Paul Eggleton [Tue, 23 Dec 2014 16:01:55 +0000 (16:01 +0000)]
oeqa/selftest: fix test_force_task so it doesn't taint the entire build
Using -C sets a taint on the specified task, and m4-native is in the
dependency chain for just about everything, which means that everything
gets forced to rebuild. We don't need the influence of this test to
extend outside of the test itself, so ensure that the taint gets cleared
at the end of the test.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Paul Eggleton [Tue, 23 Dec 2014 15:19:03 +0000 (15:19 +0000)]
oeqa/selftest: improve failure messages for devtool tests
assertTrue prints "False is not True" if it fails, which is pretty much
useless. Use a more appropriate assertion test where practical and add a
message where it isn't.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Paul Eggleton [Tue, 23 Dec 2014 14:47:45 +0000 (14:47 +0000)]
oeqa/utils: fix testcase decorator to allow calling tests individually
Without this, running tests individually failed with a traceback:
$ oe-selftest --run-test
buildoptions.ImageOptionsTests.test_incremental_image_generation
2014-12-23 14:40:37,636 - selftest - INFO - Checking that everything is in order before running the tests
2014-12-23 14:40:38,408 - selftest - INFO - Running bitbake -p
2014-12-23 14:40:40,235 - selftest - INFO - Loading tests from:
oeqa.selftest.buildoptions.ImageOptionsTests.test_incremental_image_generation
Traceback (most recent call last):
File "/home/user/poky/scripts/oe-selftest", line 179, in <module>
ret = main()
File "/home/user/poky/scripts/oe-selftest", line 164, in main
suite.addTests(loader.loadTestsFromName(test))
File "/usr/lib64/python2.7/unittest/loader.py", line 91, in loadTestsFromName
module = __import__('.'.join(parts_copy))
File "/home/user/poky/meta/lib/oeqa/selftest/buildoptions.py", line 12, in <module>
class ImageOptionsTests(oeSelfTest):
File "/home/user/poky/meta/lib/oeqa/selftest/buildoptions.py", line 14, in ImageOptionsTests
@testcase(761)
NameError: global name 'func' is not defined
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Andreas Müller [Wed, 3 Dec 2014 12:40:54 +0000 (13:40 +0100)]
pythonnative: set PYTHON_EXECUTABLE so that cmake can find it
| CMake Error at /home/a.mueller/tmp/oe-core-glibc/sysroots/x86_64-linux/usr/share/cmake-2.8/Modules/FindPackageHandleStandardArgs.cmake:108 (message):
| Could NOT find PythonInterp (missing: PYTHON_EXECUTABLE)
[RP: Added comment that is used by cmake] Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Maciej Borzecki [Mon, 22 Dec 2014 14:28:02 +0000 (15:28 +0100)]
wic: account for mmcblk device partition naming
MMC block device partitions are named differently than other block
devices and use the scheme: mmcblk<devnum>p<partnum>, ex: mmcblk0p1,
mmcblk0p2. The current code generates incorrect parition names missing
'p' infix for fstab entries. The patch resolves this problem.
Signed-off-by: Maciej Borzecki <maciej.borzecki@open-rnd.pl> Signed-off-by: Maciek Borzecki <maciek.borzecki@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Chen Qi [Tue, 21 Oct 2014 06:30:36 +0000 (14:30 +0800)]
image.bbclass: avoid boot error on read-only systemd image
New version of systemd implements a new feature of updating /etc
or /var when needed at boot. For details, please see link below.
Opointer.de/blog/projects/stateless.html
For now, at boot time, the systemd-sysusers.service would update user
database files (/etc/passwd, /etc/group, etc.) according to the configuration
files under /usr/lib/sysusers.d. This step is necessary for other systemd
services to work correctly. Examples of such services are systemd-resolved
and systemd-tmpfiles-setup.
The problem is that on a read-only file system, that is, if /etc is read-only,
the user database files could not be updated, causing failures of services.
This patch fixes this problem by adding users/groups at rootfs time.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fixed:
* The os.link() reqiures write permission on the src file (suppose the
src file belongs to another user, then you need write permission to harlink to
it since the link count would change)
* Print more info when failed to copy
The warning was like:
WARNING: Could not copy license file COPYING: [Errno 1] Operation not permitted
We couldn't know which recipe print the warning from this message.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Barker [Sun, 21 Dec 2014 22:11:50 +0000 (22:11 +0000)]
opkg-collateral: Use new lists_dir syntax
Since opkg v0.2.3, it has been possible to use the new syntax for the
'lists_dir' option in an opkg configuration file. From v0.3.0, this new syntax
is required. The new syntax matches the way other options are set.
