| Age | Commit message (Collapse) | Author | Files |
|---|
|
Ross reported the following waring when building edgerouter BSP:
WARNING: Section not found: .comment
The reason is that the testing of the existing sections in do_strip()
returned the wrong value. Please see the following code in do_strip():
for str in ${KERNEL_IMAGE_STRIP_EXTRA_SECTIONS}; do {
if [ "$headers" != *"$str"* ]; then
bbwarn "Section not found: $str";
fi
"$CROSS_COMPILE"strip -s -R $str ${KERNEL_OUTPUT}
}; done
The "*" doesn't have special meaning in the if string test, so it will
return true even the $str is a substring of $headers. Fix this issue
by replacing it with "! (echo "$headers" | grep -q "^$str$")".
Reported-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Kevin Hao <kexin.hao@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
In the same way it's done for openssh.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
regardless of init system in use
Previously it was done only if sysvinit was in DISTRO_FEATURES.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The MACHINE variable is used to handle sysroot paths within one of the
patches to python3-native. In this context, it is relocation safe and
the resulting packages should not have MACHINE specific checksums,
therefore excluding MACHINE in this context is safe.
This whole setup is ugly and ideally we should come up with a better
way of handling this but at least allow a stop gap solution for now.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This adds the markup to inform bitbake to rerun the deploy task per MACHINE.
This makes sense since DEPLOY_DIR_IMAGE contains MACHINE. Doing it this
way means the same artefact will be reused from sstate rather than rebuilt.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
multilib task signatures turned out to have issues since
SIGGEN_EXCLUDERECIPES_ABISAFE and SIGGEN_EXCLUDE_SAFE_RECIPE_DEP
did not have multilib mappings. This adds those mappings in which
in turn improves multilib task checksums to match the standard
non-mulitlib versions.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Just like we have in oe-selftest, you can add
<filename>.<Class>.<testname> in TEST_SUITES in order to
run just that test
Signed-off-by: Lucian Musat <george.l.musat@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If building for a BE8 ARM target, arch is "armeb" rather than
"arm", but ABI should still be "gnueabi". Otherwise gcc won't
build.
Signed-off-by: Peter Seebach <peter.seebach@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When looking at and trying to build a slightly customised SDK, I
realised the code could be simplified and written in a way which was
more customisable.
This patch moves various function calls into the SDK_POSTPROCESS_COMMAND
which was intended for this kind of use.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
In the json file the whole datastore is serialized which contains
absolute path to the oeqa folder. This breaks the functionality
when trying to run the tests from other machines.
Signed-off-by: Lucian Musat <george.l.musat@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
To avoid races between the sstate tasks/hooks using ${B} as the cwd, and other
tasks such as cmake_do_configure which deletes and re-creates ${B}, ensure that
all sstate hooks are run in the right directory, and run the prefunc/postfunc in WORKDIR.
(From OE-Core rev: 1d3bde02641f4b40030cf7e305ee3d7c2faabe29)
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
When image-vm.bbclass was introduced, it indirectly also introduced a
".hdddirect" image type based on boot-directdisk.bbclass. However, one
could only get that image when also enabling at least one of the
virtual machine images.
The .hdddirect images are useful by themselves. By registering
image-vm.bbclass as implementation of it, it becomes possible to
select them with:
IMAGE_FSTYPES = "hdddirect"
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
* for some reason this part of:
http://patchwork.openembedded.org/patch/102561/
wasn't ever merged.
[YOCTO #7669]
Signed-off-by: Alex Franco <alejandro.franco@linux.intel.com>
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
When building something like lib32-core-image-minimal -c populate_sdk, we
expect one sysroot with both multilibs installed. We therefore
need a single SDKTARGETSYSROOT value which doesn't change when multilibs
are enabled.
This makes the image generation code match what the meta-environment
files set the SDK up to use.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
In normal use gdk-pixbuf knows where to find its own loaders and explicitly
stating this will cause work in the future if the ABI version changes.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
If for example gdk-pixbuf and lib32-gdk-pixbuf are in an image then only one
${bindir}/gdk-pixbuf-query-loaders will be installed, so only one variant will
actually be usable.
Solve this by moving gdk-pixbuf-query-loaders into ${libdir} as it's intimately
tied to the library and rarely directly invoked by the user, and update the
callers to use the right path.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
I believe this code dates from previous times when we didn't extend
the TOOLCHAIN_TARGET* variables to cover all multilibs. We now do
this so this code acutally breaks things by removing the non-multilib
variants.
By changing this, a multilib SDK now contains both sets of base libraries
which matches the tools we ship with it. If the user wishes to customise,
this also becomes easier.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
stamp-base was only ever used by the shared workdir code in gcc. This
turned out to be problematic and has been replaced by other approaches
which don't need specialist bitbake knowledge.
stamp-base will likely get removed from bitbake but for now, remove it
from archiver to simplfy the code since gcc no longer uses it.
