drm/vmwgfx: Initial DX support

Initial DX support.
Co-authored with Sinclair Yeh, Charmaine Lee and Jakob Bornecrantz.

Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com>
Signed-off-by: Sinclair Yeh <syeh@vmware.com>
Signed-off-by: Charmaine Lee <charmainel@vmware.com>

drivers/gpu/drm/vmwgfx/Makefile

@@ -8,5 +8,6 @@ vmwgfx-y := vmwgfx_execbuf.o vmwgfx_gmr.o vmwgfx_kms.o vmwgfx_drv.o \
 	    vmwgfx_fence.o vmwgfx_dmabuf.o vmwgfx_scrn.o vmwgfx_context.o \
 	    vmwgfx_surface.o vmwgfx_prime.o vmwgfx_mob.o vmwgfx_shader.o \
 	    vmwgfx_cmdbuf_res.o vmwgfx_cmdbuf.o vmwgfx_stdu.o \
+	    vmwgfx_cotable.o vmwgfx_so.o vmwgfx_binding.o
 
 obj-$(CONFIG_DRM_VMWGFX) := vmwgfx.o

drivers/gpu/drm/vmwgfx/vmwgfx_binding.h

@@ -0,0 +1,209 @@
+/**************************************************************************
+ *
+ * Copyright © 2015 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+#ifndef _VMWGFX_BINDING_H_
+#define _VMWGFX_BINDING_H_
+
+#include "device_include/svga3d_reg.h"
+#include <linux/list.h>
+
+#define VMW_MAX_VIEW_BINDINGS 128
+
+struct vmw_private;
+struct vmw_ctx_binding_state;
+
+/*
+ * enum vmw_ctx_binding_type - abstract resource to context binding types
+ */
+enum vmw_ctx_binding_type {
+	vmw_ctx_binding_shader,
+	vmw_ctx_binding_rt,
+	vmw_ctx_binding_tex,
+	vmw_ctx_binding_cb,
+	vmw_ctx_binding_dx_shader,
+	vmw_ctx_binding_dx_rt,
+	vmw_ctx_binding_sr,
+	vmw_ctx_binding_ds,
+	vmw_ctx_binding_so,
+	vmw_ctx_binding_vb,
+	vmw_ctx_binding_ib,
+	vmw_ctx_binding_max
+};
+
+/**
+ * struct vmw_ctx_bindinfo - single binding metadata
+ *
+ * @ctx_list: List head for the context's list of bindings.
+ * @res_list: List head for a resource's list of bindings.
+ * @ctx: Non-refcounted pointer to the context that owns the binding. NULL
+ * indicates no binding present.
+ * @res: Non-refcounted pointer to the resource the binding points to. This
+ * is typically a surface or a view.
+ * @bt: Binding type.
+ * @scrubbed: Whether the binding has been scrubbed from the context.
+ */
+struct vmw_ctx_bindinfo {
+	struct list_head ctx_list;
+	struct list_head res_list;
+	struct vmw_resource *ctx;
+	struct vmw_resource *res;
+	enum vmw_ctx_binding_type bt;
+	bool scrubbed;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_tex - texture stage binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @texture_stage: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_tex {
+	struct vmw_ctx_bindinfo bi;
+	uint32 texture_stage;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_shader - Shader binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @shader_slot: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_shader {
+	struct vmw_ctx_bindinfo bi;
+	SVGA3dShaderType shader_slot;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_cb - Constant buffer binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @shader_slot: Device data used to reconstruct binding command.
+ * @offset: Device data used to reconstruct binding command.
+ * @size: Device data used to reconstruct binding command.
+ * @slot: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_cb {
+	struct vmw_ctx_bindinfo bi;
+	SVGA3dShaderType shader_slot;
+	uint32 offset;
+	uint32 size;
+	uint32 slot;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_view - View binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @shader_slot: Device data used to reconstruct binding command.
+ * @slot: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_view {
+	struct vmw_ctx_bindinfo bi;
+	SVGA3dShaderType shader_slot;
+	uint32 slot;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_so - StreamOutput binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @offset: Device data used to reconstruct binding command.
+ * @size: Device data used to reconstruct binding command.
+ * @slot: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_so {
+	struct vmw_ctx_bindinfo bi;
+	uint32 offset;
+	uint32 size;
+	uint32 slot;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_vb - Vertex buffer binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @offset: Device data used to reconstruct binding command.
+ * @stride: Device data used to reconstruct binding command.
+ * @slot: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_vb {
+	struct vmw_ctx_bindinfo bi;
+	uint32 offset;
+	uint32 stride;
+	uint32 slot;
+};
+
+/**
+ * struct vmw_ctx_bindinfo_ib - StreamOutput binding metadata
+ *
+ * @bi: struct vmw_ctx_bindinfo we derive from.
+ * @offset: Device data used to reconstruct binding command.
+ * @format: Device data used to reconstruct binding command.
+ */
+struct vmw_ctx_bindinfo_ib {
+	struct vmw_ctx_bindinfo bi;
+	uint32 offset;
+	uint32 format;
+};
+
+/**
+ * struct vmw_dx_shader_bindings - per shader type context binding state
+ *
+ * @shader: The shader binding for this shader type
+ * @const_buffer: Const buffer bindings for this shader type.
+ * @shader_res: Shader resource view bindings for this shader type.
+ * @dirty_sr: Bitmap tracking individual shader resource bindings changes
+ * that have not yet been emitted to the device.
+ * @dirty: Bitmap tracking per-binding type binding changes that have not
+ * yet been emitted to the device.
+ */
+struct vmw_dx_shader_bindings {
+	struct vmw_ctx_bindinfo_shader shader;
+	struct vmw_ctx_bindinfo_cb const_buffers[SVGA3D_DX_MAX_CONSTBUFFERS];
+	struct vmw_ctx_bindinfo_view shader_res[SVGA3D_DX_MAX_SRVIEWS];
+	DECLARE_BITMAP(dirty_sr, SVGA3D_DX_MAX_SRVIEWS);
+	unsigned long dirty;
+};
+
+extern void vmw_binding_add(struct vmw_ctx_binding_state *cbs,
+			    const struct vmw_ctx_bindinfo *ci,
+			    u32 shader_slot, u32 slot);
+extern void
+vmw_binding_state_commit(struct vmw_ctx_binding_state *to,
+			 struct vmw_ctx_binding_state *from);
+extern void vmw_binding_res_list_kill(struct list_head *head);
+extern void vmw_binding_res_list_scrub(struct list_head *head);
+extern int vmw_binding_rebind_all(struct vmw_ctx_binding_state *cbs);
+extern void vmw_binding_state_kill(struct vmw_ctx_binding_state *cbs);
+extern void vmw_binding_state_scrub(struct vmw_ctx_binding_state *cbs);
+extern struct vmw_ctx_binding_state *
+vmw_binding_state_alloc(struct vmw_private *dev_priv);
+extern void vmw_binding_state_free(struct vmw_ctx_binding_state *cbs);
+extern struct list_head *
+vmw_binding_state_list(struct vmw_ctx_binding_state *cbs);
+extern void vmw_binding_state_reset(struct vmw_ctx_binding_state *cbs);
+
+#endif

drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf.c

@@ -916,8 +916,7 @@ static void *vmw_cmdbuf_reserve_cur(struct vmw_cmdbuf_man *man,
 
 	cur = man->cur;
 	if (cur && (size + man->cur_pos > cur->size ||
-	    (ctx_id != SVGA3D_INVALID_ID &&
+		    ((cur->cb_header->flags & SVGA_CB_FLAG_DX_CONTEXT) &&
-	     (cur->cb_header->flags & SVGA_CB_FLAG_DX_CONTEXT) &&
 		     ctx_id != cur->cb_header->dxContext)))
 		__vmw_cmdbuf_cur_flush(man);
 

drivers/gpu/drm/vmwgfx/vmwgfx_cmdbuf_res.c

@@ -26,15 +26,10 @@
  **************************************************************************/
 
 #include "vmwgfx_drv.h"
+#include "vmwgfx_resource_priv.h"
 
 #define VMW_CMDBUF_RES_MAN_HT_ORDER 12
 
-enum vmw_cmdbuf_res_state {
-	VMW_CMDBUF_RES_COMMITED,
-	VMW_CMDBUF_RES_ADD,
-	VMW_CMDBUF_RES_DEL
-};
-
 /**
  * struct vmw_cmdbuf_res - Command buffer managed resource entry.
  *
@@ -132,9 +127,12 @@ void vmw_cmdbuf_res_commit(struct list_head *list)
 
 	list_for_each_entry_safe(entry, next, list, head) {
 		list_del(&entry->head);
+		if (entry->res->func->commit_notify)
+			entry->res->func->commit_notify(entry->res,
+							entry->state);
 		switch (entry->state) {
 		case VMW_CMDBUF_RES_ADD:
-			entry->state = VMW_CMDBUF_RES_COMMITED;
+			entry->state = VMW_CMDBUF_RES_COMMITTED;
 			list_add_tail(&entry->head, &entry->man->list);
 			break;
 		case VMW_CMDBUF_RES_DEL:
@@ -175,7 +173,7 @@ void vmw_cmdbuf_res_revert(struct list_head *list)
 						 &entry->hash);
 			list_del(&entry->head);
 			list_add_tail(&entry->head, &entry->man->list);
-			entry->state = VMW_CMDBUF_RES_COMMITED;
+			entry->state = VMW_CMDBUF_RES_COMMITTED;
 			break;
 		default:
 			BUG();
@@ -231,6 +229,9 @@ int vmw_cmdbuf_res_add(struct vmw_cmdbuf_res_manager *man,
  * @res_type: The resource type.
  * @user_key: The user-space id of the resource.
  * @list: The staging list.
+ * @res_p: If the resource is in an already committed state, points to the
+ * struct vmw_resource on successful return. The pointer will be
+ * non ref-counted.
  *
  * This function looks up the struct vmw_cmdbuf_res entry from the manager
  * hash table and, if it exists, removes it. Depending on its current staging
@@ -240,7 +241,8 @@ int vmw_cmdbuf_res_add(struct vmw_cmdbuf_res_manager *man,
 int vmw_cmdbuf_res_remove(struct vmw_cmdbuf_res_manager *man,
 			  enum vmw_cmdbuf_res_type res_type,
 			  u32 user_key,
-			  struct list_head *list)
+			  struct list_head *list,
+			  struct vmw_resource **res_p)
 {
 	struct vmw_cmdbuf_res *entry;
 	struct drm_hash_item *hash;
@@ -256,12 +258,14 @@ int vmw_cmdbuf_res_remove(struct vmw_cmdbuf_res_manager *man,
 	switch (entry->state) {
 	case VMW_CMDBUF_RES_ADD:
 		vmw_cmdbuf_res_free(man, entry);
+		*res_p = NULL;
 		break;
-	case VMW_CMDBUF_RES_COMMITED:
+	case VMW_CMDBUF_RES_COMMITTED:
 		(void) drm_ht_remove_item(&man->resources, &entry->hash);
 		list_del(&entry->head);
 		entry->state = VMW_CMDBUF_RES_DEL;
 		list_add_tail(&entry->head, list);
+		*res_p = entry->res;
 		break;
 	default:
 		BUG();

drivers/gpu/drm/vmwgfx/vmwgfx_context.c

@@ -27,19 +27,18 @@
 
 #include "vmwgfx_drv.h"
 #include "vmwgfx_resource_priv.h"
+#include "vmwgfx_binding.h"
 #include "ttm/ttm_placement.h"
 
 struct vmw_user_context {
 	struct ttm_base_object base;
 	struct vmw_resource res;
-	struct vmw_ctx_binding_state cbs;
+	struct vmw_ctx_binding_state *cbs;
 	struct vmw_cmdbuf_res_manager *man;
+	struct vmw_resource *cotables[SVGA_COTABLE_DX10_MAX];
+	spinlock_t cotable_lock;
 };
 
-
-
-typedef int (*vmw_scrub_func)(struct vmw_ctx_bindinfo *, bool);
-
 static void vmw_user_context_free(struct vmw_resource *res);
 static struct vmw_resource *
 vmw_user_context_base_to_res(struct ttm_base_object *base);
@@ -51,12 +50,14 @@ static int vmw_gb_context_unbind(struct vmw_resource *res,
 				 bool readback,
 				 struct ttm_validate_buffer *val_buf);
 static int vmw_gb_context_destroy(struct vmw_resource *res);
-static int vmw_context_scrub_shader(struct vmw_ctx_bindinfo *bi, bool rebind);
+static int vmw_dx_context_create(struct vmw_resource *res);
-static int vmw_context_scrub_render_target(struct vmw_ctx_bindinfo *bi,
+static int vmw_dx_context_bind(struct vmw_resource *res,
-					   bool rebind);
+			       struct ttm_validate_buffer *val_buf);
-static int vmw_context_scrub_texture(struct vmw_ctx_bindinfo *bi, bool rebind);
+static int vmw_dx_context_unbind(struct vmw_resource *res,
-static void vmw_context_binding_state_scrub(struct vmw_ctx_binding_state *cbs);
+				 bool readback,
-static void vmw_context_binding_state_kill(struct vmw_ctx_binding_state *cbs);
+				 struct ttm_validate_buffer *val_buf);
+static int vmw_dx_context_destroy(struct vmw_resource *res);
+
 static uint64_t vmw_user_context_size;
 
 static const struct vmw_user_resource_conv user_context_conv = {
@@ -93,15 +94,36 @@ static const struct vmw_res_func vmw_gb_context_func = {
 	.unbind = vmw_gb_context_unbind
 };
 
-static const vmw_scrub_func vmw_scrub_funcs[vmw_ctx_binding_max] = {
+static const struct vmw_res_func vmw_dx_context_func = {
-	[vmw_ctx_binding_shader] = vmw_context_scrub_shader,
+	.res_type = vmw_res_dx_context,
-	[vmw_ctx_binding_rt] = vmw_context_scrub_render_target,
+	.needs_backup = true,
-	[vmw_ctx_binding_tex] = vmw_context_scrub_texture };
+	.may_evict = true,
+	.type_name = "dx contexts",
+	.backup_placement = &vmw_mob_placement,
+	.create = vmw_dx_context_create,
+	.destroy = vmw_dx_context_destroy,
+	.bind = vmw_dx_context_bind,
+	.unbind = vmw_dx_context_unbind
+};
 
 /**
  * Context management:
  */
 
+static void vmw_context_cotables_unref(struct vmw_user_context *uctx)
+{
+	struct vmw_resource *res;
+	int i;
+
+	for (i = 0; i < SVGA_COTABLE_DX10_MAX; ++i) {
+		spin_lock(&uctx->cotable_lock);
+		res = uctx->cotables[i];
+		uctx->cotables[i] = NULL;
+		spin_unlock(&uctx->cotable_lock);
+		vmw_resource_unreference(&res);
+	}
+}
+
 static void vmw_hw_context_destroy(struct vmw_resource *res)
 {
 	struct vmw_user_context *uctx =
@@ -113,17 +135,19 @@ static void vmw_hw_context_destroy(struct vmw_resource *res)
 	} *cmd;
 
 
-	if (res->func->destroy == vmw_gb_context_destroy) {
+	if (res->func->destroy == vmw_gb_context_destroy ||
+	    res->func->destroy == vmw_dx_context_destroy) {
 		mutex_lock(&dev_priv->cmdbuf_mutex);
 		vmw_cmdbuf_res_man_destroy(uctx->man);
 		mutex_lock(&dev_priv->binding_mutex);
-		(void) vmw_context_binding_state_kill(&uctx->cbs);
+		vmw_binding_state_kill(uctx->cbs);
-		(void) vmw_gb_context_destroy(res);
+		(void) res->func->destroy(res);
 		mutex_unlock(&dev_priv->binding_mutex);
 		if (dev_priv->pinned_bo != NULL &&
 		    !dev_priv->query_cid_valid)
 			__vmw_execbuf_release_pinned_bo(dev_priv, NULL);
 		mutex_unlock(&dev_priv->cmdbuf_mutex);
+		vmw_context_cotables_unref(uctx);
 		return;
 	}
 
@@ -144,16 +168,20 @@ static void vmw_hw_context_destroy(struct vmw_resource *res)
 }
 
 static int vmw_gb_context_init(struct vmw_private *dev_priv,
+			       bool dx,
 			       struct vmw_resource *res,
 			       void (*res_free)(struct vmw_resource *res))
 {
-	int ret;
+	int ret, i;
 	struct vmw_user_context *uctx =
 		container_of(res, struct vmw_user_context, res);
 
+	res->backup_size = (dx ? sizeof(SVGADXContextMobFormat) :
+			    SVGA3D_CONTEXT_DATA_SIZE);
 	ret = vmw_resource_init(dev_priv, res, true,
-				res_free, &vmw_gb_context_func);
+				res_free,
-	res->backup_size = SVGA3D_CONTEXT_DATA_SIZE;
+				dx ? &vmw_dx_context_func :
+				&vmw_gb_context_func);
 	if (unlikely(ret != 0))
 		goto out_err;
 
@@ -166,12 +194,32 @@ static int vmw_gb_context_init(struct vmw_private *dev_priv,
 		}
 	}
 
-	memset(&uctx->cbs, 0, sizeof(uctx->cbs));
+	uctx->cbs = vmw_binding_state_alloc(dev_priv);
-	INIT_LIST_HEAD(&uctx->cbs.list);
+	if (IS_ERR(uctx->cbs)) {
+		ret = PTR_ERR(uctx->cbs);
+		goto out_err;
+	}
+
+	spin_lock_init(&uctx->cotable_lock);
+
+	if (dx) {
+		for (i = 0; i < SVGA_COTABLE_DX10_MAX; ++i) {
+			uctx->cotables[i] = vmw_cotable_alloc(dev_priv,
+							      &uctx->res, i);
+			if (unlikely(uctx->cotables[i] == NULL)) {
+				ret = -ENOMEM;
+				goto out_cotables;
+			}
+		}
+	}
+
+
 
 	vmw_resource_activate(res, vmw_hw_context_destroy);
 	return 0;
 
+out_cotables:
+	vmw_context_cotables_unref(uctx);
 out_err:
 	if (res_free)
 		res_free(res);
@@ -182,7 +230,8 @@ static int vmw_gb_context_init(struct vmw_private *dev_priv,
 
 static int vmw_context_init(struct vmw_private *dev_priv,
 			    struct vmw_resource *res,
-			    void (*res_free) (struct vmw_resource *res))
+			    void (*res_free)(struct vmw_resource *res),
+			    bool dx)
 {
 	int ret;
 
@@ -192,7 +241,7 @@ static int vmw_context_init(struct vmw_private *dev_priv,
 	} *cmd;
 
 	if (dev_priv->has_mob)
-		return vmw_gb_context_init(dev_priv, res, res_free);
+		return vmw_gb_context_init(dev_priv, dx, res, res_free);
 
 	ret = vmw_resource_init(dev_priv, res, false,
 				res_free, &vmw_legacy_context_func);
@@ -232,19 +281,10 @@ static int vmw_context_init(struct vmw_private *dev_priv,
 	return ret;
 }
 
-struct vmw_resource *vmw_context_alloc(struct vmw_private *dev_priv)
-{
-	struct vmw_resource *res = kmalloc(sizeof(*res), GFP_KERNEL);
-	int ret;
-
-	if (unlikely(res == NULL))
-		return NULL;
-
-	ret = vmw_context_init(dev_priv, res, NULL);
-
-	return (ret == 0) ? res : NULL;
-}
 
+/*
+ * GB context.
+ */
 
 static int vmw_gb_context_create(struct vmw_resource *res)
 {
@@ -309,7 +349,6 @@ static int vmw_gb_context_bind(struct vmw_resource *res,
 			  "binding.\n");
 		return -ENOMEM;
 	}
-
 	cmd->header.id = SVGA_3D_CMD_BIND_GB_CONTEXT;
 	cmd->header.size = sizeof(cmd->body);
 	cmd->body.cid = res->id;
@@ -346,7 +385,7 @@ static int vmw_gb_context_unbind(struct vmw_resource *res,
 	BUG_ON(bo->mem.mem_type != VMW_PL_MOB);
 
 	mutex_lock(&dev_priv->binding_mutex);
-	vmw_context_binding_state_scrub(&uctx->cbs);
+	vmw_binding_state_scrub(uctx->cbs);
 
 	submit_size = sizeof(*cmd2) + (readback ? sizeof(*cmd1) : 0);
 
@@ -419,6 +458,221 @@ static int vmw_gb_context_destroy(struct vmw_resource *res)
 	return 0;
 }
 
