Re: [Qemu-devel] [Spice-devel] Vioserial of Windows guest OS on Qemu 0.15

[Stefan Hajnoczi <stefanha@gmail.com>]

On Mon, Jan 16, 2012 at 10:02:22AM +0800, Charles.Tsai-蔡清海-研究發展部 wrote:
> Vadim,
> 
> Thank you for your prompt reply. Here are the information for our test case.
> 
> 
> 1) we use the following command line to launch the guest OS
> 
> 
> /usr/bin/kvm -S -M pc-0.14 -enable-kvm -m 1024 -smp 1,sockets=1,cores=1,threads=1 -name win_xp -uuid d9388815-ddd3-c38e-33c2-a9d5fcc7a775 -nodefconfig -nodefaults -chardev socket,id=charmonitor,path=/var/lib/libvirt/qemu/win_xp.monitor,server,nowait
> -mon chardev=charmonitor,id=monitor,mode=readline
> -rtc base=localtime
> -device virtio-serial-pci,id=virtio-serial0,bus=pci.0,multifunction=on,addr=0x5.0x0
> -drive file=/media/Images/Windows-XP.img,if=none,id=drive-ide0-0-0,format=raw
> -device ide-drive,bus=ide.0,unit=0,drive=drive-ide0-0-0,id=ide0-0-0,bootindex=1
> -netdev tap,fd=17,id=hostnet0
> -device rtl8139,netdev=hostnet0,id=net0,mac=52:54:00:e8:dc:b1,bus=pci.0,multifunction=on,addr=0x3.0x0
> -chardev pty,id=charserial0
> -device isa-serial,chardev=charserial0,id=serial0
> -chardev spicevmc,id=charchannel0,name=vdagent
> -device virtserialport,bus=virtio-serial0.0,nr=1,chardev=charchannel0,id=channel0,name=com.redhat.spice.0

I tried the virtio-serial pieces with qemu.git and a RHEL6 guest.  The
guest was able to see the "com.redhat.spice.0" port.  This suggests the
problem was either fixed since QEMU 0.15 or that it only affects the
Windows virtio drivers.

> -usb -device usb-tablet,id=input0
> -spice port=5900,addr=0.0.0.0,disable-ticketing
> -vga qxl -global qxl-vga.vram_size=67108864 -device virtio-balloon-pci,id=balloon0,bus=pci.0,multifunction=on,addr=0x4.0x0
> 
> 
> 
> 2). In Guest Windows XP OS
> 
> 
> When the following callback function of the vioserial device  is called in guest OS. The allocated resources is empty.
> 
> 
> VIOSerialEvtDevicePrepareHardware() ---This function is to get the I/O address of the vioserial device and map the physical address to the logical address space.
> 
> I added the following trace and the value of nListSize is ZERO.
> TraceEvents(TRACE_LEVEL_INFORMATION, DBG_PNP, "%s (nListSize=%d)\n", __FUNCTION__,nListSize); 
> 
> 
> So far, we have tested Qemu 0.14 without any problem but Qemu 0.15 seemed to be broken in vioserial device.
> Let me know if you need further information. Thanks.

I suggest trying qemu-kvm 1.0 or qemu-kvm.git because the issue may have
been fixed since 0.15.

If the issue is still present, you could use git-bisect(1) to find the
commit that caused behavior to change between 0.14 and 0.15.

Stefan