In case, kernel config enables compressed modules, support of
splitting via split_kernel_module_packages won't find any module.
So, first expand module pattern regex to recognize compressed
modules and then objcopy on temporary extacted to extract module
information.
(From OE-Core rev: fae400b225827400bf32380a7d599d3b2969db55)
Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fix handling of escape characters in regexs and hence fix python
Deprecation warnings which will be problematic in python 3.8.
Note that some show up as:
"""
meta/classes/package.bbclass:1293: DeprecationWarning: invalid escape sequence \.
"""
where the problem isn't on 1293 in package.bbclass but in some _prepend to a
package.bbclass function in a different file like mesa.inc, often from
do_package_split() calls.
(From OE-Core rev: 4b1c0c7d5525fc4cea9e0f02ec54e92a6fbc6199)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Before linux commit d36b691077dc59c74efec0d54ed21b86f7a2a21a, some
strings contained invalid utf-8 character which made
split_kernel_module_packages unhappy when parsing ums-isd200.ko:
Exception: UnicodeDecodeError: 'utf-8' codec can't decode byte 0xf6 in position 21: invalid start byte
(From OE-Core rev: 503b977acf6984120818cbc5cfd35ff2cffb39cd)
Signed-off-by: Pau Espin Pedrol <pespin@sysmocom.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The older style calls (plus a bashism in kernel.bbclass, fixed
separately) were introduced via the recent change to add support for
multiple kernel packages:
http://git.openembedded.org/openembedded-core/commit/?id=6c8c899849d101fd1b86aad0b8eed05c7c785924
(From OE-Core rev: e660ef68de3b3891a26ed6e10d96dc4efaf03ffc)
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some distros may want to provide alternate kernel "flavors" via feeds or
within bootable images. For example, readily available builds which
provide certain diagnostic features can enable developers and testers to
more quickly resolve issues by avoiding lengthy kernel builds.
This change allows for building multiple flavors of the kernel and
module packages by templatizing kernel package names via a new
KERNEL_PACKAGE_NAME variable in kernel.bbclass. It defaults to the old
name of "kernel", but can be overridden by certain recipes providing
alternate kernel flavors.
To maintain compatibility, recipes providing alternate kernel flavors
cannot be the "preferred provider" for virtual/kernel. This is because
OE puts the preferred provider's build and source at
"tmp-glibc/work-shared/$MACHINE/kernel-build-artifacts/" and
"tmp-glibc/work-shared/$MACHINE/kernel-source/" instead of
"tmp-glibc/work/*/$PN/" like other recipes. Therefore, recipes using the
default KERNEL_PACKAGE_NAME="kernel" follows the old semantics -- build
in the old location and may be preferred provider -- while recipes using
all other KERNEL_PACKAGE_NAME's build from the normal WORKDIR and don't
provide "virtual/kernel".
Testing:
1. Add `KERNEL_PACKAGE_NAME_pn-linux-yocto-tiny = "tiny-linux"`
to local.conf so that linux-yocto-tiny may build alongside
the main kernel (linux-yocto).
2. `bitbake linux-yocto linux-yocto-tiny` to build both kernel flavors.
3. Verified image and modules IPKs exist for both:
tmp-glibc/deploy/ipk/qemux86/kernel-* for linux-yocto
tmp-glibc/deploy/ipk/qemux86/tiny-linux* for linux-yocto-tiny
4. Verified linux-yocto is the "preferred provider", and was built in
shared directory: tmp-glibc/work-shared/qemux86/kernel-*
5. Add `CORE_IMAGE_BASE_INSTALL_append_pn-core-image-base = "tiny-linux"`
to local.conf to install both kernel flavors in core-image-base.
6. `bitbake core-image-base` to build an image.
7. Verified image contains two bzImage's under /boot/, with
"yocto-standard" (linux-yocto recipe) selected to boot via symlink.
