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, 57 insertions, 243 deletions

diff --git a/drivers/gpu/drm/i915/gt/intel_gtt.c b/drivers/gpu/drm/i915/gt/intel_gtt.c
index 2a72cce63fd9..3f1114b58b01 100644
--- a/drivers/gpu/drm/i915/gt/intel_gtt.c
+++ b/drivers/gpu/drm/i915/gt/intel_gtt.c
@@ -11,160 +11,24 @@
 #include "intel_gt.h"
 #include "intel_gtt.h"
 
-void stash_init(struct pagestash *stash)
+struct drm_i915_gem_object *alloc_pt_dma(struct i915_address_space *vm, int sz)
 {
-	pagevec_init(&stash->pvec);
-	spin_lock_init(&stash->lock);
-}
-
-static struct page *stash_pop_page(struct pagestash *stash)
-{
-	struct page *page = NULL;
-
-	spin_lock(&stash->lock);
-	if (likely(stash->pvec.nr))
-		page = stash->pvec.pages[--stash->pvec.nr];
-	spin_unlock(&stash->lock);
-
-	return page;
-}
-
-static void stash_push_pagevec(struct pagestash *stash, struct pagevec *pvec)
-{
-	unsigned int nr;
-
-	spin_lock_nested(&stash->lock, SINGLE_DEPTH_NESTING);
-
-	nr = min_t(typeof(nr), pvec->nr, pagevec_space(&stash->pvec));
-	memcpy(stash->pvec.pages + stash->pvec.nr,
-	       pvec->pages + pvec->nr - nr,
-	       sizeof(pvec->pages[0]) * nr);
-	stash->pvec.nr += nr;
-
-	spin_unlock(&stash->lock);
-
-	pvec->nr -= nr;
-}
-
-static struct page *vm_alloc_page(struct i915_address_space *vm, gfp_t gfp)
-{
-	struct pagevec stack;
-	struct page *page;
-
 	if (I915_SELFTEST_ONLY(should_fail(&vm->fault_attr, 1)))
 		i915_gem_shrink_all(vm->i915);
 
-	page = stash_pop_page(&vm->free_pages);
-	if (page)
-		return page;
-
-	if (!vm->pt_kmap_wc)
-		return alloc_page(gfp);
-
-	/* Look in our global stash of WC pages... */
-	page = stash_pop_page(&vm->i915->mm.wc_stash);
-	if (page)
-		return page;
-
-	/*
-	 * Otherwise batch allocate pages to amortize cost of set_pages_wc.
-	 *
-	 * We have to be careful as page allocation may trigger the shrinker
-	 * (via direct reclaim) which will fill up the WC stash underneath us.
-	 * So we add our WB pages into a temporary pvec on the stack and merge
-	 * them into the WC stash after all the allocations are complete.
-	 */
-	pagevec_init(&stack);
-	do {
-		struct page *page;
-
-		page = alloc_page(gfp);
-		if (unlikely(!page))
-			break;
-
-		stack.pages[stack.nr++] = page;
-	} while (pagevec_space(&stack));
-
-	if (stack.nr && !set_pages_array_wc(stack.pages, stack.nr)) {
-		page = stack.pages[--stack.nr];
-
-		/* Merge spare WC pages to the global stash */
-		if (stack.nr)
-			stash_push_pagevec(&vm->i915->mm.wc_stash, &stack);
-
-		/* Push any surplus WC pages onto the local VM stash */
-		if (stack.nr)
-			stash_push_pagevec(&vm->free_pages, &stack);
-	}
-
-	/* Return unwanted leftovers */
-	if (unlikely(stack.nr)) {
-		WARN_ON_ONCE(set_pages_array_wb(stack.pages, stack.nr));
-		__pagevec_release(&stack);
-	}
-
-	return page;
+	return i915_gem_object_create_internal(vm->i915, sz);
 }
 
-static void vm_free_pages_release(struct i915_address_space *vm,
-				  bool immediate)
+int pin_pt_dma(struct i915_address_space *vm, struct drm_i915_gem_object *obj)
 {
-	struct pagevec *pvec = &vm->free_pages.pvec;
-	struct pagevec stack;
-
-	lockdep_assert_held(&vm->free_pages.lock);
-	GEM_BUG_ON(!pagevec_count(pvec));
-
-	if (vm->pt_kmap_wc) {
-		/*
-		 * When we use WC, first fill up the global stash and then
-		 * only if full immediately free the overflow.
-		 */
-		stash_push_pagevec(&vm->i915->mm.wc_stash, pvec);
-
-		/*
-		 * As we have made some room in the VM's free_pages,
-		 * we can wait for it to fill again. Unless we are
-		 * inside i915_address_space_fini() and must
-		 * immediately release the pages!
-		 */
-		if (pvec->nr <= (immediate ? 0 : PAGEVEC_SIZE - 1))
-			return;
+	int err;
 
