1285085 - Intermittent test_set_window_size.py TestSetWindowSize.test_possible_to_request_window_larger_than_screen | crash [@ mozilla::gfx::Log<1,mozilla::gfx::CriticalLogger>::WriteLog() | mozilla::gfx::DrawTargetD2D1::Init()]

Reporter

Description

•

8 years ago

treeherder

Filed by: philringnalda@gmail.com

https://treeherder.mozilla.org/logviewer.html#?job_id=313903&repo=autoland

http://archive.mozilla.org/pub/firefox/tinderbox-builds/autoland-win64-debug/1467822443/autoland_win8_64-debug_test-marionette-bm112-tests1-windows-build23.txt.gz

Comment hidden (Intermittent Failures Robot)

22 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* autoland: 9
* mozilla-inbound: 7
* mozilla-central: 4
* mozilla-aurora: 1
* fx-team: 1

Platform breakdown:
* windows8-64: 22

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-07-04&endday=2016-07-10&tree=all

Bas Schouten (:bas.schouten)

Comment 2

•

8 years ago

Hrm, I wonder why the reason isn't getting logged here before asserting to we could see the HRESULT code. However I feel this is most likely just a device driver reset.

Whiteboard: [gfx-noted]

Comment hidden (Intermittent Failures Robot)

29 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-inbound: 12
* fx-team: 7
* mozilla-aurora: 5
* mozilla-central: 4
* autoland: 1

Platform breakdown:
* windows8-64: 29

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-07-11&endday=2016-07-17&tree=all

Andreas Tolfsen ❲:ato❳

Comment 4

•

8 years ago

I believe this happens when window.resizeTo is called with arguments larger than the available physical dimensions of the screen, only on Windows 10.

OS: Unspecified → Windows 10

Hardware: Unspecified → x86_64

Andreas Tolfsen ❲:ato❳

Comment 5

•

8 years ago

This seems to reproduce quite consistently on Windows 10 debug builds, but _is not limited_ to Win 10.

OS: Windows 10 → Windows

Andreas Tolfsen ❲:ato❳

Comment 6

•

8 years ago

https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085

Comment hidden (Intermittent Failures Robot)

36 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-inbound: 16
* autoland: 11
* mozilla-central: 4
* mozilla-aurora: 3
* try: 1
* fx-team: 1

Platform breakdown:
* windows8-64: 33
* windows7-32-vm: 1
* windows10-64: 1
* linux64: 1

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-07-18&endday=2016-07-24&tree=all

Andreas Tolfsen ❲:ato❳

Comment 8

•

8 years ago

The graphics card used on the Windows 10 try machines is NVIDIA GPU GeForce GT 610 per node.

Edwin Flores [inactive from 2016-12-01] [:eflores] [:edwin]

Comment 9

•

8 years ago

(In reply to Bas Schouten (:bas.schouten) from comment #2)
> Hrm, I wonder why the reason isn't getting logged here before asserting to
> we could see the HRESULT code. However I feel this is most likely just a
> device driver reset.

It is being logged, but to gecko.log instead:

> [GFX1]: [D2D1.1] CreateBitmapFromDxgiSurface failure Code: 0x88990003 format 0

Which appears to be D2DERR_UNSUPPORTED_OPERATION.

Edwin Flores [inactive from 2016-12-01] [:eflores] [:edwin]

Comment 10

•

8 years ago

Is it possible that the max dimensions of an IDXGISurface are larger than the max bitmap dimensions? I see that we do a size check in Init(IntSize, SurfaceFormat) against the max bitmap size, but not in Init(ID3D11Texture2D*, SurfaceFormat).

Ryan VanderMeulen [:RyanVM]

Updated

•

8 years ago

status-firefox49: --- → affected

status-firefox50: --- → affected

Comment hidden (Intermittent Failures Robot)

17 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* autoland: 7
* mozilla-aurora: 6
* mozilla-inbound: 3
* mozilla-central: 1

Platform breakdown:
* windows8-64: 15
* linux64: 2

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-07-25&endday=2016-07-31&tree=all

Henrik Skupin [:whimboo][⌚️UTC+1]

Comment 12

•

8 years ago

The marionette logging has been improved lately, so crashes like those report as `test_set_window_size.py TestSetWindowSize.test_possible_to_request_window_larger_than_screen` now.

