vmwgfx: Implement fence objects

Will be needed for queries and drm event-driven throttling. As a benefit, they help avoid stale user-space fence handles. Signed-off-by: Thomas Hellstrom <thellstrom@vmware.com> Reviewed-by: Jakob Bornecrantz <jakob@vmware.com> Signed-off-by: Dave Airlie <airlied@redhat.com>
2011-09-01 20:18:44 +00:00
parent 4f73a96bd7
commit ae2a104058
10 changed files with 1010 additions and 87 deletions
--- a/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
+++ b/drivers/gpu/drm/vmwgfx/vmwgfx_fence.c
@@ -0,0 +1,619 @@
+/**************************************************************************
+ *
+ * Copyright © 2011 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 "drmP.h"
+#include "vmwgfx_drv.h"
+
+#define VMW_FENCE_WRAP (1 << 31)
+
+struct vmw_fence_manager {
+	int num_fence_objects;
+	struct vmw_private *dev_priv;
+	spinlock_t lock;
+	u32 next_seqno;
+	struct list_head fence_list;
+	struct work_struct work;
+	u32 user_fence_size;
+	u32 fence_size;
+	bool fifo_down;
+	struct list_head cleanup_list;
+};
+
+struct vmw_user_fence {
+	struct ttm_base_object base;
+	struct vmw_fence_obj fence;
+};
+
+/**
+ * vmw_fence_destroy_locked
+ *
+ */
+
+static void vmw_fence_obj_destroy_locked(struct kref *kref)
+{
+	struct vmw_fence_obj *fence =
+		container_of(kref, struct vmw_fence_obj, kref);
+
+	struct vmw_fence_manager *fman = fence->fman;
+	unsigned int num_fences;
+
+	list_del_init(&fence->head);
+	num_fences = --fman->num_fence_objects;
+	spin_unlock_irq(&fman->lock);
+	if (fence->destroy)
+		fence->destroy(fence);
+	else
+		kfree(fence);
+
+	spin_lock_irq(&fman->lock);
+}
+
+
+/**
+ * Execute signal actions on fences recently signaled.
+ * This is done from a workqueue so we don't have to execute
+ * signal actions from atomic context.
+ */
+
+static void vmw_fence_work_func(struct work_struct *work)
+{
+	struct vmw_fence_manager *fman =
+		container_of(work, struct vmw_fence_manager, work);
+	struct list_head list;
+	struct vmw_fence_action *action, *next_action;
+
+	do {
+		INIT_LIST_HEAD(&list);
+		spin_lock_irq(&fman->lock);
+		list_splice_init(&fman->cleanup_list, &list);
+		spin_unlock_irq(&fman->lock);
+
+		if (list_empty(&list))
+			return;
+
+		/*
+		 * At this point, only we should be able to manipulate the
+		 * list heads of the actions we have on the private list.
