New issue
Advanced search Search tips
Note: Color blocks (like or ) mean that a user may not be available. Tooltip shows the reason.

Issue 800453 link

Starred by 3 users

Issue metadata

Status: Fixed
Owner:
OOO until 2019-01-24
Closed: Apr 2018
Cc:
Components:
EstimatedDays: ----
NextAction: ----
OS: Windows
Pri: 2
Type: Bug

Blocked on:
issue 754435
issue 829207

Blocking:
issue 879098



Sign in to add a comment

VMware Video Drivers blacklisted in Chrome

Reported by steve.bu...@whrsd.org, Jan 9 2018

Issue description

UserAgent: Mozilla/5.0 (Windows NT 6.1; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/63.0.3239.132 Safari/537.36

Steps to reproduce the problem:
Open chrome and check chrome://gpu and you will notice software only rendering is available.

Researching previous cases and we came across  issue 754435  and it appears a decision was made to block all VMware drivers so customers are now unable to offload rendering from CPU to GPU. 

That issue was marked as fixed so I'm opening this case for the blocked driver issue. We invested 20K+ on GPU cards that can no longer being used by Chrome - the primary application in our environment. 

What is the expected behavior?
We would like the video card to be utilized for rendering. Chrome is forced into software only mode because all vmware graphics drivers have been blacklisted.

What went wrong?
VMware video drivers are black listed

Did this work before? Yes 

Chrome version: 63.0.3239.132  Channel: stable
OS Version: 6.1 (Windows 7, Windows Server 2008 R2)
Flash Version:
 

Comment 1 by kbr@chromium.org, Jan 9 2018

Blockedon: 754435
Components: Internals>GPU>Internals

Comment 2 by kbr@chromium.org, Jan 9 2018

The Chromium team doesn't test this configuration (VMWare VMs with graphics acceleration) and customers reported that there was rendering corruption when running with graphics acceleration inside a VMWare VM. We never got further feedback from the customer that originally reported the problem so were unable to take further action on it.

If you can work with us to confirm whether rendering is still corrupted on your setup we would like to give feedback to VMWare.

Please run Chrome with the command line flag --ignore-gpu-blacklist and see whether it renders correctly, in particular the tab strip per  Issue 754435 . Please copy/paste the contents of the about:gpu page here.

Comment 3 by piman@chromium.org, Jan 9 2018

Status: Available (was: Unconfirmed)
We would be more than happy to assist. I was unable to repro the issue during initial testing on our Windows 7 32bit image. I noticed in the previous case notes that the issue may only occur on Windows 10 64bit clients so I will spin up a VM with that OS to test further. 
chrome_homepage.JPG
136 KB View Download
chrome_gpu.JPG
170 KB View Download
aboutgpu_txt.txt
9.6 KB View Download
I installed Windows 10 Enterprise version 1709 64bit (SW_DVD5_Win_Pro_Ent_Edu_N_10_1709_64BIT_English_MLF_X21-50143.ISO), Chrome 63.0.3239.132 64-bit, VMware SVGA 3D 8.15.1.50. When I launch chrome ignoring gpu blacklist I do not experience the corrupted tab strip. Unsure if there are more advanced tests that can be run in Chrome to detect rendering corruption?
chrome_gpu_WIN10x64.JPG
149 KB View Download
chrome_homepage_WIN10x64.JPG
127 KB View Download
aboutgpu_WIN10_x64_txt.txt
9.2 KB View Download
chrome_homepage_multipletabs_WIN10x64.JPG
116 KB View Download
It's important to note in the previous case it was reported 'We currently blacklist all GPU features on VMWare on Windows GPU driver <= 7.14.1.1134.  This bug is on 8.15.1.48'. To be thorough I rolled back my drivers to the impacted version by uninstalling VMware tools, restarting, installing VMware Tools 10.1.0.4449150 which provides driver version 8.15.1.48. I performed the same tests and I do not experience the corrupted tab strip. 
chrome_gpu_WIN10x64_8.15.1.48.JPG
141 KB View Download
chrome_multipletabs_WIN10x64_8.15.1.48.JPG
146 KB View Download
chrome_homepage_WIN10x64_8.15.1.48.JPG
135 KB View Download
aboutgpu_WIN10_x64_8.15.1.48_txt.txt
9.2 KB View Download
I ran a fourth test since the previous case mentioned Chrome version: 60.0.3112.90 64bit so I reverted back to that version and I'm still unable to reproduce the issues they were experiencing on driver 8.15.1.48. 
aboutgpu_WIN10_x64_8.15.1.48_60.0.3112.90_txt.txt
8.6 KB View Download
chrome_gpu_WIN10x64_8.15.1.48_60.0.3112.90.JPG
173 KB View Download
chrome_multipletabs_WIN10x64_8.15.1.48_60.0.3112.90.JPG
83.8 KB View Download
I do see many people were reporting the same symptoms on various forums on the internet and they appear to all be isolated to Workstation Pro - I have been testing on VMware's Enterprise line of products (vSphere, vCenter, Horizon Enterprise) - not Workstation Pro. I am in the process of spinning up a Windows 10 64bit image on VMWare 12.5.6 build-5528349 (version specified on previous case) to try and reproduce the issue.  

