[PATCH 74/79] no-used-pdes

Chris Wilson chris at chris-wilson.co.uk
Sun Jan 1 19:37:24 UTC 2017


---
 drivers/gpu/drm/i915/i915_gem_gtt.c | 381 +++++++++++-------------------------
 drivers/gpu/drm/i915/i915_gem_gtt.h |   6 +-
 2 files changed, 119 insertions(+), 268 deletions(-)

diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.c b/drivers/gpu/drm/i915/i915_gem_gtt.c
index c81ab02e8fb6..5a90525e6411 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.c
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.c
@@ -452,12 +452,8 @@ static void free_pt(struct i915_address_space *vm, struct i915_page_table *pt)
 static void gen8_initialize_pt(struct i915_address_space *vm,
 			       struct i915_page_table *pt)
 {
-	gen8_pte_t scratch_pte;
-
-	scratch_pte = gen8_pte_encode(vm->scratch_page.daddr,
-				      I915_CACHE_LLC);
-
-	fill_px(vm, pt, scratch_pte);
+	fill_px(vm, pt,
+		gen8_pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC));
 }
 
 static void gen6_initialize_pt(struct i915_address_space *vm,
@@ -476,39 +472,25 @@ static void gen6_initialize_pt(struct i915_address_space *vm,
 static struct i915_page_directory *alloc_pd(struct i915_address_space *vm)
 {
 	struct i915_page_directory *pd;
-	int ret = -ENOMEM;
 
-	pd = kzalloc(sizeof(*pd), GFP_KERNEL);
-	if (!pd)
+	pd = kzalloc(sizeof(*pd), GFP_KERNEL | __GFP_NOWARN);
+	if (unlikely(!pd))
 		return ERR_PTR(-ENOMEM);
 
-	pd->used_pdes = kcalloc(BITS_TO_LONGS(I915_PDES),
-				sizeof(*pd->used_pdes), GFP_KERNEL);
-	if (!pd->used_pdes)
-		goto fail_bitmap;
-
-	ret = setup_px(vm, pd);
-	if (ret)
-		goto fail_page_m;
+	if (unlikely(setup_px(vm, pd))) {
+		kfree(pd);
+		return ERR_PTR(-ENOMEM);
+	}
 
+	pd->used_pdes = 0;
 	return pd;
-
-fail_page_m:
-	kfree(pd->used_pdes);
-fail_bitmap:
-	kfree(pd);
-
-	return ERR_PTR(ret);
 }
 
 static void free_pd(struct i915_address_space *vm,
 		    struct i915_page_directory *pd)
 {
-	if (px_page(pd)) {
-		cleanup_px(vm, pd);
-		kfree(pd->used_pdes);
-		kfree(pd);
-	}
+	cleanup_px(vm, pd);
+	kfree(pd);
 }
 
 static void gen8_initialize_pd(struct i915_address_space *vm,
@@ -522,10 +504,11 @@ static void gen8_initialize_pd(struct i915_address_space *vm,
 		pd->page_table[i] = vm->scratch_pt;
 }
 
-static int __pdp_init(struct drm_i915_private *dev_priv,
+static int __pdp_init(struct i915_address_space *vm,
 		      struct i915_page_directory_pointer *pdp)
 {
-	size_t pdpes = I915_PDPES_PER_PDP(dev_priv);
+	size_t pdpes = I915_PDPES_PER_PDP(vm->i915);
+	int i;
 
 	pdp->used_pdpes = kcalloc(BITS_TO_LONGS(pdpes),
 				  sizeof(unsigned long),
@@ -533,8 +516,8 @@ static int __pdp_init(struct drm_i915_private *dev_priv,
 	if (!pdp->used_pdpes)
 		return -ENOMEM;
 
-	pdp->page_directory = kcalloc(pdpes, sizeof(*pdp->page_directory),
-				      GFP_KERNEL);
+	pdp->page_directory = kmalloc_array(pdpes, sizeof(*pdp->page_directory),
+					    GFP_KERNEL);
 	if (!pdp->page_directory) {
 		kfree(pdp->used_pdpes);
 		/* the PDP might be the statically allocated top level. Keep it
@@ -543,6 +526,9 @@ static int __pdp_init(struct drm_i915_private *dev_priv,
 		return -ENOMEM;
 	}
 