+		 */
+
+		list_for_each_entry_safe(action, next_action, &list, head) {
+			list_del_init(&action->head);
+			action->cleanup(action);
+		}
+	} while (1);
+}
+
+struct vmw_fence_manager *vmw_fence_manager_init(struct vmw_private *dev_priv)
+{
+	struct vmw_fence_manager *fman = kzalloc(sizeof(*fman), GFP_KERNEL);
+
+	if (unlikely(fman == NULL))
+		return NULL;
+
+	fman->dev_priv = dev_priv;
+	spin_lock_init(&fman->lock);
+	INIT_LIST_HEAD(&fman->fence_list);
+	INIT_LIST_HEAD(&fman->cleanup_list);
+	INIT_WORK(&fman->work, &vmw_fence_work_func);
+	fman->fifo_down = true;
+	fman->user_fence_size = ttm_round_pot(sizeof(struct vmw_user_fence));
+	fman->fence_size = ttm_round_pot(sizeof(struct vmw_fence_obj));
+
+	return fman;
+}
+
+void vmw_fence_manager_takedown(struct vmw_fence_manager *fman)
+{
+	unsigned long irq_flags;
+	bool lists_empty;
+
+	(void) cancel_work_sync(&fman->work);
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	lists_empty = list_empty(&fman->fence_list) &&
+		list_empty(&fman->cleanup_list);
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+
+	BUG_ON(!lists_empty);
+	kfree(fman);
+}
+
+static int vmw_fence_obj_init(struct vmw_fence_manager *fman,
+			      struct vmw_fence_obj *fence,
+			      u32 seqno,
+			      uint32_t mask,
+			      void (*destroy) (struct vmw_fence_obj *fence))
+{
+	unsigned long irq_flags;
+	unsigned int num_fences;
+	int ret = 0;
+
+	fence->seqno = seqno;
+	INIT_LIST_HEAD(&fence->seq_passed_actions);
+	fence->fman = fman;
+	fence->signaled = 0;
+	fence->signal_mask = mask;
+	kref_init(&fence->kref);
+	fence->destroy = destroy;
+	init_waitqueue_head(&fence->queue);
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	if (unlikely(fman->fifo_down)) {
+		ret = -EBUSY;
+		goto out_unlock;
+	}
+	list_add_tail(&fence->head, &fman->fence_list);
+	num_fences = ++fman->num_fence_objects;
+
+out_unlock:
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+	return ret;
+
+}
+
+struct vmw_fence_obj *vmw_fence_obj_reference(struct vmw_fence_obj *fence)
+{
+	kref_get(&fence->kref);
+	return fence;
+}
+
+/**
+ * vmw_fence_obj_unreference
+ *
+ * Note that this function may not be entered with disabled irqs since
+ * it may re-enable them in the destroy function.
+ *
+ */
+void vmw_fence_obj_unreference(struct vmw_fence_obj **fence_p)
+{
+	struct vmw_fence_obj *fence = *fence_p;
+	struct vmw_fence_manager *fman = fence->fman;
+
+	*fence_p = NULL;
+	spin_lock_irq(&fman->lock);
+	BUG_ON(atomic_read(&fence->kref.refcount) == 0);
+	kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
+	spin_unlock_irq(&fman->lock);
+}
+
+void vmw_fences_perform_actions(struct vmw_fence_manager *fman,
+				struct list_head *list)
+{
+	struct vmw_fence_action *action, *next_action;
+
+	list_for_each_entry_safe(action, next_action, list, head) {
+		list_del_init(&action->head);
+		if (action->seq_passed != NULL)
+			action->seq_passed(action);
+
+		/*
+		 * Add the cleanup action to the cleanup list so that
+		 * it will be performed by a worker task.
+		 */
+
+		if (action->cleanup != NULL)
+			list_add_tail(&action->head, &fman->cleanup_list);
+	}
+}
+
+void vmw_fences_update(struct vmw_fence_manager *fman, u32 seqno)
+{
+	unsigned long flags;
+	struct vmw_fence_obj *fence, *next_fence;
+	struct list_head action_list;
+
+	spin_lock_irqsave(&fman->lock, flags);
+	list_for_each_entry_safe(fence, next_fence, &fman->fence_list, head) {
+		if (seqno - fence->seqno < VMW_FENCE_WRAP) {
+			list_del_init(&fence->head);
