NXP

[i.MX6Q][QNX Neutrino 6.6.0]调试笔记------LVDS显示配置

2019-07-12 12:58发布

站内文章 / NXP MCU
16310 0 1785
QNX提供给imx6-SebreSmart的BSP默认是HDMI显示的，我板子是LVDS接显示屏的，所以需要修改显示这块。从资料介绍来看，QNX是采用了一个OpenWF的API集，具体如何实现的不关心，这里只需要关心的是如何去配置LCD的参数和时序供OpenWF使用即可
OpenWF™是一个免版税,跨平台APIs，它提供了一个底层硬件抽象接口，利用组成成分和显示硬件合成窗口系统。当有新的显示器接入，只需要通过OpenwF的显示配置API(Wfdcfg Library)提供显示器的参数。Wfdcfg Library把显示器的属性和模式提供显示驱动和屏幕，交互图如下：

 整个配置过程其实就是生成一个wfdcfg函数库供OpenWF调用
1、修改源码 wfdcfg源码目录：QNX_BSP_sabresmart/src/hardware/wfd/imx6x/wfdcfg，在该目录已经有 imx6x-hdmi目录的存在，拷贝一份hdmi为lvds
$:cp imx6x-hdmi imx6x-lvds $:cd imx6x-lvds $:mv hdmi.c mylvds.c
根据自己LCD参数修改mylvds.c static const struct mode modes[] = { // 1280x720 @ 60 Hz { .timing = { .pixel_clock_kHz = 74250, .hpixels = 1280, .hfp=110, .hsw= 40, .hbp=220, // 1650 total .vlines = 720, .vfp= 5, .vsw= 5, .vbp= 20, // 750 total .flags = 0, }, }, { // marks end of list .timing = {.pixel_clock_kHz = 0}, }, };
退出重新编译后，会在QNX_BSP_sabresmartsrchardwarewfdimx6xwfdcfgimx6x-lvds toarmdll.le.v7中生成一个wfdcfglibrary-----libwfdcfg-imx6x-lvds.so 将libwfdcfg-imx6x-lvds.so拷贝到QNX_BSP_sabresmartprebuiltarmle-v7usrlibgraphicsiMX6X
2、修改配置文件 在QNX_BSP_sabresmartprebuiltarmle-v7usrlibgraphicsiMX6X目录中有个graphics.conf的配置文件，可以配置LCD的模式、分辨率等等,其中的分辨率配置得在上面mylvds.c中支持 graphics.conf
 begin khronos begin egl display 1 egl-dlls = libGAL-gc320v0.so libEGL_viv.so glesv1-dlls = libGAL-gc320v0.so libEGL_viv.so libGLES_CM_viv.so glesv2-dlls = libGAL-gc320v0.so libEGL_viv.so libGLESv2_viv.so vg-dlls = libGAL-gc320v0.so libEGL_viv.so libOpenVG_viv.so glesv2-sc-dlls = libGLSLC.so # OpenVG emulated on the 3D hardware. Runs on all known i.MX6 platforms except SoloLite. gpu-dlls = libGalcore-3dvg.so end egl display begin wfd device 1 # 7" Chimei 1024x600 @60 display on LVDS0 wfd-dlls = libwfdcfg-imx6x-lvds.so libimx6xCSCgamma-generic.so libWFDimx6x.so # 其中这里添加刚刚生成的lvds库 display-type = lvds0 # 选择LVDS接口 # If you need IPU1 for capture, you can move WFD to IPU2 instead # IPU选择 ipu-index = 1 # ipu-index = 2 # Abort on vsync timeouts (used to detect PLL failure) vsync-failure-limit = 4 # GPU driver minimum memory pool sizes in MB (2..248) # gpu-contiguousMB = 248 gpu-contiguousMB = 64 # recommended # gpu-contiguousMB = 2 # minimum gpu-internalPoolMB = 6 gpu-sharedPoolMB = 2 #gpu-pm-force-power-on = true # Options to control what debug build of driver writes to slog debug-alloc = 0 # Logs memory allocations debug-cscgamma = 0 # Logs CSC/Gamma correction activity debug-idmac = 0 # Logs IDMAC parameters debug-isr = 0 # Logs IRQ statistics debug-pipeline = 0 # Logs pipeline activity debug-register = 0 # Logs register activity to tracelog (not slog!) debug-resize = 0 # Logs resizer (IPU IC PP) activity debug-trace = 0 # Logs entry into each WFD function debug-misc = 0 # Logs everything else end wfd device end khronos begin winmgr begin globals blit-config = gc320v0 blits-logsize = 4096 end globals begin display internal # Innolux display only video-mode = 1024 x 600 @ 60 # 适应LCD的分辨率 end display begin class framebuffer display = internal pipeline = 2 format = rgba8888 # 显示格式，这个很关键，修改过这个参数，在后面吃过大亏 usage = gc320v0 end class end winmgr
