1 files changed, 84 insertions, 0 deletions
diff --git a/results/classifier/118/permissions/1772262 b/results/classifier/118/permissions/1772262
new file mode 100644
index 000000000..28353290d
--- /dev/null
+++ b/results/classifier/118/permissions/1772262
@@ -0,0 +1,84 @@
+permissions: 0.881
+virtual: 0.852
+assembly: 0.849
+register: 0.847
+device: 0.843
+user-level: 0.835
+hypervisor: 0.834
+semantic: 0.830
+graphic: 0.825
+network: 0.816
+PID: 0.816
+architecture: 0.814
+performance: 0.807
+debug: 0.805
+boot: 0.804
+files: 0.803
+KVM: 0.795
+arm: 0.792
+ppc: 0.790
+TCG: 0.786
+vnc: 0.784
+mistranslation: 0.781
+kernel: 0.780
+VMM: 0.779
+peripherals: 0.774
+socket: 0.762
+risc-v: 0.749
+x86: 0.726
+i386: 0.626
+
+Adding -spice doesn't respect environment variable QEMU_AUDIO_DRV
+
+When -spice is added to the commandline, QEMU_AUDIO_DRV=alsa is not respected and I receive no audio from the guest using the alsa driver.  When -spice options are omitted, audio works as usual.
+
+I want to channel mouse and keyboard events to the guest using spice (with looking-glass) instead of a third-party product over a network interface (with synergy), but audio ends up over the spice protocol as well instead of using alsa despite asking otherwise.  For example, one will only hear audio when a program like spicy is running.  Naturally it would be nice if this were not needed, since spice audio is gappier than regular audio on my machine.
+
+Perhaps if -spice added an option similar to agent-mouse=off but for audio, (say audio-pipe=off or so for example,) this might mitigate the issue in a more naive fashion UX-wise, without having to force the issue with environment variables that unfortunately for me don't seem to work. :(
+
+QEMU emulator version 2.12.0
+
+Commandline
+
+        qemu-system-x86_64 \
+          -ctrl-grab \
+          -enable-kvm \
+          -cpu host,hv-time,kvm=off \
+          -smp cores=4 \
+          -m 8G \
+          -M q35 \
+          -vga none \
+          -netdev tap,id=hostnet,ifname=tap1,script=no,downscript=no \
+          -net nic,model=virtio,macaddr=52:54:FD:BF:F7:9A,netdev=hostnet \
+          -netdev user,id=usernet,smb=/media/DRIVE-C/tux/vms/share \
+          -net nic,model=virtio,macaddr=52:54:2E:40:4F:C8,netdev=usernet \
+          -usb \
+          -device usb-ehci,id=ehci \
+          -device ioh3420,bus=pcie.0,addr=1c,multifunction=on,port=1,chassis=1,id=root.1 \
+          -device vfio-pci,host=07:00.0,bus=root.1,addr=00.0,multifunction=on,x-vga=on \
+          -device ich9-intel-hda \
+          -device hda-micro \
+          -drive if=virtio,file=vm/win10/disk.img,media=disk \
+          -boot menu=on,splash=splash/boot.jpg,splash-time=5000 \
+          -name win10 \
+          -device ivshmem-plain,memdev=ivshmem \
+          -object memory-backend-file,id=ivshmem,share=on,mem-path=/dev/shm/looking-glass,size=32M \
+          -device virtio-serial-pci \
+          -device virtserialport,chardev=spicechannel0,name=com.redhat.spice.0 \
+          -chardev spicevmc,id=spicechannel0,name=vdagent \
+          -spice addr=127.0.0.1,port=5900,disable-ticketing
+
+When the last four options are not present, audio works as expected.  This is with both a windows 10 guest and a windows 7 guest.
+
+Doesn't reproduce.  The *default* driver is changed to spice when spice is enabled, but overriding using QEMU_AUDIO_DRV is not disabled.  Any chance you didn't export QEMU_AUDIO_DRV?
+
+>Any chance you didn't export QEMU_AUDIO_DRV?
+
+Okay, I just checked by calling env in the same fashion.
+
+Well lash me bootlaces, it turns out that sudo needs to be invoked with the -E flag to preserve the environment.  It's funny, the examples I read up on never mentioned this.
+
+Sorry for littering the bug tracker with a silly end-user problem.
+
+Cheers.
+