[PATCH v3 00/30] Introduce GPU SVM and Xe SVM implementation

[PATCH v3 00/30] Introduce GPU SVM and Xe SVM implementation

[Matthew Brost]

Version 3 of GPU SVM has been promoted to the proper series from an RFC.
Thanks to everyone (especially Sima and Thomas) for their numerous
reviews on revision 1, 2 and for helping to address many design issues.

This version has been tested with IGT [1] on PVC, BMG, and LNL. Also
tested with level0 (UMD) PR [2].

Major changes in v2:
- Dropped mmap write abuse
- core MM locking and retry loops instead of driver locking to avoid races
- Removed physical to virtual references
- Embedded structure/ops for drm_gpusvm_devmem
- Fixed mremap and fork issues
- Added DRM pagemap
- Included RFC documentation in the kernel doc

Major changes in v3:
- Move GPU SVM and DRM pagemap to DRM level
- Mostly addresses Thomas's feedback, lots of small changes documented
  in each individual patch change log

Known issues in v3:
- Check pages still exists, changed to threshold in this version which
  is better but still need to root cause cross process page finding on
  small user allocations.
- Dropped documentation patch, fairly large rewrite and will send out
  independently once finished.

Matt

[1] https://patchwork.freedesktop.org/series/137545/#rev3
[2] https://github.com/intel/compute-runtime/pull/782

Matthew Brost (27):
  drm/xe: Retry BO allocation
  mm/migrate: Add migrate_device_pfns
  mm/migrate: Trylock device page in do_swap_page
  drm/gpusvm: Add support for GPU Shared Virtual Memory
  drm/xe: Select DRM_GPUSVM Kconfig
  drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag
  drm/xe: Add SVM init / close / fini to faulting VMs
  drm/xe: Nuke VM's mapping upon close
  drm/xe: Add SVM range invalidation and page fault handler
  drm/gpuvm: Add DRM_GPUVA_OP_DRIVER
  drm/xe: Add (re)bind to SVM page fault handler
  drm/xe: Add SVM garbage collector
  drm/xe: Add unbind to SVM garbage collector
  drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has
    bindings
  drm/xe: Enable CPU address mirror uAPI
  drm/xe: Add migrate layer functions for SVM support
  drm/xe: Add SVM device memory mirroring
  drm/xe: Add drm_gpusvm_devmem to xe_bo
  drm/xe: Add GPUSVM device memory copy vfunc functions
  drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc
  drm/xe: Add Xe SVM devmem_release GPU SVM vfunc
  drm/xe: Add BO flags required for SVM
  drm/xe: Add SVM VRAM migration
  drm/xe: Basic SVM BO eviction
  drm/xe: Add SVM debug
  drm/xe: Add modparam for SVM notifier size
  drm/xe: Add always_migrate_to_vram modparam

Thomas Hellström (3):
  drm/pagemap: Add DRM pagemap
  drm/xe: Add dma_addr res cursor
  drm/xe: Add drm_pagemap ops to SVM

 drivers/gpu/drm/Kconfig                     |    8 +
 drivers/gpu/drm/Makefile                    |    1 +
 drivers/gpu/drm/drm_gpusvm.c                | 2220 +++++++++++++++++++
 drivers/gpu/drm/xe/Kconfig                  |   10 +
 drivers/gpu/drm/xe/Makefile                 |    1 +
 drivers/gpu/drm/xe/xe_bo.c                  |   20 +-
 drivers/gpu/drm/xe/xe_bo.h                  |    1 +
 drivers/gpu/drm/xe/xe_bo_types.h            |    4 +
 drivers/gpu/drm/xe/xe_device_types.h        |   15 +
 drivers/gpu/drm/xe/xe_gt_pagefault.c        |   17 +-
 drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c |   24 +
 drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h |    2 +
 drivers/gpu/drm/xe/xe_migrate.c             |  175 ++
 drivers/gpu/drm/xe/xe_migrate.h             |   10 +
 drivers/gpu/drm/xe/xe_module.c              |    7 +
 drivers/gpu/drm/xe/xe_module.h              |    2 +
 drivers/gpu/drm/xe/xe_pt.c                  |  393 +++-
 drivers/gpu/drm/xe/xe_pt.h                  |    5 +
 drivers/gpu/drm/xe/xe_pt_types.h            |    2 +
 drivers/gpu/drm/xe/xe_res_cursor.h          |  116 +-
 drivers/gpu/drm/xe/xe_svm.c                 |  948 ++++++++
 drivers/gpu/drm/xe/xe_svm.h                 |   83 +
 drivers/gpu/drm/xe/xe_tile.c                |    5 +
 drivers/gpu/drm/xe/xe_vm.c                  |  375 +++-
 drivers/gpu/drm/xe/xe_vm.h                  |   15 +-
 drivers/gpu/drm/xe/xe_vm_types.h            |   57 +
 include/drm/drm_gpusvm.h                    |  445 ++++
 include/drm/drm_gpuvm.h                     |    5 +
 include/drm/drm_pagemap.h                   |  103 +
 include/linux/migrate.h                     |    1 +
 include/uapi/drm/xe_drm.h                   |   19 +-
 mm/memory.c                                 |   13 +-
 mm/migrate_device.c                         |  116 +-
 33 files changed, 5061 insertions(+), 157 deletions(-)
 create mode 100644 drivers/gpu/drm/drm_gpusvm.c
 create mode 100644 drivers/gpu/drm/xe/xe_svm.c
 create mode 100644 drivers/gpu/drm/xe/xe_svm.h
 create mode 100644 include/drm/drm_gpusvm.h
 create mode 100644 include/drm/drm_pagemap.h

-- 
2.34.1

[PATCH v3 01/30] drm/xe: Retry BO allocation

[Matthew Brost]

TTM doesn't support fair eviction via WW locking, this mitigated in by
using retry loops in exec and preempt rebind worker. Extend this retry
loop to BO allocation. Once TTM supports fair eviction this patch can be
reverted.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_bo.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index e6c896ad5602..a670a0444e03 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -2142,6 +2142,7 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
 	struct xe_file *xef = to_xe_file(file);
 	struct drm_xe_gem_create *args = data;
 	struct xe_vm *vm = NULL;
+	ktime_t end = 0;
 	struct xe_bo *bo;
 	unsigned int bo_flags;
 	u32 handle;
@@ -2214,11 +2215,14 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
 		vm = xe_vm_lookup(xef, args->vm_id);
 		if (XE_IOCTL_DBG(xe, !vm))
 			return -ENOENT;
+	}
+
+retry:
+	if (vm) {
 		err = xe_vm_lock(vm, true);
 		if (err)
 			goto out_vm;
 	}
-
 	bo = xe_bo_create_user(xe, NULL, vm, args->size, args->cpu_caching,
 			       bo_flags);
 
@@ -2227,6 +2231,8 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
 
 	if (IS_ERR(bo)) {
 		err = PTR_ERR(bo);
+		if (xe_vm_validate_should_retry(NULL, err, &end))
+			goto retry;
 		goto out_vm;
 	}
 
-- 
2.34.1

[PATCH v3 02/30] mm/migrate: Add migrate_device_pfns

[Matthew Brost]

Add migrate_device_pfns which prepares an array of pre-populated device
pages for migration. This is needed for eviction of known set of
non-contiguous devices pages to cpu pages which is a common case for SVM
in DRM drivers using TTM.

v2:
 - s/migrate_device_vma_range/migrate_device_prepopulated_range
 - Drop extra mmu invalidation (Vetter)
v3:
 - s/migrate_device_prepopulated_range/migrate_device_pfns (Alistar)
 - Use helper to lock device pages (Alistar)
 - Update commit message with why this is required (Alistar)

Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 include/linux/migrate.h |  1 +
 mm/migrate_device.c     | 52 +++++++++++++++++++++++++++++------------
 2 files changed, 38 insertions(+), 15 deletions(-)

diff --git a/include/linux/migrate.h b/include/linux/migrate.h
index 002e49b2ebd9..6254746648cc 100644
--- a/include/linux/migrate.h
+++ b/include/linux/migrate.h
@@ -229,6 +229,7 @@ void migrate_vma_pages(struct migrate_vma *migrate);
 void migrate_vma_finalize(struct migrate_vma *migrate);
 int migrate_device_range(unsigned long *src_pfns, unsigned long start,
 			unsigned long npages);
+int migrate_device_pfns(unsigned long *src_pfns, unsigned long npages);
 void migrate_device_pages(unsigned long *src_pfns, unsigned long *dst_pfns,
 			unsigned long npages);
 void migrate_device_finalize(unsigned long *src_pfns,
diff --git a/mm/migrate_device.c b/mm/migrate_device.c
index 9cf26592ac93..19960743f927 100644
--- a/mm/migrate_device.c
+++ b/mm/migrate_device.c
@@ -876,6 +876,22 @@ void migrate_vma_finalize(struct migrate_vma *migrate)
 }
 EXPORT_SYMBOL(migrate_vma_finalize);
 
+static unsigned long migrate_device_pfn_lock(unsigned long pfn)
+{
+	struct folio *folio;
+
+	folio = folio_get_nontail_page(pfn_to_page(pfn));
+	if (!folio)
+		return 0;
+
+	if (!folio_trylock(folio)) {
+		folio_put(folio);
+		return 0;
+	}
+
+	return migrate_pfn(pfn) | MIGRATE_PFN_MIGRATE;
+}
+
 /**
  * migrate_device_range() - migrate device private pfns to normal memory.
  * @src_pfns: array large enough to hold migrating source device private pfns.
@@ -900,29 +916,35 @@ int migrate_device_range(unsigned long *src_pfns, unsigned long start,
 {
 	unsigned long i, pfn;
 
-	for (pfn = start, i = 0; i < npages; pfn++, i++) {
-		struct folio *folio;
+	for (pfn = start, i = 0; i < npages; pfn++, i++)
+		src_pfns[i] = migrate_device_pfn_lock(pfn);
 
-		folio = folio_get_nontail_page(pfn_to_page(pfn));
-		if (!folio) {
-			src_pfns[i] = 0;
-			continue;
-		}
+	migrate_device_unmap(src_pfns, npages, NULL);
 
-		if (!folio_trylock(folio)) {
-			src_pfns[i] = 0;
-			folio_put(folio);
-			continue;
-		}
+	return 0;
+}
+EXPORT_SYMBOL(migrate_device_range);
 
-		src_pfns[i] = migrate_pfn(pfn) | MIGRATE_PFN_MIGRATE;
-	}
+/**
+ * migrate_device_pfns() - migrate device private pfns to normal memory.
+ * @src_pfns: pre-popluated array of source device private pfns to migrate.
+ * @npages: number of pages to migrate.
+ *
+ * Similar to migrate_device_range() but supports non-contiguous pre-popluated
+ * array of device pages to migrate.
+ */
+int migrate_device_pfns(unsigned long *src_pfns, unsigned long npages)
+{
+	unsigned long i;
+
+	for (i = 0; i < npages; i++)
+		src_pfns[i] = migrate_device_pfn_lock(src_pfns[i]);
 
 	migrate_device_unmap(src_pfns, npages, NULL);
 
 	return 0;
 }
-EXPORT_SYMBOL(migrate_device_range);
+EXPORT_SYMBOL(migrate_device_pfns);
 
 /*
  * Migrate a device coherent folio back to normal memory. The caller should have
-- 
2.34.1

[PATCH v3 03/30] mm/migrate: Trylock device page in do_swap_page

[Matthew Brost]

Avoid multiple CPU page faults to the same device page racing by trying
to lock the page in do_swap_page before taking an extra reference to the
page. This prevents scenarios where multiple CPU page faults each take
an extra reference to a device page, which could abort migration in
folio_migrate_mapping. With the device page being locked in
do_swap_page, the migrate_vma_* functions need to be updated to avoid
locking the fault_page argument.

Prior to this change, a livelock scenario could occur in Xe's (Intel GPU
DRM driver) SVM implementation if enough threads faulted the same device
page.

v3:
 - Put page after unlocking page (Alistair)
 - Warn on spliting a TPH which is fault page (Alistair)
 - Warn on dst page == fault page (Alistair)

Cc: Alistair Popple <apopple@nvidia.com>
Cc: Philip Yang <Philip.Yang@amd.com>
Cc: Felix Kuehling <felix.kuehling@amd.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Suggested-by: Simona Vetter <simona.vetter@ffwll.ch>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 mm/memory.c         | 13 ++++++---
 mm/migrate_device.c | 64 ++++++++++++++++++++++++++++++++-------------
 2 files changed, 55 insertions(+), 22 deletions(-)

diff --git a/mm/memory.c b/mm/memory.c
index 75c2dfd04f72..ae8b11131dad 100644
--- a/mm/memory.c
+++ b/mm/memory.c
@@ -4267,10 +4267,15 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
 			 * Get a page reference while we know the page can't be
 			 * freed.
 			 */
-			get_page(vmf->page);
-			pte_unmap_unlock(vmf->pte, vmf->ptl);
-			ret = vmf->page->pgmap->ops->migrate_to_ram(vmf);
-			put_page(vmf->page);
+			if (trylock_page(vmf->page)) {
+				get_page(vmf->page);
+				pte_unmap_unlock(vmf->pte, vmf->ptl);
+				ret = vmf->page->pgmap->ops->migrate_to_ram(vmf);
+				unlock_page(vmf->page);
+				put_page(vmf->page);
+			} else {
+				pte_unmap_unlock(vmf->pte, vmf->ptl);
+			}
 		} else if (is_hwpoison_entry(entry)) {
 			ret = VM_FAULT_HWPOISON;
 		} else if (is_pte_marker_entry(entry)) {
diff --git a/mm/migrate_device.c b/mm/migrate_device.c
index 19960743f927..3470357d9bae 100644
--- a/mm/migrate_device.c
+++ b/mm/migrate_device.c
@@ -60,6 +60,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
 				   struct mm_walk *walk)
 {
 	struct migrate_vma *migrate = walk->private;
+	struct folio *fault_folio = migrate->fault_page ?
+		page_folio(migrate->fault_page) : NULL;
 	struct vm_area_struct *vma = walk->vma;
 	struct mm_struct *mm = vma->vm_mm;
 	unsigned long addr = start, unmapped = 0;
@@ -88,11 +90,16 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
 
 			folio_get(folio);
 			spin_unlock(ptl);
+			/* FIXME support THP */
+			if (WARN_ON_ONCE(fault_folio == folio))
+				return migrate_vma_collect_skip(start, end,
+								walk);
 			if (unlikely(!folio_trylock(folio)))
 				return migrate_vma_collect_skip(start, end,
 								walk);
 			ret = split_folio(folio);
-			folio_unlock(folio);
+			if (fault_folio != folio)
+				folio_unlock(folio);
 			folio_put(folio);
 			if (ret)
 				return migrate_vma_collect_skip(start, end,
@@ -192,7 +199,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
 		 * optimisation to avoid walking the rmap later with
 		 * try_to_migrate().
 		 */
-		if (folio_trylock(folio)) {
+		if (fault_folio == folio || folio_trylock(folio)) {
 			bool anon_exclusive;
 			pte_t swp_pte;
 
@@ -204,7 +211,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
 
 				if (folio_try_share_anon_rmap_pte(folio, page)) {
 					set_pte_at(mm, addr, ptep, pte);
-					folio_unlock(folio);
+					if (fault_folio != folio)
+						folio_unlock(folio);
 					folio_put(folio);
 					mpfn = 0;
 					goto next;
@@ -363,6 +371,8 @@ static unsigned long migrate_device_unmap(unsigned long *src_pfns,
 					  unsigned long npages,
 					  struct page *fault_page)
 {
+	struct folio *fault_folio = fault_page ?
+		page_folio(fault_page) : NULL;
 	unsigned long i, restore = 0;
 	bool allow_drain = true;
 	unsigned long unmapped = 0;
@@ -427,7 +437,8 @@ static unsigned long migrate_device_unmap(unsigned long *src_pfns,
 		remove_migration_ptes(folio, folio, 0);
 
 		src_pfns[i] = 0;
-		folio_unlock(folio);
+		if (fault_folio != folio)
+			folio_unlock(folio);
 		folio_put(folio);
 		restore--;
 	}
@@ -536,6 +547,8 @@ int migrate_vma_setup(struct migrate_vma *args)
 		return -EINVAL;
 	if (args->fault_page && !is_device_private_page(args->fault_page))
 		return -EINVAL;
+	if (args->fault_page && !PageLocked(args->fault_page))
+		return -EINVAL;
 
 	memset(args->src, 0, sizeof(*args->src) * nr_pages);
 	args->cpages = 0;
@@ -799,19 +812,13 @@ void migrate_vma_pages(struct migrate_vma *migrate)
 }
 EXPORT_SYMBOL(migrate_vma_pages);
 
-/*
- * migrate_device_finalize() - complete page migration
- * @src_pfns: src_pfns returned from migrate_device_range()
- * @dst_pfns: array of pfns allocated by the driver to migrate memory to
- * @npages: number of pages in the range
- *
- * Completes migration of the page by removing special migration entries.
- * Drivers must ensure copying of page data is complete and visible to the CPU
- * before calling this.
- */
-void migrate_device_finalize(unsigned long *src_pfns,
-			unsigned long *dst_pfns, unsigned long npages)
+static void __migrate_device_finalize(unsigned long *src_pfns,
+				      unsigned long *dst_pfns,
+				      unsigned long npages,
+				      struct page *fault_page)
 {
+	struct folio *fault_folio = fault_page ?
+		page_folio(fault_page) : NULL;
 	unsigned long i;
 
 	for (i = 0; i < npages; i++) {
@@ -824,6 +831,7 @@ void migrate_device_finalize(unsigned long *src_pfns,
 
 		if (!page) {
 			if (dst) {
+				WARN_ON_ONCE(fault_folio == dst);
 				folio_unlock(dst);
 				folio_put(dst);
 			}
@@ -834,6 +842,7 @@ void migrate_device_finalize(unsigned long *src_pfns,
 
 		if (!(src_pfns[i] & MIGRATE_PFN_MIGRATE) || !dst) {
 			if (dst) {
+				WARN_ON_ONCE(fault_folio == dst);
 				folio_unlock(dst);
 				folio_put(dst);
 			}
@@ -841,7 +850,8 @@ void migrate_device_finalize(unsigned long *src_pfns,
 		}
 
 		remove_migration_ptes(src, dst, 0);
-		folio_unlock(src);
+		if (fault_folio != src)
+			folio_unlock(src);
 
 		if (folio_is_zone_device(src))
 			folio_put(src);
@@ -849,6 +859,7 @@ void migrate_device_finalize(unsigned long *src_pfns,
 			folio_putback_lru(src);
 
 		if (dst != src) {
+			WARN_ON_ONCE(fault_folio == dst);
 			folio_unlock(dst);
 			if (folio_is_zone_device(dst))
 				folio_put(dst);
@@ -857,6 +868,22 @@ void migrate_device_finalize(unsigned long *src_pfns,
 		}
 	}
 }
+
+/*
+ * migrate_device_finalize() - complete page migration
+ * @src_pfns: src_pfns returned from migrate_device_range()
+ * @dst_pfns: array of pfns allocated by the driver to migrate memory to
+ * @npages: number of pages in the range
+ *
+ * Completes migration of the page by removing special migration entries.
+ * Drivers must ensure copying of page data is complete and visible to the CPU
+ * before calling this.
+ */
+void migrate_device_finalize(unsigned long *src_pfns,
+			unsigned long *dst_pfns, unsigned long npages)
+{
+	return __migrate_device_finalize(src_pfns, dst_pfns, npages, NULL);
+}
 EXPORT_SYMBOL(migrate_device_finalize);
 
 /**
@@ -872,7 +899,8 @@ EXPORT_SYMBOL(migrate_device_finalize);
  */
 void migrate_vma_finalize(struct migrate_vma *migrate)
 {
-	migrate_device_finalize(migrate->src, migrate->dst, migrate->npages);
+	__migrate_device_finalize(migrate->src, migrate->dst, migrate->npages,
+				  migrate->fault_page);
 }
 EXPORT_SYMBOL(migrate_vma_finalize);
 
-- 
2.34.1

[PATCH v3 04/30] drm/pagemap: Add DRM pagemap

[Matthew Brost]

From: Thomas Hellström <thomas.hellstrom@linux.intel.com>

Introduce drm_pagemap ops to map and unmap dma to VRAM resources. In the
local memory case it's a matter of merely providing an offset into the
device's physical address. For future p2p the map and unmap functions may
encode as needed.

Similar to how dma-buf works, let the memory provider (drm_pagemap) provide
the mapping functionality.

v3:
 - Move to drm level include

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 include/drm/drm_pagemap.h | 103 ++++++++++++++++++++++++++++++++++++++
 1 file changed, 103 insertions(+)
 create mode 100644 include/drm/drm_pagemap.h

diff --git a/include/drm/drm_pagemap.h b/include/drm/drm_pagemap.h
new file mode 100644
index 000000000000..b6b387b81479
--- /dev/null
+++ b/include/drm/drm_pagemap.h
@@ -0,0 +1,103 @@
+/* SPDX-License-Identifier: MIT */
+#ifndef _DRM_PAGEMAP_H_
+#define _DRM_PAGEMAP_H_
+
+#include <linux/dma-direction.h>
+#include <linux/hmm.h>
+#include <linux/types.h>
+
+struct drm_pagemap;
+struct device;
+
+/**
+ * enum drm_interconnect_protocol - Used to identify an interconnect protocol.
+ */
+enum drm_interconnect_protocol {
+	DRM_INTERCONNECT_SYSTEM,    /* DMA map is system pages. */
+	DRM_INTERCONNECT_PCIE_P2P,  /* DMA map is PCIE P2P */
+	DRM_INTERCONNECT_DRIVER,    /* DMA map is driver defined */
+	/* A driver can add private values beyond DRM_INTERCONNECT_DRIVER */
+};
+
+/**
+ * struct drm_pagemap_dma_addr - DMA address representation.
+ * @addr: The dma address or driver-defined address for driver private interconnects.
+ * @proto: The interconnect protocol.
+ * @order: The page order of the dma mapping. (Size is PAGE_SIZE << order).
+ * @dir: The DMA direction.
+ *
+ * Note: There is room for improvement here. We should be able to pack into
+ * 64 bits.
+ */
+struct drm_pagemap_dma_addr {
+	dma_addr_t addr;
+	u64 proto : 54;
+	u64 order : 8;
+	u64 dir : 2;
+};
+
+/**
+ * drm_pagemap_dma_addr_encode() - Encode a dma address with metadata
+ * @addr: The dma address or driver-defined address for driver private interconnects.
+ * @proto: The interconnect protocol.
+ * @order: The page order of the dma mapping. (Size is PAGE_SIZE << order).
+ * @dir: The DMA direction.
+ *
+ * Return: A struct drm_pagemap_dma_addr encoding the above information.
+ */
+static inline struct drm_pagemap_dma_addr
+drm_pagemap_dma_addr_encode(dma_addr_t addr,
+			    enum drm_interconnect_protocol proto,
+			    unsigned int order,
+			    enum dma_data_direction dir)
+{
+	return (struct drm_pagemap_dma_addr) {
+		.addr = addr,
+		.proto = proto,
+		.order = order,
+		.dir = dir,
+	};
+}
+
+/**
+ * struct drm_pagemap_ops: Ops for a drm-pagemap.
+ */
+struct drm_pagemap_ops {
+	/**
+	 * @map_dma: Map for dma access or provide a virtual address suitable for
+	 * @dev.
+	 * @dpagemap: The struct drm_pagemap for the page.
+	 * @dev: The dma mapper.
+	 * @page: The page to map.
+	 * @dir: The transfer direction.
+	 * @protocol: The protocol to use.
+	 */
+	struct drm_pagemap_dma_addr (*map_dma)(struct drm_pagemap *dpagemap,
+					       struct device *dev,
+					       struct page *page,
+					       unsigned int order,
+					       enum dma_data_direction dir);
+
+	/**
+	 * @unmap_dma: Unmap a dma address previously obtained using  @map_dma.
+	 * @dev: The dma unmapper.
+	 * @addr: The dma address obtained when mapping.
+	 */
+	void (*unmap_dma)(struct drm_pagemap *dpagemap,
+			  struct device *dev,
+			  struct drm_pagemap_dma_addr addr);
+
+};
+
+/**
+ * struct drm_pagemap: Additional information for a struct dev_pagemap
+ * used for device p2p handshaking.
+ * @ops: The struct drm_pagemap_ops.
+ * @dev: The struct drevice owning the device-private memory.
+ */
+struct drm_pagemap {
+	const struct drm_pagemap_ops *ops;
+	struct device *dev;
+};
+
+#endif
-- 
2.34.1

[PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Matthew Brost]

This patch introduces support for GPU Shared Virtual Memory (SVM) in the
Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
sharing of memory between the CPU and GPU, enhancing performance and
flexibility in GPU computing tasks.

The patch adds the necessary infrastructure for SVM, including data
structures and functions for managing SVM ranges and notifiers. It also
provides mechanisms for allocating, deallocating, and migrating memory
regions between system RAM and GPU VRAM.

This is largely inspired by GPUVM.

v2:
 - Take order into account in check pages
 - Clear range->pages in get pages error
 - Drop setting dirty or accessed bit in get pages (Vetter)
 - Remove mmap assert for cpu faults
 - Drop mmap write lock abuse (Vetter, Christian)
 - Decouple zdd from range (Vetter, Oak)
 - Add drm_gpusvm_range_evict, make it work with coherent pages
 - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
 - mmget/put in drm_gpusvm_evict_to_sram
 - Drop range->vram_alloation variable
 - Don't return in drm_gpusvm_evict_to_sram until all pages detached
 - Don't warn on mixing sram and device pages
 - Update kernel doc
 - Add coherent page support to get pages
 - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
 - Add struct drm_gpusvm_vram and ops (Thomas)
 - Update the range's seqno if the range is valid (Thomas)
 - Remove the is_unmapped check before hmm_range_fault (Thomas)
 - Use drm_pagemap (Thomas)
 - Drop kfree_mapping (Thomas)
 - dma mapp pages under notifier lock (Thomas)
 - Remove ctx.prefault
 - Remove ctx.mmap_locked
 - Add ctx.check_pages
 - s/vram/devmem (Thomas)
v3:
 - Fix memory leak drm_gpusvm_range_get_pages
 - Only migrate pages with same zdd on CPU fault
 - Loop over al VMAs in drm_gpusvm_range_evict
 - Make GPUSVM a drm level module
 - GPL or MIT license
 - Update main kernel doc (Thomas)
 - Prefer foo() vs foo for functions in kernel doc (Thomas)
 - Prefer functions over macros (Thomas)
 - Use unsigned long vs u64 for addresses (Thomas)
 - Use standard interval_tree (Thomas)
 - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
 - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
 - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
 - Newlines between functions defs in header file (Thomas)
 - Drop shall language in driver vfunc kernel doc (Thomas)
 - Move some static inlines from head to C file (Thomas)
 - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
 - Change check_pages to a threshold

Cc: Simona Vetter <simona.vetter@ffwll.ch>
Cc: Dave Airlie <airlied@redhat.com>
Cc: Christian König <christian.koenig@amd.com>
Cc: <dri-devel@lists.freedesktop.org>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/Kconfig      |    8 +
 drivers/gpu/drm/Makefile     |    1 +
 drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
 include/drm/drm_gpusvm.h     |  445 +++++++
 4 files changed, 2674 insertions(+)
 create mode 100644 drivers/gpu/drm/drm_gpusvm.c
 create mode 100644 include/drm/drm_gpusvm.h

diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
index 2f51546b0b88..5524cb1ed9dd 100644
--- a/drivers/gpu/drm/Kconfig
+++ b/drivers/gpu/drm/Kconfig
@@ -278,6 +278,14 @@ config DRM_GPUVM
 	  GPU-VM representation providing helpers to manage a GPUs virtual
 	  address space
 
+config DRM_GPUSVM
+	tristate
+	depends on DRM
+	depends on DEVICE_MIGRATION
+	help
+	  GPU-SVM representation providing helpers to manage a GPUs shared
+	  virtual memory
+
 config DRM_BUDDY
 	tristate
 	depends on DRM
diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
index 19fb370fbc56..ca5d44441fb6 100644
--- a/drivers/gpu/drm/Makefile
+++ b/drivers/gpu/drm/Makefile
@@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
 #
 obj-$(CONFIG_DRM_EXEC) += drm_exec.o
 obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
+obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
 
 obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
 
diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
new file mode 100644
index 000000000000..29ce2c7d1c8c
--- /dev/null
+++ b/drivers/gpu/drm/drm_gpusvm.c
@@ -0,0 +1,2220 @@
+// SPDX-License-Identifier: GPL-2.0-only OR MIT
+/*
+ * Copyright © 2024 Intel Corporation
+ *
+ * Authors:
+ *     Matthew Brost <matthew.brost@intel.com>
+ */
+
+#include <linux/dma-mapping.h>
+#include <linux/hmm.h>
+#include <linux/memremap.h>
+#include <linux/migrate.h>
+#include <linux/mm_types.h>
+#include <linux/pagemap.h>
+#include <linux/slab.h>
+
+#include <drm/drm_device.h>
+#include <drm/drm_gpusvm.h>
+#include <drm/drm_pagemap.h>
+#include <drm/drm_print.h>
+
+/**
+ * DOC: Overview
+ *
+ * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
+ *
+ * The GPU SVM layer is a component of the DRM framework designed to manage shared
+ * virtual memory between the CPU and GPU. It enables efficient data exchange and
+ * processing for GPU-accelerated applications by allowing memory sharing and
+ * synchronization between the CPU's and GPU's virtual address spaces.
+ *
+ * Key GPU SVM Components:
+ * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
+ *		GPU of changes, notifiers are sized based on a GPU SVM
+ *		initialization parameter, with a recommendation of 512M or
+ *		larger. They maintain a Red-BlacK tree and a list of ranges that
+ *		fall within the notifier interval. Notifiers are tracked within
+ *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
+ *		or removed as ranges within the interval are created or
+ *		destroyed.
+ * - Ranges: Represent memory ranges mapped in a DRM device and managed
+ *	     by GPU SVM. They are sized based on an array of chunk sizes, which
+ *	     is a GPU SVM initialization parameter, and the CPU address space.
+ *	     Upon GPU fault, the largest aligned chunk that fits within the
+ *	     faulting CPU address space is chosen for the range size. Ranges are
+ *	     expected to be dynamically allocated on GPU fault and removed on an
+ *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
+ *	     a notifier's Red-Black tree.
+ * - Operations: Define the interface for driver-specific GPU SVM operations
+ *               such as range allocation, notifier allocation, and
+ *               invalidations.
+ * - Device Memory Allocations: Embedded structure containing enough information
+ *                              for GPU SVM to migrate to / from device memory.
+ * - Device Memory Operations: Define the interface for driver-specific device
+ *                             memory operations release memory, populate pfns,
+ *                             and copy to / from device memory.
+ *
+ * This layer provides interfaces for allocating, mapping, migrating, and
+ * releasing memory ranges between the CPU and GPU. It handles all core memory
+ * management interactions (DMA mapping, HMM, and migration) and provides
+ * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
+ * to build the expected driver components for an SVM implementation as detailed
+ * below.
+ *
+ * Expected Driver Components:
+ * - GPU page fault handler: Used to create ranges and notifiers based on the
+ *			     fault address, optionally migrate the range to
+ *			     device memory, and create GPU bindings.
+ * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
+ *			Ranges are expected to be added to the garbage collector
+ *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
+ * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
+ *			ranges.
+ */
+
+/**
+ * DOC: Locking
+ *
+ * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
+ * mmap lock as needed.
+ *
+ * GPU SVM introduces a global notifier lock, which safeguards the notifier's
+ * range RB tree and list, as well as the range's DMA mappings and sequence
+ * number. GPU SVM manages all necessary locking and unlocking operations,
+ * except for the recheck range's pages being valid
+ * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
+ * lock corresponds to the 'driver->update' lock mentioned in the HMM
+ * documentation (TODO: Link). Future revisions may transition from a GPU SVM
+ * global lock to a per-notifier lock if finer-grained locking is deemed
+ * necessary.
+ *
+ * In addition to the locking mentioned above, the driver should implement a
+ * lock to safeguard core GPU SVM function calls that modify state, such as
+ * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
+ * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
+ * locking should also be possible for concurrent GPU fault processing within a
+ * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
+ * to add annotations to GPU SVM.
+ */
+
+/**
+ * DOC: Migration
+ *
+ * The migration support is quite simple, allowing migration between RAM and
+ * device memory at the range granularity. For example, GPU SVM currently does not
+ * support mixing RAM and device memory pages within a range. This means that upon GPU
+ * fault, the entire range can be migrated to device memory, and upon CPU fault, the
+ * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
+ * could be added in the future if required.
+ *
+ * The reasoning for only supporting range granularity is as follows: it
+ * simplifies the implementation, and range sizes are driver-defined and should
+ * be relatively small.
+ */
+
+/**
+ * DOC: Partial Unmapping of Ranges
+ *
+ * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
+ * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
+ * being that a subset of the range still has CPU and GPU mappings. If the
+ * backing store for the range is in device memory, a subset of the backing store has
+ * references. One option would be to split the range and device memory backing store,
+ * but the implementation for this would be quite complicated. Given that
+ * partial unmappings are rare and driver-defined range sizes are relatively
+ * small, GPU SVM does not support splitting of ranges.
+ *
+ * With no support for range splitting, upon partial unmapping of a range, the
+ * driver is expected to invalidate and destroy the entire range. If the range
+ * has device memory as its backing, the driver is also expected to migrate any
+ * remaining pages back to RAM.
+ */
+
+/**
+ * DOC: Examples
+ *
+ * This section provides three examples of how to build the expected driver
+ * components: the GPU page fault handler, the garbage collector, and the
+ * notifier callback.
+ *
+ * The generic code provided does not include logic for complex migration
+ * policies, optimized invalidations, fined grained driver locking, or other
+ * potentially required driver locking (e.g., DMA-resv locks).
+ *
+ * 1) GPU page fault handler
+ *
+ *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
+ *	{
+ *		int err = 0;
+ *
+ *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
+ *
+ *		drm_gpusvm_notifier_lock(gpusvm);
+ *		if (drm_gpusvm_range_pages_valid(range))
+ *			driver_commit_bind(gpusvm, range);
+ *		else
+ *			err = -EAGAIN;
+ *		drm_gpusvm_notifier_unlock(gpusvm);
+ *
+ *		return err;
+ *	}
+ *
+ *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
+ *			     unsigned long gpuva_start, unsigned long gpuva_end)
+ *	{
+ *		struct drm_gpusvm_ctx ctx = {};
+ *		int err;
+ *
+ *		driver_svm_lock();
+ *	retry:
+ *		// Always process UNMAPs first so view of GPU SVM ranges is current
+ *		driver_garbage_collector(gpusvm);
+ *
+ *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
+ *							gpuva_start, gpuva_end,
+ *						        &ctx);
+ *		if (IS_ERR(range)) {
+ *			err = PTR_ERR(range);
+ *			goto unlock;
+ *		}
+ *
+ *		if (driver_migration_policy(range)) {
+ *			devmem = driver_alloc_devmem();
+ *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
+ *							   devmem_allocation,
+ *							   &ctx);
+ *			if (err)	// CPU mappings may have changed
+ *				goto retry;
+ *		}
+ *
+ *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
+ *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
+ *			if (err == -EOPNOTSUPP)
+ *				drm_gpusvm_range_evict(gpusvm, range);
+ *			goto retry;
+ *		} else if (err) {
+ *			goto unlock;
+ *		}
+ *
+ *		err = driver_bind_range(gpusvm, range);
+ *		if (err == -EAGAIN)	// CPU mappings changed
+ *			goto retry
+ *
+ *	unlock:
+ *		driver_svm_unlock();
+ *		return err;
+ *	}
+ *
+ * 2) Garbage Collector.
+ *
+ *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
+ *					struct drm_gpusvm_range *range)
+ *	{
+ *		assert_driver_svm_locked(gpusvm);
+ *
+ *		// Partial unmap, migrate any remaining device memory pages back to RAM
+ *		if (range->flags.partial_unmap)
+ *			drm_gpusvm_range_evict(gpusvm, range);
+ *
+ *		driver_unbind_range(range);
+ *		drm_gpusvm_range_remove(gpusvm, range);
+ *	}
+ *
+ *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
+ *	{
+ *		assert_driver_svm_locked(gpusvm);
+ *
+ *		for_each_range_in_garbage_collector(gpusvm, range)
+ *			__driver_garbage_collector(gpusvm, range);
+ *	}
+ *
+ * 3) Notifier callback.
+ *
+ *	void driver_invalidation(struct drm_gpusvm *gpusvm,
+ *				 struct drm_gpusvm_notifier *notifier,
+ *				 const struct mmu_notifier_range *mmu_range)
+ *	{
+ *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
+ *		struct drm_gpusvm_range *range = NULL;
+ *
+ *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
+ *
+ *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
+ *					  mmu_range->end) {
+ *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
+ *
+ *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
+ *				continue;
+ *
+ *			drm_gpusvm_range_set_unmapped(range, mmu_range);
+ *			driver_garbage_collector_add(gpusvm, range);
+ *		}
+ *	}
+ */
+
+/**
+ * npages_in_range() - Calculate the number of pages in a given range
+ * @start: The start address of the range
+ * @end: The end address of the range
+ *
+ * This macro calculates the number of pages in a given memory range,
+ * specified by the start and end addresses. It divides the difference
+ * between the end and start addresses by the page size (PAGE_SIZE) to
+ * determine the number of pages in the range.
+ *
+ * Return: The number of pages in the specified range.
+ */
+static unsigned long
+npages_in_range(unsigned long start, unsigned long end)
+{
+	return (end - start) >> PAGE_SHIFT;
+}
+
+/**
+ * struct drm_gpusvm_zdd - GPU SVM zone device data
+ *
+ * @refcount: Reference count for the zdd
+ * @destroy_work: Work structure for asynchronous zdd destruction
+ * @devmem_allocation: device memory allocation
+ * @device_private_page_owner: Device private pages owner
+ *
+ * This structure serves as a generic wrapper installed in
+ * page->zone_device_data. It provides infrastructure for looking up a device
+ * memory allocation upon CPU page fault and asynchronously releasing device
+ * memory once the CPU has no page references. Asynchronous release is useful
+ * because CPU page references can be dropped in IRQ contexts, while releasing
+ * device memory likely requires sleeping locks.
+ */
+struct drm_gpusvm_zdd {
+	struct kref refcount;
+	struct work_struct destroy_work;
+	struct drm_gpusvm_devmem *devmem_allocation;
+	void *device_private_page_owner;
+};
+
+/**
+ * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
+ * @w: Pointer to the work_struct
+ *
+ * This function releases device memory, puts GPU SVM range, and frees zdd.
+ */
+static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
+{
+	struct drm_gpusvm_zdd *zdd =
+		container_of(w, struct drm_gpusvm_zdd, destroy_work);
+	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
+		zdd->devmem_allocation->ops : NULL;
+
+	if (zdd->devmem_allocation && ops->devmem_release)
+		ops->devmem_release(zdd->devmem_allocation);
+	kfree(zdd);
+}
+
+/**
+ * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
+ * @device_private_page_owner: Device private pages owner
+ *
+ * This function allocates and initializes a new zdd structure. It sets up the
+ * reference count and initializes the destroy work.
+ *
+ * Returns:
+ * Pointer to the allocated zdd on success, ERR_PTR() on failure.
+ */
+static struct drm_gpusvm_zdd *
+drm_gpusvm_zdd_alloc(void *device_private_page_owner)
+{
+	struct drm_gpusvm_zdd *zdd;
+
+	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
+	if (!zdd)
+		return NULL;
+
+	kref_init(&zdd->refcount);
+	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
+	zdd->devmem_allocation = NULL;
+	zdd->device_private_page_owner = device_private_page_owner;
+
+	return zdd;
+}
+
+/**
+ * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
+ * @zdd: Pointer to the zdd structure.
+ *
+ * This function increments the reference count of the provided zdd structure.
+ *
+ * Returns: Pointer to the zdd structure.
+ */
+static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
+{
+	kref_get(&zdd->refcount);
+	return zdd;
+}
+
+/**
+ * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
+ * @ref: Pointer to the reference count structure.
+ *
+ * This function queues the destroy_work of the zdd for asynchronous destruction.
+ */
+static void drm_gpusvm_zdd_destroy(struct kref *ref)
+{
+	struct drm_gpusvm_zdd *zdd =
+		container_of(ref, struct drm_gpusvm_zdd, refcount);
+
+	if (zdd->devmem_allocation)
+		WRITE_ONCE(zdd->devmem_allocation->detached, true);
+	schedule_work(&zdd->destroy_work);
+}
+
+/**
+ * drm_gpusvm_zdd_put() - Put a zdd reference.
+ * @zdd: Pointer to the zdd structure.
+ *
+ * This function decrements the reference count of the provided zdd structure
+ * and schedules its destruction if the count drops to zero.
+ */
+static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
+{
+	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
+}
+
+/**
+ * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
+ * @notifier: Pointer to the GPU SVM notifier structure.
+ * @start: Start address of the range
+ * @end: End address of the range
+ *
+ * Return: A pointer to the drm_gpusvm_range if found or NULL
+ */
+struct drm_gpusvm_range *
+drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
+		      unsigned long end)
+{
+	struct interval_tree_node *itree;
+
+	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
+
+	if (itree)
+		return container_of(itree, struct drm_gpusvm_range, itree);
+	else
+		return NULL;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
+
+/**
+ * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
+ * @range__: Iterator variable for the ranges
+ * @next__: Iterator variable for the ranges temporay storage
+ * @notifier__: Pointer to the GPU SVM notifier
+ * @start__: Start address of the range
+ * @end__: End address of the range
+ *
+ * This macro is used to iterate over GPU SVM ranges in a notifier while
+ * removing ranges from it.
+ */
+#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
+	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
+	     (next__) = __drm_gpusvm_range_next(range__);				\
+	     (range__) && (range__->itree.start < (end__));				\
+	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
+
+/**
+ * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
+ * @notifier: a pointer to the current drm_gpusvm_notifier
+ *
+ * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
+ *         the current notifier is the last one or if the input notifier is
+ *         NULL.
+ */
+static struct drm_gpusvm_notifier *
+__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
+{
+	if (notifier && !list_is_last(&notifier->entry,
+				      &notifier->gpusvm->notifier_list))
+		return list_next_entry(notifier, entry);
+
+	return NULL;
+}
+
+static struct drm_gpusvm_notifier *
+notifier_iter_first(struct rb_root_cached *root, unsigned long start,
+		    unsigned long last)
+{
+	struct interval_tree_node *itree;
+
+	itree = interval_tree_iter_first(root, start, last);
+
+	if (itree)
+		return container_of(itree, struct drm_gpusvm_notifier, itree);
+	else
+		return NULL;
+}
+
+/**
+ * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
+ * @notifier__: Iterator variable for the notifiers
+ * @notifier__: Pointer to the GPU SVM notifier
+ * @start__: Start address of the notifier
+ * @end__: End address of the notifier
+ *
+ * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
+ */
+#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
+
+/**
+ * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
+ * @notifier__: Iterator variable for the notifiers
+ * @next__: Iterator variable for the notifiers temporay storage
+ * @notifier__: Pointer to the GPU SVM notifier
+ * @start__: Start address of the notifier
+ * @end__: End address of the notifier
+ *
+ * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
+ * removing notifiers from it.
+ */
+#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
+	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
+
+/**
+ * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
+ * @mni: Pointer to the mmu_interval_notifier structure.
+ * @mmu_range: Pointer to the mmu_notifier_range structure.
+ * @cur_seq: Current sequence number.
+ *
+ * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
+ * notifier sequence number and calls the driver invalidate vfunc under
+ * gpusvm->notifier_lock.
+ *
+ * Returns:
+ * true if the operation succeeds, false otherwise.
+ */
+static bool
+drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
+			       const struct mmu_notifier_range *mmu_range,
+			       unsigned long cur_seq)
+{
+	struct drm_gpusvm_notifier *notifier =
+		container_of(mni, typeof(*notifier), notifier);
+	struct drm_gpusvm *gpusvm = notifier->gpusvm;
+
+	if (!mmu_notifier_range_blockable(mmu_range))
+		return false;
+
+	down_write(&gpusvm->notifier_lock);
+	mmu_interval_set_seq(mni, cur_seq);
+	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
+	up_write(&gpusvm->notifier_lock);
+
+	return true;
+}
+
+/**
+ * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
+ */
+static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
+	.invalidate = drm_gpusvm_notifier_invalidate,
+};
+
+/**
+ * drm_gpusvm_init() - Initialize the GPU SVM.
+ * @gpusvm: Pointer to the GPU SVM structure.
+ * @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.
+ * @ops: Pointer to the operations structure for GPU SVM.
+ * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
+ *               Entries should be powers of 2 in descending order with last
+ *               entry being SZ_4K.
+ * @num_chunks: Number of chunks.
+ *
+ * This function initializes the GPU SVM.
+ *
+ * Returns:
+ * 0 on success, a negative error code on failure.
+ */
+int drm_gpusvm_init(struct drm_gpusvm *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,
+		    const struct drm_gpusvm_ops *ops,
+		    const unsigned long *chunk_sizes, int num_chunks)
+{
+	if (!ops->invalidate || !num_chunks)
+		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;
+	gpusvm->ops = ops;
+	gpusvm->chunk_sizes = chunk_sizes;
+	gpusvm->num_chunks = num_chunks;
+
+	mmgrab(mm);
+	gpusvm->root = RB_ROOT_CACHED;
+	INIT_LIST_HEAD(&gpusvm->notifier_list);
+
+	init_rwsem(&gpusvm->notifier_lock);
+
+	fs_reclaim_acquire(GFP_KERNEL);
+	might_lock(&gpusvm->notifier_lock);
+	fs_reclaim_release(GFP_KERNEL);
+
+#ifdef CONFIG_LOCKDEP
+	gpusvm->lock_dep_map = NULL;
+#endif
+
+	return 0;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_init);
+
+/**
+ * drm_gpusvm_notifier_find() - Find GPU SVM notifier
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @fault_addr: Fault address
+ *
+ * This function finds the GPU SVM notifier associated with the fault address.
+ *
+ * Returns:
+ * Pointer to the GPU SVM notifier on success, NULL otherwise.
+ */
+static struct drm_gpusvm_notifier *
+drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
+			 unsigned long fault_addr)
+{
+	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
+}
+
+/**
+ * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
+ * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
+ *
+ * Return: A pointer to the containing drm_gpusvm_notifier structure.
+ */
+static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
+{
+	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
+}
+
+/**
+ * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: Pointer to the GPU SVM notifier structure
+ *
+ * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
+ */
+static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
+				       struct drm_gpusvm_notifier *notifier)
+{
+	struct rb_node *node;
+	struct list_head *head;
+
+	interval_tree_insert(&notifier->itree, &gpusvm->root);
+
+	node = rb_prev(&notifier->itree.rb);
+	if (node)
+		head = &(to_drm_gpusvm_notifier(node))->entry;
+	else
+		head = &gpusvm->notifier_list;
+
+	list_add(&notifier->entry, head);
+}
+
+/**
+ * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
+ * @gpusvm: Pointer to the GPU SVM tructure
+ * @notifier: Pointer to the GPU SVM notifier structure
+ *
+ * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
+ */
+static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
+				       struct drm_gpusvm_notifier *notifier)
+{
+	interval_tree_remove(&notifier->itree, &gpusvm->root);
+	list_del(&notifier->entry);
+}
+
+/**
+ * drm_gpusvm_fini() - Finalize the GPU SVM.
+ * @gpusvm: Pointer to the GPU SVM structure.
+ *
+ * This function finalizes the GPU SVM by cleaning up any remaining ranges and
+ * notifiers, and dropping a reference to struct MM.
+ */
+void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
+{
+	struct drm_gpusvm_notifier *notifier, *next;
+
+	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
+		struct drm_gpusvm_range *range, *__next;
+
+		/*
+		 * Remove notifier first to avoid racing with any invalidation
+		 */
+		mmu_interval_notifier_remove(&notifier->notifier);
+		notifier->flags.removed = true;
+
+		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
+					       LONG_MAX)
+			drm_gpusvm_range_remove(gpusvm, range);
+	}
+
+	mmdrop(gpusvm->mm);
+	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
+
+/**
+ * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @fault_addr: Fault address
+ *
+ * This function allocates and initializes the GPU SVM notifier structure.
+ *
+ * Returns:
+ * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
+ */
+static struct drm_gpusvm_notifier *
+drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
+{
+	struct drm_gpusvm_notifier *notifier;
+
+	if (gpusvm->ops->notifier_alloc)
+		notifier = gpusvm->ops->notifier_alloc();
+	else
+		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
+
+	if (!notifier)
+		return ERR_PTR(-ENOMEM);
+
+	notifier->gpusvm = gpusvm;
+	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
+	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
+	INIT_LIST_HEAD(&notifier->entry);
+	notifier->root = RB_ROOT_CACHED;
+	INIT_LIST_HEAD(&notifier->range_list);
+
+	return notifier;
+}
+
+/**
+ * drm_gpusvm_notifier_free() - Free GPU SVM notifier
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: Pointer to the GPU SVM notifier structure
+ *
+ * This function frees the GPU SVM notifier structure.
+ */
+static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
+				     struct drm_gpusvm_notifier *notifier)
+{
+	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
+
+	if (gpusvm->ops->notifier_free)
+		gpusvm->ops->notifier_free(notifier);
+	else
+		kfree(notifier);
+}
+
+/**
+ * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
+ * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
+ *
+ * Return: A pointer to the containing drm_gpusvm_range structure.
+ */
+static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
+{
+	return container_of(node, struct drm_gpusvm_range, itree.rb);
+}
+
+/**
+ * drm_gpusvm_range_insert() - Insert GPU SVM range
+ * @notifier: Pointer to the GPU SVM notifier structure
+ * @range: Pointer to the GPU SVM range structure
+ *
+ * This function inserts the GPU SVM range into the notifier RB tree and list.
+ */
+static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
+				    struct drm_gpusvm_range *range)
+{
+	struct rb_node *node;
+	struct list_head *head;
+
+	drm_gpusvm_notifier_lock(notifier->gpusvm);
+	interval_tree_insert(&range->itree, &notifier->root);
+
+	node = rb_prev(&range->itree.rb);
+	if (node)
+		head = &(to_drm_gpusvm_range(node))->entry;
+	else
+		head = &notifier->range_list;
+
+	list_add(&range->entry, head);
+	drm_gpusvm_notifier_unlock(notifier->gpusvm);
+}
+
+/**
+ * __drm_gpusvm_range_remove() - Remove GPU SVM range
+ * @notifier: Pointer to the GPU SVM notifier structure
+ * @range: Pointer to the GPU SVM range structure
+ *
+ * This macro removes the GPU SVM range from the notifier RB tree and list.
+ */
+static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
+				      struct drm_gpusvm_range *range)
+{
+	interval_tree_remove(&range->itree, &notifier->root);
+	list_del(&range->entry);
+}
+
+/**
+ * drm_gpusvm_range_alloc() - Allocate GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: Pointer to the GPU SVM notifier structure
+ * @fault_addr: Fault address
+ * @chunk_size: Chunk size
+ * @migrate_devmem: Flag indicating whether to migrate device memory
+ *
+ * This function allocates and initializes the GPU SVM range structure.
+ *
+ * Returns:
+ * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
+ */
+static struct drm_gpusvm_range *
+drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
+		       struct drm_gpusvm_notifier *notifier,
+		       unsigned long fault_addr, unsigned long chunk_size,
+		       bool migrate_devmem)
+{
+	struct drm_gpusvm_range *range;
+
+	if (gpusvm->ops->range_alloc)
+		range = gpusvm->ops->range_alloc(gpusvm);
+	else
+		range = kzalloc(sizeof(*range), GFP_KERNEL);
+
+	if (!range)
+		return ERR_PTR(-ENOMEM);
+
+	kref_init(&range->refcount);
+	range->gpusvm = gpusvm;
+	range->notifier = notifier;
+	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
+	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
+	INIT_LIST_HEAD(&range->entry);
+	range->notifier_seq = LONG_MAX;
+	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
+
+	return range;
+}
+
+/**
+ * drm_gpusvm_check_pages() - Check pages
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: Pointer to the GPU SVM notifier structure
+ * @start: Start address
+ * @end: End address
+ *
+ * Check if pages between start and end have been faulted in on the CPU. Use to
+ * prevent migration of pages without CPU backing store.
+ *
+ * Returns:
+ * True if pages have been faulted into CPU, False otherwise
+ */
+static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
+				   struct drm_gpusvm_notifier *notifier,
+				   unsigned long start, unsigned long end)
+{
+	struct hmm_range hmm_range = {
+		.default_flags = 0,
+		.notifier = &notifier->notifier,
+		.start = start,
+		.end = end,
+		.dev_private_owner = gpusvm->device_private_page_owner,
+	};
+	unsigned long timeout =
+		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
+	unsigned long *pfns;
+	unsigned long npages = npages_in_range(start, end);
+	int err, i;
+
+	mmap_assert_locked(gpusvm->mm);
+
+	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
+	if (!pfns)
+		return false;
+
+	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
+	hmm_range.hmm_pfns = pfns;
+
+	while (true) {
+		err = hmm_range_fault(&hmm_range);
+		if (err == -EBUSY) {
+			if (time_after(jiffies, timeout))
+				break;
+
+			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
+			continue;
+		}
+		break;
+	}
+	if (err)
+		goto err_free;
+
+	for (i = 0; i < npages;) {
+		if (!(pfns[i] & HMM_PFN_VALID)) {
+			err = -EFAULT;
+			goto err_free;
+		}
+		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
+	}
+
+err_free:
+	kvfree(pfns);
+	return err ? false : true;
+}
+
+/**
+ * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: Pointer to the GPU SVM notifier structure
+ * @vas: Pointer to the virtual memory area structure
+ * @fault_addr: Fault address
+ * @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
+ *
+ * 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
+ * memory area boundaries.
+ *
+ * Returns:
+ * Chunk size on success, LONG_MAX on failure.
+ */
+static unsigned long
+drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
+			    struct drm_gpusvm_notifier *notifier,
+			    struct vm_area_struct *vas,
+			    unsigned long fault_addr,
+			    unsigned long gpuva_start,
+			    unsigned long gpuva_end,
+			    unsigned long check_pages_threshold)
+{
+	unsigned long start, end;
+	int i = 0;
+
+retry:
+	for (; i < gpusvm->num_chunks; ++i) {
+		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
+		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
+
+		if (start >= vas->vm_start && end <= vas->vm_end &&
+		    start >= notifier->itree.start &&
+		    end <= notifier->itree.last + 1 &&
+		    start >= gpuva_start && end <= gpuva_end)
+			break;
+	}
+
+	if (i == gpusvm->num_chunks)
+		return LONG_MAX;
+
+	/*
+	 * If allocation more than page, ensure not to overlap with existing
+	 * ranges.
+	 */
+	if (end - start != SZ_4K) {
+		struct drm_gpusvm_range *range;
+
+		range = drm_gpusvm_range_find(notifier, start, end);
+		if (range) {
+			++i;
+			goto retry;
+		}
+
+		/*
+		 * XXX: Only create range on pages CPU has faulted in. Without
+		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
+		 * process-many-malloc' fails. In the failure case, each process
+		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
+		 * ranges. When migrating the SVM ranges, some processes fail in
+		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
+		 * and then upon drm_gpusvm_range_get_pages device pages from
+		 * other processes are collected + faulted in which creates all
+		 * sorts of problems. Unsure exactly how this happening, also
+		 * problem goes away if 'xe_exec_system_allocator --r
+		 * process-many-malloc' mallocs at least 64k at a time.
+		 */
+		if (end - start <= check_pages_threshold &&
+		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
+			++i;
+			goto retry;
+		}
+	}
+
+	return end - start;
+}
+
+/**
+ * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @fault_addr: Fault address
+ * @gpuva_start: Start address of GPUVA which mirrors CPU
+ * @gpuva_end: End address of GPUVA which mirrors CPU
+ * @ctx: GPU SVM context
+ *
+ * This function finds or inserts a newly allocated a GPU SVM range based on the
+ * fault address. Caller must hold a lock to protect range lookup and insertion.
+ *
+ * Returns:
+ * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
+ */
+struct drm_gpusvm_range *
+drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
+				unsigned long fault_addr,
+				unsigned long gpuva_start,
+				unsigned long gpuva_end,
+				const struct drm_gpusvm_ctx *ctx)
+{
+	struct drm_gpusvm_notifier *notifier;
+	struct drm_gpusvm_range *range;
+	struct mm_struct *mm = gpusvm->mm;
+	struct vm_area_struct *vas;
+	bool notifier_alloc = false;
+	unsigned long chunk_size;
+	int err;
+	bool migrate_devmem;
+
+	drm_gpusvm_driver_lock_held(gpusvm);
+
+	if (fault_addr < gpusvm->mm_start ||
+	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
+		return ERR_PTR(-EINVAL);
+
+	if (!mmget_not_zero(mm))
+		return ERR_PTR(-EFAULT);
+
+	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
+	if (!notifier) {
+		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
+		if (IS_ERR(notifier)) {
+			err = PTR_ERR(notifier);
+			goto err_mmunlock;
+		}
+		notifier_alloc = true;
+		err = mmu_interval_notifier_insert(&notifier->notifier,
+						   mm, notifier->itree.start,
+						   notifier->itree.last -
+						   notifier->itree.start + 1,
+						   &drm_gpusvm_notifier_ops);
+		if (err)
+			goto err_notifier;
+	}
+
+	mmap_read_lock(mm);
+
+	vas = vma_lookup(mm, fault_addr);
+	if (!vas) {
+		err = -ENOENT;
+		goto err_notifier_remove;
+	}
+
+	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
+		err = -EPERM;
+		goto err_notifier_remove;
+	}
+
+	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
+	if (range)
+		goto out_mmunlock;
+	/*
+	 * XXX: Short-circuiting migration based on migrate_vma_* current
+	 * limitations. If/when migrate_vma_* add more support, this logic will
+	 * have to change.
+	 */
+	migrate_devmem = ctx->devmem_possible &&
+		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
+
+	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
+						 fault_addr, gpuva_start,
+						 gpuva_end, migrate_devmem &&
+						 ctx->check_pages_threshold);
+	if (chunk_size == LONG_MAX) {
+		err = -EINVAL;
+		goto err_notifier_remove;
+	}
+
+	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
+				       migrate_devmem);
+	if (IS_ERR(range)) {
+		err = PTR_ERR(range);
+		goto err_notifier_remove;
+	}
+
+	drm_gpusvm_range_insert(notifier, range);
+	if (notifier_alloc)
+		drm_gpusvm_notifier_insert(gpusvm, notifier);
+
+out_mmunlock:
+	mmap_read_unlock(mm);
+	mmput(mm);
+
+	return range;
+
+err_notifier_remove:
+	mmap_read_unlock(mm);
+	if (notifier_alloc)
+		mmu_interval_notifier_remove(&notifier->notifier);
+err_notifier:
+	if (notifier_alloc)
+		drm_gpusvm_notifier_free(gpusvm, notifier);
+err_mmunlock:
+	mmput(mm);
+	return ERR_PTR(err);
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
+
+/**
+ * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ * @npages: Number of pages to unmap
+ *
+ * This function unmap pages associated with a GPU SVM range. Assumes and
+ * asserts correct locking is in place when called.
+ */
+static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
+					   struct drm_gpusvm_range *range,
+					   unsigned long npages)
+{
+	unsigned long i, j;
+	struct drm_pagemap *dpagemap = range->dpagemap;
+	struct device *dev = gpusvm->drm->dev;
+
+	lockdep_assert_held(&gpusvm->notifier_lock);
+
+	if (range->flags.has_dma_mapping) {
+		for (i = 0, j = 0; i < npages; j++) {
+			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
+
+			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
+				dma_unmap_page(dev,
+					       addr->addr,
+					       PAGE_SIZE << addr->order,
+					       addr->dir);
+			else if (dpagemap && dpagemap->ops->unmap_dma)
+				dpagemap->ops->unmap_dma(dpagemap,
+							 dev,
+							 *addr);
+			i += 1 << addr->order;
+		}
+		range->flags.has_devmem_pages = false;
+		range->flags.has_dma_mapping = false;
+		range->dpagemap = NULL;
+	}
+}
+
+/**
+ * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ *
+ * This function frees the dma address array associated with a GPU SVM range.
+ */
+static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
+					struct drm_gpusvm_range *range)
+{
+	lockdep_assert_held(&gpusvm->notifier_lock);
+
+	if (range->dma_addr) {
+		kvfree(range->dma_addr);
+		range->dma_addr = NULL;
+	}
+}
+
+/**
+ * drm_gpusvm_range_remove() - Remove GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range to be removed
+ *
+ * This function removes the specified GPU SVM range and also removes the parent
+ * GPU SVM notifier if no more ranges remain in the notifier. The caller must
+ * hold a lock to protect range and notifier removal.
+ */
+void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
+			     struct drm_gpusvm_range *range)
+{
+	unsigned long npages = npages_in_range(range->itree.start,
+					       range->itree.last + 1);
+	struct drm_gpusvm_notifier *notifier;
+
+	drm_gpusvm_driver_lock_held(gpusvm);
+
+	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
+	if (WARN_ON_ONCE(!notifier))
+		return;
+
+	drm_gpusvm_notifier_lock(gpusvm);
+	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
+	drm_gpusvm_range_free_pages(gpusvm, range);
+	__drm_gpusvm_range_remove(notifier, range);
+	drm_gpusvm_notifier_unlock(gpusvm);
+
+	drm_gpusvm_range_put(range);
+
+	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
+		if (!notifier->flags.removed)
+			mmu_interval_notifier_remove(&notifier->notifier);
+		drm_gpusvm_notifier_remove(gpusvm, notifier);
+		drm_gpusvm_notifier_free(gpusvm, notifier);
+	}
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
+
+/**
+ * drm_gpusvm_range_get() - Get a reference to GPU SVM range
+ * @range: Pointer to the GPU SVM range
+ *
+ * This function increments the reference count of the specified GPU SVM range.
+ *
+ * Returns:
+ * Pointer to the GPU SVM range.
+ */
+struct drm_gpusvm_range *
+drm_gpusvm_range_get(struct drm_gpusvm_range *range)
+{
+	kref_get(&range->refcount);
+
+	return range;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
+
+/**
+ * drm_gpusvm_range_destroy() - Destroy GPU SVM range
+ * @refcount: Pointer to the reference counter embedded in the GPU SVM range
+ *
+ * This function destroys the specified GPU SVM range when its reference count
+ * reaches zero. If a custom range-free function is provided, it is invoked to
+ * free the range; otherwise, the range is deallocated using kfree().
+ */
+static void drm_gpusvm_range_destroy(struct kref *refcount)
+{
+	struct drm_gpusvm_range *range =
+		container_of(refcount, struct drm_gpusvm_range, refcount);
+	struct drm_gpusvm *gpusvm = range->gpusvm;
+
+	if (gpusvm->ops->range_free)
+		gpusvm->ops->range_free(range);
+	else
+		kfree(range);
+}
+
+/**
+ * drm_gpusvm_range_put() - Put a reference to GPU SVM range
+ * @range: Pointer to the GPU SVM range
+ *
+ * This function decrements the reference count of the specified GPU SVM range
+ * and frees it when the count reaches zero.
+ */
+void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
+{
+	kref_put(&range->refcount, drm_gpusvm_range_destroy);
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
+
+/**
+ * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ *
+ * This function determines if a GPU SVM range pages are valid. Expected be
+ * called holding gpusvm->notifier_lock and as the last step before committing a
+ * GPU binding. This is akin to a notifier seqno check in the HMM documentation
+ * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
+ * function is required for finer grained checking (i.e., per range) if pages
+ * are valid.
+ *
+ * Returns:
+ * True if GPU SVM range has valid pages, False otherwise
+ */
+bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
+				  struct drm_gpusvm_range *range)
+{
+	lockdep_assert_held(&gpusvm->notifier_lock);
+
+	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
+
+/**
+ * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ *
+ * This function determines if a GPU SVM range pages are valid. Expected be
+ * called without holding gpusvm->notifier_lock.
+ *
+ * Returns:
+ * True if GPU SVM range has valid pages, False otherwise
+ */
+static bool
+drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
+				      struct drm_gpusvm_range *range)
+{
+	bool pages_valid;
+
+	if (!range->dma_addr)
+		return false;
+
+	drm_gpusvm_notifier_lock(gpusvm);
+	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
+	if (!pages_valid)
+		drm_gpusvm_range_free_pages(gpusvm, range);
+	drm_gpusvm_notifier_unlock(gpusvm);
+
+	return pages_valid;
+}
+
+/**
+ * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ * @ctx: GPU SVM context
+ *
+ * This function gets pages for a GPU SVM range and ensures they are mapped for
+ * DMA access.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
+			       struct drm_gpusvm_range *range,
+			       const struct drm_gpusvm_ctx *ctx)
+{
+	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
+	struct hmm_range hmm_range = {
+		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
+			HMM_PFN_REQ_WRITE),
+		.notifier = notifier,
+		.start = range->itree.start,
+		.end = range->itree.last + 1,
+		.dev_private_owner = gpusvm->device_private_page_owner,
+	};
+	struct mm_struct *mm = gpusvm->mm;
+	struct drm_gpusvm_zdd *zdd;
+	unsigned long timeout =
+		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
+	unsigned long i, j;
+	unsigned long npages = npages_in_range(range->itree.start,
+					       range->itree.last + 1);
+	unsigned long num_dma_mapped;
+	unsigned int order = 0;
+	unsigned long *pfns;
+	struct page **pages;
+	int err = 0;
+	struct dev_pagemap *pagemap;
+	struct drm_pagemap *dpagemap;
+
+retry:
+	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
+	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
+		goto set_seqno;
+
+	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
+	if (!pfns)
+		return -ENOMEM;
+
+	if (!mmget_not_zero(mm)) {
+		err = -EFAULT;
+		goto err_free;
+	}
+
+	hmm_range.hmm_pfns = pfns;
+	while (true) {
+		mmap_read_lock(mm);
+		err = hmm_range_fault(&hmm_range);
+		mmap_read_unlock(mm);
+
+		if (err == -EBUSY) {
+			if (time_after(jiffies, timeout))
+				break;
+
+			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
+			continue;
+		}
+		break;
+	}
+	mmput(mm);
+	if (err)
+		goto err_free;
+
+	pages = (struct page **)pfns;
+map_pages:
+	/*
+	 * Perform all dma mappings under the notifier lock to not
+	 * access freed pages. A notifier will either block on
+	 * the notifier lock or unmap dma.
+	 */
+	drm_gpusvm_notifier_lock(gpusvm);
+	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
+		drm_gpusvm_notifier_unlock(gpusvm);
+		kvfree(pfns);
+		goto retry;
+	}
+
+	if (!range->dma_addr) {
+		/* Unlock and restart mapping to allocate memory. */
+		drm_gpusvm_notifier_unlock(gpusvm);
+		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
+						 GFP_KERNEL);
+		if (!range->dma_addr) {
+			err = -ENOMEM;
+			goto err_free;
+		}
+		goto map_pages;
+	}
+
+	zdd = NULL;
+	num_dma_mapped = 0;
+	for (i = 0, j = 0; i < npages; ++j) {
+		struct page *page = hmm_pfn_to_page(pfns[i]);
+
+		order = hmm_pfn_to_map_order(pfns[i]);
+		if (is_device_private_page(page) || is_device_coherent_page(page)) {
+			if (zdd != page->zone_device_data && i > 0) {
+				err = -EOPNOTSUPP;
+				goto err_unmap;
+			}
+			zdd = page->zone_device_data;
+			if (pagemap != page->pgmap) {
+				if (i > 0) {
+					err = -EOPNOTSUPP;
+					goto err_unmap;
+				}
+
+				pagemap = page->pgmap;
+				dpagemap = zdd->devmem_allocation->dpagemap;
+				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
+					/*
+					 * Raced. This is not supposed to happen
+					 * since hmm_range_fault() should've migrated
+					 * this page to system.
+					 */
+					err = -EAGAIN;
+					goto err_unmap;
+				}
+			}
+			range->dma_addr[j] =
+				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
+						       page, order,
+						       DMA_BIDIRECTIONAL);
+			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
+				err = -EFAULT;
+				goto err_unmap;
+			}
+
+			pages[i] = page;
+		} else {
+			dma_addr_t addr;
+
+			if (is_zone_device_page(page) || zdd) {
+				err = -EOPNOTSUPP;
+				goto err_unmap;
+			}
+
+			addr = dma_map_page(gpusvm->drm->dev,
+					    page, 0,
+					    PAGE_SIZE << order,
+					    DMA_BIDIRECTIONAL);
+			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
+				err = -EFAULT;
+				goto err_unmap;
+			}
+
+			range->dma_addr[j] = drm_pagemap_dma_addr_encode
+				(addr, DRM_INTERCONNECT_SYSTEM, order,
+				 DMA_BIDIRECTIONAL);
+		}
+		i += 1 << order;
+		num_dma_mapped = i;
+	}
+
+	range->flags.has_dma_mapping = true;
+	if (zdd) {
+		range->flags.has_devmem_pages = true;
+		range->dpagemap = dpagemap;
+	}
+
+	drm_gpusvm_notifier_unlock(gpusvm);
+	kvfree(pfns);
+set_seqno:
+	range->notifier_seq = hmm_range.notifier_seq;
+
+	return 0;
+
+err_unmap:
+	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
+	drm_gpusvm_notifier_unlock(gpusvm);
+err_free:
+	kvfree(pfns);
+	if (err == -EAGAIN)
+		goto retry;
+	return err;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
+
+/**
+ * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ * @ctx: GPU SVM context
+ *
+ * This function unmaps pages associated with a GPU SVM range. If @in_notifier
+ * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
+ * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
+ * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
+ * security model.
+ */
+void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
+				  struct drm_gpusvm_range *range,
+				  const struct drm_gpusvm_ctx *ctx)
+{
+	unsigned long npages = npages_in_range(range->itree.start,
+					       range->itree.last + 1);
+
+	if (ctx->in_notifier)
+		lockdep_assert_held_write(&gpusvm->notifier_lock);
+	else
+		drm_gpusvm_notifier_lock(gpusvm);
+
+	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
+
+	if (!ctx->in_notifier)
+		drm_gpusvm_notifier_unlock(gpusvm);
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
+
+/**
+ * drm_gpusvm_migration_unlock_put_page() - Put a migration page
+ * @page: Pointer to the page to put
+ *
+ * This function unlocks and puts a page.
+ */
+static void drm_gpusvm_migration_unlock_put_page(struct page *page)
+{
+	unlock_page(page);
+	put_page(page);
+}
+
+/**
+ * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
+ * @npages: Number of pages
+ * @migrate_pfn: Array of migrate page frame numbers
+ *
+ * This function unlocks and puts an array of pages.
+ */
+static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
+						  unsigned long *migrate_pfn)
+{
+	unsigned long i;
+
+	for (i = 0; i < npages; ++i) {
+		struct page *page;
+
+		if (!migrate_pfn[i])
+			continue;
+
+		page = migrate_pfn_to_page(migrate_pfn[i]);
+		drm_gpusvm_migration_unlock_put_page(page);
+		migrate_pfn[i] = 0;
+	}
+}
+
+/**
+ * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
+ * @page: Pointer to the page
+ * @zdd: Pointer to the GPU SVM zone device data
+ *
+ * This function associates the given page with the specified GPU SVM zone
+ * device data and initializes it for zone device usage.
+ */
+static void drm_gpusvm_get_devmem_page(struct page *page,
+				     struct drm_gpusvm_zdd *zdd)
+{
+	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
+	zone_device_page_init(page);
+}
+
+/**
+ * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
+ * @dev: The device for which the pages are being mapped
+ * @dma_addr: Array to store DMA addresses corresponding to mapped pages
+ * @migrate_pfn: Array of migrate page frame numbers to map
+ * @npages: Number of pages to map
+ * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
+ *
+ * This function maps pages of memory for migration usage in GPU SVM. It
+ * iterates over each page frame number provided in @migrate_pfn, maps the
+ * corresponding page, and stores the DMA address in the provided @dma_addr
+ * array.
+ *
+ * Return: 0 on success, -EFAULT if an error occurs during mapping.
+ */
+static int drm_gpusvm_migrate_map_pages(struct device *dev,
+					dma_addr_t *dma_addr,
+					unsigned long *migrate_pfn,
+					unsigned long npages,
+					enum dma_data_direction dir)
+{
+	unsigned long i;
+
+	for (i = 0; i < npages; ++i) {
+		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
+
+		if (!page)
+			continue;
+
+		if (WARN_ON_ONCE(is_zone_device_page(page)))
+			return -EFAULT;
+
+		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
+		if (dma_mapping_error(dev, dma_addr[i]))
+			return -EFAULT;
+	}
+
+	return 0;
+}
+
+/**
+ * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
+ * @dev: The device for which the pages were mapped
+ * @dma_addr: Array of DMA addresses corresponding to mapped pages
+ * @npages: Number of pages to unmap
+ * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
+ *
+ * This function unmaps previously mapped pages of memory for GPU Shared Virtual
+ * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
+ * if it's valid and not already unmapped, and unmaps the corresponding page.
+ */
+static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
+					   dma_addr_t *dma_addr,
+					   unsigned long npages,
+					   enum dma_data_direction dir)
+{
+	unsigned long i;
+
+	for (i = 0; i < npages; ++i) {
+		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
+			continue;
+
+		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
+	}
+}
+
+/**
+ * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range structure
+ * @devmem_allocation: Pointer to the device memory allocation. The caller
+ *                     should hold a reference to the device memory allocation,
+ *                     which should be dropped via ops->devmem_release or upon
+ *                     the failure of this function.
+ * @ctx: GPU SVM context
+ *
+ * This function migrates the specified GPU SVM range to device memory. It performs the
+ * necessary setup and invokes the driver-specific operations for migration to
+ * device memory. Upon successful return, @devmem_allocation can safely reference @range
+ * until ops->devmem_release is called which only upon successful return.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
+				 struct drm_gpusvm_range *range,
+				 struct drm_gpusvm_devmem *devmem_allocation,
+				 const struct drm_gpusvm_ctx *ctx)
+{
+	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
+	unsigned long start = range->itree.start, end = range->itree.last + 1;
+	struct migrate_vma migrate = {
+		.start		= start,
+		.end		= end,
+		.pgmap_owner	= gpusvm->device_private_page_owner,
+		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
+	};
+	struct mm_struct *mm = gpusvm->mm;
+	unsigned long i, npages = npages_in_range(start, end);
+	struct vm_area_struct *vas;
+	struct drm_gpusvm_zdd *zdd = NULL;
+	struct page **pages;
+	dma_addr_t *dma_addr;
+	void *buf;
+	int err;
+
+	if (!range->flags.migrate_devmem)
+		return -EINVAL;
+
+	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
+	    !ops->copy_to_ram)
+		return -EOPNOTSUPP;
+
+	if (!mmget_not_zero(mm)) {
+		err = -EFAULT;
+		goto err_out;
+	}
+	mmap_read_lock(mm);
+
+	vas = vma_lookup(mm, start);
+	if (!vas) {
+		err = -ENOENT;
+		goto err_mmunlock;
+	}
+
+	if (end > vas->vm_end || start < vas->vm_start) {
+		err = -EINVAL;
+		goto err_mmunlock;
+	}
+
+	if (!vma_is_anonymous(vas)) {
+		err = -EBUSY;
+		goto err_mmunlock;
+	}
+
+	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
+		       sizeof(*pages), GFP_KERNEL);
+	if (!buf) {
+		err = -ENOMEM;
+		goto err_mmunlock;
+	}
+	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
+	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
+
+	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
+	if (!zdd) {
+		err = -ENOMEM;
+		goto err_free;
+	}
+
+	migrate.vma = vas;
+	migrate.src = buf;
+	migrate.dst = migrate.src + npages;
+
+	err = migrate_vma_setup(&migrate);
+	if (err)
+		goto err_free;
+
+	if (!migrate.cpages) {
+		err = -EFAULT;
+		goto err_free;
+	}
+
+	if (migrate.cpages != npages) {
+		err = -EBUSY;
+		goto err_finalize;
+	}
+
+	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
+	if (err)
+		goto err_finalize;
+
+	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
+					   migrate.src, npages, DMA_TO_DEVICE);
+	if (err)
+		goto err_finalize;
+
+	for (i = 0; i < npages; ++i) {
+		struct page *page = pfn_to_page(migrate.dst[i]);
+
+		pages[i] = page;
+		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
+		drm_gpusvm_get_devmem_page(page, zdd);
+	}
+
+	err = ops->copy_to_devmem(pages, dma_addr, npages);
+	if (err)
+		goto err_finalize;
+
+	/* Upon success bind devmem allocation to range and zdd */
+	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
+
+err_finalize:
+	if (err)
+		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
+	migrate_vma_pages(&migrate);
+	migrate_vma_finalize(&migrate);
+	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
+				       DMA_TO_DEVICE);
+err_free:
+	if (zdd)
+		drm_gpusvm_zdd_put(zdd);
+	kvfree(buf);
+err_mmunlock:
+	mmap_read_unlock(mm);
+	mmput(mm);
+err_out:
+	return err;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
+
+/**
+ * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
+ * @vas: Pointer to the VM area structure, can be NULL
+ * @fault_page: Fault page
+ * @npages: Number of pages to populate
+ * @mpages: Number of pages to migrate
+ * @src_mpfn: Source array of migrate PFNs
+ * @mpfn: Array of migrate PFNs to populate
+ * @addr: Start address for PFN allocation
+ *
+ * This function populates the RAM migrate page frame numbers (PFNs) for the
+ * specified VM area structure. It allocates and locks pages in the VM area for
+ * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
+ * alloc_page for allocation.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
+					       struct page *fault_page,
+					       unsigned long npages,
+					       unsigned long *mpages,
+					       unsigned long *src_mpfn,
+					       unsigned long *mpfn,
+					       unsigned long addr)
+{
+	unsigned long i;
+
+	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
+		struct page *page, *src_page;
+
+		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
+			continue;
+
+		if (fault_page) {
+			src_page = migrate_pfn_to_page(src_mpfn[i]);
+			if (src_page->zone_device_data !=
+			    fault_page->zone_device_data)
+				continue;
+		}
+
+		if (vas)
+			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
+		else
+			page = alloc_page(GFP_HIGHUSER);
+
+		if (!page)
+			goto free_pages;
+
+		mpfn[i] = migrate_pfn(page_to_pfn(page));
+	}
+
+	for (i = 0; i < npages; ++i) {
+		struct page *page = migrate_pfn_to_page(mpfn[i]);
+
+		if (!page)
+			continue;
+
+		WARN_ON_ONCE(!trylock_page(page));
+		++*mpages;
+	}
+
+	return 0;
+
+free_pages:
+	for (i = 0; i < npages; ++i) {
+		struct page *page = migrate_pfn_to_page(mpfn[i]);
+
+		if (!page)
+			continue;
+
+		put_page(page);
+		mpfn[i] = 0;
+	}
+	return -ENOMEM;
+}
+
+/**
+ * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
+ * @devmem_allocation: Pointer to the device memory allocation
+ *
+ * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
+ * migration done via migrate_device_* functions.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
+{
+	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
+	unsigned long npages, mpages = 0;
+	struct page **pages;
+	unsigned long *src, *dst;
+	dma_addr_t *dma_addr;
+	void *buf;
+	int i, err = 0;
+
+	npages = devmem_allocation->size >> PAGE_SHIFT;
+
+retry:
+	if (!mmget_not_zero(devmem_allocation->mm))
+		return -EFAULT;
+
+	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
+		       sizeof(*pages), GFP_KERNEL);
+	if (!buf) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	src = buf;
+	dst = buf + (sizeof(*src) * npages);
+	dma_addr = buf + (2 * sizeof(*src) * npages);
+	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
+
+	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
+	if (err)
+		goto err_free;
+
+	err = migrate_device_pfns(src, npages);
+	if (err)
+		goto err_free;
+
+	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
+						  src, dst, 0);
+	if (err || !mpages)
+		goto err_finalize;
+
+	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
+					   dst, npages, DMA_FROM_DEVICE);
+	if (err)
+		goto err_finalize;
+
+	for (i = 0; i < npages; ++i)
+		pages[i] = migrate_pfn_to_page(src[i]);
+
+	err = ops->copy_to_ram(pages, dma_addr, npages);
+	if (err)
+		goto err_finalize;
+
+err_finalize:
+	if (err)
+		drm_gpusvm_migration_unlock_put_pages(npages, dst);
+	migrate_device_pages(src, dst, npages);
+	migrate_device_finalize(src, dst, npages);
+	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
+				       DMA_FROM_DEVICE);
+err_free:
+	kvfree(buf);
+err_out:
+	mmput_async(devmem_allocation->mm);
+	if (!err && !READ_ONCE(devmem_allocation->detached)) {
+		cond_resched();
+		goto retry;
+	}
+
+	return err;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
+
+/**
+ * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
+ * @vas: Pointer to the VM area structure
+ * @device_private_page_owner: Device private pages owner
+ * @page: Pointer to the page for fault handling (can be NULL)
+ * @fault_addr: Fault address
+ * @size: Size of migration
+ *
+ * This internal function performs the migration of the specified GPU SVM range
+ * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
+ * invokes the driver-specific operations for migration to RAM.
+ *
+ * Returns:
+ * 0 on success, negative error code on failure.
+ */
+static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
+				       void *device_private_page_owner,
+				       struct page *page,
+				       unsigned long fault_addr,
+				       unsigned long size)
+{
+	struct migrate_vma migrate = {
+		.vma		= vas,
+		.pgmap_owner	= device_private_page_owner,
+		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
+			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
+		.fault_page	= page,
+	};
+	struct drm_gpusvm_zdd *zdd;
+	const struct drm_gpusvm_devmem_ops *ops;
+	struct device *dev;
+	unsigned long npages, mpages = 0;
+	struct page **pages;
+	dma_addr_t *dma_addr;
+	unsigned long start, end;
+	void *buf;
+	int i, err = 0;
+
+	start = ALIGN_DOWN(fault_addr, size);
+	end = ALIGN(fault_addr + 1, size);
+
+	/* Corner where VMA area struct has been partially unmapped */
+	if (start < vas->vm_start)
+		start = vas->vm_start;
+	if (end > vas->vm_end)
+		end = vas->vm_end;
+
+	migrate.start = start;
+	migrate.end = end;
+	npages = npages_in_range(start, end);
+
+	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
+		       sizeof(*pages), GFP_KERNEL);
+	if (!buf) {
+		err = -ENOMEM;
+		goto err_out;
+	}
+	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
+	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
+
+	migrate.vma = vas;
+	migrate.src = buf;
+	migrate.dst = migrate.src + npages;
+
+	err = migrate_vma_setup(&migrate);
+	if (err)
+		goto err_free;
+
+	/* Raced with another CPU fault, nothing to do */
+	if (!migrate.cpages)
+		goto err_free;
+
+	if (!page) {
+		for (i = 0; i < npages; ++i) {
+			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
+				continue;
+
+			page = migrate_pfn_to_page(migrate.src[i]);
+			break;
+		}
+
+		if (!page)
+			goto err_finalize;
+	}
+	zdd = page->zone_device_data;
+	ops = zdd->devmem_allocation->ops;
+	dev = zdd->devmem_allocation->dev;
+
+	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
+						  migrate.src, migrate.dst,
+						  start);
+	if (err)
+		goto err_finalize;
+
+	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
+					   DMA_FROM_DEVICE);
+	if (err)
+		goto err_finalize;
+
+	for (i = 0; i < npages; ++i)
+		pages[i] = migrate_pfn_to_page(migrate.src[i]);
+
+	err = ops->copy_to_ram(pages, dma_addr, npages);
+	if (err)
+		goto err_finalize;
+
+err_finalize:
+	if (err)
+		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
+	migrate_vma_pages(&migrate);
+	migrate_vma_finalize(&migrate);
+	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
+				       DMA_FROM_DEVICE);
+err_free:
+	kvfree(buf);
+err_out:
+
+	return err;
+}
+
+/**
+ * drm_gpusvm_range_evict() - Evict GPU SVM range
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @range: Pointer to the GPU SVM range to be removed
+ *
+ * This function evicts the specified GPU SVM range.
+ */
+void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
+			    struct drm_gpusvm_range *range)
+{
+	struct mm_struct *mm = gpusvm->mm;
+	struct vm_area_struct *vas;
+	unsigned long start = range->itree.start, end = range->itree.last + 1;
+
+	if (!mmget_not_zero(mm))
+		return;
+
+	mmap_read_lock(mm);
+
+	/*
+	 * Loop required to find all VMAs for the corner case when partial
+	 * unmappings have occurred.
+	 */
+again:
+	vas = find_vma(mm, start);
+	if (!vas)
+		goto unlock;
+
+	if (end <= vas->vm_start || start >= vas->vm_end)
+		goto unlock;
+
+	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
+				    NULL, start,
+				    range->itree.last - range->itree.start + 1);
+	if (vas->vm_end < end) {
+		start = vas->vm_end;
+		goto again;
+	}
+unlock:
+	mmap_read_unlock(mm);
+	mmput(mm);
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
+
+/**
+ * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
+ * @page: Pointer to the page
+ *
+ * This function is a callback used to put the GPU SVM zone device data
+ * associated with a page when it is being released.
+ */
+static void drm_gpusvm_page_free(struct page *page)
+{
+	drm_gpusvm_zdd_put(page->zone_device_data);
+}
+
+/**
+ * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
+ * @vmf: Pointer to the fault information structure
+ *
+ * This function is a page fault handler used to migrate a GPU SVM range to RAM.
+ * It retrieves the GPU SVM range information from the faulting page and invokes
+ * the internal migration function to migrate the range back to RAM.
+ *
+ * Returns:
+ * VM_FAULT_SIGBUS on failure, 0 on success.
+ */
+static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
+{
+	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
+	int err;
+
+	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
+					  zdd->device_private_page_owner,
+					  vmf->page, vmf->address,
+					  zdd->devmem_allocation->size);
+
+	return err ? VM_FAULT_SIGBUS : 0;
+}
+
+/**
+ * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
+ */
+static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
+	.page_free = drm_gpusvm_page_free,
+	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
+};
+
+/**
+ * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
+ *
+ * Returns:
+ * Pointer to the GPU SVM device page map operations structure.
+ */
+const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
+{
+	return &drm_gpusvm_pagemap_ops;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
+
+/**
+ * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
+ * @gpusvm: Pointer to the GPU SVM structure.
+ * @start: Start address
+ * @end: End address
+ *
+ * Returns:
+ * True if GPU SVM has mapping, False otherwise
+ */
+bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
+			    unsigned long end)
+{
+	struct drm_gpusvm_notifier *notifier;
+
+	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
+		struct drm_gpusvm_range *range = NULL;
+
+		drm_gpusvm_for_each_range(range, notifier, start, end)
+			return true;
+	}
+
+	return false;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
+
+/**
+ * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
+ * @range: Pointer to the GPU SVM range structure.
+ * @mmu_range: Pointer to the MMU notifier range structure.
+ *
+ * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
+ * if the range partially falls within the provided MMU notifier range.
+ */
+void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
+				   const struct mmu_notifier_range *mmu_range)
+{
+	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
+
+	range->flags.unmapped = true;
+	if (range->itree.start < mmu_range->start ||
+	    range->itree.last + 1 > mmu_range->end)
+		range->flags.partial_unmap = true;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
+
+/**
+ * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
+ *
+ * @dev: Pointer to the device structure which device memory allocation belongs to
+ * @mm: Pointer to the mm_struct for the address space
+ * @ops: Pointer to the operations structure for GPU SVM device memory
+ * @dpagemap: The struct drm_pagemap we're allocating from.
+ * @size: Size of device memory allocation
+ */
+void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
+			    struct device *dev, struct mm_struct *mm,
+			    const struct drm_gpusvm_devmem_ops *ops,
+			    struct drm_pagemap *dpagemap, size_t size)
+{
+	devmem_allocation->dev = dev;
+	devmem_allocation->mm = mm;
+	devmem_allocation->ops = ops;
+	devmem_allocation->dpagemap = dpagemap;
+	devmem_allocation->size = size;
+}
+EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
+
+MODULE_DESCRIPTION("DRM GPUSVM");
+MODULE_LICENSE("GPL");
diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
new file mode 100644
index 000000000000..2fb6980d0cf1
--- /dev/null
+++ b/include/drm/drm_gpusvm.h
@@ -0,0 +1,445 @@
+/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
+/*
+ * Copyright © 2024 Intel Corporation
+ */
+
+#ifndef __DRM_GPUSVM_H__
+#define __DRM_GPUSVM_H__
+
+#include <linux/kref.h>
+#include <linux/interval_tree.h>
+#include <linux/mmu_notifier.h>
+
+struct dev_pagemap_ops;
+struct drm_device;
+struct drm_gpusvm;
+struct drm_gpusvm_notifier;
+struct drm_gpusvm_ops;
+struct drm_gpusvm_range;
+struct drm_gpusvm_devmem;
+struct drm_pagemap;
+struct drm_pagemap_dma_addr;
+
+/**
+ * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
+ *
+ * This structure defines the operations for GPU Shared Virtual Memory (SVM)
+ * device memory. These operations are provided by the GPU driver to manage device memory
+ * allocations and perform operations such as migration between device memory and system
+ * RAM.
+ */
+struct drm_gpusvm_devmem_ops {
+	/**
+	 * @devmem_release: Release device memory allocation (optional)
+	 * @devmem_allocation: device memory allocation
+	 *
+	 * Release device memory allocation and drop a reference to device
+	 * memory allocation.
+	 */
+	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
+
+	/**
+	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
+	 * @devmem_allocation: device memory allocation
+	 * @npages: Number of pages to populate
+	 * @pfn: Array of page frame numbers to populate
+	 *
+	 * Populate device memory page frame numbers (PFN).
+	 *
+	 * Returns:
+	 * 0 on success, a negative error code on failure.
+	 */
+	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
+				   unsigned long npages, unsigned long *pfn);
+
+	/**
+	 * @copy_to_devmem: Copy to device memory (required for migration)
+	 * @pages: Pointer to array of device memory pages (destination)
+	 * @dma_addr: Pointer to array of DMA addresses (source)
+	 * @npages: Number of pages to copy
+	 *
+	 * Copy pages to device memory.
+	 *
+	 * Returns:
+	 * 0 on success, a negative error code on failure.
+	 */
+	int (*copy_to_devmem)(struct page **pages,
+			      dma_addr_t *dma_addr,
+			      unsigned long npages);
+
+	/**
+	 * @copy_to_ram: Copy to system RAM (required for migration)
+	 * @pages: Pointer to array of device memory pages (source)
+	 * @dma_addr: Pointer to array of DMA addresses (destination)
+	 * @npages: Number of pages to copy
+	 *
+	 * Copy pages to system RAM.
+	 *
+	 * Returns:
+	 * 0 on success, a negative error code on failure.
+	 */
+	int (*copy_to_ram)(struct page **pages,
+			   dma_addr_t *dma_addr,
+			   unsigned long npages);
+};
+
+/**
+ * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
+ *
+ * @dev: Pointer to the device structure which device memory allocation belongs to
+ * @mm: Pointer to the mm_struct for the address space
+ * @ops: Pointer to the operations structure for GPU SVM device memory
+ * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
+ * @size: Size of device memory allocation
+ * @detached: device memory allocations is detached from device pages
+ */
+struct drm_gpusvm_devmem {
+	struct device *dev;
+	struct mm_struct *mm;
+	const struct drm_gpusvm_devmem_ops *ops;
+	struct drm_pagemap *dpagemap;
+	size_t size;
+	bool detached;
+};
+
+/**
+ * struct drm_gpusvm_ops - Operations structure for GPU SVM
+ *
+ * This structure defines the operations for GPU Shared Virtual Memory (SVM).
+ * These operations are provided by the GPU driver to manage SVM ranges and
+ * notifiers.
+ */
+struct drm_gpusvm_ops {
+	/**
+	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
+	 *
+	 * Allocate a GPU SVM notifier.
+	 *
+	 * Returns:
+	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
+	 */
+	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
+
+	/**
+	 * @notifier_free: Free a GPU SVM notifier (optional)
+	 * @notifier: Pointer to the GPU SVM notifier to be freed
+	 *
+	 * Free a GPU SVM notifier.
+	 */
+	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
+
+	/**
+	 * @range_alloc: Allocate a GPU SVM range (optional)
+	 * @gpusvm: Pointer to the GPU SVM
+	 *
+	 * Allocate a GPU SVM range.
+	 *
+	 * Returns:
+	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
+	 */
+	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
+
+	/**
+	 * @range_free: Free a GPU SVM range (optional)
+	 * @range: Pointer to the GPU SVM range to be freed
+	 *
+	 * Free a GPU SVM range.
+	 */
+	void (*range_free)(struct drm_gpusvm_range *range);
+
+	/**
+	 * @invalidate: Invalidate GPU SVM notifier (required)
+	 * @gpusvm: Pointer to the GPU SVM
+	 * @notifier: Pointer to the GPU SVM notifier
+	 * @mmu_range: Pointer to the mmu_notifier_range structure
+	 *
+	 * Invalidate the GPU page tables. It can safely walk the notifier range
+	 * RB tree/list in this function. Called while holding the notifier lock.
+	 */
+	void (*invalidate)(struct drm_gpusvm *gpusvm,
+			   struct drm_gpusvm_notifier *notifier,
+			   const struct mmu_notifier_range *mmu_range);
+};
+
+/**
+ * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
+ *
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: MMU interval notifier
+ * @itree: Interval tree node for the notifier (inserted in GPU SVM)
+ * @entry: List entry to fast interval tree traversal
+ * @root: Cached root node of the RB tree containing ranges
+ * @range_list: List head containing of ranges in the same order they appear in
+ *              interval tree. This is useful to keep iterating ranges while
+ *              doing modifications to RB tree.
+ * @flags.removed: Flag indicating whether the MMU interval notifier has been
+ *                 removed
+ *
+ * This structure represents a GPU SVM notifier.
+ */
+struct drm_gpusvm_notifier {
+	struct drm_gpusvm *gpusvm;
+	struct mmu_interval_notifier notifier;
+	struct interval_tree_node itree;
+	struct list_head entry;
+	struct rb_root_cached root;
+	struct list_head range_list;
+	struct {
+		u32 removed : 1;
+	} flags;
+};
+
+/**
+ * struct drm_gpusvm_range - Structure representing a GPU SVM range
+ *
+ * @gpusvm: Pointer to the GPU SVM structure
+ * @notifier: Pointer to the GPU SVM notifier
+ * @refcount: Reference count for the range
+ * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
+ * @entry: List entry to fast interval tree traversal
+ * @notifier_seq: Notifier sequence number of the range's pages
+ * @dma_addr: DMA address array
+ * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
+ *            Note this is assuming only one drm_pagemap per range is allowed.
+ * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
+ * @flags.unmapped: Flag indicating if the range has been unmapped
+ * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
+ * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
+ * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
+ *
+ * This structure represents a GPU SVM range used for tracking memory ranges
+ * mapped in a DRM device.
+ */
+struct drm_gpusvm_range {
+	struct drm_gpusvm *gpusvm;
+	struct drm_gpusvm_notifier *notifier;
+	struct kref refcount;
+	struct interval_tree_node itree;
+	struct list_head entry;
+	unsigned long notifier_seq;
+	struct drm_pagemap_dma_addr *dma_addr;
+	struct drm_pagemap *dpagemap;
+	struct {
+		/* All flags below must be set upon creation */
+		u16 migrate_devmem : 1;
+		/* All flags below must be set / cleared under notifier lock */
+		u16 unmapped : 1;
+		u16 partial_unmap : 1;
+		u16 has_devmem_pages : 1;
+		u16 has_dma_mapping : 1;
+	} flags;
+};
+
+/**
+ * struct drm_gpusvm - GPU SVM structure
+ *
+ * @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
+ * @ops: Pointer to the operations structure for GPU SVM
+ * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
+ *               Entries should be powers of 2 in descending order.
+ * @num_chunks: Number of chunks
+ * @notifier_lock: Read-write semaphore for protecting notifier operations
+ * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
+ * @notifier_list: list head containing of notifiers in the same order they
+ *                 appear in interval tree. This is useful to keep iterating
+ *                 notifiers while doing modifications to RB tree.
+ *
+ * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
+ * memory ranges mapped in a DRM (Direct Rendering Manager) device.
+ *
+ * No reference counting is provided, as this is expected to be embedded in the
+ * driver VM structure along with the struct drm_gpuvm, which handles reference
+ * counting.
+ */
+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;
+	const struct drm_gpusvm_ops *ops;
+	const unsigned long *chunk_sizes;
+	int num_chunks;
+	struct rw_semaphore notifier_lock;
+	struct rb_root_cached root;
+	struct list_head notifier_list;
+#ifdef CONFIG_LOCKDEP
+	/**
+	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
+	 * drm_gpusvm_range_remove with a driver provided lock.
+	 */
+	struct lockdep_map *lock_dep_map;
+#endif
+};
+
+/**
+ * struct drm_gpusvm_ctx - DRM GPU SVM context
+ *
+ * @check_pages_threshold: Check CPU pages for present if chunk is less than or
+ *                         equal to threshold. If not present, reduce chunk
+ *                         size.
+ * @in_notifier: entering from a MMU notifier
+ * @read_only: operating on read-only memory
+ * @devmem_possible: possible to use device memory
+ *
+ * Context that is DRM GPUSVM is operating in (i.e. user arguments).
+ */
+struct drm_gpusvm_ctx {
+	unsigned long check_pages_threshold;
+	unsigned int in_notifier :1;
+	unsigned int read_only :1;
+	unsigned int devmem_possible :1;
+};
+
+int drm_gpusvm_init(struct drm_gpusvm *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,
+		    const struct drm_gpusvm_ops *ops,
+		    const unsigned long *chunk_sizes, int num_chunks);
+
+void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
+
+void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
+
+struct drm_gpusvm_range *
+drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
+				unsigned long fault_addr,
+				unsigned long gpuva_start,
+				unsigned long gpuva_end,
+				const struct drm_gpusvm_ctx *ctx);
+
+void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
+			     struct drm_gpusvm_range *range);
+
+void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
+			    struct drm_gpusvm_range *range);
+
+struct drm_gpusvm_range *
+drm_gpusvm_range_get(struct drm_gpusvm_range *range);
+
+void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
+
+bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
+				  struct drm_gpusvm_range *range);
+
+int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
+			       struct drm_gpusvm_range *range,
+			       const struct drm_gpusvm_ctx *ctx);
+
+void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
+				  struct drm_gpusvm_range *range,
+				  const struct drm_gpusvm_ctx *ctx);
+
+int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
+				 struct drm_gpusvm_range *range,
+				 struct drm_gpusvm_devmem *devmem_allocation,
+				 const struct drm_gpusvm_ctx *ctx);
+
+int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
+
+const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
+
+bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
+			    unsigned long end);
+
+struct drm_gpusvm_range *
+drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
+		      unsigned long end);
+
+void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
+				   const struct mmu_notifier_range *mmu_range);
+
+void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
+			    struct device *dev, struct mm_struct *mm,
+			    const struct drm_gpusvm_devmem_ops *ops,
+			    struct drm_pagemap *dpagemap, size_t size);
+
+#ifdef CONFIG_LOCKDEP
+/**
+ * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
+ * @gpusvm: Pointer to the GPU SVM structure.
+ * @lock: the lock used to protect the gpuva list. The locking primitive
+ * must contain a dep_map field.
+ *
+ * Call this to annotate drm_gpusvm_range_find_or_insert and
+ * drm_gpusvm_range_remove.
+ */
+#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
+	do { \
+		if (!WARN((gpusvm)->lock_dep_map, \
+			  "GPUSVM range lock should be set only once."))\
+			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
+	} while (0)
+#define drm_gpusvm_driver_lock_held(gpusvm) \
+	do { \
+		if ((gpusvm)->lock_dep_map)	\
+			lock_is_held((gpusvm)->lock_dep_map);	\
+	} while (0)
+#else
+#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
+#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
+#endif
+
+/**
+ * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
+ * @gpusvm__: Pointer to the GPU SVM structure.
+ *
+ * Abstract client usage GPU SVM notifier lock, take lock
+ */
+#define drm_gpusvm_notifier_lock(gpusvm__)	\
+	down_read(&(gpusvm__)->notifier_lock)
+
+/**
+ * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
+ * @gpusvm__: Pointer to the GPU SVM structure.
+ *
+ * Abstract client usage GPU SVM notifier lock, drop lock
+ */
+#define drm_gpusvm_notifier_unlock(gpusvm__)	\
+	up_read(&(gpusvm__)->notifier_lock)
+
+/**
+ * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
+ * @range: a pointer to the current GPU SVM range
+ *
+ * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
+ *         current range is the last one or if the input range is NULL.
+ */
+static inline struct drm_gpusvm_range *
+__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
+{
+	if (range && !list_is_last(&range->entry,
+				   &range->notifier->range_list))
+		return list_next_entry(range, entry);
+
+	return NULL;
+}
+
+/**
+ * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
+ * @range__: Iterator variable for the ranges. If set, it indicates the start of
+ *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
+ * @notifier__: Pointer to the GPU SVM notifier
+ * @start__: Start address of the range
+ * @end__: End address of the range
+ *
+ * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
+ * to use while holding the driver SVM lock or the notifier lock.
+ */
+#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
+	for ((range__) = (range__) ?:					\
+	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
+	     (range__) && (range__->itree.start < (end__));		\
+	     (range__) = __drm_gpusvm_range_next(range__))
+
+#endif /* __DRM_GPUSVM_H__ */
-- 
2.34.1

[PATCH v3 06/30] drm/xe: Select DRM_GPUSVM Kconfig

[Matthew Brost]

Xe depends on DRM_GPUSVM for SVM implementation, select it in Kconfig.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/Kconfig | 1 +
 1 file changed, 1 insertion(+)

diff --git a/drivers/gpu/drm/xe/Kconfig b/drivers/gpu/drm/xe/Kconfig
index b51a2bde73e2..3a08e16bfada 100644
--- a/drivers/gpu/drm/xe/Kconfig
+++ b/drivers/gpu/drm/xe/Kconfig
@@ -39,6 +39,7 @@ config DRM_XE
 	select DRM_TTM_HELPER
 	select DRM_EXEC
 	select DRM_GPUVM
+	select DRM_GPUSVM
 	select DRM_SCHED
 	select MMU_NOTIFIER
 	select WANT_DEV_COREDUMP
-- 
2.34.1

[PATCH v3 07/30] drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag

[Matthew Brost]

Add the DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag, which is used to
create unpopulated virtual memory areas (VMAs) without memory backing or
GPU page tables. These VMAs are referred to as CPU address mirror VMAs.
The idea is that upon a page fault or prefetch, the memory backing and
GPU page tables will be populated.

CPU address mirror VMAs only update GPUVM state; they do not have an
internal page table (PT) state, nor do they have GPU mappings.

It is expected that CPU address mirror VMAs will be mixed with buffer
object (BO) VMAs within a single VM. In other words, system allocations
and runtime allocations can be mixed within a single user-mode driver
(UMD) program.

Expected usage:

- Bind the entire virtual address (VA) space upon program load using the
  DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag.
- If a buffer object (BO) requires GPU mapping (runtime allocation),
  allocate a CPU address using mmap(PROT_NONE), bind the BO to the
  mmapped address using existing bind IOCTLs. If a CPU map of the BO is
  needed, mmap it again to the same CPU address using mmap(MAP_FIXED)
- If a BO no longer requires GPU mapping, munmap it from the CPU address
  space and them bind the mapping address with the
  DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag.
- Any malloc'd or mmapped CPU address accessed by the GPU will be
  faulted in via the SVM implementation (system allocation).
- Upon freeing any mmapped or malloc'd data, the SVM implementation will
  remove GPU mappings.

Only supporting 1 to 1 mapping between user address space and GPU
address space at the moment as that is the expected use case. uAPI
defines interface for non 1 to 1 but enforces 1 to 1, this restriction
can be lifted if use cases arrise for non 1 to 1 mappings.

