Wang Mingyu [Thu, 28 Nov 2019 09:12:26 +0000 (01:12 -0800)]
cmake: upgrade 3.15.3 -> 3.15.5
Signed-off-by: Wang Mingyu <wangmy@cn.fujitsu.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
If the pthread.h header is found, make sure library containing
"pthread_atfork" is added to the list of libraries against which to
link. On some hosts (e.g. openSUSE 15.1) "-lpthread" needs to be
explicitly added in order for the code to compile correctly.
We already had a workaround for this for native builds, but using some
external toolchains, we can be bit by this for the target recipe as
well.
Most of this is courtesy Trevor Woerner <twoerner@gmail.com>, via his
patch at stevegrubb/libcap-ng#10, the last thanks to Khem Raj.
Signed-off-by: Christopher Larson <chris_larson@mentor.com> Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Tue, 26 Nov 2019 09:18:18 +0000 (09:18 +0000)]
python: restore bytecode optimisation
The upgrade from 2.7.3 to 2.7.9 in oe-core cb02ecb introduced some
copy-paste errors which inadvertandly disabled the generation of
optimised bytecode (.pyo files).
Restore the intended behaviour and bring back .pyo files.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Nathan Rossi [Mon, 25 Nov 2019 06:58:00 +0000 (06:58 +0000)]
gcc-cross.inc: Prevent native sysroot from leaking into configargs.h
Prevent the native(sdk) sysroot path from leaking into configargs.h. The
configargs.h header is intended to be static and unchanged as the
content is used as a means of determining that a gcc plugin is built for
the same gcc. This also effects the output of 'gcc --version'. Due to
per recipe sysroots and staging, the sysroot path would be replaced with
the sysroot local to the recipe thus changing the content of
configargs.h.
The sysroot path is replaced with a generic "/host" prefix which
represents the host sysroot (e.g. native or nativesdk).
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Mark Hatle [Mon, 18 Nov 2019 17:17:30 +0000 (11:17 -0600)]
gcc-9.2: Update the relocation patch to add PREFIX/EXEC_PREFIX
Without relocating PREFIX/EXEC_PREFIX the system can not do runtime
relocation for the path to the usr/lib/gcc directory, and other components.
While this is not a normal or supported use-case it does work in the upstream
gcc. This is difficult to test with the regular OE SDKs, as it requires
running the components with the correct LD_LIBRARY_PATH and ld.so.
Without this update, gcc will typically not be able to find the gcc
provided include file for stddef.h and similar. This is due to certain
relocations being based on the PREFIX and/or EXEC_PREFIX locations which
are hardcoded at compilation time.
Signed-off-by: Mark Hatle <mark.hatle@kernel.crashing.org> Signed-off-by: Ross Burton <ross.burton@intel.com>
Add kernel-devicetree to exceptions to fix the QA issue. Because there
are already 3 kernel related criteria, simplify them by judging package
names whether start with 'kernel-'. And also refactor to remove
duplicate 'not'.
Signed-off-by: Kai Kang <kai.kang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Mon, 25 Nov 2019 17:22:21 +0000 (18:22 +0100)]
python-native: don't cause a full regeneration of the built sources
When cross-compiling Python 2 you need a native pgen binary, but the cross
recipe can't do this on it's own so we build it in python-native and install it.
The rule to build pgen was also causing a complete rebuild of all of the
generated sources, which meant that building Python 2 needs a *host* Python 2.
This can be fixed by simply building pgen, as this is all we need to install.
[ YOCTO #13645 ]
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
The LICENSE file contains all the license information so there is no
need to also include it from the png.h file (and additionally some
lines were left out from the latter).
License-Update: Remove duplicate license information Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
opkg: Trim the text part used for the license file checksum
This avoids including irrelevant information when calculating the
license checksum.
License-Update: Trim the text part used for the license file checksum Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
alsa-utils: Trim the text part used for the license file checksum
This avoids including irrelevant information when calculating the
license checksum.
License-Update: Trim the text part used for the license file checksum Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
alsa-lib: Trim the text part used for the license file checksum
This avoids including irrelevant information when calculating the
license checksum.
License-Update: Trim the text part used for the license file checksum Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Barker [Mon, 11 Nov 2019 14:16:21 +0000 (14:16 +0000)]
oeqa: archiver: Add basic tests for all archiver modes
6 new test cases are added to cover the various archiver modes
documented at the top of archiver.bbclass. Each test sets the
appropriate configuration options, runs the `do_deploy_archives` task
for the selftest-ed recipe and checks for the presence of the expected
archive file.