+/*
+ * DX context.
+ */
+
+static int vmw_dx_context_create(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	int ret;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXDefineContext body;
+	} *cmd;
+
+	if (likely(res->id != -1))
+		return 0;
+
+	ret = vmw_resource_alloc_id(res);
+	if (unlikely(ret != 0)) {
+		DRM_ERROR("Failed to allocate a context id.\n");
+		goto out_no_id;
+	}
+
+	if (unlikely(res->id >= VMWGFX_NUM_DXCONTEXT)) {
+		ret = -EBUSY;
+		goto out_no_fifo;
+	}
+
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for context "
+			  "creation.\n");
+		ret = -ENOMEM;
+		goto out_no_fifo;
+	}
+
+	cmd->header.id = SVGA_3D_CMD_DX_DEFINE_CONTEXT;
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.cid = res->id;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+	vmw_fifo_resource_inc(dev_priv);
+
+	return 0;
+
+out_no_fifo:
+	vmw_resource_release_id(res);
+out_no_id:
+	return ret;
+}
+
+static int vmw_dx_context_bind(struct vmw_resource *res,
+			       struct ttm_validate_buffer *val_buf)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXBindContext body;
+	} *cmd;
+	struct ttm_buffer_object *bo = val_buf->bo;
+
+	BUG_ON(bo->mem.mem_type != VMW_PL_MOB);
+
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for context "
+			  "binding.\n");
+		return -ENOMEM;
+	}
+
+	cmd->header.id = SVGA_3D_CMD_DX_BIND_CONTEXT;
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.cid = res->id;
+	cmd->body.mobid = bo->mem.start;
+	cmd->body.validContents = res->backup_dirty;
+	res->backup_dirty = false;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+
+
+	return 0;
+}
+
+/**
+ * vmw_dx_context_scrub_cotables - Scrub all bindings and
+ * cotables from a context
+ *
+ * @ctx: Pointer to the context resource
+ * @readback: Whether to save the otable contents on scrubbing.
+ *
+ * COtables must be unbound before their context, but unbinding requires
+ * the backup buffer being reserved, whereas scrubbing does not.
+ * This function scrubs all cotables of a context, potentially reading back
+ * the contents into their backup buffers. However, scrubbing cotables
+ * also makes the device context invalid, so scrub all bindings first so
+ * that doesn't have to be done later with an invalid context.
+ */
+void vmw_dx_context_scrub_cotables(struct vmw_resource *ctx,
+				   bool readback)
+{
+	struct vmw_user_context *uctx =
+		container_of(ctx, struct vmw_user_context, res);
+	int i;
+
+	vmw_binding_state_scrub(uctx->cbs);
+	for (i = 0; i < SVGA_COTABLE_DX10_MAX; ++i) {
+		struct vmw_resource *res;
+
+		/* Avoid racing with ongoing cotable destruction. */
+		spin_lock(&uctx->cotable_lock);
+		res = uctx->cotables[vmw_cotable_scrub_order[i]];
+		if (res)
+			res = vmw_resource_reference_unless_doomed(res);
+		spin_unlock(&uctx->cotable_lock);
+		if (!res)
+			continue;
+
+		WARN_ON(vmw_cotable_scrub(res, readback));
+		vmw_resource_unreference(&res);
+	}
+}
+
+static int vmw_dx_context_unbind(struct vmw_resource *res,
+				 bool readback,
+				 struct ttm_validate_buffer *val_buf)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct ttm_buffer_object *bo = val_buf->bo;
+	struct vmw_fence_obj *fence;
+
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXReadbackContext body;
+	} *cmd1;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXBindContext body;
+	} *cmd2;
+	uint32_t submit_size;
+	uint8_t *cmd;
+
+
+	BUG_ON(bo->mem.mem_type != VMW_PL_MOB);
+
+	mutex_lock(&dev_priv->binding_mutex);
+	vmw_dx_context_scrub_cotables(res, readback);
+
+	submit_size = sizeof(*cmd2) + (readback ? sizeof(*cmd1) : 0);
+
+	cmd = vmw_fifo_reserve(dev_priv, submit_size);
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for context "
+			  "unbinding.\n");
+		mutex_unlock(&dev_priv->binding_mutex);
+		return -ENOMEM;
+	}
+
+	cmd2 = (void *) cmd;
+	if (readback) {
+		cmd1 = (void *) cmd;
+		cmd1->header.id = SVGA_3D_CMD_DX_READBACK_CONTEXT;
+		cmd1->header.size = sizeof(cmd1->body);
+		cmd1->body.cid = res->id;
+		cmd2 = (void *) (&cmd1[1]);
+	}
+	cmd2->header.id = SVGA_3D_CMD_DX_BIND_CONTEXT;
+	cmd2->header.size = sizeof(cmd2->body);
+	cmd2->body.cid = res->id;
+	cmd2->body.mobid = SVGA3D_INVALID_ID;
+
+	vmw_fifo_commit(dev_priv, submit_size);
+	mutex_unlock(&dev_priv->binding_mutex);
+
+	/*
+	 * Create a fence object and fence the backup buffer.
+	 */
+
+	(void) vmw_execbuf_fence_commands(NULL, dev_priv,
+					  &fence, NULL);
+
+	vmw_fence_single_bo(bo, fence);
+
+	if (likely(fence != NULL))
+		vmw_fence_obj_unreference(&fence);
+
+	return 0;
+}
+
+static int vmw_dx_context_destroy(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXDestroyContext body;
+	} *cmd;
+
+	if (likely(res->id == -1))
+		return 0;
+
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for context "
+			  "destruction.\n");
+		return -ENOMEM;
+	}
+
+	cmd->header.id = SVGA_3D_CMD_DX_DESTROY_CONTEXT;
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.cid = res->id;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+	if (dev_priv->query_cid == res->id)
+		dev_priv->query_cid_valid = false;
+	vmw_resource_release_id(res);
+	vmw_fifo_resource_dec(dev_priv);
+
+	return 0;
+}
+
 /**
  * User-space context management:
  */
@@ -435,6 +689,8 @@ static void vmw_user_context_free(struct vmw_resource *res)
 	    container_of(res, struct vmw_user_context, res);
 	struct vmw_private *dev_priv = res->dev_priv;
 
+	if (ctx->cbs)
+		vmw_binding_state_free(ctx->cbs);
 	ttm_base_object_kfree(ctx, base);
 	ttm_mem_global_free(vmw_mem_glob(dev_priv),
 			    vmw_user_context_size);
@@ -465,8 +721,8 @@ int vmw_context_destroy_ioctl(struct drm_device *dev, void *data,
 	return ttm_ref_object_base_unref(tfile, arg->cid, TTM_REF_USAGE);
 }
 
-int vmw_context_define_ioctl(struct drm_device *dev, void *data,
+static int vmw_context_define(struct drm_device *dev, void *data,
-			     struct drm_file *file_priv)
+			      struct drm_file *file_priv, bool dx)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
 	struct vmw_user_context *ctx;
@@ -476,6 +732,10 @@ int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
 	int ret;
 
+	if (!dev_priv->has_dx && dx) {
+		DRM_ERROR("DX contexts not supported by device.\n");
+		return -EINVAL;
+	}
 
 	/*
 	 * Approximate idr memory usage with 128 bytes. It will be limited
@@ -516,7 +776,7 @@ int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 	 * From here on, the destructor takes over resource freeing.
 	 */
 
-	ret = vmw_context_init(dev_priv, res, vmw_user_context_free);
+	ret = vmw_context_init(dev_priv, res, vmw_user_context_free, dx);
 	if (unlikely(ret != 0))
 		goto out_unlock;
 
@@ -535,371 +795,29 @@ int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 out_unlock:
 	ttm_read_unlock(&dev_priv->reservation_sem);
 	return ret;
-
 }
 
-/**
+int vmw_context_define_ioctl(struct drm_device *dev, void *data,
- * vmw_context_scrub_shader - scrub a shader binding from a context.
+			     struct drm_file *file_priv)
- *
- * @bi: single binding information.
- * @rebind: Whether to issue a bind instead of scrub command.
- */
-static int vmw_context_scrub_shader(struct vmw_ctx_bindinfo *bi, bool rebind)
 {
-	struct vmw_private *dev_priv = bi->ctx->dev_priv;
+	return vmw_context_define(dev, data, file_priv, false);
-	struct {
-		SVGA3dCmdHeader header;
-		SVGA3dCmdSetShader body;
-	} *cmd;
-
-	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed reserving FIFO space for shader "
-			  "unbinding.\n");
-		return -ENOMEM;
 }
 
-	cmd->header.id = SVGA_3D_CMD_SET_SHADER;
+int vmw_extended_context_define_ioctl(struct drm_device *dev, void *data,
-	cmd->header.size = sizeof(cmd->body);
+				      struct drm_file *file_priv)
-	cmd->body.cid = bi->ctx->id;
-	cmd->body.type = bi->i1.shader_type;
-	cmd->body.shid = ((rebind) ? bi->res->id : SVGA3D_INVALID_ID);
-	vmw_fifo_commit(dev_priv, sizeof(*cmd));
-
-	return 0;
-}
-
-/**
- * vmw_context_scrub_render_target - scrub a render target binding
- * from a context.
- *
- * @bi: single binding information.
- * @rebind: Whether to issue a bind instead of scrub command.
- */
-static int vmw_context_scrub_render_target(struct vmw_ctx_bindinfo *bi,
-					   bool rebind)
 {
-	struct vmw_private *dev_priv = bi->ctx->dev_priv;
+	union drm_vmw_extended_context_arg *arg = (typeof(arg)) data;
-	struct {
+	struct drm_vmw_context_arg *rep = &arg->rep;
-		SVGA3dCmdHeader header;
-		SVGA3dCmdSetRenderTarget body;
-	} *cmd;
 
-	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	switch (arg->req) {
-	if (unlikely(cmd == NULL)) {
+	case drm_vmw_context_legacy:
-		DRM_ERROR("Failed reserving FIFO space for render target "
+		return vmw_context_define(dev, rep, file_priv, false);
-			  "unbinding.\n");
+	case drm_vmw_context_dx:
-		return -ENOMEM;
+		return vmw_context_define(dev, rep, file_priv, true);
-	}
-
-	cmd->header.id = SVGA_3D_CMD_SETRENDERTARGET;
-	cmd->header.size = sizeof(cmd->body);
-	cmd->body.cid = bi->ctx->id;
-	cmd->body.type = bi->i1.rt_type;
-	cmd->body.target.sid = ((rebind) ? bi->res->id : SVGA3D_INVALID_ID);
-	cmd->body.target.face = 0;
-	cmd->body.target.mipmap = 0;
-	vmw_fifo_commit(dev_priv, sizeof(*cmd));
-
-	return 0;
-}
-
-/**
- * vmw_context_scrub_texture - scrub a texture binding from a context.
- *
- * @bi: single binding information.
- * @rebind: Whether to issue a bind instead of scrub command.
- *
- * TODO: Possibly complement this function with a function that takes
- * a list of texture bindings and combines them to a single command.
- */
-static int vmw_context_scrub_texture(struct vmw_ctx_bindinfo *bi,
-				     bool rebind)
-{
-	struct vmw_private *dev_priv = bi->ctx->dev_priv;
-	struct {
-		SVGA3dCmdHeader header;
-		struct {
-			SVGA3dCmdSetTextureState c;
-			SVGA3dTextureState s1;
-		} body;
-	} *cmd;
-
-	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
-	if (unlikely(cmd == NULL)) {
-		DRM_ERROR("Failed reserving FIFO space for texture "
-			  "unbinding.\n");
-		return -ENOMEM;
-	}
-
-
-	cmd->header.id = SVGA_3D_CMD_SETTEXTURESTATE;
-	cmd->header.size = sizeof(cmd->body);
-	cmd->body.c.cid = bi->ctx->id;
-	cmd->body.s1.stage = bi->i1.texture_stage;
-	cmd->body.s1.name = SVGA3D_TS_BIND_TEXTURE;
-	cmd->body.s1.value = ((rebind) ? bi->res->id : SVGA3D_INVALID_ID);
-	vmw_fifo_commit(dev_priv, sizeof(*cmd));
-
-	return 0;
-}
-
-/**
- * vmw_context_binding_drop: Stop tracking a context binding
- *
- * @cb: Pointer to binding tracker storage.
- *
- * Stops tracking a context binding, and re-initializes its storage.
- * Typically used when the context binding is replaced with a binding to
- * another (or the same, for that matter) resource.
- */
-static void vmw_context_binding_drop(struct vmw_ctx_binding *cb)
-{
-	list_del(&cb->ctx_list);
-	if (!list_empty(&cb->res_list))
-		list_del(&cb->res_list);
-	cb->bi.ctx = NULL;
-}
-
-/**
- * vmw_context_binding_add: Start tracking a context binding
- *
- * @cbs: Pointer to the context binding state tracker.
- * @bi: Information about the binding to track.
- *
- * Performs basic checks on the binding to make sure arguments are within
- * bounds and then starts tracking the binding in the context binding
- * state structure @cbs.
- */
-int vmw_context_binding_add(struct vmw_ctx_binding_state *cbs,
-			    const struct vmw_ctx_bindinfo *bi)
-{
-	struct vmw_ctx_binding *loc;
-
-	switch (bi->bt) {
-	case vmw_ctx_binding_rt:
-		if (unlikely((unsigned)bi->i1.rt_type >= SVGA3D_RT_MAX)) {
-			DRM_ERROR("Illegal render target type %u.\n",
-				  (unsigned) bi->i1.rt_type);
-			return -EINVAL;
-		}
-		loc = &cbs->render_targets[bi->i1.rt_type];
-		break;
-	case vmw_ctx_binding_tex:
-		if (unlikely((unsigned)bi->i1.texture_stage >=
-			     SVGA3D_NUM_TEXTURE_UNITS)) {
-			DRM_ERROR("Illegal texture/sampler unit %u.\n",
-				  (unsigned) bi->i1.texture_stage);
-			return -EINVAL;
-		}
-		loc = &cbs->texture_units[bi->i1.texture_stage];
-		break;
-	case vmw_ctx_binding_shader:
-		if (unlikely((unsigned)bi->i1.shader_type >=
-			     SVGA3D_SHADERTYPE_PREDX_MAX)) {
-			DRM_ERROR("Illegal shader type %u.\n",
-				  (unsigned) bi->i1.shader_type);
-			return -EINVAL;
-		}
-		loc = &cbs->shaders[bi->i1.shader_type];
-		break;
 	default:
-		BUG();
-	}
-
-	if (loc->bi.ctx != NULL)
-		vmw_context_binding_drop(loc);
-
-	loc->bi = *bi;
-	loc->bi.scrubbed = false;
-	list_add_tail(&loc->ctx_list, &cbs->list);
-	INIT_LIST_HEAD(&loc->res_list);
-
-	return 0;
-}
-
-/**
- * vmw_context_binding_transfer: Transfer a context binding tracking entry.
- *
- * @cbs: Pointer to the persistent context binding state tracker.
- * @bi: Information about the binding to track.
- *
- */
-static void vmw_context_binding_transfer(struct vmw_ctx_binding_state *cbs,
-					 const struct vmw_ctx_bindinfo *bi)
-{
-	struct vmw_ctx_binding *loc;
-
-	switch (bi->bt) {
-	case vmw_ctx_binding_rt:
-		loc = &cbs->render_targets[bi->i1.rt_type];
 		break;
-	case vmw_ctx_binding_tex:
-		loc = &cbs->texture_units[bi->i1.texture_stage];
-		break;
-	case vmw_ctx_binding_shader:
-		loc = &cbs->shaders[bi->i1.shader_type];
-		break;
-	default:
-		BUG();
 	}
-
+	return -EINVAL;
-	if (loc->bi.ctx != NULL)
-		vmw_context_binding_drop(loc);
-
-	if (bi->res != NULL) {
-		loc->bi = *bi;
-		list_add_tail(&loc->ctx_list, &cbs->list);
-		list_add_tail(&loc->res_list, &bi->res->binding_head);
-	}
-}
-
-/**
- * vmw_context_binding_kill - Kill a binding on the device
- * and stop tracking it.
- *
- * @cb: Pointer to binding tracker storage.
- *
- * Emits FIFO commands to scrub a binding represented by @cb.
- * Then stops tracking the binding and re-initializes its storage.
- */
-static void vmw_context_binding_kill(struct vmw_ctx_binding *cb)
-{
-	if (!cb->bi.scrubbed) {
-		(void) vmw_scrub_funcs[cb->bi.bt](&cb->bi, false);
-		cb->bi.scrubbed = true;
-	}
-	vmw_context_binding_drop(cb);
-}
-
-/**
- * vmw_context_binding_state_kill - Kill all bindings associated with a
- * struct vmw_ctx_binding state structure, and re-initialize the structure.
- *
- * @cbs: Pointer to the context binding state tracker.
- *
- * Emits commands to scrub all bindings associated with the
- * context binding state tracker. Then re-initializes the whole structure.
- */
-static void vmw_context_binding_state_kill(struct vmw_ctx_binding_state *cbs)
-{
-	struct vmw_ctx_binding *entry, *next;
-
-	list_for_each_entry_safe(entry, next, &cbs->list, ctx_list)
-		vmw_context_binding_kill(entry);
-}
-
-/**
- * vmw_context_binding_state_scrub - Scrub all bindings associated with a
- * struct vmw_ctx_binding state structure.
- *
- * @cbs: Pointer to the context binding state tracker.
- *
- * Emits commands to scrub all bindings associated with the
- * context binding state tracker.
- */
-static void vmw_context_binding_state_scrub(struct vmw_ctx_binding_state *cbs)
-{
-	struct vmw_ctx_binding *entry;
-
-	list_for_each_entry(entry, &cbs->list, ctx_list) {
-		if (!entry->bi.scrubbed) {
-			(void) vmw_scrub_funcs[entry->bi.bt](&entry->bi, false);
-			entry->bi.scrubbed = true;
-		}
-	}
-}
-
-/**
- * vmw_context_binding_res_list_kill - Kill all bindings on a
- * resource binding list
- *
- * @head: list head of resource binding list
- *
- * Kills all bindings associated with a specific resource. Typically
- * called before the resource is destroyed.
- */
-void vmw_context_binding_res_list_kill(struct list_head *head)
-{
-	struct vmw_ctx_binding *entry, *next;
-
-	list_for_each_entry_safe(entry, next, head, res_list)
-		vmw_context_binding_kill(entry);
-}
-
-/**
- * vmw_context_binding_res_list_scrub - Scrub all bindings on a
- * resource binding list
- *
- * @head: list head of resource binding list
- *
- * Scrub all bindings associated with a specific resource. Typically
- * called before the resource is evicted.
- */
-void vmw_context_binding_res_list_scrub(struct list_head *head)
-{
-	struct vmw_ctx_binding *entry;
-
-	list_for_each_entry(entry, head, res_list) {
-		if (!entry->bi.scrubbed) {
-			(void) vmw_scrub_funcs[entry->bi.bt](&entry->bi, false);
-			entry->bi.scrubbed = true;
-		}
-	}
-}
-
-/**
- * vmw_context_binding_state_transfer - Commit staged binding info
- *
- * @ctx: Pointer to context to commit the staged binding info to.
- * @from: Staged binding info built during execbuf.
- *
- * Transfers binding info from a temporary structure to the persistent
- * structure in the context. This can be done once commands
- */
-void vmw_context_binding_state_transfer(struct vmw_resource *ctx,
-					struct vmw_ctx_binding_state *from)
-{
-	struct vmw_user_context *uctx =
-		container_of(ctx, struct vmw_user_context, res);
-	struct vmw_ctx_binding *entry, *next;
-
-	list_for_each_entry_safe(entry, next, &from->list, ctx_list)
-		vmw_context_binding_transfer(&uctx->cbs, &entry->bi);
-}
-
-/**
- * vmw_context_rebind_all - Rebind all scrubbed bindings of a context
- *
- * @ctx: The context resource
- *
- * Walks through the context binding list and rebinds all scrubbed
- * resources.
- */
-int vmw_context_rebind_all(struct vmw_resource *ctx)
-{
-	struct vmw_ctx_binding *entry;
-	struct vmw_user_context *uctx =
-		container_of(ctx, struct vmw_user_context, res);
-	struct vmw_ctx_binding_state *cbs = &uctx->cbs;
-	int ret;
-
-	list_for_each_entry(entry, &cbs->list, ctx_list) {
-		if (likely(!entry->bi.scrubbed))
-			continue;
-
-		if (WARN_ON(entry->bi.res == NULL || entry->bi.res->id ==
-			    SVGA3D_INVALID_ID))
-			continue;
-
-		ret = vmw_scrub_funcs[entry->bi.bt](&entry->bi, true);
-		if (unlikely(ret != 0))
-			return ret;
-
-		entry->bi.scrubbed = false;
-	}
-
-	return 0;
 }
 