+	for (i = 0; i < pdpes; i++)
+		pdp->page_directory[i] = vm->scratch_pd;
+
 	return 0;
 }
 
@@ -565,7 +551,7 @@ alloc_pdp(struct i915_address_space *vm)
 	if (!pdp)
 		return ERR_PTR(-ENOMEM);
 
-	ret = __pdp_init(vm->i915, pdp);
+	ret = __pdp_init(vm, pdp);
 	if (ret)
 		goto fail_bitmap;
 
@@ -615,22 +601,6 @@ static void gen8_initialize_pml4(struct i915_address_space *vm,
 }
 
 static void
-gen8_setup_page_directory(struct i915_hw_ppgtt *ppgtt,
-			  struct i915_page_directory_pointer *pdp,
-			  struct i915_page_directory *pd,
-			  int index)
-{
-	gen8_ppgtt_pdpe_t *page_directorypo;
-
-	if (!USES_FULL_48BIT_PPGTT(to_i915(ppgtt->base.dev)))
-		return;
-
-	page_directorypo = kmap_atomic_px(pdp);
-	page_directorypo[index] = gen8_pdpe_encode(px_dma(pd), I915_CACHE_LLC);
-	kunmap_atomic(page_directorypo);
-}
-
-static void
 gen8_setup_page_directory_pointer(struct i915_hw_ppgtt *ppgtt,
 				  struct i915_pml4 *pml4,
 				  struct i915_page_directory_pointer *pdp,
@@ -743,9 +713,6 @@ static void gen8_ppgtt_set_pde(struct i915_address_space *vm,
 	kunmap_atomic(vaddr);
 }
 
-/* Removes entries from a single page dir, releasing it if it's empty.
- * Caller can use the return value to update higher-level entries
- */
 static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
 				struct i915_page_directory *pd,
 				u64 start, u64 length)
@@ -758,15 +725,28 @@ static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
 			continue;
 
 		gen8_ppgtt_set_pde(vm, pd, vm->scratch_pt, pde);
-		__clear_bit(pde, pd->used_pdes);
+		pd->used_pdes--;
 
 		free_pt(vm, pt);
 	}
 
-	if (bitmap_empty(pd->used_pdes, I915_PDES))
-		return true;
+	return !pd->used_pdes;
+}
 
-	return false;
+static void gen8_ppgtt_set_pdpe(struct i915_address_space *vm,
+				struct i915_page_directory_pointer *pdp,
+				struct i915_page_directory *pd,
+				unsigned int pdpe)
+{
+	gen8_ppgtt_pdpe_t *vaddr;
+
+	pdp->page_directory[pdpe] = pd;
+	if (!USES_FULL_48BIT_PPGTT(vm->i915))
+		return;
+
+	vaddr = kmap_atomic_px(pdp);
+	vaddr[pdpe] = gen8_pdpe_encode(px_dma(pd), I915_CACHE_LLC);
+	kunmap_atomic(vaddr);
 }
 
 /* Removes entries from a single page dir pointer, releasing it if it's empty.
@@ -774,32 +754,20 @@ static bool gen8_ppgtt_clear_pd(struct i915_address_space *vm,
  */
 static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm,
 				 struct i915_page_directory_pointer *pdp,
-				 uint64_t start,
-				 uint64_t length)
+				 u64 start, u64 length)
 {
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_directory *pd;
-	uint64_t pdpe;
-	gen8_ppgtt_pdpe_t *pdpe_vaddr;
-	gen8_ppgtt_pdpe_t scratch_pdpe =
-		gen8_pdpe_encode(px_dma(vm->scratch_pd), I915_CACHE_LLC);
+	unsigned int pdpe;
 
 	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		if (WARN_ON(!pdp->page_directory[pdpe]))
-			break;
+		if (!gen8_ppgtt_clear_pd(vm, pd, start, length))
+			continue;
 
