gstreamer1.0-plugin-base: upgrade to version 1.14.3

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>

gstreamer1.0: upgrade to version 1.14.3

Signed-off-by: Carlos Rafael Giani <dv@pseudoterminal.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>

libva: update 2.2.0 -> 2.3.0

>From release notes:

    Bump VA-API version to 1.3.0 and libva to 2.3.0
    Add max frame size parameters for multiple pass case in legacy mode
    Add new BRC mode AVBR
    Add new interface for High Dynamic Range tone mapping
    Add missing enum to string conversions
    Add hevc subsets parameters structure
    Add Customized Noise Reduction (HVS) interfaces
    Add new BRC mode definition QVBR
    Add more complete colour properties for use in VPP

Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sanity: Add check for WSL

Users are starting to expect OE to work under WSL which it doesn't. Add a warning to
tell them about this up front and manage expectations.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

testimage: Enable autorunning of the package manager testsuites

Now that the hangs in httpservice are fixed we can let these tests
auto skip as appropriate.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/utils/httpserver: Rework to avoid hangs and improve logging

testimage.bbclass installs a SIGTERM handler which conflicts with the
use of multiprocessing here. This is paritcularly problematic if the http
service is terminated before its started and hence before its had a chance
to reset the default signal handler (as the code was written).

Instead, temporarily remove testimage's handler whilst forking the http process
which means the correct handler is installed and won't deadlock.

Also take the opportunity to add in some log messages about the server start
and shutdown so that future debugging is easier and its clearer what the code
is doing.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

testimage/testsdk/selftest: Avoid platform.distro_identifier deprecation warnings

Use our own lsb function instead as used elsewhere by the codebase.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/concurrencytest: Avoid unclosed file warnings

Avoid an unclosed file per thread warning when running selftests concurrently
by closing the result stream.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts: Remove deprecated imp module usage

The imp module is deprecated, port the code over to use importlib
as recently done for bb.utils as well.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: skip known bugs in ptest

Set SKIP_KNOWN_BUGS in run-ptest script to skip ptest cases which
are known to relate to some known bugs.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

valgrind: update upstream patch status

Two recent patches are now marked as reported in the
valgrind bugzilla.

Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

wic: bootimg-efi: add a title source parameter

Sometimes the users might want to change the title showing on UEFI
booting screen, so far it's hard-coded to 'boot'.

There is not a easy way to customize it in current design, I tried
firstly with '--configfile', but that does not work with --use-uuid,
since the later option will generate a UUID and write it to boot
config, only when the former option is not enabled.

So a new source parameter 'titile' is added in this patch, it defaults
to 'boot' to be consistent with the original title.

Signed-off-by: Ming Liu <liu.ming50@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass: add a missing blank

When calling d.appendVarFlag, a blank is needed or else it could mess
up the later appended variables.

Reported-by: Stefan Agner <stefan.agner@toradex.com>
Signed-off-by: Ming Liu <liu.ming50@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oe/package_manager: Avoid race problems when calling list_pkgs()

list_pkgs() for rpm calls RpmPM() which would try and create a copy of the
package feed. This can be called for example from buildhistory whilst some
other task may be working on an SDK or image construction, causing the package
feed to disappear part way through an operation.

Avoid the need to copy the package index just to list the installed
packages, avoiding the race.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/utils/qemurunner.py: Fix python regex warnings

Fix the warnings:

meta/lib/oeqa/utils/qemurunner.py:250: DeprecationWarning: invalid escape sequence \.
  ips = re.findall("((?:[0-9]{1,3}\.){3}[0-9]{1,3})", cmdline.split("ip=")[1])
meta/lib/oeqa/utils/qemurunner.py:343: DeprecationWarning: invalid escape sequence \-
  if re.search("root@[a-zA-Z0-9\-]+:~#", output):
poky/meta/lib/oeqa/utils/qemurunner.py:350: DeprecationWarning: invalid escape sequence \-
  if re.search("root@[a-zA-Z0-9\-]+:~#", output):
meta/lib/oeqa/utils/qemurunner.py:448: DeprecationWarning: invalid escape sequence \-
  if re.search("[a-zA-Z0-9]+@[a-zA-Z0-9\-]+:~#", data):

by correctly marking the regexs.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/selftest/context: Replace deprecated imp module usage

