Re-land^2 WebVR compositor bypass via BrowserMain context + mailbox

chrome/browser/android/vr_shell/vr_shell_gl.cc

 // Copyright 2016 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
 #include "chrome/browser/android/vr_shell/vr_shell_gl.h"
 
 #include <limits>
 #include <utility>
 
 #include "base/android/jni_android.h"
 #include "base/callback_helpers.h"
 #include "base/memory/ptr_util.h"
 #include "base/metrics/histogram_macros.h"
 #include "base/threading/thread_task_runner_handle.h"
+#include "chrome/browser/android/vr_shell/mailbox_to_surface_bridge.h"
 #include "chrome/browser/android/vr_shell/ui_elements.h"
 #include "chrome/browser/android/vr_shell/ui_scene.h"
 #include "chrome/browser/android/vr_shell/vr_controller.h"
 #include "chrome/browser/android/vr_shell/vr_gl_util.h"
 #include "chrome/browser/android/vr_shell/vr_math.h"
 #include "chrome/browser/android/vr_shell/vr_shell.h"
 #include "chrome/browser/android/vr_shell/vr_shell_delegate.h"
 #include "chrome/browser/android/vr_shell/vr_shell_renderer.h"
 #include "device/vr/android/gvr/gvr_device.h"
 #include "third_party/WebKit/public/platform/WebInputEvent.h"
@@ skipping 48 matching lines @@
 
 // The GVR viewport list has two entries (left eye and right eye) for each
 // GVR buffer.
 static constexpr int kViewportListPrimaryOffset = 0;
 static constexpr int kViewportListHeadlockedOffset = 2;
 
 // Buffer size large enough to handle the current backlog of poses which is
 // 2-3 frames.
 static constexpr unsigned kPoseRingBufferSize = 8;
 
-// Magic numbers used to mark valid pose index values encoded in frame
-// data. Must match the magic numbers used in blink's VRDisplay.cpp.
-static constexpr std::array<uint8_t, 2> kWebVrPosePixelMagicNumbers{{42, 142}};
+// Default downscale factor for computing the recommended WebVR
+// renderWidth/Height from the 1:1 pixel mapped size. Using a rather
+// aggressive downscale due to the high overhead of copying pixels
+// twice before handing off to GVR. For comparison, the polyfill
+// uses approximately 0.55 on a Pixel XL.
+static constexpr float kWebVrRecommendedResolutionScale = 0.5;

[klausw, 2017/03/10 02:49:32]

Moved to vr_shell.gl

 
 float Distance(const gvr::Vec3f& vec1, const gvr::Vec3f& vec2) {
   float xdiff = (vec1.x - vec2.x);
   float ydiff = (vec1.y - vec2.y);
   float zdiff = (vec1.z - vec2.z);
   float scale = xdiff * xdiff + ydiff * ydiff + zdiff * zdiff;
   return std::sqrt(scale);
 }
 
 // Generate a quaternion representing the rotation from the negative Z axis
@@ skipping 69 matching lines @@
       binding_(this),
       weak_vr_shell_(weak_vr_shell),
       delegate_provider_(delegate_provider),
       main_thread_task_runner_(std::move(main_thread_task_runner)),
       weak_ptr_factory_(this) {
   GvrInit(gvr_api);
 }
 
 VrShellGl::~VrShellGl() {
   vsync_task_.Cancel();
+  // TODO(mthiesse): Can we omit the Close() here? Concern is that if
+  // both ends of the connection ever live in the same process for
+  // some reason, we could receive another VSync request in response
+  // to the closing message in the destructor but fail to respond to
+  // the callback.
+  binding_.Close();
   if (!callback_.is_null()) {
     // When this VSync provider is going away we have to respond to pending
     // callbacks, so instead of providing a VSync, tell the requester to try
     // again. A VSyncProvider is guaranteed to exist, so the request in response
     // to this message will go through some other VSyncProvider.
     base::ResetAndReturn(&callback_)
         .Run(nullptr, base::TimeDelta(), -1,
-             device::mojom::VRVSyncProvider::Status::RETRY);
-  }
-  if (binding_.is_bound()) {
-    main_thread_task_runner_->PostTask(
-        FROM_HERE,
-        base::Bind(&VrShellDelegate::OnVRVsyncProviderRequest,
-                   delegate_provider_, base::Passed(binding_.Unbind())));
+             device::mojom::VRVSyncProvider::Status::CLOSING);
   }
 }
 