This patch essentially short-circuits the code in the existing VM bind
paths to avoid populating page tables when the
DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag is set.

v3:
 - Call vm_bind_ioctl_ops_fini on -ENODATA
 - Don't allow DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR on non-faulting VMs
 - s/DRM_XE_VM_BIND_FLAG_SYSTEM_ALLOCATOR/DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR (Thomas)
 - Rework commit message for expected usage (Thomas)
 - Describe state of code after patch in commit message (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_pt.c       |  76 ++++++++++++----
 drivers/gpu/drm/xe/xe_vm.c       | 150 +++++++++++++++++++------------
 drivers/gpu/drm/xe/xe_vm.h       |   8 +-
 drivers/gpu/drm/xe/xe_vm_types.h |   3 +
 include/uapi/drm/xe_drm.h        |  19 +++-
 5 files changed, 182 insertions(+), 74 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index 65c3c1688710..8c99b4fddff7 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -1069,6 +1069,11 @@ static int op_add_deps(struct xe_vm *vm, struct xe_vma_op *op,
 {
 	int err = 0;
 
+	/*
+	 * No need to check for is_cpu_addr_mirror here as vma_add_deps is a
+	 * NOP if VMA is_cpu_addr_mirror
+	 */
+
 	switch (op->base.op) {
 	case DRM_GPUVA_OP_MAP:
 		if (!op->map.immediate && xe_vm_in_fault_mode(vm))
@@ -1646,6 +1651,7 @@ static int bind_op_prepare(struct xe_vm *vm, struct xe_tile *tile,
 	struct xe_vm_pgtable_update_op *pt_op = &pt_update_ops->ops[current_op];
 	int err;
 
+	xe_tile_assert(tile, !xe_vma_is_cpu_addr_mirror(vma));
 	xe_bo_assert_held(xe_vma_bo(vma));
 
 	vm_dbg(&xe_vma_vm(vma)->xe->drm,
@@ -1713,6 +1719,7 @@ static int unbind_op_prepare(struct xe_tile *tile,
 	if (!((vma->tile_present | vma->tile_staged) & BIT(tile->id)))
 		return 0;
 
+	xe_tile_assert(tile, !xe_vma_is_cpu_addr_mirror(vma));
 	xe_bo_assert_held(xe_vma_bo(vma));
 
 	vm_dbg(&xe_vma_vm(vma)->xe->drm,
@@ -1759,15 +1766,21 @@ static int op_prepare(struct xe_vm *vm,
 
 	switch (op->base.op) {
 	case DRM_GPUVA_OP_MAP:
-		if (!op->map.immediate && xe_vm_in_fault_mode(vm))
+		if ((!op->map.immediate && xe_vm_in_fault_mode(vm)) ||
+		    op->map.is_cpu_addr_mirror)
 			break;
 
 		err = bind_op_prepare(vm, tile, pt_update_ops, op->map.vma);
 		pt_update_ops->wait_vm_kernel = true;
 		break;
 	case DRM_GPUVA_OP_REMAP:
-		err = unbind_op_prepare(tile, pt_update_ops,
-					gpuva_to_vma(op->base.remap.unmap->va));
+	{
+		struct xe_vma *old = gpuva_to_vma(op->base.remap.unmap->va);
+
+		if (xe_vma_is_cpu_addr_mirror(old))
+			break;
+
+		err = unbind_op_prepare(tile, pt_update_ops, old);
 
 		if (!err && op->remap.prev) {
 			err = bind_op_prepare(vm, tile, pt_update_ops,
@@ -1780,15 +1793,28 @@ static int op_prepare(struct xe_vm *vm,
 			pt_update_ops->wait_vm_bookkeep = true;
 		}
 		break;
+	}
 	case DRM_GPUVA_OP_UNMAP:
-		err = unbind_op_prepare(tile, pt_update_ops,
-					gpuva_to_vma(op->base.unmap.va));
+	{
+		struct xe_vma *vma = gpuva_to_vma(op->base.unmap.va);
+
+		if (xe_vma_is_cpu_addr_mirror(vma))
+			break;
+
+		err = unbind_op_prepare(tile, pt_update_ops, vma);
 		break;
+	}
 	case DRM_GPUVA_OP_PREFETCH:
-		err = bind_op_prepare(vm, tile, pt_update_ops,
-				      gpuva_to_vma(op->base.prefetch.va));
+	{
+		struct xe_vma *vma = gpuva_to_vma(op->base.prefetch.va);
+
+		if (xe_vma_is_cpu_addr_mirror(vma))
+			break;
+
+		err = bind_op_prepare(vm, tile, pt_update_ops, vma);
 		pt_update_ops->wait_vm_kernel = true;
 		break;
+	}
 	default:
 		drm_warn(&vm->xe->drm, "NOT POSSIBLE");
 	}
@@ -1858,6 +1884,8 @@ static void bind_op_commit(struct xe_vm *vm, struct xe_tile *tile,
 			   struct xe_vma *vma, struct dma_fence *fence,
 			   struct dma_fence *fence2)
 {
+	xe_tile_assert(tile, !xe_vma_is_cpu_addr_mirror(vma));
+
 	if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) {
 		dma_resv_add_fence(xe_vma_bo(vma)->ttm.base.resv, fence,
 				   pt_update_ops->wait_vm_bookkeep ?
@@ -1891,6 +1919,8 @@ static void unbind_op_commit(struct xe_vm *vm, struct xe_tile *tile,
 			     struct xe_vma *vma, struct dma_fence *fence,
 			     struct dma_fence *fence2)
 {
+	xe_tile_assert(tile, !xe_vma_is_cpu_addr_mirror(vma));
+
 	if (!xe_vma_has_no_bo(vma) && !xe_vma_bo(vma)->vm) {
 		dma_resv_add_fence(xe_vma_bo(vma)->ttm.base.resv, fence,
 				   pt_update_ops->wait_vm_bookkeep ?
@@ -1925,16 +1955,21 @@ static void op_commit(struct xe_vm *vm,
 
 	switch (op->base.op) {
 	case DRM_GPUVA_OP_MAP:
-		if (!op->map.immediate && xe_vm_in_fault_mode(vm))
+		if ((!op->map.immediate && xe_vm_in_fault_mode(vm)) ||
+		    op->map.is_cpu_addr_mirror)
 			break;
 
 		bind_op_commit(vm, tile, pt_update_ops, op->map.vma, fence,
 			       fence2);
 		break;
 	case DRM_GPUVA_OP_REMAP:
-		unbind_op_commit(vm, tile, pt_update_ops,
-				 gpuva_to_vma(op->base.remap.unmap->va), fence,
-				 fence2);
+	{
+		struct xe_vma *old = gpuva_to_vma(op->base.remap.unmap->va);
+
+		if (xe_vma_is_cpu_addr_mirror(old))
+			break;
+
+		unbind_op_commit(vm, tile, pt_update_ops, old, fence, fence2);
 
 		if (op->remap.prev)
 			bind_op_commit(vm, tile, pt_update_ops, op->remap.prev,
@@ -1943,14 +1978,25 @@ static void op_commit(struct xe_vm *vm,
 			bind_op_commit(vm, tile, pt_update_ops, op->remap.next,
 				       fence, fence2);
 		break;
+	}
 	case DRM_GPUVA_OP_UNMAP:
-		unbind_op_commit(vm, tile, pt_update_ops,
-				 gpuva_to_vma(op->base.unmap.va), fence, fence2);
+	{
+		struct xe_vma *vma = gpuva_to_vma(op->base.unmap.va);
+
+		if (!xe_vma_is_cpu_addr_mirror(vma))
+			unbind_op_commit(vm, tile, pt_update_ops, vma, fence,
+					 fence2);
 		break;
+	}
 	case DRM_GPUVA_OP_PREFETCH:
-		bind_op_commit(vm, tile, pt_update_ops,
-			       gpuva_to_vma(op->base.prefetch.va), fence, fence2);
+	{
+		struct xe_vma *vma = gpuva_to_vma(op->base.prefetch.va);
+
+		if (!xe_vma_is_cpu_addr_mirror(vma))
+			bind_op_commit(vm, tile, pt_update_ops, vma, fence,
+				       fence2);
 		break;
+	}
 	default:
 		drm_warn(&vm->xe->drm, "NOT POSSIBLE");
 	}
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index b4a44e1ea167..bef0fae881ac 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -901,9 +901,10 @@ static void xe_vma_free(struct xe_vma *vma)
 		kfree(vma);
 }
 
-#define VMA_CREATE_FLAG_READ_ONLY	BIT(0)
-#define VMA_CREATE_FLAG_IS_NULL		BIT(1)
-#define VMA_CREATE_FLAG_DUMPABLE	BIT(2)
+#define VMA_CREATE_FLAG_READ_ONLY		BIT(0)
+#define VMA_CREATE_FLAG_IS_NULL			BIT(1)
+#define VMA_CREATE_FLAG_DUMPABLE		BIT(2)
+#define VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR	BIT(3)
 
 static struct xe_vma *xe_vma_create(struct xe_vm *vm,
 				    struct xe_bo *bo,
@@ -917,6 +918,8 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm,
 	bool read_only = (flags & VMA_CREATE_FLAG_READ_ONLY);
 	bool is_null = (flags & VMA_CREATE_FLAG_IS_NULL);
 	bool dumpable = (flags & VMA_CREATE_FLAG_DUMPABLE);
+	bool is_cpu_addr_mirror =
+		(flags & VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR);
 
 	xe_assert(vm->xe, start < end);
 	xe_assert(vm->xe, end < vm->size);
@@ -925,7 +928,7 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm,
 	 * Allocate and ensure that the xe_vma_is_userptr() return
 	 * matches what was allocated.
 	 */
-	if (!bo && !is_null) {
+	if (!bo && !is_null && !is_cpu_addr_mirror) {
 		struct xe_userptr_vma *uvma = kzalloc(sizeof(*uvma), GFP_KERNEL);
 
 		if (!uvma)
@@ -937,6 +940,8 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm,
 		if (!vma)
 			return ERR_PTR(-ENOMEM);
 
+		if (is_cpu_addr_mirror)
+			vma->gpuva.flags |= XE_VMA_SYSTEM_ALLOCATOR;
 		if (is_null)
 			vma->gpuva.flags |= DRM_GPUVA_SPARSE;
 		if (bo)
@@ -979,7 +984,7 @@ static struct xe_vma *xe_vma_create(struct xe_vm *vm,
 		drm_gpuva_link(&vma->gpuva, vm_bo);
 		drm_gpuvm_bo_put(vm_bo);
 	} else /* userptr or null */ {
-		if (!is_null) {
+		if (!is_null && !is_cpu_addr_mirror) {
 			struct xe_userptr *userptr = &to_userptr_vma(vma)->userptr;
 			u64 size = end - start + 1;
 			int err;
@@ -1029,7 +1034,7 @@ static void xe_vma_destroy_late(struct xe_vma *vma)
 		 */
 		mmu_interval_notifier_remove(&userptr->notifier);
 		xe_vm_put(vm);
-	} else if (xe_vma_is_null(vma)) {
+	} else if (xe_vma_is_null(vma) || xe_vma_is_cpu_addr_mirror(vma)) {
 		xe_vm_put(vm);
 	} else {
 		xe_bo_put(xe_vma_bo(vma));
@@ -1068,7 +1073,7 @@ static void xe_vma_destroy(struct xe_vma *vma, struct dma_fence *fence)
 		spin_lock(&vm->userptr.invalidated_lock);
 		list_del(&to_userptr_vma(vma)->userptr.invalidate_link);
 		spin_unlock(&vm->userptr.invalidated_lock);
-	} else if (!xe_vma_is_null(vma)) {
+	} else if (!xe_vma_is_null(vma) && !xe_vma_is_cpu_addr_mirror(vma)) {
 		xe_bo_assert_held(xe_vma_bo(vma));
 
 		drm_gpuva_unlink(&vma->gpuva);
@@ -1968,6 +1973,8 @@ vm_bind_ioctl_ops_create(struct xe_vm *vm, struct xe_bo *bo,
 			op->map.read_only =
 				flags & DRM_XE_VM_BIND_FLAG_READONLY;
 			op->map.is_null = flags & DRM_XE_VM_BIND_FLAG_NULL;
+			op->map.is_cpu_addr_mirror = flags &
+				DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR;
 			op->map.dumpable = flags & DRM_XE_VM_BIND_FLAG_DUMPABLE;
 			op->map.pat_index = pat_index;
 		} else if (__op->op == DRM_GPUVA_OP_PREFETCH) {
@@ -2160,6 +2167,8 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 				VMA_CREATE_FLAG_IS_NULL : 0;
 			flags |= op->map.dumpable ?
 				VMA_CREATE_FLAG_DUMPABLE : 0;
+			flags |= op->map.is_cpu_addr_mirror ?
+				VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR : 0;
 
 			vma = new_vma(vm, &op->base.map, op->map.pat_index,
 				      flags);
@@ -2167,7 +2176,8 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 				return PTR_ERR(vma);
 
 			op->map.vma = vma;
-			if (op->map.immediate || !xe_vm_in_fault_mode(vm))
+			if ((op->map.immediate || !xe_vm_in_fault_mode(vm)) &&
+			    !op->map.is_cpu_addr_mirror)
 				xe_vma_ops_incr_pt_update_ops(vops,
 							      op->tile_mask);
 			break;
@@ -2176,21 +2186,24 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 		{
 			struct xe_vma *old =
 				gpuva_to_vma(op->base.remap.unmap->va);
+			bool skip = xe_vma_is_cpu_addr_mirror(old);
 
 			op->remap.start = xe_vma_start(old);
 			op->remap.range = xe_vma_size(old);
 
-			if (op->base.remap.prev) {
-				flags |= op->base.remap.unmap->va->flags &
-					XE_VMA_READ_ONLY ?
-					VMA_CREATE_FLAG_READ_ONLY : 0;
-				flags |= op->base.remap.unmap->va->flags &
-					DRM_GPUVA_SPARSE ?
-					VMA_CREATE_FLAG_IS_NULL : 0;
-				flags |= op->base.remap.unmap->va->flags &
-					XE_VMA_DUMPABLE ?
-					VMA_CREATE_FLAG_DUMPABLE : 0;
+			flags |= op->base.remap.unmap->va->flags &
+				XE_VMA_READ_ONLY ?
+				VMA_CREATE_FLAG_READ_ONLY : 0;
+			flags |= op->base.remap.unmap->va->flags &
+				DRM_GPUVA_SPARSE ?
+				VMA_CREATE_FLAG_IS_NULL : 0;
+			flags |= op->base.remap.unmap->va->flags &
+				XE_VMA_DUMPABLE ?
+				VMA_CREATE_FLAG_DUMPABLE : 0;
+			flags |= xe_vma_is_cpu_addr_mirror(old) ?
+				VMA_CREATE_FLAG_IS_SYSTEM_ALLOCATOR : 0;
 
+			if (op->base.remap.prev) {
 				vma = new_vma(vm, op->base.remap.prev,
 					      old->pat_index, flags);
 				if (IS_ERR(vma))
@@ -2202,9 +2215,10 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 				 * Userptr creates a new SG mapping so
 				 * we must also rebind.
 				 */
-				op->remap.skip_prev = !xe_vma_is_userptr(old) &&
+				op->remap.skip_prev = skip ||
+					(!xe_vma_is_userptr(old) &&
 					IS_ALIGNED(xe_vma_end(vma),
-						   xe_vma_max_pte_size(old));
+						   xe_vma_max_pte_size(old)));
 				if (op->remap.skip_prev) {
 					xe_vma_set_pte_size(vma, xe_vma_max_pte_size(old));
 					op->remap.range -=
@@ -2220,16 +2234,6 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 			}
 
 			if (op->base.remap.next) {
-				flags |= op->base.remap.unmap->va->flags &
-					XE_VMA_READ_ONLY ?
-					VMA_CREATE_FLAG_READ_ONLY : 0;
-				flags |= op->base.remap.unmap->va->flags &
-					DRM_GPUVA_SPARSE ?
-					VMA_CREATE_FLAG_IS_NULL : 0;
-				flags |= op->base.remap.unmap->va->flags &
-					XE_VMA_DUMPABLE ?
-					VMA_CREATE_FLAG_DUMPABLE : 0;
-
 				vma = new_vma(vm, op->base.remap.next,
 					      old->pat_index, flags);
 				if (IS_ERR(vma))
@@ -2241,9 +2245,10 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 				 * Userptr creates a new SG mapping so
 				 * we must also rebind.
 				 */
-				op->remap.skip_next = !xe_vma_is_userptr(old) &&
+				op->remap.skip_next = skip ||
+					(!xe_vma_is_userptr(old) &&
 					IS_ALIGNED(xe_vma_start(vma),
-						   xe_vma_max_pte_size(old));
+						   xe_vma_max_pte_size(old)));
 				if (op->remap.skip_next) {
 					xe_vma_set_pte_size(vma, xe_vma_max_pte_size(old));
 					op->remap.range -=
@@ -2256,14 +2261,27 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 					xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
 				}
 			}
-			xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
+			if (!skip)
+				xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
 			break;
 		}
 		case DRM_GPUVA_OP_UNMAP:
+		{
+			struct xe_vma *vma = gpuva_to_vma(op->base.unmap.va);
+
+			if (!xe_vma_is_cpu_addr_mirror(vma))
+				xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
+			break;
+		}
 		case DRM_GPUVA_OP_PREFETCH:
+		{
+			struct xe_vma *vma = gpuva_to_vma(op->base.prefetch.va);
+
 			/* FIXME: Need to skip some prefetch ops */
-			xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
+			if (!xe_vma_is_cpu_addr_mirror(vma))
+				xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
 			break;
+		}
 		default:
 			drm_warn(&vm->xe->drm, "NOT POSSIBLE");
 		}
@@ -2665,10 +2683,12 @@ static void vm_bind_ioctl_ops_fini(struct xe_vm *vm, struct xe_vma_ops *vops,
 	}
 	if (ufence)
 		xe_sync_ufence_put(ufence);
-	for (i = 0; i < vops->num_syncs; i++)
-		xe_sync_entry_signal(vops->syncs + i, fence);
-	xe_exec_queue_last_fence_set(wait_exec_queue, vm, fence);
-	dma_fence_put(fence);
+	if (fence) {
+		for (i = 0; i < vops->num_syncs; i++)
+			xe_sync_entry_signal(vops->syncs + i, fence);
+		xe_exec_queue_last_fence_set(wait_exec_queue, vm, fence);
+		dma_fence_put(fence);
+	}
 }
 
 static int vm_bind_ioctl_ops_execute(struct xe_vm *vm,
@@ -2691,6 +2711,8 @@ static int vm_bind_ioctl_ops_execute(struct xe_vm *vm,
 		fence = ops_execute(vm, vops);
 		if (IS_ERR(fence)) {
 			err = PTR_ERR(fence);
+			if (err == -ENODATA)
+				vm_bind_ioctl_ops_fini(vm, vops, NULL);
 			goto unlock;
 		}
 
@@ -2707,7 +2729,8 @@ ALLOW_ERROR_INJECTION(vm_bind_ioctl_ops_execute, ERRNO);
 	(DRM_XE_VM_BIND_FLAG_READONLY | \
 	 DRM_XE_VM_BIND_FLAG_IMMEDIATE | \
 	 DRM_XE_VM_BIND_FLAG_NULL | \
-	 DRM_XE_VM_BIND_FLAG_DUMPABLE)
+	 DRM_XE_VM_BIND_FLAG_DUMPABLE | \
+	 DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR)
 
 #ifdef TEST_VM_OPS_ERROR
 #define SUPPORTED_FLAGS	(SUPPORTED_FLAGS_STUB | FORCE_OP_ERROR)
@@ -2718,7 +2741,7 @@ ALLOW_ERROR_INJECTION(vm_bind_ioctl_ops_execute, ERRNO);
 #define XE_64K_PAGE_MASK 0xffffull
 #define ALL_DRM_XE_SYNCS_FLAGS (DRM_XE_SYNCS_FLAG_WAIT_FOR_OP)
 
-static int vm_bind_ioctl_check_args(struct xe_device *xe,
+static int vm_bind_ioctl_check_args(struct xe_device *xe, struct xe_vm *vm,
 				    struct drm_xe_vm_bind *args,
 				    struct drm_xe_vm_bind_op **bind_ops)
 {
@@ -2763,9 +2786,23 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe,
 		u64 obj_offset = (*bind_ops)[i].obj_offset;
 		u32 prefetch_region = (*bind_ops)[i].prefetch_mem_region_instance;
 		bool is_null = flags & DRM_XE_VM_BIND_FLAG_NULL;
+		bool is_cpu_addr_mirror = flags &
+			DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR;
 		u16 pat_index = (*bind_ops)[i].pat_index;
 		u16 coh_mode;
 
+		/* FIXME: Disabling CPU address mirror for now */
+		if (XE_IOCTL_DBG(xe, is_cpu_addr_mirror)) {
+			err = -EOPNOTSUPP;
+			goto free_bind_ops;
+		}
+
+		if (XE_IOCTL_DBG(xe, is_cpu_addr_mirror &&
+				 !xe_vm_in_fault_mode(vm))) {
+			err = -EINVAL;
+			goto free_bind_ops;
+		}
+
 		if (XE_IOCTL_DBG(xe, pat_index >= xe->pat.n_entries)) {
 			err = -EINVAL;
 			goto free_bind_ops;
@@ -2786,13 +2823,14 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe,
 
 		if (XE_IOCTL_DBG(xe, op > DRM_XE_VM_BIND_OP_PREFETCH) ||
 		    XE_IOCTL_DBG(xe, flags & ~SUPPORTED_FLAGS) ||
-		    XE_IOCTL_DBG(xe, obj && is_null) ||
-		    XE_IOCTL_DBG(xe, obj_offset && is_null) ||
+		    XE_IOCTL_DBG(xe, obj && (is_null || is_cpu_addr_mirror)) ||
+		    XE_IOCTL_DBG(xe, obj_offset && (is_null ||
+				 is_cpu_addr_mirror)) ||
 		    XE_IOCTL_DBG(xe, op != DRM_XE_VM_BIND_OP_MAP &&
-				 is_null) ||
+				 (is_null || is_cpu_addr_mirror)) ||
 		    XE_IOCTL_DBG(xe, !obj &&
 				 op == DRM_XE_VM_BIND_OP_MAP &&
-				 !is_null) ||
+				 !is_null && !is_cpu_addr_mirror) ||
 		    XE_IOCTL_DBG(xe, !obj &&
 				 op == DRM_XE_VM_BIND_OP_UNMAP_ALL) ||
 		    XE_IOCTL_DBG(xe, addr &&
@@ -2934,15 +2972,19 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 	int err;
 	int i;
 
-	err = vm_bind_ioctl_check_args(xe, args, &bind_ops);
+	vm = xe_vm_lookup(xef, args->vm_id);
+	if (XE_IOCTL_DBG(xe, !vm))
+		return -EINVAL;
+
+	err = vm_bind_ioctl_check_args(xe, vm, args, &bind_ops);
 	if (err)
-		return err;
+		goto put_vm;
 
 	if (args->exec_queue_id) {
 		q = xe_exec_queue_lookup(xef, args->exec_queue_id);
 		if (XE_IOCTL_DBG(xe, !q)) {
 			err = -ENOENT;
-			goto free_objs;
+			goto put_vm;
 		}
 
 		if (XE_IOCTL_DBG(xe, !(q->flags & EXEC_QUEUE_FLAG_VM))) {
@@ -2951,15 +2993,9 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 		}
 	}
 
-	vm = xe_vm_lookup(xef, args->vm_id);
-	if (XE_IOCTL_DBG(xe, !vm)) {
-		err = -EINVAL;
-		goto put_exec_queue;
-	}
-
 	err = down_write_killable(&vm->lock);
 	if (err)
-		goto put_vm;
+		goto put_exec_queue;
 
 	if (XE_IOCTL_DBG(xe, xe_vm_is_closed_or_banned(vm))) {
 		err = -ENOENT;
@@ -3116,12 +3152,11 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 		xe_bo_put(bos[i]);
 release_vm_lock:
 	up_write(&vm->lock);
-put_vm:
-	xe_vm_put(vm);
 put_exec_queue:
 	if (q)
 		xe_exec_queue_put(q);
-free_objs:
+put_vm:
+	xe_vm_put(vm);
 	kvfree(bos);
 	kvfree(ops);
 	if (args->num_binds > 1)
@@ -3178,6 +3213,7 @@ int xe_vm_invalidate_vma(struct xe_vma *vma)
 	int ret = 0;
 
 	xe_assert(xe, !xe_vma_is_null(vma));
+	xe_assert(xe, !xe_vma_is_cpu_addr_mirror(vma));
 	trace_xe_vma_invalidate(vma);
 
 	vm_dbg(&xe_vma_vm(vma)->xe->drm,
diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h
index 23adb7442881..0e54a0e8768d 100644
--- a/drivers/gpu/drm/xe/xe_vm.h
+++ b/drivers/gpu/drm/xe/xe_vm.h
@@ -150,6 +150,11 @@ static inline bool xe_vma_is_null(struct xe_vma *vma)
 	return vma->gpuva.flags & DRM_GPUVA_SPARSE;
 }
 
+static inline bool xe_vma_is_cpu_addr_mirror(struct xe_vma *vma)
+{
+	return vma->gpuva.flags & XE_VMA_SYSTEM_ALLOCATOR;
+}
+
 static inline bool xe_vma_has_no_bo(struct xe_vma *vma)
 {
 	return !xe_vma_bo(vma);
@@ -157,7 +162,8 @@ static inline bool xe_vma_has_no_bo(struct xe_vma *vma)
 
 static inline bool xe_vma_is_userptr(struct xe_vma *vma)
 {
-	return xe_vma_has_no_bo(vma) && !xe_vma_is_null(vma);
+	return xe_vma_has_no_bo(vma) && !xe_vma_is_null(vma) &&
+		!xe_vma_is_cpu_addr_mirror(vma);
 }
 
 /**
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index 7f9a303e51d8..f6855e4fb9e6 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -42,6 +42,7 @@ struct xe_vm_pgtable_update_op;
 #define XE_VMA_PTE_64K		(DRM_GPUVA_USERBITS << 6)
 #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)
 
 /** struct xe_userptr - User pointer */
 struct xe_userptr {
@@ -294,6 +295,8 @@ struct xe_vma_op_map {
 	bool read_only;
 	/** @is_null: is NULL binding */
 	bool is_null;
+	/** @is_cpu_addr_mirror: is CPU address mirror binding */
+	bool is_cpu_addr_mirror;
 	/** @dumpable: whether BO is dumped on GPU hang */
 	bool dumpable;
 	/** @pat_index: The pat index to use for this operation. */
diff --git a/include/uapi/drm/xe_drm.h b/include/uapi/drm/xe_drm.h
index f62689ca861a..768cf11ebde3 100644
--- a/include/uapi/drm/xe_drm.h
+++ b/include/uapi/drm/xe_drm.h
@@ -906,6 +906,12 @@ struct drm_xe_vm_destroy {
  *    will only be valid for DRM_XE_VM_BIND_OP_MAP operations, the BO
  *    handle MBZ, and the BO offset MBZ. This flag is intended to
  *    implement VK sparse bindings.
+ *  - %DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR - When the CPU address mirror flag is
+ *    set, no mappings are created rather the range is reserved for CPU address
+ *    mirroring which will be populated on GPU page faults or prefetches. Only
+ *    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.
  */
 struct drm_xe_vm_bind_op {
 	/** @extensions: Pointer to the first extension struct, if any */
@@ -958,7 +964,9 @@ struct drm_xe_vm_bind_op {
 	 * on the @pat_index. For such mappings there is no actual memory being
 	 * mapped (the address in the PTE is invalid), so the various PAT memory
 	 * attributes likely do not apply.  Simply leaving as zero is one
-	 * option (still a valid pat_index).
+	 * option (still a valid pat_index). Same applies to
+	 * DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR bindings as for such mapping
+	 * there is no actual memory being mapped.
 	 */
 	__u16 pat_index;
 
@@ -974,6 +982,14 @@ struct drm_xe_vm_bind_op {
 
 		/** @userptr: user pointer to bind on */
 		__u64 userptr;
+
+		/**
+		 * @cpu_addr_mirror_offset: Offset from GPU @addr to create
+		 * CPU address mirror mappings. MBZ with current level of
+		 * support (e.g. 1 to 1 mapping between GPU and CPU mappings
+		 * only supported).
+		 */
+		__s64 cpu_addr_mirror_offset;
 	};
 
 	/**
@@ -996,6 +1012,7 @@ struct drm_xe_vm_bind_op {
 #define DRM_XE_VM_BIND_FLAG_IMMEDIATE	(1 << 1)
 #define DRM_XE_VM_BIND_FLAG_NULL	(1 << 2)
 #define DRM_XE_VM_BIND_FLAG_DUMPABLE	(1 << 3)
+#define DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR	(1 << 4)
 	/** @flags: Bind flags */
 	__u32 flags;
 
-- 
2.34.1

[PATCH v3 08/30] drm/xe: Add SVM init / close / fini to faulting VMs

[Matthew Brost]

Add SVM init / close / fini to faulting VMs. Minimual implementation
acting as a placeholder for follow on patches.

v2:
 - Add close function
v3:
 - Better commit message (Thomas)
 - Kernel doc (Thomas)
 - Update chunk array to be unsigned long (Thomas)
 - Use new drm_gpusvm.h header location (Thomas)
 - Newlines between functions in xe_svm.h (Thomas)
 - Call drm_gpusvm_driver_set_lock in init (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/Makefile      |  1 +
 drivers/gpu/drm/xe/xe_svm.c      | 73 ++++++++++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_svm.h      | 17 ++++++++
 drivers/gpu/drm/xe/xe_vm.c       | 12 ++++++
 drivers/gpu/drm/xe/xe_vm_types.h |  7 +++
 5 files changed, 110 insertions(+)
 create mode 100644 drivers/gpu/drm/xe/xe_svm.c
 create mode 100644 drivers/gpu/drm/xe/xe_svm.h

diff --git a/drivers/gpu/drm/xe/Makefile b/drivers/gpu/drm/xe/Makefile
index 5c97ad6ed738..94d8850fdbb4 100644
--- a/drivers/gpu/drm/xe/Makefile
+++ b/drivers/gpu/drm/xe/Makefile
@@ -95,6 +95,7 @@ xe-y += xe_bb.o \
 	xe_sa.o \
 	xe_sched_job.o \
 	xe_step.o \
+	xe_svm.o \
 	xe_sync.o \
 	xe_tile.o \
 	xe_tile_sysfs.o \
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
new file mode 100644
index 000000000000..79da859f02b1
--- /dev/null
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -0,0 +1,73 @@
+// SPDX-License-Identifier: MIT
+/*
+ * Copyright © 2024 Intel Corporation
+ */
+
+#include "xe_svm.h"
+#include "xe_vm.h"
+#include "xe_vm_types.h"
+
+static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
+			      struct drm_gpusvm_notifier *notifier,
+			      const struct mmu_notifier_range *mmu_range)
+{
+	/* TODO: Implement */
+}
+
+static const struct drm_gpusvm_ops gpusvm_ops = {
+	.invalidate = xe_svm_invalidate,
+};
+
+static const unsigned long fault_chunk_sizes[] = {
+	SZ_2M,
+	SZ_64K,
+	SZ_4K,
+};
+
+/**
+ * xe_svm_init() - SVM initialize
+ * @vm: The VM.
+ *
+ * Initialize SVM state which is embedded within the VM.
+ *
+ * Return: 0 on success, negative error code on error.
+ */
+int xe_svm_init(struct xe_vm *vm)
+{
+	int err;
+
+	err = drm_gpusvm_init(&vm->svm.gpusvm, "Xe SVM", &vm->xe->drm,
+			      current->mm, NULL, 0, vm->size,
+			      SZ_512M, &gpusvm_ops, fault_chunk_sizes,
+			      ARRAY_SIZE(fault_chunk_sizes));
+	if (err)
+		return err;
+
+	drm_gpusvm_driver_set_lock(&vm->svm.gpusvm, &vm->lock);
+
+	return 0;
+}
+
+/**
+ * xe_svm_close() - SVM close
+ * @vm: The VM.
+ *
+ * Close SVM state (i.e., stop and flush all SVM actions).
+ */
+void xe_svm_close(struct xe_vm *vm)
+{
+	xe_assert(vm->xe, xe_vm_is_closed(vm));
+}
+
+/**
+ * xe_svm_fini() - SVM finalize
+ * @vm: The VM.
+ *
+ * Finalize SVM state which is embedded within the VM.
+ */
+void xe_svm_fini(struct xe_vm *vm)
+{
+	xe_assert(vm->xe, xe_vm_is_closed(vm));
+
+	drm_gpusvm_fini(&vm->svm.gpusvm);
+}
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
new file mode 100644
index 000000000000..49cfd938aa17
--- /dev/null
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -0,0 +1,17 @@
+/* SPDX-License-Identifier: MIT */
+/*
+ * Copyright © 2024 Intel Corporation
+ */
+
+#ifndef _XE_SVM_H_
+#define _XE_SVM_H_
+
+struct xe_vm;
+
+int xe_svm_init(struct xe_vm *vm);
+
+void xe_svm_fini(struct xe_vm *vm);
+
+void xe_svm_close(struct xe_vm *vm);
+
+#endif
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index bef0fae881ac..56ef76bd3e3f 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -34,6 +34,7 @@
 #include "xe_preempt_fence.h"
 #include "xe_pt.h"
 #include "xe_res_cursor.h"
+#include "xe_svm.h"
 #include "xe_sync.h"
 #include "xe_trace_bo.h"
 #include "xe_wa.h"
@@ -1504,6 +1505,12 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags)
 		}
 	}
 
+	if (flags & XE_VM_FLAG_FAULT_MODE) {
+		err = xe_svm_init(vm);
+		if (err)
+			goto err_close;
+	}
+
 	if (number_tiles > 1)
 		vm->composite_fence_ctx = dma_fence_context_alloc(1);
 
@@ -1549,6 +1556,8 @@ void xe_vm_close_and_put(struct xe_vm *vm)
 	xe_vm_close(vm);
 	if (xe_vm_in_preempt_fence_mode(vm))
 		flush_work(&vm->preempt.rebind_work);
+	if (xe_vm_in_fault_mode(vm))
+		xe_svm_close(vm);
 
 	down_write(&vm->lock);
 	for_each_tile(tile, xe, id) {
@@ -1617,6 +1626,9 @@ void xe_vm_close_and_put(struct xe_vm *vm)
 		xe_vma_destroy_unlocked(vma);
 	}
 
+	if (xe_vm_in_fault_mode(vm))
+		xe_svm_fini(vm);
+
 	up_write(&vm->lock);
 
 	down_write(&xe->usm.lock);
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index f6855e4fb9e6..aa075d5e7a3f 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -6,6 +6,7 @@
 #ifndef _XE_VM_TYPES_H_
 #define _XE_VM_TYPES_H_
 
+#include <drm/drm_gpusvm.h>
 #include <drm/drm_gpuvm.h>
 
 #include <linux/dma-resv.h>
@@ -140,6 +141,12 @@ struct xe_vm {
 	/** @gpuvm: base GPUVM used to track VMAs */
 	struct drm_gpuvm gpuvm;
 
+	/** @svm: Shared virtual memory state */
+	struct {
+		/** @svm.gpusvm: base GPUSVM used to track fault allocations */
+		struct drm_gpusvm gpusvm;
+	} svm;
+
 	struct xe_device *xe;
 
 	/* exec queue used for (un)binding vma's */
-- 
2.34.1

[PATCH v3 09/30] drm/xe: Add dma_addr res cursor

[Matthew Brost]

From: Thomas Hellström <thomas.hellstrom@linux.intel.com>

Add dma_addr res cursor which walks an array of drm_pagemap_dma_addr.
Useful for SVM ranges and programing page tables.

v3:
 - Better commit message (Thomas)
 - Use new drm_pagemap.h location

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_res_cursor.h | 116 ++++++++++++++++++++++++++++-
 drivers/gpu/drm/xe/xe_svm.h        |   4 +
 2 files changed, 118 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_res_cursor.h b/drivers/gpu/drm/xe/xe_res_cursor.h
index dca374b6521c..46486087a51d 100644
--- a/drivers/gpu/drm/xe/xe_res_cursor.h
+++ b/drivers/gpu/drm/xe/xe_res_cursor.h
@@ -26,6 +26,7 @@
 
 #include <linux/scatterlist.h>
 
+#include <drm/drm_pagemap.h>
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_range_manager.h>
 #include <drm/ttm/ttm_resource.h>
@@ -34,9 +35,13 @@
 #include "xe_bo.h"
 #include "xe_device.h"
 #include "xe_macros.h"
+#include "xe_svm.h"
 #include "xe_ttm_vram_mgr.h"
 
-/* state back for walking over vram_mgr, stolen_mgr, and gtt_mgr allocations */
+/**
+ * struct xe_res_cursor - state for walking over dma mapping, vram_mgr,
+ * stolen_mgr, and gtt_mgr allocations
+ */
 struct xe_res_cursor {
 	u64 start;
 	u64 size;
@@ -44,7 +49,17 @@ struct xe_res_cursor {
 	void *node;
 	u32 mem_type;
 	struct scatterlist *sgl;
+	/** @dma_addr: Current element in a struct drm_pagemap_dma_addr array */
+	const struct drm_pagemap_dma_addr *dma_addr;
 	struct drm_buddy *mm;
+	/**
+	 * @dma_start: DMA start address for the current segment.
+	 * This may be different to @dma_addr.addr since elements in
+	 * the array may be coalesced to a single segment.
+	 */
+	u64 dma_start;
+	/** @dma_seg_size: Size of the current segment. */
+	u64 dma_seg_size;
 };
 
 static struct drm_buddy *xe_res_get_buddy(struct ttm_resource *res)
@@ -70,6 +85,7 @@ static inline void xe_res_first(struct ttm_resource *res,
 				struct xe_res_cursor *cur)
 {
 	cur->sgl = NULL;
+	cur->dma_addr = NULL;
 	if (!res)
 		goto fallback;
 
@@ -141,6 +157,36 @@ static inline void __xe_res_sg_next(struct xe_res_cursor *cur)
 	cur->sgl = sgl;
 }
 
+/**
+ * __xe_res_dma_next() - Advance the cursor when end-of-segment is reached
+ * @cur: The cursor
+ */
+static inline void __xe_res_dma_next(struct xe_res_cursor *cur)
+{
+	const struct drm_pagemap_dma_addr *addr = cur->dma_addr;
+	u64 start = cur->start;
+
+	while (start >= cur->dma_seg_size) {
+		start -= cur->dma_seg_size;
+		addr++;
+		cur->dma_seg_size = PAGE_SIZE << addr->order;
+	}
+	cur->dma_start = addr->addr;
+
+	/* Coalesce array_elements */
+	while (cur->dma_seg_size - start < cur->remaining) {
+		if (cur->dma_start + cur->dma_seg_size != addr[1].addr ||
+		    addr->proto != addr[1].proto)
+			break;
+		addr++;
+		cur->dma_seg_size += PAGE_SIZE << addr->order;
+	}
+
+	cur->dma_addr = addr;
+	cur->start = start;
+	cur->size = cur->dma_seg_size - start;
+}
+
 /**
  * xe_res_first_sg - initialize a xe_res_cursor with a scatter gather table
  *
@@ -160,11 +206,42 @@ static inline void xe_res_first_sg(const struct sg_table *sg,
 	cur->start = start;
 	cur->remaining = size;
 	cur->size = 0;
+	cur->dma_addr = NULL;
 	cur->sgl = sg->sgl;
 	cur->mem_type = XE_PL_TT;
 	__xe_res_sg_next(cur);
 }
 
+/**
+ * xe_res_first_dma - initialize a xe_res_cursor with dma_addr array
+ *
+ * @dma_addr: struct drm_pagemap_dma_addr array to walk
+ * @start: Start of the range
+ * @size: Size of the range
+ * @cur: cursor object to initialize
+ *
+ * Start walking over the range of allocations between @start and @size.
+ */
+static inline void xe_res_first_dma(const struct drm_pagemap_dma_addr *dma_addr,
+				    u64 start, u64 size,
+				    struct xe_res_cursor *cur)
+{
+	XE_WARN_ON(!dma_addr);
+	XE_WARN_ON(!IS_ALIGNED(start, PAGE_SIZE) ||
+		   !IS_ALIGNED(size, PAGE_SIZE));
+
+	cur->node = NULL;
+	cur->start = start;
+	cur->remaining = size;
+	cur->dma_seg_size = PAGE_SIZE << dma_addr->order;
+	cur->dma_start = 0;
+	cur->size = 0;
+	cur->dma_addr = dma_addr;
+	__xe_res_dma_next(cur);
+	cur->sgl = NULL;
+	cur->mem_type = XE_PL_TT;
+}
+
 /**
  * xe_res_next - advance the cursor
  *
@@ -191,6 +268,12 @@ static inline void xe_res_next(struct xe_res_cursor *cur, u64 size)
 		return;
 	}
 
+	if (cur->dma_addr) {
+		cur->start += size;
+		__xe_res_dma_next(cur);
+		return;
+	}
+
 	if (cur->sgl) {
 		cur->start += size;
 		__xe_res_sg_next(cur);
@@ -232,6 +315,35 @@ static inline void xe_res_next(struct xe_res_cursor *cur, u64 size)
  */
 static inline u64 xe_res_dma(const struct xe_res_cursor *cur)
 {
-	return cur->sgl ? sg_dma_address(cur->sgl) + cur->start : cur->start;
+	if (cur->dma_addr)
+		return cur->dma_start + cur->start;
+	else if (cur->sgl)
+		return sg_dma_address(cur->sgl) + cur->start;
+	else
+		return cur->start;
+}
+
+/**
+ * xe_res_is_vram() - Whether the cursor current dma address points to
+ * same-device VRAM
+ * @cur: The cursor.
+ *
+ * Return: true iff the address returned by xe_res_dma() points to internal vram.
+ */
+static inline bool xe_res_is_vram(const struct xe_res_cursor *cur)
+{
+	if (cur->dma_addr)
+		return cur->dma_addr->proto == XE_INTERCONNECT_VRAM;
+
+	switch (cur->mem_type) {
+	case XE_PL_STOLEN:
+	case XE_PL_VRAM0:
+	case XE_PL_VRAM1:
+		return true;
+	default:
+		break;
+	}
+
+	return false;
 }
 #endif
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index 49cfd938aa17..4569931db622 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -6,6 +6,10 @@
 #ifndef _XE_SVM_H_
 #define _XE_SVM_H_
 
+#include <drm/drm_pagemap.h>
+
+#define XE_INTERCONNECT_VRAM DRM_INTERCONNECT_DRIVER
+
 struct xe_vm;
 
 int xe_svm_init(struct xe_vm *vm);
-- 
2.34.1

[PATCH v3 10/30] drm/xe: Nuke VM's mapping upon close

[Matthew Brost]

Clear root PT entry and invalidate entire VM's address space when
closing the VM. Will prevent the GPU from accessing any of the VM's
memory after closing.

v2:
 - s/vma/vm in kernel doc (CI)
 - Don't nuke migration VM as this occur at driver unload (CI)
v3:
 - Rebase and pull into SVM series (Thomas)
 - Wait for pending binds (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c | 24 +++++++++++++++++++++
 drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h |  2 ++
 drivers/gpu/drm/xe/xe_pt.c                  | 14 ++++++++++++
 drivers/gpu/drm/xe/xe_pt.h                  |  3 +++
 drivers/gpu/drm/xe/xe_vm.c                  | 22 +++++++++++++++++++
 5 files changed, 65 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c
index 665927b80e9e..42d7a35637f0 100644
--- a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c
+++ b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c
@@ -410,6 +410,30 @@ int xe_gt_tlb_invalidation_range(struct xe_gt *gt,
 	return send_tlb_invalidation(&gt->uc.guc, fence, action, len);
 }
 
+/**
+ * xe_gt_tlb_invalidation_vm - Issue a TLB invalidation on this GT for a VM
+ * @gt: graphics tile
+ * @vm: VM to invalidate
+ *
+ * Invalidate entire VM's address space
+ */
+void xe_gt_tlb_invalidation_vm(struct xe_gt *gt, struct xe_vm *vm)
+{
+	struct xe_gt_tlb_invalidation_fence fence;
+	u64 range = 1ull << vm->xe->info.va_bits;
+	int ret;
+
+	xe_gt_tlb_invalidation_fence_init(gt, &fence, true);
+
+	ret = xe_gt_tlb_invalidation_range(gt, &fence, 0, range, vm->usm.asid);
+	if (ret < 0) {
+		xe_gt_tlb_invalidation_fence_fini(&fence);
+		return;
+	}
+
+	xe_gt_tlb_invalidation_fence_wait(&fence);
+}
+
 /**
  * xe_gt_tlb_invalidation_vma - Issue a TLB invalidation on this GT for a VMA
  * @gt: graphics tile
diff --git a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h
index 00b1c6c01e8d..d0b78041bed6 100644
--- a/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h
+++ b/drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h
@@ -12,6 +12,7 @@
 
 struct xe_gt;
 struct xe_guc;
+struct xe_vm;
 struct xe_vma;
 
 int xe_gt_tlb_invalidation_init(struct xe_gt *gt);
@@ -20,6 +21,7 @@ int xe_gt_tlb_invalidation_ggtt(struct xe_gt *gt);
 int xe_gt_tlb_invalidation_vma(struct xe_gt *gt,
 			       struct xe_gt_tlb_invalidation_fence *fence,
 			       struct xe_vma *vma);
+void xe_gt_tlb_invalidation_vm(struct xe_gt *gt, struct xe_vm *vm);
 int xe_gt_tlb_invalidation_range(struct xe_gt *gt,
 				 struct xe_gt_tlb_invalidation_fence *fence,
 				 u64 start, u64 end, u32 asid);
diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index 8c99b4fddff7..e2e62d7eb204 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -214,6 +214,20 @@ void xe_pt_destroy(struct xe_pt *pt, u32 flags, struct llist_head *deferred)
 	xe_pt_free(pt);
 }
 
+/**
+ * xe_pt_clear() - Clear a page-table.
+ * @xe: xe device.
+ * @pt: The page-table.
+ *
+ * Clears page-table by setting to zero.
+ */
+void xe_pt_clear(struct xe_device *xe, struct xe_pt *pt)
+{
+	struct iosys_map *map = &pt->bo->vmap;
+
+	xe_map_memset(xe, map, 0, 0, SZ_4K);
+}
+
 /**
  * DOC: Pagetable building
  *
diff --git a/drivers/gpu/drm/xe/xe_pt.h b/drivers/gpu/drm/xe/xe_pt.h
index 9ab386431cad..8e43912ae8e9 100644
--- a/drivers/gpu/drm/xe/xe_pt.h
+++ b/drivers/gpu/drm/xe/xe_pt.h
@@ -13,6 +13,7 @@ struct dma_fence;
 struct xe_bo;
 struct xe_device;
 struct xe_exec_queue;
+struct xe_svm_range;
 struct xe_sync_entry;
 struct xe_tile;
 struct xe_vm;
@@ -35,6 +36,8 @@ void xe_pt_populate_empty(struct xe_tile *tile, struct xe_vm *vm,
 
 void xe_pt_destroy(struct xe_pt *pt, u32 flags, struct llist_head *deferred);
 
+void xe_pt_clear(struct xe_device *xe, struct xe_pt *pt);
+
 int xe_pt_update_ops_prepare(struct xe_tile *tile, struct xe_vma_ops *vops);
 struct dma_fence *xe_pt_update_ops_run(struct xe_tile *tile,
 				       struct xe_vma_ops *vops);
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 56ef76bd3e3f..ed49722d47ef 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -1537,8 +1537,30 @@ struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags)
 
 static void xe_vm_close(struct xe_vm *vm)
 {
+	bool migration = (vm->flags & XE_VM_FLAG_MIGRATION);
+
 	down_write(&vm->lock);
+
 	vm->size = 0;
+
+	if (!migration) {
+		struct xe_tile *tile;
+		struct xe_gt *gt;
+		u8 id;
+
+		/* Wait for pending binds */
+		dma_resv_wait_timeout(xe_vm_resv(vm),
+				      DMA_RESV_USAGE_BOOKKEEP,
+				      false, MAX_SCHEDULE_TIMEOUT);
+
+		for_each_tile(tile, vm->xe, id)
+			if (vm->pt_root[id])
+				xe_pt_clear(vm->xe, vm->pt_root[id]);
+
+		for_each_gt(gt, vm->xe, id)
+			xe_gt_tlb_invalidation_vm(gt, vm);
+	}
+
 	up_write(&vm->lock);
 }
 
-- 
2.34.1

[PATCH v3 11/30] drm/xe: Add SVM range invalidation and page fault handler

[Matthew Brost]

Add SVM range invalidation vfunc which invalidates PTEs. A new PT layer
function which accepts a SVM range is added to support this. In
addition, add the basic page fault handler which allocates a SVM range
which is used by SVM range invalidation vfunc.

v2:
 - Don't run invalidation if VM is closed
 - Cycle notifier lock in xe_svm_close
 - Drop xe_gt_tlb_invalidation_fence_fini
v3:
 - Better commit message (Thomas)
 - Add lockdep asserts (Thomas)
 - Add kernel doc (Thomas)
 - s/change/changed (Thomas)
 - Use new GPU SVM range / notifier structures
 - Ensure PTEs are zapped / dma mappings are unmapped on VM close (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_gt_pagefault.c |  17 +-
 drivers/gpu/drm/xe/xe_pt.c           |  41 +++++
 drivers/gpu/drm/xe/xe_pt.h           |   2 +
 drivers/gpu/drm/xe/xe_svm.c          | 223 ++++++++++++++++++++++++++-
 drivers/gpu/drm/xe/xe_svm.h          |  32 ++++
 drivers/gpu/drm/xe/xe_vm.c           |   4 +
 6 files changed, 313 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_gt_pagefault.c b/drivers/gpu/drm/xe/xe_gt_pagefault.c
index 2606cd396df5..7e71bf604ae8 100644
--- a/drivers/gpu/drm/xe/xe_gt_pagefault.c
+++ b/drivers/gpu/drm/xe/xe_gt_pagefault.c
@@ -18,6 +18,7 @@
 #include "xe_guc.h"
 #include "xe_guc_ct.h"
 #include "xe_migrate.h"
+#include "xe_svm.h"
 #include "xe_trace_bo.h"
 #include "xe_vm.h"
 
@@ -124,18 +125,17 @@ static int xe_pf_begin(struct drm_exec *exec, struct xe_vma *vma,
 	return 0;
 }
 
-static int handle_vma_pagefault(struct xe_tile *tile, struct pagefault *pf,
-				struct xe_vma *vma)
+static int handle_vma_pagefault(struct xe_tile *tile, struct xe_vma *vma,
+				bool atomic)
 {
 	struct xe_vm *vm = xe_vma_vm(vma);
 	struct drm_exec exec;
 	struct dma_fence *fence;
 	ktime_t end = 0;
 	int err;
-	bool atomic;
 
+	lockdep_assert_held_write(&vm->lock);
 	trace_xe_vma_pagefault(vma);
-	atomic = access_is_atomic(pf->access_type);
 
 	/* Check if VMA is valid */
 	if (vma_is_valid(tile, vma) && !atomic)
@@ -206,6 +206,7 @@ static int handle_pagefault(struct xe_gt *gt, struct pagefault *pf)
 	struct xe_vm *vm;
 	struct xe_vma *vma = NULL;
 	int err;
+	bool atomic;
 
 	/* SW isn't expected to handle TRTT faults */
 	if (pf->trva_fault)
@@ -231,7 +232,13 @@ static int handle_pagefault(struct xe_gt *gt, struct pagefault *pf)
 		goto unlock_vm;
 	}
 
-	err = handle_vma_pagefault(tile, pf, vma);
+	atomic = access_is_atomic(pf->access_type);
+
+	if (xe_vma_is_cpu_addr_mirror(vma))
+		err = xe_svm_handle_pagefault(vm, vma, tile,
+					      pf->page_addr, atomic);
+	else
+		err = handle_vma_pagefault(tile, vma, atomic);
 
 unlock_vm:
 	if (!err)
diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index e2e62d7eb204..f7c4e1b374e8 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -20,6 +20,7 @@
 #include "xe_res_cursor.h"
 #include "xe_sched_job.h"
 #include "xe_sync.h"
+#include "xe_svm.h"
 #include "xe_trace.h"
 #include "xe_ttm_stolen_mgr.h"
 #include "xe_vm.h"
@@ -844,6 +845,46 @@ bool xe_pt_zap_ptes(struct xe_tile *tile, struct xe_vma *vma)
 	return xe_walk.needs_invalidate;
 }
 
+/**
+ * xe_pt_zap_ptes_range() - Zap (zero) gpu ptes of a SVM range
+ * @tile: The tile we're zapping for.
+ * @vm: The VM we're zapping for.
+ * @range: The SVM range we're zapping for.
+ *
+ * SVM invalidation needs to be able to zap the gpu ptes of a given address
+ * range. In order to be able to do that, that function needs access to the
+ * shared page-table entries so it can either clear the leaf PTEs or
+ * clear the pointers to lower-level page-tables. The caller is required
+ * to hold the SVM notifier lock.
+ *
+ * Return: Whether ptes were actually updated and a TLB invalidation is
+ * required.
+ */
+bool xe_pt_zap_ptes_range(struct xe_tile *tile, struct xe_vm *vm,
+			  struct xe_svm_range *range)
+{
+	struct xe_pt_zap_ptes_walk xe_walk = {
+		.base = {
+			.ops = &xe_pt_zap_ptes_ops,
+			.shifts = xe_normal_pt_shifts,
+			.max_level = XE_PT_HIGHEST_LEVEL,
+		},
+		.tile = tile,
+	};
+	struct xe_pt *pt = vm->pt_root[tile->id];
+	u8 pt_mask = (range->tile_present & ~range->tile_invalidated);
+
+	xe_svm_assert_in_notifier(vm);
+
+	if (!(pt_mask & BIT(tile->id)))
+		return false;
+
+	(void)xe_pt_walk_shared(&pt->base, pt->level, range->base.itree.start,
+				range->base.itree.last + 1, &xe_walk.base);
+
+	return xe_walk.needs_invalidate;
+}
+
 static void
 xe_vm_populate_pgtable(struct xe_migrate_pt_update *pt_update, struct xe_tile *tile,
 		       struct iosys_map *map, void *data,
diff --git a/drivers/gpu/drm/xe/xe_pt.h b/drivers/gpu/drm/xe/xe_pt.h
index 8e43912ae8e9..5ecf003d513c 100644
--- a/drivers/gpu/drm/xe/xe_pt.h
+++ b/drivers/gpu/drm/xe/xe_pt.h
@@ -45,5 +45,7 @@ void xe_pt_update_ops_fini(struct xe_tile *tile, struct xe_vma_ops *vops);
 void xe_pt_update_ops_abort(struct xe_tile *tile, struct xe_vma_ops *vops);
 
 bool xe_pt_zap_ptes(struct xe_tile *tile, struct xe_vma *vma);
+bool xe_pt_zap_ptes_range(struct xe_tile *tile, struct xe_vm *vm,
+			  struct xe_svm_range *range);
 
 #endif
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 79da859f02b1..bd7b9c6ea229 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -3,18 +3,198 @@
  * Copyright © 2024 Intel Corporation
  */
 
+#include "xe_gt_tlb_invalidation.h"
+#include "xe_pt.h"
 #include "xe_svm.h"
 #include "xe_vm.h"
 #include "xe_vm_types.h"
 
+static struct xe_vm *gpusvm_to_vm(struct drm_gpusvm *gpusvm)
+{
+	return container_of(gpusvm, struct xe_vm, svm.gpusvm);
+}
+
+static struct xe_vm *range_to_vm(struct drm_gpusvm_range *r)
+{
+	return gpusvm_to_vm(r->gpusvm);
+}
+
+static struct drm_gpusvm_range *
+xe_svm_range_alloc(struct drm_gpusvm *gpusvm)
+{
+	struct xe_svm_range *range;
+
+	range = kzalloc(sizeof(*range), GFP_KERNEL);
+	if (!range)
+		return ERR_PTR(-ENOMEM);
+
+	xe_vm_get(gpusvm_to_vm(gpusvm));
+
+	return &range->base;
+}
+
+static void xe_svm_range_free(struct drm_gpusvm_range *range)
+{
+	xe_vm_put(range_to_vm(range));
+	kfree(range);
+}
+
+static struct xe_svm_range *to_xe_range(struct drm_gpusvm_range *r)
+{
+	return container_of(r, struct xe_svm_range, base);
+}
+
+static u8
+xe_svm_range_notifier_event_begin(struct xe_vm *vm, struct drm_gpusvm_range *r,
+				  const struct mmu_notifier_range *mmu_range,
+				  u64 *adj_start, u64 *adj_end)
+{
+	struct xe_svm_range *range = to_xe_range(r);
+	struct xe_device *xe = vm->xe;
+	struct xe_tile *tile;
+	u8 tile_mask = 0;
+	u8 id;
+
+	xe_svm_assert_in_notifier(vm);
+
+	/* Skip if already unmapped or if no binding exist */
+	if (range->base.flags.unmapped || !range->tile_present)
+		return 0;
+
+	/* Adjust invalidation to range boundaries */
+	if (range->base.itree.start < mmu_range->start)
+		*adj_start = range->base.itree.start;
+	if (range->base.itree.last + 1 > mmu_range->end)
+		*adj_end = range->base.itree.last + 1;
+
+	/*
+	 * XXX: Ideally would zap PTEs in one shot in xe_svm_invalidate but the
+	 * invalidation code can't correctly cope with sparse ranges or
+	 * invalidations spanning multiple ranges.
+	 */
+	for_each_tile(tile, xe, id)
+		if (xe_pt_zap_ptes_range(tile, vm, range)) {
+			tile_mask |= BIT(id);
+			range->tile_invalidated |= BIT(id);
+		}
+
+	return tile_mask;
+}
+
+static void
+xe_svm_range_notifier_event_end(struct xe_vm *vm, struct drm_gpusvm_range *r,
+				const struct mmu_notifier_range *mmu_range)
+{
+	struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
+
+	xe_svm_assert_in_notifier(vm);
+
+	drm_gpusvm_range_unmap_pages(&vm->svm.gpusvm, r, &ctx);
+	/* TODO: Add range to garbage collector if VM is not closed */
+}
+
 static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
 			      struct drm_gpusvm_notifier *notifier,
 			      const struct mmu_notifier_range *mmu_range)
 {
-	/* TODO: Implement */
+	struct xe_vm *vm = gpusvm_to_vm(gpusvm);
+	struct xe_device *xe = vm->xe;
+	struct xe_tile *tile;
+	struct drm_gpusvm_range *r, *first;
+	struct xe_gt_tlb_invalidation_fence
+		fence[XE_MAX_TILES_PER_DEVICE * XE_MAX_GT_PER_TILE];
+	u64 adj_start = mmu_range->start, adj_end = mmu_range->end;
+	u8 tile_mask = 0;
+	u8 id;
+	u32 fence_id = 0;
+	long err;
+
+	xe_svm_assert_in_notifier(vm);
+
+	/* Adjust invalidation to notifier boundaries */
+	if (adj_start < notifier->itree.start)
+		adj_start = notifier->itree.start;
+	if (adj_end > notifier->itree.last + 1)
+		adj_end = notifier->itree.last + 1;
+
+	first = drm_gpusvm_range_find(notifier, adj_start, adj_end);
+	if (!first)
+		return;
+
+	/*
+	 * PTs may be getting destroyed so not safe to touch these but PT should
+	 * be invalidated at this point in time. Regardless we still need to
+	 * ensure any dma mappings are unmapped in the here.
+	 */
+	if (xe_vm_is_closed(vm))
+		goto range_notifier_event_end;
+
+	/*
+	 * XXX: Less than ideal to always wait on VM's resv slots if an
+	 * invalidation is not required. Could walk range list twice to figure
+	 * out if an invalidations is need, but also not ideal.
+	 */
+	err = dma_resv_wait_timeout(xe_vm_resv(vm),
+				    DMA_RESV_USAGE_BOOKKEEP,
+				    false, MAX_SCHEDULE_TIMEOUT);
+	XE_WARN_ON(err <= 0);
+
+	r = first;
+	drm_gpusvm_for_each_range(r, notifier, adj_start, adj_end)
+		tile_mask |= xe_svm_range_notifier_event_begin(vm, r, mmu_range,
+							       &adj_start,
+							       &adj_end);
+	if (!tile_mask)
+		goto range_notifier_event_end;
+
+	xe_device_wmb(xe);
+
+	for_each_tile(tile, xe, id) {
+		if (tile_mask & BIT(id)) {
+			int err;
+
+			xe_gt_tlb_invalidation_fence_init(tile->primary_gt,
+							  &fence[fence_id], true);
+
+			err = xe_gt_tlb_invalidation_range(tile->primary_gt,
+							   &fence[fence_id],
+							   adj_start,
+							   adj_end,
+							   vm->usm.asid);
+			if (WARN_ON_ONCE(err < 0))
+				goto wait;
+			++fence_id;
+
+			if (!tile->media_gt)
+				continue;
+
+			xe_gt_tlb_invalidation_fence_init(tile->media_gt,
+							  &fence[fence_id], true);
+
+			err = xe_gt_tlb_invalidation_range(tile->media_gt,
+							   &fence[fence_id],
+							   adj_start,
+							   adj_end,
+							   vm->usm.asid);
+			if (WARN_ON_ONCE(err < 0))
+				goto wait;
+			++fence_id;
+		}
+	}
+
+wait:
+	for (id = 0; id < fence_id; ++id)
+		xe_gt_tlb_invalidation_fence_wait(&fence[id]);
+
+range_notifier_event_end:
+	r = first;
+	drm_gpusvm_for_each_range(r, notifier, adj_start, adj_end)
+		xe_svm_range_notifier_event_end(vm, r, mmu_range);
 }
 
 static const struct drm_gpusvm_ops gpusvm_ops = {
+	.range_alloc = xe_svm_range_alloc,
+	.range_free = xe_svm_range_free,
 	.invalidate = xe_svm_invalidate,
 };
 
@@ -71,3 +251,44 @@ void xe_svm_fini(struct xe_vm *vm)
 
 	drm_gpusvm_fini(&vm->svm.gpusvm);
 }
+
+/**
+ * xe_svm_handle_pagefault() - SVM handle page fault
+ * @vm: The VM.
+ * @vma: The CPU address mirror VMA.
+ * @tile: The tile upon the fault occurred.
+ * @fault_addr: The GPU fault address.
+ * @atomic: The fault atomic access bit.
+ *
+ * Create GPU bindings for a SVM page fault.
+ *
+ * Return: 0 on success, negative error code on error.
+ */
+int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
+			    struct xe_tile *tile, u64 fault_addr,
+			    bool atomic)
+{
+	struct drm_gpusvm_ctx ctx = { .read_only = xe_vma_read_only(vma), };
+	struct drm_gpusvm_range *r;
+	int err;
+
+	lockdep_assert_held_write(&vm->lock);
+	xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(vma));
+
+retry:
+	/* TODO: Run garbage collector */
+
+	r = drm_gpusvm_range_find_or_insert(&vm->svm.gpusvm, fault_addr,
+					    xe_vma_start(vma), xe_vma_end(vma),
+					    &ctx);
+	if (IS_ERR(r))
+		return PTR_ERR(r);
+
+	err = drm_gpusvm_range_get_pages(&vm->svm.gpusvm, r, &ctx);
+	if (err == -EFAULT || err == -EPERM)	/* Corner where CPU mappings have changed */
+		goto retry;
+
+	/* TODO: Issue bind */
+
+	return err;
+}
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index 4569931db622..c674a68c66fe 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -7,10 +7,29 @@
 #define _XE_SVM_H_
 
 #include <drm/drm_pagemap.h>
+#include <drm/drm_gpusvm.h>
 
 #define XE_INTERCONNECT_VRAM DRM_INTERCONNECT_DRIVER
 
+struct xe_tile;
 struct xe_vm;
+struct xe_vma;
+
+/** struct xe_svm_range - SVM range */
+struct xe_svm_range {
+	/** @base: base drm_gpusvm_range */
+	struct drm_gpusvm_range base;
+	/**
+	 * @tile_present: Tile mask of binding is present for this range.
+	 * Protected by GPU SVM notifier lock.
+	 */
+	u8 tile_present;
+	/**
+	 * @tile_invalidated: Tile mask of binding is invalidated for this
+	 * range. Protected by GPU SVM notifier lock.
+	 */
+	u8 tile_invalidated;
+};
 
 int xe_svm_init(struct xe_vm *vm);
 
@@ -18,4 +37,17 @@ void xe_svm_fini(struct xe_vm *vm);
 
 void xe_svm_close(struct xe_vm *vm);
 
+int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
+			    struct xe_tile *tile, u64 fault_addr,
+			    bool atomic);
+
+#define xe_svm_assert_in_notifier(vm) \
+	lockdep_assert_held_write(&vm->svm.gpusvm.notifier_lock)
+
+#define xe_svm_notifier_lock(vm__)	\
+	drm_gpusvm_notifier_lock(&(vm__)->svm.gpusvm)
+
+#define xe_svm_notifier_unlock(vm__)	\
+	drm_gpusvm_notifier_unlock(&(vm__)->svm.gpusvm)
+
 #endif
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index ed49722d47ef..3e8fd77e43b5 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -1540,6 +1540,8 @@ static void xe_vm_close(struct xe_vm *vm)
 	bool migration = (vm->flags & XE_VM_FLAG_MIGRATION);
 
 	down_write(&vm->lock);
+	if (xe_vm_in_fault_mode(vm))
+		xe_svm_notifier_lock(vm);
 
 	vm->size = 0;
 
@@ -1561,6 +1563,8 @@ static void xe_vm_close(struct xe_vm *vm)
 			xe_gt_tlb_invalidation_vm(gt, vm);
 	}
 
+	if (xe_vm_in_fault_mode(vm))
+		xe_svm_notifier_unlock(vm);
 	up_write(&vm->lock);
 }
 
-- 
2.34.1

[PATCH v3 12/30] drm/gpuvm: Add DRM_GPUVA_OP_DRIVER

[Matthew Brost]

Add DRM_GPUVA_OP_DRIVER which allows driver to define their own gpuvm
ops. Useful for driver created ops which can be passed into the bind
software pipeline.

v3:
 - s/DRM_GPUVA_OP_USER/DRM_GPUVA_OP_DRIVER (Thomas)
 - Better commit message (Thomas)

Cc: Danilo Krummrich <dakr@redhat.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 include/drm/drm_gpuvm.h | 5 +++++
 1 file changed, 5 insertions(+)

diff --git a/include/drm/drm_gpuvm.h b/include/drm/drm_gpuvm.h
index 00d4e43b76b6..2a9629377633 100644
--- a/include/drm/drm_gpuvm.h
+++ b/include/drm/drm_gpuvm.h
@@ -812,6 +812,11 @@ enum drm_gpuva_op_type {
 	 * @DRM_GPUVA_OP_PREFETCH: the prefetch op type
 	 */
 	DRM_GPUVA_OP_PREFETCH,
+
+	/**
+	 * @DRM_GPUVA_OP_DRIVER: the driver defined op type
+	 */
+	DRM_GPUVA_OP_DRIVER,
 };
 
 /**
-- 
2.34.1

[PATCH v3 13/30] drm/xe: Add (re)bind to SVM page fault handler

[Matthew Brost]

Add (re)bind to SVM page fault handler. To facilitate add support
function to VM layer which (re)binds a SVM range. Also teach PT layer to
understand (re)binds of SVM ranges.

v2:
 - Don't assert BO lock held for range binds
 - Use xe_svm_notifier_lock/unlock helper in xe_svm_close
 - Use drm_pagemap dma cursor
 - Take notifier lock in bind code to check range state
v3:
 - Use new GPU SVM range structure (Thomas)
 - Kernel doc (Thomas)
 - s/DRM_GPUVA_OP_USER/DRM_GPUVA_OP_DRIVER (Thomas)

Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_pt.c       | 170 +++++++++++++++++++++++++++----
 drivers/gpu/drm/xe/xe_pt_types.h |   2 +
 drivers/gpu/drm/xe/xe_svm.c      |  44 +++++++-
 drivers/gpu/drm/xe/xe_svm.h      |  11 ++
 drivers/gpu/drm/xe/xe_vm.c       |  92 +++++++++++++++++
 drivers/gpu/drm/xe/xe_vm.h       |   5 +
 drivers/gpu/drm/xe/xe_vm_types.h |  19 ++++
 7 files changed, 323 insertions(+), 20 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index f7c4e1b374e8..6924f3ebfe3a 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -602,6 +602,7 @@ static const struct xe_pt_walk_ops xe_pt_stage_bind_ops = {
  * range.
  * @tile: The tile we're building for.
  * @vma: The vma indicating the address range.
+ * @range: The range indicating the address range.
  * @entries: Storage for the update entries used for connecting the tree to
  * the main tree at commit time.
  * @num_entries: On output contains the number of @entries used.
@@ -617,6 +618,7 @@ static const struct xe_pt_walk_ops xe_pt_stage_bind_ops = {
  */
 static int
 xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
+		 struct xe_svm_range *range,
 		 struct xe_vm_pgtable_update *entries, u32 *num_entries)
 {
 	struct xe_device *xe = tile_to_xe(tile);
@@ -633,14 +635,38 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
 		.vm = xe_vma_vm(vma),
 		.tile = tile,
 		.curs = &curs,
-		.va_curs_start = xe_vma_start(vma),
+		.va_curs_start = range ? range->base.itree.start :
+			xe_vma_start(vma),
 		.vma = vma,
 		.wupd.entries = entries,
-		.needs_64K = (xe_vma_vm(vma)->flags & XE_VM_FLAG_64K) && is_devmem,
 	};
 	struct xe_pt *pt = xe_vma_vm(vma)->pt_root[tile->id];
 	int ret;
 
+	if (range) {
+		/* Move this entire thing to xe_svm.c? */
+		xe_svm_notifier_lock(xe_vma_vm(vma));
+		if (!xe_svm_range_pages_valid(range)) {
+			xe_svm_notifier_unlock(xe_vma_vm(vma));
+			return -EAGAIN;
+		}
+		if (xe_svm_range_has_dma_mapping(range)) {
+			xe_res_first_dma(range->base.dma_addr, 0,
+					 range->base.itree.last + 1 - range->base.itree.start,
+					 &curs);
+			is_devmem = xe_res_is_vram(&curs);
+		} else {
+			xe_assert(xe, false);
+		}
+		/*
+		 * Note, when unlocking the resource cursor dma addresses may become
+		 * stale, but the bind will be aborted anyway att commit time.
+		 */
+		xe_svm_notifier_unlock(xe_vma_vm(vma));
+	}
+
+	xe_walk.needs_64K = (xe_vma_vm(vma)->flags & XE_VM_FLAG_64K) && is_devmem;
+
 	/**
 	 * Default atomic expectations for different allocation scenarios are as follows:
 	 *
@@ -662,7 +688,7 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
 			 * gets migrated to LMEM, bind such allocations with
 			 * device atomics enabled.
 			 */
-			else if (is_devmem && !xe_bo_has_single_placement(bo))
+			else if (is_devmem)
 				xe_walk.default_pte |= XE_USM_PPGTT_PTE_AE;
 		} else {
 			xe_walk.default_pte |= XE_USM_PPGTT_PTE_AE;
@@ -678,15 +704,16 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
 
 	if (is_devmem) {
 		xe_walk.default_pte |= XE_PPGTT_PTE_DM;
-		xe_walk.dma_offset = vram_region_gpu_offset(bo->ttm.resource);
+		xe_walk.dma_offset = bo ? vram_region_gpu_offset(bo->ttm.resource) : 0;
 	}
 
 	if (!xe_vma_has_no_bo(vma) && xe_bo_is_stolen(bo))
 		xe_walk.dma_offset = xe_ttm_stolen_gpu_offset(xe_bo_device(bo));
 
-	xe_bo_assert_held(bo);
+	if (!range)
+		xe_bo_assert_held(bo);
 
-	if (!xe_vma_is_null(vma)) {
+	if (!xe_vma_is_null(vma) && !range) {
 		if (xe_vma_is_userptr(vma))
 			xe_res_first_sg(to_userptr_vma(vma)->userptr.sg, 0,
 					xe_vma_size(vma), &curs);
@@ -696,12 +723,14 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
 		else
 			xe_res_first_sg(xe_bo_sg(bo), xe_vma_bo_offset(vma),
 					xe_vma_size(vma), &curs);
-	} else {
+	} else if (!range) {
 		curs.size = xe_vma_size(vma);
 	}
 
-	ret = xe_pt_walk_range(&pt->base, pt->level, xe_vma_start(vma),
-			       xe_vma_end(vma), &xe_walk.base);
+	ret = xe_pt_walk_range(&pt->base, pt->level,
+			       range ? range->base.itree.start : xe_vma_start(vma),
+			       range ? range->base.itree.last + 1 : xe_vma_end(vma),
+			       &xe_walk.base);
 
 	*num_entries = xe_walk.wupd.num_used_entries;
 	return ret;
@@ -934,7 +963,7 @@ static void xe_pt_commit_locks_assert(struct xe_vma *vma)
 
 	lockdep_assert_held(&vm->lock);
 
-	if (!xe_vma_is_userptr(vma) && !xe_vma_is_null(vma))
+	if (!xe_vma_has_no_bo(vma))
 		dma_resv_assert_held(xe_vma_bo(vma)->ttm.base.resv);
 
 	xe_vm_assert_held(vm);
@@ -1036,12 +1065,13 @@ static void xe_pt_free_bind(struct xe_vm_pgtable_update *entries,
 
 static int
 xe_pt_prepare_bind(struct xe_tile *tile, struct xe_vma *vma,
+		   struct xe_svm_range *range,
 		   struct xe_vm_pgtable_update *entries, u32 *num_entries)
 {
 	int err;
 
 	*num_entries = 0;
-	err = xe_pt_stage_bind(tile, vma, entries, num_entries);
+	err = xe_pt_stage_bind(tile, vma, range, entries, num_entries);
 	if (!err)
 		xe_tile_assert(tile, *num_entries);
 
@@ -1147,6 +1177,8 @@ static int op_add_deps(struct xe_vm *vm, struct xe_vma_op *op,
 	case DRM_GPUVA_OP_PREFETCH:
 		err = vma_add_deps(gpuva_to_vma(op->base.prefetch.va), job);
 		break;
+	case DRM_GPUVA_OP_DRIVER:
+		break;
 	default:
 		drm_warn(&vm->xe->drm, "NOT POSSIBLE");
 	}
@@ -1371,6 +1403,34 @@ static int xe_pt_userptr_pre_commit(struct xe_migrate_pt_update *pt_update)
 	return err;
 }
 
+static int xe_pt_svm_pre_commit(struct xe_migrate_pt_update *pt_update)
+{
+	struct xe_vm *vm = pt_update->vops->vm;
+	struct xe_vma_ops *vops = pt_update->vops;
+	struct xe_vma_op *op;
+	int err;
+
+	err = xe_pt_pre_commit(pt_update);
+	if (err)
+		return err;
+
+	xe_svm_notifier_lock(vm);
+
+	list_for_each_entry(op, &vops->list, link) {
+		struct xe_svm_range *range = op->map_range.range;
+
+		xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(op->map_range.vma));
+		xe_assert(vm->xe, op->subop == XE_VMA_SUBOP_MAP_RANGE);
+
+		if (!xe_svm_range_pages_valid(range)) {
+			xe_svm_notifier_unlock(vm);
+			return -EAGAIN;
+		}
+	}
+
+	return 0;
+}
+
 struct invalidation_fence {
 	struct xe_gt_tlb_invalidation_fence base;
 	struct xe_gt *gt;
@@ -1663,12 +1723,12 @@ xe_pt_commit_prepare_unbind(struct xe_vma *vma,
 
 static void
 xe_pt_update_ops_rfence_interval(struct xe_vm_pgtable_update_ops *pt_update_ops,
-				 struct xe_vma *vma)
+				 u64 start, u64 end)
 {
+	u64 last;
 	u32 current_op = pt_update_ops->current_op;
 	struct xe_vm_pgtable_update_op *pt_op = &pt_update_ops->ops[current_op];
 	int i, level = 0;
-	u64 start, last;
 
 	for (i = 0; i < pt_op->num_entries; i++) {
 		const struct xe_vm_pgtable_update *entry = &pt_op->entries[i];
@@ -1678,8 +1738,8 @@ xe_pt_update_ops_rfence_interval(struct xe_vm_pgtable_update_ops *pt_update_ops,
 	}
 
 	/* Greedy (non-optimal) calculation but simple */
-	start = ALIGN_DOWN(xe_vma_start(vma), 0x1ull << xe_pt_shift(level));
-	last = ALIGN(xe_vma_end(vma), 0x1ull << xe_pt_shift(level)) - 1;
+	start = ALIGN_DOWN(start, 0x1ull << xe_pt_shift(level));
+	last = ALIGN(end, 0x1ull << xe_pt_shift(level)) - 1;
 
 	if (start < pt_update_ops->start)
 		pt_update_ops->start = start;
@@ -1721,7 +1781,7 @@ static int bind_op_prepare(struct xe_vm *vm, struct xe_tile *tile,
 	if (err)
 		return err;
 
-	err = xe_pt_prepare_bind(tile, vma, pt_op->entries,
+	err = xe_pt_prepare_bind(tile, vma, NULL, pt_op->entries,
 				 &pt_op->num_entries);
 	if (!err) {
 		xe_tile_assert(tile, pt_op->num_entries <=
@@ -1729,7 +1789,9 @@ static int bind_op_prepare(struct xe_vm *vm, struct xe_tile *tile,
 		xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
 					pt_op->num_entries, true);
 
-		xe_pt_update_ops_rfence_interval(pt_update_ops, vma);
+		xe_pt_update_ops_rfence_interval(pt_update_ops,
+						 xe_vma_start(vma),
+						 xe_vma_end(vma));
 		++pt_update_ops->current_op;
 		pt_update_ops->needs_userptr_lock |= xe_vma_is_userptr(vma);
 
@@ -1763,6 +1825,48 @@ static int bind_op_prepare(struct xe_vm *vm, struct xe_tile *tile,
 	return err;
 }
 
+static int bind_range_prepare(struct xe_vm *vm, struct xe_tile *tile,
+			      struct xe_vm_pgtable_update_ops *pt_update_ops,
+			      struct xe_vma *vma, struct xe_svm_range *range)
+{
+	u32 current_op = pt_update_ops->current_op;
+	struct xe_vm_pgtable_update_op *pt_op = &pt_update_ops->ops[current_op];
+	int err;
+
+	xe_tile_assert(tile, xe_vma_is_cpu_addr_mirror(vma));
+
+	vm_dbg(&xe_vma_vm(vma)->xe->drm,
+	       "Preparing bind, with range [%lx...%lx)\n",
+	       range->base.itree.start, range->base.itree.last);
+
+	pt_op->vma = NULL;
+	pt_op->bind = true;
+	pt_op->rebind = BIT(tile->id) & range->tile_present;
+
+	err = xe_pt_prepare_bind(tile, vma, range, pt_op->entries,
+				 &pt_op->num_entries);
+	if (!err) {
+		xe_tile_assert(tile, pt_op->num_entries <=
+			       ARRAY_SIZE(pt_op->entries));
+		xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
+					pt_op->num_entries, true);
+
+		xe_pt_update_ops_rfence_interval(pt_update_ops,
+						 range->base.itree.start,
+						 range->base.itree.last + 1);
+		++pt_update_ops->current_op;
+		pt_update_ops->needs_svm_lock = true;
+
+		pt_op->vma = vma;
+		xe_pt_commit_prepare_bind(vma, pt_op->entries,
+					  pt_op->num_entries, pt_op->rebind);
+	} else {
+		xe_pt_cancel_bind(vma, pt_op->entries, pt_op->num_entries);
+	}
+
+	return err;
+}
+
 static int unbind_op_prepare(struct xe_tile *tile,
 			     struct xe_vm_pgtable_update_ops *pt_update_ops,
 			     struct xe_vma *vma)
@@ -1800,7 +1904,8 @@ static int unbind_op_prepare(struct xe_tile *tile,
 
 	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
 				pt_op->num_entries, false);
-	xe_pt_update_ops_rfence_interval(pt_update_ops, vma);
+	xe_pt_update_ops_rfence_interval(pt_update_ops, xe_vma_start(vma),
+					 xe_vma_end(vma));
 	++pt_update_ops->current_op;
 	pt_update_ops->needs_userptr_lock |= xe_vma_is_userptr(vma);
 	pt_update_ops->needs_invalidation = true;
@@ -1870,6 +1975,15 @@ static int op_prepare(struct xe_vm *vm,
 		pt_update_ops->wait_vm_kernel = true;
 		break;
 	}
+	case DRM_GPUVA_OP_DRIVER:
+		if (op->subop == XE_VMA_SUBOP_MAP_RANGE) {
+			xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(op->map_range.vma));
+
+			err = bind_range_prepare(vm, tile, pt_update_ops,
+						 op->map_range.vma,
+						 op->map_range.range);
+		}
+		break;
 	default:
 		drm_warn(&vm->xe->drm, "NOT POSSIBLE");
 	}
@@ -2052,6 +2166,14 @@ static void op_commit(struct xe_vm *vm,
 				       fence2);
 		break;
 	}
+	case DRM_GPUVA_OP_DRIVER:
+	{
+		if (op->subop == XE_VMA_SUBOP_MAP_RANGE) {
+			op->map_range.range->tile_present |= BIT(tile->id);
+			op->map_range.range->tile_invalidated &= ~BIT(tile->id);
+		}
+		break;
+	}
 	default:
 		drm_warn(&vm->xe->drm, "NOT POSSIBLE");
 	}
@@ -2069,6 +2191,12 @@ static const struct xe_migrate_pt_update_ops userptr_migrate_ops = {
 	.pre_commit = xe_pt_userptr_pre_commit,
 };
 
+static const struct xe_migrate_pt_update_ops svm_migrate_ops = {
+	.populate = xe_vm_populate_pgtable,
+	.clear = xe_migrate_clear_pgtable_callback,
+	.pre_commit = xe_pt_svm_pre_commit,
+};
+
 /**
  * xe_pt_update_ops_run() - Run PT update operations
  * @tile: Tile of PT update operations
@@ -2094,7 +2222,9 @@ xe_pt_update_ops_run(struct xe_tile *tile, struct xe_vma_ops *vops)
 	struct xe_vma_op *op;
 	int err = 0, i;
 	struct xe_migrate_pt_update update = {
-		.ops = pt_update_ops->needs_userptr_lock ?
+		.ops = pt_update_ops->needs_svm_lock ?
+			&svm_migrate_ops :
+			pt_update_ops->needs_userptr_lock ?
 			&userptr_migrate_ops :
 			&migrate_ops,
 		.vops = vops,
@@ -2215,6 +2345,8 @@ xe_pt_update_ops_run(struct xe_tile *tile, struct xe_vma_ops *vops)
 				  &ifence->base.base, &mfence->base.base);
 	}
 
+	if (pt_update_ops->needs_svm_lock)
+		xe_svm_notifier_unlock(vm);
 	if (pt_update_ops->needs_userptr_lock)
 		up_read(&vm->userptr.notifier_lock);
 
diff --git a/drivers/gpu/drm/xe/xe_pt_types.h b/drivers/gpu/drm/xe/xe_pt_types.h
index 384cc04de719..69eab6f37cfe 100644
--- a/drivers/gpu/drm/xe/xe_pt_types.h
+++ b/drivers/gpu/drm/xe/xe_pt_types.h
@@ -104,6 +104,8 @@ struct xe_vm_pgtable_update_ops {
 	u32 num_ops;
 	/** @current_op: current operations */
 	u32 current_op;
+	/** @needs_svm_lock: Needs SVM lock */
+	bool needs_svm_lock;
 	/** @needs_userptr_lock: Needs userptr lock */
 	bool needs_userptr_lock;
 	/** @needs_invalidation: Needs invalidation */
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index bd7b9c6ea229..ace8c32f3428 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -252,6 +252,12 @@ void xe_svm_fini(struct xe_vm *vm)
 	drm_gpusvm_fini(&vm->svm.gpusvm);
 }
 
+static bool xe_svm_range_is_valid(struct xe_svm_range *range,
+				  struct xe_tile *tile)
+{
+	return (range->tile_present & ~range->tile_invalidated) & BIT(tile->id);
+}
+
 /**
  * xe_svm_handle_pagefault() - SVM handle page fault
  * @vm: The VM.
@@ -269,7 +275,11 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 			    bool atomic)
 {
 	struct drm_gpusvm_ctx ctx = { .read_only = xe_vma_read_only(vma), };
+	struct xe_svm_range *range;
 	struct drm_gpusvm_range *r;
+	struct drm_exec exec;
+	struct dma_fence *fence;
+	ktime_t end = 0;
 	int err;
 
 	lockdep_assert_held_write(&vm->lock);
@@ -284,11 +294,43 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	if (IS_ERR(r))
 		return PTR_ERR(r);
 
+	range = to_xe_range(r);
+	if (xe_svm_range_is_valid(range, tile))
+		return 0;
+
 	err = drm_gpusvm_range_get_pages(&vm->svm.gpusvm, r, &ctx);
 	if (err == -EFAULT || err == -EPERM)	/* Corner where CPU mappings have changed */
 		goto retry;
+	if (err)
+		goto err_out;
+
+retry_bind:
+	drm_exec_init(&exec, 0, 0);
+	drm_exec_until_all_locked(&exec) {
+		err = drm_exec_lock_obj(&exec, vm->gpuvm.r_obj);
+		drm_exec_retry_on_contention(&exec);
+		if (err) {
+			drm_exec_fini(&exec);
+			goto err_out;
+		}
+
+		fence = xe_vm_range_rebind(vm, vma, range, BIT(tile->id));
+		if (IS_ERR(fence)) {
+			drm_exec_fini(&exec);
+			err = PTR_ERR(fence);
+			if (err == -EAGAIN)
+				goto retry;
+			if (xe_vm_validate_should_retry(&exec, err, &end))
+				goto retry_bind;
+			goto err_out;
+		}
+	}
+	drm_exec_fini(&exec);
+
+	dma_fence_wait(fence, false);
+	dma_fence_put(fence);
 
-	/* TODO: Issue bind */
+err_out:
 
 	return err;
 }
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index c674a68c66fe..6cd57382509d 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -41,6 +41,17 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 			    struct xe_tile *tile, u64 fault_addr,
 			    bool atomic);
 
+static inline bool xe_svm_range_pages_valid(struct xe_svm_range *range)
+{
+	return drm_gpusvm_range_pages_valid(range->base.gpusvm, &range->base);
+}
+
+static inline bool xe_svm_range_has_dma_mapping(struct xe_svm_range *range)
+{
+	lockdep_assert_held(&range->base.gpusvm->notifier_lock);
+	return range->base.flags.has_dma_mapping;
+}
+
 #define xe_svm_assert_in_notifier(vm) \
 	lockdep_assert_held_write(&vm->svm.gpusvm.notifier_lock)
 
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 3e8fd77e43b5..67859e9c8595 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -894,6 +894,96 @@ struct dma_fence *xe_vma_rebind(struct xe_vm *vm, struct xe_vma *vma, u8 tile_ma
 	return fence;
 }
 
+static void xe_vm_populate_range_rebind(struct xe_vma_op *op,
+					struct xe_vma *vma,
+					struct xe_svm_range *range,
+					u8 tile_mask)
+{
+	INIT_LIST_HEAD(&op->link);
+	op->tile_mask = tile_mask;
+	op->base.op = DRM_GPUVA_OP_DRIVER;
+	op->subop = XE_VMA_SUBOP_MAP_RANGE;
+	op->map_range.vma = vma;
+	op->map_range.range = range;
+}
+
+static int
+xe_vm_ops_add_range_rebind(struct xe_vma_ops *vops,
+			   struct xe_vma *vma,
+			   struct xe_svm_range *range,
+			   u8 tile_mask)
+{
+	struct xe_vma_op *op;
+
+	op = kzalloc(sizeof(*op), GFP_KERNEL);
+	if (!op)
+		return -ENOMEM;
+
+	xe_vm_populate_range_rebind(op, vma, range, tile_mask);
+	list_add_tail(&op->link, &vops->list);
+	xe_vma_ops_incr_pt_update_ops(vops, tile_mask);
+
+	return 0;
+}
+
+/**
+ * xe_vm_range_rebind() - VM range (re)bind
+ * @vm: The VM which the range belongs to.
+ * @vma: The VMA which the range belongs to.
+ * @range: SVM range to rebind.
+ * @tile_mask: Tile mask to bind the range to.
+ *
+ * (re)bind SVM range setting up GPU page tables for the range.
+ *
+ * Return: dma fence for rebind to signal completion on succees, ERR_PTR on
+ * failure
+ */
+struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
+				     struct xe_vma *vma,
+				     struct xe_svm_range *range,
+				     u8 tile_mask)
+{
+	struct dma_fence *fence = NULL;
+	struct xe_vma_ops vops;
+	struct xe_vma_op *op, *next_op;
+	struct xe_tile *tile;
+	u8 id;
+	int err;
+
+	lockdep_assert_held(&vm->lock);
+	xe_vm_assert_held(vm);
+	xe_assert(vm->xe, xe_vm_in_fault_mode(vm));
+	xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(vma));
+
+	xe_vma_ops_init(&vops, vm, NULL, NULL, 0);
+	for_each_tile(tile, vm->xe, id) {
+		vops.pt_update_ops[id].wait_vm_bookkeep = true;
+		vops.pt_update_ops[tile->id].q =
+			xe_tile_migrate_exec_queue(tile);
+	}
+
+	err = xe_vm_ops_add_range_rebind(&vops, vma, range, tile_mask);
+	if (err)
+		return ERR_PTR(err);
+
+	err = xe_vma_ops_alloc(&vops, false);
+	if (err) {
+		fence = ERR_PTR(err);
+		goto free_ops;
+	}
+
+	fence = ops_execute(vm, &vops);
+
+free_ops:
+	list_for_each_entry_safe(op, next_op, &vops.list, link) {
+		list_del(&op->link);
+		kfree(op);
+	}
+	xe_vma_ops_fini(&vops);
+
+	return fence;
+}
+
 static void xe_vma_free(struct xe_vma *vma)
 {
 	if (xe_vma_is_userptr(vma))
@@ -2544,6 +2634,8 @@ static void op_trace(struct xe_vma_op *op)
 	case DRM_GPUVA_OP_PREFETCH:
 		trace_xe_vma_bind(gpuva_to_vma(op->base.prefetch.va));
 		break;
+	case DRM_GPUVA_OP_DRIVER:
+		break;
 	default:
 		XE_WARN_ON("NOT POSSIBLE");
 	}
diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h
index 0e54a0e8768d..a82fe743bbe0 100644
--- a/drivers/gpu/drm/xe/xe_vm.h
+++ b/drivers/gpu/drm/xe/xe_vm.h
@@ -21,6 +21,7 @@ struct ttm_buffer_object;
 struct xe_exec_queue;
 struct xe_file;
 struct xe_sync_entry;
+struct xe_svm_range;
 struct drm_exec;
 
 struct xe_vm *xe_vm_create(struct xe_device *xe, u32 flags);
@@ -216,6 +217,10 @@ int xe_vm_userptr_check_repin(struct xe_vm *vm);
 int xe_vm_rebind(struct xe_vm *vm, bool rebind_worker);
 struct dma_fence *xe_vma_rebind(struct xe_vm *vm, struct xe_vma *vma,
 				u8 tile_mask);
+struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
+				     struct xe_vma *vma,
+				     struct xe_svm_range *range,
+				     u8 tile_mask);
 
 int xe_vm_invalidate_vma(struct xe_vma *vma);
 
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index aa075d5e7a3f..983f724c911b 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -19,6 +19,7 @@
 #include "xe_range_fence.h"
 
 struct xe_bo;
+struct xe_svm_range;
 struct xe_sync_entry;
 struct xe_user_fence;
 struct xe_vm;
@@ -334,6 +335,14 @@ struct xe_vma_op_prefetch {
 	u32 region;
 };
 
+/** struct xe_vma_op_map_range - VMA map range operation */
+struct xe_vma_op_map_range {
+	/** @vma: VMA to map (system allocator VMA) */
+	struct xe_vma *vma;
+	/** @range: SVM range to map */
+	struct xe_svm_range *range;
+};
+
 /** enum xe_vma_op_flags - flags for VMA operation */
 enum xe_vma_op_flags {
 	/** @XE_VMA_OP_COMMITTED: VMA operation committed */
@@ -344,6 +353,12 @@ enum xe_vma_op_flags {
 	XE_VMA_OP_NEXT_COMMITTED	= BIT(2),
 };
 
+/** enum xe_vma_subop - VMA sub-operation */
+enum xe_vma_subop {
+	/** @XE_VMA_SUBOP_MAP_RANGE: Map range */
+	XE_VMA_SUBOP_MAP_RANGE,
+};
+
 /** struct xe_vma_op - VMA operation */
 struct xe_vma_op {
 	/** @base: GPUVA base operation */
@@ -352,6 +367,8 @@ struct xe_vma_op {
 	struct list_head link;
 	/** @flags: operation flags */
 	enum xe_vma_op_flags flags;
+	/** @subop: user defined sub-operation */
+	enum xe_vma_subop subop;
 	/** @tile_mask: Tile mask for operation */
 	u8 tile_mask;
 
@@ -362,6 +379,8 @@ struct xe_vma_op {
 		struct xe_vma_op_remap remap;
 		/** @prefetch: VMA prefetch operation specific data */
 		struct xe_vma_op_prefetch prefetch;
+		/** @map: VMA map range operation specific data */
+		struct xe_vma_op_map_range map_range;
 	};
 };
 
-- 
2.34.1

[PATCH v3 14/30] drm/xe: Add SVM garbage collector

[Matthew Brost]

Add basic SVM garbage collector which destroy a SVM range upon a MMU
UNMAP event. The garbage collector runs on worker or in GPU fault
handler and is required as locks in the path of reclaim are required and
cannot be taken the notifier.

v2:
 - Flush garbage collector in xe_svm_close
v3:
 - Better commit message (Thomas)
 - Kernel doc (Thomas)
 - Use list_first_entry_or_null for garbage collector loop (Thomas)
 - Don't add to garbage collector if VM is closed (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c      | 90 +++++++++++++++++++++++++++++++-
 drivers/gpu/drm/xe/xe_svm.h      |  5 ++
 drivers/gpu/drm/xe/xe_vm.c       |  4 ++
 drivers/gpu/drm/xe/xe_vm_types.h | 18 +++++++
 4 files changed, 115 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index ace8c32f3428..32b1581b1ad0 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -28,6 +28,7 @@ xe_svm_range_alloc(struct drm_gpusvm *gpusvm)
 	if (!range)
 		return ERR_PTR(-ENOMEM);
 
+	INIT_LIST_HEAD(&range->garbage_collector_link);
 	xe_vm_get(gpusvm_to_vm(gpusvm));
 
 	return &range->base;
@@ -44,6 +45,24 @@ static struct xe_svm_range *to_xe_range(struct drm_gpusvm_range *r)
 	return container_of(r, struct xe_svm_range, base);
 }
 
+static void
+xe_svm_garbage_collector_add_range(struct xe_vm *vm, struct xe_svm_range *range,
+				   const struct mmu_notifier_range *mmu_range)
+{
+	struct xe_device *xe = vm->xe;
+
+	drm_gpusvm_range_set_unmapped(&range->base, mmu_range);
+
+	spin_lock(&vm->svm.garbage_collector.lock);
+	if (list_empty(&range->garbage_collector_link))
+		list_add_tail(&range->garbage_collector_link,
+			      &vm->svm.garbage_collector.range_list);
+	spin_unlock(&vm->svm.garbage_collector.lock);
+
+	queue_work(xe_device_get_root_tile(xe)->primary_gt->usm.pf_wq,
+		   &vm->svm.garbage_collector.work);
+}
+
 static u8
 xe_svm_range_notifier_event_begin(struct xe_vm *vm, struct drm_gpusvm_range *r,
 				  const struct mmu_notifier_range *mmu_range,
@@ -90,7 +109,9 @@ xe_svm_range_notifier_event_end(struct xe_vm *vm, struct drm_gpusvm_range *r,
 	xe_svm_assert_in_notifier(vm);
 
 	drm_gpusvm_range_unmap_pages(&vm->svm.gpusvm, r, &ctx);
-	/* TODO: Add range to garbage collector if VM is not closed */
+	if (!xe_vm_is_closed(vm) && mmu_range->event == MMU_NOTIFY_UNMAP)
+		xe_svm_garbage_collector_add_range(vm, to_xe_range(r),
+						   mmu_range);
 }
 
 static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
@@ -192,6 +213,62 @@ static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
 		xe_svm_range_notifier_event_end(vm, r, mmu_range);
 }
 
+static int __xe_svm_garbage_collector(struct xe_vm *vm,
+				      struct xe_svm_range *range)
+{
+	/* TODO: Do unbind */
+
+	drm_gpusvm_range_remove(&vm->svm.gpusvm, &range->base);
+
+	return 0;
+}
+
+static int xe_svm_garbage_collector(struct xe_vm *vm)
+{
+	struct xe_svm_range *range;
+	int err;
+
+	lockdep_assert_held_write(&vm->lock);
+
+	if (xe_vm_is_closed_or_banned(vm))
+		return -ENOENT;
+
+	spin_lock(&vm->svm.garbage_collector.lock);
+	for (;;) {
+		range = list_first_entry_or_null(&vm->svm.garbage_collector.range_list,
+						 typeof(*range),
+						 garbage_collector_link);
+		if (!range)
+			break;
+
+		list_del(&range->garbage_collector_link);
+		spin_unlock(&vm->svm.garbage_collector.lock);
+
+		err = __xe_svm_garbage_collector(vm, range);
+		if (err) {
+			drm_warn(&vm->xe->drm,
+				 "Garbage collection failed: %d\n", err);
+			xe_vm_kill(vm, true);
+			return err;
+		}
+
+		spin_lock(&vm->svm.garbage_collector.lock);
+	}
+	spin_unlock(&vm->svm.garbage_collector.lock);
+
+	return 0;
+}
+
+static void xe_svm_garbage_collector_work_func(struct work_struct *w)
+{
+	struct xe_vm *vm = container_of(w, struct xe_vm,
+					svm.garbage_collector.work);
+
+	down_write(&vm->lock);
+	xe_svm_garbage_collector(vm);
+	up_write(&vm->lock);
+}
+
 static const struct drm_gpusvm_ops gpusvm_ops = {
 	.range_alloc = xe_svm_range_alloc,
 	.range_free = xe_svm_range_free,
@@ -216,6 +293,11 @@ int xe_svm_init(struct xe_vm *vm)
 {
 	int err;
 
+	spin_lock_init(&vm->svm.garbage_collector.lock);
+	INIT_LIST_HEAD(&vm->svm.garbage_collector.range_list);
+	INIT_WORK(&vm->svm.garbage_collector.work,
+		  xe_svm_garbage_collector_work_func);
+
 	err = drm_gpusvm_init(&vm->svm.gpusvm, "Xe SVM", &vm->xe->drm,
 			      current->mm, NULL, 0, vm->size,
 			      SZ_512M, &gpusvm_ops, fault_chunk_sizes,
@@ -237,6 +319,7 @@ int xe_svm_init(struct xe_vm *vm)
 void xe_svm_close(struct xe_vm *vm)
 {
 	xe_assert(vm->xe, xe_vm_is_closed(vm));
+	flush_work(&vm->svm.garbage_collector.work);
 }
 
 /**
@@ -286,7 +369,10 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(vma));
 
 retry:
-	/* TODO: Run garbage collector */
+	/* Always process UNMAPs first so view SVM ranges is current */
+	err = xe_svm_garbage_collector(vm);
+	if (err)
+		return err;
 
 	r = drm_gpusvm_range_find_or_insert(&vm->svm.gpusvm, fault_addr,
 					    xe_vma_start(vma), xe_vma_end(vma),
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index 6cd57382509d..5531868cdd6f 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -19,6 +19,11 @@ struct xe_vma;
 struct xe_svm_range {
 	/** @base: base drm_gpusvm_range */
 	struct drm_gpusvm_range base;
+	/**
+	 * @garbage_collector_link: Link into VM's garbage collect SVM range
+	 * list. Protected by VM's garbage collect lock.
+	 */
+	struct list_head garbage_collector_link;
 	/**
 	 * @tile_present: Tile mask of binding is present for this range.
 	 * Protected by GPU SVM notifier lock.
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 67859e9c8595..12e5e562c5e1 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -3123,6 +3123,10 @@ int xe_vm_bind_ioctl(struct drm_device *dev, void *data, struct drm_file *file)
 		}
 	}
 
+	/* Ensure all UNMAPs visable */
+	if (xe_vm_in_fault_mode(vm))
+		flush_work(&vm->svm.garbage_collector.work);
+
 	err = down_write_killable(&vm->lock);
 	if (err)
 		goto put_exec_queue;
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index 983f724c911b..576316729249 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -146,6 +146,24 @@ struct xe_vm {
 	struct {
 		/** @svm.gpusvm: base GPUSVM used to track fault allocations */
 		struct drm_gpusvm gpusvm;
+		/**
+		 * @svm.garbage_collector: Garbage collector which is used unmap
+		 * SVM range's GPU bindings and destroy the ranges.
+		 */
+		struct {
+			/** @svm.garbage_collector.lock: Protect's range list */
+			spinlock_t lock;
+			/**
+			 * @svm.garbage_collector.range_list: List of SVM ranges
+			 * in the garbage collector.
+			 */
+			struct list_head range_list;
+			/**
+			 * @svm.garbage_collector.work: Worker which the
+			 * garbage collector runs on.
+			 */
+			struct work_struct work;
+		} garbage_collector;
 	} svm;
 
 	struct xe_device *xe;
-- 
2.34.1

[PATCH v3 15/30] drm/xe: Add unbind to SVM garbage collector

[Matthew Brost]

Add unbind to SVM garbage collector. To facilitate add unbind support
function to VM layer which unbinds a SVM range. Also teach PY layer to
understand unbinds of SVM ranges.

v3:
 - s/INVALID_VMA/XE_INVALID_VMA (Thomas)
 - Kernel doc (Thomas)
 - New GPU SVM range structure (Thomas)
 - s/DRM_GPUVA_OP_USER/DRM_GPUVA_OP_DRIVER (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_pt.c       | 84 ++++++++++++++++++++++++++------
 drivers/gpu/drm/xe/xe_svm.c      |  9 +++-
 drivers/gpu/drm/xe/xe_vm.c       | 83 +++++++++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_vm.h       |  2 +
 drivers/gpu/drm/xe/xe_vm_types.h | 12 ++++-
 5 files changed, 172 insertions(+), 18 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index 6924f3ebfe3a..9c90dfc81fbe 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -957,10 +957,16 @@ static void xe_pt_cancel_bind(struct xe_vma *vma,
 	}
 }
 
+#define XE_INVALID_VMA	((struct xe_vma *)(0xdeaddeadull))
+
 static void xe_pt_commit_locks_assert(struct xe_vma *vma)
 {
-	struct xe_vm *vm = xe_vma_vm(vma);
+	struct xe_vm *vm;
 
+	if (vma == XE_INVALID_VMA)
+		return;
+
+	vm = xe_vma_vm(vma);
 	lockdep_assert_held(&vm->lock);
 
 	if (!xe_vma_has_no_bo(vma))
@@ -986,7 +992,8 @@ static void xe_pt_commit(struct xe_vma *vma,
 		for (j = 0; j < entries[i].qwords; j++) {
 			struct xe_pt *oldpte = entries[i].pt_entries[j].pt;
 
-			xe_pt_destroy(oldpte, xe_vma_vm(vma)->flags, deferred);
+			xe_pt_destroy(oldpte, (vma == XE_INVALID_VMA) ? 0 :
+				      xe_vma_vm(vma)->flags, deferred);
 		}
 	}
 }
@@ -1419,6 +1426,9 @@ static int xe_pt_svm_pre_commit(struct xe_migrate_pt_update *pt_update)
 	list_for_each_entry(op, &vops->list, link) {
 		struct xe_svm_range *range = op->map_range.range;
 
+		if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE)
+			continue;
+
 		xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(op->map_range.vma));
 		xe_assert(vm->xe, op->subop == XE_VMA_SUBOP_MAP_RANGE);
 
@@ -1616,7 +1626,9 @@ static const struct xe_pt_walk_ops xe_pt_stage_unbind_ops = {
  * xe_pt_stage_unbind() - Build page-table update structures for an unbind
  * operation
  * @tile: The tile we're unbinding for.
+ * @vm: The vm
  * @vma: The vma we're unbinding.
+ * @range: The range we're unbinding.
  * @entries: Caller-provided storage for the update structures.
  *
  * Builds page-table update structures for an unbind operation. The function
@@ -1626,9 +1638,14 @@ static const struct xe_pt_walk_ops xe_pt_stage_unbind_ops = {
  *
  * Return: The number of entries used.
  */
-static unsigned int xe_pt_stage_unbind(struct xe_tile *tile, struct xe_vma *vma,
+static unsigned int xe_pt_stage_unbind(struct xe_tile *tile,
+				       struct xe_vm *vm,
+				       struct xe_vma *vma,
+				       struct xe_svm_range *range,
 				       struct xe_vm_pgtable_update *entries)
 {
+	u64 start = range ? range->base.itree.start : xe_vma_start(vma);
+	u64 end = range ? range->base.itree.last + 1 : xe_vma_end(vma);
 	struct xe_pt_stage_unbind_walk xe_walk = {
 		.base = {
 			.ops = &xe_pt_stage_unbind_ops,
@@ -1636,14 +1653,14 @@ static unsigned int xe_pt_stage_unbind(struct xe_tile *tile, struct xe_vma *vma,
 			.max_level = XE_PT_HIGHEST_LEVEL,
 		},
 		.tile = tile,
-		.modified_start = xe_vma_start(vma),
-		.modified_end = xe_vma_end(vma),
+		.modified_start = start,
+		.modified_end = end,
 		.wupd.entries = entries,
 	};
-	struct xe_pt *pt = xe_vma_vm(vma)->pt_root[tile->id];
+	struct xe_pt *pt = vm->pt_root[tile->id];
 
-	(void)xe_pt_walk_shared(&pt->base, pt->level, xe_vma_start(vma),
-				xe_vma_end(vma), &xe_walk.base);
+	(void)xe_pt_walk_shared(&pt->base, pt->level, start, end,
+				&xe_walk.base);
 
 	return xe_walk.wupd.num_used_entries;
 }
@@ -1885,13 +1902,6 @@ static int unbind_op_prepare(struct xe_tile *tile,
 	       "Preparing unbind, with range [%llx...%llx)\n",
 	       xe_vma_start(vma), xe_vma_end(vma) - 1);
 
-	/*
-	 * Wait for invalidation to complete. Can corrupt internal page table
-	 * state if an invalidation is running while preparing an unbind.
-	 */
-	if (xe_vma_is_userptr(vma) && xe_vm_in_fault_mode(xe_vma_vm(vma)))
-		mmu_interval_read_begin(&to_userptr_vma(vma)->userptr.notifier);
-
 	pt_op->vma = vma;
 	pt_op->bind = false;
 	pt_op->rebind = false;
@@ -1900,7 +1910,8 @@ static int unbind_op_prepare(struct xe_tile *tile,
 	if (err)
 		return err;
 
-	pt_op->num_entries = xe_pt_stage_unbind(tile, vma, pt_op->entries);
+	pt_op->num_entries = xe_pt_stage_unbind(tile, xe_vma_vm(vma),
+						vma, NULL, pt_op->entries);
 
 	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
 				pt_op->num_entries, false);
@@ -1915,6 +1926,42 @@ static int unbind_op_prepare(struct xe_tile *tile,
 	return 0;
 }
 
+static int unbind_range_prepare(struct xe_vm *vm,
+				struct xe_tile *tile,
+				struct xe_vm_pgtable_update_ops *pt_update_ops,
+				struct xe_svm_range *range)
+{
+	u32 current_op = pt_update_ops->current_op;
+	struct xe_vm_pgtable_update_op *pt_op = &pt_update_ops->ops[current_op];
+
+	if (!(range->tile_present & BIT(tile->id)))
+		return 0;
+
+	vm_dbg(&vm->xe->drm,
+	       "Preparing unbind, with range [%lx...%lx)\n",
+	       range->base.itree.start, range->base.itree.last);
+
+	pt_op->vma = XE_INVALID_VMA;
+	pt_op->bind = false;
+	pt_op->rebind = false;
+
+	pt_op->num_entries = xe_pt_stage_unbind(tile, vm, NULL, range,
+						pt_op->entries);
+
+	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
+				pt_op->num_entries, false);
+	xe_pt_update_ops_rfence_interval(pt_update_ops, range->base.itree.start,
+					 range->base.itree.last + 1);
+	++pt_update_ops->current_op;
+	pt_update_ops->needs_svm_lock = true;
+	pt_update_ops->needs_invalidation = true;
+
+	xe_pt_commit_prepare_unbind(XE_INVALID_VMA, pt_op->entries,
+				    pt_op->num_entries);
+
+	return 0;
+}
+
 static int op_prepare(struct xe_vm *vm,
 		      struct xe_tile *tile,
 		      struct xe_vm_pgtable_update_ops *pt_update_ops,
@@ -1982,6 +2029,9 @@ static int op_prepare(struct xe_vm *vm,
 			err = bind_range_prepare(vm, tile, pt_update_ops,
 						 op->map_range.vma,
 						 op->map_range.range);
+		} else if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE) {
+			err = unbind_range_prepare(vm, tile, pt_update_ops,
+						   op->unmap_range.range);
 		}
 		break;
 	default:
@@ -2171,6 +2221,8 @@ static void op_commit(struct xe_vm *vm,
 		if (op->subop == XE_VMA_SUBOP_MAP_RANGE) {
 			op->map_range.range->tile_present |= BIT(tile->id);
 			op->map_range.range->tile_invalidated &= ~BIT(tile->id);
+		} else if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE) {
+			op->unmap_range.range->tile_present &= ~BIT(tile->id);
 		}
 		break;
 	}
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 32b1581b1ad0..58fbe1b65e8b 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -216,7 +216,14 @@ static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
 static int __xe_svm_garbage_collector(struct xe_vm *vm,
 				      struct xe_svm_range *range)
 {
-	/* TODO: Do unbind */
+	struct dma_fence *fence;
+
+	xe_vm_lock(vm, false);
+	fence = xe_vm_range_unbind(vm, range);
+	xe_vm_unlock(vm);
+	if (IS_ERR(fence))
+		return PTR_ERR(fence);
+	dma_fence_put(fence);
 
 	drm_gpusvm_range_remove(&vm->svm.gpusvm, &range->base);
 
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 12e5e562c5e1..20ea099664af 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -984,6 +984,89 @@ struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
 	return fence;
 }
 