Avoid the warning:

meta/lib/oeqa/selftest/context.py:8: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  import imp

In this case importlib is a direct replacement.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oepydevshell-internal.py: decode only when readdata is valid

fix below problem:
pydevshell raises exception when maximize the python shell window.
when click maximize, rlist of select return ready object, but the
pty.read is None, so throw exception of 'NoneType' object has no
attribute 'decode', change to only decode when readdata is valid.

[YOCTO #11875]

Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

module.bbclass: Add INSTALL_FW_PATH for out-of-tree modules

This same variable was added to the make line in kernel.bbclass in
0decf1cc1c35dd70f3b822e3b4291a810a319ba0. With the addition of the
usrmerge feature, out-of-tree modules could use this same variable.

Signed-off-by: Wes Lindauer <wesley.lindauer@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: Provide /bin/ash when usrmerge is enabled

When usrmerge is enabled, scripts that were explicitly using #!/bin/ash
will cause a QA Error like the following:

QA Issue: bar.sh contained in package foo requires /bin/ash,
but no providers found in RDEPENDS_foo? [file-rdeps].

It seems perfectly acceptable for scripts to use /bin/ash so provide
it along with /bin/sh.

Signed-off-by: Wes Lindauer <wesley.lindauer@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

checklayer: generate locked-sigs.inc under builddir

yocto-check-layer will find locked-sigs.inc under builddir,
but locked-sigs.inc is generated under current bitbake working
dir. if run yocto-check-layer outside builddir, we will met error
like "No such file or directory: *locked-sigs.inc". change to
run bitbake -S under builddir to fix this problem.

[YOCTO #12973]

Signed-off-by: Changqing Li <changqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gettext: fix CVE-2018-18751

Backport patch to fix CVE-2018-18751 for gettext. Because po-gram-gen.y
has been modified by fix-CVE-2018-18751.patch, it requires yacc which
provided by bison-native to re-create po-gram-gen.c. Please remove
bison-native from DEPENDS* when next upgrade.

Ref:
https://security-tracker.debian.org/tracker/CVE-2018-18751

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scons.bbclass: fix indention

Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scons.bbclass: Force rebuild if necessary

Signed-off-by: Andreas Müller <schnitzeltony@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tcf-agent: Disable architecture-specific features for ARC

There's no support of architecture-specific features for ARC
in tcf-agent, so disable non-generic stuff for now.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Cc: Richard Purdie <richard.purdie@linuxfoundation.org>
Cc: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

x264: Disable asm on x86

The previous commit as below only fixes the
textrel issue with musl on x86, update it also to
fix the issue such as with glibc on x86.
74ea4f280c1 x264: Disable asm on musl/x86

[YOCTO #11770]

Signed-off-by: Mingli Yu <mingli.yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpcbind: 0.2.4 -> 1.2.5

- Drop backport fixes
  0001-rpcbind-pair-all-svc_getargs-calls-with-svc_freeargs.patch
  pmapproc_dump-Fixed-typo-in-memory-leak-patch.patch
  rpcbproc_callit_com-Stop-freeing-a-static-pointer.patch

- Do not manually move binaries from bindir to sbindir,
  the upstream already moved rpcbind from bin_PROGRAMS
  to sbin_PROGRAMS in Makefile.am
  https://git.linux-nfs.org/?p=steved/rpcbind.git;a=commitdiff;h=9afccfcd5ab350d6bc72622f3d1ccfb9e54652b0

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bluez5: add mesh dependency on ell

Introduced at or before 5.50.