 void VrShellGl::Initialize() {
   scene_.reset(new UiScene);
 
   if (surfaceless_rendering_) {
     // If we're rendering surfaceless, we'll never get a java surface to render
     // into, so we can initialize GL right away.
     InitializeGl(nullptr);
@@ skipping 26 matching lines @@
     LOG(ERROR) << "gl::init::CreateGLContext failed";
     ForceExitVr();
     return;
   }
   if (!context_->MakeCurrent(surface_.get())) {
     LOG(ERROR) << "gl::GLContext::MakeCurrent() failed";
     ForceExitVr();
     return;
   }
 
-  unsigned int textures[2];
-  glGenTextures(2, textures);
+  unsigned int textures[3];
+  glGenTextures(3, textures);
   ui_texture_id_ = textures[0];
   content_texture_id_ = textures[1];
+  webvr_texture_id_ = textures[2];
   ui_surface_texture_ = gl::SurfaceTexture::Create(ui_texture_id_);
   content_surface_texture_ = gl::SurfaceTexture::Create(content_texture_id_);
+  webvr_surface_texture_ = gl::SurfaceTexture::Create(webvr_texture_id_);
   CreateUiSurface();
   CreateContentSurface();
   ui_surface_texture_->SetFrameAvailableCallback(base::Bind(
       &VrShellGl::OnUIFrameAvailable, weak_ptr_factory_.GetWeakPtr()));
   content_surface_texture_->SetFrameAvailableCallback(base::Bind(
       &VrShellGl::OnContentFrameAvailable, weak_ptr_factory_.GetWeakPtr()));
+  webvr_surface_texture_->SetFrameAvailableCallback(base::Bind(
+      &VrShellGl::OnWebVRFrameAvailable, weak_ptr_factory_.GetWeakPtr()));
+  ui_surface_texture_->SetDefaultBufferSize(ui_tex_physical_size_.width,
+                                            ui_tex_physical_size_.height);
   content_surface_texture_->SetDefaultBufferSize(
       content_tex_physical_size_.width, content_tex_physical_size_.height);
-  ui_surface_texture_->SetDefaultBufferSize(ui_tex_physical_size_.width,
-                                            ui_tex_physical_size_.height);
   InitializeRenderer();
 
+  // Pick a reasonable default size for the WebVR transfer surface
+  // based on a downscaled 1:1 render resolution. This size will also
+  // be reported to the client via CreateVRDisplayInfo as the
+  // client-recommended renderWidth/renderHeight and for the GVR
+  // framebuffer. If the client chooses a different size or resizes it
+  // while presenting, we'll resize the transfer surface and GVR
+  // framebuffer to match.
+  gvr::Sizei render_target_size =
+      gvr_api_->GetMaximumEffectiveRenderTargetSize();
+  gvr::Sizei webvr_size = {static_cast<int>(render_target_size.width *
+                                            kWebVrRecommendedResolutionScale),
+                           static_cast<int>(render_target_size.height *
+                                            kWebVrRecommendedResolutionScale)};
+
+  // TODO(klausw,crbug.com/699350): should we round the recommended
+  // size to a multiple of 2^N pixels to be friendlier to the GPU? The
+  // exact size doesn't matter.
+
+  CreateOrResizeWebVRSurface(webvr_size);
+
   vsync_task_.Reset(base::Bind(&VrShellGl::OnVSync, base::Unretained(this)));
   OnVSync();
 
   ready_to_draw_ = true;
 }
 
 void VrShellGl::CreateContentSurface() {
   content_surface_ =
       base::MakeUnique<gl::ScopedJavaSurface>(content_surface_texture_.get());
   main_thread_task_runner_->PostTask(
       FROM_HERE, base::Bind(&VrShell::ContentSurfaceChanged, weak_vr_shell_,
                             content_surface_->j_surface().obj()));
 }
 
 void VrShellGl::CreateUiSurface() {
   ui_surface_ =
       base::MakeUnique<gl::ScopedJavaSurface>(ui_surface_texture_.get());
   main_thread_task_runner_->PostTask(
       FROM_HERE, base::Bind(&VrShell::UiSurfaceChanged, weak_vr_shell_,
                             ui_surface_->j_surface().obj()));
 }
 
