Ross Burton [Mon, 3 Apr 2017 11:57:52 +0000 (12:57 +0100)]
libjpeg-turbo: don't depend on NASM for non-x86 targets
NASM is a x86-specific assembler so it is only required when building for x86
targets. Use x86-architecture and class-target overrides to depend on NASM, but
explicitly disable and don't depend on it for native as complications in the
native overrides meant NASM was enabled for x86-64 but disabled for x86 (this
will be investigated later).
Original patch by Tristan Van Berkom <tristan.vanberkom@codethink.co.uk>, more
work to solve selftest failures by Richard Purdie
<richard.purdie@linuxfoundation.org>. I just wrote a nice commit message.
Also fix some missing whitespace in _appends.
[ YOCTO #11240 ]
Signed-off-by: Ross Burton <ross.burton@intel.com>
Jussi Kukkonen [Thu, 30 Mar 2017 16:49:28 +0000 (19:49 +0300)]
quilt: Fix paths for patch and perl
Currently some shebang lines end up as
#! /usr/bin/env perl -w
env does not like the argument. Also the current sed to insert env
does not cover the copies ptests use. Fix these issues by:
- using --with-perl to insert "env"
- Replacing "-w" in shebang lines with a new "use warning;" line
Remove a EXTRA_OECONF_append_class_target from the native recipe.
Don't overwrite EXTRA_OECONF in native: the values should be correct
for native as well.
--with-patch is used within the gnu patch wrapper only: before this
commit the wrapper contained a (build host) path to native patch.
Also tweak one test so busybox mv output is accepted.
All ptests should now pass: Fixes [YOCTO #11221].
Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Chen Qi [Thu, 30 Mar 2017 08:37:36 +0000 (16:37 +0800)]
wic: improve error message
When using `wic create mkefidisk -e core-image-minimal', the following
error message appeared.
Please bake it with 'bitbake parted-native' and try again.
However, following this command doesn't do any help. The same problem
still appeared.
The problem is that when we 'bitbake parted-native', it doesn't have
anything to do with core-image-minimal. And the required tool 'parted'
is not under core-image-minimal's recipe-sysroot-native directory.
Improve the error message so that following it could get things done.
Signed-off-by: Chen Qi <Qi.Chen@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Mon, 3 Apr 2017 15:58:36 +0000 (18:58 +0300)]
build-perf-test-wrapper.sh: make it possible to specify Git branch name
Support <branch>:<commit> format for the -c argument. This makes it
possible to test older commits of a certain branch (not just the tip of
it) so that the branch name will still be correctly recorded in the test
report data.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
The javascript console log messages are used in scraping, when
converting an html test report to html email. Before this patch a
console message indicating that all charts have been drawn was not
correctly sent if the last test failed (or didn't have chart data for
some other reason) which, in turn, caused oe-build-perf-report-email.py
script to fail with a timeout.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Markus Lehtonen [Mon, 3 Apr 2017 15:58:33 +0000 (18:58 +0300)]
scripts/oe-build-perf-report: better guessing of args
When getting info from the latest commit, don't search all refs but only
branches. We don't get correct data from refs/tags/* or refs/notest/*,
for example.
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mark Hatle [Fri, 31 Mar 2017 02:30:29 +0000 (21:30 -0500)]
yocto-compat-layer.py: Fix the signature validation
The initial signatures need to be collected -after- the dependency layers have
been added to the system. Otherwise changes that happen due to dependencies,
outside of the layer being scanned, will show up as signature problems.
The add_layer function was split into two pieces so that we can process
the dependencies first, and then add the layer itself for the comparison.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mark Hatle [Fri, 31 Mar 2017 02:30:28 +0000 (21:30 -0500)]
compatlayer/__init__.py: Allow add_layer to process recursive deps
When processing a layer for dependencies, you have to process the layer
itself, it's dependencies, the dependencies dependencies and so forth until
all items have been processed.
