Description of problem: cut and paste works in a dizzy way (It doesn't work at all from guest to host, sometimes it works in reverse) Version-Release number of selected component (if applicable): fedora 26 spice-vdagent-0.17.0-2.fc26.x86_64 How reproducible: install fedora 26 and virt-manager Steps to Reproduce: 1.just install and use it. 3. Actual results: cut and paste between host and guests is a mess /dev/virtio-ports/com.redhat.spice.0 in never created Expected results: cut and paste works like a charm Additional info: i get this in journalctl spice-vdagent[2107]: Cannot access vdagent virtio channel /dev/virtio-ports/com.redhat.spice.0
- Not working from client to guest is the only issue? - When you say 'cut and paste' is it a string that you are selecting with your mouse in some application or a file? - If you are running on Wayland can you try using an X11 session to see if it works?
I can't cut and paste from the host to the clients. I work a lot with different guests, and I need to cut and paste commands or scripts from the host (my workstation) to the guests (which are spot vms I use and destroy). I was able to do it with fedora 24 but I have never been able to do it with fedora 26. I can cut from a guest and cut to the host, but I need to do the contrary. Yes, I am using Wayland and will try to switch back to X. if I check the service with the "verygoodinitdreplacement" systemd I get: systemctl status spice-vdagentd.service ● spice-vdagentd.service - Agent daemon for Spice guests Loaded: loaded (/usr/lib/systemd/system/spice-vdagentd.service; enabled; vendor preset: enabled) Active: inactive (dead) which tells me there is a problem (the service started but there was an issue). checking the boot sequence I get: Sep 09 15:15:29 <hoatname> spice-vdagent[1880]: Cannot access vdagent virtio channel /dev/virtio-ports/com.redhat.spice.0 and if i check if the device file does exist, I can't even find the directory virtio-ports in /dev.
/dev/virtio-ports/com.redhat.spice.0 not existing would be your first issue. Does that always happen? Only sometimes? However, when it's missing, I'd expect copy and paste to not be functional at all. Copy and paste working from guest to host, but not the other way round is a totally different issue, and is related to wayland.
update: using Xorg instead of Wayland fixes this issue (and I have to say that the mouse movements are smoother than before). anyway /dev/virtio-ports/com.redhat.spice.0 is NEVER created. I checked on both my workstations. So I guess we have an issue here, which does not affect the cut and paste stuff anyway. Switching back to xorg seems a workaround, nut i guess both problems should be fixed
(In reply to Davide Corrado from comment #4) > update: using Xorg instead of Wayland fixes this issue (and I have to say > that the mouse movements are smoother than before). > > anyway /dev/virtio-ports/com.redhat.spice.0 is NEVER created. I checked on > both my workstations. So I guess we have an issue here, which does not > affect the cut and paste stuff anyway. > Switching back to xorg seems a workaround, nut i guess both problems should > be fixed Just to be sure, you are checking this device in the guest VM right, not in the host/client. It would not be possible to have things working without this device.
nope, I am able to cut and paste between the host and the guests. I am doing this right now. have a look below: journalctl -b|grep virtio Hint: You are currently not seeing messages from other users and the system. Users in groups 'adm', 'systemd-journal', 'wheel' can see all messages. Pass -q to turn off this notice. Oct 06 08:13:43 niceleprotto spice-vdagent[2059]: Cannot access vdagent virtio channel /dev/virtio-ports/com.redhat.spice.0 and ls -al /dev/v* crw-rw----. 1 root tty 7, 0 Oct 6 08:13 /dev/vcs crw-rw----. 1 root tty 7, 1 Oct 6 08:13 /dev/vcs1 crw-rw----. 1 root tty 7, 2 Oct 6 08:13 /dev/vcs2 crw-rw----. 1 root tty 7, 3 Oct 6 08:13 /dev/vcs3 crw-rw----. 1 root tty 7, 4 Oct 6 08:13 /dev/vcs4 crw-rw----. 1 root tty 7, 5 Oct 6 08:13 /dev/vcs5 crw-rw----. 1 root tty 7, 6 Oct 6 08:13 /dev/vcs6 crw-rw----. 1 root tty 7, 128 Oct 6 08:13 /dev/vcsa crw-rw----. 1 root tty 7, 129 Oct 6 08:13 /dev/vcsa1 crw-rw----. 1 root tty 7, 130 Oct 6 08:13 /dev/vcsa2 crw-rw----. 1 root tty 7, 131 Oct 6 08:13 /dev/vcsa3 crw-rw----. 1 root tty 7, 132 Oct 6 08:13 /dev/vcsa4 crw-rw----. 1 root tty 7, 133 Oct 6 08:13 /dev/vcsa5 crw-rw----. 1 root tty 7, 134 Oct 6 08:13 /dev/vcsa6 crw-------. 1 root root 10, 63 Oct 6 08:13 /dev/vga_arbiter crw-------. 1 root root 10, 137 Oct 6 08:13 /dev/vhci crw-------. 1 root root 10, 238 Oct 6 08:13 /dev/vhost-net /dev/vfio: total 0 drwxr-xr-x. 2 root root 60 Oct 6 08:13 . drwxr-xr-x. 22 root root 4220 Oct 6 08:54 .. crw-------. 1 root root 10, 196 Oct 6 08:13 vfio if you dont believe me, I can record a video in which I can cut and paste. just ask
(In reply to Davide Corrado from comment #4) > update: using Xorg instead of Wayland fixes this issue (and I have to say > that the mouse movements are smoother than before). the wayland side is a bug upstream. (In reply to Davide Corrado from comment #6) > if you dont believe me, I can record a video in which I can cut and paste. > just ask My question in comment #5 was if you are checking the spice-vdagent in the VM or in the HOST. There is no need to run spice-vdagent in the HOST and in the HOST we don't have the virtio-port which would explain your concern.
sorry about that, I misunderstood then. aswering tour question, yes, I am running vdagent in the guest. spice-vdagent starts by default (fedora 26), I didn't touch it. SWo i guess I can disable the service than. So I guess my issue should be moved to wayland than. another issue I have is that if a cntr-tab in full screen in a guest, the key conbination is intercepted by wayland (I see processes running in the host). thanks for your explanation.
and I guess that vdagentd is enabled by default in case fedora runs as a guest :-)
This can be a real bug maybe selinux : 1. Can the distro get provider CISCO access like a "virtual machine" ( ... SYSLOG_PID 1672 with _GID 1000 ) into area using a device on BNC (Bayonet Neill-Concelman) 2. The logs from server Fedora also show selinux issue: unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 See also this infos bug : https://bugzilla.redhat.com/show_bug.cgi?id=1405750