-		if (gen8_ppgtt_clear_pd(vm, pd, start, length)) {
-			__clear_bit(pdpe, pdp->used_pdpes);
-			if (USES_FULL_48BIT_PPGTT(vm->i915)) {
-				pdpe_vaddr = kmap_atomic_px(pdp);
-				pdpe_vaddr[pdpe] = scratch_pdpe;
-				kunmap_atomic(pdpe_vaddr);
-			}
-			free_pd(vm, pd);
-		}
-	}
+		gen8_ppgtt_set_pdpe(vm, pdp, vm->scratch_pd, pdpe);
+		__clear_bit(pdpe, pdp->used_pdpes);
 
-	mark_tlbs_dirty(ppgtt);
+		free_pd(vm, pd);
+	}
 
 	if (bitmap_empty(pdp->used_pdpes, I915_PDPES_PER_PDP(dev_priv)))
 		return true;
@@ -807,15 +775,22 @@ static bool gen8_ppgtt_clear_pdp(struct i915_address_space *vm,
 	return false;
 }
 
+
+static void gen8_ppgtt_clear_3lvl(struct i915_address_space *vm,
+				  u64 start, u64 length)
+{
+	gen8_ppgtt_clear_pdp(vm, &i915_vm_to_ppgtt(vm)->pdp, start, length);
+}
+
 /* Removes entries from a single pml4.
  * This is the top-level structure in 4-level page tables used on gen8+.
  * Empty entries are always scratch pml4e.
  */
-static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm,
-				  struct i915_pml4 *pml4,
-				  uint64_t start,
-				  uint64_t length)
+static void gen8_ppgtt_clear_4lvl(struct i915_address_space *vm,
+				  u64 start, u64 length)
 {
+	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
+	struct i915_pml4 *pml4 = &ppgtt->pml4;
 	struct i915_page_directory_pointer *pdp;
 	uint64_t pml4e;
 	gen8_ppgtt_pml4e_t *pml4e_vaddr;
@@ -838,17 +813,6 @@ static void gen8_ppgtt_clear_pml4(struct i915_address_space *vm,
 	}
 }
 
-static void gen8_ppgtt_clear_range(struct i915_address_space *vm,
-				   uint64_t start, uint64_t length)
-{
-	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-
-	if (USES_FULL_48BIT_PPGTT(vm->i915))
-		gen8_ppgtt_clear_pml4(vm, &ppgtt->pml4, start, length);
-	else
-		gen8_ppgtt_clear_pdp(vm, &ppgtt->pdp, start, length);
-}
-
 struct sgt_dma {
 	struct scatterlist *sg;
 	dma_addr_t dma, max;
@@ -946,12 +910,9 @@ static void gen8_free_page_tables(struct i915_address_space *vm,
 	if (!px_page(pd))
 		return;
 
-	for_each_set_bit(i, pd->used_pdes, I915_PDES) {
-		if (WARN_ON(!pd->page_table[i]))
-			continue;
-
-		free_pt(vm, pd->page_table[i]);
-		pd->page_table[i] = NULL;
+	for (i = 0; i < I915_PDES; i++) {
+		if (pd->page_table[i] != vm->scratch_pt)
+			free_pt(vm, pd->page_table[i]);
 	}
 }
 
@@ -1046,7 +1007,7 @@ static void gen8_ppgtt_cleanup_3lvl(struct i915_address_space *vm,
 	int i;
 
 	for_each_set_bit(i, pdp->used_pdpes, I915_PDPES_PER_PDP(vm->i915)) {
-		if (WARN_ON(!pdp->page_directory[i]))
+		if (pdp->page_directory[i] == vm->scratch_pd)
 			continue;
 
 		gen8_free_page_tables(vm, pdp->page_directory[i]);
@@ -1086,44 +1047,28 @@ static void gen8_ppgtt_cleanup(struct i915_address_space *vm)
 	gen8_free_scratch(vm);
 }
 
