Taras Kondratiuk [Fri, 22 Mar 2019 00:56:59 +0000 (17:56 -0700)]
devtool: prevent starting git gc in background
Devtool creates a git repository for extracted sources in a temporary
directory and then moves it to a final destination after patching is
done. Unfortunately devtool is not aware that some of its git operations
may have caused git garbage collector to start in background. If timing
is just right a repository move fails because GC is removing git objects
while they are being moved.
The issue was hit on Krogoth release, but the code that moves created
git repository is still the same.
Disable gc.autodetach to make GC run synchronously and block git until
it is done.
Traceback (most recent call last):
File "<snip>/openembedded-core/scripts/devtool", line 342, in <module>
ret = main()
File "<snip>/openembedded-core/scripts/devtool", line 329, in main
ret = args.func(args, config, basepath, workspace)
File "<snip>/openembedded-core/scripts/lib/devtool/standard.py", line 352, in extract
initial_rev = _extract_source(srctree, args.keep_temp, args.branch, False, rd)
File "<snip>/openembedded-core/scripts/lib/devtool/standard.py", line 644, in _extract_source
shutil.move(srcsubdir, srctree)
File "/usr/lib/python2.7/shutil.py", line 299, in move
copytree(src, real_dst, symlinks=True)
File "/usr/lib/python2.7/shutil.py", line 208, in copytree
raise Error, errors
Error: [('/tmp/devtool5RXkuX/workdir/grpc-1.2.5/.git/objects/5e', '<snip>/build/grpc/grpc_src/.git/objects/5e', "[Errno 2] No such file or directory: '/tmp/devtool5RXkuX/workdir/grpc-1.2.5/.git/objects/5e'"), ('/tmp/devtool5RXkuX/workdir/grpc-1.2.5/.git/objects/57', '<snip>/build/grpc/grpc_src/.git/objects/57', "[Errno 2] No such file or directory: '/tmp/devtool5RXkuX/workdir/grpc-1.2.5/.git/objects/57'"), many git objects ... ]
Signed-off-by: Taras Kondratiuk <takondra@cisco.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
linux-yocto: fix wrongfully used qemuarma15 in place of qemuarmv5
Commit 6fc70eb4f3494bee2be10ee24fe3ea1c8b5ff988 changes "qemuarm" machine from
old ARMv5 to a newer Cortex-A15, moving the old platform under "qemuarmv5" name.
When adding this new "old" platform name to COMPATIBLE_MACHINE in linux-yocto
recipes, mistakes were made referring it by "qemuarma15", as discussed in [1].
This change fixes those mistakes.
Ross Burton [Thu, 21 Mar 2019 12:44:09 +0000 (12:44 +0000)]
meson: support multiple cross configuration files
For historical reasons Meson supports only a single cross configuration file but
multiple native configuration files. Add support for multiple cross files, so
that recipes such as glib can use the toolchain cross file and extend it with a
recipe-specific cross file containing values needed to build.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 21 Mar 2019 12:32:17 +0000 (12:32 +0000)]
meson: warn if the default value from meson.get_cross_property() is being used
Projects such as glib have tests that execute code to check their behaviour, and
in cross-compilation environments where they can't execute code fall back on
meson.get_cross_property() with a default value. These default values may not
be appropriate: for example if glib can't verify that vsprintf() isn't broken it
assumes that it is, and embeds the gnulib copy of vsprintf() instead of using
glibc's implementation.
Add a warning whenever the default value for get_cross_property() is used so
that we can tell this is happening and set the value appropriately.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Thu, 21 Mar 2019 12:30:42 +0000 (12:30 +0000)]
insane: improve license checksumming logic
Instead of opening files as bytes and battling decoding to UTF-8 which can throw
exceptions, open directly as strings and replace invalid codepoints. This
handles licenses in encodings which are not UTF-8 but are based on ASCII much
better.
Also instead of extracting the license lines, writing them to a file, and then
hashing the file, hash the lines directly.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Alistair Francis [Tue, 19 Mar 2019 20:47:41 +0000 (20:47 +0000)]
goarch: Add riscv64
Although RISC-V 64-bit doesn't have official golang support there are
forks that now exist with at least some support and work is ongoing in
the upstream tree. In order to be able to use the goarch class add
support for RISC-V.
For more details see here:
https://github.com/golang/go/issues/27532
Signed-off-by: Alistair Francis <alistair.francis@wdc.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Mon, 18 Mar 2019 20:04:10 +0000 (17:04 -0300)]
systemd: Update to systemd-stable v241-stable release
This changes the repository to use the systemd-stable, and update to
the latest release from v241-stable branch.
