Merge branch 'drm-next-merged' of git://people.freedesktop.org/~airlied/linux

Pull drm updates from Dave Airlie:
 "Highlights:

  Core:
   - Virtual GEM layer merged, this has been around for a long time, and
     it provides a software backed device that allows userspace to use
     it as a GEM shared memory handler.  This makes it a lot easier to
     do certain things when you have no GPU but still have to deal with
     DRI expectations.
   - atomic helper updates.
   - framebuffer modifier interface added.
   - i2c over auxch displayport fixes.
   - fb width/height confusion fixes.
   - new driver for ps8622/ps8625 bridge chips
   - lots of new panels

  i915:
   - more plane atomic conversion
   - vGPU guest support for XenGT
   - Skylake workarounds and fixes
   - Y-tiling support
   - work on dynamic pagetable allocation
   - EU count report param for gen9+
   - CHV fixes (no longer prelim)
   - remove ilk rc6
   - frontbuffer tracking for fbc
   - Displayport link rate refactoring
   - sprite colorkey refactor

  radeon:
   - Displayport MST support (not enabled by default)
   - non-ATOM native hw auxch support (DCE5+)
   - output csc support
   - new queries for userspace debug support
   - new VCE packet

  nouveau:
   - gk20a iommu support
   - gm107 graphics support
   - more gm20x bringup (waiting on signed nvidia fw).

  amdkfd:
   - multiple kgd instance support
   - use 64-bit time accessors

  msm:
   - stolen memory support
   - DSI and dual-DSI support
   - snapdragon 410 support

  exynos:
   - cleanups for atomic and pageflip

  imx-drm:
   - more media-bus formats
   - TV output prep
   - drm panel support

  tegra:
   - hw vblank counter using host1x syncpoints

  omap:
   - universal plane support
   - prep work for atomic modesetting

  rcar-du:
   - ported to atomic modesetting

  atmel-hlcdc:
   - ported to atomic modesetting
   - added suspend/resume support

  sti:
   - ported to atomic modesetting

  dwhdmi:
   - more compliant audio support
   - update rockchip phy support

  tda998x:
   - DT probing for attached crtcs
   - simplified EDID reading

  rockchip:
   - fixes

  adv7511:
   - fixes"

* 'drm-next-merged' of git://people.freedesktop.org/~airlied/linux: (689 commits)
  media-bus: Fixup RGB444_1X12, RGB565_1X16, and YUV8_1X24 media bus format
  drm/i915: Dont enable CS_PARSER_ERROR interrupts at all
  drm/i915: Move drm_framebuffer_unreference out of struct_mutex for takeover
  drm: fix trivial typo mistake
  drm: Make integer overflow checking cover universal cursor updates (v2)
  drm/nouveau/bios: fix fetching from acpi on certain systems
  drm/nouveau/gr/gm206: initial init+ctx code
  drm/nouveau/ce/gm206: enable support via gm204 code
  drm/nouveau/fifo/gm206: enable support via gm204 code
  drm/nouveau/gr/gm204: initial init+ctx code
  drm/nouveau: support for buffer moves via MaxwellDmaCopyA
  drm/nouveau/ce/gm204: initial support
  drm/nouveau: add support for gm20x fifo channels
  drm/nouveau/fifo/gm204: initial support
  drm/nouveau/gr/gk104-: prevent reading non-existent regs in intr handler
  drm/nouveau/gr/gm107: very slightly demagic part of attrib cb setup
  drm/nouveau/gr/gk104-: correct crop/zrop num_active_fbps setting
  drm/nouveau/gr/gf100-: add symbolic names for classes
  drm/nouveau/gr/gm107: support tpc "strand" ctxsw in gpccs ucode
  drm/nouveau/gr/gf100-: support mmio access with gpc offset from gpccs ucode
  ...