This stops people getting confused by the obsolete code paths which I'm
getting a lot of questions about.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Make package_qa_clean_path() return something like "work/path/to/file"
rather than "/work/path/to/file", the relative path is a little clear.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
The value of the deps varflag for tasks is (unusually) expected to be a
list object rather than a space-separated string, hence we cannot use
appendVarFlag() here. This fixes a traceback when parsing the gcc recipe
with externalsrc enabled, for example.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It turns out that there are folks out there who use externalsrc in
normal builds and don't really need to be warned; additionally within
the extensible SDK or when using devtool, it shouldn't be a warning
situation. Thus, scale it back to a note (we can't use bb.note() here
since that wouldn't actually be piped through to the bitbake UI). Also
touch up the message a little bit.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This is not actually used for anything - I thought that we would need to
use it within devtool to set global configuration, but we're able to do
everything we need within the bbappends it creates, which also saves on
parse time. If we're not going to use work-config.inc let's just drop it
completely.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Provide the ability to define a function containing extra configuration
values to be added to the local.conf file that goes into the SDK. For
example, this could be used to set up SSTATE_MIRRORS within the SDK.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
installer with sh
If you ran the extensible SDK installer file with sh (instead of bash),
then the additional call to buildtools environment setup, extension of
PATH to support running devtool, and setting of OE_SKIP_SDK_CHECK
weren't being added to the end of the script. This is because apparently
bash is happy to expand wildcards in the target of a redirection, but
bash running in POSIX sh mode won't (although it apparently does work on
the sh command line rather than within a script run as an argument to
sh). In any case using a wildcard here is a bit of a crutch which we
don't need, so replace it with the proper path to the environment setup
script.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Fixes b6e64de541b37 "Restore compatibility with previous UBI filesystems"
The multivolume UBI code creates symlinks for each volume. If the volume name
is empty, it will create a symlink that the rootfs code will attempt to
create again later, resulting in a crash like this (unless IMAGE_LINK_NAME
is blank):
ERROR: Error executing a python function in .../recipes-core/images/my-image.bb:
File: '.../oe-core/meta/lib/oe/image.py', lineno: 203, function: _create_symlinks
*** 0203: os.symlink(src, dst)
Exception: OSError: [Errno 17] File exists
To prevent this from happening, only create symlinks to volumes that have
a name, and let the rootfs script create the default symlink later.
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Add nios2 target to valid arch list definition.
Based on previous work by Walter Goossens <waltergoossens@home.nl>
at https://github.com/wgoossens/meta-nios2 .
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Walter Goossens <waltergoossens@home.nl>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Add support information for the Altera NIOS-II soft processor.
Based on previous work by Walter Goossens <waltergoossens@home.nl>
at https://github.com/wgoossens/meta-nios2 .
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Walter Goossens <waltergoossens@home.nl>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The logic is supposed to avoid following dependencies when we depend on
a target recipe which depends on a native recipe. The problem is, we were
marking the dep (the native recipe) as already processed when we avoided
traversal, meaning that even when that recipe would be pulled in via
a different dependency, we skipped it there too, and whether it was skipped
entirely depended on the non-deterministic dep processing order. If the first
one to be encountered was via the indirect target dep, it wouldn't end up in
configuredeps, otherwise it would.
As we want to avoid traversing that particular dependency relationship, not
*every* dependency on the native, we should continue, but not add it to done,
so it can be traversed from other avenues.
This fixes an intermittent bug in some of my non-GPLv3 builds, where one
dependency upon gettext-minimal-native was skipped, but others should not have
been, resulting in it being removed from configuredeps entirely, and no
gettext macros being available.
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The SIGTERM introduced another break into runexported test,
the handler is in the class TestContext while exporting it
will try to convert the singnal handler to JSON, and this
will throw an exception.
The SIGTERM handler is not needed in the exported test, so this
change will remove the handler before exporting the test. Also
when a test is exported the target is not deployed/started
anymore, it is not necessary.
[YOCTO #8239]
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
* it was causing QA checck name to be shown on separate line like this:
sblim-sfcCommon-1.0.1: sblim-sfcCommon-dev doesn't match the [a-z0-9.+-]+ regex
[pkgname]
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
* show PN as other QA checks
* strip PKGDEST prefix from shown path as other QA checks
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
* there is PN at the beginning, then possibly long list of files and
at the end we don't see which recipe has this issue, add another
line which says which PN and how many files
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Sometimes an "image" may not actually have a manifest file such as when
using a "baremetal kernel". This change allows for a user to set a
IMAGE_NO_MANIFEST flag to 1 in order to inform the code that there is no
corresponding manifest that should exist for an image.