+void VrShellGl::CreateOrResizeWebVRSurface(const gvr::Sizei& size) {
+  if (!webvr_surface_texture_) {
+    DLOG(ERROR) << "No WebVR surface texture available";
+    return;
+  }
+
+  // ContentPhysicalBoundsChanged is getting called twice with
+  // identical sizes? Avoid thrashing the existing context.
+  if (size == webvr_surface_size_) {
+    return;
+  }
+
+  if (!size.width || !size.height) {
+    // Invalid size, defer until a new size arrives on a future bounds update.
+    return;
+  }
+
+  webvr_surface_texture_->SetDefaultBufferSize(size.width, size.height);
+  webvr_surface_size_ = size;
+
+  if (mailbox_bridge_) {
+    mailbox_bridge_->ResizeSurface(size.width, size.height);
+  } else {
+    mailbox_bridge_ = base::MakeUnique<MailboxToSurfaceBridge>();
+    mailbox_bridge_->CreateSurface(webvr_surface_texture_.get());
+  }
+}
+
+void VrShellGl::SubmitWebVRFrame(int16_t frame_index,
+                                 const gpu::MailboxHolder& mailbox) {
+  TRACE_EVENT0("gpu", "VrShellGl::SubmitWebVRFrame");
+
+  // Swapping twice on a Surface without calling updateTexImage in
+  // between can lose frames, so don't draw+swap if we already have
+  // a pending frame we haven't consumed yet.
+  bool swapped = false;
+  if (pending_frames_.empty()) {
+    swapped = mailbox_bridge_->CopyMailboxToSurfaceAndSwap(mailbox);
+    if (swapped) {
+      // Tell OnWebVRFrameAvailable to expect a new frame to arrive on
+      // the SurfaceTexture, and save the associated frame index.
+      pending_frames_.emplace(frame_index);
+    }
+  }
+  // Always notify the client that we're done with the mailbox even
+  // if we haven't drawn it, so that it's eligible for destruction.
+  submit_client_->OnSubmitFrameTransferred();
+  if (!swapped) {
+    // We dropped without drawing, report this as completed rendering
+    // now to unblock the client. We're not going to receive it in
+    // OnWebVRFrameAvailable where we'd normally report that.
+    submit_client_->OnSubmitFrameRendered();
+  }
+
+  TRACE_EVENT0("gpu", "VrShellGl::glFinish");
+  // This is a load-bearing glFinish, please don't remove it without
+  // before/after timing comparisons. Goal is to clear the GPU queue
+  // of the native GL context to avoid stalls later in GVR frame
+  // acquire/submit.
+  glFinish();
+}
+
+void VrShellGl::SetSubmitClient(
+    device::mojom::VRSubmitFrameClientPtrInfo submit_client_info) {
+  submit_client_.Bind(std::move(submit_client_info));
+}
+
 void VrShellGl::OnUIFrameAvailable() {
   ui_surface_texture_->UpdateTexImage();
 }
 
 void VrShellGl::OnContentFrameAvailable() {
   content_surface_texture_->UpdateTexImage();
   received_frame_ = true;
 }
 
