Bug 213215
| Summary: | Extension GLX is broken | ||
|---|---|---|---|
| Product: | [Fedora] Fedora | Reporter: | Patrick Monnerat <patrick> |
| Component: | vnc | Assignee: | Adam Tkac <atkac> |
| Status: | CLOSED RAWHIDE | QA Contact: | David Lawrence <dkl> |
| Severity: | medium | Docs Contact: | |
| Priority: | medium | ||
| Version: | 6 | CC: | ovasik, rvokal |
| Target Milestone: | --- | ||
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | 4.1.2-6.fc7 | Doc Type: | Bug Fix |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2006-11-20 12:03:29 UTC | Type: | --- |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
|
Description
Patrick Monnerat
2006-10-31 10:40:28 UTC
Could you please give me link for glx application that doesn't run?? I tried with googleearth (http://dl.google.com/earth/GE4/GoogleEarthLinux.bin), that revealed the problem (start it from a terminal to see its error messages). Blender should do it too, but i did not check. hmm, I have 2 machines - m1 & m2 m2$ ssh -X m1 m1$ vncviewer :0 and in vncviewer I've started googleearth program and all works fine. Could you please specify more your reproduction?? I do not understand quiet well your test conditions. Where is the vnc server ?
Which graphic environment is displayed by m1's vnc window.
IMHO, I think the problem is that GLX support is not compiled into the vnc
SERVER (/usr/bin/Xvnc). The client is of little importance, since it only deals
with the RFB protocol. Rendering is done in the SERVER.
The conditions in wich the problem occurs:
Machine m1: FC6 dynamic vnc server. Configuration info:
1) /etc/services:
.
.
.
vnc-1600-1200x24 5935/tcp
2) /etc/xinetd.d/vnc-1600-1200x24:
service vnc-1600x1200x24
{
socket_type = stream
wait = no
user = root
log_on_failure += USERID
server = /usr/bin/Xvnc
server_args = -inetd -once -query localhost -geometry 1600x1200 -dpi
100 -depth 24 -fp unix/:7100 -securitytypes=none
disable = no
}
3) Restart xinetd
4) Enable local XDMCP, restart display manager.
machine m2 is an FC4 workstation: in a terminal window, submit command
vncviewer m1:35
... but again, I do not think the viewer version/os/etc. is involved in the problem.
The command on m2 above tells m1's xinetd to launch a new vnc server and start a
session with the login screen.
Once logged-in (in m1), open a terminal (on m1 via the m2 vnc window) and launch
googleearth.
Does it really make sense to have accelerated Vnc X server? That would require bunch of opengl libraries to be added and I don't see any revenue for this. It does make sense, of course: Although the animations are "jumpy", it allows to remotely run programs optimized for accelerated graphics that would'nt otherwise accept to run at all ! By the way, I applied the "BuildGlxExt" (opengl) patch on the FC4 version of VNC (4.1.1-11.fc4): the resulting vnc-server package supports successfully GLX applications under FC4 via any vnc client... I've fixed glx extensions. Could you try to reproduce with vnc-4.1.2-6.fc7, please? Thanks Tested vnc-4.1.2-6.fc7 server with blender on FC6: it works great ! FYI: GLX is a catastrophe with vnc-4.1.2-7.fc7 ... Starting blender from the menu makes a white window with a very large gray frame to blink(i.e: appear and disappear quickly), then it stays steady with a black window (without window manager items). It is possible to stop blender via Alt+Q. Starting blender from a terminal blackens the whole screen immediately. It is possible to recover via Alt+Q. By adding option -render to Xvnc, it works OK! Many thanks for your work Looks like big conflict between render & glx extensions. In next release I'm going turn off both of these extensions by default and you could start glx extensions with +glx |