 /**
@@ -912,10 +830,39 @@ int vmw_context_rebind_all(struct vmw_resource *ctx)
  */
 struct list_head *vmw_context_binding_list(struct vmw_resource *ctx)
 {
-	return &(container_of(ctx, struct vmw_user_context, res)->cbs.list);
+	struct vmw_user_context *uctx =
+		container_of(ctx, struct vmw_user_context, res);
+
+	return vmw_binding_state_list(uctx->cbs);
 }
 
 struct vmw_cmdbuf_res_manager *vmw_context_res_man(struct vmw_resource *ctx)
 {
 	return container_of(ctx, struct vmw_user_context, res)->man;
 }
+
+struct vmw_resource *vmw_context_cotable(struct vmw_resource *ctx,
+					 SVGACOTableType cotable_type)
+{
+	if (cotable_type >= SVGA_COTABLE_DX10_MAX)
+		return ERR_PTR(-EINVAL);
+
+	return vmw_resource_reference
+		(container_of(ctx, struct vmw_user_context, res)->
+		 cotables[cotable_type]);
+}
+
+/**
+ * vmw_context_binding_state -
+ * Return a pointer to a context binding state structure
+ *
+ * @ctx: The context resource
+ *
+ * Returns the current state of bindings of the given context. Note that
+ * this state becomes stale as soon as the dev_priv::binding_mutex is unlocked.
+ */
+struct vmw_ctx_binding_state *
+vmw_context_binding_state(struct vmw_resource *ctx)
+{
+	return container_of(ctx, struct vmw_user_context, res)->cbs;
+}

drivers/gpu/drm/vmwgfx/vmwgfx_cotable.c

@@ -0,0 +1,662 @@
+/**************************************************************************
+ *
+ * Copyright © 2014 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+/*
+ * Treat context OTables as resources to make use of the resource
+ * backing MOB eviction mechanism, that is used to read back the COTable
+ * whenever the backing MOB is evicted.
+ */
+
+#include "vmwgfx_drv.h"
+#include "vmwgfx_resource_priv.h"
+#include <ttm/ttm_placement.h>
+#include "vmwgfx_so.h"
+
+/**
+ * struct vmw_cotable - Context Object Table resource
+ *
+ * @res: struct vmw_resource we are deriving from.
+ * @ctx: non-refcounted pointer to the owning context.
+ * @size_read_back: Size of data read back during eviction.
+ * @seen_entries: Seen entries in command stream for this cotable.
+ * @type: The cotable type.
+ * @scrubbed: Whether the cotable has been scrubbed.
+ * @resource_list: List of resources in the cotable.
+ */
+struct vmw_cotable {
+	struct vmw_resource res;
+	struct vmw_resource *ctx;
+	size_t size_read_back;
+	int seen_entries;
+	u32 type;
+	bool scrubbed;
+	struct list_head resource_list;
+};
+
+/**
+ * struct vmw_cotable_info - Static info about cotable types
+ *
+ * @min_initial_entries: Min number of initial intries at cotable allocation
+ * for this cotable type.
+ * @size: Size of each entry.
+ */
+struct vmw_cotable_info {
+	u32 min_initial_entries;
+	u32 size;
+	void (*unbind_func)(struct vmw_private *, struct list_head *,
+			    bool);
+};
+
+static const struct vmw_cotable_info co_info[] = {
+	{1, sizeof(SVGACOTableDXRTViewEntry), &vmw_view_cotable_list_destroy},
+	{1, sizeof(SVGACOTableDXDSViewEntry), &vmw_view_cotable_list_destroy},
+	{1, sizeof(SVGACOTableDXSRViewEntry), &vmw_view_cotable_list_destroy},
+	{1, sizeof(SVGACOTableDXElementLayoutEntry), NULL},
+	{1, sizeof(SVGACOTableDXBlendStateEntry), NULL},
+	{1, sizeof(SVGACOTableDXDepthStencilEntry), NULL},
+	{1, sizeof(SVGACOTableDXRasterizerStateEntry), NULL},
+	{1, sizeof(SVGACOTableDXSamplerEntry), NULL},
+	{1, sizeof(SVGACOTableDXStreamOutputEntry), NULL},
+	{1, sizeof(SVGACOTableDXQueryEntry), NULL},
+	{1, sizeof(SVGACOTableDXShaderEntry), &vmw_dx_shader_cotable_list_scrub}
+};
+
+/*
+ * Cotables with bindings that we remove must be scrubbed first,
+ * otherwise, the device will swap in an invalid context when we remove
+ * bindings before scrubbing a cotable...
+ */
+const SVGACOTableType vmw_cotable_scrub_order[] = {
+	SVGA_COTABLE_RTVIEW,
+	SVGA_COTABLE_DSVIEW,
+	SVGA_COTABLE_SRVIEW,
+	SVGA_COTABLE_DXSHADER,
+	SVGA_COTABLE_ELEMENTLAYOUT,
+	SVGA_COTABLE_BLENDSTATE,
+	SVGA_COTABLE_DEPTHSTENCIL,
+	SVGA_COTABLE_RASTERIZERSTATE,
+	SVGA_COTABLE_SAMPLER,
+	SVGA_COTABLE_STREAMOUTPUT,
+	SVGA_COTABLE_DXQUERY,
+};
+
+static int vmw_cotable_bind(struct vmw_resource *res,
+			    struct ttm_validate_buffer *val_buf);
+static int vmw_cotable_unbind(struct vmw_resource *res,
+			      bool readback,
+			      struct ttm_validate_buffer *val_buf);
+static int vmw_cotable_create(struct vmw_resource *res);
+static int vmw_cotable_destroy(struct vmw_resource *res);
+
+static const struct vmw_res_func vmw_cotable_func = {
+	.res_type = vmw_res_cotable,
+	.needs_backup = true,
+	.may_evict = true,
+	.type_name = "context guest backed object tables",
+	.backup_placement = &vmw_mob_placement,
+	.create = vmw_cotable_create,
+	.destroy = vmw_cotable_destroy,
+	.bind = vmw_cotable_bind,
+	.unbind = vmw_cotable_unbind,
+};
+
+/**
+ * vmw_cotable - Convert a struct vmw_resource pointer to a struct
+ * vmw_cotable pointer
+ *
+ * @res: Pointer to the resource.
+ */
+static struct vmw_cotable *vmw_cotable(struct vmw_resource *res)
+{
+	return container_of(res, struct vmw_cotable, res);
+}
+
+/**
+ * vmw_cotable_destroy - Cotable resource destroy callback
+ *
+ * @res: Pointer to the cotable resource.
+ *
+ * There is no device cotable destroy command, so this function only
+ * makes sure that the resource id is set to invalid.
+ */
+static int vmw_cotable_destroy(struct vmw_resource *res)
+{
+	res->id = -1;
+	return 0;
+}
+
+/**
+ * vmw_cotable_unscrub - Undo a cotable unscrub operation
+ *
+ * @res: Pointer to the cotable resource
+ *
+ * This function issues commands to (re)bind the cotable to
+ * its backing mob, which needs to be validated and reserved at this point.
+ * This is identical to bind() except the function interface looks different.
+ */
+static int vmw_cotable_unscrub(struct vmw_resource *res)
+{
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct ttm_buffer_object *bo = &res->backup->base;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXSetCOTable body;
+	} *cmd;
+
+	WARN_ON_ONCE(bo->mem.mem_type != VMW_PL_MOB);
+	lockdep_assert_held(&bo->resv->lock.base);
+
+	cmd = vmw_fifo_reserve_dx(dev_priv, sizeof(*cmd), SVGA3D_INVALID_ID);
+	if (!cmd) {
+		DRM_ERROR("Failed reserving FIFO space for cotable "
+			  "binding.\n");
+		return -ENOMEM;
+	}
+
+	WARN_ON(vcotbl->ctx->id == SVGA3D_INVALID_ID);
+	WARN_ON(bo->mem.mem_type != VMW_PL_MOB);
+	cmd->header.id = SVGA_3D_CMD_DX_SET_COTABLE;
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.cid = vcotbl->ctx->id;
+	cmd->body.type = vcotbl->type;
+	cmd->body.mobid = bo->mem.start;
+	cmd->body.validSizeInBytes = vcotbl->size_read_back;
+
+	vmw_fifo_commit_flush(dev_priv, sizeof(*cmd));
+	vcotbl->scrubbed = false;
+
+	return 0;
+}
+
+/**
+ * vmw_cotable_bind - Undo a cotable unscrub operation
+ *
+ * @res: Pointer to the cotable resource
+ * @val_buf: Pointer to a struct ttm_validate_buffer prepared by the caller
+ * for convenience / fencing.
+ *
+ * This function issues commands to (re)bind the cotable to
+ * its backing mob, which needs to be validated and reserved at this point.
+ */
+static int vmw_cotable_bind(struct vmw_resource *res,
+			    struct ttm_validate_buffer *val_buf)
+{
+	/*
+	 * The create() callback may have changed @res->backup without
+	 * the caller noticing, and with val_buf->bo still pointing to
+	 * the old backup buffer. Although hackish, and not used currently,
+	 * take the opportunity to correct the value here so that it's not
+	 * misused in the future.
+	 */
+	val_buf->bo = &res->backup->base;
+
+	return vmw_cotable_unscrub(res);
+}
+
+/**
+ * vmw_cotable_scrub - Scrub the cotable from the device.
+ *
+ * @res: Pointer to the cotable resource.
+ * @readback: Whether initiate a readback of the cotable data to the backup
+ * buffer.
+ *
+ * In some situations (context swapouts) it might be desirable to make the
+ * device forget about the cotable without performing a full unbind. A full
+ * unbind requires reserved backup buffers and it might not be possible to
+ * reserve them due to locking order violation issues. The vmw_cotable_scrub
+ * function implements a partial unbind() without that requirement but with the
+ * following restrictions.
+ * 1) Before the cotable is again used by the GPU, vmw_cotable_unscrub() must
+ *    be called.
+ * 2) Before the cotable backing buffer is used by the CPU, or during the
+ *    resource destruction, vmw_cotable_unbind() must be called.
+ */
+int vmw_cotable_scrub(struct vmw_resource *res, bool readback)
+{
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+	size_t submit_size;
+
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXReadbackCOTable body;
+	} *cmd0;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXSetCOTable body;
+	} *cmd1;
+
+	if (vcotbl->scrubbed)
+		return 0;
+
+	if (co_info[vcotbl->type].unbind_func)
+		co_info[vcotbl->type].unbind_func(dev_priv,
+						  &vcotbl->resource_list,
+						  readback);
+	submit_size = sizeof(*cmd1);
+	if (readback)
+		submit_size += sizeof(*cmd0);
+
+	cmd1 = vmw_fifo_reserve_dx(dev_priv, submit_size, SVGA3D_INVALID_ID);
+	if (!cmd1) {
+		DRM_ERROR("Failed reserving FIFO space for cotable "
+			  "unbinding.\n");
+		return -ENOMEM;
+	}
+
+	vcotbl->size_read_back = 0;
+	if (readback) {
+		cmd0 = (void *) cmd1;
+		cmd0->header.id = SVGA_3D_CMD_DX_READBACK_COTABLE;
+		cmd0->header.size = sizeof(cmd0->body);
+		cmd0->body.cid = vcotbl->ctx->id;
+		cmd0->body.type = vcotbl->type;
+		cmd1 = (void *) &cmd0[1];
+		vcotbl->size_read_back = res->backup_size;
+	}
+	cmd1->header.id = SVGA_3D_CMD_DX_SET_COTABLE;
+	cmd1->header.size = sizeof(cmd1->body);
+	cmd1->body.cid = vcotbl->ctx->id;
+	cmd1->body.type = vcotbl->type;
+	cmd1->body.mobid = SVGA3D_INVALID_ID;
+	cmd1->body.validSizeInBytes = 0;
+	vmw_fifo_commit_flush(dev_priv, submit_size);
+	vcotbl->scrubbed = true;
+
+	/* Trigger a create() on next validate. */
+	res->id = -1;
+
+	return 0;
+}
+
+/**
+ * vmw_cotable_unbind - Cotable resource unbind callback
+ *
+ * @res: Pointer to the cotable resource.
+ * @readback: Whether to read back cotable data to the backup buffer.
+ * val_buf: Pointer to a struct ttm_validate_buffer prepared by the caller
+ * for convenience / fencing.
+ *
+ * Unbinds the cotable from the device and fences the backup buffer.
+ */
+static int vmw_cotable_unbind(struct vmw_resource *res,
+			      bool readback,
+			      struct ttm_validate_buffer *val_buf)
+{
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct ttm_buffer_object *bo = val_buf->bo;
+	struct vmw_fence_obj *fence;
+	int ret;
+
+	if (list_empty(&res->mob_head))
+		return 0;
+
+	WARN_ON_ONCE(bo->mem.mem_type != VMW_PL_MOB);
+	lockdep_assert_held(&bo->resv->lock.base);
+
+	mutex_lock(&dev_priv->binding_mutex);
+	if (!vcotbl->scrubbed)
+		vmw_dx_context_scrub_cotables(vcotbl->ctx, readback);
+	mutex_unlock(&dev_priv->binding_mutex);
+	(void) vmw_execbuf_fence_commands(NULL, dev_priv, &fence, NULL);
+	vmw_fence_single_bo(bo, fence);
+	if (likely(fence != NULL))
+		vmw_fence_obj_unreference(&fence);
+
+	return ret;
+}
+
+/**
+ * vmw_cotable_readback - Read back a cotable without unbinding.
+ *
+ * @res: The cotable resource.
+ *
+ * Reads back a cotable to its backing mob without scrubbing the MOB from
+ * the cotable. The MOB is fenced for subsequent CPU access.
+ */
+static int vmw_cotable_readback(struct vmw_resource *res)
+{
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXReadbackCOTable body;
+	} *cmd;
+	struct vmw_fence_obj *fence;
+
+	if (!vcotbl->scrubbed) {
+		cmd = vmw_fifo_reserve_dx(dev_priv, sizeof(*cmd),
+					  SVGA3D_INVALID_ID);
+		if (!cmd) {
+			DRM_ERROR("Failed reserving FIFO space for cotable "
+				  "readback.\n");
+			return -ENOMEM;
+		}
+		cmd->header.id = SVGA_3D_CMD_DX_READBACK_COTABLE;
+		cmd->header.size = sizeof(cmd->body);
+		cmd->body.cid = vcotbl->ctx->id;
+		cmd->body.type = vcotbl->type;
+		vcotbl->size_read_back = res->backup_size;
+		vmw_fifo_commit(dev_priv, sizeof(*cmd));
+	}
+
+	(void) vmw_execbuf_fence_commands(NULL, dev_priv, &fence, NULL);
+	vmw_fence_single_bo(&res->backup->base, fence);
+	vmw_fence_obj_unreference(&fence);
+
+	return 0;
+}
+
+/**
+ * vmw_cotable_resize - Resize a cotable.
+ *
+ * @res: The cotable resource.
+ * @new_size: The new size.
+ *
+ * Resizes a cotable and binds the new backup buffer.
+ * On failure the cotable is left intact.
+ * Important! This function may not fail once the MOB switch has been
+ * committed to hardware. That would put the device context in an
+ * invalid state which we can't currently recover from.
+ */
+static int vmw_cotable_resize(struct vmw_resource *res, size_t new_size)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+	struct vmw_dma_buffer *buf, *old_buf = res->backup;
+	struct ttm_buffer_object *bo, *old_bo = &res->backup->base;
+	size_t old_size = res->backup_size;
+	size_t old_size_read_back = vcotbl->size_read_back;
+	size_t cur_size_read_back;
+	struct ttm_bo_kmap_obj old_map, new_map;
+	int ret;
+	size_t i;
+
+	ret = vmw_cotable_readback(res);
+	if (ret)
+		return ret;
+
+	cur_size_read_back = vcotbl->size_read_back;
+	vcotbl->size_read_back = old_size_read_back;
+
+	/*
+	 * While device is processing, Allocate and reserve a buffer object
+	 * for the new COTable. Initially pin the buffer object to make sure
+	 * we can use tryreserve without failure.
+	 */
+	buf = kzalloc(sizeof(*buf), GFP_KERNEL);
+	if (!buf)
+		return -ENOMEM;
+
+	ret = vmw_dmabuf_init(dev_priv, buf, new_size, &vmw_mob_ne_placement,
+			      true, vmw_dmabuf_bo_free);
+	if (ret) {
+		DRM_ERROR("Failed initializing new cotable MOB.\n");
+		return ret;
+	}
+
+	bo = &buf->base;
+	WARN_ON_ONCE(ttm_bo_reserve(bo, false, true, false, NULL));
+
+	ret = ttm_bo_wait(old_bo, false, false, false);
+	if (unlikely(ret != 0)) {
+		DRM_ERROR("Failed waiting for cotable unbind.\n");
+		goto out_wait;
+	}
+
+	/*
+	 * Do a page by page copy of COTables. This eliminates slow vmap()s.
+	 * This should really be a TTM utility.
+	 */
+	for (i = 0; i < old_bo->num_pages; ++i) {
+		bool dummy;
+
+		ret = ttm_bo_kmap(old_bo, i, 1, &old_map);
+		if (unlikely(ret != 0)) {
+			DRM_ERROR("Failed mapping old COTable on resize.\n");
+			goto out_wait;
+		}
+		ret = ttm_bo_kmap(bo, i, 1, &new_map);
+		if (unlikely(ret != 0)) {
+			DRM_ERROR("Failed mapping new COTable on resize.\n");
+			goto out_map_new;
+		}
+		memcpy(ttm_kmap_obj_virtual(&new_map, &dummy),
+		       ttm_kmap_obj_virtual(&old_map, &dummy),
+		       PAGE_SIZE);
+		ttm_bo_kunmap(&new_map);
+		ttm_bo_kunmap(&old_map);
+	}
+
+	/* Unpin new buffer, and switch backup buffers. */
+	ret = ttm_bo_validate(bo, &vmw_mob_placement, false, false);
+	if (unlikely(ret != 0)) {
+		DRM_ERROR("Failed validating new COTable backup buffer.\n");
+		goto out_wait;
+	}
+
+	res->backup = buf;
+	res->backup_size = new_size;
+	vcotbl->size_read_back = cur_size_read_back;
+
+	/*
+	 * Now tell the device to switch. If this fails, then we need to
+	 * revert the full resize.
+	 */
+	ret = vmw_cotable_unscrub(res);
+	if (ret) {
+		DRM_ERROR("Failed switching COTable backup buffer.\n");
+		res->backup = old_buf;
+		res->backup_size = old_size;
+		vcotbl->size_read_back = old_size_read_back;
+		goto out_wait;
+	}
+
+	/* Let go of the old mob. */
+	list_del(&res->mob_head);
+	list_add_tail(&res->mob_head, &buf->res_list);
+	vmw_dmabuf_unreference(&old_buf);
+	res->id = vcotbl->type;
+
+	return 0;
+
+out_map_new:
+	ttm_bo_kunmap(&old_map);
+out_wait:
+	ttm_bo_unreserve(bo);
+	vmw_dmabuf_unreference(&buf);
+
+	return ret;
+}
+
+/**
+ * vmw_cotable_create - Cotable resource create callback
+ *
+ * @res: Pointer to a cotable resource.
+ *
+ * There is no separate create command for cotables, so this callback, which
+ * is called before bind() in the validation sequence is instead used for two
+ * things.
+ * 1) Unscrub the cotable if it is scrubbed and still attached to a backup
+ *    buffer, that is, if @res->mob_head is non-empty.
+ * 2) Resize the cotable if needed.
+ */
+static int vmw_cotable_create(struct vmw_resource *res)
+{
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+	size_t new_size = res->backup_size;
+	size_t needed_size;
+	int ret;
+
+	/* Check whether we need to resize the cotable */
+	needed_size = (vcotbl->seen_entries + 1) * co_info[vcotbl->type].size;
+	while (needed_size > new_size)
+		new_size *= 2;
+
+	if (likely(new_size <= res->backup_size)) {
+		if (vcotbl->scrubbed && !list_empty(&res->mob_head)) {
+			ret = vmw_cotable_unscrub(res);
+			if (ret)
+				return ret;
+		}
+		res->id = vcotbl->type;
+		return 0;
+	}
+
+	return vmw_cotable_resize(res, new_size);
+}
+
+/**
+ * vmw_hw_cotable_destroy - Cotable hw_destroy callback
+ *
+ * @res: Pointer to a cotable resource.
+ *
+ * The final (part of resource destruction) destroy callback.
+ */
+static void vmw_hw_cotable_destroy(struct vmw_resource *res)
+{
+	(void) vmw_cotable_destroy(res);
+}
+
+static size_t cotable_acc_size;
+
+/**
+ * vmw_cotable_free - Cotable resource destructor
+ *
+ * @res: Pointer to a cotable resource.
+ */
+static void vmw_cotable_free(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	kfree(res);
+	ttm_mem_global_free(vmw_mem_glob(dev_priv), cotable_acc_size);
+}
+
+/**
+ * vmw_cotable_alloc - Create a cotable resource
+ *
+ * @dev_priv: Pointer to a device private struct.
+ * @ctx: Pointer to the context resource.
+ * The cotable resource will not add a refcount.
+ * @type: The cotable type.
+ */
+struct vmw_resource *vmw_cotable_alloc(struct vmw_private *dev_priv,
+				       struct vmw_resource *ctx,
+				       u32 type)
+{
+	struct vmw_cotable *vcotbl;
+	int ret;
+	u32 num_entries;
+
+	if (unlikely(cotable_acc_size == 0))
+		cotable_acc_size = ttm_round_pot(sizeof(struct vmw_cotable));
+
+	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv),
+				   cotable_acc_size, false, true);
+	if (unlikely(ret))
+		return ERR_PTR(ret);
+
+	vcotbl = kzalloc(sizeof(*vcotbl), GFP_KERNEL);
+	if (unlikely(vcotbl == NULL)) {
+		ret = -ENOMEM;
+		goto out_no_alloc;
+	}
+
+	ret = vmw_resource_init(dev_priv, &vcotbl->res, true,
+				vmw_cotable_free, &vmw_cotable_func);
+	if (unlikely(ret != 0))
+		goto out_no_init;
+
+	INIT_LIST_HEAD(&vcotbl->resource_list);
+	vcotbl->res.id = type;
+	vcotbl->res.backup_size = PAGE_SIZE;
+	num_entries = PAGE_SIZE / co_info[type].size;
+	if (num_entries < co_info[type].min_initial_entries) {
+		vcotbl->res.backup_size = co_info[type].min_initial_entries *
+			co_info[type].size;
+		vcotbl->res.backup_size =
+			(vcotbl->res.backup_size + PAGE_SIZE - 1) & PAGE_MASK;
+	}
+
+	vcotbl->scrubbed = true;
+	vcotbl->seen_entries = -1;
+	vcotbl->type = type;
+	vcotbl->ctx = ctx;
+
+	vmw_resource_activate(&vcotbl->res, vmw_hw_cotable_destroy);
+
+	return &vcotbl->res;
+
+out_no_init:
+	kfree(vcotbl);
+out_no_alloc:
+	ttm_mem_global_free(vmw_mem_glob(dev_priv), cotable_acc_size);
+	return ERR_PTR(ret);
+}
+
+/**
+ * vmw_cotable_notify - Notify the cotable about an item creation
+ *
+ * @res: Pointer to a cotable resource.
+ * @id: Item id.
+ */
+int vmw_cotable_notify(struct vmw_resource *res, int id)
+{
+	struct vmw_cotable *vcotbl = vmw_cotable(res);
+
+	if (id < 0 || id >= SVGA_COTABLE_MAX_IDS) {
+		DRM_ERROR("Illegal COTable id. Type is %u. Id is %d\n",
+			  (unsigned) vcotbl->type, id);
+		return -EINVAL;
+	}
+
+	if (vcotbl->seen_entries < id) {
+		/* Trigger a call to create() on next validate */
+		res->id = -1;
+		vcotbl->seen_entries = id;
+	}
+
+	return 0;
+}
+
+/**
+ * vmw_cotable_add_view - add a view to the cotable's list of active views.
+ *
+ * @res: pointer struct vmw_resource representing the cotable.
+ * @head: pointer to the struct list_head member of the resource, dedicated
+ * to the cotable active resource list.
+ */
+void vmw_cotable_add_resource(struct vmw_resource *res, struct list_head *head)
+{
+	struct vmw_cotable *vcotbl =
+		container_of(res, struct vmw_cotable, res);
+
+	list_add_tail(head, &vcotbl->resource_list);
+}

drivers/gpu/drm/vmwgfx/vmwgfx_drv.c

@@ -28,6 +28,7 @@
 
 #include <drm/drmP.h>
 #include "vmwgfx_drv.h"
+#include "vmwgfx_binding.h"
 #include <drm/ttm/ttm_placement.h>
 #include <drm/ttm/ttm_bo_driver.h>
 #include <drm/ttm/ttm_object.h>
@@ -127,6 +128,9 @@
 #define DRM_IOCTL_VMW_SYNCCPU					\
 	DRM_IOW(DRM_COMMAND_BASE + DRM_VMW_SYNCCPU,		\
 		 struct drm_vmw_synccpu_arg)
+#define DRM_IOCTL_VMW_CREATE_EXTENDED_CONTEXT			\
+	DRM_IOWR(DRM_COMMAND_BASE + DRM_VMW_CREATE_EXTENDED_CONTEXT,	\
+		struct drm_vmw_context_arg)
 
