Difference between revisions of "ECE497 Lab12 The Display SubSystem (DSS)"

From eLinux.org
Jump to: navigation, search
m (Lab 11/7a OSD Setup)
m (Updated location of lab files)
Line 5: Line 5:
 
The [http://software-dl.ti.com/trainingTTO/trainingTTO_public_sw/davinci/TTO_soc_workshop_v3.05/TTO_Linux_SOC_Lab_Exercises_omap3530_v3.05.pdf Student Materials (page 7-1)] for the workshop are very complete; however the workshop is based on the DVEVM and not the BeagleBoard.  Here I'll walk you through Lab 11/7 on the Beagle.  The Video Driver  
 
The [http://software-dl.ti.com/trainingTTO/trainingTTO_public_sw/davinci/TTO_soc_workshop_v3.05/TTO_Linux_SOC_Lab_Exercises_omap3530_v3.05.pdf Student Materials (page 7-1)] for the workshop are very complete; however the workshop is based on the DVEVM and not the BeagleBoard.  Here I'll walk you through Lab 11/7 on the Beagle.  The Video Driver  
  
Here are my files for Lab 11/7 [[File:Workshop.Lab07.tar.gz]].  Lab 11/7 has four parts.
+
'''VideoThru.tar.gz''' are my files for Lab 11/7. You can get them off the class dfs site.  Lab 11/7 has four parts.
 
# OSD Setup (Take a bmp file, convert it to r16 format and display it on the GFX frame buffer (/dev/fb0).)
 
# OSD Setup (Take a bmp file, convert it to r16 format and display it on the GFX frame buffer (/dev/fb0).)
 
# Video Record (Read video from a web cam and store it in a file.)
 
# Video Record (Read video from a web cam and store it in a file.)

Revision as of 12:22, 20 April 2011


This page explains, with examples, how the Display Subsystem (DSS) on the OMAP3530 on the BeagleBoard works. It's based on the materials used in TI's DaVinci System Integration using Linux Workshop. The workshop is based on the DVEVM. I've converted those materials to the BeagleBoard. What I'm called Lab 11 is Lab 7 in these materials.

The Student Materials (page 7-1) for the workshop are very complete; however the workshop is based on the DVEVM and not the BeagleBoard. Here I'll walk you through Lab 11/7 on the Beagle. The Video Driver

VideoThru.tar.gz are my files for Lab 11/7. You can get them off the class dfs site. Lab 11/7 has four parts.

  1. OSD Setup (Take a bmp file, convert it to r16 format and display it on the GFX frame buffer (/dev/fb0).)
  2. Video Record (Read video from a web cam and store it in a file.)
  3. Video Playback (Read the file from above and display it on one of the video frame buffers (/dev/fb1).)
  4. Video Loopthru (Combine the previous two labs to read the live video and display it.)

Here is a nice set of slides that explain the OMAP3530 DSS. I found a reference to it here. It's worth going through the slides before doing these labs.

Lab 12/7a OSD Setup

This lab is a straight forward inspection lab that starts on page 200. Look at video_thread.c and video_osd.c and see how they work. The lab has you create a 720 by 60 bmp file that is converted to RGB565 format. Instead, create a 640 x 480 file that has a white background. In Lab 12/7c we will use the OMAP DSS to superimpose the video on this image. I've already changed some #defines to work with the 640 by 480 image.

The Beagle code differs from the original workshop code in the the OMAP3530 doesn't have an attribute frame. All that code has been removed.

Lab 11/7b Video Record

This is also a straight forward inspection lab. It starts on page 206. I'm using at PS3 EYE eb cam. /dev/video0 will appear when you plug the EYE into the USB of the Beagle. Look at video_thread.c and video_input.c and see how they work. I've set them up to capture a 640 by 480 image.

Ditto on the attribute frame. Capture some video here. You will use it in the next lab. The video is stored in /tmp. I suggest you move it to your working directory, otherwise it will disappear when you reboot.

