+ * @colorop: The drm_colorop object to initialize
+ * @plane: The associated drm_plane
+ * @lut_size: LUT size supported by driver
+ * @return zero on success, -E value on failure
+ */
+int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_colorop *colorop,
+					struct drm_plane *plane, uint32_t lut_size)
+{
+	struct drm_property *prop;
+	int ret;
+
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_1D_LUT);
+	if (ret)
+		return ret;
+
+	/* initialize 1D LUT only attribute */
+	/* LUT size */
+	prop = drm_property_create_range(dev, DRM_MODE_PROP_IMMUTABLE | DRM_MODE_PROP_ATOMIC,
+					 "SIZE", 0, UINT_MAX);
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->size_property = prop;
+	drm_object_attach_property(&colorop->base, colorop->size_property, lut_size);
+	colorop->size = lut_size;
+
+	/* data */
+	ret = drm_colorop_create_data_prop(dev, colorop);
+	if (ret)
+		return ret;
+
+	drm_colorop_reset(colorop);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_plane_colorop_curve_1d_lut_init);
+
 int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
 				   struct drm_plane *plane)
 {
@@ -371,6 +413,7 @@ void drm_colorop_reset(struct drm_colorop *colorop)
 
 static const char * const colorop_type_name[] = {
 	[DRM_COLOROP_1D_CURVE] = "1D Curve",
+	[DRM_COLOROP_1D_LUT] = "1D LUT",
 	[DRM_COLOROP_CTM_3X4] = "3x4 Matrix",
 };
 
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 067805276b15..529af9f8266d 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -257,6 +257,13 @@ struct drm_colorop {
 	 */
 	struct drm_property *bypass_property;
 
+	/**
+	 * @size:
+	 *
+	 * Number of entries of the custom LUT. This should be read-only.
+	 */
+	uint32_t size;
+
 	/**
 	 * @curve_1d_type_property:
 	 *
@@ -264,6 +271,13 @@ struct drm_colorop {
 	 */
 	struct drm_property *curve_1d_type_property;
 
+	/**
+	 * @size_property:
+	 *
+	 * Size property for custom LUT from userspace.
+	 */
+	struct drm_property *size_property;
+
 	/**
 	 * @data_property:
 	 *
@@ -311,6 +325,8 @@ void drm_colorop_cleanup(struct drm_colorop *colorop);
 
 int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
 				    struct drm_plane *plane, u64 supported_tfs);
+int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_colorop *colorop,
+					struct drm_plane *plane, uint32_t lut_size);
 int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
 				   struct drm_plane *plane);
 
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index 5e637ec7b64c..bec524e2fa32 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -902,6 +902,20 @@ enum drm_colorop_type {
 	 */
 	DRM_COLOROP_1D_CURVE,
 
+	/**
+	 * @DRM_COLOROP_1D_LUT:
+	 *
+	 * enum string "1D LUT"
+	 *
+	 * A simple 1D LUT of uniformly spaced &drm_color_lut32 entries,
+	 * packed into a blob via the DATA property. The driver's
+	 * expected LUT size is advertised via the SIZE property.
+	 *
+	 * The DATA blob is an array of struct drm_color_lut32 with size
+	 * of "size".
+	 */
+	DRM_COLOROP_1D_LUT,
+
 	/**
 	 * @DRM_COLOROP_CTM_3X4:
 	 *
-- 
cgit v1.2.3


From 3410108037d5b01fb35d2e4e92c17c3abdf89186 Mon Sep 17 00:00:00 2001
From: Alex Hung <alex.hung@amd.com>
Date: Fri, 14 Nov 2025 17:02:05 -0700
Subject: drm/colorop: Add multiplier type

This introduces a new drm_colorop_type: DRM_COLOROP_MULTIPLIER.

It's a simple multiplier to all pixel values. The value is
specified via a S31.32 fixed point provided via the
"MULTIPLIER" property.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-41-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic.c      |  3 +++
 drivers/gpu/drm/drm_atomic_uapi.c |  4 ++++
 drivers/gpu/drm/drm_colorop.c     | 33 +++++++++++++++++++++++++++++++++
 include/drm/drm_colorop.h         | 16 ++++++++++++++++
 include/uapi/drm/drm_mode.h       | 11 +++++++++++
 5 files changed, 67 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 142fc52bc5b2..45243d05a7fd 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -801,6 +801,9 @@ static void drm_atomic_colorop_print_state(struct drm_printer *p,
 	case DRM_COLOROP_CTM_3X4:
 		drm_printf(p, "\tdata blob id=%d\n", state->data ? state->data->base.id : 0);
 		break;
+	case DRM_COLOROP_MULTIPLIER:
+		drm_printf(p, "\tmultiplier=%llu\n", state->multiplier);
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index b9045b17fa62..874dcf8dd14f 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -728,6 +728,8 @@ static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 		state->bypass = val;
 	} else if (property == colorop->curve_1d_type_property) {
 		state->curve_1d_type = val;
+	} else if (property == colorop->multiplier_property) {
+		state->multiplier = val;
 	} else if (property == colorop->data_property) {
 		return drm_atomic_color_set_data_property(colorop, state,
 							  property, val);
@@ -753,6 +755,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 		*val = state->bypass;
 	else if (property == colorop->curve_1d_type_property)
 		*val = state->curve_1d_type;
+	else if (property == colorop->multiplier_property)
+		*val = state->multiplier;
 	else if (property == colorop->size_property)
 		*val = colorop->size;
 	else if (property == colorop->data_property)
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index eaf457d0700c..be99af98fd6c 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -66,6 +66,7 @@ static const struct drm_prop_enum_list drm_colorop_type_enum_list[] = {
 	{ DRM_COLOROP_1D_CURVE, "1D Curve" },
 	{ DRM_COLOROP_1D_LUT, "1D LUT" },
 	{ DRM_COLOROP_CTM_3X4, "3x4 Matrix"},
+	{ DRM_COLOROP_MULTIPLIER, "Multiplier"},
 };
 
 static const char * const colorop_curve_1d_type_names[] = {
@@ -325,6 +326,37 @@ int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *c
 }
 EXPORT_SYMBOL(drm_plane_colorop_ctm_3x4_init);
 