 /**
  * The core DRM version of this macro doesn't account for
@@ -168,8 +172,8 @@ static const struct drm_ioctl_desc vmw_ioctls[] = {
 		      DRM_UNLOCKED | DRM_RENDER_ALLOW),
 	VMW_IOCTL_DEF(VMW_REF_SURFACE, vmw_surface_reference_ioctl,
 		      DRM_AUTH | DRM_UNLOCKED | DRM_RENDER_ALLOW),
-	VMW_IOCTL_DEF(VMW_EXECBUF, vmw_execbuf_ioctl,
+	VMW_IOCTL_DEF(VMW_EXECBUF, NULL, DRM_AUTH | DRM_UNLOCKED |
-		      DRM_AUTH | DRM_UNLOCKED | DRM_RENDER_ALLOW),
+		      DRM_RENDER_ALLOW),
 	VMW_IOCTL_DEF(VMW_FENCE_WAIT, vmw_fence_obj_wait_ioctl,
 		      DRM_UNLOCKED | DRM_RENDER_ALLOW),
 	VMW_IOCTL_DEF(VMW_FENCE_SIGNALED,
@@ -206,6 +210,9 @@ static const struct drm_ioctl_desc vmw_ioctls[] = {
 	VMW_IOCTL_DEF(VMW_SYNCCPU,
 		      vmw_user_dmabuf_synccpu_ioctl,
 		      DRM_UNLOCKED | DRM_RENDER_ALLOW),
+	VMW_IOCTL_DEF(VMW_CREATE_EXTENDED_CONTEXT,
+		      vmw_extended_context_define_ioctl,
+		      DRM_AUTH | DRM_UNLOCKED | DRM_RENDER_ALLOW),
 };
 
 static struct pci_device_id vmw_pci_id_list[] = {
@@ -390,8 +397,10 @@ static int vmw_request_device(struct vmw_private *dev_priv)
 	}
 	vmw_fence_fifo_up(dev_priv->fman);
 	dev_priv->cman = vmw_cmdbuf_man_create(dev_priv);
-	if (IS_ERR(dev_priv->cman))
+	if (IS_ERR(dev_priv->cman)) {
 		dev_priv->cman = NULL;
+		dev_priv->has_dx = false;
+	}
 
 	ret = vmw_request_device_late(dev_priv);
 	if (ret)
@@ -848,6 +857,14 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 		}
 	}
 
+	if (dev_priv->has_mob) {
+		spin_lock(&dev_priv->cap_lock);
+		vmw_write(dev_priv, SVGA_REG_DEV_CAP, SVGA3D_DEVCAP_DX);
+		dev_priv->has_dx = !!vmw_read(dev_priv, SVGA_REG_DEV_CAP);
+		spin_unlock(&dev_priv->cap_lock);
+	}
+
+
 	ret = vmw_kms_init(dev_priv);
 	if (unlikely(ret != 0))
 		goto out_no_kms;
@@ -857,6 +874,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	if (ret)
 		goto out_no_fifo;
 
+	DRM_INFO("DX: %s\n", dev_priv->has_dx ? "yes." : "no.");
+
 	if (dev_priv->enable_fb) {
 		vmw_fifo_resource_inc(dev_priv);
 		vmw_svga_enable(dev_priv);
@@ -900,6 +919,8 @@ static int vmw_driver_load(struct drm_device *dev, unsigned long chipset)
 	for (i = vmw_res_context; i < vmw_res_max; ++i)
 		idr_destroy(&dev_priv->res_idr[i]);
 
+	if (dev_priv->ctx.staged_bindings)
+		vmw_binding_state_free(dev_priv->ctx.staged_bindings);
 	kfree(dev_priv);
 	return ret;
 }
@@ -945,6 +966,8 @@ static int vmw_driver_unload(struct drm_device *dev)
 	iounmap(dev_priv->mmio_virt);
 	arch_phys_wc_del(dev_priv->mmio_mtrr);
 	(void)ttm_bo_device_release(&dev_priv->bdev);
+	if (dev_priv->ctx.staged_bindings)
+		vmw_binding_state_free(dev_priv->ctx.staged_bindings);
 	vmw_ttm_global_release(dev_priv);
 
 	for (i = vmw_res_context; i < vmw_res_max; ++i)
@@ -1082,11 +1105,21 @@ static long vmw_generic_ioctl(struct file *filp, unsigned int cmd,
 		const struct drm_ioctl_desc *ioctl =
 			&vmw_ioctls[nr - DRM_COMMAND_BASE];
 
-		if (unlikely(ioctl->cmd != cmd)) {
+		if (nr == DRM_COMMAND_BASE + DRM_VMW_EXECBUF) {
-			DRM_ERROR("Invalid command format, ioctl %d\n",
+			ret = (long) drm_ioctl_permit(ioctl->flags, file_priv);
-				  nr - DRM_COMMAND_BASE);
+			if (unlikely(ret != 0))
-			return -EINVAL;
+				return ret;
+
+			if (unlikely((cmd & (IOC_IN | IOC_OUT)) != IOC_IN))
+				goto out_io_encoding;
+
+			return (long) vmw_execbuf_ioctl(dev, arg, file_priv,
+							_IOC_SIZE(cmd));
 		}
+
+		if (unlikely(ioctl->cmd != cmd))
+			goto out_io_encoding;
+
 		flags = ioctl->flags;
 	} else if (!drm_ioctl_flags(nr, &flags))
 		return -EINVAL;
@@ -1106,6 +1139,12 @@ static long vmw_generic_ioctl(struct file *filp, unsigned int cmd,
 		ttm_read_unlock(&vmaster->lock);
 
 	return ret;
+
+out_io_encoding:
+	DRM_ERROR("Invalid command format, ioctl %d\n",
+		  nr - DRM_COMMAND_BASE);
+
+	return -EINVAL;
 }
 
 static long vmw_unlocked_ioctl(struct file *filp, unsigned int cmd,
@@ -1156,7 +1195,6 @@ static void vmw_master_destroy(struct drm_device *dev,
 	kfree(vmaster);
 }
 
-
 static int vmw_master_set(struct drm_device *dev,
 			  struct drm_file *file_priv,
 			  bool from_open)

drivers/gpu/drm/vmwgfx/vmwgfx_drv.h

@@ -59,6 +59,8 @@
 #define VMWGFX_NUM_GB_SHADER 20000
 #define VMWGFX_NUM_GB_SURFACE 32768
 #define VMWGFX_NUM_GB_SCREEN_TARGET VMWGFX_MAX_DISPLAYS
+#define VMWGFX_NUM_DXCONTEXT 256
+#define VMWGFX_NUM_DXQUERY 512
 #define VMWGFX_NUM_MOB (VMWGFX_NUM_GB_CONTEXT +\
 			VMWGFX_NUM_GB_SHADER +\
 			VMWGFX_NUM_GB_SURFACE +\
@@ -132,6 +134,9 @@ enum vmw_res_type {
 	vmw_res_surface,
 	vmw_res_stream,
 	vmw_res_shader,
+	vmw_res_dx_context,
+	vmw_res_cotable,
+	vmw_res_view,
 	vmw_res_max
 };
 
@@ -139,7 +144,8 @@ enum vmw_res_type {
  * Resources that are managed using command streams.
  */
 enum vmw_cmdbuf_res_type {
-	vmw_cmdbuf_res_compat_shader
+	vmw_cmdbuf_res_shader,
+	vmw_cmdbuf_res_view
 };
 
 struct vmw_cmdbuf_res_manager;
@@ -162,11 +168,13 @@ struct vmw_surface {
 	struct drm_vmw_size *sizes;
 	uint32_t num_sizes;
 	bool scanout;
+	uint32_t array_size;
 	/* TODO so far just a extra pointer */
 	struct vmw_cursor_snooper snooper;
 	struct vmw_surface_offset *offsets;
 	SVGA3dTextureFilter autogen_filter;
 	uint32_t multisample_count;
+	struct list_head view_list;
 };
 
 struct vmw_marker_queue {
@@ -186,6 +194,7 @@ struct vmw_fifo_state {
 	struct mutex fifo_mutex;
 	struct rw_semaphore rwsem;
 	struct vmw_marker_queue marker_queue;
+	bool dx;
 };
 
 struct vmw_relocation {
@@ -265,73 +274,6 @@ struct vmw_piter {
 	struct page *(*page)(struct vmw_piter *);
 };
 
-/*
- * enum vmw_ctx_binding_type - abstract resource to context binding types
- */
-enum vmw_ctx_binding_type {
-	vmw_ctx_binding_shader,
-	vmw_ctx_binding_rt,
-	vmw_ctx_binding_tex,
-	vmw_ctx_binding_max
-};
-
-/**
- * struct vmw_ctx_bindinfo - structure representing a single context binding
- *
- * @ctx: Pointer to the context structure. NULL means the binding is not
- * active.
- * @res: Non ref-counted pointer to the bound resource.
- * @bt: The binding type.
- * @i1: Union of information needed to unbind.
- */
-struct vmw_ctx_bindinfo {
-	struct vmw_resource *ctx;
-	struct vmw_resource *res;
-	enum vmw_ctx_binding_type bt;
-	bool scrubbed;
-	union {
-		SVGA3dShaderType shader_type;
-		SVGA3dRenderTargetType rt_type;
-		uint32 texture_stage;
-	} i1;
-};
-
-/**
- * struct vmw_ctx_binding - structure representing a single context binding
- *                        - suitable for tracking in a context
- *
- * @ctx_list: List head for context.
- * @res_list: List head for bound resource.
- * @bi: Binding info
- */
-struct vmw_ctx_binding {
-	struct list_head ctx_list;
-	struct list_head res_list;
-	struct vmw_ctx_bindinfo bi;
-};
-
-
-/**
- * struct vmw_ctx_binding_state - context binding state
- *
- * @list: linked list of individual bindings.
- * @render_targets: Render target bindings.
- * @texture_units: Texture units/samplers bindings.
- * @shaders: Shader bindings.
- *
- * Note that this structure also provides storage space for the individual
- * struct vmw_ctx_binding objects, so that no dynamic allocation is needed
- * for individual bindings.
- *
- */
-struct vmw_ctx_binding_state {
-	struct list_head list;
-	struct vmw_ctx_binding render_targets[SVGA3D_RT_MAX];
-	struct vmw_ctx_binding texture_units[SVGA3D_NUM_TEXTURE_UNITS];
-	struct vmw_ctx_binding shaders[SVGA3D_SHADERTYPE_PREDX_MAX];
-};
-
-
 /*
  * enum vmw_display_unit_type - Describes the display unit
  */
@@ -356,6 +298,7 @@ struct vmw_sw_context{
 	uint32_t *cmd_bounce;
 	uint32_t cmd_bounce_size;
 	struct list_head resource_list;
+	struct list_head ctx_resource_list; /* For contexts and cotables */
 	struct vmw_dma_buffer *cur_query_bo;
 	struct list_head res_relocations;
 	uint32_t *buf_start;
@@ -363,8 +306,13 @@ struct vmw_sw_context{
 	struct vmw_resource *last_query_ctx;
 	bool needs_post_query_barrier;
 	struct vmw_resource *error_resource;
-	struct vmw_ctx_binding_state staged_bindings;
+	struct vmw_ctx_binding_state *staged_bindings;
+	bool staged_bindings_inuse;
 	struct list_head staged_cmd_res;
+	struct vmw_resource_val_node *dx_ctx_node;
+	struct vmw_dma_buffer *dx_query_mob;
+	struct vmw_resource *dx_query_ctx;
+	struct vmw_cmdbuf_res_manager *man;
 };
 
 struct vmw_legacy_display;
@@ -382,6 +330,26 @@ struct vmw_vga_topology_state {
 	uint32_t pos_y;
 };
 
+
+/*
+ * struct vmw_otable - Guest Memory OBject table metadata
+ *
+ * @size:           Size of the table (page-aligned).
+ * @page_table:     Pointer to a struct vmw_mob holding the page table.
+ */
+struct vmw_otable {
+	unsigned long size;
+	struct vmw_mob *page_table;
+	bool enabled;
+};
+
+struct vmw_otable_batch {
+	unsigned num_otables;
+	struct vmw_otable *otables;
+	struct vmw_resource *context;
+	struct ttm_buffer_object *otable_bo;
+};
+
 struct vmw_private {
 	struct ttm_bo_device bdev;
 	struct ttm_bo_global_ref bo_global_ref;
@@ -417,6 +385,7 @@ struct vmw_private {
 	bool has_mob;
 	spinlock_t hw_lock;
 	spinlock_t cap_lock;
+	bool has_dx;
 
 	/*
 	 * VGA registers.
@@ -552,8 +521,7 @@ struct vmw_private {
 	/*
 	 * Guest Backed stuff
 	 */
-	struct ttm_buffer_object *otable_bo;
+	struct vmw_otable_batch otable_batch;
-	struct vmw_otable *otables;
 
 	struct vmw_cmdbuf_man *cman;
 };
@@ -685,6 +653,7 @@ extern int vmw_user_stream_lookup(struct vmw_private *dev_priv,
 				  uint32_t *inout_id,
 				  struct vmw_resource **out);
 extern void vmw_resource_unreserve(struct vmw_resource *res,
+				   bool switch_backup,
 				   struct vmw_dma_buffer *new_backup,
 				   unsigned long new_backup_offset);
 extern void vmw_resource_move_notify(struct ttm_buffer_object *bo,
@@ -742,7 +711,10 @@ extern int vmw_fifo_init(struct vmw_private *dev_priv,
 extern void vmw_fifo_release(struct vmw_private *dev_priv,
 			     struct vmw_fifo_state *fifo);
 extern void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes);
+extern void *
+vmw_fifo_reserve_dx(struct vmw_private *dev_priv, uint32_t bytes, int ctx_id);
 extern void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes);
+extern void vmw_fifo_commit_flush(struct vmw_private *dev_priv, uint32_t bytes);
 extern int vmw_fifo_send_fence(struct vmw_private *dev_priv,
 			       uint32_t *seqno);
 extern void vmw_fifo_ping_host_locked(struct vmw_private *, uint32_t reason);
@@ -828,14 +800,15 @@ static inline struct page *vmw_piter_page(struct vmw_piter *viter)
  * Command submission - vmwgfx_execbuf.c
  */
 
-extern int vmw_execbuf_ioctl(struct drm_device *dev, void *data,
+extern int vmw_execbuf_ioctl(struct drm_device *dev, unsigned long data,
-			     struct drm_file *file_priv);
+			     struct drm_file *file_priv, size_t size);
 extern int vmw_execbuf_process(struct drm_file *file_priv,
 			       struct vmw_private *dev_priv,
 			       void __user *user_commands,
 			       void *kernel_commands,
 			       uint32_t command_size,
 			       uint64_t throttle_us,
+			       uint32_t dx_context_handle,
 			       struct drm_vmw_fence_rep __user
 			       *user_fence_rep,
 			       struct vmw_fence_obj **out_fence);
@@ -960,6 +933,7 @@ int vmw_dumb_destroy(struct drm_file *file_priv,
 		     uint32_t handle);
 extern int vmw_resource_pin(struct vmw_resource *res, bool interruptible);
 extern void vmw_resource_unpin(struct vmw_resource *res);
+extern enum vmw_res_type vmw_res_type(const struct vmw_resource *res);
 
 /**
  * Overlay control - vmwgfx_overlay.c
@@ -1016,27 +990,28 @@ extern void vmw_otables_takedown(struct vmw_private *dev_priv);
 
 extern const struct vmw_user_resource_conv *user_context_converter;
 
-extern struct vmw_resource *vmw_context_alloc(struct vmw_private *dev_priv);
-
 extern int vmw_context_check(struct vmw_private *dev_priv,
 			     struct ttm_object_file *tfile,
 			     int id,
 			     struct vmw_resource **p_res);
 extern int vmw_context_define_ioctl(struct drm_device *dev, void *data,
 				    struct drm_file *file_priv);
+extern int vmw_extended_context_define_ioctl(struct drm_device *dev, void *data,
+					     struct drm_file *file_priv);
 extern int vmw_context_destroy_ioctl(struct drm_device *dev, void *data,
 				     struct drm_file *file_priv);
-extern int vmw_context_binding_add(struct vmw_ctx_binding_state *cbs,
-				   const struct vmw_ctx_bindinfo *ci);
-extern void
-vmw_context_binding_state_transfer(struct vmw_resource *res,
-				   struct vmw_ctx_binding_state *cbs);
-extern void vmw_context_binding_res_list_kill(struct list_head *head);
-extern void vmw_context_binding_res_list_scrub(struct list_head *head);
-extern int vmw_context_rebind_all(struct vmw_resource *ctx);
 extern struct list_head *vmw_context_binding_list(struct vmw_resource *ctx);
 extern struct vmw_cmdbuf_res_manager *
 vmw_context_res_man(struct vmw_resource *ctx);
+extern struct vmw_resource *vmw_context_cotable(struct vmw_resource *ctx,
+						SVGACOTableType cotable_type);
+extern struct list_head *vmw_context_binding_list(struct vmw_resource *ctx);
+struct vmw_ctx_binding_state;
+extern struct vmw_ctx_binding_state *
+vmw_context_binding_state(struct vmw_resource *ctx);
+extern void vmw_dx_context_scrub_cotables(struct vmw_resource *ctx,
+					  bool readback);
+
 /*
  * Surface management - vmwgfx_surface.c
  */
@@ -1066,6 +1041,7 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 			       bool for_scanout,
 			       uint32_t num_mip_levels,
 			       uint32_t multisample_count,
+			       uint32_t array_size,
 			       struct drm_vmw_size size,
 			       struct vmw_surface **srf_out);
 
@@ -1085,11 +1061,20 @@ extern int vmw_compat_shader_add(struct vmw_private *dev_priv,
 				 SVGA3dShaderType shader_type,
 				 size_t size,
 				 struct list_head *list);
-extern int vmw_compat_shader_remove(struct vmw_cmdbuf_res_manager *man,
+extern int vmw_shader_remove(struct vmw_cmdbuf_res_manager *man,
 			     u32 user_key, SVGA3dShaderType shader_type,
 			     struct list_head *list);
+extern int vmw_dx_shader_add(struct vmw_cmdbuf_res_manager *man,
+			     struct vmw_resource *ctx,
+			     u32 user_key,
+			     SVGA3dShaderType shader_type,
+			     struct list_head *list);
+extern void vmw_dx_shader_cotable_list_scrub(struct vmw_private *dev_priv,
+					     struct list_head *list,
+					     bool readback);
+
 extern struct vmw_resource *
-vmw_compat_shader_lookup(struct vmw_cmdbuf_res_manager *man,
+vmw_shader_lookup(struct vmw_cmdbuf_res_manager *man,
 		  u32 user_key, SVGA3dShaderType shader_type);
 
 /*
@@ -1114,8 +1099,20 @@ extern int vmw_cmdbuf_res_add(struct vmw_cmdbuf_res_manager *man,
 extern int vmw_cmdbuf_res_remove(struct vmw_cmdbuf_res_manager *man,
 				 enum vmw_cmdbuf_res_type res_type,
 				 u32 user_key,
-				 struct list_head *list);
+				 struct list_head *list,
+				 struct vmw_resource **res);
 
+/*
+ * COTable management - vmwgfx_cotable.c
+ */
+extern const SVGACOTableType vmw_cotable_scrub_order[];
+extern struct vmw_resource *vmw_cotable_alloc(struct vmw_private *dev_priv,
+					      struct vmw_resource *ctx,
+					      u32 type);
+extern int vmw_cotable_notify(struct vmw_resource *res, int id);
+extern int vmw_cotable_scrub(struct vmw_resource *res, bool readback);
+extern void vmw_cotable_add_resource(struct vmw_resource *ctx,
+				     struct list_head *head);
 
 /*
  * Command buffer managerment vmwgfx_cmdbuf.c

drivers/gpu/drm/vmwgfx/vmwgfx_fifo.c

@@ -29,6 +29,11 @@
 #include <drm/drmP.h>
 #include <drm/ttm/ttm_placement.h>
 
+struct vmw_temp_set_context {
+	SVGA3dCmdHeader header;
+	SVGA3dCmdDXTempSetContext body;
+};
+
 bool vmw_fifo_have_3d(struct vmw_private *dev_priv)
 {
 	u32 __iomem *fifo_mem = dev_priv->mmio_virt;
@@ -99,6 +104,7 @@ int vmw_fifo_init(struct vmw_private *dev_priv, struct vmw_fifo_state *fifo)
 	uint32_t max;
 	uint32_t min;
 
+	fifo->dx = false;
 	fifo->static_buffer_size = VMWGFX_FIFO_STATIC_SIZE;
 	fifo->static_buffer = vmalloc(fifo->static_buffer_size);
 	if (unlikely(fifo->static_buffer == NULL))
@@ -396,15 +402,20 @@ static void *vmw_local_fifo_reserve(struct vmw_private *dev_priv,
 	return NULL;
 }
 
