]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
7 years agobase: Add MultiConfigParsed handler to deal with unstable build signatures
Richard Purdie [Thu, 6 Jul 2017 11:54:15 +0000 (12:54 +0100)]
base: Add MultiConfigParsed handler to deal with unstable build signatures

This uses the newly added MultiConfigParsed event to handle problems where
checksums in multiconfig build were not fuctioning as expected.

The issue arises around SIGGEN_EXCLUDE_SAFE_RECIPE_DEPS which contains
entries like:

* gcc-cross-${TARGET_ARCH}->virtual/${TARGET_PREFIX}libc-for-gcc
* gcc-cross-${TARGET_ARCH}->linux-libc-headers

These need to be expanded in the multiconfig data store but then placed
into the shared main datastore used by the siggen code. The only other
alternative would be a siggen instance for each multiconfig however that
seemed even more complex and invasive.

In real world usage, this issue would mean a qemux86 base config with
other armv5 and armv7 configs (e.g. beaglebone and qemuarm) would
try and build gcc-cross twice since dependencies normaly excluded
(e.g. linux-libc-headers) would now be included. This breaks sstate reuse
as well as breaking builds unless separate tmpdirs are used.

This patch adds all the entries for each multiconfig. Whilst there may
be duplicates, this shouldn't be an issue.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoimage.bbclass: create root symlinks in nativesdk target sysroot
Amarnath Valluri [Thu, 15 Jun 2017 10:32:39 +0000 (13:32 +0300)]
image.bbclass: create root symlinks in nativesdk target sysroot

To match with the target image, the generated nativesdk's target sysroot also
should have the root symlinks(/lib*, /bin/, /sbin). So, extended the
'create_merged_usr_symlinks' to reuse symlink creation part even for nativesdk
using POPULATE_SDK_PRE_TARGET_COMMAND.

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoinsane.bbclass: Add package QA check for merged /usr.
Amarnath Valluri [Wed, 14 Jun 2017 11:30:46 +0000 (14:30 +0300)]
insane.bbclass: Add package QA check for merged /usr.

This check makes sure that, when usrmerge distro feature enabled, no package
installs files to root (/bin, /sbin, /lib, /lib64) folders.

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoimage: create symlinks needed for merged /usr
Amarnath Valluri [Wed, 14 Jun 2017 11:30:45 +0000 (14:30 +0300)]
image: create symlinks needed for merged /usr

Prepare the symlinks required for merged /usr at the time of rootfs creation.

The links created in rootfs are:
/bin --> /usr/sbin
/sbin --> /usr/sbin
/lib --> /usr/lib
/lib64 --> /usr/lib64

We cannot make these symlinks as part of 'base-files' or some other package.
Because at rootfs creation, installation of the package(say kernel) that depends
on these root folders/links fails, if package manager installs this package
prior to base-files.

These symbolic links in top level folder should be present as long as
 - kernel tools use /lib/{module,firmware}
 - shell scripts uses "#!/bin/sh"

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agosystemd: changes to support merged /usr
Amarnath Valluri [Wed, 14 Jun 2017 11:30:44 +0000 (14:30 +0300)]
systemd: changes to support merged /usr

- Enable/disable the split-usr support in systemd based on 'usrmerge'
  DISTRO_FEATURE.
- Modify rootprefix to point to ${root_prefix}, rather than ${base_prefix}.
- And fixed firmware path to use ${nonarch_base_libdir} instead of hard-coded
  '/lib', because when 'usrmege' distro feature enabled this path would be
  '/usr/lib'.

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agocross.bbclass: merged /usr support
Amarnath Valluri [Wed, 14 Jun 2017 11:30:43 +0000 (14:30 +0300)]
cross.bbclass: merged /usr support

Use ${root_prefix} instead of ${base_prefix} while setting
${target_base_prefix}, otherwise we might loose the root prefix configuration
change in case of 'usrmerge' distro feature is enabled.

