Red Hat Bugzilla – Bug 812347
High network jitter causes some keystrokes to enter multiple characters instead of one
Last modified: 2013-02-21 03:47:17 EST
Description of problem:
High network jitter causes some keystrokes to enter multiple characters instead of one.
I suspect that this is caused by sending keystrokes from client "as is" and evaluating what is pressed and how long on qemu side. If this is true, a network jitter can cause that packet carrying key press event is at lowest latency range and packet carrying key release event has highest latency, making qemu see key pressed much longer than user actually presses it.
If my assumptions are good, the solution could be adding timestamps to keyboard events on client and recreate keyboard events in spice server according to them. While this change increases network latency slightly overall, it should improve user productivity by removing errors caused by network latency/jitter that are quite time-consuming to correct (because you never know when your "Delete" key press is recorded as ten...).
Version-Release number of selected component (if applicable):
Steps to Reproduce:
ping statistics to the host running the VM where I encoutered the problem:
--- <host> ping statistics ---
100 packets transmitted, 98 received, 2% packet loss, time 60467ms
rtt min/avg/max/mdev = 347.170/400.243/515.824/37.753 ms
This request was not resolved in time for the current release.
Red Hat invites you to ask your support representative to
propose this request, if still desired, for consideration in
the next release of Red Hat Enterprise Linux.
This request was erroneously removed from consideration in Red Hat Enterprise Linux 6.4, which is currently under development. This request will be evaluated for inclusion in Red Hat Enterprise Linux 6.4.
I tried to tackle this problem a while ago.
The problem is that is not solvable with timestamps, since hw keyboard events are not usually assicated with time stamp, and key repeaters are usually implemented in a simpler way that timestamp wouldn't really help either.
once solution is to send a single DOWN_UP event for regular key presses. If we detect that the client typed a short press, we can send this event. If not we send only DOWN, and wait for the release UP (and the key will be repeated).
We had other suggestion and workaround discussed in http://comments.gmane.org/gmane.comp.emulators.spice.devel/8927
Should be addressed by http://cgit.freedesktop.org/spice/spice-gtk/commit/?id=0d5ad7e22d55552ef4be8368dcfbae879fb0b77e and by the commits before this one, so this will get fixed as part of the rebase
Since the problem described in this bug report should be
resolved in a recent advisory, it has been closed with a
resolution of ERRATA.
For information on the advisory, and where to find the updated
files, follow the link below.
If the solution does not work for you, open a new bug report.