After analyzing two crash dumps created on different dates, the error DXGI_ERROR_DEVICE_HUNG (0x887a0006) consistently points to the same address in the game's executable: bf6.exe+0x4aae650.

Both dumps contain the line "COMMENT: Frostbite MiniDump", meaning they were generated by the Frostbite engine itself — not by Windows or the GPU driver. This confirms that the crash originates from within the game code......

Technical summary (WinDbg analysis)

• Exception code: 0x887a0006 = DXGI_ERROR_DEVICE_HUNG
• Crash origin: The exception was raised via KERNELBASE!RaiseException from bf6.exe+0x4aae650

No GPU drivers in stack: dxgkrnl.sys, atikmdag.sys, nvidia-kernel.sys are absent → hardware failure excluded

Comparison of two dumps (months apart, different driver versions)

Element | Value | Conclusion
Exception Code | 0x887a0006 | Same error
Where called | KERNELBASE!RaiseException+0x6c | Game initiated the crash
Where from | bf6+0x4aae650 | Identical address
Register r15 | 0x887a0006 | Error code passed as parameter
Dump Comment | Frostbite MiniDump | Dump from game engine
Process | bf6.exe | Confirmed

Request to developers

This issue is stable, reproducible across different driver versions, and points to a specific address in the game code: bf6.exe+0x4aae650.

I kindly ask the development team to investigate:

• GPU timeout handling logic in the Frostbite engine
• The module located at the specified address

If additional information or testing is needed, I'm happy to help.

Thank you.

UPDATE 

This time, I tried to disable the FSR, and here's what happened

I have captured another crash dump, and this time the stack trace provides even stronger evidence that the issue is caused by the Frostbite engine's interaction with AMD FidelityFX Super Resolution 2 (FSR2) code — regardless of in-game settings. 

• Exception code: 0x887a0006 = DXGI_ERROR_DEVICE_HUNG
• Dump comment: Frostbite MiniDump — generated by the game engine itself
• Crash origin: The exception was raised via KERNELBASE!RaiseException

 Most important: the stack trace now directly shows FSR2 functions

STACK_TEXT: ... bf6!ffxFsr2ResourceIsNull+0x22c2a20 ... bf6!ffxFsr2ResourceIsNull+0x22c1749 ... bf6!ffxFsr2ResourceIsNull+0x23845b1 ... Engine_Render_Core2_PlatformPcDx12_retail!rc2DeviceInit+0x23a06 ... Engine_Render_Core2_PlatformPcDx12_retail!rc2DeviceInit+0x1423 ... Engine_Render_Core2_PlatformPcDx12_retail!rc2GetSystemServiceProvider+0x584 ... bf6!ffxFsr2ResourceIsNull+0x22c0ab1 ... (more FSR2 calls)

This confirms:

•  FSR2 functions are called even if FSR is disabled in the game menu
•  The crash happens inside FSR2-related code

The same FAILURE_ID_HASH {823c07cc-a67a-e92f-3480-f4d74ce569af} appears again — proving this is the same underlying bug.

UPD 2

After the Javelin update, this COM permission error occurs before the system completely freezes.

CLSID {2593F8B9-4EAF-457C-B68A-50F6B8EA6B54} 

APPID {15C20B67-12E7-4BB6-92BB-7AFF07997402}.

This error is logged right before the freeze.  The hang does not occur with 100% probability, but about 2 out of 3 errors were accompanied by a hang. Anti-cheat is my only clue as it started after a recent update.