I come across a problem as follows:
there is a origin dts or dtsi files in the linux kernel dir:
arch/arm64/boot/dts/
Assume that there is a file : B.dtsi
I want to create a A file in that direction:
now arch/arm64/boot/dts/A/
i want to create dts file C.dts in A dir,then it include the file coming from dts/xx.dtsi
i can use the absolute dir path in file C.dts
include "arch/arm64/boot/dts/B.dtsi"
But is there a -i option like C compile ,the dtc tool can search the include file automatic after i set the compile option?
There is -i option when compiling dtb
use:
DTC_FLAGS ?= -i $(DCT_EXTRA_PATH)
then you should use:
#include <B.dtsi>
or
/include/ "B.dtsi"
in your dtsi files
Related
Whats wrong with the argument,
is there is no kernel_add_dts subcommand.
I get this below error whenever i try to run
$ recipetool kernel_add_dts meta-local /path/to/my.dts
recipetool: error: argument <subcommand>: invalid choice: 'kernel_add_dts' (choose from 'edit', 'create', 'newappend', 'appendfile', 'appendsrcfiles', 'appendsrcfile', 'setvar')
usage: recipetool [-d] [-q] [--color COLOR] [-h] <subcommand> ...
Use recipetool to add a new device tree to your custom layer following this syntax:
recipetool appendsrcfile -wm (MACHINE) (PATH/TO/LAYER) virtual/kernel (PATH/TO/DTS) 'arch/${ARCH}/boot/dts/(YOUR_DTS_NAME).dts'
Details:
(MACHINE): Your build machine name
(PATH/TO/LAYER): The path to the layer that you want the linux-xx_%.bbappend file with the new DTS will be created
(PATH/TO/DTS): The path to the new DTS file
(YOUR_DTS_NAME): The DTS file name
Important note:
If the default device tree name is the same as the one you are adding it is not a problem, if not, please make sure that you add it to KERNEL_DEVICETREE variable so that it will be shipped with all the DTS files in the boot partition.
KERNEL_DEVICETREE += "(NEW_DTS_NAME).dtb"
After that you can stop Uboot (if you are using Uboot) and specify the new DTS file with:
setenv fdt_file (NEW_DTS_NAME).dtb
saveenv (If you want to save it for every boot)
Please run "printenv" to make sure of the "fdt_file" variable's name.
Real run test:
recipetool appendsrcfile -wm imx8mmddr3lval /home/talel/Desktop/final_git/meta-node virtual/kernel /home/talel/Desktop/example.dts 'arch/${ARCH}/boot/dts/example.dts'
...
NOTE: Writing append file /home/talel/Desktop/final_git/meta-node/recipes-kernel/linux/linux-imx_%.bbappend
NOTE: Copying /home/talel/Desktop/example.dts to /home/talel/Desktop/final_git/meta-node/recipes-kernel/linux/linux-imx/imx8mmddr3lval/example.dts
The new bbappend file is:
$ cat /home/talel/Desktop/final_git/meta-node/recipes-kernel/linux/linux-imx_%.bbappend
SRC_URI += "file://example.dts;subdir=git/arch/${ARCH}/boot/dts"
FILESEXTRAPATHS_prepend := "${THISDIR}/${PN}:"
PACKAGE_ARCH = "${MACHINE_ARCH}"
With "virtual/kernel" it will detect what provides it (linux-imx, linux-yocto, ...) and it will create linux-imx_%.append file.
The -w flag will create "_%" as for any version number.
Solution to avoid any patch for the DTS file:
If there is patches for your Linux kernel they will fail if you are updating the DTS with new modifications that override some lines that the patch expects, so you can do it cleanly in 2 ways:
bitbake virtual/kernel -c cleansstate
bitbake virtual/kernel -c patch
Now all patches are applied, go to tmp/work/../linux-(PROVIDER)/../git and:
git add .
git commit -m "commiting old patches"
Now edit the DTS file and:
git add arch/../boot/dts/../myplatform.dts
git commit -m "changes"
git format-patch -1 -o /path/to/meta-custom/recipes-kernel/linux/files
Now add it to /path/to/meta-custom/recipes-kernel/linux/linux-(PROVIDER)_%.bbappend:
SRC_URI_append = " file://patch_file.patch"
Or, the other way is to add your final DTS after the patch is done:
SRC_URI_append = " file://myplatform.dts"
do_configure_append(){
cp ${WORKDIR}/myplatform.dts ${S}/arch/(ARCH)/boot/dts/....