Following changes are included:
c1f8ff8d0d login: mark nomodeset fb devices as master-of-seat 59f2213e45 login: HyperV requires master-of-seat to be set a09c170122 Allocate temporary strings to hold dbus paths on the heap 4f54afd5a1 Refuse dbus message paths longer than BUS_PATH_SIZE_MAX limit. b22a96ef2f NEWS: add entry about 'udevadm trigger --wait-daemon' bada94eb3e NEWS: fix release date e9f930b2f5 udev-event: make subst_format_var() always provide null-terminated string on success 66320aec80 sd-device: also store properties read from udev database to sd_device::properties_db dffc22c833 udev-rules: update log messages about OWNER= or GROUP= settings on --resolve=names=never
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Tue, 19 Mar 2019 16:36:51 +0000 (13:36 -0300)]
openssl: Upgrade 1.1.1a -> 1.1.1b
License-Update: copyright years updated Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Otavio Salvador [Tue, 19 Mar 2019 16:36:50 +0000 (13:36 -0300)]
openssl: Remove the c_rehash shell re-implementation
We had a c_rehash shell re-implementation being used for the native
package however the ca-certificates now uses the openssl rehash
internal application so there is no use for the c_rehash anymore.
Signed-off-by: Otavio Salvador <otavio@ossystems.com.br> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yeoh Ee Peng [Tue, 19 Mar 2019 01:53:53 +0000 (09:53 +0800)]
oeqa/manual/toaster: updated test id naming
All test id (eg. @alias) inside manual testcase file shall follow the same
test id naming convention from oeqa automated tests (eg. selftest,
runtime, sdk, etc), where the test id consists of
<test_module>.<test_suite>.<test_function>. Furthermore, there shall be
only 1 unique test_module per each manual testcases file, where
test_module match the file name itself.
This file was using test_module name that does not match the file name
itself. Fixed test_module name as well as the test_suite name.
Signed-off-by: Yeoh Ee Peng <ee.peng.yeoh@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tom Rini [Mon, 18 Mar 2019 00:25:48 +0000 (20:25 -0400)]
vim: Update to 8.1.1017
The most current release of vim is now 8.1.1017. The only problem
currently is that the disable_acl_header patch is still not upstream.
Cc: Wenzong Fan <wenzong.fan@windriver.com> Cc: Changqing Li <changqing.li@windriver.com> Signed-off-by: Tom Rini <trini@konsulko.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Khem Raj [Fri, 15 Mar 2019 19:20:05 +0000 (12:20 -0700)]
go: Minor 1.21.1 patch update
Minor revisions
go1.12.1 (released 2019/03/14) includes fixes to cgo, the compiler, the go command, and the fmt, net/smtp, os, path/filepath, sync, and text/template packages. See the Go 1.12.1 milestone on our issue tracker for details.
Mazliana [Mon, 11 Mar 2019 08:30:33 +0000 (16:30 +0800)]
resulttool/manualexecution: To output right test case id
We found that manualexecution does not capture test suite values
correctly if there are more than one test suite in test cases.
After verification has made we found out we should retrieved
full test cases value <test_module.test_suite.test_case> from
oeqa/manual/ json file rather than split it them into new
variables test_suite and test_cases.
Signed-off-by: Mazliana <mazliana.mohamad@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yeoh Ee Peng [Fri, 28 Dec 2018 06:54:44 +0000 (14:54 +0800)]
oeqa/manual/toaster: transfer manual toaster test cases to oeqa
As part of the solution to replace Testopia, manual test cases
need to be migrated to OEQA. These manual test case json files
will serve two use cases.
Use case#1: as input to the future commandline-based manual test
runner script, where this script will display actions and expected
result information in user friendly text, then it will capture
user input for test result and log, finally it will write test
result and log into existing standardize test result json format
from OEQA framework for automated tests.
Use case#2: QA will open and read these json file manually for
planning manual test execution. Any reader interested in
understanding manual test cases will open and read these files.
Signed-off-by: Yeoh Ee Peng <ee.peng.yeoh@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sangeeta jain [Thu, 14 Mar 2019 05:14:18 +0000 (13:14 +0800)]
oeqa/manual/kernel-dev.json: test id updated
All test id (eg. @alias) inside manual testcase file shall follow the same test id
naming convention from oeqa automated tests (eg. selftest, runtime, sdk, etc), where
the test id consists of <test_module>.<test_suite>.<test_function>. Furthermore,
there shall be only 1 unique test_module per each manual testcases file.
This file was using more than 1 unique test_module for testcases. Furthermore,
some of the testcases were defined using different test_suite where it was not needed.
This patch fix the manual testcases file to have only 1 unique test_module
as well as test_suite to simplify test id naming.