-bool VrShellGl::GetPixelEncodedFrameIndex(uint16_t* frame_index) {
-  TRACE_EVENT0("gpu", "VrShellGl::GetPixelEncodedFrameIndex");
-  if (!received_frame_) {
-    if (last_frame_index_ == (uint16_t)-1)
-      return false;
-    *frame_index = last_frame_index_;
-    return true;
+void VrShellGl::OnWebVRFrameAvailable() {
+  // A "while" loop here is a bad idea. It's legal to call
+  // UpdateTexImage repeatedly even if no frames are available, but
+  // that does *not* wait for a new frame, it just reuses the most
+  // recent one. That would mess up the count.
+  if (pending_frames_.empty()) {
+    // We're expecting a frame, but it's not here yet. Retry in OnVsync.
+    ++premature_received_frames_;
+    return;
   }
-  received_frame_ = false;
 
-  // Read the pose index encoded in a bottom left pixel as color values.
-  // See also third_party/WebKit/Source/modules/vr/VRDisplay.cpp which
-  // encodes the pose index, and device/vr/android/gvr/gvr_device.cc
-  // which tracks poses. Returns the low byte (0..255) if valid, or -1
-  // if not valid due to bad magic number.
-  uint8_t pixels[4];
-  // Assume we're reading from the framebuffer we just wrote to.
-  // That's true currently, we may need to use glReadBuffer(GL_BACK)
-  // or equivalent if the rendering setup changes in the future.
-  glReadPixels(0, 0, 1, 1, GL_RGBA, GL_UNSIGNED_BYTE, pixels);
+  webvr_surface_texture_->UpdateTexImage();
+  int frame_index = pending_frames_.front();
+  TRACE_EVENT1("gpu", "VrShellGl::OnWebVRFrameAvailable", "frame", frame_index);
+  pending_frames_.pop();
 
-  // Check for the magic number written by VRDevice.cpp on submit.
-  // This helps avoid glitches from garbage data in the render
-  // buffer that can appear during initialization or resizing. These
-  // often appear as flashes of all-black or all-white pixels.
-  if (pixels[1] == kWebVrPosePixelMagicNumbers[0] &&
-      pixels[2] == kWebVrPosePixelMagicNumbers[1]) {
-    // Pose is good.
-    *frame_index = pixels[0];
-    last_frame_index_ = pixels[0];
-    return true;
-  }
-  VLOG(1) << "WebVR: reject decoded pose index " << static_cast<int>(pixels[0])
-          << ", bad magic number " << static_cast<int>(pixels[1]) << ", "
-          << static_cast<int>(pixels[2]);
-  return false;
+  // It is legal for the WebVR client to submit a new frame now, since
+  // we've consumed the image. TODO(klausw): would timing be better if
+  // we move the "rendered" notification after draw, or suppress
+  // the next vsync until that's done?
+
+  submit_client_->OnSubmitFrameRendered();
+
+  DrawFrame(frame_index);
 }
 
 void VrShellGl::GvrInit(gvr_context* gvr_api) {
   gvr_api_ = gvr::GvrApi::WrapNonOwned(gvr_api);
   controller_.reset(new VrController(gvr_api));
 
   ViewerType viewerType;
   switch (gvr_api_->GetViewerType()) {
     case gvr::ViewerType::GVR_VIEWER_TYPE_DAYDREAM:
       viewerType = ViewerType::DAYDREAM;
       break;
     case gvr::ViewerType::GVR_VIEWER_TYPE_CARDBOARD:
       viewerType = ViewerType::CARDBOARD;
       break;
     default:
       NOTREACHED();
       viewerType = ViewerType::UNKNOWN_TYPE;
       break;
   }
   UMA_HISTOGRAM_ENUMERATION("VRViewerType", static_cast<int>(viewerType),
                             static_cast<int>(ViewerType::VIEWER_TYPE_MAX));
 }
 
 void VrShellGl::InitializeRenderer() {
-  // While WebVR is going through the compositor path, it shares
-  // the same texture ID. This will change once it gets its own
-  // surface, but store it separately to avoid future confusion.
-  // TODO(klausw,crbug.com/655722): remove this.
-  webvr_texture_id_ = content_texture_id_;
-
   gvr_api_->InitializeGl();
   webvr_head_pose_.assign(kPoseRingBufferSize,
                           gvr_api_->GetHeadSpaceFromStartSpaceRotation(
                               gvr::GvrApi::GetTimePointNow()));
 
   std::vector<gvr::BufferSpec> specs;
   // For kFramePrimaryBuffer (primary VrShell and WebVR content)
   specs.push_back(gvr_api_->CreateBufferSpec());
   render_size_primary_ = specs[kFramePrimaryBuffer].GetSize();
+  render_size_vrshell_ = render_size_primary_;
 