Signed-off-by: Paul Barker <paul@paulbarker.me.uk> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Aníbal Limón [Mon, 22 Dec 2014 23:30:46 +0000 (17:30 -0600)]
license: Validate if LICENSE is well defined.
Add check_license_format function that shows warning if LICENSE don't have
valid operators and also if have space separated entries without operator,
add check_license_format validation into base class.
[YOCTO #6758]
Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
When computing the dependency graph for the image generation, we need
to take into account the compression type and identify the base type
it relates to. This allow for a more robust graph generation even when
using composed image types.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Koen Kooi [Thu, 18 Dec 2014 08:51:04 +0000 (16:51 +0800)]
libpng 1.6.13: fix build for aarch64
The configure override was too restrictive, it needed both 'arm' and
'neon' to trigger, which breaks on aarch64. Since TUNE_FEATURES is the
only qualifier that matters, drop the 'arm' override.
Buildtested for 'genericarmv8' and 'qemux86' machines.
[RP: Added class-target override to avoid failures for libpng-native] Signed-off-by: Koen Kooi <koen.kooi@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mark Hatle [Thu, 18 Dec 2014 08:51:15 +0000 (16:51 +0800)]
linux-yocto: Fixup linux-yocto for aarch64
linux-yocto is updated to pass the sysroot path to the compiler when necessary.
linux-yocto_ver.bb are updated to reference the correct linux-yocto branchs and
SRCREVs.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Mark Hatle [Thu, 18 Dec 2014 08:51:14 +0000 (16:51 +0800)]
aarch-arm64: Update tune files
arch-arm64 is the base tune file for aarch64. Update this to allow the
system to work with both aarch32 and aarch64 (multilib).
arch-armv8 is for compatibility, it simply uses the base config for now.
feature-arm-thumb was updated, since aarch64 mode does NOT have thumb support.
We should only be processing warnings and additional arguments if thumb
support is enabled on the processor core.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Fathi Boudra [Thu, 18 Dec 2014 08:51:08 +0000 (16:51 +0800)]
libunwind: add aarch64 support
* pass --enable-debug-frame on aarch64 architecture
* include patches:
- aarch64 port (backported from upstream)
- Support-building-with-older-compilers (fix undefined reference to
`unreachable' on older compilers)
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Kai Kang [Thu, 18 Dec 2014 08:51:01 +0000 (16:51 +0800)]
glibc-locale.inc: add aarch64 to BINARY_LOCALE_ARCHES
Add aarch64 to var BINARY_LOCALE_ARCHES. In libc-packages.class, it is
used to work with other vars to check and add cross-localedef-native as
a dependecy correctly.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joe Slater [Thu, 18 Dec 2014 08:50:59 +0000 (16:50 +0800)]
siteinfo: account for 32 and 64 bit arm
Pull wordsize related items from arm-common and put in
arm-32. Leave them as they were. Copy arm-32 to arm-64
and comment out all variables. Re-define to correct
values leaving items which do not appear to be used
undefined.
Signed-off-by: Joe Slater <jslater@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Kai Kang [Thu, 18 Dec 2014 08:50:58 +0000 (16:50 +0800)]
Add machine qemuarm64
Add machine qemuarm64. The configure files are derived from linaro.
Update:
* rename genericarmv8 to qemuarm64 for coordination in oe-core
* include qemu.inc then remove common part of config
* disable using autoserial
* move arch-armv8.inc from machine/include/arm64 to machine/include/arm
[YOCTO #6487]
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In file Makefile in top directory, libs-y is assigned to to var
KBUILD_VMLINUX_MAIN. It uses script link-vmlinux.sh to link vmlinux.o,
and when execute function vmlinux_link() in link-vmlinux.sh,
KBUILD_VMLINUX_MAIN is passed to ${LD}.
If build without libgcc, the value of LIBGCC is just libgcc.a without
parent directory. linux-yocto fails to build:
| LD vmlinux.o
| aarch64-poky-linux-ld.bfd: cannot find libgcc.a: No such file or directory
Add libgcc to aarch64 kernel dependency.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Adding argparse module from Python's standard library. The module has replaced
previous python-argparse package that was build using source that are not part
of Python 2.7 release.
Signed-off-by: Maciej Borzecki <maciej.borzecki@open-rnd.pl> Signed-off-by: Maciek Borzecki <maciek.borzecki@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Barker [Thu, 18 Dec 2014 21:03:30 +0000 (21:03 +0000)]
opkg-utils: Update SRCREV
The new upstream commit incorporates the following fix to update-alternatives
from Chen Qi <Qi.Chen@windriver.com>:
update-alternatives: avoid double slash problem
The link path stored by update-alternatives should always be absolute and so we
do not need to add an additional slash when joining to $OPKG_OFFLINE_ROOT.
