Re: [Dri-devel] Bug in glxinfo.c

[Brian P. <br...@pr...>]

Gerhard den Hollander wrote:
> 
> glxinfo.c (as available from
> http://dri.sourceforge.net/resources/resources.html) has a bug on
> multiheaded displays.
> It always looks at screen 0
> 
> The fix is simple.
> diff glxinfo.c glxinfo.gdh.c
> 522c522
> <       print_screen_info(dpy, 0);
> ---
> >       print_screen_info(dpy, scrnum);
> 524c524
> <       print_visual_info(dpy, 0, mode);
> ---
> >       print_visual_info(dpy, scrnum, mode);
> 
> or with a bit more context
>    numScreens = ScreenCount(dpy);
>    for (scrnum = 0; scrnum < numScreens; scrnum++) {
> /*was print_screen_info(dpy, 0) */
>       print_screen_info(dpy, scrnum);
>       printf("\n");
> /*was print_visual_info(dpy, 0, mode);  */
>       print_visual_info(dpy, scrnum, mode);
>       if (scrnum + 1 < numScreens)
>          printf("\n\n");
>    }

Thanks.  I've fixed it and will update the new version soon.


> On a different note,
> I noticed that the glRenderer string is retrieved with
>     const char *glRenderer = (const char *) glGetString(GL_RENDERER);
> 
> And in my case returns  Mesa GLX Indirect
> 
> I am running with a dual head setup, and in that case I have no hardware
> acceleration
> could this be because my secondary head (but display :0.0) is a G200 (which
> is not yet supported ) ?

Jeff's got the G200 working here.  Not sure when he'll be checking
in that code.


> I have a voodoo3 as primary card (but display :0.1) and if I run
> singleheaded (with only the voodoo card) I do get hardware acceleration .

We haven't officially addressed multi-head support yet.  Some other
people seem to have had some success with it though.  Perhaps they
can chime in.

-Brian