Signed-off-by: Peter A. Bigot <pab@pabigot.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/utils/commands: Avoid unclosed file warnings

Avoid warnings such as:

meta/lib/oeqa/utils/commands.py:213: ResourceWarning: unclosed file <_io.BufferedReader name=4>
  return runCmd(cmd, ignore_status, timeout, output_log=output_log, **options)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/core/runner: Correctly markup regexs

Avoid the warning "DeprecationWarning: invalid escape sequence \(" by marking
the regexs correctly.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/loader: Fix deprecation warning

Clean up the warning:
meta/lib/oeqa/core/loader.py:27: DeprecationWarning: inspect.getargspec() is deprecated, use inspect.signature() or inspect.getfullargspec()
  _failed_test_args = inspect.getargspec(unittest.loader._make_failed_test).args

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runner: Sort the test result output by result class

We want to see failures/errors listed last since this is the most easily
visible part of the log on consoles or autobuilder output and makes
human processing easier rather than having to scroll up and scan for
a single failure.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runner: Always show a summary of success/fail/error/skip counts

Its useful to have the counts of success/failure/error/skipped at the end of the
results to allow for easier human reading of what happened.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

license_image: Fix rootfs file ownership issues with COPY_LIC_DIRS

We need to set ownership of the license files after copying into the rootfs.
We also need to ensure we don't overwrite the rootfs files outside rootfs
generation, hence only write there when the new rootfs function parameter
is True.

[YOCTO #12961]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

locale: Allow usage of cross-localedef for ARC

With this it's possible to build locale data for ARC
and not do it instead on the first boot.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-sdk: Disable SANITIZERS for ARC

Sanitizers are not yet supported on ARC.

Signed-off-by: Alexey Brodkin <abrodkin@synopsys.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

opkg-utils: Fix update-alternatives link relocation

Recently Debian-style support for link relocation was added to
'update-alternatives' script, but it fails under circumstances where
host rootfs root directory differs from target rootfs root directory
and two alternative packages provide a symbolic link with source
located in different directories.

An example of the case is busybox provided /bin/rev (symlinking to
/bin/busybox.nosuid) and util-linux provided /usr/bin/rev (symlinking
to /usr/bin/rev.util-linux) in which case following failure occurs
during image recipe's do_rootfs() task:

  ERROR: core-image-minimal-1.0-r0 do_rootfs: Postinstall scriptlets of ['util-linux'] have failed. If the intention is to defer them to first boot,
  then please place them into pkg_postinst_ontarget_${PN} ().
  Deferring to first boot via 'exit 1' is no longer supported.
  Details of the failure are in .../tmp/work/qemux86-poky-linux/core-image-minimal/1.0-r0/temp/log.do_rootfs.
  ERROR: core-image-minimal-1.0-r0 do_rootfs: Function failed: do_rootfs

Looking in log.do_rootfs file, following relevant lines can be observed:

  update-alternatives: renaming rev link from /bin/rev to /usr/bin/rev
  mv: cannot stat '/bin/rev': No such file or directory

Mitigate issue by applying patch which adds target root filesystem root
directory path prefix to failing 'mv' calls relevant variable references

Signed-off-by: Niko Mauno <niko.mauno@iki.fi>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

testimage: Further cleanup DEFAULT_TEST_SUITES

Now the test markup of the development tools tests is complete, this
can be further tweaked to auto run the correct tests.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runtime/cases: Improve dependencies of kernel/gcc/build tests

Mark up these tests as needing a compiler, make and kernel source code
as appropriate, the image feature requirements can then be retired.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

testimage: Simplfy DEFAULT_TEST_SUITES logic

Now that the tests have correct markup to automatically determine which images
they can run against, clean up the default test suites logic to be simpler
and not image specific.

Some cleanup of the compiler tests still needs to be completed but this
is a good first step.

The only downside to this is more noise during testing as we now see
many skipped messages for simple images like core-image-minimal.

The auto type is being removed since it currently breaks badly due to the
socat mandatory inclusion from the meta-selftest layer which is a problem
which needs to be addressed seperately.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runtime/cases: Improve test dependency information

Add the OEHasPackage decorator to a variety of tests so they determine
automatically if they should run against a given image.

