gpu: Fix some context lost marking glitches+leaks and add UMA stats

gpu/command_buffer/service/gles2_cmd_decoder.cc

 // Copyright (c) 2012 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 "gpu/command_buffer/service/gles2_cmd_decoder.h"
 
 #include <stdio.h>
 
 #include <algorithm>
 #include <list>
@@ skipping 742 matching lines @@
   base::Closure AsyncUploadTokenCompletionClosure(uint32 async_upload_token,
                                                   uint32 sync_data_shm_id,
                                                   uint32 sync_data_shm_offset);
 
 
 
   // Workarounds
   void OnFboChanged() const;
   void OnUseFramebuffer() const;
 
+  error::ContextLostReason GetContextLostReasonFromResetStatus(
+      GLenum reset_status) const;
+
   // TODO(gman): Cache these pointers?
   BufferManager* buffer_manager() {
     return group_->buffer_manager();
   }
 
   RenderbufferManager* renderbuffer_manager() {
     return group_->renderbuffer_manager();
   }
 
   FramebufferManager* framebuffer_manager() {
@@ skipping 877 matching lines @@
                        GLint fake_location,
                        uint32 shm_id,
                        uint32 shm_offset,
                        error::Error* error,
                        GLint* real_location,
                        GLuint* service_id,
                        void** result,
                        GLenum* result_type,
                        GLsizei* result_size);
 
-  void MaybeExitOnContextLost();
-  bool WasContextLost() override;
-  bool WasContextLostByRobustnessExtension() override;
-  void LoseContext(uint32 reset_status) override;
+  bool WasContextLost() const override;
+  bool WasContextLostByRobustnessExtension() const override;
+  void MarkContextLost(error::ContextLostReason reason) override;
+  bool CheckResetStatus();
 
 #if defined(OS_MACOSX)
   void ReleaseIOSurfaceForTexture(GLuint texture_id);
 #endif
 
   bool ValidateCompressedTexDimensions(
       const char* function_name,
       GLint level, GLsizei width, GLsizei height, GLenum format);
   bool ValidateCompressedTexFuncData(
       const char* function_name,
@@ skipping 182 matching lines @@
   // Cached from ContextGroup
   const Validators* validators_;
   scoped_refptr<FeatureInfo> feature_info_;
 
   int frame_number_;
 
   // Number of commands remaining to be processed in DoCommands().
   int commands_to_process_;
 
   bool has_robustness_extension_;
-  GLenum reset_status_;
+  error::ContextLostReason context_lost_reason_;
+  bool context_was_lost_;
   bool reset_by_robustness_extension_;
   bool supports_post_sub_buffer_;
 
   // These flags are used to override the state of the shared feature_info_
   // member.  Because the same FeatureInfo instance may be shared among many
   // contexts, the assumptions on the availablity of extensions in WebGL
   // contexts may be broken.  These flags override the shared state to preserve
   // WebGL semantics.
   bool force_webgl_glsl_validation_;
   bool derivatives_explicitly_enabled_;
@@ skipping 521 matching lines @@
       back_buffer_has_depth_(false),
       back_buffer_has_stencil_(false),
       surfaceless_(false),
       backbuffer_needs_clear_bits_(0),
       current_decoder_error_(error::kNoError),
       use_shader_translator_(true),
       validators_(group_->feature_info()->validators()),
       feature_info_(group_->feature_info()),
       frame_number_(0),
       has_robustness_extension_(false),
-      reset_status_(GL_NO_ERROR),
+      context_lost_reason_(error::kUnknown),
+      context_was_lost_(false),
       reset_by_robustness_extension_(false),
       supports_post_sub_buffer_(false),
       force_webgl_glsl_validation_(false),
       derivatives_explicitly_enabled_(false),
       frag_depth_explicitly_enabled_(false),
       draw_buffers_explicitly_enabled_(false),
       shader_texture_lod_explicitly_enabled_(false),
       compile_shader_always_succeeds_(false),
       lose_context_when_out_of_memory_(false),
       service_logging_(base::CommandLine::ForCurrentProcess()->HasSwitch(
@@ skipping 901 matching lines @@
     }
   }
 }
 
 // }  // anonymous namespace
 
 bool GLES2DecoderImpl::MakeCurrent() {
   if (!context_.get())
     return false;
 
-  if (!context_->MakeCurrent(surface_.get()) || WasContextLost()) {
-    LOG(ERROR) << "  GLES2DecoderImpl: Context lost during MakeCurrent.";
-
-    MaybeExitOnContextLost();
-
+  if (WasContextLost()) {
+    LOG(ERROR) << "  GLES2DecoderImpl: Trying to make lost context current.";
     return false;
   }
 
+  if (!context_->MakeCurrent(surface_.get())) {
+    LOG(ERROR) << "  GLES2DecoderImpl: Context lost during MakeCurrent.";
+    MarkContextLost(error::kMakeCurrentFailed);
+    group_->LoseContexts(error::kUnknown);
+    return false;
+  }
+
+  if (CheckResetStatus()) {
+    LOG(ERROR)
+        << "  GLES2DecoderImpl: Context reset detected after MakeCurrent.";
+    return false;

[Ken Russell (switch to Gerrit), 2015/04/27 22:35:33]

Per piman's earlier comment, were you going to broadcast this to other contexts
via group_->LoseContexts()?

