Hi devels
After some investigation i found out the following:
My accelerated blit function is only called when surface->Flip is
called. So this explains why the blending is done. For all the calls to
surface->Blit the software callback is called afterall. This happens
because although CheckState is called and the acceleration is enabled
the call to
dfb_gfxcard_state_acquire
fails, most often because it can't lock the source surface for GPU
access. Commenting out these "locking" blocks of code results in a
segfault. I assume this happens because if these "locking" function
calls fail or not called state->src and/or state->dst are not properly
setuped
The problem is that in my embedded system i don't have a GPU. Also i
don't have dedicated memory for graphics operations (such as the RAM
onboard graphics cards). So there is no need to lock these data areas.
How can i overcome this issue? Any ideas?
Best regards
Tasos parisi...@sciensis.com
_______________________________________________
directfb-dev mailing list
directfb-dev@directfb.org
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-dev
