drm/xe/dgfx: Release mmap mappings on rpm suspend

Release all mmap mappings for all vram objects which are associated
with userfault such that, while pcie function in D3hot, any access
to memory mappings will raise a userfault.

Upon userfault, in order to access memory mappings, if graphics
function is in D3 then runtime resume of dgpu will be triggered to
transition to D0.

v2:
  - Avoid iomem check before bo migration check as bo can migrate
    to system memory (Matthew Auld)
v3:
  - Delete bo userfault link during bo destroy
  - Upon bo move (vram-smem), do bo userfault link deletion in
    xe_bo_move_notify instead of xe_bo_move (Thomas Hellström)
  - Grab lock in rpm hook while deleting bo userfault link (Matthew Auld)
v4:
  - Add kernel doc and wrap vram_userfault related
    stuff in the structure (Matthew Auld)
  - Get rpm wakeref before taking dma reserve lock (Matthew Auld)
  - In suspend path apply lock for entire list op
    including list iteration (Matthew Auld)
v5:
  - Use mutex lock instead of spin lock
v6:
  - Fix review comments (Matthew Auld)

Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Matthew Auld <matthew.auld@intel.com>
Cc: Anshuman Gupta <anshuman.gupta@intel.com>
Signed-off-by: Badal Nilawar <badal.nilawar@intel.com>
Acked-by: Thomas Hellström <thomas.hellstrom@linux.intel.com> #For the xe_bo_move_notify() changes
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://lore.kernel.org/r/20240104130702.950078-1-badal.nilawar@intel.com
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>

1 files changed, 52 insertions, 4 deletions

diff --git a/drivers/gpu/drm/xe/xe_bo.c b/drivers/gpu/drm/xe/xe_bo.c
index 8e4a3b1f6b93..2e4d2157179c 100644
--- a/drivers/gpu/drm/xe/xe_bo.c
+++ b/drivers/gpu/drm/xe/xe_bo.c
@@ -586,6 +586,8 @@ static int xe_bo_move_notify(struct xe_bo *bo,
 {
 	struct ttm_buffer_object *ttm_bo = &bo->ttm;
 	struct xe_device *xe = ttm_to_xe_device(ttm_bo->bdev);
+	struct ttm_resource *old_mem = ttm_bo->resource;
+	u32 old_mem_type = old_mem ? old_mem->mem_type : XE_PL_SYSTEM;
 	int ret;
 
 	/*
@@ -605,6 +607,18 @@ static int xe_bo_move_notify(struct xe_bo *bo,
 	if (ttm_bo->base.dma_buf && !ttm_bo->base.import_attach)
 		dma_buf_move_notify(ttm_bo->base.dma_buf);
 
+	/*
+	 * TTM has already nuked the mmap for us (see ttm_bo_unmap_virtual),
+	 * so if we moved from VRAM make sure to unlink this from the userfault
+	 * tracking.
+	 */
+	if (mem_type_is_vram(old_mem_type)) {
+		mutex_lock(&xe->mem_access.vram_userfault.lock);
+		if (!list_empty(&bo->vram_userfault_link))
+			list_del_init(&bo->vram_userfault_link);
+		mutex_unlock(&xe->mem_access.vram_userfault.lock);
+	}
+
 	return 0;
 }
 
@@ -1063,6 +1077,11 @@ static void xe_ttm_bo_destroy(struct ttm_buffer_object *ttm_bo)
 	if (bo->vm && xe_bo_is_user(bo))
 		xe_vm_put(bo->vm);
 
+	mutex_lock(&xe->mem_access.vram_userfault.lock);
+	if (!list_empty(&bo->vram_userfault_link))
+		list_del(&bo->vram_userfault_link);
+	mutex_unlock(&xe->mem_access.vram_userfault.lock);
+
 	kfree(bo);
 }
 
@@ -1110,16 +1129,20 @@ static vm_fault_t xe_gem_fault(struct vm_fault *vmf)
 {
 	struct ttm_buffer_object *tbo = vmf->vma->vm_private_data;
 	struct drm_device *ddev = tbo->base.dev;
+	struct xe_device *xe = to_xe_device(ddev);
+	struct xe_bo *bo = ttm_to_xe_bo(tbo);
+	bool needs_rpm = bo->flags & XE_BO_CREATE_VRAM_MASK;
 	vm_fault_t ret;
 	int idx, r = 0;
 
+	if (needs_rpm)
+		xe_device_mem_access_get(xe);
+
 	ret = ttm_bo_vm_reserve(tbo, vmf);
 	if (ret)
-		return ret;
+		goto out;
 
 	if (drm_dev_enter(ddev, &idx)) {
-		struct xe_bo *bo = ttm_to_xe_bo(tbo);
-
 		trace_xe_bo_cpu_fault(bo);
 
 		if (should_migrate_to_system(bo)) {
@@ -1137,10 +1160,24 @@ static vm_fault_t xe_gem_fault(struct vm_fault *vmf)
 	} else {
 		ret = ttm_bo_vm_dummy_page(vmf, vmf->vma->vm_page_prot);
 	}
+
 	if (ret == VM_FAULT_RETRY && !(vmf->flags & FAULT_FLAG_RETRY_NOWAIT))
-		return ret;
+		goto out;
+	/*
+	 * ttm_bo_vm_reserve() already has dma_resv_lock.
+	 */
+	if (ret == VM_FAULT_NOPAGE && mem_type_is_vram(tbo->resource->mem_type)) {
+		mutex_lock(&xe->mem_access.vram_userfault.lock);
+		if (list_empty(&bo->vram_userfault_link))
+			list_add(&bo->vram_userfault_link, &xe->mem_access.vram_userfault.list);
+		mutex_unlock(&xe->mem_access.vram_userfault.lock);
+	}
 
 	dma_resv_unlock(tbo->base.resv);
+out:
+	if (needs_rpm)
+		xe_device_mem_access_put(xe);
+
 	return ret;
 }
 
@@ -1254,6 +1291,7 @@ struct xe_bo *___xe_bo_create_locked(struct xe_device *xe, struct xe_bo *bo,
 #ifdef CONFIG_PROC_FS
 	INIT_LIST_HEAD(&bo->client_link);
 #endif
+	INIT_LIST_HEAD(&bo->vram_userfault_link);
 
 	drm_gem_private_object_init(&xe->drm, &bo->ttm.base, size);
 
@@ -2264,6 +2302,16 @@ int xe_bo_dumb_create(struct drm_file *file_priv,
 	return err;
 }
 
+void xe_bo_runtime_pm_release_mmap_offset(struct xe_bo *bo)
+{
+	struct ttm_buffer_object *tbo = &bo->ttm;
+	struct ttm_device *bdev = tbo->bdev;
+
+	drm_vma_node_unmap(&tbo->base.vma_node, bdev->dev_mapping);
+
+	list_del_init(&bo->vram_userfault_link);
+}
+
 #if IS_ENABLED(CONFIG_DRM_XE_KUNIT_TEST)
 #include "tests/xe_bo.c"
 #endif