Signed-off-by: Paul Barker <paul@paulbarker.me.uk> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:57 +0000 (11:41 +0000)]
devtool: add QA tests
Add some QA tests for devtool (and recipetool). These aren't
comprehensive but at least they are a start, and have already helped me
catch and fix a number of regressions.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Junchun Guan [Fri, 19 Dec 2014 11:41:56 +0000 (11:41 +0000)]
scripts/devtool: Support deploy/undeploy function
Deploy recipe output files to live target machine using scp
Store the files list and target machine info in localhost if deployment
is done
Undeploy recipe output files in target machine using the previous
deployment info
[YOCTO #6654]
Signed-off-by: Junchun Guan <junchunx.guan@intel.com> Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:55 +0000 (11:41 +0000)]
scripts/devtool: add development helper tool
Provides an easy means to work on developing applications and system
components with the build system.
For example to "modify" the source for an existing recipe:
$ devtool modify -x pango /home/projects/pango
Parsing recipes..done.
NOTE: Fetching pango...
NOTE: Unpacking...
NOTE: Patching...
NOTE: Source tree extracted to /home/projects/pango
NOTE: Recipe pango now set up to build from /home/paul/projects/pango
The pango source is now extracted to /home/paul/projects/pango, managed
in git, with each patch as a commit, and a bbappend is created in the
workspace layer to use the source in /home/paul/projects/pango when
building.
Additionally, you can add a new piece of software:
$ devtool add pv /home/projects/pv
NOTE: Recipe /path/to/workspace/recipes/pv/pv.bb has been
automatically created; further editing may be required to make it
fully functional
The latter uses recipetool to create a skeleton recipe and again sets up
a bbappend to use the source in /home/projects/pv when building.
Having done a "devtool modify", can also write any changes to the
external git repository back as patches next to the recipe:
Paul Eggleton [Fri, 19 Dec 2014 11:41:54 +0000 (11:41 +0000)]
scripts: add scriptutils module
Add a utility module for scripts. This is intended to provide functions
only really useful before bitbake has been found (or only of particular
interest to scripts). At the moment this includes functions for setting
up a logger and for loading plugins.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:53 +0000 (11:41 +0000)]
scripts/recipetool: Add a recipe auto-creation script
Add a more maintainable and flexible script for creating at least the
skeleton of a recipe based on an examination of the source tree.
Commands can be added and the creation process can be extended through
plugins.
[YOCTO #6406]
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:52 +0000 (11:41 +0000)]
classes/externalsrc: set do_compile as nostamp
Most of the time what you want when using this class is for do_compile
to execute more than just once - every time the source changes would be
ideal, but that's a little tricky to accomplish. Thus, set do_compile as
nostamp to get something close. Note that in order to be effective this
also requires the change to bitbake that causes nostamp task signatures
to change on each execution.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sun, 21 Dec 2014 13:50:35 +0000 (13:50 +0000)]
kernel-devsrc: Ensure we have a dependency on the actual source
Tthe kernel populate_sysroot can come from sstate, we need the full
source here. We therefore depend on the configure task which isn't
covered by sstate to ensure we get the right set of files.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:50 +0000 (11:41 +0000)]
lib/oe: add recipeutils module
Add a module to help provide utility functions for dealing with recipes.
This would typically be used by external tools.
Substantial portions of this module were borrowed from the OE Layer
index code; other functions originally contributed by
Markus Lehtonen <markus.lehtonen@intel.com>.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:49 +0000 (11:41 +0000)]
lib/oe/patch: add support for extracting patches from git tree
When patches from a recipe have been written out to a git tree, we also
want to be able to do the reverse so we can update the patches next to
the recipe. This is implemented by adding a comment to each commit
message (using git hooks) which we can extract later on.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:46 +0000 (11:41 +0000)]
lib/oe/patch: auto-commit when falling back from git am
When PATCHTOOL = "git", if we're not able to use "git am" to apply a
patch and fall back to "git apply" or "patch", it is desirable to
actually commit the changes, attempting to preserve (and interpret) the
patch header as part of the commit message if present. As a bonus, the
code for extracting the commit message is callable externally in case it
is useful elsewhere.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:45 +0000 (11:41 +0000)]
lib/oe/patch: fall back to patch if git apply fails
When PATCHTOOL = "git", git apply doesn't support fuzzy application, so
if a patch requires that it's better to be able to apply it rather than
just failing.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Fri, 19 Dec 2014 11:41:43 +0000 (11:41 +0000)]
meta-environment: don't mark tasks as nostamp
With siggen being changed to alter the signature of nostamp tasks on the
fly, having these tasks as nostamp results in the SDK being rebuilt
every time, which is not desirable. In any case this is just legacy from
the days before we used signatures to take care of ensuring these tasks
get re-run when they need to be.