+/**
+ * drm_plane_colorop_mult_init - Initialize a DRM_COLOROP_MULTIPLIER
+ *
+ * @dev: DRM device
+ * @colorop: The drm_colorop object to initialize
+ * @plane: The associated drm_plane
+ * @return zero on success, -E value on failure
+ */
+int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colorop,
+				struct drm_plane *plane)
+{
+	struct drm_property *prop;
+	int ret;
+
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_MULTIPLIER);
+	if (ret)
+		return ret;
+
+	prop = drm_property_create_range(dev, DRM_MODE_PROP_ATOMIC, "MULTIPLIER", 0, U64_MAX);
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->multiplier_property = prop;
+	drm_object_attach_property(&colorop->base, colorop->multiplier_property, 0);
+
+	drm_colorop_reset(colorop);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_plane_colorop_mult_init);
+
 static void __drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop,
 							struct drm_colorop_state *state)
 {
@@ -415,6 +447,7 @@ static const char * const colorop_type_name[] = {
 	[DRM_COLOROP_1D_CURVE] = "1D Curve",
 	[DRM_COLOROP_1D_LUT] = "1D LUT",
 	[DRM_COLOROP_CTM_3X4] = "3x4 Matrix",
+	[DRM_COLOROP_MULTIPLIER] = "Multiplier",
 };
 
 const char *drm_get_colorop_type_name(enum drm_colorop_type type)
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 529af9f8266d..a4f6a22fa1f9 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -144,6 +144,13 @@ struct drm_colorop_state {
 	 */
 	enum drm_colorop_curve_1d_type curve_1d_type;
 
+	/**
+	 * @multiplier:
+	 *
+	 * Multiplier to 'gain' the plane. Format is S31.32 sign-magnitude.
+	 */
+	uint64_t multiplier;
+
 	/**
 	 * @data:
 	 *
@@ -271,6 +278,13 @@ struct drm_colorop {
 	 */
 	struct drm_property *curve_1d_type_property;
 
+	/**
+	 * @multiplier_property:
+	 *
+	 * Multiplier property for plane gain
+	 */
+	struct drm_property *multiplier_property;
+
 	/**
 	 * @size_property:
 	 *
@@ -329,6 +343,8 @@ int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_color
 					struct drm_plane *plane, uint32_t lut_size);
 int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
 				   struct drm_plane *plane);
+int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colorop,
+				struct drm_plane *plane);
 
 struct drm_colorop_state *
 drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop);
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index bec524e2fa32..cac25c0ca37b 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -931,6 +931,17 @@ enum drm_colorop_type {
 	 * | B |   | 8  9  10 12 |   | B |
 	 */
 	DRM_COLOROP_CTM_3X4,
+
+	/**
+	 * @DRM_COLOROP_MULTIPLIER:
+	 *
+	 * enum string "Multiplier"
+	 *
+	 * A simple multiplier, applied to all color values. The
+	 * multiplier is specified as a S31.32 via the MULTIPLIER
+	 * property.
+	 */
+	DRM_COLOROP_MULTIPLIER,
 };
 
 /**
-- 
cgit v1.2.3


From 7fa3ee8c0a79b7a8b5fae422ca29da2fde6821ba Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:02:08 -0700
Subject: drm/colorop: Define LUT_1D interpolation

We want to make sure userspace is aware of the 1D LUT
interpolation. While linear interpolation is common it
might not be supported on all HW. Give driver implementers
a way to specify their interpolation.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-44-alex.hung@amd.com
---
 .../drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c  |  6 ++--
 drivers/gpu/drm/drm_atomic.c                       |  2 ++
 drivers/gpu/drm/drm_atomic_uapi.c                  |  4 +++
 drivers/gpu/drm/drm_colorop.c                      | 37 +++++++++++++++++++++-
 include/drm/drm_colorop.h                          | 21 +++++++++++-
 include/uapi/drm/drm_mode.h                        | 13 ++++++++
 6 files changed, 79 insertions(+), 4 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
index 33907cc8e1b3..e9363814d666 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
@@ -126,7 +126,8 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES);
+	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES,
+						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR);
 	if (ret)
 		goto cleanup;
 
@@ -156,7 +157,8 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES);
+	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES,
+						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR);
 	if (ret)
 		goto cleanup;
 