Discussion threads:
http://lists.openembedded.org/pipermail/openembedded-core/2015-December/thread.html#114122http://lists.openembedded.org/pipermail/openembedded-core/2017-July/thread.html#139130
[YOCTO #11363]
(From OE-Core rev: 6c8c899849d101fd1b86aad0b8eed05c7c785924)
Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com>
Signed-off-by: Gratian Crisan <gratian.crisan@ni.com>
Signed-off-by: Haris Okanovic <haris.okanovic@ni.com>
Coauthored-by: Gratian Crisan <gratian.crisan@ni.com>
Coauthored-by: Haris Okanovic <haris.okanovic@ni.com>
Coauthored-by: Josh Hernstrom <josh.hernstrom@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit e0ed52c514.
Commit e0ed52c514 ('kernel-module-split: rrecommend kernel-image instead
of rdepend') changed kernel modules to rrecommend kernel-image instead
of rdepend on kernel. This broke existing setups, where the kernel is
omitted by setting RDEPEND_kernel-base = "".
Revert the patch, as the existing way of omitting kernel-image in images
works just fine.
(From OE-Core rev: c315b1e036b1252f35abda921c49327c950a9570)
Signed-off-by: Martin Hundebøll <mnhu@prevas.dk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Hard depending on the kernel makes it impossible to install kernel
modules without getting the kernel image installed too. This is
inconvenient in e.g. initramdisks, where the kernel is loaded from
outside the initramdisk.
Making the kernel modules rrecommend kernel-image-<version> instead of
rdepending on it, makes it possible to install kernel modules without
the kernel image by setting "kernel-image" in BAD_RECOMMENDATIONS.
(From OE-Core rev: 5dd7ddb66a6846d9bb59dc7833e8318992d0e645)
Signed-off-by: Martin Hundebøll <mnhu@prevas.dk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Python function subprocess.call() returns the return value of the
executed process. If return values are not checked, errors may
go unnoticed and bad things can happen.
Change all callers of subprocess.call() which do not check for
the return value to use subprocess.check_call() which raises
CalledProcessError if the subprocess returns with non-zero value.
https://docs.python.org/2/library/subprocess.html#using-the-subprocess-module
All users of the function were found with:
$ git grep "subprocess\.call" | \
egrep -v 'if.*subprocess\.call|=\ +subprocess\.call|return.*subprocess\.call'
Tested similar patch on top of yocto jethro. Only compile tested
core-image-minimal on poky master branch.
(From OE-Core rev: 578c8205fd14c48c6d30ef2889d86f1b4aee060a)
Signed-off-by: Mikko Rapeli <mikko.rapeli@bmw.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Replace hardcoded '/lib' in kernel modules installation path with
${nonarch_base_libdir}, which is meant exactly for this.
(From OE-Core rev: 22f5ba7154fcbe826d0a3283740903312b2aab46)
Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This makes it possible to restore the behaviour changed by commit 78cde87
"kernel-module-split: Append KERNEL_VERSION string to kernel module name".
(From OE-Core rev: 4aabdddf67859cd8518e5cb672fb391c859817e0)
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This is a complementary fix to commit 7c552996:
[ meta: remove True option to getVar calls ]
it intended to remove all True option to getVar calls, but there are
still some remaining.
Search made with the following regex: getVar ?\((.*), True\)
(From OE-Core rev: 87d03ffe03d6f01e360bfd51714be96e62506e0a)
Signed-off-by: Ming Liu <peter.x.liu@external.atlascopco.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The postinstall needs kmod-native and depmodwrapper-cross, mark these dependencies.
(From OE-Core rev: 0e4f5eb4f8443ed98d7c8aaf0b999c5618b7cf25)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
This reverts commit 0ccb2efe9837e2915c093341a662dffc1df00866.
The OVS fails to function and the kernel modules cannot be found by
any of the kernel tools such as depmod because they are installed
into the wrong directory in multilib 64bit/32bit bulids.
