I am writing an custom recipe for the Bitbake for an Makefile based project. We are able to create the RPMs with all the files place in the package but we are not able to find a way for pre and post install action.
As the application runs as service we want to stop it in pre-install step and then start it in post-install step.
But I am not able to find the same so any thoughts to achieve it.
Below is the sample recipe we written for it.
DESCRIPTION = "Simple helloworld application"
SECTION = "examples"
LICENSE = "MIT"
LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302"
PR = "r0"
SRC_URI = "file://helloworld.c"
DEPENDS = "boost"
S = "${WORKDIR}"
do_compile() {
${CC} helloworld.c -o helloworld
}
PACKAGES = "helloworld"
do_install() {
install -d ${D}${bindir}
install -m 0755 helloworld ${D}${bindir}
install -d ${D}${sysconfig}/init.d
install -m 0755 ${S}/service ${D}${sysconfig}/init.d
}
I do see INITSCRIPT_PACKAGES and INITSCRIPT_PARAMS but their description doesn't talk about being pre and post action.
So any thoughts for putting %pre and %post (in terms of RPM spec) for this purpose.
You can add post install scripts in your .bb:
pkg_postinst_PACKAGENAME() {
#!/bin/sh -e
# Commands to carry out
}
Reference: Section 5.3.16 http://www.yoctoproject.org/docs/1.7.1/mega-manual/mega-manual.html
According to the documentation the examples only runs during image creation time. There is also another function that will only run on the first boot (and never after it). It uses the meta/recipes-devtools/run-postinsts recipe to accomplish this.
I ran into the same issue. See this post for how I did the post install script. Hopefully you can glean from that answer enough to modify it for your script.
Related
I have a C file binary to add as custom layer in the yocto and run in qemu. I have created layer using bitbake-layers create-layer meta-custLayer and added using bitbake-layers add-layer meta-custLayer. The file tree of meta-custLayer is as:
The example_0.1.bb file has the following:
SUMMARY = "bitbake-layers recipe"
DESCRIPTION = "Recipe created by bitbake-layers"
LICENSE = "MIT"
LIC_FILES_CHKSUM = "file://${COREBASE}/meta/files/common-licenses/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302"
SRC_URI = "file://threading.c"
S = "${WORKDIR}"
TARGET_CC_ARCH += "${LDFLAGS}"
do_compile(){
$(CC) threading.c -o threads -lpthreads
}
do_install(){
install -d ${D}${bindir}
install -m 0755 threads
${D}${bindir}s
}
When the command bitbake example is executed, this is the output:
ERROR: example-0.1-r0 do_compile: Execution of '/home/sajil/edm_yocto/sources/poky/build/tmp/work/core2-64-poky-linux/example/0.1-r0/temp/run.do_compile.1806553' failed with exit code 127:
/home/sajil/edm_yocto/sources/poky/build/tmp/work/core2-64-poky-linux/example/0.1-r0/temp/run.do_compile.1806553: 99: CC: not found
/home/sajil/edm_yocto/sources/poky/build/tmp/work/core2-64-poky-linux/example/0.1-r0/temp/run.do_compile.1806553: 99: threading.c: not found
WARNING: exit code 127 from a shell command.
when bitbake core-image-minimal is executed, terminal shows that threads is unbuildable and runs into error:
ERROR: Nothing RPROVIDES 'threads' (but /home/sajil/edm_yocto/sources/poky/meta/recipes-core/images/core-image-minimal.bb RDEPENDS on or otherwise requires it)
NOTE: Runtime target 'threads' is unbuildable, removing...
Missing or unbuildable dependency chain was: ['threads']
ERROR: Required build target 'core-image-minimal' has no buildable providers.
Missing or unbuildable dependency chain was: ['core-image-minimal', 'threads']
I checked the path of the custom layer directory, being correct. I am clueless about the errors I am confronting.
For now my task is to run the C-file (adding it as a layer) on QEMU. I am unable to build the image for the same.
I would appreciate some help and insights!
There is multiple issues on your recipe:
Your do_compile command does not indicate where to find your .c file
You should use ${CC} instead of $(CC)
lpthread does not end with and s
do_compile() {
${CC} ${S}/threading.c -o ${S}/threads -lpthread
}
Your do_install does not provide the correct path to your binary:
do_install() {
install -d ${D}${bindir}
install -m 0755 ${WORKDIR}/threads ${D}${bindir}
}
At the end you should populate the packages:
FILES_${PN} = "${bindir}"
Edit about threads unbuildable:
threads is unbuildable because the recipe does not mention that the package is threads.