-void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes)
+void *vmw_fifo_reserve_dx(struct vmw_private *dev_priv, uint32_t bytes,
+			  int ctx_id)
 {
 	void *ret;
 
 	if (dev_priv->cman)
 		ret = vmw_cmdbuf_reserve(dev_priv->cman, bytes,
-					 SVGA3D_INVALID_ID, false, NULL);
+					 ctx_id, false, NULL);
-	else
+	else if (ctx_id == SVGA3D_INVALID_ID)
 		ret = vmw_local_fifo_reserve(dev_priv, bytes);
+	else {
+		WARN_ON("Command buffer has not been allocated.\n");
+		ret = NULL;
+	}
 	if (IS_ERR_OR_NULL(ret)) {
 		DRM_ERROR("Fifo reserve failure of %u bytes.\n",
 			  (unsigned) bytes);
@@ -466,6 +477,10 @@ static void vmw_local_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes)
 	uint32_t min = ioread32(fifo_mem + SVGA_FIFO_MIN);
 	bool reserveable = fifo_state->capabilities & SVGA_FIFO_CAP_RESERVE;
 
+	if (fifo_state->dx)
+		bytes += sizeof(struct vmw_temp_set_context);
+
+	fifo_state->dx = false;
 	BUG_ON((bytes & 3) != 0);
 	BUG_ON(bytes > fifo_state->reserved_size);
 
@@ -518,7 +533,7 @@ void vmw_fifo_commit(struct vmw_private *dev_priv, uint32_t bytes)
  * @dev_priv: Pointer to device private structure.
  * @bytes: Number of bytes to commit.
  */
-static void vmw_fifo_commit_flush(struct vmw_private *dev_priv, uint32_t bytes)
+void vmw_fifo_commit_flush(struct vmw_private *dev_priv, uint32_t bytes)
 {
 	if (dev_priv->cman)
 		vmw_cmdbuf_commit(dev_priv->cman, bytes, NULL, true);
@@ -706,3 +721,8 @@ int vmw_fifo_emit_dummy_query(struct vmw_private *dev_priv,
 
 	return vmw_fifo_emit_dummy_legacy_query(dev_priv, cid);
 }
+
+void *vmw_fifo_reserve(struct vmw_private *dev_priv, uint32_t bytes)
+{
+	return vmw_fifo_reserve_dx(dev_priv, bytes, SVGA3D_INVALID_ID);
+}

drivers/gpu/drm/vmwgfx/vmwgfx_ioctl.c

@@ -110,6 +110,9 @@ int vmw_getparam_ioctl(struct drm_device *dev, void *data,
 		param->value =
 			(dev_priv->active_display_unit == vmw_du_screen_target);
 		break;
+	case DRM_VMW_PARAM_DX:
+		param->value = dev_priv->has_dx;
+		break;
 	default:
 		DRM_ERROR("Illegal vmwgfx get param request: %d\n",
 			  param->param);
@@ -193,8 +196,8 @@ int vmw_get_cap_3d_ioctl(struct drm_device *dev, void *data,
 		uint32_t *bounce32 = (uint32_t *) bounce;
 
 		num = size / sizeof(uint32_t);
-		if (num > SVGA3D_DEVCAP_MAX)
+		if (num > SVGA3D_DEVCAP_DX)
-			num = SVGA3D_DEVCAP_MAX;
+			num = SVGA3D_DEVCAP_DX;
 
 		spin_lock(&dev_priv->cap_lock);
 		for (i = 0; i < num; ++i) {

drivers/gpu/drm/vmwgfx/vmwgfx_kms.c

@@ -528,7 +528,11 @@ static int vmw_kms_new_framebuffer_surface(struct vmw_private *dev_priv,
 		return -EINVAL;
 	}
 
-	if (unlikely(format != surface->format)) {
+	/*
+	 * For DX, surface format validation is done when surface->scanout
+	 * is set.
+	 */
+	if (!dev_priv->has_dx && format != surface->format) {
 		DRM_ERROR("Invalid surface format for requested mode.\n");
 		return -EINVAL;
 	}
@@ -754,6 +758,7 @@ static int vmw_create_dmabuf_proxy(struct drm_device *dev,
 			true, /* can be a scanout buffer */
 			1, /* num of mip levels */
 			0,
+			0,
 			content_base_size,
 			srf_out);
 	if (ret) {
@@ -769,7 +774,7 @@ static int vmw_create_dmabuf_proxy(struct drm_device *dev,
 	vmw_dmabuf_unreference(&res->backup);
 	res->backup = vmw_dmabuf_reference(dmabuf_mob);
 	res->backup_offset = 0;
-	vmw_resource_unreserve(res, NULL, 0);
+	vmw_resource_unreserve(res, false, NULL, 0);
 	mutex_unlock(&res->dev_priv->cmdbuf_mutex);
 
 	return 0;
@@ -1869,7 +1874,7 @@ void vmw_kms_helper_buffer_finish(struct vmw_private *dev_priv,
 void vmw_kms_helper_resource_revert(struct vmw_resource *res)
 {
 	vmw_kms_helper_buffer_revert(res->backup);
-	vmw_resource_unreserve(res, NULL, 0);
+	vmw_resource_unreserve(res, false, NULL, 0);
 	mutex_unlock(&res->dev_priv->cmdbuf_mutex);
 }
 
@@ -1916,7 +1921,7 @@ int vmw_kms_helper_resource_prepare(struct vmw_resource *res,
 out_revert:
 	vmw_kms_helper_buffer_revert(res->backup);
 out_unreserve:
-	vmw_resource_unreserve(res, NULL, 0);
+	vmw_resource_unreserve(res, false, NULL, 0);
 out_unlock:
 	mutex_unlock(&res->dev_priv->cmdbuf_mutex);
 	return ret;
@@ -1937,7 +1942,7 @@ void vmw_kms_helper_resource_finish(struct vmw_resource *res,
 		vmw_kms_helper_buffer_finish(res->dev_priv, NULL, res->backup,
 					     out_fence, NULL);
 
-	vmw_resource_unreserve(res, NULL, 0);
+	vmw_resource_unreserve(res, false, NULL, 0);
 	mutex_unlock(&res->dev_priv->cmdbuf_mutex);
 }
 

drivers/gpu/drm/vmwgfx/vmwgfx_mob.c

@@ -67,9 +67,23 @@ struct vmw_mob {
  * @size:           Size of the table (page-aligned).
  * @page_table:     Pointer to a struct vmw_mob holding the page table.
  */
-struct vmw_otable {
+static const struct vmw_otable pre_dx_tables[] = {
-	unsigned long size;
+	{VMWGFX_NUM_MOB * SVGA3D_OTABLE_MOB_ENTRY_SIZE, NULL, true},
-	struct vmw_mob *page_table;
+	{VMWGFX_NUM_GB_SURFACE * SVGA3D_OTABLE_SURFACE_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_CONTEXT * SVGA3D_OTABLE_CONTEXT_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_SHADER * SVGA3D_OTABLE_SHADER_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_SCREEN_TARGET * SVGA3D_OTABLE_SCREEN_TARGET_ENTRY_SIZE,
+	 NULL, VMWGFX_ENABLE_SCREEN_TARGET_OTABLE}
+};
+
+static const struct vmw_otable dx_tables[] = {
+	{VMWGFX_NUM_MOB * SVGA3D_OTABLE_MOB_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_SURFACE * SVGA3D_OTABLE_SURFACE_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_CONTEXT * SVGA3D_OTABLE_CONTEXT_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_SHADER * SVGA3D_OTABLE_SHADER_ENTRY_SIZE, NULL, true},
+	{VMWGFX_NUM_GB_SCREEN_TARGET * SVGA3D_OTABLE_SCREEN_TARGET_ENTRY_SIZE,
+	 NULL, VMWGFX_ENABLE_SCREEN_TARGET_OTABLE},
+	{VMWGFX_NUM_DXCONTEXT * sizeof(SVGAOTableDXContextEntry), NULL, true},
 };
 
 static int vmw_mob_pt_populate(struct vmw_private *dev_priv,
@@ -92,6 +106,7 @@ static void vmw_mob_pt_setup(struct vmw_mob *mob,
  */
 static int vmw_setup_otable_base(struct vmw_private *dev_priv,
 				 SVGAOTableType type,
+				 struct ttm_buffer_object *otable_bo,
 				 unsigned long offset,
 				 struct vmw_otable *otable)
 {
@@ -106,7 +121,7 @@ static int vmw_setup_otable_base(struct vmw_private *dev_priv,
 
 	BUG_ON(otable->page_table != NULL);
 
-	vsgt = vmw_bo_sg_table(dev_priv->otable_bo);
+	vsgt = vmw_bo_sg_table(otable_bo);
 	vmw_piter_start(&iter, vsgt, offset >> PAGE_SHIFT);
 	WARN_ON(!vmw_piter_next(&iter));
 
@@ -218,47 +233,21 @@ static void vmw_takedown_otable_base(struct vmw_private *dev_priv,
 	otable->page_table = NULL;
 }
 
-/*
+
- * vmw_otables_setup - Set up guest backed memory object tables
+static int vmw_otable_batch_setup(struct vmw_private *dev_priv,
- *
+				  struct vmw_otable_batch *batch)
- * @dev_priv:       Pointer to a device private structure
- *
- * Takes care of the device guest backed surface
- * initialization, by setting up the guest backed memory object tables.
- * Returns 0 on success and various error codes on failure. A succesful return
- * means the object tables can be taken down using the vmw_otables_takedown
- * function.
- */
-int vmw_otables_setup(struct vmw_private *dev_priv)
 {
 	unsigned long offset;
 	unsigned long bo_size;
-	struct vmw_otable *otables;
+	struct vmw_otable *otables = batch->otables;
 	SVGAOTableType i;
 	int ret;
 
-	otables = kzalloc(SVGA_OTABLE_DX9_MAX * sizeof(*otables),
-			  GFP_KERNEL);
-	if (unlikely(otables == NULL)) {
-		DRM_ERROR("Failed to allocate space for otable "
-			  "metadata.\n");
-		return -ENOMEM;
-	}
-
-	otables[SVGA_OTABLE_MOB].size =
-		VMWGFX_NUM_MOB * SVGA3D_OTABLE_MOB_ENTRY_SIZE;
-	otables[SVGA_OTABLE_SURFACE].size =
-		VMWGFX_NUM_GB_SURFACE * SVGA3D_OTABLE_SURFACE_ENTRY_SIZE;
-	otables[SVGA_OTABLE_CONTEXT].size =
-		VMWGFX_NUM_GB_CONTEXT * SVGA3D_OTABLE_CONTEXT_ENTRY_SIZE;
-	otables[SVGA_OTABLE_SHADER].size =
-		VMWGFX_NUM_GB_SHADER * SVGA3D_OTABLE_SHADER_ENTRY_SIZE;
-	otables[SVGA_OTABLE_SCREENTARGET].size =
-		VMWGFX_NUM_GB_SCREEN_TARGET *
-		SVGA3D_OTABLE_SCREEN_TARGET_ENTRY_SIZE;
-
 	bo_size = 0;
-	for (i = 0; i < SVGA_OTABLE_DX9_MAX; ++i) {
+	for (i = 0; i < batch->num_otables; ++i) {
+		if (!otables[i].enabled)
+			continue;
+
 		otables[i].size =
 			(otables[i].size + PAGE_SIZE - 1) & PAGE_MASK;
 		bo_size += otables[i].size;
@@ -268,46 +257,114 @@ int vmw_otables_setup(struct vmw_private *dev_priv)
 			    ttm_bo_type_device,
 			    &vmw_sys_ne_placement,
 			    0, false, NULL,
-			    &dev_priv->otable_bo);
+			    &batch->otable_bo);
 
 	if (unlikely(ret != 0))
 		goto out_no_bo;
 
-	ret = ttm_bo_reserve(dev_priv->otable_bo, false, true, false, NULL);
+	ret = ttm_bo_reserve(batch->otable_bo, false, true, false, NULL);
 	BUG_ON(ret != 0);
-	ret = vmw_bo_driver.ttm_tt_populate(dev_priv->otable_bo->ttm);
+	ret = vmw_bo_driver.ttm_tt_populate(batch->otable_bo->ttm);
 	if (unlikely(ret != 0))
 		goto out_unreserve;
-	ret = vmw_bo_map_dma(dev_priv->otable_bo);
+	ret = vmw_bo_map_dma(batch->otable_bo);
 	if (unlikely(ret != 0))
 		goto out_unreserve;
 
-	ttm_bo_unreserve(dev_priv->otable_bo);
+	ttm_bo_unreserve(batch->otable_bo);
 
 	offset = 0;
-	for (i = 0; i < SVGA_OTABLE_DX9_MAX - VMW_OTABLE_SETUP_SUB; ++i) {
+	for (i = 0; i < batch->num_otables; ++i) {
-		ret = vmw_setup_otable_base(dev_priv, i, offset,
+		if (!batch->otables[i].enabled)
+			continue;
+
+		ret = vmw_setup_otable_base(dev_priv, i, batch->otable_bo,
+					    offset,
 					    &otables[i]);
 		if (unlikely(ret != 0))
 			goto out_no_setup;
 		offset += otables[i].size;
 	}
 
-	dev_priv->otables = otables;
 	return 0;
 
 out_unreserve:
-	ttm_bo_unreserve(dev_priv->otable_bo);
+	ttm_bo_unreserve(batch->otable_bo);
 out_no_setup:
-	for (i = 0; i < SVGA_OTABLE_DX9_MAX - VMW_OTABLE_SETUP_SUB; ++i)
+	for (i = 0; i < batch->num_otables; ++i) {
-		vmw_takedown_otable_base(dev_priv, i, &otables[i]);
+		if (batch->otables[i].enabled)
+			vmw_takedown_otable_base(dev_priv, i,
+						 &batch->otables[i]);
+	}
 
-	ttm_bo_unref(&dev_priv->otable_bo);
+	ttm_bo_unref(&batch->otable_bo);
 out_no_bo:
-	kfree(otables);
 	return ret;
 }
 
+/*
+ * vmw_otables_setup - Set up guest backed memory object tables
+ *
+ * @dev_priv:       Pointer to a device private structure
+ *
+ * Takes care of the device guest backed surface
+ * initialization, by setting up the guest backed memory object tables.
+ * Returns 0 on success and various error codes on failure. A successful return
+ * means the object tables can be taken down using the vmw_otables_takedown
+ * function.
+ */
+int vmw_otables_setup(struct vmw_private *dev_priv)
+{
+	struct vmw_otable **otables = &dev_priv->otable_batch.otables;
+	int ret;
+
+	if (dev_priv->has_dx) {
+		*otables = kmalloc(sizeof(dx_tables), GFP_KERNEL);
+		if (*otables == NULL)
+			return -ENOMEM;
+
+		memcpy(*otables, dx_tables, sizeof(dx_tables));
+		dev_priv->otable_batch.num_otables = ARRAY_SIZE(dx_tables);
+	} else {
+		*otables = kmalloc(sizeof(pre_dx_tables), GFP_KERNEL);
+		if (*otables == NULL)
+			return -ENOMEM;
+
+		memcpy(*otables, pre_dx_tables, sizeof(pre_dx_tables));
+		dev_priv->otable_batch.num_otables = ARRAY_SIZE(pre_dx_tables);
+	}
+
+	ret = vmw_otable_batch_setup(dev_priv, &dev_priv->otable_batch);
+	if (unlikely(ret != 0))
+		goto out_setup;
+
+	return 0;
+
+out_setup:
+	kfree(*otables);
+	return ret;
+}
+
+static void vmw_otable_batch_takedown(struct vmw_private *dev_priv,
+			       struct vmw_otable_batch *batch)
+{
+	SVGAOTableType i;
+	struct ttm_buffer_object *bo = batch->otable_bo;
+	int ret;
+
+	for (i = 0; i < batch->num_otables; ++i)
+		if (batch->otables[i].enabled)
+			vmw_takedown_otable_base(dev_priv, i,
+						 &batch->otables[i]);
+
+	ret = ttm_bo_reserve(bo, false, true, false, NULL);
+	BUG_ON(ret != 0);
+
+	vmw_fence_single_bo(bo, NULL);
+	ttm_bo_unreserve(bo);
+
+	ttm_bo_unref(&batch->otable_bo);
+}
 
 /*
  * vmw_otables_takedown - Take down guest backed memory object tables
@@ -318,26 +375,10 @@ int vmw_otables_setup(struct vmw_private *dev_priv)
  */
 void vmw_otables_takedown(struct vmw_private *dev_priv)
 {
-	SVGAOTableType i;
+	vmw_otable_batch_takedown(dev_priv, &dev_priv->otable_batch);
-	struct ttm_buffer_object *bo = dev_priv->otable_bo;
+	kfree(dev_priv->otable_batch.otables);
-	int ret;
-
-	for (i = 0; i < SVGA_OTABLE_DX9_MAX - VMW_OTABLE_SETUP_SUB; ++i)
-		vmw_takedown_otable_base(dev_priv, i,
-					 &dev_priv->otables[i]);
-
-	ret = ttm_bo_reserve(bo, false, true, false, NULL);
-	BUG_ON(ret != 0);
-
-	vmw_fence_single_bo(bo, NULL);
-	ttm_bo_unreserve(bo);
-
-	ttm_bo_unref(&dev_priv->otable_bo);
-	kfree(dev_priv->otables);
-	dev_priv->otables = NULL;
 }
 
-
 /*
  * vmw_mob_calculate_pt_pages - Calculate the number of page table pages
  * needed for a guest backed memory object.

drivers/gpu/drm/vmwgfx/vmwgfx_resource.c

@@ -31,6 +31,7 @@
 #include <drm/ttm/ttm_placement.h>
 #include <drm/drmP.h>
 #include "vmwgfx_resource_priv.h"
+#include "vmwgfx_binding.h"
 
 #define VMW_RES_EVICT_ERR_COUNT 10
 
@@ -144,10 +145,10 @@ static void vmw_resource_release(struct kref *kref)
 	}
 
 	if (likely(res->hw_destroy != NULL)) {
-		res->hw_destroy(res);
 		mutex_lock(&dev_priv->binding_mutex);
-		vmw_context_binding_res_list_kill(&res->binding_head);
+		vmw_binding_res_list_kill(&res->binding_head);
 		mutex_unlock(&dev_priv->binding_mutex);
+		res->hw_destroy(res);
 	}
 
 	id = res->id;
@@ -1149,14 +1150,16 @@ static int vmw_resource_do_validate(struct vmw_resource *res,
  * command submission.
  *
  * @res:               Pointer to the struct vmw_resource to unreserve.
+ * @switch_backup:     Backup buffer has been switched.
  * @new_backup:        Pointer to new backup buffer if command submission
- *                     switched.
+ *                     switched. May be NULL.
- * @new_backup_offset: New backup offset if @new_backup is !NULL.
+ * @new_backup_offset: New backup offset if @switch_backup is true.
  *
  * Currently unreserving a resource means putting it back on the device's
  * resource lru list, so that it can be evicted if necessary.
  */
 void vmw_resource_unreserve(struct vmw_resource *res,
+			    bool switch_backup,
 			    struct vmw_dma_buffer *new_backup,
 			    unsigned long new_backup_offset)
 {
@@ -1165,19 +1168,22 @@ void vmw_resource_unreserve(struct vmw_resource *res,
 	if (!list_empty(&res->lru_head))
 		return;
 
-	if (new_backup && new_backup != res->backup) {
+	if (switch_backup && new_backup != res->backup) {
-
 		if (res->backup) {
 			lockdep_assert_held(&res->backup->base.resv->lock.base);
 			list_del_init(&res->mob_head);
 			vmw_dmabuf_unreference(&res->backup);
 		}
 
+		if (new_backup) {
 			res->backup = vmw_dmabuf_reference(new_backup);
 			lockdep_assert_held(&new_backup->base.resv->lock.base);
 			list_add_tail(&res->mob_head, &new_backup->res_list);
+		} else {
+			res->backup = NULL;
 		}
-	if (new_backup)
+	}
+	if (switch_backup)
 		res->backup_offset = new_backup_offset;
 
 	if (!res->func->may_evict || res->id == -1 || res->pin_count)
@@ -1269,9 +1275,13 @@ int vmw_resource_reserve(struct vmw_resource *res, bool interruptible,
 	if (res->func->needs_backup && res->backup == NULL &&
 	    !no_backup) {
 		ret = vmw_resource_buf_alloc(res, interruptible);
-		if (unlikely(ret != 0))
+		if (unlikely(ret != 0)) {
+			DRM_ERROR("Failed to allocate a backup buffer "
+				  "of size %lu. bytes\n",
+				  (unsigned long) res->backup_size);
 			return ret;
 		}
+	}
 
 	return 0;
 }
@@ -1354,7 +1364,7 @@ int vmw_resource_validate(struct vmw_resource *res)
 	struct ttm_validate_buffer val_buf;
 	unsigned err_count = 0;
 
-	if (likely(!res->func->may_evict))
+	if (!res->func->create)
 		return 0;
 
 	val_buf.bo = NULL;
@@ -1624,7 +1634,7 @@ int vmw_resource_pin(struct vmw_resource *res, bool interruptible)
 	res->pin_count++;
 
 out_no_validate:
-	vmw_resource_unreserve(res, NULL, 0UL);
+	vmw_resource_unreserve(res, false, NULL, 0UL);
 out_no_reserve:
 	mutex_unlock(&dev_priv->cmdbuf_mutex);
 	ttm_write_unlock(&dev_priv->reservation_sem);
@@ -1660,8 +1670,18 @@ void vmw_resource_unpin(struct vmw_resource *res)
 		ttm_bo_unreserve(&vbo->base);
 	}
 
-	vmw_resource_unreserve(res, NULL, 0UL);
+	vmw_resource_unreserve(res, false, NULL, 0UL);
 
 	mutex_unlock(&dev_priv->cmdbuf_mutex);
 	ttm_read_unlock(&dev_priv->reservation_sem);
 }
+
+/**
+ * vmw_res_type - Return the resource type
+ *
+ * @res: Pointer to the resource
+ */
+enum vmw_res_type vmw_res_type(const struct vmw_resource *res)
+{
+	return res->func->res_type;
+}

drivers/gpu/drm/vmwgfx/vmwgfx_resource_priv.h

@@ -30,6 +30,12 @@
 
 #include "vmwgfx_drv.h"
 