(From OE-Core rev: 85cec1e3df68e932c7b210956ef5f17b85f3616f)
Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The KERNEL_VERSION string is added to kernel module package names in order
to make the kernel modules for different kernel versions distinct packages
instead of different versions of the same package. With this change, when
a new kernel is installed together with its kernel modules (e.g. by upgrade
of the packages kernel and kernel-modules) using some package manager such
as apt-get or rpm, the kernel modules for the older kernel will not be
removed. This enables a fall back to the older kernel if the new one fails.
Also, for backwards compatibility and to enable kernel version agnostic
dependencies to kernel modules, create a virtual package with the old
(shorter) kernel module package name using RPROVIDES.
(From OE-Core rev: 78cde87bb6e71ec5b603426879267874900d09f3)
Signed-off-by: Ola Redell <ola.redell@retotech.se>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
getVar() now defaults to expanding by default, thus remove the True
option from getVar() calls with a regex search and replace.
Search made with the following regex: getVar ?\(( ?[^,()]*), True\)
(From OE-Core rev: 7c552996597faaee2fbee185b250c0ee30ea3b5f)
Signed-off-by: Joshua Lock <joshua.g.lock@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
With the recent changes, executing depmod is not needed
anymore.
This simplifies and removes a lot of unnecessary code.
(From OE-Core rev: 8296e258b36a6238605e068e13c1982b1d12fe53)
Signed-off-by: André Draszik <git@andred.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The information retrieved via depmod is incomplete with
regards to kernel modules that are dependencies, in
particular where two kernel modules are built from
different source trees / recipes, which leads to incomplete
dependency information for packages created.
So far, our packages created didn't contain dependencies on
packages created by other recipes, as we solely use depmod
for that, and depmod can only work well after *all* kernel
modules have been copied into one place - it doesn't work
well in a staged approach.
Now that all .ko have correct dependency information at packaging
time, we can use that information to properly track dependencies
across recipies, and can combine the information from the
.modinfo elf section with the information from depmod.
(From OE-Core rev: e4af1fa3aee7f1cf00ca27944b10b886f41f2fda)
Signed-off-by: André Draszik <git@andred.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When splitting kernel modules into individual packages, such packages take
their names from the module name. This is OK under most of the circumstances.
However, it may lead to package naming collisions if there exists two
modules with the same name.
Situations like this can occur when building testing modules. For instance,
there exists testing versions of the modules for non-volatile memory that
are built with different linker options but bear the same module name. If
one wants to package such modules, it is be good to be able to name
packages differently. This can be done by prefixing the package name with
a KERNEL_MODULE_PACKAGE_PREFIX that can be set by the recipes that inherit
from module.bbclass.
Cc: Megha Dey <megha.dey@intel.com>
(From OE-Core rev: 4f941e8c5ee8e95291c3beff0a2798aa13f8dfc8)
Signed-off-by: Ricardo Neri <ricardo.neri-calderon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When using KERNEL_MODULE_AUTOLOAD, autoload was empty, causing
"modprobe None" to get appended to the postinst script.
(From OE-Core rev: 2e3768075878b8fc0f7642cae89ab7b9d03fe52c)
Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
With the rpm package backend enabled, running:
bitbake <image>
bitbake virtual/kernel -c clean
bitbake <image> -c rootfs -f
results in an image with incorrect kernel module dependency information.
The problem is that the System.map and kernel-abiversion files are needed
for depmod and after the recent kernel changes, these are no longer in
sstate.
Its reasonable to require the kernel to unpack/build if you're
about to build a module against it. It is not reasonable to require this
just to build a rootfs.
Therefore stash the needed files specifically for depmod.
Also fix some STAGING_KERNEL_DIR references which were incorrect, found
whilst sorting through his change.
(From OE-Core rev: b851504dcf5e147c9efb1c7b6a4d22c1a1a87cd7)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
commit 3b3f7e785e279 [kernel: Rearrange for 1.8] began the process of
moving the kernel source and build artefacts out of sstate control and
into a shared location.
This changed triggered some workflow issues, as well as bugs related
to the kernel source containing build output, and hence being dirty and
breaking kernel rebuilds.