Summary: Intermittent runner.py | application crashed [@ mozilla::gfx::Log<1,mozilla::gfx::CriticalLogger>::WriteLog(std::basic_string<char,std::char_traits<char>,std::allocator<char> > const &)] → Intermittent test_set_window_size.py TestSetWindowSize.test_possible_to_request_window_larger_than_screen | crash [@ mozilla::gfx::Log<1,mozilla::gfx::CriticalLogger>::WriteLog(std::basic_string<char,std::char_traits<char>,std::allocator<char> > const &)]

Henrik Skupin [:whimboo][⌚️UTC+1]

Comment 13

•

8 years ago

The assertion which is shown from this line of code:

https://dxr.mozilla.org/mozilla-central/rev/fe895421dfbe1f1f8f1fc6a39bb20774423a6d74/gfx/2d/Logging.h#510

Assertion failure: [GFX1]: [D2D1.1] CreateBitmapFromDxgiSurface failure Code: 0x88990003 format 0, at c:\builds\moz2_slave\autoland-w64-d-000000000000000\build\src\obj-firefox\dist\include\mozilla/gfx/Logging.h:509
#01: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x777577]
#02: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xfd88cd]
#03: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xfe9386]
#04: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x119b696]
#05: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11a90be]
#06: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11a658f]
#07: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x118305c]
#08: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x117e036]
#09: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1118449]
#10: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11779f6]
#11: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1183707]
#12: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11880c7]
#13: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x115f849]
#14: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x115843c]
#15: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1156f87]
#16: DumpFrameArray[C:\slave\test\build\application\firefox\xul.dll +0x335e5db]
#17: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f60a23]
#18: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f607ae]
#19: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f60572]
#20: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f61b16]
#21: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f5e58d]
#22: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x301cc6b]
#23: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x301f9af]
#24: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x30271b1]
#25: CallWindowProcCrashProtected[C:\slave\test\build\application\firefox\xul.dll +0x1bd11e]
#26: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3027028]
#27: IsRectEmpty[C:\windows\system32\USER32.dll +0x3e95]
#28: WaitMessage[C:\windows\system32\USER32.dll +0x2a62]
#29: PostMessageW[C:\windows\system32\USER32.dll +0x294d]
#30: KiUserCallbackDispatcher[C:\windows\SYSTEM32\ntdll.dll +0x4b47]
#31: SetTimer[C:\windows\system32\USER32.dll +0x203a]
#32: SetTimer[C:\windows\system32\USER32.dll +0x204c]
#33: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3003911]
#34: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f91d5c]
#35: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f9602f]
#36: XRE_AddStaticComponent[C:\slave\test\build\application\firefox\xul.dll +0x254283]
#37: NS_StringSetIsVoid[C:\slave\test\build\application\firefox\xul.dll +0x28eca7]
#38: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x799df3]
#39: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x762a5e]
#40: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x7626f6]
#41: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f9882b]
#42: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x300738c]
#43: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3b06082]
#44: XRE_GetFileFromPath[C:\slave\test\build\application\firefox\xul.dll +0x3b89e41]
#45: XRE_GetFileFromPath[C:\slave\test\build\application\firefox\xul.dll +0x3b86c93]
#46: XRE_main[C:\slave\test\build\application\firefox\xul.dll +0x3b8b470]
#47: ???[C:\slave\test\build\application\firefox\firefox.exe +0x1f49]
#48: ???[C:\slave\test\build\application\firefox\firefox.exe +0x173a]
#49: ???[C:\slave\test\build\application\firefox\firefox.exe +0x23b3]
#50: TargetNtUnmapViewOfSection[C:\slave\test\build\application\firefox\firefox.exe +0x404b4]
#51: BaseThreadInitThunk[C:\windows\system32\KERNEL32.DLL +0x167e]
#52: RtlUserThreadStart[C:\windows\SYSTEM32\ntdll.dll +0x1c3f1]

There is also another assertion from the graphics logger which might be related:

Hit MOZ_CRASH(GFX: An assert from the graphics logger) at c:\builds\moz2_slave\autoland-w64-d-000000000000000\build\src\obj-firefox\dist\include\mozilla/gfx/Logging.h:510
#01: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x777577]
#02: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xfd88cd]
#03: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xfe9386]
#04: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x119b696]
#05: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11a90be]
#06: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11a658f]
#07: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x118305c]
#08: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x117e036]
#09: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1118449]
#10: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11779f6]
#11: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1183707]
#12: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11880c7]
#13: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x115f849]
#14: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x115843c]
#15: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1156f87]
#16: DumpFrameArray[C:\slave\test\build\application\firefox\xul.dll +0x335e5db]
#17: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f60a23]
#18: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f607ae]
#19: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f60572]
#20: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f61b16]
#21: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f5e58d]
#22: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x301cc6b]
#23: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x301f9af]
#24: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x30271b1]
#25: CallWindowProcCrashProtected[C:\slave\test\build\application\firefox\xul.dll +0x1bd11e]
#26: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3027028]
#27: IsRectEmpty[C:\windows\system32\USER32.dll +0x3e95]
#28: WaitMessage[C:\windows\system32\USER32.dll +0x2a62]
#29: PostMessageW[C:\windows\system32\USER32.dll +0x294d]
#30: KiUserCallbackDispatcher[C:\windows\SYSTEM32\ntdll.dll +0x4b47]
#31: SetTimer[C:\windows\system32\USER32.dll +0x203a]
#32: SetTimer[C:\windows\system32\USER32.dll +0x204c]
#33: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3003911]
#34: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f91d5c]
#35: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f9602f]
#36: XRE_AddStaticComponent[C:\slave\test\build\application\firefox\xul.dll +0x254283]
#37: NS_StringSetIsVoid[C:\slave\test\build\application\firefox\xul.dll +0x28eca7]
#38: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x799df3]
#39: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x762a5e]
#40: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x7626f6]
#41: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f9882b]
#42: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x300738c]
#43: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3b06082]
#44: XRE_GetFileFromPath[C:\slave\test\build\application\firefox\xul.dll +0x3b89e41]
#45: XRE_GetFileFromPath[C:\slave\test\build\application\firefox\xul.dll +0x3b86c93]
#46: XRE_main[C:\slave\test\build\application\firefox\xul.dll +0x3b8b470]
#47: ???[C:\slave\test\build\application\firefox\firefox.exe +0x1f49]
#48: ???[C:\slave\test\build\application\firefox\firefox.exe +0x173a]
#49: ???[C:\slave\test\build\application\firefox\firefox.exe +0x23b3]
#50: TargetNtUnmapViewOfSection[C:\slave\test\build\application\firefox\firefox.exe +0x404b4]
#51: BaseThreadInitThunk[C:\windows\system32\KERNEL32.DLL +0x167e]
#52: RtlUserThreadStart[C:\windows\SYSTEM32\ntdll.dll +0x1c3f1]

Comment hidden (Intermittent Failures Robot)

11 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* autoland: 6
* mozilla-beta: 3
* mozilla-central: 1
* mozilla-aurora: 1

Platform breakdown:
* windows8-64: 11

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-08-15&endday=2016-08-21&tree=all

Comment hidden (Intermittent Failures Robot)

28 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* autoland: 11
* mozilla-inbound: 6
* mozilla-central: 6
* mozilla-aurora: 2
* fx-team: 2
* oak: 1

Platform breakdown:
* windows8-64: 27
* osx-10-10: 1

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-08-22&endday=2016-08-28&tree=all

BMO Automation

Comment 17

•

8 years ago

Bulk assigning P3 to all open intermittent bugs without a priority set in Firefox components per bug 1298978.

Priority: -- → P3

Ryan VanderMeulen [:RyanVM]

Comment 18

•

8 years ago

Sadly, this was not fixed by bug 1051567.

status-firefox51: --- → affected

Henrik Skupin [:whimboo][⌚️UTC+1]

Comment 19

•

8 years ago

(In reply to Ryan VanderMeulen [:RyanVM] from comment #18)
> Sadly, this was not fixed by bug 1051567.

Because the fix on the other bug was mainly for OSX and Linux inside the IPC component. The assertion and crash as shown here are different and are graphics related as the stack shows: "mozilla::gfx::DrawTargetD2D1::Init".

https://hg.mozilla.org/mozilla-central/file/cd0957de64a0186a148f5f781044456b94d231c6/gfx/2d/DrawTargetD2D1.cpp#l995

As OrangeFactor shows this bug is Windows 8 x64 only and seems to only happen for debug builds.