To ensure tests can do this we need to move target operations such
as scp commands into the tests and out of the class startup/teardown.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rng-tools: fix typo of unrecognised options: --disable-jitterntropy

- Fix typo of `--disable-jitterntropy' in both of PACKAGECONFIG and configure --help

- Enable jitterntropy for musl

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

maintainers.inc: Add entry for libjitterentropy (copied from rng-tools)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libjitterentropy: add 2.1.2

>From the following kernel commit we know: Fedora has integrated the jitter
entropy daemon to work around slow boot problems, especially on VM's that
don't support virtio-rng

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=81e69df38e2911b642ec121dec319fad2a4782f3

Do the same thing for Yocto.

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rng-tools: 5 -> 6.6

- Update SRC_URI, previously upstream is dead, latest update is 2014

- Drop read_error_msg.patch

- Rework do_configure and do_compile

- Enable jitterentropy

License-Update: Upstream changed, but license is still gplv2

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

machine/qemu*: fix kernel finish crng init more and more slowly

Just adding `-device virtio-rng-pci' to the QEMU invocation will
add the device with a default host backend. As of QEMU 1.3+,
the default backend is to use the host's /dev/random as a
source of entropy. [1]

When the entropy pool is empty, reads from /dev/random will
block until additional environmental noise is gathered. [2]

For Yocto, if call runqemu frequently, it will consume lots
of host's /dev/random, and kernel finish crng init in guest get
more and more slowly.

Here are 4 times runqemu boot test:

[    3.464432] random: crng init done
[   20.874030] random: crng init done
[   23.583589] random: crng init done
[   23.858945] random: crng init done

Modify entropy source to /dev/urandom device on the host which
returns random bytes using a pseudorandom number generator seeded
from the entropy pool. Reads from this device do not block and
kernel finish crng init in guest will not delay.

Of course, the side effect is obviously, we lost the quality of
randomness, but the modification is only on runqemu script
rather than real embedded device, and it benefits oeqa efficiency
in which many cases call runqemu especially multiple oeqa builds
on one host.

After apply the fix:

[    3.364670] random: crng init done
[    4.619061] random: crng init done
[    3.403897] random: crng init done
[    3.450717] random: crng init done

[1] https://wiki.qemu.org/Features/VirtIORNG
[2] http://man7.org/linux/man-pages/man4/random.4.html

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rng-tools: Fix crazy defaults

Since commit [f1dc9ac rng-tools: Fix crazy defaults] fixed
init based on sysvinit, this fix rngd.service based on systemd.

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

testsdk: fix skipped testcase output "UNKNOWN" status while multiprocess execution

Usually skipped testcase output "SKIPPED"
[snip serial execution]
|RESULTS - buildgalculator.GalculatorTest.test_galculator - Testcase -1: SKIPPED (0.01s)
|RESULTS - python.PythonTest.test_python3 - Testcase -1: SKIPPED (0.01s)
[snip serial execution]

But if enable multiprocess execution, skipped testcase output "UNKNOWN" status
[snip enable multiprocess execution]
|RESULTS - buildgalculator.GalculatorTest.test_galculator - Testcase -1: UNKNOWN
|RESULTS - python.PythonTest.test_python3 - Testcase -1: UNKNOWN
[snip enable multiprocess execution]

Here is my investigation:

There is a class pairs TestProtocolClient and TestProtocolServer
provided by python3-subunit. The TestProtocolClient generates a
subunit stream of TestResult from a test run, and TestProtocolServer
parses the stream of subunit TestResult.

The class ProtocolTestCase is a unittest.TestCase adapter and it
uses TestProtocolServer to parse the stream of subunit TestResult.

In Yocto testsdk, it forks multiple processes to execute testcases
and use TestProtocolClient to generate TestResult stream; and then
it creates multiple threads to use ProtocolTestCase to parse stream
of subunit TestResult through pipe; finally it passes multiple
ProtocolTestCase as TestCase instance to main process and output
status result.

