* there is no upgrade from 2.6.X to 3.X.Y
last part of PV is used as kernel priority for u-a, but X is usually higher then Y in 3.x.x
so use all 3 parts in one bigger number
* and make it weak assignment if this scheme doesn't work for some
recipe
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
* /etc/modules-load.d/foo.conf is used by systemd like /etc/modutils/foo was with sysvinit
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
The parsing condition was broken causing the packages not being
expanded to include <PN>-systemd packages automatically.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
* append systemd functionalty moved into systemd_add_rdepends
* append base-pgk in case receipe does not set RDEPENDS for *-systemd package
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
* Add files found in SYSTEMD_SERVICE and add service files referenced by 'Also='
(and 'Conflicts=' in case of one service for links to /dev/null) recursively.
* For *.socket files the corresponding '*.service' and '*@.service are packed
* In case a file set in SYSTEMD_SERVICE does not exist, build is aborted with an
error message.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
All further manipulations need to be performed only once. This modication
avoids unwanted behaviour when inheriting rm_work.bbclass [1]
[1] http://lists.linuxtogo.org/pipermail/openembedded-devel/2012-February/038067.html
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
after recipe parsing the following plausibilities are checked:
* ERROR if empty SYSTEMD_PACKAGES
* ERROR if SYSTEMD_PACKAGES entry does not contain '-systemd' AND
package_name != ${PN}
* ERROR if SYSTEMD_PACKAGES entry contains '-systemd' but entry without
'-systemd' (base pkg) does not exist
* WARNING if SYSTEMD_PACKAGES containing ${PN} to force package maintainers to
split out systemd-specific files into *-systemd
* error if SYSTEMD_SERVICE(_<systemd-package>) is empty for one or more entries
in SYSTEMD_PACKAGES
The code in populate_packages_prepend to set SYSTEMD_PACKAGES to ${PN} as
default was removed.
Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Acked-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Packages that uses the systemd class need to have systemd build
otherwise they might fail to detect systemd as available or fail to
generate rootfs.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
* Symptom: kernel symlink in /boot is not created.
* Rename in order to create the expected files:
* /var/lib/opkg/info/kernel-3.1.4.postinst
* /var/lib/opkg/info/kernel-3.1.4.postrm
Signed-off-by: Andrea Adami <andrea.adami@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
The systemd runtime dependency needs to be add dynamically onto each
package that has the postinst generate thus removing the SYSTEMDPN use
et all.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
This generates the postinst/prerm/postrm routines for the package
specified in SYSTEMD_PACKAGES avoding code duplication and
centralizing the handling of it for easy maintainence work.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
In case of multiple GIT repositories are used, SRCREV_FORMAT will be
respected while filling it with the proper GIT information to each
revision.
This new feature needed heavy changes in the code so basically it was
a rewrote version that keeps compatibility with previous usage.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Save the kernel image name into sysroot so it can be used during image
build if need. This is used by O.S. Systems products and probably
useful to others.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
* from org.openembedded.dev
* reworked in meta-zaurus
* added missing quotes
* explain more clearly why the compiler/linker flags are reset
Signed-off-by: Andrea Adami <andrea.adami@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
For kernel's that use a split source/object build the copy
of defconfig to {S} in the base kernel class is problematic.
The previous solution for this was to override the do_configure
of the base kernel class in a subclass. While this is still
a viable/valid option, it does mean that changes to the base
do_configure will be missed.
The solution to this is to copy a defconfig to {B} which is
typically the same as {S}, so most kernel recipes won't see or
care about this change.
With this change in place, linux-yocto.bbclass can drop its
override of do_configure.
Tested with linux-yocto and oe linux recipes.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
KERNEL_LD was using ${LD} in it's definition, which is not correct for
different ABIs such as x32 or i386 on x86_64 machine. This brings it
into sync with the corresponding gcc settings, likewise the same with
the KERNEL_AR variable.
[RP: Updated commit message]
Signed-off-by: Nitin A Kamble <nitin.a.kamble@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
* The menuconfig target exists in places other than the kernel that use kernel style config.
Signed-off-by: Noor, Ahsan <noor_ahsan@mentor.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Fixes [YOCTO 1136]
linux-yocto adds some configure steps that are necessary to prepare the source
tree after the do_patch task. This causes a "-c menuconfig" to fail in a clean
build tree. Typical use of menuconfig should be to modify the config provided
by the recipe being built. It therefor makes sense for the menuconfig task to
come after the configure task. This also happens to fix the issue seen with the
linux-yocto kernel recipe.
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
The same version of this class is now in openembedded-core.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
It will now additionally create a new file which will
contain list of licenses of installed packages in a given
image.
The fix to get the package name is there because the packages
which contains epoch PE in the version string were not
being constructed correctly since opkg-cl would dump
'epoch:version' for Version string but the real ipk
did not have 'epoch' in its name. e.g.
for armv5te/libopkg0_0.1.8+svnr609-r2_armv5te.ipk
Package: libopkg0
Version: 1:0.1.8+svnr609-r2
as you can see version contained in ipk is not
containing the 'epoch'
so we have to eliminate 'epoch' from version we get from
opkg-cl to for the installed ipk name correctly.
Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
The perf component is built with the kernel, so ensure that it gets the
debug information associated with it.
Also bump the PR in the linux-yocto-* to ensure they get rebuilt.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Change to only depend on virtual/*/depmod.
Change all calls to only use depmod.
Do not install depmod as depmod-2.6
Bump PR in module-init-tools-cross.
Signed-off-by: Anders Darander <anders@chargestorm.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This removes a couple of megabytes of QA warnings!
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
do_install was putting files directly into the sysroot which means sstate
had no knowledge of them. This meant they didn't get cleaned along with the
other files from the task amongst other issues.
This patch puts them in ${D} where they were supposted to be.
Tested-by: Tom Rini <tom_rini@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
Without this, images will fail now that kernel-abiversion is back.
Signed-off-by: Tom Rini <tom_rini@mentor.com>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
This fixes external module recipes that need $KERNEL_VERSION. It got removed by
commit a9d41062e24a6b99661b3a5256f369b557433607
Author: Darren Hart <dvhart@linux.intel.com>
Date: Tue Mar 8 17:09:10 2011 -0800
kernel/bbclass: rework kernel and module classes to allow for building out-of-tree modules
seemingly as an oversight.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
- KERNEL_VERSION may contain characters unsuitable for package
names, e.g. underscores. Use legitimize_package_name to replace
those characters.
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
These are good for general use since other distros (SHR, SlugOS) are copying them into their layers nowadays. Let's stop this duplication bu moving them here.
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
When 'kernel-misc' is empty do_rootfs will fail when including 'kernel-modules'
The proper fix would be to have a whitelist for the 'kernel-modules' meta-pkg
Signed-off-by: Koen Kooi <koen@dominion.thruhere.net>