Signed-off-by: sangeeta jain <sangeeta.jain@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sangeeta jain [Thu, 14 Mar 2019 04:31:17 +0000 (12:31 +0800)]
oeqa/manual/compliance-test.json: test id updated and missing test cases added
Two changes made in oeqa/manual/compliance-test.json:
1. All test id (eg. @alias) inside manual testcase file shall follow the same test id
naming convention from oeqa automated tests (eg. selftest, runtime, sdk, etc), where
the test id consists of <test_module>.<test_suite>.<test_function>. Furthermore,
there shall be only 1 unique test_module per each manual testcases file.
This file was using more than 1 unique test_module for testcases. Furthermore, some of
the testcases were defined using different test_suite where it was not needed. This patch
fix the manual testcases file to have only 1 unique test_module as well as test_suite to
simplify test id naming.
2. Added 2 missing test cases for Beaglebone Stress Test.
Signed-off-by: sangeeta jain <sangeeta.jain@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
sangeeta jain [Thu, 14 Mar 2019 02:55:23 +0000 (10:55 +0800)]
oeqa/manual/bsp-hw.json: test id updated and obsolete test cases removed
Two changes made in oeqa/manual/bsp-hw.json:
1. All test id (eg. @alias) inside manual testcase file shall follow the same test id naming
convention from oeqa automated tests (eg. selftest, runtime, sdk, etc), where the
test id consists of <test_module>.<test_suite>.<test_function>. Furthermore,
there shall be only 1 unique test_module per each manual testcases file.
This file was using more than 1 unique test_module for testcases.
Furthermore, some of the testcases were defined using different test_suite where it was not
needed. This patch fix the manual testcases file to have only 1 unique
test_module as well as test_suite to simplify test id naming.
2. As per review by Intel and Windriver team, 7 testcases were found obsolete. Removed 7 testcases.
Signed-off-by: sangeeta jain <sangeeta.jain@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Yi Zhao [Wed, 13 Mar 2019 03:26:31 +0000 (11:26 +0800)]
shadow: drop inappropriate patch
The 0001-useradd-copy-extended-attributes-of-home.patch (oe-core commit: eed66e85af5ca6bbdd80cc3d5cf8453e8d8880bc) introduced a runtime failure
when enable SELinux.
When enable SELinux, The directory /home/user will get the extended
attributes of /etc/skel. However, the SELinux lable for /etc/skel is
etc_t which is also copied to /home/user. It will cause the user can not
write their home directory because the SELinux lable for /home/user
should be user_home_dir_t.
See discussion: http://lists.openembedded.org/pipermail/openembedded-core/2018-January/146039.html
The solution at the moment is to drop this patch.
Signed-off-by: Yi Zhao <yi.zhao@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Ross Burton [Wed, 13 Mar 2019 00:37:02 +0000 (00:37 +0000)]
glib-2.0: clean up EXTRA_OEMESON
Instead of spreading the ptest enabling over the recipe, just add a
PACKAGECONFIG to enable the installed tests. Whether this is enabled respects
PTEST_ENABLED by default, so it always disabled in native and nativesdk builds.
This means we can remove the conditional dependency on dbus, all the conditional
enabling of installed tests, and clean up EXTRA_OEMESON.
Signed-off-by: Ross Burton <ross.burton@intel.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Tom Rini [Tue, 12 Mar 2019 20:07:21 +0000 (16:07 -0400)]
inetutils: Handle rsh, etc more consistently
- We cannot build rsh, rshd, rlogind and rcp on musl. This is handled
gracefully in the configure scripts and spelled out with EXTRA_OECONF.
Expand this to include rexec to cover all of the related functionality.
- Rework adding in the xinetd.d files for these services to only do so
when we even have the services being built. This leads to no rsh/rshd
sub-packages on musl at all.
- If we use the normal alternatives mechanism to allow for this or
netkit-rsh to provide rsh/rshd functionality we end up with QA issues
on musl as we have unused ALTERNATIVES logic. Switch to making use of
RPROVIDES / RCONFLICTS logic instead and make it match the netkit-rsh
packaging names.