1 files changed, 135 insertions, 59 deletions

diff --git a/drivers/gpu/drm/i915/i915_gem_execbuffer.c b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
index 38a742532c4f..a3190e793ed4 100644
--- a/drivers/gpu/drm/i915/i915_gem_execbuffer.c
+++ b/drivers/gpu/drm/i915/i915_gem_execbuffer.c
@@ -251,7 +251,6 @@ static inline int use_cpu_reloc(struct drm_i915_gem_object *obj)
 {
 	return (HAS_LLC(obj->base.dev) ||
 		obj->base.write_domain == I915_GEM_DOMAIN_CPU ||
-		!obj->map_and_fenceable ||
 		obj->cache_level != I915_CACHE_NONE);
 }
 
@@ -337,6 +336,51 @@ relocate_entry_gtt(struct drm_i915_gem_object *obj,
 	return 0;
 }
 
+static void
+clflush_write32(void *addr, uint32_t value)
+{
+	/* This is not a fast path, so KISS. */
+	drm_clflush_virt_range(addr, sizeof(uint32_t));
+	*(uint32_t *)addr = value;
+	drm_clflush_virt_range(addr, sizeof(uint32_t));
+}
+
+static int
+relocate_entry_clflush(struct drm_i915_gem_object *obj,
+		       struct drm_i915_gem_relocation_entry *reloc,
+		       uint64_t target_offset)
+{
+	struct drm_device *dev = obj->base.dev;
+	uint32_t page_offset = offset_in_page(reloc->offset);
+	uint64_t delta = (int)reloc->delta + target_offset;
+	char *vaddr;
+	int ret;
+
+	ret = i915_gem_object_set_to_gtt_domain(obj, true);
+	if (ret)
+		return ret;
+
+	vaddr = kmap_atomic(i915_gem_object_get_page(obj,
+				reloc->offset >> PAGE_SHIFT));
+	clflush_write32(vaddr + page_offset, lower_32_bits(delta));
+
+	if (INTEL_INFO(dev)->gen >= 8) {
+		page_offset = offset_in_page(page_offset + sizeof(uint32_t));
+
+		if (page_offset == 0) {
+			kunmap_atomic(vaddr);
+			vaddr = kmap_atomic(i915_gem_object_get_page(obj,
+			    (reloc->offset + sizeof(uint32_t)) >> PAGE_SHIFT));
+		}
+
+		clflush_write32(vaddr + page_offset, upper_32_bits(delta));
+	}
+
+	kunmap_atomic(vaddr);
+
+	return 0;
+}
+
 static int
 i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 				   struct eb_vmas *eb,
@@ -426,8 +470,14 @@ i915_gem_execbuffer_relocate_entry(struct drm_i915_gem_object *obj,
 
 	if (use_cpu_reloc(obj))
 		ret = relocate_entry_cpu(obj, reloc, target_offset);
-	else
+	else if (obj->map_and_fenceable)
 		ret = relocate_entry_gtt(obj, reloc, target_offset);
+	else if (cpu_has_clflush)
+		ret = relocate_entry_clflush(obj, reloc, target_offset);
+	else {
+		WARN_ONCE(1, "Impossible case in relocation handling\n");
+		ret = -ENODEV;
+	}
 
 	if (ret)
 		return ret;
@@ -525,6 +575,12 @@ i915_gem_execbuffer_relocate(struct eb_vmas *eb)
 	return ret;
 }
 
+static bool only_mappable_for_reloc(unsigned int flags)
+{
+	return (flags & (EXEC_OBJECT_NEEDS_FENCE | __EXEC_OBJECT_NEEDS_MAP)) ==
+		__EXEC_OBJECT_NEEDS_MAP;
+}
+
 static int
 i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 				struct intel_engine_cs *ring,
@@ -536,14 +592,21 @@ i915_gem_execbuffer_reserve_vma(struct i915_vma *vma,
 	int ret;
 