Here you have some options:
Rename your recipe to threads_0.1.bb (I recommend you to use a less generic name)
use the PACKAGES variable in your recipe. You will need to modify the FILES_* variable too. (a bit complicated if you are new to Yocto)
Use the current recipe name, and change the IMAGE_INSTALL += "threads" to IMAGE_INSTALL += "example"
I am a beginner in yocto. Till now i have learnt about how to build a yocto image and add recipes through layers.openembedded. But i am not able to figure out e.g. i have developed a python3-flask project in my PC and then i want to copy/transfer that project into my yocto os. how can i do that? do i have to make something like executable of that project and then copy that into my os using some recipe?
I have seen this recipe but i am not able to understand what does LIC_FILES_CHSUM means and where to get it? where should i put these file e.g. setup.py? in the same directory as my .bb file?
and on building where my project would be copied in the yocto os?
DESCRIPTION = "Simple Python setuptools hello world application"
SECTION = "examples"
LICENSE = "MIT"
LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302"
SRC_URI = "file://setup.py \
file://python-helloworld.py \
file://helloworld/__init__.py \
file://helloworld/main.py"
S = "${WORKDIR}"
inherit setuptools
do_install_append () {
install -d ${D}${bindir}
install -m 0755 python-helloworld.py ${D}${bindir}
}
The LICENSE field should obviously correspond to the license you picked for your SW. LIC_FILES_CHKSUM needs a file that holds the actual license defined in LICENSE and its md5sum passed as an "argument". It is just a way to monitor if the license ever changes. Then its md5sum changes and then the recipe fail to build because the license probably changed and requires the maintainer's attention.
THISDIR being the directory containing the recipe (not exactly but enough for the example), files in SRC_URI will be searched for in FILESPATH (which is first ${THISDIR}/<recipe_name>-<recipe_version> then ${THISDIR}/<recipe-name> then ${THISDIR}/files). So you need to put files (the ones starting with file://) declared in SRC_URI into one of those directories.
Best would actually to have your SW in a git repo somewhere. This better follows the philosophy of having SW source code separate from your build system (what if you decide you want to use a different build system later?).
When you're wondering what some variables or tasks are doing, I highly recommend looking them up on the mega manual: https://www.yoctoproject.org/docs/latest/mega-manual/mega-manual.html. This works extremely well as a "dictionary". The documentation is great albeit dense and long but it is rare not to found what you look for as a beginner.
Update
LIC_FILES_CHKSUM can be found in yocto meta/files/common-licenses in the Source Directory. You can also take your own license and check the checksum with md5sum filename and copy that to bb file. If your SW is to be distributed, it is highly recommended to have the license mentioned somewhere in your sources. So better use that part of the code in LIC_FILES_CHKSUM than using the ones in ${COMMON_LICENSE_DIR} so that people will know if and when there is a license change in your project.
The below recipe works fine and you can find your app in /usr/bin directory.
DESCRIPTION = "Simple Python setuptools application"
SECTION = "examples"
LICENSE = "MIT"
LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302"
SRC_URI = "file://setup.py \
file://packagetest.py \
file://example/__init__.py \
file://example/file1.py \
file://example/file2.py \
file://example/file3.py"
S = "${WORKDIR}"
inherit pypi setuptools3
do_install_append () {
install -d ${D}${bindir}
install -m 0755 packagetest.py ${D}${bindir}
}
For the past couple of days I have been trying to generate a viable CMake SDK with Yocto Project. I'm trying to generate SDK based on an image file which is given below:
#To build SDK, use bitbake meta-toolchain
DESCRIPTION = "Embeddev-LXDE image."
