[PATCH 12/16] drm/i915: accurate page size tracking for the ppgtt

[Matthew Auld]

Now that we support multiple page sizes for the ppgtt, it would be
useful to track the real usage for debugging purposes.

Signed-off-by: Matthew Auld <matthew.auld at intel.com>
---
 drivers/gpu/drm/i915/i915_gem_gtt.c    | 16 ++++++++++++++++
 drivers/gpu/drm/i915/i915_gem_object.h | 10 ++++++++++
 2 files changed, 26 insertions(+)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index 27c80a2adc29..d1d41b491b3f 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -215,6 +215,8 @@ static int ppgtt_bind_vma(struct i915_vma *vma,
 static void ppgtt_unbind_vma(struct i915_vma *vma)
 {
 	vma->vm->clear_range(vma->vm, vma->node.start, vma->size);
+
+	vma->page_sizes.gtt = 0;
 }
 
 static gen8_pte_t gen8_pte_encode(dma_addr_t addr,
@@ -920,6 +922,8 @@ static void gen8_ppgtt_insert_3lvl(struct i915_address_space *vm,
 
 	gen8_ppgtt_insert_pte_entries(ppgtt, &ppgtt->pdp, &iter, &idx,
 				      cache_level);
+
+	vma->page_sizes.gtt = I915_GTT_PAGE_SIZE;
 }
 
 static void gen8_ppgtt_insert_huge_entries(struct i915_vma *vma,
@@ -999,8 +1003,10 @@ static void gen8_ppgtt_insert_huge_entries(struct i915_vma *vma,
 				vaddr = kmap_atomic_px(pd);
 				vaddr[idx.pde] |= GEN8_PDE_IPS_64K;
 				kunmap_atomic(vaddr);
+				page_size = I915_GTT_PAGE_SIZE_64K;
 			}
 		}
+		vma->page_sizes.gtt |= page_size;
 	} while (iter->sg);
 }
 
@@ -1025,6 +1031,8 @@ static void gen8_ppgtt_insert_4lvl(struct i915_address_space *vm,
 		while (gen8_ppgtt_insert_pte_entries(ppgtt, pdps[idx.pml4e++],
 						     &iter, &idx, cache_level))
 			GEM_BUG_ON(idx.pml4e >= GEN8_PML4ES_PER_PML4);
+
+		vma->page_sizes.gtt = I915_GTT_PAGE_SIZE;
 	}
 }
 
@@ -1741,6 +1749,8 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
 		}
 	} while (1);
 	kunmap_atomic(vaddr);
+
+	vma->page_sizes.gtt = I915_GTT_PAGE_SIZE;
 }
 
 static int gen6_alloc_va_range(struct i915_address_space *vm,
@@ -2435,6 +2445,8 @@ static int ggtt_bind_vma(struct i915_vma *vma,
 	vma->vm->insert_entries(vma->vm, vma, cache_level, pte_flags);
 	intel_runtime_pm_put(i915);
 
+	vma->page_sizes.gtt = I915_GTT_PAGE_SIZE;
+
 	/*
 	 * Without aliasing PPGTT there's no difference between
 	 * GLOBAL/LOCAL_BIND, it's all the same ptes. Hence unconditionally
@@ -2452,6 +2464,8 @@ static void ggtt_unbind_vma(struct i915_vma *vma)
 	intel_runtime_pm_get(i915);
 	vma->vm->clear_range(vma->vm, vma->node.start, vma->size);
 	intel_runtime_pm_put(i915);
+
+	vma->page_sizes.gtt = 0;
 }
 
 static int aliasing_gtt_bind_vma(struct i915_vma *vma,
@@ -2524,6 +2538,8 @@ static void aliasing_gtt_unbind_vma(struct i915_vma *vma)
 
 		vm->clear_range(vm, vma->node.start, vma->size);
 	}
+
+	vma->page_sizes.gtt = 0;
 }
 
 void i915_gem_gtt_finish_pages(struct drm_i915_gem_object *obj,
diff --git a/drivers/gpu/drm/i915/i915_gem_object.h b/drivers/gpu/drm/i915/i915_gem_object.h
index b487552b5846..5a0ea04eddd4 100644
--- a/drivers/gpu/drm/i915/i915_gem_object.h
+++ b/drivers/gpu/drm/i915/i915_gem_object.h
@@ -145,6 +145,7 @@ struct drm_i915_gem_object {
 		struct sg_table *pages;
 		void *mapping;
 
+		/* TODO: whack some of this into the error state */
 		struct i915_page_sizes {
 			/**
 			 * The sg mask of the pages sg_table. i.e the mask of
@@ -160,6 +161,15 @@ struct drm_i915_gem_object {
 			 * opportunistically.
 			 */
 			unsigned int sg;
+
+			/**
+			 * The actual gtt page size usage. Since we can have
+			 * multiple vma associated with this object we need to
+			 * prevent any trampling of state, hence a copy of this
+			 * struct also lives in each vma, therefore the gtt
+			 * value here should only be read/write through the vma.
+			 */
+			unsigned int gtt;
 		} page_sizes;
 
 		struct i915_gem_object_page_iter {
-- 
2.9.4