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 45243d05a7fd..e49d2442fa12 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -796,6 +796,8 @@ static void drm_atomic_colorop_print_state(struct drm_printer *p,
 		break;
 	case DRM_COLOROP_1D_LUT:
 		drm_printf(p, "\tsize=%d\n", colorop->size);
+		drm_printf(p, "\tinterpolation=%s\n",
+			   drm_get_colorop_lut1d_interpolation_name(colorop->lut1d_interpolation));
 		drm_printf(p, "\tdata blob id=%d\n", state->data ? state->data->base.id : 0);
 		break;
 	case DRM_COLOROP_CTM_3X4:
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 874dcf8dd14f..64e49338e284 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -726,6 +726,8 @@ static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 {
 	if (property == colorop->bypass_property) {
 		state->bypass = val;
+	} else if (property == colorop->lut1d_interpolation_property) {
+		colorop->lut1d_interpolation = val;
 	} else if (property == colorop->curve_1d_type_property) {
 		state->curve_1d_type = val;
 	} else if (property == colorop->multiplier_property) {
@@ -753,6 +755,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 		*val = colorop->type;
 	else if (property == colorop->bypass_property)
 		*val = state->bypass;
+	else if (property == colorop->lut1d_interpolation_property)
+		*val = colorop->lut1d_interpolation;
 	else if (property == colorop->curve_1d_type_property)
 		*val = state->curve_1d_type;
 	else if (property == colorop->multiplier_property)
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index be99af98fd6c..f9717d809829 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -78,6 +78,10 @@ static const char * const colorop_curve_1d_type_names[] = {
 	[DRM_COLOROP_1D_CURVE_BT2020_OETF] = "BT.2020 OETF",
 };
 
+static const struct drm_prop_enum_list drm_colorop_lut1d_interpolation_list[] = {
+	{ DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR, "Linear" },
+};
+
 /* Init Helpers */
 
 static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *colorop,
@@ -273,10 +277,12 @@ static int drm_colorop_create_data_prop(struct drm_device *dev, struct drm_color
  * @colorop: The drm_colorop object to initialize
  * @plane: The associated drm_plane
  * @lut_size: LUT size supported by driver
+ * @interpolation: 1D LUT interpolation type
  * @return zero on success, -E value on failure
  */
 int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_colorop *colorop,
-					struct drm_plane *plane, uint32_t lut_size)
+					struct drm_plane *plane, uint32_t lut_size,
+					enum drm_colorop_lut1d_interpolation_type interpolation)
 {
 	struct drm_property *prop;
 	int ret;
@@ -296,6 +302,17 @@ int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_color
 	drm_object_attach_property(&colorop->base, colorop->size_property, lut_size);
 	colorop->size = lut_size;
 
+	/* interpolation */
+	prop = drm_property_create_enum(dev, 0, "LUT1D_INTERPOLATION",
+					drm_colorop_lut1d_interpolation_list,
+					ARRAY_SIZE(drm_colorop_lut1d_interpolation_list));
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->lut1d_interpolation_property = prop;
+	drm_object_attach_property(&colorop->base, prop, interpolation);
+	colorop->lut1d_interpolation = interpolation;
+
 	/* data */
 	ret = drm_colorop_create_data_prop(dev, colorop);
 	if (ret)
@@ -449,6 +466,9 @@ static const char * const colorop_type_name[] = {
 	[DRM_COLOROP_CTM_3X4] = "3x4 Matrix",
 	[DRM_COLOROP_MULTIPLIER] = "Multiplier",
 };
+static const char * const colorop_lut1d_interpolation_name[] = {
+	[DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR] = "Linear",
+};
 
 const char *drm_get_colorop_type_name(enum drm_colorop_type type)
 {
@@ -466,6 +486,21 @@ const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type ty
 	return colorop_curve_1d_type_names[type];
 }
 
+/**
+ * drm_get_colorop_lut1d_interpolation_name: return a string for interpolation type
+ * @type: interpolation type to compute name of
+ *
+ * In contrast to the other drm_get_*_name functions this one here returns a
+ * const pointer and hence is threadsafe.
+ */
+const char *drm_get_colorop_lut1d_interpolation_name(enum drm_colorop_lut1d_interpolation_type type)
+{
+	if (WARN_ON(type >= ARRAY_SIZE(colorop_lut1d_interpolation_name)))
+		return "unknown";
+
+	return colorop_lut1d_interpolation_name[type];
+}
+
 /**
  * drm_colorop_set_next_property - sets the next pointer
  * @colorop: drm colorop
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index a4f6a22fa1f9..0e1a5e9d26f3 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -271,6 +271,21 @@ struct drm_colorop {
 	 */
 	uint32_t size;
 
+	/**
+	 * @lut1d_interpolation:
+	 *
+	 * Read-only
+	 * Interpolation for DRM_COLOROP_1D_LUT
+	 */
+	enum drm_colorop_lut1d_interpolation_type lut1d_interpolation;
+
+	/**
+	 * @lut1d_interpolation_property:
+	 *
+	 * Read-only property for DRM_COLOROP_1D_LUT interpolation
+	 */
+	struct drm_property *lut1d_interpolation_property;
+
 	/**
 	 * @curve_1d_type_property:
 	 *
@@ -340,7 +355,8 @@ void drm_colorop_cleanup(struct drm_colorop *colorop);
 int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
 				    struct drm_plane *plane, u64 supported_tfs);
 int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_colorop *colorop,
-					struct drm_plane *plane, uint32_t lut_size);
+					struct drm_plane *plane, uint32_t lut_size,
+					enum drm_colorop_lut1d_interpolation_type interpolation);
 int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
 				   struct drm_plane *plane);
 int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colorop,
@@ -394,6 +410,9 @@ const char *drm_get_colorop_type_name(enum drm_colorop_type type);
  */
 const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type type);
 
+const char *
+drm_get_colorop_lut1d_interpolation_name(enum drm_colorop_lut1d_interpolation_type type);
+
 void drm_colorop_set_next_property(struct drm_colorop *colorop, struct drm_colorop *next);
 