Signed-off-by: Randy Witt <randy.e.witt@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
${B} is the default cwd of tasks, so there might be race issues such as:
| mkdir: cannot create directory `${B}': File exists
[snip]
NOTE: recipe perf-1.0-r9: task do_configure: Failed
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
We want do_rootfs to rerun if the fstype or compression commands or
dependencies change for any of our configured fstypes (IMAGE_FSTYPES).
IMAGE_TYPEDEP isn't explicitly handled, as it's traversed already, so the end
result will change if it does, and we only really care about the results, not
how we got there. This uses oe.image.Image()._get_image_types() to get the
info about the image and compression types in use.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
When installing backport modules they stumble upon each other, complaining with
the following error:
ERROR: The recipe python-backports-ssl is trying to install files into a shared
area when those files already exist.
This is the correct behavior since thats just the way they were designed, all backport
modules provide an __init__.py file (the same among all packages), and without it they
simply wouldnt work.
distutils handles FILES_${PN}* variables for python packages, but it uses wildcards
to include the required files, hence removing the __init__.py files from each backport
package during build time is impossible since it doenst actually contain that value,
this patch simply removes the __init__.py* files from the staging area if they already
exist on sysroot, this way, these are not included in FILES_${PN} anymore, fixing the
issue mentioned above.
[YOCTO #8207]
Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Support for multiple ubi images has broken dozens of machine deployment scripts
in two ways:
Previously, ubi filesystems would be named $IMAGE_NAME as one would expect. The
current version would append "_rootfs" to that name for no reason. Fix it so that
the name for ubi images remains unchanged if there is only one image to build.
Machines would append to IMAGE_CMD_ubi, adding extra image processing of their
own. This is broken now that IMAGE_CMD_ubi became a variable instead of a function.
Make IMAGE_CMD_ubi a function again, this also makes for more logical quotes (I
was surprised to find that " within " would even work).
Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
On qemumips, some tests are slow to the point of the autobuilder appearing
hung. We therefore specifically exclude buildsudoku but to do this, we need
to list the tests we expect to work on core-image-lsb-sdk.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
aarch64 sets OLDEST_KERNEL to 3.14. This stops the aarch64 SDK installing on
anything with an older kernel which is clearly incorrect.
I attempted to extract the correct non-overridden version from the data store
but it proved problematic and I was running into data store issues. Those
are a separate problem but there isn't time to fix this right now.
Instead just code the SDK kernel version separately to work around this for
now (and fix the autobuilder tests and SDK usage).
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
Currently if the SDK fails to extract, no error is shown and the test is marked
as passed! Clearly this is incorrect, fix it to correctly raise an error.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
This reverts commit 56c5fa106eeccda2ca92dbeb73ff01ba40992e7d.
(This fix was incorrect and there is a better fix due to be merged)
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
|
It is useful in a few different contexts to see which files have been
written out by an sstate task; break out a function that lets us get the
path to the manifest file easily.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
Only fetch remote signatures if verification has been enabled, as otherwise the
fetcher throws errors that sstate.bbclass can't ignore.
[ YOCTO #8265 ]
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
This reverts commit 8ca8e2e5bf4a9f01dc48300149a8e1d71d715400.
Further testing showed the kernel does not support root=UUID; it is
something typically handled by the initramfs.
Because boot-directdisk.bbclass cannot know whether it is using a
suitable initramfs, root=UUID cannot be the default. Instead, it will
have to be set in image recipes on a case-by-case basis.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
As most upstreams don't have installable test suites it's fairly common to copy
files directly out of a source tree for ptests, but this results in files in the
recipe being owned by the user running bitbake:
WARNING: QA Issue: .../sed/4.2.2-r0/packages-split/sed-ptest/usr/lib64/
sed/ptest/testsuite/bug-regex21 is owned by uid 1000, which is the same
as the user running bitbake. This may be due to host contamination
[host-user-contaminated]
Instead of needing to fix this in every recipe that has this problem simply
chown the files to root:root in do_install_ptest_base.
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
provide the nfs-client IMAGE_FEATURES, to ease a user to only
install nfs client related files to image
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
We really do not want our (external) source tree to be removed. There
might be multiple values in the 'cleandirs' varflag pointing to our
source tree - causing it to be wiped out. This patch improves the
filtering of 'cleandirs' by examining the expanded values inside it. Any
(expanded) values pointing to our source tree get removed.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The sstate URI is relative to begin with, so it's completely valid to want to
alter it in a way that keeps it relative using a mirror, and I have real world
cases where this is causing issues. Remove the warning.
Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
|
|
The do_install_ptest_base task wasn't running in the fakeroot environment so
files installed by do_install_ptest were often not owned by root:root but the
user running the builds.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
|
