Merge tag 'drm-misc-next-2026-02-26' of https://gitlab.freedesktop.org/drm/misc/kernel into drm-next

drm-misc-next for v7.1:

UAPI Changes:

connector:
- Add panel_type property

fourcc:
- Add ARM interleaved 64k modifier

nouveau:
- Query Z-Cull info with DRM_IOCTL_NOUVEAU_GET_ZCULL_INFO

Cross-subsystem Changes:

coreboot:
- Clean up coreboot framebuffer support

dma-buf:
- Provide revoke mechanism for shared buffers
- Rename move_notify callback to invalidate_mappings and update users.
- Always enable move_notify
- Support dma_fence_was_initialized() test
- Protect dma_fence_ops by RCU and improve locking
- Fix sparse warnings

Core Changes:

atomic:
- Allocate drm_private_state via callback and convert drivers

atomic-helper:
- Use system_percpu_wq

buddy:
- Make buddy allocator available to all DRM drivers
- Document flags and structures

colorop:
- Add destroy helper and convert drivers

fbdev-emulation:
- Clean up

gem:
- Fix drm_gem_objects_lookup() error cleanup

Driver Changes:

amdgpu:
- Set panel_type to OELD for eDP

atmel-hlcdc:
- Support sana5d65 LCD controller

bridge:
- anx7625: Support USB-C plus DT bindings
- connector: Fix EDID detection
- dw-hdmi-qp: Support Vendor-Specfic and SDP Infoframes; improve others
- fsl-ldb: Fix visual artifacts plus related DT property 'enable-termination-resistor'
- imx8qxp-pixel-link: Improve bridge reference handling
- lt9611: Support Port-B-only input plus DT bindings
- tda998x: Support DRM_BRIDGE_ATTACH_NO_CONNECTOR; Clean up
- Support TH1520 HDMI plus DT bindings
- Clean up

imagination:
- Clean up

komeda:
- Fix integer overflow in AFBC checks

mcde:
- Improve bridge handling

nouveau:
- Provide Z-cull info to user space
- gsp: Support GA100
- Shutdown on PCI device shutdown
- Clean up

panel:
- panel-jdi-lt070me05000: Use mipi-dsi multi functions
- panel-edp: Support Add AUO B116XAT04.1 (HW: 1A); Support CMN N116BCL-EAK (C2); Support FriendlyELEC plus DT changes
- Fix Kconfig dependencies

panthor:
- Add tracepoints for power and IRQs

rcar-du:
- dsi: fix VCLK calculation

rockchip:
- vop2: Use drm_ logging functions
- Support DisplayPort on RK3576

sysfb:
- corebootdrm: Support system framebuffer on coreboot firmware; detect orientation
- Clean up pixel-format lookup

sun4i:
- Clean up

tilcdc:
- Use DT bindings scheme
- Use managed DRM interfaces
- Support DRM_BRIDGE_ATTACH_NO_CONNECTOR
- Clean up a lot of obsolete code

v3d:
- Clean up

vc4:
- Use system_percpu_wq
- Clean up

verisilicon:
- Support DC8200 plus DT bindings

virtgpu:
- Support PRIME imports with enabled 3D

Signed-off-by: Dave Airlie <airlied@redhat.com>

From: Thomas Zimmermann <tzimmermann@suse.de>
Link: https://patch.msgid.link/20260226143615.GA47200@linux.fritz.box