Mark Hatle [Fri, 31 Mar 2017 02:30:27 +0000 (21:30 -0500)]
yocto-compat-layer.py: Add --dependency argument
When processing a large number of items, there are times that it would be
nice to be able to pass in a series of layers that can be used as dependencies
for the layer that is being scanned. This avoids the significant overhead
of processing all of the layers to compatibility.
Signed-off-by: Mark Hatle <mark.hatle@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Tue, 4 Apr 2017 14:40:01 +0000 (15:40 +0100)]
gpgme: fix configure if 'import distutils' causes output on stderr
There are a number of reasons that importing a module could cause output on
stderr that isn't a fatal error (compatibilty problems with inputrc, or encoding
warnings) so backport a patch from autoconf-archive to only check the exit code
instead of asserting that stderr is empty.
[ YOCTO #11231 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
selftest/devtool: use kernel provider instead of recipe allowing other kernel providers
To generalize the unit test for other kernel providers, use virtual/kernel instead
of linux-yocto, allowing to run the selftest on non-poky distros, i.e. poky-tiny.
Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Martin Jansa [Tue, 4 Apr 2017 08:02:02 +0000 (10:02 +0200)]
x86-base.inc, qemuarm.conf: prefer 4.10 version of linux-yocto
* 4.8 version was removed in:
commit 466e6e45ca04a07ebe1b1f52de747f077b362d54
Author: Bruce Ashfield <bruce.ashfield@windriver.com>
Date: Tue Mar 28 08:58:07 2017 -0400
linux-yocto: drop 4.8 recipes
We have 4.1 (LTSI), 4.4 (LTS), 4.9 (LTS/LTSI) and 4.10 available in
master. 4.8 is no longer required, so we drop the recipe to keep
the version selection under control.
* causing each build to start with 188 lines of this stuff:
NOTE: preferred version 4.8% of linux-yocto not available (for item virtual/kernel)
NOTE: versions of linux-yocto available: 4.1.38+gitAUTOINC+7140ddb86e_4d2c95e78c 4.10.5+gitAUTOINC+01f18cba44_b97dcd4f06 4.4.56+gitAUTOINC+271b0c8d51_01aaede0a2 4.9.17+gitAUTOINC+8b97a445fa_3ff3760c2a
NOTE: preferred version 4.8% of linux-yocto not available (for item kernel-module-raid456)
NOTE: versions of linux-yocto available: 4.1.38+gitAUTOINC+7140ddb86e_4d2c95e78c 4.10.5+gitAUTOINC+01f18cba44_b97dcd4f06 4.4.56+gitAUTOINC+271b0c8d51_01aaede0a2 4.9.17+gitAUTOINC+8b97a445fa_3ff3760c2a
Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ming Liu [Thu, 30 Mar 2017 13:20:13 +0000 (15:20 +0200)]
kernel.bbclass: move in addtask kernel_link_images statement from linux-yocto.inc
Add kernel_link_images task in kernel.bbclass instead of adding it in
linux-yocto.inc, or else the recipes inheriting kernel.bbclass might
run into implicit dependency issues.
Signed-off-by: Ming Liu <peter.x.liu@external.atlascopco.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
brian avery [Tue, 4 Apr 2017 23:23:24 +0000 (16:23 -0700)]
devtool: point runqemu to correct native bindir
devtool/runqemu.py was relying on STAGING_BINDIR_NATIVE to find the host
tools it needed like qemu-system-<arch>. In the post RSS world, this no
longer exists. This patch points it to
{STAGING_DIR}/{BUILD_ARCH}/{bindir_native}.
[YOCTO #11223]
Signed-off-by: brian avery <brian.avery@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
brian avery [Tue, 4 Apr 2017 22:22:11 +0000 (15:22 -0700)]
meta-ide-support: exclude this recipe from rm_work
The results of meta-ide-support are used by utilities like runqemu and
runqemu-extract-sdk. Since the usr/bin that meta-ide-support creates is
subsequently removed bu rm_work, we exclude this recipe from the rm_work
list.