+static void xe_vm_populate_range_unbind(struct xe_vma_op *op,
+					struct xe_svm_range *range)
+{
+	INIT_LIST_HEAD(&op->link);
+	op->tile_mask = range->tile_present;
+	op->base.op = DRM_GPUVA_OP_DRIVER;
+	op->subop = XE_VMA_SUBOP_UNMAP_RANGE;
+	op->unmap_range.range = range;
+}
+
+static int
+xe_vm_ops_add_range_unbind(struct xe_vma_ops *vops,
+			   struct xe_svm_range *range)
+{
+	struct xe_vma_op *op;
+
+	op = kzalloc(sizeof(*op), GFP_KERNEL);
+	if (!op)
+		return -ENOMEM;
+
+	xe_vm_populate_range_unbind(op, range);
+	list_add_tail(&op->link, &vops->list);
+	xe_vma_ops_incr_pt_update_ops(vops, range->tile_present);
+
+	return 0;
+}
+
+/**
+ * xe_vm_range_unbind() - VM range unbind
+ * @vm: The VM which the range belongs to.
+ * @range: SVM range to rebind.
+ *
+ * Unbind SVM range removing the GPU page tables for the range.
+ *
+ * Return: dma fence for unbind to signal completion on succees, ERR_PTR on
+ * failure
+ */
+struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm,
+				     struct xe_svm_range *range)
+{
+	struct dma_fence *fence = NULL;
+	struct xe_vma_ops vops;
+	struct xe_vma_op *op, *next_op;
+	struct xe_tile *tile;
+	u8 id;
+	int err;
+
+	lockdep_assert_held(&vm->lock);
+	xe_vm_assert_held(vm);
+	xe_assert(vm->xe, xe_vm_in_fault_mode(vm));
+
+	if (!range->tile_present)
+		return dma_fence_get_stub();
+
+	xe_vma_ops_init(&vops, vm, NULL, NULL, 0);
+	for_each_tile(tile, vm->xe, id) {
+		vops.pt_update_ops[id].wait_vm_bookkeep = true;
+		vops.pt_update_ops[tile->id].q =
+			xe_tile_migrate_exec_queue(tile);
+	}
+
+	err = xe_vm_ops_add_range_unbind(&vops, range);
+	if (err)
+		return ERR_PTR(err);
+
+	err = xe_vma_ops_alloc(&vops, false);
+	if (err) {
+		fence = ERR_PTR(err);
+		goto free_ops;
+	}
+
+	fence = ops_execute(vm, &vops);
+
+free_ops:
+	list_for_each_entry_safe(op, next_op, &vops.list, link) {
+		list_del(&op->link);
+		kfree(op);
+	}
+	xe_vma_ops_fini(&vops);
+
+	return fence;
+}
+
 static void xe_vma_free(struct xe_vma *vma)
 {
 	if (xe_vma_is_userptr(vma))
diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h
index a82fe743bbe0..3b6316dd9fd6 100644
--- a/drivers/gpu/drm/xe/xe_vm.h
+++ b/drivers/gpu/drm/xe/xe_vm.h
@@ -221,6 +221,8 @@ struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
 				     struct xe_vma *vma,
 				     struct xe_svm_range *range,
 				     u8 tile_mask);
+struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm,
+				     struct xe_svm_range *range);
 
 int xe_vm_invalidate_vma(struct xe_vma *vma);
 
diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
index 576316729249..19576ac095e7 100644
--- a/drivers/gpu/drm/xe/xe_vm_types.h
+++ b/drivers/gpu/drm/xe/xe_vm_types.h
@@ -361,6 +361,12 @@ struct xe_vma_op_map_range {
 	struct xe_svm_range *range;
 };
 
+/** struct xe_vma_op_unmap_range - VMA unmap range operation */
+struct xe_vma_op_unmap_range {
+	/** @range: SVM range to unmap */
+	struct xe_svm_range *range;
+};
+
 /** enum xe_vma_op_flags - flags for VMA operation */
 enum xe_vma_op_flags {
 	/** @XE_VMA_OP_COMMITTED: VMA operation committed */
@@ -375,6 +381,8 @@ enum xe_vma_op_flags {
 enum xe_vma_subop {
 	/** @XE_VMA_SUBOP_MAP_RANGE: Map range */
 	XE_VMA_SUBOP_MAP_RANGE,
+	/** @XE_VMA_SUBOP_UNMAP_RANGE: Unmap range */
+	XE_VMA_SUBOP_UNMAP_RANGE,
 };
 
 /** struct xe_vma_op - VMA operation */
@@ -397,8 +405,10 @@ struct xe_vma_op {
 		struct xe_vma_op_remap remap;
 		/** @prefetch: VMA prefetch operation specific data */
 		struct xe_vma_op_prefetch prefetch;
-		/** @map: VMA map range operation specific data */
+		/** @map_range: VMA map range operation specific data */
 		struct xe_vma_op_map_range map_range;
+		/** @unmap_range: VMA unmap range operation specific data */
+		struct xe_vma_op_map_range unmap_range;
 	};
 };
 
-- 
2.34.1

[PATCH v3 16/30] drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has bindings

[Matthew Brost]

uAPI is designed with the use case that only mapping a BO to a malloc'd
address will unbind a CPU-address mirror VMA. Therefore, allowing a
CPU-address mirror VMA to unbind when the GPU has bindings in the range
being unbound does not make much sense. This behavior is not supported,
as it simplifies the code. This decision can always be revisited if a
use case arises.

v3:
 - s/arrises/arises (Thomas)
 - s/system allocator/GPU address mirror (Thomas)
 - Kernel doc (Thomas)
 - Newline between function defs (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c |  5 +++++
 drivers/gpu/drm/xe/xe_svm.h |  2 ++
 drivers/gpu/drm/xe/xe_vm.c  | 16 ++++++++++++++++
 3 files changed, 23 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 58fbe1b65e8b..8237b534a65b 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -427,3 +427,8 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 
 	return err;
 }
+
+bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
+{
+	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
+}
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index 5531868cdd6f..ddce1069ba1e 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -46,6 +46,8 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 			    struct xe_tile *tile, u64 fault_addr,
 			    bool atomic);
 
+bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end);
+
 static inline bool xe_svm_range_pages_valid(struct xe_svm_range *range)
 {
 	return drm_gpusvm_range_pages_valid(range->base.gpusvm, &range->base);
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 20ea099664af..613a68498a24 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -2398,6 +2398,17 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 			struct xe_vma *old =
 				gpuva_to_vma(op->base.remap.unmap->va);
 			bool skip = xe_vma_is_cpu_addr_mirror(old);
+			u64 start = xe_vma_start(old), end = xe_vma_end(old);
+
+			if (op->base.remap.prev)
+				start = op->base.remap.prev->va.addr +
+					op->base.remap.prev->va.range;
+			if (op->base.remap.next)
+				end = op->base.remap.next->va.addr;
+
+			if (xe_vma_is_cpu_addr_mirror(old) &&
+			    xe_svm_has_mapping(vm, start, end))
+				return -EBUSY;
 
 			op->remap.start = xe_vma_start(old);
 			op->remap.range = xe_vma_size(old);
@@ -2480,6 +2491,11 @@ static int vm_bind_ioctl_ops_parse(struct xe_vm *vm, struct drm_gpuva_ops *ops,
 		{
 			struct xe_vma *vma = gpuva_to_vma(op->base.unmap.va);
 
+			if (xe_vma_is_cpu_addr_mirror(vma) &&
+			    xe_svm_has_mapping(vm, xe_vma_start(vma),
+					       xe_vma_end(vma)))
+				return -EBUSY;
+
 			if (!xe_vma_is_cpu_addr_mirror(vma))
 				xe_vma_ops_incr_pt_update_ops(vops, op->tile_mask);
 			break;
-- 
2.34.1

[PATCH v3 17/30] drm/xe: Enable CPU address mirror uAPI

[Matthew Brost]

Support for CPU address mirror bindings in SRAM fully in place, enable the
implementation.

v3:
 - s/system allocator/CPU address mirror (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c | 10 ++++++++++
 drivers/gpu/drm/xe/xe_vm.c  |  6 ------
 2 files changed, 10 insertions(+), 6 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 8237b534a65b..3cbf29732249 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -428,6 +428,16 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	return err;
 }
 
+/**
+ * xe_svm_has_mapping() - SVM has mappings
+ * @vm: The VM.
+ * @start: Start address.
+ * @end: End address.
+ *
+ * Check if an address range has SVM mappings.
+ *
+ * Return: True is address range has a SVM mapping, False otherwise
+ */
 bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
 {
 	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
index 613a68498a24..12cbb140404d 100644
--- a/drivers/gpu/drm/xe/xe_vm.c
+++ b/drivers/gpu/drm/xe/xe_vm.c
@@ -3020,12 +3020,6 @@ static int vm_bind_ioctl_check_args(struct xe_device *xe, struct xe_vm *vm,
 		u16 pat_index = (*bind_ops)[i].pat_index;
 		u16 coh_mode;
 
-		/* FIXME: Disabling CPU address mirror for now */
-		if (XE_IOCTL_DBG(xe, is_cpu_addr_mirror)) {
-			err = -EOPNOTSUPP;
-			goto free_bind_ops;
-		}
-
 		if (XE_IOCTL_DBG(xe, is_cpu_addr_mirror &&
 				 !xe_vm_in_fault_mode(vm))) {
 			err = -EINVAL;
-- 
2.34.1

[PATCH v3 18/30] drm/xe: Add migrate layer functions for SVM support

[Matthew Brost]

Add functions which migrate to / from VRAM accepting a single DPA
argument (VRAM) and array of dma addresses (SRAM). Used for SVM
migrations.

v2:
 - Don't unlock job_mutex in error path of xe_migrate_vram
v3:
 - Kernel doc (Thomas)
 - Better commit message (Thomas)
 - s/dword/num_dword (Thomas)
 - Return error to large of migration (Thomas)

Signed-off-by: Oak Zeng <oak.zeng@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_migrate.c | 175 ++++++++++++++++++++++++++++++++
 drivers/gpu/drm/xe/xe_migrate.h |  10 ++
 2 files changed, 185 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_migrate.c b/drivers/gpu/drm/xe/xe_migrate.c
index 1b97d90aadda..0e5ec7724d1c 100644
--- a/drivers/gpu/drm/xe/xe_migrate.c
+++ b/drivers/gpu/drm/xe/xe_migrate.c
@@ -1544,6 +1544,181 @@ void xe_migrate_wait(struct xe_migrate *m)
 		dma_fence_wait(m->fence, false);
 }
 
+static u32 pte_update_cmd_size(u64 size)
+{
+	u32 num_dword;
+	u64 entries = DIV_ROUND_UP(size, XE_PAGE_SIZE);
+
+	XE_WARN_ON(size > MAX_PREEMPTDISABLE_TRANSFER);
+	/*
+	 * MI_STORE_DATA_IMM command is used to update page table. Each
+	 * instruction can update maximumly 0x1ff pte entries. To update
+	 * n (n <= 0x1ff) pte entries, we need:
+	 * 1 dword for the MI_STORE_DATA_IMM command header (opcode etc)
+	 * 2 dword for the page table's physical location
+	 * 2*n dword for value of pte to fill (each pte entry is 2 dwords)
+	 */
+	num_dword = (1 + 2) * DIV_ROUND_UP(entries, 0x1ff);
+	num_dword += entries * 2;
+
+	return num_dword;
+}
+
+static void build_pt_update_batch_sram(struct xe_migrate *m,
+				       struct xe_bb *bb, u32 pt_offset,
+				       dma_addr_t *sram_addr, u32 size)
+{
+	u16 pat_index = tile_to_xe(m->tile)->pat.idx[XE_CACHE_WB];
+	u32 ptes;
+	int i = 0;
+
+	ptes = DIV_ROUND_UP(size, XE_PAGE_SIZE);
+	while (ptes) {
+		u32 chunk = min(0x1ffU, ptes);
+
+		bb->cs[bb->len++] = MI_STORE_DATA_IMM | MI_SDI_NUM_QW(chunk);
+		bb->cs[bb->len++] = pt_offset;
+		bb->cs[bb->len++] = 0;
+
+		pt_offset += chunk * 8;
+		ptes -= chunk;
+
+		while (chunk--) {
+			u64 addr = sram_addr[i++] & PAGE_MASK;
+
+			xe_tile_assert(m->tile, addr);
+			addr = m->q->vm->pt_ops->pte_encode_addr(m->tile->xe,
+								 addr, pat_index,
+								 0, false, 0);
+			bb->cs[bb->len++] = lower_32_bits(addr);
+			bb->cs[bb->len++] = upper_32_bits(addr);
+		}
+	}
+}
+
+enum xe_migrate_copy_dir {
+	XE_MIGRATE_COPY_TO_VRAM,
+	XE_MIGRATE_COPY_TO_SRAM,
+};
+
+static struct dma_fence *xe_migrate_vram(struct xe_migrate *m,
+					 unsigned long npages,
+					 dma_addr_t *sram_addr, u64 vram_addr,
+					 const enum xe_migrate_copy_dir dir)
+{
+	struct xe_gt *gt = m->tile->primary_gt;
+	struct xe_device *xe = gt_to_xe(gt);
+	struct dma_fence *fence = NULL;
+	u32 batch_size = 2;
+	u64 src_L0_ofs, dst_L0_ofs;
+	u64 round_update_size;
+	struct xe_sched_job *job;
+	struct xe_bb *bb;
+	u32 update_idx, pt_slot = 0;
+	int err;
+
+	if (npages * PAGE_SIZE > MAX_PREEMPTDISABLE_TRANSFER)
+		return ERR_PTR(-EINVAL);
+
+	round_update_size = npages * PAGE_SIZE;
+	batch_size += pte_update_cmd_size(round_update_size);
+	batch_size += EMIT_COPY_DW;
+
+	bb = xe_bb_new(gt, batch_size, true);
+	if (IS_ERR(bb)) {
+		err = PTR_ERR(bb);
+		return ERR_PTR(err);
+	}
+
+	build_pt_update_batch_sram(m, bb, pt_slot * XE_PAGE_SIZE,
+				   sram_addr, round_update_size);
+
+	if (dir == XE_MIGRATE_COPY_TO_VRAM) {
+		src_L0_ofs = xe_migrate_vm_addr(pt_slot, 0);
+		dst_L0_ofs = xe_migrate_vram_ofs(xe, vram_addr, false);
+
+	} else {
+		src_L0_ofs = xe_migrate_vram_ofs(xe, vram_addr, false);
+		dst_L0_ofs = xe_migrate_vm_addr(pt_slot, 0);
+	}
+
+	bb->cs[bb->len++] = MI_BATCH_BUFFER_END;
+	update_idx = bb->len;
+
+	emit_copy(gt, bb, src_L0_ofs, dst_L0_ofs, round_update_size,
+		  XE_PAGE_SIZE);
+
+	job = xe_bb_create_migration_job(m->q, bb,
+					 xe_migrate_batch_base(m, true),
+					 update_idx);
+	if (IS_ERR(job)) {
+		err = PTR_ERR(job);
+		goto err;
+	}
+
+	xe_sched_job_add_migrate_flush(job, 0);
+
+	mutex_lock(&m->job_mutex);
+	xe_sched_job_arm(job);
+	fence = dma_fence_get(&job->drm.s_fence->finished);
+	xe_sched_job_push(job);
+
+	dma_fence_put(m->fence);
+	m->fence = dma_fence_get(fence);
+	mutex_unlock(&m->job_mutex);
+
+	xe_bb_free(bb, fence);
+
+	return fence;
+
+err:
+	xe_bb_free(bb, NULL);
+
+	return ERR_PTR(err);
+}
+
+/**
+ * xe_migrate_to_vram() - Migrate to VRAM
+ * @m: The migration context.
+ * @npages: Number of pages to migrate.
+ * @src_addr: Array of dma addresses (source of migrate)
+ * @dst_addr: Device physical address of VRAM (destination of migrate)
+ *
+ * Copy from an array dma addresses to a VRAM device physical address
+ *
+ * Return: dma fence for migrate to signal completion on succees, ERR_PTR on
+ * failure
+ */
+struct dma_fence *xe_migrate_to_vram(struct xe_migrate *m,
+				     unsigned long npages,
+				     dma_addr_t *src_addr,
+				     u64 dst_addr)
+{
+	return xe_migrate_vram(m, npages, src_addr, dst_addr,
+			       XE_MIGRATE_COPY_TO_VRAM);
+}
+
+/**
+ * xe_migrate_from_vram() - Migrate from VRAM
+ * @m: The migration context.
+ * @npages: Number of pages to migrate.
+ * @src_addr: Device physical address of VRAM (source of migrate)
+ * @dst_addr: Array of dma addresses (destination of migrate)
+ *
+ * Copy from a VRAM device physical address to an array dma addresses
+ *
+ * Return: dma fence for migrate to signal completion on succees, ERR_PTR on
+ * failure
+ */
+struct dma_fence *xe_migrate_from_vram(struct xe_migrate *m,
+				       unsigned long npages,
+				       u64 src_addr,
+				       dma_addr_t *dst_addr)
+{
+	return xe_migrate_vram(m, npages, dst_addr, src_addr,
+			       XE_MIGRATE_COPY_TO_SRAM);
+}
+
 #if IS_ENABLED(CONFIG_DRM_XE_KUNIT_TEST)
 #include "tests/xe_migrate.c"
 #endif
diff --git a/drivers/gpu/drm/xe/xe_migrate.h b/drivers/gpu/drm/xe/xe_migrate.h
index 0109866e398a..6ff9a963425c 100644
--- a/drivers/gpu/drm/xe/xe_migrate.h
+++ b/drivers/gpu/drm/xe/xe_migrate.h
@@ -95,6 +95,16 @@ struct xe_migrate_pt_update {
 
 struct xe_migrate *xe_migrate_init(struct xe_tile *tile);
 
+struct dma_fence *xe_migrate_to_vram(struct xe_migrate *m,
+				     unsigned long npages,
+				     dma_addr_t *src_addr,
+				     u64 dst_addr);
+
+struct dma_fence *xe_migrate_from_vram(struct xe_migrate *m,
+				       unsigned long npages,
+				       u64 src_addr,
+				       dma_addr_t *dst_addr);
+
 struct dma_fence *xe_migrate_copy(struct xe_migrate *m,
 				  struct xe_bo *src_bo,
 				  struct xe_bo *dst_bo,
-- 
2.34.1

[PATCH v3 19/30] drm/xe: Add SVM device memory mirroring

[Matthew Brost]

Add SVM device memory mirroring which enables device pages for
migration. Enabled via CONFIG_XE_DEVMEM_MIRROR Kconfig. Kconfig option
defaults to enabled. If not enabled, SVM will work sans migration and
KMD memory footprint will be less.

v3:
 - Add CONFIG_XE_DEVMEM_MIRROR

Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Oak Zeng <oak.zeng@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/Kconfig           |  9 ++++
 drivers/gpu/drm/xe/xe_device_types.h |  8 ++++
 drivers/gpu/drm/xe/xe_svm.c          | 62 +++++++++++++++++++++++++++-
 drivers/gpu/drm/xe/xe_svm.h          |  3 ++
 drivers/gpu/drm/xe/xe_tile.c         |  5 +++
 5 files changed, 85 insertions(+), 2 deletions(-)

diff --git a/drivers/gpu/drm/xe/Kconfig b/drivers/gpu/drm/xe/Kconfig
index 3a08e16bfada..1989c99a9f5a 100644
--- a/drivers/gpu/drm/xe/Kconfig
+++ b/drivers/gpu/drm/xe/Kconfig
@@ -60,6 +60,15 @@ config DRM_XE_DISPLAY
 	help
 	  Disable this option only if you want to compile out display support.
 
+config CONFIG_XE_DEVMEM_MIRROR
+	bool "Enable device memory mirror"
+	depends on DRM_XE
+	select CONFIG_GET_FREE_REGION
+	default y
+	help
+	  Disable this option only if you want to compile out without device
+	  memory mirror. Will reduce KMD memory footprint when disabled.
+
 config DRM_XE_FORCE_PROBE
 	string "Force probe xe for selected Intel hardware IDs"
 	depends on DRM_XE
diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index 8a7b15972413..b949a960cebb 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -102,6 +102,14 @@ struct xe_mem_region {
 	resource_size_t actual_physical_size;
 	/** @mapping: pointer to VRAM mappable space */
 	void __iomem *mapping;
+	/** @pagemap: Used to remap device memory as ZONE_DEVICE */
+	struct dev_pagemap pagemap;
+	/**
+	 * @hpa_base: base host physical address
+	 *
+	 * This is generated when remap device memory as ZONE_DEVICE
+	 */
+	resource_size_t hpa_base;
 };
 
 /**
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 3cbf29732249..2860e54bb035 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -19,6 +19,11 @@ static struct xe_vm *range_to_vm(struct drm_gpusvm_range *r)
 	return gpusvm_to_vm(r->gpusvm);
 }
 
+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)
 {
@@ -306,8 +311,8 @@ 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, NULL, 0, vm->size,
-			      SZ_512M, &gpusvm_ops, fault_chunk_sizes,
+			      current->mm, xe_svm_devm_owner(vm->xe), 0,
+			      vm->size, SZ_512M, &gpusvm_ops, fault_chunk_sizes,
 			      ARRAY_SIZE(fault_chunk_sizes));
 	if (err)
 		return err;
@@ -442,3 +447,56 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
 {
 	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
 }
+
+#if IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR)
+/**
+ * xe_devm_add: Remap and provide memmap backing for device memory
+ * @tile: tile that the memory region belongs to
+ * @mr: memory region to remap
+ *
+ * This remap device memory to host physical address space and create
+ * struct page to back device memory
+ *
+ * Return: 0 on success standard error code otherwise
+ */
+int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
+{
+	struct xe_device *xe = tile_to_xe(tile);
+	struct device *dev = &to_pci_dev(xe->drm.dev)->dev;
+	struct resource *res;
+	void *addr;
+	int ret;
+
+	res = devm_request_free_mem_region(dev, &iomem_resource,
+					   mr->usable_size);
+	if (IS_ERR(res)) {
+		ret = PTR_ERR(res);
+		return ret;
+	}
+
+	mr->pagemap.type = MEMORY_DEVICE_PRIVATE;
+	mr->pagemap.range.start = res->start;
+	mr->pagemap.range.end = res->end;
+	mr->pagemap.nr_range = 1;
+	mr->pagemap.ops = drm_gpusvm_pagemap_ops_get();
+	mr->pagemap.owner = xe_svm_devm_owner(xe);
+	addr = devm_memremap_pages(dev, &mr->pagemap);
+	if (IS_ERR(addr)) {
+		devm_release_mem_region(dev, res->start, resource_size(res));
+		ret = PTR_ERR(addr);
+		drm_err(&xe->drm, "Failed to remap tile %d memory, errno %d\n",
+				tile->id, ret);
+		return ret;
+	}
+	mr->hpa_base = res->start;
+
+	drm_info(&xe->drm, "Added tile %d memory [%llx-%llx] to devm, remapped to %pr\n",
+		 tile->id, mr->io_start, mr->io_start + mr->usable_size, res);
+	return 0;
+}
+#else
+int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
+{
+	return 0;
+}
+#endif
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index ddce1069ba1e..85832e6ed3f5 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -11,6 +11,7 @@
 
 #define XE_INTERCONNECT_VRAM DRM_INTERCONNECT_DRIVER
 
+struct xe_mem_region;
 struct xe_tile;
 struct xe_vm;
 struct xe_vma;
@@ -36,6 +37,8 @@ struct xe_svm_range {
 	u8 tile_invalidated;
 };
 
+int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr);
+
 int xe_svm_init(struct xe_vm *vm);
 
 void xe_svm_fini(struct xe_vm *vm);
diff --git a/drivers/gpu/drm/xe/xe_tile.c b/drivers/gpu/drm/xe/xe_tile.c
index 07cf7cfe4abd..bbb430392dfb 100644
--- a/drivers/gpu/drm/xe/xe_tile.c
+++ b/drivers/gpu/drm/xe/xe_tile.c
@@ -13,6 +13,7 @@
 #include "xe_migrate.h"
 #include "xe_pcode.h"
 #include "xe_sa.h"
+#include "xe_svm.h"
 #include "xe_tile.h"
 #include "xe_tile_sysfs.h"
 #include "xe_ttm_vram_mgr.h"
@@ -164,6 +165,7 @@ static int tile_ttm_mgr_init(struct xe_tile *tile)
  */
 int xe_tile_init_noalloc(struct xe_tile *tile)
 {
+	struct xe_device *xe = tile_to_xe(tile);
 	int err;
 
 	err = tile_ttm_mgr_init(tile);
@@ -176,6 +178,9 @@ int xe_tile_init_noalloc(struct xe_tile *tile)
 
 	xe_wa_apply_tile_workarounds(tile);
 
+	if (xe->info.has_usm && IS_DGFX(xe))
+		xe_devm_add(tile, &tile->mem.vram);
+
 	err = xe_tile_sysfs_init(tile);
 
 	return 0;
-- 
2.34.1

[PATCH v3 20/30] drm/xe: Add drm_gpusvm_devmem to xe_bo

[Matthew Brost]

Add drm_gpusvm_devmem to xe_bo. Required to enable SVM migrations.

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_bo_types.h | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_bo_types.h b/drivers/gpu/drm/xe/xe_bo_types.h
index 46dc9e4e3e46..6d53ccde0256 100644
--- a/drivers/gpu/drm/xe/xe_bo_types.h
+++ b/drivers/gpu/drm/xe/xe_bo_types.h
@@ -8,6 +8,7 @@
 
 #include <linux/iosys-map.h>
 
+#include <drm/drm_gpusvm.h>
 #include <drm/ttm/ttm_bo.h>
 #include <drm/ttm/ttm_device.h>
 #include <drm/ttm/ttm_placement.h>
@@ -74,6 +75,9 @@ struct xe_bo {
 	 */
 	u16 cpu_caching;
 
+	/** @devmem_allocation: SVM device memory allocation */
+	struct drm_gpusvm_devmem devmem_allocation;
+
 	/** @vram_userfault_link: Link into @mem_access.vram_userfault.list */
 		struct list_head vram_userfault_link;
 
-- 
2.34.1

[PATCH v3 21/30] drm/xe: Add drm_pagemap ops to SVM

[Matthew Brost]

From: Thomas Hellström <thomas.hellstrom@linux.intel.com>

Add support for mapping device pages to Xe SVM by attaching drm_pagemap
to a memory region, which is then linked to a GPU SVM devmem allocation.
This enables GPU SVM to derive the device page address.

v3:
 - Better commit message (Thomas)
 - New drm_pagemap.h location

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_device_types.h |  7 +++++++
 drivers/gpu/drm/xe/xe_svm.c          | 30 ++++++++++++++++++++++++++++
 2 files changed, 37 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
index b949a960cebb..b76d08df13ef 100644
--- a/drivers/gpu/drm/xe/xe_device_types.h
+++ b/drivers/gpu/drm/xe/xe_device_types.h
@@ -10,8 +10,10 @@
 
 #include <drm/drm_device.h>
 #include <drm/drm_file.h>
+#include <drm/drm_pagemap.h>
 #include <drm/ttm/ttm_device.h>
 
+
 #include "xe_devcoredump_types.h"
 #include "xe_heci_gsc.h"
 #include "xe_lmtt_types.h"
@@ -104,6 +106,11 @@ struct xe_mem_region {
 	void __iomem *mapping;
 	/** @pagemap: Used to remap device memory as ZONE_DEVICE */
 	struct dev_pagemap pagemap;
+	/**
+	 * @dpagemap: The struct drm_pagemap of the ZONE_DEVICE memory
+	 * pages of this tile.
+	 */
+	struct drm_pagemap dpagemap;
 	/**
 	 * @hpa_base: base host physical address
 	 *
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 2860e54bb035..faeacf0ccdaa 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -449,6 +449,32 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
 }
 
 #if IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR)
+static struct drm_pagemap_dma_addr
+xe_drm_pagemap_map_dma(struct drm_pagemap *dpagemap,
+		       struct device *dev,
+		       struct page *page,
+		       unsigned int order,
+		       enum dma_data_direction dir)
+{
+	struct device *pgmap_dev = dpagemap->dev;
+	enum drm_interconnect_protocol prot;
+	dma_addr_t addr;
+
+	if (pgmap_dev == dev) {
+		addr = xe_mem_region_page_to_dpa(page_to_mr(page), page);
+		prot = XE_INTERCONNECT_VRAM;
+	} else {
+		addr = DMA_MAPPING_ERROR;
+		prot = 0;
+	}
+
+	return drm_pagemap_dma_addr_encode(addr, prot, order, dir);
+}
+
+static const struct drm_pagemap_ops xe_drm_pagemap_ops = {
+	.map_dma = xe_drm_pagemap_map_dma,
+};
+
 /**
  * xe_devm_add: Remap and provide memmap backing for device memory
  * @tile: tile that the memory region belongs to
@@ -481,6 +507,10 @@ int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
 	mr->pagemap.ops = drm_gpusvm_pagemap_ops_get();
 	mr->pagemap.owner = xe_svm_devm_owner(xe);
 	addr = devm_memremap_pages(dev, &mr->pagemap);
+
+	mr->dpagemap.dev = dev;
+	mr->dpagemap.ops = &xe_drm_pagemap_ops;
+
 	if (IS_ERR(addr)) {
 		devm_release_mem_region(dev, res->start, resource_size(res));
 		ret = PTR_ERR(addr);
-- 
2.34.1

[PATCH v3 22/30] drm/xe: Add GPUSVM device memory copy vfunc functions

[Matthew Brost]

Add GPUSVM device memory copy vfunc functions and connect to migration
layer. Used for device memory migration.

v2:
 - Allow NULL device pages in xe_svm_copy
 - Use new drm_gpusvm_devmem_ops
v3:
 - Prefix defines with XE_ (Thomas)
 - Change copy chunk size to 8M
 - Add a bunch of comments to xe_svm_copy to clarify behavior (Thomas)
 - Better commit message (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c | 179 ++++++++++++++++++++++++++++++++++++
 1 file changed, 179 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index faeacf0ccdaa..dcd99520a6d9 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -4,6 +4,7 @@
  */
 
 #include "xe_gt_tlb_invalidation.h"
+#include "xe_migrate.h"
 #include "xe_pt.h"
 #include "xe_svm.h"
 #include "xe_vm.h"
@@ -281,6 +282,184 @@ static void xe_svm_garbage_collector_work_func(struct work_struct *w)
 	up_write(&vm->lock);
 }
 
+static struct xe_mem_region *page_to_mr(struct page *page)
+{
+	return container_of(page->pgmap, struct xe_mem_region, pagemap);
+}
+
+static struct xe_tile *mr_to_tile(struct xe_mem_region *mr)
+{
+	return container_of(mr, struct xe_tile, mem.vram);
+}
+
+static u64 xe_mem_region_page_to_dpa(struct xe_mem_region *mr,
+				     struct page *page)
+{
+	u64 dpa;
+	struct xe_tile *tile = mr_to_tile(mr);
+	u64 pfn = page_to_pfn(page);
+	u64 offset;
+
+	xe_tile_assert(tile, is_device_private_page(page));
+	xe_tile_assert(tile, (pfn << PAGE_SHIFT) >= mr->hpa_base);
+
+	offset = (pfn << PAGE_SHIFT) - mr->hpa_base;
+	dpa = mr->dpa_base + offset;
+
+	return dpa;
+}
+
+enum xe_svm_copy_dir {
+	XE_SVM_COPY_TO_VRAM,
+	XE_SVM_COPY_TO_SRAM,
+};
+
+static int xe_svm_copy(struct page **pages, dma_addr_t *dma_addr,
+		       unsigned long npages, const enum xe_svm_copy_dir dir)
+{
+	struct xe_mem_region *mr = NULL;
+	struct xe_tile *tile;
+	struct dma_fence *fence = NULL;
+	unsigned long i;
+#define XE_VRAM_ADDR_INVALID	~0x0ull
+	u64 vram_addr = XE_VRAM_ADDR_INVALID;
+	int err = 0, pos = 0;
+	bool sram = dir == XE_SVM_COPY_TO_SRAM;
+
+	/*
+	 * This flow is complex: it locates physically contiguous device pages,
+	 * derives the starting physical address, and performs a single GPU copy
+	 * to for every 8M chunk in a DMA address array. Both device pages and
+	 * DMA addresses may be sparsely populated. If either is NULL, a copy is
+	 * triggered based on the current search state. The last GPU copy is
+	 * waited on to ensure all copies are complete.
+	 */
+
+	for (i = 0; i < npages; ++i) {
+		struct page *spage = pages[i];
+		struct dma_fence *__fence;
+		u64 __vram_addr;
+		bool match = false, chunk, last;
+
+#define XE_MIGRATE_CHUNK_SIZE	SZ_8M
+		chunk = (i - pos) == (XE_MIGRATE_CHUNK_SIZE / PAGE_SIZE);
+		last = (i + 1) == npages;
+
+		/* No CPU page and no device pages queue'd to copy */
+		if (!dma_addr[i] && vram_addr == XE_VRAM_ADDR_INVALID)
+			continue;
+
+		if (!mr && spage) {
+			mr = page_to_mr(spage);
+			tile = mr_to_tile(mr);
+		}
+		XE_WARN_ON(spage && page_to_mr(spage) != mr);
+
+		/*
+		 * CPU page and device page valid, capture physical address on
+		 * first device page, check if physical contiguous on subsequent
+		 * device pages.
+		 */
+		if (dma_addr[i] && spage) {
+			__vram_addr = xe_mem_region_page_to_dpa(mr, spage);
+			if (vram_addr == XE_VRAM_ADDR_INVALID) {
+				vram_addr = __vram_addr;
+				pos = i;
+			}
+
+			match = vram_addr + PAGE_SIZE * (i - pos) == __vram_addr;
+		}
+
+		/*
+		 * Mismatched physical address, 8M copy chunk, or last page -
+		 * trigger a copy.
+		 */
+		if (!match || chunk || last) {
+			/*
+			 * Extra page for first copy if last page and matching
+			 * physical address.
+			 */
+			int incr = (match && last) ? 1 : 0;
+
+			if (vram_addr != XE_VRAM_ADDR_INVALID) {
+				if (sram)
+					__fence = xe_migrate_from_vram(tile->migrate,
+								       i - pos + incr,
+								       vram_addr,
+								       dma_addr + pos);
+				else
+					__fence = xe_migrate_to_vram(tile->migrate,
+								     i - pos + incr,
+								     dma_addr + pos,
+								     vram_addr);
+				if (IS_ERR(__fence)) {
+					err = PTR_ERR(__fence);
+					goto err_out;
+				}
+
+				dma_fence_put(fence);
+				fence = __fence;
+			}
+
+			/* Setup physical address of next device page */
+			if (dma_addr[i] && spage) {
+				vram_addr = __vram_addr;
+				pos = i;
+			} else {
+				vram_addr = XE_VRAM_ADDR_INVALID;
+			}
+
+			/* Extra mismatched device page, copy it */
+			if (!match && last && vram_addr != XE_VRAM_ADDR_INVALID) {
+				if (sram)
+					__fence = xe_migrate_from_vram(tile->migrate, 1,
+								       vram_addr,
+								       dma_addr + pos);
+				else
+					__fence = xe_migrate_to_vram(tile->migrate, 1,
+								     dma_addr + pos,
+								     vram_addr);
+				if (IS_ERR(__fence)) {
+					err = PTR_ERR(__fence);
+					goto err_out;
+				}
+
+				dma_fence_put(fence);
+				fence = __fence;
+			}
+		}
+	}
+
+err_out:
+	/* Wait for all copies to complete */
+	if (fence) {
+		dma_fence_wait(fence, false);
+		dma_fence_put(fence);
+	}
+
+	return err;
+#undef XE_MIGRATE_CHUNK_SIZE
+#undef XE_VRAM_ADDR_INVALID
+}
+
+static int xe_svm_copy_to_devmem(struct page **pages, dma_addr_t *dma_addr,
+				 unsigned long npages)
+{
+	return xe_svm_copy(pages, dma_addr, npages, XE_SVM_COPY_TO_VRAM);
+}
+
+static int xe_svm_copy_to_ram(struct page **pages, dma_addr_t *dma_addr,
+			      unsigned long npages)
+{
+	return xe_svm_copy(pages, dma_addr, npages, XE_SVM_COPY_TO_SRAM);
+}
+
+__maybe_unused
+static const struct drm_gpusvm_devmem_ops gpusvm_devmem_ops = {
+	.copy_to_devmem = xe_svm_copy_to_devmem,
+	.copy_to_ram = xe_svm_copy_to_ram,
+};
+
 static const struct drm_gpusvm_ops gpusvm_ops = {
 	.range_alloc = xe_svm_range_alloc,
 	.range_free = xe_svm_range_free,
-- 
2.34.1

[PATCH v3 23/30] drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc

[Matthew Brost]

Get device pfns from BO's buddy blocks. Used in migrate_* core MM
functions called in GPU SVM to migrate between device and system memory.

v2:
 - Use new drm_gpusvm_devmem_ops
v3:
 - Better commit message (Thomas)

Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
Signed-off-by: Oak Zeng <oak.zeng@intel.com>
Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c | 40 +++++++++++++++++++++++++++++++++++++
 1 file changed, 40 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index dcd99520a6d9..16d526b2d19a 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -7,6 +7,7 @@
 #include "xe_migrate.h"
 #include "xe_pt.h"
 #include "xe_svm.h"
+#include "xe_ttm_vram_mgr.h"
 #include "xe_vm.h"
 #include "xe_vm_types.h"
 
@@ -454,8 +455,47 @@ static int xe_svm_copy_to_ram(struct page **pages, dma_addr_t *dma_addr,
 	return xe_svm_copy(pages, dma_addr, npages, XE_SVM_COPY_TO_SRAM);
 }
 
+static struct xe_bo *to_xe_bo(struct drm_gpusvm_devmem *devmem_allocation)
+{
+	return container_of(devmem_allocation, struct xe_bo, devmem_allocation);
+}
+
+static u64 block_offset_to_pfn(struct xe_mem_region *mr, u64 offset)
+{
+	return PHYS_PFN(offset + mr->hpa_base);
+}
+
+static struct drm_buddy *tile_to_buddy(struct xe_tile *tile)
+{
+	return &tile->mem.vram_mgr->mm;
+}
+
+static int xe_svm_populate_devmem_pfn(struct drm_gpusvm_devmem *devmem_allocation,
+				      unsigned long npages, unsigned long *pfn)
+{
+	struct xe_bo *bo = to_xe_bo(devmem_allocation);
+	struct ttm_resource *res = bo->ttm.resource;
+	struct list_head *blocks = &to_xe_ttm_vram_mgr_resource(res)->blocks;
+	struct drm_buddy_block *block;
+	int j = 0;
+
+	list_for_each_entry(block, blocks, link) {
+		struct xe_mem_region *mr = block->private;
+		struct xe_tile *tile = mr_to_tile(mr);
+		struct drm_buddy *buddy = tile_to_buddy(tile);
+		u64 block_pfn = block_offset_to_pfn(mr, drm_buddy_block_offset(block));
+		int i;
+
+		for (i = 0; i < drm_buddy_block_size(buddy, block) >> PAGE_SHIFT; ++i)
+			pfn[j++] = block_pfn + i;
+	}
+
+	return 0;
+}
+
 __maybe_unused
 static const struct drm_gpusvm_devmem_ops gpusvm_devmem_ops = {
+	.populate_devmem_pfn = xe_svm_populate_devmem_pfn,
 	.copy_to_devmem = xe_svm_copy_to_devmem,
 	.copy_to_ram = xe_svm_copy_to_ram,
 };
-- 
2.34.1

[PATCH v3 24/30] drm/xe: Add Xe SVM devmem_release GPU SVM vfunc

[Matthew Brost]

Implement with a simple BO put which releases the device memory.

v2:
 - Use new drm_gpusvm_devmem_ops
v3:
 - Better commit message (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c | 7 +++++++
 1 file changed, 7 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 16d526b2d19a..466a0b8cbfa8 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -3,6 +3,7 @@
  * Copyright © 2024 Intel Corporation
  */
 
+#include "xe_bo.h"
 #include "xe_gt_tlb_invalidation.h"
 #include "xe_migrate.h"
 #include "xe_pt.h"
@@ -460,6 +461,11 @@ static struct xe_bo *to_xe_bo(struct drm_gpusvm_devmem *devmem_allocation)
 	return container_of(devmem_allocation, struct xe_bo, devmem_allocation);
 }
 
+static void xe_svm_devmem_release(struct drm_gpusvm_devmem *devmem_allocation)
+{
+	xe_bo_put(to_xe_bo(devmem_allocation));
+}
+
 static u64 block_offset_to_pfn(struct xe_mem_region *mr, u64 offset)
 {
 	return PHYS_PFN(offset + mr->hpa_base);
@@ -495,6 +501,7 @@ static int xe_svm_populate_devmem_pfn(struct drm_gpusvm_devmem *devmem_allocatio
 
 __maybe_unused
 static const struct drm_gpusvm_devmem_ops gpusvm_devmem_ops = {
+	.devmem_release = xe_svm_devmem_release,
 	.populate_devmem_pfn = xe_svm_populate_devmem_pfn,
 	.copy_to_devmem = xe_svm_copy_to_devmem,
 	.copy_to_ram = xe_svm_copy_to_ram,
-- 
2.34.1

[PATCH v3 25/30] drm/xe: Add BO flags required for SVM

[Matthew Brost]

Add XE_BO_FLAG_CPU_ADDR_MIRROR to indicate BO is tied to SVM range.
While these BO's are kernel allocations, we need a VM reference in this
case which this flag indicates. In addition, we do not support CCS on
these BO's either. The later can be revisited later.

v2:
 - Take VM ref for system allocator BOs
v3:
 - s/XE_BO_FLAG_SYSTEM_ALLOC/XE_BO_FLAG_CPU_ADDR_MIRROR (Thomas)
 - Better commit message (Thomas)
 - Drop XE_BO_FLAG_SKIP_CLEAR for now
 - Add comment about possibly supporting CCS (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_bo.c | 12 ++++++++----
 drivers/gpu/drm/xe/xe_bo.h |  1 +
 2 files changed, 9 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index a670a0444e03..0c1134d6209c 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -1222,7 +1222,7 @@ static void xe_ttm_bo_destroy(struct ttm_buffer_object *ttm_bo)
 		xe_drm_client_remove_bo(bo);
 #endif
 
-	if (bo->vm && xe_bo_is_user(bo))
+	if (bo->vm && (xe_bo_is_user(bo) || bo->flags & XE_BO_FLAG_CPU_ADDR_MIRROR))
 		xe_vm_put(bo->vm);
 
 	mutex_lock(&xe->mem_access.vram_userfault.lock);
@@ -1418,7 +1418,8 @@ struct xe_bo *___xe_bo_create_locked(struct xe_device *xe, struct xe_bo *bo,
 	int err;
 
 	/* Only kernel objects should set GT */
-	xe_assert(xe, !tile || type == ttm_bo_type_kernel);
+	xe_assert(xe, !tile || type == ttm_bo_type_kernel ||
+		  flags & XE_BO_FLAG_CPU_ADDR_MIRROR);
 
 	if (XE_WARN_ON(!size)) {
 		xe_bo_free(bo);
@@ -1614,7 +1615,7 @@ __xe_bo_create_locked(struct xe_device *xe,
 	 * by having all the vm's bo refereferences released at vm close
 	 * time.
 	 */
-	if (vm && xe_bo_is_user(bo))
+	if (vm && (xe_bo_is_user(bo) || bo->flags & XE_BO_FLAG_CPU_ADDR_MIRROR))
 		xe_vm_get(vm);
 	bo->vm = vm;
 
@@ -2465,8 +2466,11 @@ bool xe_bo_needs_ccs_pages(struct xe_bo *bo)
 	 * system memory (i.e., it allows XE_PL_TT placement), FlatCCS
 	 * can't be used since there's no CCS storage associated with
 	 * non-VRAM addresses.
+	 *
+	 * XXX: Can we support CCS with CPU address mirroring?
 	 */
-	if (IS_DGFX(xe) && (bo->flags & XE_BO_FLAG_SYSTEM))
+	if (IS_DGFX(xe) && ((bo->flags & XE_BO_FLAG_SYSTEM) ||
+	    (bo->flags & XE_BO_FLAG_CPU_ADDR_MIRROR)))
 		return false;
 
 	return true;
diff --git a/drivers/gpu/drm/xe/xe_bo.h b/drivers/gpu/drm/xe/xe_bo.h
index d9386ab03140..a281e0187ebe 100644
--- a/drivers/gpu/drm/xe/xe_bo.h
+++ b/drivers/gpu/drm/xe/xe_bo.h
@@ -47,6 +47,7 @@
 					 XE_BO_FLAG_GGTT1 | \
 					 XE_BO_FLAG_GGTT2 | \
 					 XE_BO_FLAG_GGTT3)
+#define XE_BO_FLAG_CPU_ADDR_MIRROR	BIT(22)
 
 /* this one is trigger internally only */
 #define XE_BO_FLAG_INTERNAL_TEST	BIT(30)
-- 
2.34.1

[PATCH v3 26/30] drm/xe: Add SVM VRAM migration

[Matthew Brost]

Migration is implemented with range granularity, with VRAM backing being
a VM private TTM BO (i.e., shares dma-resv with VM). The lifetime of the
TTM BO is limited to when the SVM range is in VRAM (i.e., when a VRAM
SVM range is migrated to SRAM, the TTM BO is destroyed).

The design choice for using TTM BO for VRAM backing store, as opposed to
direct buddy allocation, is as follows:

- DRM buddy allocations are not at page granularity, offering no
  advantage over a BO.
- Unified eviction is required (SVM VRAM and TTM BOs need to be able to
  evict each other).
- For exhaustive eviction [1], SVM VRAM allocations will almost certainly
  require a dma-resv.
- Likely allocation size is 2M which makes of size of BO (872)
  acceptable per allocation (872 / 2M == .0004158).

With this, using TTM BO for VRAM backing store seems to be an obvious
choice as it allows leveraging of the TTM eviction code.

Current migration policy is migrate any SVM range greater than or equal
to 64k once.

[1] https://patchwork.freedesktop.org/series/133643/

v2:
 - Rebase on latest GPU SVM
 - Retry page fault on get pages returning mixed allocation
 - Use drm_gpusvm_devmem
v3:
 - Use new BO flags
 - New range structure (Thomas)
 - Hide migration behind Kconfig
 - Kernel doc (Thomas)
 - Use check_pages_threshold

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c | 102 ++++++++++++++++++++++++++++++++++--
 drivers/gpu/drm/xe/xe_svm.h |   5 ++
 2 files changed, 103 insertions(+), 4 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 466a0b8cbfa8..a417d8942da4 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -226,6 +226,9 @@ static int __xe_svm_garbage_collector(struct xe_vm *vm,
 {
 	struct dma_fence *fence;
 
+	if (IS_DGFX(vm->xe) && range->base.flags.partial_unmap)
+		drm_gpusvm_range_evict(&vm->svm.gpusvm, &range->base);
+
 	xe_vm_lock(vm, false);
 	fence = xe_vm_range_unbind(vm, range);
 	xe_vm_unlock(vm);
@@ -499,7 +502,6 @@ static int xe_svm_populate_devmem_pfn(struct drm_gpusvm_devmem *devmem_allocatio
 	return 0;
 }
 
-__maybe_unused
 static const struct drm_gpusvm_devmem_ops gpusvm_devmem_ops = {
 	.devmem_release = xe_svm_devmem_release,
 	.populate_devmem_pfn = xe_svm_populate_devmem_pfn,
@@ -579,6 +581,64 @@ static bool xe_svm_range_is_valid(struct xe_svm_range *range,
 	return (range->tile_present & ~range->tile_invalidated) & BIT(tile->id);
 }
 
+static struct xe_mem_region *tile_to_mr(struct xe_tile *tile)
+{
+	return &tile->mem.vram;
+}
+
+static struct xe_bo *xe_svm_alloc_vram(struct xe_vm *vm, struct xe_tile *tile,
+				       struct xe_svm_range *range,
+				       const struct drm_gpusvm_ctx *ctx)
+{
+	struct xe_mem_region *mr = tile_to_mr(tile);
+	struct drm_buddy_block *block;
+	struct list_head *blocks;
+	struct xe_bo *bo;
+	ktime_t end = 0;
+	int err;
+
+retry:
+	xe_vm_lock(vm, false);
+	bo = xe_bo_create(tile_to_xe(tile), tile, vm, range->base.itree.last + 1 -
+			  range->base.itree.start, ttm_bo_type_device,
+			  XE_BO_FLAG_VRAM_IF_DGFX(tile) |
+			  XE_BO_FLAG_CPU_ADDR_MIRROR);
+	xe_vm_unlock(vm);
+	if (IS_ERR(bo)) {
+		err = PTR_ERR(bo);
+		if (xe_vm_validate_should_retry(NULL, err, &end))
+			goto retry;
+		return bo;
+	}
+
+	drm_gpusvm_devmem_init(&bo->devmem_allocation,
+			       vm->xe->drm.dev, vm->svm.gpusvm.mm,
+			       &gpusvm_devmem_ops,
+			       &tile->mem.vram.dpagemap,
+			       range->base.itree.last + 1 -
+			       range->base.itree.start);
+
+	blocks = &to_xe_ttm_vram_mgr_resource(bo->ttm.resource)->blocks;
+	list_for_each_entry(block, blocks, link)
+		block->private = mr;
+
+	/*
+	 * Take ref because as soon as drm_gpusvm_migrate_to_devmem succeeds the
+	 * creation ref can be dropped upon CPU fault or unmap.
+	 */
+	xe_bo_get(bo);
+
+	err = drm_gpusvm_migrate_to_devmem(&vm->svm.gpusvm, &range->base,
+					   &bo->devmem_allocation, ctx);
+	if (err) {
+		xe_bo_put(bo);	/* Local ref */
+		xe_bo_put(bo);	/* Creation ref */
+		return ERR_PTR(err);
+	}
+
+	return bo;
+}
+
 /**
  * xe_svm_handle_pagefault() - SVM handle page fault
  * @vm: The VM.
@@ -587,7 +647,8 @@ static bool xe_svm_range_is_valid(struct xe_svm_range *range,
  * @fault_addr: The GPU fault address.
  * @atomic: The fault atomic access bit.
  *
- * Create GPU bindings for a SVM page fault.
+ * Create GPU bindings for a SVM page fault. Optionally migrate to device
+ * memory.
  *
  * Return: 0 on success, negative error code on error.
  */
@@ -595,11 +656,18 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 			    struct xe_tile *tile, u64 fault_addr,
 			    bool atomic)
 {
-	struct drm_gpusvm_ctx ctx = { .read_only = xe_vma_read_only(vma), };
+	struct drm_gpusvm_ctx ctx = {
+		.read_only = xe_vma_read_only(vma),
+		.devmem_possible = IS_DGFX(vm->xe) &&
+			IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR),
+		.check_pages_threshold = IS_DGFX(vm->xe) &&
+			IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR) ? SZ_64K : 0,
+	};
 	struct xe_svm_range *range;
 	struct drm_gpusvm_range *r;
 	struct drm_exec exec;
 	struct dma_fence *fence;
+	struct xe_bo *bo = NULL;
 	ktime_t end = 0;
 	int err;
 
@@ -607,6 +675,9 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(vma));
 
 retry:
+	xe_bo_put(bo);
+	bo = NULL;
+
 	/* Always process UNMAPs first so view SVM ranges is current */
 	err = xe_svm_garbage_collector(vm);
 	if (err)
@@ -622,9 +693,31 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	if (xe_svm_range_is_valid(range, tile))
 		return 0;
 
+	/* XXX: Add migration policy, for now migrate range once */
+	if (!range->migrated && range->base.flags.migrate_devmem &&
+	    (range->base.itree.last + 1 - range->base.itree.start) >= SZ_64K) {
+		range->migrated = true;
+
+		bo = xe_svm_alloc_vram(vm, tile, range, &ctx);
+		if (IS_ERR(bo)) {
+			drm_info(&vm->xe->drm,
+				 "VRAM allocation failed, falling back to retrying, asid=%u, errno %ld\n",
+				 vm->usm.asid, PTR_ERR(bo));
+			bo = NULL;
+			goto retry;
+		}
+	}
+
 	err = drm_gpusvm_range_get_pages(&vm->svm.gpusvm, r, &ctx);
-	if (err == -EFAULT || err == -EPERM)	/* Corner where CPU mappings have changed */
+	/* Corner where CPU mappings have changed */
+	if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {
+		if (err == -EOPNOTSUPP)
+			drm_gpusvm_range_evict(&vm->svm.gpusvm, &range->base);
+		drm_info(&vm->xe->drm,
+			 "Get pages failed, falling back to retrying, asid=%u, gpusvm=0x%016llx, errno %d\n",
+			 vm->usm.asid, (u64)&vm->svm.gpusvm, err);
 		goto retry;
+	}
 	if (err)
 		goto err_out;
 
@@ -655,6 +748,7 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	dma_fence_put(fence);
 
 err_out:
+	xe_bo_put(bo);
 
 	return err;
 }
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index 85832e6ed3f5..d549dd9e8641 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -35,6 +35,11 @@ struct xe_svm_range {
 	 * range. Protected by GPU SVM notifier lock.
 	 */
 	u8 tile_invalidated;
+	/**
+	 * @migrated: Range has been migrated to device memory, protected by
+	 * GPU fault handler locking.
+	 */
+	u8 migrated	:1;
 };
 
 int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr);
-- 
2.34.1

[PATCH v3 27/30] drm/xe: Basic SVM BO eviction

[Matthew Brost]

Wire xe_bo_move to GPU SVM migration via new helper xe_svm_bo_evict.

v2:
 - Use xe_svm_bo_evict
 - Drop bo->range
v3:
 - Kernel doc (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_svm.c | 14 ++++++++++++++
 drivers/gpu/drm/xe/xe_svm.h |  2 ++
 2 files changed, 16 insertions(+)

diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index a417d8942da4..8780a0b2c81e 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -768,6 +768,20 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
 	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
 }
 
+/**
+ * xe_svm_bo_evict() - SVM evict BO to system memory
+ * @bo: BO to evict
+ *
+ * SVM evict BO to system memory. GPU SVM layer ensures all device pages
+ * are evicted before returning.
+ *
+ * Return: 0 on success standard error code otherwise
+ */
+int xe_svm_bo_evict(struct xe_bo *bo)
+{
+	return drm_gpusvm_evict_to_ram(&bo->devmem_allocation);
+}
+
 #if IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR)
 static struct drm_pagemap_dma_addr
 xe_drm_pagemap_map_dma(struct drm_pagemap *dpagemap,
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index d549dd9e8641..9e9d333bb9d3 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -56,6 +56,8 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 
 bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end);
 
+int xe_svm_bo_evict(struct xe_bo *bo);
+
 static inline bool xe_svm_range_pages_valid(struct xe_svm_range *range)
 {
 	return drm_gpusvm_range_pages_valid(range->base.gpusvm, &range->base);
-- 
2.34.1

[PATCH v3 28/30] drm/xe: Add SVM debug

[Matthew Brost]

Add some useful SVM debug logging fro SVM range which prints the range's
state.

v2:
 - Upadte logging with latest structure layout
v3:
 - Better commit message (Thomas)
 - New range structure (Thomas)
 - s/COLLECTOT/s/COLLECTOR (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
---
 drivers/gpu/drm/xe/xe_pt.c  |  8 ++++
 drivers/gpu/drm/xe/xe_svm.c | 95 +++++++++++++++++++++++++++++++++----
 drivers/gpu/drm/xe/xe_svm.h |  2 +
 3 files changed, 96 insertions(+), 9 deletions(-)

diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
index 9c90dfc81fbe..23a1696a67c0 100644
--- a/drivers/gpu/drm/xe/xe_pt.c
+++ b/drivers/gpu/drm/xe/xe_pt.c
@@ -647,6 +647,7 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
 		/* Move this entire thing to xe_svm.c? */
 		xe_svm_notifier_lock(xe_vma_vm(vma));
 		if (!xe_svm_range_pages_valid(range)) {
+			xe_svm_range_debug(range, "BIND PREPARE - RETRY");
 			xe_svm_notifier_unlock(xe_vma_vm(vma));
 			return -EAGAIN;
 		}
@@ -655,6 +656,10 @@ xe_pt_stage_bind(struct xe_tile *tile, struct xe_vma *vma,
 					 range->base.itree.last + 1 - range->base.itree.start,
 					 &curs);
 			is_devmem = xe_res_is_vram(&curs);
+			if (is_devmem)
+				xe_svm_range_debug(range, "BIND PREPARE - DMA VRAM");
+			else
+				xe_svm_range_debug(range, "BIND PREPARE - DMA");
 		} else {
 			xe_assert(xe, false);
 		}
@@ -1429,10 +1434,13 @@ static int xe_pt_svm_pre_commit(struct xe_migrate_pt_update *pt_update)
 		if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE)
 			continue;
 
+		xe_svm_range_debug(range, "PRE-COMMIT");
+
 		xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(op->map_range.vma));
 		xe_assert(vm->xe, op->subop == XE_VMA_SUBOP_MAP_RANGE);
 
 		if (!xe_svm_range_pages_valid(range)) {
+			xe_svm_range_debug(range, "PRE-COMMIT - RETRY");
 			xe_svm_notifier_unlock(vm);
 			return -EAGAIN;
 		}
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 8780a0b2c81e..530ba34a86dc 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -12,6 +12,18 @@
 #include "xe_vm.h"
 #include "xe_vm_types.h"
 
+static bool xe_svm_range_in_vram(struct xe_svm_range *range)
+{
+	/* Not reliable without notifier lock */
+	return range->base.flags.has_devmem_pages;
+}
+
+static bool xe_svm_range_has_vram_binding(struct xe_svm_range *range)
+{
+	/* Not reliable without notifier lock */
+	return xe_svm_range_in_vram(range) && range->tile_present;
+}
+
 static struct xe_vm *gpusvm_to_vm(struct drm_gpusvm *gpusvm)
 {
 	return container_of(gpusvm, struct xe_vm, svm.gpusvm);
@@ -22,6 +34,23 @@ static struct xe_vm *range_to_vm(struct drm_gpusvm_range *r)
 	return gpusvm_to_vm(r->gpusvm);
 }
 
+#define range_debug(r__, operaton__)					\
+	vm_dbg(&range_to_vm(&(r__)->base)->xe->drm,			\
+	       "%s: asid=%u, gpusvm=0x%016llx, vram=%d,%d, seqno=%lu, " \
+	       "start=0x%014lx, end=0x%014lx, size=%lu",		\
+	       (operaton__), range_to_vm(&(r__)->base)->usm.asid,	\
+	       (u64)(r__)->base.gpusvm,					\
+	       xe_svm_range_in_vram((r__)) ? 1 : 0,			\
+	       xe_svm_range_has_vram_binding((r__)) ? 1 : 0,		\
+	       (r__)->base.notifier_seq,				\
+	       (r__)->base.itree.start, (r__)->base.itree.last + 1,	\
+	       (r__)->base.itree.last + 1 - (r__)->base.itree.start)
+
+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;
@@ -59,6 +88,8 @@ xe_svm_garbage_collector_add_range(struct xe_vm *vm, struct xe_svm_range *range,
 {
 	struct xe_device *xe = vm->xe;
 
+	range_debug(range, "GARBAGE COLLECTOR ADD");
+
 	drm_gpusvm_range_set_unmapped(&range->base, mmu_range);
 
 	spin_lock(&vm->svm.garbage_collector.lock);
@@ -84,10 +115,14 @@ xe_svm_range_notifier_event_begin(struct xe_vm *vm, struct drm_gpusvm_range *r,
 
 	xe_svm_assert_in_notifier(vm);
 
+	range_debug(range, "NOTIFIER");
+
 	/* Skip if already unmapped or if no binding exist */
 	if (range->base.flags.unmapped || !range->tile_present)
 		return 0;
 
+	range_debug(range, "NOTIFIER - EXECUTE");
+
 	/* Adjust invalidation to range boundaries */
 	if (range->base.itree.start < mmu_range->start)
 		*adj_start = range->base.itree.start;
@@ -140,6 +175,11 @@ static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
 
 	xe_svm_assert_in_notifier(vm);
 
+	vm_dbg(&gpusvm_to_vm(gpusvm)->xe->drm,
+	       "INVALIDATE: asid=%u, gpusvm=0x%016llx, seqno=%lu, start=0x%016lx, end=0x%016lx, event=%d",
+	       vm->usm.asid, (u64)gpusvm, notifier->notifier.invalidate_seq,
+	       mmu_range->start, mmu_range->end, mmu_range->event);
+
 	/* Adjust invalidation to notifier boundaries */
 	if (adj_start < notifier->itree.start)
 		adj_start = notifier->itree.start;
@@ -226,8 +266,12 @@ static int __xe_svm_garbage_collector(struct xe_vm *vm,
 {
 	struct dma_fence *fence;
 
-	if (IS_DGFX(vm->xe) && range->base.flags.partial_unmap)
+	range_debug(range, "GARBAGE COLLECTOR");
+
+	if (IS_DGFX(vm->xe) && range->base.flags.partial_unmap) {
+		range_debug(range, "GARBAGE COLLECTOR - EVICT");
 		drm_gpusvm_range_evict(&vm->svm.gpusvm, &range->base);
+	}
 
 	xe_vm_lock(vm, false);
 	fence = xe_vm_range_unbind(vm, range);
@@ -387,16 +431,23 @@ static int xe_svm_copy(struct page **pages, dma_addr_t *dma_addr,
 			int incr = (match && last) ? 1 : 0;
 
 			if (vram_addr != XE_VRAM_ADDR_INVALID) {
-				if (sram)
+				if (sram) {
+					vm_dbg(&tile->xe->drm,
+					       "COPY TO SRAM - 0x%016llx -> 0x%016llx, NPAGES=%ld",
+					       vram_addr, dma_addr[pos], i - pos + incr);
 					__fence = xe_migrate_from_vram(tile->migrate,
 								       i - pos + incr,
 								       vram_addr,
 								       dma_addr + pos);
-				else
+				} else {
+					vm_dbg(&tile->xe->drm,
+					       "COPY TO VRAM - 0x%016llx -> 0x%016llx, NPAGES=%ld",
+					       dma_addr[pos], vram_addr, i - pos + incr);
 					__fence = xe_migrate_to_vram(tile->migrate,
 								     i - pos + incr,
 								     dma_addr + pos,
 								     vram_addr);
+				}
 				if (IS_ERR(__fence)) {
 					err = PTR_ERR(__fence);
 					goto err_out;
@@ -416,14 +467,21 @@ static int xe_svm_copy(struct page **pages, dma_addr_t *dma_addr,
 
 			/* Extra mismatched device page, copy it */
 			if (!match && last && vram_addr != XE_VRAM_ADDR_INVALID) {
-				if (sram)
+				if (sram) {
+					vm_dbg(&tile->xe->drm,
+					       "COPY TO SRAM - 0x%016llx -> 0x%016llx, NPAGES=%d",
+					       vram_addr, dma_addr[pos], 1);
 					__fence = xe_migrate_from_vram(tile->migrate, 1,
 								       vram_addr,
 								       dma_addr + pos);
-				else
+				} else {
+					vm_dbg(&tile->xe->drm,
+					       "COPY TO VRAM - 0x%016llx -> 0x%016llx, NPAGES=%d",
+					       dma_addr[pos], vram_addr, 1);
 					__fence = xe_migrate_to_vram(tile->migrate, 1,
 								     dma_addr + pos,
 								     vram_addr);
+				}
 				if (IS_ERR(__fence)) {
 					err = PTR_ERR(__fence);
 					goto err_out;
@@ -591,12 +649,14 @@ static struct xe_bo *xe_svm_alloc_vram(struct xe_vm *vm, struct xe_tile *tile,
 				       const struct drm_gpusvm_ctx *ctx)
 {
 	struct xe_mem_region *mr = tile_to_mr(tile);
+	struct drm_buddy *buddy = tile_to_buddy(tile);
 	struct drm_buddy_block *block;
 	struct list_head *blocks;
 	struct xe_bo *bo;
 	ktime_t end = 0;
 	int err;
 
+	range_debug(range, "ALLOCATE VRAM");
 retry:
 	xe_vm_lock(vm, false);
 	bo = xe_bo_create(tile_to_xe(tile), tile, vm, range->base.itree.last + 1 -
@@ -619,8 +679,13 @@ static struct xe_bo *xe_svm_alloc_vram(struct xe_vm *vm, struct xe_tile *tile,
 			       range->base.itree.start);
 
 	blocks = &to_xe_ttm_vram_mgr_resource(bo->ttm.resource)->blocks;
-	list_for_each_entry(block, blocks, link)
+	list_for_each_entry(block, blocks, link) {
+		vm_dbg(&vm->xe->drm, "ALLOC VRAM: asid=%u, gpusvm=0x%016llx, pfn=%llu, npages=%llu",
+		       vm->usm.asid, (u64)&vm->svm.gpusvm,
+		       block_offset_to_pfn(mr, drm_buddy_block_offset(block)),
+		       drm_buddy_block_size(buddy, block) >> PAGE_SHIFT);
 		block->private = mr;
+	}
 
 	/*
 	 * Take ref because as soon as drm_gpusvm_migrate_to_devmem succeeds the
@@ -693,6 +758,8 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	if (xe_svm_range_is_valid(range, tile))
 		return 0;
 
+	range_debug(range, "PAGE FAULT");
+
 	/* XXX: Add migration policy, for now migrate range once */
 	if (!range->migrated && range->base.flags.migrate_devmem &&
 	    (range->base.itree.last + 1 - range->base.itree.start) >= SZ_64K) {
@@ -708,18 +775,26 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 		}
 	}
 
+	range_debug(range, "GET PAGES");
 	err = drm_gpusvm_range_get_pages(&vm->svm.gpusvm, r, &ctx);
 	/* Corner where CPU mappings have changed */
 	if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {
-		if (err == -EOPNOTSUPP)
+		if (err == -EOPNOTSUPP) {
+			range_debug(range, "PAGE FAULT - EVICT PAGES");
 			drm_gpusvm_range_evict(&vm->svm.gpusvm, &range->base);
+		}
 		drm_info(&vm->xe->drm,
 			 "Get pages failed, falling back to retrying, asid=%u, gpusvm=0x%016llx, errno %d\n",
 			 vm->usm.asid, (u64)&vm->svm.gpusvm, err);
+		range_debug(range, "PAGE FAULT - RETRY PAGES");
 		goto retry;
 	}
-	if (err)
+	if (err) {
+		range_debug(range, "PAGE FAULT - FAIL PAGE COLLECT");
 		goto err_out;
+	}
+
+	range_debug(range, "PAGE FAULT - BIND");
 
 retry_bind:
 	drm_exec_init(&exec, 0, 0);
@@ -735,8 +810,10 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 		if (IS_ERR(fence)) {
 			drm_exec_fini(&exec);
 			err = PTR_ERR(fence);
-			if (err == -EAGAIN)
+			if (err == -EAGAIN) {
+				range_debug(range, "PAGE FAULT - RETRY BIND");
 				goto retry;
+			}
 			if (xe_vm_validate_should_retry(&exec, err, &end))
 				goto retry_bind;
 			goto err_out;
diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
index 9e9d333bb9d3..f9751fb2f0ea 100644
--- a/drivers/gpu/drm/xe/xe_svm.h
+++ b/drivers/gpu/drm/xe/xe_svm.h
@@ -56,6 +56,8 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 
 bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end);
 
+void xe_svm_range_debug(struct xe_svm_range *range, const char *operation);
+
 int xe_svm_bo_evict(struct xe_bo *bo);
 
 static inline bool xe_svm_range_pages_valid(struct xe_svm_range *range)
-- 
2.34.1

[PATCH v3 29/30] drm/xe: Add modparam for SVM notifier size

[Matthew Brost]

Useful to experiment with notifier size and how it affects performance.

v3:
 - Pull missing changes including in following patch (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_module.c | 4 ++++
 drivers/gpu/drm/xe/xe_module.h | 1 +
 drivers/gpu/drm/xe/xe_svm.c    | 4 +++-
 3 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_module.c b/drivers/gpu/drm/xe/xe_module.c
index 0f2c20e9204a..2126e99ede01 100644
--- a/drivers/gpu/drm/xe/xe_module.c
+++ b/drivers/gpu/drm/xe/xe_module.c
@@ -25,9 +25,13 @@ struct xe_modparam xe_modparam = {
 	.max_vfs = IS_ENABLED(CONFIG_DRM_XE_DEBUG) ? ~0 : 0,
 #endif
 	.wedged_mode = 1,
+	.svm_notifier_size = 512,
 	/* the rest are 0 by default */
 };
 
+module_param_named(svm_notifier_size, xe_modparam.svm_notifier_size, uint, 0600);
+MODULE_PARM_DESC(svm_notifier_size, "Set the svm notifier size(in MiB), must be pow2");
+
 module_param_named_unsafe(force_execlist, xe_modparam.force_execlist, bool, 0444);
 MODULE_PARM_DESC(force_execlist, "Force Execlist submission");
 
diff --git a/drivers/gpu/drm/xe/xe_module.h b/drivers/gpu/drm/xe/xe_module.h
index 161a5e6f717f..5a3bfea8b7b4 100644
--- a/drivers/gpu/drm/xe/xe_module.h
+++ b/drivers/gpu/drm/xe/xe_module.h
@@ -22,6 +22,7 @@ struct xe_modparam {
 	unsigned int max_vfs;
 #endif
 	int wedged_mode;
+	u32 svm_notifier_size;
 };
 
 extern struct xe_modparam xe_modparam;
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 530ba34a86dc..9586c711386f 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -6,6 +6,7 @@
 #include "xe_bo.h"
 #include "xe_gt_tlb_invalidation.h"
 #include "xe_migrate.h"
+#include "xe_module.h"
 #include "xe_pt.h"
 #include "xe_svm.h"
 #include "xe_ttm_vram_mgr.h"
@@ -598,7 +599,8 @@ int xe_svm_init(struct xe_vm *vm)
 
 	err = drm_gpusvm_init(&vm->svm.gpusvm, "Xe SVM", &vm->xe->drm,
 			      current->mm, xe_svm_devm_owner(vm->xe), 0,
-			      vm->size, SZ_512M, &gpusvm_ops, fault_chunk_sizes,
+			      vm->size, xe_modparam.svm_notifier_size * SZ_1M,
+			      &gpusvm_ops, fault_chunk_sizes,
 			      ARRAY_SIZE(fault_chunk_sizes));
 	if (err)
 		return err;
-- 
2.34.1

[PATCH v3 30/30] drm/xe: Add always_migrate_to_vram modparam

[Matthew Brost]

Used to show we can bounce memory multiple times which will happen once
a real migration policy is implemented. Can be removed once migration
policy is implemented.

v3:
 - Pull some changes into the previous patch (Thomas)
 - Spell out power of 2 (Thomas)
 - Better commit message (Thomas)

Signed-off-by: Matthew Brost <matthew.brost@intel.com>
---
 drivers/gpu/drm/xe/xe_module.c | 5 ++++-
 drivers/gpu/drm/xe/xe_module.h | 1 +
 drivers/gpu/drm/xe/xe_svm.c    | 3 +++
 3 files changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/gpu/drm/xe/xe_module.c b/drivers/gpu/drm/xe/xe_module.c
index 2126e99ede01..192047b3419b 100644
--- a/drivers/gpu/drm/xe/xe_module.c
+++ b/drivers/gpu/drm/xe/xe_module.c
@@ -30,7 +30,10 @@ struct xe_modparam xe_modparam = {
 };
 
 module_param_named(svm_notifier_size, xe_modparam.svm_notifier_size, uint, 0600);
-MODULE_PARM_DESC(svm_notifier_size, "Set the svm notifier size(in MiB), must be pow2");
+MODULE_PARM_DESC(svm_notifier_size, "Set the svm notifier size(in MiB), must be power of 2");
+
+module_param_named(always_migrate_to_vram, xe_modparam.always_migrate_to_vram, bool, 0444);
+MODULE_PARM_DESC(always_migrate_to_vram, "Always migrate to VRAM on GPU fault");
 
 module_param_named_unsafe(force_execlist, xe_modparam.force_execlist, bool, 0444);
 MODULE_PARM_DESC(force_execlist, "Force Execlist submission");
diff --git a/drivers/gpu/drm/xe/xe_module.h b/drivers/gpu/drm/xe/xe_module.h
index 5a3bfea8b7b4..84339e509c80 100644
--- a/drivers/gpu/drm/xe/xe_module.h
+++ b/drivers/gpu/drm/xe/xe_module.h
@@ -12,6 +12,7 @@
 struct xe_modparam {
 	bool force_execlist;
 	bool probe_display;
+	bool always_migrate_to_vram;
 	u32 force_vram_bar_size;
 	int guc_log_level;
 	char *guc_firmware_path;
diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
index 9586c711386f..cc5998903469 100644
--- a/drivers/gpu/drm/xe/xe_svm.c
+++ b/drivers/gpu/drm/xe/xe_svm.c
@@ -823,6 +823,9 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
 	}
 	drm_exec_fini(&exec);
 
+	if (xe_modparam.always_migrate_to_vram)
+		range->migrated = false;
+
 	dma_fence_wait(fence, false);
 	dma_fence_put(fence);
 
-- 
2.34.1

✓ CI.Patch_applied: success for Introduce GPU SVM and Xe SVM implementation (rev3)

[Patchwork]

== Series Details ==

Series: Introduce GPU SVM and Xe SVM implementation (rev3)
URL   : https://patchwork.freedesktop.org/series/137870/
State : success

== Summary ==

=== Applying kernel patches on branch 'drm-tip' with base: ===
Base commit: 69607bbcaccc drm-tip: 2024y-12m-17d-21h-13m-30s UTC integration manifest
=== git am output follows ===
Applying: drm/xe: Retry BO allocation
Applying: mm/migrate: Add migrate_device_pfns
Applying: mm/migrate: Trylock device page in do_swap_page
Applying: drm/pagemap: Add DRM pagemap
Applying: drm/gpusvm: Add support for GPU Shared Virtual Memory
Applying: drm/xe: Select DRM_GPUSVM Kconfig
Applying: drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag
Applying: drm/xe: Add SVM init / close / fini to faulting VMs
Applying: drm/xe: Add dma_addr res cursor
Applying: drm/xe: Nuke VM's mapping upon close
Applying: drm/xe: Add SVM range invalidation and page fault handler
Applying: drm/gpuvm: Add DRM_GPUVA_OP_DRIVER
Applying: drm/xe: Add (re)bind to SVM page fault handler
Applying: drm/xe: Add SVM garbage collector
Applying: drm/xe: Add unbind to SVM garbage collector
Applying: drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has bindings
Applying: drm/xe: Enable CPU address mirror uAPI
Applying: drm/xe: Add migrate layer functions for SVM support
Applying: drm/xe: Add SVM device memory mirroring
Applying: drm/xe: Add drm_gpusvm_devmem to xe_bo
Applying: drm/xe: Add drm_pagemap ops to SVM
Applying: drm/xe: Add GPUSVM device memory copy vfunc functions
Applying: drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc
Applying: drm/xe: Add Xe SVM devmem_release GPU SVM vfunc
Applying: drm/xe: Add BO flags required for SVM
Applying: drm/xe: Add SVM VRAM migration
Applying: drm/xe: Basic SVM BO eviction
Applying: drm/xe: Add SVM debug
Applying: drm/xe: Add modparam for SVM notifier size
Applying: drm/xe: Add always_migrate_to_vram modparam

✗ CI.checkpatch: warning for Introduce GPU SVM and Xe SVM implementation (rev3)

[Patchwork]

== Series Details ==

Series: Introduce GPU SVM and Xe SVM implementation (rev3)
URL   : https://patchwork.freedesktop.org/series/137870/
State : warning

== Summary ==

+ KERNEL=/kernel
+ git clone https://gitlab.freedesktop.org/drm/maintainer-tools mt
Cloning into 'mt'...
warning: redirecting to https://gitlab.freedesktop.org/drm/maintainer-tools.git/
+ git -C mt rev-list -n1 origin/master
30ab6715fc09baee6cc14cb3c89ad8858688d474
+ cd /kernel
+ git config --global --add safe.directory /kernel
+ git log -n1
commit 631a11cb5d11f3ea9a8b9129fcd507d13c012b55
Author: Matthew Brost <matthew.brost@intel.com>
Date:   Tue Dec 17 15:33:48 2024 -0800

    drm/xe: Add always_migrate_to_vram modparam
    
    Used to show we can bounce memory multiple times which will happen once
    a real migration policy is implemented. Can be removed once migration
    policy is implemented.
    
    v3:
     - Pull some changes into the previous patch (Thomas)
     - Spell out power of 2 (Thomas)
     - Better commit message (Thomas)
    
    Signed-off-by: Matthew Brost <matthew.brost@intel.com>
+ /mt/dim checkpatch 69607bbcaccc29ae254d974a5e547d873e7c0ac0 drm-intel
8d81d4e9325b drm/xe: Retry BO allocation
213054c066a8 mm/migrate: Add migrate_device_pfns
be63b3df8327 mm/migrate: Trylock device page in do_swap_page
-:212: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis
#212: FILE: mm/migrate_device.c:883:
+void migrate_device_finalize(unsigned long *src_pfns,
+			unsigned long *dst_pfns, unsigned long npages)

total: 0 errors, 0 warnings, 1 checks, 172 lines checked
d19b1014cd49 drm/pagemap: Add DRM pagemap
-:24: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating?
#24: 
new file mode 100644

total: 0 errors, 1 warnings, 0 checks, 103 lines checked
e16e8523ac82 drm/gpusvm: Add support for GPU Shared Virtual Memory
-:58: WARNING:COMMIT_LOG_LONG_LINE: Prefer a maximum 75 chars per line (possible unwrapped commit description?)
#58: 
 - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)

-:106: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating?
#106: 
new file mode 100644

-:302: WARNING:LONG_LINE_COMMENT: line length of 103 exceeds 100 columns
#302: FILE: drivers/gpu/drm/drm_gpusvm.c:192:
+ *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed

-:527: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'range__' - possible side-effects?
#527: FILE: drivers/gpu/drm/drm_gpusvm.c:417:
+#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
+	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
+	     (next__) = __drm_gpusvm_range_next(range__);				\
+	     (range__) && (range__->itree.start < (end__));				\
+	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))

-:527: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'next__' - possible side-effects?
#527: FILE: drivers/gpu/drm/drm_gpusvm.c:417:
+#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
+	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
+	     (next__) = __drm_gpusvm_range_next(range__);				\
+	     (range__) && (range__->itree.start < (end__));				\
+	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))

-:527: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'end__' - possible side-effects?
#527: FILE: drivers/gpu/drm/drm_gpusvm.c:417:
+#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
+	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
+	     (next__) = __drm_gpusvm_range_next(range__);				\
+	     (range__) && (range__->itree.start < (end__));				\
+	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))

-:574: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'notifier__' - possible side-effects?
#574: FILE: drivers/gpu/drm/drm_gpusvm.c:464:
+#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))

-:574: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'end__' - possible side-effects?
#574: FILE: drivers/gpu/drm/drm_gpusvm.c:464:
+#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))

-:590: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'notifier__' - possible side-effects?
#590: FILE: drivers/gpu/drm/drm_gpusvm.c:480:
+#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
+	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))

-:590: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'next__' - possible side-effects?
#590: FILE: drivers/gpu/drm/drm_gpusvm.c:480:
+#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
+	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))

-:590: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'end__' - possible side-effects?
#590: FILE: drivers/gpu/drm/drm_gpusvm.c:480:
+#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
+	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
+	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
+	     (notifier__) && (notifier__->itree.start < (end__));			\
+	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))

-:1666: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis
#1666: FILE: drivers/gpu/drm/drm_gpusvm.c:1556:
+static void drm_gpusvm_get_devmem_page(struct page *page,
+				     struct drm_gpusvm_zdd *zdd)

-:2713: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'gpusvm' - possible side-effects?
#2713: FILE: include/drm/drm_gpusvm.h:377:
+#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
+	do { \
+		if (!WARN((gpusvm)->lock_dep_map, \
+			  "GPUSVM range lock should be set only once."))\
+			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
+	} while (0)

-:2719: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'gpusvm' - possible side-effects?
#2719: FILE: include/drm/drm_gpusvm.h:383:
+#define drm_gpusvm_driver_lock_held(gpusvm) \
+	do { \
+		if ((gpusvm)->lock_dep_map)	\
+			lock_is_held((gpusvm)->lock_dep_map);	\
+	} while (0)

-:2725: WARNING:MACRO_ARG_UNUSED: Argument 'gpusvm' is not used in function-like macro
#2725: FILE: include/drm/drm_gpusvm.h:389:
+#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)

-:2725: WARNING:MACRO_ARG_UNUSED: Argument 'lock' is not used in function-like macro
#2725: FILE: include/drm/drm_gpusvm.h:389:
+#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)

-:2726: WARNING:MACRO_ARG_UNUSED: Argument 'gpusvm' is not used in function-like macro
#2726: FILE: include/drm/drm_gpusvm.h:390:
+#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)

-:2775: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'range__' - possible side-effects?
#2775: FILE: include/drm/drm_gpusvm.h:439:
+#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
+	for ((range__) = (range__) ?:					\
+	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
+	     (range__) && (range__->itree.start < (end__));		\
+	     (range__) = __drm_gpusvm_range_next(range__))

-:2775: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'end__' - possible side-effects?
#2775: FILE: include/drm/drm_gpusvm.h:439:
+#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
+	for ((range__) = (range__) ?:					\
+	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
+	     (range__) && (range__->itree.start < (end__));		\
+	     (range__) = __drm_gpusvm_range_next(range__))

total: 0 errors, 6 warnings, 13 checks, 2686 lines checked
de6d65928e70 drm/xe: Select DRM_GPUSVM Kconfig
3f7c7cf5672c drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag
-:48: WARNING:COMMIT_LOG_LONG_LINE: Prefer a maximum 75 chars per line (possible unwrapped commit description?)
#48: 
 - s/DRM_XE_VM_BIND_FLAG_SYSTEM_ALLOCATOR/DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR (Thomas)

-:503: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis
#503: FILE: drivers/gpu/drm/xe/xe_vm.c:2828:
+		    XE_IOCTL_DBG(xe, obj_offset && (is_null ||
+				 is_cpu_addr_mirror)) ||

total: 0 errors, 1 warnings, 1 checks, 559 lines checked
b3765a07ecd6 drm/xe: Add SVM init / close / fini to faulting VMs
-:34: WARNING:FILE_PATH_CHANGES: added, moved or deleted file(s), does MAINTAINERS need updating?
#34: 
new file mode 100644

total: 0 errors, 1 warnings, 0 checks, 152 lines checked
83b085e62f12 drm/xe: Add dma_addr res cursor
d1a713d05f4a drm/xe: Nuke VM's mapping upon close
5a576b0fc97c drm/xe: Add SVM range invalidation and page fault handler
-:444: CHECK:MACRO_ARG_PRECEDENCE: Macro argument 'vm' may be better as '(vm)' to avoid precedence issues
#444: FILE: drivers/gpu/drm/xe/xe_svm.h:44:
+#define xe_svm_assert_in_notifier(vm) \
+	lockdep_assert_held_write(&vm->svm.gpusvm.notifier_lock)

total: 0 errors, 0 warnings, 1 checks, 414 lines checked
fa60830fdd44 drm/gpuvm: Add DRM_GPUVA_OP_DRIVER
7eb37f7e2bb0 drm/xe: Add (re)bind to SVM page fault handler
2173bac33e8f drm/xe: Add SVM garbage collector
273c0ce75c6d drm/xe: Add unbind to SVM garbage collector
758c590ea9a9 drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has bindings
2a6468453c4a drm/xe: Enable CPU address mirror uAPI
a0247119f7a9 drm/xe: Add migrate layer functions for SVM support
bc6859d4b8b9 drm/xe: Add SVM device memory mirroring
-:26: WARNING:CONFIG_DESCRIPTION: please write a help paragraph that fully describes the config symbol
#26: FILE: drivers/gpu/drm/xe/Kconfig:63:
+config CONFIG_XE_DEVMEM_MIRROR
+	bool "Enable device memory mirror"
+	depends on DRM_XE
+	select CONFIG_GET_FREE_REGION
+	default y
+	help
+	  Disable this option only if you want to compile out without device
+	  memory mirror. Will reduce KMD memory footprint when disabled.
+

-:126: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis
#126: FILE: drivers/gpu/drm/xe/xe_svm.c:488:
+		drm_err(&xe->drm, "Failed to remap tile %d memory, errno %d\n",
+				tile->id, ret);

total: 0 errors, 1 warnings, 1 checks, 144 lines checked
410317a6f4d1 drm/xe: Add drm_gpusvm_devmem to xe_bo
2e11e0a73b4c drm/xe: Add drm_pagemap ops to SVM
-:32: CHECK:LINE_SPACING: Please don't use multiple blank lines
#32: FILE: drivers/gpu/drm/xe/xe_device_types.h:16:
 
+

total: 0 errors, 0 warnings, 1 checks, 63 lines checked
eac9baf0d856 drm/xe: Add GPUSVM device memory copy vfunc functions
03b28bc36cf7 drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc
96aeeebec17f drm/xe: Add Xe SVM devmem_release GPU SVM vfunc
6ced5f3a86b9 drm/xe: Add BO flags required for SVM
-:62: CHECK:PARENTHESIS_ALIGNMENT: Alignment should match open parenthesis
#62: FILE: drivers/gpu/drm/xe/xe_bo.c:2473:
+	if (IS_DGFX(xe) && ((bo->flags & XE_BO_FLAG_SYSTEM) ||
+	    (bo->flags & XE_BO_FLAG_CPU_ADDR_MIRROR)))

total: 0 errors, 0 warnings, 1 checks, 44 lines checked
b8be9d03aa5d drm/xe: Add SVM VRAM migration
fca5c34367c4 drm/xe: Basic SVM BO eviction
ebbd217308b0 drm/xe: Add SVM debug
-:86: CHECK:MACRO_ARG_REUSE: Macro argument reuse 'r__' - possible side-effects?
#86: FILE: drivers/gpu/drm/xe/xe_svm.c:37:
+#define range_debug(r__, operaton__)					\
+	vm_dbg(&range_to_vm(&(r__)->base)->xe->drm,			\
+	       "%s: asid=%u, gpusvm=0x%016llx, vram=%d,%d, seqno=%lu, " \
+	       "start=0x%014lx, end=0x%014lx, size=%lu",		\
+	       (operaton__), range_to_vm(&(r__)->base)->usm.asid,	\
+	       (u64)(r__)->base.gpusvm,					\
+	       xe_svm_range_in_vram((r__)) ? 1 : 0,			\
+	       xe_svm_range_has_vram_binding((r__)) ? 1 : 0,		\
+	       (r__)->base.notifier_seq,				\
+	       (r__)->base.itree.start, (r__)->base.itree.last + 1,	\
+	       (r__)->base.itree.last + 1 - (r__)->base.itree.start)

total: 0 errors, 0 warnings, 1 checks, 248 lines checked
617d19fda52a drm/xe: Add modparam for SVM notifier size
631a11cb5d11 drm/xe: Add always_migrate_to_vram modparam

✗ CI.KUnit: failure for Introduce GPU SVM and Xe SVM implementation (rev3)

[Patchwork]

== Series Details ==

Series: Introduce GPU SVM and Xe SVM implementation (rev3)
URL   : https://patchwork.freedesktop.org/series/137870/
State : failure

== Summary ==

+ trap cleanup EXIT
+ /kernel/tools/testing/kunit/kunit.py run --kunitconfig /kernel/drivers/gpu/drm/xe/.kunitconfig
ERROR:root:../drivers/gpu/drm/drm_gpusvm.c: In function ‘drm_gpusvm_get_devmem_page’:
../drivers/gpu/drm/drm_gpusvm.c:1559:9: error: implicit declaration of function ‘zone_device_page_init’ [-Werror=implicit-function-declaration]
 1559 |         zone_device_page_init(page);
      |         ^~~~~~~~~~~~~~~~~~~~~
../lib/iomap.c:156:5: warning: no previous prototype for ‘ioread64_lo_hi’ [-Wmissing-prototypes]
  156 | u64 ioread64_lo_hi(const void __iomem *addr)
      |     ^~~~~~~~~~~~~~
../lib/iomap.c:163:5: warning: no previous prototype for ‘ioread64_hi_lo’ [-Wmissing-prototypes]
  163 | u64 ioread64_hi_lo(const void __iomem *addr)
      |     ^~~~~~~~~~~~~~
../lib/iomap.c:170:5: warning: no previous prototype for ‘ioread64be_lo_hi’ [-Wmissing-prototypes]
  170 | u64 ioread64be_lo_hi(const void __iomem *addr)
      |     ^~~~~~~~~~~~~~~~
../lib/iomap.c:178:5: warning: no previous prototype for ‘ioread64be_hi_lo’ [-Wmissing-prototypes]
  178 | u64 ioread64be_hi_lo(const void __iomem *addr)
      |     ^~~~~~~~~~~~~~~~
../lib/iomap.c:264:6: warning: no previous prototype for ‘iowrite64_lo_hi’ [-Wmissing-prototypes]
  264 | void iowrite64_lo_hi(u64 val, void __iomem *addr)
      |      ^~~~~~~~~~~~~~~
../lib/iomap.c:272:6: warning: no previous prototype for ‘iowrite64_hi_lo’ [-Wmissing-prototypes]
  272 | void iowrite64_hi_lo(u64 val, void __iomem *addr)
      |      ^~~~~~~~~~~~~~~
../lib/iomap.c:280:6: warning: no previous prototype for ‘iowrite64be_lo_hi’ [-Wmissing-prototypes]
  280 | void iowrite64be_lo_hi(u64 val, void __iomem *addr)
      |      ^~~~~~~~~~~~~~~~~
../lib/iomap.c:288:6: warning: no previous prototype for ‘iowrite64be_hi_lo’ [-Wmissing-prototypes]
  288 | void iowrite64be_hi_lo(u64 val, void __iomem *addr)
      |      ^~~~~~~~~~~~~~~~~
cc1: some warnings being treated as errors
make[6]: *** [../scripts/Makefile.build:194: drivers/gpu/drm/drm_gpusvm.o] Error 1
make[6]: *** Waiting for unfinished jobs....
make[5]: *** [../scripts/Makefile.build:440: drivers/gpu/drm] Error 2
make[4]: *** [../scripts/Makefile.build:440: drivers/gpu] Error 2
make[3]: *** [../scripts/Makefile.build:440: drivers] Error 2
make[2]: *** [/kernel/Makefile:1989: .] Error 2
make[1]: *** [/kernel/Makefile:251: __sub-make] Error 2
make: *** [Makefile:251: __sub-make] Error 2

[04:08:41] Configuring KUnit Kernel ...
Generating .config ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig
[04:08:45] Building KUnit Kernel ...
Populating config with:
$ make ARCH=um O=.kunit olddefconfig
Building with:
$ make all compile_commands.json ARCH=um O=.kunit --jobs=48
+ cleanup
++ stat -c %u:%g /kernel
+ chown -R 1003:1003 /kernel

Re: [PATCH v3 27/30] drm/xe: Basic SVM BO eviction

[Matthew Brost]

On Tue, Dec 17, 2024 at 03:33:45PM -0800, Matthew Brost wrote:
> Wire xe_bo_move to GPU SVM migration via new helper xe_svm_bo_evict.
> 

Somehow lost the xe_bo.c changes in this rev which call xe_svm_bo_evict.
Ignore this patch.

Matt

> v2:
>  - Use xe_svm_bo_evict
>  - Drop bo->range
> v3:
>  - Kernel doc (Thomas)
> 
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_svm.c | 14 ++++++++++++++
>  drivers/gpu/drm/xe/xe_svm.h |  2 ++
>  2 files changed, 16 insertions(+)
> 
> diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
> index a417d8942da4..8780a0b2c81e 100644
> --- a/drivers/gpu/drm/xe/xe_svm.c
> +++ b/drivers/gpu/drm/xe/xe_svm.c
> @@ -768,6 +768,20 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
>  	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
>  }
>  
> +/**
> + * xe_svm_bo_evict() - SVM evict BO to system memory
> + * @bo: BO to evict
> + *
> + * SVM evict BO to system memory. GPU SVM layer ensures all device pages
> + * are evicted before returning.
> + *
> + * Return: 0 on success standard error code otherwise
> + */
> +int xe_svm_bo_evict(struct xe_bo *bo)
> +{
> +	return drm_gpusvm_evict_to_ram(&bo->devmem_allocation);
> +}
> +
>  #if IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR)
>  static struct drm_pagemap_dma_addr
>  xe_drm_pagemap_map_dma(struct drm_pagemap *dpagemap,
> diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
> index d549dd9e8641..9e9d333bb9d3 100644
> --- a/drivers/gpu/drm/xe/xe_svm.h
> +++ b/drivers/gpu/drm/xe/xe_svm.h
> @@ -56,6 +56,8 @@ int xe_svm_handle_pagefault(struct xe_vm *vm, struct xe_vma *vma,
>  
>  bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end);
>  
> +int xe_svm_bo_evict(struct xe_bo *bo);
> +
>  static inline bool xe_svm_range_pages_valid(struct xe_svm_range *range)
>  {
>  	return drm_gpusvm_range_pages_valid(range->base.gpusvm, &range->base);
> -- 
> 2.34.1
> 

Re: [PATCH v3 19/30] drm/xe: Add SVM device memory mirroring

[Ghimiray, Himal Prasad]

On 18-12-2024 05:03, Matthew Brost wrote:
> Add SVM device memory mirroring which enables device pages for
> migration. Enabled via CONFIG_XE_DEVMEM_MIRROR Kconfig. Kconfig option
> defaults to enabled. If not enabled, SVM will work sans migration and
> KMD memory footprint will be less.
> 
> v3:
>   - Add CONFIG_XE_DEVMEM_MIRROR
> 
> Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
> Signed-off-by: Oak Zeng <oak.zeng@intel.com>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>   drivers/gpu/drm/xe/Kconfig           |  9 ++++
>   drivers/gpu/drm/xe/xe_device_types.h |  8 ++++
>   drivers/gpu/drm/xe/xe_svm.c          | 62 +++++++++++++++++++++++++++-
>   drivers/gpu/drm/xe/xe_svm.h          |  3 ++
>   drivers/gpu/drm/xe/xe_tile.c         |  5 +++
>   5 files changed, 85 insertions(+), 2 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/Kconfig b/drivers/gpu/drm/xe/Kconfig
> index 3a08e16bfada..1989c99a9f5a 100644
> --- a/drivers/gpu/drm/xe/Kconfig
> +++ b/drivers/gpu/drm/xe/Kconfig
> @@ -60,6 +60,15 @@ config DRM_XE_DISPLAY
>   	help
>   	  Disable this option only if you want to compile out display support.
>   
> +config CONFIG_XE_DEVMEM_MIRROR


%s/CONFIG_XE_DEVMEM_MIRROR/XE_DEVMEM_MIRROR


> +	bool "Enable device memory mirror"
> +	depends on DRM_XE
> +	select CONFIG_GET_FREE_REGION
> +	default y
> +	help
> +	  Disable this option only if you want to compile out without device
> +	  memory mirror. Will reduce KMD memory footprint when disabled.
> +
>   config DRM_XE_FORCE_PROBE
>   	string "Force probe xe for selected Intel hardware IDs"
>   	depends on DRM_XE
> diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
> index 8a7b15972413..b949a960cebb 100644
> --- a/drivers/gpu/drm/xe/xe_device_types.h
> +++ b/drivers/gpu/drm/xe/xe_device_types.h
> @@ -102,6 +102,14 @@ struct xe_mem_region {
>   	resource_size_t actual_physical_size;
>   	/** @mapping: pointer to VRAM mappable space */
>   	void __iomem *mapping;
> +	/** @pagemap: Used to remap device memory as ZONE_DEVICE */
> +	struct dev_pagemap pagemap;
> +	/**
> +	 * @hpa_base: base host physical address
> +	 *
> +	 * This is generated when remap device memory as ZONE_DEVICE
> +	 */
> +	resource_size_t hpa_base;
>   };
>   
>   /**
> diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
> index 3cbf29732249..2860e54bb035 100644
> --- a/drivers/gpu/drm/xe/xe_svm.c
> +++ b/drivers/gpu/drm/xe/xe_svm.c
> @@ -19,6 +19,11 @@ static struct xe_vm *range_to_vm(struct drm_gpusvm_range *r)
>   	return gpusvm_to_vm(r->gpusvm);
>   }
>   
> +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)
>   {
> @@ -306,8 +311,8 @@ 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, NULL, 0, vm->size,
> -			      SZ_512M, &gpusvm_ops, fault_chunk_sizes,
> +			      current->mm, xe_svm_devm_owner(vm->xe), 0,
> +			      vm->size, SZ_512M, &gpusvm_ops, fault_chunk_sizes,
>   			      ARRAY_SIZE(fault_chunk_sizes));
>   	if (err)
>   		return err;
> @@ -442,3 +447,56 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
>   {
>   	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
>   }
> +
> +#if IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR)
> +/**
> + * xe_devm_add: Remap and provide memmap backing for device memory
> + * @tile: tile that the memory region belongs to
> + * @mr: memory region to remap
> + *
> + * This remap device memory to host physical address space and create
> + * struct page to back device memory
> + *
> + * Return: 0 on success standard error code otherwise
> + */
> +int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
> +{
> +	struct xe_device *xe = tile_to_xe(tile);
> +	struct device *dev = &to_pci_dev(xe->drm.dev)->dev;
> +	struct resource *res;
> +	void *addr;
> +	int ret;
> +
> +	res = devm_request_free_mem_region(dev, &iomem_resource,
> +					   mr->usable_size);
> +	if (IS_ERR(res)) {
> +		ret = PTR_ERR(res);
> +		return ret;
> +	}
> +
> +	mr->pagemap.type = MEMORY_DEVICE_PRIVATE;
> +	mr->pagemap.range.start = res->start;
> +	mr->pagemap.range.end = res->end;
> +	mr->pagemap.nr_range = 1;
> +	mr->pagemap.ops = drm_gpusvm_pagemap_ops_get();
> +	mr->pagemap.owner = xe_svm_devm_owner(xe);
> +	addr = devm_memremap_pages(dev, &mr->pagemap);
> +	if (IS_ERR(addr)) {
> +		devm_release_mem_region(dev, res->start, resource_size(res));
> +		ret = PTR_ERR(addr);
> +		drm_err(&xe->drm, "Failed to remap tile %d memory, errno %d\n",
> +				tile->id, ret);
> +		return ret;
> +	}
> +	mr->hpa_base = res->start;
> +
> +	drm_info(&xe->drm, "Added tile %d memory [%llx-%llx] to devm, remapped to %pr\n",
> +		 tile->id, mr->io_start, mr->io_start + mr->usable_size, res);
> +	return 0;
> +}
> +#else
> +int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
> +{
> +	return 0;
> +}
> +#endif
> diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
> index ddce1069ba1e..85832e6ed3f5 100644
> --- a/drivers/gpu/drm/xe/xe_svm.h
> +++ b/drivers/gpu/drm/xe/xe_svm.h
> @@ -11,6 +11,7 @@
>   
>   #define XE_INTERCONNECT_VRAM DRM_INTERCONNECT_DRIVER
>   
> +struct xe_mem_region;
>   struct xe_tile;
>   struct xe_vm;
>   struct xe_vma;
> @@ -36,6 +37,8 @@ struct xe_svm_range {
>   	u8 tile_invalidated;
>   };
>   
> +int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr);
> +
>   int xe_svm_init(struct xe_vm *vm);
>   
>   void xe_svm_fini(struct xe_vm *vm);
> diff --git a/drivers/gpu/drm/xe/xe_tile.c b/drivers/gpu/drm/xe/xe_tile.c
> index 07cf7cfe4abd..bbb430392dfb 100644
> --- a/drivers/gpu/drm/xe/xe_tile.c
> +++ b/drivers/gpu/drm/xe/xe_tile.c
> @@ -13,6 +13,7 @@
>   #include "xe_migrate.h"
>   #include "xe_pcode.h"
>   #include "xe_sa.h"
> +#include "xe_svm.h"
>   #include "xe_tile.h"
>   #include "xe_tile_sysfs.h"
>   #include "xe_ttm_vram_mgr.h"
> @@ -164,6 +165,7 @@ static int tile_ttm_mgr_init(struct xe_tile *tile)
>    */
>   int xe_tile_init_noalloc(struct xe_tile *tile)
>   {
> +	struct xe_device *xe = tile_to_xe(tile);
>   	int err;
>   
>   	err = tile_ttm_mgr_init(tile);
> @@ -176,6 +178,9 @@ int xe_tile_init_noalloc(struct xe_tile *tile)
>   
>   	xe_wa_apply_tile_workarounds(tile);
>   
> +	if (xe->info.has_usm && IS_DGFX(xe))
> +		xe_devm_add(tile, &tile->mem.vram);
> +
>   	err = xe_tile_sysfs_init(tile);
>   
>   	return 0;

Re: [PATCH v3 19/30] drm/xe: Add SVM device memory mirroring

[Matthew Brost]

On Sat, Dec 21, 2024 at 12:09:55AM +0530, Ghimiray, Himal Prasad wrote:
> 
> 
> On 18-12-2024 05:03, Matthew Brost wrote:
> > Add SVM device memory mirroring which enables device pages for
> > migration. Enabled via CONFIG_XE_DEVMEM_MIRROR Kconfig. Kconfig option
> > defaults to enabled. If not enabled, SVM will work sans migration and
> > KMD memory footprint will be less.
> > 
> > v3:
> >   - Add CONFIG_XE_DEVMEM_MIRROR
> > 
> > Signed-off-by: Niranjana Vishwanathapura <niranjana.vishwanathapura@intel.com>
> > Signed-off-by: Oak Zeng <oak.zeng@intel.com>
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > ---
> >   drivers/gpu/drm/xe/Kconfig           |  9 ++++
> >   drivers/gpu/drm/xe/xe_device_types.h |  8 ++++
> >   drivers/gpu/drm/xe/xe_svm.c          | 62 +++++++++++++++++++++++++++-
> >   drivers/gpu/drm/xe/xe_svm.h          |  3 ++
> >   drivers/gpu/drm/xe/xe_tile.c         |  5 +++
> >   5 files changed, 85 insertions(+), 2 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/xe/Kconfig b/drivers/gpu/drm/xe/Kconfig
> > index 3a08e16bfada..1989c99a9f5a 100644
> > --- a/drivers/gpu/drm/xe/Kconfig
> > +++ b/drivers/gpu/drm/xe/Kconfig
> > @@ -60,6 +60,15 @@ config DRM_XE_DISPLAY
> >   	help
> >   	  Disable this option only if you want to compile out display support.
> > +config CONFIG_XE_DEVMEM_MIRROR
> 
> 
> %s/CONFIG_XE_DEVMEM_MIRROR/XE_DEVMEM_MIRROR
> 

Yep, missed this. This version is broken with proper Kconfig enabled
too. Working on fixing this up now.