 	flags = 0;
-	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP)
-		flags |= PIN_GLOBAL | PIN_MAPPABLE;
-	if (entry->flags & EXEC_OBJECT_NEEDS_GTT)
-		flags |= PIN_GLOBAL;
-	if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS)
-		flags |= BATCH_OFFSET_BIAS | PIN_OFFSET_BIAS;
+	if (!drm_mm_node_allocated(&vma->node)) {
+		if (entry->flags & __EXEC_OBJECT_NEEDS_MAP)
+			flags |= PIN_GLOBAL | PIN_MAPPABLE;
+		if (entry->flags & EXEC_OBJECT_NEEDS_GTT)
+			flags |= PIN_GLOBAL;
+		if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS)
+			flags |= BATCH_OFFSET_BIAS | PIN_OFFSET_BIAS;
+	}
 
 	ret = i915_gem_object_pin(obj, vma->vm, entry->alignment, flags);
+	if ((ret == -ENOSPC  || ret == -E2BIG) &&
+	    only_mappable_for_reloc(entry->flags))
+		ret = i915_gem_object_pin(obj, vma->vm,
+					  entry->alignment,
+					  flags & ~(PIN_GLOBAL | PIN_MAPPABLE));
 	if (ret)
 		return ret;
 
@@ -605,13 +668,14 @@ eb_vma_misplaced(struct i915_vma *vma)
 	    vma->node.start & (entry->alignment - 1))
 		return true;
 
-	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP && !obj->map_and_fenceable)
-		return true;
-
 	if (entry->flags & __EXEC_OBJECT_NEEDS_BIAS &&
 	    vma->node.start < BATCH_OFFSET_BIAS)
 		return true;
 
+	/* avoid costly ping-pong once a batch bo ended up non-mappable */
+	if (entry->flags & __EXEC_OBJECT_NEEDS_MAP && !obj->map_and_fenceable)
+		return !only_mappable_for_reloc(entry->flags);
+
 	return false;
 }
 
@@ -971,7 +1035,7 @@ i915_gem_execbuffer_move_to_active(struct list_head *vmas,
 			obj->dirty = 1;
 			i915_gem_request_assign(&obj->last_write_req, req);
 
-			intel_fb_obj_invalidate(obj, ring);
+			intel_fb_obj_invalidate(obj, ring, ORIGIN_CS);
 
 			/* update for the implicit flush after a batch */
 			obj->base.write_domain &= ~I915_GEM_GPU_DOMAINS;
@@ -1076,16 +1140,15 @@ i915_gem_execbuffer_parse(struct intel_engine_cs *ring,
 			  struct drm_i915_gem_object *batch_obj,
 			  u32 batch_start_offset,
 			  u32 batch_len,
-			  bool is_master,
-			  u32 *flags)
+			  bool is_master)
 {
 	struct drm_i915_private *dev_priv = to_i915(batch_obj->base.dev);
 	struct drm_i915_gem_object *shadow_batch_obj;
-	bool need_reloc = false;
+	struct i915_vma *vma;
 	int ret;
 
 	shadow_batch_obj = i915_gem_batch_pool_get(&dev_priv->mm.batch_pool,
-						   batch_obj->base.size);
+						   PAGE_ALIGN(batch_len));
 	if (IS_ERR(shadow_batch_obj))
 		return shadow_batch_obj;
 
@@ -1095,40 +1158,30 @@ i915_gem_execbuffer_parse(struct intel_engine_cs *ring,
 			      batch_start_offset,
 			      batch_len,
 			      is_master);
-	if (ret) {
-		if (ret == -EACCES)
-			return batch_obj;
-	} else {
-		struct i915_vma *vma;
+	if (ret)
+		goto err;
 
-		memset(shadow_exec_entry, 0, sizeof(*shadow_exec_entry));
+	ret = i915_gem_obj_ggtt_pin(shadow_batch_obj, 0, 0);
+	if (ret)
+		goto err;
 