 #endif /* __DRM_COLOROP_H__ */
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index cac25c0ca37b..4b38da880fc7 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -944,6 +944,19 @@ enum drm_colorop_type {
 	DRM_COLOROP_MULTIPLIER,
 };
 
+/**
+ * enum drm_colorop_lut1d_interpolation_type - type of interpolation for 1D LUTs
+ */
+enum drm_colorop_lut1d_interpolation_type {
+	/**
+	 * @DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR:
+	 *
+	 * Linear interpolation. Values between points of the LUT will be
+	 * linearly interpolated.
+	 */
+	DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR,
+};
+
 /**
  * struct drm_plane_size_hint - Plane size hints
  * @width: The width of the plane in pixel
-- 
cgit v1.2.3


From 2468963482d844f0657a52f791f15ceb953e5880 Mon Sep 17 00:00:00 2001
From: Harry Wentland <harry.wentland@amd.com>
Date: Fri, 14 Nov 2025 17:02:09 -0700
Subject: drm/colorop: allow non-bypass colorops

Not all HW will be able to do bypass on all color
operations. Introduce an 32 bits 'flags' for all colorop
init functions and DRM_COLOROP_FLAG_ALLOW_BYPASS for creating
the BYPASS property when it's true.

Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Harry Wentland <harry.wentland@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Simon Ser <contact@emersion.fr>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-45-alex.hung@amd.com
---
 .../drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c  | 22 +++++++----
 drivers/gpu/drm/drm_atomic.c                       |  3 +-
 drivers/gpu/drm/drm_colorop.c                      | 45 +++++++++++++---------
 drivers/gpu/drm/vkms/vkms_colorop.c                | 10 +++--
 include/drm/drm_colorop.h                          | 23 ++++++-----
 5 files changed, 63 insertions(+), 40 deletions(-)

(limited to 'include')

diff --git a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
index e9363814d666..cd3b7e802416 100644
--- a/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
+++ b/drivers/gpu/drm/amd/display/amdgpu_dm/amdgpu_dm_colorop.c
@@ -65,7 +65,9 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, amdgpu_dm_supported_degam_tfs);
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane,
+					      amdgpu_dm_supported_degam_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -81,7 +83,7 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_mult_init(dev, ops[i], plane);
+	ret = drm_plane_colorop_mult_init(dev, ops[i], plane, DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -96,7 +98,7 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane);
+	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane, DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -111,7 +113,9 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, amdgpu_dm_supported_shaper_tfs);
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane,
+					      amdgpu_dm_supported_shaper_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -127,7 +131,8 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 	}
 
 	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES,
-						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR);
+						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR,
+						  DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -142,7 +147,9 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, amdgpu_dm_supported_blnd_tfs);
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane,
+					      amdgpu_dm_supported_blnd_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -158,7 +165,8 @@ int amdgpu_dm_initialize_default_pipeline(struct drm_plane *plane, struct drm_pr
 	}
 
 	ret = drm_plane_colorop_curve_1d_lut_init(dev, ops[i], plane, MAX_COLOR_LUT_ENTRIES,
-						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR);
+						  DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR,
+						  DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index e49d2442fa12..436d2a948e41 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -787,7 +787,8 @@ static void drm_atomic_colorop_print_state(struct drm_printer *p,
 
 	drm_printf(p, "colorop[%u]:\n", colorop->base.id);
 	drm_printf(p, "\ttype=%s\n", drm_get_colorop_type_name(colorop->type));
-	drm_printf(p, "\tbypass=%u\n", state->bypass);
+	if (colorop->bypass_property)
+		drm_printf(p, "\tbypass=%u\n", state->bypass);
 
 	switch (colorop->type) {
 	case DRM_COLOROP_1D_CURVE:
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index f9717d809829..fd666394f31f 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -85,7 +85,8 @@ static const struct drm_prop_enum_list drm_colorop_lut1d_interpolation_list[] =
 /* Init Helpers */
 
 static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *colorop,
-				  struct drm_plane *plane, enum drm_colorop_type type)
+				  struct drm_plane *plane, enum drm_colorop_type type,
+				  uint32_t flags)
 {
 	struct drm_mode_config *config = &dev->mode_config;
 	struct drm_property *prop;
@@ -121,16 +122,18 @@ static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *co
 				   colorop->type_property,
 				   colorop->type);
 
-	/* bypass */
-	prop = drm_property_create_bool(dev, DRM_MODE_PROP_ATOMIC,
-					"BYPASS");
-	if (!prop)
-		return -ENOMEM;
-
-	colorop->bypass_property = prop;
-	drm_object_attach_property(&colorop->base,
-				   colorop->bypass_property,
-				   1);
+	if (flags & DRM_COLOROP_FLAG_ALLOW_BYPASS) {
+		/* bypass */
+		prop = drm_property_create_bool(dev, DRM_MODE_PROP_ATOMIC,
+						"BYPASS");
+		if (!prop)
+			return -ENOMEM;
+
+		colorop->bypass_property = prop;
+		drm_object_attach_property(&colorop->base,
+					colorop->bypass_property,
+					1);
+	}
 