Matt

> 
> > +	bool "Enable device memory mirror"
> > +	depends on DRM_XE
> > +	select CONFIG_GET_FREE_REGION
> > +	default y
> > +	help
> > +	  Disable this option only if you want to compile out without device
> > +	  memory mirror. Will reduce KMD memory footprint when disabled.
> > +
> >   config DRM_XE_FORCE_PROBE
> >   	string "Force probe xe for selected Intel hardware IDs"
> >   	depends on DRM_XE
> > diff --git a/drivers/gpu/drm/xe/xe_device_types.h b/drivers/gpu/drm/xe/xe_device_types.h
> > index 8a7b15972413..b949a960cebb 100644
> > --- a/drivers/gpu/drm/xe/xe_device_types.h
> > +++ b/drivers/gpu/drm/xe/xe_device_types.h
> > @@ -102,6 +102,14 @@ struct xe_mem_region {
> >   	resource_size_t actual_physical_size;
> >   	/** @mapping: pointer to VRAM mappable space */
> >   	void __iomem *mapping;
> > +	/** @pagemap: Used to remap device memory as ZONE_DEVICE */
> > +	struct dev_pagemap pagemap;
> > +	/**
> > +	 * @hpa_base: base host physical address
> > +	 *
> > +	 * This is generated when remap device memory as ZONE_DEVICE
> > +	 */
> > +	resource_size_t hpa_base;
> >   };
> >   /**
> > diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
> > index 3cbf29732249..2860e54bb035 100644
> > --- a/drivers/gpu/drm/xe/xe_svm.c
> > +++ b/drivers/gpu/drm/xe/xe_svm.c
> > @@ -19,6 +19,11 @@ static struct xe_vm *range_to_vm(struct drm_gpusvm_range *r)
> >   	return gpusvm_to_vm(r->gpusvm);
> >   }
> > +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)
> >   {
> > @@ -306,8 +311,8 @@ 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, NULL, 0, vm->size,
> > -			      SZ_512M, &gpusvm_ops, fault_chunk_sizes,
> > +			      current->mm, xe_svm_devm_owner(vm->xe), 0,
> > +			      vm->size, SZ_512M, &gpusvm_ops, fault_chunk_sizes,
> >   			      ARRAY_SIZE(fault_chunk_sizes));
> >   	if (err)
> >   		return err;
> > @@ -442,3 +447,56 @@ bool xe_svm_has_mapping(struct xe_vm *vm, u64 start, u64 end)
> >   {
> >   	return drm_gpusvm_has_mapping(&vm->svm.gpusvm, start, end);
> >   }
> > +
> > +#if IS_ENABLED(CONFIG_XE_DEVMEM_MIRROR)
> > +/**
> > + * xe_devm_add: Remap and provide memmap backing for device memory
> > + * @tile: tile that the memory region belongs to
> > + * @mr: memory region to remap
> > + *
> > + * This remap device memory to host physical address space and create
> > + * struct page to back device memory
> > + *
> > + * Return: 0 on success standard error code otherwise
> > + */
> > +int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
> > +{
> > +	struct xe_device *xe = tile_to_xe(tile);
> > +	struct device *dev = &to_pci_dev(xe->drm.dev)->dev;
> > +	struct resource *res;
> > +	void *addr;
> > +	int ret;
> > +
> > +	res = devm_request_free_mem_region(dev, &iomem_resource,
> > +					   mr->usable_size);
> > +	if (IS_ERR(res)) {
> > +		ret = PTR_ERR(res);
> > +		return ret;
> > +	}
> > +
> > +	mr->pagemap.type = MEMORY_DEVICE_PRIVATE;
> > +	mr->pagemap.range.start = res->start;
> > +	mr->pagemap.range.end = res->end;
> > +	mr->pagemap.nr_range = 1;
> > +	mr->pagemap.ops = drm_gpusvm_pagemap_ops_get();
> > +	mr->pagemap.owner = xe_svm_devm_owner(xe);
> > +	addr = devm_memremap_pages(dev, &mr->pagemap);
> > +	if (IS_ERR(addr)) {
> > +		devm_release_mem_region(dev, res->start, resource_size(res));
> > +		ret = PTR_ERR(addr);
> > +		drm_err(&xe->drm, "Failed to remap tile %d memory, errno %d\n",
> > +				tile->id, ret);
> > +		return ret;
> > +	}
> > +	mr->hpa_base = res->start;
> > +
> > +	drm_info(&xe->drm, "Added tile %d memory [%llx-%llx] to devm, remapped to %pr\n",
> > +		 tile->id, mr->io_start, mr->io_start + mr->usable_size, res);
> > +	return 0;
> > +}
> > +#else
> > +int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr)
> > +{
> > +	return 0;
> > +}
> > +#endif
> > diff --git a/drivers/gpu/drm/xe/xe_svm.h b/drivers/gpu/drm/xe/xe_svm.h
> > index ddce1069ba1e..85832e6ed3f5 100644
> > --- a/drivers/gpu/drm/xe/xe_svm.h
> > +++ b/drivers/gpu/drm/xe/xe_svm.h
> > @@ -11,6 +11,7 @@
> >   #define XE_INTERCONNECT_VRAM DRM_INTERCONNECT_DRIVER
> > +struct xe_mem_region;
> >   struct xe_tile;
> >   struct xe_vm;
> >   struct xe_vma;
> > @@ -36,6 +37,8 @@ struct xe_svm_range {
> >   	u8 tile_invalidated;
> >   };
> > +int xe_devm_add(struct xe_tile *tile, struct xe_mem_region *mr);
> > +
> >   int xe_svm_init(struct xe_vm *vm);
> >   void xe_svm_fini(struct xe_vm *vm);
> > diff --git a/drivers/gpu/drm/xe/xe_tile.c b/drivers/gpu/drm/xe/xe_tile.c
> > index 07cf7cfe4abd..bbb430392dfb 100644
> > --- a/drivers/gpu/drm/xe/xe_tile.c
> > +++ b/drivers/gpu/drm/xe/xe_tile.c
> > @@ -13,6 +13,7 @@
> >   #include "xe_migrate.h"
> >   #include "xe_pcode.h"
> >   #include "xe_sa.h"
> > +#include "xe_svm.h"
> >   #include "xe_tile.h"
> >   #include "xe_tile_sysfs.h"
> >   #include "xe_ttm_vram_mgr.h"
> > @@ -164,6 +165,7 @@ static int tile_ttm_mgr_init(struct xe_tile *tile)
> >    */
> >   int xe_tile_init_noalloc(struct xe_tile *tile)
> >   {
> > +	struct xe_device *xe = tile_to_xe(tile);
> >   	int err;
> >   	err = tile_ttm_mgr_init(tile);
> > @@ -176,6 +178,9 @@ int xe_tile_init_noalloc(struct xe_tile *tile)
> >   	xe_wa_apply_tile_workarounds(tile);
> > +	if (xe->info.has_usm && IS_DGFX(xe))
> > +		xe_devm_add(tile, &tile->mem.vram);
> > +
> >   	err = xe_tile_sysfs_init(tile);
> >   	return 0;
> 

Re: [PATCH v3 15/30] drm/xe: Add unbind to SVM garbage collector

[Ghimiray, Himal Prasad]

On 18-12-2024 05:03, Matthew Brost wrote:
> Add unbind to SVM garbage collector. To facilitate add unbind support
> function to VM layer which unbinds a SVM range. Also teach PY layer to
> understand unbinds of SVM ranges.
> 
> v3:
>   - s/INVALID_VMA/XE_INVALID_VMA (Thomas)
>   - Kernel doc (Thomas)
>   - New GPU SVM range structure (Thomas)
>   - s/DRM_GPUVA_OP_USER/DRM_GPUVA_OP_DRIVER (Thomas)
> 
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>   drivers/gpu/drm/xe/xe_pt.c       | 84 ++++++++++++++++++++++++++------
>   drivers/gpu/drm/xe/xe_svm.c      |  9 +++-
>   drivers/gpu/drm/xe/xe_vm.c       | 83 +++++++++++++++++++++++++++++++
>   drivers/gpu/drm/xe/xe_vm.h       |  2 +
>   drivers/gpu/drm/xe/xe_vm_types.h | 12 ++++-
>   5 files changed, 172 insertions(+), 18 deletions(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
> index 6924f3ebfe3a..9c90dfc81fbe 100644
> --- a/drivers/gpu/drm/xe/xe_pt.c
> +++ b/drivers/gpu/drm/xe/xe_pt.c
> @@ -957,10 +957,16 @@ static void xe_pt_cancel_bind(struct xe_vma *vma,
>   	}
>   }
>   
> +#define XE_INVALID_VMA	((struct xe_vma *)(0xdeaddeadull))
> +
>   static void xe_pt_commit_locks_assert(struct xe_vma *vma)
>   {
> -	struct xe_vm *vm = xe_vma_vm(vma);
> +	struct xe_vm *vm;
>   
> +	if (vma == XE_INVALID_VMA)
> +		return;
> +
> +	vm = xe_vma_vm(vma);
>   	lockdep_assert_held(&vm->lock);
>   
>   	if (!xe_vma_has_no_bo(vma))
> @@ -986,7 +992,8 @@ static void xe_pt_commit(struct xe_vma *vma,
>   		for (j = 0; j < entries[i].qwords; j++) {
>   			struct xe_pt *oldpte = entries[i].pt_entries[j].pt;
>   
> -			xe_pt_destroy(oldpte, xe_vma_vm(vma)->flags, deferred);
> +			xe_pt_destroy(oldpte, (vma == XE_INVALID_VMA) ? 0 :
> +				      xe_vma_vm(vma)->flags, deferred);
>   		}
>   	}
>   }
> @@ -1419,6 +1426,9 @@ static int xe_pt_svm_pre_commit(struct xe_migrate_pt_update *pt_update)
>   	list_for_each_entry(op, &vops->list, link) {
>   		struct xe_svm_range *range = op->map_range.range;
>   
> +		if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE)
> +			continue;
> +
>   		xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(op->map_range.vma));
>   		xe_assert(vm->xe, op->subop == XE_VMA_SUBOP_MAP_RANGE);
>   
> @@ -1616,7 +1626,9 @@ static const struct xe_pt_walk_ops xe_pt_stage_unbind_ops = {
>    * xe_pt_stage_unbind() - Build page-table update structures for an unbind
>    * operation
>    * @tile: The tile we're unbinding for.
> + * @vm: The vm
>    * @vma: The vma we're unbinding.
> + * @range: The range we're unbinding.
>    * @entries: Caller-provided storage for the update structures.
>    *
>    * Builds page-table update structures for an unbind operation. The function
> @@ -1626,9 +1638,14 @@ static const struct xe_pt_walk_ops xe_pt_stage_unbind_ops = {
>    *
>    * Return: The number of entries used.
>    */
> -static unsigned int xe_pt_stage_unbind(struct xe_tile *tile, struct xe_vma *vma,
> +static unsigned int xe_pt_stage_unbind(struct xe_tile *tile,
> +				       struct xe_vm *vm,
> +				       struct xe_vma *vma,
> +				       struct xe_svm_range *range,
>   				       struct xe_vm_pgtable_update *entries)
>   {
> +	u64 start = range ? range->base.itree.start : xe_vma_start(vma);
> +	u64 end = range ? range->base.itree.last + 1 : xe_vma_end(vma);
>   	struct xe_pt_stage_unbind_walk xe_walk = {
>   		.base = {
>   			.ops = &xe_pt_stage_unbind_ops,
> @@ -1636,14 +1653,14 @@ static unsigned int xe_pt_stage_unbind(struct xe_tile *tile, struct xe_vma *vma,
>   			.max_level = XE_PT_HIGHEST_LEVEL,
>   		},
>   		.tile = tile,
> -		.modified_start = xe_vma_start(vma),
> -		.modified_end = xe_vma_end(vma),
> +		.modified_start = start,
> +		.modified_end = end,
>   		.wupd.entries = entries,
>   	};
> -	struct xe_pt *pt = xe_vma_vm(vma)->pt_root[tile->id];
> +	struct xe_pt *pt = vm->pt_root[tile->id];
>   
> -	(void)xe_pt_walk_shared(&pt->base, pt->level, xe_vma_start(vma),
> -				xe_vma_end(vma), &xe_walk.base);
> +	(void)xe_pt_walk_shared(&pt->base, pt->level, start, end,
> +				&xe_walk.base);
>   
>   	return xe_walk.wupd.num_used_entries;
>   }
> @@ -1885,13 +1902,6 @@ static int unbind_op_prepare(struct xe_tile *tile,
>   	       "Preparing unbind, with range [%llx...%llx)\n",
>   	       xe_vma_start(vma), xe_vma_end(vma) - 1);
>   
> -	/*
> -	 * Wait for invalidation to complete. Can corrupt internal page table
> -	 * state if an invalidation is running while preparing an unbind.
> -	 */
> -	if (xe_vma_is_userptr(vma) && xe_vm_in_fault_mode(xe_vma_vm(vma)))
> -		mmu_interval_read_begin(&to_userptr_vma(vma)->userptr.notifier);
> -
>   	pt_op->vma = vma;
>   	pt_op->bind = false;
>   	pt_op->rebind = false;
> @@ -1900,7 +1910,8 @@ static int unbind_op_prepare(struct xe_tile *tile,
>   	if (err)
>   		return err;
>   
> -	pt_op->num_entries = xe_pt_stage_unbind(tile, vma, pt_op->entries);
> +	pt_op->num_entries = xe_pt_stage_unbind(tile, xe_vma_vm(vma),
> +						vma, NULL, pt_op->entries);
>   
>   	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
>   				pt_op->num_entries, false);
> @@ -1915,6 +1926,42 @@ static int unbind_op_prepare(struct xe_tile *tile,
>   	return 0;
>   }
>   
> +static int unbind_range_prepare(struct xe_vm *vm,
> +				struct xe_tile *tile,
> +				struct xe_vm_pgtable_update_ops *pt_update_ops,
> +				struct xe_svm_range *range)
> +{
> +	u32 current_op = pt_update_ops->current_op;
> +	struct xe_vm_pgtable_update_op *pt_op = &pt_update_ops->ops[current_op];
> +
> +	if (!(range->tile_present & BIT(tile->id)))
> +		return 0;
> +
> +	vm_dbg(&vm->xe->drm,
> +	       "Preparing unbind, with range [%lx...%lx)\n",
> +	       range->base.itree.start, range->base.itree.last);
> +
> +	pt_op->vma = XE_INVALID_VMA;
> +	pt_op->bind = false;
> +	pt_op->rebind = false;
> +
> +	pt_op->num_entries = xe_pt_stage_unbind(tile, vm, NULL, range,
> +						pt_op->entries);
> +
> +	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
> +				pt_op->num_entries, false);
> +	xe_pt_update_ops_rfence_interval(pt_update_ops, range->base.itree.start,
> +					 range->base.itree.last + 1);
> +	++pt_update_ops->current_op;
> +	pt_update_ops->needs_svm_lock = true;
> +	pt_update_ops->needs_invalidation = true;
> +
> +	xe_pt_commit_prepare_unbind(XE_INVALID_VMA, pt_op->entries,
> +				    pt_op->num_entries);
> +
> +	return 0;
> +}
> +
>   static int op_prepare(struct xe_vm *vm,
>   		      struct xe_tile *tile,
>   		      struct xe_vm_pgtable_update_ops *pt_update_ops,
> @@ -1982,6 +2029,9 @@ static int op_prepare(struct xe_vm *vm,
>   			err = bind_range_prepare(vm, tile, pt_update_ops,
>   						 op->map_range.vma,
>   						 op->map_range.range);
> +		} else if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE) {
> +			err = unbind_range_prepare(vm, tile, pt_update_ops,
> +						   op->unmap_range.range);
>   		}
>   		break;
>   	default:
> @@ -2171,6 +2221,8 @@ static void op_commit(struct xe_vm *vm,
>   		if (op->subop == XE_VMA_SUBOP_MAP_RANGE) {
>   			op->map_range.range->tile_present |= BIT(tile->id);
>   			op->map_range.range->tile_invalidated &= ~BIT(tile->id);
> +		} else if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE) {
> +			op->unmap_range.range->tile_present &= ~BIT(tile->id);
>   		}
>   		break;
>   	}
> diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
> index 32b1581b1ad0..58fbe1b65e8b 100644
> --- a/drivers/gpu/drm/xe/xe_svm.c
> +++ b/drivers/gpu/drm/xe/xe_svm.c
> @@ -216,7 +216,14 @@ static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
>   static int __xe_svm_garbage_collector(struct xe_vm *vm,
>   				      struct xe_svm_range *range)
>   {
> -	/* TODO: Do unbind */
> +	struct dma_fence *fence;
> +
> +	xe_vm_lock(vm, false);
> +	fence = xe_vm_range_unbind(vm, range);
> +	xe_vm_unlock(vm);
> +	if (IS_ERR(fence))
> +		return PTR_ERR(fence);
> +	dma_fence_put(fence);
>   
>   	drm_gpusvm_range_remove(&vm->svm.gpusvm, &range->base);
>   
> diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
> index 12e5e562c5e1..20ea099664af 100644
> --- a/drivers/gpu/drm/xe/xe_vm.c
> +++ b/drivers/gpu/drm/xe/xe_vm.c
> @@ -984,6 +984,89 @@ struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
>   	return fence;
>   }
>   
> +static void xe_vm_populate_range_unbind(struct xe_vma_op *op,
> +					struct xe_svm_range *range)
> +{
> +	INIT_LIST_HEAD(&op->link);
> +	op->tile_mask = range->tile_present;
> +	op->base.op = DRM_GPUVA_OP_DRIVER;
> +	op->subop = XE_VMA_SUBOP_UNMAP_RANGE;
> +	op->unmap_range.range = range;
> +}
> +
> +static int
> +xe_vm_ops_add_range_unbind(struct xe_vma_ops *vops,
> +			   struct xe_svm_range *range)
> +{
> +	struct xe_vma_op *op;
> +
> +	op = kzalloc(sizeof(*op), GFP_KERNEL);
> +	if (!op)
> +		return -ENOMEM;
> +
> +	xe_vm_populate_range_unbind(op, range);
> +	list_add_tail(&op->link, &vops->list);
> +	xe_vma_ops_incr_pt_update_ops(vops, range->tile_present);
> +
> +	return 0;
> +}
> +
> +/**
> + * xe_vm_range_unbind() - VM range unbind
> + * @vm: The VM which the range belongs to.
> + * @range: SVM range to rebind.
> + *
> + * Unbind SVM range removing the GPU page tables for the range.
> + *
> + * Return: dma fence for unbind to signal completion on succees, ERR_PTR on
> + * failure
> + */
> +struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm,
> +				     struct xe_svm_range *range)
> +{
> +	struct dma_fence *fence = NULL;
> +	struct xe_vma_ops vops;
> +	struct xe_vma_op *op, *next_op;
> +	struct xe_tile *tile;
> +	u8 id;
> +	int err;
> +
> +	lockdep_assert_held(&vm->lock);
> +	xe_vm_assert_held(vm);
> +	xe_assert(vm->xe, xe_vm_in_fault_mode(vm));
> +
> +	if (!range->tile_present)
> +		return dma_fence_get_stub();
> +
> +	xe_vma_ops_init(&vops, vm, NULL, NULL, 0);
> +	for_each_tile(tile, vm->xe, id) {
> +		vops.pt_update_ops[id].wait_vm_bookkeep = true;
> +		vops.pt_update_ops[tile->id].q =
> +			xe_tile_migrate_exec_queue(tile);
> +	}
> +
> +	err = xe_vm_ops_add_range_unbind(&vops, range);
> +	if (err)
> +		return ERR_PTR(err);
> +
> +	err = xe_vma_ops_alloc(&vops, false);
> +	if (err) {
> +		fence = ERR_PTR(err);
> +		goto free_ops;
> +	}
> +
> +	fence = ops_execute(vm, &vops);
> +
> +free_ops:
> +	list_for_each_entry_safe(op, next_op, &vops.list, link) {
> +		list_del(&op->link);
> +		kfree(op);
> +	}
> +	xe_vma_ops_fini(&vops);
> +
> +	return fence;
> +}
> +
>   static void xe_vma_free(struct xe_vma *vma)
>   {
>   	if (xe_vma_is_userptr(vma))
> diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h
> index a82fe743bbe0..3b6316dd9fd6 100644
> --- a/drivers/gpu/drm/xe/xe_vm.h
> +++ b/drivers/gpu/drm/xe/xe_vm.h
> @@ -221,6 +221,8 @@ struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
>   				     struct xe_vma *vma,
>   				     struct xe_svm_range *range,
>   				     u8 tile_mask);
> +struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm,
> +				     struct xe_svm_range *range);
>   
>   int xe_vm_invalidate_vma(struct xe_vma *vma);
>   
> diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
> index 576316729249..19576ac095e7 100644
> --- a/drivers/gpu/drm/xe/xe_vm_types.h
> +++ b/drivers/gpu/drm/xe/xe_vm_types.h
> @@ -361,6 +361,12 @@ struct xe_vma_op_map_range {
>   	struct xe_svm_range *range;
>   };
>   
> +/** struct xe_vma_op_unmap_range - VMA unmap range operation */
> +struct xe_vma_op_unmap_range {
> +	/** @range: SVM range to unmap */
> +	struct xe_svm_range *range;
> +};
> +
>   /** enum xe_vma_op_flags - flags for VMA operation */
>   enum xe_vma_op_flags {
>   	/** @XE_VMA_OP_COMMITTED: VMA operation committed */
> @@ -375,6 +381,8 @@ enum xe_vma_op_flags {
>   enum xe_vma_subop {
>   	/** @XE_VMA_SUBOP_MAP_RANGE: Map range */
>   	XE_VMA_SUBOP_MAP_RANGE,
> +	/** @XE_VMA_SUBOP_UNMAP_RANGE: Unmap range */
> +	XE_VMA_SUBOP_UNMAP_RANGE,
>   };
>   
>   /** struct xe_vma_op - VMA operation */
> @@ -397,8 +405,10 @@ struct xe_vma_op {
>   		struct xe_vma_op_remap remap;
>   		/** @prefetch: VMA prefetch operation specific data */
>   		struct xe_vma_op_prefetch prefetch;
> -		/** @map: VMA map range operation specific data */
> +		/** @map_range: VMA map range operation specific data */
>   		struct xe_vma_op_map_range map_range;
> +		/** @unmap_range: VMA unmap range operation specific data */
> +		struct xe_vma_op_map_range unmap_range;

:%s/xe_vma_op_map_range unmap_range/xe_vma_op_unmap_range  unmap_range/
>   	};
>   };
>   

Re: [PATCH v3 15/30] drm/xe: Add unbind to SVM garbage collector

[Matthew Brost]

On Sat, Dec 21, 2024 at 12:20:09AM +0530, Ghimiray, Himal Prasad wrote:
> 
> 
> On 18-12-2024 05:03, Matthew Brost wrote:
> > Add unbind to SVM garbage collector. To facilitate add unbind support
> > function to VM layer which unbinds a SVM range. Also teach PY layer to
> > understand unbinds of SVM ranges.
> > 
> > v3:
> >   - s/INVALID_VMA/XE_INVALID_VMA (Thomas)
> >   - Kernel doc (Thomas)
> >   - New GPU SVM range structure (Thomas)
> >   - s/DRM_GPUVA_OP_USER/DRM_GPUVA_OP_DRIVER (Thomas)
> > 
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > ---
> >   drivers/gpu/drm/xe/xe_pt.c       | 84 ++++++++++++++++++++++++++------
> >   drivers/gpu/drm/xe/xe_svm.c      |  9 +++-
> >   drivers/gpu/drm/xe/xe_vm.c       | 83 +++++++++++++++++++++++++++++++
> >   drivers/gpu/drm/xe/xe_vm.h       |  2 +
> >   drivers/gpu/drm/xe/xe_vm_types.h | 12 ++++-
> >   5 files changed, 172 insertions(+), 18 deletions(-)
> > 
> > diff --git a/drivers/gpu/drm/xe/xe_pt.c b/drivers/gpu/drm/xe/xe_pt.c
> > index 6924f3ebfe3a..9c90dfc81fbe 100644
> > --- a/drivers/gpu/drm/xe/xe_pt.c
> > +++ b/drivers/gpu/drm/xe/xe_pt.c
> > @@ -957,10 +957,16 @@ static void xe_pt_cancel_bind(struct xe_vma *vma,
> >   	}
> >   }
> > +#define XE_INVALID_VMA	((struct xe_vma *)(0xdeaddeadull))
> > +
> >   static void xe_pt_commit_locks_assert(struct xe_vma *vma)
> >   {
> > -	struct xe_vm *vm = xe_vma_vm(vma);
> > +	struct xe_vm *vm;
> > +	if (vma == XE_INVALID_VMA)
> > +		return;
> > +
> > +	vm = xe_vma_vm(vma);
> >   	lockdep_assert_held(&vm->lock);
> >   	if (!xe_vma_has_no_bo(vma))
> > @@ -986,7 +992,8 @@ static void xe_pt_commit(struct xe_vma *vma,
> >   		for (j = 0; j < entries[i].qwords; j++) {
> >   			struct xe_pt *oldpte = entries[i].pt_entries[j].pt;
> > -			xe_pt_destroy(oldpte, xe_vma_vm(vma)->flags, deferred);
> > +			xe_pt_destroy(oldpte, (vma == XE_INVALID_VMA) ? 0 :
> > +				      xe_vma_vm(vma)->flags, deferred);
> >   		}
> >   	}
> >   }
> > @@ -1419,6 +1426,9 @@ static int xe_pt_svm_pre_commit(struct xe_migrate_pt_update *pt_update)
> >   	list_for_each_entry(op, &vops->list, link) {
> >   		struct xe_svm_range *range = op->map_range.range;
> > +		if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE)
> > +			continue;
> > +
> >   		xe_assert(vm->xe, xe_vma_is_cpu_addr_mirror(op->map_range.vma));
> >   		xe_assert(vm->xe, op->subop == XE_VMA_SUBOP_MAP_RANGE);
> > @@ -1616,7 +1626,9 @@ static const struct xe_pt_walk_ops xe_pt_stage_unbind_ops = {
> >    * xe_pt_stage_unbind() - Build page-table update structures for an unbind
> >    * operation
> >    * @tile: The tile we're unbinding for.
> > + * @vm: The vm
> >    * @vma: The vma we're unbinding.
> > + * @range: The range we're unbinding.
> >    * @entries: Caller-provided storage for the update structures.
> >    *
> >    * Builds page-table update structures for an unbind operation. The function
> > @@ -1626,9 +1638,14 @@ static const struct xe_pt_walk_ops xe_pt_stage_unbind_ops = {
> >    *
> >    * Return: The number of entries used.
> >    */
> > -static unsigned int xe_pt_stage_unbind(struct xe_tile *tile, struct xe_vma *vma,
> > +static unsigned int xe_pt_stage_unbind(struct xe_tile *tile,
> > +				       struct xe_vm *vm,
> > +				       struct xe_vma *vma,
> > +				       struct xe_svm_range *range,
> >   				       struct xe_vm_pgtable_update *entries)
> >   {
> > +	u64 start = range ? range->base.itree.start : xe_vma_start(vma);
> > +	u64 end = range ? range->base.itree.last + 1 : xe_vma_end(vma);
> >   	struct xe_pt_stage_unbind_walk xe_walk = {
> >   		.base = {
> >   			.ops = &xe_pt_stage_unbind_ops,
> > @@ -1636,14 +1653,14 @@ static unsigned int xe_pt_stage_unbind(struct xe_tile *tile, struct xe_vma *vma,
> >   			.max_level = XE_PT_HIGHEST_LEVEL,
> >   		},
> >   		.tile = tile,
> > -		.modified_start = xe_vma_start(vma),
> > -		.modified_end = xe_vma_end(vma),
> > +		.modified_start = start,
> > +		.modified_end = end,
> >   		.wupd.entries = entries,
> >   	};
> > -	struct xe_pt *pt = xe_vma_vm(vma)->pt_root[tile->id];
> > +	struct xe_pt *pt = vm->pt_root[tile->id];
> > -	(void)xe_pt_walk_shared(&pt->base, pt->level, xe_vma_start(vma),
> > -				xe_vma_end(vma), &xe_walk.base);
> > +	(void)xe_pt_walk_shared(&pt->base, pt->level, start, end,
> > +				&xe_walk.base);
> >   	return xe_walk.wupd.num_used_entries;
> >   }
> > @@ -1885,13 +1902,6 @@ static int unbind_op_prepare(struct xe_tile *tile,
> >   	       "Preparing unbind, with range [%llx...%llx)\n",
> >   	       xe_vma_start(vma), xe_vma_end(vma) - 1);
> > -	/*
> > -	 * Wait for invalidation to complete. Can corrupt internal page table
> > -	 * state if an invalidation is running while preparing an unbind.
> > -	 */
> > -	if (xe_vma_is_userptr(vma) && xe_vm_in_fault_mode(xe_vma_vm(vma)))
> > -		mmu_interval_read_begin(&to_userptr_vma(vma)->userptr.notifier);
> > -
> >   	pt_op->vma = vma;
> >   	pt_op->bind = false;
> >   	pt_op->rebind = false;
> > @@ -1900,7 +1910,8 @@ static int unbind_op_prepare(struct xe_tile *tile,
> >   	if (err)
> >   		return err;
> > -	pt_op->num_entries = xe_pt_stage_unbind(tile, vma, pt_op->entries);
> > +	pt_op->num_entries = xe_pt_stage_unbind(tile, xe_vma_vm(vma),
> > +						vma, NULL, pt_op->entries);
> >   	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
> >   				pt_op->num_entries, false);
> > @@ -1915,6 +1926,42 @@ static int unbind_op_prepare(struct xe_tile *tile,
> >   	return 0;
> >   }
> > +static int unbind_range_prepare(struct xe_vm *vm,
> > +				struct xe_tile *tile,
> > +				struct xe_vm_pgtable_update_ops *pt_update_ops,
> > +				struct xe_svm_range *range)
> > +{
> > +	u32 current_op = pt_update_ops->current_op;
> > +	struct xe_vm_pgtable_update_op *pt_op = &pt_update_ops->ops[current_op];
> > +
> > +	if (!(range->tile_present & BIT(tile->id)))
> > +		return 0;
> > +
> > +	vm_dbg(&vm->xe->drm,
> > +	       "Preparing unbind, with range [%lx...%lx)\n",
> > +	       range->base.itree.start, range->base.itree.last);
> > +
> > +	pt_op->vma = XE_INVALID_VMA;
> > +	pt_op->bind = false;
> > +	pt_op->rebind = false;
> > +
> > +	pt_op->num_entries = xe_pt_stage_unbind(tile, vm, NULL, range,
> > +						pt_op->entries);
> > +
> > +	xe_vm_dbg_print_entries(tile_to_xe(tile), pt_op->entries,
> > +				pt_op->num_entries, false);
> > +	xe_pt_update_ops_rfence_interval(pt_update_ops, range->base.itree.start,
> > +					 range->base.itree.last + 1);
> > +	++pt_update_ops->current_op;
> > +	pt_update_ops->needs_svm_lock = true;
> > +	pt_update_ops->needs_invalidation = true;
> > +
> > +	xe_pt_commit_prepare_unbind(XE_INVALID_VMA, pt_op->entries,
> > +				    pt_op->num_entries);
> > +
> > +	return 0;
> > +}
> > +
> >   static int op_prepare(struct xe_vm *vm,
> >   		      struct xe_tile *tile,
> >   		      struct xe_vm_pgtable_update_ops *pt_update_ops,
> > @@ -1982,6 +2029,9 @@ static int op_prepare(struct xe_vm *vm,
> >   			err = bind_range_prepare(vm, tile, pt_update_ops,
> >   						 op->map_range.vma,
> >   						 op->map_range.range);
> > +		} else if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE) {
> > +			err = unbind_range_prepare(vm, tile, pt_update_ops,
> > +						   op->unmap_range.range);
> >   		}
> >   		break;
> >   	default:
> > @@ -2171,6 +2221,8 @@ static void op_commit(struct xe_vm *vm,
> >   		if (op->subop == XE_VMA_SUBOP_MAP_RANGE) {
> >   			op->map_range.range->tile_present |= BIT(tile->id);
> >   			op->map_range.range->tile_invalidated &= ~BIT(tile->id);
> > +		} else if (op->subop == XE_VMA_SUBOP_UNMAP_RANGE) {
> > +			op->unmap_range.range->tile_present &= ~BIT(tile->id);
> >   		}
> >   		break;
> >   	}
> > diff --git a/drivers/gpu/drm/xe/xe_svm.c b/drivers/gpu/drm/xe/xe_svm.c
> > index 32b1581b1ad0..58fbe1b65e8b 100644
> > --- a/drivers/gpu/drm/xe/xe_svm.c
> > +++ b/drivers/gpu/drm/xe/xe_svm.c
> > @@ -216,7 +216,14 @@ static void xe_svm_invalidate(struct drm_gpusvm *gpusvm,
> >   static int __xe_svm_garbage_collector(struct xe_vm *vm,
> >   				      struct xe_svm_range *range)
> >   {
> > -	/* TODO: Do unbind */
> > +	struct dma_fence *fence;
> > +
> > +	xe_vm_lock(vm, false);
> > +	fence = xe_vm_range_unbind(vm, range);
> > +	xe_vm_unlock(vm);
> > +	if (IS_ERR(fence))
> > +		return PTR_ERR(fence);
> > +	dma_fence_put(fence);
> >   	drm_gpusvm_range_remove(&vm->svm.gpusvm, &range->base);
> > diff --git a/drivers/gpu/drm/xe/xe_vm.c b/drivers/gpu/drm/xe/xe_vm.c
> > index 12e5e562c5e1..20ea099664af 100644
> > --- a/drivers/gpu/drm/xe/xe_vm.c
> > +++ b/drivers/gpu/drm/xe/xe_vm.c
> > @@ -984,6 +984,89 @@ struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
> >   	return fence;
> >   }
> > +static void xe_vm_populate_range_unbind(struct xe_vma_op *op,
> > +					struct xe_svm_range *range)
> > +{
> > +	INIT_LIST_HEAD(&op->link);
> > +	op->tile_mask = range->tile_present;
> > +	op->base.op = DRM_GPUVA_OP_DRIVER;
> > +	op->subop = XE_VMA_SUBOP_UNMAP_RANGE;
> > +	op->unmap_range.range = range;
> > +}
> > +
> > +static int
> > +xe_vm_ops_add_range_unbind(struct xe_vma_ops *vops,
> > +			   struct xe_svm_range *range)
> > +{
> > +	struct xe_vma_op *op;
> > +
> > +	op = kzalloc(sizeof(*op), GFP_KERNEL);
> > +	if (!op)
> > +		return -ENOMEM;
> > +
> > +	xe_vm_populate_range_unbind(op, range);
> > +	list_add_tail(&op->link, &vops->list);
> > +	xe_vma_ops_incr_pt_update_ops(vops, range->tile_present);
> > +
> > +	return 0;
> > +}
> > +
> > +/**
> > + * xe_vm_range_unbind() - VM range unbind
> > + * @vm: The VM which the range belongs to.
> > + * @range: SVM range to rebind.
> > + *
> > + * Unbind SVM range removing the GPU page tables for the range.
> > + *
> > + * Return: dma fence for unbind to signal completion on succees, ERR_PTR on
> > + * failure
> > + */
> > +struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm,
> > +				     struct xe_svm_range *range)
> > +{
> > +	struct dma_fence *fence = NULL;
> > +	struct xe_vma_ops vops;
> > +	struct xe_vma_op *op, *next_op;
> > +	struct xe_tile *tile;
> > +	u8 id;
> > +	int err;
> > +
> > +	lockdep_assert_held(&vm->lock);
> > +	xe_vm_assert_held(vm);
> > +	xe_assert(vm->xe, xe_vm_in_fault_mode(vm));
> > +
> > +	if (!range->tile_present)
> > +		return dma_fence_get_stub();
> > +
> > +	xe_vma_ops_init(&vops, vm, NULL, NULL, 0);
> > +	for_each_tile(tile, vm->xe, id) {
> > +		vops.pt_update_ops[id].wait_vm_bookkeep = true;
> > +		vops.pt_update_ops[tile->id].q =
> > +			xe_tile_migrate_exec_queue(tile);
> > +	}
> > +
> > +	err = xe_vm_ops_add_range_unbind(&vops, range);
> > +	if (err)
> > +		return ERR_PTR(err);
> > +
> > +	err = xe_vma_ops_alloc(&vops, false);
> > +	if (err) {
> > +		fence = ERR_PTR(err);
> > +		goto free_ops;
> > +	}
> > +
> > +	fence = ops_execute(vm, &vops);
> > +
> > +free_ops:
> > +	list_for_each_entry_safe(op, next_op, &vops.list, link) {
> > +		list_del(&op->link);
> > +		kfree(op);
> > +	}
> > +	xe_vma_ops_fini(&vops);
> > +
> > +	return fence;
> > +}
> > +
> >   static void xe_vma_free(struct xe_vma *vma)
> >   {
> >   	if (xe_vma_is_userptr(vma))
> > diff --git a/drivers/gpu/drm/xe/xe_vm.h b/drivers/gpu/drm/xe/xe_vm.h
> > index a82fe743bbe0..3b6316dd9fd6 100644
> > --- a/drivers/gpu/drm/xe/xe_vm.h
> > +++ b/drivers/gpu/drm/xe/xe_vm.h
> > @@ -221,6 +221,8 @@ struct dma_fence *xe_vm_range_rebind(struct xe_vm *vm,
> >   				     struct xe_vma *vma,
> >   				     struct xe_svm_range *range,
> >   				     u8 tile_mask);
> > +struct dma_fence *xe_vm_range_unbind(struct xe_vm *vm,
> > +				     struct xe_svm_range *range);
> >   int xe_vm_invalidate_vma(struct xe_vma *vma);
> > diff --git a/drivers/gpu/drm/xe/xe_vm_types.h b/drivers/gpu/drm/xe/xe_vm_types.h
> > index 576316729249..19576ac095e7 100644
> > --- a/drivers/gpu/drm/xe/xe_vm_types.h
> > +++ b/drivers/gpu/drm/xe/xe_vm_types.h
> > @@ -361,6 +361,12 @@ struct xe_vma_op_map_range {
> >   	struct xe_svm_range *range;
> >   };
> > +/** struct xe_vma_op_unmap_range - VMA unmap range operation */
> > +struct xe_vma_op_unmap_range {
> > +	/** @range: SVM range to unmap */
> > +	struct xe_svm_range *range;
> > +};
> > +
> >   /** enum xe_vma_op_flags - flags for VMA operation */
> >   enum xe_vma_op_flags {
> >   	/** @XE_VMA_OP_COMMITTED: VMA operation committed */
> > @@ -375,6 +381,8 @@ enum xe_vma_op_flags {
> >   enum xe_vma_subop {
> >   	/** @XE_VMA_SUBOP_MAP_RANGE: Map range */
> >   	XE_VMA_SUBOP_MAP_RANGE,
> > +	/** @XE_VMA_SUBOP_UNMAP_RANGE: Unmap range */
> > +	XE_VMA_SUBOP_UNMAP_RANGE,
> >   };
> >   /** struct xe_vma_op - VMA operation */
> > @@ -397,8 +405,10 @@ struct xe_vma_op {
> >   		struct xe_vma_op_remap remap;
> >   		/** @prefetch: VMA prefetch operation specific data */
> >   		struct xe_vma_op_prefetch prefetch;
> > -		/** @map: VMA map range operation specific data */
> > +		/** @map_range: VMA map range operation specific data */
> >   		struct xe_vma_op_map_range map_range;
> > +		/** @unmap_range: VMA unmap range operation specific data */
> > +		struct xe_vma_op_map_range unmap_range;
> 
> :%s/xe_vma_op_map_range unmap_range/xe_vma_op_unmap_range  unmap_range/

Yep, thanks.

Matt

> >   	};
> >   };
> 

Re: [PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Matthew Brost]

On Tue, Dec 17, 2024 at 03:33:23PM -0800, Matthew Brost wrote:
> This patch introduces support for GPU Shared Virtual Memory (SVM) in the
> Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
> sharing of memory between the CPU and GPU, enhancing performance and
> flexibility in GPU computing tasks.
> 
> The patch adds the necessary infrastructure for SVM, including data
> structures and functions for managing SVM ranges and notifiers. It also
> provides mechanisms for allocating, deallocating, and migrating memory
> regions between system RAM and GPU VRAM.
> 
> This is largely inspired by GPUVM.
> 
> v2:
>  - Take order into account in check pages
>  - Clear range->pages in get pages error
>  - Drop setting dirty or accessed bit in get pages (Vetter)
>  - Remove mmap assert for cpu faults
>  - Drop mmap write lock abuse (Vetter, Christian)
>  - Decouple zdd from range (Vetter, Oak)
>  - Add drm_gpusvm_range_evict, make it work with coherent pages
>  - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
>  - mmget/put in drm_gpusvm_evict_to_sram
>  - Drop range->vram_alloation variable
>  - Don't return in drm_gpusvm_evict_to_sram until all pages detached
>  - Don't warn on mixing sram and device pages
>  - Update kernel doc
>  - Add coherent page support to get pages
>  - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
>  - Add struct drm_gpusvm_vram and ops (Thomas)
>  - Update the range's seqno if the range is valid (Thomas)
>  - Remove the is_unmapped check before hmm_range_fault (Thomas)
>  - Use drm_pagemap (Thomas)
>  - Drop kfree_mapping (Thomas)
>  - dma mapp pages under notifier lock (Thomas)
>  - Remove ctx.prefault
>  - Remove ctx.mmap_locked
>  - Add ctx.check_pages
>  - s/vram/devmem (Thomas)
> v3:
>  - Fix memory leak drm_gpusvm_range_get_pages
>  - Only migrate pages with same zdd on CPU fault
>  - Loop over al VMAs in drm_gpusvm_range_evict
>  - Make GPUSVM a drm level module
>  - GPL or MIT license
>  - Update main kernel doc (Thomas)
>  - Prefer foo() vs foo for functions in kernel doc (Thomas)
>  - Prefer functions over macros (Thomas)
>  - Use unsigned long vs u64 for addresses (Thomas)
>  - Use standard interval_tree (Thomas)
>  - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
>  - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
>  - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
>  - Newlines between functions defs in header file (Thomas)
>  - Drop shall language in driver vfunc kernel doc (Thomas)
>  - Move some static inlines from head to C file (Thomas)
>  - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
>  - Change check_pages to a threshold
> 
> Cc: Simona Vetter <simona.vetter@ffwll.ch>
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Christian König <christian.koenig@amd.com>
> Cc: <dri-devel@lists.freedesktop.org>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
>  drivers/gpu/drm/Kconfig      |    8 +
>  drivers/gpu/drm/Makefile     |    1 +
>  drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
>  include/drm/drm_gpusvm.h     |  445 +++++++
>  4 files changed, 2674 insertions(+)
>  create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>  create mode 100644 include/drm/drm_gpusvm.h
> 
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 2f51546b0b88..5524cb1ed9dd 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -278,6 +278,14 @@ config DRM_GPUVM
>  	  GPU-VM representation providing helpers to manage a GPUs virtual
>  	  address space
>  
> +config DRM_GPUSVM
> +	tristate
> +	depends on DRM
> +	depends on DEVICE_MIGRATION
> +	help
> +	  GPU-SVM representation providing helpers to manage a GPUs shared
> +	  virtual memory
> +
>  config DRM_BUDDY
>  	tristate
>  	depends on DRM
> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> index 19fb370fbc56..ca5d44441fb6 100644
> --- a/drivers/gpu/drm/Makefile
> +++ b/drivers/gpu/drm/Makefile
> @@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
>  #
>  obj-$(CONFIG_DRM_EXEC) += drm_exec.o
>  obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> +obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
>  
>  obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
>  
> diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> new file mode 100644
> index 000000000000..29ce2c7d1c8c
> --- /dev/null
> +++ b/drivers/gpu/drm/drm_gpusvm.c
> @@ -0,0 +1,2220 @@
> +// SPDX-License-Identifier: GPL-2.0-only OR MIT
> +/*
> + * Copyright © 2024 Intel Corporation
> + *
> + * Authors:
> + *     Matthew Brost <matthew.brost@intel.com>
> + */
> +
> +#include <linux/dma-mapping.h>
> +#include <linux/hmm.h>
> +#include <linux/memremap.h>
> +#include <linux/migrate.h>
> +#include <linux/mm_types.h>
> +#include <linux/pagemap.h>
> +#include <linux/slab.h>
> +
> +#include <drm/drm_device.h>
> +#include <drm/drm_gpusvm.h>
> +#include <drm/drm_pagemap.h>
> +#include <drm/drm_print.h>
> +
> +/**
> + * DOC: Overview
> + *
> + * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
> + *
> + * The GPU SVM layer is a component of the DRM framework designed to manage shared
> + * virtual memory between the CPU and GPU. It enables efficient data exchange and
> + * processing for GPU-accelerated applications by allowing memory sharing and
> + * synchronization between the CPU's and GPU's virtual address spaces.
> + *
> + * Key GPU SVM Components:
> + * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
> + *		GPU of changes, notifiers are sized based on a GPU SVM
> + *		initialization parameter, with a recommendation of 512M or
> + *		larger. They maintain a Red-BlacK tree and a list of ranges that
> + *		fall within the notifier interval. Notifiers are tracked within
> + *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
> + *		or removed as ranges within the interval are created or
> + *		destroyed.
> + * - Ranges: Represent memory ranges mapped in a DRM device and managed
> + *	     by GPU SVM. They are sized based on an array of chunk sizes, which
> + *	     is a GPU SVM initialization parameter, and the CPU address space.
> + *	     Upon GPU fault, the largest aligned chunk that fits within the
> + *	     faulting CPU address space is chosen for the range size. Ranges are
> + *	     expected to be dynamically allocated on GPU fault and removed on an
> + *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
> + *	     a notifier's Red-Black tree.
> + * - Operations: Define the interface for driver-specific GPU SVM operations
> + *               such as range allocation, notifier allocation, and
> + *               invalidations.
> + * - Device Memory Allocations: Embedded structure containing enough information
> + *                              for GPU SVM to migrate to / from device memory.
> + * - Device Memory Operations: Define the interface for driver-specific device
> + *                             memory operations release memory, populate pfns,
> + *                             and copy to / from device memory.
> + *
> + * This layer provides interfaces for allocating, mapping, migrating, and
> + * releasing memory ranges between the CPU and GPU. It handles all core memory
> + * management interactions (DMA mapping, HMM, and migration) and provides
> + * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
> + * to build the expected driver components for an SVM implementation as detailed
> + * below.
> + *
> + * Expected Driver Components:
> + * - GPU page fault handler: Used to create ranges and notifiers based on the
> + *			     fault address, optionally migrate the range to
> + *			     device memory, and create GPU bindings.
> + * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
> + *			Ranges are expected to be added to the garbage collector
> + *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
> + * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
> + *			ranges.
> + */
> +
> +/**
> + * DOC: Locking
> + *
> + * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
> + * mmap lock as needed.
> + *
> + * GPU SVM introduces a global notifier lock, which safeguards the notifier's
> + * range RB tree and list, as well as the range's DMA mappings and sequence
> + * number. GPU SVM manages all necessary locking and unlocking operations,
> + * except for the recheck range's pages being valid
> + * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
> + * lock corresponds to the 'driver->update' lock mentioned in the HMM
> + * documentation (TODO: Link). Future revisions may transition from a GPU SVM
> + * global lock to a per-notifier lock if finer-grained locking is deemed
> + * necessary.
> + *
> + * In addition to the locking mentioned above, the driver should implement a
> + * lock to safeguard core GPU SVM function calls that modify state, such as
> + * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
> + * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
> + * locking should also be possible for concurrent GPU fault processing within a
> + * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
> + * to add annotations to GPU SVM.
> + */
> +
> +/**
> + * DOC: Migration
> + *
> + * The migration support is quite simple, allowing migration between RAM and
> + * device memory at the range granularity. For example, GPU SVM currently does not
> + * support mixing RAM and device memory pages within a range. This means that upon GPU
> + * fault, the entire range can be migrated to device memory, and upon CPU fault, the
> + * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
> + * could be added in the future if required.
> + *
> + * The reasoning for only supporting range granularity is as follows: it
> + * simplifies the implementation, and range sizes are driver-defined and should
> + * be relatively small.
> + */
> +
> +/**
> + * DOC: Partial Unmapping of Ranges
> + *
> + * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
> + * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
> + * being that a subset of the range still has CPU and GPU mappings. If the
> + * backing store for the range is in device memory, a subset of the backing store has
> + * references. One option would be to split the range and device memory backing store,
> + * but the implementation for this would be quite complicated. Given that
> + * partial unmappings are rare and driver-defined range sizes are relatively
> + * small, GPU SVM does not support splitting of ranges.
> + *
> + * With no support for range splitting, upon partial unmapping of a range, the
> + * driver is expected to invalidate and destroy the entire range. If the range
> + * has device memory as its backing, the driver is also expected to migrate any
> + * remaining pages back to RAM.
> + */
> +
> +/**
> + * DOC: Examples
> + *
> + * This section provides three examples of how to build the expected driver
> + * components: the GPU page fault handler, the garbage collector, and the
> + * notifier callback.
> + *
> + * The generic code provided does not include logic for complex migration
> + * policies, optimized invalidations, fined grained driver locking, or other
> + * potentially required driver locking (e.g., DMA-resv locks).
> + *
> + * 1) GPU page fault handler
> + *
> + *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
> + *	{
> + *		int err = 0;
> + *
> + *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
> + *
> + *		drm_gpusvm_notifier_lock(gpusvm);
> + *		if (drm_gpusvm_range_pages_valid(range))
> + *			driver_commit_bind(gpusvm, range);
> + *		else
> + *			err = -EAGAIN;
> + *		drm_gpusvm_notifier_unlock(gpusvm);
> + *
> + *		return err;
> + *	}
> + *
> + *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
> + *			     unsigned long gpuva_start, unsigned long gpuva_end)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = {};
> + *		int err;
> + *
> + *		driver_svm_lock();
> + *	retry:
> + *		// Always process UNMAPs first so view of GPU SVM ranges is current
> + *		driver_garbage_collector(gpusvm);
> + *
> + *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
> + *							gpuva_start, gpuva_end,
> + *						        &ctx);
> + *		if (IS_ERR(range)) {
> + *			err = PTR_ERR(range);
> + *			goto unlock;
> + *		}
> + *
> + *		if (driver_migration_policy(range)) {
> + *			devmem = driver_alloc_devmem();
> + *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
> + *							   devmem_allocation,
> + *							   &ctx);
> + *			if (err)	// CPU mappings may have changed
> + *				goto retry;
> + *		}
> + *
> + *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
> + *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
> + *			if (err == -EOPNOTSUPP)
> + *				drm_gpusvm_range_evict(gpusvm, range);
> + *			goto retry;
> + *		} else if (err) {
> + *			goto unlock;
> + *		}
> + *
> + *		err = driver_bind_range(gpusvm, range);
> + *		if (err == -EAGAIN)	// CPU mappings changed
> + *			goto retry
> + *
> + *	unlock:
> + *		driver_svm_unlock();
> + *		return err;
> + *	}
> + *
> + * 2) Garbage Collector.
> + *
> + *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
> + *					struct drm_gpusvm_range *range)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		// Partial unmap, migrate any remaining device memory pages back to RAM
> + *		if (range->flags.partial_unmap)
> + *			drm_gpusvm_range_evict(gpusvm, range);
> + *
> + *		driver_unbind_range(range);
> + *		drm_gpusvm_range_remove(gpusvm, range);
> + *	}
> + *
> + *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		for_each_range_in_garbage_collector(gpusvm, range)
> + *			__driver_garbage_collector(gpusvm, range);
> + *	}
> + *
> + * 3) Notifier callback.
> + *
> + *	void driver_invalidation(struct drm_gpusvm *gpusvm,
> + *				 struct drm_gpusvm_notifier *notifier,
> + *				 const struct mmu_notifier_range *mmu_range)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
> + *		struct drm_gpusvm_range *range = NULL;
> + *
> + *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
> + *
> + *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
> + *					  mmu_range->end) {
> + *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
> + *
> + *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
> + *				continue;
> + *
> + *			drm_gpusvm_range_set_unmapped(range, mmu_range);
> + *			driver_garbage_collector_add(gpusvm, range);
> + *		}
> + *	}
> + */
> +
> +/**
> + * npages_in_range() - Calculate the number of pages in a given range
> + * @start: The start address of the range
> + * @end: The end address of the range
> + *
> + * This macro calculates the number of pages in a given memory range,
> + * specified by the start and end addresses. It divides the difference
> + * between the end and start addresses by the page size (PAGE_SIZE) to
> + * determine the number of pages in the range.
> + *
> + * Return: The number of pages in the specified range.
> + */
> +static unsigned long
> +npages_in_range(unsigned long start, unsigned long end)
> +{
> +	return (end - start) >> PAGE_SHIFT;
> +}
> +
> +/**
> + * struct drm_gpusvm_zdd - GPU SVM zone device data
> + *
> + * @refcount: Reference count for the zdd
> + * @destroy_work: Work structure for asynchronous zdd destruction
> + * @devmem_allocation: device memory allocation
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This structure serves as a generic wrapper installed in
> + * page->zone_device_data. It provides infrastructure for looking up a device
> + * memory allocation upon CPU page fault and asynchronously releasing device
> + * memory once the CPU has no page references. Asynchronous release is useful
> + * because CPU page references can be dropped in IRQ contexts, while releasing
> + * device memory likely requires sleeping locks.
> + */
> +struct drm_gpusvm_zdd {
> +	struct kref refcount;
> +	struct work_struct destroy_work;
> +	struct drm_gpusvm_devmem *devmem_allocation;
> +	void *device_private_page_owner;
> +};
> +
> +/**
> + * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
> + * @w: Pointer to the work_struct
> + *
> + * This function releases device memory, puts GPU SVM range, and frees zdd.
> + */
> +static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(w, struct drm_gpusvm_zdd, destroy_work);
> +	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
> +		zdd->devmem_allocation->ops : NULL;
> +
> +	if (zdd->devmem_allocation && ops->devmem_release)
> +		ops->devmem_release(zdd->devmem_allocation);
> +	kfree(zdd);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This function allocates and initializes a new zdd structure. It sets up the
> + * reference count and initializes the destroy work.
> + *
> + * Returns:
> + * Pointer to the allocated zdd on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_zdd *
> +drm_gpusvm_zdd_alloc(void *device_private_page_owner)
> +{
> +	struct drm_gpusvm_zdd *zdd;
> +
> +	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
> +	if (!zdd)
> +		return NULL;
> +
> +	kref_init(&zdd->refcount);
> +	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
> +	zdd->devmem_allocation = NULL;
> +	zdd->device_private_page_owner = device_private_page_owner;
> +
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function increments the reference count of the provided zdd structure.
> + *
> + * Returns: Pointer to the zdd structure.
> + */
> +static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_get(&zdd->refcount);
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
> + * @ref: Pointer to the reference count structure.
> + *
> + * This function queues the destroy_work of the zdd for asynchronous destruction.
> + */
> +static void drm_gpusvm_zdd_destroy(struct kref *ref)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(ref, struct drm_gpusvm_zdd, refcount);
> +
> +	if (zdd->devmem_allocation)
> +		WRITE_ONCE(zdd->devmem_allocation->detached, true);
> +	schedule_work(&zdd->destroy_work);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_put() - Put a zdd reference.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function decrements the reference count of the provided zdd structure
> + * and schedules its destruction if the count drops to zero.
> + */
> +static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
> +}
> +
> +/**
> + * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
> + * @notifier: Pointer to the GPU SVM notifier structure.
> + * @start: Start address of the range
> + * @end: End address of the range
> + *
> + * Return: A pointer to the drm_gpusvm_range if found or NULL
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_range, itree);
> +	else
> +		return NULL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
> +
> +/**
> + * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges
> + * @next__: Iterator variable for the ranges temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier while
> + * removing ranges from it.
> + */
> +#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
> +	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
> +	     (next__) = __drm_gpusvm_range_next(range__);				\
> +	     (range__) && (range__->itree.start < (end__));				\
> +	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
> +
> +/**
> + * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
> + * @notifier: a pointer to the current drm_gpusvm_notifier
> + *
> + * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
> + *         the current notifier is the last one or if the input notifier is
> + *         NULL.
> + */
> +static struct drm_gpusvm_notifier *
> +__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
> +{
> +	if (notifier && !list_is_last(&notifier->entry,
> +				      &notifier->gpusvm->notifier_list))
> +		return list_next_entry(notifier, entry);
> +
> +	return NULL;
> +}
> +
> +static struct drm_gpusvm_notifier *
> +notifier_iter_first(struct rb_root_cached *root, unsigned long start,
> +		    unsigned long last)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(root, start, last);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_notifier, itree);
> +	else
> +		return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
> + */
> +#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @next__: Iterator variable for the notifiers temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
> + * removing notifiers from it.
> + */
> +#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
> +	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
> + * @mni: Pointer to the mmu_interval_notifier structure.
> + * @mmu_range: Pointer to the mmu_notifier_range structure.
> + * @cur_seq: Current sequence number.
> + *
> + * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
> + * notifier sequence number and calls the driver invalidate vfunc under
> + * gpusvm->notifier_lock.
> + *
> + * Returns:
> + * true if the operation succeeds, false otherwise.
> + */
> +static bool
> +drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
> +			       const struct mmu_notifier_range *mmu_range,
> +			       unsigned long cur_seq)
> +{
> +	struct drm_gpusvm_notifier *notifier =
> +		container_of(mni, typeof(*notifier), notifier);
> +	struct drm_gpusvm *gpusvm = notifier->gpusvm;
> +
> +	if (!mmu_notifier_range_blockable(mmu_range))
> +		return false;
> +
> +	down_write(&gpusvm->notifier_lock);
> +	mmu_interval_set_seq(mni, cur_seq);
> +	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
> +	up_write(&gpusvm->notifier_lock);
> +
> +	return true;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
> + */
> +static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
> +	.invalidate = drm_gpusvm_notifier_invalidate,
> +};
> +
> +/**
> + * drm_gpusvm_init() - Initialize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @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.
> + * @ops: Pointer to the operations structure for GPU SVM.
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order with last
> + *               entry being SZ_4K.
> + * @num_chunks: Number of chunks.
> + *
> + * This function initializes the GPU SVM.
> + *
> + * Returns:
> + * 0 on success, a negative error code on failure.
> + */
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks)
> +{
> +	if (!ops->invalidate || !num_chunks)
> +		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;
> +	gpusvm->ops = ops;
> +	gpusvm->chunk_sizes = chunk_sizes;
> +	gpusvm->num_chunks = num_chunks;
> +
> +	mmgrab(mm);
> +	gpusvm->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&gpusvm->notifier_list);
> +
> +	init_rwsem(&gpusvm->notifier_lock);
> +
> +	fs_reclaim_acquire(GFP_KERNEL);
> +	might_lock(&gpusvm->notifier_lock);
> +	fs_reclaim_release(GFP_KERNEL);
> +
> +#ifdef CONFIG_LOCKDEP
> +	gpusvm->lock_dep_map = NULL;
> +#endif
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_init);
> +
> +/**
> + * drm_gpusvm_notifier_find() - Find GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function finds the GPU SVM notifier associated with the fault address.
> + *
> + * Returns:
> + * Pointer to the GPU SVM notifier on success, NULL otherwise.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
> +			 unsigned long fault_addr)
> +{
> +	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
> +}
> +
> +/**
> + * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_notifier structure.
> + */
> +static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	interval_tree_insert(&notifier->itree, &gpusvm->root);
> +
> +	node = rb_prev(&notifier->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_notifier(node))->entry;
> +	else
> +		head = &gpusvm->notifier_list;
> +
> +	list_add(&notifier->entry, head);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM tructure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	interval_tree_remove(&notifier->itree, &gpusvm->root);
> +	list_del(&notifier->entry);
> +}
> +
> +/**
> + * drm_gpusvm_fini() - Finalize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + *
> + * This function finalizes the GPU SVM by cleaning up any remaining ranges and
> + * notifiers, and dropping a reference to struct MM.
> + */
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
> +{
> +	struct drm_gpusvm_notifier *notifier, *next;
> +
> +	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
> +		struct drm_gpusvm_range *range, *__next;
> +
> +		/*
> +		 * Remove notifier first to avoid racing with any invalidation
> +		 */
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +		notifier->flags.removed = true;
> +
> +		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
> +					       LONG_MAX)
> +			drm_gpusvm_range_remove(gpusvm, range);
> +	}
> +
> +	mmdrop(gpusvm->mm);
> +	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
> +
> +/**
> + * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function allocates and initializes the GPU SVM notifier structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	if (gpusvm->ops->notifier_alloc)
> +		notifier = gpusvm->ops->notifier_alloc();
> +	else
> +		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
> +
> +	if (!notifier)
> +		return ERR_PTR(-ENOMEM);
> +
> +	notifier->gpusvm = gpusvm;
> +	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
> +	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
> +	INIT_LIST_HEAD(&notifier->entry);
> +	notifier->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&notifier->range_list);
> +
> +	return notifier;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_free() - Free GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function frees the GPU SVM notifier structure.
> + */
> +static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
> +				     struct drm_gpusvm_notifier *notifier)
> +{
> +	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
> +
> +	if (gpusvm->ops->notifier_free)
> +		gpusvm->ops->notifier_free(notifier);
> +	else
> +		kfree(notifier);
> +}
> +
> +/**
> + * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_range structure.
> + */
> +static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_range, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_range_insert() - Insert GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function inserts the GPU SVM range into the notifier RB tree and list.
> + */
> +static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
> +				    struct drm_gpusvm_range *range)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	drm_gpusvm_notifier_lock(notifier->gpusvm);
> +	interval_tree_insert(&range->itree, &notifier->root);
> +
> +	node = rb_prev(&range->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_range(node))->entry;
> +	else
> +		head = &notifier->range_list;
> +
> +	list_add(&range->entry, head);
> +	drm_gpusvm_notifier_unlock(notifier->gpusvm);
> +}
> +
> +/**
> + * __drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This macro removes the GPU SVM range from the notifier RB tree and list.
> + */
> +static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
> +				      struct drm_gpusvm_range *range)
> +{
> +	interval_tree_remove(&range->itree, &notifier->root);
> +	list_del(&range->entry);
> +}
> +
> +/**
> + * drm_gpusvm_range_alloc() - Allocate GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @fault_addr: Fault address
> + * @chunk_size: Chunk size
> + * @migrate_devmem: Flag indicating whether to migrate device memory
> + *
> + * This function allocates and initializes the GPU SVM range structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_range *
> +drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
> +		       struct drm_gpusvm_notifier *notifier,
> +		       unsigned long fault_addr, unsigned long chunk_size,
> +		       bool migrate_devmem)
> +{
> +	struct drm_gpusvm_range *range;
> +
> +	if (gpusvm->ops->range_alloc)
> +		range = gpusvm->ops->range_alloc(gpusvm);
> +	else
> +		range = kzalloc(sizeof(*range), GFP_KERNEL);
> +
> +	if (!range)
> +		return ERR_PTR(-ENOMEM);
> +
> +	kref_init(&range->refcount);
> +	range->gpusvm = gpusvm;
> +	range->notifier = notifier;
> +	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
> +	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
> +	INIT_LIST_HEAD(&range->entry);
> +	range->notifier_seq = LONG_MAX;
> +	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
> +
> +	return range;
> +}
> +
> +/**
> + * drm_gpusvm_check_pages() - Check pages
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @start: Start address
> + * @end: End address
> + *
> + * Check if pages between start and end have been faulted in on the CPU. Use to
> + * prevent migration of pages without CPU backing store.
> + *
> + * Returns:
> + * True if pages have been faulted into CPU, False otherwise
> + */
> +static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
> +				   struct drm_gpusvm_notifier *notifier,
> +				   unsigned long start, unsigned long end)
> +{
> +	struct hmm_range hmm_range = {
> +		.default_flags = 0,
> +		.notifier = &notifier->notifier,
> +		.start = start,
> +		.end = end,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long *pfns;
> +	unsigned long npages = npages_in_range(start, end);
> +	int err, i;
> +
> +	mmap_assert_locked(gpusvm->mm);
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return false;
> +
> +	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +	hmm_range.hmm_pfns = pfns;
> +
> +	while (true) {
> +		err = hmm_range_fault(&hmm_range);
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	if (err)
> +		goto err_free;
> +
> +	for (i = 0; i < npages;) {
> +		if (!(pfns[i] & HMM_PFN_VALID)) {
> +			err = -EFAULT;
> +			goto err_free;
> +		}
> +		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
> +	}
> +
> +err_free:
> +	kvfree(pfns);
> +	return err ? false : true;
> +}
> +
> +/**
> + * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @vas: Pointer to the virtual memory area structure
> + * @fault_addr: Fault address
> + * @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
> + *
> + * 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
> + * memory area boundaries.
> + *
> + * Returns:
> + * Chunk size on success, LONG_MAX on failure.
> + */
> +static unsigned long
> +drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_notifier *notifier,
> +			    struct vm_area_struct *vas,
> +			    unsigned long fault_addr,
> +			    unsigned long gpuva_start,
> +			    unsigned long gpuva_end,
> +			    unsigned long check_pages_threshold)
> +{
> +	unsigned long start, end;
> +	int i = 0;
> +
> +retry:
> +	for (; i < gpusvm->num_chunks; ++i) {
> +		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
> +		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
> +
> +		if (start >= vas->vm_start && end <= vas->vm_end &&
> +		    start >= notifier->itree.start &&
> +		    end <= notifier->itree.last + 1 &&
> +		    start >= gpuva_start && end <= gpuva_end)
> +			break;
> +	}
> +
> +	if (i == gpusvm->num_chunks)
> +		return LONG_MAX;
> +
> +	/*
> +	 * If allocation more than page, ensure not to overlap with existing
> +	 * ranges.
> +	 */
> +	if (end - start != SZ_4K) {
> +		struct drm_gpusvm_range *range;
> +
> +		range = drm_gpusvm_range_find(notifier, start, end);
> +		if (range) {
> +			++i;
> +			goto retry;
> +		}
> +
> +		/*
> +		 * XXX: Only create range on pages CPU has faulted in. Without
> +		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
> +		 * process-many-malloc' fails. In the failure case, each process
> +		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
> +		 * ranges. When migrating the SVM ranges, some processes fail in
> +		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
> +		 * and then upon drm_gpusvm_range_get_pages device pages from
> +		 * other processes are collected + faulted in which creates all
> +		 * sorts of problems. Unsure exactly how this happening, also
> +		 * problem goes away if 'xe_exec_system_allocator --r
> +		 * process-many-malloc' mallocs at least 64k at a time.
> +		 */
> +		if (end - start <= check_pages_threshold &&
> +		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
> +			++i;
> +			goto retry;
> +		}
> +	}
> +
> +	return end - start;
> +}
> +
> +/**
> + * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + * @gpuva_start: Start address of GPUVA which mirrors CPU
> + * @gpuva_end: End address of GPUVA which mirrors CPU
> + * @ctx: GPU SVM context
> + *
> + * This function finds or inserts a newly allocated a GPU SVM range based on the
> + * fault address. Caller must hold a lock to protect range lookup and insertion.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +	struct drm_gpusvm_range *range;
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	bool notifier_alloc = false;
> +	unsigned long chunk_size;
> +	int err;
> +	bool migrate_devmem;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	if (fault_addr < gpusvm->mm_start ||
> +	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
> +		return ERR_PTR(-EINVAL);
> +
> +	if (!mmget_not_zero(mm))
> +		return ERR_PTR(-EFAULT);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
> +	if (!notifier) {
> +		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
> +		if (IS_ERR(notifier)) {
> +			err = PTR_ERR(notifier);
> +			goto err_mmunlock;
> +		}
> +		notifier_alloc = true;
> +		err = mmu_interval_notifier_insert(&notifier->notifier,
> +						   mm, notifier->itree.start,
> +						   notifier->itree.last -
> +						   notifier->itree.start + 1,
> +						   &drm_gpusvm_notifier_ops);
> +		if (err)
> +			goto err_notifier;
> +	}
> +
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, fault_addr);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_notifier_remove;
> +	}
> +
> +	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
> +		err = -EPERM;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
> +	if (range)
> +		goto out_mmunlock;
> +	/*
> +	 * XXX: Short-circuiting migration based on migrate_vma_* current
> +	 * limitations. If/when migrate_vma_* add more support, this logic will
> +	 * have to change.
> +	 */
> +	migrate_devmem = ctx->devmem_possible &&
> +		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
> +
> +	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
> +						 fault_addr, gpuva_start,
> +						 gpuva_end, migrate_devmem &&
> +						 ctx->check_pages_threshold);
> +	if (chunk_size == LONG_MAX) {
> +		err = -EINVAL;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
> +				       migrate_devmem);
> +	if (IS_ERR(range)) {
> +		err = PTR_ERR(range);
> +		goto err_notifier_remove;
> +	}
> +
> +	drm_gpusvm_range_insert(notifier, range);
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_insert(gpusvm, notifier);
> +
> +out_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +
> +	return range;
> +
> +err_notifier_remove:
> +	mmap_read_unlock(mm);
> +	if (notifier_alloc)
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +err_notifier:
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +err_mmunlock:
> +	mmput(mm);
> +	return ERR_PTR(err);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
> +
> +/**
> + * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @npages: Number of pages to unmap
> + *
> + * This function unmap pages associated with a GPU SVM range. Assumes and
> + * asserts correct locking is in place when called.
> + */
> +static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +					   struct drm_gpusvm_range *range,
> +					   unsigned long npages)
> +{
> +	unsigned long i, j;
> +	struct drm_pagemap *dpagemap = range->dpagemap;
> +	struct device *dev = gpusvm->drm->dev;
> +
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->flags.has_dma_mapping) {
> +		for (i = 0, j = 0; i < npages; j++) {
> +			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
> +
> +			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
> +				dma_unmap_page(dev,
> +					       addr->addr,
> +					       PAGE_SIZE << addr->order,
> +					       addr->dir);
> +			else if (dpagemap && dpagemap->ops->unmap_dma)
> +				dpagemap->ops->unmap_dma(dpagemap,
> +							 dev,
> +							 *addr);
> +			i += 1 << addr->order;
> +		}
> +		range->flags.has_devmem_pages = false;
> +		range->flags.has_dma_mapping = false;
> +		range->dpagemap = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function frees the dma address array associated with a GPU SVM range.
> + */
> +static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
> +					struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->dma_addr) {
> +		kvfree(range->dma_addr);
> +		range->dma_addr = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function removes the specified GPU SVM range and also removes the parent
> + * GPU SVM notifier if no more ranges remain in the notifier. The caller must
> + * hold a lock to protect range and notifier removal.
> + */
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
> +	if (WARN_ON_ONCE(!notifier))
> +		return;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +	drm_gpusvm_range_free_pages(gpusvm, range);
> +	__drm_gpusvm_range_remove(notifier, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	drm_gpusvm_range_put(range);
> +
> +	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
> +		if (!notifier->flags.removed)
> +			mmu_interval_notifier_remove(&notifier->notifier);
> +		drm_gpusvm_notifier_remove(gpusvm, notifier);
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +	}
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
> +
> +/**
> + * drm_gpusvm_range_get() - Get a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function increments the reference count of the specified GPU SVM range.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range)
> +{
> +	kref_get(&range->refcount);
> +
> +	return range;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
> +
> +/**
> + * drm_gpusvm_range_destroy() - Destroy GPU SVM range
> + * @refcount: Pointer to the reference counter embedded in the GPU SVM range
> + *
> + * This function destroys the specified GPU SVM range when its reference count
> + * reaches zero. If a custom range-free function is provided, it is invoked to
> + * free the range; otherwise, the range is deallocated using kfree().
> + */
> +static void drm_gpusvm_range_destroy(struct kref *refcount)
> +{
> +	struct drm_gpusvm_range *range =
> +		container_of(refcount, struct drm_gpusvm_range, refcount);
> +	struct drm_gpusvm *gpusvm = range->gpusvm;
> +
> +	if (gpusvm->ops->range_free)
> +		gpusvm->ops->range_free(range);
> +	else
> +		kfree(range);
> +}
> +
> +/**
> + * drm_gpusvm_range_put() - Put a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function decrements the reference count of the specified GPU SVM range
> + * and frees it when the count reaches zero.
> + */
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
> +{
> +	kref_put(&range->refcount, drm_gpusvm_range_destroy);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
> +
> +/**
> + * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called holding gpusvm->notifier_lock and as the last step before committing a
> + * GPU binding. This is akin to a notifier seqno check in the HMM documentation
> + * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
> + * function is required for finer grained checking (i.e., per range) if pages
> + * are valid.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
> +
> +/**
> + * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called without holding gpusvm->notifier_lock.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +static bool
> +drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
> +				      struct drm_gpusvm_range *range)
> +{
> +	bool pages_valid;
> +
> +	if (!range->dma_addr)
> +		return false;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
> +	if (!pages_valid)
> +		drm_gpusvm_range_free_pages(gpusvm, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	return pages_valid;
> +}
> +
> +/**
> + * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function gets pages for a GPU SVM range and ensures they are mapped for
> + * DMA access.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
> +	struct hmm_range hmm_range = {
> +		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
> +			HMM_PFN_REQ_WRITE),
> +		.notifier = notifier,
> +		.start = range->itree.start,
> +		.end = range->itree.last + 1,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct drm_gpusvm_zdd *zdd;
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long i, j;
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	unsigned long num_dma_mapped;
> +	unsigned int order = 0;
> +	unsigned long *pfns;
> +	struct page **pages;
> +	int err = 0;
> +	struct dev_pagemap *pagemap;
> +	struct drm_pagemap *dpagemap;
> +
> +retry:
> +	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
> +		goto set_seqno;
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return -ENOMEM;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	hmm_range.hmm_pfns = pfns;
> +	while (true) {
> +		mmap_read_lock(mm);
> +		err = hmm_range_fault(&hmm_range);
> +		mmap_read_unlock(mm);
> +
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	mmput(mm);
> +	if (err)
> +		goto err_free;
> +
> +	pages = (struct page **)pfns;
> +map_pages:
> +	/*
> +	 * Perform all dma mappings under the notifier lock to not
> +	 * access freed pages. A notifier will either block on
> +	 * the notifier lock or unmap dma.
> +	 */
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		kvfree(pfns);
> +		goto retry;
> +	}
> +
> +	if (!range->dma_addr) {
> +		/* Unlock and restart mapping to allocate memory. */
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
> +						 GFP_KERNEL);
> +		if (!range->dma_addr) {
> +			err = -ENOMEM;
> +			goto err_free;
> +		}
> +		goto map_pages;
> +	}
> +
> +	zdd = NULL;
> +	num_dma_mapped = 0;
> +	for (i = 0, j = 0; i < npages; ++j) {
> +		struct page *page = hmm_pfn_to_page(pfns[i]);
> +
> +		order = hmm_pfn_to_map_order(pfns[i]);
> +		if (is_device_private_page(page) || is_device_coherent_page(page)) {
> +			if (zdd != page->zone_device_data && i > 0) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +			zdd = page->zone_device_data;
> +			if (pagemap != page->pgmap) {
> +				if (i > 0) {
> +					err = -EOPNOTSUPP;
> +					goto err_unmap;
> +				}
> +
> +				pagemap = page->pgmap;
> +				dpagemap = zdd->devmem_allocation->dpagemap;
> +				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
> +					/*
> +					 * Raced. This is not supposed to happen
> +					 * since hmm_range_fault() should've migrated
> +					 * this page to system.
> +					 */
> +					err = -EAGAIN;
> +					goto err_unmap;
> +				}
> +			}
> +			range->dma_addr[j] =
> +				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
> +						       page, order,
> +						       DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			pages[i] = page;
> +		} else {
> +			dma_addr_t addr;
> +
> +			if (is_zone_device_page(page) || zdd) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +
> +			addr = dma_map_page(gpusvm->drm->dev,
> +					    page, 0,
> +					    PAGE_SIZE << order,
> +					    DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			range->dma_addr[j] = drm_pagemap_dma_addr_encode
> +				(addr, DRM_INTERCONNECT_SYSTEM, order,
> +				 DMA_BIDIRECTIONAL);
> +		}
> +		i += 1 << order;
> +		num_dma_mapped = i;
> +	}
> +
> +	range->flags.has_dma_mapping = true;
> +	if (zdd) {
> +		range->flags.has_devmem_pages = true;
> +		range->dpagemap = dpagemap;
> +	}
> +
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +	kvfree(pfns);
> +set_seqno:
> +	range->notifier_seq = hmm_range.notifier_seq;
> +
> +	return 0;
> +
> +err_unmap:
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +err_free:
> +	kvfree(pfns);
> +	if (err == -EAGAIN)
> +		goto retry;
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
> +
> +/**
> + * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function unmaps pages associated with a GPU SVM range. If @in_notifier
> + * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
> + * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
> + * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
> + * security model.
> + */
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +
> +	if (ctx->in_notifier)
> +		lockdep_assert_held_write(&gpusvm->notifier_lock);
> +	else
> +		drm_gpusvm_notifier_lock(gpusvm);
> +
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +
> +	if (!ctx->in_notifier)
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_page() - Put a migration page
> + * @page: Pointer to the page to put
> + *
> + * This function unlocks and puts a page.
> + */
> +static void drm_gpusvm_migration_unlock_put_page(struct page *page)
> +{
> +	unlock_page(page);
> +	put_page(page);
> +}
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
> + * @npages: Number of pages
> + * @migrate_pfn: Array of migrate page frame numbers
> + *
> + * This function unlocks and puts an array of pages.
> + */
> +static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
> +						  unsigned long *migrate_pfn)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page;
> +
> +		if (!migrate_pfn[i])
> +			continue;
> +
> +		page = migrate_pfn_to_page(migrate_pfn[i]);
> +		drm_gpusvm_migration_unlock_put_page(page);
> +		migrate_pfn[i] = 0;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
> + * @page: Pointer to the page
> + * @zdd: Pointer to the GPU SVM zone device data
> + *
> + * This function associates the given page with the specified GPU SVM zone
> + * device data and initializes it for zone device usage.
> + */
> +static void drm_gpusvm_get_devmem_page(struct page *page,
> +				     struct drm_gpusvm_zdd *zdd)
> +{
> +	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
> +	zone_device_page_init(page);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
> + * @dev: The device for which the pages are being mapped
> + * @dma_addr: Array to store DMA addresses corresponding to mapped pages
> + * @migrate_pfn: Array of migrate page frame numbers to map
> + * @npages: Number of pages to map
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function maps pages of memory for migration usage in GPU SVM. It
> + * iterates over each page frame number provided in @migrate_pfn, maps the
> + * corresponding page, and stores the DMA address in the provided @dma_addr
> + * array.
> + *
> + * Return: 0 on success, -EFAULT if an error occurs during mapping.
> + */
> +static int drm_gpusvm_migrate_map_pages(struct device *dev,
> +					dma_addr_t *dma_addr,
> +					unsigned long *migrate_pfn,
> +					unsigned long npages,
> +					enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		if (WARN_ON_ONCE(is_zone_device_page(page)))
> +			return -EFAULT;
> +
> +		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
> +		if (dma_mapping_error(dev, dma_addr[i]))
> +			return -EFAULT;
> +	}
> +
> +	return 0;
> +}
> +
> +/**
> + * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
> + * @dev: The device for which the pages were mapped
> + * @dma_addr: Array of DMA addresses corresponding to mapped pages
> + * @npages: Number of pages to unmap
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function unmaps previously mapped pages of memory for GPU Shared Virtual
> + * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
> + * if it's valid and not already unmapped, and unmaps the corresponding page.
> + */
> +static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
> +					   dma_addr_t *dma_addr,
> +					   unsigned long npages,
> +					   enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
> +			continue;
> +
> +		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @devmem_allocation: Pointer to the device memory allocation. The caller
> + *                     should hold a reference to the device memory allocation,
> + *                     which should be dropped via ops->devmem_release or upon
> + *                     the failure of this function.
> + * @ctx: GPU SVM context
> + *
> + * This function migrates the specified GPU SVM range to device memory. It performs the
> + * necessary setup and invokes the driver-specific operations for migration to
> + * device memory. Upon successful return, @devmem_allocation can safely reference @range
> + * until ops->devmem_release is called which only upon successful return.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +	struct migrate_vma migrate = {
> +		.start		= start,
> +		.end		= end,
> +		.pgmap_owner	= gpusvm->device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	unsigned long i, npages = npages_in_range(start, end);
> +	struct vm_area_struct *vas;
> +	struct drm_gpusvm_zdd *zdd = NULL;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int err;
> +
> +	if (!range->flags.migrate_devmem)
> +		return -EINVAL;
> +
> +	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
> +	    !ops->copy_to_ram)
> +		return -EOPNOTSUPP;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_out;
> +	}
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, start);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_mmunlock;
> +	}
> +
> +	if (end > vas->vm_end || start < vas->vm_start) {
> +		err = -EINVAL;
> +		goto err_mmunlock;
> +	}
> +
> +	if (!vma_is_anonymous(vas)) {
> +		err = -EBUSY;
> +		goto err_mmunlock;
> +	}
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_mmunlock;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
> +	if (!zdd) {
> +		err = -ENOMEM;
> +		goto err_free;
> +	}
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	if (!migrate.cpages) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	if (migrate.cpages != npages) {
> +		err = -EBUSY;
> +		goto err_finalize;
> +	}
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   migrate.src, npages, DMA_TO_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = pfn_to_page(migrate.dst[i]);
> +
> +		pages[i] = page;
> +		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
> +		drm_gpusvm_get_devmem_page(page, zdd);
> +	}
> +
> +	err = ops->copy_to_devmem(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +	/* Upon success bind devmem allocation to range and zdd */
> +	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_TO_DEVICE);
> +err_free:
> +	if (zdd)
> +		drm_gpusvm_zdd_put(zdd);
> +	kvfree(buf);
> +err_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +err_out:
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
> +
> +/**
> + * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
> + * @vas: Pointer to the VM area structure, can be NULL
> + * @fault_page: Fault page
> + * @npages: Number of pages to populate
> + * @mpages: Number of pages to migrate
> + * @src_mpfn: Source array of migrate PFNs
> + * @mpfn: Array of migrate PFNs to populate
> + * @addr: Start address for PFN allocation
> + *
> + * This function populates the RAM migrate page frame numbers (PFNs) for the
> + * specified VM area structure. It allocates and locks pages in the VM area for
> + * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
> + * alloc_page for allocation.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> +					       struct page *fault_page,
> +					       unsigned long npages,
> +					       unsigned long *mpages,
> +					       unsigned long *src_mpfn,
> +					       unsigned long *mpfn,
> +					       unsigned long addr)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
> +		struct page *page, *src_page;
> +
> +		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
> +			continue;
> +
> +		if (fault_page) {
> +			src_page = migrate_pfn_to_page(src_mpfn[i]);
> +			if (src_page->zone_device_data !=

if (src_page && ...)

This pops once the Kconfig issue Himal raised was fixed.

Matt

> +			    fault_page->zone_device_data)
> +				continue;
> +		}
> +
> +		if (vas)
> +			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
> +		else
> +			page = alloc_page(GFP_HIGHUSER);
> +
> +		if (!page)
> +			goto free_pages;
> +
> +		mpfn[i] = migrate_pfn(page_to_pfn(page));
> +	}
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		WARN_ON_ONCE(!trylock_page(page));
> +		++*mpages;
> +	}
> +
> +	return 0;
> +
> +free_pages:
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		put_page(page);
> +		mpfn[i] = 0;
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
> + * @devmem_allocation: Pointer to the device memory allocation
> + *
> + * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
> + * migration done via migrate_device_* functions.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	unsigned long *src, *dst;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int i, err = 0;
> +
> +	npages = devmem_allocation->size >> PAGE_SHIFT;
> +
> +retry:
> +	if (!mmget_not_zero(devmem_allocation->mm))
> +		return -EFAULT;
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	src = buf;
> +	dst = buf + (sizeof(*src) * npages);
> +	dma_addr = buf + (2 * sizeof(*src) * npages);
> +	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
> +	if (err)
> +		goto err_free;
> +
> +	err = migrate_device_pfns(src, npages);
> +	if (err)
> +		goto err_free;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
> +						  src, dst, 0);
> +	if (err || !mpages)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   dst, npages, DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, dst);
> +	migrate_device_pages(src, dst, npages);
> +	migrate_device_finalize(src, dst, npages);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +	mmput_async(devmem_allocation->mm);
> +	if (!err && !READ_ONCE(devmem_allocation->detached)) {
> +		cond_resched();
> +		goto retry;
> +	}
> +
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
> +
> +/**
> + * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
> + * @vas: Pointer to the VM area structure
> + * @device_private_page_owner: Device private pages owner
> + * @page: Pointer to the page for fault handling (can be NULL)
> + * @fault_addr: Fault address
> + * @size: Size of migration
> + *
> + * This internal function performs the migration of the specified GPU SVM range
> + * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
> + * invokes the driver-specific operations for migration to RAM.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
> +				       void *device_private_page_owner,
> +				       struct page *page,
> +				       unsigned long fault_addr,
> +				       unsigned long size)
> +{
> +	struct migrate_vma migrate = {
> +		.vma		= vas,
> +		.pgmap_owner	= device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
> +			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
> +		.fault_page	= page,
> +	};
> +	struct drm_gpusvm_zdd *zdd;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct device *dev;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	unsigned long start, end;
> +	void *buf;
> +	int i, err = 0;
> +
> +	start = ALIGN_DOWN(fault_addr, size);
> +	end = ALIGN(fault_addr + 1, size);
> +
> +	/* Corner where VMA area struct has been partially unmapped */
> +	if (start < vas->vm_start)
> +		start = vas->vm_start;
> +	if (end > vas->vm_end)
> +		end = vas->vm_end;
> +
> +	migrate.start = start;
> +	migrate.end = end;
> +	npages = npages_in_range(start, end);
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	/* Raced with another CPU fault, nothing to do */
> +	if (!migrate.cpages)
> +		goto err_free;
> +
> +	if (!page) {
> +		for (i = 0; i < npages; ++i) {
> +			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
> +				continue;
> +
> +			page = migrate_pfn_to_page(migrate.src[i]);
> +			break;
> +		}
> +
> +		if (!page)
> +			goto err_finalize;
> +	}
> +	zdd = page->zone_device_data;
> +	ops = zdd->devmem_allocation->ops;
> +	dev = zdd->devmem_allocation->dev;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
> +						  migrate.src, migrate.dst,
> +						  start);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
> +					   DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(migrate.src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +
> +	return err;
> +}
> +
> +/**
> + * drm_gpusvm_range_evict() - Evict GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function evicts the specified GPU SVM range.
> + */
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range)
> +{
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +
> +	if (!mmget_not_zero(mm))
> +		return;
> +
> +	mmap_read_lock(mm);
> +
> +	/*
> +	 * Loop required to find all VMAs for the corner case when partial
> +	 * unmappings have occurred.
> +	 */
> +again:
> +	vas = find_vma(mm, start);
> +	if (!vas)
> +		goto unlock;
> +
> +	if (end <= vas->vm_start || start >= vas->vm_end)
> +		goto unlock;
> +
> +	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
> +				    NULL, start,
> +				    range->itree.last - range->itree.start + 1);
> +	if (vas->vm_end < end) {
> +		start = vas->vm_end;
> +		goto again;
> +	}
> +unlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
> +
> +/**
> + * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
> + * @page: Pointer to the page
> + *
> + * This function is a callback used to put the GPU SVM zone device data
> + * associated with a page when it is being released.
> + */
> +static void drm_gpusvm_page_free(struct page *page)
> +{
> +	drm_gpusvm_zdd_put(page->zone_device_data);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
> + * @vmf: Pointer to the fault information structure
> + *
> + * This function is a page fault handler used to migrate a GPU SVM range to RAM.
> + * It retrieves the GPU SVM range information from the faulting page and invokes
> + * the internal migration function to migrate the range back to RAM.
> + *
> + * Returns:
> + * VM_FAULT_SIGBUS on failure, 0 on success.
> + */
> +static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
> +{
> +	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
> +	int err;
> +
> +	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
> +					  zdd->device_private_page_owner,
> +					  vmf->page, vmf->address,
> +					  zdd->devmem_allocation->size);
> +
> +	return err ? VM_FAULT_SIGBUS : 0;
> +}
> +
> +/**
> + * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
> + */
> +static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
> +	.page_free = drm_gpusvm_page_free,
> +	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
> +};
> +
> +/**
> + * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
> + *
> + * Returns:
> + * Pointer to the GPU SVM device page map operations structure.
> + */
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
> +{
> +	return &drm_gpusvm_pagemap_ops;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
> +
> +/**
> + * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @start: Start address
> + * @end: End address
> + *
> + * Returns:
> + * True if GPU SVM has mapping, False otherwise
> + */
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
> +		struct drm_gpusvm_range *range = NULL;
> +
> +		drm_gpusvm_for_each_range(range, notifier, start, end)
> +			return true;
> +	}
> +
> +	return false;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
> +
> +/**
> + * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
> + * @range: Pointer to the GPU SVM range structure.
> + * @mmu_range: Pointer to the MMU notifier range structure.
> + *
> + * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
> + * if the range partially falls within the provided MMU notifier range.
> + */
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range)
> +{
> +	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
> +
> +	range->flags.unmapped = true;
> +	if (range->itree.start < mmu_range->start ||
> +	    range->itree.last + 1 > mmu_range->end)
> +		range->flags.partial_unmap = true;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
> +
> +/**
> + * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap we're allocating from.
> + * @size: Size of device memory allocation
> + */
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size)
> +{
> +	devmem_allocation->dev = dev;
> +	devmem_allocation->mm = mm;
> +	devmem_allocation->ops = ops;
> +	devmem_allocation->dpagemap = dpagemap;
> +	devmem_allocation->size = size;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
> +
> +MODULE_DESCRIPTION("DRM GPUSVM");
> +MODULE_LICENSE("GPL");
> diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> new file mode 100644
> index 000000000000..2fb6980d0cf1
> --- /dev/null
> +++ b/include/drm/drm_gpusvm.h
> @@ -0,0 +1,445 @@
> +/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
> +/*
> + * Copyright © 2024 Intel Corporation
> + */
> +
> +#ifndef __DRM_GPUSVM_H__
> +#define __DRM_GPUSVM_H__
> +
> +#include <linux/kref.h>
> +#include <linux/interval_tree.h>
> +#include <linux/mmu_notifier.h>
> +
> +struct dev_pagemap_ops;
> +struct drm_device;
> +struct drm_gpusvm;
> +struct drm_gpusvm_notifier;
> +struct drm_gpusvm_ops;
> +struct drm_gpusvm_range;
> +struct drm_gpusvm_devmem;
> +struct drm_pagemap;
> +struct drm_pagemap_dma_addr;
> +
> +/**
> + * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM)
> + * device memory. These operations are provided by the GPU driver to manage device memory
> + * allocations and perform operations such as migration between device memory and system
> + * RAM.
> + */
> +struct drm_gpusvm_devmem_ops {
> +	/**
> +	 * @devmem_release: Release device memory allocation (optional)
> +	 * @devmem_allocation: device memory allocation
> +	 *
> +	 * Release device memory allocation and drop a reference to device
> +	 * memory allocation.
> +	 */
> +	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +	/**
> +	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
> +	 * @devmem_allocation: device memory allocation
> +	 * @npages: Number of pages to populate
> +	 * @pfn: Array of page frame numbers to populate
> +	 *
> +	 * Populate device memory page frame numbers (PFN).
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
> +				   unsigned long npages, unsigned long *pfn);
> +
> +	/**
> +	 * @copy_to_devmem: Copy to device memory (required for migration)
> +	 * @pages: Pointer to array of device memory pages (destination)
> +	 * @dma_addr: Pointer to array of DMA addresses (source)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to device memory.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_devmem)(struct page **pages,
> +			      dma_addr_t *dma_addr,
> +			      unsigned long npages);
> +
> +	/**
> +	 * @copy_to_ram: Copy to system RAM (required for migration)
> +	 * @pages: Pointer to array of device memory pages (source)
> +	 * @dma_addr: Pointer to array of DMA addresses (destination)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to system RAM.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_ram)(struct page **pages,
> +			   dma_addr_t *dma_addr,
> +			   unsigned long npages);
> +};
> +
> +/**
> + * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
> + * @size: Size of device memory allocation
> + * @detached: device memory allocations is detached from device pages
> + */
> +struct drm_gpusvm_devmem {
> +	struct device *dev;
> +	struct mm_struct *mm;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct drm_pagemap *dpagemap;
> +	size_t size;
> +	bool detached;
> +};
> +
> +/**
> + * struct drm_gpusvm_ops - Operations structure for GPU SVM
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM).
> + * These operations are provided by the GPU driver to manage SVM ranges and
> + * notifiers.
> + */
> +struct drm_gpusvm_ops {
> +	/**
> +	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
> +	 *
> +	 * Allocate a GPU SVM notifier.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
> +
> +	/**
> +	 * @notifier_free: Free a GPU SVM notifier (optional)
> +	 * @notifier: Pointer to the GPU SVM notifier to be freed
> +	 *
> +	 * Free a GPU SVM notifier.
> +	 */
> +	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
> +
> +	/**
> +	 * @range_alloc: Allocate a GPU SVM range (optional)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 *
> +	 * Allocate a GPU SVM range.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
> +
> +	/**
> +	 * @range_free: Free a GPU SVM range (optional)
> +	 * @range: Pointer to the GPU SVM range to be freed
> +	 *
> +	 * Free a GPU SVM range.
> +	 */
> +	void (*range_free)(struct drm_gpusvm_range *range);
> +
> +	/**
> +	 * @invalidate: Invalidate GPU SVM notifier (required)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 * @notifier: Pointer to the GPU SVM notifier
> +	 * @mmu_range: Pointer to the mmu_notifier_range structure
> +	 *
> +	 * Invalidate the GPU page tables. It can safely walk the notifier range
> +	 * RB tree/list in this function. Called while holding the notifier lock.
> +	 */
> +	void (*invalidate)(struct drm_gpusvm *gpusvm,
> +			   struct drm_gpusvm_notifier *notifier,
> +			   const struct mmu_notifier_range *mmu_range);
> +};
> +
> +/**
> + * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: MMU interval notifier
> + * @itree: Interval tree node for the notifier (inserted in GPU SVM)
> + * @entry: List entry to fast interval tree traversal
> + * @root: Cached root node of the RB tree containing ranges
> + * @range_list: List head containing of ranges in the same order they appear in
> + *              interval tree. This is useful to keep iterating ranges while
> + *              doing modifications to RB tree.
> + * @flags.removed: Flag indicating whether the MMU interval notifier has been
> + *                 removed
> + *
> + * This structure represents a GPU SVM notifier.
> + */
> +struct drm_gpusvm_notifier {
> +	struct drm_gpusvm *gpusvm;
> +	struct mmu_interval_notifier notifier;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	struct rb_root_cached root;
> +	struct list_head range_list;
> +	struct {
> +		u32 removed : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm_range - Structure representing a GPU SVM range
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier
> + * @refcount: Reference count for the range
> + * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
> + * @entry: List entry to fast interval tree traversal
> + * @notifier_seq: Notifier sequence number of the range's pages
> + * @dma_addr: DMA address array
> + * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
> + *            Note this is assuming only one drm_pagemap per range is allowed.
> + * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
> + * @flags.unmapped: Flag indicating if the range has been unmapped
> + * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
> + * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
> + * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
> + *
> + * This structure represents a GPU SVM range used for tracking memory ranges
> + * mapped in a DRM device.
> + */
> +struct drm_gpusvm_range {
> +	struct drm_gpusvm *gpusvm;
> +	struct drm_gpusvm_notifier *notifier;
> +	struct kref refcount;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	unsigned long notifier_seq;
> +	struct drm_pagemap_dma_addr *dma_addr;
> +	struct drm_pagemap *dpagemap;
> +	struct {
> +		/* All flags below must be set upon creation */
> +		u16 migrate_devmem : 1;
> +		/* All flags below must be set / cleared under notifier lock */
> +		u16 unmapped : 1;
> +		u16 partial_unmap : 1;
> +		u16 has_devmem_pages : 1;
> +		u16 has_dma_mapping : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm - GPU SVM structure
> + *
> + * @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
> + * @ops: Pointer to the operations structure for GPU SVM
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order.
> + * @num_chunks: Number of chunks
> + * @notifier_lock: Read-write semaphore for protecting notifier operations
> + * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
> + * @notifier_list: list head containing of notifiers in the same order they
> + *                 appear in interval tree. This is useful to keep iterating
> + *                 notifiers while doing modifications to RB tree.
> + *
> + * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
> + * memory ranges mapped in a DRM (Direct Rendering Manager) device.
> + *
> + * No reference counting is provided, as this is expected to be embedded in the
> + * driver VM structure along with the struct drm_gpuvm, which handles reference
> + * counting.
> + */
> +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;
> +	const struct drm_gpusvm_ops *ops;
> +	const unsigned long *chunk_sizes;
> +	int num_chunks;
> +	struct rw_semaphore notifier_lock;
> +	struct rb_root_cached root;
> +	struct list_head notifier_list;
> +#ifdef CONFIG_LOCKDEP
> +	/**
> +	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
> +	 * drm_gpusvm_range_remove with a driver provided lock.
> +	 */
> +	struct lockdep_map *lock_dep_map;
> +#endif
> +};
> +
> +/**
> + * struct drm_gpusvm_ctx - DRM GPU SVM context
> + *
> + * @check_pages_threshold: Check CPU pages for present if chunk is less than or
> + *                         equal to threshold. If not present, reduce chunk
> + *                         size.
> + * @in_notifier: entering from a MMU notifier
> + * @read_only: operating on read-only memory
> + * @devmem_possible: possible to use device memory
> + *
> + * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> + */
> +struct drm_gpusvm_ctx {
> +	unsigned long check_pages_threshold;
> +	unsigned int in_notifier :1;
> +	unsigned int read_only :1;
> +	unsigned int devmem_possible :1;
> +};
> +
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks);
> +
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
> +
> +void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
> +
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range);
> +
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
> +
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end);
> +
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range);
> +
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size);
> +
> +#ifdef CONFIG_LOCKDEP
> +/**
> + * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @lock: the lock used to protect the gpuva list. The locking primitive
> + * must contain a dep_map field.
> + *
> + * Call this to annotate drm_gpusvm_range_find_or_insert and
> + * drm_gpusvm_range_remove.
> + */
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
> +	do { \
> +		if (!WARN((gpusvm)->lock_dep_map, \
> +			  "GPUSVM range lock should be set only once."))\
> +			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
> +	} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) \
> +	do { \
> +		if ((gpusvm)->lock_dep_map)	\
> +			lock_is_held((gpusvm)->lock_dep_map);	\
> +	} while (0)
> +#else
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
> +#endif
> +
> +/**
> + * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, take lock
> + */
> +#define drm_gpusvm_notifier_lock(gpusvm__)	\
> +	down_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, drop lock
> + */
> +#define drm_gpusvm_notifier_unlock(gpusvm__)	\
> +	up_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
> + * @range: a pointer to the current GPU SVM range
> + *
> + * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
> + *         current range is the last one or if the input range is NULL.
> + */
> +static inline struct drm_gpusvm_range *
> +__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
> +{
> +	if (range && !list_is_last(&range->entry,
> +				   &range->notifier->range_list))
> +		return list_next_entry(range, entry);
> +
> +	return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges. If set, it indicates the start of
> + *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
> + * to use while holding the driver SVM lock or the notifier lock.
> + */
> +#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
> +	for ((range__) = (range__) ?:					\
> +	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
> +	     (range__) && (range__->itree.start < (end__));		\
> +	     (range__) = __drm_gpusvm_range_next(range__))
> +
> +#endif /* __DRM_GPUSVM_H__ */
> -- 
> 2.34.1
> 

Re: [PATCH v3 00/30] Introduce GPU SVM and Xe SVM implementation

[Gwan-gyeong Mun]

Hi Matthew Brost,

After applying this patch series and the following to the latest 
drm-tip, while testing[1] with the mentioned IGT, I faced a kernel oops[3].
It makes prevent progressing of the mentioned igt tests.
Could you please check the following oops log?

(1) apply comments of "[v3,05/30] drm/gpusvm: Add support for GPU Shared 
Virtual Memory"
(2) apply comments of "[v3,15/30] drm/xe: Add unbind to SVM garbage 
collector"
(3) drop "[v3,27/30] drm/xe: Basic SVM BO eviction" patch

The kernel config used, the entire dmesg, and detailed information can 
be found in [2].

br,

G.G.

[1] used igt command: xe_exec_system_allocator --run-subtest once-malloc
[2] https://gitlab.freedesktop.org/elongbug/drm-tip/-/snippets/7823

[3] kernel oops dmesg
[   51.365230] Console: switching to colour VGA+ 80x25
[   51.367772] [IGT] xe_exec_system_allocator: executing
[   51.383611] [IGT] xe_exec_system_allocator: starting subtest once-malloc
[   51.386066] xe 0000:00:04.0: [drm:vm_bind_ioctl_ops_create [xe]] 
op=0, addr=0x0000000000000000, range=0x0001000000000000, 
bo_offset_or_userptr=0x0000000000000000
[   51.386171] xe 0000:00:04.0: [drm:vm_bind_ioctl_ops_create [xe]] MAP: 
addr=0x0000000000000000, range=0x0001000000000000
[   51.389429] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] PAGE 
FAULT: asid=1, gpusvm=0xffff8881775e9188, vram=0,0, 
seqno=9223372036854775807, start=0x005584e8400000, end=0x005584e8410000, 
size=65536
[   51.389529] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] 
ALLOCATE VRAM: asid=1, gpusvm=0xffff8881775e9188, vram=0,0, 
seqno=9223372036854775807, start=0x005584e8400000, end=0x005584e8410000, 
size=65536
[   51.389935] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] ALLOC 
VRAM: asid=1, gpusvm=0xffff8881775e9188, pfn=3126960, npages=16
[   51.390048] xe 0000:00:04.0: [drm:xe_svm_invalidate [xe]] INVALIDATE: 
asid=1, gpusvm=0xffff8881775e9188, seqno=3, start=0x00005584e8400000, 
end=0x00005584e8410000, event=6
[   51.390440] xe 0000:00:04.0: [drm:xe_svm_invalidate [xe]] NOTIFIER: 
asid=1, gpusvm=0xffff8881775e9188, vram=0,0, seqno=9223372036854775807, 
start=0x005584e8400000, end=0x005584e8410000, size=65536
[   51.390948] Oops: general protection fault, probably for 
non-canonical address 0x3fff88842fc80000: 0000 [#1] PREEMPT SMP NOPTI
[   51.391624] CPU: 1 UID: 0 PID: 76 Comm: kworker/u17:0 Not tainted 
6.13.0-rc4-drm-tip-test+ #48
[   51.392088] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), 
BIOS 1.15.0-1 04/01/2014
[   51.392527] Workqueue: xe_gt_page_fault_work_queue pf_queue_work_func 
[xe]
[   51.392947] RIP: 0010:zone_device_page_init+0x5d/0x240
[   51.393228] Code: 04 dd ff e8 d5 d2 a1 00 5a 85 c0 0f 85 ba 00 00 00 
e8 d7 bb df ff 85 c0 0f 84 9d 01 00 00 48 8b 45 38 a8 03 0f 85 ec 00 00 
00 <65> 48 ff 00 e8 aa d2 a1 00 85 c0 0f 85 0d 01 00 00 48 c7 c7 20 cb
[   51.394247] RSP: 0018:ffffc9000039fb48 EFLAGS: 00010246
[   51.394570] RAX: 4000000000000000 RBX: ffffea000bedac00 RCX: 
0000000000000000
[   51.394950] RDX: 0000000000000046 RSI: ffffffff824c67b4 RDI: 
ffffffff824e58f5
[   51.395328] RBP: ffffea000bedac08 R08: 0000000000000015 R09: 
0000000000000004
[   51.395709] R10: 0000000000000001 R11: 0000000000000004 R12: 
0000000000000001
[   51.396093] R13: ffff888170fd8d40 R14: ffff88817f922640 R15: 
ffffea000bedac00
[   51.396472] FS:  0000000000000000(0000) GS:ffff88842fc80000(0000) 
knlGS:0000000000000000
[   51.396925] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   51.397237] CR2: 0000563d1f7ecbe4 CR3: 000000017c212000 CR4: 
0000000000750ef0
[   51.397618] PKRU: 55555554
[   51.397768] Call Trace:
[   51.397904]  <TASK>
[   51.398024]  ? __die_body.cold+0x19/0x26
[   51.398238]  ? die_addr+0x38/0x60
[   51.398420]  ? exc_general_protection+0x19e/0x450
[   51.398678]  ? asm_exc_general_protection+0x22/0x30
[   51.398942]  ? zone_device_page_init+0x5d/0x240
[   51.399188]  ? zone_device_page_init+0x49/0x240
[   51.399433]  drm_gpusvm_migrate_to_devmem+0x379/0x9e0 [drm_gpusvm]
[   51.399768]  xe_svm_handle_pagefault+0x62c/0xa60 [xe]
[   51.400110]  ? xe_vm_find_overlapping_vma+0xa4/0x1d0 [xe]
[   51.400475]  pf_queue_work_func+0x1ba/0x450 [xe]
[   51.400777]  process_one_work+0x1fe/0x580
[   51.400996]  worker_thread+0x1d1/0x3b0
[   51.401201]  ? __pfx_worker_thread+0x10/0x10
[   51.401433]  kthread+0xeb/0x120
[   51.401609]  ? __pfx_kthread+0x10/0x10
[   51.401813]  ret_from_fork+0x2d/0x50
[   51.402008]  ? __pfx_kthread+0x10/0x10
[   51.402211]  ret_from_fork_asm+0x1a/0x30
[   51.402427]  </TASK>
[   51.402551] Modules linked in: xe drm_ttm_helper gpu_sched 
drm_suballoc_helper drm_gpuvm drm_exec drm_gpusvm i2c_algo_bit drm_buddy 
video wmi ttm drm_display_helper drm_kms_helper crct10dif_pclmul 
crc32_pclmul e1000 ghash_clmulni_intel i2c_piix4 i2c_smbus fuse
[   51.403779] ---[ end trace 0000000000000000 ]---
[   51.404106] RIP: 0010:zone_device_page_init+0x5d/0x240
[   51.404393] Code: 04 dd ff e8 d5 d2 a1 00 5a 85 c0 0f 85 ba 00 00 00 
e8 d7 bb df ff 85 c0 0f 84 9d 01 00 00 48 8b 45 38 a8 03 0f 85 ec 00 00 
00 <65> 48 ff 00 e8 aa d2 a1 00 85 c0 0f 85 0d 01 00 00 48 c7 c7 20 cb
[   51.405408] RSP: 0018:ffffc9000039fb48 EFLAGS: 00010246
[   51.405725] RAX: 4000000000000000 RBX: ffffea000bedac00 RCX: 
0000000000000000
[   51.406110] RDX: 0000000000000046 RSI: ffffffff824c67b4 RDI: 
ffffffff824e58f5
[   51.406518] RBP: ffffea000bedac08 R08: 0000000000000015 R09: 
0000000000000004
[   51.406905] R10: 0000000000000001 R11: 0000000000000004 R12: 
0000000000000001
[   51.407312] R13: ffff888170fd8d40 R14: ffff88817f922640 R15: 
ffffea000bedac00
[   51.407691] FS:  0000000000000000(0000) GS:ffff88842fc80000(0000) 
knlGS:0000000000000000
[   51.408135] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
[   51.408484] CR2: 0000563d1f7ecbe4 CR3: 000000017c212000 CR4: 
0000000000750ef0
[   51.408877] PKRU: 55555554
[   51.409047] BUG: sleeping function called from invalid context at 
./include/linux/percpu-rwsem.h:49
[   51.409528] in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 
76, name: kworker/u17:0
[   51.409976] preempt_count: 0, expected: 0
[   51.410212] RCU nest depth: 1, expected: 0
[   51.410435] INFO: lockdep is turned off.
[   51.410648] CPU: 1 UID: 0 PID: 76 Comm: kworker/u17:0 Tainted: G 
D            6.13.0-rc4-drm-tip-test+ #48
[   51.411180] Tainted: [D]=DIE
[   51.411338] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), 
BIOS 1.15.0-1 04/01/2014
[   51.411859] Workqueue: xe_gt_page_fault_work_queue pf_queue_work_func 
[xe]
[   51.412269] Call Trace:
[   51.412404]  <TASK>
[   51.412525]  dump_stack_lvl+0x69/0xa0
[   51.412724]  __might_resched.cold+0xe5/0x120
[   51.412956]  exit_signals+0x1a/0x360
[   51.413150]  do_exit+0x122/0xbd0
[   51.413328]  ? __pfx_worker_thread+0x10/0x10
[   51.413562]  make_task_dead+0x88/0x90
[   51.413783]  rewind_stack_and_make_dead+0x16/0x20
[   51.414045] RIP: 0000:0x0
[   51.414191] Code: Unable to access opcode bytes at 0xffffffffffffffd6.
[   51.414595] RSP: 0000:0000000000000000 EFLAGS: 00000000 ORIG_RAX: 
0000000000000000
[   51.414993] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 
0000000000000000
[   51.415369] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 
0000000000000000
[   51.415746] RBP: 0000000000000000 R08: 0000000000000000 R09: 
0000000000000000
[   51.416123] R10: 0000000000000000 R11: 0000000000000000 R12: 
0000000000000000
[   51.416501] R13: 0000000000000000 R14: 0000000000000000 R15: 
0000000000000000
[   51.416899]  </TASK>


On 12/18/24 1:33 AM, Matthew Brost wrote:
> Version 3 of GPU SVM has been promoted to the proper series from an RFC.
> Thanks to everyone (especially Sima and Thomas) for their numerous
> reviews on revision 1, 2 and for helping to address many design issues.
> 
> This version has been tested with IGT [1] on PVC, BMG, and LNL. Also
> tested with level0 (UMD) PR [2].
> 
> Major changes in v2:
> - Dropped mmap write abuse
> - core MM locking and retry loops instead of driver locking to avoid races
> - Removed physical to virtual references
> - Embedded structure/ops for drm_gpusvm_devmem
> - Fixed mremap and fork issues
> - Added DRM pagemap
> - Included RFC documentation in the kernel doc
> 
> Major changes in v3:
> - Move GPU SVM and DRM pagemap to DRM level
> - Mostly addresses Thomas's feedback, lots of small changes documented
>    in each individual patch change log
> 
> Known issues in v3:
> - Check pages still exists, changed to threshold in this version which
>    is better but still need to root cause cross process page finding on
>    small user allocations.
> - Dropped documentation patch, fairly large rewrite and will send out
>    independently once finished.
> 
> Matt
> 
> [1] https://patchwork.freedesktop.org/series/137545/#rev3
> [2] https://github.com/intel/compute-runtime/pull/782
> 
> Matthew Brost (27):
>    drm/xe: Retry BO allocation
>    mm/migrate: Add migrate_device_pfns
>    mm/migrate: Trylock device page in do_swap_page
>    drm/gpusvm: Add support for GPU Shared Virtual Memory
>    drm/xe: Select DRM_GPUSVM Kconfig
>    drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag
>    drm/xe: Add SVM init / close / fini to faulting VMs
>    drm/xe: Nuke VM's mapping upon close
>    drm/xe: Add SVM range invalidation and page fault handler
>    drm/gpuvm: Add DRM_GPUVA_OP_DRIVER
>    drm/xe: Add (re)bind to SVM page fault handler
>    drm/xe: Add SVM garbage collector
>    drm/xe: Add unbind to SVM garbage collector
>    drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has
>      bindings
>    drm/xe: Enable CPU address mirror uAPI
>    drm/xe: Add migrate layer functions for SVM support
>    drm/xe: Add SVM device memory mirroring
>    drm/xe: Add drm_gpusvm_devmem to xe_bo
>    drm/xe: Add GPUSVM device memory copy vfunc functions
>    drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc
>    drm/xe: Add Xe SVM devmem_release GPU SVM vfunc
>    drm/xe: Add BO flags required for SVM
>    drm/xe: Add SVM VRAM migration
>    drm/xe: Basic SVM BO eviction
>    drm/xe: Add SVM debug
>    drm/xe: Add modparam for SVM notifier size
>    drm/xe: Add always_migrate_to_vram modparam
> 
> Thomas Hellström (3):
>    drm/pagemap: Add DRM pagemap
>    drm/xe: Add dma_addr res cursor
>    drm/xe: Add drm_pagemap ops to SVM
> 
>   drivers/gpu/drm/Kconfig                     |    8 +
>   drivers/gpu/drm/Makefile                    |    1 +
>   drivers/gpu/drm/drm_gpusvm.c                | 2220 +++++++++++++++++++
>   drivers/gpu/drm/xe/Kconfig                  |   10 +
>   drivers/gpu/drm/xe/Makefile                 |    1 +
>   drivers/gpu/drm/xe/xe_bo.c                  |   20 +-
>   drivers/gpu/drm/xe/xe_bo.h                  |    1 +
>   drivers/gpu/drm/xe/xe_bo_types.h            |    4 +
>   drivers/gpu/drm/xe/xe_device_types.h        |   15 +
>   drivers/gpu/drm/xe/xe_gt_pagefault.c        |   17 +-
>   drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c |   24 +
>   drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h |    2 +
>   drivers/gpu/drm/xe/xe_migrate.c             |  175 ++
>   drivers/gpu/drm/xe/xe_migrate.h             |   10 +
>   drivers/gpu/drm/xe/xe_module.c              |    7 +
>   drivers/gpu/drm/xe/xe_module.h              |    2 +
>   drivers/gpu/drm/xe/xe_pt.c                  |  393 +++-
>   drivers/gpu/drm/xe/xe_pt.h                  |    5 +
>   drivers/gpu/drm/xe/xe_pt_types.h            |    2 +
>   drivers/gpu/drm/xe/xe_res_cursor.h          |  116 +-
>   drivers/gpu/drm/xe/xe_svm.c                 |  948 ++++++++
>   drivers/gpu/drm/xe/xe_svm.h                 |   83 +
>   drivers/gpu/drm/xe/xe_tile.c                |    5 +
>   drivers/gpu/drm/xe/xe_vm.c                  |  375 +++-
>   drivers/gpu/drm/xe/xe_vm.h                  |   15 +-
>   drivers/gpu/drm/xe/xe_vm_types.h            |   57 +
>   include/drm/drm_gpusvm.h                    |  445 ++++
>   include/drm/drm_gpuvm.h                     |    5 +
>   include/drm/drm_pagemap.h                   |  103 +
>   include/linux/migrate.h                     |    1 +
>   include/uapi/drm/xe_drm.h                   |   19 +-
>   mm/memory.c                                 |   13 +-
>   mm/migrate_device.c                         |  116 +-
>   33 files changed, 5061 insertions(+), 157 deletions(-)
>   create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>   create mode 100644 drivers/gpu/drm/xe/xe_svm.c
>   create mode 100644 drivers/gpu/drm/xe/xe_svm.h
>   create mode 100644 include/drm/drm_gpusvm.h
>   create mode 100644 include/drm/drm_pagemap.h
> 

Re: [PATCH v3 01/30] drm/xe: Retry BO allocation

[Gwan-gyeong Mun]

On 12/18/24 1:33 AM, Matthew Brost wrote:
> TTM doesn't support fair eviction via WW locking, this mitigated in by
> using retry loops in exec and preempt rebind worker. Extend this retry
> loop to BO allocation. Once TTM supports fair eviction this patch can be
> reverted.
> 
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>   drivers/gpu/drm/xe/xe_bo.c | 8 +++++++-
>   1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
> index e6c896ad5602..a670a0444e03 100644
> --- a/drivers/gpu/drm/xe/xe_bo.c
> +++ b/drivers/gpu/drm/xe/xe_bo.c
> @@ -2142,6 +2142,7 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
>   	struct xe_file *xef = to_xe_file(file);
>   	struct drm_xe_gem_create *args = data;
>   	struct xe_vm *vm = NULL;
> +	ktime_t end = 0;
>   	struct xe_bo *bo;
>   	unsigned int bo_flags;
>   	u32 handle;
> @@ -2214,11 +2215,14 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
>   		vm = xe_vm_lookup(xef, args->vm_id);
>   		if (XE_IOCTL_DBG(xe, !vm))
>   			return -ENOENT;
> +	}
> +
> +retry:
> +	if (vm) {
>   		err = xe_vm_lock(vm, true);
>   		if (err)
>   			goto out_vm;
>   	}
> -
>   	bo = xe_bo_create_user(xe, NULL, vm, args->size, args->cpu_caching,
>   			       bo_flags);
>   
> @@ -2227,6 +2231,8 @@ int xe_gem_create_ioctl(struct drm_device *dev, void *data,
>   
>   	if (IS_ERR(bo)) {
>   		err = PTR_ERR(bo);
> +		if (xe_vm_validate_should_retry(NULL, err, &end))
> +			goto retry;
>   		goto out_vm;
>   	}
>   
looks good to me.