6 files changed, 426 insertions, 19 deletions

diff --git a/include/linux/coreboot.h b/include/linux/coreboot.h
new file mode 100644
index 000000000000..5d40ca7a1d89
--- /dev/null
+++ b/include/linux/coreboot.h
@@ -0,0 +1,90 @@
+/* SPDX-License-Identifier: GPL-2.0-only */
+/*
+ * coreboot.h
+ *
+ * Coreboot device and driver interfaces.
+ *
+ * Copyright 2014 Gerd Hoffmann <kraxel@redhat.com>
+ * Copyright 2017 Google Inc.
+ * Copyright 2017 Samuel Holland <samuel@sholland.org>
+ */
+
+#ifndef _LINUX_COREBOOT_H
+#define _LINUX_COREBOOT_H
+
+#include <linux/compiler_attributes.h>
+#include <linux/stddef.h>
+#include <linux/types.h>
+
+typedef __aligned(4) u64 cb_u64;
+
+/* List of coreboot entry structures that is used */
+
+#define CB_TAG_FRAMEBUFFER 0x12
+#define LB_TAG_CBMEM_ENTRY 0x31
+
+/* Generic */
+struct coreboot_table_entry {
+	u32 tag;
+	u32 size;
+};
+
+/* Points to a CBMEM entry */
+struct lb_cbmem_ref {
+	u32 tag;
+	u32 size;
+
+	cb_u64 cbmem_addr;
+};
+
+/* Corresponds to LB_TAG_CBMEM_ENTRY */
+struct lb_cbmem_entry {
+	u32 tag;
+	u32 size;
+
+	cb_u64 address;
+	u32 entry_size;
+	u32 id;
+};
+
+#define LB_FRAMEBUFFER_ORIENTATION_NORMAL	0
+#define LB_FRAMEBUFFER_ORIENTATION_BOTTOM_UP	1
+#define LB_FRAMEBUFFER_ORIENTATION_LEFT_UP	2
+#define LB_FRAMEBUFFER_ORIENTATION_RIGHT_UP	3
+
+/* Describes framebuffer setup by coreboot */
+struct lb_framebuffer {
+	u32 tag;
+	u32 size;
+
+	cb_u64 physical_address;
+	u32 x_resolution;
+	u32 y_resolution;
+	u32 bytes_per_line;
+	u8  bits_per_pixel;
+	u8  red_mask_pos;
+	u8  red_mask_size;
+	u8  green_mask_pos;
+	u8  green_mask_size;
+	u8  blue_mask_pos;
+	u8  blue_mask_size;
+	u8  reserved_mask_pos;
+	u8  reserved_mask_size;
+	u8  orientation;
+};
+
+/*
+ * True if the coreboot-provided data is large enough to hold information
+ * on the linear framebuffer. False otherwise.
+ */
+#define LB_FRAMEBUFFER_HAS_LFB(__fb) \
+	((__fb)->size >= offsetofend(struct lb_framebuffer, reserved_mask_size))
+
+/*
+ * True if the coreboot-provided data is large enough to hold information
+ * on the display orientation. False otherwise.
+ */
+#define LB_FRAMEBUFFER_HAS_ORIENTATION(__fb) \
+	((__fb)->size >= offsetofend(struct lb_framebuffer, orientation))
+
+#endif /* _LINUX_COREBOOT_H */
diff --git a/include/linux/dma-buf.h b/include/linux/dma-buf.h
index 133b9e637b55..166933b82e27 100644
--- a/include/linux/dma-buf.h
+++ b/include/linux/dma-buf.h
@@ -407,7 +407,7 @@ struct dma_buf {
 	 *   through the device.
 	 *
 	 * - Dynamic importers should set fences for any access that they can't
-	 *   disable immediately from their &dma_buf_attach_ops.move_notify
+	 *   disable immediately from their &dma_buf_attach_ops.invalidate_mappings
 	 *   callback.
 	 *
 	 * IMPORTANT:
@@ -446,7 +446,7 @@ struct dma_buf_attach_ops {
 	bool allow_peer2peer;
 
 	/**
-	 * @move_notify: [optional] notification that the DMA-buf is moving
+	 * @invalidate_mappings: [optional] notification that the DMA-buf is moving
 	 *
 	 * If this callback is provided the framework can avoid pinning the
 	 * backing store while mappings exists.
@@ -456,14 +456,10 @@ struct dma_buf_attach_ops {
 	 * called with this lock held as well. This makes sure that no mapping
 	 * is created concurrently with an ongoing move operation.
 	 *
-	 * Mappings stay valid and are not directly affected by this callback.
-	 * But the DMA-buf can now be in a different physical location, so all
-	 * mappings should be destroyed and re-created as soon as possible.
-	 *
-	 * New mappings can be created after this callback returns, and will
-	 * point to the new location of the DMA-buf.
+	 * See the kdoc for dma_buf_invalidate_mappings() for details on the
+	 * required behavior.
 	 */
-	void (*move_notify)(struct dma_buf_attachment *attach);
+	void (*invalidate_mappings)(struct dma_buf_attachment *attach);
 };
 
