Creating the Yocto/OE environment

Now that we have developed the main application, what remains for us is to integrate it with our Yocto RPI environment.

Modifying the image

The first step consists of integrating the nodejs package (located at meta-openembedded/meta-oe/recipes-devtools/nodejs) with the Raspberry Pi environment. For this, we have to add the following line to raspberry-pack-image:

# Base this image on core-image-minimal 
include recipes-core/images/core-image-minimal.bb 
DESCRIPTION = "Image for raspberry-pi" 
IMAGE_FEATURES += "ssh-server-dropbear splash" 
# Include modules in rootfs 
IMAGE_INSTALL += "  
    kernel-modules  
    gpio-packt     
    i2c-tools     
    spitools     
    nunchuck     
    v4l-utils     
    nodejs         
"

We have now integrated nodejs with our Poky distro.

Creating the recipe file

The second step consists of creating the recipe file; let's call this recipe webserver-packt_01.bb.

The idea is to have the following architecture inside the recipes-custom folder:

$ tree webserver-packt/
webserver-packt/
├── webserver-packt
│   └── server.tar.gz
└── webserver-packt_0.1.bb

Note

The server.tar.gz tarball contains the main application (nodejs + HTML5).

To create this recipe, we will use the recipetool command , as follows:

$ source oe-init-build-env rpi-build
$ recipetool create -d -o webserver-packt_0.1.bb ../meta-packt_rpi/recipes-custom/webserver-packt/webserver-packt/
$ cp webserver-packt_0.1.bb ../meta-packt_rpi/recipes-custom/webserver-packt/

Note

The complete documentation of recipetool command is online at http://www.yoctoproject.org/docs/1.8/mega-manual/mega-manual.html#new-recipe-creating-the-base-recipe-using-recipetool .

Now we just need to copy the entire archive (server.tar.gz) containing the sources to ${datadir} (/usr/share/); this step is carried out in the do_install() block, as shown here:

DESCRIPTION = "Panel to monitor rpi temperature" 
LICENSE = "GPLv2" 
LIC_FILES_CHKSUM = "file://${COREBASE}/meta/COPYING.GPLv2;md5=751419260aa954499f7abaabaa882bbe" 
 
# Package Release 
PR = "r0" 
 
# Use local tarball 
SRC_URI = "  
  file://server.tar.gz  
  " 
   
RDEPENDS_${PN} += "bash" 
 
# Make sure our source directory (for the build) matches the directory structure in the tarball 
S = "${WORKDIR}" 
 
do_install() { 
  install -d ${D}${datadir}/server-packt 
  cp -a ${S}/server ${D}${datadir}/server-packt 
} 
 
# Package files 
FILES_${PN} += "${datadir}/server-packt"

Explanation

Let's take a look at what the variables do:

  • The RDEPENDS_{PN} variable lists a package's runtime dependencies that must be installed in order for the built package to run correctly. In our case, bash is necessary for nodejs.
  • * The FILES_{PN} variable contains the list of directories or files that are placed in the packages. In our case, server-packt is necessary.

We now have an environment ready to be used. If we want to deploy the application on our Raspberry Pi, we can do so with the following commands.

The host commands are as follows:

$ bitbake webserver-packt
$ scp tmp/deploy/rpm/arm1176jzfshf_vfp/webserver-packt-0.1- r0.arm1176jzfshf_vfp.rpm  root@ip_address_of_rpi:/home/

These are the Raspberry Pi commands:

$ rpm -ivh webserver-packt-0.1-r0.arm1176jzfshf_vfp.rpm
$ rm webserver-packt-0.1-r0.arm1176jzfshf_vfp.rpm
$ node /usr/share/server-packt/server/server.js

Autostarting - the init file

If we want to have a standalone application, it would be interesting if our application can start during the boot sequence of our Raspberry Pi. For this, we will use an init script, inheriting directly from System V.