Signed-off-by: Paul Barker <paul@betafive.co.uk> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Sat, 23 Nov 2019 18:15:28 +0000 (10:15 -0800)]
systemd-bootchart: Backport RISC-V support
This is a combined backport from upstream patches for added risc-v
support
Upstream code has been re-organised before risc-v support was added to
its mix of two commits
primarily
https://github.com/systemd/systemd/commit/171b53380085b1288b03b19a2b978f36a5c003d0
Ernst Sjöstrand [Mon, 25 Nov 2019 05:41:41 +0000 (05:41 +0000)]
do_image_cpio: Improve reproducibility
Find lists files in the order that the filesystem has registered them,
which can vary. Adding a sort should have minimal performance impact.
Also use the --reproducible option to cpio.
Signed-off-by: Ernst Sjöstrand <ernst.sjostrand@verisure.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nathan Rossi [Mon, 25 Nov 2019 06:57:41 +0000 (06:57 +0000)]
oeqa/selftest/cases/gcc.py: Split 'gcc' and 'g++' testsuites
Split the GccCrossSelfTest testcase into separate testcases for 'gcc'
and 'g++' respectively. In order to split them use the "check-gcc-*"
language make check targets.
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Nathan Rossi [Mon, 25 Nov 2019 06:56:56 +0000 (06:56 +0000)]
oeqa/utils/qemurunner.py: Handle QEMU machines with a single serial
Not all QEMU machines are capable of having more than one serial port,
this is due to the machine emulating a physical device/board.
Rework QemuRunner to handle machines that only have 1 serial port, where
the serial port shares output of the kernel log buffer and a login
console. In this case the output is mixed but enables the machine to
boot and have QemuRunner detect the login prompt.
QemuTarget uses SERIAL_CONSOLES to determine the number of available
serial ports.
Signed-off-by: Nathan Rossi <nathan@nathanrossi.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Stefan Agner [Fri, 15 Nov 2019 16:09:44 +0000 (16:09 +0000)]
base-files: set ptmxmode to 666
Make sure that the (newer) /dev/pts/ptmx is accessible by users. This
is useful e.g. when running containers which symlink /dev/ptmx to
/dev/pts/ptmx on start. The default mode (000) does not allow to
create ptys inside the container.
Using 666 when symlinking /dev/ptmx is also recommended by the kernel
documentation when /dev/ptmx is symlinked:
https://www.kernel.org/doc/Documentation/filesystems/devpts.txt
Khem Raj [Mon, 25 Nov 2019 08:44:52 +0000 (00:44 -0800)]
gcc-runtime: Package libstdc++.a-gdb.py
There is python gdb script for static libstdc++ archives as well
fixes
ERROR: gcc-runtime-9.2.0-r0 do_package: QA Issue: gcc-runtime: Files/directories were installed but not shipped in any package:
/usr/lib/libstdc++.a-gdb.py
Signed-off-by: Khem Raj <raj.khem@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Added a new package: pulseaudio-pa-info. It contains the new pa-info
script.
BlueZ 4 support was removed in this version. That's not visible in the
recipe, but I noticed that the BlueZ 4 modules were still being built in
12.2, since they hadn't been explicitly disabled in the recipe.
Signed-off-by: Tanu Kaskinen <tanuk@iki.fi> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Joshua Watt [Thu, 21 Nov 2019 16:59:11 +0000 (10:59 -0600)]
classes/cmake: Use relative RPATHs
In most cases, the RPATH is stripped out when the ELF file is packages,
but by then the damage is done from a reproducible perspective because
this absolute path is hashed as part of the build-id generated at link
time ([1] has a good explanation). Fortunately, newer cmake has an
option to generated relative RPATHs that use $ORIGIN to set the path, so
set it in the toolchain file.
Joshua Watt [Thu, 21 Nov 2019 16:58:24 +0000 (10:58 -0600)]
librsvg: Fix build reproducibility
librsvg was encoding the path to the build directory in order to find a
font file for testing. This wouldn't work in OE anyway since the build
directory isn't present at that exact location on the target, so remove
the offending path to make the build reproducible.
Signed-off-by: Joshua Watt <JPEWhacker@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Stefan Agner [Fri, 22 Nov 2019 16:24:41 +0000 (16:24 +0000)]
rpcbind: use upstream systemd service
Use upstream systemd service files instead of our own service files.
This also makes sure that /run/rpcbind.sock is used which fixes the
following systemd warning:
/usr/lib/systemd/system/rpcbind.socket:5: ListenStream= references a
path below legacy directory /var/run/, updating /var/run/rpcbind.sock
\xe2\x86\x92 /run/rpcbind.sock; please update the unit file accordingly.