 /**
@@ -578,7 +574,8 @@ struct sg_table *dma_buf_map_attachment(struct dma_buf_attachment *,
 					enum dma_data_direction);
 void dma_buf_unmap_attachment(struct dma_buf_attachment *, struct sg_table *,
 				enum dma_data_direction);
-void dma_buf_move_notify(struct dma_buf *dma_buf);
+void dma_buf_invalidate_mappings(struct dma_buf *dma_buf);
+bool dma_buf_attach_revocable(struct dma_buf_attachment *attach);
 int dma_buf_begin_cpu_access(struct dma_buf *dma_buf,
 			     enum dma_data_direction dir);
 int dma_buf_end_cpu_access(struct dma_buf *dma_buf,
diff --git a/include/linux/dma-fence-array.h b/include/linux/dma-fence-array.h
index 079b3dec0a16..370b3d2bba37 100644
--- a/include/linux/dma-fence-array.h
+++ b/include/linux/dma-fence-array.h
@@ -38,7 +38,6 @@ struct dma_fence_array_cb {
 struct dma_fence_array {
 	struct dma_fence base;
 
-	spinlock_t lock;
 	unsigned num_fences;
 	atomic_t num_pending;
 	struct dma_fence **fences;
diff --git a/include/linux/dma-fence-chain.h b/include/linux/dma-fence-chain.h
index 5cd3ba53b4a1..df3beadf1515 100644
--- a/include/linux/dma-fence-chain.h
+++ b/include/linux/dma-fence-chain.h
@@ -46,7 +46,6 @@ struct dma_fence_chain {
 		 */
 		struct irq_work work;
 	};
-	spinlock_t lock;
 };
 
 
diff --git a/include/linux/dma-fence.h b/include/linux/dma-fence.h
index d4c92fd35092..3dc93f068bf6 100644
--- a/include/linux/dma-fence.h
+++ b/include/linux/dma-fence.h
@@ -34,7 +34,8 @@ struct seq_file;
  * @ops: dma_fence_ops associated with this fence
  * @rcu: used for releasing fence with kfree_rcu
  * @cb_list: list of all callbacks to call
- * @lock: spin_lock_irqsave used for locking
+ * @extern_lock: external spin_lock_irqsave used for locking (deprecated)
+ * @inline_lock: alternative internal spin_lock_irqsave used for locking
  * @context: execution context this fence belongs to, returned by
  *           dma_fence_context_alloc()
  * @seqno: the sequence number of this fence inside the execution context,
@@ -48,6 +49,8 @@ struct seq_file;
  * atomic ops (bit_*), so taking the spinlock will not be needed most
  * of the time.
  *
+ * DMA_FENCE_FLAG_INITIALIZED_BIT - fence was initialized
+ * DMA_FENCE_FLAG_INLINE_LOCK_BIT - use inline spinlock instead of external one
  * DMA_FENCE_FLAG_SIGNALED_BIT - fence is already signaled
  * DMA_FENCE_FLAG_TIMESTAMP_BIT - timestamp recorded for fence signaling
  * DMA_FENCE_FLAG_ENABLE_SIGNAL_BIT - enable_signaling might have been called
@@ -65,8 +68,11 @@ struct seq_file;
  * been completed, or never called at all.
  */
 struct dma_fence {
-	spinlock_t *lock;
-	const struct dma_fence_ops *ops;
+	union {
+		spinlock_t *extern_lock;
+		spinlock_t inline_lock;
+	};
+	const struct dma_fence_ops __rcu *ops;
 	/*
 	 * We clear the callback list on kref_put so that by the time we
 	 * release the fence it is unused. No one should be adding to the
@@ -98,6 +104,8 @@ struct dma_fence {
 };
 
