Closed Bug 806786 Opened 7 years ago Closed 4 years ago

crash in mozilla::layers::LayerManagerD3D10::Render @ GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport with Intel HD Graphics/NVIDIA dual GPUs and Intel driver version 8.15.10.2202 and 2302

Categories

(Core :: Graphics, defect, critical)

All
Windows 7
defect
Not set
critical

Tracking

()

RESOLVED INCOMPLETE
Tracking Status
firefox19 + ---
firefox20 + ---

People

(Reporter: scoobidiver, Assigned: bas.schouten)

References

(Blocks 3 open bugs)

Details

(Keywords: crash)

Crash Data

It started slightly spiking across all versions around October 15th, probably after a custom graphics driver update.
Intel driver version 8.15.10.2202 is the first version not gfx-blocked for Intel HD Graphics.

Signature 	GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport(void*&) More Reports Search
UUID	5edfe139-7ecc-4cef-bdc2-940dc2121029
Date Processed	2012-10-29 21:39:08
Uptime	1255
Last Crash	21.0 minutes before submission
Install Age	33.2 minutes since version was first installed.
Install Time	2012-10-29 21:05:41
Product	Firefox
Version	19.0a1
Build ID	20121029030553
Release Channel	nightly
OS	Windows NT
OS Version	6.1.7601 Service Pack 1
Build Architecture	x86
Build Architecture Info	GenuineIntel family 6 model 37 stepping 5
Crash Reason	EXCEPTION_ACCESS_VIOLATION_WRITE
Crash Address	0xc0
App Notes 	
AdapterVendorID: 0x8086, AdapterDeviceID: 0x0046, AdapterSubsysID: 04881025, AdapterDriverVersion: 8.15.10.2202
Has dual GPUs. GPU #2: AdapterVendorID2: 0x10de, AdapterDeviceID2: 0x0df4, AdapterSubsysID2: 04881025, AdapterDriverVersion2: 8.17.12.5997D2D? D2D+ DWrite? DWrite+ D3D10 Layers? D3D10 Layers+ 
EMCheckCompatibility	True
Adapter Vendor ID	0x8086
Adapter Device ID	0x0046
Total Virtual Memory	4294836224
Available Virtual Memory	3432722432
System Memory Use Percentage	39
Available Page File	5638119424
Available Physical Memory	2393341952

Frame 	Module 	Signature 	Source
0 	igd10umd32.dll 	GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport 	
1 	igd10umd32.dll 	GHAL3D::CGen4StateProcessorStream::GetColorCalculatorUnitState 	
2 	igd10umd32.dll 	GHAL3D::CGen4StateProcessorStream::GetPipelineStatePointers 	
3 	igd10umd32.dll 	GHAL3D::CGen4PrivateStateProcessor::BindPipeline 	
4 	igd10umd32.dll 	D3D10::CDevice::BindPipeline 	
5 	igd10umd32.dll 	D3D10::CDevice::FlushState 	
6 	igd10umd32.dll 	D3D10::CBltContext::SendPrimitive 	
7 	igd10umd32.dll 	D3D10::CDXGIBlt::Blt 	
8 	igd10umd32.dll 	D3D10::CDevice::Blt 	
9 	igd10umd32.dll 	DXGIAPI::Blt 	
10 	nvumdshim.dll 	nvumdshim.dll@0x1b496 	
11 	nvumdshim.dll 	nvumdshim.dll@0x1f042 	
12 	dxgi.dll 	CD3D10Device::Blt 	
13 	dxgi.dll 	CDXGISwapChainBase<IDXGISwapChain>::PresentImpl 	
14 	dxgi.dll 	CDXGISwapChainBase<IDXGISwapChain>::Present 	
15 	nvdxgiwrap.dll 	nvdxgiwrap.dll@0x1b88 	
16 	xul.dll 	mozilla::layers::LayerManagerD3D10::Render 	gfx/layers/d3d10/LayerManagerD3D10.cpp:821
17 	xul.dll 	mozilla::layers::LayerManagerD3D10::EndTransaction 	gfx/layers/d3d10/LayerManagerD3D10.cpp:381
18 	xul.dll 	mozilla::layers::LayerManagerD3D10::EndEmptyTransaction 	gfx/layers/opengl/LayerManagerOGL.cpp:625
19 	xul.dll 	PresShell::Paint 	layout/base/nsPresShell.cpp:5275
20 	xul.dll 	nsViewManager::Refresh 	view/src/nsViewManager.cpp:372
21 	xul.dll 	nsViewManager::PaintWindow 	view/src/nsViewManager.cpp:705
22 	xul.dll 	nsView::PaintWindow 	view/src/nsView.cpp:1039
23 	xul.dll 	nsWindow::OnPaint 	widget/windows/nsWindowGfx.cpp:534
24 	xul.dll 	nsWindow::ProcessMessage 	widget/windows/nsWindow.cpp:4751
25 	xul.dll 	nsWindow::WindowProcInternal 	widget/windows/nsWindow.cpp:4360
26 	xul.dll 	CallWindowProcCrashProtected 	xpcom/base/nsCrashOnException.cpp:32
27 	xul.dll 	nsWindow::WindowProc 	widget/windows/nsWindow.cpp:4306
28 	user32.dll 	InternalCallWinProc 	

