drm/i915/gt: Switch to object allocations for page directories

The GEM object is grossly overweight for the practicality of tracking
large numbers of individual pages, yet it is currently our only
abstraction for tracking DMA allocations. Since those allocations need
to be reserved upfront before an operation, and that we need to break
away from simple system memory, we need to ditch using plain struct page
wrappers.

In the process, we drop the WC mapping as we ended up clflushing
everything anyway due to various issues across a wider range of
platforms. Though in a future step, we need to drop the kmap_atomic
approach which suggests we need to pre-map all the pages and keep them
mapped.

v2: Verify our large scratch page is suitably DMA aligned; and manually
clear the scratch since we are allocating plain struct pages full of
prior content.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200729164219.5737-2-chris@chris-wilson.co.uk
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>

1 files changed, 34 insertions, 60 deletions

diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.h b/drivers/gpu/drm/i915/gt/intel_gtt.h
index 0d9f29aea6b4..6abab2d37b6f 100644
--- a/drivers/gpu/drm/i915/gt/intel_gtt.h
+++ b/drivers/gpu/drm/i915/gt/intel_gtt.h
@@ -134,31 +134,19 @@ typedef u64 gen8_pte_t;
 #define GEN8_PDE_IPS_64K BIT(11)
 #define GEN8_PDE_PS_2M   BIT(7)
 
+enum i915_cache_level;
+
+struct drm_i915_file_private;
+struct drm_i915_gem_object;
 struct i915_fence_reg;
+struct i915_vma;
+struct intel_gt;
 
 #define for_each_sgt_daddr(__dp, __iter, __sgt) \
 	__for_each_sgt_daddr(__dp, __iter, __sgt, I915_GTT_PAGE_SIZE)
 
-struct i915_page_dma {
-	struct page *page;
-	union {
-		dma_addr_t daddr;
-
-		/*
-		 * For gen6/gen7 only. This is the offset in the GGTT
-		 * where the page directory entries for PPGTT begin
-		 */
-		u32 ggtt_offset;
-	};
-};
-
-struct i915_page_scratch {
-	struct i915_page_dma base;
-	u64 encode;
-};
-
 struct i915_page_table {
-	struct i915_page_dma base;
+	struct drm_i915_gem_object *base;
 	union {
 		atomic_t used;
 		struct i915_page_table *stash;
@@ -179,12 +167,14 @@ struct i915_page_directory {
 	other)
 
 #define px_base(px) \
-	__px_choose_expr(px, struct i915_page_dma *, __x, \
-	__px_choose_expr(px, struct i915_page_scratch *, &__x->base, \
-	__px_choose_expr(px, struct i915_page_table *, &__x->base, \
-	__px_choose_expr(px, struct i915_page_directory *, &__x->pt.base, \
-	(void)0))))
-#define px_dma(px) (px_base(px)->daddr)
+	__px_choose_expr(px, struct drm_i915_gem_object *, __x, \
+	__px_choose_expr(px, struct i915_page_table *, __x->base, \
+	__px_choose_expr(px, struct i915_page_directory *, __x->pt.base, \
+	(void)0)))
+
+struct page *__px_page(struct drm_i915_gem_object *p);
+dma_addr_t __px_dma(struct drm_i915_gem_object *p);
+#define px_dma(px) (__px_dma(px_base(px)))
 
 #define px_pt(px) \
 	__px_choose_expr(px, struct i915_page_table *, __x, \
@@ -192,13 +182,6 @@ struct i915_page_directory {
 	(void)0))
 #define px_used(px) (&px_pt(px)->used)
 
-enum i915_cache_level;
-
-struct drm_i915_file_private;
-struct drm_i915_gem_object;
-struct i915_vma;
-struct intel_gt;
-
 struct i915_vm_pt_stash {
 	/* preallocated chains of page tables/directories */
 	struct i915_page_table *pt[2];
@@ -222,13 +205,6 @@ struct i915_vma_ops {
 	void (*clear_pages)(struct i915_vma *vma);
 };
 
-struct pagestash {
-	spinlock_t lock;
-	struct pagevec pvec;
-};
-
-void stash_init(struct pagestash *stash);
-
 struct i915_address_space {
 	struct kref ref;
 	struct rcu_work rcu;
@@ -265,20 +241,15 @@ struct i915_address_space {
 #define VM_CLASS_GGTT 0
 #define VM_CLASS_PPGTT 1
 
-	struct i915_page_scratch scratch[4];
+	struct drm_i915_gem_object *scratch[4];
 	/**
 	 * List of vma currently bound.
 	 */
 	struct list_head bound_list;
 
-	struct pagestash free_pages;
-
 	/* Global GTT */
 	bool is_ggtt:1;
 
-	/* Some systems require uncached updates of the page directories */
-	bool pt_kmap_wc:1;
-
 	/* Some systems support read-only mappings for GGTT and/or PPGTT */
 	bool has_read_only:1;
 