Here the assertion from the gecko.log:

http://mozilla-releng-blobs.s3.amazonaws.com/blobs/mozilla-central/sha512/e8946172d44c679c3ce470598f414676c68e131f2ee773729b5a5e2dc519867d05848c5c9655027cf2d1e6b303b4ab0441374b76b7b877aee21c9445c8b87666

1472570217300	Marionette	TRACE	conn7 -> [0,6,"setWindowSize",{"width":100000,"height":100000}]
1472570217310	Marionette	TRACE	conn7 <- [1,6,null,{"width":16384,"height":16384}]
[GFX1-]: Detected rendering device reset on refresh: 6
[GFX1-]: Detected rendering device reset on refresh: 6
[GFX1]: [D2D1.1] CreateBitmapFromDxgiSurface failure Code: 0x88990003 format 0
Assertion failure: [GFX1]: [D2D1.1] CreateBitmapFromDxgiSurface failure Code: 0x88990003 format 0, at c:\builds\moz2_slave\m-cen-w64-d-000000000000000000\build\src\obj-firefox\dist\include\mozilla/gfx/Logging.h:510
#01: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x785d27]
#02: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xfc3da9]
#03: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0xfd4e9a]
#04: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1188c96]
#05: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1196dea]
#06: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x119454f]
#07: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x11699ac]
#08: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1163f3e]
#09: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1105251]
#10: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x115d00a]
#11: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x116a593]
#12: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x116f30b]
#13: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x114a31d]
#14: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x1163ab4]
#15: soundtouch::SoundTouch::operator=[C:\slave\test\build\application\firefox\xul.dll +0x116321b]
#16: DumpFrameArray[C:\slave\test\build\application\firefox\xul.dll +0x3374053]
#17: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f6f3e3]
#18: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f6f16e]
#19: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f6ef32]
#20: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f704d6]
#21: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2f6cf55]
#22: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x302be9f]
#23: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x302ebe3]
#24: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x30361fd]
#25: CallWindowProcCrashProtected[C:\slave\test\build\application\firefox\xul.dll +0x1c566a]
#26: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3036074]
#27: IsRectEmpty[C:\windows\system32\USER32.dll +0x3e95]
#28: WaitMessage[C:\windows\system32\USER32.dll +0x2a62]
#29: PostMessageW[C:\windows\system32\USER32.dll +0x294d]
#30: KiUserCallbackDispatcher[C:\windows\SYSTEM32\ntdll.dll +0x4b47]
#31: SetTimer[C:\windows\system32\USER32.dll +0x203a]
#32: SetTimer[C:\windows\system32\USER32.dll +0x204c]
#33: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3012f61]
#34: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2fa0ab0]
#35: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2fa4d9f]
#36: XRE_AddStaticComponent[C:\slave\test\build\application\firefox\xul.dll +0x25d127]
#37: NS_StringSetIsVoid[C:\slave\test\build\application\firefox\xul.dll +0x297fdb]
#38: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x7a874f]
#39: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x7711ea]
#40: mozilla::net::LoadInfo::TriggeringPrincipal[C:\slave\test\build\application\firefox\xul.dll +0x770e82]
#41: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x2fa75ab]
#42: imgLoader::SupportImageWithMimeType[C:\slave\test\build\application\firefox\xul.dll +0x3016a54]
#43: workerlz4_maxCompressedSize[C:\slave\test\build\application\firefox\xul.dll +0x3b3378e]
#44: XRE_GetFileFromPath[C:\slave\test\build\application\firefox\xul.dll +0x3bb9919]
#45: XRE_GetFileFromPath[C:\slave\test\build\application\firefox\xul.dll +0x3bb66ab]
#46: XRE_main[C:\slave\test\build\application\firefox\xul.dll +0x3bbaee0]
#47: ???[C:\slave\test\build\application\firefox\firefox.exe +0x1f49]
#48: ???[C:\slave\test\build\application\firefox\firefox.exe +0x173a]
#49: ???[C:\slave\test\build\application\firefox\firefox.exe +0x23b3]
#50: TargetNtUnmapViewOfSection[C:\slave\test\build\application\firefox\firefox.exe +0x404b4]
#51: BaseThreadInitThunk[C:\windows\system32\KERNEL32.DLL +0x167e]
#52: RtlUserThreadStart[C:\windows\SYSTEM32\ntdll.dll +0x1c3f1]