 enum dma_fence_flag_bits {
+	DMA_FENCE_FLAG_INITIALIZED_BIT,
+	DMA_FENCE_FLAG_INLINE_LOCK_BIT,
 	DMA_FENCE_FLAG_SEQNO64_BIT,
 	DMA_FENCE_FLAG_SIGNALED_BIT,
 	DMA_FENCE_FLAG_TIMESTAMP_BIT,
@@ -218,6 +226,10 @@ struct dma_fence_ops {
 	 * timed out. Can also return other error values on custom implementations,
 	 * which should be treated as if the fence is signaled. For example a hardware
 	 * lockup could be reported like that.
+	 *
+	 * Implementing this callback prevents the fence from detaching after
+	 * signaling and so it is necessary for the module providing the
+	 * dma_fence_ops to stay loaded as long as the dma_fence exists.
 	 */
 	signed long (*wait)(struct dma_fence *fence,
 			    bool intr, signed long timeout);
@@ -229,6 +241,13 @@ struct dma_fence_ops {
 	 * Can be called from irq context.  This callback is optional. If it is
 	 * NULL, then dma_fence_free() is instead called as the default
 	 * implementation.
+	 *
+	 * Implementing this callback prevents the fence from detaching after
+	 * signaling and so it is necessary for the module providing the
+	 * dma_fence_ops to stay loaded as long as the dma_fence exists.
+	 *
+	 * If the callback is implemented the memory backing the dma_fence
+	 * object must be freed RCU safe.
 	 */
 	void (*release)(struct dma_fence *fence);
 
@@ -264,6 +283,19 @@ void dma_fence_free(struct dma_fence *fence);
 void dma_fence_describe(struct dma_fence *fence, struct seq_file *seq);
 
 /**
+ * dma_fence_was_initialized - test if fence was initialized
+ * @fence: fence to test
+ *
+ * Return: True if fence was ever initialized, false otherwise. Works correctly
+ * only when memory backing the fence structure is zero initialized on
+ * allocation.
+ */
+static inline bool dma_fence_was_initialized(struct dma_fence *fence)
+{
+	return fence && test_bit(DMA_FENCE_FLAG_INITIALIZED_BIT, &fence->flags);
+}
+
+/**
  * dma_fence_put - decreases refcount of the fence
  * @fence: fence to reduce refcount of
  */
@@ -351,6 +383,45 @@ dma_fence_get_rcu_safe(struct dma_fence __rcu **fencep)
 	} while (1);
 }
 
+/**
+ * dma_fence_spinlock - return pointer to the spinlock protecting the fence
+ * @fence: the fence to get the lock from
+ *
+ * Return either the pointer to the embedded or the external spin lock.
+ */
+static inline spinlock_t *dma_fence_spinlock(struct dma_fence *fence)
+{
+	return test_bit(DMA_FENCE_FLAG_INLINE_LOCK_BIT, &fence->flags) ?
+		&fence->inline_lock : fence->extern_lock;
+}
+
+/**
+ * dma_fence_lock_irqsave - irqsave lock the fence
+ * @fence: the fence to lock
+ * @flags: where to store the CPU flags.
+ *
+ * Lock the fence, preventing it from changing to the signaled state.
+ */
+#define dma_fence_lock_irqsave(fence, flags)	\
+	spin_lock_irqsave(dma_fence_spinlock(fence), flags)
+
+/**
+ * dma_fence_unlock_irqrestore - unlock the fence and irqrestore
+ * @fence: the fence to unlock
+ * @flags the CPU flags to restore
+ *
+ * Unlock the fence, allowing it to change it's state to signaled again.
+ */
+#define dma_fence_unlock_irqrestore(fence, flags)	\
+	spin_unlock_irqrestore(dma_fence_spinlock(fence), flags)
+
+/**
+ * dma_fence_assert_held - lockdep assertion that fence is locked
+ * @fence: the fence which should be locked
+ */
+#define dma_fence_assert_held(fence)	\
+	lockdep_assert_held(dma_fence_spinlock(fence));
+
 #ifdef CONFIG_LOCKDEP
 bool dma_fence_begin_signalling(void);
 void dma_fence_end_signalling(bool cookie);
@@ -439,13 +510,19 @@ dma_fence_test_signaled_flag(struct dma_fence *fence)
 static inline bool
 dma_fence_is_signaled_locked(struct dma_fence *fence)
 {
+	const struct dma_fence_ops *ops;
+
 	if (dma_fence_test_signaled_flag(fence))
 		return true;
 