More reports at:
https://crash-stats.mozilla.com/report/list?signature=GHAL3D%3A%3ACGen4StateProcessorStream%3A%3AGetColorCalculatorViewport%28void*%26%29
Hardware: x86 → All
Blocks: 605779
Blocks: 605749
It's #13 top browser crasher in the first day of 19.0.

Here are correlations per Intel driver version:
    100% (131/131) vs.   4% (2532/61447) igd10umd32.dll
         89% (116/131) vs.   1% (336/61447) 8.15.10.2202
         11% (15/131) vs.   0% (202/61447) 8.15.10.2302
Keywords: topcrash
Summary: crash in mozilla::layers::LayerManagerD3D10::Render @ GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport with Intel HD Graphics/NVIDIA dual GPUs and Intel driver version 8.15.10.2202 → crash in mozilla::layers::LayerManagerD3D10::Render @ GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport with Intel HD Graphics/NVIDIA dual GPUs and Intel driver version 8.15.10.2202 and 2302
Depends on: 843273
Spoke with KaiRo, we're going to wait to see if this remains a top crash with more than a day of data.
It's #12 top browser crasher one day later.
Do we have correlations to any particular chipsets and URLs?
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #4)
> Do we have correlations to any particular chipsets and URLs?
For the chipset, see the summary and also bug 843273.
(In reply to Scoobidiver from comment #5)
> (In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #4)
> > Do we have correlations to any particular chipsets and URLs?
> For the chipset, see the summary and also bug 843273.

Sorry but it's still not clear to me. Am I to assume that this affects *all* Intel HD graphics chipsets and *all* NVidia dual GPU chipsets?

> Vendor:  0x8086
> Devices: 0x0046
I don't know what vendor and devices this refers to. It would be great if QA could get that information up front.
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #6) 
> Sorry but it's still not clear to me. Am I to assume that this affects *all*
> Intel HD graphics chipsets and *all* NVidia dual GPU chipsets?
It affects the first generation of HD Graphics. I don't know what NVIDIA dual GPU chipsets are and haven't seen that because dual GPUs are composed of one powerful GPU and a low consumption one (Intel in that case).

> > Vendor:  0x8086
> > Devices: 0x0046
> I don't know what vendor and devices this refers to. It would be great if QA
> could get that information up front.
0x8086 means Intel and 0x0046, HD Graphics, and both are in the summary.
(In reply to Scoobidiver from comment #7)
> 0x8086 means Intel and 0x0046, HD Graphics, and both are in the summary.