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 20 Dec 2014 11:16:40 +0000 (11:16 +0000)]
core-image-sato-sdk: Use kernel-devsrc to ensure kernel module builds work
With the move of the kernel source to the kernel-devsrc, we need this package
to ensure the kernel module sanity tests made on core-image-sato-sdk continue
to pass. Therefore update the name of the package included in the image.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 20 Dec 2014 11:19:51 +0000 (11:19 +0000)]
boot-directdisk: Use kernel from DEPLOY_DIR_IMAGE, not sysroot
After the recent kernel changes, this avoids failures like:
DEBUG: Executing shell function build_boot_dd
install: cannot stat '/home/pokybuild/yocto-autobuilder/yocto-worker/build-appliance/build/build/tmp/sysroots/qemux86-64/usr/src/kernel/bzImage': No such file or directory
WARNING: exit code 1 from a shell command.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 19 Dec 2014 17:46:27 +0000 (17:46 +0000)]
kernel: Fix non linux-yocto builds
After the recent kernel changes, non linux-yocto builds stopped working
properly for two reasons:
a) ${S} was being reset to ${WORKDIR}/git for example and STAGING_KERNEL_DIR
did not contain the source
b) Most builds were using ${B} == ${S}
This patch adds a fixup to the unpack function to handle the case where
${S} != ${STAGING_KERNEL_DIR} and also set up the infrastrcture so that
B != S for kernel builds from now on. The kernel build system is one of the
best for supporting this and there is no good reason not to take advantage
of it.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sun, 21 Dec 2014 12:30:53 +0000 (12:30 +0000)]
kernel-devsrc: Ensure we don't race against do_make_scripts from module-base.bbclass
do_install for kernel-devsrc can race against do_make_scripts from
module-base.bbclass. Since there is a lock there to guard against concurrency
already, we can just use it here to avoid a race.
Ultimately, this can all likely be much more streamlined but this resolves
the immediate build failures.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sun, 21 Dec 2014 12:29:03 +0000 (12:29 +0000)]
kernel-devsrc: Handle ppc crtsaves.o explictly for now
Resolve kernel module build failures for qemuppc by including crtsaves.o.
I'm not particularly happy to be doing this, it should perhaps be contained
in the kernel-dev package. Until the overlap between kernel-devsrc and
kernel-dev is resolved, this at least removed the regressions.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 19 Dec 2014 14:12:43 +0000 (14:12 +0000)]
kernel-devsrc: Inherit module-base
As a "normal" recipe, mulitlib would try and extend it for multilibs.
By inheriting module-base, we can avoid this since we now look more
'kernel' like.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sat, 20 Dec 2014 13:55:13 +0000 (13:55 +0000)]
linux-libc-headers: Upgrade 3.17 -> 3.17.7
Whilst we don't usually take linux-libc-headers point release, 3.17.3
fixes a userspace header issue which caused gdb to fail to build on
mips. We therefore may as well upgrade to the latest point release.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Wed, 10 Dec 2014 09:22:10 +0000 (04:22 -0500)]
linux-yocto/3.14: update to 3.14.26, integrate ltsi and -rt updates
Updating the 3.14 tree to the latest korg 3.14.26, as well as
integrating 3.14 LTSI content, and refreshing preempt-rt. Minor
conflict resolutions were performed between ltsi, stable and -rt
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Bruce Ashfield [Wed, 10 Dec 2014 09:20:19 +0000 (04:20 -0500)]
linux-yocto/3.10: update to v3.10.62
Updating to the latest korg -stable update for the 3.10 series. Minor
merge conflict resolution was done with the standard/ltsi and
standard/preempt-rt branches.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Bruce Ashfield [Wed, 15 Oct 2014 14:11:49 +0000 (10:11 -0400)]
linux-yocto/3.17: update to v3.17.1
Updating to the kver to the latest korg stable update. This fixes some
important bugs (as well as the usual mix of more minor ones), in particular
libata bugs and SSD corruption issues are fixed.
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Richard Purdie [Thu, 18 Dec 2014 11:41:58 +0000 (11:41 +0000)]
kernel: Clean ${S} before unpack
Currently unpack just forces sources over the current files. This change
ensures ${S} is cleaned out before sources are unpacked. This resolves
issues seen when upgrading to the new kernel class changes.
Ultimately, this should probably move to base.bbclass but one step at a
time, this solves an immediate problem blocking the other patches from
merging.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