-	if (fence->ops->signaled && fence->ops->signaled(fence)) {
+	rcu_read_lock();
+	ops = rcu_dereference(fence->ops);
+	if (ops && ops->signaled && ops->signaled(fence)) {
+		rcu_read_unlock();
 		dma_fence_signal_locked(fence);
 		return true;
 	}
+	rcu_read_unlock();
 
 	return false;
 }
@@ -469,13 +546,19 @@ dma_fence_is_signaled_locked(struct dma_fence *fence)
 static inline bool
 dma_fence_is_signaled(struct dma_fence *fence)
 {
+	const struct dma_fence_ops *ops;
+
 	if (dma_fence_test_signaled_flag(fence))
 		return true;
 
-	if (fence->ops->signaled && fence->ops->signaled(fence)) {
+	rcu_read_lock();
+	ops = rcu_dereference(fence->ops);
+	if (ops && ops->signaled && ops->signaled(fence)) {
+		rcu_read_unlock();
 		dma_fence_signal(fence);
 		return true;
 	}
+	rcu_read_unlock();
 
 	return false;
 }
@@ -680,7 +763,7 @@ extern const struct dma_fence_ops dma_fence_chain_ops;
  */
 static inline bool dma_fence_is_array(struct dma_fence *fence)
 {
-	return fence->ops == &dma_fence_array_ops;
+	return rcu_access_pointer(fence->ops) == &dma_fence_array_ops;
 }
 
 /**
@@ -691,7 +774,7 @@ static inline bool dma_fence_is_array(struct dma_fence *fence)
  */
 static inline bool dma_fence_is_chain(struct dma_fence *fence)
 {
-	return fence->ops == &dma_fence_chain_ops;
+	return rcu_access_pointer(fence->ops) == &dma_fence_chain_ops;
 }
 