LICENSE="CLOSED"
IMAGE_INSTALL = "packagegroup-core-boot \
packagegroup-core-x11 \
packagegroup-lxde-base \
kernel-modules \
"
IMAGE_INSTALL_append = " nano git cmake qtbase qtchooser dbus packagegroup-core-ssh-openssh xterm"
#Framebuffer driver for tft
IMAGE_INSTALL_append = " xf86-video-fbdev"
IMAGE_INSTALL_append = " apt dpkg sudo tzdata glibc-utils localedef networkmanager pointercal xinit xkeyboard-config base-passwd liberation-fonts pkgconfig"
IMAGE_INSTALL_append = " wiringpi"
#Maybe consider connman instead of networkmanager
#vc-graphics is problematic with userland..
inherit populate_sdk
## SDK stuff, to build sdk use bitbake rpi-embeddev-lxde-image -c populate_sdk
# Add all static packages: SDKIMAGE_FEATURES += "staticdev-pkgs"
SDKIMAGE_FEATURES += "staticdev-pkgs"
SDKIMAGE_FEATURES += "dev-pkgs"
TOOLCHAIN_TARGET_TASK_append = " wiringpi-dev"
TOOLCHAIN_HOST_TASK_append = " nativesdk-cmake"
##
inherit distro_features_check
REQUIRED_DISTRO_FEATURES = "x11"
IMAGE_LINGUAS ?= " "
LICENSE = "MIT"
export IMAGE_BASENAME = "rpi-embeddev-lxde-image"
inherit core-image
ENABLE_SPI_BUS = "1"
ENABLE_I2C = "1"
# qtwebengine qtwebkit ...
do_image_prepend() {
}
I create my SDK with bitbake rpi-embeddev-lxde-image -c populate_sdk.
I would like to describe the exact problem. The problem is that nativesdk-cmake is not correctly installed in the SDK. Cmake 3.10.2 recipe gives:
do_install_append_class-nativesdk() {
mkdir -p ${D}${datadir}/cmake
install -m 644 ${WORKDIR}/OEToolchainConfig.cmake ${D}${datadir}/cmake/
mkdir -p ${D}${SDKPATHNATIVE}/environment-setup.d
install -m 644 ${WORKDIR}/environment.d-cmake.sh ${D}${SDKPATHNATIVE}/environment-setup.d/cmake.sh
}
FILES_${PN}_append_class-nativesdk = " ${SDKPATHNATIVE}"
FILES_${PN} += "${datadir}/cmake-${CMAKE_MAJOR_VERSION}"
FILES_${PN}-doc += "${docdir}/cmake-${CMAKE_MAJOR_VERSION}"
BBCLASSEXTEND = "nativesdk"
Tracing the root of the problem, I have seen that the cmake/ directory that should be created is created in:
/home/<user>/poky/build/tmp/work/x86_64-nativesdk-pokysdk-linux/nativesdk-cmake/3.10.2-r0/image/opt/poky/2.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/share/cmake/
However, this cmake directory is not valid in /opt/poky/2.4+snapshot/sysroots/x86_64-pokysdk-linux/usr/share/, -where it is actually needed- when I install the SDK to /opt, unfortunately.
Do I need to know anything else or do anything else regarding how to properly generate SDK?
I am really stuck here any help is much appreciated indeed.
Thanks in advance.
EDIT: I moved TOOLCHAIN_TASK statements to layer.conf and used bitbake meta-toolchain which also didn't work.
EDIT2: I used cmake version 3.6 with PREFERRED_PROVIDER_cmake = "3.6.1", which also did not work.
There seems to be a bug in the imminent Yocto Project 2.5, Sumo, release. Here, sysroots/x86_64-chargestorm-linux/usr/share/cmake/OEToolchainConfig.cmake seems to be omitted.
The temporary solution for this is to add
TOOLCHAIN_HOST_TASK += "nativesdk-cmake-dev"
And yes, using a release is always helpful, especially if your new to some project. Just remember to always use the same release branches for all your included layers. And personally, I'd not start a new project based on Morty, which was released 1.5 years ago when writing this, as it's likely to go out of official Yocto Project support rather soon.
As an aside, it seems that it is still a bug (Or would this be mis-feature at this point...?) in Sumo. Just got bit by this...workaround looks to be the same as described.
(Note: This is a release at this point in time... X-D)
Problem solved by using a release (I used "morty", and not the "master branch") for every layer and poky itself. This is apparently very important.
I like to remove some unused drivers for my RPI2 + custom board. For that I am creating an own configuration via:
bitbake linux-raspberrypi -c menuconfig
and save the new kernel preset to the file defconfig.
After this I created an append file for the linux-raspberryp recipe.