https://www.reddit.com/r/vmware/comments/73jgfd/chrome_tabs_bar_completely_white_issue/

https://www.reddit.com/r/vmware/comments/6u8cfi/chrome_tab_bar_whited_out_in_vmware_xpost_from/

https://www.reddit.com/r/chrome/comments/6u88mc/tab_bar_whited_out_in_vmware/

Installed Vmware Workstation Pro (VMware-workstation-full-12.5.6-5528349.exe) with VMware tools that contains driver version 8.15.1.48 (VMware-tools-windows-10.1.0-4449150.iso).
Disabled chrome auto update and installed Chrome version 60.0.3112.90. Launched chrome disabling gpu blacklist and I'm still unable to reproduce the issue others have reported. 

There must be other factors involved to trigger this issue than the version of chrome, driver version, and OS version/architecture. 

We have a need to use the vmware drivers that are currently blacklisted so we can take advantage of Virtual Shared Graphics Acceleration (vSGA) and offload rendering from CPU to GPU. In this pod we have 600 concurrent VMs where Chrome is unable to offload rendering due to these restrictions. How should we proceed? Is there a way to system-wide enforce the ignore blacklist flag? Even if we create desktop shortcuts with that tag included this won't update shortcuts, links that launch Chrome (from outlook for example), pins to taskbar + start menu, etc.





aboutgpu_WIN10_x64_WP_8.15.1.48_60.0.3112.90_txt.txt
8.6 KB View Download
chrome_version_WIN10x64_WP_8.15.1.48_60.0.3112.90.JPG
108 KB View Download
chrome_multitabs_WIN10x64_WP_8.15.1.48_60.0.3112.90.JPG
184 KB View Download
I implemented a logon script in our test environment to flag the ignore-gpu-blacklist option for enabled_labs_experiments. End users in the testing group are reporting that sites do not render correctly when the black list is ignored. While we do not experience the missing tab strip we do have an issue with the latest version of chrome and VMware video driver 8.15.1.50. They confirmed when they navigate to 'chrome://gpu' and disable the override the sites load as expected. I do see a newer VMware video driver is available and we will upgrade to test that shortly. I will reach out VMware support and open a case for rendering issues. Examples attached.

Comment 11 by kbr@chromium.org, Mar 7 2018

Cc: kbr@chromium.org vmi...@chromium.org

Comment 12 by kbr@chromium.org, Mar 7 2018

Cc: zmo@chromium.org
Note to zmo@: this is a situation where it would be really helpful to be able to turn off certain software_rendering_list.json entries, in the same way that we can turn off individual GPU driver bug workarounds.

Comment 13 by kbr@chromium.org, Mar 7 2018

https://chromium-review.googlesource.com/952778 up for consideration. Waiting for confirmation from someone that it has the desired effect.

Comment 14 by kbr@chromium.org, Apr 5 2018

Cc: perezju@chromium.org nednguyen@chromium.org
Components: Tests>Telemetry
Owner: kbr@chromium.org
Status: Assigned (was: Available)
Have received confirmation from VMWare and, indirectly, the submitter, that at least on current drivers, re-enabling GPU acceleration works.