Severity: normal → critical

Crash Signature: [@ mozilla::gfx::Log<1,mozilla::gfx::CriticalLogger>::WriteLog() | mozilla::gfx::DrawTargetD2D1::Init()]

Keywords: crash

OS: Windows → Windows 8

Summary: Intermittent test_set_window_size.py TestSetWindowSize.test_possible_to_request_window_larger_than_screen | crash [@ mozilla::gfx::Log<1,mozilla::gfx::CriticalLogger>::WriteLog(std::basic_string<char,std::char_traits<char>,std::allocator<char> > const &)] → Intermittent test_set_window_size.py TestSetWindowSize.test_possible_to_request_window_larger_than_screen | crash [@ mozilla::gfx::Log<1,mozilla::gfx::CriticalLogger>::WriteLog() | mozilla::gfx::DrawTargetD2D1::Init()]

Henrik Skupin [:whimboo][⌚️UTC+1]

Comment 20

•

8 years ago

Edwin, is the last comment helpful to see better what's happening?

Flags: needinfo?(edwin)

Milan Sreckovic [:milan] (needinfo for best results)

Comment 21

•

8 years ago

There is a driver reset during the test run, with an out of memory reason.  By the time we hit this assert, it's already too late - the reset has already happened.  This comes just before the driver reset detection:
1472570217300	Marionette	TRACE	conn7 -> [0,6,"setWindowSize",{"width":100000,"height":100000}]
1472570217310	Marionette	TRACE	conn7 <- [1,6,null,{"width":16384,"height":16384}]
and it feels like either 100,000^2 or 16k^2 could be something that gets us to OOM.

Comment hidden (Intermittent Failures Robot)

22 automation job failures were associated with this bug in the last 7 days.

Repository breakdown:
* mozilla-aurora: 6
* mozilla-inbound: 5
* autoland: 5
* mozilla-beta: 4
* mozilla-central: 2

Platform breakdown:
* windows8-64: 22

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1285085&startday=2016-08-29&endday=2016-09-04&tree=all

Henrik Skupin [:whimboo][⌚️UTC+1]

Comment 23

•

8 years ago

(In reply to Milan Sreckovic [:milan] from comment #21)
> and it feels like either 100,000^2 or 16k^2 could be something that gets us
> to OOM.

Milan, is the test correct in using such a high resolution? Should it be reduced to something like twice the current screen size? Or should we correctly handle that resolution change in Firefox?

Milan Sreckovic [:milan] (needinfo for best results)

Comment 24

•

8 years ago

Bas, in DrawTargetD2D1::Init, is there a way to know what size bitmap we're trying to create before the call to mDC->CreateBitmapFromDxgiSurface?  Right now, if that call succeeds, we do:
aTexture->GetDesc(&desc);
mSize.width = desc.Width;
mSize.height = desc.Height;
to get the size, but can we get the size before that call, or even if the call fails (if it fails, would the size be 0, invalid, or what we tried to get but couldn't?)

Flags: needinfo?(bas)

Bas Schouten (:bas.schouten)

Comment 25

•

8 years ago

There is indeed, see also: https://msdn.microsoft.com/en-us/library/windows/desktop/bb174566(v=vs.85).aspx

Flags: needinfo?(bas)

Bas Schouten (:bas.schouten)

Comment 26

•

8 years ago

(In reply to Milan Sreckovic [:milan] from comment #24)
> Bas, in DrawTargetD2D1::Init, is there a way to know what size bitmap we're
> trying to create before the call to mDC->CreateBitmapFromDxgiSurface?  Right
> now, if that call succeeds, we do:
> aTexture->GetDesc(&desc);
> mSize.width = desc.Width;
> mSize.height = desc.Height;
> to get the size, but can we get the size before that call, or even if the
> call fails (if it fails, would the size be 0, invalid, or what we tried to
> get but couldn't?)

Also, I'm not sure, I've never seen this call fail but I'd imagine it's either uninitialized memory or 0.