Lab 11/7c Video Playback

Here we really depart from the workshop code. The lab starts on page 209. video_osd.c writes the the image your created in lab 7a on frame buffer 0 (/dev/fb0). This is called the graphics buffer (FBVID_GFX). video_output.c reads the video from lab 7b and writes it to frame buffer 1 (/dev/fb1), which is one of the two video buffers.

You need to boot with the following arguments for this to work.

setenv mmcargs 'setenv bootargs console=${console} vram=${vram} omapfb.mode=dvi:${dvimode} omapfb.debug=y omapdss.def_disp=${defaultdisplay} root=${mmcroot} rootfstype=${mmcrootfstype} ${optargs}'
setenv optargs omapfb.vram=0:10M,1:10M vram=20M 
setenv bootargs console=ttyS2,115200n8 root=/dev/mmcblk0p2 rw rootwait omapfb.mode=1024x768MR-16@60 omapfb.debug=y omapdss.def_disp=dvi omapfb.vram=0:10M,1:10M vram=20M
saveenv

This reserves 20M for the video ram and allocate 10M for each for framebuffers 0 and 1. The SPEd image should have all this set, though the values may be different. Therefor you may not need to change your boot arguments.

The following (from Media:ShowOSD.txt) sets up the DSS so graphics buffer is in front of the video buffer. The graphics are at 50% transparency and while is transparent. This will let the video show through the graphics.

showOSD


#!/bin/bash
# /etc/init.d/gpe-dm stop
# Here are the bootargs that are needed
# set mmcargs 'setenv bootargs console=${console} vram=${vram} omapfb.mode=dvi:${dvimode} omapfb.debug=y omapdss.def_disp=${defaultdisplay} root=${mmcroot} rootfstype=${mmcrootfstype} ${optargs}'
# setenv optargs omapfb.vram=0:10M,1:10M vram=20M 
# omapfb.test=y
# saveenv

video_size=640,480

ovl0=/sys/devices/platform/omapdss/overlay0
ovl1=/sys/devices/platform/omapdss/overlay1
ovl2=/sys/devices/platform/omapdss/overlay2

mgr0=/sys/devices/platform/omapdss/manager0
mgr1=/sys/devices/platform/omapdss/manager1

lcd=/sys/devices/platform/omapdss/display0
tv=/sys/devices/platform/omapdss/display1

fb0=/sys/class/graphics/fb0
fb1=/sys/class/graphics/fb1
fb2=/sys/class/graphics/fb2

# Disable the overlays
echo "0" > $ovl0/enabled
echo "0" > $ovl1/enabled

# Connect fb0 to ovl0
echo "0" > $fb0/overlays
echo "1" > $fb1/overlays
echo $video_size > $fb1/virtual_size

# Point both overlays to the lcd manager
echo "lcd" > $ovl0/manager
echo "lcd" > $ovl1/manager

# for param in bits_per_pixel modes virtual_size
# do
# 	cat $fb0/$param > $fb1/$param
# done

# echo $ovl0/output_size > $ovl1/output_size
echo 0,0 > $ovl0/position
echo $video_size > $ovl1/output_size
echo 100,100 > $ovl1/position

echo 128 > $ovl0/global_alpha
echo   1 > $mgr0/alpha_blending_enabled

# Turn on transparency, make white (65535) transparent
echo   1 > $mgr0/trans_key_enabled
echo 65535 > $mgr0/trans_key_value
# echo video-source > $mgr0/trans_key_type
echo gfx-destination > $mgr0/trans_key_type

# Set the video mode to YUV (not need since the c-code does it)
# http://groups.google.com/group/beagleboard/browse_thread/thread/9bc347f5f0853aa1/907f1ac3554b1a19?lnk=gst&q=fbset#907f1ac3554b1a19
# /usr/sbin/fbset -fb /dev/fb1 -nonstd 8

echo "1" > $ovl0/enabled
echo "1" > $ovl1/enabled

Turning off gpe

Running above will write on top of what Angstrom is doing. You can turn off the display manager by using:

/etc/init.d/gpe-dm stop

Replace stop with start if you want to run it again.

Lab 11/7d Video Loopthru

Documentation to come.