Unfortunately I'm having great difficulty landing https://chromium-review.googlesource.com/952778 . It breaks telemetry_unittests and telemetry_perf_unittests on our Windows bots because, apparently, they're running old VMWare VMs and Chrome's GPU process crashes during startup. There's no stack trace from the GPU process in the crash logs, and the unexpected crash causes an assertion failure inside Telemetry's SystemInfo collection. Here's a copy/paste from one of my updates from that CL:

-----
The VMs we use for testing must be reporting a VMWare GPU (even though Swarming reports it as "gpu:none"), and Chrome's GPU process is crashing while attempting to start in this configuration. Telemetry is making this call to Browser.GetSystemInfo():
https://cs.chromium.org/chromium/src/third_party/catapult/telemetry/telemetry/internal/browser/browser.py?q=%22Detailed+OS+version%22&sq=package:chromium&l=114

and that's resulting in a GPU process crash (we don't see the symbolized stack trace for this), causing a DCHECK somewhere underneath chrome!content::GetFeatureStatus:
https://cs.chromium.org/chromium/src/content/browser/gpu/compositor_util.cc?type=cs&q=GetFeatureStatus&l=284

From code inspection, I think the DCHECK getting hit is here, inside GetFeatureData:
https://cs.chromium.org/chromium/src/content/browser/gpu/compositor_util.cc?sq=package:chromium&l=74
-----

I attempted to reproduce this on my Mac development machine by introducing a DCHECK inside the GPU process's info collection but that caused a different assertion failure in the renderer process:

[51188:50947:0403/171531.408391:FATAL:thumbnail_database.cc(272)] database is locked
  	0   libbase.dylib                       0x0000000104b7178c base::debug::StackTrace::StackTrace(unsigned long) + 28
  	1   libbase.dylib                       0x0000000104b9cb5f logging::LogMessage::~LogMessage() + 223
  	2   libchrome_dll.dylib                 0x0000000106634e9d history::(anonymous namespace)::DatabaseErrorCallback(sql::Connection*, base::FilePath const&, history::HistoryBackendClient*, int, sql::Statement*) + 237
  	3   libsql.dylib                        0x0000000111ffa4d7 sql::Connection::OnSqliteError(int, sql::Statement*, char const*) const + 983
  	4   libsql.dylib                        0x0000000111ff61de sql::Connection::OpenInternal(std::__1::basic_string<char, std::__1::char_traits<char>, std::__1::allocator<char> > const&, sql::Connection::Retry) + 1822
  	5   libsql.dylib                        0x0000000111ff5aa0 sql::Connection::Open(base::FilePath const&) + 320
  	6   libchrome_dll.dylib                 0x0000000106634d9a history::ThumbnailDatabase::OpenDatabase(sql::Connection*, base::FilePath const&) + 330
  	7   libchrome_dll.dylib                 0x0000000106631ac5 history::ThumbnailDatabase::InitImpl(base::FilePath const&) + 37
  	8   libchrome_dll.dylib                 0x0000000106631a46 history::ThumbnailDatabase::Init(base::FilePath const&) + 22
  	9   libchrome_dll.dylib                 0x0000000106606677 history::HistoryBackend::InitImpl(history::HistoryDatabaseParams const&) + 935
  	10  libchrome_dll.dylib                 0x0000000106606182 history::HistoryBackend::Init(bool, history::HistoryDatabaseParams const&) + 98
  	11  libchrome_dll.dylib                 0x000000010662cd41 base::internal::Invoker<base::internal::BindState<void (history::HistoryBackend::*)(bool, history::HistoryDatabaseParams const&), scoped_refptr<history::HistoryBackend>, bool, history::HistoryDatabaseParams>, void ()>::Run(base::internal::BindStateBase*) + 145
  	12  libbase.dylib                       0x0000000104b72104 base::debug::TaskAnnotator::RunTask(char const*, base::PendingTask*) + 260
  	13  libbase.dylib                       0x0000000104bb57c9 base::internal::IncomingTaskQueue::RunTask(base::PendingTask*) + 121
  	14  libbase.dylib                       0x0000000104bba1e7 base::MessageLoop::RunTask(base::PendingTask*) + 599
  	15  libbase.dylib                       0x0000000104bba5aa base::MessageLoop::DeferOrRunPendingTask(base::PendingTask) + 186
  	16  libbase.dylib                       0x0000000104bba81c base::MessageLoop::DoWork() + 572
  	17  libbase.dylib                       0x0000000104bbc823 base::MessagePumpDefault::Run(base::MessagePump::Delegate*) + 51
  	18  libbase.dylib                       0x0000000104bb9b09 base::MessageLoop::Run(bool) + 169
  	19  libbase.dylib                       0x0000000104bfb4e9 base::RunLoop::Run() + 249
  	20  libbase.dylib                       0x0000000104c44f2e base::Thread::Run(base::RunLoop*) + 206
  	21  libbase.dylib                       0x0000000104c455dc base::Thread::ThreadMain() + 908
  	22  libbase.dylib                       0x0000000104c401ff base::(anonymous namespace)::ThreadFunc(void*) + 95
  	23  libsystem_pthread.dylib             0x00007fff5b4f76c1 _pthread_body + 340
  	24  libsystem_pthread.dylib             0x00007fff5b4f756d _pthread_body + 0
  	25  libsystem_pthread.dylib             0x00007fff5b4f6c5d thread_start + 13