 /**
diff --git a/include/linux/gpu_buddy.h b/include/linux/gpu_buddy.h
new file mode 100644
index 000000000000..f1fb6eff604a
--- /dev/null
+++ b/include/linux/gpu_buddy.h
@@ -0,0 +1,239 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2021 Intel Corporation
+ */
+
+#ifndef __GPU_BUDDY_H__
+#define __GPU_BUDDY_H__
+
+#include <linux/bitops.h>
+#include <linux/list.h>
+#include <linux/slab.h>
+#include <linux/sched.h>
+#include <linux/rbtree.h>
+
+/**
+ * GPU_BUDDY_RANGE_ALLOCATION - Allocate within a specific address range
+ *
+ * When set, allocation is restricted to the range [start, end) specified
+ * in gpu_buddy_alloc_blocks(). Without this flag, start/end are ignored
+ * and allocation can use any free space.
+ */
+#define GPU_BUDDY_RANGE_ALLOCATION		BIT(0)
+
+/**
+ * GPU_BUDDY_TOPDOWN_ALLOCATION - Allocate from top of address space
+ *
+ * Allocate starting from high addresses and working down. Useful for
+ * separating different allocation types (e.g., kernel vs userspace)
+ * to reduce fragmentation.
+ */
+#define GPU_BUDDY_TOPDOWN_ALLOCATION		BIT(1)
+
+/**
+ * GPU_BUDDY_CONTIGUOUS_ALLOCATION - Require physically contiguous blocks
+ *
+ * The allocation must be satisfied with a single contiguous block.
+ * If the requested size cannot be allocated contiguously, the
+ * allocation fails with -ENOSPC.
+ */
+#define GPU_BUDDY_CONTIGUOUS_ALLOCATION		BIT(2)
+
+/**
+ * GPU_BUDDY_CLEAR_ALLOCATION - Prefer pre-cleared (zeroed) memory
+ *
+ * Attempt to allocate from the clear tree first. If insufficient clear
+ * memory is available, falls back to dirty memory. Useful when the
+ * caller needs zeroed memory and wants to avoid GPU clear operations.
+ */
+#define GPU_BUDDY_CLEAR_ALLOCATION		BIT(3)
+
+/**
+ * GPU_BUDDY_CLEARED - Mark returned blocks as cleared
+ *
+ * Used with gpu_buddy_free_list() to indicate that the memory being
+ * freed has been cleared (zeroed). The blocks will be placed in the
+ * clear tree for future GPU_BUDDY_CLEAR_ALLOCATION requests.
+ */
+#define GPU_BUDDY_CLEARED			BIT(4)
+
+/**
+ * GPU_BUDDY_TRIM_DISABLE - Disable automatic block trimming
+ *
+ * By default, if an allocation is smaller than the allocated block,
+ * excess memory is trimmed and returned to the free pool. This flag
+ * disables trimming, keeping the full power-of-two block size.
+ */
+#define GPU_BUDDY_TRIM_DISABLE			BIT(5)
+
+enum gpu_buddy_free_tree {
+	GPU_BUDDY_CLEAR_TREE = 0,
+	GPU_BUDDY_DIRTY_TREE,
+	GPU_BUDDY_MAX_FREE_TREES,
+};
+
+#define for_each_free_tree(tree) \
+	for ((tree) = 0; (tree) < GPU_BUDDY_MAX_FREE_TREES; (tree)++)
+
+/**
+ * struct gpu_buddy_block - Block within a buddy allocator
+ *
+ * Each block in the buddy allocator is represented by this structure.
+ * Blocks are organized in a binary tree where each parent block can be
+ * split into two children (left and right buddies). The allocator manages
+ * blocks at various orders (power-of-2 sizes) from chunk_size up to the
+ * largest contiguous region.
+ *
+ * @private: Private data owned by the allocator user (e.g., driver-specific data)
+ * @link: List node for user ownership while block is allocated
+ */
+struct gpu_buddy_block {
+/* private: */
+	/*
+	 * Header bit layout:
+	 * - Bits 63:12: block offset within the address space
+	 * - Bits 11:10: state (ALLOCATED, FREE, or SPLIT)
+	 * - Bit 9: clear bit (1 if memory is zeroed)
+	 * - Bits 8:6: reserved
+	 * - Bits 5:0: order (log2 of size relative to chunk_size)
+	 */
+#define GPU_BUDDY_HEADER_OFFSET GENMASK_ULL(63, 12)
+#define GPU_BUDDY_HEADER_STATE  GENMASK_ULL(11, 10)
+#define   GPU_BUDDY_ALLOCATED	   (1 << 10)
+#define   GPU_BUDDY_FREE	   (2 << 10)
+#define   GPU_BUDDY_SPLIT	   (3 << 10)
+#define GPU_BUDDY_HEADER_CLEAR  GENMASK_ULL(9, 9)
+/* Free to be used, if needed in the future */
+#define GPU_BUDDY_HEADER_UNUSED GENMASK_ULL(8, 6)
+#define GPU_BUDDY_HEADER_ORDER  GENMASK_ULL(5, 0)
+	u64 header;
+
+	struct gpu_buddy_block *left;
