Audio subsystem versus unplugging uaudio

Hi, Testing ohci recently, I came across a problem where our audio stack doesn't handle removing an audio device whilst it's open. I connected a uaudio device (dmesg below), ran mplayer: :; mplayer -ao oss:device=/dev/audio1 track.ogg and removed the uaudio device in the middle of the track.

Re: Automated report: NetBSD-current/i386 test failure

Yesterday, the NetBSD Test Fixture wrote: > The newly failing test case is: > > rump/rumpkern/t_vm:busypage This one is still failing. The rump kernel panics with: [ 1.1400050] panic: kernel diagnostic assertion "(pg->flags & PG_FAKE) == 0" failed: file "/tmp/build/2020.12.07.10.02.51