+enum vmw_cmdbuf_res_state {
+	VMW_CMDBUF_RES_COMMITTED,
+	VMW_CMDBUF_RES_ADD,
+	VMW_CMDBUF_RES_DEL
+};
+
 /**
  * struct vmw_user_resource_conv - Identify a derived user-exported resource
  * type and provide a function to convert its ttm_base_object pointer to
@@ -55,8 +61,10 @@ struct vmw_user_resource_conv {
  * @bind:              Bind a hardware resource to persistent buffer storage.
  * @unbind:            Unbind a hardware resource from persistent
  *                     buffer storage.
+ * @commit_notify:     If the resource is a command buffer managed resource,
+ *                     callback to notify that a define or remove command
+ *                     has been committed to the device.
  */
-
 struct vmw_res_func {
 	enum vmw_res_type res_type;
 	bool needs_backup;
@@ -71,6 +79,8 @@ struct vmw_res_func {
 	int (*unbind) (struct vmw_resource *res,
 		       bool readback,
 		       struct ttm_validate_buffer *val_buf);
+	void (*commit_notify)(struct vmw_resource *res,
+			      enum vmw_cmdbuf_res_state state);
 };
 
 int vmw_resource_alloc_id(struct vmw_resource *res);

drivers/gpu/drm/vmwgfx/vmwgfx_shader.c

@@ -27,12 +27,15 @@
 
 #include "vmwgfx_drv.h"
 #include "vmwgfx_resource_priv.h"
+#include "vmwgfx_binding.h"
 #include "ttm/ttm_placement.h"
 
 struct vmw_shader {
 	struct vmw_resource res;
 	SVGA3dShaderType type;
 	uint32_t size;
+	uint8_t num_input_sig;
+	uint8_t num_output_sig;
 };
 
 struct vmw_user_shader {
@@ -40,8 +43,18 @@ struct vmw_user_shader {
 	struct vmw_shader shader;
 };
 
+struct vmw_dx_shader {
+	struct vmw_resource res;
+	struct vmw_resource *ctx;
+	struct vmw_resource *cotable;
+	u32 id;
+	bool committed;
+	struct list_head cotable_head;
+};
+
 static uint64_t vmw_user_shader_size;
 static uint64_t vmw_shader_size;
+static size_t vmw_shader_dx_size;
 
 static void vmw_user_shader_free(struct vmw_resource *res);
 static struct vmw_resource *
@@ -55,6 +68,18 @@ static int vmw_gb_shader_unbind(struct vmw_resource *res,
 				 struct ttm_validate_buffer *val_buf);
 static int vmw_gb_shader_destroy(struct vmw_resource *res);
 
+static int vmw_dx_shader_create(struct vmw_resource *res);
+static int vmw_dx_shader_bind(struct vmw_resource *res,
+			       struct ttm_validate_buffer *val_buf);
+static int vmw_dx_shader_unbind(struct vmw_resource *res,
+				 bool readback,
+				 struct ttm_validate_buffer *val_buf);
+static void vmw_dx_shader_commit_notify(struct vmw_resource *res,
+					enum vmw_cmdbuf_res_state state);
+static bool vmw_shader_id_ok(u32 user_key, SVGA3dShaderType shader_type);
+static u32 vmw_shader_key(u32 user_key, SVGA3dShaderType shader_type);
+static uint64_t vmw_user_shader_size;
+
 static const struct vmw_user_resource_conv user_shader_conv = {
 	.object_type = VMW_RES_SHADER,
 	.base_obj_to_res = vmw_user_shader_base_to_res,
@@ -77,6 +102,24 @@ static const struct vmw_res_func vmw_gb_shader_func = {
 	.unbind = vmw_gb_shader_unbind
 };
 
+static const struct vmw_res_func vmw_dx_shader_func = {
+	.res_type = vmw_res_shader,
+	.needs_backup = true,
+	.may_evict = false,
+	.type_name = "dx shaders",
+	.backup_placement = &vmw_mob_placement,
+	.create = vmw_dx_shader_create,
+	/*
+	 * The destroy callback is only called with a committed resource on
+	 * context destroy, in which case we destroy the cotable anyway,
+	 * so there's no need to destroy DX shaders separately.
+	 */
+	.destroy = NULL,
+	.bind = vmw_dx_shader_bind,
+	.unbind = vmw_dx_shader_unbind,
+	.commit_notify = vmw_dx_shader_commit_notify,
+};
+
 /**
  * Shader management:
  */
@@ -87,25 +130,42 @@ vmw_res_to_shader(struct vmw_resource *res)
 	return container_of(res, struct vmw_shader, res);
 }
 
+/**
+ * vmw_res_to_dx_shader - typecast a struct vmw_resource to a
+ * struct vmw_dx_shader
+ *
+ * @res: Pointer to the struct vmw_resource.
+ */
+static inline struct vmw_dx_shader *
+vmw_res_to_dx_shader(struct vmw_resource *res)
+{
+	return container_of(res, struct vmw_dx_shader, res);
+}
+
 static void vmw_hw_shader_destroy(struct vmw_resource *res)
 {
-	(void) vmw_gb_shader_destroy(res);
+	if (likely(res->func->destroy))
+		(void) res->func->destroy(res);
+	else
+		res->id = -1;
 }
 
+
 static int vmw_gb_shader_init(struct vmw_private *dev_priv,
 			      struct vmw_resource *res,
 			      uint32_t size,
 			      uint64_t offset,
 			      SVGA3dShaderType type,
+			      uint8_t num_input_sig,
+			      uint8_t num_output_sig,
 			      struct vmw_dma_buffer *byte_code,
 			      void (*res_free) (struct vmw_resource *res))
 {
 	struct vmw_shader *shader = vmw_res_to_shader(res);
 	int ret;
 
-	ret = vmw_resource_init(dev_priv, res, true,
+	ret = vmw_resource_init(dev_priv, res, true, res_free,
-				res_free, &vmw_gb_shader_func);
+				&vmw_gb_shader_func);
-
 
 	if (unlikely(ret != 0)) {
 		if (res_free)
@@ -122,11 +182,17 @@ static int vmw_gb_shader_init(struct vmw_private *dev_priv,
 	}
 	shader->size = size;
 	shader->type = type;
+	shader->num_input_sig = num_input_sig;
+	shader->num_output_sig = num_output_sig;
 
 	vmw_resource_activate(res, vmw_hw_shader_destroy);
 	return 0;
 }
 
+/*
+ * GB shader code:
+ */
+
 static int vmw_gb_shader_create(struct vmw_resource *res)
 {
 	struct vmw_private *dev_priv = res->dev_priv;
@@ -259,7 +325,7 @@ static int vmw_gb_shader_destroy(struct vmw_resource *res)
 		return 0;
 
 	mutex_lock(&dev_priv->binding_mutex);
-	vmw_context_binding_res_list_scrub(&res->binding_head);
+	vmw_binding_res_list_scrub(&res->binding_head);
 
 	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
 	if (unlikely(cmd == NULL)) {
@@ -280,6 +346,321 @@ static int vmw_gb_shader_destroy(struct vmw_resource *res)
 	return 0;
 }
 
+/*
+ * DX shader code:
+ */
+
+/**
+ * vmw_dx_shader_commit_notify - Notify that a shader operation has been
+ * committed to hardware from a user-supplied command stream.
+ *
+ * @res: Pointer to the shader resource.
+ * @state: Indicating whether a creation or removal has been committed.
+ *
+ */
+static void vmw_dx_shader_commit_notify(struct vmw_resource *res,
+					enum vmw_cmdbuf_res_state state)
+{
+	struct vmw_dx_shader *shader = vmw_res_to_dx_shader(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	if (state == VMW_CMDBUF_RES_ADD) {
+		mutex_lock(&dev_priv->binding_mutex);
+		vmw_cotable_add_resource(shader->cotable,
+					 &shader->cotable_head);
+		shader->committed = true;
+		res->id = shader->id;
+		mutex_unlock(&dev_priv->binding_mutex);
+	} else {
+		mutex_lock(&dev_priv->binding_mutex);
+		list_del_init(&shader->cotable_head);
+		shader->committed = false;
+		res->id = -1;
+		mutex_unlock(&dev_priv->binding_mutex);
+	}
+}
+
+/**
+ * vmw_dx_shader_unscrub - Have the device reattach a MOB to a DX shader.
+ *
+ * @res: The shader resource
+ *
+ * This function reverts a scrub operation.
+ */
+static int vmw_dx_shader_unscrub(struct vmw_resource *res)
+{
+	struct vmw_dx_shader *shader = vmw_res_to_dx_shader(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXBindShader body;
+	} *cmd;
+
+	if (!list_empty(&shader->cotable_head) || !shader->committed)
+		return 0;
+
+	cmd = vmw_fifo_reserve_dx(dev_priv, sizeof(*cmd),
+				  shader->ctx->id);
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for shader "
+			  "scrubbing.\n");
+		return -ENOMEM;
+	}
+
+	cmd->header.id = SVGA_3D_CMD_DX_BIND_SHADER;
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.cid = shader->ctx->id;
+	cmd->body.shid = shader->id;
+	cmd->body.mobid = res->backup->base.mem.start;
+	cmd->body.offsetInBytes = res->backup_offset;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+
+	vmw_cotable_add_resource(shader->cotable, &shader->cotable_head);
+
+	return 0;
+}
+
+/**
+ * vmw_dx_shader_create - The DX shader create callback
+ *
+ * @res: The DX shader resource
+ *
+ * The create callback is called as part of resource validation and
+ * makes sure that we unscrub the shader if it's previously been scrubbed.
+ */
+static int vmw_dx_shader_create(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct vmw_dx_shader *shader = vmw_res_to_dx_shader(res);
+	int ret = 0;
+
+	WARN_ON_ONCE(!shader->committed);
+
+	if (!list_empty(&res->mob_head)) {
+		mutex_lock(&dev_priv->binding_mutex);
+		ret = vmw_dx_shader_unscrub(res);
+		mutex_unlock(&dev_priv->binding_mutex);
+	}
+
+	res->id = shader->id;
+	return ret;
+}
+
+/**
+ * vmw_dx_shader_bind - The DX shader bind callback
+ *
+ * @res: The DX shader resource
+ * @val_buf: Pointer to the validate buffer.
+ *
+ */
+static int vmw_dx_shader_bind(struct vmw_resource *res,
+			      struct ttm_validate_buffer *val_buf)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct ttm_buffer_object *bo = val_buf->bo;
+
+	BUG_ON(bo->mem.mem_type != VMW_PL_MOB);
+	mutex_lock(&dev_priv->binding_mutex);
+	vmw_dx_shader_unscrub(res);
+	mutex_unlock(&dev_priv->binding_mutex);
+
+	return 0;
+}
+
+/**
+ * vmw_dx_shader_scrub - Have the device unbind a MOB from a DX shader.
+ *
+ * @res: The shader resource
+ *
+ * This function unbinds a MOB from the DX shader without requiring the
+ * MOB dma_buffer to be reserved. The driver still considers the MOB bound.
+ * However, once the driver eventually decides to unbind the MOB, it doesn't
+ * need to access the context.
+ */
+static int vmw_dx_shader_scrub(struct vmw_resource *res)
+{
+	struct vmw_dx_shader *shader = vmw_res_to_dx_shader(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDXBindShader body;
+	} *cmd;
+
+	if (list_empty(&shader->cotable_head))
+		return 0;
+
+	WARN_ON_ONCE(!shader->committed);
+	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
+	if (unlikely(cmd == NULL)) {
+		DRM_ERROR("Failed reserving FIFO space for shader "
+			  "scrubbing.\n");
+		return -ENOMEM;
+	}
+
+	cmd->header.id = SVGA_3D_CMD_DX_BIND_SHADER;
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.cid = shader->ctx->id;
+	cmd->body.shid = res->id;
+	cmd->body.mobid = SVGA3D_INVALID_ID;
+	cmd->body.offsetInBytes = 0;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+	res->id = -1;
+	list_del_init(&shader->cotable_head);
+
+	return 0;
+}
+
+/**
+ * vmw_dx_shader_unbind - The dx shader unbind callback.
+ *
+ * @res: The shader resource
+ * @readback: Whether this is a readback unbind. Currently unused.
+ * @val_buf: MOB buffer information.
+ */
+static int vmw_dx_shader_unbind(struct vmw_resource *res,
+				bool readback,
+				struct ttm_validate_buffer *val_buf)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct vmw_fence_obj *fence;
+	int ret;
+
+	BUG_ON(res->backup->base.mem.mem_type != VMW_PL_MOB);
+
+	mutex_lock(&dev_priv->binding_mutex);
+	ret = vmw_dx_shader_scrub(res);
+	mutex_unlock(&dev_priv->binding_mutex);
+
+	if (ret)
+		return ret;
+
+	(void) vmw_execbuf_fence_commands(NULL, dev_priv,
+					  &fence, NULL);
+	vmw_fence_single_bo(val_buf->bo, fence);
+
+	if (likely(fence != NULL))
+		vmw_fence_obj_unreference(&fence);
+
+	return 0;
+}
+
+/**
+ * vmw_dx_shader_cotable_list_scrub - The cotable unbind_func callback for
+ * DX shaders.
+ *
+ * @dev_priv: Pointer to device private structure.
+ * @list: The list of cotable resources.
+ * @readback: Whether the call was part of a readback unbind.
+ *
+ * Scrubs all shader MOBs so that any subsequent shader unbind or shader
+ * destroy operation won't need to swap in the context.
+ */
+void vmw_dx_shader_cotable_list_scrub(struct vmw_private *dev_priv,
+				      struct list_head *list,
+				      bool readback)
+{
+	struct vmw_dx_shader *entry, *next;
+
+	WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex));
+
+	list_for_each_entry_safe(entry, next, list, cotable_head) {
+		WARN_ON(vmw_dx_shader_scrub(&entry->res));
+		if (!readback)
+			entry->committed = false;
+	}
+}
+
+/**
+ * vmw_dx_shader_res_free - The DX shader free callback
+ *
+ * @res: The shader resource
+ *
+ * Frees the DX shader resource and updates memory accounting.
+ */
+static void vmw_dx_shader_res_free(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct vmw_dx_shader *shader = vmw_res_to_dx_shader(res);
+
+	vmw_resource_unreference(&shader->cotable);
+	kfree(shader);
+	ttm_mem_global_free(vmw_mem_glob(dev_priv), vmw_shader_dx_size);
+}
+
+/**
+ * vmw_dx_shader_add - Add a shader resource as a command buffer managed
+ * resource.
+ *
+ * @man: The command buffer resource manager.
+ * @ctx: Pointer to the context resource.
+ * @user_key: The id used for this shader.
+ * @shader_type: The shader type.
+ * @list: The list of staged command buffer managed resources.
+ */
+int vmw_dx_shader_add(struct vmw_cmdbuf_res_manager *man,
+		      struct vmw_resource *ctx,
+		      u32 user_key,
+		      SVGA3dShaderType shader_type,
+		      struct list_head *list)
+{
+	struct vmw_dx_shader *shader;
+	struct vmw_resource *res;
+	struct vmw_private *dev_priv = ctx->dev_priv;
+	int ret;
+
+	if (!vmw_shader_dx_size)
+		vmw_shader_dx_size = ttm_round_pot(sizeof(*shader));
+
+	if (!vmw_shader_id_ok(user_key, shader_type))
+		return -EINVAL;
+
+	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), vmw_shader_dx_size,
+				   false, true);
+	if (ret) {
+		if (ret != -ERESTARTSYS)
+			DRM_ERROR("Out of graphics memory for shader "
+				  "creation.\n");
+		return ret;
+	}
+
+	shader = kmalloc(sizeof(*shader), GFP_KERNEL);
+	if (!shader) {
+		ttm_mem_global_free(vmw_mem_glob(dev_priv), vmw_shader_dx_size);
+		return -ENOMEM;
+	}
+
+	res = &shader->res;
+	shader->ctx = ctx;
+	shader->cotable = vmw_context_cotable(ctx, SVGA_COTABLE_DXSHADER);
+	shader->id = user_key;
+	shader->committed = false;
+	INIT_LIST_HEAD(&shader->cotable_head);
+	ret = vmw_resource_init(dev_priv, res, true,
+				vmw_dx_shader_res_free, &vmw_dx_shader_func);
+	if (ret)
+		goto out_resource_init;
+
+	/*
+	 * The user_key name-space is not per shader type for DX shaders,
+	 * so when hashing, use a single zero shader type.
+	 */
+	ret = vmw_cmdbuf_res_add(man, vmw_cmdbuf_res_shader,
+				 vmw_shader_key(user_key, 0),
+				 res, list);
+	if (ret)
+		goto out_resource_init;
+
+	res->id = shader->id;
+	vmw_resource_activate(res, vmw_hw_shader_destroy);
+
+out_resource_init:
+	vmw_resource_unreference(&res);
+
+	return ret;
+}
+
+
+
 /**
  * User-space shader management:
  */
@@ -341,6 +722,8 @@ static int vmw_user_shader_alloc(struct vmw_private *dev_priv,
 				 size_t shader_size,
 				 size_t offset,
 				 SVGA3dShaderType shader_type,
+				 uint8_t num_input_sig,
+				 uint8_t num_output_sig,
 				 struct ttm_object_file *tfile,
 				 u32 *handle)
 {
@@ -383,7 +766,8 @@ static int vmw_user_shader_alloc(struct vmw_private *dev_priv,
 	 */
 
 	ret = vmw_gb_shader_init(dev_priv, res, shader_size,
-				 offset, shader_type, buffer,
+				 offset, shader_type, num_input_sig,
+				 num_output_sig, buffer,
 				 vmw_user_shader_free);
 	if (unlikely(ret != 0))
 		goto out;
@@ -449,7 +833,7 @@ static struct vmw_resource *vmw_shader_alloc(struct vmw_private *dev_priv,
 	 * From here on, the destructor takes over resource freeing.
 	 */
 	ret = vmw_gb_shader_init(dev_priv, res, shader_size,
-				 offset, shader_type, buffer,
+				 offset, shader_type, 0, 0, buffer,
 				 vmw_shader_free);
 
 out_err:
@@ -457,19 +841,20 @@ static struct vmw_resource *vmw_shader_alloc(struct vmw_private *dev_priv,
 }
 
 
-int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
+static int vmw_shader_define(struct drm_device *dev, struct drm_file *file_priv,
-			     struct drm_file *file_priv)
+			     enum drm_vmw_shader_type shader_type_drm,
+			     u32 buffer_handle, size_t size, size_t offset,
+			     uint8_t num_input_sig, uint8_t num_output_sig,
+			     uint32_t *shader_handle)
 {
 	struct vmw_private *dev_priv = vmw_priv(dev);
-	struct drm_vmw_shader_create_arg *arg =
-		(struct drm_vmw_shader_create_arg *)data;
 	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
 	struct vmw_dma_buffer *buffer = NULL;
 	SVGA3dShaderType shader_type;
 	int ret;
 
-	if (arg->buffer_handle != SVGA3D_INVALID_ID) {
+	if (buffer_handle != SVGA3D_INVALID_ID) {
-		ret = vmw_user_dmabuf_lookup(tfile, arg->buffer_handle,
+		ret = vmw_user_dmabuf_lookup(tfile, buffer_handle,
 					     &buffer);
 		if (unlikely(ret != 0)) {
 			DRM_ERROR("Could not find buffer for shader "
@@ -478,23 +863,20 @@ int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
 		}
 
 		if ((u64)buffer->base.num_pages * PAGE_SIZE <
-		    (u64)arg->size + (u64)arg->offset) {
+		    (u64)size + (u64)offset) {
 			DRM_ERROR("Illegal buffer- or shader size.\n");
 			ret = -EINVAL;
 			goto out_bad_arg;
 		}
 	}
 
-	switch (arg->shader_type) {
+	switch (shader_type_drm) {
 	case drm_vmw_shader_type_vs:
 		shader_type = SVGA3D_SHADERTYPE_VS;
 		break;
 	case drm_vmw_shader_type_ps:
 		shader_type = SVGA3D_SHADERTYPE_PS;
 		break;
-	case drm_vmw_shader_type_gs:
-		shader_type = SVGA3D_SHADERTYPE_GS;
-		break;
 	default:
 		DRM_ERROR("Illegal shader type.\n");
 		ret = -EINVAL;
@@ -505,8 +887,9 @@ int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
 	if (unlikely(ret != 0))
 		goto out_bad_arg;
 
-	ret = vmw_user_shader_alloc(dev_priv, buffer, arg->size, arg->offset,
+	ret = vmw_user_shader_alloc(dev_priv, buffer, size, offset,
-				    shader_type, tfile, &arg->shader_handle);
+				    shader_type, num_input_sig,
+				    num_output_sig, tfile, shader_handle);
 
 	ttm_read_unlock(&dev_priv->reservation_sem);
 out_bad_arg:
@@ -515,7 +898,7 @@ int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
 }
 
 /**
- * vmw_compat_shader_id_ok - Check whether a compat shader user key and
+ * vmw_shader_id_ok - Check whether a compat shader user key and
  * shader type are within valid bounds.
  *
  * @user_key: User space id of the shader.
@@ -523,13 +906,13 @@ int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
  *
  * Returns true if valid false if not.
  */
-static bool vmw_compat_shader_id_ok(u32 user_key, SVGA3dShaderType shader_type)
+static bool vmw_shader_id_ok(u32 user_key, SVGA3dShaderType shader_type)
 {
 	return user_key <= ((1 << 20) - 1) && (unsigned) shader_type < 16;
 }
 
 /**
- * vmw_compat_shader_key - Compute a hash key suitable for a compat shader.
+ * vmw_shader_key - Compute a hash key suitable for a compat shader.
  *
  * @user_key: User space id of the shader.
  * @shader_type: Shader type.
@@ -537,13 +920,13 @@ static bool vmw_compat_shader_id_ok(u32 user_key, SVGA3dShaderType shader_type)
  * Returns a hash key suitable for a command buffer managed resource
  * manager hash table.
  */
-static u32 vmw_compat_shader_key(u32 user_key, SVGA3dShaderType shader_type)
+static u32 vmw_shader_key(u32 user_key, SVGA3dShaderType shader_type)
 {
 	return user_key | (shader_type << 20);
 }
 