[YOCTO #11119]
Signed-off-by: brian avery <brian.avery@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
brian avery [Tue, 4 Apr 2017 22:22:10 +0000 (15:22 -0700)]
scripts: change way we find native tools (pseudo)
oe-find-native-sysroot: Recipe Specific Sysroots have eliminated the
large STAGING_DIR_NATIVE. Now, we will rely on the meta-ide-support
sysroot that is what was populating the large STAGING_DIR_NATIVE in
previous versions anyway. We now look for RECIPE_SYSROOT_NATIVE.
[YOCTO #11119]
Signed-off-by: brian avery <brian.avery@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Serhii Popovych [Fri, 31 Mar 2017 11:46:58 +0000 (11:46 +0000)]
automake: Adjust shebang lines to remove interpreter path hardcode
If build host perl (and other tools) is old and we use some kind
of toolchain to provide recent perl/python/etc to the OE build
we still locked to use build host perl due to hardcoded shebang
lines in automake scripts.
Behaviour was observed with Enterprise Linux 6 and devtoolset
toolchain from SCL (Software Collections) used to provide recent
version of perl (not provided with default buildtools-tarball).
Pass /usr/bin/env perl in ac_cv_path_PERL configuration variables
for class-native and class-nativesdk. Use patch to automake to replace
-w option in shebang line with modern way to enable warnings on perl
(i.e. "use warnings").
Also add nativesdk-autoconf to RDEPENDS to bring runtime dependencies
inline with other targets.
Note that ac_cv_path_PERL must be valid perl interpreter path
since configure will check perl version and Flock implementation.
It is not possible currently to use nativeperl from native
sysroot because automake does not DEPENDS on perl-native (and
doing so fails due to circular dependencies). Only possible
solution is to overwrite shebangs with nativeperl somewhere at
do_install() and update RDEPENDS for class-native. Or add perl
symlinks to nativeperl in sysroot.
For now it seems good to use perl found by /usr/bin/env from
automake-native.
Serhii Popovych [Fri, 31 Mar 2017 11:46:57 +0000 (11:46 +0000)]
autoconf: Adjust shebang lines to remove interpreter path hardcode
If build host perl (and other tools) is old and we use some kind
of toolchain to provide recent perl/python/etc to the OE build
we still locked to use build host perl due to hardcoded shebang
lines in autoconf scripts.
Behaviour was observed with Enterprise Linux 6 and devtoolset
toolchain from SCL (Software Collections) used to provide recent
version of perl (not provided with default buildtools-tarball).
Pass /usr/bin/env perl in ac_cv_path_PERL configuration variables
for class-native and class-nativesdk. Use patch to autoconf to replace
-w option in shebang line with modern way to enable warnings on perl
(i.e. "use warnings").
Also add nativesdk-m4 and nativesdk-gnu-config to RDEPENDS to bring
runtime dependencies inline with other targets.
Note that ac_cv_path_PERL must be valid perl interpreter path
since configure will check perl version and Flock implementation.
It is not possible currently to use nativeperl from native
sysroot because autoconf does not DEPENDS on perl-native (and
doing so fails due to circular dependencies). Only possible
solution is to overwrite shebangs with nativeperl somewhere at
do_install() and update RDEPENDS for class-native. Or add perl
symlinks to nativeperl in sysroot.
For now it seems good to use perl found by /usr/bin/env from
autoconf-native.
Markus Lehtonen [Fri, 31 Mar 2017 14:07:30 +0000 (17:07 +0300)]
scripts/contrib: add oe-build-perf-report-email
Script for sending build perf test reports as an email. Mangles an html
report, generated by oe-build-perf-report, into a format suitable for
html emails. Supports multipart emails where a plaintext alternative can
be included in the same email.
Dependencies required to be installed on the host:
- phantomjs
- optipng
[YOCTO #10931]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Markus Lehtonen [Fri, 31 Mar 2017 14:07:29 +0000 (17:07 +0300)]
scripts: add oe-build-perf-report script
A new tool for pretty-printing build perf test results stored in a Git
repository. The scripts is able to produce either simple plaintext
report showing the difference between two commits, or, an html report
that also displays trendcharts of the test results. The script uses
Jinja2 templates for generating HTML reports so it requires
python3-jinja2 to be installed on the system.