I don't think I can find the time to debug this on one of the affected Windows VMs. It will take hours to set up the debugging environment.

The easiest workaround would be if I could pass --disable-gpu to the browser on Windows inside the telemetry_perf_unittests and telemetry_unittests, which are test suites that always run on VMs.

Ken: you can pass  --disable-gpu to the browser by adding "--extra-browser-args=--disable-gpu" to telemetry_perf_unittests & telemetry_unittests on that config.

Comment 17 by kbr@chromium.org, Apr 5 2018

Blockedon: 829207

Comment 18 by kbr@chromium.org, Apr 5 2018

Cc: capn@chromium.org sugoi@chromium.org
Here is Chrome's about:gpu from the target VM. It unfortunately isn't useful because SwiftShader's info collection overrides the device-level information. It does however show the line numbers where Chrome fails to gather GPU information and bails out in a way which causes the SystemInfo collection to fail:

[1972:3112:0404/212115.312:ERROR:gl_surface_egl.cc(845)] : eglInitialize D3D9 failed with error EGL_NOT_INITIALIZED
[1972:3112:0404/212115.312:ERROR:gl_initializer_win.cc(232)] : GLSurfaceEGL::InitializeOneOff failed.
GpuProcessHostUIShim: The GPU process crashed!
GpuProcessHostUIShim: The GPU process exited normally. Everything is okay.

I'll follow with another comment containing Firefox's about:support.

------

Graphics Feature Status
Canvas: Software only, hardware acceleration unavailable
CheckerImaging: Enabled
Flash: Software only. Hardware acceleration disabled
Flash Stage3D: Software only. Hardware acceleration disabled
Flash Stage3D Baseline profile: Software only. Hardware acceleration disabled
Compositing: Software only. Hardware acceleration disabled
Multiple Raster Threads: Enabled
Native GpuMemoryBuffers: Software only. Hardware acceleration disabled
Rasterization: Software only. Hardware acceleration disabled
Surface Synchronization: Enabled
Video Decode: Software only. Hardware acceleration disabled
Viz Service Display Compositor: Disabled
WebGL: Software only, hardware acceleration unavailable
WebGL2: Software only, hardware acceleration unavailable
Problems Detected
Native GpuMemoryBuffers have been disabled, either via about:flags or command line.
Disabled Features: native_gpu_memory_buffers
Viz service display compositor is not enabled by default.
Disabled Features: viz_display_compositor
Version Information
Data exported
2018-04-05T04:21:47.129Z
Chrome version
Chrome/67.0.3388.0
Operating system
Windows NT 6.1.7601 SP1
Software rendering list URL
https://chromium.googlesource.com/chromium/src/+/ff69a10a306f029b25759e78e0c3cc98a3fa840c/gpu/config/software_rendering_list.json
Driver bug list URL
https://chromium.googlesource.com/chromium/src/+/ff69a10a306f029b25759e78e0c3cc98a3fa840c/gpu/config/gpu_driver_bug_list.json
ANGLE commit id
f81d7c5bb715
2D graphics backend
Skia/67 c0528e2418e09765807f1db510733c573232f7e2-
Command Line
"C:\Users\chrome-bot\AppData\Local\Chromium\Application\chrome.exe" --flag-switches-begin --flag-switches-end
Driver Information
Initialization time
84
In-process GPU
false
Passthrough Command Decoder
false
Direct Composition
false
Supports overlays
false
Sandboxed
true
GPU0
VENDOR = 0xffff, DEVICE= 0xfffe *ACTIVE*
Optimus
false
Optimus
false
AMD switchable
false
Desktop compositing
none
Driver vendor


Driver version
4.0.0.3
Driver date


Pixel shader version
1.00
Vertex shader version
1.00
Max. MSAA samples
4
Machine model name


