Open Bug 1746932 Opened 3 years ago Updated 1 year ago

Crash in [@ MitLibTriggerFailFast]

Categories

(Core :: Security: Process Sandboxing, defect, P3)

Unspecified
Windows 10
defect

Tracking

()

People

(Reporter: gsvelto, Unassigned)

References

Details

(Keywords: crash)

Crash Data

Crash report: https://crash-stats.mozilla.org/report/index/1eecd47b-f240-472f-8438-dfa010211220

Reason: EXCEPTION_STACK_BUFFER_OVERRUN / FAST_FAIL_PAYLOAD_RESTRICTION_VIOLATION

Top 10 frames of crashing thread:

0 payloadrestrictions.dll MitLibTriggerFailFast 
1 payloadrestrictions.dll MitLibValidateAccessToProtectedPage 
2 ntdll.dll RtlpCallVectoredHandlers 
3 ntdll.dll RtlDispatchException 
4 ntdll.dll KiUserExceptionDispatch 
5 firefox.exe base::win::PEImage::VerifyMagic const security/sandbox/chromium/base/win/pe_image.cc:568
6 firefox.exe ResolveNTFunctionPtr security/sandbox/chromium/sandbox/win/src/win_utils.cc:609
7 firefox.exe sandbox::GetProcessBaseAddress security/sandbox/chromium/sandbox/win/src/win_utils.cc:524
8 firefox.exe sandbox::TargetProcess::Create security/sandbox/chromium/sandbox/win/src/target_process.cc:223
9 firefox.exe sandbox::BrokerServicesBase::SpawnTarget security/sandbox/chromium/sandbox/win/src/broker_services.cc:620

Really odd crashes, I'm not sure what's going on here. These are all caught by WER so we might have had them for a while without having been able to intercept them in the past.

This looks like a crash caused by EAF (= Export Address Filtering). Payloadrestrictions.dll is a system module and is injected when the process enables additional exploit protections.

See Also: → 1509748
Severity: S2 → S4
Priority: -- → P3

I ran 91.4.1esr on Windows 10 with EAF+. Payloadrestrictions.dll was loaded but everything worked fine without crash. Our statement https://support.mozilla.org/en-US/kb/compatibility-exploit-protection-windows-10 is still valid. We recommend not enabling EAF for firefox.exe.

We seem to be catching a lot of volume on this crash signature in ESR. I don't know why Bugzilla only shows 115.11.0esr, but crash-stats suggests that the peak in volume actually started with 115.9.1esr:

1 	115.9.1esr 	5141 	31.39 %
2 	115.10.0esr 	4572 	27.91 %
3 	115.11.0esr 	3640 	22.22 %
4 	115.9.0esr 	655 	4.00 %

This is probably preexisting volume that we are now catching (again?) thanks to the WER patches from bug 1872920.

You need to log in before you can comment on or make changes to this bug.