The first step is to create a .init file within our structure. We will call this file server.init. The tree will look like this:

$ tree webserver-packt/
webserver-packt/
├── webserver-packt
│   ├── server.init
│   └── server.tar.gz
└── webserver-packt_0.1.bb

So, the script will start the application automatically after the installation and every time the Raspberry Pi starts up. Here is the init script from System V:

#!/bin/bash                                           
### BEGIN INIT INFO 
# Provides:          server.init 
# Required-Start:    $remote_fs $syslog 
# Required-Stop:Å    $remote_fs $syslog      
# Default-Start:     2 3 4 5 
# Default-Stop:      0 1 6 
# Short-Description: Start daemon at boot time 
# Description:       Enable service provided by daemon. 
### END INIT INFO 
 
DAEMON=node 
NAME=server.init 
DESC="Nodejs app" 
ARGS="/usr/share/server-packt/server/server.js" 
 
set -e 
 
usage() 
{ 
        echo "----------------------------------" 
        echo "Usage: $0 (stop|start|restart)" 
        echo "----------------------------------" 
} 
  
 
service_start() 
{ 
    echo -n "starting $DESC: $NAME... " 
    start-stop-daemon -S -x $DAEMON -- $ARGS & 
    echo "done." 
} 
 
 
echo "" 
  
service_stop() 
{ 
    echo -n "stopping $DESC: $NAME... " 
    start-stop-daemon -K -x $DAEMON 
    echo "done." 
} 
  
case $1 in 
        stop) 
                service_stop 
                echo "" 
        ;; 
        start) 
                service_start 
                echo "" 
        ;; 
        restart) 
                service_stop 
                service_start 
                echo "" 
        ;; 
        *) 
                usage 
esac 
exit 0

Explanation

Let's look at how the code works:

  • The DAEMON variable specifies the binary to launch (/usr/bin/node)
  • The NAME variable specifies the script name
  • The DESC variable contains some information about DAEMON (node)
  • The ARGS variable contains the arguments to pass to DAEMON (node)

If you want to test this script on the Raspberry Pi, use these commands:

$ scp server.init root@ip_address_of_rpi:/etc/init.d
$ cd /etc/init.d
$ update-rc.d server.init defaults

This command stops the application:

/etc/init.d/server.init stop

This one restarts it:

/etc/init.d/server.init restart

Autostarting - the recipe file

The inclusion of our init file in our webserver-packt.bb recipe consists of:

  • Adding the server.init file to SRC_URI.
  • Adding do_install() to the step for our init script.
  • Including update-rc.d. Adding this to Yocto is done by adding inherit update-rc.d to our recipe.

Here is webserver-packt.bb after being updated:

DESCRIPTION = "Panel to monitor rpi temperature" 
LICENSE = "GPLv2" 
LIC_FILES_CHKSUM = "file://${COREBASE}/meta/COPYING.GPLv2;md5=751419260aa954499f7abaabaa882bbe" 
 
# Package Release 
PR = "r0" 
 
# Use local tarball 
SRC_URI = "  
    file://server.tar.gz  
    file://server.init  
    " 
     
RDEPENDS_${PN} += "bash" 
 
# Make sure our source directory (for the build) matches the directory structure in the tarball 
S = "${WORKDIR}" 
 
do_install() { 
    install -d ${D}${datadir}/server-packt 
    cp -a ${S}/server ${D}${datadir}/server-packt 
     
    install -d ${D}${sysconfdir}/init.d/ 
    install -m 0755 ${WORKDIR}/server.init ${D}${sysconfdir}/init.d/server-packt-init 
} 
 
# Package files 
FILES_${PN} += "${datadir}/server-packt" 
 
inherit update-rc.d 
 
INITSCRIPT_NAME = "server-packt-init" 
INITSCRIPT_PARAMS = "start 99 5 2 . stop 19 0 1 6 ."

Explanation

Here is an explanation of our updated webserver-packt.bb file:

  • The INITSCRIPT_NAME variable represents the filename of the initialization script, as installed to /etc/init.d.
  • * The INITSCRIPT_PARAMS variable specifies the options to pass to update-rc.d. In our case, the script has a runlevel of 99, is started at init levels 2 and 5, and is stopped at levels 0, 1, and 6.

Note

If you want more information about update-rc.d, you can visit the official documentation of the Yocto Project: http://www.yoctoproject.org/docs/1.8/mega-manual/mega-manual.html#ref-classes-update-rc.d

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset
3.145.178.240