The problem point is TestProtocolServer parses `skip:' directive
after reading a `test:' directive. Without `test:' directive,
`skip:' directive will be ignored. All above requires SkipTest should
be raised inside a test method rather than setUpClass method.

Throwing SkipTest inside setUp works correctly

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libaio: Extend to native

lvm2 currently requires libaio. So building lvm2-native will result in
the following error.

  ERROR: Required build target 'lvm2-native' has no buildable providers.
  Missing or unbuildable dependency chain was: ['lvm2-native', 'libaio-native']

Extend libaio to native to fix this issue.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libsm: extend to nativesdk

After enable tk via PACKAGECONFIG for python3,
there comes below error:
Missing or unbuildable dependency chain was: ['buildtools-tarball', 'nativesdk-python3-testtools', 'nativesdk-python3-pbr', 'nativesdk-python3', 'nativesdk-tk', 'nativesdk-libxt', 'nativesdk-libsm']

Extend libsm to nativesdk to fix this issue.

Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libice: extend to nativesdk

After enable tk via PACKAGECONFIG for python3,
there comes below error:
Missing or unbuildable dependency chain was: ['buildtools-tarball', 'nativesdk-python3-git', 'nativesdk-python3-gitdb', 'nativesdk-python3-async', 'nativesdk-python3', 'nativesdk-tk', 'nativesdk-libxt', 'nativesdk-libsm', 'nativesdk-libice']

Extend libice to nativesdk to fix this issue.

Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

fontconfig: extend to nativesdk

After enable tk via PACKAGECONFIG for python3,
there comes below error:
Missing or unbuildable dependency chain was: ['nativesdk-packagegroup-sdk-host', 'nativesdk-opkg', 'nativesdk-libsolv', 'nativesdk-rpm', 'nativesdk-python3', 'nativesdk-tk', 'nativesdk-xft', 'nativesdk-fontconfig']

Extend fontconfig to nativesdk to fix this issue.

Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxft: extend to nativesdk

After enable tk via PACKAGECONFIG for python3,
there comes below error:
Missing or unbuildable dependency chain was: ['buildtools-tarball', 'nativesdk-python3-testtools', 'nativesdk-python3-pbr', 'nativesdk-python3', 'nativesdk-tk', 'nativesdk-xft']

Extend libxft to nativesdk to fix this issue as
PROVIDES = "xft" in libxft recipe.

Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxt: extend to nativesdk

After enable tk via PACKAGECONFIG for python3,
there comes below error:
Missing or unbuildable dependency chain was: ['buildtools-tarball', 'nativesdk-python3-pytz', 'nativesdk-python3', 'nativesdk-tk', 'nativesdk-libxt']

Extend libxt to nativesdk to fix this issue.

Signed-off-by: Mingli Yu <Mingli.Yu@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

testimage: Remove duplicate dependencies

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runtime/utils/targetbuildproject: Use a subdir within ~/

Without doing this, the code can and sometimes does try and do "rm ~/"
which I think we'd all prefer it didn't.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/utils/buildproject: Only clean files if we've done something

We should only be wiping out things on target if the tests have
actually run.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

distcc: Handle upstream branch deletion for old releases

distcc upstream dropped the 3.2 branch which we reference in older project releases
the revisions are there, just the branch is not. In order to be able to continue
to build those old releases, adjust any mirror tarball to contain the missing branch.
We have updated the  newer point releases but the previous tags were failing to build.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/manual/build-appliance: Drop test which is already automated

The nightly-build-appliance target already tests this particular build.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base.bbclass: Display name of licenses which caused SkipRecipe

Display the name of the restricted licenses which caused the recipe to
be skipped. This makes it easy to determine which license or licenses
are missing and need to be checked and whitelisted.

Signed-off-by: Nathan Rossi <nathan@nathanrossi.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

openssl-1.1.1: remove build path from version info

The openssl build system generates buildinf.h containing the full
compiler command line used to compile objects. This breaks
reproducibility, as the compile command is baked into libcrypto, where
it is used when running `openssl version -f`.

Add stripped build variables for the compiler and cflags lines, and use
those when generating buildinfo.h.