So I created the file
linux-raspberrypi%.bbappend
and filled it with:
FILESEXTRAPATHS_prepend := "${THISDIR}/linux-raspberrypi:"
SRC_URI += "file://defconfig"
PACKAGE_ARCH = "raspberrypi2"
I put the defconfig file to:
<meta-mylayer>/recipes-kernel/linux/linux-raspberrypi/raspberrypi2/defconfig
When recompiling the kernel via:
bitbake linux-raspberrypi -c clean
bitbake linux-raspberrypi
The standard RPI2 configuration is taken.
Any idea how to overcome this problem?
I am working on the "actual" pyro branch of meta-raspberrypi and yocto.
Well, unfortunately, the easiest way is probably to patch the kernel source... Or copy your defconfig over the in kernel-tree one.
The meta-raspberrypi layer does some unfortunate things in their kernel recipes, and even though this has become better with time, they're still not really nice...
If you take a look at recipes-kernel/linux/linux-raspberrypi.inc, the following lines explains the issue:
KERNEL_DEFCONFIG_raspberrypi2 ?= "bcm2709_defconfig"
do_kernel_configme_prepend() {
install -m 0644 ${S}/arch/${ARCH}/configs/${KERNEL_DEFCONFIG} ${WORKDIR}/defconfig || die "No default configuration for ${MACHINE} / ${KERNEL_DEFCONFIG} available."
}
Thus,they're copying the in-tree defconfig to ${WORKDIR}/defconfig, thereby overwriting your own defconfig.
You in you .bbappend, you could try to add:
do_kernel_configme_prepend() {
install -m 0644 ${WORKDIR}/defconfig ${S}/arch/${ARCH}/configs/${KERNEL_DEFCONFIG} || die "No default configuration for ${MACHINE} / ${KERNEL_DEFCONFIG} available."
}
Thus, first overwriting the in-kernel-tree one with your own defconfig.
Please take a look at how to use devtool to modify source code for the jethro:
http://www.yoctoproject.org/docs/2.0/dev-manual/dev-manual.html#using-devtool-in-your-workflow
I would start by having a fork at the git repository that it is using;
http://git.yoctoproject.org/cgit/cgit.cgi/meta-raspberrypi/tree/recipes-kernel/linux/linux-raspberrypi_4.9.bb
Using devtool in Yocto;
in your build directory: create a my-linux-raspberry folder;
mkdir linux-raspberry-test
devtool modify -x linux-raspberry ./my-linux-raspberry
This will put unpack the source code into my-linux-raspberry for you to modify; It also create the git repository in there;
Then, modify the code in my-linux-raspberry; To test build, run devtool build linux-raspberry; Once you are satisfied, add this git repository to your fork;
git add .
git commit -m "my-linux-raspberry"
devtool update-recipe linux-raspberry
Optional: run devtool reset linux-raspberry to remove the bbappend file;
have a bit of a problem creating a recipe for yocto. More specifically i have to install a library from git that normally installs like this:
./bootstrap
./configure --sysconfdir=/etc
make
sudo make install
My question is how can I add this to the recipe functions do_configure, do_compile, do_install. Haven't found much information or examples online.
Update 1:
This is the library that i want to integrate into yocto
https://github.com/NXPNFCLinux/linux_libnfc-nci
It's just a regular autotools based library. The main issues, that someone ought to fix, are to make the build create versioned libraries and to add a LICENSE or COPYING file.
However, a quick recipe could look like:
SUMMARY = "Linux NFC stack for NCI based NXP NFC Controllers"
HOMEPAGE = ""
LICENSE = "Apache-2.0"
LIC_FILES_CHKSUM = "file://src/include/linux_nfc_api.h;endline=17;md5=42fdb99b3ff2c12f594b22a774cb7308"
SECTION = "libs"
SRC_URI = "git://github.com/NXPNFCLinux/linux_libnfc-nci.git"
SRCREV = "118ea118cecda55c1b6a87d151a77b04515687df"
PV = "2.0+git${SRCPV}"
S = "${WORKDIR}/git"
inherit autotools
FILES_${PN} += "${libdir}/libnfc_nci_linux-1.so"
# Make sure it isn’t in the dev package’s files list
FILES_SOLIBSDEV = "${libdir}/libnfc_nci_linux.so"
A versioned library would allow us to remove the last three lines.