qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[Qemu-devel] After patch to fix non repeating keys, keys get stuck on Wi


From: Leonardo Soares Müller
Subject: [Qemu-devel] After patch to fix non repeating keys, keys get stuck on Windows guests
Date: Mon, 4 Mar 2019 15:39:53 +0000

Using QEMU version 3.1.50 (v3.1.0-2338-g1ba530a4ec-dirty), after the
commit 35921860156e39f17ffd7e18d0f84d2396a6e8f4
kbd-state: don't block auto-repeat events

It was noticed that keys can get stuck on an Intel GVT-g (iGVT-g)
Windows 10 guest. For example, pressing "e" will result on
"eeeeeeeeeeeeeeeeeeeeee" until the key "e" is pressed again. I tried to
reproduce this on simple Linux guests and simple Windows 10 guests but
failed to do so.

The QEMU command line is pretty long:

sudo env QEMU_AUDIO_DRV=pa QEMU_AUDIO_DAC_FIXED_FREQ=96000
QEMU_AUDIO_ADC_FIXED_FREQ=96000
QEMU_PA_SERVER=/run/user/1000/pulse/native QEMU_AUDIO_ADC_VOICES=0
PULSE_LATENCY_MSEC=10 qemu-system-x86_64 -name 'Windows 10' -k pt-br
-nodefaults -device nec-usb-xhci,id=necxhci -hda
/home/usuario/.local/share/libvirt/images/redm.qcow2 -enable-kvm -cpu
Skylake-Client,hv_time,hv_relaxed,hv_vapic,hv_spinlocks=0x1fff,hv_runtime
-smp cores=2,threads=2 -m 3072M -bios /usr/local/share/qemu/bios.bin
-device usb-tablet,id=tablet -device usb-audio,id=usbaudvir,buffer=6144
-device usb-host,vendorid=0x0079,id=redragon -device
usb-host,vendorid=0x0458,id=mousegenius -vga none -monitor vc -serial
stdio -display gtk,gl=on -device
vfio-pci,sysfsdev=/sys/bus/pci/devices/0000:00:02.0/123f09b0-4c00-11e8-a6ca-f3c21e47e012,x-igd-opregion=on,rombar=0,display=on,addr=0x3,id=iHD520
-machine kernel_irqchip=on -global PIIX4_PM.disable_s3=1 -global
PIIX4_PM.disable_s4=1 -M pc,usb=false -netdev user,id=net0 -device
e1000,netdev=net0,id=n0,addr=0x8 -device
usb-host,vendorid=0x0480,productid=0x0200,id=toshiba

"redragon" is a gamepad, "toshiba" an external HDD and "mousegenius" is
a mouse. The problem was noticed both with devices connected and
disconnected, and when they were removed using device_del.

While I faced this on multiple situations (even using notepad, for
example), this bug is easy to trigger with games. It seems the usb-audio
emulation plus iGVT-g load plus high CPU load makes the key release
event never to be sent to the guest, making the key get stuck on the
guest until it is pressed again.

reply via email to

[Prev in Thread] Current Thread [Next in Thread]