-/**
- * gen8_ppgtt_alloc_pagetabs() - Allocate page tables for VA range.
- * @vm:	Master vm structure.
- * @pd:	Page directory for this address range.
- * @start:	Starting virtual address to begin allocations.
- * @length:	Size of the allocations.
- *
- * Allocate the required number of page tables. Extremely similar to
- * gen8_ppgtt_alloc_page_directories(). The main difference is here we are limited by
- * the page directory boundary (instead of the page directory pointer). That
- * boundary is 1GB virtual. Therefore, unlike gen8_ppgtt_alloc_page_directories(), it is
- * possible, and likely that the caller will need to use multiple calls of this
- * function to achieve the appropriate allocation.
- *
- * Return: 0 if success; negative error code otherwise.
- */
-static int gen8_ppgtt_alloc_pagetabs(struct i915_address_space *vm,
-				     struct i915_page_directory *pd,
-				     u64 start, u64 length)
+static int gen8_ppgtt_alloc_pd(struct i915_address_space *vm,
+			       struct i915_page_directory *pd,
+			       u64 start, u64 length)
 {
 	struct i915_page_table *pt;
 	u64 from = start;
-	uint32_t pde;
+	unsigned int pde;
 
 	gen8_for_each_pde(pt, pd, start, length, pde) {
-		/* Don't reallocate page tables */
-		if (!test_bit(pde, pd->used_pdes)) {
+		if (pt == vm->scratch_pt) {
 			pt = alloc_pt(vm);
 			if (IS_ERR(pt))
 				goto unwind;
 
 			gen8_initialize_pt(vm, pt);
-			pd->page_table[pde] = pt;
+
+			gen8_ppgtt_set_pde(vm, pd, pt, pde);
+			pd->used_pdes++;
 		}
+
 		pt->used_ptes += gen8_pte_count(start, length);
-		trace_i915_page_table_entry_alloc(vm, pde, start, GEN8_PDE_SHIFT);
 	}
-
 	return 0;
 
 unwind:
@@ -1200,7 +1145,7 @@ gen8_ppgtt_alloc_page_directories(struct i915_address_space *vm,
  *		caller to free on error.
  *
  * Allocate the required number of page directory pointers. Extremely similar to
- * gen8_ppgtt_alloc_page_directories() and gen8_ppgtt_alloc_pagetabs().
+ * gen8_ppgtt_alloc_page_directories() and gen8_ppgtt_alloc_pd().
  * The main difference is here we are limited by the pml4 boundary (instead of
  * the page directory pointer).
  *
@@ -1252,9 +1197,8 @@ free_gen8_temp_bitmaps(unsigned long *new_pds)
 /* Fills in the page directory bitmap, and the array of page tables bitmap. Both
  * of these are based on the number of PDPEs in the system.
  */
-static
-int __must_check alloc_gen8_temp_bitmaps(unsigned long **new_pds,
-					 uint32_t pdpes)
+static int __must_check
+alloc_gen8_temp_bitmaps(unsigned long **new_pds, uint32_t pdpes)
 {
 	unsigned long *pds;
 
@@ -1274,8 +1218,6 @@ static int gen8_alloc_va_range_3lvl(struct i915_address_space *vm,
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	unsigned long *new_page_dirs;
 	struct i915_page_directory *pd;
-	const uint64_t orig_start = start;
-	const uint64_t orig_length = length;
 	uint32_t pdpe;
 	uint32_t pdpes = I915_PDPES_PER_PDP(dev_priv);
 	int ret;
@@ -1294,51 +1236,16 @@ static int gen8_alloc_va_range_3lvl(struct i915_address_space *vm,
 
 	/* For every page directory referenced, allocate page tables */
 	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		ret = gen8_ppgtt_alloc_pagetabs(vm, pd, start, length);
+		ret = gen8_ppgtt_alloc_pd(vm, pd, start, length);
 		if (ret)
 			goto err_out;
-	}
 
-	start = orig_start;
-	length = orig_length;
+		if (test_and_set_bit(pdpe, pdp->used_pdpes))
+			gen8_ppgtt_set_pdpe(vm, pdp, pd, pdpe);
+	}
 