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agobitbake.conf: support for merged usr with DISTRO_FEATURE usrmerge
Amarnath Valluri [Wed, 14 Jun 2017 13:05:55 +0000 (16:05 +0300)]
bitbake.conf: support for merged usr with DISTRO_FEATURE usrmerge

A new configuration variable ${root_prefix} added, which shall be used by all
base_{lib,bin,sbin}dir variables. When usrmerge DISTRO_FEATURE is enabled
${root_prefix} points to ${exec_prefix} otherwise to ${base_prefix}

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agospeex: update SRC_URI
Ross Burton [Thu, 6 Jul 2017 11:43:23 +0000 (12:43 +0100)]
speex: update SRC_URI

The US download mirror appears to be down, so use the main host instead.

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoavahi-ui: reduce local pending patches
Dengke Du [Wed, 5 Jul 2017 08:33:29 +0000 (04:33 -0400)]
avahi-ui: reduce local pending patches

[Yocto #11548]

Signed-off-by: Dengke Du <dengke.du@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomirrors: Add HTTP mirrors for ftp://sourceware.org
Soren Brinkmann [Tue, 4 Jul 2017 17:29:07 +0000 (10:29 -0700)]
mirrors: Add HTTP mirrors for ftp://sourceware.org

Signed-off-by: Soren Brinkmann <soren.brinkmann@xilinx.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolocal.conf.sample: drop image-swab reference
Ming Liu [Tue, 4 Jul 2017 08:32:21 +0000 (10:32 +0200)]
local.conf.sample: drop image-swab reference

The image swabber had been dropped by commit e18657df:
[ meta: Drop swabber ]

We need also drop its reference in local.conf.

Signed-off-by: Ming Liu <peter.x.liu@external.atlascopco.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoltp: add acl, attr, curl and util-linux runtime dependencies
Fathi Boudra [Wed, 5 Jul 2017 11:13:11 +0000 (14:13 +0300)]
ltp: add acl, attr, curl and util-linux runtime dependencies

* Use a list and re-order alphabetically to make it easier to read and
  update.
* Add missing runtime dependencies:
  - acl (getfacl/setfacl commands are required)
  - attr (getfattr/setfattr commands are required)
  - iproute2 (required for netns tests)
  - curl (curl command is required)
  - util-linux (ipcs command is required)

Reported-by: Naresh Kamboju <naresh.kamboju@linaro.org>
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoltp: Reduce local Pending patches
Dengke Du [Tue, 4 Jul 2017 03:36:21 +0000 (23:36 -0400)]
ltp: Reduce local Pending patches

Rebase the patch on latest ltp upstream:

    0007-fix-__WORDSIZE-undeclared-when-building-with-musl.patch

[Yocto #11548]

Signed-off-by: Dengke Du <dengke.du@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoltp: syscalls/add_key02: fix for nonempty NULL payload
Fathi Boudra [Mon, 3 Jul 2017 11:45:34 +0000 (14:45 +0300)]
ltp: syscalls/add_key02: fix for nonempty NULL payload

Add upstream patch to fix syscalls/add_key02 test for nonempty
NULL payload:
https://github.com/linux-test-project/ltp/commit/25045624e941ee76a13febd36187e23c6c435507

Signed-off-by: Milosz Wasilewski <milosz.wasilewski@linaro.org>
Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibgfortran: Add missing fincludes
Ricardo Ribalda Delgado [Mon, 3 Jul 2017 21:57:43 +0000 (23:57 +0200)]
libgfortran: Add missing fincludes

ERROR: libgfortran-6.3.0-r0 do_package: QA Issue: libgfortran:
Files/directories were installed but not shipped in any package:
/usr/lib/gcc/x86_64-poky-linux/6.3.0/finclude
/usr/lib/gcc/x86_64-poky-linux/6.3.0/finclude/ieee_features.mod
/usr/lib/gcc/x86_64-poky-linux/6.3.0/finclude/ieee_arithmetic.mod
/usr/lib/gcc/x86_64-poky-linux/6.3.0/finclude/ieee_exceptions.mod

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibgfortran: Add missing dependency gcc-cross
Ricardo Ribalda Delgado [Mon, 3 Jul 2017 21:57:42 +0000 (23:57 +0200)]
libgfortran: Add missing dependency gcc-cross

Due to the fact that the recipe uses INHIBIT_DEFAULT_DEPS, we need to
manually add the cross compiler as a dependency.

Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agosystemd: Do not use xlocale.h
Khem Raj [Mon, 3 Jul 2017 15:59:27 +0000 (08:59 -0700)]
systemd: Do not use xlocale.h

glibc specific header which has been removed from glibc 2.26+

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomesa: Upgrade to 17.1.4 release
Otavio Salvador [Mon, 3 Jul 2017 20:02:10 +0000 (17:02 -0300)]
mesa: Upgrade to 17.1.4 release

This includes following upstream bug fixes:

Bug 77240 - khrplatform.h not installed if EGL is disabled
Bug 95530 - Stellaris - colored overlay of sectors doesn't render on i965
Bug 96958 - [SKL] Improper rendering in Europa Universalis IV
Bug 99467 - [radv] DOOM 2016 + wine. Green screen everywhere (but can be started)
Bug 101071 - compiling glsl fails with undefined reference to `pthread_create'
Bug 101252 - eglGetDisplay() is not thread safe
Bug 101294 - radeonsi minecraft forge splash freeze since 17.1
Bug 101451 - [G33] ES2-CTS.functional.clipping.polygon regression

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomesa: Avoid platform probing when building without EGL
Otavio Salvador [Mon, 3 Jul 2017 20:02:09 +0000 (17:02 -0300)]
mesa: Avoid platform probing when building without EGL

The 17.1.2 release has changed the platform setting and when not
explicitly disabled it assumes x11 support.

Fixes:

| checking for x11-xcb xcb xcb-dri2 >= 1.8 xcb-xfixes... no
| configure: error: Package requirements (x11-xcb xcb xcb-dri2 >= 1.8 xcb-xfixes) were not met:
|
| No package 'x11-xcb' found
| No package 'xcb' found
| No package 'xcb-dri2' found
| No package 'xcb-xfixes' found
|
| Consider adjusting the PKG_CONFIG_PATH environment variable if you
| installed software in a non-standard prefix.
|
| Alternatively, you may set the environment variables XCB_DRI2_CFLAGS
| and XCB_DRI2_LIBS to avoid the need to call pkg-config.
| See the pkg-config man page for more details.

The issue has been exposed by meta-freescale BSP. Fix tested with
imx6qsabresd machine.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agosanity.bbclass: fix AttributeError in mirror format checks
Mikko Ylinen [Mon, 3 Jul 2017 11:49:55 +0000 (14:49 +0300)]
sanity.bbclass: fix AttributeError in mirror format checks

mirrors is a list after split() and results in:

AttributeError: 'list' object has no attribute 'strip'

when the 'mirror values are pairs' check fails.

Signed-off-by: Mikko Ylinen <mikko.ylinen@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agooe-pkgdata-util: package-info: Allow extra variables to be displayed
Peter Kjellerstedt [Mon, 3 Jul 2017 10:14:59 +0000 (12:14 +0200)]
oe-pkgdata-util: package-info: Allow extra variables to be displayed

By specifying the -e <var> option one or multiple times, extra
variables available in the pkgdata can be displayed, e.,g,
`oe-pkgdata-util package-info -e SUMMARY -e LICENSE ...`. The extra
variables displayed are quoted since some of them may contain
whitespace.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoexpat: upgrade to 2.2.1
Dengke Du [Mon, 3 Jul 2017 09:36:52 +0000 (05:36 -0400)]
expat: upgrade to 2.2.1

The COPYING file in expat has the following changes:

    2001-20016 to 2001-2017

Signed-off-by: Dengke Du <dengke.du@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogrep: upgrade to 3.1
Dengke Du [Mon, 3 Jul 2017 08:45:47 +0000 (04:45 -0400)]
grep: upgrade to 3.1

Signed-off-by: Dengke Du <dengke.du@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoclasses/populate_sdk_base: Fix SDK manifest generation
Aníbal Limón [Fri, 30 Jun 2017 21:07:30 +0000 (16:07 -0500)]
classes/populate_sdk_base: Fix SDK manifest generation

The write_{host,target}_sdk_manifest don't need to be set as
SDK_POSTPROCESS_COMMAND because are already append to
POPULATE_SDK_POST_{HOST,TARGET}_COMMAND and the if doesn't
make sense because are only added for populate sdk task.