Machine model version


GL_VENDOR
Google Inc.
GL_RENDERER
Google SwiftShader
GL_VERSION
OpenGL ES 2.0 SwiftShader 4.0.0.3
GL_EXTENSIONS
GL_OES_compressed_ETC1_RGB8_texture GL_OES_depth24 GL_OES_depth32 GL_OES_depth_texture GL_OES_depth_texture_cube_map GL_OES_EGL_image GL_OES_EGL_image_external GL_OES_EGL_sync GL_OES_element_index_uint GL_OES_framebuffer_object GL_OES_packed_depth_stencil GL_OES_rgb8_rgba8 GL_OES_standard_derivatives GL_OES_surfaceless_context GL_OES_texture_float GL_OES_texture_float_linear GL_OES_texture_half_float GL_OES_texture_half_float_linear GL_OES_texture_npot GL_OES_texture_3D GL_OES_vertex_half_float GL_EXT_blend_minmax GL_EXT_color_buffer_half_float GL_EXT_draw_buffers GL_EXT_instanced_arrays GL_EXT_occlusion_query_boolean GL_EXT_read_format_bgra GL_EXT_texture_compression_dxt1 GL_EXT_texture_filter_anisotropic GL_EXT_texture_format_BGRA8888 GL_EXT_texture_rg GL_ARB_texture_rectangle GL_ANGLE_framebuffer_blit GL_ANGLE_framebuffer_multisample GL_ANGLE_instanced_arrays GL_ANGLE_texture_compression_dxt3 GL_ANGLE_texture_compression_dxt5 GL_APPLE_texture_format_BGRA8888 GL_CHROMIUM_color_buffer_float_rgba GL_CHROMIUM_texture_filtering_hint GL_NV_fence GL_NV_framebuffer_blit GL_NV_read_depth GL_EXT_color_buffer_float
Disabled Extensions


Disabled WebGL Extensions


Window system binding vendor
Google Inc.
Window system binding version
1.4 SwiftShader 4.0.0.3
Window system binding extensions
EGL_KHR_create_context EGL_KHR_get_all_proc_addresses EGL_KHR_gl_texture_2D_image EGL_KHR_gl_texture_cubemap_image EGL_KHR_gl_renderbuffer_image EGL_KHR_fence_sync EGL_KHR_image_base EGL_KHR_surfaceless_context EGL_ANDROID_framebuffer_target EGL_ANDROID_recordable
Direct rendering
Yes
Reset notification strategy
0x0000
GPU process crash count
0
Compositor Information
Tile Update Mode
One-copy
Partial Raster
Enabled
GpuMemoryBuffers Status
ATC
Software only
ATCIA
Software only
DXT1
Software only
DXT5
Software only
ETC1
Software only
R_8
Software only
R_16
Software only
RG_88
Software only
BGR_565
Software only
RGBA_4444
Software only
RGBX_8888
GPU_READ, SCANOUT
RGBA_8888
GPU_READ, SCANOUT
BGRX_8888
Software only
BGRX_1010102
Software only
RGBX_1010102
Software only
BGRA_8888
Software only
RGBA_F16
Software only
YVU_420
Software only
YUV_420_BIPLANAR
Software only
UYVY_422
Software only
Display(s) Information
Info
Display[2528732444] bounds=[0,0 1500x1000], workarea=[0,0 1500x960], scale=1, external.
Color space information
{primaries:BT709, transfer:IEC61966_2_1, matrix:RGB, range:FULL}
Bits per color component
8
Bits per pixel
24
Video Acceleration Information

Diagnostics
0
b3DAccelerationEnabled
true
b3DAccelerationExists
true
bAGPEnabled
true
bAGPExistenceValid
true
bAGPExists
false
bCanRenderWindow
true
bDDAccelerationEnabled
true
bDriverBeta
false
bDriverDebug
false
bDriverSigned
false
bDriverSignedValid
false
bNoHardware
true
dwBpp
16
dwDDIVersion
9
dwHeight
1000
dwRefreshRate
60
dwWHQLLevel
0
dwWidth
1500
iAdapter
0
lDriverSize
0
lMiniVddSize
0
szAGPStatusEnglish
Not Available
szAGPStatusLocalized
Not Available
szChipType


szD3DStatusEnglish
Enabled
szD3DStatusLocalized
Enabled
szDACType


szDDIVersionEnglish
9Ex
szDDIVersionLocalized
9Ex
szDDStatusEnglish
Not Available
szDDStatusLocalized
Not Available
szDXVAHDEnglish
Not Supported
szDXVAModes