Comment hidden (mozreview-request)

Review commit: https://reviewboard.mozilla.org/r/77050/diff/#index_header
See other reviews: https://reviewboard.mozilla.org/r/77050/

Milan Sreckovic [:milan] (needinfo for best results)

Comment 28

•

8 years ago

I get a driver reset caught at a different place in the code, when I run this test locally, when swap chain Present call returns E_OUTOFMEMORY (which is fair, we are trying to allocate a gig of video memory for the window alone.)

(In reply to Henrik Skupin (:whimboo) from comment #23)
> ...
> Milan, is the test correct in using such a high resolution? Should it be
> reduced to something like twice the current screen size? Or should we
> correctly handle that resolution change in Firefox?

When we force 100kx100k or 16kx16k, all we're really testing is Firefox against driver resets - 'cause a driver reset is probably going to happen :)  I would do something like you suggest - do a 4x on the screen size, for example (so, your windows is now 16 times larger than the screen), and see if this clears up.

Flags: needinfo?(edwin)

Milan Sreckovic [:milan] (needinfo for best results)

Updated

•

8 years ago

Attachment #8788910 - Attachment is obsolete: true

Attachment #8788910 - Flags: review?(bas)

Comment hidden (mozreview-request)

Comment on attachment 8788910 [details]
Bug 1285085: Testing for a windows 4x larger than the screen in each dimension should be enough for this test.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/77050/diff/1-2/

Henrik Skupin [:whimboo][⌚️UTC+1]

Comment 30

•

8 years ago

mozreview-review

Comment on attachment 8788910 [details]
Bug 1285085: Testing for a windows 4x larger than the screen in each dimension should be enough for this test.

https://reviewboard.mozilla.org/r/77050/#review75534

Please note that I cannot give review for the graphics changes. My r+ is only for the modified test with the nit fixed.

::: testing/marionette/harness/marionette/tests/unit/test_set_window_size.py:58
(Diff revision 2)
>                           "New width is %s but should be %s" % (size['width'], width))
>          self.assertEqual(size['height'], height,
>                           "New height is %s but should be %s" % (size['height'], height))
>  
>      def test_possible_to_request_window_larger_than_screen(self):
> -        self.marionette.set_window_size(100000, 100000)
> +        self.marionette.set_window_size(4*self.max_width, 4*self.max_height)

nit: please add spaces between the operands and operator in both cases.

Attachment #8788910 - Flags: review?(hskupin) → review+

Comment hidden (mozreview-request)

Comment on attachment 8788910 [details]
Bug 1285085: Testing for a windows 4x larger than the screen in each dimension should be enough for this test.

Review request updated; see interdiff: https://reviewboard.mozilla.org/r/77050/diff/2-3/

Milan Sreckovic [:milan] (needinfo for best results)

Comment 32

•

8 years ago

(In reply to Henrik Skupin (:whimboo) from comment #30)
> Comment on attachment 8788910 [details]
> Bug 1285085: Testing for a windows 4x larger than the screen in each
> dimension should be enough for this test.
> 
> https://reviewboard.mozilla.org/r/77050/#review75534
> 
> Please note that I cannot give review for the graphics changes. My r+ is
> only for the modified test with the nit fixed.

Yes, thanks - I don't think we need the graphics patch for now.

Pulsebot

Comment 33

•

8 years ago

Pushed by msreckovic@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/5d39f44cafe1
Testing for a windows 4x larger than the screen in each dimension should be enough for this test. r=whimboo

Phil Ringnalda (:philor)

Comment 34

•

8 years ago

bugherder

https://hg.mozilla.org/mozilla-central/rev/5d39f44cafe1

Status: NEW → RESOLVED

Closed: 8 years ago

status-firefox51: affected → fixed

Resolution: --- → FIXED

Target Milestone: --- → mozilla51

Sylvestre Ledru [:Sylvestre]

Comment 35

•

8 years ago

This is something we want to uplift to 50?

status-firefox49: affected → wontfix

Ryan VanderMeulen [:RyanVM]

Comment 36

•

8 years ago

bugherder uplift

https://hg.mozilla.org/releases/mozilla-aurora/rev/b1a01538f4be

status-firefox50: affected → fixed

Flags: in-testsuite+