problems with glMapBufferOES

[Brian Beuken]

Hi
I'm having some strange issues with glMapBufferOES
my command

Code: Select all

	framedata = (unsigned char*)glMapBufferOES(GL_ARRAY_BUFFER, GL_WRITE_ONLY_OES);

always returns null, I do know that glMapBufferOES is an extension and I have checked that there is a proc address for it,

Code: Select all

glMapBuffer  =  (PFNGLMAPBUFFEROESPROC)eglGetProcAddress("glMapBufferOES");

which returned a value.. but whether I call it via a proc address or direct, the framedata is never set and I cannot proceed!

Anyone know why? Am I missing an init step somewhere? I have the bcm_host_init(); at the start of my code which should fire up all the BCM specific stuff?

[Brian Beuken]

to partially answer my own questions, I tracked this down to a problem with the way the buffers are created

Code: Select all

	glBufferData(GL_ARRAY_BUFFER, sizeMemory, NULL, GL_STREAM_DRAW);

seems to prevent glMapBufferOES from working
but

Code: Select all

	glBufferData(GL_ARRAY_BUFFER, sizeMemory, NULL, GL_DYNAMIC_DRAW);

allows it...I've not got my code up and running yet to see what impact Dynamic has over stream though.
There is some comments on line that suggest stream_draw may have a bug in it.

[jamesh]

Has this been reported on the RPF's github?

[Brian Beuken]

not sure.. I just stumbled on it, and it took quite some googling to find any reference to any issues, but it appears to relate to broadcom drivers rather than OGLES itself