[YOCTO #10931]
Signed-off-by: Markus Lehtonen <markus.lehtonen@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Juro Bystricky [Fri, 31 Mar 2017 21:28:13 +0000 (14:28 -0700)]
ncurses_6.0: Improve reproducibility
Build static libraries without the binutils "ar" -U option.
This option deliberately breaks deterministic mode.
The option seems to be a relic from 2015, intended as a workaround
for some unspecified build problems.
[YOCTO#11247]
Signed-off-by: Juro Bystricky <juro.bystricky@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Fri, 31 Mar 2017 13:39:16 +0000 (14:39 +0100)]
dpkg: Add PACKAGECONFIG for liblzma and enable it
liblzma is part of xz and we already build it but configure it out. This makes
no sense. Enabling it means we gain multithreaded compression and it speeds
dpkg-deb up massively. It also removes the fork overhead of separate xz processes.
Turning the existing config into a PACKAGECONFIG and turning it on by default
therefore makes best use of what we have available.
The manual RDEPENDS are no longer needed since it uses liblzma which is picked
up by the shlibs code magically.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ming Liu [Thu, 30 Mar 2017 07:59:58 +0000 (09:59 +0200)]
kernel.bbclass: introduce INITRAMFS_IMAGE_NAME
It defaults to ${INITRAMFS_IMAGE}-${MACHINE} if INITRAMFS_IMAGE is not
empty.
This allows the end users to be able to override the initramfs image
name with a customized value.
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>
Daniel Díaz [Thu, 23 Mar 2017 21:34:44 +0000 (15:34 -0600)]
base-files: resize only serial tty's in profile
We don't want to run resize on non serial consoles. There's
been an earlier attempt (6557787), so this builds upon that.
The problem we're seeing is that if there is text buffered in
the virtual console (like from a desperate user trying to
enter login details), resize will get stuck while calling
ioctl(tty, TCSETAW);
Since serial consoles are named (not just numbered), this
change limits resize's reach even further to run only on
/dev/tty[A-z] (thus avoiding /dev/tty[0-9]).
Signed-off-by: Daniel Díaz <daniel.diaz@linaro.org> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
lsb: Create ${base_prefix}/lib64 correctly when needed
There were two remaining cases that could end up creating /lib64
rather than ${base_prefix}/lib64. The difference matters when building
with usrmerge.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Use ${D}${var} rather than ${D}/${var} for variables where ${var}
contains an absolute path.
Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Robert Yang [Mon, 27 Mar 2017 10:39:13 +0000 (03:39 -0700)]
oe-run-native: OLDPATH -> OLD_PATH
It's a typo.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Enrico Jorns [Wed, 29 Mar 2017 09:07:33 +0000 (11:07 +0200)]
systemd: make vconsole a PACKAGECONFIG option
Allowing to remove the systemd-vconsole-setup package without specifying
the --disable-vconsole configure option for systemd will make the system
boot with the failure prompt
| systemd-udevd[142]: failed to execute '/lib/systemd/systemd-vconsole-setup' '/lib/systemd/systemd-vconsole-setup': No such file or directory
| systemd-udevd[96]: Process '/lib/systemd/systemd-vconsole-setup' failed with exit code 2.
as the 90-vconsole.rules will still be installed with having a
RUN+="/lib/systemd/systemd-vconsole-setup" in it that attempts to
execute a non-existing binary.
Signed-off-by: Enrico Jorns <ejo@pengutronix.de> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ming Liu [Thu, 30 Mar 2017 13:17:04 +0000 (15:17 +0200)]
kernel.bbclass: fix a typo
In a addtask statement, do_strip should be strip.
