1 files changed, 46 insertions, 0 deletions
diff --git a/results/classifier/118/graphic/2073 b/results/classifier/118/graphic/2073
new file mode 100644
index 000000000..c71b0f7ab
--- /dev/null
+++ b/results/classifier/118/graphic/2073
@@ -0,0 +1,46 @@
+graphic: 0.885
+permissions: 0.821
+device: 0.809
+semantic: 0.544
+register: 0.522
+performance: 0.459
+kernel: 0.423
+mistranslation: 0.423
+PID: 0.393
+ppc: 0.387
+vnc: 0.381
+arm: 0.357
+VMM: 0.335
+user-level: 0.329
+boot: 0.323
+x86: 0.299
+i386: 0.296
+architecture: 0.277
+socket: 0.268
+KVM: 0.266
+risc-v: 0.266
+debug: 0.246
+network: 0.237
+TCG: 0.188
+virtual: 0.188
+peripherals: 0.172
+assembly: 0.168
+hypervisor: 0.124
+files: 0.100
+
+Audio: missing ability to disable microphone input from host?
+Description of problem:
+**It appears there is no way to disable the microphone / input to the audio backend device(s).**
+
+
+There are at least two cases where this matters:
+1. The host has no microphone input (e.g. only HDMI audio output with video).
+2. The host has a microphone input, but the user doesn't want the guest VM to have access to the microphone/input.
+
+I tried the option in.channels=0, as that seemed the most obvious way, though that doesn't work.
+
+For -audio dsound, it appears that CLSID_DirectSoundCapture is unconditionally acquired.
+
+There will also be later periodic warning/text outputs from QEMU "Could not create a backend for voice virtio.in", if you're running on a host system with no audio input device.
+
+Adding a couple backend checks for channels > 0 may work well.  Not sure if it matters that audio front end device in the VM still thinks there is an audio input.