Reviewed-by: Gwan-gyeong Mun <gwan-gyeong.mun@intel.com>

Re: [PATCH v3 01/30] drm/xe: Retry BO allocation

[Summers, Stuart]

On Tue, 2024-12-17 at 15:33 -0800, Matthew Brost wrote:
> TTM doesn't support fair eviction via WW locking, this mitigated in
> by
> using retry loops in exec and preempt rebind worker. Extend this
> retry
> loop to BO allocation. Once TTM supports fair eviction this patch can
> be
> reverted.
> 
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>  drivers/gpu/drm/xe/xe_bo.c | 8 +++++++-
>  1 file changed, 7 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
> index e6c896ad5602..a670a0444e03 100644
> --- a/drivers/gpu/drm/xe/xe_bo.c
> +++ b/drivers/gpu/drm/xe/xe_bo.c
> @@ -2142,6 +2142,7 @@ int xe_gem_create_ioctl(struct drm_device *dev,
> void *data,
>         struct xe_file *xef = to_xe_file(file);
>         struct drm_xe_gem_create *args = data;
>         struct xe_vm *vm = NULL;
> +       ktime_t end = 0;
>         struct xe_bo *bo;
>         unsigned int bo_flags;
>         u32 handle;
> @@ -2214,11 +2215,14 @@ int xe_gem_create_ioctl(struct drm_device
> *dev, void *data,
>                 vm = xe_vm_lookup(xef, args->vm_id);
>                 if (XE_IOCTL_DBG(xe, !vm))
>                         return -ENOENT;
> +       }
> +
> +retry:
> +       if (vm) {
>                 err = xe_vm_lock(vm, true);
>                 if (err)
>                         goto out_vm;
>         }
> -

Feel like we could keep this line break, but not a blocker. Looks good,
thanks!

Reviewed-by: Stuart Summers <stuart.summers@intel.com>

>         bo = xe_bo_create_user(xe, NULL, vm, args->size, args-
> >cpu_caching,
>                                bo_flags);
>  
> @@ -2227,6 +2231,8 @@ int xe_gem_create_ioctl(struct drm_device *dev,
> void *data,
>  
>         if (IS_ERR(bo)) {
>                 err = PTR_ERR(bo);
> +               if (xe_vm_validate_should_retry(NULL, err, &end))
> +                       goto retry;
>                 goto out_vm;
>         }
>  

Re: [PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Matthew Brost]

On Tue, Dec 17, 2024 at 03:33:23PM -0800, Matthew Brost wrote:
> This patch introduces support for GPU Shared Virtual Memory (SVM) in the
> Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
> sharing of memory between the CPU and GPU, enhancing performance and
> flexibility in GPU computing tasks.
> 
> The patch adds the necessary infrastructure for SVM, including data
> structures and functions for managing SVM ranges and notifiers. It also
> provides mechanisms for allocating, deallocating, and migrating memory
> regions between system RAM and GPU VRAM.
> 
> This is largely inspired by GPUVM.
> 
> v2:
>  - Take order into account in check pages
>  - Clear range->pages in get pages error
>  - Drop setting dirty or accessed bit in get pages (Vetter)
>  - Remove mmap assert for cpu faults
>  - Drop mmap write lock abuse (Vetter, Christian)
>  - Decouple zdd from range (Vetter, Oak)
>  - Add drm_gpusvm_range_evict, make it work with coherent pages
>  - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
>  - mmget/put in drm_gpusvm_evict_to_sram
>  - Drop range->vram_alloation variable
>  - Don't return in drm_gpusvm_evict_to_sram until all pages detached
>  - Don't warn on mixing sram and device pages
>  - Update kernel doc
>  - Add coherent page support to get pages
>  - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
>  - Add struct drm_gpusvm_vram and ops (Thomas)
>  - Update the range's seqno if the range is valid (Thomas)
>  - Remove the is_unmapped check before hmm_range_fault (Thomas)
>  - Use drm_pagemap (Thomas)
>  - Drop kfree_mapping (Thomas)
>  - dma mapp pages under notifier lock (Thomas)
>  - Remove ctx.prefault
>  - Remove ctx.mmap_locked
>  - Add ctx.check_pages
>  - s/vram/devmem (Thomas)
> v3:
>  - Fix memory leak drm_gpusvm_range_get_pages
>  - Only migrate pages with same zdd on CPU fault
>  - Loop over al VMAs in drm_gpusvm_range_evict
>  - Make GPUSVM a drm level module
>  - GPL or MIT license
>  - Update main kernel doc (Thomas)
>  - Prefer foo() vs foo for functions in kernel doc (Thomas)
>  - Prefer functions over macros (Thomas)
>  - Use unsigned long vs u64 for addresses (Thomas)
>  - Use standard interval_tree (Thomas)
>  - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
>  - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
>  - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
>  - Newlines between functions defs in header file (Thomas)
>  - Drop shall language in driver vfunc kernel doc (Thomas)
>  - Move some static inlines from head to C file (Thomas)
>  - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
>  - Change check_pages to a threshold
> 
> Cc: Simona Vetter <simona.vetter@ffwll.ch>
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Christian König <christian.koenig@amd.com>
> Cc: <dri-devel@lists.freedesktop.org>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
>  drivers/gpu/drm/Kconfig      |    8 +
>  drivers/gpu/drm/Makefile     |    1 +
>  drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
>  include/drm/drm_gpusvm.h     |  445 +++++++
>  4 files changed, 2674 insertions(+)
>  create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>  create mode 100644 include/drm/drm_gpusvm.h
> 
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 2f51546b0b88..5524cb1ed9dd 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -278,6 +278,14 @@ config DRM_GPUVM
>  	  GPU-VM representation providing helpers to manage a GPUs virtual
>  	  address space
>  
> +config DRM_GPUSVM
> +	tristate
> +	depends on DRM
> +	depends on DEVICE_MIGRATION
> +	help
> +	  GPU-SVM representation providing helpers to manage a GPUs shared
> +	  virtual memory
> +
>  config DRM_BUDDY
>  	tristate
>  	depends on DRM
> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> index 19fb370fbc56..ca5d44441fb6 100644
> --- a/drivers/gpu/drm/Makefile
> +++ b/drivers/gpu/drm/Makefile
> @@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
>  #
>  obj-$(CONFIG_DRM_EXEC) += drm_exec.o
>  obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> +obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
>  
>  obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
>  
> diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> new file mode 100644
> index 000000000000..29ce2c7d1c8c
> --- /dev/null
> +++ b/drivers/gpu/drm/drm_gpusvm.c
> @@ -0,0 +1,2220 @@
> +// SPDX-License-Identifier: GPL-2.0-only OR MIT
> +/*
> + * Copyright © 2024 Intel Corporation
> + *
> + * Authors:
> + *     Matthew Brost <matthew.brost@intel.com>
> + */
> +
> +#include <linux/dma-mapping.h>
> +#include <linux/hmm.h>
> +#include <linux/memremap.h>
> +#include <linux/migrate.h>
> +#include <linux/mm_types.h>
> +#include <linux/pagemap.h>
> +#include <linux/slab.h>
> +
> +#include <drm/drm_device.h>
> +#include <drm/drm_gpusvm.h>
> +#include <drm/drm_pagemap.h>
> +#include <drm/drm_print.h>
> +
> +/**
> + * DOC: Overview
> + *
> + * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
> + *
> + * The GPU SVM layer is a component of the DRM framework designed to manage shared
> + * virtual memory between the CPU and GPU. It enables efficient data exchange and
> + * processing for GPU-accelerated applications by allowing memory sharing and
> + * synchronization between the CPU's and GPU's virtual address spaces.
> + *
> + * Key GPU SVM Components:
> + * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
> + *		GPU of changes, notifiers are sized based on a GPU SVM
> + *		initialization parameter, with a recommendation of 512M or
> + *		larger. They maintain a Red-BlacK tree and a list of ranges that
> + *		fall within the notifier interval. Notifiers are tracked within
> + *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
> + *		or removed as ranges within the interval are created or
> + *		destroyed.
> + * - Ranges: Represent memory ranges mapped in a DRM device and managed
> + *	     by GPU SVM. They are sized based on an array of chunk sizes, which
> + *	     is a GPU SVM initialization parameter, and the CPU address space.
> + *	     Upon GPU fault, the largest aligned chunk that fits within the
> + *	     faulting CPU address space is chosen for the range size. Ranges are
> + *	     expected to be dynamically allocated on GPU fault and removed on an
> + *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
> + *	     a notifier's Red-Black tree.
> + * - Operations: Define the interface for driver-specific GPU SVM operations
> + *               such as range allocation, notifier allocation, and
> + *               invalidations.
> + * - Device Memory Allocations: Embedded structure containing enough information
> + *                              for GPU SVM to migrate to / from device memory.
> + * - Device Memory Operations: Define the interface for driver-specific device
> + *                             memory operations release memory, populate pfns,
> + *                             and copy to / from device memory.
> + *
> + * This layer provides interfaces for allocating, mapping, migrating, and
> + * releasing memory ranges between the CPU and GPU. It handles all core memory
> + * management interactions (DMA mapping, HMM, and migration) and provides
> + * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
> + * to build the expected driver components for an SVM implementation as detailed
> + * below.
> + *
> + * Expected Driver Components:
> + * - GPU page fault handler: Used to create ranges and notifiers based on the
> + *			     fault address, optionally migrate the range to
> + *			     device memory, and create GPU bindings.
> + * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
> + *			Ranges are expected to be added to the garbage collector
> + *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
> + * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
> + *			ranges.
> + */
> +
> +/**
> + * DOC: Locking
> + *
> + * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
> + * mmap lock as needed.
> + *
> + * GPU SVM introduces a global notifier lock, which safeguards the notifier's
> + * range RB tree and list, as well as the range's DMA mappings and sequence
> + * number. GPU SVM manages all necessary locking and unlocking operations,
> + * except for the recheck range's pages being valid
> + * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
> + * lock corresponds to the 'driver->update' lock mentioned in the HMM
> + * documentation (TODO: Link). Future revisions may transition from a GPU SVM
> + * global lock to a per-notifier lock if finer-grained locking is deemed
> + * necessary.
> + *
> + * In addition to the locking mentioned above, the driver should implement a
> + * lock to safeguard core GPU SVM function calls that modify state, such as
> + * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
> + * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
> + * locking should also be possible for concurrent GPU fault processing within a
> + * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
> + * to add annotations to GPU SVM.
> + */
> +
> +/**
> + * DOC: Migration
> + *
> + * The migration support is quite simple, allowing migration between RAM and
> + * device memory at the range granularity. For example, GPU SVM currently does not
> + * support mixing RAM and device memory pages within a range. This means that upon GPU
> + * fault, the entire range can be migrated to device memory, and upon CPU fault, the
> + * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
> + * could be added in the future if required.
> + *
> + * The reasoning for only supporting range granularity is as follows: it
> + * simplifies the implementation, and range sizes are driver-defined and should
> + * be relatively small.
> + */
> +
> +/**
> + * DOC: Partial Unmapping of Ranges
> + *
> + * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
> + * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
> + * being that a subset of the range still has CPU and GPU mappings. If the
> + * backing store for the range is in device memory, a subset of the backing store has
> + * references. One option would be to split the range and device memory backing store,
> + * but the implementation for this would be quite complicated. Given that
> + * partial unmappings are rare and driver-defined range sizes are relatively
> + * small, GPU SVM does not support splitting of ranges.
> + *
> + * With no support for range splitting, upon partial unmapping of a range, the
> + * driver is expected to invalidate and destroy the entire range. If the range
> + * has device memory as its backing, the driver is also expected to migrate any
> + * remaining pages back to RAM.
> + */
> +
> +/**
> + * DOC: Examples
> + *
> + * This section provides three examples of how to build the expected driver
> + * components: the GPU page fault handler, the garbage collector, and the
> + * notifier callback.
> + *
> + * The generic code provided does not include logic for complex migration
> + * policies, optimized invalidations, fined grained driver locking, or other
> + * potentially required driver locking (e.g., DMA-resv locks).
> + *
> + * 1) GPU page fault handler
> + *
> + *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
> + *	{
> + *		int err = 0;
> + *
> + *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
> + *
> + *		drm_gpusvm_notifier_lock(gpusvm);
> + *		if (drm_gpusvm_range_pages_valid(range))
> + *			driver_commit_bind(gpusvm, range);
> + *		else
> + *			err = -EAGAIN;
> + *		drm_gpusvm_notifier_unlock(gpusvm);
> + *
> + *		return err;
> + *	}
> + *
> + *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
> + *			     unsigned long gpuva_start, unsigned long gpuva_end)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = {};
> + *		int err;
> + *
> + *		driver_svm_lock();
> + *	retry:
> + *		// Always process UNMAPs first so view of GPU SVM ranges is current
> + *		driver_garbage_collector(gpusvm);
> + *
> + *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
> + *							gpuva_start, gpuva_end,
> + *						        &ctx);
> + *		if (IS_ERR(range)) {
> + *			err = PTR_ERR(range);
> + *			goto unlock;
> + *		}
> + *
> + *		if (driver_migration_policy(range)) {
> + *			devmem = driver_alloc_devmem();
> + *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
> + *							   devmem_allocation,
> + *							   &ctx);
> + *			if (err)	// CPU mappings may have changed
> + *				goto retry;
> + *		}
> + *
> + *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
> + *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
> + *			if (err == -EOPNOTSUPP)
> + *				drm_gpusvm_range_evict(gpusvm, range);
> + *			goto retry;
> + *		} else if (err) {
> + *			goto unlock;
> + *		}
> + *
> + *		err = driver_bind_range(gpusvm, range);
> + *		if (err == -EAGAIN)	// CPU mappings changed
> + *			goto retry
> + *
> + *	unlock:
> + *		driver_svm_unlock();
> + *		return err;
> + *	}
> + *
> + * 2) Garbage Collector.
> + *
> + *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
> + *					struct drm_gpusvm_range *range)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		// Partial unmap, migrate any remaining device memory pages back to RAM
> + *		if (range->flags.partial_unmap)
> + *			drm_gpusvm_range_evict(gpusvm, range);
> + *
> + *		driver_unbind_range(range);
> + *		drm_gpusvm_range_remove(gpusvm, range);
> + *	}
> + *
> + *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		for_each_range_in_garbage_collector(gpusvm, range)
> + *			__driver_garbage_collector(gpusvm, range);
> + *	}
> + *
> + * 3) Notifier callback.
> + *
> + *	void driver_invalidation(struct drm_gpusvm *gpusvm,
> + *				 struct drm_gpusvm_notifier *notifier,
> + *				 const struct mmu_notifier_range *mmu_range)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
> + *		struct drm_gpusvm_range *range = NULL;
> + *
> + *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
> + *
> + *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
> + *					  mmu_range->end) {
> + *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
> + *
> + *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
> + *				continue;
> + *
> + *			drm_gpusvm_range_set_unmapped(range, mmu_range);
> + *			driver_garbage_collector_add(gpusvm, range);
> + *		}
> + *	}
> + */
> +
> +/**
> + * npages_in_range() - Calculate the number of pages in a given range
> + * @start: The start address of the range
> + * @end: The end address of the range
> + *
> + * This macro calculates the number of pages in a given memory range,
> + * specified by the start and end addresses. It divides the difference
> + * between the end and start addresses by the page size (PAGE_SIZE) to
> + * determine the number of pages in the range.
> + *
> + * Return: The number of pages in the specified range.
> + */
> +static unsigned long
> +npages_in_range(unsigned long start, unsigned long end)
> +{
> +	return (end - start) >> PAGE_SHIFT;
> +}
> +
> +/**
> + * struct drm_gpusvm_zdd - GPU SVM zone device data
> + *
> + * @refcount: Reference count for the zdd
> + * @destroy_work: Work structure for asynchronous zdd destruction
> + * @devmem_allocation: device memory allocation
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This structure serves as a generic wrapper installed in
> + * page->zone_device_data. It provides infrastructure for looking up a device
> + * memory allocation upon CPU page fault and asynchronously releasing device
> + * memory once the CPU has no page references. Asynchronous release is useful
> + * because CPU page references can be dropped in IRQ contexts, while releasing
> + * device memory likely requires sleeping locks.
> + */
> +struct drm_gpusvm_zdd {
> +	struct kref refcount;
> +	struct work_struct destroy_work;
> +	struct drm_gpusvm_devmem *devmem_allocation;
> +	void *device_private_page_owner;
> +};
> +
> +/**
> + * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
> + * @w: Pointer to the work_struct
> + *
> + * This function releases device memory, puts GPU SVM range, and frees zdd.
> + */
> +static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(w, struct drm_gpusvm_zdd, destroy_work);
> +	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
> +		zdd->devmem_allocation->ops : NULL;
> +
> +	if (zdd->devmem_allocation && ops->devmem_release)
> +		ops->devmem_release(zdd->devmem_allocation);
> +	kfree(zdd);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This function allocates and initializes a new zdd structure. It sets up the
> + * reference count and initializes the destroy work.
> + *
> + * Returns:
> + * Pointer to the allocated zdd on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_zdd *
> +drm_gpusvm_zdd_alloc(void *device_private_page_owner)
> +{
> +	struct drm_gpusvm_zdd *zdd;
> +
> +	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
> +	if (!zdd)
> +		return NULL;
> +
> +	kref_init(&zdd->refcount);
> +	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
> +	zdd->devmem_allocation = NULL;
> +	zdd->device_private_page_owner = device_private_page_owner;
> +
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function increments the reference count of the provided zdd structure.
> + *
> + * Returns: Pointer to the zdd structure.
> + */
> +static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_get(&zdd->refcount);
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
> + * @ref: Pointer to the reference count structure.
> + *
> + * This function queues the destroy_work of the zdd for asynchronous destruction.
> + */
> +static void drm_gpusvm_zdd_destroy(struct kref *ref)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(ref, struct drm_gpusvm_zdd, refcount);
> +
> +	if (zdd->devmem_allocation)
> +		WRITE_ONCE(zdd->devmem_allocation->detached, true);
> +	schedule_work(&zdd->destroy_work);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_put() - Put a zdd reference.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function decrements the reference count of the provided zdd structure
> + * and schedules its destruction if the count drops to zero.
> + */
> +static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
> +}
> +
> +/**
> + * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
> + * @notifier: Pointer to the GPU SVM notifier structure.
> + * @start: Start address of the range
> + * @end: End address of the range
> + *
> + * Return: A pointer to the drm_gpusvm_range if found or NULL
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_range, itree);
> +	else
> +		return NULL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
> +
> +/**
> + * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges
> + * @next__: Iterator variable for the ranges temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier while
> + * removing ranges from it.
> + */
> +#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
> +	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
> +	     (next__) = __drm_gpusvm_range_next(range__);				\
> +	     (range__) && (range__->itree.start < (end__));				\
> +	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
> +
> +/**
> + * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
> + * @notifier: a pointer to the current drm_gpusvm_notifier
> + *
> + * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
> + *         the current notifier is the last one or if the input notifier is
> + *         NULL.
> + */
> +static struct drm_gpusvm_notifier *
> +__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
> +{
> +	if (notifier && !list_is_last(&notifier->entry,
> +				      &notifier->gpusvm->notifier_list))
> +		return list_next_entry(notifier, entry);
> +
> +	return NULL;
> +}
> +
> +static struct drm_gpusvm_notifier *
> +notifier_iter_first(struct rb_root_cached *root, unsigned long start,
> +		    unsigned long last)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(root, start, last);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_notifier, itree);
> +	else
> +		return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
> + */
> +#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @next__: Iterator variable for the notifiers temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
> + * removing notifiers from it.
> + */
> +#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
> +	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
> + * @mni: Pointer to the mmu_interval_notifier structure.
> + * @mmu_range: Pointer to the mmu_notifier_range structure.
> + * @cur_seq: Current sequence number.
> + *
> + * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
> + * notifier sequence number and calls the driver invalidate vfunc under
> + * gpusvm->notifier_lock.
> + *
> + * Returns:
> + * true if the operation succeeds, false otherwise.
> + */
> +static bool
> +drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
> +			       const struct mmu_notifier_range *mmu_range,
> +			       unsigned long cur_seq)
> +{
> +	struct drm_gpusvm_notifier *notifier =
> +		container_of(mni, typeof(*notifier), notifier);
> +	struct drm_gpusvm *gpusvm = notifier->gpusvm;
> +
> +	if (!mmu_notifier_range_blockable(mmu_range))
> +		return false;
> +
> +	down_write(&gpusvm->notifier_lock);
> +	mmu_interval_set_seq(mni, cur_seq);
> +	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
> +	up_write(&gpusvm->notifier_lock);
> +
> +	return true;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
> + */
> +static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
> +	.invalidate = drm_gpusvm_notifier_invalidate,
> +};
> +
> +/**
> + * drm_gpusvm_init() - Initialize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @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.
> + * @ops: Pointer to the operations structure for GPU SVM.
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order with last
> + *               entry being SZ_4K.
> + * @num_chunks: Number of chunks.
> + *
> + * This function initializes the GPU SVM.
> + *
> + * Returns:
> + * 0 on success, a negative error code on failure.
> + */
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks)
> +{
> +	if (!ops->invalidate || !num_chunks)
> +		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;
> +	gpusvm->ops = ops;
> +	gpusvm->chunk_sizes = chunk_sizes;
> +	gpusvm->num_chunks = num_chunks;
> +
> +	mmgrab(mm);
> +	gpusvm->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&gpusvm->notifier_list);
> +
> +	init_rwsem(&gpusvm->notifier_lock);
> +
> +	fs_reclaim_acquire(GFP_KERNEL);
> +	might_lock(&gpusvm->notifier_lock);
> +	fs_reclaim_release(GFP_KERNEL);
> +
> +#ifdef CONFIG_LOCKDEP
> +	gpusvm->lock_dep_map = NULL;
> +#endif
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_init);
> +
> +/**
> + * drm_gpusvm_notifier_find() - Find GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function finds the GPU SVM notifier associated with the fault address.
> + *
> + * Returns:
> + * Pointer to the GPU SVM notifier on success, NULL otherwise.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
> +			 unsigned long fault_addr)
> +{
> +	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
> +}
> +
> +/**
> + * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_notifier structure.
> + */
> +static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	interval_tree_insert(&notifier->itree, &gpusvm->root);
> +
> +	node = rb_prev(&notifier->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_notifier(node))->entry;
> +	else
> +		head = &gpusvm->notifier_list;
> +
> +	list_add(&notifier->entry, head);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM tructure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	interval_tree_remove(&notifier->itree, &gpusvm->root);
> +	list_del(&notifier->entry);
> +}
> +
> +/**
> + * drm_gpusvm_fini() - Finalize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + *
> + * This function finalizes the GPU SVM by cleaning up any remaining ranges and
> + * notifiers, and dropping a reference to struct MM.
> + */
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
> +{
> +	struct drm_gpusvm_notifier *notifier, *next;
> +
> +	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
> +		struct drm_gpusvm_range *range, *__next;
> +
> +		/*
> +		 * Remove notifier first to avoid racing with any invalidation
> +		 */
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +		notifier->flags.removed = true;
> +
> +		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
> +					       LONG_MAX)
> +			drm_gpusvm_range_remove(gpusvm, range);
> +	}
> +
> +	mmdrop(gpusvm->mm);
> +	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
> +
> +/**
> + * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function allocates and initializes the GPU SVM notifier structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	if (gpusvm->ops->notifier_alloc)
> +		notifier = gpusvm->ops->notifier_alloc();
> +	else
> +		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
> +
> +	if (!notifier)
> +		return ERR_PTR(-ENOMEM);
> +
> +	notifier->gpusvm = gpusvm;
> +	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
> +	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
> +	INIT_LIST_HEAD(&notifier->entry);
> +	notifier->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&notifier->range_list);
> +
> +	return notifier;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_free() - Free GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function frees the GPU SVM notifier structure.
> + */
> +static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
> +				     struct drm_gpusvm_notifier *notifier)
> +{
> +	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
> +
> +	if (gpusvm->ops->notifier_free)
> +		gpusvm->ops->notifier_free(notifier);
> +	else
> +		kfree(notifier);
> +}
> +
> +/**
> + * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_range structure.
> + */
> +static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_range, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_range_insert() - Insert GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function inserts the GPU SVM range into the notifier RB tree and list.
> + */
> +static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
> +				    struct drm_gpusvm_range *range)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	drm_gpusvm_notifier_lock(notifier->gpusvm);
> +	interval_tree_insert(&range->itree, &notifier->root);
> +
> +	node = rb_prev(&range->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_range(node))->entry;
> +	else
> +		head = &notifier->range_list;
> +
> +	list_add(&range->entry, head);
> +	drm_gpusvm_notifier_unlock(notifier->gpusvm);
> +}
> +
> +/**
> + * __drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This macro removes the GPU SVM range from the notifier RB tree and list.
> + */
> +static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
> +				      struct drm_gpusvm_range *range)
> +{
> +	interval_tree_remove(&range->itree, &notifier->root);
> +	list_del(&range->entry);
> +}
> +
> +/**
> + * drm_gpusvm_range_alloc() - Allocate GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @fault_addr: Fault address
> + * @chunk_size: Chunk size
> + * @migrate_devmem: Flag indicating whether to migrate device memory
> + *
> + * This function allocates and initializes the GPU SVM range structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_range *
> +drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
> +		       struct drm_gpusvm_notifier *notifier,
> +		       unsigned long fault_addr, unsigned long chunk_size,
> +		       bool migrate_devmem)
> +{
> +	struct drm_gpusvm_range *range;
> +
> +	if (gpusvm->ops->range_alloc)
> +		range = gpusvm->ops->range_alloc(gpusvm);
> +	else
> +		range = kzalloc(sizeof(*range), GFP_KERNEL);
> +
> +	if (!range)
> +		return ERR_PTR(-ENOMEM);
> +
> +	kref_init(&range->refcount);
> +	range->gpusvm = gpusvm;
> +	range->notifier = notifier;
> +	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
> +	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
> +	INIT_LIST_HEAD(&range->entry);
> +	range->notifier_seq = LONG_MAX;
> +	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
> +
> +	return range;
> +}
> +
> +/**
> + * drm_gpusvm_check_pages() - Check pages
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @start: Start address
> + * @end: End address
> + *
> + * Check if pages between start and end have been faulted in on the CPU. Use to
> + * prevent migration of pages without CPU backing store.
> + *
> + * Returns:
> + * True if pages have been faulted into CPU, False otherwise
> + */
> +static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
> +				   struct drm_gpusvm_notifier *notifier,
> +				   unsigned long start, unsigned long end)
> +{
> +	struct hmm_range hmm_range = {
> +		.default_flags = 0,
> +		.notifier = &notifier->notifier,
> +		.start = start,
> +		.end = end,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long *pfns;
> +	unsigned long npages = npages_in_range(start, end);
> +	int err, i;
> +
> +	mmap_assert_locked(gpusvm->mm);
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return false;
> +
> +	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +	hmm_range.hmm_pfns = pfns;
> +
> +	while (true) {
> +		err = hmm_range_fault(&hmm_range);
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	if (err)
> +		goto err_free;
> +
> +	for (i = 0; i < npages;) {
> +		if (!(pfns[i] & HMM_PFN_VALID)) {
> +			err = -EFAULT;
> +			goto err_free;
> +		}
> +		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
> +	}
> +
> +err_free:
> +	kvfree(pfns);
> +	return err ? false : true;
> +}
> +
> +/**
> + * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @vas: Pointer to the virtual memory area structure
> + * @fault_addr: Fault address
> + * @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
> + *
> + * 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
> + * memory area boundaries.
> + *
> + * Returns:
> + * Chunk size on success, LONG_MAX on failure.
> + */
> +static unsigned long
> +drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_notifier *notifier,
> +			    struct vm_area_struct *vas,
> +			    unsigned long fault_addr,
> +			    unsigned long gpuva_start,
> +			    unsigned long gpuva_end,
> +			    unsigned long check_pages_threshold)
> +{
> +	unsigned long start, end;
> +	int i = 0;
> +
> +retry:
> +	for (; i < gpusvm->num_chunks; ++i) {
> +		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
> +		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
> +
> +		if (start >= vas->vm_start && end <= vas->vm_end &&
> +		    start >= notifier->itree.start &&
> +		    end <= notifier->itree.last + 1 &&
> +		    start >= gpuva_start && end <= gpuva_end)
> +			break;
> +	}
> +
> +	if (i == gpusvm->num_chunks)
> +		return LONG_MAX;
> +
> +	/*
> +	 * If allocation more than page, ensure not to overlap with existing
> +	 * ranges.
> +	 */
> +	if (end - start != SZ_4K) {
> +		struct drm_gpusvm_range *range;
> +
> +		range = drm_gpusvm_range_find(notifier, start, end);
> +		if (range) {
> +			++i;
> +			goto retry;
> +		}
> +
> +		/*
> +		 * XXX: Only create range on pages CPU has faulted in. Without
> +		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
> +		 * process-many-malloc' fails. In the failure case, each process
> +		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
> +		 * ranges. When migrating the SVM ranges, some processes fail in
> +		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
> +		 * and then upon drm_gpusvm_range_get_pages device pages from
> +		 * other processes are collected + faulted in which creates all
> +		 * sorts of problems. Unsure exactly how this happening, also
> +		 * problem goes away if 'xe_exec_system_allocator --r
> +		 * process-many-malloc' mallocs at least 64k at a time.
> +		 */
> +		if (end - start <= check_pages_threshold &&
> +		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
> +			++i;
> +			goto retry;
> +		}
> +	}
> +
> +	return end - start;
> +}
> +
> +/**
> + * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + * @gpuva_start: Start address of GPUVA which mirrors CPU
> + * @gpuva_end: End address of GPUVA which mirrors CPU
> + * @ctx: GPU SVM context
> + *
> + * This function finds or inserts a newly allocated a GPU SVM range based on the
> + * fault address. Caller must hold a lock to protect range lookup and insertion.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +	struct drm_gpusvm_range *range;
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	bool notifier_alloc = false;
> +	unsigned long chunk_size;
> +	int err;
> +	bool migrate_devmem;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	if (fault_addr < gpusvm->mm_start ||
> +	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
> +		return ERR_PTR(-EINVAL);
> +
> +	if (!mmget_not_zero(mm))
> +		return ERR_PTR(-EFAULT);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
> +	if (!notifier) {
> +		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
> +		if (IS_ERR(notifier)) {
> +			err = PTR_ERR(notifier);
> +			goto err_mmunlock;
> +		}
> +		notifier_alloc = true;
> +		err = mmu_interval_notifier_insert(&notifier->notifier,
> +						   mm, notifier->itree.start,
> +						   notifier->itree.last -
> +						   notifier->itree.start + 1,
> +						   &drm_gpusvm_notifier_ops);
> +		if (err)
> +			goto err_notifier;
> +	}
> +
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, fault_addr);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_notifier_remove;
> +	}
> +
> +	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
> +		err = -EPERM;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
> +	if (range)
> +		goto out_mmunlock;
> +	/*
> +	 * XXX: Short-circuiting migration based on migrate_vma_* current
> +	 * limitations. If/when migrate_vma_* add more support, this logic will
> +	 * have to change.
> +	 */
> +	migrate_devmem = ctx->devmem_possible &&
> +		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
> +
> +	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
> +						 fault_addr, gpuva_start,
> +						 gpuva_end, migrate_devmem &&

Another issue which Himal's Kconfig fix exposed.

s/migrate_devmem && ctx->check_pages_threshold/ctx->check_pages_threshold

Matt

> +						 ctx->check_pages_threshold);
> +	if (chunk_size == LONG_MAX) {
> +		err = -EINVAL;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
> +				       migrate_devmem);
> +	if (IS_ERR(range)) {
> +		err = PTR_ERR(range);
> +		goto err_notifier_remove;
> +	}
> +
> +	drm_gpusvm_range_insert(notifier, range);
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_insert(gpusvm, notifier);
> +
> +out_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +
> +	return range;
> +
> +err_notifier_remove:
> +	mmap_read_unlock(mm);
> +	if (notifier_alloc)
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +err_notifier:
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +err_mmunlock:
> +	mmput(mm);
> +	return ERR_PTR(err);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
> +
> +/**
> + * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @npages: Number of pages to unmap
> + *
> + * This function unmap pages associated with a GPU SVM range. Assumes and
> + * asserts correct locking is in place when called.
> + */
> +static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +					   struct drm_gpusvm_range *range,
> +					   unsigned long npages)
> +{
> +	unsigned long i, j;
> +	struct drm_pagemap *dpagemap = range->dpagemap;
> +	struct device *dev = gpusvm->drm->dev;
> +
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->flags.has_dma_mapping) {
> +		for (i = 0, j = 0; i < npages; j++) {
> +			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
> +
> +			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
> +				dma_unmap_page(dev,
> +					       addr->addr,
> +					       PAGE_SIZE << addr->order,
> +					       addr->dir);
> +			else if (dpagemap && dpagemap->ops->unmap_dma)
> +				dpagemap->ops->unmap_dma(dpagemap,
> +							 dev,
> +							 *addr);
> +			i += 1 << addr->order;
> +		}
> +		range->flags.has_devmem_pages = false;
> +		range->flags.has_dma_mapping = false;
> +		range->dpagemap = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function frees the dma address array associated with a GPU SVM range.
> + */
> +static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
> +					struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->dma_addr) {
> +		kvfree(range->dma_addr);
> +		range->dma_addr = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function removes the specified GPU SVM range and also removes the parent
> + * GPU SVM notifier if no more ranges remain in the notifier. The caller must
> + * hold a lock to protect range and notifier removal.
> + */
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
> +	if (WARN_ON_ONCE(!notifier))
> +		return;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +	drm_gpusvm_range_free_pages(gpusvm, range);
> +	__drm_gpusvm_range_remove(notifier, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	drm_gpusvm_range_put(range);
> +
> +	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
> +		if (!notifier->flags.removed)
> +			mmu_interval_notifier_remove(&notifier->notifier);
> +		drm_gpusvm_notifier_remove(gpusvm, notifier);
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +	}
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
> +
> +/**
> + * drm_gpusvm_range_get() - Get a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function increments the reference count of the specified GPU SVM range.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range)
> +{
> +	kref_get(&range->refcount);
> +
> +	return range;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
> +
> +/**
> + * drm_gpusvm_range_destroy() - Destroy GPU SVM range
> + * @refcount: Pointer to the reference counter embedded in the GPU SVM range
> + *
> + * This function destroys the specified GPU SVM range when its reference count
> + * reaches zero. If a custom range-free function is provided, it is invoked to
> + * free the range; otherwise, the range is deallocated using kfree().
> + */
> +static void drm_gpusvm_range_destroy(struct kref *refcount)
> +{
> +	struct drm_gpusvm_range *range =
> +		container_of(refcount, struct drm_gpusvm_range, refcount);
> +	struct drm_gpusvm *gpusvm = range->gpusvm;
> +
> +	if (gpusvm->ops->range_free)
> +		gpusvm->ops->range_free(range);
> +	else
> +		kfree(range);
> +}
> +
> +/**
> + * drm_gpusvm_range_put() - Put a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function decrements the reference count of the specified GPU SVM range
> + * and frees it when the count reaches zero.
> + */
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
> +{
> +	kref_put(&range->refcount, drm_gpusvm_range_destroy);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
> +
> +/**
> + * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called holding gpusvm->notifier_lock and as the last step before committing a
> + * GPU binding. This is akin to a notifier seqno check in the HMM documentation
> + * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
> + * function is required for finer grained checking (i.e., per range) if pages
> + * are valid.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
> +
> +/**
> + * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called without holding gpusvm->notifier_lock.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +static bool
> +drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
> +				      struct drm_gpusvm_range *range)
> +{
> +	bool pages_valid;
> +
> +	if (!range->dma_addr)
> +		return false;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
> +	if (!pages_valid)
> +		drm_gpusvm_range_free_pages(gpusvm, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	return pages_valid;
> +}
> +
> +/**
> + * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function gets pages for a GPU SVM range and ensures they are mapped for
> + * DMA access.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
> +	struct hmm_range hmm_range = {
> +		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
> +			HMM_PFN_REQ_WRITE),
> +		.notifier = notifier,
> +		.start = range->itree.start,
> +		.end = range->itree.last + 1,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct drm_gpusvm_zdd *zdd;
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long i, j;
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	unsigned long num_dma_mapped;
> +	unsigned int order = 0;
> +	unsigned long *pfns;
> +	struct page **pages;
> +	int err = 0;
> +	struct dev_pagemap *pagemap;
> +	struct drm_pagemap *dpagemap;
> +
> +retry:
> +	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
> +		goto set_seqno;
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return -ENOMEM;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	hmm_range.hmm_pfns = pfns;
> +	while (true) {
> +		mmap_read_lock(mm);
> +		err = hmm_range_fault(&hmm_range);
> +		mmap_read_unlock(mm);
> +
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	mmput(mm);
> +	if (err)
> +		goto err_free;
> +
> +	pages = (struct page **)pfns;
> +map_pages:
> +	/*
> +	 * Perform all dma mappings under the notifier lock to not
> +	 * access freed pages. A notifier will either block on
> +	 * the notifier lock or unmap dma.
> +	 */
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		kvfree(pfns);
> +		goto retry;
> +	}
> +
> +	if (!range->dma_addr) {
> +		/* Unlock and restart mapping to allocate memory. */
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
> +						 GFP_KERNEL);
> +		if (!range->dma_addr) {
> +			err = -ENOMEM;
> +			goto err_free;
> +		}
> +		goto map_pages;
> +	}
> +
> +	zdd = NULL;
> +	num_dma_mapped = 0;
> +	for (i = 0, j = 0; i < npages; ++j) {
> +		struct page *page = hmm_pfn_to_page(pfns[i]);
> +
> +		order = hmm_pfn_to_map_order(pfns[i]);
> +		if (is_device_private_page(page) || is_device_coherent_page(page)) {
> +			if (zdd != page->zone_device_data && i > 0) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +			zdd = page->zone_device_data;
> +			if (pagemap != page->pgmap) {
> +				if (i > 0) {
> +					err = -EOPNOTSUPP;
> +					goto err_unmap;
> +				}
> +
> +				pagemap = page->pgmap;
> +				dpagemap = zdd->devmem_allocation->dpagemap;
> +				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
> +					/*
> +					 * Raced. This is not supposed to happen
> +					 * since hmm_range_fault() should've migrated
> +					 * this page to system.
> +					 */
> +					err = -EAGAIN;
> +					goto err_unmap;
> +				}
> +			}
> +			range->dma_addr[j] =
> +				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
> +						       page, order,
> +						       DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			pages[i] = page;
> +		} else {
> +			dma_addr_t addr;
> +
> +			if (is_zone_device_page(page) || zdd) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +
> +			addr = dma_map_page(gpusvm->drm->dev,
> +					    page, 0,
> +					    PAGE_SIZE << order,
> +					    DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			range->dma_addr[j] = drm_pagemap_dma_addr_encode
> +				(addr, DRM_INTERCONNECT_SYSTEM, order,
> +				 DMA_BIDIRECTIONAL);
> +		}
> +		i += 1 << order;
> +		num_dma_mapped = i;
> +	}
> +
> +	range->flags.has_dma_mapping = true;
> +	if (zdd) {
> +		range->flags.has_devmem_pages = true;
> +		range->dpagemap = dpagemap;
> +	}
> +
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +	kvfree(pfns);
> +set_seqno:
> +	range->notifier_seq = hmm_range.notifier_seq;
> +
> +	return 0;
> +
> +err_unmap:
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +err_free:
> +	kvfree(pfns);
> +	if (err == -EAGAIN)
> +		goto retry;
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
> +
> +/**
> + * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function unmaps pages associated with a GPU SVM range. If @in_notifier
> + * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
> + * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
> + * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
> + * security model.
> + */
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +
> +	if (ctx->in_notifier)
> +		lockdep_assert_held_write(&gpusvm->notifier_lock);
> +	else
> +		drm_gpusvm_notifier_lock(gpusvm);
> +
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +
> +	if (!ctx->in_notifier)
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_page() - Put a migration page
> + * @page: Pointer to the page to put
> + *
> + * This function unlocks and puts a page.
> + */
> +static void drm_gpusvm_migration_unlock_put_page(struct page *page)
> +{
> +	unlock_page(page);
> +	put_page(page);
> +}
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
> + * @npages: Number of pages
> + * @migrate_pfn: Array of migrate page frame numbers
> + *
> + * This function unlocks and puts an array of pages.
> + */
> +static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
> +						  unsigned long *migrate_pfn)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page;
> +
> +		if (!migrate_pfn[i])
> +			continue;
> +
> +		page = migrate_pfn_to_page(migrate_pfn[i]);
> +		drm_gpusvm_migration_unlock_put_page(page);
> +		migrate_pfn[i] = 0;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
> + * @page: Pointer to the page
> + * @zdd: Pointer to the GPU SVM zone device data
> + *
> + * This function associates the given page with the specified GPU SVM zone
> + * device data and initializes it for zone device usage.
> + */
> +static void drm_gpusvm_get_devmem_page(struct page *page,
> +				     struct drm_gpusvm_zdd *zdd)
> +{
> +	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
> +	zone_device_page_init(page);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
> + * @dev: The device for which the pages are being mapped
> + * @dma_addr: Array to store DMA addresses corresponding to mapped pages
> + * @migrate_pfn: Array of migrate page frame numbers to map
> + * @npages: Number of pages to map
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function maps pages of memory for migration usage in GPU SVM. It
> + * iterates over each page frame number provided in @migrate_pfn, maps the
> + * corresponding page, and stores the DMA address in the provided @dma_addr
> + * array.
> + *
> + * Return: 0 on success, -EFAULT if an error occurs during mapping.
> + */
> +static int drm_gpusvm_migrate_map_pages(struct device *dev,
> +					dma_addr_t *dma_addr,
> +					unsigned long *migrate_pfn,
> +					unsigned long npages,
> +					enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		if (WARN_ON_ONCE(is_zone_device_page(page)))
> +			return -EFAULT;
> +
> +		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
> +		if (dma_mapping_error(dev, dma_addr[i]))
> +			return -EFAULT;
> +	}
> +
> +	return 0;
> +}
> +
> +/**
> + * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
> + * @dev: The device for which the pages were mapped
> + * @dma_addr: Array of DMA addresses corresponding to mapped pages
> + * @npages: Number of pages to unmap
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function unmaps previously mapped pages of memory for GPU Shared Virtual
> + * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
> + * if it's valid and not already unmapped, and unmaps the corresponding page.
> + */
> +static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
> +					   dma_addr_t *dma_addr,
> +					   unsigned long npages,
> +					   enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
> +			continue;
> +
> +		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @devmem_allocation: Pointer to the device memory allocation. The caller
> + *                     should hold a reference to the device memory allocation,
> + *                     which should be dropped via ops->devmem_release or upon
> + *                     the failure of this function.
> + * @ctx: GPU SVM context
> + *
> + * This function migrates the specified GPU SVM range to device memory. It performs the
> + * necessary setup and invokes the driver-specific operations for migration to
> + * device memory. Upon successful return, @devmem_allocation can safely reference @range
> + * until ops->devmem_release is called which only upon successful return.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +	struct migrate_vma migrate = {
> +		.start		= start,
> +		.end		= end,
> +		.pgmap_owner	= gpusvm->device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	unsigned long i, npages = npages_in_range(start, end);
> +	struct vm_area_struct *vas;
> +	struct drm_gpusvm_zdd *zdd = NULL;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int err;
> +
> +	if (!range->flags.migrate_devmem)
> +		return -EINVAL;
> +
> +	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
> +	    !ops->copy_to_ram)
> +		return -EOPNOTSUPP;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_out;
> +	}
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, start);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_mmunlock;
> +	}
> +
> +	if (end > vas->vm_end || start < vas->vm_start) {
> +		err = -EINVAL;
> +		goto err_mmunlock;
> +	}
> +
> +	if (!vma_is_anonymous(vas)) {
> +		err = -EBUSY;
> +		goto err_mmunlock;
> +	}
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_mmunlock;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
> +	if (!zdd) {
> +		err = -ENOMEM;
> +		goto err_free;
> +	}
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	if (!migrate.cpages) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	if (migrate.cpages != npages) {
> +		err = -EBUSY;
> +		goto err_finalize;
> +	}
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   migrate.src, npages, DMA_TO_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = pfn_to_page(migrate.dst[i]);
> +
> +		pages[i] = page;
> +		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
> +		drm_gpusvm_get_devmem_page(page, zdd);
> +	}
> +
> +	err = ops->copy_to_devmem(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +	/* Upon success bind devmem allocation to range and zdd */
> +	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_TO_DEVICE);
> +err_free:
> +	if (zdd)
> +		drm_gpusvm_zdd_put(zdd);
> +	kvfree(buf);
> +err_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +err_out:
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
> +
> +/**
> + * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
> + * @vas: Pointer to the VM area structure, can be NULL
> + * @fault_page: Fault page
> + * @npages: Number of pages to populate
> + * @mpages: Number of pages to migrate
> + * @src_mpfn: Source array of migrate PFNs
> + * @mpfn: Array of migrate PFNs to populate
> + * @addr: Start address for PFN allocation
> + *
> + * This function populates the RAM migrate page frame numbers (PFNs) for the
> + * specified VM area structure. It allocates and locks pages in the VM area for
> + * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
> + * alloc_page for allocation.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> +					       struct page *fault_page,
> +					       unsigned long npages,
> +					       unsigned long *mpages,
> +					       unsigned long *src_mpfn,
> +					       unsigned long *mpfn,
> +					       unsigned long addr)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
> +		struct page *page, *src_page;
> +
> +		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
> +			continue;
> +
> +		if (fault_page) {
> +			src_page = migrate_pfn_to_page(src_mpfn[i]);
> +			if (src_page->zone_device_data !=
> +			    fault_page->zone_device_data)
> +				continue;
> +		}
> +
> +		if (vas)
> +			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
> +		else
> +			page = alloc_page(GFP_HIGHUSER);
> +
> +		if (!page)
> +			goto free_pages;
> +
> +		mpfn[i] = migrate_pfn(page_to_pfn(page));
> +	}
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		WARN_ON_ONCE(!trylock_page(page));
> +		++*mpages;
> +	}
> +
> +	return 0;
> +
> +free_pages:
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		put_page(page);
> +		mpfn[i] = 0;
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
> + * @devmem_allocation: Pointer to the device memory allocation
> + *
> + * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
> + * migration done via migrate_device_* functions.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	unsigned long *src, *dst;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int i, err = 0;
> +
> +	npages = devmem_allocation->size >> PAGE_SHIFT;
> +
> +retry:
> +	if (!mmget_not_zero(devmem_allocation->mm))
> +		return -EFAULT;
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	src = buf;
> +	dst = buf + (sizeof(*src) * npages);
> +	dma_addr = buf + (2 * sizeof(*src) * npages);
> +	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
> +	if (err)
> +		goto err_free;
> +
> +	err = migrate_device_pfns(src, npages);
> +	if (err)
> +		goto err_free;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
> +						  src, dst, 0);
> +	if (err || !mpages)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   dst, npages, DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, dst);
> +	migrate_device_pages(src, dst, npages);
> +	migrate_device_finalize(src, dst, npages);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +	mmput_async(devmem_allocation->mm);
> +	if (!err && !READ_ONCE(devmem_allocation->detached)) {
> +		cond_resched();
> +		goto retry;
> +	}
> +
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
> +
> +/**
> + * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
> + * @vas: Pointer to the VM area structure
> + * @device_private_page_owner: Device private pages owner
> + * @page: Pointer to the page for fault handling (can be NULL)
> + * @fault_addr: Fault address
> + * @size: Size of migration
> + *
> + * This internal function performs the migration of the specified GPU SVM range
> + * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
> + * invokes the driver-specific operations for migration to RAM.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
> +				       void *device_private_page_owner,
> +				       struct page *page,
> +				       unsigned long fault_addr,
> +				       unsigned long size)
> +{
> +	struct migrate_vma migrate = {
> +		.vma		= vas,
> +		.pgmap_owner	= device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
> +			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
> +		.fault_page	= page,
> +	};
> +	struct drm_gpusvm_zdd *zdd;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct device *dev;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	unsigned long start, end;
> +	void *buf;
> +	int i, err = 0;
> +
> +	start = ALIGN_DOWN(fault_addr, size);
> +	end = ALIGN(fault_addr + 1, size);
> +
> +	/* Corner where VMA area struct has been partially unmapped */
> +	if (start < vas->vm_start)
> +		start = vas->vm_start;
> +	if (end > vas->vm_end)
> +		end = vas->vm_end;
> +
> +	migrate.start = start;
> +	migrate.end = end;
> +	npages = npages_in_range(start, end);
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	/* Raced with another CPU fault, nothing to do */
> +	if (!migrate.cpages)
> +		goto err_free;
> +
> +	if (!page) {
> +		for (i = 0; i < npages; ++i) {
> +			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
> +				continue;
> +
> +			page = migrate_pfn_to_page(migrate.src[i]);
> +			break;
> +		}
> +
> +		if (!page)
> +			goto err_finalize;
> +	}
> +	zdd = page->zone_device_data;
> +	ops = zdd->devmem_allocation->ops;
> +	dev = zdd->devmem_allocation->dev;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
> +						  migrate.src, migrate.dst,
> +						  start);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
> +					   DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(migrate.src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +
> +	return err;
> +}
> +
> +/**
> + * drm_gpusvm_range_evict() - Evict GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function evicts the specified GPU SVM range.
> + */
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range)
> +{
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +
> +	if (!mmget_not_zero(mm))
> +		return;
> +
> +	mmap_read_lock(mm);
> +
> +	/*
> +	 * Loop required to find all VMAs for the corner case when partial
> +	 * unmappings have occurred.
> +	 */
> +again:
> +	vas = find_vma(mm, start);
> +	if (!vas)
> +		goto unlock;
> +
> +	if (end <= vas->vm_start || start >= vas->vm_end)
> +		goto unlock;
> +
> +	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
> +				    NULL, start,
> +				    range->itree.last - range->itree.start + 1);
> +	if (vas->vm_end < end) {
> +		start = vas->vm_end;
> +		goto again;
> +	}
> +unlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
> +
> +/**
> + * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
> + * @page: Pointer to the page
> + *
> + * This function is a callback used to put the GPU SVM zone device data
> + * associated with a page when it is being released.
> + */
> +static void drm_gpusvm_page_free(struct page *page)
> +{
> +	drm_gpusvm_zdd_put(page->zone_device_data);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
> + * @vmf: Pointer to the fault information structure
> + *
> + * This function is a page fault handler used to migrate a GPU SVM range to RAM.
> + * It retrieves the GPU SVM range information from the faulting page and invokes
> + * the internal migration function to migrate the range back to RAM.
> + *
> + * Returns:
> + * VM_FAULT_SIGBUS on failure, 0 on success.
> + */
> +static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
> +{
> +	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
> +	int err;
> +
> +	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
> +					  zdd->device_private_page_owner,
> +					  vmf->page, vmf->address,
> +					  zdd->devmem_allocation->size);
> +
> +	return err ? VM_FAULT_SIGBUS : 0;
> +}
> +
> +/**
> + * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
> + */
> +static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
> +	.page_free = drm_gpusvm_page_free,
> +	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
> +};
> +
> +/**
> + * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
> + *
> + * Returns:
> + * Pointer to the GPU SVM device page map operations structure.
> + */
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
> +{
> +	return &drm_gpusvm_pagemap_ops;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
> +
> +/**
> + * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @start: Start address
> + * @end: End address
> + *
> + * Returns:
> + * True if GPU SVM has mapping, False otherwise
> + */
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
> +		struct drm_gpusvm_range *range = NULL;
> +
> +		drm_gpusvm_for_each_range(range, notifier, start, end)
> +			return true;
> +	}
> +
> +	return false;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
> +
> +/**
> + * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
> + * @range: Pointer to the GPU SVM range structure.
> + * @mmu_range: Pointer to the MMU notifier range structure.
> + *
> + * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
> + * if the range partially falls within the provided MMU notifier range.
> + */
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range)
> +{
> +	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
> +
> +	range->flags.unmapped = true;
> +	if (range->itree.start < mmu_range->start ||
> +	    range->itree.last + 1 > mmu_range->end)
> +		range->flags.partial_unmap = true;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
> +
> +/**
> + * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap we're allocating from.
> + * @size: Size of device memory allocation
> + */
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size)
> +{
> +	devmem_allocation->dev = dev;
> +	devmem_allocation->mm = mm;
> +	devmem_allocation->ops = ops;
> +	devmem_allocation->dpagemap = dpagemap;
> +	devmem_allocation->size = size;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
> +
> +MODULE_DESCRIPTION("DRM GPUSVM");
> +MODULE_LICENSE("GPL");
> diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> new file mode 100644
> index 000000000000..2fb6980d0cf1
> --- /dev/null
> +++ b/include/drm/drm_gpusvm.h
> @@ -0,0 +1,445 @@
> +/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
> +/*
> + * Copyright © 2024 Intel Corporation
> + */
> +
> +#ifndef __DRM_GPUSVM_H__
> +#define __DRM_GPUSVM_H__
> +
> +#include <linux/kref.h>
> +#include <linux/interval_tree.h>
> +#include <linux/mmu_notifier.h>
> +
> +struct dev_pagemap_ops;
> +struct drm_device;
> +struct drm_gpusvm;
> +struct drm_gpusvm_notifier;
> +struct drm_gpusvm_ops;
> +struct drm_gpusvm_range;
> +struct drm_gpusvm_devmem;
> +struct drm_pagemap;
> +struct drm_pagemap_dma_addr;
> +
> +/**
> + * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM)
> + * device memory. These operations are provided by the GPU driver to manage device memory
> + * allocations and perform operations such as migration between device memory and system
> + * RAM.
> + */
> +struct drm_gpusvm_devmem_ops {
> +	/**
> +	 * @devmem_release: Release device memory allocation (optional)
> +	 * @devmem_allocation: device memory allocation
> +	 *
> +	 * Release device memory allocation and drop a reference to device
> +	 * memory allocation.
> +	 */
> +	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +	/**
> +	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
> +	 * @devmem_allocation: device memory allocation
> +	 * @npages: Number of pages to populate
> +	 * @pfn: Array of page frame numbers to populate
> +	 *
> +	 * Populate device memory page frame numbers (PFN).
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
> +				   unsigned long npages, unsigned long *pfn);
> +
> +	/**
> +	 * @copy_to_devmem: Copy to device memory (required for migration)
> +	 * @pages: Pointer to array of device memory pages (destination)
> +	 * @dma_addr: Pointer to array of DMA addresses (source)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to device memory.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_devmem)(struct page **pages,
> +			      dma_addr_t *dma_addr,
> +			      unsigned long npages);
> +
> +	/**
> +	 * @copy_to_ram: Copy to system RAM (required for migration)
> +	 * @pages: Pointer to array of device memory pages (source)
> +	 * @dma_addr: Pointer to array of DMA addresses (destination)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to system RAM.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_ram)(struct page **pages,
> +			   dma_addr_t *dma_addr,
> +			   unsigned long npages);
> +};
> +
> +/**
> + * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
> + * @size: Size of device memory allocation
> + * @detached: device memory allocations is detached from device pages
> + */
> +struct drm_gpusvm_devmem {
> +	struct device *dev;
> +	struct mm_struct *mm;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct drm_pagemap *dpagemap;
> +	size_t size;
> +	bool detached;
> +};
> +
> +/**
> + * struct drm_gpusvm_ops - Operations structure for GPU SVM
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM).
> + * These operations are provided by the GPU driver to manage SVM ranges and
> + * notifiers.
> + */
> +struct drm_gpusvm_ops {
> +	/**
> +	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
> +	 *
> +	 * Allocate a GPU SVM notifier.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
> +
> +	/**
> +	 * @notifier_free: Free a GPU SVM notifier (optional)
> +	 * @notifier: Pointer to the GPU SVM notifier to be freed
> +	 *
> +	 * Free a GPU SVM notifier.
> +	 */
> +	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
> +
> +	/**
> +	 * @range_alloc: Allocate a GPU SVM range (optional)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 *
> +	 * Allocate a GPU SVM range.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
> +
> +	/**
> +	 * @range_free: Free a GPU SVM range (optional)
> +	 * @range: Pointer to the GPU SVM range to be freed
> +	 *
> +	 * Free a GPU SVM range.
> +	 */
> +	void (*range_free)(struct drm_gpusvm_range *range);
> +
> +	/**
> +	 * @invalidate: Invalidate GPU SVM notifier (required)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 * @notifier: Pointer to the GPU SVM notifier
> +	 * @mmu_range: Pointer to the mmu_notifier_range structure
> +	 *
> +	 * Invalidate the GPU page tables. It can safely walk the notifier range
> +	 * RB tree/list in this function. Called while holding the notifier lock.
> +	 */
> +	void (*invalidate)(struct drm_gpusvm *gpusvm,
> +			   struct drm_gpusvm_notifier *notifier,
> +			   const struct mmu_notifier_range *mmu_range);
> +};
> +
> +/**
> + * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: MMU interval notifier
> + * @itree: Interval tree node for the notifier (inserted in GPU SVM)
> + * @entry: List entry to fast interval tree traversal
> + * @root: Cached root node of the RB tree containing ranges
> + * @range_list: List head containing of ranges in the same order they appear in
> + *              interval tree. This is useful to keep iterating ranges while
> + *              doing modifications to RB tree.
> + * @flags.removed: Flag indicating whether the MMU interval notifier has been
> + *                 removed
> + *
> + * This structure represents a GPU SVM notifier.
> + */
> +struct drm_gpusvm_notifier {
> +	struct drm_gpusvm *gpusvm;
> +	struct mmu_interval_notifier notifier;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	struct rb_root_cached root;
> +	struct list_head range_list;
> +	struct {
> +		u32 removed : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm_range - Structure representing a GPU SVM range
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier
> + * @refcount: Reference count for the range
> + * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
> + * @entry: List entry to fast interval tree traversal
> + * @notifier_seq: Notifier sequence number of the range's pages
> + * @dma_addr: DMA address array
> + * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
> + *            Note this is assuming only one drm_pagemap per range is allowed.
> + * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
> + * @flags.unmapped: Flag indicating if the range has been unmapped
> + * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
> + * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
> + * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
> + *
> + * This structure represents a GPU SVM range used for tracking memory ranges
> + * mapped in a DRM device.
> + */
> +struct drm_gpusvm_range {
> +	struct drm_gpusvm *gpusvm;
> +	struct drm_gpusvm_notifier *notifier;
> +	struct kref refcount;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	unsigned long notifier_seq;
> +	struct drm_pagemap_dma_addr *dma_addr;
> +	struct drm_pagemap *dpagemap;
> +	struct {
> +		/* All flags below must be set upon creation */
> +		u16 migrate_devmem : 1;
> +		/* All flags below must be set / cleared under notifier lock */
> +		u16 unmapped : 1;
> +		u16 partial_unmap : 1;
> +		u16 has_devmem_pages : 1;
> +		u16 has_dma_mapping : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm - GPU SVM structure
> + *
> + * @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
> + * @ops: Pointer to the operations structure for GPU SVM
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order.
> + * @num_chunks: Number of chunks
> + * @notifier_lock: Read-write semaphore for protecting notifier operations
> + * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
> + * @notifier_list: list head containing of notifiers in the same order they
> + *                 appear in interval tree. This is useful to keep iterating
> + *                 notifiers while doing modifications to RB tree.
> + *
> + * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
> + * memory ranges mapped in a DRM (Direct Rendering Manager) device.
> + *
> + * No reference counting is provided, as this is expected to be embedded in the
> + * driver VM structure along with the struct drm_gpuvm, which handles reference
> + * counting.
> + */
> +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;
> +	const struct drm_gpusvm_ops *ops;
> +	const unsigned long *chunk_sizes;
> +	int num_chunks;
> +	struct rw_semaphore notifier_lock;
> +	struct rb_root_cached root;
> +	struct list_head notifier_list;
> +#ifdef CONFIG_LOCKDEP
> +	/**
> +	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
> +	 * drm_gpusvm_range_remove with a driver provided lock.
> +	 */
> +	struct lockdep_map *lock_dep_map;
> +#endif
> +};
> +
> +/**
> + * struct drm_gpusvm_ctx - DRM GPU SVM context
> + *
> + * @check_pages_threshold: Check CPU pages for present if chunk is less than or
> + *                         equal to threshold. If not present, reduce chunk
> + *                         size.
> + * @in_notifier: entering from a MMU notifier
> + * @read_only: operating on read-only memory
> + * @devmem_possible: possible to use device memory
> + *
> + * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> + */
> +struct drm_gpusvm_ctx {
> +	unsigned long check_pages_threshold;
> +	unsigned int in_notifier :1;
> +	unsigned int read_only :1;
> +	unsigned int devmem_possible :1;
> +};
> +
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks);
> +
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
> +
> +void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
> +
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range);
> +
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
> +
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end);
> +
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range);
> +
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size);
> +
> +#ifdef CONFIG_LOCKDEP
> +/**
> + * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @lock: the lock used to protect the gpuva list. The locking primitive
> + * must contain a dep_map field.
> + *
> + * Call this to annotate drm_gpusvm_range_find_or_insert and
> + * drm_gpusvm_range_remove.
> + */
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
> +	do { \
> +		if (!WARN((gpusvm)->lock_dep_map, \
> +			  "GPUSVM range lock should be set only once."))\
> +			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
> +	} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) \
> +	do { \
> +		if ((gpusvm)->lock_dep_map)	\
> +			lock_is_held((gpusvm)->lock_dep_map);	\
> +	} while (0)
> +#else
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
> +#endif
> +
> +/**
> + * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, take lock
> + */
> +#define drm_gpusvm_notifier_lock(gpusvm__)	\
> +	down_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, drop lock
> + */
> +#define drm_gpusvm_notifier_unlock(gpusvm__)	\
> +	up_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
> + * @range: a pointer to the current GPU SVM range
> + *
> + * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
> + *         current range is the last one or if the input range is NULL.
> + */
> +static inline struct drm_gpusvm_range *
> +__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
> +{
> +	if (range && !list_is_last(&range->entry,
> +				   &range->notifier->range_list))
> +		return list_next_entry(range, entry);
> +
> +	return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges. If set, it indicates the start of
> + *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
> + * to use while holding the driver SVM lock or the notifier lock.
> + */
> +#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
> +	for ((range__) = (range__) ?:					\
> +	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
> +	     (range__) && (range__->itree.start < (end__));		\
> +	     (range__) = __drm_gpusvm_range_next(range__))
> +
> +#endif /* __DRM_GPUSVM_H__ */
> -- 
> 2.34.1
> 

Re: [PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Matthew Brost]

On Tue, Dec 17, 2024 at 03:33:23PM -0800, Matthew Brost wrote:
> This patch introduces support for GPU Shared Virtual Memory (SVM) in the
> Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
> sharing of memory between the CPU and GPU, enhancing performance and
> flexibility in GPU computing tasks.
> 
> The patch adds the necessary infrastructure for SVM, including data
> structures and functions for managing SVM ranges and notifiers. It also
> provides mechanisms for allocating, deallocating, and migrating memory
> regions between system RAM and GPU VRAM.
> 
> This is largely inspired by GPUVM.
> 
> v2:
>  - Take order into account in check pages
>  - Clear range->pages in get pages error
>  - Drop setting dirty or accessed bit in get pages (Vetter)
>  - Remove mmap assert for cpu faults
>  - Drop mmap write lock abuse (Vetter, Christian)
>  - Decouple zdd from range (Vetter, Oak)
>  - Add drm_gpusvm_range_evict, make it work with coherent pages
>  - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
>  - mmget/put in drm_gpusvm_evict_to_sram
>  - Drop range->vram_alloation variable
>  - Don't return in drm_gpusvm_evict_to_sram until all pages detached
>  - Don't warn on mixing sram and device pages
>  - Update kernel doc
>  - Add coherent page support to get pages
>  - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
>  - Add struct drm_gpusvm_vram and ops (Thomas)
>  - Update the range's seqno if the range is valid (Thomas)
>  - Remove the is_unmapped check before hmm_range_fault (Thomas)
>  - Use drm_pagemap (Thomas)
>  - Drop kfree_mapping (Thomas)
>  - dma mapp pages under notifier lock (Thomas)
>  - Remove ctx.prefault
>  - Remove ctx.mmap_locked
>  - Add ctx.check_pages
>  - s/vram/devmem (Thomas)
> v3:
>  - Fix memory leak drm_gpusvm_range_get_pages
>  - Only migrate pages with same zdd on CPU fault
>  - Loop over al VMAs in drm_gpusvm_range_evict
>  - Make GPUSVM a drm level module
>  - GPL or MIT license
>  - Update main kernel doc (Thomas)
>  - Prefer foo() vs foo for functions in kernel doc (Thomas)
>  - Prefer functions over macros (Thomas)
>  - Use unsigned long vs u64 for addresses (Thomas)
>  - Use standard interval_tree (Thomas)
>  - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
>  - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
>  - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
>  - Newlines between functions defs in header file (Thomas)
>  - Drop shall language in driver vfunc kernel doc (Thomas)
>  - Move some static inlines from head to C file (Thomas)
>  - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
>  - Change check_pages to a threshold
> 
> Cc: Simona Vetter <simona.vetter@ffwll.ch>
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Christian König <christian.koenig@amd.com>
> Cc: <dri-devel@lists.freedesktop.org>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
>  drivers/gpu/drm/Kconfig      |    8 +
>  drivers/gpu/drm/Makefile     |    1 +
>  drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
>  include/drm/drm_gpusvm.h     |  445 +++++++
>  4 files changed, 2674 insertions(+)
>  create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>  create mode 100644 include/drm/drm_gpusvm.h
> 
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 2f51546b0b88..5524cb1ed9dd 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -278,6 +278,14 @@ config DRM_GPUVM
>  	  GPU-VM representation providing helpers to manage a GPUs virtual
>  	  address space
>  
> +config DRM_GPUSVM
> +	tristate
> +	depends on DRM
> +	depends on DEVICE_MIGRATION
> +	help
> +	  GPU-SVM representation providing helpers to manage a GPUs shared
> +	  virtual memory
> +
>  config DRM_BUDDY
>  	tristate
>  	depends on DRM
> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> index 19fb370fbc56..ca5d44441fb6 100644
> --- a/drivers/gpu/drm/Makefile
> +++ b/drivers/gpu/drm/Makefile
> @@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
>  #
>  obj-$(CONFIG_DRM_EXEC) += drm_exec.o
>  obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> +obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
>  
>  obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
>  
> diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> new file mode 100644
> index 000000000000..29ce2c7d1c8c
> --- /dev/null
> +++ b/drivers/gpu/drm/drm_gpusvm.c
> @@ -0,0 +1,2220 @@
> +// SPDX-License-Identifier: GPL-2.0-only OR MIT
> +/*
> + * Copyright © 2024 Intel Corporation
> + *
> + * Authors:
> + *     Matthew Brost <matthew.brost@intel.com>
> + */
> +
> +#include <linux/dma-mapping.h>
> +#include <linux/hmm.h>
> +#include <linux/memremap.h>
> +#include <linux/migrate.h>
> +#include <linux/mm_types.h>
> +#include <linux/pagemap.h>
> +#include <linux/slab.h>
> +
> +#include <drm/drm_device.h>
> +#include <drm/drm_gpusvm.h>
> +#include <drm/drm_pagemap.h>
> +#include <drm/drm_print.h>
> +
> +/**
> + * DOC: Overview
> + *
> + * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
> + *
> + * The GPU SVM layer is a component of the DRM framework designed to manage shared
> + * virtual memory between the CPU and GPU. It enables efficient data exchange and
> + * processing for GPU-accelerated applications by allowing memory sharing and
> + * synchronization between the CPU's and GPU's virtual address spaces.
> + *
> + * Key GPU SVM Components:
> + * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
> + *		GPU of changes, notifiers are sized based on a GPU SVM
> + *		initialization parameter, with a recommendation of 512M or
> + *		larger. They maintain a Red-BlacK tree and a list of ranges that
> + *		fall within the notifier interval. Notifiers are tracked within
> + *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
> + *		or removed as ranges within the interval are created or
> + *		destroyed.
> + * - Ranges: Represent memory ranges mapped in a DRM device and managed
> + *	     by GPU SVM. They are sized based on an array of chunk sizes, which
> + *	     is a GPU SVM initialization parameter, and the CPU address space.
> + *	     Upon GPU fault, the largest aligned chunk that fits within the
> + *	     faulting CPU address space is chosen for the range size. Ranges are
> + *	     expected to be dynamically allocated on GPU fault and removed on an
> + *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
> + *	     a notifier's Red-Black tree.
> + * - Operations: Define the interface for driver-specific GPU SVM operations
> + *               such as range allocation, notifier allocation, and
> + *               invalidations.
> + * - Device Memory Allocations: Embedded structure containing enough information
> + *                              for GPU SVM to migrate to / from device memory.
> + * - Device Memory Operations: Define the interface for driver-specific device
> + *                             memory operations release memory, populate pfns,
> + *                             and copy to / from device memory.
> + *
> + * This layer provides interfaces for allocating, mapping, migrating, and
> + * releasing memory ranges between the CPU and GPU. It handles all core memory
> + * management interactions (DMA mapping, HMM, and migration) and provides
> + * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
> + * to build the expected driver components for an SVM implementation as detailed
> + * below.
> + *
> + * Expected Driver Components:
> + * - GPU page fault handler: Used to create ranges and notifiers based on the
> + *			     fault address, optionally migrate the range to
> + *			     device memory, and create GPU bindings.
> + * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
> + *			Ranges are expected to be added to the garbage collector
> + *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
> + * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
> + *			ranges.
> + */
> +
> +/**
> + * DOC: Locking
> + *
> + * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
> + * mmap lock as needed.
> + *
> + * GPU SVM introduces a global notifier lock, which safeguards the notifier's
> + * range RB tree and list, as well as the range's DMA mappings and sequence
> + * number. GPU SVM manages all necessary locking and unlocking operations,
> + * except for the recheck range's pages being valid
> + * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
> + * lock corresponds to the 'driver->update' lock mentioned in the HMM
> + * documentation (TODO: Link). Future revisions may transition from a GPU SVM
> + * global lock to a per-notifier lock if finer-grained locking is deemed
> + * necessary.
> + *
> + * In addition to the locking mentioned above, the driver should implement a
> + * lock to safeguard core GPU SVM function calls that modify state, such as
> + * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
> + * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
> + * locking should also be possible for concurrent GPU fault processing within a
> + * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
> + * to add annotations to GPU SVM.
> + */
> +
> +/**
> + * DOC: Migration
> + *
> + * The migration support is quite simple, allowing migration between RAM and
> + * device memory at the range granularity. For example, GPU SVM currently does not
> + * support mixing RAM and device memory pages within a range. This means that upon GPU
> + * fault, the entire range can be migrated to device memory, and upon CPU fault, the
> + * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
> + * could be added in the future if required.
> + *
> + * The reasoning for only supporting range granularity is as follows: it
> + * simplifies the implementation, and range sizes are driver-defined and should
> + * be relatively small.
> + */
> +
> +/**
> + * DOC: Partial Unmapping of Ranges
> + *
> + * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
> + * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
> + * being that a subset of the range still has CPU and GPU mappings. If the
> + * backing store for the range is in device memory, a subset of the backing store has
> + * references. One option would be to split the range and device memory backing store,
> + * but the implementation for this would be quite complicated. Given that
> + * partial unmappings are rare and driver-defined range sizes are relatively
> + * small, GPU SVM does not support splitting of ranges.
> + *
> + * With no support for range splitting, upon partial unmapping of a range, the
> + * driver is expected to invalidate and destroy the entire range. If the range
> + * has device memory as its backing, the driver is also expected to migrate any
> + * remaining pages back to RAM.
> + */
> +
> +/**
> + * DOC: Examples
> + *
> + * This section provides three examples of how to build the expected driver
> + * components: the GPU page fault handler, the garbage collector, and the
> + * notifier callback.
> + *
> + * The generic code provided does not include logic for complex migration
> + * policies, optimized invalidations, fined grained driver locking, or other
> + * potentially required driver locking (e.g., DMA-resv locks).
> + *
> + * 1) GPU page fault handler
> + *
> + *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
> + *	{
> + *		int err = 0;
> + *
> + *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
> + *
> + *		drm_gpusvm_notifier_lock(gpusvm);
> + *		if (drm_gpusvm_range_pages_valid(range))
> + *			driver_commit_bind(gpusvm, range);
> + *		else
> + *			err = -EAGAIN;
> + *		drm_gpusvm_notifier_unlock(gpusvm);
> + *
> + *		return err;
> + *	}
> + *
> + *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
> + *			     unsigned long gpuva_start, unsigned long gpuva_end)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = {};
> + *		int err;
> + *
> + *		driver_svm_lock();
> + *	retry:
> + *		// Always process UNMAPs first so view of GPU SVM ranges is current
> + *		driver_garbage_collector(gpusvm);
> + *
> + *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
> + *							gpuva_start, gpuva_end,
> + *						        &ctx);
> + *		if (IS_ERR(range)) {
> + *			err = PTR_ERR(range);
> + *			goto unlock;
> + *		}
> + *
> + *		if (driver_migration_policy(range)) {
> + *			devmem = driver_alloc_devmem();
> + *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
> + *							   devmem_allocation,
> + *							   &ctx);
> + *			if (err)	// CPU mappings may have changed
> + *				goto retry;
> + *		}
> + *
> + *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
> + *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
> + *			if (err == -EOPNOTSUPP)
> + *				drm_gpusvm_range_evict(gpusvm, range);
> + *			goto retry;
> + *		} else if (err) {
> + *			goto unlock;
> + *		}
> + *
> + *		err = driver_bind_range(gpusvm, range);
> + *		if (err == -EAGAIN)	// CPU mappings changed
> + *			goto retry
> + *
> + *	unlock:
> + *		driver_svm_unlock();
> + *		return err;
> + *	}
> + *
> + * 2) Garbage Collector.
> + *
> + *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
> + *					struct drm_gpusvm_range *range)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		// Partial unmap, migrate any remaining device memory pages back to RAM
> + *		if (range->flags.partial_unmap)
> + *			drm_gpusvm_range_evict(gpusvm, range);
> + *
> + *		driver_unbind_range(range);
> + *		drm_gpusvm_range_remove(gpusvm, range);
> + *	}
> + *
> + *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		for_each_range_in_garbage_collector(gpusvm, range)
> + *			__driver_garbage_collector(gpusvm, range);
> + *	}
> + *
> + * 3) Notifier callback.
> + *
> + *	void driver_invalidation(struct drm_gpusvm *gpusvm,
> + *				 struct drm_gpusvm_notifier *notifier,
> + *				 const struct mmu_notifier_range *mmu_range)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
> + *		struct drm_gpusvm_range *range = NULL;
> + *
> + *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
> + *
> + *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
> + *					  mmu_range->end) {
> + *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
> + *
> + *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
> + *				continue;
> + *
> + *			drm_gpusvm_range_set_unmapped(range, mmu_range);
> + *			driver_garbage_collector_add(gpusvm, range);
> + *		}
> + *	}
> + */
> +
> +/**
> + * npages_in_range() - Calculate the number of pages in a given range
> + * @start: The start address of the range
> + * @end: The end address of the range
> + *
> + * This macro calculates the number of pages in a given memory range,
> + * specified by the start and end addresses. It divides the difference
> + * between the end and start addresses by the page size (PAGE_SIZE) to
> + * determine the number of pages in the range.
> + *
> + * Return: The number of pages in the specified range.
> + */
> +static unsigned long
> +npages_in_range(unsigned long start, unsigned long end)
> +{
> +	return (end - start) >> PAGE_SHIFT;
> +}
> +
> +/**
> + * struct drm_gpusvm_zdd - GPU SVM zone device data
> + *
> + * @refcount: Reference count for the zdd
> + * @destroy_work: Work structure for asynchronous zdd destruction
> + * @devmem_allocation: device memory allocation
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This structure serves as a generic wrapper installed in
> + * page->zone_device_data. It provides infrastructure for looking up a device
> + * memory allocation upon CPU page fault and asynchronously releasing device
> + * memory once the CPU has no page references. Asynchronous release is useful
> + * because CPU page references can be dropped in IRQ contexts, while releasing
> + * device memory likely requires sleeping locks.
> + */
> +struct drm_gpusvm_zdd {
> +	struct kref refcount;
> +	struct work_struct destroy_work;
> +	struct drm_gpusvm_devmem *devmem_allocation;
> +	void *device_private_page_owner;
> +};
> +
> +/**
> + * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
> + * @w: Pointer to the work_struct
> + *
> + * This function releases device memory, puts GPU SVM range, and frees zdd.
> + */
> +static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(w, struct drm_gpusvm_zdd, destroy_work);
> +	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
> +		zdd->devmem_allocation->ops : NULL;
> +
> +	if (zdd->devmem_allocation && ops->devmem_release)
> +		ops->devmem_release(zdd->devmem_allocation);
> +	kfree(zdd);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This function allocates and initializes a new zdd structure. It sets up the
> + * reference count and initializes the destroy work.
> + *
> + * Returns:
> + * Pointer to the allocated zdd on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_zdd *
> +drm_gpusvm_zdd_alloc(void *device_private_page_owner)
> +{
> +	struct drm_gpusvm_zdd *zdd;
> +
> +	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
> +	if (!zdd)
> +		return NULL;
> +
> +	kref_init(&zdd->refcount);
> +	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
> +	zdd->devmem_allocation = NULL;
> +	zdd->device_private_page_owner = device_private_page_owner;
> +
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function increments the reference count of the provided zdd structure.
> + *
> + * Returns: Pointer to the zdd structure.
> + */
> +static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_get(&zdd->refcount);
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
> + * @ref: Pointer to the reference count structure.
> + *
> + * This function queues the destroy_work of the zdd for asynchronous destruction.
> + */
> +static void drm_gpusvm_zdd_destroy(struct kref *ref)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(ref, struct drm_gpusvm_zdd, refcount);
> +
> +	if (zdd->devmem_allocation)
> +		WRITE_ONCE(zdd->devmem_allocation->detached, true);
> +	schedule_work(&zdd->destroy_work);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_put() - Put a zdd reference.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function decrements the reference count of the provided zdd structure
> + * and schedules its destruction if the count drops to zero.
> + */
> +static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
> +}
> +
> +/**
> + * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
> + * @notifier: Pointer to the GPU SVM notifier structure.
> + * @start: Start address of the range
> + * @end: End address of the range
> + *
> + * Return: A pointer to the drm_gpusvm_range if found or NULL
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_range, itree);
> +	else
> +		return NULL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
> +
> +/**
> + * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges
> + * @next__: Iterator variable for the ranges temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier while
> + * removing ranges from it.
> + */
> +#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
> +	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
> +	     (next__) = __drm_gpusvm_range_next(range__);				\
> +	     (range__) && (range__->itree.start < (end__));				\
> +	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
> +
> +/**
> + * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
> + * @notifier: a pointer to the current drm_gpusvm_notifier
> + *
> + * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
> + *         the current notifier is the last one or if the input notifier is
> + *         NULL.
> + */
> +static struct drm_gpusvm_notifier *
> +__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
> +{
> +	if (notifier && !list_is_last(&notifier->entry,
> +				      &notifier->gpusvm->notifier_list))
> +		return list_next_entry(notifier, entry);
> +
> +	return NULL;
> +}
> +
> +static struct drm_gpusvm_notifier *
> +notifier_iter_first(struct rb_root_cached *root, unsigned long start,
> +		    unsigned long last)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(root, start, last);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_notifier, itree);
> +	else
> +		return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
> + */
> +#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @next__: Iterator variable for the notifiers temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
> + * removing notifiers from it.
> + */
> +#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
> +	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
> + * @mni: Pointer to the mmu_interval_notifier structure.
> + * @mmu_range: Pointer to the mmu_notifier_range structure.
> + * @cur_seq: Current sequence number.
> + *
> + * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
> + * notifier sequence number and calls the driver invalidate vfunc under
> + * gpusvm->notifier_lock.
> + *
> + * Returns:
> + * true if the operation succeeds, false otherwise.
> + */
> +static bool
> +drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
> +			       const struct mmu_notifier_range *mmu_range,
> +			       unsigned long cur_seq)
> +{
> +	struct drm_gpusvm_notifier *notifier =
> +		container_of(mni, typeof(*notifier), notifier);
> +	struct drm_gpusvm *gpusvm = notifier->gpusvm;
> +
> +	if (!mmu_notifier_range_blockable(mmu_range))
> +		return false;
> +
> +	down_write(&gpusvm->notifier_lock);
> +	mmu_interval_set_seq(mni, cur_seq);
> +	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
> +	up_write(&gpusvm->notifier_lock);
> +
> +	return true;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
> + */
> +static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
> +	.invalidate = drm_gpusvm_notifier_invalidate,
> +};
> +
> +/**
> + * drm_gpusvm_init() - Initialize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @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.
> + * @ops: Pointer to the operations structure for GPU SVM.
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order with last
> + *               entry being SZ_4K.
> + * @num_chunks: Number of chunks.
> + *
> + * This function initializes the GPU SVM.
> + *
> + * Returns:
> + * 0 on success, a negative error code on failure.
> + */
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks)
> +{
> +	if (!ops->invalidate || !num_chunks)
> +		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;
> +	gpusvm->ops = ops;
> +	gpusvm->chunk_sizes = chunk_sizes;
> +	gpusvm->num_chunks = num_chunks;
> +
> +	mmgrab(mm);
> +	gpusvm->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&gpusvm->notifier_list);
> +
> +	init_rwsem(&gpusvm->notifier_lock);
> +
> +	fs_reclaim_acquire(GFP_KERNEL);
> +	might_lock(&gpusvm->notifier_lock);
> +	fs_reclaim_release(GFP_KERNEL);
> +
> +#ifdef CONFIG_LOCKDEP
> +	gpusvm->lock_dep_map = NULL;
> +#endif
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_init);
> +
> +/**
> + * drm_gpusvm_notifier_find() - Find GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function finds the GPU SVM notifier associated with the fault address.
> + *
> + * Returns:
> + * Pointer to the GPU SVM notifier on success, NULL otherwise.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
> +			 unsigned long fault_addr)
> +{
> +	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
> +}
> +
> +/**
> + * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_notifier structure.
> + */
> +static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	interval_tree_insert(&notifier->itree, &gpusvm->root);
> +
> +	node = rb_prev(&notifier->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_notifier(node))->entry;
> +	else
> +		head = &gpusvm->notifier_list;
> +
> +	list_add(&notifier->entry, head);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM tructure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	interval_tree_remove(&notifier->itree, &gpusvm->root);
> +	list_del(&notifier->entry);
> +}
> +
> +/**
> + * drm_gpusvm_fini() - Finalize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + *
> + * This function finalizes the GPU SVM by cleaning up any remaining ranges and
> + * notifiers, and dropping a reference to struct MM.
> + */
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
> +{
> +	struct drm_gpusvm_notifier *notifier, *next;
> +
> +	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
> +		struct drm_gpusvm_range *range, *__next;
> +
> +		/*
> +		 * Remove notifier first to avoid racing with any invalidation
> +		 */
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +		notifier->flags.removed = true;
> +
> +		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
> +					       LONG_MAX)
> +			drm_gpusvm_range_remove(gpusvm, range);
> +	}
> +
> +	mmdrop(gpusvm->mm);
> +	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
> +
> +/**
> + * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function allocates and initializes the GPU SVM notifier structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	if (gpusvm->ops->notifier_alloc)
> +		notifier = gpusvm->ops->notifier_alloc();
> +	else
> +		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
> +
> +	if (!notifier)
> +		return ERR_PTR(-ENOMEM);
> +
> +	notifier->gpusvm = gpusvm;
> +	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
> +	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
> +	INIT_LIST_HEAD(&notifier->entry);
> +	notifier->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&notifier->range_list);
> +
> +	return notifier;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_free() - Free GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function frees the GPU SVM notifier structure.
> + */
> +static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
> +				     struct drm_gpusvm_notifier *notifier)
> +{
> +	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
> +
> +	if (gpusvm->ops->notifier_free)
> +		gpusvm->ops->notifier_free(notifier);
> +	else
> +		kfree(notifier);
> +}
> +
> +/**
> + * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_range structure.
> + */
> +static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_range, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_range_insert() - Insert GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function inserts the GPU SVM range into the notifier RB tree and list.
> + */
> +static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
> +				    struct drm_gpusvm_range *range)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	drm_gpusvm_notifier_lock(notifier->gpusvm);
> +	interval_tree_insert(&range->itree, &notifier->root);
> +
> +	node = rb_prev(&range->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_range(node))->entry;
> +	else
> +		head = &notifier->range_list;
> +
> +	list_add(&range->entry, head);
> +	drm_gpusvm_notifier_unlock(notifier->gpusvm);
> +}
> +
> +/**
> + * __drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This macro removes the GPU SVM range from the notifier RB tree and list.
> + */
> +static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
> +				      struct drm_gpusvm_range *range)
> +{
> +	interval_tree_remove(&range->itree, &notifier->root);
> +	list_del(&range->entry);
> +}
> +
> +/**
> + * drm_gpusvm_range_alloc() - Allocate GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @fault_addr: Fault address
> + * @chunk_size: Chunk size
> + * @migrate_devmem: Flag indicating whether to migrate device memory
> + *
> + * This function allocates and initializes the GPU SVM range structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_range *
> +drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
> +		       struct drm_gpusvm_notifier *notifier,
> +		       unsigned long fault_addr, unsigned long chunk_size,
> +		       bool migrate_devmem)
> +{
> +	struct drm_gpusvm_range *range;
> +
> +	if (gpusvm->ops->range_alloc)
> +		range = gpusvm->ops->range_alloc(gpusvm);
> +	else
> +		range = kzalloc(sizeof(*range), GFP_KERNEL);
> +
> +	if (!range)
> +		return ERR_PTR(-ENOMEM);
> +
> +	kref_init(&range->refcount);
> +	range->gpusvm = gpusvm;
> +	range->notifier = notifier;
> +	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
> +	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
> +	INIT_LIST_HEAD(&range->entry);
> +	range->notifier_seq = LONG_MAX;
> +	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
> +
> +	return range;
> +}
> +
> +/**
> + * drm_gpusvm_check_pages() - Check pages
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @start: Start address
> + * @end: End address
> + *
> + * Check if pages between start and end have been faulted in on the CPU. Use to
> + * prevent migration of pages without CPU backing store.
> + *
> + * Returns:
> + * True if pages have been faulted into CPU, False otherwise
> + */
> +static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
> +				   struct drm_gpusvm_notifier *notifier,
> +				   unsigned long start, unsigned long end)
> +{
> +	struct hmm_range hmm_range = {
> +		.default_flags = 0,
> +		.notifier = &notifier->notifier,
> +		.start = start,
> +		.end = end,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long *pfns;
> +	unsigned long npages = npages_in_range(start, end);
> +	int err, i;
> +
> +	mmap_assert_locked(gpusvm->mm);
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return false;
> +
> +	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +	hmm_range.hmm_pfns = pfns;
> +
> +	while (true) {
> +		err = hmm_range_fault(&hmm_range);
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	if (err)
> +		goto err_free;
> +
> +	for (i = 0; i < npages;) {
> +		if (!(pfns[i] & HMM_PFN_VALID)) {
> +			err = -EFAULT;
> +			goto err_free;
> +		}
> +		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
> +	}
> +
> +err_free:
> +	kvfree(pfns);
> +	return err ? false : true;
> +}
> +
> +/**
> + * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @vas: Pointer to the virtual memory area structure
> + * @fault_addr: Fault address
> + * @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
> + *
> + * 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
> + * memory area boundaries.
> + *
> + * Returns:
> + * Chunk size on success, LONG_MAX on failure.
> + */
> +static unsigned long
> +drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_notifier *notifier,
> +			    struct vm_area_struct *vas,
> +			    unsigned long fault_addr,
> +			    unsigned long gpuva_start,
> +			    unsigned long gpuva_end,
> +			    unsigned long check_pages_threshold)
> +{
> +	unsigned long start, end;
> +	int i = 0;
> +
> +retry:
> +	for (; i < gpusvm->num_chunks; ++i) {
> +		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
> +		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
> +
> +		if (start >= vas->vm_start && end <= vas->vm_end &&
> +		    start >= notifier->itree.start &&
> +		    end <= notifier->itree.last + 1 &&
> +		    start >= gpuva_start && end <= gpuva_end)
> +			break;
> +	}
> +
> +	if (i == gpusvm->num_chunks)
> +		return LONG_MAX;
> +
> +	/*
> +	 * If allocation more than page, ensure not to overlap with existing
> +	 * ranges.
> +	 */
> +	if (end - start != SZ_4K) {
> +		struct drm_gpusvm_range *range;
> +
> +		range = drm_gpusvm_range_find(notifier, start, end);
> +		if (range) {
> +			++i;
> +			goto retry;
> +		}
> +
> +		/*
> +		 * XXX: Only create range on pages CPU has faulted in. Without
> +		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
> +		 * process-many-malloc' fails. In the failure case, each process
> +		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
> +		 * ranges. When migrating the SVM ranges, some processes fail in
> +		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
> +		 * and then upon drm_gpusvm_range_get_pages device pages from
> +		 * other processes are collected + faulted in which creates all
> +		 * sorts of problems. Unsure exactly how this happening, also
> +		 * problem goes away if 'xe_exec_system_allocator --r
> +		 * process-many-malloc' mallocs at least 64k at a time.
> +		 */
> +		if (end - start <= check_pages_threshold &&
> +		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
> +			++i;
> +			goto retry;
> +		}
> +	}
> +
> +	return end - start;
> +}
> +
> +/**
> + * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + * @gpuva_start: Start address of GPUVA which mirrors CPU
> + * @gpuva_end: End address of GPUVA which mirrors CPU
> + * @ctx: GPU SVM context
> + *
> + * This function finds or inserts a newly allocated a GPU SVM range based on the
> + * fault address. Caller must hold a lock to protect range lookup and insertion.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +	struct drm_gpusvm_range *range;
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	bool notifier_alloc = false;
> +	unsigned long chunk_size;
> +	int err;
> +	bool migrate_devmem;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	if (fault_addr < gpusvm->mm_start ||
> +	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
> +		return ERR_PTR(-EINVAL);
> +
> +	if (!mmget_not_zero(mm))
> +		return ERR_PTR(-EFAULT);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
> +	if (!notifier) {
> +		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
> +		if (IS_ERR(notifier)) {
> +			err = PTR_ERR(notifier);
> +			goto err_mmunlock;
> +		}
> +		notifier_alloc = true;
> +		err = mmu_interval_notifier_insert(&notifier->notifier,
> +						   mm, notifier->itree.start,
> +						   notifier->itree.last -
> +						   notifier->itree.start + 1,
> +						   &drm_gpusvm_notifier_ops);
> +		if (err)
> +			goto err_notifier;
> +	}
> +
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, fault_addr);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_notifier_remove;
> +	}
> +
> +	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
> +		err = -EPERM;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
> +	if (range)
> +		goto out_mmunlock;
> +	/*
> +	 * XXX: Short-circuiting migration based on migrate_vma_* current
> +	 * limitations. If/when migrate_vma_* add more support, this logic will
> +	 * have to change.
> +	 */
> +	migrate_devmem = ctx->devmem_possible &&
> +		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
> +
> +	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
> +						 fault_addr, gpuva_start,
> +						 gpuva_end, migrate_devmem &&
> +						 ctx->check_pages_threshold);
> +	if (chunk_size == LONG_MAX) {
> +		err = -EINVAL;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
> +				       migrate_devmem);
> +	if (IS_ERR(range)) {
> +		err = PTR_ERR(range);
> +		goto err_notifier_remove;
> +	}
> +
> +	drm_gpusvm_range_insert(notifier, range);
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_insert(gpusvm, notifier);
> +
> +out_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +
> +	return range;
> +
> +err_notifier_remove:
> +	mmap_read_unlock(mm);
> +	if (notifier_alloc)
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +err_notifier:
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +err_mmunlock:
> +	mmput(mm);
> +	return ERR_PTR(err);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
> +
> +/**
> + * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @npages: Number of pages to unmap
> + *
> + * This function unmap pages associated with a GPU SVM range. Assumes and
> + * asserts correct locking is in place when called.
> + */
> +static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +					   struct drm_gpusvm_range *range,
> +					   unsigned long npages)
> +{
> +	unsigned long i, j;
> +	struct drm_pagemap *dpagemap = range->dpagemap;
> +	struct device *dev = gpusvm->drm->dev;
> +
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->flags.has_dma_mapping) {
> +		for (i = 0, j = 0; i < npages; j++) {
> +			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
> +
> +			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
> +				dma_unmap_page(dev,
> +					       addr->addr,
> +					       PAGE_SIZE << addr->order,
> +					       addr->dir);
> +			else if (dpagemap && dpagemap->ops->unmap_dma)
> +				dpagemap->ops->unmap_dma(dpagemap,
> +							 dev,
> +							 *addr);
> +			i += 1 << addr->order;
> +		}
> +		range->flags.has_devmem_pages = false;
> +		range->flags.has_dma_mapping = false;
> +		range->dpagemap = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function frees the dma address array associated with a GPU SVM range.
> + */
> +static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
> +					struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->dma_addr) {
> +		kvfree(range->dma_addr);
> +		range->dma_addr = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function removes the specified GPU SVM range and also removes the parent
> + * GPU SVM notifier if no more ranges remain in the notifier. The caller must
> + * hold a lock to protect range and notifier removal.
> + */
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
> +	if (WARN_ON_ONCE(!notifier))
> +		return;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +	drm_gpusvm_range_free_pages(gpusvm, range);
> +	__drm_gpusvm_range_remove(notifier, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	drm_gpusvm_range_put(range);
> +
> +	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
> +		if (!notifier->flags.removed)
> +			mmu_interval_notifier_remove(&notifier->notifier);
> +		drm_gpusvm_notifier_remove(gpusvm, notifier);
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +	}
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
> +
> +/**
> + * drm_gpusvm_range_get() - Get a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function increments the reference count of the specified GPU SVM range.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range)
> +{
> +	kref_get(&range->refcount);
> +
> +	return range;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
> +
> +/**
> + * drm_gpusvm_range_destroy() - Destroy GPU SVM range
> + * @refcount: Pointer to the reference counter embedded in the GPU SVM range
> + *
> + * This function destroys the specified GPU SVM range when its reference count
> + * reaches zero. If a custom range-free function is provided, it is invoked to
> + * free the range; otherwise, the range is deallocated using kfree().
> + */
> +static void drm_gpusvm_range_destroy(struct kref *refcount)
> +{
> +	struct drm_gpusvm_range *range =
> +		container_of(refcount, struct drm_gpusvm_range, refcount);
> +	struct drm_gpusvm *gpusvm = range->gpusvm;
> +
> +	if (gpusvm->ops->range_free)
> +		gpusvm->ops->range_free(range);
> +	else
> +		kfree(range);
> +}
> +
> +/**
> + * drm_gpusvm_range_put() - Put a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function decrements the reference count of the specified GPU SVM range
> + * and frees it when the count reaches zero.
> + */
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
> +{
> +	kref_put(&range->refcount, drm_gpusvm_range_destroy);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
> +
> +/**
> + * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called holding gpusvm->notifier_lock and as the last step before committing a
> + * GPU binding. This is akin to a notifier seqno check in the HMM documentation
> + * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
> + * function is required for finer grained checking (i.e., per range) if pages
> + * are valid.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
> +
> +/**
> + * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called without holding gpusvm->notifier_lock.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +static bool
> +drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
> +				      struct drm_gpusvm_range *range)
> +{
> +	bool pages_valid;
> +
> +	if (!range->dma_addr)
> +		return false;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
> +	if (!pages_valid)
> +		drm_gpusvm_range_free_pages(gpusvm, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	return pages_valid;
> +}
> +
> +/**
> + * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function gets pages for a GPU SVM range and ensures they are mapped for
> + * DMA access.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
> +	struct hmm_range hmm_range = {
> +		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
> +			HMM_PFN_REQ_WRITE),
> +		.notifier = notifier,
> +		.start = range->itree.start,
> +		.end = range->itree.last + 1,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct drm_gpusvm_zdd *zdd;
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long i, j;
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	unsigned long num_dma_mapped;
> +	unsigned int order = 0;
> +	unsigned long *pfns;
> +	struct page **pages;
> +	int err = 0;
> +	struct dev_pagemap *pagemap;
> +	struct drm_pagemap *dpagemap;
> +
> +retry:
> +	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
> +		goto set_seqno;
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return -ENOMEM;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	hmm_range.hmm_pfns = pfns;
> +	while (true) {
> +		mmap_read_lock(mm);
> +		err = hmm_range_fault(&hmm_range);
> +		mmap_read_unlock(mm);
> +
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	mmput(mm);
> +	if (err)
> +		goto err_free;
> +
> +	pages = (struct page **)pfns;
> +map_pages:
> +	/*
> +	 * Perform all dma mappings under the notifier lock to not
> +	 * access freed pages. A notifier will either block on
> +	 * the notifier lock or unmap dma.
> +	 */
> +	drm_gpusvm_notifier_lock(gpusvm);

Found this in some local testing with
'xe_exec_system_allocator --r threads-many-stride-free-race'

+       if (range->flags.unmapped) {
+               drm_gpusvm_notifier_unlock(gpusvm);
+               err = -EFAULT;
+               goto err_free;
+       }

Odd race where the range is unmapped in notifier, subsequently we
successfully get the pages and bind the range, a future notifier comes
in and we skip invalidating the GPU page tables as we view the range as
unmapped but still removed the dma mapping, and boom DMAR error.

The above change to return -EFAULT ensures the GPU fault handler loop is
restarted with a newly created range rather than binding the old one
which has been unmapped in the notifier.