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>
Daniel Schultz [Thu, 30 Mar 2017 07:07:13 +0000 (09:07 +0200)]
wic: partition: Update fsck parameters
These parameters are copied from the ext image class.
-D will let fsck perform further directory optimizations
-v might be helpful for debugging purpose
Signed-off-by: Daniel Schultz <d.schultz@phytec.de> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ed Bartosh [Thu, 30 Mar 2017 18:17:20 +0000 (21:17 +0300)]
wic: remove prepare_empty_partition_squashfs
There is not much sense in creation of empty squashfs
partition. It's also not possible to create empty squashfs
partition of specified size.
Even more, prepare_empty_partition_squashfs method is
absolutely broken. It raises exception when called and
even its signature differs from the rest of of similar
methods. It means that nobody uses it and it's safe
to remove it.
Removed prepare_empty_partition_squashfs method and
testing of empty squashfs partition.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Thu, 30 Mar 2017 11:05:52 +0000 (14:05 +0300)]
wic: don't silently skip unknown fstypes
Fixed wic code that loops through hard-coded list of known fstypes
to find prepare_rootfs_<fstype> or prepare_empty_partition_<fstype>
methods and silently skipping unknown fstypes.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Sun, 26 Mar 2017 17:39:22 +0000 (20:39 +0300)]
wic: remove unused code from runner module
Removed unused APIs 'outs' and 'quiet'.
Removed 'catch' parameter from runner.runtool API as wic
uses only one value of it. Removed the code that handles
unused values of 'catch' parameter.
[YOCTO #10618]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Sun, 26 Mar 2017 17:27:48 +0000 (20:27 +0300)]
wic: use wic-tools STAGING_DATADIR as bootimg_dir
If bootloader artifacts are not found in default bootimg_dir
use wic-tools sysroot for the same purpose. This should
prevent wic from failing if bootloader artifacts can't be
found in image native sysroot.
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Sun, 26 Mar 2017 17:20:31 +0000 (20:20 +0300)]
wic: remove fsimage plugin
Removed fsimage plugin and prepare_rootfs_from_fs_image API as
they duplicate functionality of rawcopy plugin. Fsimage plugin makes
wic to remove the image artifact from deployment directory, which
can cause problems too.
[YOCTO #10618]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Ed Bartosh [Sun, 26 Mar 2017 17:08:46 +0000 (20:08 +0300)]
filemap: remove FilemapSeek class
FIEMAP API was added to Linux kernel 2.6.28 back in 2008
SEEK_HOLE and SEEK_DATA API was added much letter.
As FIEMAP is used by filemap module as a default API it's
safe to remove FileMpSeek class as it's never used.
[YOCTO #10618]
Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Mariano Lopez [Thu, 30 Mar 2017 22:18:14 +0000 (15:18 -0700)]
oeqa/runtime/cases: Migrate underscore cases
There were two missing cases to be migrated to the new framework: _qemutiny and
_ptest.
qemutiny was straightforward.
ptest on the other hand wasn't working even in previous releases; it has been
migrated from smart to dnf, and how ptest packages are gathered to be
installed, adapted to use unicode, and removed a lot of code that wasn't needed
anymore.
Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
In 2.4 development cycle openssl 1.1 will replace openssl 1.0 as the
default openssl version. Openssl 1.0 will stay but will be renamed
to openssl10, and eventually it will be removed (hopefully much
sooner than the official end of support date of Dec 2019, as we do not
want an unsupported openssl version in supported Yocto releases).
There are several recipes that are not API compatible with 1.1; some
of them will eventually be fixed, but others will never be (such as Qt4).
To avoid breaking such recipes when openssl 1.1 is added to oe-core,
let's provide "openssl10" already now and change the recipes to depend
on that where necessary; Qt4 is a particularly pressing issue as it is
causing failures on the autobuilder with my work in progress
openssl 1.1 branch, and so I'm not able to see what else would fail
later in the build process.
Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Serhii Popovych [Tue, 28 Mar 2017 19:02:38 +0000 (19:02 +0000)]
perl-native: Bring build dependencies inline with target perl package
Make sure we have all build time dependencies of perl-native
satisfied before we start building it.