}
and copy your myplatform.dts to /path/to/meta-custom/recipes-kernel/linux/files.
Now, that's your final DTS file.
Remove what recipetool added:
Actually, no undo subcommand in recipetool, you just need to delete the files that recipetool deployed, recipetool copy the file you specified and create a bbappend file, so remove those two files.
Example: you used recipetool to add example.dts file, recipetool copied example.dts to:
meta-custom/recipes-kernel/linux/(MACHINE)/example.dts
and created bbappend file in which it added example.dts to SRC_URI variable.
If you need to keep the bbappend file because you are using it in other way, just modify it and remove the line added by recipetool which contains:
SRC_URI ... "file://example.dts ..."
I am new, yocto build at imx6q embedded system.
I want to overwrite linux system files after do_rootfs. For example, target system files are below.
/etc/network/interface
/etc/issue
/etc/init.d/rcS
/home/root/mytest.sh
so, i made custom layer and custom recipe.
helloworld binary is copy ok.
but, do_mytask function is not called.
what's wrong with my code?
or any other method for my purpose.
#
# This file was derived from the 'Hello World!' example recipe in the
# Yocto Project Development Manual.
#
SUMMARY = "Simple helloworld application"
SECTION = "examples"
LICENSE = "MIT"
LIC_FILES_CHKSUM = "file://${COMMON_LICENSE_DIR}/MIT;md5=0835ade698e0bcf8506ecda2f7b4f302"
SRC_URI = "file://helloworld.c"
S = "${WORKDIR}"
do_compile() {
${CC} helloworld.c -o helloworld
}
do_install() {
install -d ${D}${bindir}
install -m 0755 helloworld ${D}${bindir}
}
FILESEXTRAPATHS_prepend := "${THISDIR}/files:"
SRC_URI += " \
file://interfaces \
file://issue \
file://mytest.sh \
"
addtask mytask after do_rootfs before do_image
do_mytask() {
install -d ${D}/etc/network
cp -af ${WORKDIR}/interfaces ${D}/etc/network/interfaces
cp -af ${WORKDIR}/issue ${D}/etc/issue
}
You'll need to extend the recipes that provide the files you want to replace.
Using /etc/network/interfaces as an example, the first step is to figure out which recipe installs that file.
From the bitbake prompt:
$ oe-pkgdata-util find-path /etc/network/interfaces
init-ifupdown: /etc/network/interfaces
So this tells us that /etc/network/interfaces is installed by the init-ifupdown receipe.
A file search shows that init-ifupdown is part of poky:
$ find . -name init-ifupdown*.bb
./poky/meta/recipes-core/init-ifupdown/init-ifupdown_1.0.bb
Now, since you need to modify the output of init-ifupdown, you'll need to extend init-ifupdown by creating a similarly named .bbappend in your own layer.
You might create the new .bbappend at
my-layer/receipes-core/init-ifupdown/init-ifupdown_%.bbappend
The % is a wildcard that ensures the .bbappend will apply to all future versions of the init-ifupdown recipe, which is probably what you want.
Place your custom interfaces file in a folder below the .bbappend:
my-layer/receipes-core/init-ifupdown/files/interfaces
The .bbappend then only needs to contain a single line to enable bitbake to pick up the new interfaces file:
FILESEXTRAPATHS_prepend := "${THISDIR}/files:"
Finally, repeat the above with each system file you'd like to replace.
It depends on the file to modify. For example, if you search 'interfaces' in poky directories, you'll find it in 'meta/recipes-core/init-ifupdown/init-ifupdown-${PV}/'. You just need to create a recipe named init-ifupdown-${PV}.bbappend in your meta, recreating the path seen in poky (recipes-core/init-ifupdown/). This recipe can contain a single line :
FILESEXTRAPATHS_prepend := "${THISDIR}/files:"
Then you create a 'files' folder with the 'interfaces' file you want to have.