3、修改Buildfile 在Buildfile的Graphics support部分，添加lvds库文件 [search=../install/armle-v7/usr/lib/graphics/iMX6X] /usr/lib/graphics/iMX6X/libwfdcfg-imx6x-lvds.so=libwfdcfg-imx6x-lvds.so
在QNX_BSP根目录 make clean  然后 make,生成新的QNX-IFS，拷贝到SD卡中启动板子： Welcome to QNX Neutrino Initial Program Loader for Freescale i.MX6Q Sabre-Smart (ARM Cortex-A9 MPCore) SDMMC download... load image done. Found image @ 0x18000008 Jumping to startup @ 0x108068F8 SCU_CONFIG = 00005503, 4 cpus L2 cache enabled CPU0: L1 Icache: 1024x32 CPU0: L1 Dcache: 1024x32 WB CPU0: VFP-d32 FPSID=41033094 CPU0: NEON MVFR0=10110222 MVFR1=01111111 CPU0: 412fc09a: Cortex A9 MPCore rev 10 792MHz Detected i.MX6 Dual/Quad, revision Unknown Revision PLL1 - System : 792MHz PLL2 - System Bus : 528MHz PLL3 - OTG USB : 480MHz PLL4 - Audio : 36MHz PLL5 - Video : 339MHz PLL7 - Host USB : 480MHz PLL8 - Enet : 50MHz IPG clock : 66000000Hz IPG per clock : 66000000Hz UART clock : 80000000Hz CSPI clock : 60000000Hz AHB clock : 132000000Hz AXI clock : 264000000Hz EMI_SLOW clock: 29333333Hz DDR clock : 528000000Hz USDHC1 clock : 198000000Hz USDHC2 clock : 198000000Hz USDHC3 clock : 198000000Hz USDHC4 clock : 198000000Hz ENFC clock : 24000000Hz CPU1: L1 Icache: 1024x32 CPU1: L1 Dcache: 1024x32 WB CPU1: VFP-d32 FPSID=41033094 CPU1: NEON MVFR0=10110222 MVFR1=01111111 CPU1: 412fc09a: Cortex A9 MPCore rev 10 792MHz CPU2: L1 Icache: 1024x32 CPU2: L1 Dcache: 1024x32 WB CPU2: VFP-d32 FPSID=41033094 CPU2: NEON MVFR0=10110222 MVFR1=01111111 CPU2: 412fc09a: Cortex A9 MPCore rev 10 792MHz CPU3: L1 Icache: 1024x32 CPU3: L1 Dcache: 1024x32 WB CPU3: VFP-d32 FPSID=41033094 CPU3: NEON MVFR0=10110222 MVFR1=01111111 CPU3: 412fc09a: Cortex A9 MPCore rev 10 792MHz cpu_startnext: cpu1 -> fc411d58 cpu_startnext: cpu2 -> fc411d58 cpu_startnext: cpu3 -> fc411d58 System page at phys:10023000 user:fc410000 kern:fc410000 Starting next program at vfe054e3c cpu_startnext: cpu0 -> fe054e3c Welcome to QNX Neutrino 6.6.0 on the i.mx6 Smart-Device (ARM Cortex-A9 MPCore) Starting watchdog... Starting SD3 memory card driver... Starting SD1 (microSD connector) memory card driver... Unable to access "/dev/sd20" (2) setting env variables. Starting Screen Graphics... Mempool Map addr range[28100000-2c700000] Mempool Map paddr range[4b600000-4fc00000] [Interrupt] Attached irqLine 41 with id 6. [Interrupt] Attached irqLine 42 with id 7. Attached resmgr to /dev/galcore with id:0. done. Launching devb-umass... done #
从上面打印信息来看Screen已经启动OK了，运行OpenGL的测试程序gles2-gears，可以看到三个齿轮旋转， {MOD}差没什么都正常就说明Wfdcfg Library是OK了

热门文章