There is a flaky failure of GpuProcessHostBrowserTest.Shutdown when running with --enable-viz.

This was first seen on the CQ, and I haven't reproduced it locally yet. Though stevenjb@ did. There's been one failure on the FYI bot.

Stack Trace:

#0 0x00000205329c base::debug::StackTrace::StackTrace()
#1 0x000001d41d52 content::(anonymous namespace)::DumpStackTraceSignalHandler()
#2 0x7f3b7ef36cb0 <unknown>
#3 0x7f3b820d6402 __pthread_cond_wait
#4 0x00000208e778 base::ConditionVariable::Wait()
#5 0x00000208ed4e base::WaitableEvent::TimedWaitUntil()
#6 0x00000208ebff base::WaitableEvent::Wait()
#7 0x000000cdded8 gpu::GpuChannelHost::Send()
#8 0x000000cdbd2e gpu::CommandBufferProxyImpl::WaitForGetOffsetInRange()
#9 0x000000c7cd5b gpu::CommandBufferHelper::Finish()
#10 0x000002e55406 gpu::gles2::GLES2Implementation::FinishHelper()
#11 0x000002c78be0 cc::ResourceProvider::~ResourceProvider()
#12 0x000002c7764e cc::LayerTreeResourceProvider::~LayerTreeResourceProvider()
#13 0x000002ca1c72 cc::LayerTreeHostImpl::ReleaseLayerTreeFrameSink()
#14 0x000002cc8979 cc::SingleThreadProxy::Stop()
#15 0x000002c91a33 cc::LayerTreeHost::~LayerTreeHost()
#16 0x000002c91fce cc::LayerTreeHost::~LayerTreeHost()
#17 0x000001ec39eb ui::Compositor::~Compositor()
#18 0x000001ec3e1e ui::Compositor::~Compositor()
#19 0x000002037b7d aura::WindowTreeHostPlatform::~WindowTreeHostPlatform()
#20 0x000004302706 wm::WMTestHelper::~WMTestHelper()
#21 0x00000430279e wm::WMTestHelper::~WMTestHelper()
#22 0x000001d8ced9 content::Shell::PlatformExit()
#23 0x000001d8387b content::Shell::CloseAllWindows()
#24 0x000001d387b0 content::ContentBrowserTest::PostRunTestOnMainThread()
#25 0x000001d41ac0 content::BrowserTestBase::ProxyRunTestOnMainThreadLoop()
#26 0x000001d9bda9 content::ShellBrowserMainParts::PreMainMessageLoopRun()
#27 0x0000017c6dc1 content::BrowserMainLoop::PreMainMessageLoopRun()
#28 0x000001aea567 content::StartupTaskRunner::RunAllTasksNow()
#29 0x0000017c532c content::BrowserMainLoop::CreateStartupTasks()
#30 0x0000017c9cf8 content::BrowserMainRunnerImpl::Initialize()
#31 0x000001d846e9 ShellBrowserMain()
#32 0x000001d7d2c0 content::ShellMainDelegate::RunProcess()
#33 0x000001709f11 content::RunNamedProcessTypeMain()
#34 0x00000170a885 content::ContentMainRunnerImpl::Run()
#35 0x0000030d2c66 service_manager::Main()
#36 0x000000f06a54 content::ContentMain()
#37 0x000001d41774 content::BrowserTestBase::SetUp()
#38 0x000001d385b9 content::ContentBrowserTest::SetUp()
#39 0x000000a20c01 testing::Test::Run()
#40 0x000000a21630 testing::TestInfo::Run()
#41 0x000000a21b17 testing::TestCase::Run()
#42 0x000000a280c7 testing::internal::UnitTestImpl::RunAllTests()
#43 0x000000a27d47 testing::UnitTest::Run()
#44 0x000001d684d2 base::TestSuite::Run()
#45 0x000001d3bd8a content::ContentTestLauncherDelegate::RunTestSuite()
#46 0x000001d54667 content::LaunchTests()
#47 0x000001d3bd50 main
#48 0x7f3b7ef21f45 __libc_start_main
#49 0x000000481013 <unknown>

Example failing run: https://build.chromium.org/p/chromium.chromiumos/builders/Linux%20ChromiumOS%20Tests%20%281%29/builds/46591

It's possible that the receiver of GpuCommandBufferMsg_WaitForGetOffsetInRange has already shutdown at this point in the test. But investigation is needed.