For 'issue', like others found in the /etc directory (profile, fstab, ...), it's the same procedure, with the sources in poky/meta/recipes-core/base-files/.
For init.d scripts, use the 'update-rc' class.
You recipe is not "image recipe" (and it shouldn't be for hello world) thus you cannot use tasks do_rootfs and do_image in this case. A bit of clarification: image recipe is .bb file that you use to build image with bitbake or devtool (in your case some containing imx6q, you can find them with bitbake-layers show-recipes "*-image-*").
It looks like you are looking really is a way to override do_install of some recipe that installs that mentioned files. Then find what recipe installs those files and create bbappend file in your top layer. This bbappend file may contain do_install_append task where you can place your install <file> <dir> lines (note, using cp as not recommended, everything should be done with install tool).
Adding an extra comment based on Carsten Hansen original answer for folks working with Xilinx/Petalinux.
Under Petalinux environment we don't really have the command: oe-pkgdata-util, so the strategy is to do a search in the Xilinx SDK folder. You might have it installed on Linux under /opt according to the documentation. If you do a:
grep -r syslog-startup.conf .
you will see busybox recipe being the one that does the installation of the syslog-startup.conf.
You can create the override recipe called busybox_%.bbappend under:
../project-spec/meta-user/recipes-core/busybox/
Put the modified syslog-startup.conf file under:
../project-spec/meta-user/recipes-core/busybox/files/syslog-startup.conf
Rebuild via petalinux-build. You can also force the creation of the rootfs via petalinux-build -c rootfs and the system should populate your new file.
I have a pretty big file structure of a project which I need to convert into a multiplatform cmake project. Now it seams that cmake requires ever single cpp file be added individually to the executable. But is there a script that automates this? That snoopes through the file structure and just adds every source file automatically? Since the project will probably get a lot more source files and I probably wont be able to manually add every single one.
You could use execute_process() with a cmake -P script that uses globbing to recursively scan for source files which writes to an included file in your CMakeLists.txt i.e. something like:
"CMakeLists.txt":
execute_process(COMMAND ${CMAKE_COMMAND}
-D "RDIR=${CMAKE_CURRENT_SOURCE_DIR}"
-P "scansources.cmake"
WORKING_DIRECTORY "${CMAKE_CURRENT_SOURCE_DIR}")
include("sources.cmake")
add_executable(myexe ${sources})
"scansources.cmake" (generates "sources.cmake"):
file(GLOB_RECURSE sourcelist
*.c
*.cc
*.cpp
*.cxx)
string(REGEX REPLACE "${RDIR}/" "" relative_sources "${sourcelist}")
string(REPLACE ";" "\n" sources_string "${relative_sources}")
set(sources_string "set(sources\n${sources_string})")
file(WRITE sources.cmake "${sources_string}")
The reason why this works is because execute_process() occurs at configure time.
You could, of course, generate sources.cmake via some other tool or IDE then you wouldn't need scansources.cmake or execute_process().
tl;dr version:
Is it possible with CMake (>= 2.8) to generate zip files from some files and put the packed zip file in a specific location?
longer version:
I have a CMakeLists.txt that builds my project into a .exe file, and this exe file will read data from a zip file. The content to be packed in the zip file is in my git repository so that it can be edited, too. But, the program needs this data in a zip file. So it would be good if the CMake script could take the data, put it in a zip file, and place it next to the exe. I already heard of CPack, but I did not find any easy examples and am not sure if this is even the right tool for my task.
Is this possible? If yes, how?
Since version 3.2 CMake has the functionality to generate a zip file built-in. The CMake command-line mode sub-command tar supports both the creation of zip and 7zip archives.
