Bugzilla

So most of the crashes have

`XPCOMSpinEventLoopStack: AsyncShutdown Spinner for quit-application-granted`  (or more but always containing the "quit-application-granted" piece.

Now there is a misalignment between `ShutdownPhase::AppShutdownConfirmed` mapping to the "quit-application" notification and the `AsyncShutdown` exposing `quitApplicationGranted` which maps (not very surprisingly) to "quit-application-granted" instead.

Any blocker using `quitApplicationGranted` that might cause (directly or indirectly) a content process to begin its shutdown might race with our shutdown check based on `ShutdownPhase::AppShutdownConfirmed`.

I think the right course of action would probably be to align this better. It might actually be enough to [map this barrier](https://searchfox.org/mozilla-central/rev/be3db2aef8d3916c891527794a2d5e2f2dc9fab1/toolkit/components/asyncshutdown/AsyncShutdown.sys.mjs#1109)  to "quit-application", as there seems to be no mapping for that one, anyways. There might be a slight risk of mixing those with other things that are already happening during "quit-application", changing the relative order, but I think it is worth to check this for fallout.

So most of the crashes have

`XPCOMSpinEventLoopStack: AsyncShutdown Spinner for quit-application-granted`  (or more but always containing the "quit-application-granted" piece.)

Now there is a misalignment between `ShutdownPhase::AppShutdownConfirmed` mapping to the "quit-application" notification and the `AsyncShutdown` exposing `quitApplicationGranted` which maps (not very surprisingly) to "quit-application-granted" instead.

Any blocker using `quitApplicationGranted` that might cause (directly or indirectly) a content process to begin its shutdown might race with our shutdown check based on `ShutdownPhase::AppShutdownConfirmed`.

I think the right course of action would probably be to align this better. It might actually be enough to [map this barrier](https://searchfox.org/mozilla-central/rev/be3db2aef8d3916c891527794a2d5e2f2dc9fab1/toolkit/components/asyncshutdown/AsyncShutdown.sys.mjs#1109)  to "quit-application", as there seems to be no mapping for that one, anyways. There might be a slight risk of mixing those with other things that are already happening during "quit-application", changing the relative order, but I think it is worth to check this for fallout.