Hi again all,

This is a question about source and destination surface buffers: These buffers
have an "offset" and a "pitch". Now, my framebuffer device is seen by the
accelerator not as a linear framebuffer but rather as a two dimensional array
with a "virtual" width and height. 

You could say that this imposes that the only possible value for a surface
buffer's pitch field is the framebuffer's pitch, and that the driver has to
process the offset given by DFB into (x,y) coordinates in the framebuffer
virtual space, and use this as a two dimensional offset.

My question is how do I tell DFB's surface manager that my framebuffer is
limited in this way (alignment won't do, it has to be the one exact value, the
way I see it at the moment anyway. I suspect the call to
"dfb_gfxcard_adjust_heap_offset" in fbdev.c has something to do with this), and
how will this limitation affect DFB performance/behavior?

A more basic question is – when are multiple surfaces created and managed on the
same linear framebuffer? Running the DFB examples (even though I haven't yet
been able to go through them all nor to use each to the fullest), I was only
able to see buffer offsets which were 0, and pitch values were sometimes 0 and
sometimes the framebuffer device's line_length field which is the proper
frmaebuffer pitch I require. I am not sure how I should interpret a pitch value
of 0 (perhaps this means the source pitch should be inferred from the
destination pitch in these cases, which was non zero).

Thanks
S

_______________________________________________
directfb-users mailing list
[email protected]
http://mail.directfb.org/cgi-bin/mailman/listinfo/directfb-users

Reply via email to