[YOCTO #11741]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agovalgrind: Remove -no-pie from cflags
Khem Raj [Wed, 14 Jun 2017 14:55:21 +0000 (07:55 -0700)]
valgrind: Remove -no-pie from cflags

It has been moved to distro security include file

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoicu: Fix build with glibc 2.26
Khem Raj [Sun, 25 Jun 2017 05:55:03 +0000 (22:55 -0700)]
icu: Fix build with glibc 2.26

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoepiphany: Fix build errors when compiling with security flags
Khem Raj [Thu, 29 Jun 2017 04:59:05 +0000 (21:59 -0700)]
epiphany: Fix build errors when compiling with security flags

We need to have some level of optimization level for FORTIFY_SOURCE
to work therefore set --enable-debug=no

Add a patch to fix build error due to unused result of fread

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoqemu: Replace use of struct ucontext with ucontext_t
Khem Raj [Thu, 29 Jun 2017 04:59:04 +0000 (21:59 -0700)]
qemu: Replace use of struct ucontext with ucontext_t

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agostrace: upgrade to 4.17
Khem Raj [Thu, 29 Jun 2017 04:59:03 +0000 (21:59 -0700)]
strace: upgrade to 4.17

Fix build with upcoming glibc 2.26

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agovalgrind: Fix build with glibc 2.26
Khem Raj [Thu, 29 Jun 2017 04:59:02 +0000 (21:59 -0700)]
valgrind: Fix build with glibc 2.26

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agobluez: Correct the timer count for bcm43xx firmware download
Jun Zhu [Thu, 29 Jun 2017 02:14:55 +0000 (10:14 +0800)]
bluez: Correct the timer count for bcm43xx firmware download

bcm43xx failed as time out for firmware downloading.
The root cause is that it need wait 50ms to download firmware,
but the value of 50us is set to the timer.

Signed-off-by: Jun Zhu <junzhu@nxp.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agobinutils: update SRCREV to fix powerpc gold link bug
Randy MacLeod [Wed, 28 Jun 2017 01:35:06 +0000 (21:35 -0400)]
binutils: update SRCREV to fix powerpc gold link bug

Update to the latest commit on the 2.28 branch to pick up:
   a3e00a8 [GOLD] Avoid duplicate PLT stub symbols on ppc32
along with several backported bug fixes for arm*, etc.

This resolves the systemd link error for qemuppc
with the error message:
   .../powerpc-oe-linux-ld.gold:
   error: invalid STB_LOCAL symbol in external symbols
   error: linker defined: multiple definition of '00000001.plt_call.memcpy+8000'
   command line: previous definition here

described in:

[YOCTO #11696]

Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoyocto-compat-layer.py: make signature check code reusable
Patrick Ohly [Tue, 27 Jun 2017 15:33:43 +0000 (17:33 +0200)]
yocto-compat-layer.py: make signature check code reusable

This moves the main content of test_signature into a helper
function. It can be reused by arbitrary tests that need to do
a before/after signature comparison. Long-term this might even
be useful in oeqa itself.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoyocto-compat-layer.py: allow README with suffix
Patrick Ohly [Tue, 27 Jun 2017 15:33:42 +0000 (17:33 +0200)]
yocto-compat-layer.py: allow README with suffix

It may be useful to append a suffix denoting the file format. For
example, README.rst is rendered differently when viewed on Github, and
also helps editors to switch to a mode more suitable for the format.

The tests uses a file pattern to find the README file(s) and treats
the one with the shortest name as the main one which must not be
empty.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoyocto-compat-layer.py: add test_world
Patrick Ohly [Tue, 27 Jun 2017 15:33:41 +0000 (17:33 +0200)]
yocto-compat-layer.py: add test_world

"test_signatures" ignores wold build breakage for the sake of
reporting differences also when a world build is broken. Therefore we
need a dedicated test that a world build at least theoretically can
proceed without obvious parse time problems (dependencies, parse
errors, dangling .bbappends, etc.).

This is similar to the BSP test_machine_world. The difference is
that test_world doesn't change the MACHINE.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoyocto-compat-layer.py: apply test_signatures to all layers
Patrick Ohly [Tue, 27 Jun 2017 15:33:40 +0000 (17:33 +0200)]
yocto-compat-layer.py: apply test_signatures to all layers

Software layers were previously allowed to change signatures, but
that's not desired for those layers either. The rule that a layer
which is "Yocto Compatible 2.0" must not change signatures unless
explicitly requested holds for all kinds of layers.

