1189782 - tb 38.1.0 D3D11 related crash in driver for Nvidia Geforce 740M on startup when HWA, layers.acceleration and gfx.direct2d, are enabled

Status: RESOLVED INVALID

(Thunderbird :: General, defect)

Description

[u538344]

User Agent: Mozilla/5.0 (Windows NT 6.3; WOW64; rv:39.0) Gecko/20100101 Firefox/39.0
Build ID: 20150630154324

Steps to reproduce:

Settings layers.acceleration.disabled and gfx.direct2d.disabled to true fixes the crash but is hardly an acceptable solution. I've been getting this crash for weeks every time I thought I'd try the 38 series until I got a hunch about hw accel being the culprit.

Windows 8.1 64bit, Nvidia Geforce 740M with driver 353.62, Intel HD Graphics 4000

--

Update to 38.1.0
start tb


Actual results:

Instant crash (uploaded a million crash reports, bp-ee9e8976-2f3a-483b-b715-9db092150728 is one of the crashIDs)


Expected results:

TB should have opened

Comment 1

[Wayne Mery (:wsmwk)]

So YOU are the one who created that million crash reports!  :)

There's some chance this might be bug 1187764. If so, you will find that 40.0 beta downloadable from https://www.mozilla.org/en-US/thunderbird/channel/ will work.  Could you try it for us?

If it fails, please post additional crash IDs.
If it works, please let us know.

Comment 2

[Wayne Mery (:wsmwk)]

 nvwgf2um.dll@0xf73ab 
bp-ee9e8976-2f3a-483b-b715-9db092150728
Ø 0 	nvwgf2um.dll	nvwgf2um.dll@0xf73ab	
Ø 1 	nvwgf2um.dll	nvwgf2um.dll@0xf9bce	
Ø 2 	nvwgf2um.dll	nvwgf2um.dll@0xa51c9	
Ø 3 	nvumdshim.dll	nvumdshim.dll@0x2273e	
Ø 4 	nvumdshim.dll	nvumdshim.dll@0x30b63	
Ø 5 	nvumdshim.dll	nvumdshim.dll@0x1b9e8	
6 	d3d11.dll	d3d11.dll@0xcd6a9	
7 	d3d11.dll	CContext::ID3D11DeviceContext2_CopySubresourceRegion_<1>(ID3D11DeviceContext2*, ID3D11Resource*, unsigned int, unsigned int, unsigned int, unsigned int, ID3D11Resource*, unsigned int, D3D11_BOX const*)	
8 	d3d11.dll	NDXGI::CDevice::Blt(IDXGIResource*, tagRECT const*, tagRECT const*, unsigned int, IDXGIResource*, tagRECT const*, SUBRESOURCE_BLT_MAP const*, unsigned __int64, unsigned int, unsigned int)	
Ø 9 	dxgi.dll	dxgi.dll@0x40a93	
Ø 10 	dxgi.dll	dxgi.dll@0x1d67b	
Ø 11 	dxgi.dll	dxgi.dll@0x10bad	
Ø 12 	dxgi.dll	dxgi.dll@0x4271b	
13 	xul.dll	mozilla::layers::CompositorD3D11::EndFrame()	gfx/layers/d3d11/CompositorD3D11.cpp
14 	xul.dll	mozilla::layers::LayerManagerComposite::Render()	gfx/layers/composite/LayerManagerComposite.cpp
15 	xul.dll	mozilla::layers::LayerManagerComposite::EndTransaction(void (*)(mozilla::layers::PaintedLayer*, gfxContext*, nsIntRegion const&, mozilla::layers::DrawRegionClip, nsIntRegion const&, void*), void*, mozilla::layers::LayerManager::EndTransactionFlags)	gfx/layers/composite/LayerManagerComposite.cpp

Comment 3

[u538344]

40.0b1 also crashes with the hw accel options enabled; it doesn't fix the problem.

Comment 4

[Wayne Mery (:wsmwk)]

Need crash IDs please.

Comment 5

[u538344]

Right, my bad, sorry.

It seems gfx.direct2d isn't related actually btw. Only layers.acceleration causes the crash.

bp-311f6874-ff26-4735-8966-50f122150731

Comment 6

[GMA]

https://downloadcenter.intel.com/download/24973/Intel-HD-Graphics-Driver-for-Windows-7-8-8-1-64-bit

Can you update to the latest Intel 10.18.10.4226 driver for HD4000 and see if it still crashes?

Comment 7

[u538344]

(In reply to GMA from comment #6)
> https://downloadcenter.intel.com/download/24973/Intel-HD-Graphics-Driver-for-
> Windows-7-8-8-1-64-bit
> 
> Can you update to the latest Intel 10.18.10.4226 driver for HD4000 and see
> if it still crashes?

No change

Comment 8

[q1]

I am going to guess that line 1131 is the culprit:

  1118      nsIntRegionRectIterator iter(mInvalidRegion);
  1119      const IntRect* r;
  1120      uint32_t i = 0;
  1121      while ((r = iter.Next()) != nullptr) {
  1122        RECT rect;
  1123        rect.left = r->x;
  ...
  1128        rects.push_back(rect);
  1129      }
  1130
  1131      params.pDirtyRects = &rects.front();
  1132      chain->Present1(presentInterval, mDisableSequenceForNextFrame ? DXGI_PRESENT_DO_NOT_SEQUENCE : 0, &params);

because it assumes that rects is nonempty.

Comment 9

[q1]

(In reply to q1 from comment #8)
> I am going to guess that line 1131 is the culprit:

That would be in CompositorD3D11.cpp, function EndFrame.

Comment 10

[NVD]

http://www.geforce.com/whats-new/articles/geforce-355-60-whql-driver-released

Please test GeForce 355.60 WHQL driver.

Comment 11

[u538344]

(In reply to NVD from comment #10)
> http://www.geforce.com/whats-new/articles/geforce-355-60-whql-driver-released
> 
> Please test GeForce 355.60 WHQL driver.

No change

Comment 12

[NVD]

http://www.geforce.com/whats-new/articles/geforce-358-50-whql-driver-released

Please test 358.50 WHQL driver.

Comment 13

[Wayne Mery (:wsmwk)]

marnick, 
Do you still crash in version 45 with HWA enabled?
If you do, is it  better with (in this order)
1. current beta version from http://www.mozilla.org/en-US/thunderbird/channel/ ?
2. latest NVIDIA driver (As of Dec 16, 2016 driver version is up to 376.33) ?

Comment 14

[u538344]

I confirm I can set both preferences to true today and nothing bad happens. Problem seems to be gone.