 	/* next */
 	prop = drm_property_create_object(dev, DRM_MODE_PROP_IMMUTABLE | DRM_MODE_PROP_ATOMIC,
@@ -195,10 +198,11 @@ EXPORT_SYMBOL(drm_colorop_pipeline_destroy);
  * @supported_tfs: A bitfield of supported drm_plane_colorop_curve_1d_init enum values,
  *                 created using BIT(curve_type) and combined with the OR '|'
  *                 operator.
+ * @flags: bitmask of misc, see DRM_COLOROP_FLAG_* defines.
  * @return zero on success, -E value on failure
  */
 int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
-				    struct drm_plane *plane, u64 supported_tfs)
+				    struct drm_plane *plane, u64 supported_tfs, uint32_t flags)
 {
 	struct drm_prop_enum_list enum_list[DRM_COLOROP_1D_CURVE_COUNT];
 	int i, len;
@@ -219,7 +223,7 @@ int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *
 		return -EINVAL;
 	}
 
-	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_1D_CURVE);
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_1D_CURVE, flags);
 	if (ret)
 		return ret;
 
@@ -278,16 +282,18 @@ static int drm_colorop_create_data_prop(struct drm_device *dev, struct drm_color
  * @plane: The associated drm_plane
  * @lut_size: LUT size supported by driver
  * @interpolation: 1D LUT interpolation type
+ * @flags: bitmask of misc, see DRM_COLOROP_FLAG_* defines.
  * @return zero on success, -E value on failure
  */
 int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_colorop *colorop,
 					struct drm_plane *plane, uint32_t lut_size,
-					enum drm_colorop_lut1d_interpolation_type interpolation)
+					enum drm_colorop_lut1d_interpolation_type interpolation,
+					uint32_t flags)
 {
 	struct drm_property *prop;
 	int ret;
 
-	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_1D_LUT);
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_1D_LUT, flags);
 	if (ret)
 		return ret;
 
@@ -325,11 +331,11 @@ int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_color
 EXPORT_SYMBOL(drm_plane_colorop_curve_1d_lut_init);
 
 int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
-				   struct drm_plane *plane)
+				   struct drm_plane *plane, uint32_t flags)
 {
 	int ret;
 
-	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_CTM_3X4);
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_CTM_3X4, flags);
 	if (ret)
 		return ret;
 
@@ -349,15 +355,16 @@ EXPORT_SYMBOL(drm_plane_colorop_ctm_3x4_init);
  * @dev: DRM device
  * @colorop: The drm_colorop object to initialize
  * @plane: The associated drm_plane
+ * @flags: bitmask of misc, see DRM_COLOROP_FLAG_* defines.
  * @return zero on success, -E value on failure
  */
 int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colorop,
-				struct drm_plane *plane)
+				struct drm_plane *plane, uint32_t flags)
 {
 	struct drm_property *prop;
 	int ret;
 
-	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_MULTIPLIER);
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_MULTIPLIER, flags);
 	if (ret)
 		return ret;
 
diff --git a/drivers/gpu/drm/vkms/vkms_colorop.c b/drivers/gpu/drm/vkms/vkms_colorop.c
index 7aceb0813b2d..5c3ffc78aea0 100644
--- a/drivers/gpu/drm/vkms/vkms_colorop.c
+++ b/drivers/gpu/drm/vkms/vkms_colorop.c
@@ -31,7 +31,8 @@ static int vkms_initialize_color_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, supported_tfs);
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, supported_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -48,7 +49,7 @@ static int vkms_initialize_color_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane);
+	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane, DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -64,7 +65,7 @@ static int vkms_initialize_color_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane);
+	ret = drm_plane_colorop_ctm_3x4_init(dev, ops[i], plane, DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
@@ -80,7 +81,8 @@ static int vkms_initialize_color_pipeline(struct drm_plane *plane, struct drm_pr
 		goto cleanup;
 	}
 
-	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, supported_tfs);
+	ret = drm_plane_colorop_curve_1d_init(dev, ops[i], plane, supported_tfs,
+					      DRM_COLOROP_FLAG_ALLOW_BYPASS);
 	if (ret)
 		goto cleanup;
 
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 0e1a5e9d26f3..828888861ad9 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -31,6 +31,9 @@
 #include <drm/drm_mode.h>
 #include <drm/drm_property.h>
 
+/* DRM colorop flags */
+#define DRM_COLOROP_FLAG_ALLOW_BYPASS	(1<<0)	/* Allow bypass on the drm_colorop */
+
 /**
  * enum drm_colorop_curve_1d_type - type of 1D curve
  *
@@ -256,11 +259,12 @@ struct drm_colorop {
 	 * @bypass_property:
 	 *
 	 * Boolean property to control enablement of the color
-	 * operation. Setting bypass to "true" shall always be supported
-	 * in order to allow compositors to quickly fall back to
-	 * alternate methods of color processing. This is important
-	 * since setting color operations can fail due to unique
-	 * HW constraints.
+	 * operation. Only present if DRM_COLOROP_FLAG_ALLOW_BYPASS
+	 * flag is set. When present, setting bypass to "true" shall
+	 * always be supported to allow compositors to quickly fall
+	 * back to alternate methods of color processing. This is
+	 * important since setting color operations can fail due to
+	 * unique HW constraints.
 	 */
 	struct drm_property *bypass_property;
 
@@ -353,14 +357,15 @@ void drm_colorop_pipeline_destroy(struct drm_device *dev);
 void drm_colorop_cleanup(struct drm_colorop *colorop);
 
 int drm_plane_colorop_curve_1d_init(struct drm_device *dev, struct drm_colorop *colorop,
-				    struct drm_plane *plane, u64 supported_tfs);
+				    struct drm_plane *plane, u64 supported_tfs, uint32_t flags);
 int drm_plane_colorop_curve_1d_lut_init(struct drm_device *dev, struct drm_colorop *colorop,
 					struct drm_plane *plane, uint32_t lut_size,
-					enum drm_colorop_lut1d_interpolation_type interpolation);
+					enum drm_colorop_lut1d_interpolation_type interpolation,
+					uint32_t flags);
 int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *colorop,
-				   struct drm_plane *plane);
+				   struct drm_plane *plane, uint32_t flags);
 int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colorop,
-				struct drm_plane *plane);
+				struct drm_plane *plane, uint32_t flags);
 
 struct drm_colorop_state *
 drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop);
-- 
cgit v1.2.3


