View Issue Details

IDProjectCategoryView StatusLast Update
0000286madVRbugpublic2015-04-29 23:19
Reporterkasper93 Assigned Tomadshi  
PrioritynormalSeverityminorReproducibilityalways
Status closedResolutionfixed 
Summary0000286: Access violation reading location exception when initializing madVR x64 instance
DescriptionHi,

Here is stack trace https://gist.github.com/kasper93/f891086bbd726056c510 looks like somewhere in monitor enumeration code? You should be able to reproduce, but if not you could provide symbol file and I can provide better stack trace.

Exception is thrown every time (I see it in debugger), but most of the time it doesn't terminate process. Unfortunately sometimes it does. (I couldn't find pattern here)

FWIW, I also see other exceptions under debugger, for example during connection. I can provide more information, but those doesn't seem to be causing any issues in normal usage, so probably are handled internally.

Note that on x86 build I don't see any exceptions.
TagsNo tags attached.
madVR Version0.87.21
Media Player (with version info)MPC-HC 1.7.8.162
Splitter (with version info)Internal LAV
Decoder (with version info)Internal LAV
DecodingSoftware
Deinterlacingnone (progressive)
DXVA2 Scaling Activeno
Aero / Desktop CompositionOn
Problem occurs with modeall modes
GPU ManufacturerAMD
GPU ModelHD5870
GPU Driver Versionnot relevant

Activities

madshi

2015-04-29 17:21

administrator   ~0000971

Interesting. Unfortunately the mvrSettings64.dll is compiled with Delphi, so I can't provide you with PDB symbols for that. However, the DLL should come have debug infos linked in which work with madVR's own freeze detection. In the moment when that crash is active (outside of a debugger), do you get a Windows crash box? If so, can you try to press Ctrl+Alt+Shift+Pause/Break in that moment? With a bit of luck you might then get a freeze report on your desktop which might contain a full callstack of the crashing thread. Could also be that the freeze report feature doesn't work in that situation, though...

Those other crashes, do they also appear to be in mvrSettings64.dll or in madHcNet64.dll? Or are they in madVR64.ax?

kasper93

2015-04-29 22:23

reporter   ~0000972

Last edited: 2015-04-29 22:24

> do you get a Windows crash box?
Yes.

> If so, can you try to press Ctrl+Alt+Shift+Pause/Break in that moment?
No luck here, nothing is generated.

> Those other crashes, do they also appear to be in mvrSettings64.dll or in madHcNet64.dll? Or are they in madVR64.ax?

I made a list of exceptions during opening file. It seems to be the same problem everywhere with LocalFree(). And yes only in mvrSettings64.dll or madHcNet64.dll. Although there is one unrelated in madVR64.ax.

First-chance exception at 0x00007FF82455517C (ntdll.dll) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0x000000000F9BF5FF.
https://gist.github.com/kasper93/0dda58b57fefc64c8a73

First-chance exception at 0x00007FF82455517C (ntdll.dll) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0x000000000F9BF69F.
https://gist.github.com/kasper93/3f5cf6f19a406c667ac4

First-chance exception at 0x000000004A4C4ACA (madVR64.ax) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0xFFFFFFFFFFFFFFFF.
https://gist.github.com/kasper93/5b740b8c92b37b3cd218

First-chance exception at 0x00007FF82455517C (ntdll.dll) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0x0000000012D2C64F.
https://gist.github.com/kasper93/c41cc627d07cb336cf8f

First-chance exception at 0x00007FF82455517C (ntdll.dll) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0x000000000FA74C3F.
https://gist.github.com/kasper93/2dc5ad8fbdc4f71f8f5d

First-chance exception at 0x00007FF82455517C (ntdll.dll) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0x0000000012D2826F.
https://gist.github.com/kasper93/d7eba1d25d3aa9adabaf

First-chance exception at 0x00007FF82455517C (ntdll.dll) in mpc-hc64.exe: 0xC0000005: Access violation reading location 0x0000000012D2785F.
https://gist.github.com/kasper93/ec33e2d99da600507b2a

madshi

2015-04-29 23:19

administrator   ~0000973

Thanks. I've found some things I've not properly converted to 64bit. The crashes should all be gone in the next build.

Issue History

Date Modified Username Field Change
2015-04-29 17:16 kasper93 New Issue
2015-04-29 17:21 madshi Note Added: 0000971
2015-04-29 17:21 madshi Assigned To => madshi
2015-04-29 17:21 madshi Status new => feedback
2015-04-29 22:23 kasper93 Note Added: 0000972
2015-04-29 22:23 kasper93 Status feedback => assigned
2015-04-29 22:24 kasper93 Note Edited: 0000972
2015-04-29 23:19 madshi Note Added: 0000973
2015-04-29 23:19 madshi Status assigned => closed
2015-04-29 23:19 madshi Resolution open => fixed