[no sievers, 2015/04/27 23:24:41]

Good catch. Missed this one since it didn't call MarkContextLost() directly.

+  }
+
   ProcessFinishedAsyncTransfers();
 
   // Rebind the FBO if it was unbound by the context.
   if (workarounds().unbind_fbo_on_context_switch)
     RestoreFramebufferBindings();
 
   framebuffer_state_.clear_state_dirty = true;
 
   return true;
 }
@@ skipping 6950 matching lines @@
       offscreen_saved_color_texture_->AllocateStorage(
           offscreen_size_, offscreen_saved_color_format_, false);
 
       offscreen_saved_frame_buffer_->AttachRenderTexture(
           offscreen_saved_color_texture_.get());
       if (offscreen_size_.width() != 0 && offscreen_size_.height() != 0) {
         if (offscreen_saved_frame_buffer_->CheckStatus() !=
             GL_FRAMEBUFFER_COMPLETE) {
           LOG(ERROR) << "GLES2DecoderImpl::ResizeOffscreenFrameBuffer failed "
                      << "because offscreen saved FBO was incomplete.";
-          LoseContext(GL_UNKNOWN_CONTEXT_RESET_ARB);
+          MarkContextLost(error::kUnknown);
+          group_->LoseContexts(error::kUnknown);
           return;
         }
 
         // Clear the offscreen color texture.
         // TODO(piman): Is this still necessary?
         {
           ScopedFrameBufferBinder binder(this,
                                          offscreen_saved_frame_buffer_->id());
           glClearColor(0, 0, 0, 0);
           state_.SetDeviceColorMask(GL_TRUE, GL_TRUE, GL_TRUE, GL_TRUE);
@@ skipping 36 matching lines @@
 
       // Ensure the side effects of the copy are visible to the parent
       // context. There is no need to do this for ANGLE because it uses a
       // single D3D device for all contexts.
       if (!feature_info_->gl_version_info().is_angle)
         glFlush();
     }
   } else {
     if (!surface_->SwapBuffers()) {
       LOG(ERROR) << "Context lost because SwapBuffers failed.";
-      LoseContext(GL_UNKNOWN_CONTEXT_RESET_ARB);
+      if (!CheckResetStatus()) {
+        MarkContextLost(error::kUnknown);
+        group_->LoseContexts(error::kUnknown);

[Ken Russell (switch to Gerrit), 2015/04/27 22:35:33]

Per our offline conversation, one comment about broadcasting LoseContext
notifications.

I'm not sure how likely it is that guilty context notifications will be reported
promptly. Even if the driver properly assigns GPU resets to the right context,
if enough buffering of commands happens in the driver, the guilty notification
might not be discovered until the next time the guilty context is made current.
In the interim, "innocent" or "unknown" notifications may be delivered to other
contexts. Since we lose all other contexts in the share group immediately with
unknown status as soon as we get any reset notification, we may squelch guilty
notifications.

I don't think this is a big deal, at least right now. WebGL's DoS defenses
aren't predicated on receiving guilty notifications. It's something to keep in
mind though.

+      }
     }
   }
 
   // This may be a slow command.  Exit command processing to allow for
   // context preemption and GPU watchdog checks.
   ExitCommandProcessingEarly();
 }
 
 void GLES2DecoderImpl::DoSwapInterval(int interval) {
   context_->SetSwapInterval(interval);
@@ skipping 239 matching lines @@
   Program* program = NULL;
   program = GetProgram(program_id);
   if (!program || !program->IsValid()) {
     return error::kNoError;
   }
   program->GetTransformFeedbackVaryings(bucket);
   return error::kNoError;
 }
 
 error::ContextLostReason GLES2DecoderImpl::GetContextLostReason() {
-  switch (reset_status_) {
+  return context_lost_reason_;
+}
+
+error::ContextLostReason GLES2DecoderImpl::GetContextLostReasonFromResetStatus(
+    GLenum reset_status) const {
+  switch (reset_status) {
     case GL_NO_ERROR:
       // TODO(kbr): improve the precision of the error code in this case.
       // Consider delegating to context for error code if MakeCurrent fails.
       return error::kUnknown;
     case GL_GUILTY_CONTEXT_RESET_ARB:
       return error::kGuilty;
     case GL_INNOCENT_CONTEXT_RESET_ARB:
       return error::kInnocent;
     case GL_UNKNOWN_CONTEXT_RESET_ARB:
       return error::kUnknown;
   }
 