+			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
+			INIT_LIST_HEAD(&action_list);
+			list_splice_init(&fence->seq_passed_actions,
+					 &action_list);
+			vmw_fences_perform_actions(fman, &action_list);
+			wake_up_all(&fence->queue);
+		}
+
+	}
+	if (!list_empty(&fman->cleanup_list))
+		(void) schedule_work(&fman->work);
+	spin_unlock_irqrestore(&fman->lock, flags);
+}
+
+
+bool vmw_fence_obj_signaled(struct vmw_fence_obj *fence,
+			    uint32_t flags)
+{
+	struct vmw_fence_manager *fman = fence->fman;
+	unsigned long irq_flags;
+	uint32_t signaled;
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	signaled = fence->signaled;
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+
+	flags &= fence->signal_mask;
+	if ((signaled & flags) == flags)
+		return 1;
+
+	if ((signaled & DRM_VMW_FENCE_FLAG_EXEC) == 0) {
+		struct vmw_private *dev_priv = fman->dev_priv;
+		__le32 __iomem *fifo_mem = dev_priv->mmio_virt;
+		u32 seqno;
+
+		seqno = ioread32(fifo_mem + SVGA_FIFO_FENCE);
+		vmw_fences_update(fman, seqno);
+	}
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	signaled = fence->signaled;
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+
+	return ((signaled & flags) == flags);
+}
+
+int vmw_fence_obj_wait(struct vmw_fence_obj *fence,
+		       uint32_t flags, bool lazy,
+		       bool interruptible, unsigned long timeout)
+{
+	struct vmw_private *dev_priv = fence->fman->dev_priv;
+	long ret;
+
+	if (likely(vmw_fence_obj_signaled(fence, flags)))
+		return 0;
+
+	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
+	vmw_seqno_waiter_add(dev_priv);
+
+	if (interruptible)
+		ret = wait_event_interruptible_timeout
+			(fence->queue,
+			 vmw_fence_obj_signaled(fence, flags),
+			 timeout);
+	else
+		ret = wait_event_timeout
+			(fence->queue,
+			 vmw_fence_obj_signaled(fence, flags),
+			 timeout);
+
+	vmw_seqno_waiter_remove(dev_priv);
+
+	if (unlikely(ret == 0))
+		ret = -EBUSY;
+	else if (likely(ret > 0))
+		ret = 0;
+
+	return ret;
+}
+
+void vmw_fence_obj_flush(struct vmw_fence_obj *fence)
+{
+	struct vmw_private *dev_priv = fence->fman->dev_priv;
+
+	vmw_fifo_ping_host(dev_priv, SVGA_SYNC_GENERIC);
+}
+
+static void vmw_fence_destroy(struct vmw_fence_obj *fence)
+{
+	struct vmw_fence_manager *fman = fence->fman;
+
+	kfree(fence);
+	/*
+	 * Free kernel space accounting.
+	 */
+	ttm_mem_global_free(vmw_mem_glob(fman->dev_priv),
+			    fman->fence_size);
+}
+
+int vmw_fence_create(struct vmw_fence_manager *fman,
+		     uint32_t seqno,
+		     uint32_t mask,
+		     struct vmw_fence_obj **p_fence)
+{
+	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
+	struct vmw_fence_obj *fence;
+	int ret;
+
+	ret = ttm_mem_global_alloc(mem_glob, fman->fence_size,
+				   false, false);
+	if (unlikely(ret != 0))
+		return ret;
+
+	fence = kzalloc(sizeof(*fence), GFP_KERNEL);
+	if (unlikely(fence == NULL)) {
+		ret = -ENOMEM;
+		goto out_no_object;
+	}
+
+	ret = vmw_fence_obj_init(fman, fence, seqno, mask,
+				 vmw_fence_destroy);
+	if (unlikely(ret != 0))
+		goto out_err_init;
+
+	*p_fence = fence;
+	return 0;
+
+out_err_init:
+	kfree(fence);
+out_no_object:
+	ttm_mem_global_free(mem_glob, fman->fence_size);
+	return ret;
+}
+
+
+static void vmw_user_fence_destroy(struct vmw_fence_obj *fence)
+{
+	struct vmw_user_fence *ufence =
+		container_of(fence, struct vmw_user_fence, fence);
+	struct vmw_fence_manager *fman = fence->fman;
+
+	kfree(ufence);
+	/*
+	 * Free kernel space accounting.
