Open Bug 2000475 Opened 11 months ago Updated 6 days ago

Crash in [@ RaiseFailFastException | CFlat::Abandonment::FailWithException]

Categories

(Core :: Widget: Win32, defect)

Other
Windows 11
defect

Tracking

()

Tracking Status
firefox-esr115 --- unaffected
firefox-esr140 --- unaffected
firefox145 --- wontfix
firefox146 --- wontfix
firefox147 --- fix-optional
firefox148 --- fix-optional

People

(Reporter: release-mgmt-account-bot, Unassigned)

References

(Blocks 1 open bug, Regression)

Details

(Keywords: crash, regression)

Crash Data

Crash report: https://crash-stats.mozilla.org/report/index/e6b43fac-c0f2-40e4-a0db-335a30251026

Reason: unknown 0xe0464645

Top 10 frames of crashing thread:

0  KERNELBASE.dll  RaiseFailFastException  
1  CoreMessaging.dll  CFlat::Abandonment::FailWithException  
2  CoreMessaging.dll  CFlat::Abandonment::Fail  
3  CoreMessaging.dll  CFlat::Abandonment::UseOfInvalidatedExportAdapter  
4  CoreMessaging.dll  Microsoft::CoreUI::Messaging::MessagingValidationException::FromResult  
5  CoreMessaging.dll  Microsoft::CoreUI::Dispatch::WaitCollection::UnregisterWait  
6  CoreMessaging.dll  Microsoft::CoreUI::Dispatch::WaitController::DisableEventLoopWait  
7  CoreMessaging.dll  Microsoft::CoreUI::Dispatch::Win32EventLoopBridge::Reset  
8  CoreMessaging.dll  Microsoft::CoreUI::Dispatch::EventLoop::Reset  
9  CoreMessaging.dll  Microsoft::CoreUI::Messaging::MessageSession::DoNonblockingCleanupOnFinalRelease  

By querying Nightly crashes reported within the last 2 months, here are some insights about the signature:

  • First crash report: 2025-10-18
  • Process type: Multiple distinct types
  • Is startup crash: No
  • Has user comments: No
  • Is null crash: No

The first Gecko frame way up the stack is mozilla::widget::IMEContext::AssociateDefaultContext().

Component: General → Widget: Win32

I have no idea only from the stack because we tried to enable IME on the window when an editable element in a remote process gets focus and TSF in Windows crashes during its initialization.

The crash occurred during Microsoft::CoreUI::Messaging::MessageSession::OnFinalRelease(bool, CFlat::Ref<System::IntPtr>). According to its name, sounds like that double free or something might have happened. If so, we can do nothing.

Component: Widget: Win32 → DOM: UI Events & Focus Handling
Keywords: inputmethod

Well, the first Gecko frame in the other crash reports of the same signature are vary. It seems that this is a generic issue of Windows.

As far as I checked, the most first frame is WindowsUIUtils::SetIsTitlebarCollapsed. I think the crashes occurred at various API call immediately after startup.

Component: DOM: UI Events & Focus Handling → Widget: Win32
Keywords: inputmethod

Yeah, most of these go through WindowsUIUtils::SetIsTitlebarCollapsed(), specifically the call to put_ExtendsContentIntoTitleBar().

Severity: -- → S3
Keywords: regression
Regressed by: 1947324

Set release status flags based on info from the regressing bug 1947324

Set release status flags based on info from the regressing bug 1947324

https://crash-stats.mozilla.org/report/index/85253364-f46a-4d90-932a-c63e90251209

This is another scenario. This crashes when we make a window for the fullscreen mode. The crash occurred during this call, ::SetWindowLongPtrW(aWnd, GWL_STYLE, aStyles.style). Then, nsWindow gets received a message for the window's style change and crash during a call of previous wndproc. Before that, it calls nsWindow::ProcessMessage so that anything could have already occurred such as destruction of the window. So, I wonder if the last call is safe.

My sense is that the signature is generic, as comment 3 suggests. Looks to be related to exception reporting, but maybe we can narrow it down.

I think the Gecko signature is the right frame to search for here. For example, searching further up the stack, the Win32 function TextInputHost::Initialize includes stuff that isn't related. Crash frequency suggests this hasn't changed in the 6 months we have data for.

The emergence of the (unrelated) signature in comment 7 is probably related to the WinApp SDK work that went in in early October. I've filed bug 2005322 for that.

See Also: → 2011618

https://crash-stats.mozilla.org/report/index/315bead2-8518-4b58-9329-00d270260717
https://crash-stats.mozilla.org/report/index/1db658ac-8e07-4fd7-8fb1-5a1b50260717
https://crash-stats.mozilla.org/report/index/87c51b72-9303-400c-9143-351880260717
https://crash-stats.mozilla.org/report/index/e0a24c1e-564b-429b-8c58-cf5c10260717
https://crash-stats.mozilla.org/report/index/d2422098-b0ba-4d38-b5a1-d35b00260709

Started occuring for me after moving my windows install from a win11/cascade lake machine to a win11/zen5 epyc machine about 3 weeks ago. Went from never happening to around once every 2 days. Seems like its the same issue every time, may be hardware dependent? The intel machine has had either a GTX 1080 or an Arc B570 in the past 2 months and the AMD machine has an RTX 3090. Not sure if the hardware info may be useful.

https://crash-stats.mozilla.org/report/index/7342c2e6-2505-430f-a3f6-263de0260821

154.0 still affected. Opening a new private window reliably triggers the issue for me on the zen5 machine listed above.

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