Behaviour was observed with RHEL6 build hosts where perl-native
build fails at linging stage when attempting to link to the build
hosts old Berkley DB because virtual/db-native is missing in
native sysroot.
Add dependencies to the native packages taken from perl recipe.
Serhii Popovych [Tue, 28 Mar 2017 19:31:59 +0000 (19:31 +0000)]
distutils.bbclass: Fix path to python for scripts
Commit 11229de (distutils: Replacing path to native python by path
to python in the image to support python packages with
console-script setup) replaces path to python-native with path
to the python interpreter in the target image.
However on nativesdk builds ${bindir} expanded to the
${SDKPATHNATIVE}${prefix_nativesdk} making shebang line
pointing to the absolute path to env(1) in SDKPATHNATIVE
which may not be present if coreutils isn't part of nativesdk.
On the other hand commit cf63d90 (bitbake.conf: Define USRBINPATH
globally instead of individually) introduces USRBINPATH variable
which has correct value regarding build class and intended for
this use case.
Max Krummenacher [Fri, 24 Mar 2017 15:01:14 +0000 (16:01 +0100)]
openssl: Fix regression when building for thumb2
Commit 'c8da8ce openssl: Fix build with clang' introduced a regression.
do_compile fails when building with gcc/thumb2.
Note that I did not test if it still builds with clang.
Prevents the following when building with thumb2:
| ghash-armv4.S: Assembler messages:
| ghash-armv4.S:88: Error: thumb conditional instruction should be in IT block -- `ldrplb r12,[r2,r3]'
| ghash-armv4.S:98: conditional infixes are deprecated in unified syntax
| ghash-armv4.S:98: Error: thumb conditional instruction should be in IT block -- `ldrplb r8,[r0,r3]'
| ghash-armv4.S:105: Error: thumb conditional instruction should be in IT block -- `eorpl r12,r12,r8'
| ghash-armv4.S:107: Error: thumb conditional instruction should be in IT block -- `andpl r14,r12,#0xf0'
| ghash-armv4.S:108: Error: thumb conditional instruction should be in IT block -- `andpl r12,r12,#0x0f'
| ghash-armv4.S:144: conditional infixes are deprecated in unified syntax
| ghash-armv4.S:144: Error: thumb conditional instruction should be in IT block -- `ldrneb r12,[r2,#15]'
| ghash-armv4.S:231: conditional infixes are deprecated in unified syntax
| ghash-armv4.S:231: Error: thumb conditional instruction should be in IT block -- `ldrplb r12,[r0,r3]'
| ghash-armv4.S:248: Error: thumb conditional instruction should be in IT block -- `andpl r14,r12,#0xf0'
| ghash-armv4.S:249: Error: thumb conditional instruction should be in IT block -- `andpl r12,r12,#0x0f'
Signed-off-by: Max Krummenacher <max.krummenacher@toradex.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 30 Mar 2017 11:26:02 +0000 (12:26 +0100)]
nasm: remove COMPATIBLE_HOST
nasm can build on every architecture, it just can't generate anything but X86
code. As we can't know what the user intends to do with nasm, remove the
COMPATIBLE_HOST line.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Patrick Ohly [Thu, 30 Mar 2017 15:09:36 +0000 (17:09 +0200)]
net-tools: enable native and nativesdk variant
net-tools-native is needed by swtpm-wrappers (in meta-security)
because swtpm_setup.sh calls netstat, which cannot be assumed to be
present in all Linux installations (for example, it is not in OpenSUSE
minimal base).
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 30 Mar 2017 13:34:10 +0000 (14:34 +0100)]
base.bbclass: ensure HOSTTOOLS links point to executables
Use the new executable argument to bb.utils.which() to ensure that the symlinks
point to executable files and not for example directories with the right name
which happened to be on $PATH.