Matt

> +	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		kvfree(pfns);
> +		goto retry;
> +	}
> +
> +	if (!range->dma_addr) {
> +		/* Unlock and restart mapping to allocate memory. */
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
> +						 GFP_KERNEL);
> +		if (!range->dma_addr) {
> +			err = -ENOMEM;
> +			goto err_free;
> +		}
> +		goto map_pages;
> +	}
> +
> +	zdd = NULL;
> +	num_dma_mapped = 0;
> +	for (i = 0, j = 0; i < npages; ++j) {
> +		struct page *page = hmm_pfn_to_page(pfns[i]);
> +
> +		order = hmm_pfn_to_map_order(pfns[i]);
> +		if (is_device_private_page(page) || is_device_coherent_page(page)) {
> +			if (zdd != page->zone_device_data && i > 0) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +			zdd = page->zone_device_data;
> +			if (pagemap != page->pgmap) {
> +				if (i > 0) {
> +					err = -EOPNOTSUPP;
> +					goto err_unmap;
> +				}
> +
> +				pagemap = page->pgmap;
> +				dpagemap = zdd->devmem_allocation->dpagemap;
> +				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
> +					/*
> +					 * Raced. This is not supposed to happen
> +					 * since hmm_range_fault() should've migrated
> +					 * this page to system.
> +					 */
> +					err = -EAGAIN;
> +					goto err_unmap;
> +				}
> +			}
> +			range->dma_addr[j] =
> +				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
> +						       page, order,
> +						       DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			pages[i] = page;
> +		} else {
> +			dma_addr_t addr;
> +
> +			if (is_zone_device_page(page) || zdd) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +
> +			addr = dma_map_page(gpusvm->drm->dev,
> +					    page, 0,
> +					    PAGE_SIZE << order,
> +					    DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			range->dma_addr[j] = drm_pagemap_dma_addr_encode
> +				(addr, DRM_INTERCONNECT_SYSTEM, order,
> +				 DMA_BIDIRECTIONAL);
> +		}
> +		i += 1 << order;
> +		num_dma_mapped = i;
> +	}
> +
> +	range->flags.has_dma_mapping = true;
> +	if (zdd) {
> +		range->flags.has_devmem_pages = true;
> +		range->dpagemap = dpagemap;
> +	}
> +
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +	kvfree(pfns);
> +set_seqno:
> +	range->notifier_seq = hmm_range.notifier_seq;
> +
> +	return 0;
> +
> +err_unmap:
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +err_free:
> +	kvfree(pfns);
> +	if (err == -EAGAIN)
> +		goto retry;
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
> +
> +/**
> + * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function unmaps pages associated with a GPU SVM range. If @in_notifier
> + * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
> + * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
> + * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
> + * security model.
> + */
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +
> +	if (ctx->in_notifier)
> +		lockdep_assert_held_write(&gpusvm->notifier_lock);
> +	else
> +		drm_gpusvm_notifier_lock(gpusvm);
> +
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +
> +	if (!ctx->in_notifier)
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_page() - Put a migration page
> + * @page: Pointer to the page to put
> + *
> + * This function unlocks and puts a page.
> + */
> +static void drm_gpusvm_migration_unlock_put_page(struct page *page)
> +{
> +	unlock_page(page);
> +	put_page(page);
> +}
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
> + * @npages: Number of pages
> + * @migrate_pfn: Array of migrate page frame numbers
> + *
> + * This function unlocks and puts an array of pages.
> + */
> +static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
> +						  unsigned long *migrate_pfn)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page;
> +
> +		if (!migrate_pfn[i])
> +			continue;
> +
> +		page = migrate_pfn_to_page(migrate_pfn[i]);
> +		drm_gpusvm_migration_unlock_put_page(page);
> +		migrate_pfn[i] = 0;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
> + * @page: Pointer to the page
> + * @zdd: Pointer to the GPU SVM zone device data
> + *
> + * This function associates the given page with the specified GPU SVM zone
> + * device data and initializes it for zone device usage.
> + */
> +static void drm_gpusvm_get_devmem_page(struct page *page,
> +				     struct drm_gpusvm_zdd *zdd)
> +{
> +	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
> +	zone_device_page_init(page);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
> + * @dev: The device for which the pages are being mapped
> + * @dma_addr: Array to store DMA addresses corresponding to mapped pages
> + * @migrate_pfn: Array of migrate page frame numbers to map
> + * @npages: Number of pages to map
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function maps pages of memory for migration usage in GPU SVM. It
> + * iterates over each page frame number provided in @migrate_pfn, maps the
> + * corresponding page, and stores the DMA address in the provided @dma_addr
> + * array.
> + *
> + * Return: 0 on success, -EFAULT if an error occurs during mapping.
> + */
> +static int drm_gpusvm_migrate_map_pages(struct device *dev,
> +					dma_addr_t *dma_addr,
> +					unsigned long *migrate_pfn,
> +					unsigned long npages,
> +					enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		if (WARN_ON_ONCE(is_zone_device_page(page)))
> +			return -EFAULT;
> +
> +		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
> +		if (dma_mapping_error(dev, dma_addr[i]))
> +			return -EFAULT;
> +	}
> +
> +	return 0;
> +}
> +
> +/**
> + * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
> + * @dev: The device for which the pages were mapped
> + * @dma_addr: Array of DMA addresses corresponding to mapped pages
> + * @npages: Number of pages to unmap
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function unmaps previously mapped pages of memory for GPU Shared Virtual
> + * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
> + * if it's valid and not already unmapped, and unmaps the corresponding page.
> + */
> +static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
> +					   dma_addr_t *dma_addr,
> +					   unsigned long npages,
> +					   enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
> +			continue;
> +
> +		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @devmem_allocation: Pointer to the device memory allocation. The caller
> + *                     should hold a reference to the device memory allocation,
> + *                     which should be dropped via ops->devmem_release or upon
> + *                     the failure of this function.
> + * @ctx: GPU SVM context
> + *
> + * This function migrates the specified GPU SVM range to device memory. It performs the
> + * necessary setup and invokes the driver-specific operations for migration to
> + * device memory. Upon successful return, @devmem_allocation can safely reference @range
> + * until ops->devmem_release is called which only upon successful return.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +	struct migrate_vma migrate = {
> +		.start		= start,
> +		.end		= end,
> +		.pgmap_owner	= gpusvm->device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	unsigned long i, npages = npages_in_range(start, end);
> +	struct vm_area_struct *vas;
> +	struct drm_gpusvm_zdd *zdd = NULL;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int err;
> +
> +	if (!range->flags.migrate_devmem)
> +		return -EINVAL;
> +
> +	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
> +	    !ops->copy_to_ram)
> +		return -EOPNOTSUPP;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_out;
> +	}
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, start);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_mmunlock;
> +	}
> +
> +	if (end > vas->vm_end || start < vas->vm_start) {
> +		err = -EINVAL;
> +		goto err_mmunlock;
> +	}
> +
> +	if (!vma_is_anonymous(vas)) {
> +		err = -EBUSY;
> +		goto err_mmunlock;
> +	}
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_mmunlock;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
> +	if (!zdd) {
> +		err = -ENOMEM;
> +		goto err_free;
> +	}
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	if (!migrate.cpages) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	if (migrate.cpages != npages) {
> +		err = -EBUSY;
> +		goto err_finalize;
> +	}
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   migrate.src, npages, DMA_TO_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = pfn_to_page(migrate.dst[i]);
> +
> +		pages[i] = page;
> +		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
> +		drm_gpusvm_get_devmem_page(page, zdd);
> +	}
> +
> +	err = ops->copy_to_devmem(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +	/* Upon success bind devmem allocation to range and zdd */
> +	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_TO_DEVICE);
> +err_free:
> +	if (zdd)
> +		drm_gpusvm_zdd_put(zdd);
> +	kvfree(buf);
> +err_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +err_out:
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
> +
> +/**
> + * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
> + * @vas: Pointer to the VM area structure, can be NULL
> + * @fault_page: Fault page
> + * @npages: Number of pages to populate
> + * @mpages: Number of pages to migrate
> + * @src_mpfn: Source array of migrate PFNs
> + * @mpfn: Array of migrate PFNs to populate
> + * @addr: Start address for PFN allocation
> + *
> + * This function populates the RAM migrate page frame numbers (PFNs) for the
> + * specified VM area structure. It allocates and locks pages in the VM area for
> + * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
> + * alloc_page for allocation.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> +					       struct page *fault_page,
> +					       unsigned long npages,
> +					       unsigned long *mpages,
> +					       unsigned long *src_mpfn,
> +					       unsigned long *mpfn,
> +					       unsigned long addr)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
> +		struct page *page, *src_page;
> +
> +		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
> +			continue;
> +
> +		if (fault_page) {
> +			src_page = migrate_pfn_to_page(src_mpfn[i]);
> +			if (src_page->zone_device_data !=
> +			    fault_page->zone_device_data)
> +				continue;
> +		}
> +
> +		if (vas)
> +			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
> +		else
> +			page = alloc_page(GFP_HIGHUSER);
> +
> +		if (!page)
> +			goto free_pages;
> +
> +		mpfn[i] = migrate_pfn(page_to_pfn(page));
> +	}
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		WARN_ON_ONCE(!trylock_page(page));
> +		++*mpages;
> +	}
> +
> +	return 0;
> +
> +free_pages:
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		put_page(page);
> +		mpfn[i] = 0;
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
> + * @devmem_allocation: Pointer to the device memory allocation
> + *
> + * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
> + * migration done via migrate_device_* functions.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	unsigned long *src, *dst;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int i, err = 0;
> +
> +	npages = devmem_allocation->size >> PAGE_SHIFT;
> +
> +retry:
> +	if (!mmget_not_zero(devmem_allocation->mm))
> +		return -EFAULT;
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	src = buf;
> +	dst = buf + (sizeof(*src) * npages);
> +	dma_addr = buf + (2 * sizeof(*src) * npages);
> +	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
> +	if (err)
> +		goto err_free;
> +
> +	err = migrate_device_pfns(src, npages);
> +	if (err)
> +		goto err_free;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
> +						  src, dst, 0);
> +	if (err || !mpages)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   dst, npages, DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, dst);
> +	migrate_device_pages(src, dst, npages);
> +	migrate_device_finalize(src, dst, npages);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +	mmput_async(devmem_allocation->mm);
> +	if (!err && !READ_ONCE(devmem_allocation->detached)) {
> +		cond_resched();
> +		goto retry;
> +	}
> +
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
> +
> +/**
> + * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
> + * @vas: Pointer to the VM area structure
> + * @device_private_page_owner: Device private pages owner
> + * @page: Pointer to the page for fault handling (can be NULL)
> + * @fault_addr: Fault address
> + * @size: Size of migration
> + *
> + * This internal function performs the migration of the specified GPU SVM range
> + * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
> + * invokes the driver-specific operations for migration to RAM.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
> +				       void *device_private_page_owner,
> +				       struct page *page,
> +				       unsigned long fault_addr,
> +				       unsigned long size)
> +{
> +	struct migrate_vma migrate = {
> +		.vma		= vas,
> +		.pgmap_owner	= device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
> +			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
> +		.fault_page	= page,
> +	};
> +	struct drm_gpusvm_zdd *zdd;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct device *dev;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	unsigned long start, end;
> +	void *buf;
> +	int i, err = 0;
> +
> +	start = ALIGN_DOWN(fault_addr, size);
> +	end = ALIGN(fault_addr + 1, size);
> +
> +	/* Corner where VMA area struct has been partially unmapped */
> +	if (start < vas->vm_start)
> +		start = vas->vm_start;
> +	if (end > vas->vm_end)
> +		end = vas->vm_end;
> +
> +	migrate.start = start;
> +	migrate.end = end;
> +	npages = npages_in_range(start, end);
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	/* Raced with another CPU fault, nothing to do */
> +	if (!migrate.cpages)
> +		goto err_free;
> +
> +	if (!page) {
> +		for (i = 0; i < npages; ++i) {
> +			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
> +				continue;
> +
> +			page = migrate_pfn_to_page(migrate.src[i]);
> +			break;
> +		}
> +
> +		if (!page)
> +			goto err_finalize;
> +	}
> +	zdd = page->zone_device_data;
> +	ops = zdd->devmem_allocation->ops;
> +	dev = zdd->devmem_allocation->dev;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
> +						  migrate.src, migrate.dst,
> +						  start);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
> +					   DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(migrate.src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +
> +	return err;
> +}
> +
> +/**
> + * drm_gpusvm_range_evict() - Evict GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function evicts the specified GPU SVM range.
> + */
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range)
> +{
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +
> +	if (!mmget_not_zero(mm))
> +		return;
> +
> +	mmap_read_lock(mm);
> +
> +	/*
> +	 * Loop required to find all VMAs for the corner case when partial
> +	 * unmappings have occurred.
> +	 */
> +again:
> +	vas = find_vma(mm, start);
> +	if (!vas)
> +		goto unlock;
> +
> +	if (end <= vas->vm_start || start >= vas->vm_end)
> +		goto unlock;
> +
> +	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
> +				    NULL, start,
> +				    range->itree.last - range->itree.start + 1);
> +	if (vas->vm_end < end) {
> +		start = vas->vm_end;
> +		goto again;
> +	}
> +unlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
> +
> +/**
> + * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
> + * @page: Pointer to the page
> + *
> + * This function is a callback used to put the GPU SVM zone device data
> + * associated with a page when it is being released.
> + */
> +static void drm_gpusvm_page_free(struct page *page)
> +{
> +	drm_gpusvm_zdd_put(page->zone_device_data);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
> + * @vmf: Pointer to the fault information structure
> + *
> + * This function is a page fault handler used to migrate a GPU SVM range to RAM.
> + * It retrieves the GPU SVM range information from the faulting page and invokes
> + * the internal migration function to migrate the range back to RAM.
> + *
> + * Returns:
> + * VM_FAULT_SIGBUS on failure, 0 on success.
> + */
> +static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
> +{
> +	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
> +	int err;
> +
> +	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
> +					  zdd->device_private_page_owner,
> +					  vmf->page, vmf->address,
> +					  zdd->devmem_allocation->size);
> +
> +	return err ? VM_FAULT_SIGBUS : 0;
> +}
> +
> +/**
> + * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
> + */
> +static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
> +	.page_free = drm_gpusvm_page_free,
> +	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
> +};
> +
> +/**
> + * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
> + *
> + * Returns:
> + * Pointer to the GPU SVM device page map operations structure.
> + */
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
> +{
> +	return &drm_gpusvm_pagemap_ops;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
> +
> +/**
> + * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @start: Start address
> + * @end: End address
> + *
> + * Returns:
> + * True if GPU SVM has mapping, False otherwise
> + */
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
> +		struct drm_gpusvm_range *range = NULL;
> +
> +		drm_gpusvm_for_each_range(range, notifier, start, end)
> +			return true;
> +	}
> +
> +	return false;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
> +
> +/**
> + * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
> + * @range: Pointer to the GPU SVM range structure.
> + * @mmu_range: Pointer to the MMU notifier range structure.
> + *
> + * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
> + * if the range partially falls within the provided MMU notifier range.
> + */
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range)
> +{
> +	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
> +
> +	range->flags.unmapped = true;
> +	if (range->itree.start < mmu_range->start ||
> +	    range->itree.last + 1 > mmu_range->end)
> +		range->flags.partial_unmap = true;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
> +
> +/**
> + * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap we're allocating from.
> + * @size: Size of device memory allocation
> + */
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size)
> +{
> +	devmem_allocation->dev = dev;
> +	devmem_allocation->mm = mm;
> +	devmem_allocation->ops = ops;
> +	devmem_allocation->dpagemap = dpagemap;
> +	devmem_allocation->size = size;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
> +
> +MODULE_DESCRIPTION("DRM GPUSVM");
> +MODULE_LICENSE("GPL");
> diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> new file mode 100644
> index 000000000000..2fb6980d0cf1
> --- /dev/null
> +++ b/include/drm/drm_gpusvm.h
> @@ -0,0 +1,445 @@
> +/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
> +/*
> + * Copyright © 2024 Intel Corporation
> + */
> +
> +#ifndef __DRM_GPUSVM_H__
> +#define __DRM_GPUSVM_H__
> +
> +#include <linux/kref.h>
> +#include <linux/interval_tree.h>
> +#include <linux/mmu_notifier.h>
> +
> +struct dev_pagemap_ops;
> +struct drm_device;
> +struct drm_gpusvm;
> +struct drm_gpusvm_notifier;
> +struct drm_gpusvm_ops;
> +struct drm_gpusvm_range;
> +struct drm_gpusvm_devmem;
> +struct drm_pagemap;
> +struct drm_pagemap_dma_addr;
> +
> +/**
> + * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM)
> + * device memory. These operations are provided by the GPU driver to manage device memory
> + * allocations and perform operations such as migration between device memory and system
> + * RAM.
> + */
> +struct drm_gpusvm_devmem_ops {
> +	/**
> +	 * @devmem_release: Release device memory allocation (optional)
> +	 * @devmem_allocation: device memory allocation
> +	 *
> +	 * Release device memory allocation and drop a reference to device
> +	 * memory allocation.
> +	 */
> +	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +	/**
> +	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
> +	 * @devmem_allocation: device memory allocation
> +	 * @npages: Number of pages to populate
> +	 * @pfn: Array of page frame numbers to populate
> +	 *
> +	 * Populate device memory page frame numbers (PFN).
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
> +				   unsigned long npages, unsigned long *pfn);
> +
> +	/**
> +	 * @copy_to_devmem: Copy to device memory (required for migration)
> +	 * @pages: Pointer to array of device memory pages (destination)
> +	 * @dma_addr: Pointer to array of DMA addresses (source)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to device memory.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_devmem)(struct page **pages,
> +			      dma_addr_t *dma_addr,
> +			      unsigned long npages);
> +
> +	/**
> +	 * @copy_to_ram: Copy to system RAM (required for migration)
> +	 * @pages: Pointer to array of device memory pages (source)
> +	 * @dma_addr: Pointer to array of DMA addresses (destination)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to system RAM.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_ram)(struct page **pages,
> +			   dma_addr_t *dma_addr,
> +			   unsigned long npages);
> +};
> +
> +/**
> + * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
> + * @size: Size of device memory allocation
> + * @detached: device memory allocations is detached from device pages
> + */
> +struct drm_gpusvm_devmem {
> +	struct device *dev;
> +	struct mm_struct *mm;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct drm_pagemap *dpagemap;
> +	size_t size;
> +	bool detached;
> +};
> +
> +/**
> + * struct drm_gpusvm_ops - Operations structure for GPU SVM
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM).
> + * These operations are provided by the GPU driver to manage SVM ranges and
> + * notifiers.
> + */
> +struct drm_gpusvm_ops {
> +	/**
> +	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
> +	 *
> +	 * Allocate a GPU SVM notifier.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
> +
> +	/**
> +	 * @notifier_free: Free a GPU SVM notifier (optional)
> +	 * @notifier: Pointer to the GPU SVM notifier to be freed
> +	 *
> +	 * Free a GPU SVM notifier.
> +	 */
> +	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
> +
> +	/**
> +	 * @range_alloc: Allocate a GPU SVM range (optional)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 *
> +	 * Allocate a GPU SVM range.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
> +
> +	/**
> +	 * @range_free: Free a GPU SVM range (optional)
> +	 * @range: Pointer to the GPU SVM range to be freed
> +	 *
> +	 * Free a GPU SVM range.
> +	 */
> +	void (*range_free)(struct drm_gpusvm_range *range);
> +
> +	/**
> +	 * @invalidate: Invalidate GPU SVM notifier (required)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 * @notifier: Pointer to the GPU SVM notifier
> +	 * @mmu_range: Pointer to the mmu_notifier_range structure
> +	 *
> +	 * Invalidate the GPU page tables. It can safely walk the notifier range
> +	 * RB tree/list in this function. Called while holding the notifier lock.
> +	 */
> +	void (*invalidate)(struct drm_gpusvm *gpusvm,
> +			   struct drm_gpusvm_notifier *notifier,
> +			   const struct mmu_notifier_range *mmu_range);
> +};
> +
> +/**
> + * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: MMU interval notifier
> + * @itree: Interval tree node for the notifier (inserted in GPU SVM)
> + * @entry: List entry to fast interval tree traversal
> + * @root: Cached root node of the RB tree containing ranges
> + * @range_list: List head containing of ranges in the same order they appear in
> + *              interval tree. This is useful to keep iterating ranges while
> + *              doing modifications to RB tree.
> + * @flags.removed: Flag indicating whether the MMU interval notifier has been
> + *                 removed
> + *
> + * This structure represents a GPU SVM notifier.
> + */
> +struct drm_gpusvm_notifier {
> +	struct drm_gpusvm *gpusvm;
> +	struct mmu_interval_notifier notifier;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	struct rb_root_cached root;
> +	struct list_head range_list;
> +	struct {
> +		u32 removed : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm_range - Structure representing a GPU SVM range
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier
> + * @refcount: Reference count for the range
> + * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
> + * @entry: List entry to fast interval tree traversal
> + * @notifier_seq: Notifier sequence number of the range's pages
> + * @dma_addr: DMA address array
> + * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
> + *            Note this is assuming only one drm_pagemap per range is allowed.
> + * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
> + * @flags.unmapped: Flag indicating if the range has been unmapped
> + * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
> + * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
> + * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
> + *
> + * This structure represents a GPU SVM range used for tracking memory ranges
> + * mapped in a DRM device.
> + */
> +struct drm_gpusvm_range {
> +	struct drm_gpusvm *gpusvm;
> +	struct drm_gpusvm_notifier *notifier;
> +	struct kref refcount;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	unsigned long notifier_seq;
> +	struct drm_pagemap_dma_addr *dma_addr;
> +	struct drm_pagemap *dpagemap;
> +	struct {
> +		/* All flags below must be set upon creation */
> +		u16 migrate_devmem : 1;
> +		/* All flags below must be set / cleared under notifier lock */
> +		u16 unmapped : 1;
> +		u16 partial_unmap : 1;
> +		u16 has_devmem_pages : 1;
> +		u16 has_dma_mapping : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm - GPU SVM structure
> + *
> + * @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
> + * @ops: Pointer to the operations structure for GPU SVM
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order.
> + * @num_chunks: Number of chunks
> + * @notifier_lock: Read-write semaphore for protecting notifier operations
> + * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
> + * @notifier_list: list head containing of notifiers in the same order they
> + *                 appear in interval tree. This is useful to keep iterating
> + *                 notifiers while doing modifications to RB tree.
> + *
> + * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
> + * memory ranges mapped in a DRM (Direct Rendering Manager) device.
> + *
> + * No reference counting is provided, as this is expected to be embedded in the
> + * driver VM structure along with the struct drm_gpuvm, which handles reference
> + * counting.
> + */
> +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;
> +	const struct drm_gpusvm_ops *ops;
> +	const unsigned long *chunk_sizes;
> +	int num_chunks;
> +	struct rw_semaphore notifier_lock;
> +	struct rb_root_cached root;
> +	struct list_head notifier_list;
> +#ifdef CONFIG_LOCKDEP
> +	/**
> +	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
> +	 * drm_gpusvm_range_remove with a driver provided lock.
> +	 */
> +	struct lockdep_map *lock_dep_map;
> +#endif
> +};
> +
> +/**
> + * struct drm_gpusvm_ctx - DRM GPU SVM context
> + *
> + * @check_pages_threshold: Check CPU pages for present if chunk is less than or
> + *                         equal to threshold. If not present, reduce chunk
> + *                         size.
> + * @in_notifier: entering from a MMU notifier
> + * @read_only: operating on read-only memory
> + * @devmem_possible: possible to use device memory
> + *
> + * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> + */
> +struct drm_gpusvm_ctx {
> +	unsigned long check_pages_threshold;
> +	unsigned int in_notifier :1;
> +	unsigned int read_only :1;
> +	unsigned int devmem_possible :1;
> +};
> +
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks);
> +
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
> +
> +void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
> +
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range);
> +
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
> +
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end);
> +
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range);
> +
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size);
> +
> +#ifdef CONFIG_LOCKDEP
> +/**
> + * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @lock: the lock used to protect the gpuva list. The locking primitive
> + * must contain a dep_map field.
> + *
> + * Call this to annotate drm_gpusvm_range_find_or_insert and
> + * drm_gpusvm_range_remove.
> + */
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
> +	do { \
> +		if (!WARN((gpusvm)->lock_dep_map, \
> +			  "GPUSVM range lock should be set only once."))\
> +			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
> +	} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) \
> +	do { \
> +		if ((gpusvm)->lock_dep_map)	\
> +			lock_is_held((gpusvm)->lock_dep_map);	\
> +	} while (0)
> +#else
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
> +#endif
> +
> +/**
> + * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, take lock
> + */
> +#define drm_gpusvm_notifier_lock(gpusvm__)	\
> +	down_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, drop lock
> + */
> +#define drm_gpusvm_notifier_unlock(gpusvm__)	\
> +	up_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
> + * @range: a pointer to the current GPU SVM range
> + *
> + * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
> + *         current range is the last one or if the input range is NULL.
> + */
> +static inline struct drm_gpusvm_range *
> +__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
> +{
> +	if (range && !list_is_last(&range->entry,
> +				   &range->notifier->range_list))
> +		return list_next_entry(range, entry);
> +
> +	return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges. If set, it indicates the start of
> + *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
> + * to use while holding the driver SVM lock or the notifier lock.
> + */
> +#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
> +	for ((range__) = (range__) ?:					\
> +	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
> +	     (range__) && (range__->itree.start < (end__));		\
> +	     (range__) = __drm_gpusvm_range_next(range__))
> +
> +#endif /* __DRM_GPUSVM_H__ */
> -- 
> 2.34.1
> 

Re: [PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Gwan-gyeong Mun]

On 12/18/24 1:33 AM, Matthew Brost wrote:
> This patch introduces support for GPU Shared Virtual Memory (SVM) in the
> Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
> sharing of memory between the CPU and GPU, enhancing performance and
> flexibility in GPU computing tasks.
> 
> The patch adds the necessary infrastructure for SVM, including data
> structures and functions for managing SVM ranges and notifiers. It also
> provides mechanisms for allocating, deallocating, and migrating memory
> regions between system RAM and GPU VRAM.
> 
> This is largely inspired by GPUVM.
> 
> v2:
>   - Take order into account in check pages
>   - Clear range->pages in get pages error
>   - Drop setting dirty or accessed bit in get pages (Vetter)
>   - Remove mmap assert for cpu faults
>   - Drop mmap write lock abuse (Vetter, Christian)
>   - Decouple zdd from range (Vetter, Oak)
>   - Add drm_gpusvm_range_evict, make it work with coherent pages
>   - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
>   - mmget/put in drm_gpusvm_evict_to_sram
>   - Drop range->vram_alloation variable
>   - Don't return in drm_gpusvm_evict_to_sram until all pages detached
>   - Don't warn on mixing sram and device pages
>   - Update kernel doc
>   - Add coherent page support to get pages
>   - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
>   - Add struct drm_gpusvm_vram and ops (Thomas)
>   - Update the range's seqno if the range is valid (Thomas)
>   - Remove the is_unmapped check before hmm_range_fault (Thomas)
>   - Use drm_pagemap (Thomas)
>   - Drop kfree_mapping (Thomas)
>   - dma mapp pages under notifier lock (Thomas)
>   - Remove ctx.prefault
>   - Remove ctx.mmap_locked
>   - Add ctx.check_pages
>   - s/vram/devmem (Thomas)
> v3:
>   - Fix memory leak drm_gpusvm_range_get_pages
>   - Only migrate pages with same zdd on CPU fault
>   - Loop over al VMAs in drm_gpusvm_range_evict
>   - Make GPUSVM a drm level module
>   - GPL or MIT license
>   - Update main kernel doc (Thomas)
>   - Prefer foo() vs foo for functions in kernel doc (Thomas)
>   - Prefer functions over macros (Thomas)
>   - Use unsigned long vs u64 for addresses (Thomas)
>   - Use standard interval_tree (Thomas)
>   - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
>   - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
>   - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
>   - Newlines between functions defs in header file (Thomas)
>   - Drop shall language in driver vfunc kernel doc (Thomas)
>   - Move some static inlines from head to C file (Thomas)
>   - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
>   - Change check_pages to a threshold
> 
> Cc: Simona Vetter <simona.vetter@ffwll.ch>
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Christian König <christian.koenig@amd.com>
> Cc: <dri-devel@lists.freedesktop.org>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
>   drivers/gpu/drm/Kconfig      |    8 +
>   drivers/gpu/drm/Makefile     |    1 +
>   drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
>   include/drm/drm_gpusvm.h     |  445 +++++++
>   4 files changed, 2674 insertions(+)
>   create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>   create mode 100644 include/drm/drm_gpusvm.h
> 
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 2f51546b0b88..5524cb1ed9dd 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -278,6 +278,14 @@ config DRM_GPUVM
>   	  GPU-VM representation providing helpers to manage a GPUs virtual
>   	  address space
>   
> +config DRM_GPUSVM
> +	tristate
> +	depends on DRM
> +	depends on DEVICE_MIGRATION
> +	help
> +	  GPU-SVM representation providing helpers to manage a GPUs shared
> +	  virtual memory
> +
>   config DRM_BUDDY
>   	tristate
>   	depends on DRM
> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> index 19fb370fbc56..ca5d44441fb6 100644
> --- a/drivers/gpu/drm/Makefile
> +++ b/drivers/gpu/drm/Makefile
> @@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
>   #
>   obj-$(CONFIG_DRM_EXEC) += drm_exec.o
>   obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> +obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
>   
>   obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
>   
> diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> new file mode 100644
> index 000000000000..29ce2c7d1c8c
> --- /dev/null
> +++ b/drivers/gpu/drm/drm_gpusvm.c
> @@ -0,0 +1,2220 @@
> +// SPDX-License-Identifier: GPL-2.0-only OR MIT
> +/*
> + * Copyright © 2024 Intel Corporation
> + *
> + * Authors:
> + *     Matthew Brost <matthew.brost@intel.com>
> + */
> +
> +#include <linux/dma-mapping.h>
> +#include <linux/hmm.h>
> +#include <linux/memremap.h>
> +#include <linux/migrate.h>
> +#include <linux/mm_types.h>
> +#include <linux/pagemap.h>
> +#include <linux/slab.h>
> +
> +#include <drm/drm_device.h>
> +#include <drm/drm_gpusvm.h>
> +#include <drm/drm_pagemap.h>
> +#include <drm/drm_print.h>
> +
> +/**
> + * DOC: Overview
> + *
> + * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
> + *
> + * The GPU SVM layer is a component of the DRM framework designed to manage shared
> + * virtual memory between the CPU and GPU. It enables efficient data exchange and
> + * processing for GPU-accelerated applications by allowing memory sharing and
> + * synchronization between the CPU's and GPU's virtual address spaces.
> + *
> + * Key GPU SVM Components:
> + * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
> + *		GPU of changes, notifiers are sized based on a GPU SVM
> + *		initialization parameter, with a recommendation of 512M or
> + *		larger. They maintain a Red-BlacK tree and a list of ranges that
> + *		fall within the notifier interval. Notifiers are tracked within
> + *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
> + *		or removed as ranges within the interval are created or
> + *		destroyed.
> + * - Ranges: Represent memory ranges mapped in a DRM device and managed
> + *	     by GPU SVM. They are sized based on an array of chunk sizes, which
> + *	     is a GPU SVM initialization parameter, and the CPU address space.
> + *	     Upon GPU fault, the largest aligned chunk that fits within the
> + *	     faulting CPU address space is chosen for the range size. Ranges are
> + *	     expected to be dynamically allocated on GPU fault and removed on an
> + *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
> + *	     a notifier's Red-Black tree.
> + * - Operations: Define the interface for driver-specific GPU SVM operations
> + *               such as range allocation, notifier allocation, and
> + *               invalidations.
> + * - Device Memory Allocations: Embedded structure containing enough information
> + *                              for GPU SVM to migrate to / from device memory.
> + * - Device Memory Operations: Define the interface for driver-specific device
> + *                             memory operations release memory, populate pfns,
> + *                             and copy to / from device memory.
> + *
> + * This layer provides interfaces for allocating, mapping, migrating, and
> + * releasing memory ranges between the CPU and GPU. It handles all core memory
> + * management interactions (DMA mapping, HMM, and migration) and provides
> + * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
> + * to build the expected driver components for an SVM implementation as detailed
> + * below.
> + *
> + * Expected Driver Components:
> + * - GPU page fault handler: Used to create ranges and notifiers based on the
> + *			     fault address, optionally migrate the range to
> + *			     device memory, and create GPU bindings.
> + * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
> + *			Ranges are expected to be added to the garbage collector
> + *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
> + * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
> + *			ranges.
> + */
> +
> +/**
> + * DOC: Locking
> + *
> + * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
> + * mmap lock as needed.
> + *
> + * GPU SVM introduces a global notifier lock, which safeguards the notifier's
> + * range RB tree and list, as well as the range's DMA mappings and sequence
> + * number. GPU SVM manages all necessary locking and unlocking operations,
> + * except for the recheck range's pages being valid
> + * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
> + * lock corresponds to the 'driver->update' lock mentioned in the HMM
> + * documentation (TODO: Link). Future revisions may transition from a GPU SVM
> + * global lock to a per-notifier lock if finer-grained locking is deemed
> + * necessary.
> + *
> + * In addition to the locking mentioned above, the driver should implement a
> + * lock to safeguard core GPU SVM function calls that modify state, such as
> + * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
> + * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
> + * locking should also be possible for concurrent GPU fault processing within a
> + * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
> + * to add annotations to GPU SVM.
> + */
> +
> +/**
> + * DOC: Migration
> + *
> + * The migration support is quite simple, allowing migration between RAM and
> + * device memory at the range granularity. For example, GPU SVM currently does not
> + * support mixing RAM and device memory pages within a range. This means that upon GPU
> + * fault, the entire range can be migrated to device memory, and upon CPU fault, the
> + * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
> + * could be added in the future if required.
> + *
> + * The reasoning for only supporting range granularity is as follows: it
> + * simplifies the implementation, and range sizes are driver-defined and should
> + * be relatively small.
> + */
> +
> +/**
> + * DOC: Partial Unmapping of Ranges
> + *
> + * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
> + * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
> + * being that a subset of the range still has CPU and GPU mappings. If the
> + * backing store for the range is in device memory, a subset of the backing store has
> + * references. One option would be to split the range and device memory backing store,
> + * but the implementation for this would be quite complicated. Given that
> + * partial unmappings are rare and driver-defined range sizes are relatively
> + * small, GPU SVM does not support splitting of ranges.
> + *
> + * With no support for range splitting, upon partial unmapping of a range, the
> + * driver is expected to invalidate and destroy the entire range. If the range
> + * has device memory as its backing, the driver is also expected to migrate any
> + * remaining pages back to RAM.
> + */
> +
> +/**
> + * DOC: Examples
> + *
> + * This section provides three examples of how to build the expected driver
> + * components: the GPU page fault handler, the garbage collector, and the
> + * notifier callback.
> + *
> + * The generic code provided does not include logic for complex migration
> + * policies, optimized invalidations, fined grained driver locking, or other
> + * potentially required driver locking (e.g., DMA-resv locks).
> + *
> + * 1) GPU page fault handler
> + *
> + *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
> + *	{
> + *		int err = 0;
> + *
> + *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
> + *
> + *		drm_gpusvm_notifier_lock(gpusvm);
> + *		if (drm_gpusvm_range_pages_valid(range))
> + *			driver_commit_bind(gpusvm, range);
> + *		else
> + *			err = -EAGAIN;
> + *		drm_gpusvm_notifier_unlock(gpusvm);
> + *
> + *		return err;
> + *	}
> + *
> + *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
> + *			     unsigned long gpuva_start, unsigned long gpuva_end)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = {};
> + *		int err;
> + *
> + *		driver_svm_lock();
> + *	retry:
> + *		// Always process UNMAPs first so view of GPU SVM ranges is current
> + *		driver_garbage_collector(gpusvm);
> + *
> + *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
> + *							gpuva_start, gpuva_end,
> + *						        &ctx);
> + *		if (IS_ERR(range)) {
> + *			err = PTR_ERR(range);
> + *			goto unlock;
> + *		}
> + *
> + *		if (driver_migration_policy(range)) {
> + *			devmem = driver_alloc_devmem();
> + *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
> + *							   devmem_allocation,
> + *							   &ctx);
> + *			if (err)	// CPU mappings may have changed
> + *				goto retry;
> + *		}
> + *
> + *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
> + *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
> + *			if (err == -EOPNOTSUPP)
> + *				drm_gpusvm_range_evict(gpusvm, range);
> + *			goto retry;
> + *		} else if (err) {
> + *			goto unlock;
> + *		}
> + *
> + *		err = driver_bind_range(gpusvm, range);
> + *		if (err == -EAGAIN)	// CPU mappings changed
> + *			goto retry
> + *
> + *	unlock:
> + *		driver_svm_unlock();
> + *		return err;
> + *	}
> + *
> + * 2) Garbage Collector.
> + *
> + *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
> + *					struct drm_gpusvm_range *range)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		// Partial unmap, migrate any remaining device memory pages back to RAM
> + *		if (range->flags.partial_unmap)
> + *			drm_gpusvm_range_evict(gpusvm, range);
> + *
> + *		driver_unbind_range(range);
> + *		drm_gpusvm_range_remove(gpusvm, range);
> + *	}
> + *
> + *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		for_each_range_in_garbage_collector(gpusvm, range)
> + *			__driver_garbage_collector(gpusvm, range);
> + *	}
> + *
> + * 3) Notifier callback.
> + *
> + *	void driver_invalidation(struct drm_gpusvm *gpusvm,
> + *				 struct drm_gpusvm_notifier *notifier,
> + *				 const struct mmu_notifier_range *mmu_range)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
> + *		struct drm_gpusvm_range *range = NULL;
> + *
> + *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
> + *
> + *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
> + *					  mmu_range->end) {
> + *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
> + *
> + *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
> + *				continue;
> + *
> + *			drm_gpusvm_range_set_unmapped(range, mmu_range);
> + *			driver_garbage_collector_add(gpusvm, range);
> + *		}
> + *	}
> + */
> +
> +/**
> + * npages_in_range() - Calculate the number of pages in a given range
> + * @start: The start address of the range
> + * @end: The end address of the range
> + *
> + * This macro calculates the number of pages in a given memory range,
> + * specified by the start and end addresses. It divides the difference
> + * between the end and start addresses by the page size (PAGE_SIZE) to
> + * determine the number of pages in the range.
> + *
> + * Return: The number of pages in the specified range.
> + */
> +static unsigned long
> +npages_in_range(unsigned long start, unsigned long end)
> +{
> +	return (end - start) >> PAGE_SHIFT;
> +}
> +
> +/**
> + * struct drm_gpusvm_zdd - GPU SVM zone device data
> + *
> + * @refcount: Reference count for the zdd
> + * @destroy_work: Work structure for asynchronous zdd destruction
> + * @devmem_allocation: device memory allocation
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This structure serves as a generic wrapper installed in
> + * page->zone_device_data. It provides infrastructure for looking up a device
> + * memory allocation upon CPU page fault and asynchronously releasing device
> + * memory once the CPU has no page references. Asynchronous release is useful
> + * because CPU page references can be dropped in IRQ contexts, while releasing
> + * device memory likely requires sleeping locks.
> + */
> +struct drm_gpusvm_zdd {
> +	struct kref refcount;
> +	struct work_struct destroy_work;
> +	struct drm_gpusvm_devmem *devmem_allocation;
> +	void *device_private_page_owner;
> +};
> +
> +/**
> + * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
> + * @w: Pointer to the work_struct
> + *
> + * This function releases device memory, puts GPU SVM range, and frees zdd.
> + */
> +static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(w, struct drm_gpusvm_zdd, destroy_work);
> +	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
> +		zdd->devmem_allocation->ops : NULL;
> +
> +	if (zdd->devmem_allocation && ops->devmem_release)
> +		ops->devmem_release(zdd->devmem_allocation);
> +	kfree(zdd);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This function allocates and initializes a new zdd structure. It sets up the
> + * reference count and initializes the destroy work.
> + *
> + * Returns:
> + * Pointer to the allocated zdd on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_zdd *
> +drm_gpusvm_zdd_alloc(void *device_private_page_owner)
> +{
> +	struct drm_gpusvm_zdd *zdd;
> +
> +	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
> +	if (!zdd)
> +		return NULL;
> +
> +	kref_init(&zdd->refcount);
> +	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
> +	zdd->devmem_allocation = NULL;
> +	zdd->device_private_page_owner = device_private_page_owner;
> +
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function increments the reference count of the provided zdd structure.
> + *
> + * Returns: Pointer to the zdd structure.
> + */
> +static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_get(&zdd->refcount);
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
> + * @ref: Pointer to the reference count structure.
> + *
> + * This function queues the destroy_work of the zdd for asynchronous destruction.
> + */
> +static void drm_gpusvm_zdd_destroy(struct kref *ref)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(ref, struct drm_gpusvm_zdd, refcount);
> +
> +	if (zdd->devmem_allocation)
> +		WRITE_ONCE(zdd->devmem_allocation->detached, true);
> +	schedule_work(&zdd->destroy_work);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_put() - Put a zdd reference.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function decrements the reference count of the provided zdd structure
> + * and schedules its destruction if the count drops to zero.
> + */
> +static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
> +}
> +
> +/**
> + * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
> + * @notifier: Pointer to the GPU SVM notifier structure.
> + * @start: Start address of the range
> + * @end: End address of the range
> + *
> + * Return: A pointer to the drm_gpusvm_range if found or NULL
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_range, itree);
> +	else
> +		return NULL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
> +
> +/**
> + * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges
> + * @next__: Iterator variable for the ranges temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier while
> + * removing ranges from it.
> + */
> +#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
> +	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
> +	     (next__) = __drm_gpusvm_range_next(range__);				\
> +	     (range__) && (range__->itree.start < (end__));				\
> +	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
> +
> +/**
> + * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
> + * @notifier: a pointer to the current drm_gpusvm_notifier
> + *
> + * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
> + *         the current notifier is the last one or if the input notifier is
> + *         NULL.
> + */
> +static struct drm_gpusvm_notifier *
> +__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
> +{
> +	if (notifier && !list_is_last(&notifier->entry,
> +				      &notifier->gpusvm->notifier_list))
> +		return list_next_entry(notifier, entry);
> +
> +	return NULL;
> +}
> +
> +static struct drm_gpusvm_notifier *
> +notifier_iter_first(struct rb_root_cached *root, unsigned long start,
> +		    unsigned long last)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(root, start, last);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_notifier, itree);
> +	else
> +		return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
> + */
> +#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @next__: Iterator variable for the notifiers temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
> + * removing notifiers from it.
> + */
> +#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
> +	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
> + * @mni: Pointer to the mmu_interval_notifier structure.
> + * @mmu_range: Pointer to the mmu_notifier_range structure.
> + * @cur_seq: Current sequence number.
> + *
> + * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
> + * notifier sequence number and calls the driver invalidate vfunc under
> + * gpusvm->notifier_lock.
> + *
> + * Returns:
> + * true if the operation succeeds, false otherwise.
> + */
> +static bool
> +drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
> +			       const struct mmu_notifier_range *mmu_range,
> +			       unsigned long cur_seq)
> +{
> +	struct drm_gpusvm_notifier *notifier =
> +		container_of(mni, typeof(*notifier), notifier);
> +	struct drm_gpusvm *gpusvm = notifier->gpusvm;
> +
> +	if (!mmu_notifier_range_blockable(mmu_range))
> +		return false;
> +
> +	down_write(&gpusvm->notifier_lock);
> +	mmu_interval_set_seq(mni, cur_seq);
> +	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
> +	up_write(&gpusvm->notifier_lock);
> +
> +	return true;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
> + */
> +static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
> +	.invalidate = drm_gpusvm_notifier_invalidate,
> +};
> +
> +/**
> + * drm_gpusvm_init() - Initialize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @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.
> + * @ops: Pointer to the operations structure for GPU SVM.
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order with last
> + *               entry being SZ_4K.
> + * @num_chunks: Number of chunks.
> + *
> + * This function initializes the GPU SVM.
> + *
> + * Returns:
> + * 0 on success, a negative error code on failure.
> + */
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks)
> +{
> +	if (!ops->invalidate || !num_chunks)
> +		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;
> +	gpusvm->ops = ops;
> +	gpusvm->chunk_sizes = chunk_sizes;
> +	gpusvm->num_chunks = num_chunks;
> +
> +	mmgrab(mm);
> +	gpusvm->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&gpusvm->notifier_list);
> +
> +	init_rwsem(&gpusvm->notifier_lock);
> +
> +	fs_reclaim_acquire(GFP_KERNEL);
> +	might_lock(&gpusvm->notifier_lock);
> +	fs_reclaim_release(GFP_KERNEL);
> +
> +#ifdef CONFIG_LOCKDEP
> +	gpusvm->lock_dep_map = NULL;
> +#endif
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_init);
> +
> +/**
> + * drm_gpusvm_notifier_find() - Find GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function finds the GPU SVM notifier associated with the fault address.
> + *
> + * Returns:
> + * Pointer to the GPU SVM notifier on success, NULL otherwise.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
> +			 unsigned long fault_addr)
> +{
> +	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
> +}
> +
> +/**
> + * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_notifier structure.
> + */
> +static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	interval_tree_insert(&notifier->itree, &gpusvm->root);
> +
> +	node = rb_prev(&notifier->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_notifier(node))->entry;
> +	else
> +		head = &gpusvm->notifier_list;
> +
> +	list_add(&notifier->entry, head);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM tructure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	interval_tree_remove(&notifier->itree, &gpusvm->root);
> +	list_del(&notifier->entry);
> +}
> +
> +/**
> + * drm_gpusvm_fini() - Finalize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + *
> + * This function finalizes the GPU SVM by cleaning up any remaining ranges and
> + * notifiers, and dropping a reference to struct MM.
> + */
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
> +{
> +	struct drm_gpusvm_notifier *notifier, *next;
> +
> +	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
> +		struct drm_gpusvm_range *range, *__next;
> +
> +		/*
> +		 * Remove notifier first to avoid racing with any invalidation
> +		 */
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +		notifier->flags.removed = true;
> +
> +		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
> +					       LONG_MAX)
> +			drm_gpusvm_range_remove(gpusvm, range);
> +	}
> +
> +	mmdrop(gpusvm->mm);
> +	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
> +
> +/**
> + * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function allocates and initializes the GPU SVM notifier structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	if (gpusvm->ops->notifier_alloc)
> +		notifier = gpusvm->ops->notifier_alloc();
> +	else
> +		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
> +
> +	if (!notifier)
> +		return ERR_PTR(-ENOMEM);
> +
> +	notifier->gpusvm = gpusvm;
> +	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
> +	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
> +	INIT_LIST_HEAD(&notifier->entry);
> +	notifier->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&notifier->range_list);
> +
> +	return notifier;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_free() - Free GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function frees the GPU SVM notifier structure.
> + */
> +static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
> +				     struct drm_gpusvm_notifier *notifier)
> +{
> +	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
> +
> +	if (gpusvm->ops->notifier_free)
> +		gpusvm->ops->notifier_free(notifier);
> +	else
> +		kfree(notifier);
> +}
> +
> +/**
> + * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_range structure.
> + */
> +static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_range, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_range_insert() - Insert GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function inserts the GPU SVM range into the notifier RB tree and list.
> + */
> +static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
> +				    struct drm_gpusvm_range *range)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	drm_gpusvm_notifier_lock(notifier->gpusvm);
> +	interval_tree_insert(&range->itree, &notifier->root);
> +
> +	node = rb_prev(&range->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_range(node))->entry;
> +	else
> +		head = &notifier->range_list;
> +
> +	list_add(&range->entry, head);
> +	drm_gpusvm_notifier_unlock(notifier->gpusvm);
> +}
> +
> +/**
> + * __drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This macro removes the GPU SVM range from the notifier RB tree and list.
> + */
> +static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
> +				      struct drm_gpusvm_range *range)
> +{
> +	interval_tree_remove(&range->itree, &notifier->root);
> +	list_del(&range->entry);
> +}
> +
> +/**
> + * drm_gpusvm_range_alloc() - Allocate GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @fault_addr: Fault address
> + * @chunk_size: Chunk size
> + * @migrate_devmem: Flag indicating whether to migrate device memory
> + *
> + * This function allocates and initializes the GPU SVM range structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_range *
> +drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
> +		       struct drm_gpusvm_notifier *notifier,
> +		       unsigned long fault_addr, unsigned long chunk_size,
> +		       bool migrate_devmem)
> +{
> +	struct drm_gpusvm_range *range;
> +
> +	if (gpusvm->ops->range_alloc)
> +		range = gpusvm->ops->range_alloc(gpusvm);
> +	else
> +		range = kzalloc(sizeof(*range), GFP_KERNEL);
> +
> +	if (!range)
> +		return ERR_PTR(-ENOMEM);
> +
> +	kref_init(&range->refcount);
> +	range->gpusvm = gpusvm;
> +	range->notifier = notifier;
> +	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
> +	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
> +	INIT_LIST_HEAD(&range->entry);
> +	range->notifier_seq = LONG_MAX;
> +	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
> +
> +	return range;
> +}
> +
> +/**
> + * drm_gpusvm_check_pages() - Check pages
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @start: Start address
> + * @end: End address
> + *
> + * Check if pages between start and end have been faulted in on the CPU. Use to
> + * prevent migration of pages without CPU backing store.
> + *
> + * Returns:
> + * True if pages have been faulted into CPU, False otherwise
> + */
> +static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
> +				   struct drm_gpusvm_notifier *notifier,
> +				   unsigned long start, unsigned long end)
> +{
> +	struct hmm_range hmm_range = {
> +		.default_flags = 0,
> +		.notifier = &notifier->notifier,
> +		.start = start,
> +		.end = end,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long *pfns;
> +	unsigned long npages = npages_in_range(start, end);
> +	int err, i;
> +
> +	mmap_assert_locked(gpusvm->mm);
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return false;
> +
> +	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +	hmm_range.hmm_pfns = pfns;
> +
> +	while (true) {
> +		err = hmm_range_fault(&hmm_range);
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	if (err)
> +		goto err_free;
> +
> +	for (i = 0; i < npages;) {
> +		if (!(pfns[i] & HMM_PFN_VALID)) {
> +			err = -EFAULT;
> +			goto err_free;
> +		}
> +		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
> +	}
> +
> +err_free:
> +	kvfree(pfns);
> +	return err ? false : true;
> +}
> +
> +/**
> + * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @vas: Pointer to the virtual memory area structure
Hi Matthew Brost,

mm core mainly uses vma as the naming of the structure vm_area_struct 
variable.
can you consider changing vas to vma for the same recognition as mm core?
( vas naming is used in this function, as well as many of the functions 
introduced in this patch. )
> + * @fault_addr: Fault address
> + * @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
> + *
> + * 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
> + * memory area boundaries.
> + *
> + * Returns:
> + * Chunk size on success, LONG_MAX on failure.
> + */
> +static unsigned long
> +drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_notifier *notifier,
> +			    struct vm_area_struct *vas,
> +			    unsigned long fault_addr,
> +			    unsigned long gpuva_start,
> +			    unsigned long gpuva_end,
> +			    unsigned long check_pages_threshold)
> +{
> +	unsigned long start, end;
> +	int i = 0;
> +
> +retry:
> +	for (; i < gpusvm->num_chunks; ++i) {
> +		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
> +		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
> +
> +		if (start >= vas->vm_start && end <= vas->vm_end &&
> +		    start >= notifier->itree.start &&
> +		    end <= notifier->itree.last + 1 &&
> +		    start >= gpuva_start && end <= gpuva_end)
> +			break;
> +	}
> +
> +	if (i == gpusvm->num_chunks)
> +		return LONG_MAX;
> +
> +	/*
> +	 * If allocation more than page, ensure not to overlap with existing
> +	 * ranges.
> +	 */
> +	if (end - start != SZ_4K) {
> +		struct drm_gpusvm_range *range;
> +
> +		range = drm_gpusvm_range_find(notifier, start, end);
> +		if (range) {
> +			++i;
> +			goto retry;
> +		}
> +
> +		/*
> +		 * XXX: Only create range on pages CPU has faulted in. Without
> +		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
> +		 * process-many-malloc' fails. In the failure case, each process
> +		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
> +		 * ranges. When migrating the SVM ranges, some processes fail in
> +		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
> +		 * and then upon drm_gpusvm_range_get_pages device pages from
> +		 * other processes are collected + faulted in which creates all
> +		 * sorts of problems. Unsure exactly how this happening, also
> +		 * problem goes away if 'xe_exec_system_allocator --r
> +		 * process-many-malloc' mallocs at least 64k at a time.
> +		 */
> +		if (end - start <= check_pages_threshold &&
> +		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
> +			++i;
> +			goto retry;
> +		}
> +	}
> +
> +	return end - start;
> +}
> +
> +/**
> + * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + * @gpuva_start: Start address of GPUVA which mirrors CPU
> + * @gpuva_end: End address of GPUVA which mirrors CPU
> + * @ctx: GPU SVM context
> + *
> + * This function finds or inserts a newly allocated a GPU SVM range based on the
> + * fault address. Caller must hold a lock to protect range lookup and insertion.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +	struct drm_gpusvm_range *range;
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	bool notifier_alloc = false;
> +	unsigned long chunk_size;
> +	int err;
> +	bool migrate_devmem;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	if (fault_addr < gpusvm->mm_start ||
> +	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
> +		return ERR_PTR(-EINVAL);
> +
> +	if (!mmget_not_zero(mm))
> +		return ERR_PTR(-EFAULT);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
> +	if (!notifier) {
> +		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
> +		if (IS_ERR(notifier)) {
> +			err = PTR_ERR(notifier);
> +			goto err_mmunlock;
> +		}
> +		notifier_alloc = true;
> +		err = mmu_interval_notifier_insert(&notifier->notifier,
> +						   mm, notifier->itree.start,
> +						   notifier->itree.last -
> +						   notifier->itree.start + 1,
> +						   &drm_gpusvm_notifier_ops);
> +		if (err)
> +			goto err_notifier;
> +	}
> +
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, fault_addr);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_notifier_remove;
> +	}
> +
> +	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
> +		err = -EPERM;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
> +	if (range)
> +		goto out_mmunlock;
> +	/*
> +	 * XXX: Short-circuiting migration based on migrate_vma_* current
> +	 * limitations. If/when migrate_vma_* add more support, this logic will
> +	 * have to change.
> +	 */
> +	migrate_devmem = ctx->devmem_possible &&
> +		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
> +
> +	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
> +						 fault_addr, gpuva_start,
> +						 gpuva_end, migrate_devmem &&
> +						 ctx->check_pages_threshold);
> +	if (chunk_size == LONG_MAX) {
> +		err = -EINVAL;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
> +				       migrate_devmem);
> +	if (IS_ERR(range)) {
> +		err = PTR_ERR(range);
> +		goto err_notifier_remove;
> +	}
> +
> +	drm_gpusvm_range_insert(notifier, range);
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_insert(gpusvm, notifier);
> +
> +out_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +
> +	return range;
> +
> +err_notifier_remove:
> +	mmap_read_unlock(mm);
> +	if (notifier_alloc)
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +err_notifier:
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +err_mmunlock:
> +	mmput(mm);
> +	return ERR_PTR(err);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
> +
> +/**
> + * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @npages: Number of pages to unmap
> + *
> + * This function unmap pages associated with a GPU SVM range. Assumes and
> + * asserts correct locking is in place when called.
> + */
> +static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +					   struct drm_gpusvm_range *range,
> +					   unsigned long npages)
> +{
> +	unsigned long i, j;
> +	struct drm_pagemap *dpagemap = range->dpagemap;
> +	struct device *dev = gpusvm->drm->dev;
> +
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->flags.has_dma_mapping) {
> +		for (i = 0, j = 0; i < npages; j++) {
> +			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
> +
> +			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
> +				dma_unmap_page(dev,
> +					       addr->addr,
> +					       PAGE_SIZE << addr->order,
> +					       addr->dir);
> +			else if (dpagemap && dpagemap->ops->unmap_dma)
> +				dpagemap->ops->unmap_dma(dpagemap,
> +							 dev,
> +							 *addr);
> +			i += 1 << addr->order;
> +		}
> +		range->flags.has_devmem_pages = false;
> +		range->flags.has_dma_mapping = false;
> +		range->dpagemap = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function frees the dma address array associated with a GPU SVM range.
> + */
> +static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
> +					struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->dma_addr) {
> +		kvfree(range->dma_addr);
> +		range->dma_addr = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function removes the specified GPU SVM range and also removes the parent
> + * GPU SVM notifier if no more ranges remain in the notifier. The caller must
> + * hold a lock to protect range and notifier removal.
> + */
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
> +	if (WARN_ON_ONCE(!notifier))
> +		return;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +	drm_gpusvm_range_free_pages(gpusvm, range);
> +	__drm_gpusvm_range_remove(notifier, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	drm_gpusvm_range_put(range);
> +
> +	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
> +		if (!notifier->flags.removed)
> +			mmu_interval_notifier_remove(&notifier->notifier);
> +		drm_gpusvm_notifier_remove(gpusvm, notifier);
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +	}
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
> +
> +/**
> + * drm_gpusvm_range_get() - Get a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function increments the reference count of the specified GPU SVM range.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range)
> +{
> +	kref_get(&range->refcount);
> +
> +	return range;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
> +
> +/**
> + * drm_gpusvm_range_destroy() - Destroy GPU SVM range
> + * @refcount: Pointer to the reference counter embedded in the GPU SVM range
> + *
> + * This function destroys the specified GPU SVM range when its reference count
> + * reaches zero. If a custom range-free function is provided, it is invoked to
> + * free the range; otherwise, the range is deallocated using kfree().
> + */
> +static void drm_gpusvm_range_destroy(struct kref *refcount)
> +{
> +	struct drm_gpusvm_range *range =
> +		container_of(refcount, struct drm_gpusvm_range, refcount);
> +	struct drm_gpusvm *gpusvm = range->gpusvm;
> +
> +	if (gpusvm->ops->range_free)
> +		gpusvm->ops->range_free(range);
> +	else
> +		kfree(range);
> +}
> +
> +/**
> + * drm_gpusvm_range_put() - Put a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function decrements the reference count of the specified GPU SVM range
> + * and frees it when the count reaches zero.
> + */
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
> +{
> +	kref_put(&range->refcount, drm_gpusvm_range_destroy);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
> +
> +/**
> + * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called holding gpusvm->notifier_lock and as the last step before committing a
> + * GPU binding. This is akin to a notifier seqno check in the HMM documentation
> + * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
> + * function is required for finer grained checking (i.e., per range) if pages
> + * are valid.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
> +
> +/**
> + * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called without holding gpusvm->notifier_lock.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +static bool
> +drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
> +				      struct drm_gpusvm_range *range)
> +{
> +	bool pages_valid;
> +
> +	if (!range->dma_addr)
> +		return false;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
> +	if (!pages_valid)
> +		drm_gpusvm_range_free_pages(gpusvm, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	return pages_valid;
> +}
> +
> +/**
> + * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function gets pages for a GPU SVM range and ensures they are mapped for
> + * DMA access.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
> +	struct hmm_range hmm_range = {
> +		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
> +			HMM_PFN_REQ_WRITE),
> +		.notifier = notifier,
> +		.start = range->itree.start,
> +		.end = range->itree.last + 1,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct drm_gpusvm_zdd *zdd;
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long i, j;
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	unsigned long num_dma_mapped;
> +	unsigned int order = 0;
> +	unsigned long *pfns;
> +	struct page **pages;
> +	int err = 0;
> +	struct dev_pagemap *pagemap;
> +	struct drm_pagemap *dpagemap;
> +
> +retry:
> +	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
> +		goto set_seqno;
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return -ENOMEM;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	hmm_range.hmm_pfns = pfns;
> +	while (true) {
> +		mmap_read_lock(mm);
> +		err = hmm_range_fault(&hmm_range);
> +		mmap_read_unlock(mm);
> +
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	mmput(mm);
> +	if (err)
> +		goto err_free;
> +
> +	pages = (struct page **)pfns;
> +map_pages:
> +	/*
> +	 * Perform all dma mappings under the notifier lock to not
> +	 * access freed pages. A notifier will either block on
> +	 * the notifier lock or unmap dma.
> +	 */
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		kvfree(pfns);
> +		goto retry;
> +	}
> +
> +	if (!range->dma_addr) {
> +		/* Unlock and restart mapping to allocate memory. */
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
> +						 GFP_KERNEL);
> +		if (!range->dma_addr) {
> +			err = -ENOMEM;
> +			goto err_free;
> +		}
> +		goto map_pages;
> +	}
> +
> +	zdd = NULL;
> +	num_dma_mapped = 0;
> +	for (i = 0, j = 0; i < npages; ++j) {
> +		struct page *page = hmm_pfn_to_page(pfns[i]);
> +
> +		order = hmm_pfn_to_map_order(pfns[i]);
> +		if (is_device_private_page(page) || is_device_coherent_page(page)) {
> +			if (zdd != page->zone_device_data && i > 0) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +			zdd = page->zone_device_data;
> +			if (pagemap != page->pgmap) {
> +				if (i > 0) {
> +					err = -EOPNOTSUPP;
> +					goto err_unmap;
> +				}
> +
> +				pagemap = page->pgmap;
> +				dpagemap = zdd->devmem_allocation->dpagemap;
> +				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
> +					/*
> +					 * Raced. This is not supposed to happen
> +					 * since hmm_range_fault() should've migrated
> +					 * this page to system.
> +					 */
> +					err = -EAGAIN;
> +					goto err_unmap;
> +				}
> +			}
> +			range->dma_addr[j] =
> +				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
> +						       page, order,
> +						       DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			pages[i] = page;
> +		} else {
> +			dma_addr_t addr;
> +
> +			if (is_zone_device_page(page) || zdd) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +
> +			addr = dma_map_page(gpusvm->drm->dev,
> +					    page, 0,
> +					    PAGE_SIZE << order,
> +					    DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			range->dma_addr[j] = drm_pagemap_dma_addr_encode
> +				(addr, DRM_INTERCONNECT_SYSTEM, order,
> +				 DMA_BIDIRECTIONAL);
> +		}
> +		i += 1 << order;
> +		num_dma_mapped = i;
> +	}
> +
> +	range->flags.has_dma_mapping = true;
> +	if (zdd) {
> +		range->flags.has_devmem_pages = true;
> +		range->dpagemap = dpagemap;
> +	}
> +
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +	kvfree(pfns);
> +set_seqno:
> +	range->notifier_seq = hmm_range.notifier_seq;
> +
> +	return 0;
> +
> +err_unmap:
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +err_free:
> +	kvfree(pfns);
> +	if (err == -EAGAIN)
> +		goto retry;
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
> +
> +/**
> + * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function unmaps pages associated with a GPU SVM range. If @in_notifier
> + * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
> + * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
> + * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
> + * security model.
> + */
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +
> +	if (ctx->in_notifier)
> +		lockdep_assert_held_write(&gpusvm->notifier_lock);
> +	else
> +		drm_gpusvm_notifier_lock(gpusvm);
> +
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +
> +	if (!ctx->in_notifier)
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_page() - Put a migration page
> + * @page: Pointer to the page to put
> + *
> + * This function unlocks and puts a page.
> + */
> +static void drm_gpusvm_migration_unlock_put_page(struct page *page)
> +{
> +	unlock_page(page);
> +	put_page(page);
> +}
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
> + * @npages: Number of pages
> + * @migrate_pfn: Array of migrate page frame numbers
> + *
> + * This function unlocks and puts an array of pages.
> + */
> +static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
> +						  unsigned long *migrate_pfn)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page;
> +
> +		if (!migrate_pfn[i])
> +			continue;
> +
> +		page = migrate_pfn_to_page(migrate_pfn[i]);
> +		drm_gpusvm_migration_unlock_put_page(page);
> +		migrate_pfn[i] = 0;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
> + * @page: Pointer to the page
> + * @zdd: Pointer to the GPU SVM zone device data
> + *
> + * This function associates the given page with the specified GPU SVM zone
> + * device data and initializes it for zone device usage.
> + */
> +static void drm_gpusvm_get_devmem_page(struct page *page,
> +				     struct drm_gpusvm_zdd *zdd)
> +{
> +	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
> +	zone_device_page_init(page);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
> + * @dev: The device for which the pages are being mapped
> + * @dma_addr: Array to store DMA addresses corresponding to mapped pages
> + * @migrate_pfn: Array of migrate page frame numbers to map
> + * @npages: Number of pages to map
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function maps pages of memory for migration usage in GPU SVM. It
> + * iterates over each page frame number provided in @migrate_pfn, maps the
> + * corresponding page, and stores the DMA address in the provided @dma_addr
> + * array.
> + *
> + * Return: 0 on success, -EFAULT if an error occurs during mapping.
> + */
> +static int drm_gpusvm_migrate_map_pages(struct device *dev,
> +					dma_addr_t *dma_addr,
> +					unsigned long *migrate_pfn,
> +					unsigned long npages,
> +					enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		if (WARN_ON_ONCE(is_zone_device_page(page)))
> +			return -EFAULT;
> +
> +		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
> +		if (dma_mapping_error(dev, dma_addr[i]))
> +			return -EFAULT;
> +	}
> +
> +	return 0;
> +}
> +
> +/**
> + * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
> + * @dev: The device for which the pages were mapped
> + * @dma_addr: Array of DMA addresses corresponding to mapped pages
> + * @npages: Number of pages to unmap
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function unmaps previously mapped pages of memory for GPU Shared Virtual
> + * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
> + * if it's valid and not already unmapped, and unmaps the corresponding page.
> + */
> +static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
> +					   dma_addr_t *dma_addr,
> +					   unsigned long npages,
> +					   enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
> +			continue;
> +
> +		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @devmem_allocation: Pointer to the device memory allocation. The caller
> + *                     should hold a reference to the device memory allocation,
> + *                     which should be dropped via ops->devmem_release or upon
> + *                     the failure of this function.
> + * @ctx: GPU SVM context
> + *
> + * This function migrates the specified GPU SVM range to device memory. It performs the
> + * necessary setup and invokes the driver-specific operations for migration to
> + * device memory. Upon successful return, @devmem_allocation can safely reference @range
> + * until ops->devmem_release is called which only upon successful return.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +	struct migrate_vma migrate = {
> +		.start		= start,
> +		.end		= end,
> +		.pgmap_owner	= gpusvm->device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	unsigned long i, npages = npages_in_range(start, end);
> +	struct vm_area_struct *vas;
> +	struct drm_gpusvm_zdd *zdd = NULL;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int err;
> +
> +	if (!range->flags.migrate_devmem)
> +		return -EINVAL;
> +
> +	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
> +	    !ops->copy_to_ram)
> +		return -EOPNOTSUPP;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_out;
> +	}
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, start);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_mmunlock;
> +	}
> +
> +	if (end > vas->vm_end || start < vas->vm_start) {
> +		err = -EINVAL;
> +		goto err_mmunlock;
> +	}
> +
> +	if (!vma_is_anonymous(vas)) {
> +		err = -EBUSY;
> +		goto err_mmunlock;
> +	}
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_mmunlock;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
> +	if (!zdd) {
> +		err = -ENOMEM;
> +		goto err_free;
> +	}
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	if (!migrate.cpages) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	if (migrate.cpages != npages) {
> +		err = -EBUSY;
> +		goto err_finalize;
> +	}
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   migrate.src, npages, DMA_TO_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = pfn_to_page(migrate.dst[i]);
> +
> +		pages[i] = page;
> +		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
> +		drm_gpusvm_get_devmem_page(page, zdd);
> +	}
> +
> +	err = ops->copy_to_devmem(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +	/* Upon success bind devmem allocation to range and zdd */
> +	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_TO_DEVICE);
> +err_free:
> +	if (zdd)
> +		drm_gpusvm_zdd_put(zdd);
> +	kvfree(buf);
> +err_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +err_out:
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
> +
> +/**
> + * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
> + * @vas: Pointer to the VM area structure, can be NULL
> + * @fault_page: Fault page
> + * @npages: Number of pages to populate
> + * @mpages: Number of pages to migrate
> + * @src_mpfn: Source array of migrate PFNs
> + * @mpfn: Array of migrate PFNs to populate
> + * @addr: Start address for PFN allocation
> + *
> + * This function populates the RAM migrate page frame numbers (PFNs) for the
> + * specified VM area structure. It allocates and locks pages in the VM area for
> + * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
> + * alloc_page for allocation.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> +					       struct page *fault_page,
> +					       unsigned long npages,
> +					       unsigned long *mpages,
> +					       unsigned long *src_mpfn,
> +					       unsigned long *mpfn,
> +					       unsigned long addr)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
> +		struct page *page, *src_page;
> +
> +		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
> +			continue;
> +
> +		if (fault_page) {
> +			src_page = migrate_pfn_to_page(src_mpfn[i]);
> +			if (src_page->zone_device_data !=
> +			    fault_page->zone_device_data)
> +				continue;
> +		}
> +
> +		if (vas)
> +			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
> +		else
> +			page = alloc_page(GFP_HIGHUSER);
> +
> +		if (!page)
> +			goto free_pages;
> +
> +		mpfn[i] = migrate_pfn(page_to_pfn(page));
> +	}
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		WARN_ON_ONCE(!trylock_page(page));
> +		++*mpages;
> +	}
> +
> +	return 0;
> +
> +free_pages:
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		put_page(page);
> +		mpfn[i] = 0;
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
> + * @devmem_allocation: Pointer to the device memory allocation
> + *
> + * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
> + * migration done via migrate_device_* functions.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	unsigned long *src, *dst;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int i, err = 0;
> +
> +	npages = devmem_allocation->size >> PAGE_SHIFT;
> +
> +retry:
> +	if (!mmget_not_zero(devmem_allocation->mm))
> +		return -EFAULT;
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	src = buf;
> +	dst = buf + (sizeof(*src) * npages);
> +	dma_addr = buf + (2 * sizeof(*src) * npages);
> +	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
> +	if (err)
> +		goto err_free;
> +
> +	err = migrate_device_pfns(src, npages);
> +	if (err)
> +		goto err_free;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
> +						  src, dst, 0);
> +	if (err || !mpages)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   dst, npages, DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, dst);
> +	migrate_device_pages(src, dst, npages);
> +	migrate_device_finalize(src, dst, npages);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +	mmput_async(devmem_allocation->mm);
> +	if (!err && !READ_ONCE(devmem_allocation->detached)) {
> +		cond_resched();
> +		goto retry;
> +	}
> +
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
> +
> +/**
> + * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
> + * @vas: Pointer to the VM area structure
> + * @device_private_page_owner: Device private pages owner
> + * @page: Pointer to the page for fault handling (can be NULL)
> + * @fault_addr: Fault address
> + * @size: Size of migration
> + *
> + * This internal function performs the migration of the specified GPU SVM range
> + * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
> + * invokes the driver-specific operations for migration to RAM.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
> +				       void *device_private_page_owner,
> +				       struct page *page,
> +				       unsigned long fault_addr,
> +				       unsigned long size)
> +{
> +	struct migrate_vma migrate = {
> +		.vma		= vas,
> +		.pgmap_owner	= device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
> +			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
> +		.fault_page	= page,
> +	};
> +	struct drm_gpusvm_zdd *zdd;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct device *dev;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	unsigned long start, end;
> +	void *buf;
> +	int i, err = 0;
> +
> +	start = ALIGN_DOWN(fault_addr, size);
> +	end = ALIGN(fault_addr + 1, size);
> +
> +	/* Corner where VMA area struct has been partially unmapped */
> +	if (start < vas->vm_start)
> +		start = vas->vm_start;
> +	if (end > vas->vm_end)
> +		end = vas->vm_end;
> +
> +	migrate.start = start;
> +	migrate.end = end;
> +	npages = npages_in_range(start, end);
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	/* Raced with another CPU fault, nothing to do */
> +	if (!migrate.cpages)
> +		goto err_free;
> +
> +	if (!page) {
> +		for (i = 0; i < npages; ++i) {
> +			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
> +				continue;
> +
> +			page = migrate_pfn_to_page(migrate.src[i]);
> +			break;
> +		}
> +
> +		if (!page)
> +			goto err_finalize;
If you goto to err_finalize through this error case, the uninitialized 
dev will be passed to drm_gpusvm_migrate_unmap_pages().

Br,
G.G.
> +	}
> +	zdd = page->zone_device_data;
> +	ops = zdd->devmem_allocation->ops;
> +	dev = zdd->devmem_allocation->dev;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
> +						  migrate.src, migrate.dst,
> +						  start);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
> +					   DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(migrate.src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +
> +	return err;
> +}
> +
> +/**
> + * drm_gpusvm_range_evict() - Evict GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function evicts the specified GPU SVM range.
> + */
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range)
> +{
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +
> +	if (!mmget_not_zero(mm))
> +		return;
> +
> +	mmap_read_lock(mm);
> +
> +	/*
> +	 * Loop required to find all VMAs for the corner case when partial
> +	 * unmappings have occurred.
> +	 */
> +again:
> +	vas = find_vma(mm, start);
> +	if (!vas)
> +		goto unlock;
> +
> +	if (end <= vas->vm_start || start >= vas->vm_end)
> +		goto unlock;
> +
> +	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
> +				    NULL, start,
> +				    range->itree.last - range->itree.start + 1);
> +	if (vas->vm_end < end) {
> +		start = vas->vm_end;
> +		goto again;
> +	}
> +unlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
> +
> +/**
> + * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
> + * @page: Pointer to the page
> + *
> + * This function is a callback used to put the GPU SVM zone device data
> + * associated with a page when it is being released.
> + */
> +static void drm_gpusvm_page_free(struct page *page)
> +{
> +	drm_gpusvm_zdd_put(page->zone_device_data);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
> + * @vmf: Pointer to the fault information structure
> + *
> + * This function is a page fault handler used to migrate a GPU SVM range to RAM.
> + * It retrieves the GPU SVM range information from the faulting page and invokes
> + * the internal migration function to migrate the range back to RAM.
> + *
> + * Returns:
> + * VM_FAULT_SIGBUS on failure, 0 on success.
> + */
> +static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
> +{
> +	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
> +	int err;
> +
> +	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
> +					  zdd->device_private_page_owner,
> +					  vmf->page, vmf->address,
> +					  zdd->devmem_allocation->size);
> +
> +	return err ? VM_FAULT_SIGBUS : 0;
> +}
> +
> +/**
> + * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
> + */
> +static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
> +	.page_free = drm_gpusvm_page_free,
> +	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
> +};
> +
> +/**
> + * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
> + *
> + * Returns:
> + * Pointer to the GPU SVM device page map operations structure.
> + */
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
> +{
> +	return &drm_gpusvm_pagemap_ops;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
> +
> +/**
> + * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @start: Start address
> + * @end: End address
> + *
> + * Returns:
> + * True if GPU SVM has mapping, False otherwise
> + */
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
> +		struct drm_gpusvm_range *range = NULL;
> +
> +		drm_gpusvm_for_each_range(range, notifier, start, end)
> +			return true;
> +	}
> +
> +	return false;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
> +
> +/**
> + * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
> + * @range: Pointer to the GPU SVM range structure.
> + * @mmu_range: Pointer to the MMU notifier range structure.
> + *
> + * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
> + * if the range partially falls within the provided MMU notifier range.
> + */
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range)
> +{
> +	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
> +
> +	range->flags.unmapped = true;
> +	if (range->itree.start < mmu_range->start ||
> +	    range->itree.last + 1 > mmu_range->end)
> +		range->flags.partial_unmap = true;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
> +
> +/**
> + * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap we're allocating from.
> + * @size: Size of device memory allocation
> + */
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size)
> +{
> +	devmem_allocation->dev = dev;
> +	devmem_allocation->mm = mm;
> +	devmem_allocation->ops = ops;
> +	devmem_allocation->dpagemap = dpagemap;
> +	devmem_allocation->size = size;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
> +
> +MODULE_DESCRIPTION("DRM GPUSVM");
> +MODULE_LICENSE("GPL");
> diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> new file mode 100644
> index 000000000000..2fb6980d0cf1
> --- /dev/null
> +++ b/include/drm/drm_gpusvm.h
> @@ -0,0 +1,445 @@
> +/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
> +/*
> + * Copyright © 2024 Intel Corporation
> + */
> +
> +#ifndef __DRM_GPUSVM_H__
> +#define __DRM_GPUSVM_H__
> +
> +#include <linux/kref.h>
> +#include <linux/interval_tree.h>
> +#include <linux/mmu_notifier.h>
> +
> +struct dev_pagemap_ops;
> +struct drm_device;
> +struct drm_gpusvm;
> +struct drm_gpusvm_notifier;
> +struct drm_gpusvm_ops;
> +struct drm_gpusvm_range;
> +struct drm_gpusvm_devmem;
> +struct drm_pagemap;
> +struct drm_pagemap_dma_addr;
> +
> +/**
> + * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM)
> + * device memory. These operations are provided by the GPU driver to manage device memory
> + * allocations and perform operations such as migration between device memory and system
> + * RAM.
> + */
> +struct drm_gpusvm_devmem_ops {
> +	/**
> +	 * @devmem_release: Release device memory allocation (optional)
> +	 * @devmem_allocation: device memory allocation
> +	 *
> +	 * Release device memory allocation and drop a reference to device
> +	 * memory allocation.
> +	 */
> +	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +	/**
> +	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
> +	 * @devmem_allocation: device memory allocation
> +	 * @npages: Number of pages to populate
> +	 * @pfn: Array of page frame numbers to populate
> +	 *
> +	 * Populate device memory page frame numbers (PFN).
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
> +				   unsigned long npages, unsigned long *pfn);
> +
> +	/**
> +	 * @copy_to_devmem: Copy to device memory (required for migration)
> +	 * @pages: Pointer to array of device memory pages (destination)
> +	 * @dma_addr: Pointer to array of DMA addresses (source)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to device memory.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_devmem)(struct page **pages,
> +			      dma_addr_t *dma_addr,
> +			      unsigned long npages);
> +
> +	/**
> +	 * @copy_to_ram: Copy to system RAM (required for migration)
> +	 * @pages: Pointer to array of device memory pages (source)
> +	 * @dma_addr: Pointer to array of DMA addresses (destination)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to system RAM.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_ram)(struct page **pages,
> +			   dma_addr_t *dma_addr,
> +			   unsigned long npages);
> +};
> +
> +/**
> + * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
> + * @size: Size of device memory allocation
> + * @detached: device memory allocations is detached from device pages
> + */
> +struct drm_gpusvm_devmem {
> +	struct device *dev;
> +	struct mm_struct *mm;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct drm_pagemap *dpagemap;
> +	size_t size;
> +	bool detached;
> +};
> +
> +/**
> + * struct drm_gpusvm_ops - Operations structure for GPU SVM
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM).
> + * These operations are provided by the GPU driver to manage SVM ranges and
> + * notifiers.
> + */
> +struct drm_gpusvm_ops {
> +	/**
> +	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
> +	 *
> +	 * Allocate a GPU SVM notifier.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
> +
> +	/**
> +	 * @notifier_free: Free a GPU SVM notifier (optional)
> +	 * @notifier: Pointer to the GPU SVM notifier to be freed
> +	 *
> +	 * Free a GPU SVM notifier.
> +	 */
> +	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
> +
> +	/**
> +	 * @range_alloc: Allocate a GPU SVM range (optional)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 *
> +	 * Allocate a GPU SVM range.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
> +
> +	/**
> +	 * @range_free: Free a GPU SVM range (optional)
> +	 * @range: Pointer to the GPU SVM range to be freed
> +	 *
> +	 * Free a GPU SVM range.
> +	 */
> +	void (*range_free)(struct drm_gpusvm_range *range);
> +
> +	/**
> +	 * @invalidate: Invalidate GPU SVM notifier (required)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 * @notifier: Pointer to the GPU SVM notifier
> +	 * @mmu_range: Pointer to the mmu_notifier_range structure
> +	 *
> +	 * Invalidate the GPU page tables. It can safely walk the notifier range
> +	 * RB tree/list in this function. Called while holding the notifier lock.
> +	 */
> +	void (*invalidate)(struct drm_gpusvm *gpusvm,
> +			   struct drm_gpusvm_notifier *notifier,
> +			   const struct mmu_notifier_range *mmu_range);
> +};
> +
> +/**
> + * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: MMU interval notifier
> + * @itree: Interval tree node for the notifier (inserted in GPU SVM)
> + * @entry: List entry to fast interval tree traversal
> + * @root: Cached root node of the RB tree containing ranges
> + * @range_list: List head containing of ranges in the same order they appear in
> + *              interval tree. This is useful to keep iterating ranges while
> + *              doing modifications to RB tree.
> + * @flags.removed: Flag indicating whether the MMU interval notifier has been
> + *                 removed
> + *
> + * This structure represents a GPU SVM notifier.
> + */
> +struct drm_gpusvm_notifier {
> +	struct drm_gpusvm *gpusvm;
> +	struct mmu_interval_notifier notifier;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	struct rb_root_cached root;
> +	struct list_head range_list;
> +	struct {
> +		u32 removed : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm_range - Structure representing a GPU SVM range
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier
> + * @refcount: Reference count for the range
> + * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
> + * @entry: List entry to fast interval tree traversal
> + * @notifier_seq: Notifier sequence number of the range's pages
> + * @dma_addr: DMA address array
> + * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
> + *            Note this is assuming only one drm_pagemap per range is allowed.
> + * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
> + * @flags.unmapped: Flag indicating if the range has been unmapped
> + * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
> + * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
> + * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
> + *
> + * This structure represents a GPU SVM range used for tracking memory ranges
> + * mapped in a DRM device.
> + */
> +struct drm_gpusvm_range {
> +	struct drm_gpusvm *gpusvm;
> +	struct drm_gpusvm_notifier *notifier;
> +	struct kref refcount;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	unsigned long notifier_seq;
> +	struct drm_pagemap_dma_addr *dma_addr;
> +	struct drm_pagemap *dpagemap;
> +	struct {
> +		/* All flags below must be set upon creation */
> +		u16 migrate_devmem : 1;
> +		/* All flags below must be set / cleared under notifier lock */
> +		u16 unmapped : 1;
> +		u16 partial_unmap : 1;
> +		u16 has_devmem_pages : 1;
> +		u16 has_dma_mapping : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm - GPU SVM structure
> + *
> + * @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
> + * @ops: Pointer to the operations structure for GPU SVM
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order.
> + * @num_chunks: Number of chunks
> + * @notifier_lock: Read-write semaphore for protecting notifier operations
> + * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
> + * @notifier_list: list head containing of notifiers in the same order they
> + *                 appear in interval tree. This is useful to keep iterating
> + *                 notifiers while doing modifications to RB tree.
> + *
> + * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
> + * memory ranges mapped in a DRM (Direct Rendering Manager) device.
> + *
> + * No reference counting is provided, as this is expected to be embedded in the
> + * driver VM structure along with the struct drm_gpuvm, which handles reference
> + * counting.
> + */
> +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;
> +	const struct drm_gpusvm_ops *ops;
> +	const unsigned long *chunk_sizes;
> +	int num_chunks;
> +	struct rw_semaphore notifier_lock;
> +	struct rb_root_cached root;
> +	struct list_head notifier_list;
> +#ifdef CONFIG_LOCKDEP
> +	/**
> +	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
> +	 * drm_gpusvm_range_remove with a driver provided lock.
> +	 */
> +	struct lockdep_map *lock_dep_map;
> +#endif
> +};
> +
> +/**
> + * struct drm_gpusvm_ctx - DRM GPU SVM context
> + *
> + * @check_pages_threshold: Check CPU pages for present if chunk is less than or
> + *                         equal to threshold. If not present, reduce chunk
> + *                         size.
> + * @in_notifier: entering from a MMU notifier
> + * @read_only: operating on read-only memory
> + * @devmem_possible: possible to use device memory
> + *
> + * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> + */
> +struct drm_gpusvm_ctx {
> +	unsigned long check_pages_threshold;
> +	unsigned int in_notifier :1;
> +	unsigned int read_only :1;
> +	unsigned int devmem_possible :1;
> +};
> +
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks);
> +
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
> +
> +void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
> +
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range);
> +
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
> +
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end);
> +
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range);
> +
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size);
> +
> +#ifdef CONFIG_LOCKDEP
> +/**
> + * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @lock: the lock used to protect the gpuva list. The locking primitive
> + * must contain a dep_map field.
> + *
> + * Call this to annotate drm_gpusvm_range_find_or_insert and
> + * drm_gpusvm_range_remove.
> + */
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
> +	do { \
> +		if (!WARN((gpusvm)->lock_dep_map, \
> +			  "GPUSVM range lock should be set only once."))\
> +			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
> +	} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) \
> +	do { \
> +		if ((gpusvm)->lock_dep_map)	\
> +			lock_is_held((gpusvm)->lock_dep_map);	\
> +	} while (0)
> +#else
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
> +#endif
> +
> +/**
> + * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, take lock
> + */
> +#define drm_gpusvm_notifier_lock(gpusvm__)	\
> +	down_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, drop lock
> + */
> +#define drm_gpusvm_notifier_unlock(gpusvm__)	\
> +	up_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
> + * @range: a pointer to the current GPU SVM range
> + *
> + * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
> + *         current range is the last one or if the input range is NULL.
> + */
> +static inline struct drm_gpusvm_range *
> +__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
> +{
> +	if (range && !list_is_last(&range->entry,
> +				   &range->notifier->range_list))
> +		return list_next_entry(range, entry);
> +
> +	return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges. If set, it indicates the start of
> + *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
> + * to use while holding the driver SVM lock or the notifier lock.
> + */
> +#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
> +	for ((range__) = (range__) ?:					\
> +	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
> +	     (range__) && (range__->itree.start < (end__));		\
> +	     (range__) = __drm_gpusvm_range_next(range__))
> +
> +#endif /* __DRM_GPUSVM_H__ */

Re: [PATCH v3 00/30] Introduce GPU SVM and Xe SVM implementation

[Gwan-gyeong Mun]

Hi,
This kernel oops, which I reported before, was caused by my incorrect 
modification (incorrect applying of review comments) of this patch 
"[v3,19/30] drm/xe: Add SVM device memory mirroring"
( the kernel oops occurred because the xe_drm_pagemap_map_dma() and 
xe_devm_add() functions were built in the form of empty functions. )

This issue disappeared after proper patch modifications were applied.
So please ignore the previously reported this kernel oops.

Br,

G.G.

On 1/7/25 2:19 PM, Gwan-gyeong Mun wrote:
> Hi Matthew Brost,
> 
> After applying this patch series and the following to the latest drm- 
> tip, while testing[1] with the mentioned IGT, I faced a kernel oops[3].
> It makes prevent progressing of the mentioned igt tests.
> Could you please check the following oops log?
> 
> (1) apply comments of "[v3,05/30] drm/gpusvm: Add support for GPU Shared 
> Virtual Memory"
> (2) apply comments of "[v3,15/30] drm/xe: Add unbind to SVM garbage 
> collector"
> (3) drop "[v3,27/30] drm/xe: Basic SVM BO eviction" patch
> 
> The kernel config used, the entire dmesg, and detailed information can 
> be found in [2].
> 
> br,
> 
> G.G.
> 
> [1] used igt command: xe_exec_system_allocator --run-subtest once-malloc
> [2] https://gitlab.freedesktop.org/elongbug/drm-tip/-/snippets/7823
> 
> [3] kernel oops dmesg
> [   51.365230] Console: switching to colour VGA+ 80x25
> [   51.367772] [IGT] xe_exec_system_allocator: executing
> [   51.383611] [IGT] xe_exec_system_allocator: starting subtest once-malloc
> [   51.386066] xe 0000:00:04.0: [drm:vm_bind_ioctl_ops_create [xe]] 
> op=0, addr=0x0000000000000000, range=0x0001000000000000, 
> bo_offset_or_userptr=0x0000000000000000
> [   51.386171] xe 0000:00:04.0: [drm:vm_bind_ioctl_ops_create [xe]] MAP: 
> addr=0x0000000000000000, range=0x0001000000000000
> [   51.389429] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] PAGE 
> FAULT: asid=1, gpusvm=0xffff8881775e9188, vram=0,0, 
> seqno=9223372036854775807, start=0x005584e8400000, end=0x005584e8410000, 
> size=65536
> [   51.389529] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] 
> ALLOCATE VRAM: asid=1, gpusvm=0xffff8881775e9188, vram=0,0, 
> seqno=9223372036854775807, start=0x005584e8400000, end=0x005584e8410000, 
> size=65536
> [   51.389935] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] ALLOC 
> VRAM: asid=1, gpusvm=0xffff8881775e9188, pfn=3126960, npages=16
> [   51.390048] xe 0000:00:04.0: [drm:xe_svm_invalidate [xe]] INVALIDATE: 
> asid=1, gpusvm=0xffff8881775e9188, seqno=3, start=0x00005584e8400000, 
> end=0x00005584e8410000, event=6
> [   51.390440] xe 0000:00:04.0: [drm:xe_svm_invalidate [xe]] NOTIFIER: 
> asid=1, gpusvm=0xffff8881775e9188, vram=0,0, seqno=9223372036854775807, 
> start=0x005584e8400000, end=0x005584e8410000, size=65536
> [   51.390948] Oops: general protection fault, probably for non- 
> canonical address 0x3fff88842fc80000: 0000 [#1] PREEMPT SMP NOPTI
> [   51.391624] CPU: 1 UID: 0 PID: 76 Comm: kworker/u17:0 Not tainted 
> 6.13.0-rc4-drm-tip-test+ #48
> [   51.392088] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), 
> BIOS 1.15.0-1 04/01/2014
> [   51.392527] Workqueue: xe_gt_page_fault_work_queue pf_queue_work_func 
> [xe]
> [   51.392947] RIP: 0010:zone_device_page_init+0x5d/0x240
> [   51.393228] Code: 04 dd ff e8 d5 d2 a1 00 5a 85 c0 0f 85 ba 00 00 00 
> e8 d7 bb df ff 85 c0 0f 84 9d 01 00 00 48 8b 45 38 a8 03 0f 85 ec 00 00 
> 00 <65> 48 ff 00 e8 aa d2 a1 00 85 c0 0f 85 0d 01 00 00 48 c7 c7 20 cb
> [   51.394247] RSP: 0018:ffffc9000039fb48 EFLAGS: 00010246
> [   51.394570] RAX: 4000000000000000 RBX: ffffea000bedac00 RCX: 
> 0000000000000000
> [   51.394950] RDX: 0000000000000046 RSI: ffffffff824c67b4 RDI: 
> ffffffff824e58f5
> [   51.395328] RBP: ffffea000bedac08 R08: 0000000000000015 R09: 
> 0000000000000004
> [   51.395709] R10: 0000000000000001 R11: 0000000000000004 R12: 
> 0000000000000001
> [   51.396093] R13: ffff888170fd8d40 R14: ffff88817f922640 R15: 
> ffffea000bedac00
> [   51.396472] FS:  0000000000000000(0000) GS:ffff88842fc80000(0000) 
> knlGS:0000000000000000
> [   51.396925] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   51.397237] CR2: 0000563d1f7ecbe4 CR3: 000000017c212000 CR4: 
> 0000000000750ef0
> [   51.397618] PKRU: 55555554
> [   51.397768] Call Trace:
> [   51.397904]  <TASK>
> [   51.398024]  ? __die_body.cold+0x19/0x26
> [   51.398238]  ? die_addr+0x38/0x60
> [   51.398420]  ? exc_general_protection+0x19e/0x450
> [   51.398678]  ? asm_exc_general_protection+0x22/0x30
> [   51.398942]  ? zone_device_page_init+0x5d/0x240
> [   51.399188]  ? zone_device_page_init+0x49/0x240
> [   51.399433]  drm_gpusvm_migrate_to_devmem+0x379/0x9e0 [drm_gpusvm]
> [   51.399768]  xe_svm_handle_pagefault+0x62c/0xa60 [xe]
> [   51.400110]  ? xe_vm_find_overlapping_vma+0xa4/0x1d0 [xe]
> [   51.400475]  pf_queue_work_func+0x1ba/0x450 [xe]
> [   51.400777]  process_one_work+0x1fe/0x580
> [   51.400996]  worker_thread+0x1d1/0x3b0
> [   51.401201]  ? __pfx_worker_thread+0x10/0x10
> [   51.401433]  kthread+0xeb/0x120
> [   51.401609]  ? __pfx_kthread+0x10/0x10
> [   51.401813]  ret_from_fork+0x2d/0x50
> [   51.402008]  ? __pfx_kthread+0x10/0x10
> [   51.402211]  ret_from_fork_asm+0x1a/0x30
> [   51.402427]  </TASK>
> [   51.402551] Modules linked in: xe drm_ttm_helper gpu_sched 
> drm_suballoc_helper drm_gpuvm drm_exec drm_gpusvm i2c_algo_bit drm_buddy 
> video wmi ttm drm_display_helper drm_kms_helper crct10dif_pclmul 
> crc32_pclmul e1000 ghash_clmulni_intel i2c_piix4 i2c_smbus fuse
> [   51.403779] ---[ end trace 0000000000000000 ]---
> [   51.404106] RIP: 0010:zone_device_page_init+0x5d/0x240
> [   51.404393] Code: 04 dd ff e8 d5 d2 a1 00 5a 85 c0 0f 85 ba 00 00 00 
> e8 d7 bb df ff 85 c0 0f 84 9d 01 00 00 48 8b 45 38 a8 03 0f 85 ec 00 00 
> 00 <65> 48 ff 00 e8 aa d2 a1 00 85 c0 0f 85 0d 01 00 00 48 c7 c7 20 cb
> [   51.405408] RSP: 0018:ffffc9000039fb48 EFLAGS: 00010246
> [   51.405725] RAX: 4000000000000000 RBX: ffffea000bedac00 RCX: 
> 0000000000000000
> [   51.406110] RDX: 0000000000000046 RSI: ffffffff824c67b4 RDI: 
> ffffffff824e58f5
> [   51.406518] RBP: ffffea000bedac08 R08: 0000000000000015 R09: 
> 0000000000000004
> [   51.406905] R10: 0000000000000001 R11: 0000000000000004 R12: 
> 0000000000000001
> [   51.407312] R13: ffff888170fd8d40 R14: ffff88817f922640 R15: 
> ffffea000bedac00
> [   51.407691] FS:  0000000000000000(0000) GS:ffff88842fc80000(0000) 
> knlGS:0000000000000000
> [   51.408135] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> [   51.408484] CR2: 0000563d1f7ecbe4 CR3: 000000017c212000 CR4: 
> 0000000000750ef0
> [   51.408877] PKRU: 55555554
> [   51.409047] BUG: sleeping function called from invalid context at ./ 
> include/linux/percpu-rwsem.h:49
> [   51.409528] in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid: 
> 76, name: kworker/u17:0
> [   51.409976] preempt_count: 0, expected: 0
> [   51.410212] RCU nest depth: 1, expected: 0
> [   51.410435] INFO: lockdep is turned off.
> [   51.410648] CPU: 1 UID: 0 PID: 76 Comm: kworker/u17:0 Tainted: G 
> D            6.13.0-rc4-drm-tip-test+ #48
> [   51.411180] Tainted: [D]=DIE
> [   51.411338] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), 
> BIOS 1.15.0-1 04/01/2014
> [   51.411859] Workqueue: xe_gt_page_fault_work_queue pf_queue_work_func 
> [xe]
> [   51.412269] Call Trace:
> [   51.412404]  <TASK>
> [   51.412525]  dump_stack_lvl+0x69/0xa0
> [   51.412724]  __might_resched.cold+0xe5/0x120
> [   51.412956]  exit_signals+0x1a/0x360
> [   51.413150]  do_exit+0x122/0xbd0
> [   51.413328]  ? __pfx_worker_thread+0x10/0x10
> [   51.413562]  make_task_dead+0x88/0x90
> [   51.413783]  rewind_stack_and_make_dead+0x16/0x20
> [   51.414045] RIP: 0000:0x0
> [   51.414191] Code: Unable to access opcode bytes at 0xffffffffffffffd6.
> [   51.414595] RSP: 0000:0000000000000000 EFLAGS: 00000000 ORIG_RAX: 
> 0000000000000000
> [   51.414993] RAX: 0000000000000000 RBX: 0000000000000000 RCX: 
> 0000000000000000
> [   51.415369] RDX: 0000000000000000 RSI: 0000000000000000 RDI: 
> 0000000000000000
> [   51.415746] RBP: 0000000000000000 R08: 0000000000000000 R09: 
> 0000000000000000
> [   51.416123] R10: 0000000000000000 R11: 0000000000000000 R12: 
> 0000000000000000
> [   51.416501] R13: 0000000000000000 R14: 0000000000000000 R15: 
> 0000000000000000
> [   51.416899]  </TASK>
> 
> 
> On 12/18/24 1:33 AM, Matthew Brost wrote:
>> Version 3 of GPU SVM has been promoted to the proper series from an RFC.
>> Thanks to everyone (especially Sima and Thomas) for their numerous
>> reviews on revision 1, 2 and for helping to address many design issues.
>>
>> This version has been tested with IGT [1] on PVC, BMG, and LNL. Also
>> tested with level0 (UMD) PR [2].
>>
>> Major changes in v2:
>> - Dropped mmap write abuse
>> - core MM locking and retry loops instead of driver locking to avoid 
>> races
>> - Removed physical to virtual references
>> - Embedded structure/ops for drm_gpusvm_devmem
>> - Fixed mremap and fork issues
>> - Added DRM pagemap
>> - Included RFC documentation in the kernel doc
>>
>> Major changes in v3:
>> - Move GPU SVM and DRM pagemap to DRM level
>> - Mostly addresses Thomas's feedback, lots of small changes documented
>>    in each individual patch change log
>>
>> Known issues in v3:
>> - Check pages still exists, changed to threshold in this version which
>>    is better but still need to root cause cross process page finding on
>>    small user allocations.
>> - Dropped documentation patch, fairly large rewrite and will send out
>>    independently once finished.
>>
>> Matt
>>
>> [1] https://patchwork.freedesktop.org/series/137545/#rev3
>> [2] https://github.com/intel/compute-runtime/pull/782
>>
>> Matthew Brost (27):
>>    drm/xe: Retry BO allocation
>>    mm/migrate: Add migrate_device_pfns
>>    mm/migrate: Trylock device page in do_swap_page
>>    drm/gpusvm: Add support for GPU Shared Virtual Memory
>>    drm/xe: Select DRM_GPUSVM Kconfig
>>    drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag
>>    drm/xe: Add SVM init / close / fini to faulting VMs
>>    drm/xe: Nuke VM's mapping upon close
>>    drm/xe: Add SVM range invalidation and page fault handler
>>    drm/gpuvm: Add DRM_GPUVA_OP_DRIVER
>>    drm/xe: Add (re)bind to SVM page fault handler
>>    drm/xe: Add SVM garbage collector
>>    drm/xe: Add unbind to SVM garbage collector
>>    drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has
>>      bindings
>>    drm/xe: Enable CPU address mirror uAPI
>>    drm/xe: Add migrate layer functions for SVM support
>>    drm/xe: Add SVM device memory mirroring
>>    drm/xe: Add drm_gpusvm_devmem to xe_bo
>>    drm/xe: Add GPUSVM device memory copy vfunc functions
>>    drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc
>>    drm/xe: Add Xe SVM devmem_release GPU SVM vfunc
>>    drm/xe: Add BO flags required for SVM
>>    drm/xe: Add SVM VRAM migration
>>    drm/xe: Basic SVM BO eviction
>>    drm/xe: Add SVM debug
>>    drm/xe: Add modparam for SVM notifier size
>>    drm/xe: Add always_migrate_to_vram modparam
>>
>> Thomas Hellström (3):
>>    drm/pagemap: Add DRM pagemap
>>    drm/xe: Add dma_addr res cursor
>>    drm/xe: Add drm_pagemap ops to SVM
>>
>>   drivers/gpu/drm/Kconfig                     |    8 +
>>   drivers/gpu/drm/Makefile                    |    1 +
>>   drivers/gpu/drm/drm_gpusvm.c                | 2220 +++++++++++++++++++
>>   drivers/gpu/drm/xe/Kconfig                  |   10 +
>>   drivers/gpu/drm/xe/Makefile                 |    1 +
>>   drivers/gpu/drm/xe/xe_bo.c                  |   20 +-
>>   drivers/gpu/drm/xe/xe_bo.h                  |    1 +
>>   drivers/gpu/drm/xe/xe_bo_types.h            |    4 +
>>   drivers/gpu/drm/xe/xe_device_types.h        |   15 +
>>   drivers/gpu/drm/xe/xe_gt_pagefault.c        |   17 +-
>>   drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c |   24 +
>>   drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h |    2 +
>>   drivers/gpu/drm/xe/xe_migrate.c             |  175 ++
>>   drivers/gpu/drm/xe/xe_migrate.h             |   10 +
>>   drivers/gpu/drm/xe/xe_module.c              |    7 +
>>   drivers/gpu/drm/xe/xe_module.h              |    2 +
>>   drivers/gpu/drm/xe/xe_pt.c                  |  393 +++-
>>   drivers/gpu/drm/xe/xe_pt.h                  |    5 +
>>   drivers/gpu/drm/xe/xe_pt_types.h            |    2 +
>>   drivers/gpu/drm/xe/xe_res_cursor.h          |  116 +-
>>   drivers/gpu/drm/xe/xe_svm.c                 |  948 ++++++++
>>   drivers/gpu/drm/xe/xe_svm.h                 |   83 +
>>   drivers/gpu/drm/xe/xe_tile.c                |    5 +
>>   drivers/gpu/drm/xe/xe_vm.c                  |  375 +++-
>>   drivers/gpu/drm/xe/xe_vm.h                  |   15 +-
>>   drivers/gpu/drm/xe/xe_vm_types.h            |   57 +
>>   include/drm/drm_gpusvm.h                    |  445 ++++
>>   include/drm/drm_gpuvm.h                     |    5 +
>>   include/drm/drm_pagemap.h                   |  103 +
>>   include/linux/migrate.h                     |    1 +
>>   include/uapi/drm/xe_drm.h                   |   19 +-
>>   mm/memory.c                                 |   13 +-
>>   mm/migrate_device.c                         |  116 +-
>>   33 files changed, 5061 insertions(+), 157 deletions(-)
>>   create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>>   create mode 100644 drivers/gpu/drm/xe/xe_svm.c
>>   create mode 100644 drivers/gpu/drm/xe/xe_svm.h
>>   create mode 100644 include/drm/drm_gpusvm.h
>>   create mode 100644 include/drm/drm_pagemap.h
>>
> 

Re: [PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Matthew Brost]