However, as this is something that software layers might not be able
to do right away, testing for signature changes in software layers can
be disabled. It's on by default, as that was Richard's
recommendation. Whether that should change needs further discussion as
part of finalizing "Yocto Compatible 2.0".

As it might still change, the tool now has both a with/without
parameter so that users of the tool can choose the desired behavior
without being affected by future changes to the default.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoyocto-compat-layer.py: tolerate broken world builds during signature diff
Patrick Ohly [Tue, 27 Jun 2017 15:33:39 +0000 (17:33 +0200)]
yocto-compat-layer.py: tolerate broken world builds during signature diff

The "test_signatures" test ignored a broken world build when getting
signatures, but the code which then tried to analyze a difference
found by the test didn't, which prevented printing the difference.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoyocto-compat-layer.py: avoid adding layers more than once
Patrick Ohly [Tue, 27 Jun 2017 15:33:38 +0000 (17:33 +0200)]
yocto-compat-layer.py: avoid adding layers more than once

add_layer_dependencies() might get called more than once, or one of
the layer dependencies might already be present. The function should
not add layers again because doing so can cause warnings like:

  WARNING: Duplicate inclusion for .../meta-openembedded/meta-oe/conf/distro/include/meta_oe_security_flags.inc in .../meta-openembedded/meta-oe/conf/layer.conf

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agosysstat:11.5.5 -> 11.5.6
susanbian [Tue, 27 Jun 2017 07:01:06 +0000 (15:01 +0800)]
sysstat:11.5.5 -> 11.5.6

Upgrade sysstat from 11.5.5 to 11.5.6.

Signed-off-by: susanbian <bianyq@cn.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoopenssl: Upgrade 1.0.2k -> 1.0.2l
Changhyeok Bae [Sun, 18 Jun 2017 15:23:34 +0000 (15:23 +0000)]
openssl: Upgrade 1.0.2k -> 1.0.2l

1. Dropped obsolete patches, because the new version contains them:
  - fix-cipher-des-ede3-cfb1.patch
  - openssl-avoid-NULL-pointer-dereference-in-EVP_DigestInit_ex.patch
2. LICENSE checksum change due to copyright years and wording tweak.
3. Test binaries (x86-64) are included in source code. So remove those
only for ptest.

Signed-off-by: Changhyeok Bae <changhyeok.bae@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibepoxy: Upgrade 1.4.2 -> 1.4.3
Jussi Kukkonen [Tue, 27 Jun 2017 13:16:25 +0000 (16:16 +0300)]
libepoxy: Upgrade 1.4.2 -> 1.4.3

Imports the current EGL API registry from Khronos.

Makes EGL support optional: this is reflected in the recipe but
egl is enabled by default as before.

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogtk+3: Update the patches to work with old versions of patch
Peter Kjellerstedt [Tue, 4 Jul 2017 00:06:46 +0000 (02:06 +0200)]
gtk+3: Update the patches to work with old versions of patch

The patch "0003-Add-disable-opengl-configure-option.patch" used Git
rename syntax, which is not supported by patch 2.6.x and older.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogtk+3: Upgrade 3.22.15 -> 3.22.16
Jussi Kukkonen [Tue, 27 Jun 2017 13:16:24 +0000 (16:16 +0300)]
gtk+3: Upgrade 3.22.15 -> 3.22.16

A fair amount of bug fixes and improvements.

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogtk+3: Update UPSTREAM_CHECK_REGEX
Jussi Kukkonen [Tue, 27 Jun 2017 13:16:23 +0000 (16:16 +0300)]
gtk+3: Update UPSTREAM_CHECK_REGEX

Avoid versions >= 3.90 (development versions for 4.0).

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocmake: Use find_program if find_host_program is not available
Maxime Roussin-Bélanger [Tue, 27 Jun 2017 14:47:31 +0000 (10:47 -0400)]
cmake: Use find_program if find_host_program is not available

CMake does not define the `find_host_program` command we've
been using in the cross-compiling code path.  It was
provided by a widely used Android toolchain file.  For
compatibility, continue to use `find_host_program` if
available, but otherwise use just `find_program`.