-		vma = i915_gem_obj_to_ggtt(shadow_batch_obj);
-		vma->exec_entry = shadow_exec_entry;
-		vma->exec_entry->flags = __EXEC_OBJECT_PURGEABLE;
-		drm_gem_object_reference(&shadow_batch_obj->base);
-		i915_gem_execbuffer_reserve_vma(vma, ring, &need_reloc);
-		list_add_tail(&vma->exec_list, &eb->vmas);
+	memset(shadow_exec_entry, 0, sizeof(*shadow_exec_entry));
 
-		shadow_batch_obj->base.pending_read_domains =
-			batch_obj->base.pending_read_domains;
+	vma = i915_gem_obj_to_ggtt(shadow_batch_obj);
+	vma->exec_entry = shadow_exec_entry;
+	vma->exec_entry->flags = __EXEC_OBJECT_PURGEABLE | __EXEC_OBJECT_HAS_PIN;
+	drm_gem_object_reference(&shadow_batch_obj->base);
+	list_add_tail(&vma->exec_list, &eb->vmas);
 
-		/*
-		 * Set the DISPATCH_SECURE bit to remove the NON_SECURE
-		 * bit from MI_BATCH_BUFFER_START commands issued in the
-		 * dispatch_execbuffer implementations. We specifically
-		 * don't want that set when the command parser is
-		 * enabled.
-		 *
-		 * FIXME: with aliasing ppgtt, buffers that should only
-		 * be in ggtt still end up in the aliasing ppgtt. remove
-		 * this check when that is fixed.
-		 */
-		if (USES_FULL_PPGTT(dev))
-			*flags |= I915_DISPATCH_SECURE;
-	}
+	shadow_batch_obj->base.pending_read_domains = I915_GEM_DOMAIN_COMMAND;
+
+	return shadow_batch_obj;
 
-	return ret ? ERR_PTR(ret) : shadow_batch_obj;
+err:
+	if (ret == -EACCES) /* unhandled chained batch */
+		return batch_obj;
+	else
+		return ERR_PTR(ret);
 }
 
 int
@@ -1138,7 +1191,7 @@ i915_gem_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
 			       struct drm_i915_gem_execbuffer2 *args,
 			       struct list_head *vmas,
 			       struct drm_i915_gem_object *batch_obj,
-			       u64 exec_start, u32 flags)
+			       u64 exec_start, u32 dispatch_flags)
 {
 	struct drm_clip_rect *cliprects = NULL;
 	struct drm_i915_private *dev_priv = dev->dev_private;
@@ -1198,6 +1251,13 @@ i915_gem_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
 	if (ret)
 		goto error;
 
+	if (ctx->ppgtt)
+		WARN(ctx->ppgtt->pd_dirty_rings & (1<<ring->id),
+			"%s didn't clear reload\n", ring->name);
+	else if (dev_priv->mm.aliasing_ppgtt)
+		WARN(dev_priv->mm.aliasing_ppgtt->pd_dirty_rings &
+			(1<<ring->id), "%s didn't clear reload\n", ring->name);
+
 	instp_mode = args->flags & I915_EXEC_CONSTANTS_MASK;
 	instp_mask = I915_EXEC_CONSTANTS_MASK;
 	switch (instp_mode) {
@@ -1266,19 +1326,19 @@ i915_gem_ringbuffer_submission(struct drm_device *dev, struct drm_file *file,
 
 			ret = ring->dispatch_execbuffer(ring,
 							exec_start, exec_len,
-							flags);
+							dispatch_flags);
 			if (ret)
 				goto error;
 		}
 	} else {
 		ret = ring->dispatch_execbuffer(ring,
 						exec_start, exec_len,
-						flags);
+						dispatch_flags);
 		if (ret)
 			return ret;
 	}
 
-	trace_i915_gem_ring_dispatch(intel_ring_get_request(ring), flags);
+	trace_i915_gem_ring_dispatch(intel_ring_get_request(ring), dispatch_flags);
 
 	i915_gem_execbuffer_move_to_active(vmas, ring);
 	i915_gem_execbuffer_retire_commands(dev, file, ring, batch_obj);
@@ -1353,7 +1413,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	struct i915_address_space *vm;
 	const u32 ctx_id = i915_execbuffer2_get_context_id(*args);
 	u64 exec_start = args->batch_start_offset;
-	u32 flags;
+	u32 dispatch_flags;
 	int ret;
 	bool need_relocs;
 
