Freescale Semiconductor i.MX 6SoloLite Linux, Справочное руководство

Страница: 87 / 199

components shown in light gray are the standard components in the X-Windows System
without acceleration. The components shown in orange are those added to support X-
Windows System acceleration and are briefly described here.
The i.MX X Driver library module (
vivante-drv.so
) is loaded by the X server and
contains the high-level implementation of the X-Windows acceleration interface for i.MX
platforms containing the GC320 2D GPU core. The entire linearly contiguous frame
buffer memory in
/dev/fb0
is used for allocating pixmaps for X both on screen and off
screen. The driver supports a custom X extension which allows X clients to query the
GPU address of any X pixmap stored in frame buffer memory.
The libGAL.so library module (
libGAL.so
) contains the register-level programming
interface to the GC320 GPU module. This includes the storing of register programming
commands into packets which can be streamed to the device. The functions in the
libGAL.so library are called by the i.MX X Driver code.
11.3.2 i.MX 6 Driver for X-Windows System
The i.MX X Driver, referred to as vivante-drv.so, implements the EXA interface of the X
server in providing acceleration.
The implementation details are as follows:
• The implementation builds upon the source from the fbdev frame buffer driver for X,
so that it can be the fallback when the acceleration is disabled.
• The implementation is based on X server EXA version 2.5.0.
• The EXA solid fill operation is accelerated, except for source/target drawables
containing less than 1024x1024 pixels, in which case software failure may occur.
• The EXA copy operation is accelerated, except for source/target drawables
containing less than 1024x1024 pixels, in which case software failure may occur.
• EXA putimage (upload into video memory) is accelerated, except for source
drawables containing less than 400x400 pixels, in which case software failure may
occur.
• For EXA solid fill, only solid plane masks and only GXcopy raster-op operations are
accelerated.
• For EXA copy operation, the raster-op operations (GXandInverted, GXnor,
GXorReverse, GXorInverted, GXnand ) are not accelerated.
• EXA composite allows for many options and combinations of source/mask/target for
rendering. Commonly used EXA composite operations are accelerated.
The following types of EXA composite operations are accelerated:
Chapter 11 X Windows Acceleration
i.MX 6SoloLite Linux Reference Manual, Rev. L3.0.35_4.1.0, 09/2013
Freescale Semiconductor, Inc. 87