-		/*
-		 * We have to drop the lock to allow ourselves to sleep,
-		 * so take a copy of the pvec and clear the stash for
-		 * others to use it as we sleep.
-		 */
-		stack = *pvec;
-		pagevec_reinit(pvec);
-		spin_unlock(&vm->free_pages.lock);
-
-		pvec = &stack;
-		set_pages_array_wb(pvec->pages, pvec->nr);
-
-		spin_lock(&vm->free_pages.lock);
-	}
+	err = i915_gem_object_pin_pages(obj);
+	if (err)
+		return err;
 
-	__pagevec_release(pvec);
-}
-
-static void vm_free_page(struct i915_address_space *vm, struct page *page)
-{
-	/*
-	 * On !llc, we need to change the pages back to WB. We only do so
-	 * in bulk, so we rarely need to change the page attributes here,
-	 * but doing so requires a stop_machine() from deep inside arch/x86/mm.
-	 * To make detection of the possible sleep more likely, use an
-	 * unconditional might_sleep() for everybody.
-	 */
-	might_sleep();
-	spin_lock(&vm->free_pages.lock);
-	while (!pagevec_space(&vm->free_pages.pvec))
-		vm_free_pages_release(vm, false);
-	GEM_BUG_ON(pagevec_count(&vm->free_pages.pvec) >= PAGEVEC_SIZE);
-	pagevec_add(&vm->free_pages.pvec, page);
-	spin_unlock(&vm->free_pages.lock);
+	i915_gem_object_make_unshrinkable(obj);
+	return 0;
 }
 
 void __i915_vm_close(struct i915_address_space *vm)
@@ -194,14 +58,7 @@ void __i915_vm_close(struct i915_address_space *vm)
 
 void i915_address_space_fini(struct i915_address_space *vm)
 {
-	spin_lock(&vm->free_pages.lock);
-	if (pagevec_count(&vm->free_pages.pvec))
-		vm_free_pages_release(vm, true);
-	GEM_BUG_ON(pagevec_count(&vm->free_pages.pvec));
-	spin_unlock(&vm->free_pages.lock);
-
 	drm_mm_takedown(&vm->mm);
-
 	mutex_destroy(&vm->mutex);
 }
 
@@ -246,8 +103,6 @@ void i915_address_space_init(struct i915_address_space *vm, int subclass)
 	drm_mm_init(&vm->mm, 0, vm->total);
 	vm->mm.head_node.color = I915_COLOR_UNEVICTABLE;
 
-	stash_init(&vm->free_pages);
-
 	INIT_LIST_HEAD(&vm->bound_list);
 }
 
@@ -264,64 +119,50 @@ void clear_pages(struct i915_vma *vma)
 	memset(&vma->page_sizes, 0, sizeof(vma->page_sizes));
 }
 
-static int __setup_page_dma(struct i915_address_space *vm,
-			    struct i915_page_dma *p,
-			    gfp_t gfp)
-{
-	p->page = vm_alloc_page(vm, gfp | I915_GFP_ALLOW_FAIL);
-	if (unlikely(!p->page))
-		return -ENOMEM;
-
-	p->daddr = dma_map_page_attrs(vm->dma,
-				      p->page, 0, PAGE_SIZE,
-				      PCI_DMA_BIDIRECTIONAL,
-				      DMA_ATTR_SKIP_CPU_SYNC |
-				      DMA_ATTR_NO_WARN);
-	if (unlikely(dma_mapping_error(vm->dma, p->daddr))) {
-		vm_free_page(vm, p->page);
-		return -ENOMEM;
-	}
-
-	return 0;
-}
-
-int setup_page_dma(struct i915_address_space *vm, struct i915_page_dma *p)
+dma_addr_t __px_dma(struct drm_i915_gem_object *p)
 {
-	return __setup_page_dma(vm, p, __GFP_HIGHMEM);
+	GEM_BUG_ON(!i915_gem_object_has_pages(p));
+	return sg_dma_address(p->mm.pages->sgl);
 }
 
-void cleanup_page_dma(struct i915_address_space *vm, struct i915_page_dma *p)
+struct page *__px_page(struct drm_i915_gem_object *p)
 {
-	dma_unmap_page(vm->dma, p->daddr, PAGE_SIZE, PCI_DMA_BIDIRECTIONAL);
-	vm_free_page(vm, p->page);
+	GEM_BUG_ON(!i915_gem_object_has_pages(p));
+	return sg_page(p->mm.pages->sgl);
 }
 
 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)
 {
-	kunmap_atomic(memset64(kmap_atomic(p->page), val, count));
+	struct page *page = __px_page(p);
+	void *vaddr;
+
+	vaddr = kmap(page);
+	memset64(vaddr, val, count);
+	clflush_cache_range(vaddr, PAGE_SIZE);
+	kunmap(page);
 }
 
-static void poison_scratch_page(struct page *page, unsigned long size)
+static void poison_scratch_page(struct drm_i915_gem_object *scratch)
 {
-	if (!IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
-		return;
+	struct sgt_iter sgt;
+	struct page *page;
+	u8 val;
 
