Created attachment 201355 [details]
The reason of this panic is the same as for bug #235081 : unstable smartphone USB cable connection.
The assertion is simply bogus. There is no coordination that ensures the check in g_resize_provider() is still valid in g_resize_provider_event(). I think this should probably just be a if () -> return case.
I expect the similar assertion in g_new_provider_event() is similarly bogus.
I encountered this issue on some test hardware with a virtual IPMI USB CDROM that did not attach correctly.