This is based on a similar patch for older openssl versions:
https://patchwork.openembedded.org/patch/147229/

Signed-off-by: Martin Hundebøll <martin@geanix.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cpan.bbclass: avoid newer MakeMaker run perl.real

In newer versions of ExtUtils::MakeMaker, the location of the perl
binary to run is determined differently which results in finding
${bindir}/perl.real instead of ${bindir}/perl

Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

busybox: make busybox.links.{suid, nosuid} reproducible

The busybox.link.* files are generated from autoconf.h and applets.h,
which are both auto-generated by the build system. The contents of the
two files might be in different order, and so the link files are not
reproducble as is.

Fix this by sorting the lists using `sort`.

Signed-off-by: Martin Hundebøll <martin@geanix.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

oeqa/runtime: Add tests for gstreamer CLI

We should not be finding GStreamer issues right around
release time but that's what we seem to do.

Port gstreamer tests from Intel Ref Kit.
https://github.com/intel/intel-iot-refkit/blob/master/meta-iotqa/
lib/oeqa/runtime/multimedia/gstreamer/gstreamer_cli.py

Add tests for gstreamer commandline tools (gst-inspect-1.0
& gst-launch-1.0).

[YOCTO #11418]

Signed-off-by: Yeoh Ee Peng <ee.peng.yeoh@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/autobuilder-worker-prereq-tests: Shore up qemu testing

Check that yocto-autobuilder-helper has been cloned to the correct location.
Check that vnc is running using the same script the autobuilder does.
Set the DISPLAY enviroment variable to :1 the same way we do when building
normally.
Make the VM's serial console available so we can log in and power off allowing
the tests to continue.

Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/runqemu: Replace subprocess.run() for compatibilty

subprocess.run() was introduced in Python 3.5. We currently support down to
Python 3.4 so I've replaced it with subprocess.check_call() which is available
in that version.

Signed-off-by: Michael Halstead <mhalstead@linuxfoundation.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/runqemu: Add serialstdio mode

Its currently not possible to have a console available whilst using qemu in graphics
mode. This is causing some issues for testing autobuilder bringup so all a "serialstdio"
mode to runqemu to accomodate this.

The existing serialstdio internal variable is renamed to allow the new user visible option.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sqlite3_3: Update to 3.25.2

Update SQLite3 from 3.23.1 to 3.25.2 for UPSERT, window
functions and improved ALTER TABLE support.

For a detailed list of changes since 3.23.1, see
* https://www.sqlite.org/releaselog/3_25_2.html
* https://www.sqlite.org/releaselog/3_24_0.html

Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd-systemctl-native: handle Install wildcards

Handle the %i wildcard appearing in a dependency in the Install section
of a template unit, e.g.

  $ cat foo@.service

  [Install]
  WantedBy=bar@%i.target

Using the real systemctl something like:

  $ systemctl enable foo@baz.service

will create a symlink in /etc/systemd/system/bar@baz.target.wants.
Detect wildcards in templates and make the appropriate substitution.

Signed-off-by: Brad Bishop <bradleyb@fuzziesquirrel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pseudo: fix link with new sqlite3

If sqlite3 is built with FTS5 it uses log() from libm, but that linkage is lost
if we manually statically link so explicitly link to libm.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glide: disable ptest on mips

It fails to run task compile_ptest_base of glide on mips:

| vet config not found

The root cause is that go lang lacks '-buildmode=pie' support for mips.
So disable glide ptest on mips for now.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rootfs: ensure that packages in IMAGE_INSTALL_DEBUGFS are existing

Ensure that additional packages in IMAGE_INSTALL_DEBUGFS are available
before installation.

Signed-off-by: Andrej Valek <andrej.valek@siemens.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

metadata_scm: Drop orphaned monotone functions

Its been 'a while' since we used monotone, drop these revision
function remnants.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

metadata_scm.bbclass: simplify base_detect_revision()

The previous code looks strange because of historical reason, simply return
base_get_metadata_git_revision/branch() is enough, the later functions has
handled everything like '<unknown>' and strip().

