Denis Oliver Kropp wrote: > Keith Williams wrote: > >>>> Thanks. The /usr/local/etc move did the trick.....on to implementing >>>> layers/screens. >>>> >>>> >>> Great! >>> >>> It's so much easier without the FBDev API sitting in your way :) >>> >>> >>> >> OK, after transitioning to devmem from Linux framebuffer, I thought that >> I would not enable any acceleration and make sure that I set everything >> up OK with layers and screens. >> >> Unfortunately, I have yet to get the system to come up. Origianlly, I >> was having problems with the fact that the length information in >> directfb is not entered in as hex, as the addresses are--I couldn't find >> this documented anywhere. After fixing this, the system catches a >> signal 11. Running with --dfb:debug shows the following: >> >> Core/GfxState: switch from (nil) (0) to 0x30eec [1] >> Core/GfxState: mod_hw 0x00033fff, modified 0x00000000 >> Core/GfxState: mod_hw 0x00033fff, set 0x00000000 >> Core/GfxState: mod_hw 0x00000000, set 0x00000000 >> >> Caught signal 11 (at (nil), invalid address) >> >> Can anyone point me toward what this looks like it might be? >> > > You probably did not set functions like FillRectangle or Blit after indicating > support in CheckState and setting necessary state values in SetState. > > Actually, the problem was in the ...checkstate function. I had commented out all of the acceleration functions and the setting of the call-backs, but checkstate would always toggle the state->accel bit no matter what. So, DirectFB would try to call something.
_______________________________________________ directfb-users mailing list directfb-users@directfb.org http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users