 /**
- * vmw_compat_shader_remove - Stage a compat shader for removal.
+ * vmw_shader_remove - Stage a compat shader for removal.
  *
  * @man: Pointer to the compat shader manager identifying the shader namespace.
  * @user_key: The key that is used to identify the shader. The key is
@@ -551,17 +934,18 @@ static u32 vmw_compat_shader_key(u32 user_key, SVGA3dShaderType shader_type)
  * @shader_type: Shader type.
  * @list: Caller's list of staged command buffer resource actions.
  */
-int vmw_compat_shader_remove(struct vmw_cmdbuf_res_manager *man,
+int vmw_shader_remove(struct vmw_cmdbuf_res_manager *man,
 		      u32 user_key, SVGA3dShaderType shader_type,
 		      struct list_head *list)
 {
-	if (!vmw_compat_shader_id_ok(user_key, shader_type))
+	struct vmw_resource *dummy;
+
+	if (!vmw_shader_id_ok(user_key, shader_type))
 		return -EINVAL;
 
-	return vmw_cmdbuf_res_remove(man, vmw_cmdbuf_res_compat_shader,
+	return vmw_cmdbuf_res_remove(man, vmw_cmdbuf_res_shader,
-				     vmw_compat_shader_key(user_key,
+				     vmw_shader_key(user_key, shader_type),
-							   shader_type),
+				     list, &dummy);
-				     list);
 }
 
 /**
@@ -591,7 +975,7 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv,
 	int ret;
 	struct vmw_resource *res;
 
-	if (!vmw_compat_shader_id_ok(user_key, shader_type))
+	if (!vmw_shader_id_ok(user_key, shader_type))
 		return -EINVAL;
 
 	/* Allocate and pin a DMA buffer */
@@ -628,8 +1012,8 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv,
 	if (unlikely(ret != 0))
 		goto no_reserve;
 
-	ret = vmw_cmdbuf_res_add(man, vmw_cmdbuf_res_compat_shader,
+	ret = vmw_cmdbuf_res_add(man, vmw_cmdbuf_res_shader,
-				 vmw_compat_shader_key(user_key, shader_type),
+				 vmw_shader_key(user_key, shader_type),
 				 res, list);
 	vmw_resource_unreference(&res);
 no_reserve:
@@ -639,7 +1023,7 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv,
 }
 
 /**
- * vmw_compat_shader_lookup - Look up a compat shader
+ * vmw_shader_lookup - Look up a compat shader
  *
  * @man: Pointer to the command buffer managed resource manager identifying
  * the shader namespace.
@@ -650,14 +1034,26 @@ int vmw_compat_shader_add(struct vmw_private *dev_priv,
  * found. An error pointer otherwise.
  */
 struct vmw_resource *
-vmw_compat_shader_lookup(struct vmw_cmdbuf_res_manager *man,
+vmw_shader_lookup(struct vmw_cmdbuf_res_manager *man,
 		  u32 user_key,
 		  SVGA3dShaderType shader_type)
 {
-	if (!vmw_compat_shader_id_ok(user_key, shader_type))
+	if (!vmw_shader_id_ok(user_key, shader_type))
 		return ERR_PTR(-EINVAL);
 
-	return vmw_cmdbuf_res_lookup(man, vmw_cmdbuf_res_compat_shader,
+	return vmw_cmdbuf_res_lookup(man, vmw_cmdbuf_res_shader,
-				     vmw_compat_shader_key(user_key,
+				     vmw_shader_key(user_key, shader_type));
-							   shader_type));
+}
+
+int vmw_shader_define_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *file_priv)
+{
+	struct drm_vmw_shader_create_arg *arg =
+		(struct drm_vmw_shader_create_arg *)data;
+
+	return vmw_shader_define(dev, file_priv, arg->shader_type,
+				 arg->buffer_handle,
+				 arg->size, arg->offset,
+				 0, 0,
+				 &arg->shader_handle);
 }

drivers/gpu/drm/vmwgfx/vmwgfx_so.c

@@ -0,0 +1,555 @@
+/**************************************************************************
+ * Copyright © 2014 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#include "vmwgfx_drv.h"
+#include "vmwgfx_resource_priv.h"
+#include "vmwgfx_so.h"
+#include "vmwgfx_binding.h"
+
+/*
+ * The currently only reason we need to keep track of views is that if we
+ * destroy a hardware surface, all views pointing to it must also be destroyed,
+ * otherwise the device will error.
+ * So in particuar if a surface is evicted, we must destroy all views pointing
+ * to it, and all context bindings of that view. Similarly we must restore
+ * the view bindings, views and surfaces pointed to by the views when a
+ * context is referenced in the command stream.
+ */
+
+/**
+ * struct vmw_view - view metadata
+ *
+ * @res: The struct vmw_resource we derive from
+ * @ctx: Non-refcounted pointer to the context this view belongs to.
+ * @srf: Refcounted pointer to the surface pointed to by this view.
+ * @cotable: Refcounted pointer to the cotable holding this view.
+ * @srf_head: List head for the surface-to-view list.
+ * @cotable_head: List head for the cotable-to_view list.
+ * @view_type: View type.
+ * @view_id: User-space per context view id. Currently used also as per
+ * context device view id.
+ * @cmd_size: Size of the SVGA3D define view command that we've copied from the
+ * command stream.
+ * @committed: Whether the view is actually created or pending creation at the
+ * device level.
+ * @cmd: The SVGA3D define view command copied from the command stream.
+ */
+struct vmw_view {
+	struct rcu_head rcu;
+	struct vmw_resource res;
+	struct vmw_resource *ctx;      /* Immutable */
+	struct vmw_resource *srf;      /* Immutable */
+	struct vmw_resource *cotable;  /* Immutable */
+	struct list_head srf_head;     /* Protected by binding_mutex */
+	struct list_head cotable_head; /* Protected by binding_mutex */
+	unsigned view_type;            /* Immutable */
+	unsigned view_id;              /* Immutable */
+	u32 cmd_size;                  /* Immutable */
+	bool committed;                /* Protected by binding_mutex */
+	u32 cmd[1];                    /* Immutable */
+};
+
+static int vmw_view_create(struct vmw_resource *res);
+static int vmw_view_destroy(struct vmw_resource *res);
+static void vmw_hw_view_destroy(struct vmw_resource *res);
+static void vmw_view_commit_notify(struct vmw_resource *res,
+				   enum vmw_cmdbuf_res_state state);
+
+static const struct vmw_res_func vmw_view_func = {
+	.res_type = vmw_res_view,
+	.needs_backup = false,
+	.may_evict = false,
+	.type_name = "DX view",
+	.backup_placement = NULL,
+	.create = vmw_view_create,
+	.commit_notify = vmw_view_commit_notify,
+};
+
+/**
+ * struct vmw_view - view define command body stub
+ *
+ * @view_id: The device id of the view being defined
+ * @sid: The surface id of the view being defined
+ *
+ * This generic struct is used by the code to change @view_id and @sid of a
+ * saved view define command.
+ */
+struct vmw_view_define {
+	uint32 view_id;
+	uint32 sid;
+};
+
+/**
+ * vmw_view - Convert a struct vmw_resource to a struct vmw_view
+ *
+ * @res: Pointer to the resource to convert.
+ *
+ * Returns a pointer to a struct vmw_view.
+ */
+static struct vmw_view *vmw_view(struct vmw_resource *res)
+{
+	return container_of(res, struct vmw_view, res);
+}
+
+/**
+ * vmw_view_commit_notify - Notify that a view operation has been committed to
+ * hardware from a user-supplied command stream.
+ *
+ * @res: Pointer to the view resource.
+ * @state: Indicating whether a creation or removal has been committed.
+ *
+ */
+static void vmw_view_commit_notify(struct vmw_resource *res,
+				   enum vmw_cmdbuf_res_state state)
+{
+	struct vmw_view *view = vmw_view(res);
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	mutex_lock(&dev_priv->binding_mutex);
+	if (state == VMW_CMDBUF_RES_ADD) {
+		struct vmw_surface *srf = vmw_res_to_srf(view->srf);
+
+		list_add_tail(&view->srf_head, &srf->view_list);
+		vmw_cotable_add_resource(view->cotable, &view->cotable_head);
+		view->committed = true;
+		res->id = view->view_id;
+
+	} else {
+		list_del_init(&view->cotable_head);
+		list_del_init(&view->srf_head);
+		view->committed = false;
+		res->id = -1;
+	}
+	mutex_unlock(&dev_priv->binding_mutex);
+}
+
+/**
+ * vmw_view_create - Create a hardware view.
+ *
+ * @res: Pointer to the view resource.
+ *
+ * Create a hardware view. Typically used if that view has previously been
+ * destroyed by an eviction operation.
+ */
+static int vmw_view_create(struct vmw_resource *res)
+{
+	struct vmw_view *view = vmw_view(res);
+	struct vmw_surface *srf = vmw_res_to_srf(view->srf);
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct {
+		SVGA3dCmdHeader header;
+		struct vmw_view_define body;
+	} *cmd;
+
+	mutex_lock(&dev_priv->binding_mutex);
+	if (!view->committed) {
+		mutex_unlock(&dev_priv->binding_mutex);
+		return 0;
+	}
+
+	cmd = vmw_fifo_reserve_dx(res->dev_priv, view->cmd_size,
+				  view->ctx->id);
+	if (!cmd) {
+		DRM_ERROR("Failed reserving FIFO space for view creation.\n");
+		mutex_unlock(&dev_priv->binding_mutex);
+		return -ENOMEM;
+	}
+	memcpy(cmd, &view->cmd, view->cmd_size);
+	WARN_ON(cmd->body.view_id != view->view_id);
+	/* Sid may have changed due to surface eviction. */
+	WARN_ON(view->srf->id == SVGA3D_INVALID_ID);
+	cmd->body.sid = view->srf->id;
+	vmw_fifo_commit(res->dev_priv, view->cmd_size);
+	res->id = view->view_id;
+	list_add_tail(&view->srf_head, &srf->view_list);
+	vmw_cotable_add_resource(view->cotable, &view->cotable_head);
+	mutex_unlock(&dev_priv->binding_mutex);
+
+	return 0;
+}
+
+/**
+ * vmw_view_destroy - Destroy a hardware view.
+ *
+ * @res: Pointer to the view resource.
+ *
+ * Destroy a hardware view. Typically used on unexpected termination of the
+ * owning process or if the surface the view is pointing to is destroyed.
+ */
+static int vmw_view_destroy(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+	struct vmw_view *view = vmw_view(res);
+	struct {
+		SVGA3dCmdHeader header;
+		union vmw_view_destroy body;
+	} *cmd;
+
+	WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex));
+	vmw_binding_res_list_scrub(&res->binding_head);
+
+	if (!view->committed || res->id == -1)
+		return 0;
+
+	cmd = vmw_fifo_reserve_dx(dev_priv, sizeof(*cmd), view->ctx->id);
+	if (!cmd) {
+		DRM_ERROR("Failed reserving FIFO space for view "
+			  "destruction.\n");
+		return -ENOMEM;
+	}
+
+	cmd->header.id = vmw_view_destroy_cmds[view->view_type];
+	cmd->header.size = sizeof(cmd->body);
+	cmd->body.view_id = view->view_id;
+	vmw_fifo_commit(dev_priv, sizeof(*cmd));
+	res->id = -1;
+	list_del_init(&view->cotable_head);
+	list_del_init(&view->srf_head);
+
+	return 0;
+}
+
+/**
+ * vmw_hw_view_destroy - Destroy a hardware view as part of resource cleanup.
+ *
+ * @res: Pointer to the view resource.
+ *
+ * Destroy a hardware view if it's still present.
+ */
+static void vmw_hw_view_destroy(struct vmw_resource *res)
+{
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	mutex_lock(&dev_priv->binding_mutex);
+	WARN_ON(vmw_view_destroy(res));
+	res->id = -1;
+	mutex_unlock(&dev_priv->binding_mutex);
+}
+
+/**
+ * vmw_view_key - Compute a view key suitable for the cmdbuf resource manager
+ *
+ * @user_key: The user-space id used for the view.
+ * @view_type: The view type.
+ *
+ * Destroy a hardware view if it's still present.
+ */
+static u32 vmw_view_key(u32 user_key, enum vmw_view_type view_type)
+{
+	return user_key | (view_type << 20);
+}
+
+/**
+ * vmw_view_id_ok - Basic view id and type range checks.
+ *
+ * @user_key: The user-space id used for the view.
+ * @view_type: The view type.
+ *
+ * Checks that the view id and type (typically provided by user-space) is
+ * valid.
+ */
+static bool vmw_view_id_ok(u32 user_key, enum vmw_view_type view_type)
+{
+	return (user_key < SVGA_COTABLE_MAX_IDS &&
+		view_type < vmw_view_max);
+}
+
+/**
+ * vmw_view_res_free - resource res_free callback for view resources
+ *
+ * @res: Pointer to a struct vmw_resource
+ *
+ * Frees memory and memory accounting held by a struct vmw_view.
+ */
+static void vmw_view_res_free(struct vmw_resource *res)
+{
+	struct vmw_view *view = vmw_view(res);
+	size_t size = offsetof(struct vmw_view, cmd) + view->cmd_size;
+	struct vmw_private *dev_priv = res->dev_priv;
+
+	vmw_resource_unreference(&view->cotable);
+	vmw_resource_unreference(&view->srf);
+	kfree_rcu(view, rcu);
+	ttm_mem_global_free(vmw_mem_glob(dev_priv), size);
+}
+
+/**
+ * vmw_view_add - Create a view resource and stage it for addition
+ * as a command buffer managed resource.
+ *
+ * @man: Pointer to the compat shader manager identifying the shader namespace.
+ * @ctx: Pointer to a struct vmw_resource identifying the active context.
+ * @srf: Pointer to a struct vmw_resource identifying the surface the view
+ * points to.
+ * @view_type: The view type deduced from the view create command.
+ * @user_key: The key that is used to identify the shader. The key is
+ * unique to the view type and to the context.
+ * @cmd: Pointer to the view create command in the command stream.
+ * @cmd_size: Size of the view create command in the command stream.
+ * @list: Caller's list of staged command buffer resource actions.
+ */
+int vmw_view_add(struct vmw_cmdbuf_res_manager *man,
+		 struct vmw_resource *ctx,
+		 struct vmw_resource *srf,
+		 enum vmw_view_type view_type,
+		 u32 user_key,
+		 const void *cmd,
+		 size_t cmd_size,
+		 struct list_head *list)
+{
+	static const size_t vmw_view_define_sizes[] = {
+		[vmw_view_sr] = sizeof(SVGA3dCmdDXDefineShaderResourceView),
+		[vmw_view_rt] = sizeof(SVGA3dCmdDXDefineRenderTargetView),
+		[vmw_view_ds] = sizeof(SVGA3dCmdDXDefineDepthStencilView)
+	};
+
+	struct vmw_private *dev_priv = ctx->dev_priv;
+	struct vmw_resource *res;
+	struct vmw_view *view;
+	size_t size;
+	int ret;
+
+	if (cmd_size != vmw_view_define_sizes[view_type] +
+	    sizeof(SVGA3dCmdHeader)) {
+		DRM_ERROR("Illegal view create command size.\n");
+		return -EINVAL;
+	}
+
+	if (!vmw_view_id_ok(user_key, view_type)) {
+		DRM_ERROR("Illegal view add view id.\n");
+		return -EINVAL;
+	}
+
+	size = offsetof(struct vmw_view, cmd) + cmd_size;
+
+	ret = ttm_mem_global_alloc(vmw_mem_glob(dev_priv), size, false, true);
+	if (ret) {
+		if (ret != -ERESTARTSYS)
+			DRM_ERROR("Out of graphics memory for view"
+				  " creation.\n");
+		return ret;
+	}
+
+	view = kmalloc(size, GFP_KERNEL);
+	if (!view) {
+		ttm_mem_global_free(vmw_mem_glob(dev_priv), size);
+		return -ENOMEM;
+	}
+
+	res = &view->res;
+	view->ctx = ctx;
+	view->srf = vmw_resource_reference(srf);
+	view->cotable = vmw_context_cotable(ctx, vmw_view_cotables[view_type]);
+	view->view_type = view_type;
+	view->view_id = user_key;
+	view->cmd_size = cmd_size;
+	view->committed = false;
+	INIT_LIST_HEAD(&view->srf_head);
+	INIT_LIST_HEAD(&view->cotable_head);
+	memcpy(&view->cmd, cmd, cmd_size);
+	ret = vmw_resource_init(dev_priv, res, true,
+				vmw_view_res_free, &vmw_view_func);
+	if (ret)
+		goto out_resource_init;
+
+	ret = vmw_cmdbuf_res_add(man, vmw_cmdbuf_res_view,
+				 vmw_view_key(user_key, view_type),
+				 res, list);
+	if (ret)
+		goto out_resource_init;
+
+	res->id = view->view_id;
+	vmw_resource_activate(res, vmw_hw_view_destroy);
+
+out_resource_init:
+	vmw_resource_unreference(&res);
+
+	return ret;
+}
+
+/**
+ * vmw_view_remove - Stage a view for removal.
+ *
+ * @man: Pointer to the view manager identifying the shader namespace.
+ * @user_key: The key that is used to identify the view. The key is
+ * unique to the view type.
+ * @view_type: View type
+ * @list: Caller's list of staged command buffer resource actions.
+ * @res_p: If the resource is in an already committed state, points to the
+ * struct vmw_resource on successful return. The pointer will be
+ * non ref-counted.
+ */
+int vmw_view_remove(struct vmw_cmdbuf_res_manager *man,
+		    u32 user_key, enum vmw_view_type view_type,
+		    struct list_head *list,
+		    struct vmw_resource **res_p)
+{
+	if (!vmw_view_id_ok(user_key, view_type)) {
+		DRM_ERROR("Illegal view remove view id.\n");
+		return -EINVAL;
+	}
+
+	return vmw_cmdbuf_res_remove(man, vmw_cmdbuf_res_view,
+				     vmw_view_key(user_key, view_type),
+				     list, res_p);
+}
+
+/**
+ * vmw_view_cotable_list_destroy - Evict all views belonging to a cotable.
+ *
+ * @dev_priv: Pointer to a device private struct.
+ * @list: List of views belonging to a cotable.
+ * @readback: Unused. Needed for function interface only.
+ *
+ * This function evicts all views belonging to a cotable.
+ * It must be called with the binding_mutex held, and the caller must hold
+ * a reference to the view resource. This is typically called before the
+ * cotable is paged out.
+ */
+void vmw_view_cotable_list_destroy(struct vmw_private *dev_priv,
+				   struct list_head *list,
+				   bool readback)
+{
+	struct vmw_view *entry, *next;
+
+	WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex));
+
+	list_for_each_entry_safe(entry, next, list, cotable_head)
+		WARN_ON(vmw_view_destroy(&entry->res));
+}
+
+/**
+ * vmw_view_surface_list_destroy - Evict all views pointing to a surface
+ *
+ * @dev_priv: Pointer to a device private struct.
+ * @list: List of views pointing to a surface.
+ *
+ * This function evicts all views pointing to a surface. This is typically
+ * called before the surface is evicted.
+ */
+void vmw_view_surface_list_destroy(struct vmw_private *dev_priv,
+				   struct list_head *list)
+{
+	struct vmw_view *entry, *next;
+
+	WARN_ON_ONCE(!mutex_is_locked(&dev_priv->binding_mutex));
+
+	list_for_each_entry_safe(entry, next, list, srf_head)
+		WARN_ON(vmw_view_destroy(&entry->res));
+}
+
+/**
+ * vmw_view_srf - Return a non-refcounted pointer to the surface a view is
+ * pointing to.
+ *
+ * @res: pointer to a view resource.
+ *
+ * Note that the view itself is holding a reference, so as long
+ * the view resource is alive, the surface resource will be.
+ */
+struct vmw_resource *vmw_view_srf(struct vmw_resource *res)
+{
+	return vmw_view(res)->srf;
+}
+
+/**
+ * vmw_view_lookup - Look up a view.
+ *
+ * @man: The context's cmdbuf ref manager.
+ * @view_type: The view type.
+ * @user_key: The view user id.
+ *
+ * returns a refcounted pointer to a view or an error pointer if not found.
+ */
+struct vmw_resource *vmw_view_lookup(struct vmw_cmdbuf_res_manager *man,
+				     enum vmw_view_type view_type,
+				     u32 user_key)
+{
+	return vmw_cmdbuf_res_lookup(man, vmw_cmdbuf_res_view,
+				     vmw_view_key(user_key, view_type));
+}
+
+const u32 vmw_view_destroy_cmds[] = {
+	[vmw_view_sr] = SVGA_3D_CMD_DX_DESTROY_SHADERRESOURCE_VIEW,
+	[vmw_view_rt] = SVGA_3D_CMD_DX_DESTROY_RENDERTARGET_VIEW,
+	[vmw_view_ds] = SVGA_3D_CMD_DX_DESTROY_DEPTHSTENCIL_VIEW,
+};
+
+const SVGACOTableType vmw_view_cotables[] = {
+	[vmw_view_sr] = SVGA_COTABLE_SRVIEW,
+	[vmw_view_rt] = SVGA_COTABLE_RTVIEW,
+	[vmw_view_ds] = SVGA_COTABLE_DSVIEW,
+};
+
+const SVGACOTableType vmw_so_cotables[] = {
+	[vmw_so_el] = SVGA_COTABLE_ELEMENTLAYOUT,
+	[vmw_so_bs] = SVGA_COTABLE_BLENDSTATE,
+	[vmw_so_ds] = SVGA_COTABLE_DEPTHSTENCIL,
+	[vmw_so_rs] = SVGA_COTABLE_RASTERIZERSTATE,
+	[vmw_so_ss] = SVGA_COTABLE_SAMPLER,
+	[vmw_so_so] = SVGA_COTABLE_STREAMOUTPUT
+};
+
+
+/* To remove unused function warning */
+static void vmw_so_build_asserts(void) __attribute__((used));
+
+
+/*
+ * This function is unused at run-time, and only used to dump various build
+ * asserts important for code optimization assumptions.
+ */
+static void vmw_so_build_asserts(void)
+{
+	/* Assert that our vmw_view_cmd_to_type() function is correct. */
+	BUILD_BUG_ON(SVGA_3D_CMD_DX_DESTROY_SHADERRESOURCE_VIEW !=
+		     SVGA_3D_CMD_DX_DEFINE_SHADERRESOURCE_VIEW + 1);
+	BUILD_BUG_ON(SVGA_3D_CMD_DX_DEFINE_RENDERTARGET_VIEW !=
+		     SVGA_3D_CMD_DX_DEFINE_SHADERRESOURCE_VIEW + 2);
+	BUILD_BUG_ON(SVGA_3D_CMD_DX_DESTROY_RENDERTARGET_VIEW !=
+		     SVGA_3D_CMD_DX_DEFINE_SHADERRESOURCE_VIEW + 3);
+	BUILD_BUG_ON(SVGA_3D_CMD_DX_DEFINE_DEPTHSTENCIL_VIEW !=
+		     SVGA_3D_CMD_DX_DEFINE_SHADERRESOURCE_VIEW + 4);
+	BUILD_BUG_ON(SVGA_3D_CMD_DX_DESTROY_DEPTHSTENCIL_VIEW !=
+		     SVGA_3D_CMD_DX_DEFINE_SHADERRESOURCE_VIEW + 5);
+
+	/* Assert that our "one body fits all" assumption is valid */
+	BUILD_BUG_ON(sizeof(union vmw_view_destroy) != sizeof(u32));
+
+	/* Assert that the view key space can hold all view ids. */
+	BUILD_BUG_ON(SVGA_COTABLE_MAX_IDS >= ((1 << 20) - 1));
+
+	/*
+	 * Assert that the offset of sid in all view define commands
+	 * is what we assume it to be.
+	 */
+	BUILD_BUG_ON(offsetof(struct vmw_view_define, sid) !=
+		     offsetof(SVGA3dCmdDXDefineShaderResourceView, sid));
+	BUILD_BUG_ON(offsetof(struct vmw_view_define, sid) !=
+		     offsetof(SVGA3dCmdDXDefineRenderTargetView, sid));
+	BUILD_BUG_ON(offsetof(struct vmw_view_define, sid) !=
+		     offsetof(SVGA3dCmdDXDefineDepthStencilView, sid));
+}

