Brian Paul wrote:
> The DRI trunk code has a few problems with thread safety and
> swapbuffers (all drivers but tdfx, probably). I've fixed this on
> the 3.5 branch. You might try that.
hmm... I picked the branch on a bad day it seems...
#1 0x81ff980 in ?? ()
#2 0x406ccc7a in _tnl_run_pipeline (ctx=0x814ac00) at t_pipeline.c:155
#3 0x40728db0 in radeonWrapRunPipeline (ctx=0x814ac00) at radeon_tris.c:435
#4 0x406c0f5e in _tnl_DrawArrays (mode=7, start=8, count=4) at t_array_api.c:222
[...]
my application is calling:
glDrawArrays(GL_QUADS, i*4, 4);
at this point. Looking at t_pipeline.c:
151 if (s->active && running) {
152 if (s->changed_inputs)
153 changed_inputs |= s->outputs;
154
155 running = s->run( ctx, s );
156
157 s->changed_inputs = 0;
158 VB->importable_data &= ~s->outputs;
159 }
(gdb) print *s
$7 = {name = 0x0, check_state = 0, run_state = 0, active = 0 '\000',
inputs = 0, outputs = 0, changed_inputs = 0, privatePtr = 0x0,
destroy = 0, check = 0, run = 0}
But my original problem is solved, I can use two threads with each of
them rendering to their own window. :-)
--
Marcelo
_______________________________________________
Dri-devel mailing list
[EMAIL PROTECTED]
http://lists.sourceforge.net/lists/listinfo/dri-devel