@@ -286,6 +257,9 @@ struct i915_address_space {
 	u8 pd_shift;
 	u8 scratch_order;
 
+	struct drm_i915_gem_object *
+		(*alloc_pt_dma)(struct i915_address_space *vm, int sz);
+
 	u64 (*pte_encode)(dma_addr_t addr,
 			  enum i915_cache_level level,
 			  u32 flags); /* Create a valid PTE */
@@ -501,9 +475,9 @@ i915_pd_entry(const struct i915_page_directory * const pdp,
 static inline dma_addr_t
 i915_page_dir_dma_addr(const struct i915_ppgtt *ppgtt, const unsigned int n)
 {
-	struct i915_page_dma *pt = ppgtt->pd->entry[n];
+	struct i915_page_table *pt = ppgtt->pd->entry[n];
 
-	return px_dma(pt ?: px_base(&ppgtt->vm.scratch[ppgtt->vm.top]));
+	return __px_dma(pt ? px_base(pt) : ppgtt->vm.scratch[ppgtt->vm.top]);
 }
 
 void ppgtt_init(struct i915_ppgtt *ppgtt, struct intel_gt *gt);
@@ -528,13 +502,10 @@ struct i915_ppgtt *i915_ppgtt_create(struct intel_gt *gt);
 void i915_ggtt_suspend(struct i915_ggtt *gtt);
 void i915_ggtt_resume(struct i915_ggtt *ggtt);
 
-int setup_page_dma(struct i915_address_space *vm, struct i915_page_dma *p);
-void cleanup_page_dma(struct i915_address_space *vm, struct i915_page_dma *p);
-
-#define kmap_atomic_px(px) kmap_atomic(px_base(px)->page)
+#define kmap_atomic_px(px) kmap_atomic(__px_page(px_base(px)))
 
 void
-fill_page_dma(const struct i915_page_dma *p, const u64 val, unsigned int count);
+fill_page_dma(struct drm_i915_gem_object *p, const u64 val, unsigned int count);
 
 #define fill_px(px, v) fill_page_dma(px_base(px), (v), PAGE_SIZE / sizeof(u64))
 #define fill32_px(px, v) do {						\
@@ -542,37 +513,38 @@ fill_page_dma(const struct i915_page_dma *p, const u64 val, unsigned int count);
 	fill_px((px), v__ << 32 | v__);					\
 } while (0)
 
-int setup_scratch_page(struct i915_address_space *vm, gfp_t gfp);
-void cleanup_scratch_page(struct i915_address_space *vm);
+int setup_scratch_page(struct i915_address_space *vm);
 void free_scratch(struct i915_address_space *vm);
 
+struct drm_i915_gem_object *alloc_pt_dma(struct i915_address_space *vm, int sz);
 struct i915_page_table *alloc_pt(struct i915_address_space *vm);
 struct i915_page_directory *alloc_pd(struct i915_address_space *vm);
 struct i915_page_directory *__alloc_pd(size_t sz);
 
-void free_pd(struct i915_address_space *vm, struct i915_page_dma *pd);
+int pin_pt_dma(struct i915_address_space *vm, struct drm_i915_gem_object *obj);
 
-#define free_px(vm, px) free_pd(vm, px_base(px))
+void free_pt(struct i915_address_space *vm, struct i915_page_table *pt);
+#define free_px(vm, px) free_pt(vm, px_pt(px))
 
 void
 __set_pd_entry(struct i915_page_directory * const pd,
 	       const unsigned short idx,
-	       struct i915_page_dma * const to,
+	       struct i915_page_table *pt,
 	       u64 (*encode)(const dma_addr_t, const enum i915_cache_level));
 
 #define set_pd_entry(pd, idx, to) \
-	__set_pd_entry((pd), (idx), px_base(to), gen8_pde_encode)
+	__set_pd_entry((pd), (idx), px_pt(to), gen8_pde_encode)
 
 void
 clear_pd_entry(struct i915_page_directory * const pd,
 	       const unsigned short idx,
-	       const struct i915_page_scratch * const scratch);
+	       const struct drm_i915_gem_object * const scratch);
 
 bool
 release_pd_entry(struct i915_page_directory * const pd,
 		 const unsigned short idx,
 		 struct i915_page_table * const pt,
-		 const struct i915_page_scratch * const scratch);
+		 const struct drm_i915_gem_object * const scratch);
 void gen6_ggtt_invalidate(struct i915_ggtt *ggtt);
 
 int ggtt_set_pages(struct i915_vma *vma);
@@ -594,6 +566,8 @@ void setup_private_pat(struct intel_uncore *uncore);
 int i915_vm_alloc_pt_stash(struct i915_address_space *vm,
 			   struct i915_vm_pt_stash *stash,
 			   u64 size);
+int i915_vm_pin_pt_stash(struct i915_address_space *vm,
+			 struct i915_vm_pt_stash *stash);
 void i915_vm_free_pt_stash(struct i915_address_space *vm,
 			   struct i915_vm_pt_stash *stash);