   // For kFrameHeadlockedBuffer (for WebVR insecure content warning).
   // Set this up at fixed resolution, the (smaller) FOV gets set below.
   specs.push_back(gvr_api_->CreateBufferSpec());
   specs.back().SetSize(kHeadlockedBufferDimensions);
   render_size_headlocked_ = specs[kFrameHeadlockedBuffer].GetSize();
 
   swap_chain_.reset(new gvr::SwapChain(gvr_api_->CreateSwapChain(specs)));
 
   vr_shell_renderer_.reset(new VrShellRenderer());
@@ skipping 256 matching lines @@
                             std::unique_ptr<blink::WebInputEvent> event) {
   DCHECK(input_target != InputTarget::NONE);
   auto&& target = input_target == InputTarget::CONTENT
                       ? &VrShell::ProcessContentGesture
                       : &VrShell::ProcessUIGesture;
   main_thread_task_runner_->PostTask(
       FROM_HERE,
       base::Bind(target, weak_vr_shell_, base::Passed(std::move(event))));
 }
 
-void VrShellGl::DrawFrame() {
-  TRACE_EVENT0("gpu", "VrShellGl::DrawFrame");
+void VrShellGl::DrawFrame(int16_t frame_index) {
+  TRACE_EVENT1("gpu", "VrShellGl::DrawFrame", "frame", frame_index);
 
   // Reset the viewport list to just the pair of viewports for the
   // primary buffer each frame. Head-locked viewports get added by
   // DrawVrShell if needed.
   buffer_viewport_list_->SetToRecommendedBufferViewports();
 
+  // If needed, resize the primary buffer for use with WebVR.
+  if (web_vr_mode_) {
+    if (render_size_primary_ != webvr_surface_size_) {
+      if (!webvr_surface_size_.width) {
+        return;
+      }
+      render_size_primary_ = webvr_surface_size_;
+      swap_chain_->ResizeBuffer(kFramePrimaryBuffer, render_size_primary_);
+    }
+  } else {
+    if (render_size_primary_ != render_size_vrshell_) {
+      render_size_primary_ = render_size_vrshell_;
+      swap_chain_->ResizeBuffer(kFramePrimaryBuffer, render_size_primary_);
+    }
+  }
+
+  TRACE_EVENT_BEGIN0("gpu", "VrShellGl::AcquireFrame");
   gvr::Frame frame = swap_chain_->AcquireFrame();
+  TRACE_EVENT_END0("gpu", "VrShellGl::AcquireFrame");
   if (!frame.is_valid()) {
     return;
   }
   frame.BindBuffer(kFramePrimaryBuffer);
   if (web_vr_mode_) {
     DrawWebVr();
   }
 
-  uint16_t frame_index;
   gvr::Mat4f head_pose;
 