Signed-off-by: Stefan Agner <stefan.agner@toradex.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Stefan Ghinea [Thu, 21 Nov 2019 15:28:04 +0000 (17:28 +0200)]
ghostscript: CVE-2019-14869
A flaw was found in all versions of ghostscript 9.x before 9.28,
where the `.charkeys` procedure, where it did not properly secure
its privileged calls, enabling scripts to bypass `-dSAFER` restrictions.
An attacker could abuse this flaw by creating a specially crafted
PostScript file that could escalate privileges within the Ghostscript
and access files outside of restricted areas or execute commands.
Wenlin Kang [Tue, 5 Nov 2019 10:04:20 +0000 (18:04 +0800)]
sysstat: remove check for chkconfig
For cross-platform, chkconfig can't work, so should remove check for it.
This can only be reproduced on some platform with chkconfig(e.g. CentOS
Linux release 7.2.1511), and need with --enable-install-cron and without
--enable-copy-only.
Fixed:
| if [ "n" == "n" ]; then \
| if [ -x "/usr/sbin/chkconfig" ]; then \
| cd /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/init.d && /usr/sbin/chkconfig --add sysstat; \
| else \
| [ -d /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc2.d ] || mkdir -p /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc2.d; \
| [ -d /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc3.d ] || mkdir -p /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc3.d; \
| [ -d /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc5.d ] || mkdir -p /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc5.d; \
| cd /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc2.d && ln -s -f ../init.d/sysstat S01sysstat; \
| cd /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc3.d && ln -s -f ../init.d/sysstat S01sysstat; \
| cd /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d/rc5.d && ln -s -f ../init.d/sysstat S01sysstat; \
| fi \
| fi \
| elif [ -d /path/to/tmp/work/corei7-64-wrs-linux/sysstat/11.1.5-r0/image/etc/rc.d ]; then \
| ...
| fi
| error reading information on service sysstat: No such file or directory
| Makefile:382: recipe for target 'install_all' failed
Signed-off-by: Wenlin Kang <wenlin.kang@windriver.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Fri, 17 Nov 2017 20:19:32 +0000 (20:19 +0000)]
texi2html: remove
The last user of this obsolete recipe (abandoned upstream in 2010, removed from
oe-core build dependencies in 2012) has now been deleted from oe-core, so delete
the recipe too.
Signed-off-by: Ross Burton <ross.burton@intel.com>
Ross Burton [Tue, 19 Nov 2019 20:38:24 +0000 (20:38 +0000)]
packagegroup-core-sdk: remove intltool
Intltool is deprecated these days, as gettext can handle almost everything
intltool could. Remove it from the SDK packagegroups, if it is needed then the
user can add it explicitly.
Signed-off-by: Ross Burton <ross.burton@intel.com>
The homepage: https://www.kernel.org/doc/man-pages/download.html
specifies that the latest tarballs are published one level up
from where we currently take them.
Signed-off-by: Alexander Kanavin <alex.kanavin@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Alex Kube [Fri, 25 Oct 2019 19:19:15 +0000 (23:49 +0430)]
go: Extract common environment setup.
Add default values for go environment variables to go-common.inc.
Override where appropriate in other go*.inc files, and use
host/target tuples from goarch for setting CC flags.
Signed-off-by: Alex Kube <alexander.j.kube@gmail.com> Signed-off-by: Ross Burton <ross.burton@intel.com>
Richard Purdie [Fri, 22 Nov 2019 15:12:44 +0000 (15:12 +0000)]
package_ipk: Remove pointless comment to trigger rebuild
The opkg-utils change didn't trigger a full repackage due to its mention
in layer.conf for update-alternatives hash exclusion. Do it manually
to invalidate broken ipks.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Thu, 21 Nov 2019 14:59:06 +0000 (14:59 +0000)]
sanity: Add check for tar older than 1.28
Older versions break opkg-build when reproducible builds are enabled.
Rather than trying to be selective based on which features are enabled,
lets just make this a minimum version.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
opkg-build was failing on hosts where tar < 1.28 and reproducibile builds
were enabled but it was doing this silently and generating corrupted
(empty) ipk files. Add a fix for this (submitted upstream).
The fix requires bash but if you're building ipk files this shoudn't be
a problem.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Paul Eggleton [Tue, 19 Nov 2019 10:52:52 +0000 (23:52 +1300)]
devtool: fix devtool upgrade with reproducible_builds class
If the reproducible_build class is inherited then there may be a
"source-date-epoch" subdirectory in a fetched source tree; devtool
upgrade was not expecting that in the upgraded source. Take a small
snippet of code from recipetool create which already handles this,
and make it a shared function that can be used in both places.
Additionally, fix an assumption that the source is always in a
subdirectory in the cleanup code that blocked debugging this.
[YOCTO #13635]
Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