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

shadow: improve reproducibility by hard-coding shell path

The shadow configure script tries really hard to detect the running
shell to make sure it doesn't do unsupported calls.

On my system the shell is detected as /bin/sh, while a build in an
ubuntu docker it resolves to /bin/bash. And since the shell path is
baked into the target binaries through config.h, the build becomes
inreproducible.

Fix reproducibility by hard-coding the shell to be /bin/sh

Signed-off-by: Martin Hundebøll <martin@geanix.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

go 1.9/1.11: fix textrel qa warning for non mips arch

While building go itself, the go build system does not support
to set `-buildmode=pie' from environment.

Add GOBUILDMODE to support it which make PIE executables the default
build mode, as PIE executables are required as of Yocto

But mips doesn't support -buildmode=pie, so skip the QA checking for
mips and its variants

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pkgconfig: export variables

Export the PKG_CONFIG_PATH and PKG_CONFIG_LIBDIR variables in case the
parent shell hasn't done so.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ppce500v2: remove spe

*The PowerPC SPE support is split off to a separate powerpcspe port

Signed-off-by: Chunrong Guo <chunrong.guo@nxp.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

strace: Bump to version 4.25

V4L2_CID_USER_IMX_BASE value differs in v4.14 and v4.19 LTS kernels
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/include/uapi/linux/v4l2-controls.h?h=v4.14.79#n191
https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/tree/include/uapi/linux/v4l2-controls.h?h=v4.19.1#n191

Building strace v4.24 against 4.19 kernel headers results in following
build failure

  | arm-vaisala-linux-gnueabi-ranlib libstrace.a
  | In file included from ../strace-4.24/static_assert.h:31,
  |                  from ../strace-4.24/xlat/v4l2_pix_fmts.h:4,
  |                  from ../strace-4.24/v4l2.c:78:
  | ../strace-4.24/xlat/v4l2_control_id_bases.h:36:1: error: static assertion failed: "V4L2_CID_USER_IMX_BASE != (V4L2_CID_BASE + 0x1090)"

This Linux kernel ABI breakage that occurred wrt V4L2_CID_USER_IMX_BASE
has been observed and addressed in strace upstream commit
https://github.com/strace/strace/commit/b89a69dec27cf638df0e17db80ed937c3e1abf77

Bump strace to version 4.25 which also addresses ABI breakage so that
strace once again builds against both v4.14 and v4.19 LTS kernel headers

Signed-off-by: Niko Mauno <niko.mauno@vaisala.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gpgme: 1.11.1 -> 1.12.0

Rebase 0007-qt-python-Add-variables-to-tests.patch

Rework 0007-qt-python-Add-variables-to-tests.patch -> 0007-python-Add-variables-to-tests.patch

Drop newly added option --disable-lang-qt-test which is merged to --disable-gpg-test by
upstream commit [26820ba qt: Respect --disable-gpg-test for tests]

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

man-db: 2.8.3 -> 2.8.4

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

ncurses: upgrade 6.1 -> 6.1+20180714

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gnupg: upgrade 2.2.9 -> 2.2.10

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

apr: upgrade 1.6.3 -> 1.6.5

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xf86-video-intel: update to latest

For changes, please see:

https://cgit.freedesktop.org/xorg/driver/xf86-video-intel/log/?qt=range&q=e4fe79cf0d9a05ee3f3a027148ef0aeb2b1b34e1...0932a6b37ba6d5c9e916a1cb6ab89c3205b81a0c

Enable sna by default and remove upstreamed patches.