   // When using async reprojection, we need to know which pose was used in
-  // the WebVR app for drawing this frame. Due to unknown amounts of
-  // buffering in the compositor and SurfaceTexture, we read the pose number
-  // from a corner pixel. There's no point in doing this for legacy
-  // distortion rendering since that doesn't need a pose, and reading back
-  // pixels is an expensive operation. TODO(klausw,crbug.com/655722): stop
-  // doing this once we have working no-compositor rendering for WebVR.
-  if (web_vr_mode_ && gvr_api_->GetAsyncReprojectionEnabled() &&
-      GetPixelEncodedFrameIndex(&frame_index)) {
+  // the WebVR app for drawing this frame. Only needed if reprojection is
+  // in use.
+  if (web_vr_mode_ && gvr_api_->GetAsyncReprojectionEnabled()) {
     static_assert(!((kPoseRingBufferSize - 1) & kPoseRingBufferSize),
                   "kPoseRingBufferSize must be a power of 2");
     head_pose = webvr_head_pose_[frame_index % kPoseRingBufferSize];
     // Process all pending_bounds_ changes targeted for before this frame, being
     // careful of wrapping frame indices.
     static constexpr unsigned max =
         std::numeric_limits<decltype(frame_index_)>::max();
     static_assert(max > kPoseRingBufferSize * 2,
                   "To detect wrapping, kPoseRingBufferSize must be smaller "
                   "than half of frame_index_ range.");
     while (!pending_bounds_.empty()) {
       uint16_t index = pending_bounds_.front().first;
       // If index is less than the frame_index it's possible we've wrapped, so
       // we extend the range and 'un-wrap' to account for this.
       if (index < frame_index)
         index += max;
       // If the pending bounds change is for an upcoming frame within our buffer
       // size, wait to apply it. Otherwise, apply it immediately. This
       // guarantees that even if we miss many frames, the queue can't fill up
       // with stale bounds.
       if (index > frame_index && index <= frame_index + kPoseRingBufferSize)
         break;
 
-      const BoundsPair& bounds = pending_bounds_.front().second;
-      webvr_left_viewport_->SetSourceUv(bounds.first);
-      webvr_right_viewport_->SetSourceUv(bounds.second);
+      const WebVrBounds& bounds = pending_bounds_.front().second;
+      webvr_left_viewport_->SetSourceUv(bounds.left_bounds);
+      webvr_right_viewport_->SetSourceUv(bounds.right_bounds);
+      CreateOrResizeWebVRSurface(bounds.source_size);
       pending_bounds_.pop();
     }
     buffer_viewport_list_->SetBufferViewport(GVR_LEFT_EYE,
                                              *webvr_left_viewport_);
     buffer_viewport_list_->SetBufferViewport(GVR_RIGHT_EYE,
                                              *webvr_right_viewport_);
   } else {
     gvr::ClockTimePoint target_time = gvr::GvrApi::GetTimePointNow();
     target_time.monotonic_system_time_nanos += kPredictionTimeWithoutVsyncNanos;
     head_pose = gvr_api_->GetHeadSpaceFromStartSpaceRotation(target_time);
   }
 
   gvr::Vec3f position = GetTranslation(head_pose);
   if (position.x == 0.0f && position.y == 0.0f && position.z == 0.0f) {
     // This appears to be a 3DOF pose without a neck model. Add one.
     // The head pose has redundant data. Assume we're only using the
     // object_from_reference_matrix, we're not updating position_external.
     // TODO: Not sure what object_from_reference_matrix is. The new api removed
     // it. For now, removing it seems working fine.
     gvr_api_->ApplyNeckModel(head_pose, 1.0f);
   }
 
   // Update the render position of all UI elements (including desktop).
   scene_->UpdateTransforms(TimeInMicroseconds());
 
-  UpdateController(GetForwardVector(head_pose));
+  {
+    TRACE_EVENT0("gpu", "VrShellGl::UpdateController");
+    UpdateController(GetForwardVector(head_pose));
+  }
 
-  DrawVrShell(head_pose, frame);
+  // Finish drawing in the primary buffer, and draw the headlocked buffer
+  // if needed. This must be the last drawing call, this method will
+  // return with no frame being bound.
+  DrawVrShellAndUnbind(head_pose, frame);
 
-  frame.Unbind();
-  frame.Submit(*buffer_viewport_list_, head_pose);
+  {
+    TRACE_EVENT0("gpu", "VrShellGl::Submit");
+    frame.Submit(*buffer_viewport_list_, head_pose);
+  }
 
   // No need to swap buffers for surfaceless rendering.
   if (!surfaceless_rendering_) {
     // TODO(mthiesse): Support asynchronous SwapBuffers.
+    TRACE_EVENT0("gpu", "VrShellGl::SwapBuffers");
     surface_->SwapBuffers();
   }
 }
 