From db971856bbe0263d0ba78d641ea66d4fcdfc8fc3 Mon Sep 17 00:00:00 2001
From: Alex Hung <alex.hung@amd.com>
Date: Fri, 14 Nov 2025 17:02:10 -0700
Subject: drm/colorop: Add 3D LUT support to color pipeline

It is to be used to enable HDR by allowing userpace to create and pass
3D LUTs to kernel and hardware.

new drm_colorop_type: DRM_COLOROP_3D_LUT.

Reviewed-by: Simon Ser <contact@emersion.fr>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Reviewed-by: Daniel Stone <daniels@collabora.com>
Reviewed-by: Melissa Wen <mwen@igalia.com>
Reviewed-by: Sebastian Wick <sebastian.wick@redhat.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-46-alex.hung@amd.com
---
 drivers/gpu/drm/drm_atomic.c      |  6 ++++
 drivers/gpu/drm/drm_atomic_uapi.c |  8 +++++
 drivers/gpu/drm/drm_colorop.c     | 72 +++++++++++++++++++++++++++++++++++++++
 include/drm/drm_colorop.h         | 23 +++++++++++++
 include/uapi/drm/drm_mode.h       | 34 ++++++++++++++++++
 5 files changed, 143 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c
index 436d2a948e41..67e095e398a3 100644
--- a/drivers/gpu/drm/drm_atomic.c
+++ b/drivers/gpu/drm/drm_atomic.c
@@ -807,6 +807,12 @@ static void drm_atomic_colorop_print_state(struct drm_printer *p,
 	case DRM_COLOROP_MULTIPLIER:
 		drm_printf(p, "\tmultiplier=%llu\n", state->multiplier);
 		break;
+	case DRM_COLOROP_3D_LUT:
+		drm_printf(p, "\tsize=%d\n", colorop->size);
+		drm_printf(p, "\tinterpolation=%s\n",
+			   drm_get_colorop_lut3d_interpolation_name(colorop->lut3d_interpolation));
+		drm_printf(p, "\tdata blob id=%d\n", state->data ? state->data->base.id : 0);
+		break;
 	default:
 		break;
 	}
diff --git a/drivers/gpu/drm/drm_atomic_uapi.c b/drivers/gpu/drm/drm_atomic_uapi.c
index 64e49338e284..7320db4b8489 100644
--- a/drivers/gpu/drm/drm_atomic_uapi.c
+++ b/drivers/gpu/drm/drm_atomic_uapi.c
@@ -705,6 +705,10 @@ static int drm_atomic_color_set_data_property(struct drm_colorop *colorop,
 	case DRM_COLOROP_CTM_3X4:
 		size = sizeof(struct drm_color_ctm_3x4);
 		break;
+	case DRM_COLOROP_3D_LUT:
+		size = colorop->size * colorop->size * colorop->size *
+		       sizeof(struct drm_color_lut32);
+		break;
 	default:
 		/* should never get here */
 		return -EINVAL;
@@ -732,6 +736,8 @@ static int drm_atomic_colorop_set_property(struct drm_colorop *colorop,
 		state->curve_1d_type = val;
 	} else if (property == colorop->multiplier_property) {
 		state->multiplier = val;
+	} else if (property == colorop->lut3d_interpolation_property) {
+		colorop->lut3d_interpolation = val;
 	} else if (property == colorop->data_property) {
 		return drm_atomic_color_set_data_property(colorop, state,
 							  property, val);
@@ -763,6 +769,8 @@ drm_atomic_colorop_get_property(struct drm_colorop *colorop,
 		*val = state->multiplier;
 	else if (property == colorop->size_property)
 		*val = colorop->size;
+	else if (property == colorop->lut3d_interpolation_property)
+		*val = colorop->lut3d_interpolation;
 	else if (property == colorop->data_property)
 		*val = (state->data) ? state->data->base.id : 0;
 	else
diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index fd666394f31f..272a73c28c0e 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -67,6 +67,7 @@ static const struct drm_prop_enum_list drm_colorop_type_enum_list[] = {
 	{ DRM_COLOROP_1D_LUT, "1D LUT" },
 	{ DRM_COLOROP_CTM_3X4, "3x4 Matrix"},
 	{ DRM_COLOROP_MULTIPLIER, "Multiplier"},
+	{ DRM_COLOROP_3D_LUT, "3D LUT"},
 };
 
 static const char * const colorop_curve_1d_type_names[] = {
@@ -82,6 +83,11 @@ static const struct drm_prop_enum_list drm_colorop_lut1d_interpolation_list[] =
 	{ DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR, "Linear" },
 };
 
+
+static const struct drm_prop_enum_list drm_colorop_lut3d_interpolation_list[] = {
+	{ DRM_COLOROP_LUT3D_INTERPOLATION_TETRAHEDRAL, "Tetrahedral" },
+};
+
 /* Init Helpers */
 
 static int drm_plane_colorop_init(struct drm_device *dev, struct drm_colorop *colorop,
@@ -381,6 +387,51 @@ int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colo
 }
 EXPORT_SYMBOL(drm_plane_colorop_mult_init);
 