   NOTREACHED();
   return error::kUnknown;
 }
 
-void GLES2DecoderImpl::MaybeExitOnContextLost() {
+bool GLES2DecoderImpl::WasContextLost() const {
+  return context_was_lost_;
+}
+
+bool GLES2DecoderImpl::WasContextLostByRobustnessExtension() const {
+  return WasContextLost() && reset_by_robustness_extension_;
+}
+
+void GLES2DecoderImpl::MarkContextLost(error::ContextLostReason reason) {
+  // Only lose the context once.
+  if (WasContextLost())
+    return;
+
+  // Don't make GL calls in here, the context might not be current.
+  context_lost_reason_ = reason;
+  current_decoder_error_ = error::kLostContext;
+  context_was_lost_ = true;
+
   // Some D3D drivers cannot recover from device lost in the GPU process
   // sandbox. Allow a new GPU process to launch.
   if (workarounds().exit_on_context_lost) {
     LOG(ERROR) << "Exiting GPU process because some drivers cannot reset"
                << " a D3D device in the Chrome GPU process sandbox.";
 #if defined(OS_WIN)
     base::win::SetShouldCrashOnProcessDetach(false);
 #endif
     exit(0);
   }
 }
 
-bool GLES2DecoderImpl::WasContextLost() {
-  if (reset_status_ != GL_NO_ERROR) {
-    MaybeExitOnContextLost();
+bool GLES2DecoderImpl::CheckResetStatus() {
+  DCHECK(!WasContextLost());
+  DCHECK(context_->IsCurrent(NULL));
+
+  if (IsRobustnessSupported()) {
+    // If the reason for the call was a GL error, we can try to determine the
+    // reset status more accurately.
+    GLenum driver_status = glGetGraphicsResetStatusARB();
+    if (driver_status == GL_NO_ERROR)
+      return false;
+
+    LOG(ERROR) << (surface_->IsOffscreen() ? "Offscreen" : "Onscreen")
+               << " context lost via ARB/EXT_robustness. Reset status = "
+               << GLES2Util::GetStringEnum(driver_status);
+
+    // Don't pretend we know which client was responsible.
+    if (workarounds().use_virtualized_gl_contexts)
+      driver_status = GL_UNKNOWN_CONTEXT_RESET_ARB;
+
+    switch (driver_status) {
+      case GL_GUILTY_CONTEXT_RESET_ARB:
+        MarkContextLost(error::kGuilty);
+        break;
+      case GL_INNOCENT_CONTEXT_RESET_ARB:
+        MarkContextLost(error::kInnocent);
+        break;
+      case GL_UNKNOWN_CONTEXT_RESET_ARB:
+        MarkContextLost(error::kUnknown);
+        break;
+      default:
+        NOTREACHED();
+        return false;
+    }
+    reset_by_robustness_extension_ = true;
     return true;
   }
-  if (IsRobustnessSupported()) {
-    GLenum status = glGetGraphicsResetStatusARB();
-    if (status != GL_NO_ERROR) {
-      // The graphics card was reset. Signal a lost context to the application.
-      reset_status_ = status;
-      reset_by_robustness_extension_ = true;
-      LOG(ERROR) << (surface_->IsOffscreen() ? "Offscreen" : "Onscreen")
-                 << " context lost via ARB/EXT_robustness. Reset status = "
-                 << GLES2Util::GetStringEnum(status);
-      MaybeExitOnContextLost();
-      return true;
-    }
-  }
   return false;
 }
 
-bool GLES2DecoderImpl::WasContextLostByRobustnessExtension() {
-  return WasContextLost() && reset_by_robustness_extension_;
-}
-
-void GLES2DecoderImpl::LoseContext(uint32 reset_status) {
-  // Only loses the context once.
-  if (reset_status_ != GL_NO_ERROR) {
-    return;
-  }
-
-  if (workarounds().use_virtualized_gl_contexts) {
-    // If the context is virtual, the real context being guilty does not ensure
-    // that the virtual context is guilty.
-    if (reset_status == GL_GUILTY_CONTEXT_RESET_ARB) {
-      reset_status = GL_UNKNOWN_CONTEXT_RESET_ARB;
-    }
-  } else if (reset_status == GL_UNKNOWN_CONTEXT_RESET_ARB &&
-             IsRobustnessSupported()) {
-    // If the reason for the call was a GL error, we can try to determine the
-    // reset status more accurately.
-    GLenum driver_status = glGetGraphicsResetStatusARB();
-    if (driver_status == GL_GUILTY_CONTEXT_RESET_ARB ||
-        driver_status == GL_INNOCENT_CONTEXT_RESET_ARB) {
-      reset_status = driver_status;
-    }
-  }
-
-  // Marks this context as lost.
-  reset_status_ = reset_status;
-  current_decoder_error_ = error::kLostContext;
-}
-
 error::Error GLES2DecoderImpl::HandleInsertSyncPointCHROMIUM(
     uint32 immediate_data_size,
     const void* cmd_data) {
   return error::kUnknownCommand;
 }
 