+	 */
+	ttm_mem_global_free(vmw_mem_glob(fman->dev_priv),
+			    fman->user_fence_size);
+}
+
+static void vmw_user_fence_base_release(struct ttm_base_object **p_base)
+{
+	struct ttm_base_object *base = *p_base;
+	struct vmw_user_fence *ufence =
+		container_of(base, struct vmw_user_fence, base);
+	struct vmw_fence_obj *fence = &ufence->fence;
+
+	*p_base = NULL;
+	vmw_fence_obj_unreference(&fence);
+}
+
+int vmw_user_fence_create(struct drm_file *file_priv,
+			  struct vmw_fence_manager *fman,
+			  uint32_t seqno,
+			  uint32_t mask,
+			  struct vmw_fence_obj **p_fence,
+			  uint32_t *p_handle)
+{
+	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	struct vmw_user_fence *ufence;
+	struct vmw_fence_obj *tmp;
+	struct ttm_mem_global *mem_glob = vmw_mem_glob(fman->dev_priv);
+	int ret;
+
+	/*
+	 * Kernel memory space accounting, since this object may
+	 * be created by a user-space request.
+	 */
+
+	ret = ttm_mem_global_alloc(mem_glob, fman->user_fence_size,
+				   false, false);
+	if (unlikely(ret != 0))
+		return ret;
+
+	ufence = kzalloc(sizeof(*ufence), GFP_KERNEL);
+	if (unlikely(ufence == NULL)) {
+		ret = -ENOMEM;
+		goto out_no_object;
+	}
+
+	ret = vmw_fence_obj_init(fman, &ufence->fence, seqno,
+				 mask, vmw_user_fence_destroy);
+	if (unlikely(ret != 0)) {
+		kfree(ufence);
+		goto out_no_object;
+	}
+
+	/*
+	 * The base object holds a reference which is freed in
+	 * vmw_user_fence_base_release.
+	 */
+	tmp = vmw_fence_obj_reference(&ufence->fence);
+	ret = ttm_base_object_init(tfile, &ufence->base, false,
+				   VMW_RES_FENCE,
+				   &vmw_user_fence_base_release, NULL);
+
+
+	if (unlikely(ret != 0)) {
+		/*
+		 * Free the base object's reference
+		 */
+		vmw_fence_obj_unreference(&tmp);
+		goto out_err;
+	}
+
+	*p_fence = &ufence->fence;
+	*p_handle = ufence->base.hash.key;
+
+	return 0;
+out_err:
+	tmp = &ufence->fence;
+	vmw_fence_obj_unreference(&tmp);
+out_no_object:
+	ttm_mem_global_free(mem_glob, fman->user_fence_size);
+	return ret;
+}
+
+
+/**
+ * vmw_fence_fifo_down - signal all unsignaled fence objects.
+ */
+
+void vmw_fence_fifo_down(struct vmw_fence_manager *fman)
+{
+	unsigned long irq_flags;
+	struct list_head action_list;
+	int ret;
+
+	/*
+	 * The list may be altered while we traverse it, so always
+	 * restart when we've released the fman->lock.
+	 */
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	fman->fifo_down = true;
+	while (!list_empty(&fman->fence_list)) {
+		struct vmw_fence_obj *fence =
+			list_entry(fman->fence_list.prev, struct vmw_fence_obj,
+				   head);
+		kref_get(&fence->kref);
+		spin_unlock_irq(&fman->lock);
+
+		ret = vmw_fence_obj_wait(fence, fence->signal_mask,
+					 false, false,
+					 VMW_FENCE_WAIT_TIMEOUT);
+
+		if (unlikely(ret != 0)) {
+			list_del_init(&fence->head);
+			fence->signaled |= DRM_VMW_FENCE_FLAG_EXEC;
+			INIT_LIST_HEAD(&action_list);
+			list_splice_init(&fence->seq_passed_actions,
+					 &action_list);
+			vmw_fences_perform_actions(fman, &action_list);
+			wake_up_all(&fence->queue);
+		}
+
+		spin_lock_irq(&fman->lock);
+
+		BUG_ON(!list_empty(&fence->head));
+		kref_put(&fence->kref, vmw_fence_obj_destroy_locked);
+	}
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+}
+
+void vmw_fence_fifo_up(struct vmw_fence_manager *fman)
+{
+	unsigned long irq_flags;
+
+	spin_lock_irqsave(&fman->lock, irq_flags);
+	fman->fifo_down = false;
+	spin_unlock_irqrestore(&fman->lock, irq_flags);
+}
+
+
+int vmw_fence_obj_wait_ioctl(struct drm_device *dev, void *data,
+			     struct drm_file *file_priv)
+{
+	struct drm_vmw_fence_wait_arg *arg =
+	    (struct drm_vmw_fence_wait_arg *)data;
+	unsigned long timeout;
+	struct ttm_base_object *base;
+	struct vmw_fence_obj *fence;
+	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	int ret;
+	uint64_t wait_timeout = ((uint64_t)arg->timeout_us * HZ);
+
+	/*
+	 * 64-bit division not present on 32-bit systems, so do an
+	 * approximation. (Divide by 1000000).