drivers/gpu/drm/vmwgfx/vmwgfx_so.h

@@ -0,0 +1,160 @@
+/**************************************************************************
+ * Copyright © 2014 VMware, Inc., Palo Alto, CA., USA
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+#ifndef VMW_SO_H
+#define VMW_SO_H
+
+enum vmw_view_type {
+	vmw_view_sr,
+	vmw_view_rt,
+	vmw_view_ds,
+	vmw_view_max,
+};
+
+enum vmw_so_type {
+	vmw_so_el,
+	vmw_so_bs,
+	vmw_so_ds,
+	vmw_so_rs,
+	vmw_so_ss,
+	vmw_so_so,
+	vmw_so_max,
+};
+
+/**
+ * union vmw_view_destroy - view destruction command body
+ *
+ * @rtv: RenderTarget view destruction command body
+ * @srv: ShaderResource view destruction command body
+ * @dsv: DepthStencil view destruction command body
+ * @view_id: A single u32 view id.
+ *
+ * The assumption here is that all union members are really represented by a
+ * single u32 in the command stream. If that's not the case,
+ * the size of this union will not equal the size of an u32, and the
+ * assumption is invalid, and we detect that at compile time in the
+ * vmw_so_build_asserts() function.
+ */
+union vmw_view_destroy {
+	struct SVGA3dCmdDXDestroyRenderTargetView rtv;
+	struct SVGA3dCmdDXDestroyShaderResourceView srv;
+	struct SVGA3dCmdDXDestroyDepthStencilView dsv;
+	u32 view_id;
+};
+
+/* Map enum vmw_view_type to view destroy command ids*/
+extern const u32 vmw_view_destroy_cmds[];
+
+/* Map enum vmw_view_type to SVGACOTableType */
+extern const SVGACOTableType vmw_view_cotables[];
+
+/* Map enum vmw_so_type to SVGACOTableType */
+extern const SVGACOTableType vmw_so_cotables[];
+
+/*
+ * vmw_view_cmd_to_type - Return the view type for a create or destroy command
+ *
+ * @id: The SVGA3D command id.
+ *
+ * For a given view create or destroy command id, return the corresponding
+ * enum vmw_view_type. If the command is unknown, return vmw_view_max.
+ * The validity of the simplified calculation is verified in the
+ * vmw_so_build_asserts() function.
+ */
+static inline enum vmw_view_type vmw_view_cmd_to_type(u32 id)
+{
+	u32 tmp = (id - SVGA_3D_CMD_DX_DEFINE_SHADERRESOURCE_VIEW) / 2;
+
+	if (tmp > (u32)vmw_view_max)
+		return vmw_view_max;
+
+	return (enum vmw_view_type) tmp;
+}
+
+/*
+ * vmw_so_cmd_to_type - Return the state object type for a
+ * create or destroy command
+ *
+ * @id: The SVGA3D command id.
+ *
+ * For a given state object create or destroy command id,
+ * return the corresponding enum vmw_so_type. If the command is uknown,
+ * return vmw_so_max. We should perhaps optimize this function using
+ * a similar strategy as vmw_view_cmd_to_type().
+ */
+static inline enum vmw_so_type vmw_so_cmd_to_type(u32 id)
+{
+	switch (id) {
+	case SVGA_3D_CMD_DX_DEFINE_ELEMENTLAYOUT:
+	case SVGA_3D_CMD_DX_DESTROY_ELEMENTLAYOUT:
+		return vmw_so_el;
+	case SVGA_3D_CMD_DX_DEFINE_BLEND_STATE:
+	case SVGA_3D_CMD_DX_DESTROY_BLEND_STATE:
+		return vmw_so_bs;
+	case SVGA_3D_CMD_DX_DEFINE_DEPTHSTENCIL_STATE:
+	case SVGA_3D_CMD_DX_DESTROY_DEPTHSTENCIL_STATE:
+		return vmw_so_ds;
+	case SVGA_3D_CMD_DX_DEFINE_RASTERIZER_STATE:
+	case SVGA_3D_CMD_DX_DESTROY_RASTERIZER_STATE:
+		return vmw_so_rs;
+	case SVGA_3D_CMD_DX_DEFINE_SAMPLER_STATE:
+	case SVGA_3D_CMD_DX_DESTROY_SAMPLER_STATE:
+		return vmw_so_ss;
+	case SVGA_3D_CMD_DX_DEFINE_STREAMOUTPUT:
+	case SVGA_3D_CMD_DX_DESTROY_STREAMOUTPUT:
+		return vmw_so_so;
+	default:
+		break;
+	}
+	return vmw_so_max;
+}
+
+/*
+ * View management - vmwgfx_so.c
+ */
+extern int vmw_view_add(struct vmw_cmdbuf_res_manager *man,
+			struct vmw_resource *ctx,
+			struct vmw_resource *srf,
+			enum vmw_view_type view_type,
+			u32 user_key,
+			const void *cmd,
+			size_t cmd_size,
+			struct list_head *list);
+
+extern int vmw_view_remove(struct vmw_cmdbuf_res_manager *man,
+			   u32 user_key, enum vmw_view_type view_type,
+			   struct list_head *list,
+			   struct vmw_resource **res_p);
+
+extern void vmw_view_surface_list_destroy(struct vmw_private *dev_priv,
+					  struct list_head *view_list);
+extern void vmw_view_cotable_list_destroy(struct vmw_private *dev_priv,
+					  struct list_head *list,
+					  bool readback);
+extern struct vmw_resource *vmw_view_srf(struct vmw_resource *res);
+extern struct vmw_resource *vmw_view_lookup(struct vmw_cmdbuf_res_manager *man,
+					    enum vmw_view_type view_type,
+					    u32 user_key);
+#endif

drivers/gpu/drm/vmwgfx/vmwgfx_stdu.c

@@ -561,6 +561,7 @@ static int vmw_stdu_crtc_set_config(struct drm_mode_set *set)
 				true, /* a scanout buffer */
 				content_srf.mip_levels[0],
 				content_srf.multisample_count,
+				0,
 				display_base_size,
 				&display_srf);
 		if (unlikely(ret != 0)) {

drivers/gpu/drm/vmwgfx/vmwgfx_surface.c

@@ -27,9 +27,12 @@
 
 #include "vmwgfx_drv.h"
 #include "vmwgfx_resource_priv.h"
+#include "vmwgfx_so.h"
+#include "vmwgfx_binding.h"
 #include <ttm/ttm_placement.h>
 #include "device_include/svga3d_surfacedefs.h"
 
+
 /**
  * struct vmw_user_surface - User-space visible surface resource
  *
@@ -593,6 +596,7 @@ static int vmw_surface_init(struct vmw_private *dev_priv,
 	 * surface validate.
 	 */
 
+	INIT_LIST_HEAD(&srf->view_list);
 	vmw_resource_activate(res, vmw_hw_surface_destroy);
 	return ret;
 }
@@ -723,6 +727,7 @@ int vmw_surface_define_ioctl(struct drm_device *dev, void *data,
 	desc = svga3dsurface_get_desc(req->format);
 	if (unlikely(desc->block_desc == SVGA3DBLOCKDESC_NONE)) {
 		DRM_ERROR("Invalid surface format for surface creation.\n");
+		DRM_ERROR("Format requested is: %d\n", req->format);
 		return -EINVAL;
 	}
 
@@ -1018,12 +1023,16 @@ static int vmw_gb_surface_create(struct vmw_resource *res)
 {
 	struct vmw_private *dev_priv = res->dev_priv;
 	struct vmw_surface *srf = vmw_res_to_srf(res);
-	uint32_t cmd_len, submit_len;
+	uint32_t cmd_len, cmd_id, submit_len;
 	int ret;
 	struct {
 		SVGA3dCmdHeader header;
 		SVGA3dCmdDefineGBSurface body;
 	} *cmd;
+	struct {
+		SVGA3dCmdHeader header;
+		SVGA3dCmdDefineGBSurface_v2 body;
+	} *cmd2;
 
 	if (likely(res->id != -1))
 		return 0;
@@ -1040,9 +1049,19 @@ static int vmw_gb_surface_create(struct vmw_resource *res)
 		goto out_no_fifo;
 	}
 
+	if (srf->array_size > 0) {
+		/* has_dx checked on creation time. */
+		cmd_id = SVGA_3D_CMD_DEFINE_GB_SURFACE_V2;
+		cmd_len = sizeof(cmd2->body);
+		submit_len = sizeof(*cmd2);
+	} else {
+		cmd_id = SVGA_3D_CMD_DEFINE_GB_SURFACE;
 		cmd_len = sizeof(cmd->body);
 		submit_len = sizeof(*cmd);
+	}
+
 	cmd = vmw_fifo_reserve(dev_priv, submit_len);
+	cmd2 = (typeof(cmd2))cmd;
 	if (unlikely(cmd == NULL)) {
 		DRM_ERROR("Failed reserving FIFO space for surface "
 			  "creation.\n");
@@ -1050,17 +1069,33 @@ static int vmw_gb_surface_create(struct vmw_resource *res)
 		goto out_no_fifo;
 	}
 
-	cmd->header.id = SVGA_3D_CMD_DEFINE_GB_SURFACE;
+	if (srf->array_size > 0) {
+		cmd2->header.id = cmd_id;
+		cmd2->header.size = cmd_len;
+		cmd2->body.sid = srf->res.id;
+		cmd2->body.surfaceFlags = srf->flags;
+		cmd2->body.format = cpu_to_le32(srf->format);
+		cmd2->body.numMipLevels = srf->mip_levels[0];
+		cmd2->body.multisampleCount = srf->multisample_count;
+		cmd2->body.autogenFilter = srf->autogen_filter;
+		cmd2->body.size.width = srf->base_size.width;
+		cmd2->body.size.height = srf->base_size.height;
+		cmd2->body.size.depth = srf->base_size.depth;
+		cmd2->body.arraySize = srf->array_size;
+	} else {
+		cmd->header.id = cmd_id;
 		cmd->header.size = cmd_len;
 		cmd->body.sid = srf->res.id;
 		cmd->body.surfaceFlags = srf->flags;
-	cmd->body.format = srf->format;
+		cmd->body.format = cpu_to_le32(srf->format);
 		cmd->body.numMipLevels = srf->mip_levels[0];
 		cmd->body.multisampleCount = srf->multisample_count;
 		cmd->body.autogenFilter = srf->autogen_filter;
 		cmd->body.size.width = srf->base_size.width;
 		cmd->body.size.height = srf->base_size.height;
 		cmd->body.size.depth = srf->base_size.depth;
+	}
+
 	vmw_fifo_commit(dev_priv, submit_len);
 
 	return 0;
@@ -1188,6 +1223,7 @@ static int vmw_gb_surface_unbind(struct vmw_resource *res,
 static int vmw_gb_surface_destroy(struct vmw_resource *res)
 {
 	struct vmw_private *dev_priv = res->dev_priv;
+	struct vmw_surface *srf = vmw_res_to_srf(res);
 	struct {
 		SVGA3dCmdHeader header;
 		SVGA3dCmdDestroyGBSurface body;
@@ -1197,7 +1233,8 @@ static int vmw_gb_surface_destroy(struct vmw_resource *res)
 		return 0;
 
 	mutex_lock(&dev_priv->binding_mutex);
-	vmw_context_binding_res_list_scrub(&res->binding_head);
+	vmw_view_surface_list_destroy(dev_priv, &srf->view_list);
+	vmw_binding_res_list_scrub(&res->binding_head);
 
 	cmd = vmw_fifo_reserve(dev_priv, sizeof(*cmd));
 	if (unlikely(cmd == NULL)) {
@@ -1259,6 +1296,7 @@ int vmw_gb_surface_define_ioctl(struct drm_device *dev, void *data,
 			req->drm_surface_flags & drm_vmw_surface_flag_scanout,
 			req->mip_levels,
 			req->multisample_count,
+			req->array_size,
 			req->base_size,
 			&srf);
 	if (unlikely(ret != 0))
@@ -1275,10 +1313,17 @@ int vmw_gb_surface_define_ioctl(struct drm_device *dev, void *data,
 	res = &user_srf->srf.res;
 
 
-	if (req->buffer_handle != SVGA3D_INVALID_ID)
+	if (req->buffer_handle != SVGA3D_INVALID_ID) {
 		ret = vmw_user_dmabuf_lookup(tfile, req->buffer_handle,
 					     &res->backup);
-	else if (req->drm_surface_flags & drm_vmw_surface_flag_create_buffer)
+		if (ret == 0 && res->backup->base.num_pages * PAGE_SIZE <
+		    res->backup_size) {
+			DRM_ERROR("Surface backup buffer is too small.\n");
+			vmw_dmabuf_unreference(&res->backup);
+			ret = -EINVAL;
+			goto out_unlock;
+		}
+	} else if (req->drm_surface_flags & drm_vmw_surface_flag_create_buffer)
 		ret = vmw_user_dmabuf_alloc(dev_priv, tfile,
 					    res->backup_size,
 					    req->drm_surface_flags &
@@ -1378,6 +1423,7 @@ int vmw_gb_surface_reference_ioctl(struct drm_device *dev, void *data,
 	rep->creq.drm_surface_flags = 0;
 	rep->creq.multisample_count = srf->multisample_count;
 	rep->creq.autogen_filter = srf->autogen_filter;
+	rep->creq.array_size = srf->array_size;
 	rep->creq.buffer_handle = backup_handle;
 	rep->creq.base_size = srf->base_size;
 	rep->crep.handle = user_srf->prime.base.hash.key;
@@ -1404,6 +1450,7 @@ int vmw_gb_surface_reference_ioctl(struct drm_device *dev, void *data,
  * @for_scanout: true if inteded to be used for scanout buffer
  * @num_mip_levels:  number of MIP levels
  * @multisample_count:
+ * @array_size: Surface array size.
  * @size: width, heigh, depth of the surface requested
  * @user_srf_out: allocated user_srf.  Set to NULL on failure.
  *
@@ -1419,6 +1466,7 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 			       bool for_scanout,
 			       uint32_t num_mip_levels,
 			       uint32_t multisample_count,
+			       uint32_t array_size,
 			       struct drm_vmw_size size,
 			       struct vmw_surface **srf_out)
 {
@@ -1426,7 +1474,7 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 	struct vmw_user_surface *user_srf;
 	struct vmw_surface *srf;
 	int ret;
-
+	u32 num_layers;
 
 	*srf_out = NULL;
 
@@ -1445,6 +1493,12 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 		}
 	}
 
+	/* array_size must be null for non-GL3 host. */
+	if (array_size > 0 && !dev_priv->has_dx) {
+		DRM_ERROR("Tried to create DX surface on non-DX host.\n");
+		return -EINVAL;
+	}
+
 	ret = ttm_read_lock(&dev_priv->reservation_sem, true);
 	if (unlikely(ret != 0))
 		return ret;
@@ -1481,10 +1535,21 @@ int vmw_surface_gb_priv_define(struct drm_device *dev,
 	srf->autogen_filter    = SVGA3D_TEX_FILTER_NONE;
 	srf->multisample_count = multisample_count;
 
-	srf->res.backup_size   = svga3dsurface_get_serialized_size(srf->format,
+	if (array_size)
+		num_layers = array_size;
+	else if (svga3d_flags & SVGA3D_SURFACE_CUBEMAP)
+		num_layers = SVGA3D_MAX_SURFACE_FACES;
+	else
+		num_layers = 1;
+
+	srf->res.backup_size   =
+		svga3dsurface_get_serialized_size(srf->format,
 						  srf->base_size,
 						  srf->mip_levels[0],
-					srf->flags & SVGA3D_SURFACE_CUBEMAP);
+						  num_layers);
+
+	if (srf->flags & SVGA3D_SURFACE_BIND_STREAM_OUTPUT)
+		srf->res.backup_size += sizeof(SVGA3dDXSOState);
 
 	if (dev_priv->active_display_unit == vmw_du_screen_target &&
 	    for_scanout)

include/uapi/drm/vmwgfx_drm.h

@@ -64,6 +64,7 @@
 #define DRM_VMW_GB_SURFACE_CREATE    23
 #define DRM_VMW_GB_SURFACE_REF       24
 #define DRM_VMW_SYNCCPU              25
+#define DRM_VMW_CREATE_EXTENDED_CONTEXT 26
 
 /*************************************************************************/
 /**
@@ -89,6 +90,7 @@
 #define DRM_VMW_PARAM_MAX_MOB_MEMORY   9
 #define DRM_VMW_PARAM_MAX_MOB_SIZE     10
 #define DRM_VMW_PARAM_SCREEN_TARGET    11
+#define DRM_VMW_PARAM_DX               12
 
 /**
  * enum drm_vmw_handle_type - handle type for ref ioctls
@@ -297,7 +299,7 @@ union drm_vmw_surface_reference_arg {
  * Argument to the DRM_VMW_EXECBUF Ioctl.
  */
 
-#define DRM_VMW_EXECBUF_VERSION 1
+#define DRM_VMW_EXECBUF_VERSION 2
 
 struct drm_vmw_execbuf_arg {
 	uint64_t commands;
@@ -306,6 +308,8 @@ struct drm_vmw_execbuf_arg {
 	uint64_t fence_rep;
 	uint32_t version;
 	uint32_t flags;
+	uint32_t context_handle;
+	uint32_t pad64;
 };
 
 /**
@@ -826,7 +830,6 @@ struct drm_vmw_update_layout_arg {
 enum drm_vmw_shader_type {
 	drm_vmw_shader_type_vs = 0,
 	drm_vmw_shader_type_ps,
-	drm_vmw_shader_type_gs
 };
 
 
@@ -908,6 +911,8 @@ enum drm_vmw_surface_flags {
  * @buffer_handle     Buffer handle of backup buffer. SVGA3D_INVALID_ID
  *                    if none.
  * @base_size         Size of the base mip level for all faces.
+ * @array_size        Must be zero for non-DX hardware, and if non-zero
+ *                    svga3d_flags must have proper bind flags setup.
  *
  * Input argument to the  DRM_VMW_GB_SURFACE_CREATE Ioctl.
  * Part of output argument for the DRM_VMW_GB_SURFACE_REF Ioctl.
@@ -920,7 +925,7 @@ struct drm_vmw_gb_surface_create_req {
 	uint32_t multisample_count;
 	uint32_t autogen_filter;
 	uint32_t buffer_handle;
-	uint32_t pad64;
+	uint32_t array_size;
 	struct drm_vmw_size base_size;
 };
 
@@ -1060,4 +1065,28 @@ struct drm_vmw_synccpu_arg {
 	uint32_t pad64;
 };
 
+/*************************************************************************/
+/**
+ * DRM_VMW_CREATE_EXTENDED_CONTEXT - Create a host context.
+ *
+ * Allocates a device unique context id, and queues a create context command
+ * for the host. Does not wait for host completion.
+ */
+enum drm_vmw_extended_context {
+	drm_vmw_context_legacy,
+	drm_vmw_context_dx
+};
+
+/**
+ * union drm_vmw_extended_context_arg
+ *
+ * @req: Context type.
+ * @rep: Context identifier.
+ *
+ * Argument to the DRM_VMW_CREATE_EXTENDED_CONTEXT Ioctl.
+ */
+union drm_vmw_extended_context_arg {
+	enum drm_vmw_extended_context req;
+	struct drm_vmw_context_arg rep;
+};
 #endif