To solve these issues, and to make it clear that the kernel is not under
sstate control, we move the source and build outputs to:
work-shared/MACHINE/kernel-source
work-shared/MACHINE/kernel-build-artifacts
Where kernel-build-artifacts is the kernel build output and
kernel-source is kept "pristine". The build-artifacts contain everything
that is required to build external modules against the kernel source,
and includes the defconfig, the kernel-abiversion, System.map files and
output from "make scripts".
External module builds should either pass O= on the command line, or
set KBUILD_OUTPUT to point to the build-artifacts. module-base.bbclass
takes care of setting KBUILD_OUTPUT, so most existing external module
recipes are transparently adapted to the new source/build layout.
recipes that depend on the kernel source must have a depedency on the
do_shared_workdir task:
do_configure[depends] += "virtual/kernel:do_shared_workdir"
With this dependency added, the STAGING_KERNEL_DIR will be populated and
available to the rest of the build.
(From OE-Core rev: 6a1ff0e7eacef595738f2fed086986fd622ec32a)
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* new KERNEL_MODULE_AUTOLOAD syntax doesn't support modules where basename and
module name don't match (usually - and _), e.g.:
module_autoload_bq27x00_battery = "bq27x00-battery"
* sometimes it's useful to load modules in particular order and
module_autoload allowed to just list multiple modules, e.g.:
module_autoload_snd-soc-neo1973-wm8753 = "snd-soc-s3c24xx snd_soc_s3c24xx_i2s snd-soc-dfbmcs320 snd-soc-wm8753 snd-soc-neo1973-wm8753"
or
module_autoload_g_ether = "s3c2410_udc g_ether"
restore this possibility which is useful for incorrect dependencies
between modules
(From OE-Core rev: e9cd8ba3dda624615b68c601eac04427d9483f14)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The current module_autoload_* and module_conf_* variables are error
both ugly and error prone. They aren't registered in the task checksums
so changes to them aren't reflected in the build. This turns out to
be near impossible to fix with the current variable format in any
sensible way :(.
This patch replace module_autoload with the list of variables in
KERNEL_MODULE_AUTOLOAD which is a much simpler and usable API. An
error is printed if an old style variable is encountered. It should
be simple to convert to this.
module_conf_* are harder to deal with since there is data associated
with it, it isn't simply a flag. We need a list of variables that are set
in order to be able to correctly handle the task checksum so we add
KERNEL_MODULE_PROBECONF for this purpose and error if the user hasn't
added a module to it when they should have.
[YOCTO #5786]
(From OE-Core rev: 6f8b5be646be0f3e15e215907547f11d2a23d81b)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
During rootfs construction, image.bbclass will call depmod after all the
modules are installed. There's no need to run it from the postinst when
operating in offline root mode.
(From OE-Core rev: e8db81e4655ab7535db04aa3c8d7f9868ced6039)
Signed-off-by: Phil Blundell <pb@pbcl.net>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The has_key() attribute has been removed in python 3 since there is better
syntax available. Use the improved syntax.
(From OE-Core rev: 3dff13793e875ff58cc38c4a960caca9b6969843)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* it was generating invalid RDEPENDS when KERNEL_MODULES_META_PACKAGE had
RDEPENDS set already without trailing space
(From OE-Core rev: 5ff26f61b6860e56c255b2b6a2b0215be75b1db9)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
* this way we can reuse the same functionality also for external modules
including module_autoload_foo and module_conf_foo functionality
* MODULE_PACKAGES variable was removed (splited modules are now returned
by do_split_packages
* KERNEL_MODULES_META_PACKAGE is used to append all splitted packages
to RDEPENDS. In kernel.bbclass it's old "kernel-modules" in
module.bbclass it defaults to ${PN} for upgrade path from
single PN with all modules to PN depending on all new kernel-module-*
(From OE-Core rev: 51928b6b5ca0a46a9dcd754483a19af58b95fa18)
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
