Difference between revisions of "RZ-G/RZG2 yocto"

From eLinux.org
Jump to: navigation, search
(moved some content to page RZ-G/RZG2_BSP)
(moved some content to page RZ-G/RZG2_BSP)
 
(No difference)

Latest revision as of 11:46, 18 February 2021

BSP Differences Between renesas.com Download and Github Version

(moved to RZ-G/RZG2_BSP )

Online vs Offline Yocto build

The RZ/G VLP download includes a number of packages named 'OSS package xxx' that contain most of the source code that will be used during the Linux BSP build. The user can download those packages and then run an 'offline' Yocto build, which means that Yocto won't download any source code during the build. However, dealing with the 'OSS' packages is quite inconvenient - the user needs to download 10 to 15 large binary files, then merge them into one very large file, and then transfer that file to the build environment. There are also some limitations to the 'offline' build, for example some packages cannot be built that way.

For that reason we recommend doing an 'online' Yocto build, which is the most typical way those builds are done. There is no need to download the 'OSS' packages in this case. The only thing that is required is checking your local.conf file and making sure that a couple of variables are set correctly. The two variables are 'DL_DIR' - needs to be commented out, and BB_NO_NETWORK - needs to be set to 0.

#DL_DIR = "${TOPDIR}/oss_packages"BB_NO_NETWORK = "0" 
BB_NO_NETWORK = "0"

Hello World Example

Yocto has an example of how to create a simple Hello World application using the SDK that you created/installed on your machine.

https://www.yoctoproject.org/docs/2.4/sdk-manual/sdk-manual.html#sdk-working-projects

Helpful Packages to Include

These packages are not part of the default Yocto BSP build, but they might be helpful to add to your build. Simple add them to your local.conf file.

devmem2 can be used to read RZ/G registers from command line

IMAGE_INSTALL_append = " devmem2"

i2c-tools can be used to read/write to I2C devices from the command line

IMAGE_INSTALL_append = " i2c-tools"

libgpiod can be used to read/write to GPIOs from the command line and from application code

IMAGE_INSTALL_append = " libgpiod" 


Making the Yocto SDK and the '-sdk' root filesystem images smaller

When building the Yocto SDK or one of the '-sdk' images, e.g. 'core-image-weson-sdk', the default configuration includes the source code for most packages which makes the generated images install very big. Generally, you don't need this code outside of Yocto build environment, but some of the Renesas BSP recipes include it. Adding this line below will make your SDK much smaller.

# Do not include "dbg-pkgs" in SDK Installs 
SDKIMAGE_FEATURES_remove = " dbg-pkgs" 

If the SDK is not going to be used for building out-of-tree kernel modules, the kernel source code does not need to be included in it. Removing that reduces the size of the SDK significantly. The line below also removes the 'ltp' package, which is a Linux testing framework that is not needed in most cases.

IMAGE_INSTALL_remove = " kernel-devsrc ltp"  


Common Yocto Build Issues

1) Fixing build issues It often happens that some packages fail while building, e.g.:

ERROR: python3-pytorch-1.5.0-r0 do_configure: oe_runmake failed 
ERROR: python3-pytorch-1.5.0-r0 do_configure: Function failed: do_configure (log file is located at ...) 
ERROR: Logfile of failure stored in: ... 

This kind of errors can be easily fixed by cleaning up :

bitbake –c cleansstate python3-pythorch 

And retrying.

2) Fix bitbake error Invoking bitbake may result in a weird error:

OSError: Cannot initialize new instance of inotify, Errno=Too many open files (EMFILE) 

It is possible to overcome this error by giving this command:

sudo sh -c "echo 8192 > /proc/sys/fs/inotify/max_user_instances" 


Build DTB only with Yocto

If you have modified your Device Tree Source file (.dts) and just want to rebuild only that file (not the entire kernel) to create a new Device Tree Binary (.dtb) file, you can do the following steps.
Open a "development shell" for the Linux kernel:

bitbake linux-renesas -c devshell

Once in the shell, enter this command to have the kernel rebuild the Device Tree Binaries:

make dtbs 

To exit the 'devshell' window, just type exit.

exit

Even though you have rebuilt your device files inside the devshell, the output files are still located deep inside the Yocto directories (build/tmp/work/hihope_rzg2m-poky-linux/linux-renesas/xxx/xxx/xxx/xxx)
Therefore, execute the following command (outside of devshell) in order to copy them to the default "deploy" directory.

bitbake linux-renesas -fc deploy

Using Header Files and Shared Libraries in Applications

When you add a library to your Yocto build such as libgpio ( IMAGE_INSTALL_append = " libgpiod" ) and then build an SDK with it (bitbake core-image-weston-sdk -c populate_sdk), the appropriate share library file (.so) and header files (.h) will be added to the 'sysroot' directory in you toolchain.

For example:

$ find /opt/poky/2.4.3 -name gpiod.h
/opt/poky/2.4.3/sysroots/aarch64-poky-linux/usr/include/gpiod.h

$ find /opt/poky/2.4.3 -name "libgpiod.*"
/opt/poky/2.4.3/sysroots/aarch64-poky-linux/usr/lib64/libgpiod.so
/opt/poky/2.4.3/sysroots/aarch64-poky-linux/usr/lib64/libgpiod.so.0
/opt/poky/2.4.3/sysroots/aarch64-poky-linux/usr/lib64/libgpiod.so.0.3.1

Then in your application, you can add #include <gpiod.h> to have access to those library functions.

However, when you build your application with gcc, you need to: 1) Tell gcc where to find the header file.

Use the gcc option " -I " to add an include patch to your gcc command line.