 	/* Allocations have completed successfully, so set the bitmaps, and do
 	 * the mappings. */
-	gen8_for_each_pdpe(pd, pdp, start, length, pdpe) {
-		gen8_pde_t *const page_directory = kmap_atomic_px(pd);
-		struct i915_page_table *pt;
-		uint64_t pd_len = length;
-		uint64_t pd_start = start;
-		uint32_t pde;
-
-		/* Every pd should be allocated, we just did that above. */
-		WARN_ON(!pd);
-
-		gen8_for_each_pde(pt, pd, pd_start, pd_len, pde) {
-			/* Same reasoning as pd */
-			WARN_ON(!pt);
-			WARN_ON(!pd_len);
-			WARN_ON(!gen8_pte_count(pd_start, pd_len));
-
-			/* Our pde is now pointing to the pagetable, pt */
-			__set_bit(pde, pd->used_pdes);
-
-			/* Map the PDE to the page table */
-			page_directory[pde] = gen8_pde_encode(px_dma(pt),
-							      I915_CACHE_LLC);
-			trace_i915_page_table_entry_map(&ppgtt->base, pde, pt,
-							gen8_pte_index(start),
-							gen8_pte_count(start, length));
-
-			/* NB: We haven't yet mapped ptes to pages. At this
-			 * point we're still relying on insert_entries() */
-		}
-
-		kunmap_atomic(page_directory);
-		__set_bit(pdpe, pdp->used_pdpes);
-		gen8_setup_page_directory(ppgtt, pdp, pd, pdpe);
-	}
-
 	free_gen8_temp_bitmaps(new_page_dirs);
 	mark_tlbs_dirty(ppgtt);
 	return 0;
@@ -1431,7 +1338,7 @@ static void gen8_dump_pdp(struct i915_hw_ppgtt *ppgtt,
 			uint32_t  pte;
 			gen8_pte_t *pt_vaddr;
 
-			if (!test_bit(pde, pd->used_pdes))
+			if (pd->page_table[pde] == ppgtt->base.scratch_pt)
 				continue;
 
 			pt_vaddr = kmap_atomic_px(pt);
@@ -1534,7 +1441,6 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 	ppgtt->base.start = 0;
 	ppgtt->base.cleanup = gen8_ppgtt_cleanup;
 	ppgtt->base.allocate_va_range = gen8_alloc_va_range;
-	ppgtt->base.clear_range = gen8_ppgtt_clear_range;
 	ppgtt->base.unbind_vma = ppgtt_unbind_vma;
 	ppgtt->base.bind_vma = ppgtt_bind_vma;
 	ppgtt->debug_dump = gen8_dump_ppgtt;
@@ -1556,8 +1462,9 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 		ppgtt->switch_mm = gen8_48b_mm_switch;
 
 		ppgtt->base.insert_entries = gen8_ppgtt_insert_4lvl;
+		ppgtt->base.clear_range = gen8_ppgtt_clear_4lvl;
 	} else {
-		ret = __pdp_init(dev_priv, &ppgtt->pdp);
+		ret = __pdp_init(&ppgtt->base, &ppgtt->pdp);
 		if (ret)
 			goto free_scratch;
 
@@ -1574,6 +1481,7 @@ static int gen8_ppgtt_init(struct i915_hw_ppgtt *ppgtt)
 		}
 
 		ppgtt->base.insert_entries = gen8_ppgtt_insert_3lvl;
+		ppgtt->base.clear_range = gen8_ppgtt_clear_3lvl;
 	}
 
 	if (intel_vgpu_active(dev_priv))
@@ -1808,35 +1716,40 @@ static void gen6_ppgtt_enable(struct drm_i915_private *dev_priv)
 