 error::Error GLES2DecoderImpl::HandleWaitSyncPointCHROMIUM(
     uint32 immediate_data_size,
     const void* cmd_data) {
   const gles2::cmds::WaitSyncPointCHROMIUM& c =
@@ skipping 1353 matching lines @@
     if (GetBackbufferServiceId() != 0 &&  // emulated backbuffer
         bufs[0] == GL_BACK) {
       mapped_buf = GL_COLOR_ATTACHMENT0;
     }
     glDrawBuffersARB(count, &mapped_buf);
     group_->set_draw_buffer(bufs[0]);
   }
 }
 
 void GLES2DecoderImpl::DoLoseContextCHROMIUM(GLenum current, GLenum other) {
-  group_->LoseContexts(other);
-  reset_status_ = current;
-  current_decoder_error_ = error::kLostContext;
+  MarkContextLost(GetContextLostReasonFromResetStatus(current));
+  group_->LoseContexts(GetContextLostReasonFromResetStatus(other));
+  reset_by_robustness_extension_ = true;
 }
 
 void GLES2DecoderImpl::DoMatrixLoadfCHROMIUM(GLenum matrix_mode,
                                              const GLfloat* matrix) {
   DCHECK(matrix_mode == GL_PATH_PROJECTION_CHROMIUM ||
          matrix_mode == GL_PATH_MODELVIEW_CHROMIUM);
   if (!features().chromium_path_rendering) {
     LOCAL_SET_GL_ERROR(GL_INVALID_OPERATION,
                        "glMatrixLoadfCHROMIUM",
                        "function not available");
@@ skipping 497 matching lines @@
   }
   buffer->RemoveMappedRange();
   GLboolean rt = glUnmapBuffer(target);
   if (rt == GL_FALSE) {
     // At this point, we have already done the necessary validation, so
     // GL_FALSE indicates data corruption.
     // TODO(zmo): We could redo the map / copy data / unmap to recover, but
     // the second unmap could still return GL_FALSE. For now, we simply lose
     // the contexts in the share group.
     LOG(ERROR) << "glUnmapBuffer unexpectedly returned GL_FALSE";
-    group_->LoseContexts(GL_INNOCENT_CONTEXT_RESET_ARB);
-    reset_status_ = GL_GUILTY_CONTEXT_RESET_ARB;
+    // Need to lose current context before broadcasting!
+    MarkContextLost(error::kGuilty);
+    group_->LoseContexts(error::kInnocent);
     return error::kLostContext;
   }
   return error::kNoError;
 }
 
 void GLES2DecoderImpl::OnTextureRefDetachedFromFramebuffer(
     TextureRef* texture_ref) {
   Texture* texture = texture_ref->texture();
   DoDidUseTexImageIfNeeded(texture, texture->target());
 }
 
+// Note that GL_LOST_CONTEXT is specific to GLES.
+// For desktop GL we have to query the reset status proactively.
 void GLES2DecoderImpl::OnContextLostError() {
-  group_->LoseContexts(GL_UNKNOWN_CONTEXT_RESET_ARB);
+  if (!WasContextLost()) {
+    // Need to lose current context before broadcasting!
+    CheckResetStatus();
+    group_->LoseContexts(error::kUnknown);
+    reset_by_robustness_extension_ = true;
+  }
 }
 
 void GLES2DecoderImpl::OnOutOfMemoryError() {
-  if (lose_context_when_out_of_memory_) {
-    group_->LoseContexts(GL_UNKNOWN_CONTEXT_RESET_ARB);
+  if (lose_context_when_out_of_memory_ && !WasContextLost()) {
+    error::ContextLostReason other = error::kOutOfMemory;
+    if (CheckResetStatus()) {
+      other = error::kUnknown;
+    } else {
+      // Need to lose current context before broadcasting!
+      MarkContextLost(error::kOutOfMemory);
+    }
+    group_->LoseContexts(other);
   }
 }
 
 // Include the auto-generated part of this file. We split this because it means
 // we can easily edit the non-auto generated parts right here in this file
 // instead of having to edit some template or the code generator.
 #include "gpu/command_buffer/service/gles2_cmd_decoder_autogen.h"
 
 }  // namespace gles2
 }  // namespace gpu