szDescription
RDPDD Chained DD
szDeviceId
0x0000
szDeviceIdentifier
{D7B70EE0-4340-11CF-B063-0020AFC2CD35}
szDeviceName
\\.\DISPLAY1
szDisplayMemoryEnglish
160 MB
szDisplayMemoryLocalized
160 MB
szDisplayModeEnglish
1500 x 1000 (16 bit) (60Hz)
szDisplayModeLocalized
1500 x 1000 (16 bit) (60Hz)
szDriverAssemblyVersion


szDriverAttributes


szDriverDateEnglish


szDriverDateLocalized


szDriverLanguageEnglish


szDriverLanguageLocalized


szDriverModelEnglish
unknown
szDriverModelLocalized
unknown
szDriverName


szDriverNodeStrongName
Unknown
szDriverSignDate


szDriverVersion


szKeyDeviceID
Enum\
szKeyDeviceKey
\REGISTRY\Machine\System\CurrentControlSet\Services\RDPDD\Device0
szManufacturer


szMiniVdd


szMiniVddDateEnglish


szMiniVddDateLocalized


szMonitorMaxRes


szMonitorName


szNotesEnglish
No problems found.
szNotesLocalized
No problems found.
szOverlayEnglish
Not Supported
szRankOfInstalledDriver
Unknown
szRegHelpText


szRevision


szRevisionId
0x0000
szSubSysId
0x00000000
szTestResultD3D7English
Not run
szTestResultD3D7Localized
Not run
szTestResultD3D8English
Not run
szTestResultD3D8Localized
Not run
szTestResultD3D9English
Not run
szTestResultD3D9Localized
Not run
szTestResultDDEnglish
Not run
szTestResultDDLocalized
Not run
szVdd


szVendorId
0x0000


Log Messages
[1972:3112:0404/212115.312:ERROR:gl_surface_egl.cc(845)] : eglInitialize D3D9 failed with error EGL_NOT_INITIALIZED
[1972:3112:0404/212115.312:ERROR:gl_initializer_win.cc(232)] : GLSurfaceEGL::InitializeOneOff failed.
GpuProcessHostUIShim: The GPU process crashed!
GpuProcessHostUIShim: The GPU process exited normally. Everything is okay.


Comment 19 by kbr@chromium.org, Apr 5 2018

Here's Firefox's about:support from the same machine. It shows the following:

Vendor ID (GPU #2) 0x15ad
Device ID (GPU #2) 0x0405
Driver Date (GPU #2) 12-2-2014
Driver Version (GPU #2) 8.14.1.6

------

Application Basics
Name
Firefox
Version
42.0
Build ID
20151029151421
Update History
Show Update History
Update Channel
release
User Agent
Mozilla/5.0 (Windows NT 6.1; WOW64; rv:42.0) Gecko/20100101 Firefox/42.0
Profile Folder
Show Folder
Enabled Plugins
about:plugins
Build Configuration
about:buildconfig
Memory Use
about:memory
Registered Service Workers
about:serviceworkers
Multiprocess Windows
0/1 (default: false)
Safe Mode
false
Crash Reports for the Last 3 Days
Report ID
Submitted
All Crash Reports
This application has not been configured to display crash reports.
Extensions
Name
Version
Enabled
ID
Graphics
Adapter Description
RDPDD Chained DD
Adapter Description (GPU #2)
VMware SVGA 3D
Adapter Drivers
RDPDD
Adapter Drivers (GPU #2)
vm3dum64 vm3dum
Adapter RAM
Unknown
Adapter RAM (GPU #2)
64
Asynchronous Pan/Zoom
none
Device ID
0x0000
Device ID (GPU #2)
0x0405
Direct2D Enabled
Blocked for your graphics card because of unresolved driver issues.
DirectWrite Enabled
false (6.2.9200.17461)
Driver Date (GPU #2)
12-2-2014
Driver Version (GPU #2)
8.14.1.6
GPU #2 Active
false
GPU Accelerated Windows
0/1 Basic (OMTC) Blocked for your graphics card because of unresolved driver issues.
Subsys ID
00000000
Subsys ID (GPU #2)
040515ad
Supports Hardware H264 Decoding
false
Vendor ID
0x0000
Vendor ID (GPU #2)
0x15ad
WebGL Renderer
Blocked for your graphics card because of unresolved driver issues.
windowLayerManagerRemote
true
AzureCanvasBackend
skia
AzureContentBackend
cairo
AzureFallbackCanvasBackend
cairo
AzureSkiaAccelerated
0
Important Modified Preferences
Name
Value
browser.cache.disk.capacity
358400
browser.cache.disk.filesystem_reported
1
browser.cache.disk.smart_size_cached_value
358400
browser.cache.disk.smart_size.first_run
false
browser.cache.disk.smart_size.use_old_max
false
browser.cache.frecency_experiment
3
browser.download.importedFromSqlite
true
browser.places.smartBookmarksVersion
7
browser.sessionstore.upgradeBackup.latestBuildID
20151029151421
browser.startup.homepage_override.buildID
20151029151421
browser.startup.homepage_override.mstone
42.0
dom.apps.reset-permissions
true
dom.mozApps.used
true
dom.w3c_touch_events.expose
false
extensions.lastAppVersion
42.0
gfx.driver-init.appVersion
42.0
gfx.driver-init.deviceID
0x0000
gfx.driver-init.driverVersion