"SDKTARGETSYSROOT" will be defined by the yocto SDK.

-I $SDKTARGETSYSROOT/usr/include

2) Tell gcc that the share library libgpio is required by the application using the -l argument. For libgpio, that would be "-lgpio"

For example:

aarch64-poky-linux-gcc -lgpiod -march=armv8-a -mtune=cortex-a57.cortex-a53 --sysroot=$SDKTARGETSYSROOT -I $SDKTARGETSYSROOT -O2 -pipe -g -feliminate-unused-debug-types -Wl,-O1 -Wl,--hash-style=gnu -Wl,--as-needed hello1.c -o hello1

Building Webviewer

Webviewer is the RZ/G2 lightweight GPU optimized browser we recommend for HTML5 GUIs. It is not included in the VLP by default but you need Gecko to be there as pre-requisite (see VLP release note about how to add it).

git clone https://github.com/webdino/meta-browser

in the user_work directory (or whenever the other VLP meta layers are). Change the local.conf (/user_work/build/conf) to have "firefox" replaced with "webviewer":

IMAGE_INSTALL_append = “ webviewer “ 
IMAGE_INSTALL_append = “ ttf-sazanami-gothic ttf-sazanami-mincho “ 
PACKAGECONFIG_append_pn-webviewer = “ egl “ 
PACKAGECONFIG_append_pn-webviewer = “ openmax “ 
PACKAGECONFIG_append_pn-webviewer = “ webgl “ 
PACKAGECONFIG_append_pn-webviewer = “ canvas-gpu “ 
PACKAGECONFIG_append_pn-webviewer = “ stylo “ 

Add the meta-layer to the bblayers.conf:

BBLAYERS += " ${TOPDIR}/../meta-browser " 
Optionally you can include the Gem Tanzanite demo: 
BBLAYERS += " ${TOPDIR}/../meta-gecko-embedded/meta-demo " 

And in the local.conf:

IMAGE_INSTALL_append = " gem-tanzanite " 

The file /usr/bin/gem-tanzanite.sh shall be modified as well, replacing firefox with webviewer. Then the demo can then be started by typing the command:

gem-tanzanite

Note that the demo can be navigated only by using a touch-enabled monitor (no mouse). If you know what you're doing you can hack the app.js in the js folder to turn the touch events into mouse events.

Fix the VLP64 v1.0.5-RT SDK Toolchain

( moved to RZ-G/RZG2_BSP )

Fix RZ/G2H eMMC boot for VLP64 v1.0.5-RT and v1.0.6

( moved to RZ-G/RZG2_BSP )

How to add any files in rootfs with yocto

If you have a file or files and you want this file to be present on the root file system. There are two steps for this:

1. Create a recipe
2. Add this recipe into your local.conf file

There are different ways to add new recipes to Yocto. One way is to simply create a new recipe_version.bb file within one of the existing layers used by Yocto. Another preferred method for adding recipes to the build environment is to place them within a new layer.

For this post, let's consider you want to copy a script file to bin folder of root file system. Also create your own layer to add recipe to the build.

1. Create your own layer called "meta-custom" using the bitbake-layers create-layer command.

$cd build
$bitbake-layers create-layer ../meta-custom

The command automates layer creation by setting up a subdirectory with a layer.conf configuration file, a recipes-example subdirectory that contains an example.bb recipe, a licensing file, and a README.

../meta-custom
├── conf
│   └── layer.conf
├── COPYING.MIT
├── README
└── recipes-example
    └── example
        └── example_0.1.bb

After that you can add the newly created layer to the Yocto Project environment in conf/bblayers.conf:

${TOPDIR}/../meta-custom \

There is also a command to add a new layer to bblayer.conf: bitbake-layers add-layer.

But this includes the meta layer with absolute paths, which can be avoided by adding it manually.


2. Create your recipe called 'recipes-app' folder inside your own 'meta-custom' layer


3. In 'recipes-app' folder create a new folder named as per your requirement, for example 'echo-hello-world'


4. cd into the 'echo-hello-world' and then create a folder named 'files' into that


5. copy your script file "hello.sh" into the files folder. Also add a licensing file.

#!/bin/bash
var="Hello World"
# print it 
echo "$var"

6. cd back to the 'echo-hello-world' folder and create a file named 'echo-hello-world_1.0.bb'(recipe_version.bb)

../meta-custom
├── conf
│   ├── bitbake-cookerdaemon.log
│   └── layer.conf
├── COPYING.MIT
├── README
├── recipes-app
│   └── echo-hello-world
│       ├── echo-hello-world_1.0.bb
│       └── files
│           ├── COPYING
│           └── hello.sh
└── recipes-example
    └── example
        └── example.bb

7. copy the recipe in echo-hello-world_1.0.bb

SUMMARY = "Hello World echo script"
LICENSE = "MIT"
LIC_FILES_CHKSUM = "file://COPYING;md5=a70b80ed7aa06ee4bc60901a9e98d373"

SRC_URI = " \
file://COPYING \
file://hello.sh \
"

S = "${WORKDIR}"

do_install () {
    # create the /usr/bin folder in the rootfs with default permissions
    install -d ${D}${bindir}
    # install the application into the /usr/bin folder
    install -m 0755 ${S}/hello.sh ${D}${bindir}
}
RDEPENDS_${PN} += "bash"

8. Add the following line into local.conf :

IMAGE_INSTALL_append = " echo-hello-world"


After board boot up, you can see your script file in /usr/bin as:

root@hihope-rzg2m:/usr/bin# ls | grep hello.sh 
hello.sh
root@hihope-rzg2m:/usr/bin# ./hello.sh 
Hello World