On Fri, Jan 17, 2025 at 10:26:16AM +0200, Gwan-gyeong Mun wrote:
> 
> 
> On 12/18/24 1:33 AM, Matthew Brost wrote:
> > This patch introduces support for GPU Shared Virtual Memory (SVM) in the
> > Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
> > sharing of memory between the CPU and GPU, enhancing performance and
> > flexibility in GPU computing tasks.
> > 
> > The patch adds the necessary infrastructure for SVM, including data
> > structures and functions for managing SVM ranges and notifiers. It also
> > provides mechanisms for allocating, deallocating, and migrating memory
> > regions between system RAM and GPU VRAM.
> > 
> > This is largely inspired by GPUVM.
> > 
> > v2:
> >   - Take order into account in check pages
> >   - Clear range->pages in get pages error
> >   - Drop setting dirty or accessed bit in get pages (Vetter)
> >   - Remove mmap assert for cpu faults
> >   - Drop mmap write lock abuse (Vetter, Christian)
> >   - Decouple zdd from range (Vetter, Oak)
> >   - Add drm_gpusvm_range_evict, make it work with coherent pages
> >   - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
> >   - mmget/put in drm_gpusvm_evict_to_sram
> >   - Drop range->vram_alloation variable
> >   - Don't return in drm_gpusvm_evict_to_sram until all pages detached
> >   - Don't warn on mixing sram and device pages
> >   - Update kernel doc
> >   - Add coherent page support to get pages
> >   - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
> >   - Add struct drm_gpusvm_vram and ops (Thomas)
> >   - Update the range's seqno if the range is valid (Thomas)
> >   - Remove the is_unmapped check before hmm_range_fault (Thomas)
> >   - Use drm_pagemap (Thomas)
> >   - Drop kfree_mapping (Thomas)
> >   - dma mapp pages under notifier lock (Thomas)
> >   - Remove ctx.prefault
> >   - Remove ctx.mmap_locked
> >   - Add ctx.check_pages
> >   - s/vram/devmem (Thomas)
> > v3:
> >   - Fix memory leak drm_gpusvm_range_get_pages
> >   - Only migrate pages with same zdd on CPU fault
> >   - Loop over al VMAs in drm_gpusvm_range_evict
> >   - Make GPUSVM a drm level module
> >   - GPL or MIT license
> >   - Update main kernel doc (Thomas)
> >   - Prefer foo() vs foo for functions in kernel doc (Thomas)
> >   - Prefer functions over macros (Thomas)
> >   - Use unsigned long vs u64 for addresses (Thomas)
> >   - Use standard interval_tree (Thomas)
> >   - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
> >   - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
> >   - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
> >   - Newlines between functions defs in header file (Thomas)
> >   - Drop shall language in driver vfunc kernel doc (Thomas)
> >   - Move some static inlines from head to C file (Thomas)
> >   - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
> >   - Change check_pages to a threshold
> > 
> > Cc: Simona Vetter <simona.vetter@ffwll.ch>
> > Cc: Dave Airlie <airlied@redhat.com>
> > Cc: Christian König <christian.koenig@amd.com>
> > Cc: <dri-devel@lists.freedesktop.org>
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> > ---
> >   drivers/gpu/drm/Kconfig      |    8 +
> >   drivers/gpu/drm/Makefile     |    1 +
> >   drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
> >   include/drm/drm_gpusvm.h     |  445 +++++++
> >   4 files changed, 2674 insertions(+)
> >   create mode 100644 drivers/gpu/drm/drm_gpusvm.c
> >   create mode 100644 include/drm/drm_gpusvm.h
> > 
> > diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> > index 2f51546b0b88..5524cb1ed9dd 100644
> > --- a/drivers/gpu/drm/Kconfig
> > +++ b/drivers/gpu/drm/Kconfig
> > @@ -278,6 +278,14 @@ config DRM_GPUVM
> >   	  GPU-VM representation providing helpers to manage a GPUs virtual
> >   	  address space
> > +config DRM_GPUSVM
> > +	tristate
> > +	depends on DRM
> > +	depends on DEVICE_MIGRATION
> > +	help
> > +	  GPU-SVM representation providing helpers to manage a GPUs shared
> > +	  virtual memory
> > +
> >   config DRM_BUDDY
> >   	tristate
> >   	depends on DRM
> > diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> > index 19fb370fbc56..ca5d44441fb6 100644
> > --- a/drivers/gpu/drm/Makefile
> > +++ b/drivers/gpu/drm/Makefile
> > @@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
> >   #
> >   obj-$(CONFIG_DRM_EXEC) += drm_exec.o
> >   obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> > +obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
> >   obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
> > diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> > new file mode 100644
> > index 000000000000..29ce2c7d1c8c
> > --- /dev/null
> > +++ b/drivers/gpu/drm/drm_gpusvm.c
> > @@ -0,0 +1,2220 @@
> > +// SPDX-License-Identifier: GPL-2.0-only OR MIT
> > +/*
> > + * Copyright © 2024 Intel Corporation
> > + *
> > + * Authors:
> > + *     Matthew Brost <matthew.brost@intel.com>
> > + */
> > +
> > +#include <linux/dma-mapping.h>
> > +#include <linux/hmm.h>
> > +#include <linux/memremap.h>
> > +#include <linux/migrate.h>
> > +#include <linux/mm_types.h>
> > +#include <linux/pagemap.h>
> > +#include <linux/slab.h>
> > +
> > +#include <drm/drm_device.h>
> > +#include <drm/drm_gpusvm.h>
> > +#include <drm/drm_pagemap.h>
> > +#include <drm/drm_print.h>
> > +
> > +/**
> > + * DOC: Overview
> > + *
> > + * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
> > + *
> > + * The GPU SVM layer is a component of the DRM framework designed to manage shared
> > + * virtual memory between the CPU and GPU. It enables efficient data exchange and
> > + * processing for GPU-accelerated applications by allowing memory sharing and
> > + * synchronization between the CPU's and GPU's virtual address spaces.
> > + *
> > + * Key GPU SVM Components:
> > + * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
> > + *		GPU of changes, notifiers are sized based on a GPU SVM
> > + *		initialization parameter, with a recommendation of 512M or
> > + *		larger. They maintain a Red-BlacK tree and a list of ranges that
> > + *		fall within the notifier interval. Notifiers are tracked within
> > + *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
> > + *		or removed as ranges within the interval are created or
> > + *		destroyed.
> > + * - Ranges: Represent memory ranges mapped in a DRM device and managed
> > + *	     by GPU SVM. They are sized based on an array of chunk sizes, which
> > + *	     is a GPU SVM initialization parameter, and the CPU address space.
> > + *	     Upon GPU fault, the largest aligned chunk that fits within the
> > + *	     faulting CPU address space is chosen for the range size. Ranges are
> > + *	     expected to be dynamically allocated on GPU fault and removed on an
> > + *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
> > + *	     a notifier's Red-Black tree.
> > + * - Operations: Define the interface for driver-specific GPU SVM operations
> > + *               such as range allocation, notifier allocation, and
> > + *               invalidations.
> > + * - Device Memory Allocations: Embedded structure containing enough information
> > + *                              for GPU SVM to migrate to / from device memory.
> > + * - Device Memory Operations: Define the interface for driver-specific device
> > + *                             memory operations release memory, populate pfns,
> > + *                             and copy to / from device memory.
> > + *
> > + * This layer provides interfaces for allocating, mapping, migrating, and
> > + * releasing memory ranges between the CPU and GPU. It handles all core memory
> > + * management interactions (DMA mapping, HMM, and migration) and provides
> > + * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
> > + * to build the expected driver components for an SVM implementation as detailed
> > + * below.
> > + *
> > + * Expected Driver Components:
> > + * - GPU page fault handler: Used to create ranges and notifiers based on the
> > + *			     fault address, optionally migrate the range to
> > + *			     device memory, and create GPU bindings.
> > + * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
> > + *			Ranges are expected to be added to the garbage collector
> > + *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
> > + * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
> > + *			ranges.
> > + */
> > +
> > +/**
> > + * DOC: Locking
> > + *
> > + * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
> > + * mmap lock as needed.
> > + *
> > + * GPU SVM introduces a global notifier lock, which safeguards the notifier's
> > + * range RB tree and list, as well as the range's DMA mappings and sequence
> > + * number. GPU SVM manages all necessary locking and unlocking operations,
> > + * except for the recheck range's pages being valid
> > + * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
> > + * lock corresponds to the 'driver->update' lock mentioned in the HMM
> > + * documentation (TODO: Link). Future revisions may transition from a GPU SVM
> > + * global lock to a per-notifier lock if finer-grained locking is deemed
> > + * necessary.
> > + *
> > + * In addition to the locking mentioned above, the driver should implement a
> > + * lock to safeguard core GPU SVM function calls that modify state, such as
> > + * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
> > + * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
> > + * locking should also be possible for concurrent GPU fault processing within a
> > + * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
> > + * to add annotations to GPU SVM.
> > + */
> > +
> > +/**
> > + * DOC: Migration
> > + *
> > + * The migration support is quite simple, allowing migration between RAM and
> > + * device memory at the range granularity. For example, GPU SVM currently does not
> > + * support mixing RAM and device memory pages within a range. This means that upon GPU
> > + * fault, the entire range can be migrated to device memory, and upon CPU fault, the
> > + * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
> > + * could be added in the future if required.
> > + *
> > + * The reasoning for only supporting range granularity is as follows: it
> > + * simplifies the implementation, and range sizes are driver-defined and should
> > + * be relatively small.
> > + */
> > +
> > +/**
> > + * DOC: Partial Unmapping of Ranges
> > + *
> > + * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
> > + * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
> > + * being that a subset of the range still has CPU and GPU mappings. If the
> > + * backing store for the range is in device memory, a subset of the backing store has
> > + * references. One option would be to split the range and device memory backing store,
> > + * but the implementation for this would be quite complicated. Given that
> > + * partial unmappings are rare and driver-defined range sizes are relatively
> > + * small, GPU SVM does not support splitting of ranges.
> > + *
> > + * With no support for range splitting, upon partial unmapping of a range, the
> > + * driver is expected to invalidate and destroy the entire range. If the range
> > + * has device memory as its backing, the driver is also expected to migrate any
> > + * remaining pages back to RAM.
> > + */
> > +
> > +/**
> > + * DOC: Examples
> > + *
> > + * This section provides three examples of how to build the expected driver
> > + * components: the GPU page fault handler, the garbage collector, and the
> > + * notifier callback.
> > + *
> > + * The generic code provided does not include logic for complex migration
> > + * policies, optimized invalidations, fined grained driver locking, or other
> > + * potentially required driver locking (e.g., DMA-resv locks).
> > + *
> > + * 1) GPU page fault handler
> > + *
> > + *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
> > + *	{
> > + *		int err = 0;
> > + *
> > + *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
> > + *
> > + *		drm_gpusvm_notifier_lock(gpusvm);
> > + *		if (drm_gpusvm_range_pages_valid(range))
> > + *			driver_commit_bind(gpusvm, range);
> > + *		else
> > + *			err = -EAGAIN;
> > + *		drm_gpusvm_notifier_unlock(gpusvm);
> > + *
> > + *		return err;
> > + *	}
> > + *
> > + *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
> > + *			     unsigned long gpuva_start, unsigned long gpuva_end)
> > + *	{
> > + *		struct drm_gpusvm_ctx ctx = {};
> > + *		int err;
> > + *
> > + *		driver_svm_lock();
> > + *	retry:
> > + *		// Always process UNMAPs first so view of GPU SVM ranges is current
> > + *		driver_garbage_collector(gpusvm);
> > + *
> > + *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
> > + *							gpuva_start, gpuva_end,
> > + *						        &ctx);
> > + *		if (IS_ERR(range)) {
> > + *			err = PTR_ERR(range);
> > + *			goto unlock;
> > + *		}
> > + *
> > + *		if (driver_migration_policy(range)) {
> > + *			devmem = driver_alloc_devmem();
> > + *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
> > + *							   devmem_allocation,
> > + *							   &ctx);
> > + *			if (err)	// CPU mappings may have changed
> > + *				goto retry;
> > + *		}
> > + *
> > + *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
> > + *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
> > + *			if (err == -EOPNOTSUPP)
> > + *				drm_gpusvm_range_evict(gpusvm, range);
> > + *			goto retry;
> > + *		} else if (err) {
> > + *			goto unlock;
> > + *		}
> > + *
> > + *		err = driver_bind_range(gpusvm, range);
> > + *		if (err == -EAGAIN)	// CPU mappings changed
> > + *			goto retry
> > + *
> > + *	unlock:
> > + *		driver_svm_unlock();
> > + *		return err;
> > + *	}
> > + *
> > + * 2) Garbage Collector.
> > + *
> > + *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
> > + *					struct drm_gpusvm_range *range)
> > + *	{
> > + *		assert_driver_svm_locked(gpusvm);
> > + *
> > + *		// Partial unmap, migrate any remaining device memory pages back to RAM
> > + *		if (range->flags.partial_unmap)
> > + *			drm_gpusvm_range_evict(gpusvm, range);
> > + *
> > + *		driver_unbind_range(range);
> > + *		drm_gpusvm_range_remove(gpusvm, range);
> > + *	}
> > + *
> > + *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
> > + *	{
> > + *		assert_driver_svm_locked(gpusvm);
> > + *
> > + *		for_each_range_in_garbage_collector(gpusvm, range)
> > + *			__driver_garbage_collector(gpusvm, range);
> > + *	}
> > + *
> > + * 3) Notifier callback.
> > + *
> > + *	void driver_invalidation(struct drm_gpusvm *gpusvm,
> > + *				 struct drm_gpusvm_notifier *notifier,
> > + *				 const struct mmu_notifier_range *mmu_range)
> > + *	{
> > + *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
> > + *		struct drm_gpusvm_range *range = NULL;
> > + *
> > + *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
> > + *
> > + *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
> > + *					  mmu_range->end) {
> > + *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
> > + *
> > + *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
> > + *				continue;
> > + *
> > + *			drm_gpusvm_range_set_unmapped(range, mmu_range);
> > + *			driver_garbage_collector_add(gpusvm, range);
> > + *		}
> > + *	}
> > + */
> > +
> > +/**
> > + * npages_in_range() - Calculate the number of pages in a given range
> > + * @start: The start address of the range
> > + * @end: The end address of the range
> > + *
> > + * This macro calculates the number of pages in a given memory range,
> > + * specified by the start and end addresses. It divides the difference
> > + * between the end and start addresses by the page size (PAGE_SIZE) to
> > + * determine the number of pages in the range.
> > + *
> > + * Return: The number of pages in the specified range.
> > + */
> > +static unsigned long
> > +npages_in_range(unsigned long start, unsigned long end)
> > +{
> > +	return (end - start) >> PAGE_SHIFT;
> > +}
> > +
> > +/**
> > + * struct drm_gpusvm_zdd - GPU SVM zone device data
> > + *
> > + * @refcount: Reference count for the zdd
> > + * @destroy_work: Work structure for asynchronous zdd destruction
> > + * @devmem_allocation: device memory allocation
> > + * @device_private_page_owner: Device private pages owner
> > + *
> > + * This structure serves as a generic wrapper installed in
> > + * page->zone_device_data. It provides infrastructure for looking up a device
> > + * memory allocation upon CPU page fault and asynchronously releasing device
> > + * memory once the CPU has no page references. Asynchronous release is useful
> > + * because CPU page references can be dropped in IRQ contexts, while releasing
> > + * device memory likely requires sleeping locks.
> > + */
> > +struct drm_gpusvm_zdd {
> > +	struct kref refcount;
> > +	struct work_struct destroy_work;
> > +	struct drm_gpusvm_devmem *devmem_allocation;
> > +	void *device_private_page_owner;
> > +};
> > +
> > +/**
> > + * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
> > + * @w: Pointer to the work_struct
> > + *
> > + * This function releases device memory, puts GPU SVM range, and frees zdd.
> > + */
> > +static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
> > +{
> > +	struct drm_gpusvm_zdd *zdd =
> > +		container_of(w, struct drm_gpusvm_zdd, destroy_work);
> > +	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
> > +		zdd->devmem_allocation->ops : NULL;
> > +
> > +	if (zdd->devmem_allocation && ops->devmem_release)
> > +		ops->devmem_release(zdd->devmem_allocation);
> > +	kfree(zdd);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
> > + * @device_private_page_owner: Device private pages owner
> > + *
> > + * This function allocates and initializes a new zdd structure. It sets up the
> > + * reference count and initializes the destroy work.
> > + *
> > + * Returns:
> > + * Pointer to the allocated zdd on success, ERR_PTR() on failure.
> > + */
> > +static struct drm_gpusvm_zdd *
> > +drm_gpusvm_zdd_alloc(void *device_private_page_owner)
> > +{
> > +	struct drm_gpusvm_zdd *zdd;
> > +
> > +	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
> > +	if (!zdd)
> > +		return NULL;
> > +
> > +	kref_init(&zdd->refcount);
> > +	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
> > +	zdd->devmem_allocation = NULL;
> > +	zdd->device_private_page_owner = device_private_page_owner;
> > +
> > +	return zdd;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
> > + * @zdd: Pointer to the zdd structure.
> > + *
> > + * This function increments the reference count of the provided zdd structure.
> > + *
> > + * Returns: Pointer to the zdd structure.
> > + */
> > +static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
> > +{
> > +	kref_get(&zdd->refcount);
> > +	return zdd;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
> > + * @ref: Pointer to the reference count structure.
> > + *
> > + * This function queues the destroy_work of the zdd for asynchronous destruction.
> > + */
> > +static void drm_gpusvm_zdd_destroy(struct kref *ref)
> > +{
> > +	struct drm_gpusvm_zdd *zdd =
> > +		container_of(ref, struct drm_gpusvm_zdd, refcount);
> > +
> > +	if (zdd->devmem_allocation)
> > +		WRITE_ONCE(zdd->devmem_allocation->detached, true);
> > +	schedule_work(&zdd->destroy_work);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_zdd_put() - Put a zdd reference.
> > + * @zdd: Pointer to the zdd structure.
> > + *
> > + * This function decrements the reference count of the provided zdd structure
> > + * and schedules its destruction if the count drops to zero.
> > + */
> > +static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
> > +{
> > +	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
> > + * @notifier: Pointer to the GPU SVM notifier structure.
> > + * @start: Start address of the range
> > + * @end: End address of the range
> > + *
> > + * Return: A pointer to the drm_gpusvm_range if found or NULL
> > + */
> > +struct drm_gpusvm_range *
> > +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> > +		      unsigned long end)
> > +{
> > +	struct interval_tree_node *itree;
> > +
> > +	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
> > +
> > +	if (itree)
> > +		return container_of(itree, struct drm_gpusvm_range, itree);
> > +	else
> > +		return NULL;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
> > +
> > +/**
> > + * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
> > + * @range__: Iterator variable for the ranges
> > + * @next__: Iterator variable for the ranges temporay storage
> > + * @notifier__: Pointer to the GPU SVM notifier
> > + * @start__: Start address of the range
> > + * @end__: End address of the range
> > + *
> > + * This macro is used to iterate over GPU SVM ranges in a notifier while
> > + * removing ranges from it.
> > + */
> > +#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
> > +	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
> > +	     (next__) = __drm_gpusvm_range_next(range__);				\
> > +	     (range__) && (range__->itree.start < (end__));				\
> > +	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
> > +
> > +/**
> > + * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
> > + * @notifier: a pointer to the current drm_gpusvm_notifier
> > + *
> > + * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
> > + *         the current notifier is the last one or if the input notifier is
> > + *         NULL.
> > + */
> > +static struct drm_gpusvm_notifier *
> > +__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
> > +{
> > +	if (notifier && !list_is_last(&notifier->entry,
> > +				      &notifier->gpusvm->notifier_list))
> > +		return list_next_entry(notifier, entry);
> > +
> > +	return NULL;
> > +}
> > +
> > +static struct drm_gpusvm_notifier *
> > +notifier_iter_first(struct rb_root_cached *root, unsigned long start,
> > +		    unsigned long last)
> > +{
> > +	struct interval_tree_node *itree;
> > +
> > +	itree = interval_tree_iter_first(root, start, last);
> > +
> > +	if (itree)
> > +		return container_of(itree, struct drm_gpusvm_notifier, itree);
> > +	else
> > +		return NULL;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
> > + * @notifier__: Iterator variable for the notifiers
> > + * @notifier__: Pointer to the GPU SVM notifier
> > + * @start__: Start address of the notifier
> > + * @end__: End address of the notifier
> > + *
> > + * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
> > + */
> > +#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
> > +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
> > +	     (notifier__) && (notifier__->itree.start < (end__));			\
> > +	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
> > +
> > +/**
> > + * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
> > + * @notifier__: Iterator variable for the notifiers
> > + * @next__: Iterator variable for the notifiers temporay storage
> > + * @notifier__: Pointer to the GPU SVM notifier
> > + * @start__: Start address of the notifier
> > + * @end__: End address of the notifier
> > + *
> > + * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
> > + * removing notifiers from it.
> > + */
> > +#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
> > +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
> > +	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
> > +	     (notifier__) && (notifier__->itree.start < (end__));			\
> > +	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
> > +
> > +/**
> > + * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
> > + * @mni: Pointer to the mmu_interval_notifier structure.
> > + * @mmu_range: Pointer to the mmu_notifier_range structure.
> > + * @cur_seq: Current sequence number.
> > + *
> > + * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
> > + * notifier sequence number and calls the driver invalidate vfunc under
> > + * gpusvm->notifier_lock.
> > + *
> > + * Returns:
> > + * true if the operation succeeds, false otherwise.
> > + */
> > +static bool
> > +drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
> > +			       const struct mmu_notifier_range *mmu_range,
> > +			       unsigned long cur_seq)
> > +{
> > +	struct drm_gpusvm_notifier *notifier =
> > +		container_of(mni, typeof(*notifier), notifier);
> > +	struct drm_gpusvm *gpusvm = notifier->gpusvm;
> > +
> > +	if (!mmu_notifier_range_blockable(mmu_range))
> > +		return false;
> > +
> > +	down_write(&gpusvm->notifier_lock);
> > +	mmu_interval_set_seq(mni, cur_seq);
> > +	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
> > +	up_write(&gpusvm->notifier_lock);
> > +
> > +	return true;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
> > + */
> > +static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
> > +	.invalidate = drm_gpusvm_notifier_invalidate,
> > +};
> > +
> > +/**
> > + * drm_gpusvm_init() - Initialize the GPU SVM.
> > + * @gpusvm: Pointer to the GPU SVM structure.
> > + * @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.
> > + * @ops: Pointer to the operations structure for GPU SVM.
> > + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> > + *               Entries should be powers of 2 in descending order with last
> > + *               entry being SZ_4K.
> > + * @num_chunks: Number of chunks.
> > + *
> > + * This function initializes the GPU SVM.
> > + *
> > + * Returns:
> > + * 0 on success, a negative error code on failure.
> > + */
> > +int drm_gpusvm_init(struct drm_gpusvm *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,
> > +		    const struct drm_gpusvm_ops *ops,
> > +		    const unsigned long *chunk_sizes, int num_chunks)
> > +{
> > +	if (!ops->invalidate || !num_chunks)
> > +		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;
> > +	gpusvm->ops = ops;
> > +	gpusvm->chunk_sizes = chunk_sizes;
> > +	gpusvm->num_chunks = num_chunks;
> > +
> > +	mmgrab(mm);
> > +	gpusvm->root = RB_ROOT_CACHED;
> > +	INIT_LIST_HEAD(&gpusvm->notifier_list);
> > +
> > +	init_rwsem(&gpusvm->notifier_lock);
> > +
> > +	fs_reclaim_acquire(GFP_KERNEL);
> > +	might_lock(&gpusvm->notifier_lock);
> > +	fs_reclaim_release(GFP_KERNEL);
> > +
> > +#ifdef CONFIG_LOCKDEP
> > +	gpusvm->lock_dep_map = NULL;
> > +#endif
> > +
> > +	return 0;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_init);
> > +
> > +/**
> > + * drm_gpusvm_notifier_find() - Find GPU SVM notifier
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @fault_addr: Fault address
> > + *
> > + * This function finds the GPU SVM notifier associated with the fault address.
> > + *
> > + * Returns:
> > + * Pointer to the GPU SVM notifier on success, NULL otherwise.
> > + */
> > +static struct drm_gpusvm_notifier *
> > +drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
> > +			 unsigned long fault_addr)
> > +{
> > +	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
> > +}
> > +
> > +/**
> > + * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
> > + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
> > + *
> > + * Return: A pointer to the containing drm_gpusvm_notifier structure.
> > + */
> > +static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
> > +{
> > +	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + *
> > + * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
> > + */
> > +static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
> > +				       struct drm_gpusvm_notifier *notifier)
> > +{
> > +	struct rb_node *node;
> > +	struct list_head *head;
> > +
> > +	interval_tree_insert(&notifier->itree, &gpusvm->root);
> > +
> > +	node = rb_prev(&notifier->itree.rb);
> > +	if (node)
> > +		head = &(to_drm_gpusvm_notifier(node))->entry;
> > +	else
> > +		head = &gpusvm->notifier_list;
> > +
> > +	list_add(&notifier->entry, head);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
> > + * @gpusvm: Pointer to the GPU SVM tructure
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + *
> > + * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
> > + */
> > +static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
> > +				       struct drm_gpusvm_notifier *notifier)
> > +{
> > +	interval_tree_remove(&notifier->itree, &gpusvm->root);
> > +	list_del(&notifier->entry);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_fini() - Finalize the GPU SVM.
> > + * @gpusvm: Pointer to the GPU SVM structure.
> > + *
> > + * This function finalizes the GPU SVM by cleaning up any remaining ranges and
> > + * notifiers, and dropping a reference to struct MM.
> > + */
> > +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
> > +{
> > +	struct drm_gpusvm_notifier *notifier, *next;
> > +
> > +	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
> > +		struct drm_gpusvm_range *range, *__next;
> > +
> > +		/*
> > +		 * Remove notifier first to avoid racing with any invalidation
> > +		 */
> > +		mmu_interval_notifier_remove(&notifier->notifier);
> > +		notifier->flags.removed = true;
> > +
> > +		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
> > +					       LONG_MAX)
> > +			drm_gpusvm_range_remove(gpusvm, range);
> > +	}
> > +
> > +	mmdrop(gpusvm->mm);
> > +	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
> > +
> > +/**
> > + * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @fault_addr: Fault address
> > + *
> > + * This function allocates and initializes the GPU SVM notifier structure.
> > + *
> > + * Returns:
> > + * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
> > + */
> > +static struct drm_gpusvm_notifier *
> > +drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
> > +{
> > +	struct drm_gpusvm_notifier *notifier;
> > +
> > +	if (gpusvm->ops->notifier_alloc)
> > +		notifier = gpusvm->ops->notifier_alloc();
> > +	else
> > +		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
> > +
> > +	if (!notifier)
> > +		return ERR_PTR(-ENOMEM);
> > +
> > +	notifier->gpusvm = gpusvm;
> > +	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
> > +	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
> > +	INIT_LIST_HEAD(&notifier->entry);
> > +	notifier->root = RB_ROOT_CACHED;
> > +	INIT_LIST_HEAD(&notifier->range_list);
> > +
> > +	return notifier;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_notifier_free() - Free GPU SVM notifier
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + *
> > + * This function frees the GPU SVM notifier structure.
> > + */
> > +static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
> > +				     struct drm_gpusvm_notifier *notifier)
> > +{
> > +	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
> > +
> > +	if (gpusvm->ops->notifier_free)
> > +		gpusvm->ops->notifier_free(notifier);
> > +	else
> > +		kfree(notifier);
> > +}
> > +
> > +/**
> > + * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
> > + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
> > + *
> > + * Return: A pointer to the containing drm_gpusvm_range structure.
> > + */
> > +static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
> > +{
> > +	return container_of(node, struct drm_gpusvm_range, itree.rb);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_insert() - Insert GPU SVM range
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + * @range: Pointer to the GPU SVM range structure
> > + *
> > + * This function inserts the GPU SVM range into the notifier RB tree and list.
> > + */
> > +static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
> > +				    struct drm_gpusvm_range *range)
> > +{
> > +	struct rb_node *node;
> > +	struct list_head *head;
> > +
> > +	drm_gpusvm_notifier_lock(notifier->gpusvm);
> > +	interval_tree_insert(&range->itree, &notifier->root);
> > +
> > +	node = rb_prev(&range->itree.rb);
> > +	if (node)
> > +		head = &(to_drm_gpusvm_range(node))->entry;
> > +	else
> > +		head = &notifier->range_list;
> > +
> > +	list_add(&range->entry, head);
> > +	drm_gpusvm_notifier_unlock(notifier->gpusvm);
> > +}
> > +
> > +/**
> > + * __drm_gpusvm_range_remove() - Remove GPU SVM range
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + * @range: Pointer to the GPU SVM range structure
> > + *
> > + * This macro removes the GPU SVM range from the notifier RB tree and list.
> > + */
> > +static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
> > +				      struct drm_gpusvm_range *range)
> > +{
> > +	interval_tree_remove(&range->itree, &notifier->root);
> > +	list_del(&range->entry);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_alloc() - Allocate GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + * @fault_addr: Fault address
> > + * @chunk_size: Chunk size
> > + * @migrate_devmem: Flag indicating whether to migrate device memory
> > + *
> > + * This function allocates and initializes the GPU SVM range structure.
> > + *
> > + * Returns:
> > + * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
> > + */
> > +static struct drm_gpusvm_range *
> > +drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
> > +		       struct drm_gpusvm_notifier *notifier,
> > +		       unsigned long fault_addr, unsigned long chunk_size,
> > +		       bool migrate_devmem)
> > +{
> > +	struct drm_gpusvm_range *range;
> > +
> > +	if (gpusvm->ops->range_alloc)
> > +		range = gpusvm->ops->range_alloc(gpusvm);
> > +	else
> > +		range = kzalloc(sizeof(*range), GFP_KERNEL);
> > +
> > +	if (!range)
> > +		return ERR_PTR(-ENOMEM);
> > +
> > +	kref_init(&range->refcount);
> > +	range->gpusvm = gpusvm;
> > +	range->notifier = notifier;
> > +	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
> > +	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
> > +	INIT_LIST_HEAD(&range->entry);
> > +	range->notifier_seq = LONG_MAX;
> > +	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
> > +
> > +	return range;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_check_pages() - Check pages
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + * @start: Start address
> > + * @end: End address
> > + *
> > + * Check if pages between start and end have been faulted in on the CPU. Use to
> > + * prevent migration of pages without CPU backing store.
> > + *
> > + * Returns:
> > + * True if pages have been faulted into CPU, False otherwise
> > + */
> > +static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
> > +				   struct drm_gpusvm_notifier *notifier,
> > +				   unsigned long start, unsigned long end)
> > +{
> > +	struct hmm_range hmm_range = {
> > +		.default_flags = 0,
> > +		.notifier = &notifier->notifier,
> > +		.start = start,
> > +		.end = end,
> > +		.dev_private_owner = gpusvm->device_private_page_owner,
> > +	};
> > +	unsigned long timeout =
> > +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> > +	unsigned long *pfns;
> > +	unsigned long npages = npages_in_range(start, end);
> > +	int err, i;
> > +
> > +	mmap_assert_locked(gpusvm->mm);
> > +
> > +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> > +	if (!pfns)
> > +		return false;
> > +
> > +	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> > +	hmm_range.hmm_pfns = pfns;
> > +
> > +	while (true) {
> > +		err = hmm_range_fault(&hmm_range);
> > +		if (err == -EBUSY) {
> > +			if (time_after(jiffies, timeout))
> > +				break;
> > +
> > +			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> > +			continue;
> > +		}
> > +		break;
> > +	}
> > +	if (err)
> > +		goto err_free;
> > +
> > +	for (i = 0; i < npages;) {
> > +		if (!(pfns[i] & HMM_PFN_VALID)) {
> > +			err = -EFAULT;
> > +			goto err_free;
> > +		}
> > +		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
> > +	}
> > +
> > +err_free:
> > +	kvfree(pfns);
> > +	return err ? false : true;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: Pointer to the GPU SVM notifier structure
> > + * @vas: Pointer to the virtual memory area structure
> Hi Matthew Brost,
> 
> mm core mainly uses vma as the naming of the structure vm_area_struct
> variable.
> can you consider changing vas to vma for the same recognition as mm core?
> ( vas naming is used in this function, as well as many of the functions
> introduced in this patch. )

This is intentional. We are not in core MM rather a GPU driver. GPUs
also have VMAs, grep for vma in DRM subsystem. These are GPU VMAs while
a vm_area_struct is a CPU VMA. To avoid confusing GPU VMAs and CPU VMAs,
I decided on 'vas' for the latter in this code.

> > + * @fault_addr: Fault address
> > + * @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
> > + *
> > + * 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
> > + * memory area boundaries.
> > + *
> > + * Returns:
> > + * Chunk size on success, LONG_MAX on failure.
> > + */
> > +static unsigned long
> > +drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
> > +			    struct drm_gpusvm_notifier *notifier,
> > +			    struct vm_area_struct *vas,
> > +			    unsigned long fault_addr,
> > +			    unsigned long gpuva_start,
> > +			    unsigned long gpuva_end,
> > +			    unsigned long check_pages_threshold)
> > +{
> > +	unsigned long start, end;
> > +	int i = 0;
> > +
> > +retry:
> > +	for (; i < gpusvm->num_chunks; ++i) {
> > +		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
> > +		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
> > +
> > +		if (start >= vas->vm_start && end <= vas->vm_end &&
> > +		    start >= notifier->itree.start &&
> > +		    end <= notifier->itree.last + 1 &&
> > +		    start >= gpuva_start && end <= gpuva_end)
> > +			break;
> > +	}
> > +
> > +	if (i == gpusvm->num_chunks)
> > +		return LONG_MAX;
> > +
> > +	/*
> > +	 * If allocation more than page, ensure not to overlap with existing
> > +	 * ranges.
> > +	 */
> > +	if (end - start != SZ_4K) {
> > +		struct drm_gpusvm_range *range;
> > +
> > +		range = drm_gpusvm_range_find(notifier, start, end);
> > +		if (range) {
> > +			++i;
> > +			goto retry;
> > +		}
> > +
> > +		/*
> > +		 * XXX: Only create range on pages CPU has faulted in. Without
> > +		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
> > +		 * process-many-malloc' fails. In the failure case, each process
> > +		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
> > +		 * ranges. When migrating the SVM ranges, some processes fail in
> > +		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
> > +		 * and then upon drm_gpusvm_range_get_pages device pages from
> > +		 * other processes are collected + faulted in which creates all
> > +		 * sorts of problems. Unsure exactly how this happening, also
> > +		 * problem goes away if 'xe_exec_system_allocator --r
> > +		 * process-many-malloc' mallocs at least 64k at a time.
> > +		 */
> > +		if (end - start <= check_pages_threshold &&
> > +		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
> > +			++i;
> > +			goto retry;
> > +		}
> > +	}
> > +
> > +	return end - start;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @fault_addr: Fault address
> > + * @gpuva_start: Start address of GPUVA which mirrors CPU
> > + * @gpuva_end: End address of GPUVA which mirrors CPU
> > + * @ctx: GPU SVM context
> > + *
> > + * This function finds or inserts a newly allocated a GPU SVM range based on the
> > + * fault address. Caller must hold a lock to protect range lookup and insertion.
> > + *
> > + * Returns:
> > + * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
> > + */
> > +struct drm_gpusvm_range *
> > +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> > +				unsigned long fault_addr,
> > +				unsigned long gpuva_start,
> > +				unsigned long gpuva_end,
> > +				const struct drm_gpusvm_ctx *ctx)
> > +{
> > +	struct drm_gpusvm_notifier *notifier;
> > +	struct drm_gpusvm_range *range;
> > +	struct mm_struct *mm = gpusvm->mm;
> > +	struct vm_area_struct *vas;
> > +	bool notifier_alloc = false;
> > +	unsigned long chunk_size;
> > +	int err;
> > +	bool migrate_devmem;
> > +
> > +	drm_gpusvm_driver_lock_held(gpusvm);
> > +
> > +	if (fault_addr < gpusvm->mm_start ||
> > +	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
> > +		return ERR_PTR(-EINVAL);
> > +
> > +	if (!mmget_not_zero(mm))
> > +		return ERR_PTR(-EFAULT);
> > +
> > +	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
> > +	if (!notifier) {
> > +		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
> > +		if (IS_ERR(notifier)) {
> > +			err = PTR_ERR(notifier);
> > +			goto err_mmunlock;
> > +		}
> > +		notifier_alloc = true;
> > +		err = mmu_interval_notifier_insert(&notifier->notifier,
> > +						   mm, notifier->itree.start,
> > +						   notifier->itree.last -
> > +						   notifier->itree.start + 1,
> > +						   &drm_gpusvm_notifier_ops);
> > +		if (err)
> > +			goto err_notifier;
> > +	}
> > +
> > +	mmap_read_lock(mm);
> > +
> > +	vas = vma_lookup(mm, fault_addr);
> > +	if (!vas) {
> > +		err = -ENOENT;
> > +		goto err_notifier_remove;
> > +	}
> > +
> > +	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
> > +		err = -EPERM;
> > +		goto err_notifier_remove;
> > +	}
> > +
> > +	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
> > +	if (range)
> > +		goto out_mmunlock;
> > +	/*
> > +	 * XXX: Short-circuiting migration based on migrate_vma_* current
> > +	 * limitations. If/when migrate_vma_* add more support, this logic will
> > +	 * have to change.
> > +	 */
> > +	migrate_devmem = ctx->devmem_possible &&
> > +		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
> > +
> > +	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
> > +						 fault_addr, gpuva_start,
> > +						 gpuva_end, migrate_devmem &&
> > +						 ctx->check_pages_threshold);
> > +	if (chunk_size == LONG_MAX) {
> > +		err = -EINVAL;
> > +		goto err_notifier_remove;
> > +	}
> > +
> > +	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
> > +				       migrate_devmem);
> > +	if (IS_ERR(range)) {
> > +		err = PTR_ERR(range);
> > +		goto err_notifier_remove;
> > +	}
> > +
> > +	drm_gpusvm_range_insert(notifier, range);
> > +	if (notifier_alloc)
> > +		drm_gpusvm_notifier_insert(gpusvm, notifier);
> > +
> > +out_mmunlock:
> > +	mmap_read_unlock(mm);
> > +	mmput(mm);
> > +
> > +	return range;
> > +
> > +err_notifier_remove:
> > +	mmap_read_unlock(mm);
> > +	if (notifier_alloc)
> > +		mmu_interval_notifier_remove(&notifier->notifier);
> > +err_notifier:
> > +	if (notifier_alloc)
> > +		drm_gpusvm_notifier_free(gpusvm, notifier);
> > +err_mmunlock:
> > +	mmput(mm);
> > +	return ERR_PTR(err);
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
> > +
> > +/**
> > + * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + * @npages: Number of pages to unmap
> > + *
> > + * This function unmap pages associated with a GPU SVM range. Assumes and
> > + * asserts correct locking is in place when called.
> > + */
> > +static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> > +					   struct drm_gpusvm_range *range,
> > +					   unsigned long npages)
> > +{
> > +	unsigned long i, j;
> > +	struct drm_pagemap *dpagemap = range->dpagemap;
> > +	struct device *dev = gpusvm->drm->dev;
> > +
> > +	lockdep_assert_held(&gpusvm->notifier_lock);
> > +
> > +	if (range->flags.has_dma_mapping) {
> > +		for (i = 0, j = 0; i < npages; j++) {
> > +			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
> > +
> > +			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
> > +				dma_unmap_page(dev,
> > +					       addr->addr,
> > +					       PAGE_SIZE << addr->order,
> > +					       addr->dir);
> > +			else if (dpagemap && dpagemap->ops->unmap_dma)
> > +				dpagemap->ops->unmap_dma(dpagemap,
> > +							 dev,
> > +							 *addr);
> > +			i += 1 << addr->order;
> > +		}
> > +		range->flags.has_devmem_pages = false;
> > +		range->flags.has_dma_mapping = false;
> > +		range->dpagemap = NULL;
> > +	}
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + *
> > + * This function frees the dma address array associated with a GPU SVM range.
> > + */
> > +static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
> > +					struct drm_gpusvm_range *range)
> > +{
> > +	lockdep_assert_held(&gpusvm->notifier_lock);
> > +
> > +	if (range->dma_addr) {
> > +		kvfree(range->dma_addr);
> > +		range->dma_addr = NULL;
> > +	}
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_remove() - Remove GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range to be removed
> > + *
> > + * This function removes the specified GPU SVM range and also removes the parent
> > + * GPU SVM notifier if no more ranges remain in the notifier. The caller must
> > + * hold a lock to protect range and notifier removal.
> > + */
> > +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> > +			     struct drm_gpusvm_range *range)
> > +{
> > +	unsigned long npages = npages_in_range(range->itree.start,
> > +					       range->itree.last + 1);
> > +	struct drm_gpusvm_notifier *notifier;
> > +
> > +	drm_gpusvm_driver_lock_held(gpusvm);
> > +
> > +	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
> > +	if (WARN_ON_ONCE(!notifier))
> > +		return;
> > +
> > +	drm_gpusvm_notifier_lock(gpusvm);
> > +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> > +	drm_gpusvm_range_free_pages(gpusvm, range);
> > +	__drm_gpusvm_range_remove(notifier, range);
> > +	drm_gpusvm_notifier_unlock(gpusvm);
> > +
> > +	drm_gpusvm_range_put(range);
> > +
> > +	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
> > +		if (!notifier->flags.removed)
> > +			mmu_interval_notifier_remove(&notifier->notifier);
> > +		drm_gpusvm_notifier_remove(gpusvm, notifier);
> > +		drm_gpusvm_notifier_free(gpusvm, notifier);
> > +	}
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
> > +
> > +/**
> > + * drm_gpusvm_range_get() - Get a reference to GPU SVM range
> > + * @range: Pointer to the GPU SVM range
> > + *
> > + * This function increments the reference count of the specified GPU SVM range.
> > + *
> > + * Returns:
> > + * Pointer to the GPU SVM range.
> > + */
> > +struct drm_gpusvm_range *
> > +drm_gpusvm_range_get(struct drm_gpusvm_range *range)
> > +{
> > +	kref_get(&range->refcount);
> > +
> > +	return range;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
> > +
> > +/**
> > + * drm_gpusvm_range_destroy() - Destroy GPU SVM range
> > + * @refcount: Pointer to the reference counter embedded in the GPU SVM range
> > + *
> > + * This function destroys the specified GPU SVM range when its reference count
> > + * reaches zero. If a custom range-free function is provided, it is invoked to
> > + * free the range; otherwise, the range is deallocated using kfree().
> > + */
> > +static void drm_gpusvm_range_destroy(struct kref *refcount)
> > +{
> > +	struct drm_gpusvm_range *range =
> > +		container_of(refcount, struct drm_gpusvm_range, refcount);
> > +	struct drm_gpusvm *gpusvm = range->gpusvm;
> > +
> > +	if (gpusvm->ops->range_free)
> > +		gpusvm->ops->range_free(range);
> > +	else
> > +		kfree(range);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_put() - Put a reference to GPU SVM range
> > + * @range: Pointer to the GPU SVM range
> > + *
> > + * This function decrements the reference count of the specified GPU SVM range
> > + * and frees it when the count reaches zero.
> > + */
> > +void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
> > +{
> > +	kref_put(&range->refcount, drm_gpusvm_range_destroy);
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
> > +
> > +/**
> > + * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + *
> > + * This function determines if a GPU SVM range pages are valid. Expected be
> > + * called holding gpusvm->notifier_lock and as the last step before committing a
> > + * GPU binding. This is akin to a notifier seqno check in the HMM documentation
> > + * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
> > + * function is required for finer grained checking (i.e., per range) if pages
> > + * are valid.
> > + *
> > + * Returns:
> > + * True if GPU SVM range has valid pages, False otherwise
> > + */
> > +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> > +				  struct drm_gpusvm_range *range)
> > +{
> > +	lockdep_assert_held(&gpusvm->notifier_lock);
> > +
> > +	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
> > +
> > +/**
> > + * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + *
> > + * This function determines if a GPU SVM range pages are valid. Expected be
> > + * called without holding gpusvm->notifier_lock.
> > + *
> > + * Returns:
> > + * True if GPU SVM range has valid pages, False otherwise
> > + */
> > +static bool
> > +drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
> > +				      struct drm_gpusvm_range *range)
> > +{
> > +	bool pages_valid;
> > +
> > +	if (!range->dma_addr)
> > +		return false;
> > +
> > +	drm_gpusvm_notifier_lock(gpusvm);
> > +	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
> > +	if (!pages_valid)
> > +		drm_gpusvm_range_free_pages(gpusvm, range);
> > +	drm_gpusvm_notifier_unlock(gpusvm);
> > +
> > +	return pages_valid;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + * @ctx: GPU SVM context
> > + *
> > + * This function gets pages for a GPU SVM range and ensures they are mapped for
> > + * DMA access.
> > + *
> > + * Returns:
> > + * 0 on success, negative error code on failure.
> > + */
> > +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> > +			       struct drm_gpusvm_range *range,
> > +			       const struct drm_gpusvm_ctx *ctx)
> > +{
> > +	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
> > +	struct hmm_range hmm_range = {
> > +		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
> > +			HMM_PFN_REQ_WRITE),
> > +		.notifier = notifier,
> > +		.start = range->itree.start,
> > +		.end = range->itree.last + 1,
> > +		.dev_private_owner = gpusvm->device_private_page_owner,
> > +	};
> > +	struct mm_struct *mm = gpusvm->mm;
> > +	struct drm_gpusvm_zdd *zdd;
> > +	unsigned long timeout =
> > +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> > +	unsigned long i, j;
> > +	unsigned long npages = npages_in_range(range->itree.start,
> > +					       range->itree.last + 1);
> > +	unsigned long num_dma_mapped;
> > +	unsigned int order = 0;
> > +	unsigned long *pfns;
> > +	struct page **pages;
> > +	int err = 0;
> > +	struct dev_pagemap *pagemap;
> > +	struct drm_pagemap *dpagemap;
> > +
> > +retry:
> > +	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> > +	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
> > +		goto set_seqno;
> > +
> > +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> > +	if (!pfns)
> > +		return -ENOMEM;
> > +
> > +	if (!mmget_not_zero(mm)) {
> > +		err = -EFAULT;
> > +		goto err_free;
> > +	}
> > +
> > +	hmm_range.hmm_pfns = pfns;
> > +	while (true) {
> > +		mmap_read_lock(mm);
> > +		err = hmm_range_fault(&hmm_range);
> > +		mmap_read_unlock(mm);
> > +
> > +		if (err == -EBUSY) {
> > +			if (time_after(jiffies, timeout))
> > +				break;
> > +
> > +			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> > +			continue;
> > +		}
> > +		break;
> > +	}
> > +	mmput(mm);
> > +	if (err)
> > +		goto err_free;
> > +
> > +	pages = (struct page **)pfns;
> > +map_pages:
> > +	/*
> > +	 * Perform all dma mappings under the notifier lock to not
> > +	 * access freed pages. A notifier will either block on
> > +	 * the notifier lock or unmap dma.
> > +	 */
> > +	drm_gpusvm_notifier_lock(gpusvm);
> > +	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
> > +		drm_gpusvm_notifier_unlock(gpusvm);
> > +		kvfree(pfns);
> > +		goto retry;
> > +	}
> > +
> > +	if (!range->dma_addr) {
> > +		/* Unlock and restart mapping to allocate memory. */
> > +		drm_gpusvm_notifier_unlock(gpusvm);
> > +		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
> > +						 GFP_KERNEL);
> > +		if (!range->dma_addr) {
> > +			err = -ENOMEM;
> > +			goto err_free;
> > +		}
> > +		goto map_pages;
> > +	}
> > +
> > +	zdd = NULL;
> > +	num_dma_mapped = 0;
> > +	for (i = 0, j = 0; i < npages; ++j) {
> > +		struct page *page = hmm_pfn_to_page(pfns[i]);
> > +
> > +		order = hmm_pfn_to_map_order(pfns[i]);
> > +		if (is_device_private_page(page) || is_device_coherent_page(page)) {
> > +			if (zdd != page->zone_device_data && i > 0) {
> > +				err = -EOPNOTSUPP;
> > +				goto err_unmap;
> > +			}
> > +			zdd = page->zone_device_data;
> > +			if (pagemap != page->pgmap) {
> > +				if (i > 0) {
> > +					err = -EOPNOTSUPP;
> > +					goto err_unmap;
> > +				}
> > +
> > +				pagemap = page->pgmap;
> > +				dpagemap = zdd->devmem_allocation->dpagemap;
> > +				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
> > +					/*
> > +					 * Raced. This is not supposed to happen
> > +					 * since hmm_range_fault() should've migrated
> > +					 * this page to system.
> > +					 */
> > +					err = -EAGAIN;
> > +					goto err_unmap;
> > +				}
> > +			}
> > +			range->dma_addr[j] =
> > +				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
> > +						       page, order,
> > +						       DMA_BIDIRECTIONAL);
> > +			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
> > +				err = -EFAULT;
> > +				goto err_unmap;
> > +			}
> > +
> > +			pages[i] = page;
> > +		} else {
> > +			dma_addr_t addr;
> > +
> > +			if (is_zone_device_page(page) || zdd) {
> > +				err = -EOPNOTSUPP;
> > +				goto err_unmap;
> > +			}
> > +
> > +			addr = dma_map_page(gpusvm->drm->dev,
> > +					    page, 0,
> > +					    PAGE_SIZE << order,
> > +					    DMA_BIDIRECTIONAL);
> > +			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
> > +				err = -EFAULT;
> > +				goto err_unmap;
> > +			}
> > +
> > +			range->dma_addr[j] = drm_pagemap_dma_addr_encode
> > +				(addr, DRM_INTERCONNECT_SYSTEM, order,
> > +				 DMA_BIDIRECTIONAL);
> > +		}
> > +		i += 1 << order;
> > +		num_dma_mapped = i;
> > +	}
> > +
> > +	range->flags.has_dma_mapping = true;
> > +	if (zdd) {
> > +		range->flags.has_devmem_pages = true;
> > +		range->dpagemap = dpagemap;
> > +	}
> > +
> > +	drm_gpusvm_notifier_unlock(gpusvm);
> > +	kvfree(pfns);
> > +set_seqno:
> > +	range->notifier_seq = hmm_range.notifier_seq;
> > +
> > +	return 0;
> > +
> > +err_unmap:
> > +	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
> > +	drm_gpusvm_notifier_unlock(gpusvm);
> > +err_free:
> > +	kvfree(pfns);
> > +	if (err == -EAGAIN)
> > +		goto retry;
> > +	return err;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
> > +
> > +/**
> > + * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + * @ctx: GPU SVM context
> > + *
> > + * This function unmaps pages associated with a GPU SVM range. If @in_notifier
> > + * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
> > + * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
> > + * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
> > + * security model.
> > + */
> > +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> > +				  struct drm_gpusvm_range *range,
> > +				  const struct drm_gpusvm_ctx *ctx)
> > +{
> > +	unsigned long npages = npages_in_range(range->itree.start,
> > +					       range->itree.last + 1);
> > +
> > +	if (ctx->in_notifier)
> > +		lockdep_assert_held_write(&gpusvm->notifier_lock);
> > +	else
> > +		drm_gpusvm_notifier_lock(gpusvm);
> > +
> > +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> > +
> > +	if (!ctx->in_notifier)
> > +		drm_gpusvm_notifier_unlock(gpusvm);
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
> > +
> > +/**
> > + * drm_gpusvm_migration_unlock_put_page() - Put a migration page
> > + * @page: Pointer to the page to put
> > + *
> > + * This function unlocks and puts a page.
> > + */
> > +static void drm_gpusvm_migration_unlock_put_page(struct page *page)
> > +{
> > +	unlock_page(page);
> > +	put_page(page);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
> > + * @npages: Number of pages
> > + * @migrate_pfn: Array of migrate page frame numbers
> > + *
> > + * This function unlocks and puts an array of pages.
> > + */
> > +static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
> > +						  unsigned long *migrate_pfn)
> > +{
> > +	unsigned long i;
> > +
> > +	for (i = 0; i < npages; ++i) {
> > +		struct page *page;
> > +
> > +		if (!migrate_pfn[i])
> > +			continue;
> > +
> > +		page = migrate_pfn_to_page(migrate_pfn[i]);
> > +		drm_gpusvm_migration_unlock_put_page(page);
> > +		migrate_pfn[i] = 0;
> > +	}
> > +}
> > +
> > +/**
> > + * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
> > + * @page: Pointer to the page
> > + * @zdd: Pointer to the GPU SVM zone device data
> > + *
> > + * This function associates the given page with the specified GPU SVM zone
> > + * device data and initializes it for zone device usage.
> > + */
> > +static void drm_gpusvm_get_devmem_page(struct page *page,
> > +				     struct drm_gpusvm_zdd *zdd)
> > +{
> > +	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
> > +	zone_device_page_init(page);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
> > + * @dev: The device for which the pages are being mapped
> > + * @dma_addr: Array to store DMA addresses corresponding to mapped pages
> > + * @migrate_pfn: Array of migrate page frame numbers to map
> > + * @npages: Number of pages to map
> > + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> > + *
> > + * This function maps pages of memory for migration usage in GPU SVM. It
> > + * iterates over each page frame number provided in @migrate_pfn, maps the
> > + * corresponding page, and stores the DMA address in the provided @dma_addr
> > + * array.
> > + *
> > + * Return: 0 on success, -EFAULT if an error occurs during mapping.
> > + */
> > +static int drm_gpusvm_migrate_map_pages(struct device *dev,
> > +					dma_addr_t *dma_addr,
> > +					unsigned long *migrate_pfn,
> > +					unsigned long npages,
> > +					enum dma_data_direction dir)
> > +{
> > +	unsigned long i;
> > +
> > +	for (i = 0; i < npages; ++i) {
> > +		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
> > +
> > +		if (!page)
> > +			continue;
> > +
> > +		if (WARN_ON_ONCE(is_zone_device_page(page)))
> > +			return -EFAULT;
> > +
> > +		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
> > +		if (dma_mapping_error(dev, dma_addr[i]))
> > +			return -EFAULT;
> > +	}
> > +
> > +	return 0;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
> > + * @dev: The device for which the pages were mapped
> > + * @dma_addr: Array of DMA addresses corresponding to mapped pages
> > + * @npages: Number of pages to unmap
> > + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> > + *
> > + * This function unmaps previously mapped pages of memory for GPU Shared Virtual
> > + * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
> > + * if it's valid and not already unmapped, and unmaps the corresponding page.
> > + */
> > +static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
> > +					   dma_addr_t *dma_addr,
> > +					   unsigned long npages,
> > +					   enum dma_data_direction dir)
> > +{
> > +	unsigned long i;
> > +
> > +	for (i = 0; i < npages; ++i) {
> > +		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
> > +			continue;
> > +
> > +		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
> > +	}
> > +}
> > +
> > +/**
> > + * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range structure
> > + * @devmem_allocation: Pointer to the device memory allocation. The caller
> > + *                     should hold a reference to the device memory allocation,
> > + *                     which should be dropped via ops->devmem_release or upon
> > + *                     the failure of this function.
> > + * @ctx: GPU SVM context
> > + *
> > + * This function migrates the specified GPU SVM range to device memory. It performs the
> > + * necessary setup and invokes the driver-specific operations for migration to
> > + * device memory. Upon successful return, @devmem_allocation can safely reference @range
> > + * until ops->devmem_release is called which only upon successful return.
> > + *
> > + * Returns:
> > + * 0 on success, negative error code on failure.
> > + */
> > +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> > +				 struct drm_gpusvm_range *range,
> > +				 struct drm_gpusvm_devmem *devmem_allocation,
> > +				 const struct drm_gpusvm_ctx *ctx)
> > +{
> > +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> > +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> > +	struct migrate_vma migrate = {
> > +		.start		= start,
> > +		.end		= end,
> > +		.pgmap_owner	= gpusvm->device_private_page_owner,
> > +		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
> > +	};
> > +	struct mm_struct *mm = gpusvm->mm;
> > +	unsigned long i, npages = npages_in_range(start, end);
> > +	struct vm_area_struct *vas;
> > +	struct drm_gpusvm_zdd *zdd = NULL;
> > +	struct page **pages;
> > +	dma_addr_t *dma_addr;
> > +	void *buf;
> > +	int err;
> > +
> > +	if (!range->flags.migrate_devmem)
> > +		return -EINVAL;
> > +
> > +	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
> > +	    !ops->copy_to_ram)
> > +		return -EOPNOTSUPP;
> > +
> > +	if (!mmget_not_zero(mm)) {
> > +		err = -EFAULT;
> > +		goto err_out;
> > +	}
> > +	mmap_read_lock(mm);
> > +
> > +	vas = vma_lookup(mm, start);
> > +	if (!vas) {
> > +		err = -ENOENT;
> > +		goto err_mmunlock;
> > +	}
> > +
> > +	if (end > vas->vm_end || start < vas->vm_start) {
> > +		err = -EINVAL;
> > +		goto err_mmunlock;
> > +	}
> > +
> > +	if (!vma_is_anonymous(vas)) {
> > +		err = -EBUSY;
> > +		goto err_mmunlock;
> > +	}
> > +
> > +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> > +		       sizeof(*pages), GFP_KERNEL);
> > +	if (!buf) {
> > +		err = -ENOMEM;
> > +		goto err_mmunlock;
> > +	}
> > +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> > +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> > +
> > +	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
> > +	if (!zdd) {
> > +		err = -ENOMEM;
> > +		goto err_free;
> > +	}
> > +
> > +	migrate.vma = vas;
> > +	migrate.src = buf;
> > +	migrate.dst = migrate.src + npages;
> > +
> > +	err = migrate_vma_setup(&migrate);
> > +	if (err)
> > +		goto err_free;
> > +
> > +	if (!migrate.cpages) {
> > +		err = -EFAULT;
> > +		goto err_free;
> > +	}
> > +
> > +	if (migrate.cpages != npages) {
> > +		err = -EBUSY;
> > +		goto err_finalize;
> > +	}
> > +
> > +	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> > +					   migrate.src, npages, DMA_TO_DEVICE);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +	for (i = 0; i < npages; ++i) {
> > +		struct page *page = pfn_to_page(migrate.dst[i]);
> > +
> > +		pages[i] = page;
> > +		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
> > +		drm_gpusvm_get_devmem_page(page, zdd);
> > +	}
> > +
> > +	err = ops->copy_to_devmem(pages, dma_addr, npages);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +	/* Upon success bind devmem allocation to range and zdd */
> > +	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
> > +
> > +err_finalize:
> > +	if (err)
> > +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> > +	migrate_vma_pages(&migrate);
> > +	migrate_vma_finalize(&migrate);
> > +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> > +				       DMA_TO_DEVICE);
> > +err_free:
> > +	if (zdd)
> > +		drm_gpusvm_zdd_put(zdd);
> > +	kvfree(buf);
> > +err_mmunlock:
> > +	mmap_read_unlock(mm);
> > +	mmput(mm);
> > +err_out:
> > +	return err;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
> > +
> > +/**
> > + * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
> > + * @vas: Pointer to the VM area structure, can be NULL
> > + * @fault_page: Fault page
> > + * @npages: Number of pages to populate
> > + * @mpages: Number of pages to migrate
> > + * @src_mpfn: Source array of migrate PFNs
> > + * @mpfn: Array of migrate PFNs to populate
> > + * @addr: Start address for PFN allocation
> > + *
> > + * This function populates the RAM migrate page frame numbers (PFNs) for the
> > + * specified VM area structure. It allocates and locks pages in the VM area for
> > + * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
> > + * alloc_page for allocation.
> > + *
> > + * Returns:
> > + * 0 on success, negative error code on failure.
> > + */
> > +static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> > +					       struct page *fault_page,
> > +					       unsigned long npages,
> > +					       unsigned long *mpages,
> > +					       unsigned long *src_mpfn,
> > +					       unsigned long *mpfn,
> > +					       unsigned long addr)
> > +{
> > +	unsigned long i;
> > +
> > +	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
> > +		struct page *page, *src_page;
> > +
> > +		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
> > +			continue;
> > +
> > +		if (fault_page) {
> > +			src_page = migrate_pfn_to_page(src_mpfn[i]);
> > +			if (src_page->zone_device_data !=
> > +			    fault_page->zone_device_data)
> > +				continue;
> > +		}
> > +
> > +		if (vas)
> > +			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
> > +		else
> > +			page = alloc_page(GFP_HIGHUSER);
> > +
> > +		if (!page)
> > +			goto free_pages;
> > +
> > +		mpfn[i] = migrate_pfn(page_to_pfn(page));
> > +	}
> > +
> > +	for (i = 0; i < npages; ++i) {
> > +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> > +
> > +		if (!page)
> > +			continue;
> > +
> > +		WARN_ON_ONCE(!trylock_page(page));
> > +		++*mpages;
> > +	}
> > +
> > +	return 0;
> > +
> > +free_pages:
> > +	for (i = 0; i < npages; ++i) {
> > +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> > +
> > +		if (!page)
> > +			continue;
> > +
> > +		put_page(page);
> > +		mpfn[i] = 0;
> > +	}
> > +	return -ENOMEM;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
> > + * @devmem_allocation: Pointer to the device memory allocation
> > + *
> > + * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
> > + * migration done via migrate_device_* functions.
> > + *
> > + * Returns:
> > + * 0 on success, negative error code on failure.
> > + */
> > +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
> > +{
> > +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> > +	unsigned long npages, mpages = 0;
> > +	struct page **pages;
> > +	unsigned long *src, *dst;
> > +	dma_addr_t *dma_addr;
> > +	void *buf;
> > +	int i, err = 0;
> > +
> > +	npages = devmem_allocation->size >> PAGE_SHIFT;
> > +
> > +retry:
> > +	if (!mmget_not_zero(devmem_allocation->mm))
> > +		return -EFAULT;
> > +
> > +	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
> > +		       sizeof(*pages), GFP_KERNEL);
> > +	if (!buf) {
> > +		err = -ENOMEM;
> > +		goto err_out;
> > +	}
> > +	src = buf;
> > +	dst = buf + (sizeof(*src) * npages);
> > +	dma_addr = buf + (2 * sizeof(*src) * npages);
> > +	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
> > +
> > +	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
> > +	if (err)
> > +		goto err_free;
> > +
> > +	err = migrate_device_pfns(src, npages);
> > +	if (err)
> > +		goto err_free;
> > +
> > +	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
> > +						  src, dst, 0);
> > +	if (err || !mpages)
> > +		goto err_finalize;
> > +
> > +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> > +					   dst, npages, DMA_FROM_DEVICE);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +	for (i = 0; i < npages; ++i)
> > +		pages[i] = migrate_pfn_to_page(src[i]);
> > +
> > +	err = ops->copy_to_ram(pages, dma_addr, npages);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +err_finalize:
> > +	if (err)
> > +		drm_gpusvm_migration_unlock_put_pages(npages, dst);
> > +	migrate_device_pages(src, dst, npages);
> > +	migrate_device_finalize(src, dst, npages);
> > +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> > +				       DMA_FROM_DEVICE);
> > +err_free:
> > +	kvfree(buf);
> > +err_out:
> > +	mmput_async(devmem_allocation->mm);
> > +	if (!err && !READ_ONCE(devmem_allocation->detached)) {
> > +		cond_resched();
> > +		goto retry;
> > +	}
> > +
> > +	return err;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
> > +
> > +/**
> > + * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
> > + * @vas: Pointer to the VM area structure
> > + * @device_private_page_owner: Device private pages owner
> > + * @page: Pointer to the page for fault handling (can be NULL)
> > + * @fault_addr: Fault address
> > + * @size: Size of migration
> > + *
> > + * This internal function performs the migration of the specified GPU SVM range
> > + * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
> > + * invokes the driver-specific operations for migration to RAM.
> > + *
> > + * Returns:
> > + * 0 on success, negative error code on failure.
> > + */
> > +static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
> > +				       void *device_private_page_owner,
> > +				       struct page *page,
> > +				       unsigned long fault_addr,
> > +				       unsigned long size)
> > +{
> > +	struct migrate_vma migrate = {
> > +		.vma		= vas,
> > +		.pgmap_owner	= device_private_page_owner,
> > +		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
> > +			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
> > +		.fault_page	= page,
> > +	};
> > +	struct drm_gpusvm_zdd *zdd;
> > +	const struct drm_gpusvm_devmem_ops *ops;
> > +	struct device *dev;
> > +	unsigned long npages, mpages = 0;
> > +	struct page **pages;
> > +	dma_addr_t *dma_addr;
> > +	unsigned long start, end;
> > +	void *buf;
> > +	int i, err = 0;
> > +
> > +	start = ALIGN_DOWN(fault_addr, size);
> > +	end = ALIGN(fault_addr + 1, size);
> > +
> > +	/* Corner where VMA area struct has been partially unmapped */
> > +	if (start < vas->vm_start)
> > +		start = vas->vm_start;
> > +	if (end > vas->vm_end)
> > +		end = vas->vm_end;
> > +
> > +	migrate.start = start;
> > +	migrate.end = end;
> > +	npages = npages_in_range(start, end);
> > +
> > +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> > +		       sizeof(*pages), GFP_KERNEL);
> > +	if (!buf) {
> > +		err = -ENOMEM;
> > +		goto err_out;
> > +	}
> > +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> > +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> > +
> > +	migrate.vma = vas;
> > +	migrate.src = buf;
> > +	migrate.dst = migrate.src + npages;
> > +
> > +	err = migrate_vma_setup(&migrate);
> > +	if (err)
> > +		goto err_free;
> > +
> > +	/* Raced with another CPU fault, nothing to do */
> > +	if (!migrate.cpages)
> > +		goto err_free;
> > +
> > +	if (!page) {
> > +		for (i = 0; i < npages; ++i) {
> > +			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
> > +				continue;
> > +
> > +			page = migrate_pfn_to_page(migrate.src[i]);
> > +			break;
> > +		}
> > +
> > +		if (!page)
> > +			goto err_finalize;
> If you goto to err_finalize through this error case, the uninitialized dev
> will be passed to drm_gpusvm_migrate_unmap_pages().
> 

Yep. The code works as drm_gpusvm_migrate_unmap_pages loops over the dma
mappings which will not exist. I suppose I could short circuit this on
if (dev) if you prefer.

Matt

> Br,
> G.G.
> > +	}
> > +	zdd = page->zone_device_data;
> > +	ops = zdd->devmem_allocation->ops;
> > +	dev = zdd->devmem_allocation->dev;
> > +
> > +	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
> > +						  migrate.src, migrate.dst,
> > +						  start);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
> > +					   DMA_FROM_DEVICE);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +	for (i = 0; i < npages; ++i)
> > +		pages[i] = migrate_pfn_to_page(migrate.src[i]);
> > +
> > +	err = ops->copy_to_ram(pages, dma_addr, npages);
> > +	if (err)
> > +		goto err_finalize;
> > +
> > +err_finalize:
> > +	if (err)
> > +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> > +	migrate_vma_pages(&migrate);
> > +	migrate_vma_finalize(&migrate);
> > +	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
> > +				       DMA_FROM_DEVICE);
> > +err_free:
> > +	kvfree(buf);
> > +err_out:
> > +
> > +	return err;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_range_evict() - Evict GPU SVM range
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @range: Pointer to the GPU SVM range to be removed
> > + *
> > + * This function evicts the specified GPU SVM range.
> > + */
> > +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> > +			    struct drm_gpusvm_range *range)
> > +{
> > +	struct mm_struct *mm = gpusvm->mm;
> > +	struct vm_area_struct *vas;
> > +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> > +
> > +	if (!mmget_not_zero(mm))
> > +		return;
> > +
> > +	mmap_read_lock(mm);
> > +
> > +	/*
> > +	 * Loop required to find all VMAs for the corner case when partial
> > +	 * unmappings have occurred.
> > +	 */
> > +again:
> > +	vas = find_vma(mm, start);
> > +	if (!vas)
> > +		goto unlock;
> > +
> > +	if (end <= vas->vm_start || start >= vas->vm_end)
> > +		goto unlock;
> > +
> > +	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
> > +				    NULL, start,
> > +				    range->itree.last - range->itree.start + 1);
> > +	if (vas->vm_end < end) {
> > +		start = vas->vm_end;
> > +		goto again;
> > +	}
> > +unlock:
> > +	mmap_read_unlock(mm);
> > +	mmput(mm);
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
> > +
> > +/**
> > + * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
> > + * @page: Pointer to the page
> > + *
> > + * This function is a callback used to put the GPU SVM zone device data
> > + * associated with a page when it is being released.
> > + */
> > +static void drm_gpusvm_page_free(struct page *page)
> > +{
> > +	drm_gpusvm_zdd_put(page->zone_device_data);
> > +}
> > +
> > +/**
> > + * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
> > + * @vmf: Pointer to the fault information structure
> > + *
> > + * This function is a page fault handler used to migrate a GPU SVM range to RAM.
> > + * It retrieves the GPU SVM range information from the faulting page and invokes
> > + * the internal migration function to migrate the range back to RAM.
> > + *
> > + * Returns:
> > + * VM_FAULT_SIGBUS on failure, 0 on success.
> > + */
> > +static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
> > +{
> > +	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
> > +	int err;
> > +
> > +	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
> > +					  zdd->device_private_page_owner,
> > +					  vmf->page, vmf->address,
> > +					  zdd->devmem_allocation->size);
> > +
> > +	return err ? VM_FAULT_SIGBUS : 0;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
> > + */
> > +static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
> > +	.page_free = drm_gpusvm_page_free,
> > +	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
> > +};
> > +
> > +/**
> > + * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
> > + *
> > + * Returns:
> > + * Pointer to the GPU SVM device page map operations structure.
> > + */
> > +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
> > +{
> > +	return &drm_gpusvm_pagemap_ops;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
> > +
> > +/**
> > + * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
> > + * @gpusvm: Pointer to the GPU SVM structure.
> > + * @start: Start address
> > + * @end: End address
> > + *
> > + * Returns:
> > + * True if GPU SVM has mapping, False otherwise
> > + */
> > +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> > +			    unsigned long end)
> > +{
> > +	struct drm_gpusvm_notifier *notifier;
> > +
> > +	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
> > +		struct drm_gpusvm_range *range = NULL;
> > +
> > +		drm_gpusvm_for_each_range(range, notifier, start, end)
> > +			return true;
> > +	}
> > +
> > +	return false;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
> > +
> > +/**
> > + * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
> > + * @range: Pointer to the GPU SVM range structure.
> > + * @mmu_range: Pointer to the MMU notifier range structure.
> > + *
> > + * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
> > + * if the range partially falls within the provided MMU notifier range.
> > + */
> > +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> > +				   const struct mmu_notifier_range *mmu_range)
> > +{
> > +	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
> > +
> > +	range->flags.unmapped = true;
> > +	if (range->itree.start < mmu_range->start ||
> > +	    range->itree.last + 1 > mmu_range->end)
> > +		range->flags.partial_unmap = true;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
> > +
> > +/**
> > + * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
> > + *
> > + * @dev: Pointer to the device structure which device memory allocation belongs to
> > + * @mm: Pointer to the mm_struct for the address space
> > + * @ops: Pointer to the operations structure for GPU SVM device memory
> > + * @dpagemap: The struct drm_pagemap we're allocating from.
> > + * @size: Size of device memory allocation
> > + */
> > +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> > +			    struct device *dev, struct mm_struct *mm,
> > +			    const struct drm_gpusvm_devmem_ops *ops,
> > +			    struct drm_pagemap *dpagemap, size_t size)
> > +{
> > +	devmem_allocation->dev = dev;
> > +	devmem_allocation->mm = mm;
> > +	devmem_allocation->ops = ops;
> > +	devmem_allocation->dpagemap = dpagemap;
> > +	devmem_allocation->size = size;
> > +}
> > +EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
> > +
> > +MODULE_DESCRIPTION("DRM GPUSVM");
> > +MODULE_LICENSE("GPL");
> > diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> > new file mode 100644
> > index 000000000000..2fb6980d0cf1
> > --- /dev/null
> > +++ b/include/drm/drm_gpusvm.h
> > @@ -0,0 +1,445 @@
> > +/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
> > +/*
> > + * Copyright © 2024 Intel Corporation
> > + */
> > +
> > +#ifndef __DRM_GPUSVM_H__
> > +#define __DRM_GPUSVM_H__
> > +
> > +#include <linux/kref.h>
> > +#include <linux/interval_tree.h>
> > +#include <linux/mmu_notifier.h>
> > +
> > +struct dev_pagemap_ops;
> > +struct drm_device;
> > +struct drm_gpusvm;
> > +struct drm_gpusvm_notifier;
> > +struct drm_gpusvm_ops;
> > +struct drm_gpusvm_range;
> > +struct drm_gpusvm_devmem;
> > +struct drm_pagemap;
> > +struct drm_pagemap_dma_addr;
> > +
> > +/**
> > + * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
> > + *
> > + * This structure defines the operations for GPU Shared Virtual Memory (SVM)
> > + * device memory. These operations are provided by the GPU driver to manage device memory
> > + * allocations and perform operations such as migration between device memory and system
> > + * RAM.
> > + */
> > +struct drm_gpusvm_devmem_ops {
> > +	/**
> > +	 * @devmem_release: Release device memory allocation (optional)
> > +	 * @devmem_allocation: device memory allocation
> > +	 *
> > +	 * Release device memory allocation and drop a reference to device
> > +	 * memory allocation.
> > +	 */
> > +	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
> > +
> > +	/**
> > +	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
> > +	 * @devmem_allocation: device memory allocation
> > +	 * @npages: Number of pages to populate
> > +	 * @pfn: Array of page frame numbers to populate
> > +	 *
> > +	 * Populate device memory page frame numbers (PFN).
> > +	 *
> > +	 * Returns:
> > +	 * 0 on success, a negative error code on failure.
> > +	 */
> > +	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
> > +				   unsigned long npages, unsigned long *pfn);
> > +
> > +	/**
> > +	 * @copy_to_devmem: Copy to device memory (required for migration)
> > +	 * @pages: Pointer to array of device memory pages (destination)
> > +	 * @dma_addr: Pointer to array of DMA addresses (source)
> > +	 * @npages: Number of pages to copy
> > +	 *
> > +	 * Copy pages to device memory.
> > +	 *
> > +	 * Returns:
> > +	 * 0 on success, a negative error code on failure.
> > +	 */
> > +	int (*copy_to_devmem)(struct page **pages,
> > +			      dma_addr_t *dma_addr,
> > +			      unsigned long npages);
> > +
> > +	/**
> > +	 * @copy_to_ram: Copy to system RAM (required for migration)
> > +	 * @pages: Pointer to array of device memory pages (source)
> > +	 * @dma_addr: Pointer to array of DMA addresses (destination)
> > +	 * @npages: Number of pages to copy
> > +	 *
> > +	 * Copy pages to system RAM.
> > +	 *
> > +	 * Returns:
> > +	 * 0 on success, a negative error code on failure.
> > +	 */
> > +	int (*copy_to_ram)(struct page **pages,
> > +			   dma_addr_t *dma_addr,
> > +			   unsigned long npages);
> > +};
> > +
> > +/**
> > + * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
> > + *
> > + * @dev: Pointer to the device structure which device memory allocation belongs to
> > + * @mm: Pointer to the mm_struct for the address space
> > + * @ops: Pointer to the operations structure for GPU SVM device memory
> > + * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
> > + * @size: Size of device memory allocation
> > + * @detached: device memory allocations is detached from device pages
> > + */
> > +struct drm_gpusvm_devmem {
> > +	struct device *dev;
> > +	struct mm_struct *mm;
> > +	const struct drm_gpusvm_devmem_ops *ops;
> > +	struct drm_pagemap *dpagemap;
> > +	size_t size;
> > +	bool detached;
> > +};
> > +
> > +/**
> > + * struct drm_gpusvm_ops - Operations structure for GPU SVM
> > + *
> > + * This structure defines the operations for GPU Shared Virtual Memory (SVM).
> > + * These operations are provided by the GPU driver to manage SVM ranges and
> > + * notifiers.
> > + */
> > +struct drm_gpusvm_ops {
> > +	/**
> > +	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
> > +	 *
> > +	 * Allocate a GPU SVM notifier.
> > +	 *
> > +	 * Returns:
> > +	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
> > +	 */
> > +	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
> > +
> > +	/**
> > +	 * @notifier_free: Free a GPU SVM notifier (optional)
> > +	 * @notifier: Pointer to the GPU SVM notifier to be freed
> > +	 *
> > +	 * Free a GPU SVM notifier.
> > +	 */
> > +	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
> > +
> > +	/**
> > +	 * @range_alloc: Allocate a GPU SVM range (optional)
> > +	 * @gpusvm: Pointer to the GPU SVM
> > +	 *
> > +	 * Allocate a GPU SVM range.
> > +	 *
> > +	 * Returns:
> > +	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
> > +	 */
> > +	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
> > +
> > +	/**
> > +	 * @range_free: Free a GPU SVM range (optional)
> > +	 * @range: Pointer to the GPU SVM range to be freed
> > +	 *
> > +	 * Free a GPU SVM range.
> > +	 */
> > +	void (*range_free)(struct drm_gpusvm_range *range);
> > +
> > +	/**
> > +	 * @invalidate: Invalidate GPU SVM notifier (required)
> > +	 * @gpusvm: Pointer to the GPU SVM
> > +	 * @notifier: Pointer to the GPU SVM notifier
> > +	 * @mmu_range: Pointer to the mmu_notifier_range structure
> > +	 *
> > +	 * Invalidate the GPU page tables. It can safely walk the notifier range
> > +	 * RB tree/list in this function. Called while holding the notifier lock.
> > +	 */
> > +	void (*invalidate)(struct drm_gpusvm *gpusvm,
> > +			   struct drm_gpusvm_notifier *notifier,
> > +			   const struct mmu_notifier_range *mmu_range);
> > +};
> > +
> > +/**
> > + * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
> > + *
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: MMU interval notifier
> > + * @itree: Interval tree node for the notifier (inserted in GPU SVM)
> > + * @entry: List entry to fast interval tree traversal
> > + * @root: Cached root node of the RB tree containing ranges
> > + * @range_list: List head containing of ranges in the same order they appear in
> > + *              interval tree. This is useful to keep iterating ranges while
> > + *              doing modifications to RB tree.
> > + * @flags.removed: Flag indicating whether the MMU interval notifier has been
> > + *                 removed
> > + *
> > + * This structure represents a GPU SVM notifier.
> > + */
> > +struct drm_gpusvm_notifier {
> > +	struct drm_gpusvm *gpusvm;
> > +	struct mmu_interval_notifier notifier;
> > +	struct interval_tree_node itree;
> > +	struct list_head entry;
> > +	struct rb_root_cached root;
> > +	struct list_head range_list;
> > +	struct {
> > +		u32 removed : 1;
> > +	} flags;
> > +};
> > +
> > +/**
> > + * struct drm_gpusvm_range - Structure representing a GPU SVM range
> > + *
> > + * @gpusvm: Pointer to the GPU SVM structure
> > + * @notifier: Pointer to the GPU SVM notifier
> > + * @refcount: Reference count for the range
> > + * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
> > + * @entry: List entry to fast interval tree traversal
> > + * @notifier_seq: Notifier sequence number of the range's pages
> > + * @dma_addr: DMA address array
> > + * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
> > + *            Note this is assuming only one drm_pagemap per range is allowed.
> > + * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
> > + * @flags.unmapped: Flag indicating if the range has been unmapped
> > + * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
> > + * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
> > + * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
> > + *
> > + * This structure represents a GPU SVM range used for tracking memory ranges
> > + * mapped in a DRM device.
> > + */
> > +struct drm_gpusvm_range {
> > +	struct drm_gpusvm *gpusvm;
> > +	struct drm_gpusvm_notifier *notifier;
> > +	struct kref refcount;
> > +	struct interval_tree_node itree;
> > +	struct list_head entry;
> > +	unsigned long notifier_seq;
> > +	struct drm_pagemap_dma_addr *dma_addr;
> > +	struct drm_pagemap *dpagemap;
> > +	struct {
> > +		/* All flags below must be set upon creation */
> > +		u16 migrate_devmem : 1;
> > +		/* All flags below must be set / cleared under notifier lock */
> > +		u16 unmapped : 1;
> > +		u16 partial_unmap : 1;
> > +		u16 has_devmem_pages : 1;
> > +		u16 has_dma_mapping : 1;
> > +	} flags;
> > +};
> > +
> > +/**
> > + * struct drm_gpusvm - GPU SVM structure
> > + *
> > + * @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
> > + * @ops: Pointer to the operations structure for GPU SVM
> > + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> > + *               Entries should be powers of 2 in descending order.
> > + * @num_chunks: Number of chunks
> > + * @notifier_lock: Read-write semaphore for protecting notifier operations
> > + * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
> > + * @notifier_list: list head containing of notifiers in the same order they
> > + *                 appear in interval tree. This is useful to keep iterating
> > + *                 notifiers while doing modifications to RB tree.
> > + *
> > + * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
> > + * memory ranges mapped in a DRM (Direct Rendering Manager) device.
> > + *
> > + * No reference counting is provided, as this is expected to be embedded in the
> > + * driver VM structure along with the struct drm_gpuvm, which handles reference
> > + * counting.
> > + */
> > +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;
> > +	const struct drm_gpusvm_ops *ops;
> > +	const unsigned long *chunk_sizes;
> > +	int num_chunks;
> > +	struct rw_semaphore notifier_lock;
> > +	struct rb_root_cached root;
> > +	struct list_head notifier_list;
> > +#ifdef CONFIG_LOCKDEP
> > +	/**
> > +	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
> > +	 * drm_gpusvm_range_remove with a driver provided lock.
> > +	 */
> > +	struct lockdep_map *lock_dep_map;
> > +#endif
> > +};
> > +
> > +/**
> > + * struct drm_gpusvm_ctx - DRM GPU SVM context
> > + *
> > + * @check_pages_threshold: Check CPU pages for present if chunk is less than or
> > + *                         equal to threshold. If not present, reduce chunk
> > + *                         size.
> > + * @in_notifier: entering from a MMU notifier
> > + * @read_only: operating on read-only memory
> > + * @devmem_possible: possible to use device memory
> > + *
> > + * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> > + */
> > +struct drm_gpusvm_ctx {
> > +	unsigned long check_pages_threshold;
> > +	unsigned int in_notifier :1;
> > +	unsigned int read_only :1;
> > +	unsigned int devmem_possible :1;
> > +};
> > +
> > +int drm_gpusvm_init(struct drm_gpusvm *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,
> > +		    const struct drm_gpusvm_ops *ops,
> > +		    const unsigned long *chunk_sizes, int num_chunks);
> > +
> > +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
> > +
> > +void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
> > +
> > +struct drm_gpusvm_range *
> > +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> > +				unsigned long fault_addr,
> > +				unsigned long gpuva_start,
> > +				unsigned long gpuva_end,
> > +				const struct drm_gpusvm_ctx *ctx);
> > +
> > +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> > +			     struct drm_gpusvm_range *range);
> > +
> > +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> > +			    struct drm_gpusvm_range *range);
> > +
> > +struct drm_gpusvm_range *
> > +drm_gpusvm_range_get(struct drm_gpusvm_range *range);
> > +
> > +void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
> > +
> > +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> > +				  struct drm_gpusvm_range *range);
> > +
> > +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> > +			       struct drm_gpusvm_range *range,
> > +			       const struct drm_gpusvm_ctx *ctx);
> > +
> > +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> > +				  struct drm_gpusvm_range *range,
> > +				  const struct drm_gpusvm_ctx *ctx);
> > +
> > +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> > +				 struct drm_gpusvm_range *range,
> > +				 struct drm_gpusvm_devmem *devmem_allocation,
> > +				 const struct drm_gpusvm_ctx *ctx);
> > +
> > +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
> > +
> > +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
> > +
> > +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> > +			    unsigned long end);
> > +
> > +struct drm_gpusvm_range *
> > +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> > +		      unsigned long end);
> > +
> > +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> > +				   const struct mmu_notifier_range *mmu_range);
> > +
> > +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> > +			    struct device *dev, struct mm_struct *mm,
> > +			    const struct drm_gpusvm_devmem_ops *ops,
> > +			    struct drm_pagemap *dpagemap, size_t size);
> > +
> > +#ifdef CONFIG_LOCKDEP
> > +/**
> > + * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
> > + * @gpusvm: Pointer to the GPU SVM structure.
> > + * @lock: the lock used to protect the gpuva list. The locking primitive
> > + * must contain a dep_map field.
> > + *
> > + * Call this to annotate drm_gpusvm_range_find_or_insert and
> > + * drm_gpusvm_range_remove.
> > + */
> > +#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
> > +	do { \
> > +		if (!WARN((gpusvm)->lock_dep_map, \
> > +			  "GPUSVM range lock should be set only once."))\
> > +			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
> > +	} while (0)
> > +#define drm_gpusvm_driver_lock_held(gpusvm) \
> > +	do { \
> > +		if ((gpusvm)->lock_dep_map)	\
> > +			lock_is_held((gpusvm)->lock_dep_map);	\
> > +	} while (0)
> > +#else
> > +#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
> > +#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
> > +#endif
> > +
> > +/**
> > + * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
> > + * @gpusvm__: Pointer to the GPU SVM structure.
> > + *
> > + * Abstract client usage GPU SVM notifier lock, take lock
> > + */
> > +#define drm_gpusvm_notifier_lock(gpusvm__)	\
> > +	down_read(&(gpusvm__)->notifier_lock)
> > +
> > +/**
> > + * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
> > + * @gpusvm__: Pointer to the GPU SVM structure.
> > + *
> > + * Abstract client usage GPU SVM notifier lock, drop lock
> > + */
> > +#define drm_gpusvm_notifier_unlock(gpusvm__)	\
> > +	up_read(&(gpusvm__)->notifier_lock)
> > +
> > +/**
> > + * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
> > + * @range: a pointer to the current GPU SVM range
> > + *
> > + * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
> > + *         current range is the last one or if the input range is NULL.
> > + */
> > +static inline struct drm_gpusvm_range *
> > +__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
> > +{
> > +	if (range && !list_is_last(&range->entry,
> > +				   &range->notifier->range_list))
> > +		return list_next_entry(range, entry);
> > +
> > +	return NULL;
> > +}
> > +
> > +/**
> > + * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
> > + * @range__: Iterator variable for the ranges. If set, it indicates the start of
> > + *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
> > + * @notifier__: Pointer to the GPU SVM notifier
> > + * @start__: Start address of the range
> > + * @end__: End address of the range
> > + *
> > + * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
> > + * to use while holding the driver SVM lock or the notifier lock.
> > + */
> > +#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
> > +	for ((range__) = (range__) ?:					\
> > +	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
> > +	     (range__) && (range__->itree.start < (end__));		\
> > +	     (range__) = __drm_gpusvm_range_next(range__))
> > +
> > +#endif /* __DRM_GPUSVM_H__ */
> 

Re: [PATCH v3 00/30] Introduce GPU SVM and Xe SVM implementation

[Matthew Brost]

On Fri, Jan 17, 2025 at 11:47:41AM +0200, Gwan-gyeong Mun wrote:
> Hi,
> This kernel oops, which I reported before, was caused by my incorrect
> modification (incorrect applying of review comments) of this patch
> "[v3,19/30] drm/xe: Add SVM device memory mirroring"
> ( the kernel oops occurred because the xe_drm_pagemap_map_dma() and
> xe_devm_add() functions were built in the form of empty functions. )
> 
> This issue disappeared after proper patch modifications were applied.
> So please ignore the previously reported this kernel oops.
> 

My post unfortunately had some bugs masked by Kconfig issue Himal
pointed out. If you want to test of this code I suggest you pull this
branch [1] - it should be stable with all fixes I mention on this rev.

Matt

[1] https://gitlab.freedesktop.org/mbrost/xe-kernel-driver-svm-stable-1-13-25/-/tree/stable-1.13.25?ref_type=heads

> Br,
> 
> G.G.
> 
> On 1/7/25 2:19 PM, Gwan-gyeong Mun wrote:
> > Hi Matthew Brost,
> > 
> > After applying this patch series and the following to the latest drm-
> > tip, while testing[1] with the mentioned IGT, I faced a kernel oops[3].
> > It makes prevent progressing of the mentioned igt tests.
> > Could you please check the following oops log?
> > 
> > (1) apply comments of "[v3,05/30] drm/gpusvm: Add support for GPU Shared
> > Virtual Memory"
> > (2) apply comments of "[v3,15/30] drm/xe: Add unbind to SVM garbage
> > collector"
> > (3) drop "[v3,27/30] drm/xe: Basic SVM BO eviction" patch
> > 
> > The kernel config used, the entire dmesg, and detailed information can
> > be found in [2].
> > 
> > br,
> > 
> > G.G.
> > 
> > [1] used igt command: xe_exec_system_allocator --run-subtest once-malloc
> > [2] https://gitlab.freedesktop.org/elongbug/drm-tip/-/snippets/7823
> > 
> > [3] kernel oops dmesg
> > [   51.365230] Console: switching to colour VGA+ 80x25
> > [   51.367772] [IGT] xe_exec_system_allocator: executing
> > [   51.383611] [IGT] xe_exec_system_allocator: starting subtest once-malloc
> > [   51.386066] xe 0000:00:04.0: [drm:vm_bind_ioctl_ops_create [xe]]
> > op=0, addr=0x0000000000000000, range=0x0001000000000000,
> > bo_offset_or_userptr=0x0000000000000000
> > [   51.386171] xe 0000:00:04.0: [drm:vm_bind_ioctl_ops_create [xe]] MAP:
> > addr=0x0000000000000000, range=0x0001000000000000
> > [   51.389429] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] PAGE
> > FAULT: asid=1, gpusvm=0xffff8881775e9188, vram=0,0,
> > seqno=9223372036854775807, start=0x005584e8400000, end=0x005584e8410000,
> > size=65536
> > [   51.389529] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]]
> > ALLOCATE VRAM: asid=1, gpusvm=0xffff8881775e9188, vram=0,0,
> > seqno=9223372036854775807, start=0x005584e8400000, end=0x005584e8410000,
> > size=65536
> > [   51.389935] xe 0000:00:04.0: [drm:xe_svm_handle_pagefault [xe]] ALLOC
> > VRAM: asid=1, gpusvm=0xffff8881775e9188, pfn=3126960, npages=16
> > [   51.390048] xe 0000:00:04.0: [drm:xe_svm_invalidate [xe]] INVALIDATE:
> > asid=1, gpusvm=0xffff8881775e9188, seqno=3, start=0x00005584e8400000,
> > end=0x00005584e8410000, event=6
> > [   51.390440] xe 0000:00:04.0: [drm:xe_svm_invalidate [xe]] NOTIFIER:
> > asid=1, gpusvm=0xffff8881775e9188, vram=0,0, seqno=9223372036854775807,
> > start=0x005584e8400000, end=0x005584e8410000, size=65536
> > [   51.390948] Oops: general protection fault, probably for non-
> > canonical address 0x3fff88842fc80000: 0000 [#1] PREEMPT SMP NOPTI
> > [   51.391624] CPU: 1 UID: 0 PID: 76 Comm: kworker/u17:0 Not tainted
> > 6.13.0-rc4-drm-tip-test+ #48
> > [   51.392088] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
> > BIOS 1.15.0-1 04/01/2014
> > [   51.392527] Workqueue: xe_gt_page_fault_work_queue pf_queue_work_func
> > [xe]
> > [   51.392947] RIP: 0010:zone_device_page_init+0x5d/0x240
> > [   51.393228] Code: 04 dd ff e8 d5 d2 a1 00 5a 85 c0 0f 85 ba 00 00 00
> > e8 d7 bb df ff 85 c0 0f 84 9d 01 00 00 48 8b 45 38 a8 03 0f 85 ec 00 00
> > 00 <65> 48 ff 00 e8 aa d2 a1 00 85 c0 0f 85 0d 01 00 00 48 c7 c7 20 cb
> > [   51.394247] RSP: 0018:ffffc9000039fb48 EFLAGS: 00010246
> > [   51.394570] RAX: 4000000000000000 RBX: ffffea000bedac00 RCX:
> > 0000000000000000
> > [   51.394950] RDX: 0000000000000046 RSI: ffffffff824c67b4 RDI:
> > ffffffff824e58f5
> > [   51.395328] RBP: ffffea000bedac08 R08: 0000000000000015 R09:
> > 0000000000000004
> > [   51.395709] R10: 0000000000000001 R11: 0000000000000004 R12:
> > 0000000000000001
> > [   51.396093] R13: ffff888170fd8d40 R14: ffff88817f922640 R15:
> > ffffea000bedac00
> > [   51.396472] FS:  0000000000000000(0000) GS:ffff88842fc80000(0000)
> > knlGS:0000000000000000
> > [   51.396925] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [   51.397237] CR2: 0000563d1f7ecbe4 CR3: 000000017c212000 CR4:
> > 0000000000750ef0
> > [   51.397618] PKRU: 55555554
> > [   51.397768] Call Trace:
> > [   51.397904]  <TASK>
> > [   51.398024]  ? __die_body.cold+0x19/0x26
> > [   51.398238]  ? die_addr+0x38/0x60
> > [   51.398420]  ? exc_general_protection+0x19e/0x450
> > [   51.398678]  ? asm_exc_general_protection+0x22/0x30
> > [   51.398942]  ? zone_device_page_init+0x5d/0x240
> > [   51.399188]  ? zone_device_page_init+0x49/0x240
> > [   51.399433]  drm_gpusvm_migrate_to_devmem+0x379/0x9e0 [drm_gpusvm]
> > [   51.399768]  xe_svm_handle_pagefault+0x62c/0xa60 [xe]
> > [   51.400110]  ? xe_vm_find_overlapping_vma+0xa4/0x1d0 [xe]
> > [   51.400475]  pf_queue_work_func+0x1ba/0x450 [xe]
> > [   51.400777]  process_one_work+0x1fe/0x580
> > [   51.400996]  worker_thread+0x1d1/0x3b0
> > [   51.401201]  ? __pfx_worker_thread+0x10/0x10
> > [   51.401433]  kthread+0xeb/0x120
> > [   51.401609]  ? __pfx_kthread+0x10/0x10
> > [   51.401813]  ret_from_fork+0x2d/0x50
> > [   51.402008]  ? __pfx_kthread+0x10/0x10
> > [   51.402211]  ret_from_fork_asm+0x1a/0x30
> > [   51.402427]  </TASK>
> > [   51.402551] Modules linked in: xe drm_ttm_helper gpu_sched
> > drm_suballoc_helper drm_gpuvm drm_exec drm_gpusvm i2c_algo_bit drm_buddy
> > video wmi ttm drm_display_helper drm_kms_helper crct10dif_pclmul
> > crc32_pclmul e1000 ghash_clmulni_intel i2c_piix4 i2c_smbus fuse
> > [   51.403779] ---[ end trace 0000000000000000 ]---
> > [   51.404106] RIP: 0010:zone_device_page_init+0x5d/0x240
> > [   51.404393] Code: 04 dd ff e8 d5 d2 a1 00 5a 85 c0 0f 85 ba 00 00 00
> > e8 d7 bb df ff 85 c0 0f 84 9d 01 00 00 48 8b 45 38 a8 03 0f 85 ec 00 00
> > 00 <65> 48 ff 00 e8 aa d2 a1 00 85 c0 0f 85 0d 01 00 00 48 c7 c7 20 cb
> > [   51.405408] RSP: 0018:ffffc9000039fb48 EFLAGS: 00010246
> > [   51.405725] RAX: 4000000000000000 RBX: ffffea000bedac00 RCX:
> > 0000000000000000
> > [   51.406110] RDX: 0000000000000046 RSI: ffffffff824c67b4 RDI:
> > ffffffff824e58f5
> > [   51.406518] RBP: ffffea000bedac08 R08: 0000000000000015 R09:
> > 0000000000000004
> > [   51.406905] R10: 0000000000000001 R11: 0000000000000004 R12:
> > 0000000000000001
> > [   51.407312] R13: ffff888170fd8d40 R14: ffff88817f922640 R15:
> > ffffea000bedac00
> > [   51.407691] FS:  0000000000000000(0000) GS:ffff88842fc80000(0000)
> > knlGS:0000000000000000
> > [   51.408135] CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033
> > [   51.408484] CR2: 0000563d1f7ecbe4 CR3: 000000017c212000 CR4:
> > 0000000000750ef0
> > [   51.408877] PKRU: 55555554
> > [   51.409047] BUG: sleeping function called from invalid context at ./
> > include/linux/percpu-rwsem.h:49
> > [   51.409528] in_atomic(): 0, irqs_disabled(): 0, non_block: 0, pid:
> > 76, name: kworker/u17:0
> > [   51.409976] preempt_count: 0, expected: 0
> > [   51.410212] RCU nest depth: 1, expected: 0
> > [   51.410435] INFO: lockdep is turned off.
> > [   51.410648] CPU: 1 UID: 0 PID: 76 Comm: kworker/u17:0 Tainted: G
> > D            6.13.0-rc4-drm-tip-test+ #48
> > [   51.411180] Tainted: [D]=DIE
> > [   51.411338] Hardware name: QEMU Standard PC (i440FX + PIIX, 1996),
> > BIOS 1.15.0-1 04/01/2014
> > [   51.411859] Workqueue: xe_gt_page_fault_work_queue pf_queue_work_func
> > [xe]
> > [   51.412269] Call Trace:
> > [   51.412404]  <TASK>
> > [   51.412525]  dump_stack_lvl+0x69/0xa0
> > [   51.412724]  __might_resched.cold+0xe5/0x120
> > [   51.412956]  exit_signals+0x1a/0x360
> > [   51.413150]  do_exit+0x122/0xbd0
> > [   51.413328]  ? __pfx_worker_thread+0x10/0x10
> > [   51.413562]  make_task_dead+0x88/0x90
> > [   51.413783]  rewind_stack_and_make_dead+0x16/0x20
> > [   51.414045] RIP: 0000:0x0
> > [   51.414191] Code: Unable to access opcode bytes at 0xffffffffffffffd6.
> > [   51.414595] RSP: 0000:0000000000000000 EFLAGS: 00000000 ORIG_RAX:
> > 0000000000000000
> > [   51.414993] RAX: 0000000000000000 RBX: 0000000000000000 RCX:
> > 0000000000000000
> > [   51.415369] RDX: 0000000000000000 RSI: 0000000000000000 RDI:
> > 0000000000000000
> > [   51.415746] RBP: 0000000000000000 R08: 0000000000000000 R09:
> > 0000000000000000
> > [   51.416123] R10: 0000000000000000 R11: 0000000000000000 R12:
> > 0000000000000000
> > [   51.416501] R13: 0000000000000000 R14: 0000000000000000 R15:
> > 0000000000000000
> > [   51.416899]  </TASK>
> > 
> > 
> > On 12/18/24 1:33 AM, Matthew Brost wrote:
> > > Version 3 of GPU SVM has been promoted to the proper series from an RFC.
> > > Thanks to everyone (especially Sima and Thomas) for their numerous
> > > reviews on revision 1, 2 and for helping to address many design issues.
> > > 
> > > This version has been tested with IGT [1] on PVC, BMG, and LNL. Also
> > > tested with level0 (UMD) PR [2].
> > > 
> > > Major changes in v2:
> > > - Dropped mmap write abuse
> > > - core MM locking and retry loops instead of driver locking to avoid
> > > races
> > > - Removed physical to virtual references
> > > - Embedded structure/ops for drm_gpusvm_devmem
> > > - Fixed mremap and fork issues
> > > - Added DRM pagemap
> > > - Included RFC documentation in the kernel doc
> > > 
> > > Major changes in v3:
> > > - Move GPU SVM and DRM pagemap to DRM level
> > > - Mostly addresses Thomas's feedback, lots of small changes documented
> > >    in each individual patch change log
> > > 
> > > Known issues in v3:
> > > - Check pages still exists, changed to threshold in this version which
> > >    is better but still need to root cause cross process page finding on
> > >    small user allocations.
> > > - Dropped documentation patch, fairly large rewrite and will send out
> > >    independently once finished.
> > > 
> > > Matt
> > > 
> > > [1] https://patchwork.freedesktop.org/series/137545/#rev3
> > > [2] https://github.com/intel/compute-runtime/pull/782
> > > 
> > > Matthew Brost (27):
> > >    drm/xe: Retry BO allocation
> > >    mm/migrate: Add migrate_device_pfns
> > >    mm/migrate: Trylock device page in do_swap_page
> > >    drm/gpusvm: Add support for GPU Shared Virtual Memory
> > >    drm/xe: Select DRM_GPUSVM Kconfig
> > >    drm/xe/uapi: Add DRM_XE_VM_BIND_FLAG_CPU_ADDR_MIRROR flag
> > >    drm/xe: Add SVM init / close / fini to faulting VMs
> > >    drm/xe: Nuke VM's mapping upon close
> > >    drm/xe: Add SVM range invalidation and page fault handler
> > >    drm/gpuvm: Add DRM_GPUVA_OP_DRIVER
> > >    drm/xe: Add (re)bind to SVM page fault handler
> > >    drm/xe: Add SVM garbage collector
> > >    drm/xe: Add unbind to SVM garbage collector
> > >    drm/xe: Do not allow CPU address mirror VMA unbind if the GPU has
> > >      bindings
> > >    drm/xe: Enable CPU address mirror uAPI
> > >    drm/xe: Add migrate layer functions for SVM support
> > >    drm/xe: Add SVM device memory mirroring
> > >    drm/xe: Add drm_gpusvm_devmem to xe_bo
> > >    drm/xe: Add GPUSVM device memory copy vfunc functions
> > >    drm/xe: Add Xe SVM populate_devmem_pfn GPU SVM vfunc
> > >    drm/xe: Add Xe SVM devmem_release GPU SVM vfunc
> > >    drm/xe: Add BO flags required for SVM
> > >    drm/xe: Add SVM VRAM migration
> > >    drm/xe: Basic SVM BO eviction
> > >    drm/xe: Add SVM debug
> > >    drm/xe: Add modparam for SVM notifier size
> > >    drm/xe: Add always_migrate_to_vram modparam
> > > 
> > > Thomas Hellström (3):
> > >    drm/pagemap: Add DRM pagemap
> > >    drm/xe: Add dma_addr res cursor
> > >    drm/xe: Add drm_pagemap ops to SVM
> > > 
> > >   drivers/gpu/drm/Kconfig                     |    8 +
> > >   drivers/gpu/drm/Makefile                    |    1 +
> > >   drivers/gpu/drm/drm_gpusvm.c                | 2220 +++++++++++++++++++
> > >   drivers/gpu/drm/xe/Kconfig                  |   10 +
> > >   drivers/gpu/drm/xe/Makefile                 |    1 +
> > >   drivers/gpu/drm/xe/xe_bo.c                  |   20 +-
> > >   drivers/gpu/drm/xe/xe_bo.h                  |    1 +
> > >   drivers/gpu/drm/xe/xe_bo_types.h            |    4 +
> > >   drivers/gpu/drm/xe/xe_device_types.h        |   15 +
> > >   drivers/gpu/drm/xe/xe_gt_pagefault.c        |   17 +-
> > >   drivers/gpu/drm/xe/xe_gt_tlb_invalidation.c |   24 +
> > >   drivers/gpu/drm/xe/xe_gt_tlb_invalidation.h |    2 +
> > >   drivers/gpu/drm/xe/xe_migrate.c             |  175 ++
> > >   drivers/gpu/drm/xe/xe_migrate.h             |   10 +
> > >   drivers/gpu/drm/xe/xe_module.c              |    7 +
> > >   drivers/gpu/drm/xe/xe_module.h              |    2 +
> > >   drivers/gpu/drm/xe/xe_pt.c                  |  393 +++-
> > >   drivers/gpu/drm/xe/xe_pt.h                  |    5 +
> > >   drivers/gpu/drm/xe/xe_pt_types.h            |    2 +
> > >   drivers/gpu/drm/xe/xe_res_cursor.h          |  116 +-
> > >   drivers/gpu/drm/xe/xe_svm.c                 |  948 ++++++++
> > >   drivers/gpu/drm/xe/xe_svm.h                 |   83 +
> > >   drivers/gpu/drm/xe/xe_tile.c                |    5 +
> > >   drivers/gpu/drm/xe/xe_vm.c                  |  375 +++-
> > >   drivers/gpu/drm/xe/xe_vm.h                  |   15 +-
> > >   drivers/gpu/drm/xe/xe_vm_types.h            |   57 +
> > >   include/drm/drm_gpusvm.h                    |  445 ++++
> > >   include/drm/drm_gpuvm.h                     |    5 +
> > >   include/drm/drm_pagemap.h                   |  103 +
> > >   include/linux/migrate.h                     |    1 +
> > >   include/uapi/drm/xe_drm.h                   |   19 +-
> > >   mm/memory.c                                 |   13 +-
> > >   mm/migrate_device.c                         |  116 +-
> > >   33 files changed, 5061 insertions(+), 157 deletions(-)
> > >   create mode 100644 drivers/gpu/drm/drm_gpusvm.c
> > >   create mode 100644 drivers/gpu/drm/xe/xe_svm.c
> > >   create mode 100644 drivers/gpu/drm/xe/xe_svm.h
> > >   create mode 100644 include/drm/drm_gpusvm.h
> > >   create mode 100644 include/drm/drm_pagemap.h
> > > 
> > 
> 

Re: [PATCH v3 05/30] drm/gpusvm: Add support for GPU Shared Virtual Memory

[Gwan-gyeong Mun]