Cc: Khem Raj <raj.khem@gmail.com> Signed-off-by: Tom Rini <trini@konsulko.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
while compiled with -Werror=maybe-uninitialized/-Werror=format-overflow=,
it failed
[snip]
| Incremental.c: In function 'Incremental_container':
| Incremental.c:1593:3: error: 'mdfd' may be used uninitialized in this function [-Werror=maybe-uninitialized]
| close(mdfd);
| ^~~~~~~~~~~
[snip]
super-intel.c: In function 'apply_takeover_update':
| super-intel.c:9615:15: error: '%d' directive writing between 1 and 11 bytes into a region of size 7 [-Werror=format-overflow=]
| " MISSING_%d", du->index);
| ^~
Signed-off-by: Changqing Li <changqing.li@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Changqing Li [Tue, 12 Mar 2019 05:40:24 +0000 (13:40 +0800)]
glibc: fix do_populate_sdk fail when multilib used
fix below error:
file /usr/include/bits/procfs-id.h conflicts between attempted installs of lib32-libc6-dev-2.29-r0.armv7vet2hf_vfp and libc6-dev-2.29-r0.aarch64
file /usr/include/bits/procfs.h conflicts between attempted installs of lib32-libc6-dev-2.29-r0.armv7vet2hf_vfp and libc6-dev-2.29-r0.aarch64
file /usr/include/bits/shmlba.h conflicts between attempted installs of lib32-libc6-dev-2.29-r0.armv7vet2hf_vfp and libc6-dev-2.29-r0.aarch64
Signed-off-by: Changqing Li <changqing.li@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Jonathan Rajotte [Tue, 12 Mar 2019 21:13:29 +0000 (21:13 +0000)]
lttng-tools ptest: add missing dependencies
Multiple tests are failing due to missing dependencies on a bare
core-image-minimal build with only lttng-tools ptest present.
"getconf LONG_BIT" is used to get the bitness of the host to run the
correct consumerd. Depend on glibc-utils.
perl-module-overloading is needed to validate trace. babelstats.pl
"head -c" is used to generate a random string. It is easier to depends
on coreutils to get a feature complete "head" than to enable to "fancy
head" feature of busybox. See utils.sh randstring.
"taskset" is used in a couple of scripts to ensure that events generated
finish in the same buffer. Depend on util-linux.
[RP: Tweak for glibc/musl handling] Signed-off-by: Jonathan Rajotte <jonathan.rajotte-julien@efficios.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Robert Yang [Mon, 11 Mar 2019 10:16:55 +0000 (18:16 +0800)]
oe-init-build-env: Error out when failed to locate cwd
Ubuntu's /bin/sh symlinks to /bin/dash by default, so
subprocess.check_call(oe-init-build-env, cwd=builddir) would be failed since
pwd is builddir, and there is no $builddir/oe-init-build-env, this would
lead to other confusing errors, check and error it out earlier to make it
easier to locate the problem.
We don't meet the problem when manually run ". oe-init-build-env" is because
Ubuntu's default login shell is bash, but subprocess.check_call() doesn't
respect to login shell, so the error only happens in situations like
subprocess.check_call().
And also print errors to stderr as oe-buildenv-internal does.
Signed-off-by: Robert Yang <liezhi.yang@windriver.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Douglas Royds [Mon, 11 Mar 2019 05:51:51 +0000 (18:51 +1300)]
cmake: Reduce verbosity for make invocation
Since the dawn of time, we have set CMAKE_VERBOSE_MAKEFILE=1 in cmake.bbclass.
Back in 2016, we also explicitly set VERBOSE=1 in cmake_do_compile(),
to ensure that make (and ninja) output were verbose in log.do_compile.
Turning off CMAKE_VERBOSE_MAKEFILE=1 means that make (or ninja)
invocations from the command-line are non-verbose,
giving CMake's default human-readable output on the terminal instead.
The user can still invoke VERBOSE=1 make if they do want verbose output.
This has no effect on the verbose output that goes into the logs.
Signed-off-by: Douglas Royds <douglas.royds@taitradio.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Bruce Ashfield [Mon, 11 Mar 2019 04:20:51 +0000 (00:20 -0400)]
linux-yocto: fix SRC_URI patching with unified repos
As reported by Paul Barker, my attempt to allow the patching
of kernel meta data broke other repository configurations, since
the meta data patch routing was matching too broadly and the
same repo would end up being patched twice.
Using his suggested fix, we are up and running with both types
of repos again.
Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
which makes me wonder if cc-option fails unless -mfpu-vfp is on the commandline too.
Since we have a gcc which accepts the armv7-a arch, just remove the cc-option
wrapper unconditionally here.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
]
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org> Signed-off-by: Bruce Ashfield <bruce.ashfield@gmail.com> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Richard Purdie [Sun, 10 Mar 2019 07:46:10 +0000 (07:46 +0000)]
linux-yocto-4.19: Fix systemtap issue on armv7
Add temporary patch to fix the systemtap issue on armv7 with the 4.19 kernel
until Bruce is able to merge this into the repos, similarly to the fix for
5.0 kernel.
Alternative patches are being discussed upstream but this fixes the issue for
us for now in a simple way since we know the gcc versions we use.
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