+	 */
+
+	wait_timeout = (wait_timeout >> 20) + (wait_timeout >> 24) -
+	  (wait_timeout >> 26);
+
+	if (!arg->cookie_valid) {
+		arg->cookie_valid = 1;
+		arg->kernel_cookie = jiffies + wait_timeout;
+	}
+
+	base = ttm_base_object_lookup(tfile, arg->handle);
+	if (unlikely(base == NULL)) {
+		printk(KERN_ERR "Wait invalid fence object handle "
+		       "0x%08lx.\n",
+		       (unsigned long)arg->handle);
+		return -EINVAL;
+	}
+
+	fence = &(container_of(base, struct vmw_user_fence, base)->fence);
+
+	timeout = jiffies;
+	if (time_after_eq(timeout, (unsigned long)arg->kernel_cookie)) {
+		ret = ((vmw_fence_obj_signaled(fence, arg->flags)) ?
+		       0 : -EBUSY);
+		goto out;
+	}
+
+	timeout = (unsigned long)arg->kernel_cookie - timeout;
+
+	ret = vmw_fence_obj_wait(fence, arg->flags, arg->lazy, true, timeout);
+
+out:
+	ttm_base_object_unref(&base);
+
+	/*
+	 * Optionally unref the fence object.
+	 */
+
+	if (ret == 0 && (arg->wait_options & DRM_VMW_WAIT_OPTION_UNREF))
+		return ttm_ref_object_base_unref(tfile, arg->handle,
+						 TTM_REF_USAGE);
+	return ret;
+}
+
+int vmw_fence_obj_signaled_ioctl(struct drm_device *dev, void *data,
+				 struct drm_file *file_priv)
+{
+	struct drm_vmw_fence_signaled_arg *arg =
+		(struct drm_vmw_fence_signaled_arg *) data;
+	struct ttm_base_object *base;
+	struct vmw_fence_obj *fence;
+	struct vmw_fence_manager *fman;
+	struct ttm_object_file *tfile = vmw_fpriv(file_priv)->tfile;
+	struct vmw_private *dev_priv = vmw_priv(dev);
+
+	base = ttm_base_object_lookup(tfile, arg->handle);
+	if (unlikely(base == NULL)) {
+		printk(KERN_ERR "Fence signaled invalid fence object handle "
+		       "0x%08lx.\n",
+		       (unsigned long)arg->handle);
+		return -EINVAL;
+	}
+
+	fence = &(container_of(base, struct vmw_user_fence, base)->fence);
+	fman = fence->fman;
+
+	arg->signaled = vmw_fence_obj_signaled(fence, arg->flags);
+	spin_lock_irq(&fman->lock);
+
+	arg->signaled_flags = fence->signaled;
+	arg->passed_seqno = dev_priv->last_read_seqno;
+	spin_unlock_irq(&fman->lock);
+
+	ttm_base_object_unref(&base);
+
+	return 0;
+}
+
+
+int vmw_fence_obj_unref_ioctl(struct drm_device *dev, void *data,
+			      struct drm_file *file_priv)
+{
+	struct drm_vmw_fence_arg *arg =
+		(struct drm_vmw_fence_arg *) data;
+
+	return ttm_ref_object_base_unref(vmw_fpriv(file_priv)->tfile,
+					 arg->handle,
+					 TTM_REF_USAGE);
+}