[ YOCTO #11256 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 30 Mar 2017 14:08:33 +0000 (15:08 +0100)]
glibc-package: Allow 32 and 64 bit headers to exist on arm
With this change (combined with the previous linux-libc-header fix), a
combined sysroot for 32 and 64 bit arm works meaning our SDK works
correctly for that multilib setup.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 30 Mar 2017 14:05:33 +0000 (15:05 +0100)]
linux-libc-headers: Allow arm kernel headers to coexist on 32/64 bit
Its rather sad we need to do this but in order to make combined 32 and 64 bit
SDKs work, we need a common set of headers and this is the delta that
allows things to work. It only applies on arm.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 30 Mar 2017 14:02:21 +0000 (15:02 +0100)]
multilib_header: Update wrapper to handle arm 32/64 bit
Having arm 32/64 bit headers coexisting turns out to be tricky. Unfortunately
our wrapper works using wordsize.h and this differs on arm so we can't use it.
Therefore replicate the logic here for arm. I did look into writing our
own wordsize.h but we also need to remap kernel headers on arm and
since wordsize.h comes from libc, that doesn't work for kernel headers.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 30 Mar 2017 09:26:56 +0000 (10:26 +0100)]
sanity: Drop obsolete TMPDIR ABI conversions
When we get to version 12 we have a hard break as we can't convert to newer
versions. There is no point in running the old conversions on an old tmpdir
only to hit that block. Remove all the old conversions to avoid that and
make things clearer.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Patrick Ohly [Wed, 29 Mar 2017 11:29:50 +0000 (13:29 +0200)]
populate_sdk_ext.bbclass: enhance compatibility with rm_work.bbclass
"bitbake -c populate_dsk_ext" must not trigger do_rm_work, because it
is impossible to declare that the additional tasks activated by "-c
populate_dsk_ext" must run before do_rm_work. When do_populate_dsk_ext
and do_rm_work are both active, the resulting race condition breaks
do_populate_dsk_ext.
The existing bitbake dependencies can't be used for that, because
"addtask populate_dsk_ext before do_rm_work" would then always execute
populate_dsk_ext also in normal builds.
do_populate_dsk_ext triggers do_rm_work indirectly through the
dependency on do_build of the SDK_TARGETs. Using the new
do_build_without_rm_work instead (when available, with do_build as
before if not) avoids the problem.
However, one has to be careful to not trigger do_rm_work in the same
build in some other way. "bitbake core-image-sato:do_populate_sdk_ext
core-image-sato:do_build" still fails, for example. Doing one after
the other works.
Fixes: [YOCTO 11042] Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Some classes, for example populate_sdk_ext.bbclass, must be able to
trigger a full build of a recipe without also triggering
do_rm_work. They cannot depend on do_build anymore, because that would
trigger do_rm_work.
Instead, do_build_without_rm_work can be used. It has the exact same
dependencies as do_build, minus do_rm_work and do_rm_work_all.
This may also be useful in a test build of a recipe where one wants
to preserve the work directory without having to modify configuration
settings:
bitbake foobar:do_build_without_rm_work
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Patrick Ohly [Wed, 29 Mar 2017 12:52:06 +0000 (14:52 +0200)]
ovmf: avoid linking with gold
Linking with gold fails due to an internal error in gold. The ovmf
linker is gcc, which has a -fuse-ld=bfd option to choose the linker
which (for ovmf) is known to work.
Like the choice of the compilers, this is done in ovmf-native. To keep
that recipe independent of DISTRO_FEATURES, choosing bfd is done
unconditionally.
Signed-off-by: Patrick Ohly <patrick.ohly@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sysroot-relativelinks: also consider links to dirs on the host
Dead symlinks, or symlinks to existing files will show up in 'files' of an
os.walk, but symlinks to existing directories show up in 'dirs', so we need to
consider both.
As one example where this is an issue, the symlink from /usr/lib/ssl/certs was
left pointing to /etc/ssl/certs rather than the relative path when the sdk was
built on hosts where the latter exists.
Signed-off-by: Christopher Larson <chris_larson@mentor.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