On 12/18/24 1:33 AM, Matthew Brost wrote:
> This patch introduces support for GPU Shared Virtual Memory (SVM) in the
> Direct Rendering Manager (DRM) subsystem. SVM allows for seamless
> sharing of memory between the CPU and GPU, enhancing performance and
> flexibility in GPU computing tasks.
> 
> The patch adds the necessary infrastructure for SVM, including data
> structures and functions for managing SVM ranges and notifiers. It also
> provides mechanisms for allocating, deallocating, and migrating memory
> regions between system RAM and GPU VRAM.
> 
> This is largely inspired by GPUVM.
> 
> v2:
>   - Take order into account in check pages
>   - Clear range->pages in get pages error
>   - Drop setting dirty or accessed bit in get pages (Vetter)
>   - Remove mmap assert for cpu faults
>   - Drop mmap write lock abuse (Vetter, Christian)
>   - Decouple zdd from range (Vetter, Oak)
>   - Add drm_gpusvm_range_evict, make it work with coherent pages
>   - Export drm_gpusvm_evict_to_sram, only use in BO evict path (Vetter)
>   - mmget/put in drm_gpusvm_evict_to_sram
>   - Drop range->vram_alloation variable
>   - Don't return in drm_gpusvm_evict_to_sram until all pages detached
>   - Don't warn on mixing sram and device pages
>   - Update kernel doc
>   - Add coherent page support to get pages
>   - Use DMA_FROM_DEVICE rather than DMA_BIDIRECTIONAL
>   - Add struct drm_gpusvm_vram and ops (Thomas)
>   - Update the range's seqno if the range is valid (Thomas)
>   - Remove the is_unmapped check before hmm_range_fault (Thomas)
>   - Use drm_pagemap (Thomas)
>   - Drop kfree_mapping (Thomas)
>   - dma mapp pages under notifier lock (Thomas)
>   - Remove ctx.prefault
>   - Remove ctx.mmap_locked
>   - Add ctx.check_pages
>   - s/vram/devmem (Thomas)
> v3:
>   - Fix memory leak drm_gpusvm_range_get_pages
>   - Only migrate pages with same zdd on CPU fault
>   - Loop over al VMAs in drm_gpusvm_range_evict
>   - Make GPUSVM a drm level module
>   - GPL or MIT license
>   - Update main kernel doc (Thomas)
>   - Prefer foo() vs foo for functions in kernel doc (Thomas)
>   - Prefer functions over macros (Thomas)
>   - Use unsigned long vs u64 for addresses (Thomas)
>   - Use standard interval_tree (Thomas)
>   - s/drm_gpusvm_migration_put_page/drm_gpusvm_migration_unlock_put_page (Thomas)
>   - Drop err_out label in drm_gpusvm_range_find_or_insert (Thomas)
>   - Fix kernel doc in drm_gpusvm_range_free_pages (Thomas)
>   - Newlines between functions defs in header file (Thomas)
>   - Drop shall language in driver vfunc kernel doc (Thomas)
>   - Move some static inlines from head to C file (Thomas)
>   - Don't allocate pages under page lock in drm_gpusvm_migrate_populate_ram_pfn (Thomas)
>   - Change check_pages to a threshold
> 
> Cc: Simona Vetter <simona.vetter@ffwll.ch>
> Cc: Dave Airlie <airlied@redhat.com>
> Cc: Christian König <christian.koenig@amd.com>
> Cc: <dri-devel@lists.freedesktop.org>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
>   drivers/gpu/drm/Kconfig      |    8 +
>   drivers/gpu/drm/Makefile     |    1 +
>   drivers/gpu/drm/drm_gpusvm.c | 2220 ++++++++++++++++++++++++++++++++++
>   include/drm/drm_gpusvm.h     |  445 +++++++
>   4 files changed, 2674 insertions(+)
>   create mode 100644 drivers/gpu/drm/drm_gpusvm.c
>   create mode 100644 include/drm/drm_gpusvm.h
> 
> diff --git a/drivers/gpu/drm/Kconfig b/drivers/gpu/drm/Kconfig
> index 2f51546b0b88..5524cb1ed9dd 100644
> --- a/drivers/gpu/drm/Kconfig
> +++ b/drivers/gpu/drm/Kconfig
> @@ -278,6 +278,14 @@ config DRM_GPUVM
>   	  GPU-VM representation providing helpers to manage a GPUs virtual
>   	  address space
>   
> +config DRM_GPUSVM
> +	tristate
> +	depends on DRM
> +	depends on DEVICE_MIGRATION
> +	help
> +	  GPU-SVM representation providing helpers to manage a GPUs shared
> +	  virtual memory
> +
>   config DRM_BUDDY
>   	tristate
>   	depends on DRM
> diff --git a/drivers/gpu/drm/Makefile b/drivers/gpu/drm/Makefile
> index 19fb370fbc56..ca5d44441fb6 100644
> --- a/drivers/gpu/drm/Makefile
> +++ b/drivers/gpu/drm/Makefile
> @@ -104,6 +104,7 @@ obj-$(CONFIG_DRM_PANEL_BACKLIGHT_QUIRKS) += drm_panel_backlight_quirks.o
>   #
>   obj-$(CONFIG_DRM_EXEC) += drm_exec.o
>   obj-$(CONFIG_DRM_GPUVM) += drm_gpuvm.o
> +obj-$(CONFIG_DRM_GPUSVM) += drm_gpusvm.o
>   
>   obj-$(CONFIG_DRM_BUDDY) += drm_buddy.o
>   
> diff --git a/drivers/gpu/drm/drm_gpusvm.c b/drivers/gpu/drm/drm_gpusvm.c
> new file mode 100644
> index 000000000000..29ce2c7d1c8c
> --- /dev/null
> +++ b/drivers/gpu/drm/drm_gpusvm.c
> @@ -0,0 +1,2220 @@
> +// SPDX-License-Identifier: GPL-2.0-only OR MIT
> +/*
> + * Copyright © 2024 Intel Corporation
> + *
> + * Authors:
> + *     Matthew Brost <matthew.brost@intel.com>
> + */
> +
> +#include <linux/dma-mapping.h>
> +#include <linux/hmm.h>
> +#include <linux/memremap.h>
> +#include <linux/migrate.h>
> +#include <linux/mm_types.h>
> +#include <linux/pagemap.h>
> +#include <linux/slab.h>
> +
> +#include <drm/drm_device.h>
> +#include <drm/drm_gpusvm.h>
> +#include <drm/drm_pagemap.h>
> +#include <drm/drm_print.h>
> +
> +/**
> + * DOC: Overview
> + *
> + * GPU Shared Virtual Memory (GPU SVM) layer for the Direct Rendering Manager (DRM)
> + *
> + * The GPU SVM layer is a component of the DRM framework designed to manage shared
> + * virtual memory between the CPU and GPU. It enables efficient data exchange and
> + * processing for GPU-accelerated applications by allowing memory sharing and
> + * synchronization between the CPU's and GPU's virtual address spaces.
> + *
> + * Key GPU SVM Components:
> + * - Notifiers: Notifiers: Used for tracking memory intervals and notifying the
> + *		GPU of changes, notifiers are sized based on a GPU SVM
> + *		initialization parameter, with a recommendation of 512M or
> + *		larger. They maintain a Red-BlacK tree and a list of ranges that
> + *		fall within the notifier interval. Notifiers are tracked within
> + *		a GPU SVM Red-BlacK tree and list and are dynamically inserted
> + *		or removed as ranges within the interval are created or
> + *		destroyed.
> + * - Ranges: Represent memory ranges mapped in a DRM device and managed
> + *	     by GPU SVM. They are sized based on an array of chunk sizes, which
> + *	     is a GPU SVM initialization parameter, and the CPU address space.
> + *	     Upon GPU fault, the largest aligned chunk that fits within the
> + *	     faulting CPU address space is chosen for the range size. Ranges are
> + *	     expected to be dynamically allocated on GPU fault and removed on an
> + *	     MMU notifier UNMAP event. As mentioned above, ranges are tracked in
> + *	     a notifier's Red-Black tree.
> + * - Operations: Define the interface for driver-specific GPU SVM operations
> + *               such as range allocation, notifier allocation, and
> + *               invalidations.
> + * - Device Memory Allocations: Embedded structure containing enough information
> + *                              for GPU SVM to migrate to / from device memory.
> + * - Device Memory Operations: Define the interface for driver-specific device
> + *                             memory operations release memory, populate pfns,
> + *                             and copy to / from device memory.
> + *
> + * This layer provides interfaces for allocating, mapping, migrating, and
> + * releasing memory ranges between the CPU and GPU. It handles all core memory
> + * management interactions (DMA mapping, HMM, and migration) and provides
> + * driver-specific virtual functions (vfuncs). This infrastructure is sufficient
> + * to build the expected driver components for an SVM implementation as detailed
> + * below.
> + *
> + * Expected Driver Components:
> + * - GPU page fault handler: Used to create ranges and notifiers based on the
> + *			     fault address, optionally migrate the range to
> + *			     device memory, and create GPU bindings.
> + * - Garbage collector: Used to unmap and destroy GPU bindings for ranges.
> + *			Ranges are expected to be added to the garbage collector
> + *			upon a MMU_NOTIFY_UNMAP event in notifier callback.
> + * - Notifier callback: Used to invalidate and DMA unmap GPU bindings for
> + *			ranges.
> + */
> +
> +/**
> + * DOC: Locking
> + *
> + * GPU SVM handles locking for core MM interactions, i.e., it locks/unlocks the
> + * mmap lock as needed.
> + *
> + * GPU SVM introduces a global notifier lock, which safeguards the notifier's
> + * range RB tree and list, as well as the range's DMA mappings and sequence
> + * number. GPU SVM manages all necessary locking and unlocking operations,
> + * except for the recheck range's pages being valid
> + * (drm_gpusvm_range_pages_valid) when the driver is committing GPU bindings. This
> + * lock corresponds to the 'driver->update' lock mentioned in the HMM
> + * documentation (TODO: Link). Future revisions may transition from a GPU SVM
> + * global lock to a per-notifier lock if finer-grained locking is deemed
> + * necessary.
> + *
> + * In addition to the locking mentioned above, the driver should implement a
> + * lock to safeguard core GPU SVM function calls that modify state, such as
> + * drm_gpusvm_range_find_or_insert and drm_gpusvm_range_remove. This lock is
> + * denoted as 'driver_svm_lock' in code examples. Finer grained driver side
> + * locking should also be possible for concurrent GPU fault processing within a
> + * single GPU SVM. The 'driver_svm_lock' can be via drm_gpusvm_driver_set_lock
> + * to add annotations to GPU SVM.
> + */
> +
> +/**
> + * DOC: Migration
> + *
> + * The migration support is quite simple, allowing migration between RAM and
> + * device memory at the range granularity. For example, GPU SVM currently does not
> + * support mixing RAM and device memory pages within a range. This means that upon GPU
> + * fault, the entire range can be migrated to device memory, and upon CPU fault, the
> + * entire range is migrated to RAM. Mixed RAM and device memory storage within a range
> + * could be added in the future if required.
> + *
> + * The reasoning for only supporting range granularity is as follows: it
> + * simplifies the implementation, and range sizes are driver-defined and should
> + * be relatively small.
> + */
> +
> +/**
> + * DOC: Partial Unmapping of Ranges
> + *
> + * Partial unmapping of ranges (e.g., 1M out of 2M is unmapped by CPU resulting
> + * in MMU_NOTIFY_UNMAP event) presents several challenges, with the main one
> + * being that a subset of the range still has CPU and GPU mappings. If the
> + * backing store for the range is in device memory, a subset of the backing store has
> + * references. One option would be to split the range and device memory backing store,
> + * but the implementation for this would be quite complicated. Given that
> + * partial unmappings are rare and driver-defined range sizes are relatively
> + * small, GPU SVM does not support splitting of ranges.
> + *
> + * With no support for range splitting, upon partial unmapping of a range, the
> + * driver is expected to invalidate and destroy the entire range. If the range
> + * has device memory as its backing, the driver is also expected to migrate any
> + * remaining pages back to RAM.
> + */
> +
> +/**
> + * DOC: Examples
> + *
> + * This section provides three examples of how to build the expected driver
> + * components: the GPU page fault handler, the garbage collector, and the
> + * notifier callback.
> + *
> + * The generic code provided does not include logic for complex migration
> + * policies, optimized invalidations, fined grained driver locking, or other
> + * potentially required driver locking (e.g., DMA-resv locks).
> + *
> + * 1) GPU page fault handler
> + *
> + *	int driver_bind_range(struct drm_gpusvm *gpusvm, struct drm_gpusvm_range *range)
> + *	{
> + *		int err = 0;
> + *
> + *		driver_alloc_and_setup_memory_for_bind(gpusvm, range);
> + *
> + *		drm_gpusvm_notifier_lock(gpusvm);
> + *		if (drm_gpusvm_range_pages_valid(range))
> + *			driver_commit_bind(gpusvm, range);
> + *		else
> + *			err = -EAGAIN;
> + *		drm_gpusvm_notifier_unlock(gpusvm);
> + *
> + *		return err;
> + *	}
> + *
> + *	int driver_gpu_fault(struct drm_gpusvm *gpusvm, unsigned long fault_addr,
> + *			     unsigned long gpuva_start, unsigned long gpuva_end)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = {};
> + *		int err;
> + *
> + *		driver_svm_lock();
> + *	retry:
> + *		// Always process UNMAPs first so view of GPU SVM ranges is current
> + *		driver_garbage_collector(gpusvm);
> + *
> + *		range = drm_gpusvm_range_find_or_insert(gpusvm, fault_addr,
> + *							gpuva_start, gpuva_end,
> + *						        &ctx);
> + *		if (IS_ERR(range)) {
> + *			err = PTR_ERR(range);
> + *			goto unlock;
> + *		}
> + *
> + *		if (driver_migration_policy(range)) {
> + *			devmem = driver_alloc_devmem();
> + *			err = drm_gpusvm_migrate_to_devmem(gpusvm, range,
> + *							   devmem_allocation,
> + *							   &ctx);
> + *			if (err)	// CPU mappings may have changed
> + *				goto retry;
> + *		}
> + *
> + *		err = drm_gpusvm_range_get_pages(gpusvm, range, &ctx);
> + *		if (err == -EOPNOTSUPP || err == -EFAULT || err == -EPERM) {	// CPU mappings changed
> + *			if (err == -EOPNOTSUPP)
> + *				drm_gpusvm_range_evict(gpusvm, range);
> + *			goto retry;
> + *		} else if (err) {
> + *			goto unlock;
> + *		}
> + *
> + *		err = driver_bind_range(gpusvm, range);
> + *		if (err == -EAGAIN)	// CPU mappings changed
> + *			goto retry
> + *
> + *	unlock:
> + *		driver_svm_unlock();
> + *		return err;
> + *	}
> + *
> + * 2) Garbage Collector.
> + *
> + *	void __driver_garbage_collector(struct drm_gpusvm *gpusvm,
> + *					struct drm_gpusvm_range *range)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		// Partial unmap, migrate any remaining device memory pages back to RAM
> + *		if (range->flags.partial_unmap)
> + *			drm_gpusvm_range_evict(gpusvm, range);
> + *
> + *		driver_unbind_range(range);
> + *		drm_gpusvm_range_remove(gpusvm, range);
> + *	}
> + *
> + *	void driver_garbage_collector(struct drm_gpusvm *gpusvm)
> + *	{
> + *		assert_driver_svm_locked(gpusvm);
> + *
> + *		for_each_range_in_garbage_collector(gpusvm, range)
> + *			__driver_garbage_collector(gpusvm, range);
> + *	}
> + *
> + * 3) Notifier callback.
> + *
> + *	void driver_invalidation(struct drm_gpusvm *gpusvm,
> + *				 struct drm_gpusvm_notifier *notifier,
> + *				 const struct mmu_notifier_range *mmu_range)
> + *	{
> + *		struct drm_gpusvm_ctx ctx = { .in_notifier = true, };
> + *		struct drm_gpusvm_range *range = NULL;
> + *
> + *		driver_invalidate_device_pages(gpusvm, mmu_range->start, mmu_range->end);
> + *
> + *		drm_gpusvm_for_each_range(range, notifier, mmu_range->start,
> + *					  mmu_range->end) {
> + *			drm_gpusvm_range_unmap_pages(gpusvm, range, &ctx);
> + *
> + *			if (mmu_range->event != MMU_NOTIFY_UNMAP)
> + *				continue;
> + *
> + *			drm_gpusvm_range_set_unmapped(range, mmu_range);
> + *			driver_garbage_collector_add(gpusvm, range);
> + *		}
> + *	}
> + */
> +
> +/**
> + * npages_in_range() - Calculate the number of pages in a given range
> + * @start: The start address of the range
> + * @end: The end address of the range
> + *
> + * This macro calculates the number of pages in a given memory range,
> + * specified by the start and end addresses. It divides the difference
> + * between the end and start addresses by the page size (PAGE_SIZE) to
> + * determine the number of pages in the range.
> + *
> + * Return: The number of pages in the specified range.
> + */
> +static unsigned long
> +npages_in_range(unsigned long start, unsigned long end)
> +{
> +	return (end - start) >> PAGE_SHIFT;
> +}
> +
> +/**
> + * struct drm_gpusvm_zdd - GPU SVM zone device data
> + *
> + * @refcount: Reference count for the zdd
> + * @destroy_work: Work structure for asynchronous zdd destruction
> + * @devmem_allocation: device memory allocation
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This structure serves as a generic wrapper installed in
> + * page->zone_device_data. It provides infrastructure for looking up a device
> + * memory allocation upon CPU page fault and asynchronously releasing device
> + * memory once the CPU has no page references. Asynchronous release is useful
> + * because CPU page references can be dropped in IRQ contexts, while releasing
> + * device memory likely requires sleeping locks.
> + */
> +struct drm_gpusvm_zdd {
> +	struct kref refcount;
> +	struct work_struct destroy_work;
> +	struct drm_gpusvm_devmem *devmem_allocation;
> +	void *device_private_page_owner;
> +};
> +
> +/**
> + * drm_gpusvm_zdd_destroy_work_func() - Work function for destroying a zdd
> + * @w: Pointer to the work_struct
> + *
> + * This function releases device memory, puts GPU SVM range, and frees zdd.
> + */
> +static void drm_gpusvm_zdd_destroy_work_func(struct work_struct *w)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(w, struct drm_gpusvm_zdd, destroy_work);
> +	const struct drm_gpusvm_devmem_ops *ops = zdd->devmem_allocation ?
> +		zdd->devmem_allocation->ops : NULL;
> +
> +	if (zdd->devmem_allocation && ops->devmem_release)
> +		ops->devmem_release(zdd->devmem_allocation);
> +	kfree(zdd);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_alloc() - Allocate a zdd structure.
> + * @device_private_page_owner: Device private pages owner
> + *
> + * This function allocates and initializes a new zdd structure. It sets up the
> + * reference count and initializes the destroy work.
> + *
> + * Returns:
> + * Pointer to the allocated zdd on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_zdd *
> +drm_gpusvm_zdd_alloc(void *device_private_page_owner)
> +{
> +	struct drm_gpusvm_zdd *zdd;
> +
> +	zdd = kmalloc(sizeof(*zdd), GFP_KERNEL);
> +	if (!zdd)
> +		return NULL;
> +
> +	kref_init(&zdd->refcount);
> +	INIT_WORK(&zdd->destroy_work, drm_gpusvm_zdd_destroy_work_func);
> +	zdd->devmem_allocation = NULL;
> +	zdd->device_private_page_owner = device_private_page_owner;
> +
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_get() - Get a reference to a zdd structure.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function increments the reference count of the provided zdd structure.
> + *
> + * Returns: Pointer to the zdd structure.
> + */
> +static struct drm_gpusvm_zdd *drm_gpusvm_zdd_get(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_get(&zdd->refcount);
> +	return zdd;
> +}
> +
> +/**
> + * drm_gpusvm_zdd_destroy() - Destroy a zdd structure.
> + * @ref: Pointer to the reference count structure.
> + *
> + * This function queues the destroy_work of the zdd for asynchronous destruction.
> + */
> +static void drm_gpusvm_zdd_destroy(struct kref *ref)
> +{
> +	struct drm_gpusvm_zdd *zdd =
> +		container_of(ref, struct drm_gpusvm_zdd, refcount);
> +
> +	if (zdd->devmem_allocation)
> +		WRITE_ONCE(zdd->devmem_allocation->detached, true);
> +	schedule_work(&zdd->destroy_work);
> +}
> +
> +/**
> + * drm_gpusvm_zdd_put() - Put a zdd reference.
> + * @zdd: Pointer to the zdd structure.
> + *
> + * This function decrements the reference count of the provided zdd structure
> + * and schedules its destruction if the count drops to zero.
> + */
> +static void drm_gpusvm_zdd_put(struct drm_gpusvm_zdd *zdd)
> +{
> +	kref_put(&zdd->refcount, drm_gpusvm_zdd_destroy);
> +}
> +
> +/**
> + * drm_gpusvm_range_find() - Find GPU SVM range from GPU SVM notifier
> + * @notifier: Pointer to the GPU SVM notifier structure.
> + * @start: Start address of the range
> + * @end: End address of the range
> + *
> + * Return: A pointer to the drm_gpusvm_range if found or NULL
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(&notifier->root, start, end - 1);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_range, itree);
> +	else
> +		return NULL;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find);
> +
> +/**
> + * drm_gpusvm_for_each_range_safe() - Safely iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges
> + * @next__: Iterator variable for the ranges temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier while
> + * removing ranges from it.
> + */
> +#define drm_gpusvm_for_each_range_safe(range__, next__, notifier__, start__, end__)	\
> +	for ((range__) = drm_gpusvm_range_find((notifier__), (start__), (end__)),	\
> +	     (next__) = __drm_gpusvm_range_next(range__);				\
> +	     (range__) && (range__->itree.start < (end__));				\
> +	     (range__) = (next__), (next__) = __drm_gpusvm_range_next(range__))
> +
> +/**
> + * __drm_gpusvm_notifier_next() - get the next drm_gpusvm_notifier in the list
> + * @notifier: a pointer to the current drm_gpusvm_notifier
> + *
> + * Return: A pointer to the next drm_gpusvm_notifier if available, or NULL if
> + *         the current notifier is the last one or if the input notifier is
> + *         NULL.
> + */
> +static struct drm_gpusvm_notifier *
> +__drm_gpusvm_notifier_next(struct drm_gpusvm_notifier *notifier)
> +{
> +	if (notifier && !list_is_last(&notifier->entry,
> +				      &notifier->gpusvm->notifier_list))
> +		return list_next_entry(notifier, entry);
> +
> +	return NULL;
> +}
> +
> +static struct drm_gpusvm_notifier *
> +notifier_iter_first(struct rb_root_cached *root, unsigned long start,
> +		    unsigned long last)
> +{
> +	struct interval_tree_node *itree;
> +
> +	itree = interval_tree_iter_first(root, start, last);
> +
> +	if (itree)
> +		return container_of(itree, struct drm_gpusvm_notifier, itree);
> +	else
> +		return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_notifier() - Iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm.
> + */
> +#define drm_gpusvm_for_each_notifier(notifier__, gpusvm__, start__, end__)		\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1);	\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_for_each_notifier_safe() - Safely iterate over GPU SVM notifiers in a gpusvm
> + * @notifier__: Iterator variable for the notifiers
> + * @next__: Iterator variable for the notifiers temporay storage
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the notifier
> + * @end__: End address of the notifier
> + *
> + * This macro is used to iterate over GPU SVM notifiers in a gpusvm while
> + * removing notifiers from it.
> + */
> +#define drm_gpusvm_for_each_notifier_safe(notifier__, next__, gpusvm__, start__, end__)	\
> +	for ((notifier__) = notifier_iter_first(&(gpusvm__)->root, (start__), (end__) - 1),	\
> +	     (next__) = __drm_gpusvm_notifier_next(notifier__);				\
> +	     (notifier__) && (notifier__->itree.start < (end__));			\
> +	     (notifier__) = (next__), (next__) = __drm_gpusvm_notifier_next(notifier__))
> +
> +/**
> + * drm_gpusvm_notifier_invalidate() - Invalidate a GPU SVM notifier.
> + * @mni: Pointer to the mmu_interval_notifier structure.
> + * @mmu_range: Pointer to the mmu_notifier_range structure.
> + * @cur_seq: Current sequence number.
> + *
> + * This function serves as a generic MMU notifier for GPU SVM. It sets the MMU
> + * notifier sequence number and calls the driver invalidate vfunc under
> + * gpusvm->notifier_lock.
> + *
> + * Returns:
> + * true if the operation succeeds, false otherwise.
> + */
> +static bool
> +drm_gpusvm_notifier_invalidate(struct mmu_interval_notifier *mni,
> +			       const struct mmu_notifier_range *mmu_range,
> +			       unsigned long cur_seq)
> +{
> +	struct drm_gpusvm_notifier *notifier =
> +		container_of(mni, typeof(*notifier), notifier);
> +	struct drm_gpusvm *gpusvm = notifier->gpusvm;
> +
> +	if (!mmu_notifier_range_blockable(mmu_range))
> +		return false;
> +
> +	down_write(&gpusvm->notifier_lock);
> +	mmu_interval_set_seq(mni, cur_seq);
> +	gpusvm->ops->invalidate(gpusvm, notifier, mmu_range);
> +	up_write(&gpusvm->notifier_lock);
> +
> +	return true;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_ops - MMU interval notifier operations for GPU SVM
> + */
> +static const struct mmu_interval_notifier_ops drm_gpusvm_notifier_ops = {
> +	.invalidate = drm_gpusvm_notifier_invalidate,
> +};
> +
> +/**
> + * drm_gpusvm_init() - Initialize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @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.
> + * @ops: Pointer to the operations structure for GPU SVM.
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order with last
> + *               entry being SZ_4K.
> + * @num_chunks: Number of chunks.
> + *
> + * This function initializes the GPU SVM.
> + *
> + * Returns:
> + * 0 on success, a negative error code on failure.
> + */
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks)
> +{
> +	if (!ops->invalidate || !num_chunks)
> +		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;
> +	gpusvm->ops = ops;
> +	gpusvm->chunk_sizes = chunk_sizes;
> +	gpusvm->num_chunks = num_chunks;
> +
> +	mmgrab(mm);
> +	gpusvm->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&gpusvm->notifier_list);
> +
> +	init_rwsem(&gpusvm->notifier_lock);
> +
> +	fs_reclaim_acquire(GFP_KERNEL);
> +	might_lock(&gpusvm->notifier_lock);
> +	fs_reclaim_release(GFP_KERNEL);
> +
> +#ifdef CONFIG_LOCKDEP
> +	gpusvm->lock_dep_map = NULL;
> +#endif
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_init);
> +
> +/**
> + * drm_gpusvm_notifier_find() - Find GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function finds the GPU SVM notifier associated with the fault address.
> + *
> + * Returns:
> + * Pointer to the GPU SVM notifier on success, NULL otherwise.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_find(struct drm_gpusvm *gpusvm,
> +			 unsigned long fault_addr)
> +{
> +	return notifier_iter_first(&gpusvm->root, fault_addr, fault_addr + 1);
> +}
> +
> +/**
> + * to_drm_gpusvm_notifier() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_notifier struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_notifier structure.
> + */
> +static struct drm_gpusvm_notifier *to_drm_gpusvm_notifier(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_notifier, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_insert() - Insert GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function inserts the GPU SVM notifier into the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_insert(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	interval_tree_insert(&notifier->itree, &gpusvm->root);
> +
> +	node = rb_prev(&notifier->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_notifier(node))->entry;
> +	else
> +		head = &gpusvm->notifier_list;
> +
> +	list_add(&notifier->entry, head);
> +}
> +
> +/**
> + * drm_gpusvm_notifier_remove() - Remove GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM tructure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function removes the GPU SVM notifier from the GPU SVM RB tree and list.
> + */
> +static void drm_gpusvm_notifier_remove(struct drm_gpusvm *gpusvm,
> +				       struct drm_gpusvm_notifier *notifier)
> +{
> +	interval_tree_remove(&notifier->itree, &gpusvm->root);
> +	list_del(&notifier->entry);
> +}
> +
> +/**
> + * drm_gpusvm_fini() - Finalize the GPU SVM.
> + * @gpusvm: Pointer to the GPU SVM structure.
> + *
> + * This function finalizes the GPU SVM by cleaning up any remaining ranges and
> + * notifiers, and dropping a reference to struct MM.
> + */
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm)
> +{
> +	struct drm_gpusvm_notifier *notifier, *next;
> +
> +	drm_gpusvm_for_each_notifier_safe(notifier, next, gpusvm, 0, LONG_MAX) {
> +		struct drm_gpusvm_range *range, *__next;
> +
> +		/*
> +		 * Remove notifier first to avoid racing with any invalidation
> +		 */
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +		notifier->flags.removed = true;
> +
> +		drm_gpusvm_for_each_range_safe(range, __next, notifier, 0,
> +					       LONG_MAX)
> +			drm_gpusvm_range_remove(gpusvm, range);
> +	}
> +
> +	mmdrop(gpusvm->mm);
> +	WARN_ON(!RB_EMPTY_ROOT(&gpusvm->root.rb_root));
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_fini);
> +
> +/**
> + * drm_gpusvm_notifier_alloc() - Allocate GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + *
> + * This function allocates and initializes the GPU SVM notifier structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM notifier on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_notifier *
> +drm_gpusvm_notifier_alloc(struct drm_gpusvm *gpusvm, unsigned long fault_addr)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	if (gpusvm->ops->notifier_alloc)
> +		notifier = gpusvm->ops->notifier_alloc();
> +	else
> +		notifier = kzalloc(sizeof(*notifier), GFP_KERNEL);
> +
> +	if (!notifier)
> +		return ERR_PTR(-ENOMEM);
> +
> +	notifier->gpusvm = gpusvm;
> +	notifier->itree.start = ALIGN_DOWN(fault_addr, gpusvm->notifier_size);
> +	notifier->itree.last = ALIGN(fault_addr + 1, gpusvm->notifier_size) - 1;
> +	INIT_LIST_HEAD(&notifier->entry);
> +	notifier->root = RB_ROOT_CACHED;
> +	INIT_LIST_HEAD(&notifier->range_list);
> +
> +	return notifier;
> +}
> +
> +/**
> + * drm_gpusvm_notifier_free() - Free GPU SVM notifier
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + *
> + * This function frees the GPU SVM notifier structure.
> + */
> +static void drm_gpusvm_notifier_free(struct drm_gpusvm *gpusvm,
> +				     struct drm_gpusvm_notifier *notifier)
> +{
> +	WARN_ON(!RB_EMPTY_ROOT(&notifier->root.rb_root));
> +
> +	if (gpusvm->ops->notifier_free)
> +		gpusvm->ops->notifier_free(notifier);
> +	else
> +		kfree(notifier);
> +}
> +
> +/**
> + * to_drm_gpusvm_range() - retrieve the container struct for a given rbtree node
> + * @node: a pointer to the rbtree node embedded within a drm_gpusvm_range struct
> + *
> + * Return: A pointer to the containing drm_gpusvm_range structure.
> + */
> +static struct drm_gpusvm_range *to_drm_gpusvm_range(struct rb_node *node)
> +{
> +	return container_of(node, struct drm_gpusvm_range, itree.rb);
> +}
> +
> +/**
> + * drm_gpusvm_range_insert() - Insert GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function inserts the GPU SVM range into the notifier RB tree and list.
> + */
> +static void drm_gpusvm_range_insert(struct drm_gpusvm_notifier *notifier,
> +				    struct drm_gpusvm_range *range)
> +{
> +	struct rb_node *node;
> +	struct list_head *head;
> +
> +	drm_gpusvm_notifier_lock(notifier->gpusvm);
> +	interval_tree_insert(&range->itree, &notifier->root);
> +
> +	node = rb_prev(&range->itree.rb);
> +	if (node)
> +		head = &(to_drm_gpusvm_range(node))->entry;
> +	else
> +		head = &notifier->range_list;
> +
> +	list_add(&range->entry, head);
> +	drm_gpusvm_notifier_unlock(notifier->gpusvm);
> +}
> +
> +/**
> + * __drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This macro removes the GPU SVM range from the notifier RB tree and list.
> + */
> +static void __drm_gpusvm_range_remove(struct drm_gpusvm_notifier *notifier,
> +				      struct drm_gpusvm_range *range)
> +{
> +	interval_tree_remove(&range->itree, &notifier->root);
> +	list_del(&range->entry);
> +}
> +
> +/**
> + * drm_gpusvm_range_alloc() - Allocate GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @fault_addr: Fault address
> + * @chunk_size: Chunk size
> + * @migrate_devmem: Flag indicating whether to migrate device memory
> + *
> + * This function allocates and initializes the GPU SVM range structure.
> + *
> + * Returns:
> + * Pointer to the allocated GPU SVM range on success, ERR_PTR() on failure.
> + */
> +static struct drm_gpusvm_range *
> +drm_gpusvm_range_alloc(struct drm_gpusvm *gpusvm,
> +		       struct drm_gpusvm_notifier *notifier,
> +		       unsigned long fault_addr, unsigned long chunk_size,
> +		       bool migrate_devmem)
> +{
> +	struct drm_gpusvm_range *range;
> +
> +	if (gpusvm->ops->range_alloc)
> +		range = gpusvm->ops->range_alloc(gpusvm);
> +	else
> +		range = kzalloc(sizeof(*range), GFP_KERNEL);
> +
> +	if (!range)
> +		return ERR_PTR(-ENOMEM);
> +
> +	kref_init(&range->refcount);
> +	range->gpusvm = gpusvm;
> +	range->notifier = notifier;
> +	range->itree.start = ALIGN_DOWN(fault_addr, chunk_size);
> +	range->itree.last = ALIGN(fault_addr + 1, chunk_size) - 1;
> +	INIT_LIST_HEAD(&range->entry);
> +	range->notifier_seq = LONG_MAX;
> +	range->flags.migrate_devmem = migrate_devmem ? 1 : 0;
> +
> +	return range;
> +}
> +
> +/**
> + * drm_gpusvm_check_pages() - Check pages
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @start: Start address
> + * @end: End address
> + *
> + * Check if pages between start and end have been faulted in on the CPU. Use to
> + * prevent migration of pages without CPU backing store.
> + *
> + * Returns:
> + * True if pages have been faulted into CPU, False otherwise
> + */
> +static bool drm_gpusvm_check_pages(struct drm_gpusvm *gpusvm,
> +				   struct drm_gpusvm_notifier *notifier,
> +				   unsigned long start, unsigned long end)
> +{
> +	struct hmm_range hmm_range = {
> +		.default_flags = 0,
> +		.notifier = &notifier->notifier,
> +		.start = start,
> +		.end = end,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long *pfns;
> +	unsigned long npages = npages_in_range(start, end);
> +	int err, i;
> +
> +	mmap_assert_locked(gpusvm->mm);
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return false;
> +
> +	hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +	hmm_range.hmm_pfns = pfns;
> +
> +	while (true) {
> +		err = hmm_range_fault(&hmm_range);
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(&notifier->notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	if (err)
> +		goto err_free;
> +
> +	for (i = 0; i < npages;) {
> +		if (!(pfns[i] & HMM_PFN_VALID)) {
> +			err = -EFAULT;
> +			goto err_free;
> +		}
> +		i += 0x1 << hmm_pfn_to_map_order(pfns[i]);
> +	}
> +
> +err_free:
> +	kvfree(pfns);
> +	return err ? false : true;
> +}
> +
> +/**
> + * drm_gpusvm_range_chunk_size() - Determine chunk size for GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier structure
> + * @vas: Pointer to the virtual memory area structure
> + * @fault_addr: Fault address
> + * @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
> + *
> + * 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
> + * memory area boundaries.
> + *
> + * Returns:
> + * Chunk size on success, LONG_MAX on failure.
> + */
> +static unsigned long
> +drm_gpusvm_range_chunk_size(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_notifier *notifier,
> +			    struct vm_area_struct *vas,
> +			    unsigned long fault_addr,
> +			    unsigned long gpuva_start,
> +			    unsigned long gpuva_end,
> +			    unsigned long check_pages_threshold)
> +{
> +	unsigned long start, end;
> +	int i = 0;
> +
> +retry:
> +	for (; i < gpusvm->num_chunks; ++i) {
> +		start = ALIGN_DOWN(fault_addr, gpusvm->chunk_sizes[i]);
> +		end = ALIGN(fault_addr + 1, gpusvm->chunk_sizes[i]);
> +
> +		if (start >= vas->vm_start && end <= vas->vm_end &&
> +		    start >= notifier->itree.start &&
> +		    end <= notifier->itree.last + 1 &&
> +		    start >= gpuva_start && end <= gpuva_end)
> +			break;
> +	}
> +
> +	if (i == gpusvm->num_chunks)
> +		return LONG_MAX;
> +
> +	/*
> +	 * If allocation more than page, ensure not to overlap with existing
> +	 * ranges.
> +	 */
> +	if (end - start != SZ_4K) {
> +		struct drm_gpusvm_range *range;
> +
> +		range = drm_gpusvm_range_find(notifier, start, end);
> +		if (range) {
> +			++i;
> +			goto retry;
> +		}
> +
> +		/*
> +		 * XXX: Only create range on pages CPU has faulted in. Without
> +		 * this check, or prefault, on BMG 'xe_exec_system_allocator --r
> +		 * process-many-malloc' fails. In the failure case, each process
> +		 * mallocs 16k but the CPU VMA is ~128k which results in 64k SVM
> +		 * ranges. When migrating the SVM ranges, some processes fail in
> +		 * drm_gpusvm_migrate_to_devmem with 'migrate.cpages != npages'
> +		 * and then upon drm_gpusvm_range_get_pages device pages from
> +		 * other processes are collected + faulted in which creates all
> +		 * sorts of problems. Unsure exactly how this happening, also
> +		 * problem goes away if 'xe_exec_system_allocator --r
> +		 * process-many-malloc' mallocs at least 64k at a time.
> +		 */
> +		if (end - start <= check_pages_threshold &&
> +		    !drm_gpusvm_check_pages(gpusvm, notifier, start, end)) {
> +			++i;
> +			goto retry;
> +		}
> +	}
> +
> +	return end - start;
> +}
> +
> +/**
> + * drm_gpusvm_range_find_or_insert() - Find or insert GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @fault_addr: Fault address
> + * @gpuva_start: Start address of GPUVA which mirrors CPU
> + * @gpuva_end: End address of GPUVA which mirrors CPU
> + * @ctx: GPU SVM context
> + *
> + * This function finds or inserts a newly allocated a GPU SVM range based on the
> + * fault address. Caller must hold a lock to protect range lookup and insertion.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range on success, ERR_PTR() on failure.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +	struct drm_gpusvm_range *range;
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	bool notifier_alloc = false;
> +	unsigned long chunk_size;
> +	int err;
> +	bool migrate_devmem;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	if (fault_addr < gpusvm->mm_start ||
> +	    fault_addr > gpusvm->mm_start + gpusvm->mm_range)
> +		return ERR_PTR(-EINVAL);
> +
> +	if (!mmget_not_zero(mm))
> +		return ERR_PTR(-EFAULT);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, fault_addr);
> +	if (!notifier) {
> +		notifier = drm_gpusvm_notifier_alloc(gpusvm, fault_addr);
> +		if (IS_ERR(notifier)) {
> +			err = PTR_ERR(notifier);
> +			goto err_mmunlock;
> +		}
> +		notifier_alloc = true;
> +		err = mmu_interval_notifier_insert(&notifier->notifier,
> +						   mm, notifier->itree.start,
> +						   notifier->itree.last -
> +						   notifier->itree.start + 1,
> +						   &drm_gpusvm_notifier_ops);
> +		if (err)
> +			goto err_notifier;
> +	}
> +
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, fault_addr);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_notifier_remove;
> +	}
> +
> +	if (!ctx->read_only && !(vas->vm_flags & VM_WRITE)) {
> +		err = -EPERM;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_find(notifier, fault_addr, fault_addr + 1);
> +	if (range)
> +		goto out_mmunlock;
> +	/*
> +	 * XXX: Short-circuiting migration based on migrate_vma_* current
> +	 * limitations. If/when migrate_vma_* add more support, this logic will
> +	 * have to change.
> +	 */
> +	migrate_devmem = ctx->devmem_possible &&
> +		vma_is_anonymous(vas) && !is_vm_hugetlb_page(vas);
> +
> +	chunk_size = drm_gpusvm_range_chunk_size(gpusvm, notifier, vas,
> +						 fault_addr, gpuva_start,
> +						 gpuva_end, migrate_devmem &&
> +						 ctx->check_pages_threshold);
> +	if (chunk_size == LONG_MAX) {
> +		err = -EINVAL;
> +		goto err_notifier_remove;
> +	}
> +
> +	range = drm_gpusvm_range_alloc(gpusvm, notifier, fault_addr, chunk_size,
> +				       migrate_devmem);
> +	if (IS_ERR(range)) {
> +		err = PTR_ERR(range);
> +		goto err_notifier_remove;
> +	}
> +
> +	drm_gpusvm_range_insert(notifier, range);
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_insert(gpusvm, notifier);
> +
> +out_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +
> +	return range;
> +
> +err_notifier_remove:
> +	mmap_read_unlock(mm);
> +	if (notifier_alloc)
> +		mmu_interval_notifier_remove(&notifier->notifier);
> +err_notifier:
> +	if (notifier_alloc)
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +err_mmunlock:
> +	mmput(mm);
> +	return ERR_PTR(err);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_find_or_insert);
> +
> +/**
> + * __drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range (internal)
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @npages: Number of pages to unmap
> + *
> + * This function unmap pages associated with a GPU SVM range. Assumes and
> + * asserts correct locking is in place when called.
> + */
> +static void __drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +					   struct drm_gpusvm_range *range,
> +					   unsigned long npages)
> +{
> +	unsigned long i, j;
> +	struct drm_pagemap *dpagemap = range->dpagemap;
> +	struct device *dev = gpusvm->drm->dev;
> +
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->flags.has_dma_mapping) {
> +		for (i = 0, j = 0; i < npages; j++) {
> +			struct drm_pagemap_dma_addr *addr = &range->dma_addr[j];
> +
> +			if (addr->proto == DRM_INTERCONNECT_SYSTEM)
> +				dma_unmap_page(dev,
> +					       addr->addr,
> +					       PAGE_SIZE << addr->order,
> +					       addr->dir);
> +			else if (dpagemap && dpagemap->ops->unmap_dma)
> +				dpagemap->ops->unmap_dma(dpagemap,
> +							 dev,
> +							 *addr);
> +			i += 1 << addr->order;
> +		}
> +		range->flags.has_devmem_pages = false;
> +		range->flags.has_dma_mapping = false;
> +		range->dpagemap = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_free_pages() - Free pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function frees the dma address array associated with a GPU SVM range.
> + */
> +static void drm_gpusvm_range_free_pages(struct drm_gpusvm *gpusvm,
> +					struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	if (range->dma_addr) {
> +		kvfree(range->dma_addr);
> +		range->dma_addr = NULL;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_range_remove() - Remove GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function removes the specified GPU SVM range and also removes the parent
> + * GPU SVM notifier if no more ranges remain in the notifier. The caller must
> + * hold a lock to protect range and notifier removal.
> + */
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_driver_lock_held(gpusvm);
> +
> +	notifier = drm_gpusvm_notifier_find(gpusvm, range->itree.start);
> +	if (WARN_ON_ONCE(!notifier))
> +		return;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +	drm_gpusvm_range_free_pages(gpusvm, range);
> +	__drm_gpusvm_range_remove(notifier, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	drm_gpusvm_range_put(range);
> +
> +	if (RB_EMPTY_ROOT(&notifier->root.rb_root)) {
> +		if (!notifier->flags.removed)
> +			mmu_interval_notifier_remove(&notifier->notifier);
> +		drm_gpusvm_notifier_remove(gpusvm, notifier);
> +		drm_gpusvm_notifier_free(gpusvm, notifier);
> +	}
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_remove);
> +
> +/**
> + * drm_gpusvm_range_get() - Get a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function increments the reference count of the specified GPU SVM range.
> + *
> + * Returns:
> + * Pointer to the GPU SVM range.
> + */
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range)
> +{
> +	kref_get(&range->refcount);
> +
> +	return range;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get);
> +
> +/**
> + * drm_gpusvm_range_destroy() - Destroy GPU SVM range
> + * @refcount: Pointer to the reference counter embedded in the GPU SVM range
> + *
> + * This function destroys the specified GPU SVM range when its reference count
> + * reaches zero. If a custom range-free function is provided, it is invoked to
> + * free the range; otherwise, the range is deallocated using kfree().
> + */
> +static void drm_gpusvm_range_destroy(struct kref *refcount)
> +{
> +	struct drm_gpusvm_range *range =
> +		container_of(refcount, struct drm_gpusvm_range, refcount);
> +	struct drm_gpusvm *gpusvm = range->gpusvm;
> +
> +	if (gpusvm->ops->range_free)
> +		gpusvm->ops->range_free(range);
> +	else
> +		kfree(range);
> +}
> +
> +/**
> + * drm_gpusvm_range_put() - Put a reference to GPU SVM range
> + * @range: Pointer to the GPU SVM range
> + *
> + * This function decrements the reference count of the specified GPU SVM range
> + * and frees it when the count reaches zero.
> + */
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range)
> +{
> +	kref_put(&range->refcount, drm_gpusvm_range_destroy);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_put);
> +
> +/**
> + * drm_gpusvm_range_pages_valid() - GPU SVM range pages valid
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called holding gpusvm->notifier_lock and as the last step before committing a
> + * GPU binding. This is akin to a notifier seqno check in the HMM documentation
> + * but due to wider notifiers (i.e., notifiers which span multiple ranges) this
> + * function is required for finer grained checking (i.e., per range) if pages
> + * are valid.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range)
> +{
> +	lockdep_assert_held(&gpusvm->notifier_lock);
> +
> +	return range->flags.has_devmem_pages || range->flags.has_dma_mapping;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_pages_valid);
> +
> +/**
> + * drm_gpusvm_range_pages_valid_unlocked() - GPU SVM range pages valid unlocked
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + *
> + * This function determines if a GPU SVM range pages are valid. Expected be
> + * called without holding gpusvm->notifier_lock.
> + *
> + * Returns:
> + * True if GPU SVM range has valid pages, False otherwise
> + */
> +static bool
> +drm_gpusvm_range_pages_valid_unlocked(struct drm_gpusvm *gpusvm,
> +				      struct drm_gpusvm_range *range)
> +{
> +	bool pages_valid;
> +
> +	if (!range->dma_addr)
> +		return false;
> +
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	pages_valid = drm_gpusvm_range_pages_valid(gpusvm, range);
> +	if (!pages_valid)
> +		drm_gpusvm_range_free_pages(gpusvm, range);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +
> +	return pages_valid;
> +}
> +
> +/**
> + * drm_gpusvm_range_get_pages() - Get pages for a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function gets pages for a GPU SVM range and ensures they are mapped for
> + * DMA access.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx)
> +{
> +	struct mmu_interval_notifier *notifier = &range->notifier->notifier;
> +	struct hmm_range hmm_range = {
> +		.default_flags = HMM_PFN_REQ_FAULT | (ctx->read_only ? 0 :
> +			HMM_PFN_REQ_WRITE),
> +		.notifier = notifier,
> +		.start = range->itree.start,
> +		.end = range->itree.last + 1,
> +		.dev_private_owner = gpusvm->device_private_page_owner,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct drm_gpusvm_zdd *zdd;
> +	unsigned long timeout =
> +		jiffies + msecs_to_jiffies(HMM_RANGE_DEFAULT_TIMEOUT);
> +	unsigned long i, j;
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +	unsigned long num_dma_mapped;
> +	unsigned int order = 0;
> +	unsigned long *pfns;
> +	struct page **pages;
> +	int err = 0;
> +	struct dev_pagemap *pagemap;
> +	struct drm_pagemap *dpagemap;
> +
> +retry:
> +	hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +	if (drm_gpusvm_range_pages_valid_unlocked(gpusvm, range))
> +		goto set_seqno;
> +
> +	pfns = kvmalloc_array(npages, sizeof(*pfns), GFP_KERNEL);
> +	if (!pfns)
> +		return -ENOMEM;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	hmm_range.hmm_pfns = pfns;
> +	while (true) {
> +		mmap_read_lock(mm);
> +		err = hmm_range_fault(&hmm_range);
> +		mmap_read_unlock(mm);
> +
> +		if (err == -EBUSY) {
> +			if (time_after(jiffies, timeout))
> +				break;
> +
> +			hmm_range.notifier_seq = mmu_interval_read_begin(notifier);
> +			continue;
> +		}
> +		break;
> +	}
> +	mmput(mm);
> +	if (err)
> +		goto err_free;
> +
> +	pages = (struct page **)pfns;
What is the purpose of this code?
> +map_pages:
> +	/*
> +	 * Perform all dma mappings under the notifier lock to not
> +	 * access freed pages. A notifier will either block on
> +	 * the notifier lock or unmap dma.
> +	 */
> +	drm_gpusvm_notifier_lock(gpusvm);
> +	if (mmu_interval_read_retry(notifier, hmm_range.notifier_seq)) {
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		kvfree(pfns);
> +		goto retry;
> +	}
> +
> +	if (!range->dma_addr) {
> +		/* Unlock and restart mapping to allocate memory. */
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +		range->dma_addr = kvmalloc_array(npages, sizeof(*range->dma_addr),
> +						 GFP_KERNEL);
> +		if (!range->dma_addr) {
> +			err = -ENOMEM;
> +			goto err_free;
> +		}
> +		goto map_pages;
> +	}
> +
> +	zdd = NULL;
> +	num_dma_mapped = 0;
> +	for (i = 0, j = 0; i < npages; ++j) {
> +		struct page *page = hmm_pfn_to_page(pfns[i]);
> +
> +		order = hmm_pfn_to_map_order(pfns[i]);
> +		if (is_device_private_page(page) || is_device_coherent_page(page)) {
> +			if (zdd != page->zone_device_data && i > 0) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +			zdd = page->zone_device_data;
> +			if (pagemap != page->pgmap) {
> +				if (i > 0) {
> +					err = -EOPNOTSUPP;
> +					goto err_unmap;
> +				}
> +
> +				pagemap = page->pgmap;
> +				dpagemap = zdd->devmem_allocation->dpagemap;
> +				if (drm_WARN_ON(gpusvm->drm, !dpagemap)) {
> +					/*
> +					 * Raced. This is not supposed to happen
> +					 * since hmm_range_fault() should've migrated
> +					 * this page to system.
> +					 */
> +					err = -EAGAIN;
> +					goto err_unmap;
> +				}
> +			}
> +			range->dma_addr[j] =
> +				dpagemap->ops->map_dma(dpagemap, gpusvm->drm->dev,
> +						       page, order,
> +						       DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, range->dma_addr[j].addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			pages[i] = page;
what is the effect of this code?
> +		} else {
> +			dma_addr_t addr;
> +
> +			if (is_zone_device_page(page) || zdd) {
> +				err = -EOPNOTSUPP;
> +				goto err_unmap;
> +			}
> +
> +			addr = dma_map_page(gpusvm->drm->dev,
> +					    page, 0,
> +					    PAGE_SIZE << order,
> +					    DMA_BIDIRECTIONAL);
> +			if (dma_mapping_error(gpusvm->drm->dev, addr)) {
> +				err = -EFAULT;
> +				goto err_unmap;
> +			}
> +
> +			range->dma_addr[j] = drm_pagemap_dma_addr_encode
> +				(addr, DRM_INTERCONNECT_SYSTEM, order,
> +				 DMA_BIDIRECTIONAL);
> +		}
> +		i += 1 << order;
> +		num_dma_mapped = i;
> +	}
> +
> +	range->flags.has_dma_mapping = true;
> +	if (zdd) {
> +		range->flags.has_devmem_pages = true;
> +		range->dpagemap = dpagemap;
> +	}
> +
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +	kvfree(pfns);
> +set_seqno:
> +	range->notifier_seq = hmm_range.notifier_seq;
> +
> +	return 0;
> +
> +err_unmap:
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, num_dma_mapped);
> +	drm_gpusvm_notifier_unlock(gpusvm);
> +err_free:
> +	kvfree(pfns);
> +	if (err == -EAGAIN)
> +		goto retry;
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_get_pages);
> +
> +/**
> + * drm_gpusvm_range_unmap_pages() - Unmap pages associated with a GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @ctx: GPU SVM context
> + *
> + * This function unmaps pages associated with a GPU SVM range. If @in_notifier
> + * is set, it is assumed that gpusvm->notifier_lock is held in write mode; if it
> + * is clear, it acquires gpusvm->notifier_lock in read mode. Must be called on
> + * each GPU SVM range attached to notifier in gpusvm->ops->invalidate for IOMMU
> + * security model.
> + */
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx)
> +{
> +	unsigned long npages = npages_in_range(range->itree.start,
> +					       range->itree.last + 1);
> +
> +	if (ctx->in_notifier)
> +		lockdep_assert_held_write(&gpusvm->notifier_lock);
> +	else
> +		drm_gpusvm_notifier_lock(gpusvm);
> +
> +	__drm_gpusvm_range_unmap_pages(gpusvm, range, npages);
> +
> +	if (!ctx->in_notifier)
> +		drm_gpusvm_notifier_unlock(gpusvm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_unmap_pages);
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_page() - Put a migration page
> + * @page: Pointer to the page to put
> + *
> + * This function unlocks and puts a page.
> + */
> +static void drm_gpusvm_migration_unlock_put_page(struct page *page)
> +{
> +	unlock_page(page);
> +	put_page(page);
> +}
> +
> +/**
> + * drm_gpusvm_migration_unlock_put_pages() - Put migration pages
> + * @npages: Number of pages
> + * @migrate_pfn: Array of migrate page frame numbers
> + *
> + * This function unlocks and puts an array of pages.
> + */
> +static void drm_gpusvm_migration_unlock_put_pages(unsigned long npages,
> +						  unsigned long *migrate_pfn)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page;
> +
> +		if (!migrate_pfn[i])
> +			continue;
> +
> +		page = migrate_pfn_to_page(migrate_pfn[i]);
> +		drm_gpusvm_migration_unlock_put_page(page);
> +		migrate_pfn[i] = 0;
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_get_devmem_page() - Get a reference to a device memory page
> + * @page: Pointer to the page
> + * @zdd: Pointer to the GPU SVM zone device data
> + *
> + * This function associates the given page with the specified GPU SVM zone
> + * device data and initializes it for zone device usage.
> + */
> +static void drm_gpusvm_get_devmem_page(struct page *page,
> +				     struct drm_gpusvm_zdd *zdd)
> +{
> +	page->zone_device_data = drm_gpusvm_zdd_get(zdd);
> +	zone_device_page_init(page);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_map_pages() - Map migration pages for GPU SVM migration
> + * @dev: The device for which the pages are being mapped
> + * @dma_addr: Array to store DMA addresses corresponding to mapped pages
> + * @migrate_pfn: Array of migrate page frame numbers to map
> + * @npages: Number of pages to map
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function maps pages of memory for migration usage in GPU SVM. It
> + * iterates over each page frame number provided in @migrate_pfn, maps the
> + * corresponding page, and stores the DMA address in the provided @dma_addr
> + * array.
> + *
> + * Return: 0 on success, -EFAULT if an error occurs during mapping.
> + */
> +static int drm_gpusvm_migrate_map_pages(struct device *dev,
> +					dma_addr_t *dma_addr,
> +					unsigned long *migrate_pfn,
> +					unsigned long npages,
> +					enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(migrate_pfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		if (WARN_ON_ONCE(is_zone_device_page(page)))
> +			return -EFAULT;
> +
> +		dma_addr[i] = dma_map_page(dev, page, 0, PAGE_SIZE, dir);
> +		if (dma_mapping_error(dev, dma_addr[i]))
> +			return -EFAULT;
> +	}
> +
> +	return 0;
> +}
> +
> +/**
> + * drm_gpusvm_migrate_unmap_pages() - Unmap pages previously mapped for GPU SVM migration
> + * @dev: The device for which the pages were mapped
> + * @dma_addr: Array of DMA addresses corresponding to mapped pages
> + * @npages: Number of pages to unmap
> + * @dir: Direction of data transfer (e.g., DMA_BIDIRECTIONAL)
> + *
> + * This function unmaps previously mapped pages of memory for GPU Shared Virtual
> + * Memory (SVM). It iterates over each DMA address provided in @dma_addr, checks
> + * if it's valid and not already unmapped, and unmaps the corresponding page.
> + */
> +static void drm_gpusvm_migrate_unmap_pages(struct device *dev,
> +					   dma_addr_t *dma_addr,
> +					   unsigned long npages,
> +					   enum dma_data_direction dir)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i) {
> +		if (!dma_addr[i] || dma_mapping_error(dev, dma_addr[i]))
> +			continue;
> +
> +		dma_unmap_page(dev, dma_addr[i], PAGE_SIZE, dir);
> +	}
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_devmem() - Migrate GPU SVM range to device memory
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range structure
> + * @devmem_allocation: Pointer to the device memory allocation. The caller
> + *                     should hold a reference to the device memory allocation,
> + *                     which should be dropped via ops->devmem_release or upon
> + *                     the failure of this function.
> + * @ctx: GPU SVM context
> + *
> + * This function migrates the specified GPU SVM range to device memory. It performs the
> + * necessary setup and invokes the driver-specific operations for migration to
> + * device memory. Upon successful return, @devmem_allocation can safely reference @range
> + * until ops->devmem_release is called which only upon successful return.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +	struct migrate_vma migrate = {
> +		.start		= start,
> +		.end		= end,
> +		.pgmap_owner	= gpusvm->device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_SYSTEM,
> +	};
> +	struct mm_struct *mm = gpusvm->mm;
> +	unsigned long i, npages = npages_in_range(start, end);
> +	struct vm_area_struct *vas;
> +	struct drm_gpusvm_zdd *zdd = NULL;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int err;
> +
> +	if (!range->flags.migrate_devmem)
> +		return -EINVAL;
> +
> +	if (!ops->populate_devmem_pfn || !ops->copy_to_devmem ||
> +	    !ops->copy_to_ram)
> +		return -EOPNOTSUPP;
> +
> +	if (!mmget_not_zero(mm)) {
> +		err = -EFAULT;
> +		goto err_out;
> +	}
> +	mmap_read_lock(mm);
> +
> +	vas = vma_lookup(mm, start);
> +	if (!vas) {
> +		err = -ENOENT;
> +		goto err_mmunlock;
> +	}
> +
> +	if (end > vas->vm_end || start < vas->vm_start) {
> +		err = -EINVAL;
> +		goto err_mmunlock;
> +	}
> +
> +	if (!vma_is_anonymous(vas)) {
> +		err = -EBUSY;
> +		goto err_mmunlock;
> +	}
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_mmunlock;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	zdd = drm_gpusvm_zdd_alloc(gpusvm->device_private_page_owner);
> +	if (!zdd) {
> +		err = -ENOMEM;
> +		goto err_free;
> +	}
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	if (!migrate.cpages) {
> +		err = -EFAULT;
> +		goto err_free;
> +	}
> +
> +	if (migrate.cpages != npages) {
> +		err = -EBUSY;
> +		goto err_finalize;
> +	}
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, migrate.dst);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   migrate.src, npages, DMA_TO_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = pfn_to_page(migrate.dst[i]);
> +
> +		pages[i] = page;
> +		migrate.dst[i] = migrate_pfn(migrate.dst[i]);
> +		drm_gpusvm_get_devmem_page(page, zdd);
> +	}
> +
> +	err = ops->copy_to_devmem(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +	/* Upon success bind devmem allocation to range and zdd */
> +	WRITE_ONCE(zdd->devmem_allocation, devmem_allocation);	/* Owns ref */
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_TO_DEVICE);
> +err_free:
> +	if (zdd)
> +		drm_gpusvm_zdd_put(zdd);
> +	kvfree(buf);
> +err_mmunlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +err_out:
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_migrate_to_devmem);
> +
> +/**
> + * drm_gpusvm_migrate_populate_ram_pfn() - Populate RAM PFNs for a VM area
> + * @vas: Pointer to the VM area structure, can be NULL
> + * @fault_page: Fault page
> + * @npages: Number of pages to populate
> + * @mpages: Number of pages to migrate
> + * @src_mpfn: Source array of migrate PFNs
> + * @mpfn: Array of migrate PFNs to populate
> + * @addr: Start address for PFN allocation
> + *
> + * This function populates the RAM migrate page frame numbers (PFNs) for the
> + * specified VM area structure. It allocates and locks pages in the VM area for
> + * RAM usage. If vas is non-NULL use alloc_page_vma for allocation, if NULL use
> + * alloc_page for allocation.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int drm_gpusvm_migrate_populate_ram_pfn(struct vm_area_struct *vas,
> +					       struct page *fault_page,
> +					       unsigned long npages,
> +					       unsigned long *mpages,
> +					       unsigned long *src_mpfn,
> +					       unsigned long *mpfn,
> +					       unsigned long addr)
> +{
> +	unsigned long i;
> +
> +	for (i = 0; i < npages; ++i, addr += PAGE_SIZE) {
> +		struct page *page, *src_page;
> +
> +		if (!(src_mpfn[i] & MIGRATE_PFN_MIGRATE))
> +			continue;
> +
> +		if (fault_page) {
> +			src_page = migrate_pfn_to_page(src_mpfn[i]);
> +			if (src_page->zone_device_data !=
> +			    fault_page->zone_device_data)
> +				continue;
> +		}
> +
> +		if (vas)
> +			page = alloc_page_vma(GFP_HIGHUSER, vas, addr);
> +		else
> +			page = alloc_page(GFP_HIGHUSER);
> +
> +		if (!page)
> +			goto free_pages;
> +
> +		mpfn[i] = migrate_pfn(page_to_pfn(page));
> +	}
> +
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		WARN_ON_ONCE(!trylock_page(page));
> +		++*mpages;
> +	}
> +
> +	return 0;
> +
> +free_pages:
> +	for (i = 0; i < npages; ++i) {
> +		struct page *page = migrate_pfn_to_page(mpfn[i]);
> +
> +		if (!page)
> +			continue;
> +
> +		put_page(page);
> +		mpfn[i] = 0;
> +	}
> +	return -ENOMEM;
> +}
> +
> +/**
> + * drm_gpusvm_evict_to_ram() - Evict GPU SVM range to RAM
> + * @devmem_allocation: Pointer to the device memory allocation
> + *
> + * Similar to __drm_gpusvm_migrate_to_ram but does not require mmap lock and
> + * migration done via migrate_device_* functions.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
Hi,

This patch series does not have any code that calls this 
drm_gpusvm_evict_to_ram() API (it is not used yet).
Will you use drm_gpusvm_evict_to_ram() in future patches? How did you 
verify the operation of this code at this stage of the patch series?
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation)
> +{
> +	const struct drm_gpusvm_devmem_ops *ops = devmem_allocation->ops;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	unsigned long *src, *dst;
> +	dma_addr_t *dma_addr;
> +	void *buf;
> +	int i, err = 0;
> +
> +	npages = devmem_allocation->size >> PAGE_SHIFT;
> +
> +retry:
> +	if (!mmget_not_zero(devmem_allocation->mm))
> +		return -EFAULT;
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	src = buf;
> +	dst = buf + (sizeof(*src) * npages);
> +	dma_addr = buf + (2 * sizeof(*src) * npages);
> +	pages = buf + (2 * sizeof(*src) + sizeof(*dma_addr)) * npages;
> +
> +	err = ops->populate_devmem_pfn(devmem_allocation, npages, src);
> +	if (err)
> +		goto err_free;
> +
> +	err = migrate_device_pfns(src, npages);
> +	if (err)
> +		goto err_free;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(NULL, NULL, npages, &mpages,
> +						  src, dst, 0);
> +	if (err || !mpages)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(devmem_allocation->dev, dma_addr,
> +					   dst, npages, DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, dst);
> +	migrate_device_pages(src, dst, npages);
> +	migrate_device_finalize(src, dst, npages);
> +	drm_gpusvm_migrate_unmap_pages(devmem_allocation->dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +	mmput_async(devmem_allocation->mm);
> +	if (!err && !READ_ONCE(devmem_allocation->detached)) {
> +		cond_resched();
> +		goto retry;
> +	}
> +
> +	return err;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_evict_to_ram);
> +
> +/**
> + * __drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (internal)
> + * @vas: Pointer to the VM area structure
> + * @device_private_page_owner: Device private pages owner
> + * @page: Pointer to the page for fault handling (can be NULL)
> + * @fault_addr: Fault address
> + * @size: Size of migration
> + *
> + * This internal function performs the migration of the specified GPU SVM range
> + * to RAM. It sets up the migration, populates + dma maps RAM PFNs, and
> + * invokes the driver-specific operations for migration to RAM.
> + *
> + * Returns:
> + * 0 on success, negative error code on failure.
> + */
> +static int __drm_gpusvm_migrate_to_ram(struct vm_area_struct *vas,
> +				       void *device_private_page_owner,
> +				       struct page *page,
> +				       unsigned long fault_addr,
> +				       unsigned long size)
> +{
> +	struct migrate_vma migrate = {
> +		.vma		= vas,
> +		.pgmap_owner	= device_private_page_owner,
> +		.flags		= MIGRATE_VMA_SELECT_DEVICE_PRIVATE |
> +			MIGRATE_VMA_SELECT_DEVICE_COHERENT,
> +		.fault_page	= page,
> +	};
> +	struct drm_gpusvm_zdd *zdd;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct device *dev;
> +	unsigned long npages, mpages = 0;
> +	struct page **pages;
> +	dma_addr_t *dma_addr;
> +	unsigned long start, end;
> +	void *buf;
> +	int i, err = 0;
> +
> +	start = ALIGN_DOWN(fault_addr, size);
> +	end = ALIGN(fault_addr + 1, size);
> +
> +	/* Corner where VMA area struct has been partially unmapped */
> +	if (start < vas->vm_start)
> +		start = vas->vm_start;
> +	if (end > vas->vm_end)
> +		end = vas->vm_end;
> +
> +	migrate.start = start;
> +	migrate.end = end;
> +	npages = npages_in_range(start, end);
> +
> +	buf = kvcalloc(npages, 2 * sizeof(*migrate.src) + sizeof(*dma_addr) +
> +		       sizeof(*pages), GFP_KERNEL);
> +	if (!buf) {
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +	dma_addr = buf + (2 * sizeof(*migrate.src) * npages);
> +	pages = buf + (2 * sizeof(*migrate.src) + sizeof(*dma_addr)) * npages;
> +
> +	migrate.vma = vas;
> +	migrate.src = buf;
> +	migrate.dst = migrate.src + npages;
> +
> +	err = migrate_vma_setup(&migrate);
> +	if (err)
> +		goto err_free;
> +
> +	/* Raced with another CPU fault, nothing to do */
> +	if (!migrate.cpages)
> +		goto err_free;
> +
> +	if (!page) {
> +		for (i = 0; i < npages; ++i) {
> +			if (!(migrate.src[i] & MIGRATE_PFN_MIGRATE))
> +				continue;
> +
> +			page = migrate_pfn_to_page(migrate.src[i]);
> +			break;
> +		}
> +
> +		if (!page)
> +			goto err_finalize;
> +	}
> +	zdd = page->zone_device_data;
> +	ops = zdd->devmem_allocation->ops;
> +	dev = zdd->devmem_allocation->dev;
> +
> +	err = drm_gpusvm_migrate_populate_ram_pfn(vas, page, npages, &mpages,
> +						  migrate.src, migrate.dst,
> +						  start);
> +	if (err)
> +		goto err_finalize;
> +
> +	err = drm_gpusvm_migrate_map_pages(dev, dma_addr, migrate.dst, npages,
> +					   DMA_FROM_DEVICE);
> +	if (err)
> +		goto err_finalize;
> +
> +	for (i = 0; i < npages; ++i)
> +		pages[i] = migrate_pfn_to_page(migrate.src[i]);
> +
> +	err = ops->copy_to_ram(pages, dma_addr, npages);
> +	if (err)
> +		goto err_finalize;
> +
> +err_finalize:
> +	if (err)
> +		drm_gpusvm_migration_unlock_put_pages(npages, migrate.dst);
> +	migrate_vma_pages(&migrate);
you must check if the migration was successful after calling 
migrate_vma_pages().
please refer use of svm_migrate_unsuccessful_pages() of kfd_migrate.c

Br,
G.G.
> +	migrate_vma_finalize(&migrate);
> +	drm_gpusvm_migrate_unmap_pages(dev, dma_addr, npages,
> +				       DMA_FROM_DEVICE);
> +err_free:
> +	kvfree(buf);
> +err_out:
> +
> +	return err;
> +}
> +
> +/**
> + * drm_gpusvm_range_evict() - Evict GPU SVM range
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @range: Pointer to the GPU SVM range to be removed
> + *
> + * This function evicts the specified GPU SVM range.
> + */
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range)
> +{
> +	struct mm_struct *mm = gpusvm->mm;
> +	struct vm_area_struct *vas;
> +	unsigned long start = range->itree.start, end = range->itree.last + 1;
> +
> +	if (!mmget_not_zero(mm))
> +		return;
> +
> +	mmap_read_lock(mm);
> +
> +	/*
> +	 * Loop required to find all VMAs for the corner case when partial
> +	 * unmappings have occurred.
> +	 */
> +again:
> +	vas = find_vma(mm, start);
> +	if (!vas)
> +		goto unlock;
> +
> +	if (end <= vas->vm_start || start >= vas->vm_end)
> +		goto unlock;
> +
> +	__drm_gpusvm_migrate_to_ram(vas, gpusvm->device_private_page_owner,
> +				    NULL, start,
> +				    range->itree.last - range->itree.start + 1);
> +	if (vas->vm_end < end) {
> +		start = vas->vm_end;
> +		goto again;
> +	}
> +unlock:
> +	mmap_read_unlock(mm);
> +	mmput(mm);
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_evict);
> +
> +/**
> + * drm_gpusvm_page_free() - Put GPU SVM zone device data associated with a page
> + * @page: Pointer to the page
> + *
> + * This function is a callback used to put the GPU SVM zone device data
> + * associated with a page when it is being released.
> + */
> +static void drm_gpusvm_page_free(struct page *page)
> +{
> +	drm_gpusvm_zdd_put(page->zone_device_data);
> +}
> +
> +/**
> + * drm_gpusvm_migrate_to_ram() - Migrate GPU SVM range to RAM (page fault handler)
> + * @vmf: Pointer to the fault information structure
> + *
> + * This function is a page fault handler used to migrate a GPU SVM range to RAM.
> + * It retrieves the GPU SVM range information from the faulting page and invokes
> + * the internal migration function to migrate the range back to RAM.
> + *
> + * Returns:
> + * VM_FAULT_SIGBUS on failure, 0 on success.
> + */
> +static vm_fault_t drm_gpusvm_migrate_to_ram(struct vm_fault *vmf)
> +{
> +	struct drm_gpusvm_zdd *zdd = vmf->page->zone_device_data;
> +	int err;
> +
> +	err = __drm_gpusvm_migrate_to_ram(vmf->vma,
> +					  zdd->device_private_page_owner,
> +					  vmf->page, vmf->address,
> +					  zdd->devmem_allocation->size);
> +
> +	return err ? VM_FAULT_SIGBUS : 0;
> +}
> +
> +/**
> + * drm_gpusvm_pagemap_ops() - Device page map operations for GPU SVM
> + */
> +static const struct dev_pagemap_ops drm_gpusvm_pagemap_ops = {
> +	.page_free = drm_gpusvm_page_free,
> +	.migrate_to_ram = drm_gpusvm_migrate_to_ram,
> +};
> +
> +/**
> + * drm_gpusvm_pagemap_ops_get() - Retrieve GPU SVM device page map operations
> + *
> + * Returns:
> + * Pointer to the GPU SVM device page map operations structure.
> + */
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void)
> +{
> +	return &drm_gpusvm_pagemap_ops;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_pagemap_ops_get);
> +
> +/**
> + * drm_gpusvm_has_mapping() - Check if GPU SVM has mapping for the given address range
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @start: Start address
> + * @end: End address
> + *
> + * Returns:
> + * True if GPU SVM has mapping, False otherwise
> + */
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end)
> +{
> +	struct drm_gpusvm_notifier *notifier;
> +
> +	drm_gpusvm_for_each_notifier(notifier, gpusvm, start, end) {
> +		struct drm_gpusvm_range *range = NULL;
> +
> +		drm_gpusvm_for_each_range(range, notifier, start, end)
> +			return true;
> +	}
> +
> +	return false;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_has_mapping);
> +
> +/**
> + * drm_gpusvm_range_set_unmapped() - Mark a GPU SVM range as unmapped
> + * @range: Pointer to the GPU SVM range structure.
> + * @mmu_range: Pointer to the MMU notifier range structure.
> + *
> + * This function marks a GPU SVM range as unmapped and sets the partial_unmap flag
> + * if the range partially falls within the provided MMU notifier range.
> + */
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range)
> +{
> +	lockdep_assert_held_write(&range->gpusvm->notifier_lock);
> +
> +	range->flags.unmapped = true;
> +	if (range->itree.start < mmu_range->start ||
> +	    range->itree.last + 1 > mmu_range->end)
> +		range->flags.partial_unmap = true;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_range_set_unmapped);
> +
> +/**
> + * drm_gpusvm_devmem_init() - Initialize a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap we're allocating from.
> + * @size: Size of device memory allocation
> + */
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size)
> +{
> +	devmem_allocation->dev = dev;
> +	devmem_allocation->mm = mm;
> +	devmem_allocation->ops = ops;
> +	devmem_allocation->dpagemap = dpagemap;
> +	devmem_allocation->size = size;
> +}
> +EXPORT_SYMBOL_GPL(drm_gpusvm_devmem_init);
> +
> +MODULE_DESCRIPTION("DRM GPUSVM");
> +MODULE_LICENSE("GPL");
> diff --git a/include/drm/drm_gpusvm.h b/include/drm/drm_gpusvm.h
> new file mode 100644
> index 000000000000..2fb6980d0cf1
> --- /dev/null
> +++ b/include/drm/drm_gpusvm.h
> @@ -0,0 +1,445 @@
> +/* SPDX-License-Identifier: GPL-2.0-only OR MIT */
> +/*
> + * Copyright © 2024 Intel Corporation
> + */
> +
> +#ifndef __DRM_GPUSVM_H__
> +#define __DRM_GPUSVM_H__
> +
> +#include <linux/kref.h>
> +#include <linux/interval_tree.h>
> +#include <linux/mmu_notifier.h>
> +
> +struct dev_pagemap_ops;
> +struct drm_device;
> +struct drm_gpusvm;
> +struct drm_gpusvm_notifier;
> +struct drm_gpusvm_ops;
> +struct drm_gpusvm_range;
> +struct drm_gpusvm_devmem;
> +struct drm_pagemap;
> +struct drm_pagemap_dma_addr;
> +
> +/**
> + * struct drm_gpusvm_devmem_ops - Operations structure for GPU SVM device memory
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM)
> + * device memory. These operations are provided by the GPU driver to manage device memory
> + * allocations and perform operations such as migration between device memory and system
> + * RAM.
> + */
> +struct drm_gpusvm_devmem_ops {
> +	/**
> +	 * @devmem_release: Release device memory allocation (optional)
> +	 * @devmem_allocation: device memory allocation
> +	 *
> +	 * Release device memory allocation and drop a reference to device
> +	 * memory allocation.
> +	 */
> +	void (*devmem_release)(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +	/**
> +	 * @populate_devmem_pfn: Populate device memory PFN (required for migration)
> +	 * @devmem_allocation: device memory allocation
> +	 * @npages: Number of pages to populate
> +	 * @pfn: Array of page frame numbers to populate
> +	 *
> +	 * Populate device memory page frame numbers (PFN).
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*populate_devmem_pfn)(struct drm_gpusvm_devmem *devmem_allocation,
> +				   unsigned long npages, unsigned long *pfn);
> +
> +	/**
> +	 * @copy_to_devmem: Copy to device memory (required for migration)
> +	 * @pages: Pointer to array of device memory pages (destination)
> +	 * @dma_addr: Pointer to array of DMA addresses (source)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to device memory.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_devmem)(struct page **pages,
> +			      dma_addr_t *dma_addr,
> +			      unsigned long npages);
> +
> +	/**
> +	 * @copy_to_ram: Copy to system RAM (required for migration)
> +	 * @pages: Pointer to array of device memory pages (source)
> +	 * @dma_addr: Pointer to array of DMA addresses (destination)
> +	 * @npages: Number of pages to copy
> +	 *
> +	 * Copy pages to system RAM.
> +	 *
> +	 * Returns:
> +	 * 0 on success, a negative error code on failure.
> +	 */
> +	int (*copy_to_ram)(struct page **pages,
> +			   dma_addr_t *dma_addr,
> +			   unsigned long npages);
> +};
> +
> +/**
> + * struct drm_gpusvm_devmem - Structure representing a GPU SVM device memory allocation
> + *
> + * @dev: Pointer to the device structure which device memory allocation belongs to
> + * @mm: Pointer to the mm_struct for the address space
> + * @ops: Pointer to the operations structure for GPU SVM device memory
> + * @dpagemap: The struct drm_pagemap of the pages this allocation belongs to.
> + * @size: Size of device memory allocation
> + * @detached: device memory allocations is detached from device pages
> + */
> +struct drm_gpusvm_devmem {
> +	struct device *dev;
> +	struct mm_struct *mm;
> +	const struct drm_gpusvm_devmem_ops *ops;
> +	struct drm_pagemap *dpagemap;
> +	size_t size;
> +	bool detached;
> +};
> +
> +/**
> + * struct drm_gpusvm_ops - Operations structure for GPU SVM
> + *
> + * This structure defines the operations for GPU Shared Virtual Memory (SVM).
> + * These operations are provided by the GPU driver to manage SVM ranges and
> + * notifiers.
> + */
> +struct drm_gpusvm_ops {
> +	/**
> +	 * @notifier_alloc: Allocate a GPU SVM notifier (optional)
> +	 *
> +	 * Allocate a GPU SVM notifier.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM notifier on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_notifier *(*notifier_alloc)(void);
> +
> +	/**
> +	 * @notifier_free: Free a GPU SVM notifier (optional)
> +	 * @notifier: Pointer to the GPU SVM notifier to be freed
> +	 *
> +	 * Free a GPU SVM notifier.
> +	 */
> +	void (*notifier_free)(struct drm_gpusvm_notifier *notifier);
> +
> +	/**
> +	 * @range_alloc: Allocate a GPU SVM range (optional)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 *
> +	 * Allocate a GPU SVM range.
> +	 *
> +	 * Returns:
> +	 * Pointer to the allocated GPU SVM range on success, NULL on failure.
> +	 */
> +	struct drm_gpusvm_range *(*range_alloc)(struct drm_gpusvm *gpusvm);
> +
> +	/**
> +	 * @range_free: Free a GPU SVM range (optional)
> +	 * @range: Pointer to the GPU SVM range to be freed
> +	 *
> +	 * Free a GPU SVM range.
> +	 */
> +	void (*range_free)(struct drm_gpusvm_range *range);
> +
> +	/**
> +	 * @invalidate: Invalidate GPU SVM notifier (required)
> +	 * @gpusvm: Pointer to the GPU SVM
> +	 * @notifier: Pointer to the GPU SVM notifier
> +	 * @mmu_range: Pointer to the mmu_notifier_range structure
> +	 *
> +	 * Invalidate the GPU page tables. It can safely walk the notifier range
> +	 * RB tree/list in this function. Called while holding the notifier lock.
> +	 */
> +	void (*invalidate)(struct drm_gpusvm *gpusvm,
> +			   struct drm_gpusvm_notifier *notifier,
> +			   const struct mmu_notifier_range *mmu_range);
> +};
> +
> +/**
> + * struct drm_gpusvm_notifier - Structure representing a GPU SVM notifier
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: MMU interval notifier
> + * @itree: Interval tree node for the notifier (inserted in GPU SVM)
> + * @entry: List entry to fast interval tree traversal
> + * @root: Cached root node of the RB tree containing ranges
> + * @range_list: List head containing of ranges in the same order they appear in
> + *              interval tree. This is useful to keep iterating ranges while
> + *              doing modifications to RB tree.
> + * @flags.removed: Flag indicating whether the MMU interval notifier has been
> + *                 removed
> + *
> + * This structure represents a GPU SVM notifier.
> + */
> +struct drm_gpusvm_notifier {
> +	struct drm_gpusvm *gpusvm;
> +	struct mmu_interval_notifier notifier;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	struct rb_root_cached root;
> +	struct list_head range_list;
> +	struct {
> +		u32 removed : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm_range - Structure representing a GPU SVM range
> + *
> + * @gpusvm: Pointer to the GPU SVM structure
> + * @notifier: Pointer to the GPU SVM notifier
> + * @refcount: Reference count for the range
> + * @itree: Interval tree node for the range (inserted in GPU SVM notifier)
> + * @entry: List entry to fast interval tree traversal
> + * @notifier_seq: Notifier sequence number of the range's pages
> + * @dma_addr: DMA address array
> + * @dpagemap: The struct drm_pagemap of the device pages we're dma-mapping.
> + *            Note this is assuming only one drm_pagemap per range is allowed.
> + * @flags.migrate_devmem: Flag indicating whether the range can be migrated to device memory
> + * @flags.unmapped: Flag indicating if the range has been unmapped
> + * @flags.partial_unmap: Flag indicating if the range has been partially unmapped
> + * @flags.has_devmem_pages: Flag indicating if the range has devmem pages
> + * @flags.has_dma_mapping: Flag indicating if the range has a DMA mapping
> + *
> + * This structure represents a GPU SVM range used for tracking memory ranges
> + * mapped in a DRM device.
> + */
> +struct drm_gpusvm_range {
> +	struct drm_gpusvm *gpusvm;
> +	struct drm_gpusvm_notifier *notifier;
> +	struct kref refcount;
> +	struct interval_tree_node itree;
> +	struct list_head entry;
> +	unsigned long notifier_seq;
> +	struct drm_pagemap_dma_addr *dma_addr;
> +	struct drm_pagemap *dpagemap;
> +	struct {
> +		/* All flags below must be set upon creation */
> +		u16 migrate_devmem : 1;
> +		/* All flags below must be set / cleared under notifier lock */
> +		u16 unmapped : 1;
> +		u16 partial_unmap : 1;
> +		u16 has_devmem_pages : 1;
> +		u16 has_dma_mapping : 1;
> +	} flags;
> +};
> +
> +/**
> + * struct drm_gpusvm - GPU SVM structure
> + *
> + * @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
> + * @ops: Pointer to the operations structure for GPU SVM
> + * @chunk_sizes: Pointer to the array of chunk sizes used in range allocation.
> + *               Entries should be powers of 2 in descending order.
> + * @num_chunks: Number of chunks
> + * @notifier_lock: Read-write semaphore for protecting notifier operations
> + * @root: Cached root node of the Red-Black tree containing GPU SVM notifiers
> + * @notifier_list: list head containing of notifiers in the same order they
> + *                 appear in interval tree. This is useful to keep iterating
> + *                 notifiers while doing modifications to RB tree.
> + *
> + * This structure represents a GPU SVM (Shared Virtual Memory) used for tracking
> + * memory ranges mapped in a DRM (Direct Rendering Manager) device.
> + *
> + * No reference counting is provided, as this is expected to be embedded in the
> + * driver VM structure along with the struct drm_gpuvm, which handles reference
> + * counting.
> + */
> +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;
> +	const struct drm_gpusvm_ops *ops;
> +	const unsigned long *chunk_sizes;
> +	int num_chunks;
> +	struct rw_semaphore notifier_lock;
> +	struct rb_root_cached root;
> +	struct list_head notifier_list;
> +#ifdef CONFIG_LOCKDEP
> +	/**
> +	 * @lock_dep_map: Annotates drm_gpusvm_range_find_or_insert and
> +	 * drm_gpusvm_range_remove with a driver provided lock.
> +	 */
> +	struct lockdep_map *lock_dep_map;
> +#endif
> +};
> +
> +/**
> + * struct drm_gpusvm_ctx - DRM GPU SVM context
> + *
> + * @check_pages_threshold: Check CPU pages for present if chunk is less than or
> + *                         equal to threshold. If not present, reduce chunk
> + *                         size.
> + * @in_notifier: entering from a MMU notifier
> + * @read_only: operating on read-only memory
> + * @devmem_possible: possible to use device memory
> + *
> + * Context that is DRM GPUSVM is operating in (i.e. user arguments).
> + */
> +struct drm_gpusvm_ctx {
> +	unsigned long check_pages_threshold;
> +	unsigned int in_notifier :1;
> +	unsigned int read_only :1;
> +	unsigned int devmem_possible :1;
> +};
> +
> +int drm_gpusvm_init(struct drm_gpusvm *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,
> +		    const struct drm_gpusvm_ops *ops,
> +		    const unsigned long *chunk_sizes, int num_chunks);
> +
> +void drm_gpusvm_fini(struct drm_gpusvm *gpusvm);
> +
> +void drm_gpusvm_free(struct drm_gpusvm *gpusvm);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find_or_insert(struct drm_gpusvm *gpusvm,
> +				unsigned long fault_addr,
> +				unsigned long gpuva_start,
> +				unsigned long gpuva_end,
> +				const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_remove(struct drm_gpusvm *gpusvm,
> +			     struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_evict(struct drm_gpusvm *gpusvm,
> +			    struct drm_gpusvm_range *range);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_get(struct drm_gpusvm_range *range);
> +
> +void drm_gpusvm_range_put(struct drm_gpusvm_range *range);
> +
> +bool drm_gpusvm_range_pages_valid(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range);
> +
> +int drm_gpusvm_range_get_pages(struct drm_gpusvm *gpusvm,
> +			       struct drm_gpusvm_range *range,
> +			       const struct drm_gpusvm_ctx *ctx);
> +
> +void drm_gpusvm_range_unmap_pages(struct drm_gpusvm *gpusvm,
> +				  struct drm_gpusvm_range *range,
> +				  const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_migrate_to_devmem(struct drm_gpusvm *gpusvm,
> +				 struct drm_gpusvm_range *range,
> +				 struct drm_gpusvm_devmem *devmem_allocation,
> +				 const struct drm_gpusvm_ctx *ctx);
> +
> +int drm_gpusvm_evict_to_ram(struct drm_gpusvm_devmem *devmem_allocation);
> +
> +const struct dev_pagemap_ops *drm_gpusvm_pagemap_ops_get(void);
> +
> +bool drm_gpusvm_has_mapping(struct drm_gpusvm *gpusvm, unsigned long start,
> +			    unsigned long end);
> +
> +struct drm_gpusvm_range *
> +drm_gpusvm_range_find(struct drm_gpusvm_notifier *notifier, unsigned long start,
> +		      unsigned long end);
> +
> +void drm_gpusvm_range_set_unmapped(struct drm_gpusvm_range *range,
> +				   const struct mmu_notifier_range *mmu_range);
> +
> +void drm_gpusvm_devmem_init(struct drm_gpusvm_devmem *devmem_allocation,
> +			    struct device *dev, struct mm_struct *mm,
> +			    const struct drm_gpusvm_devmem_ops *ops,
> +			    struct drm_pagemap *dpagemap, size_t size);
> +
> +#ifdef CONFIG_LOCKDEP
> +/**
> + * drm_gpusvm_driver_set_lock() - Set the lock protecting accesses to GPU SVM
> + * @gpusvm: Pointer to the GPU SVM structure.
> + * @lock: the lock used to protect the gpuva list. The locking primitive
> + * must contain a dep_map field.
> + *
> + * Call this to annotate drm_gpusvm_range_find_or_insert and
> + * drm_gpusvm_range_remove.
> + */
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) \
> +	do { \
> +		if (!WARN((gpusvm)->lock_dep_map, \
> +			  "GPUSVM range lock should be set only once."))\
> +			(gpusvm)->lock_dep_map = &(lock)->dep_map;	\
> +	} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) \
> +	do { \
> +		if ((gpusvm)->lock_dep_map)	\
> +			lock_is_held((gpusvm)->lock_dep_map);	\
> +	} while (0)
> +#else
> +#define drm_gpusvm_driver_set_lock(gpusvm, lock) do {} while (0)
> +#define drm_gpusvm_driver_lock_held(gpusvm) do {} while (0)
> +#endif
> +
> +/**
> + * drm_gpusvm_notifier_lock() - Lock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, take lock
> + */
> +#define drm_gpusvm_notifier_lock(gpusvm__)	\
> +	down_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * drm_gpusvm_notifier_unlock() - Unlock GPU SVM notifier
> + * @gpusvm__: Pointer to the GPU SVM structure.
> + *
> + * Abstract client usage GPU SVM notifier lock, drop lock
> + */
> +#define drm_gpusvm_notifier_unlock(gpusvm__)	\
> +	up_read(&(gpusvm__)->notifier_lock)
> +
> +/**
> + * __drm_gpusvm_range_next() - Get the next GPU SVM range in the list
> + * @range: a pointer to the current GPU SVM range
> + *
> + * Return: A pointer to the next drm_gpusvm_range if available, or NULL if the
> + *         current range is the last one or if the input range is NULL.
> + */
> +static inline struct drm_gpusvm_range *
> +__drm_gpusvm_range_next(struct drm_gpusvm_range *range)
> +{
> +	if (range && !list_is_last(&range->entry,
> +				   &range->notifier->range_list))
> +		return list_next_entry(range, entry);
> +
> +	return NULL;
> +}
> +
> +/**
> + * drm_gpusvm_for_each_range() - Iterate over GPU SVM ranges in a notifier
> + * @range__: Iterator variable for the ranges. If set, it indicates the start of
> + *	     the iterator. If NULL, call drm_gpusvm_range_find() to get the range.
> + * @notifier__: Pointer to the GPU SVM notifier
> + * @start__: Start address of the range
> + * @end__: End address of the range
> + *
> + * This macro is used to iterate over GPU SVM ranges in a notifier. It is safe
> + * to use while holding the driver SVM lock or the notifier lock.
> + */
> +#define drm_gpusvm_for_each_range(range__, notifier__, start__, end__)	\
> +	for ((range__) = (range__) ?:					\
> +	     drm_gpusvm_range_find((notifier__), (start__), (end__));	\
> +	     (range__) && (range__->itree.start < (end__));		\
> +	     (range__) = __drm_gpusvm_range_next(range__))
> +
> +#endif /* __DRM_GPUSVM_H__ */

Re: [PATCH v3 04/30] drm/pagemap: Add DRM pagemap

[Gwan-gyeong Mun]

On 12/18/24 1:33 AM, Matthew Brost wrote:
> From: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> 
> Introduce drm_pagemap ops to map and unmap dma to VRAM resources. In the
> local memory case it's a matter of merely providing an offset into the
> device's physical address. For future p2p the map and unmap functions may
> encode as needed.
> 
> Similar to how dma-buf works, let the memory provider (drm_pagemap) provide
> the mapping functionality.
> 
> v3:
>   - Move to drm level include
> 
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> ---
>   include/drm/drm_pagemap.h | 103 ++++++++++++++++++++++++++++++++++++++
>   1 file changed, 103 insertions(+)
>   create mode 100644 include/drm/drm_pagemap.h
> 
> diff --git a/include/drm/drm_pagemap.h b/include/drm/drm_pagemap.h
> new file mode 100644
> index 000000000000..b6b387b81479
> --- /dev/null
> +++ b/include/drm/drm_pagemap.h
> @@ -0,0 +1,103 @@
> +/* SPDX-License-Identifier: MIT */
> +#ifndef _DRM_PAGEMAP_H_
> +#define _DRM_PAGEMAP_H_
> +
> +#include <linux/dma-direction.h>
> +#include <linux/hmm.h>
> +#include <linux/types.h>
> +
> +struct drm_pagemap;
> +struct device;
> +
> +/**
> + * enum drm_interconnect_protocol - Used to identify an interconnect protocol.
> + */
> +enum drm_interconnect_protocol {
> +	DRM_INTERCONNECT_SYSTEM,    /* DMA map is system pages. */
> +	DRM_INTERCONNECT_PCIE_P2P,  /* DMA map is PCIE P2P */
> +	DRM_INTERCONNECT_DRIVER,    /* DMA map is driver defined */
> +	/* A driver can add private values beyond DRM_INTERCONNECT_DRIVER */
> +};
> +
> +/**
> + * struct drm_pagemap_dma_addr - DMA address representation.
> + * @addr: The dma address or driver-defined address for driver private interconnects.
> + * @proto: The interconnect protocol.
> + * @order: The page order of the dma mapping. (Size is PAGE_SIZE << order).
> + * @dir: The DMA direction.
> + *
> + * Note: There is room for improvement here. We should be able to pack into
> + * 64 bits.
> + */
> +struct drm_pagemap_dma_addr {
> +	dma_addr_t addr;
> +	u64 proto : 54;
> +	u64 order : 8;
> +	u64 dir : 2;
> +};
> +
> +/**
> + * drm_pagemap_dma_addr_encode() - Encode a dma address with metadata
> + * @addr: The dma address or driver-defined address for driver private interconnects.
> + * @proto: The interconnect protocol.
> + * @order: The page order of the dma mapping. (Size is PAGE_SIZE << order).
> + * @dir: The DMA direction.
> + *
> + * Return: A struct drm_pagemap_dma_addr encoding the above information.
> + */
> +static inline struct drm_pagemap_dma_addr
> +drm_pagemap_dma_addr_encode(dma_addr_t addr,
> +			    enum drm_interconnect_protocol proto,
> +			    unsigned int order,
> +			    enum dma_data_direction dir)
> +{
> +	return (struct drm_pagemap_dma_addr) {
> +		.addr = addr,
> +		.proto = proto,
> +		.order = order,
> +		.dir = dir,
> +	};
> +}
> +
> +/**
> + * struct drm_pagemap_ops: Ops for a drm-pagemap.
> + */
Hi,
In the following comment, the protocol should be removed and an 
explanation of order should be added.

Br,
G.G.
> +struct drm_pagemap_ops {
> +	/**
> +	 * @map_dma: Map for dma access or provide a virtual address suitable for
> +	 * @dev.
> +	 * @dpagemap: The struct drm_pagemap for the page.
> +	 * @dev: The dma mapper.
> +	 * @page: The page to map.
> +	 * @dir: The transfer direction.
> +	 * @protocol: The protocol to use.
> +	 */
> +	struct drm_pagemap_dma_addr (*map_dma)(struct drm_pagemap *dpagemap,
> +					       struct device *dev,
> +					       struct page *page,
> +					       unsigned int order,
> +					       enum dma_data_direction dir);
> +
> +	/**
> +	 * @unmap_dma: Unmap a dma address previously obtained using  @map_dma.
> +	 * @dev: The dma unmapper.
> +	 * @addr: The dma address obtained when mapping.
> +	 */
> +	void (*unmap_dma)(struct drm_pagemap *dpagemap,
> +			  struct device *dev,
> +			  struct drm_pagemap_dma_addr addr);
> +
> +};
> +
> +/**
> + * struct drm_pagemap: Additional information for a struct dev_pagemap
> + * used for device p2p handshaking.
> + * @ops: The struct drm_pagemap_ops.
> + * @dev: The struct drevice owning the device-private memory.
> + */
> +struct drm_pagemap {
> +	const struct drm_pagemap_ops *ops;
> +	struct device *dev;
> +};
> +
> +#endif

Re: [PATCH v3 03/30] mm/migrate: Trylock device page in do_swap_page

[Thomas Hellström]

On Tue, 2024-12-17 at 15:33 -0800, Matthew Brost wrote:
> Avoid multiple CPU page faults to the same device page racing by
> trying
> to lock the page in do_swap_page before taking an extra reference to
> the
> page. This prevents scenarios where multiple CPU page faults each
> take
> an extra reference to a device page, which could abort migration in
> folio_migrate_mapping. With the device page being locked in
> do_swap_page, the migrate_vma_* functions need to be updated to avoid
> locking the fault_page argument.
> 
> Prior to this change, a livelock scenario could occur in Xe's (Intel
> GPU
> DRM driver) SVM implementation if enough threads faulted the same
> device
> page.
> 
> v3:
>  - Put page after unlocking page (Alistair)
>  - Warn on spliting a TPH which is fault page (Alistair)
>  - Warn on dst page == fault page (Alistair)
> 
> Cc: Alistair Popple <apopple@nvidia.com>
> Cc: Philip Yang <Philip.Yang@amd.com>
> Cc: Felix Kuehling <felix.kuehling@amd.com>
> Cc: Christian König <christian.koenig@amd.com>
> Cc: Andrew Morton <akpm@linux-foundation.org>
> Suggested-by: Simona Vetter <simona.vetter@ffwll.ch>
> Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> ---
>  mm/memory.c         | 13 ++++++---
>  mm/migrate_device.c | 64 ++++++++++++++++++++++++++++++++-----------
> --
>  2 files changed, 55 insertions(+), 22 deletions(-)
> 
> diff --git a/mm/memory.c b/mm/memory.c
> index 75c2dfd04f72..ae8b11131dad 100644
> --- a/mm/memory.c
> +++ b/mm/memory.c
> @@ -4267,10 +4267,15 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
>  			 * Get a page reference while we know the
> page can't be
>  			 * freed.
>  			 */
> -			get_page(vmf->page);
> -			pte_unmap_unlock(vmf->pte, vmf->ptl);
> -			ret = vmf->page->pgmap->ops-
> >migrate_to_ram(vmf);
> -			put_page(vmf->page);
> +			if (trylock_page(vmf->page)) {
> +				get_page(vmf->page);

A couple of questions that I and Matt have been discussing around this
change and that we aren't completely cleare about. Perhaps somebody
else has some input:

1) The livelock occurs because we do a page reference check in
migrate_vma_check_page(). However judging from the documentation it
still uses a page refcount to determine whether it's pinned. 

If we were to use folio_maybe_dma_pinned() which uses the pin-count
instead problem would be solved? However if there already is a refcount
that we don't know about, that holder could possibly upgrade it to a
pin-count. Does anybody know why not folio_maybe_dma_pinned() is used?



> +				pte_unmap_unlock(vmf->pte, vmf-
> >ptl);
> +				ret = vmf->page->pgmap->ops-
> >migrate_to_ram(vmf);


2) Second question is allocating memory under a page-lock. There
appears to be code avoiding that, arguing that page-locks may be taken
during reclaim. At least unless GFP_NOFS is used. Here we'd allow
GFP_KERNEL allocations under page-locks but can argue that it's only
device_private pages that are safe. Although some drivers do GFP_KERNEL
locks also under ordinary page-locks in migrate_to_ram(), probably
arguing that those pages are taken off LRU. Seems a bit fragile, is
there any known policy regarding this?

Thanks,
Thomas


> +				unlock_page(vmf->page);
> +				put_page(vmf->page);
> +			} else {
> +				pte_unmap_unlock(vmf->pte, vmf-
> >ptl);
> +			}
>  		} else if (is_hwpoison_entry(entry)) {
>  			ret = VM_FAULT_HWPOISON;
>  		} else if (is_pte_marker_entry(entry)) {
> diff --git a/mm/migrate_device.c b/mm/migrate_device.c
> index 19960743f927..3470357d9bae 100644
> --- a/mm/migrate_device.c
> +++ b/mm/migrate_device.c
> @@ -60,6 +60,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
>  				   struct mm_walk *walk)
>  {
>  	struct migrate_vma *migrate = walk->private;
> +	struct folio *fault_folio = migrate->fault_page ?
> +		page_folio(migrate->fault_page) : NULL;
>  	struct vm_area_struct *vma = walk->vma;
>  	struct mm_struct *mm = vma->vm_mm;
>  	unsigned long addr = start, unmapped = 0;
> @@ -88,11 +90,16 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
>  
>  			folio_get(folio);
>  			spin_unlock(ptl);
> +			/* FIXME support THP */
> +			if (WARN_ON_ONCE(fault_folio == folio))
> +				return
> migrate_vma_collect_skip(start, end,
> +								walk
> );
>  			if (unlikely(!folio_trylock(folio)))
>  				return
> migrate_vma_collect_skip(start, end,
>  								walk
> );
>  			ret = split_folio(folio);
> -			folio_unlock(folio);
> +			if (fault_folio != folio)
> +				folio_unlock(folio);
>  			folio_put(folio);
>  			if (ret)
>  				return
> migrate_vma_collect_skip(start, end,
> @@ -192,7 +199,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
>  		 * optimisation to avoid walking the rmap later with
>  		 * try_to_migrate().
>  		 */
> -		if (folio_trylock(folio)) {
> +		if (fault_folio == folio || folio_trylock(folio)) {
>  			bool anon_exclusive;
>  			pte_t swp_pte;
>  
> @@ -204,7 +211,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
>  
>  				if
> (folio_try_share_anon_rmap_pte(folio, page)) {
>  					set_pte_at(mm, addr, ptep,
> pte);
> -					folio_unlock(folio);
> +					if (fault_folio != folio)
> +						folio_unlock(folio);
>  					folio_put(folio);
>  					mpfn = 0;
>  					goto next;
> @@ -363,6 +371,8 @@ static unsigned long
> migrate_device_unmap(unsigned long *src_pfns,
>  					  unsigned long npages,
>  					  struct page *fault_page)
>  {
> +	struct folio *fault_folio = fault_page ?
> +		page_folio(fault_page) : NULL;
>  	unsigned long i, restore = 0;
>  	bool allow_drain = true;
>  	unsigned long unmapped = 0;
> @@ -427,7 +437,8 @@ static unsigned long
> migrate_device_unmap(unsigned long *src_pfns,
>  		remove_migration_ptes(folio, folio, 0);
>  
>  		src_pfns[i] = 0;
> -		folio_unlock(folio);
> +		if (fault_folio != folio)
> +			folio_unlock(folio);
>  		folio_put(folio);
>  		restore--;
>  	}
> @@ -536,6 +547,8 @@ int migrate_vma_setup(struct migrate_vma *args)
>  		return -EINVAL;
>  	if (args->fault_page && !is_device_private_page(args-
> >fault_page))
>  		return -EINVAL;
> +	if (args->fault_page && !PageLocked(args->fault_page))
> +		return -EINVAL;
>  
>  	memset(args->src, 0, sizeof(*args->src) * nr_pages);
>  	args->cpages = 0;
> @@ -799,19 +812,13 @@ void migrate_vma_pages(struct migrate_vma
> *migrate)
>  }
>  EXPORT_SYMBOL(migrate_vma_pages);
>  
> -/*
> - * migrate_device_finalize() - complete page migration
> - * @src_pfns: src_pfns returned from migrate_device_range()
> - * @dst_pfns: array of pfns allocated by the driver to migrate
> memory to
> - * @npages: number of pages in the range
> - *
> - * Completes migration of the page by removing special migration
> entries.
> - * Drivers must ensure copying of page data is complete and visible
> to the CPU
> - * before calling this.
> - */
> -void migrate_device_finalize(unsigned long *src_pfns,
> -			unsigned long *dst_pfns, unsigned long
> npages)
> +static void __migrate_device_finalize(unsigned long *src_pfns,
> +				      unsigned long *dst_pfns,
> +				      unsigned long npages,
> +				      struct page *fault_page)
>  {
> +	struct folio *fault_folio = fault_page ?
> +		page_folio(fault_page) : NULL;
>  	unsigned long i;
>  
>  	for (i = 0; i < npages; i++) {
> @@ -824,6 +831,7 @@ void migrate_device_finalize(unsigned long
> *src_pfns,
>  
>  		if (!page) {
>  			if (dst) {
> +				WARN_ON_ONCE(fault_folio == dst);
>  				folio_unlock(dst);
>  				folio_put(dst);
>  			}
> @@ -834,6 +842,7 @@ void migrate_device_finalize(unsigned long
> *src_pfns,
>  
>  		if (!(src_pfns[i] & MIGRATE_PFN_MIGRATE) || !dst) {
>  			if (dst) {
> +				WARN_ON_ONCE(fault_folio == dst);
>  				folio_unlock(dst);
>  				folio_put(dst);
>  			}
> @@ -841,7 +850,8 @@ void migrate_device_finalize(unsigned long
> *src_pfns,
>  		}
>  
>  		remove_migration_ptes(src, dst, 0);
> -		folio_unlock(src);
> +		if (fault_folio != src)
> +			folio_unlock(src);
>  
>  		if (folio_is_zone_device(src))
>  			folio_put(src);
> @@ -849,6 +859,7 @@ void migrate_device_finalize(unsigned long
> *src_pfns,
>  			folio_putback_lru(src);
>  
>  		if (dst != src) {
> +			WARN_ON_ONCE(fault_folio == dst);
>  			folio_unlock(dst);
>  			if (folio_is_zone_device(dst))
>  				folio_put(dst);
> @@ -857,6 +868,22 @@ void migrate_device_finalize(unsigned long
> *src_pfns,
>  		}
>  	}
>  }
> +
> +/*
> + * migrate_device_finalize() - complete page migration
> + * @src_pfns: src_pfns returned from migrate_device_range()
> + * @dst_pfns: array of pfns allocated by the driver to migrate
> memory to
> + * @npages: number of pages in the range
> + *
> + * Completes migration of the page by removing special migration
> entries.
> + * Drivers must ensure copying of page data is complete and visible
> to the CPU
> + * before calling this.
> + */
> +void migrate_device_finalize(unsigned long *src_pfns,
> +			unsigned long *dst_pfns, unsigned long
> npages)
> +{
> +	return __migrate_device_finalize(src_pfns, dst_pfns, npages,
> NULL);
> +}
>  EXPORT_SYMBOL(migrate_device_finalize);
>  
>  /**
> @@ -872,7 +899,8 @@ EXPORT_SYMBOL(migrate_device_finalize);
>   */
>  void migrate_vma_finalize(struct migrate_vma *migrate)
>  {
> -	migrate_device_finalize(migrate->src, migrate->dst, migrate-
> >npages);
> +	__migrate_device_finalize(migrate->src, migrate->dst,
> migrate->npages,
> +				  migrate->fault_page);
>  }
>  EXPORT_SYMBOL(migrate_vma_finalize);
>  

Re: [PATCH v3 03/30] mm/migrate: Trylock device page in do_swap_page

[Matthew Brost]

On Tue, Jan 28, 2025 at 06:26:29PM +0100, Thomas Hellström wrote:
> On Tue, 2024-12-17 at 15:33 -0800, Matthew Brost wrote:
> > Avoid multiple CPU page faults to the same device page racing by
> > trying
> > to lock the page in do_swap_page before taking an extra reference to
> > the
> > page. This prevents scenarios where multiple CPU page faults each
> > take
> > an extra reference to a device page, which could abort migration in
> > folio_migrate_mapping. With the device page being locked in
> > do_swap_page, the migrate_vma_* functions need to be updated to avoid
> > locking the fault_page argument.
> > 
> > Prior to this change, a livelock scenario could occur in Xe's (Intel
> > GPU
> > DRM driver) SVM implementation if enough threads faulted the same
> > device
> > page.
> > 
> > v3:
> >  - Put page after unlocking page (Alistair)
> >  - Warn on spliting a TPH which is fault page (Alistair)
> >  - Warn on dst page == fault page (Alistair)
> > 
> > Cc: Alistair Popple <apopple@nvidia.com>
> > Cc: Philip Yang <Philip.Yang@amd.com>
> > Cc: Felix Kuehling <felix.kuehling@amd.com>
> > Cc: Christian König <christian.koenig@amd.com>
> > Cc: Andrew Morton <akpm@linux-foundation.org>
> > Suggested-by: Simona Vetter <simona.vetter@ffwll.ch>
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > ---
> >  mm/memory.c         | 13 ++++++---
> >  mm/migrate_device.c | 64 ++++++++++++++++++++++++++++++++-----------
> > --
> >  2 files changed, 55 insertions(+), 22 deletions(-)
> > 
> > diff --git a/mm/memory.c b/mm/memory.c
> > index 75c2dfd04f72..ae8b11131dad 100644
> > --- a/mm/memory.c
> > +++ b/mm/memory.c
> > @@ -4267,10 +4267,15 @@ vm_fault_t do_swap_page(struct vm_fault *vmf)
> >  			 * Get a page reference while we know the
> > page can't be
> >  			 * freed.
> >  			 */
> > -			get_page(vmf->page);
> > -			pte_unmap_unlock(vmf->pte, vmf->ptl);
> > -			ret = vmf->page->pgmap->ops-
> > >migrate_to_ram(vmf);
> > -			put_page(vmf->page);
> > +			if (trylock_page(vmf->page)) {
> > +				get_page(vmf->page);
> 
> A couple of questions that I and Matt have been discussing around this
> change and that we aren't completely cleare about. Perhaps somebody
> else has some input:
> 
> 1) The livelock occurs because we do a page reference check in
> migrate_vma_check_page(). However judging from the documentation it
> still uses a page refcount to determine whether it's pinned. 
> 

Also in folio_migrate_mapping & __folio_migrate_mapping via
folio_ref_count & folio_ref_freeze respectfully to clear. I have same
questions as Thomas too.

Matt 

> If we were to use folio_maybe_dma_pinned() which uses the pin-count
> instead problem would be solved? However if there already is a refcount
> that we don't know about, that holder could possibly upgrade it to a
> pin-count. Does anybody know why not folio_maybe_dma_pinned() is used?
> 
> 
> 
> > +				pte_unmap_unlock(vmf->pte, vmf-
> > >ptl);
> > +				ret = vmf->page->pgmap->ops-
> > >migrate_to_ram(vmf);
> 
> 
> 2) Second question is allocating memory under a page-lock. There
> appears to be code avoiding that, arguing that page-locks may be taken
> during reclaim. At least unless GFP_NOFS is used. Here we'd allow
> GFP_KERNEL allocations under page-locks but can argue that it's only
> device_private pages that are safe. Although some drivers do GFP_KERNEL
> locks also under ordinary page-locks in migrate_to_ram(), probably
> arguing that those pages are taken off LRU. Seems a bit fragile, is
> there any known policy regarding this?
> 
> Thanks,
> Thomas
> 
> 
> > +				unlock_page(vmf->page);
> > +				put_page(vmf->page);
> > +			} else {
> > +				pte_unmap_unlock(vmf->pte, vmf-
> > >ptl);
> > +			}
> >  		} else if (is_hwpoison_entry(entry)) {
> >  			ret = VM_FAULT_HWPOISON;
> >  		} else if (is_pte_marker_entry(entry)) {
> > diff --git a/mm/migrate_device.c b/mm/migrate_device.c
> > index 19960743f927..3470357d9bae 100644
> > --- a/mm/migrate_device.c
> > +++ b/mm/migrate_device.c
> > @@ -60,6 +60,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> >  				   struct mm_walk *walk)
> >  {
> >  	struct migrate_vma *migrate = walk->private;
> > +	struct folio *fault_folio = migrate->fault_page ?
> > +		page_folio(migrate->fault_page) : NULL;
> >  	struct vm_area_struct *vma = walk->vma;
> >  	struct mm_struct *mm = vma->vm_mm;
> >  	unsigned long addr = start, unmapped = 0;
> > @@ -88,11 +90,16 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> >  
> >  			folio_get(folio);
> >  			spin_unlock(ptl);
> > +			/* FIXME support THP */
> > +			if (WARN_ON_ONCE(fault_folio == folio))
> > +				return
> > migrate_vma_collect_skip(start, end,
> > +								walk
> > );
> >  			if (unlikely(!folio_trylock(folio)))
> >  				return
> > migrate_vma_collect_skip(start, end,
> >  								walk
> > );
> >  			ret = split_folio(folio);
> > -			folio_unlock(folio);
> > +			if (fault_folio != folio)
> > +				folio_unlock(folio);
> >  			folio_put(folio);
> >  			if (ret)
> >  				return
> > migrate_vma_collect_skip(start, end,
> > @@ -192,7 +199,7 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> >  		 * optimisation to avoid walking the rmap later with
> >  		 * try_to_migrate().
> >  		 */
> > -		if (folio_trylock(folio)) {
> > +		if (fault_folio == folio || folio_trylock(folio)) {
> >  			bool anon_exclusive;
> >  			pte_t swp_pte;
> >  
> > @@ -204,7 +211,8 @@ static int migrate_vma_collect_pmd(pmd_t *pmdp,
> >  
> >  				if
> > (folio_try_share_anon_rmap_pte(folio, page)) {
> >  					set_pte_at(mm, addr, ptep,
> > pte);
> > -					folio_unlock(folio);
> > +					if (fault_folio != folio)
> > +						folio_unlock(folio);
> >  					folio_put(folio);
> >  					mpfn = 0;
> >  					goto next;
> > @@ -363,6 +371,8 @@ static unsigned long
> > migrate_device_unmap(unsigned long *src_pfns,
> >  					  unsigned long npages,
> >  					  struct page *fault_page)
> >  {
> > +	struct folio *fault_folio = fault_page ?
> > +		page_folio(fault_page) : NULL;
> >  	unsigned long i, restore = 0;
> >  	bool allow_drain = true;
> >  	unsigned long unmapped = 0;
> > @@ -427,7 +437,8 @@ static unsigned long
> > migrate_device_unmap(unsigned long *src_pfns,
> >  		remove_migration_ptes(folio, folio, 0);
> >  
> >  		src_pfns[i] = 0;
> > -		folio_unlock(folio);
> > +		if (fault_folio != folio)
> > +			folio_unlock(folio);
> >  		folio_put(folio);
> >  		restore--;
> >  	}
> > @@ -536,6 +547,8 @@ int migrate_vma_setup(struct migrate_vma *args)
> >  		return -EINVAL;
> >  	if (args->fault_page && !is_device_private_page(args-
> > >fault_page))
> >  		return -EINVAL;
> > +	if (args->fault_page && !PageLocked(args->fault_page))
> > +		return -EINVAL;
> >  
> >  	memset(args->src, 0, sizeof(*args->src) * nr_pages);
> >  	args->cpages = 0;
> > @@ -799,19 +812,13 @@ void migrate_vma_pages(struct migrate_vma
> > *migrate)
> >  }
> >  EXPORT_SYMBOL(migrate_vma_pages);
> >  
> > -/*
> > - * migrate_device_finalize() - complete page migration
> > - * @src_pfns: src_pfns returned from migrate_device_range()
> > - * @dst_pfns: array of pfns allocated by the driver to migrate
> > memory to
> > - * @npages: number of pages in the range
> > - *
> > - * Completes migration of the page by removing special migration
> > entries.
> > - * Drivers must ensure copying of page data is complete and visible
> > to the CPU
> > - * before calling this.
> > - */
> > -void migrate_device_finalize(unsigned long *src_pfns,
> > -			unsigned long *dst_pfns, unsigned long
> > npages)
> > +static void __migrate_device_finalize(unsigned long *src_pfns,
> > +				      unsigned long *dst_pfns,
> > +				      unsigned long npages,
> > +				      struct page *fault_page)
> >  {
> > +	struct folio *fault_folio = fault_page ?
> > +		page_folio(fault_page) : NULL;
> >  	unsigned long i;
> >  
> >  	for (i = 0; i < npages; i++) {
> > @@ -824,6 +831,7 @@ void migrate_device_finalize(unsigned long
> > *src_pfns,
> >  
> >  		if (!page) {
> >  			if (dst) {
> > +				WARN_ON_ONCE(fault_folio == dst);
> >  				folio_unlock(dst);
> >  				folio_put(dst);
> >  			}
> > @@ -834,6 +842,7 @@ void migrate_device_finalize(unsigned long
> > *src_pfns,
> >  
> >  		if (!(src_pfns[i] & MIGRATE_PFN_MIGRATE) || !dst) {
> >  			if (dst) {
> > +				WARN_ON_ONCE(fault_folio == dst);
> >  				folio_unlock(dst);
> >  				folio_put(dst);
> >  			}
> > @@ -841,7 +850,8 @@ void migrate_device_finalize(unsigned long
> > *src_pfns,
> >  		}
> >  
> >  		remove_migration_ptes(src, dst, 0);
> > -		folio_unlock(src);
> > +		if (fault_folio != src)
> > +			folio_unlock(src);
> >  
> >  		if (folio_is_zone_device(src))
> >  			folio_put(src);
> > @@ -849,6 +859,7 @@ void migrate_device_finalize(unsigned long
> > *src_pfns,
> >  			folio_putback_lru(src);
> >  
> >  		if (dst != src) {
> > +			WARN_ON_ONCE(fault_folio == dst);
> >  			folio_unlock(dst);
> >  			if (folio_is_zone_device(dst))
> >  				folio_put(dst);
> > @@ -857,6 +868,22 @@ void migrate_device_finalize(unsigned long
> > *src_pfns,
> >  		}
> >  	}
> >  }
> > +
> > +/*
> > + * migrate_device_finalize() - complete page migration
> > + * @src_pfns: src_pfns returned from migrate_device_range()
> > + * @dst_pfns: array of pfns allocated by the driver to migrate
> > memory to
> > + * @npages: number of pages in the range
> > + *
> > + * Completes migration of the page by removing special migration
> > entries.
> > + * Drivers must ensure copying of page data is complete and visible
> > to the CPU
> > + * before calling this.
> > + */
> > +void migrate_device_finalize(unsigned long *src_pfns,
> > +			unsigned long *dst_pfns, unsigned long
> > npages)
> > +{
> > +	return __migrate_device_finalize(src_pfns, dst_pfns, npages,
> > NULL);
> > +}
> >  EXPORT_SYMBOL(migrate_device_finalize);
> >  
> >  /**
> > @@ -872,7 +899,8 @@ EXPORT_SYMBOL(migrate_device_finalize);
> >   */
> >  void migrate_vma_finalize(struct migrate_vma *migrate)
> >  {
> > -	migrate_device_finalize(migrate->src, migrate->dst, migrate-
> > >npages);
> > +	__migrate_device_finalize(migrate->src, migrate->dst,
> > migrate->npages,
> > +				  migrate->fault_page);
> >  }
> >  EXPORT_SYMBOL(migrate_vma_finalize);
> >  
> 

Re: [PATCH v3 04/30] drm/pagemap: Add DRM pagemap

[Matthew Brost]

On Fri, Jan 24, 2025 at 09:19:41AM +0200, Gwan-gyeong Mun wrote:
> 
> 
> On 12/18/24 1:33 AM, Matthew Brost wrote:
> > From: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> > 
> > Introduce drm_pagemap ops to map and unmap dma to VRAM resources. In the
> > local memory case it's a matter of merely providing an offset into the
> > device's physical address. For future p2p the map and unmap functions may
> > encode as needed.
> > 
> > Similar to how dma-buf works, let the memory provider (drm_pagemap) provide
> > the mapping functionality.
> > 
> > v3:
> >   - Move to drm level include
> > 
> > Signed-off-by: Matthew Brost <matthew.brost@intel.com>
> > Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
> > ---
> >   include/drm/drm_pagemap.h | 103 ++++++++++++++++++++++++++++++++++++++
> >   1 file changed, 103 insertions(+)
> >   create mode 100644 include/drm/drm_pagemap.h
> > 
> > diff --git a/include/drm/drm_pagemap.h b/include/drm/drm_pagemap.h
> > new file mode 100644
> > index 000000000000..b6b387b81479
> > --- /dev/null
> > +++ b/include/drm/drm_pagemap.h
> > @@ -0,0 +1,103 @@
> > +/* SPDX-License-Identifier: MIT */
> > +#ifndef _DRM_PAGEMAP_H_
> > +#define _DRM_PAGEMAP_H_
> > +
> > +#include <linux/dma-direction.h>
> > +#include <linux/hmm.h>
> > +#include <linux/types.h>
> > +
> > +struct drm_pagemap;
> > +struct device;
> > +
> > +/**
> > + * enum drm_interconnect_protocol - Used to identify an interconnect protocol.
> > + */
> > +enum drm_interconnect_protocol {
> > +	DRM_INTERCONNECT_SYSTEM,    /* DMA map is system pages. */
> > +	DRM_INTERCONNECT_PCIE_P2P,  /* DMA map is PCIE P2P */
> > +	DRM_INTERCONNECT_DRIVER,    /* DMA map is driver defined */
> > +	/* A driver can add private values beyond DRM_INTERCONNECT_DRIVER */
> > +};
> > +
> > +/**
> > + * struct drm_pagemap_dma_addr - DMA address representation.
> > + * @addr: The dma address or driver-defined address for driver private interconnects.
> > + * @proto: The interconnect protocol.
> > + * @order: The page order of the dma mapping. (Size is PAGE_SIZE << order).
> > + * @dir: The DMA direction.
> > + *
> > + * Note: There is room for improvement here. We should be able to pack into
> > + * 64 bits.
> > + */
> > +struct drm_pagemap_dma_addr {
> > +	dma_addr_t addr;
> > +	u64 proto : 54;
> > +	u64 order : 8;
> > +	u64 dir : 2;
> > +};
> > +
> > +/**
> > + * drm_pagemap_dma_addr_encode() - Encode a dma address with metadata
> > + * @addr: The dma address or driver-defined address for driver private interconnects.
> > + * @proto: The interconnect protocol.
> > + * @order: The page order of the dma mapping. (Size is PAGE_SIZE << order).
> > + * @dir: The DMA direction.
> > + *
> > + * Return: A struct drm_pagemap_dma_addr encoding the above information.
> > + */
> > +static inline struct drm_pagemap_dma_addr
> > +drm_pagemap_dma_addr_encode(dma_addr_t addr,
> > +			    enum drm_interconnect_protocol proto,
> > +			    unsigned int order,
> > +			    enum dma_data_direction dir)
> > +{
> > +	return (struct drm_pagemap_dma_addr) {
> > +		.addr = addr,
> > +		.proto = proto,
> > +		.order = order,
> > +		.dir = dir,
> > +	};
> > +}
> > +
> > +/**
> > + * struct drm_pagemap_ops: Ops for a drm-pagemap.
> > + */
> Hi,
> In the following comment, the protocol should be removed and an explanation
> of order should be added.
> 

Thanks. There a couple of kernel doc errors here too. Will fix.

Matt

> Br,
> G.G.
> > +struct drm_pagemap_ops {
> > +	/**
> > +	 * @map_dma: Map for dma access or provide a virtual address suitable for
> > +	 * @dev.
> > +	 * @dpagemap: The struct drm_pagemap for the page.
> > +	 * @dev: The dma mapper.
> > +	 * @page: The page to map.
> > +	 * @dir: The transfer direction.
> > +	 * @protocol: The protocol to use.
> > +	 */
> > +	struct drm_pagemap_dma_addr (*map_dma)(struct drm_pagemap *dpagemap,
> > +					       struct device *dev,
> > +					       struct page *page,
> > +					       unsigned int order,
> > +					       enum dma_data_direction dir);
> > +
> > +	/**
> > +	 * @unmap_dma: Unmap a dma address previously obtained using  @map_dma.
> > +	 * @dev: The dma unmapper.
> > +	 * @addr: The dma address obtained when mapping.
> > +	 */
> > +	void (*unmap_dma)(struct drm_pagemap *dpagemap,
> > +			  struct device *dev,
> > +			  struct drm_pagemap_dma_addr addr);
> > +
> > +};
> > +
> > +/**
> > + * struct drm_pagemap: Additional information for a struct dev_pagemap
> > + * used for device p2p handshaking.
> > + * @ops: The struct drm_pagemap_ops.
> > + * @dev: The struct drevice owning the device-private memory.
> > + */
> > +struct drm_pagemap {
> > +	const struct drm_pagemap_ops *ops;
> > +	struct device *dev;
> > +};
> > +
> > +#endif
>