For example, if the current CMake source directory contains the file testfile.txt and the directory testdir, you can use the following CMake commands to create a zip file containing both items:
add_custom_target(create_zip COMMAND
${CMAKE_COMMAND} -E tar "cfv" "archive.zip" --format=zip
"${CMAKE_CURRENT_SOURCE_DIR}/testfile.txt"
"${CMAKE_CURRENT_SOURCE_DIR}/testdir")
As a work-around for earlier CMake versions, you can use the jar command that is part of a standard Java JRE installation.
find_package(Java)
execute_process(
COMMAND
"${Java_JAR_EXECUTABLE}" "cfM" "archive.zip"
"-C" "${CMAKE_CURRENT_SOURCE_DIR}" "testfile.txt"
"-C" "${CMAKE_CURRENT_SOURCE_DIR}" "testdir"
RESULT_VARIABLE _result
)
The zip file will be generated in the current CMake binary dir (CMAKE_CURRENT_BINARY_DIR).
It's never late to show real answer:
function(create_zip output_file input_files working_dir)
add_custom_command(
COMMAND ${CMAKE_COMMAND} -E tar "cf" "${output_file}" --format=zip -- ${input_files}
WORKING_DIRECTORY "${working_dir}"
OUTPUT "${output_file}"
DEPENDS ${input_files}
COMMENT "Zipping to ${output_file}."
)
endfunction()
Use like
file(GLOB ZIP_FILES "${CMAKE_CURRENT_SOURCE_DIR}/zip/*")
create_zip("${CMAKE_CURRENT_BINARY_DIR}/native_data.zip" "${ZIP_FILES}" "${CMAKE_CURRENT_SOURCE_DIR}/zip")
This will pack all files from zip/ subdirectory into native_data.zip (in build directory). Then either include your archive (path will differ in different CMakeLists.txt!) as source file or add it as target:
add_custom_target("project-data" ALL DEPENDS "${CMAKE_CURRENT_BINARY_DIR}/native_data.zip")
Install will not differ a lot from usual:
install(FILES "${CMAKE_CURRENT_BINARY_DIR}/native_data.zip" DESTINATION ${DATADIR} RENAME "data000.zip") # Install our zip (as data000.zip)
I assume you already have a zip-tool installed (WinZip or 7z, etc.). You could write a find_zip-tool script which will search for WinZip, or 7Z, etc...
Snippet for WinZip:
FIND_PROGRAM(ZIP_EXECUTABLE wzzip PATHS "$ENV{ProgramFiles}/WinZip")
IF(ZIP_EXECUTABLE)
SET(ZIP_COMMAND "\"${ZIP_EXECUTABLE}\" -P \"<ARCHIVE>\" #<FILELIST>")
ENDIF(ZIP_EXECUTABLE)
Snippet for 7-zip:
FIND_PROGRAM(ZIP_EXECUTABLE 7z PATHS "$ENV{ProgramFiles}/7-Zip")
IF(ZIP_EXECUTABLE)
SET(ZIP_COMMAND "\"${ZIP_EXECUTABLE}\" a -tzip \"<ARCHIVE>\" #<FILELIST>")
ENDIF(ZIP_EXECUTABLE)
Take a look at the file
<cmake-install-dir>\share\cmake-2.8\Modules\CPackZIP.cmake
it shows how CPack searches for a Zip_Executable and prepares some "useful" default flags.
After that, I would suggest to execute_process, similar to sakra's answer
As of version 3.18, CMake now directly supports creating zip or archive files using the file() command with ARCHIVE_CREATE:
file(ARCHIVE_CREATE OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/MyData.zip
PATHS ${CMAKE_CURRENT_SOURCE_DIR}/data
FORMAT zip
)
Be sure to specify a full path for the OUTPUT zipped filename, or the file may not be generated. Also, the PATHS option accepts files or directories to be placed in the zip file, but it does not accept wildcards at the time of writing.
This command supports several archive formats and compression flavors. So, you can use the same command to create tarballs as well:
file(ARCHIVE_CREATE OUTPUT ${CMAKE_CURRENT_BINARY_DIR}/MyData.tar.gz
PATHS ${CMAKE_CURRENT_SOURCE_DIR}/data
FORMAT gnutar
COMPRESSION GZip
)
Since this is the top search result for creating zip files with CMake, here is a CPack solution for completeness. The basic idea is that you make calls to install() and then tell it what to name the resulting zip file. It will be placed in the build directory, though there may be a way to change that. Then you can create the zip file with make package or cpack.