@@ -1364,15 +1424,15 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	if (ret)
 		return ret;
 
-	flags = 0;
+	dispatch_flags = 0;
 	if (args->flags & I915_EXEC_SECURE) {
 		if (!file->is_master || !capable(CAP_SYS_ADMIN))
 		    return -EPERM;
 
-		flags |= I915_DISPATCH_SECURE;
+		dispatch_flags |= I915_DISPATCH_SECURE;
 	}
 	if (args->flags & I915_EXEC_IS_PINNED)
-		flags |= I915_DISPATCH_PINNED;
+		dispatch_flags |= I915_DISPATCH_PINNED;
 
 	if ((args->flags & I915_EXEC_RING_MASK) > LAST_USER_RING) {
 		DRM_DEBUG("execbuf with unknown ring: %d\n",
@@ -1494,12 +1554,27 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 						      batch_obj,
 						      args->batch_start_offset,
 						      args->batch_len,
-						      file->is_master,
-						      &flags);
+						      file->is_master);
 		if (IS_ERR(batch_obj)) {
 			ret = PTR_ERR(batch_obj);
 			goto err;
 		}
+
+		/*
+		 * Set the DISPATCH_SECURE bit to remove the NON_SECURE
+		 * bit from MI_BATCH_BUFFER_START commands issued in the
+		 * dispatch_execbuffer implementations. We specifically
+		 * don't want that set when the command parser is
+		 * enabled.
+		 *
+		 * FIXME: with aliasing ppgtt, buffers that should only
+		 * be in ggtt still end up in the aliasing ppgtt. remove
+		 * this check when that is fixed.
+		 */
+		if (USES_FULL_PPGTT(dev))
+			dispatch_flags |= I915_DISPATCH_SECURE;
+
+		exec_start = 0;
 	}
 
 	batch_obj->base.pending_read_domains |= I915_GEM_DOMAIN_COMMAND;
@@ -1507,14 +1582,14 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	/* snb/ivb/vlv conflate the "batch in ppgtt" bit with the "non-secure
 	 * batch" bit. Hence we need to pin secure batches into the global gtt.
 	 * hsw should have this fixed, but bdw mucks it up again. */
-	if (flags & I915_DISPATCH_SECURE) {
+	if (dispatch_flags & I915_DISPATCH_SECURE) {
 		/*
 		 * So on first glance it looks freaky that we pin the batch here
 		 * outside of the reservation loop. But:
 		 * - The batch is already pinned into the relevant ppgtt, so we
 		 *   already have the backing storage fully allocated.
 		 * - No other BO uses the global gtt (well contexts, but meh),
-		 *   so we don't really have issues with mutliple objects not
+		 *   so we don't really have issues with multiple objects not
 		 *   fitting due to fragmentation.
 		 * So this is actually safe.
 		 */
@@ -1527,7 +1602,8 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 		exec_start += i915_gem_obj_offset(batch_obj, vm);
 
 	ret = dev_priv->gt.do_execbuf(dev, file, ring, ctx, args,
-				      &eb->vmas, batch_obj, exec_start, flags);
+				      &eb->vmas, batch_obj, exec_start,
+				      dispatch_flags);
 
 	/*
 	 * FIXME: We crucially rely upon the active tracking for the (ppgtt)
@@ -1535,7 +1611,7 @@ i915_gem_do_execbuffer(struct drm_device *dev, void *data,
 	 * needs to be adjusted to also track the ggtt batch vma properly as
 	 * active.
 	 */
-	if (flags & I915_DISPATCH_SECURE)
+	if (dispatch_flags & I915_DISPATCH_SECURE)
 		i915_gem_object_ggtt_unpin(batch_obj);
 err:
 	/* the request owns the ref now */