Yup, I can see that. I guess I'm trying to figure out specifically which version of Intels HD GPUs are affected. There have been many iterations over the years.
(In reply to Anthony Hughes, Mozilla QA (:ashughes) from comment #8)
> I guess I'm trying to figure out specifically which version of Intels HD GPUs are
> affected.
This one (only 0x0046 because 0x0042 isn't included in dual GPUs): https://en.wikipedia.org/wiki/Comparison_of_Intel_graphics_processing_units#Fifth_generation
(In reply to Scoobidiver from comment #9)
> This one (only 0x0046 because 0x0042 isn't included in dual GPUs):
> https://en.wikipedia.org/wiki/
> Comparison_of_Intel_graphics_processing_units#Fifth_generation

Thanks, I've asked Juan to check to see if QA has hardware on hand meeting this criteria.
I haven't been able to find a machine with these specs. I tried to add a gfx card to a couple of machines that had an existing nvidia card, but there wasn't any slot for it.
(In reply to juan becerra [:juanb] from comment #11)
> I haven't been able to find a machine with these specs. I tried to add a gfx
> card to a couple of machines that had an existing nvidia card, but there
> wasn't any slot for it.

Can we order an affected model? It was difficult to find, but here's one: http://www.amazon.com/Lenovo-ThinkPad-Digitizer-Professional-Wireless/dp/B007JMK336
(In reply to Alex Keybl [:akeybl] from comment #12)
> (In reply to juan becerra [:juanb] from comment #11)
> > I haven't been able to find a machine with these specs. I tried to add a gfx
> > card to a couple of machines that had an existing nvidia card, but there
> > wasn't any slot for it.
> 
> Can we order an affected model? It was difficult to find, but here's one:
> http://www.amazon.com/Lenovo-ThinkPad-Digitizer-Professional-Wireless/dp/
> B007JMK336

Nevermind, this isn't an option according to http://www.pcmag.com/article2/0,2817,2360199,00.asp

"the X201 runs on integrated graphics only"
Found Sony VAIO VPC-Z116GXS, but it's the wrong generation Intel CPU. May be very difficult to get our hands on an impacted config.

Let's see if https://bugzilla.mozilla.org/show_bug.cgi?id=843273#c1 is the right way to go.
Benoit should prepare a built-in blacklist patch in bug 843273 and we can see how that affects this crash's volume after beta 4.
Clarkdale/Arrandale HD graphics users that want hardware acceleration back after bug 843274 has turned it off, please install the latest 8.15.10.2993 drivers. These drivers will only install on Clarkdale/Arrandale hardware, not newer hardware like Sandy Bridge or Ivy Bridge.

32bit
http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&ProdId=3231&DwnldID=22519

64bit
http://downloadcenter.intel.com/Detail_Desc.aspx?agr=Y&ProdId=3231&DwnldID=22521
Bas, given that this is a tracked bug I feel more comfortable when somebody is assigned to it.  If you (or others) know of somebody else that should own it, great, let's reassign.
Assignee: nobody → bas
(In reply to Milan Sreckovic [:milan] from comment #17)
> Bas, given that this is a tracked bug I feel more comfortable when somebody
> is assigned to it.  If you (or others) know of somebody else that should own
> it, great, let's reassign.

Since we decided to solve this by blacklisting Benoit Jacob might be a better owner, but I think there's no action items here right now anyway.
There's one crash in 20.0 despite Direct2D not being forced-enabled: bp-f587e4a6-321e-408b-ac7a-f102a2130405. Maybe the GPU was switched from the NVIDIA one to the Intel one.
Keywords: topcrash
Crash Signature: [@ GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport(void*&)] → [@ GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport(void*&)] [@ GHAL3D::CGen4StateProcessorStream::GetColorCalculatorViewport]
Closing this bug as we no longer see crashes in these signatures. Please reopen if you have any new leads.
Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.