-	GEM_BUG_ON(!IS_ALIGNED(size, PAGE_SIZE));
+	val = 0;
+	if (IS_ENABLED(CONFIG_DRM_I915_DEBUG_GEM))
+		val = POISON_FREE;
 
-	do {
+	for_each_sgt_page(page, sgt, scratch->mm.pages) {
 		void *vaddr;
 
 		vaddr = kmap(page);
-		memset(vaddr, POISON_FREE, PAGE_SIZE);
+		memset(vaddr, val, PAGE_SIZE);
 		kunmap(page);
-
-		page = pfn_to_page(page_to_pfn(page) + 1);
-		size -= PAGE_SIZE;
-	} while (size);
+	}
 }
 
-int setup_scratch_page(struct i915_address_space *vm, gfp_t gfp)
+int setup_scratch_page(struct i915_address_space *vm)
 {
 	unsigned long size;
 
@@ -338,21 +179,27 @@ int setup_scratch_page(struct i915_address_space *vm, gfp_t gfp)
 	 */
 	size = I915_GTT_PAGE_SIZE_4K;
 	if (i915_vm_is_4lvl(vm) &&
-	    HAS_PAGE_SIZES(vm->i915, I915_GTT_PAGE_SIZE_64K)) {
+	    HAS_PAGE_SIZES(vm->i915, I915_GTT_PAGE_SIZE_64K))
 		size = I915_GTT_PAGE_SIZE_64K;
-		gfp |= __GFP_NOWARN;
-	}
-	gfp |= __GFP_ZERO | __GFP_RETRY_MAYFAIL;
 
 	do {
-		unsigned int order = get_order(size);
-		struct page *page;
-		dma_addr_t addr;
+		struct drm_i915_gem_object *obj;
 
-		page = alloc_pages(gfp, order);
-		if (unlikely(!page))
+		obj = vm->alloc_pt_dma(vm, size);
+		if (IS_ERR(obj))
 			goto skip;
 
+		if (pin_pt_dma(vm, obj))
+			goto skip_obj;
+
+		/* We need a single contiguous page for our scratch */
+		if (obj->mm.page_sizes.sg < size)
+			goto skip_obj;
+
+		/* And it needs to be correspondingly aligned */
+		if (__px_dma(obj) & (size - 1))
+			goto skip_obj;
+
 		/*
 		 * Use a non-zero scratch page for debugging.
 		 *
@@ -362,61 +209,28 @@ int setup_scratch_page(struct i915_address_space *vm, gfp_t gfp)
 		 * should it ever be accidentally used, the effect should be
 		 * fairly benign.
 		 */
-		poison_scratch_page(page, size);
-
-		addr = dma_map_page_attrs(vm->dma,
-					  page, 0, size,
-					  PCI_DMA_BIDIRECTIONAL,
-					  DMA_ATTR_SKIP_CPU_SYNC |
-					  DMA_ATTR_NO_WARN);
-		if (unlikely(dma_mapping_error(vm->dma, addr)))
-			goto free_page;
-
-		if (unlikely(!IS_ALIGNED(addr, size)))
-			goto unmap_page;
-
-		vm->scratch[0].base.page = page;
-		vm->scratch[0].base.daddr = addr;
-		vm->scratch_order = order;
+		poison_scratch_page(obj);
+
+		vm->scratch[0] = obj;
+		vm->scratch_order = get_order(size);
 		return 0;
 
-unmap_page:
-		dma_unmap_page(vm->dma, addr, size, PCI_DMA_BIDIRECTIONAL);
-free_page:
-		__free_pages(page, order);
+skip_obj:
+		i915_gem_object_put(obj);
 skip:
 		if (size == I915_GTT_PAGE_SIZE_4K)
 			return -ENOMEM;
 
 		size = I915_GTT_PAGE_SIZE_4K;
-		gfp &= ~__GFP_NOWARN;
 	} while (1);
 }
 
-void cleanup_scratch_page(struct i915_address_space *vm)
-{
-	struct i915_page_dma *p = px_base(&vm->scratch[0]);
-	unsigned int order = vm->scratch_order;
-
-	dma_unmap_page(vm->dma, p->daddr, BIT(order) << PAGE_SHIFT,
-		       PCI_DMA_BIDIRECTIONAL);
-	__free_pages(p->page, order);
-}
-
 void free_scratch(struct i915_address_space *vm)
 {
 	int i;
 
-	if (!px_dma(&vm->scratch[0])) /* set to 0 on clones */
-		return;
-
-	for (i = 1; i <= vm->top; i++) {
-		if (!px_dma(&vm->scratch[i]))
-			break;
-		cleanup_page_dma(vm, px_base(&vm->scratch[i]));
-	}
-
-	cleanup_scratch_page(vm);
+	for (i = 0; i <= vm->top; i++)
+		i915_gem_object_put(vm->scratch[i]);
 }
 
 void gtt_write_workarounds(struct intel_gt *gt)