-void VrShellGl::DrawVrShell(const gvr::Mat4f& head_pose, gvr::Frame& frame) {
+void VrShellGl::DrawVrShellAndUnbind(const gvr::Mat4f& head_pose,
+                                     gvr::Frame& frame) {
   TRACE_EVENT0("gpu", "VrShellGl::DrawVrShell");
   std::vector<const ContentRectangle*> head_locked_elements;
   std::vector<const ContentRectangle*> world_elements;
   for (const auto& rect : scene_->GetUiElements()) {
     if (!rect->IsVisible())
       continue;
     if (rect->lock_to_fov) {
       head_locked_elements.push_back(rect.get());
     } else {
       world_elements.push_back(rect.get());
@@ skipping 16 matching lines @@
 
     const Colorf& backgroundColor = scene_->GetBackgroundColor();
     glClearColor(backgroundColor.r, backgroundColor.g, backgroundColor.b,
                  backgroundColor.a);
     glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
   }
   if (!world_elements.empty()) {
     DrawUiView(&head_pose, world_elements, render_size_primary_,
                kViewportListPrimaryOffset);
   }
+  frame.Unbind();  // Done with the primary buffer.
 
   if (!head_locked_elements.empty()) {
     // Add head-locked viewports. The list gets reset to just
     // the recommended viewports (for the primary buffer) each frame.
     buffer_viewport_list_->SetBufferViewport(
         kViewportListHeadlockedOffset + GVR_LEFT_EYE,
         *headlocked_left_viewport_);
     buffer_viewport_list_->SetBufferViewport(
         kViewportListHeadlockedOffset + GVR_RIGHT_EYE,
         *headlocked_right_viewport_);
 
     // Bind the headlocked framebuffer.
-    // TODO(mthiesse): We don't unbind this? Maybe some cleanup is in order
-    // here.
     frame.BindBuffer(kFrameHeadlockedBuffer);
     glClearColor(0.0f, 0.0f, 0.0f, 0.0f);
     glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
     DrawUiView(nullptr, head_locked_elements, render_size_headlocked_,
                kViewportListHeadlockedOffset);
+    frame.Unbind();  // Done with the headlocked buffer.
   }
 }
 
-gvr::Sizei VrShellGl::GetWebVRCompositorSurfaceSize() {
-  // This is a stopgap while we're using the WebVR compositor rendering path.
-  // TODO(klausw,crbug.com/655722): Remove this method and member once we're
-  // using a separate WebVR render surface.
-  return content_tex_physical_size_;
-}
-
 void VrShellGl::DrawUiView(const gvr::Mat4f* head_pose,
                            const std::vector<const ContentRectangle*>& elements,
                            const gvr::Sizei& render_size,
                            int viewport_offset) {
   TRACE_EVENT0("gpu", "VrShellGl::DrawUiView");
 
   gvr::Mat4f view_matrix;
   if (head_pose) {
     view_matrix = *head_pose;
   } else {
@@ skipping 185 matching lines @@
 void VrShellGl::DrawWebVr() {
   TRACE_EVENT0("gpu", "VrShellGl::DrawWebVr");
   // Don't need face culling, depth testing, blending, etc. Turn it all off.
   glDisable(GL_CULL_FACE);
   glDepthMask(GL_FALSE);
   glDisable(GL_DEPTH_TEST);
   glDisable(GL_SCISSOR_TEST);
   glDisable(GL_BLEND);
   glDisable(GL_POLYGON_OFFSET_FILL);
 
-  glViewport(0, 0, render_size_primary_.width, render_size_primary_.height);
+  glViewport(0, 0, webvr_surface_size_.width, webvr_surface_size_.height);
   vr_shell_renderer_->GetWebVrRenderer()->Draw(webvr_texture_id_);
 }
 
 void VrShellGl::OnTriggerEvent() {
   // Set a flag to handle this on the render thread at the next frame.
   touch_pending_ = true;
 }
 
 void VrShellGl::OnPause() {
   vsync_task_.Cancel();
@@ skipping 10 matching lines @@
     OnVSync();
   }
 }
 
