To further clarify the answer, when you call XQueryExtension() from a 3D application, this call gets intercepted by VirtualGL. If the query is about extension "GLX", then VirtualGL always returns True. If the query is about another extension, then VirtualGL passes the call to the server's Xlib unmodified, as if VirtualGL had not intercepted the function at all. Xlib will send the request to whatever display is specified in the dpy argument to the function, which is the "2D X display", not the "3D X display".
On 3/25/10 8:31 PM, DRC wrote: > OK, never mind the question about which version of VirtualGL, because > the interposed version of XQueryExtension() hasn't changed since 2.1. > > VirtualGL generally sends all glX*() requests to the 3D X server and the > rest of the X calls to the 2D X server. VGL intercepts some plain X11 > calls, such as those related to window creation/destruction/resizing, so > that it can maintain a proper correlation between the Pbuffers that are > used for 3D rendering and their corresponding windows. In general, > though, all X11 commands except for GLX commands are sent to the 2D X > server. > > In the specific case of XQueryExtension(), VirtualGL intercepts it > solely for the purpose of returning True if the application calls that > function with an argument of "GLX". Some applications try to call > XQueryExtension() to determine whether GLX is available and will fall > back to using a non-OpenGL method of drawing if GLX is not available. > Thus, VirtualGL has to intercept the function to tell such applications > that GLX is available, even if the 2D X server does not support it. > > Please keep me in the loop about integration with QEMU, as that project > interests me as well. > > On 3/25/10 12:46 PM, Rodrex Lee wrote: >> Dear, >> >> We are currently evaluating virtualGL for its extension on QEMU.But we >> are currently facing some issues realted to it. >> Xlib: Extension "GLX" missing on display. >> In order to solve the above issue, I have few questions related to >> virtualGL. >> 1) The overrided XQueryExtension function in virtualGL calls real >> XQueryExtension function of XLib[Accelerated server] which eventually >> tests whether the GLX extension is present in the Xserver.Now it must >> check about this extension in Xserver of accelerated server, but when we >> debugged the same,we find it that it went to Xserver of user's >> display[client remote side].how? >> 2) How the Xlib of server side decides that it has to send the X11 >> events to 2DXserver of client side, not to the Xserver of server side? >> Thanks & Regards, >> Rodrex Lee. >> >> >> >> ------------------------------------------------------------------------------ >> Download Intel® Parallel Studio Eval >> Try the new software tools for yourself. Speed compiling, find bugs >> proactively, and fine-tune applications for parallel performance. >> See why Intel Parallel Studio got high marks during beta. >> http://p.sf.net/sfu/intel-sw-dev >> >> >> >> _______________________________________________ >> VirtualGL-Users mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/virtualgl-users ------------------------------------------------------------------------------ Download Intel® Parallel Studio Eval Try the new software tools for yourself. Speed compiling, find bugs proactively, and fine-tune applications for parallel performance. See why Intel Parallel Studio got high marks during beta. http://p.sf.net/sfu/intel-sw-dev _______________________________________________ VirtualGL-Users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/virtualgl-users