+	struct gpu_buddy_block *right;
+	struct gpu_buddy_block *parent;
+/* public: */
+	void *private; /* owned by creator */
+
+	/*
+	 * While the block is allocated by the user through gpu_buddy_alloc*,
+	 * the user has ownership of the link, for example to maintain within
+	 * a list, if so desired. As soon as the block is freed with
+	 * gpu_buddy_free* ownership is given back to the mm.
+	 */
+	union {
+/* private: */
+		struct rb_node rb;
+/* public: */
+		struct list_head link;
+	};
+/* private: */
+	struct list_head tmp_link;
+};
+
+/* Order-zero must be at least SZ_4K */
+#define GPU_BUDDY_MAX_ORDER (63 - 12)
+
+/**
+ * struct gpu_buddy - GPU binary buddy allocator
+ *
+ * The buddy allocator provides efficient power-of-two memory allocation
+ * with fast allocation and free operations. It is commonly used for GPU
+ * memory management where allocations can be split into power-of-two
+ * block sizes.
+ *
+ * Locking should be handled by the user; a simple mutex around
+ * gpu_buddy_alloc_blocks() and gpu_buddy_free_block()/gpu_buddy_free_list()
+ * should suffice.
+ *
+ * @n_roots: Number of root blocks in the roots array.
+ * @max_order: Maximum block order (log2 of largest block size / chunk_size).
+ * @chunk_size: Minimum allocation granularity in bytes. Must be at least SZ_4K.
+ * @size: Total size of the address space managed by this allocator in bytes.
+ * @avail: Total free space currently available for allocation in bytes.
+ * @clear_avail: Free space available in the clear tree (zeroed memory) in bytes.
+ *               This is a subset of @avail.
+ */
+struct gpu_buddy {
+/* private: */
+	/*
+	 * Array of red-black trees for free block management.
+	 * Indexed as free_trees[clear/dirty][order] where:
+	 * - Index 0 (GPU_BUDDY_CLEAR_TREE): blocks with zeroed content
+	 * - Index 1 (GPU_BUDDY_DIRTY_TREE): blocks with unknown content
+	 * Each tree holds free blocks of the corresponding order.
+	 */
+	struct rb_root **free_trees;
+	/*
+	 * Array of root blocks representing the top-level blocks of the
+	 * binary tree(s). Multiple roots exist when the total size is not
+	 * a power of two, with each root being the largest power-of-two
+	 * that fits in the remaining space.
+	 */
+	struct gpu_buddy_block **roots;
+/* public: */
+	unsigned int n_roots;
+	unsigned int max_order;
+	u64 chunk_size;
+	u64 size;
+	u64 avail;
+	u64 clear_avail;
+};
+
+static inline u64
+gpu_buddy_block_offset(const struct gpu_buddy_block *block)
+{
+	return block->header & GPU_BUDDY_HEADER_OFFSET;
+}
+
+static inline unsigned int
+gpu_buddy_block_order(struct gpu_buddy_block *block)
+{
+	return block->header & GPU_BUDDY_HEADER_ORDER;
+}
+
+static inline bool
+gpu_buddy_block_is_free(struct gpu_buddy_block *block)
+{
+	return (block->header & GPU_BUDDY_HEADER_STATE) == GPU_BUDDY_FREE;
+}
+
+static inline bool
+gpu_buddy_block_is_clear(struct gpu_buddy_block *block)
+{
+	return block->header & GPU_BUDDY_HEADER_CLEAR;
+}
+
+static inline u64
+gpu_buddy_block_size(struct gpu_buddy *mm,
+		     struct gpu_buddy_block *block)
+{
+	return mm->chunk_size << gpu_buddy_block_order(block);
+}
+
+int gpu_buddy_init(struct gpu_buddy *mm, u64 size, u64 chunk_size);
+
+void gpu_buddy_fini(struct gpu_buddy *mm);
+
+int gpu_buddy_alloc_blocks(struct gpu_buddy *mm,
+			   u64 start, u64 end, u64 size,
+			   u64 min_page_size,
+			   struct list_head *blocks,
+			   unsigned long flags);
+
+int gpu_buddy_block_trim(struct gpu_buddy *mm,
+			 u64 *start,
+			 u64 new_size,
+			 struct list_head *blocks);
+
+void gpu_buddy_reset_clear(struct gpu_buddy *mm, bool is_clear);
+
+void gpu_buddy_free_block(struct gpu_buddy *mm, struct gpu_buddy_block *block);
+
+void gpu_buddy_free_list(struct gpu_buddy *mm,
+			 struct list_head *objects,
+			 unsigned int flags);
+
+void gpu_buddy_print(struct gpu_buddy *mm);
+void gpu_buddy_block_print(struct gpu_buddy *mm,
+			   struct gpu_buddy_block *block);
+#endif