gfx.driver-init.feature-d2d
false
gfx.driver-init.feature-d3d11
false
gfx.driver-init.status
2
media.gmp-eme-adobe.abi
x86-msvc
media.gmp-eme-adobe.lastUpdate
1522901507
media.gmp-eme-adobe.version
15
media.gmp-gmpopenh264.abi
x86-msvc
media.gmp-gmpopenh264.lastUpdate
1522901508
media.gmp-gmpopenh264.version
1.5.3
media.gmp-manager.buildID
20151029151421
media.gmp-manager.lastCheck
1522901506
media.hardware-video-decoding.failed
false
network.cookie.prefsMigrated
true
network.predictor.cleaned-up
true
places.database.lastMaintenance
1522901583
places.history.expiration.transient_current_max_pages
104858
plugin.disable_full_page_plugin_for_types
application/pdf
plugin.importedState
true
privacy.cpd.offlineApps
true
privacy.cpd.siteSettings
true
privacy.sanitize.migrateFx3Prefs
true
privacy.sanitize.timeSpan
0
security.warn_viewing_mixed
false
storage.vacuum.last.index
0
storage.vacuum.last.places.sqlite
1522901583
user.js Preferences
Your profile folder contains a user.js file, which includes preferences that were not created by Firefox.
Important Locked Preferences
Name
Value
JavaScript
Incremental GC
true
Accessibility
Activated
false
Prevent Accessibility
0
Library Versions


Expected minimum version
Version in use
NSPR
4.10.10
4.10.10
NSS
3.19.4 Basic ECC
3.19.4 Basic ECC
NSSSMIME
3.19.4 Basic ECC
3.19.4 Basic ECC
NSSSSL
3.19.4 Basic ECC
3.19.4 Basic ECC
NSSUTIL
3.19.4
3.19.4
Experimental Features
Name
ID
Description
Active
End Date
Homepage


Hello Ken,

Is there anything we can do on our end to assist with the testing? I know you have the automated bots but see references to local testing (https://www.chromium.org/developers/testing/). Not sure if this is an instance where we are able to help or not but wanted to let the team know we are willing to run tests on our VMs with the sandbox code and provide any information you need.

-Steve

Comment 21 by kbr@chromium.org, Apr 5 2018

Cc: sunn...@chromium.org
Status: Started (was: Assigned)
Thanks Steve for your offer. The main issue is that our automated test machines are affected by this change, and our test harnesses aren't as robust as the normal operation of the browser (falling back to SwiftShader after 3 failed attempts to bring up the GPU process). This surely depends on the GPU hardware on the host, and I don't know what that is for these test machines.

We trust that VMWare users have configured their systems so that GPU acceleration will work, and be useful. We've worked around the problem and are proceeding with re-enabling GPU acceleration for current versions of VMWare's driver. We'll appreciate your help testing Chrome Canary in VMWare once this fix is incorporated into it. Thanks.

