Hi,
The DISPC's GFX pipeline supports BITMAP1/2/4/8 according to the overview of 11.2.1 in the TRM (SPRUHZ6D). However, the errata i596 (in SPRZ429H) says that BITMAP1/2/4 aren't supported. That's OK, it's BITMAP8 I'm interested in.
The omapdrm driver (/kernel/drivers/gpu/drm/omapdrm/omap_fbdev.c) uses a hardwired ARGB32 format in omap_fbdev_create() to support alpha blending, but in my application I need palletised 8bpp.
Looking at the table for the FORMAT bits of DISPC_GFX_ATTRIBUTES the BITMAP8 setting seems to be missing but it's possible to just guess the settings since there are only 16 combinations to choose from. Or are there? Table 11-94 lists 17 ticks in the GFX column.
Halting the console output (ie. so there's some text on the screen) the FORMAT bits can be tried one at a time, and the text changes width correspondingly - ie. a 16bpp format is twice width, albeit totally the wrong colour. None of the settings correspond to 4x width which is what I would expect for 8bpp.
So - which register bits control BITMAP8?
Thanks,
James.