+int drm_plane_colorop_3dlut_init(struct drm_device *dev, struct drm_colorop *colorop,
+				 struct drm_plane *plane,
+				 uint32_t lut_size,
+				 enum drm_colorop_lut3d_interpolation_type interpolation,
+				 uint32_t flags)
+{
+	struct drm_property *prop;
+	int ret;
+
+	ret = drm_plane_colorop_init(dev, colorop, plane, DRM_COLOROP_3D_LUT, flags);
+	if (ret)
+		return ret;
+
+	/* LUT size */
+	prop = drm_property_create_range(dev, DRM_MODE_PROP_IMMUTABLE  | DRM_MODE_PROP_ATOMIC,
+					 "SIZE", 0, UINT_MAX);
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->size_property = prop;
+	drm_object_attach_property(&colorop->base, colorop->size_property, lut_size);
+	colorop->size = lut_size;
+
+	/* interpolation */
+	prop = drm_property_create_enum(dev, 0, "LUT3D_INTERPOLATION",
+					drm_colorop_lut3d_interpolation_list,
+					ARRAY_SIZE(drm_colorop_lut3d_interpolation_list));
+	if (!prop)
+		return -ENOMEM;
+
+	colorop->lut3d_interpolation_property = prop;
+	drm_object_attach_property(&colorop->base, prop, interpolation);
+	colorop->lut3d_interpolation = interpolation;
+
+	/* data */
+	ret = drm_colorop_create_data_prop(dev, colorop);
+	if (ret)
+		return ret;
+
+	drm_colorop_reset(colorop);
+
+	return 0;
+}
+EXPORT_SYMBOL(drm_plane_colorop_3dlut_init);
+
 static void __drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop,
 							struct drm_colorop_state *state)
 {
@@ -472,7 +523,13 @@ static const char * const colorop_type_name[] = {
 	[DRM_COLOROP_1D_LUT] = "1D LUT",
 	[DRM_COLOROP_CTM_3X4] = "3x4 Matrix",
 	[DRM_COLOROP_MULTIPLIER] = "Multiplier",
+	[DRM_COLOROP_3D_LUT] = "3D LUT",
+};
+
+static const char * const colorop_lu3d_interpolation_name[] = {
+	[DRM_COLOROP_LUT3D_INTERPOLATION_TETRAHEDRAL] = "Tetrahedral",
 };
+
 static const char * const colorop_lut1d_interpolation_name[] = {
 	[DRM_COLOROP_LUT1D_INTERPOLATION_LINEAR] = "Linear",
 };
@@ -508,6 +565,21 @@ const char *drm_get_colorop_lut1d_interpolation_name(enum drm_colorop_lut1d_inte
 	return colorop_lut1d_interpolation_name[type];
 }
 
+/**
+ * drm_get_colorop_lut3d_interpolation_name - return a string for interpolation type
+ * @type: interpolation type to compute name of
+ *
+ * In contrast to the other drm_get_*_name functions this one here returns a
+ * const pointer and hence is threadsafe.
+ */
+const char *drm_get_colorop_lut3d_interpolation_name(enum drm_colorop_lut3d_interpolation_type type)
+{
+	if (WARN_ON(type >= ARRAY_SIZE(colorop_lu3d_interpolation_name)))
+		return "unknown";
+
+	return colorop_lu3d_interpolation_name[type];
+}
+
 /**
  * drm_colorop_set_next_property - sets the next pointer
  * @colorop: drm colorop
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 828888861ad9..9773e30e15ae 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -283,6 +283,14 @@ struct drm_colorop {
 	 */
 	enum drm_colorop_lut1d_interpolation_type lut1d_interpolation;
 
+	/**
+	 * @lut3d_interpolation:
+	 *
+	 * Read-only
+	 * Interpolation for DRM_COLOROP_3D_LUT
+	 */
+	enum drm_colorop_lut3d_interpolation_type lut3d_interpolation;
+
 	/**
 	 * @lut1d_interpolation_property:
 	 *
@@ -311,6 +319,13 @@ struct drm_colorop {
 	 */
 	struct drm_property *size_property;
 
+	/**
+	 * @lut3d_interpolation_property:
+	 *
+	 * Read-only property for DRM_COLOROP_3D_LUT interpolation
+	 */
+	struct drm_property *lut3d_interpolation_property;
+
 	/**
 	 * @data_property:
 	 *
@@ -366,6 +381,11 @@ int drm_plane_colorop_ctm_3x4_init(struct drm_device *dev, struct drm_colorop *c
 				   struct drm_plane *plane, uint32_t flags);
 int drm_plane_colorop_mult_init(struct drm_device *dev, struct drm_colorop *colorop,
 				struct drm_plane *plane, uint32_t flags);
+int drm_plane_colorop_3dlut_init(struct drm_device *dev, struct drm_colorop *colorop,
+				 struct drm_plane *plane,
+				 uint32_t lut_size,
+				 enum drm_colorop_lut3d_interpolation_type interpolation,
+				 uint32_t flags);
 
 struct drm_colorop_state *
 drm_atomic_helper_colorop_duplicate_state(struct drm_colorop *colorop);
@@ -418,6 +438,9 @@ const char *drm_get_colorop_curve_1d_type_name(enum drm_colorop_curve_1d_type ty
 const char *
 drm_get_colorop_lut1d_interpolation_name(enum drm_colorop_lut1d_interpolation_type type);
 