Comment 22 by kbr@chromium.org, Apr 5 2018

Note also that sunnyps@ had a concern that Chrome's GPU process crashing on the Windows bots might adversely impact test runtimes. I compared a couple of shards of telemetry_perf_unittests from my patch:

https://ci.chromium.org/buildbot/tryserver.chromium.win/win7_chromium_rel_ng/138173
https://chromium-swarm.appspot.com/task?id=3cadd08ee11bce10&refresh=10&show_raw=1
https://chromium-swarm.appspot.com/task?id=3cadd08f7e859d10&refresh=10&show_raw=1

to another current tryjob without my patch:

https://ci.chromium.org/buildbot/tryserver.chromium.win/win7_chromium_rel_ng/138721
https://chromium-swarm.appspot.com/task?id=3cb0491e5a4e4e10&refresh=10&show_raw=1
https://chromium-swarm.appspot.com/task?id=3cb04920127cff10&refresh=10&show_raw=1

The run times are basically identical, around 1m30s for each.

Project Member

Comment 23 by bugdroid1@chromium.org, Apr 12 2018

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/225fb01ffc9063debd4616743dcc165ec47da982

commit 225fb01ffc9063debd4616743dcc165ec47da982
Author: Ken Russell <kbr@chromium.org>
Date: Thu Apr 12 18:01:04 2018

Re-enable GPU acceleration for recent VMWare drivers.

Make GpuFeatureData collection robust to GPU process crashes. On many of
the VMWare Windows VMs used for testing, Chrome's GPU process crashes
during startup. This was causing DevTools' SystemInfo collection to break,
the browser to crash, and Telemetry to fail to start the browser, causing
failures of the telemetry_perf_unittests and telemetry_unittests
suites. Instead, robustly report that all GPU features are disabled, which
causes the associated tests to be skipped.

Bug:  800453 
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Change-Id: If80125798064316e0033fddd2d43f04748c76778
Reviewed-on: https://chromium-review.googlesource.com/952778
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Victor Miura <vmiura@chromium.org>
Cr-Commit-Position: refs/heads/master@{#550265}
[modify] https://crrev.com/225fb01ffc9063debd4616743dcc165ec47da982/content/browser/gpu/compositor_util.cc
[modify] https://crrev.com/225fb01ffc9063debd4616743dcc165ec47da982/gpu/config/software_rendering_list.json

Comment 24 by kbr@chromium.org, Apr 12 2018

Status: Fixed (was: Started)
This is finally done. Submitter, others, please test Chrome Canary soon once it is past revision 550265 and report any bugs either here, or file a new bug and email me the bug ID directly. Thanks.

Labels: Needs-Feedback
kbr@ Could you please help us with the repro steps to verify the fix from TE-end

Comment 26 by kbr@chromium.org, Apr 13 2018

Sorry, I don't know the VMWare VM configuration that reproduced this. Maybe the submitter can comment, but I don't think we will be able to verify this fix.

We were never impacted by the original issue reported on 754435 that resulted in VMWare being blacklisted. We tried reproducing on different versions of Chrome, drivers, and architectures, in both Horizon and Workstation products but were unsuccessful. It is our understanding users impacted by the issue would only have to launch Chrome and the tab strip at the top of the browser  would be corrupted. 
Project Member

Comment 28 by bugdroid1@chromium.org, Apr 17 2018

Labels: merge-merged-testbranch
The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/225fb01ffc9063debd4616743dcc165ec47da982

commit 225fb01ffc9063debd4616743dcc165ec47da982
Author: Ken Russell <kbr@chromium.org>
Date: Thu Apr 12 18:01:04 2018

Re-enable GPU acceleration for recent VMWare drivers.

Make GpuFeatureData collection robust to GPU process crashes. On many of
the VMWare Windows VMs used for testing, Chrome's GPU process crashes
during startup. This was causing DevTools' SystemInfo collection to break,
the browser to crash, and Telemetry to fail to start the browser, causing
failures of the telemetry_perf_unittests and telemetry_unittests
suites. Instead, robustly report that all GPU features are disabled, which
causes the associated tests to be skipped.

Bug:  800453 
Cq-Include-Trybots: luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;master.tryserver.chromium.win:win_optional_gpu_tests_rel
Change-Id: If80125798064316e0033fddd2d43f04748c76778
Reviewed-on: https://chromium-review.googlesource.com/952778
Commit-Queue: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Victor Miura <vmiura@chromium.org>
Cr-Commit-Position: refs/heads/master@{#550265}
[modify] https://crrev.com/225fb01ffc9063debd4616743dcc165ec47da982/content/browser/gpu/compositor_util.cc
[modify] https://crrev.com/225fb01ffc9063debd4616743dcc165ec47da982/gpu/config/software_rendering_list.json

Blocking: 879098

Sign in to add a comment