# Version 1: Subtractive
# Include everything in the project source directory.
# Put it at the top level of the zip via `DESTINATION .`
# Subtract things we don't want.
# The trailing slash after "${PROJECT_SOURCE_DIR}/" prevents
# an extra layer of directories.
install(DIRECTORY "${PROJECT_SOURCE_DIR}/"
DESTINATION .
PATTERN ".git*" EXCLUDE
PATTERN ".DS_Store" EXCLUDE
PATTERN "examples" EXCLUDE
PATTERN "docs" EXCLUDE
PATTERN "README.md" EXCLUDE
)
# Version 2: Additive
# Include only the list of things we specify.
# Put it at the top level of the zip via `DESTINATION .`
# install(FILES
# ${SRCS}
# "Notes.txt"
# DESTINATION .
# )
# Tell CPack to create a zip file.
set(CPACK_GENERATOR "ZIP")
# Tell CPack what to name the zip file. It will append `.zip`.
set(CPACK_PACKAGE_FILE_NAME "${CMAKE_PROJECT_NAME}")
# Tell CPack not to put everything inside an enclosing directory.
set(CPACK_INCLUDE_TOPLEVEL_DIRECTORY OFF)
# Apparently this should be always on but isn't for backwards compatibility.
set(CPACK_VERBATIM_VARIABLES YES)
include(CPack)
Essentially what I did was create custom target
add_custom_target(STAGE_FILES)
With this target I copy the files and directories to the CMAKE_CURRENT_BINARY_DIR
add_custom_command(
TARGET STAGE_FILES
COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_SOURCE_DIR}/assets/video ${CMAKE_CURRENT_BINARY_DIR}/video
COMMAND ${CMAKE_COMMAND} -E copy_directory ${CMAKE_SOURCE_DIR}/assets/data ${CMAKE_CURRENT_BINARY_DIR}/data
COMMAND ${CMAKE_COMMAND} -E copy ${CMAKE_SOURCE_DIR}/assets/strings_en.csv ${CMAKE_CURRENT_BINARY_DIR}/
COMMAND ${CMAKE_COMMAND} -E copy ${CMAKE_SOURCE_DIR}/assets/strings_rules_en.csv ${CMAKE_CURRENT_BINARY_DIR}/
COMMAND ${CMAKE_COMMAND} -E tar "cfv" "data.zip" --format=zip --files-from=${CMAKE_SOURCE_DIR}/assets/to_zip.txt
COMMAND ${CMAKE_COMMAND} -E remove_directory ${CMAKE_CURRENT_BINARY_DIR}/data
COMMAND ${CMAKE_COMMAND} -E rename ${CMAKE_CURRENT_BINARY_DIR}/data.zip ${CMAKE_CURRENT_BINARY_DIR}/data
)
The important line
COMMAND ${CMAKE_COMMAND} -E tar "cfv" "data.zip" --format=zip --files-from=${CMAKE_SOURCE_DIR}/assets/to_zip.txt
inside my
to_zip.txt
I specify all the files I want to include in my zip
data/
video/
...
I can now execute the command
make STAGE_FILES
which will copy and zip everything i need
As you know the getch() and getche() functions don't work with the cygwin, a linux oriented one.
But can I include the conio.h header file of borland c and call the functions getch in my makefiles?
Will it work and can anyone tell me how to include the header files from different directories in cywgin.
I have a header file strcal.h in directory c:/makk/string/.
How do I include that header file in my makefile?
gcc -I/string small.c
It is not working and my current directory is makk.
In stdio.h, there is a getchar() function which is what you need. You can't just bring across the Borland header file since that just declares the function, it doesn't define it. Standard C has no need for getch().
To include header files in different areas, you use the -I directives of gcc to set up search paths.
So, if you have a /xyz/myheader.h file, you can do something like:
gcc -I /xyz myprogram.c
To get at c:/makk/string/strcal.h, you may have to use gcc -I /cygdrive/c/makk/string or, if you know you're actually in that makk directory, you can use -I string (note the lack of leading / since you want a relative path, not an absolute one).