+const char *
+drm_get_colorop_lut3d_interpolation_name(enum drm_colorop_lut3d_interpolation_type type);
+
 void drm_colorop_set_next_property(struct drm_colorop *colorop, struct drm_colorop *next);
 
 #endif /* __DRM_COLOROP_H__ */
diff --git a/include/uapi/drm/drm_mode.h b/include/uapi/drm/drm_mode.h
index 4b38da880fc7..cbbbfc1dfe2b 100644
--- a/include/uapi/drm/drm_mode.h
+++ b/include/uapi/drm/drm_mode.h
@@ -942,6 +942,40 @@ enum drm_colorop_type {
 	 * property.
 	 */
 	DRM_COLOROP_MULTIPLIER,
+
+	/**
+	 * @DRM_COLOROP_3D_LUT:
+	 *
+	 * enum string "3D LUT"
+	 *
+	 * A 3D LUT of &drm_color_lut32 entries,
+	 * packed into a blob via the DATA property. The driver's expected
+	 * LUT size is advertised via the SIZE property, i.e., a 3D LUT with
+	 * 17x17x17 entries will have SIZE set to 17.
+	 *
+	 * The DATA blob is a 3D array of struct drm_color_lut32 with dimension
+	 * length of "size".
+	 * The LUT elements are traversed like so:
+	 *
+	 *   for B in range 0..n
+	 *     for G in range 0..n
+	 *       for R in range 0..n
+	 *        index = R + n * (G + n * B)
+	 *         color = lut3d[index]
+	 */
+	DRM_COLOROP_3D_LUT,
+};
+
+/**
+ * enum drm_colorop_lut3d_interpolation_type - type of 3DLUT interpolation
+ */
+enum drm_colorop_lut3d_interpolation_type {
+	/**
+	 * @DRM_COLOROP_LUT3D_INTERPOLATION_TETRAHEDRAL:
+	 *
+	 * Tetrahedral 3DLUT interpolation
+	 */
+	DRM_COLOROP_LUT3D_INTERPOLATION_TETRAHEDRAL,
 };
 
 /**
-- 
cgit v1.2.3


From 8e304a45116a567a33cfbd9f104356622c6aefbc Mon Sep 17 00:00:00 2001
From: Alex Hung <alex.hung@amd.com>
Date: Fri, 14 Nov 2025 17:02:15 -0700
Subject: drm/colorop: Add DRM_COLOROP_1D_CURVE_GAMMA22 to 1D Curve

Add "DRM_COLOROP_1D_CURVE_GAMMA22" and DRM_COLOROP_1D_CURVE_GAMMA22_INV
subtypes to drm_colorop of DRM_COLOROP_1D_CURVE.

Reviewed-by: Harry Wentland <harry.wentland@amd.com>
Signed-off-by: Alex Hung <alex.hung@amd.com>
Signed-off-by: Simon Ser <contact@emersion.fr>
Link: https://patch.msgid.link/20251115000237.3561250-51-alex.hung@amd.com
---
 drivers/gpu/drm/drm_colorop.c |  2 ++
 include/drm/drm_colorop.h     | 18 ++++++++++++++++++
 2 files changed, 20 insertions(+)

(limited to 'include')

diff --git a/drivers/gpu/drm/drm_colorop.c b/drivers/gpu/drm/drm_colorop.c
index 272a73c28c0e..44eb823585d2 100644
--- a/drivers/gpu/drm/drm_colorop.c
+++ b/drivers/gpu/drm/drm_colorop.c
@@ -77,6 +77,8 @@ static const char * const colorop_curve_1d_type_names[] = {
 	[DRM_COLOROP_1D_CURVE_PQ_125_INV_EOTF] = "PQ 125 Inverse EOTF",
 	[DRM_COLOROP_1D_CURVE_BT2020_INV_OETF] = "BT.2020 Inverse OETF",
 	[DRM_COLOROP_1D_CURVE_BT2020_OETF] = "BT.2020 OETF",
+	[DRM_COLOROP_1D_CURVE_GAMMA22] = "Gamma 2.2",
+	[DRM_COLOROP_1D_CURVE_GAMMA22_INV] = "Gamma 2.2 Inverse",
 };
 
 static const struct drm_prop_enum_list drm_colorop_lut1d_interpolation_list[] = {
diff --git a/include/drm/drm_colorop.h b/include/drm/drm_colorop.h
index 9773e30e15ae..a3a32f9f918c 100644
--- a/include/drm/drm_colorop.h
+++ b/include/drm/drm_colorop.h
@@ -108,6 +108,24 @@ enum drm_colorop_curve_1d_type {
 	 */
 	DRM_COLOROP_1D_CURVE_BT2020_OETF,
 
+	/**
+	 * @DRM_COLOROP_1D_CURVE_GAMMA22:
+	 *
+	 * enum string "Gamma 2.2"
+	 *
+	 * A gamma 2.2 power function. This applies a power curve with
+	 * gamma value of 2.2 to the input values.
+	 */
+	DRM_COLOROP_1D_CURVE_GAMMA22,
+
+	/**
+	 * @DRM_COLOROP_1D_CURVE_GAMMA22_INV:
+	 *
+	 * enum string "Gamma 2.2 Inverse"
+	 *
+	 * The inverse of &DRM_COLOROP_1D_CURVE_GAMMA22
+	 */
+	DRM_COLOROP_1D_CURVE_GAMMA22_INV,
 	/**
 	 * @DRM_COLOROP_1D_CURVE_COUNT:
 	 *
-- 
cgit v1.2.3