 /* PPGTT support for Sandybdrige/Gen6 and later */
 static void gen6_ppgtt_clear_range(struct i915_address_space *vm,
-				   uint64_t start,
-				   uint64_t length)
+				   u64 start, u64 length)
 {
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
-	gen6_pte_t *pt_vaddr, scratch_pte;
 	unsigned first_entry = start >> PAGE_SHIFT;
 	unsigned num_entries = length >> PAGE_SHIFT;
-	unsigned act_pt = first_entry / GEN6_PTES;
+	unsigned pde = first_entry / GEN6_PTES;
 	unsigned first_pte = first_entry % GEN6_PTES;
-	unsigned last_pte, i;
-
-	scratch_pte = vm->pte_encode(vm->scratch_page.daddr,
-				     I915_CACHE_LLC, 0);
+	gen6_pte_t scratch_pte =
+	       	vm->pte_encode(vm->scratch_page.daddr, I915_CACHE_LLC, 0);
 
 	while (num_entries) {
-		last_pte = first_pte + num_entries;
-		if (last_pte > GEN6_PTES)
-			last_pte = GEN6_PTES;
-
-		pt_vaddr = kmap_atomic_px(ppgtt->pd.page_table[act_pt]);
-
-		for (i = first_pte; i < last_pte; i++)
-			pt_vaddr[i] = scratch_pte;
+		struct i915_page_table *pt = ppgtt->pd.page_table[pde];
+		unsigned int last_pte =
+		       	min_t(unsigned int, first_pte + num_entries, GEN6_PTES);
 
-		kunmap_atomic(pt_vaddr);
+		GEM_BUG_ON(pt->used_ptes < last_pte - first_pte);
+		pt->used_ptes -= last_pte - first_pte;
+		if (!pt->used_ptes) {
+			free_pt(vm, pt);
+			ppgtt->pd.page_table[pde] = vm->scratch_pt;
+			gen6_write_pde(&ppgtt->pd, pde, vm->scratch_pt);
+		} else {
+			gen6_pte_t *vaddr;
+			unsigned int i;
+
+			vaddr = kmap_atomic_px(pt);
+			for (i = first_pte; i < last_pte; i++)
+				vaddr[i] = scratch_pte;
+			kunmap_atomic(vaddr);
+		}
 
 		num_entries -= last_pte - first_pte;
 		first_pte = 0;
-		act_pt++;
+		pde++;
 	}
 }
 
@@ -1880,81 +1793,34 @@ static void gen6_ppgtt_insert_entries(struct i915_address_space *vm,
 }
 
 static int gen6_alloc_va_range(struct i915_address_space *vm,
-			       uint64_t start_in, uint64_t length_in)
+			       u64 start, u64 length)
 {
-	DECLARE_BITMAP(new_page_tables, I915_PDES);
-	struct drm_i915_private *dev_priv = vm->i915;
-	struct i915_ggtt *ggtt = &dev_priv->ggtt;
 	struct i915_hw_ppgtt *ppgtt = i915_vm_to_ppgtt(vm);
 	struct i915_page_table *pt;
-	uint32_t start, length, start_save, length_save;
-	uint32_t pde;
-	int ret;
-
-	start = start_save = start_in;
-	length = length_save = length_in;
-
-	bitmap_zero(new_page_tables, I915_PDES);
-
-	/* The allocation is done in two stages so that we can bail out with
-	 * minimal amount of pain. The first stage finds new page tables that
-	 * need allocation. The second stage marks use ptes within the page
-	 * tables.
-	 */
-	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde) {
-		if (pt != vm->scratch_pt)
-			continue;
-
-		pt = alloc_pt(vm);
-		if (IS_ERR(pt)) {
-			ret = PTR_ERR(pt);
-			goto unwind_out;
-		}
-
-		gen6_initialize_pt(vm, pt);
-
-		ppgtt->pd.page_table[pde] = pt;
-		__set_bit(pde, new_page_tables);
-		trace_i915_page_table_entry_alloc(vm, pde, start, GEN6_PDE_SHIFT);
-	}
-
-	start = start_save;
-	length = length_save;
+	u64 from = start;
+	unsigned int pde;
 
 	gen6_for_each_pde(pt, &ppgtt->pd, start, length, pde) {
-		DECLARE_BITMAP(tmp_bitmap, GEN6_PTES);
+		if (pt == vm->scratch_pt) {
+			pt = alloc_pt(vm);
+			if (IS_ERR(pt))
+				goto unwind_out;
 