Also include a patch from fedora to fix compile issues when using
qemux86 which doesn't enable sse2 leading to gcc refusing to
inline vertex_emit_2s in emit_vertex because they are defined as:

  static __attribute__((always_inline)) void
  vertex_emit_2s(struct sna *sna, int16_t x, int16_t y)

  __attribute__((target("sse2,fpmath=sse"))) __attribute__((always_inline))
  static void emit_vertex(/* omitted */)

leading to errors like:

| In file included from ../../../git/src/sna/gen4_vertex.c:34:
| ../../../git/src/sna/gen4_vertex.c: In function 'emit_vertex':
| ../../../git/src/sna/sna_render_inline.h:40:26: error: inlining failed in call to always_inline 'vertex_emit_2s': target specific option mismatch
|  static force_inline void vertex_emit_2s(struct sna *sna, int16_t x, int16_t y)
|                           ^~~~~~~~~~~~~~
| ../../../git/src/sna/gen4_vertex.c:308:25: note: called from here
|  #define OUT_VERTEX(x,y) vertex_emit_2s(sna, x,y) /* XXX assert(!too_large(x, y)); */
|                          ^~~~~~~~~~~~~~~~~~~~~~~~
| ../../../git/src/sna/gen4_vertex.c:360:2: note: in expansion of macro 'OUT_VERTEX'
|   OUT_VERTEX(dstX, dstY);
|   ^~~~~~~~~~

Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman: update to 1.36

- updated connman to v1.36
- removed mainstreamed patches
- includes.patch has been rabased and transformed into git format

Signed-off-by: Oleksandr Kravchuk <dev@sashko.rv.ua>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xserver-xorg: update to 1.20.3

1.20.3 fixes arm booting in testimage

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

kernel: use olddefconfig as the primary target for KERNEL_CONFIG_COMMAND

As was warned by commit 312ee68752fa [kconfig: announce removal of
oldnoconfig if used], oldnoconfig has been removed from the 4.20 kernel.

So we switch our default mode to olddefconfig.

commit fb16d8912 [kconfig: replace 'oldnoconfig' with 'olddefconfig',
and keep the old name as an alias] introduced olddefconfig in the 3.10
kernel, we shuffle oldnoconfig to the fallback target.

The fallback mode allows kernels between 3.10 and the currently listed
oldest kernel of 3.2 to continue to configure.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bind: update to latest LTS 9.11.5

includes:
CVE-2018-5738

drop patch for CVE-2018-5740 now included in update

see: https://ftp.isc.org/isc/bind9/9.11.5/RELEASE-NOTES-bind-9.11.5.html

Add RECIPE_NO_UPDATE_REASON for lts

Signed-off-by: Armin Kuster <akuster808@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dhcp: fix issue with new bind changes

Signed-off-by: Armin Kuster <akuster808@gmail.com>

xf86-input-libinput: update to 0.28.1

Signed-off-by: Armin Kuster <akuster808@gmail.com>

xkeyboard-config: update to 2.25

Signed-off-by: Armin Kuster <akuster808@gmail.com>

libxcb: update to 1.13.1

Signed-off-by: Armin Kuster <akuster808@gmail.com>

libx11: update to 1.6.7 and sytle cleanup

refresh Fix-hanging-issue-in-_XReply.patch

Signed-off-by: Armin Kuster <akuster808@gmail.com>

libsm: update to 1.2.3

reoder things per style

Signed-off-by: Armin Kuster <akuster808@gmail.com>

openssh: Update to 7.9 primarily bug fixes

see: https://lwn.net/Articles/768991/

Signed-off-by: Armin Kuster <akuster808@gmail.com>

libpcre2: update to 10.32

LIC_FILES_CHKSUM changed do to typo fixes and tidies for 10.32

see: https://www.pcre.org/changelog.txt

Signed-off-by: Armin Kuster <akuster808@gmail.com>

gnutls: update to 3.6.4

Notable change:

libgnutls: Added the final (RFC8446) version numbering of the TLS1.3 protocol.
see: https://lists.gnupg.org/pipermail/gnutls-help/2018-September/004457.html

Signed-off-by: Armin Kuster <akuster808@gmail.com>
--
[v2]
Fix typo in version in subject

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nspr: Update to 4.20

RISC-v support now included in this release so drop patch.
and refresh patches

Signed-off-by: Armin Kuster <akuster@mvista.com>

curl: actually apply latest CVE patches

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

unzip: actually apply CVE-2018-18384

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>