 void VrShellGl::SetWebVrMode(bool enabled) {
   web_vr_mode_ = enabled;
 }
 
 void VrShellGl::UpdateWebVRTextureBounds(int16_t frame_index,
                                          const gvr::Rectf& left_bounds,
-                                         const gvr::Rectf& right_bounds) {
+                                         const gvr::Rectf& right_bounds,
+                                         const gvr::Sizei& source_size) {
   if (frame_index < 0) {
     webvr_left_viewport_->SetSourceUv(left_bounds);
     webvr_right_viewport_->SetSourceUv(right_bounds);
   } else {
     pending_bounds_.emplace(
-        std::make_pair(frame_index, std::make_pair(left_bounds, right_bounds)));
+        frame_index, WebVrBounds(left_bounds, right_bounds, source_size));
   }
 }
 
 void VrShellGl::ContentBoundsChanged(int width, int height) {
   TRACE_EVENT0("gpu", "VrShellGl::ContentBoundsChanged");
   content_tex_css_width_ = width;
   content_tex_css_height_ = height;
 }
 
 void VrShellGl::ContentPhysicalBoundsChanged(int width, int height) {
@@ skipping 13 matching lines @@
     ui_surface_texture_->SetDefaultBufferSize(width, height);
   ui_tex_physical_size_.width = width;
   ui_tex_physical_size_.height = height;
 }
 
 base::WeakPtr<VrShellGl> VrShellGl::GetWeakPtr() {
   return weak_ptr_factory_.GetWeakPtr();
 }
 
 void VrShellGl::OnVSync() {
+  while (premature_received_frames_ > 0) {
+    TRACE_EVENT0("gpu", "VrShellGl::OnWebVRFrameAvailableRetry");
+    --premature_received_frames_;
+    OnWebVRFrameAvailable();
+  }
+
   base::TimeTicks now = base::TimeTicks::Now();
   base::TimeTicks target;
 
   // Don't send VSyncs until we have a timebase/interval.
   if (vsync_interval_.is_zero())
     return;
   target = now + vsync_interval_;
   int64_t intervals = (target - vsync_timebase_) / vsync_interval_;
   target = vsync_timebase_ + intervals * vsync_interval_;
   task_runner_->PostDelayedTask(FROM_HERE, vsync_task_.callback(),
                                 target - now);
 
   base::TimeDelta time = intervals * vsync_interval_;
   if (!callback_.is_null()) {
     SendVSync(time, base::ResetAndReturn(&callback_));
   } else {
     pending_vsync_ = true;
     pending_time_ = time;
   }
-  DrawFrame();
+  if (!web_vr_mode_) {
+    DrawFrame(-1);
+  }
 }
 
 void VrShellGl::OnRequest(device::mojom::VRVSyncProviderRequest request) {
   binding_.Close();
   binding_.Bind(std::move(request));
 }
 
 void VrShellGl::GetVSync(const GetVSyncCallback& callback) {
   if (!pending_vsync_) {
     if (!callback_.is_null()) {
@@ skipping 49 matching lines @@
 void VrShellGl::ResetPose() {
   // Should never call RecenterTracking when using with Daydream viewers. On
   // those devices recentering should only be done via the controller.
   if (gvr_api_ && gvr_api_->GetViewerType() == GVR_VIEWER_TYPE_CARDBOARD)
     gvr_api_->RecenterTracking();
 }
 
 void VrShellGl::CreateVRDisplayInfo(
     const base::Callback<void(device::mojom::VRDisplayInfoPtr)>& callback,
     uint32_t device_id) {
+  // This assumes that the initial webvr_surface_size_ was set to the
+  // appropriate recommended render resolution as the default size during
+  // InitializeGl. Revisit if the initialization order changes.
   device::mojom::VRDisplayInfoPtr info = VrShell::CreateVRDisplayInfo(
-      gvr_api_.get(), content_tex_physical_size_, device_id);
+      gvr_api_.get(), webvr_surface_size_, device_id);
   main_thread_task_runner_->PostTask(
       FROM_HERE,
       base::Bind(&RunVRDisplayInfoCallback, callback, base::Passed(&info)));
 }
 
 }  // namespace vr_shell