-		bitmap_zero(tmp_bitmap, GEN6_PTES);
-		bitmap_set(tmp_bitmap, gen6_pte_index(start),
-			   gen6_pte_count(start, length));
+			gen6_initialize_pt(vm, pt);
+			ppgtt->pd.page_table[pde] = pt;
 
-		if (__test_and_clear_bit(pde, new_page_tables))
 			gen6_write_pde(&ppgtt->pd, pde, pt);
+			mark_tlbs_dirty(ppgtt);
+		}
 
-		trace_i915_page_table_entry_map(vm, pde, pt,
-					 gen6_pte_index(start),
-					 gen6_pte_count(start, length));
+		pt->used_ptes += gen6_pte_count(start, length);
 	}
 
-	WARN_ON(!bitmap_empty(new_page_tables, I915_PDES));
-
-	/* Make sure write is complete before other code can use this page
-	 * table. Also require for WC mapped PTEs */
-	readl(ggtt->gsm);
-
-	mark_tlbs_dirty(ppgtt);
 	return 0;
 
 unwind_out:
-	for_each_set_bit(pde, new_page_tables, I915_PDES) {
-		struct i915_page_table *pt = ppgtt->pd.page_table[pde];
-
-		ppgtt->pd.page_table[pde] = vm->scratch_pt;
-		free_pt(vm, pt);
-	}
-
-	mark_tlbs_dirty(ppgtt);
-	return ret;
+	gen6_ppgtt_clear_range(vm, from, start);
+	return -ENOMEM;
 }
 
 static int gen6_init_scratch(struct i915_address_space *vm)
@@ -2721,17 +2587,6 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 			goto err;
 		}
 
-		if (ppgtt->base.allocate_va_range) {
-			ret = ppgtt->base.allocate_va_range(&ppgtt->base, 0,
-							    ppgtt->base.total);
-			if (ret)
-				goto err_ppgtt;
-		}
-
-		ppgtt->base.clear_range(&ppgtt->base,
-					ppgtt->base.start,
-					ppgtt->base.total);
-
 		dev_priv->mm.aliasing_ppgtt = ppgtt;
 		WARN_ON(ggtt->base.bind_vma != ggtt_bind_vma);
 		ggtt->base.bind_vma = aliasing_gtt_bind_vma;
@@ -2739,8 +2594,6 @@ int i915_gem_init_ggtt(struct drm_i915_private *dev_priv)
 
 	return 0;
 
-err_ppgtt:
-	i915_ppgtt_put(ppgtt);
 err:
 	drm_mm_remove_node(&ggtt->error_capture);
 	return ret;
diff --git a/drivers/gpu/drm/i915/i915_gem_gtt.h b/drivers/gpu/drm/i915/i915_gem_gtt.h
index 9cb1c0e6f194..55a7a1d443c3 100644
--- a/drivers/gpu/drm/i915/i915_gem_gtt.h
+++ b/drivers/gpu/drm/i915/i915_gem_gtt.h
@@ -222,8 +222,8 @@ struct i915_page_table {
 struct i915_page_directory {
 	struct i915_page_dma base;
 
-	unsigned long *used_pdes;
 	struct i915_page_table *page_table[I915_PDES]; /* PDEs */
+	unsigned int used_pdes;
 };
 
 struct i915_page_directory_pointer {
@@ -513,9 +513,7 @@ static inline size_t gen8_pte_count(uint64_t address, uint64_t length)
 static inline dma_addr_t
 i915_page_dir_dma_addr(const struct i915_hw_ppgtt *ppgtt, const unsigned n)
 {
-	return test_bit(n, ppgtt->pdp.used_pdpes) ?
-		px_dma(ppgtt->pdp.page_directory[n]) :
-		px_dma(ppgtt->base.scratch_pd);
+	return px_dma(ppgtt->pdp.page_directory[n]);
 }
 
 static inline struct i915_ggtt *
-- 
2.11.0



More information about the Intel-gfx-trybot mailing list