Signed-off-by: Maxime Roussin-Bélanger <maxime.roussinbelanger@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoinsane: remove obsolete gcc 4.5 check
Ross Burton [Wed, 28 Jun 2017 14:50:45 +0000 (15:50 +0100)]
insane: remove obsolete gcc 4.5 check

As gcc 4.5 is very old now (released in 2010, gcc 4.6 released in 2011)
this check can be removed now.

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agosanity.bbclass: remove ASSUME_PROVIDED checks that can't succeed
Ross Burton [Wed, 28 Jun 2017 14:38:36 +0000 (15:38 +0100)]
sanity.bbclass: remove ASSUME_PROVIDED checks that can't succeed

qemu-arm and libsdl-native are not in HOSTTOOLS, so there's no point in
checking that they're on PATH.

Also qemu uses pkg-config to find SDL, so libsdl-native isn't required.

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agometa/lib/oe/sdk.py: support added for executing pre-target commands
Amarnath Valluri [Wed, 14 Jun 2017 11:30:47 +0000 (14:30 +0300)]
meta/lib/oe/sdk.py: support added for executing pre-target commands

Added a new POPULATE_SDK_PRE_TARGET_COMMAND variable, which can contain
functions need to be executed at pre traget sysroot creation phase.

classes/populate_sdk_base.bbclass: Added POPULATE_SDK_PRE_TARGET_COMMAND to sdk
command variables list.

Signed-off-by: Amarnath Valluri <amarnath.valluri@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agomkefidsk: fix bash/dash shell quoting problem
Saul Wold [Thu, 29 Jun 2017 20:46:18 +0000 (13:46 -0700)]
mkefidsk: fix bash/dash shell quoting problem

mkefidsk currently writes a startup.nsh with embedded control characters.
This happens because \b etc are control sequences to the shell echo
command when using dash. The resulting startup.nsh causes the bootup
to fail, and the user is dropped into the EFI shell to manually run
startup.nsh.

Patch originally provided by Troy D. Hanson <troy.hanson@jhuapl.edu>

[YOCTO #9665]

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agoscripts/contrib/patchreview: add new script
Ross Burton [Thu, 29 Jun 2017 11:43:10 +0000 (12:43 +0100)]
scripts/contrib/patchreview: add new script

This script analyses the patches we apply and can sanity check or output
statistics.

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
7 years agolz4: Add patch to fix re-builds
Jussi Kukkonen [Mon, 26 Jun 2017 13:14:24 +0000 (16:14 +0300)]
lz4: Add patch to fix re-builds

Configure fails when rebuilding lz4 as LIBDIR is used in two different
places for two different things and we override it with a environment
variable.

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoscripts/test-reexec: fix several incomplete regexp
Ming Liu [Tue, 27 Jun 2017 10:39:14 +0000 (12:39 +0200)]
scripts/test-reexec: fix several incomplete regexp

Some "=" are missing in the regexp, which leads it match multiple results
if the variables have overrides, for instance:
...
SSTATE_DIR="xxxx"
SSTATE_DIR_qemux86="yyyy"
...

it will match both of them without "=".

Signed-off-by: Ming Liu <peter.x.liu@external.atlascopco.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoscripts/oe-find-native-sysroot: fix a incomplete regexp
Ming Liu [Tue, 27 Jun 2017 10:38:49 +0000 (12:38 +0200)]
scripts/oe-find-native-sysroot: fix a incomplete regexp

A "=" is missing in the regexp, which leads it match multiple results
if STAGING_DIR_NATIVE has overrides, for instance:
...
STAGING_DIR_NATIVE="xxxx"
STAGING_DIR_NATIVE_qemux86="yyyy"
...

it will match both of them without "=".

Signed-off-by: Ming Liu <peter.x.liu@external.atlascopco.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoruncmd.py: unit testing for runCmd()
Patrick Ohly [Tue, 27 Jun 2017 11:03:41 +0000 (13:03 +0200)]
runcmd.py: unit testing for runCmd()

This covers the traditional API as well as the new output_log feature.
While testing, it was noticed that killing hanging commands does not
work when a shell is used to run the command(s). This might be worth
fixing.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocommands.py: live output logging + result.error encoding fix
Patrick Ohly [Tue, 27 Jun 2017 11:03:40 +0000 (13:03 +0200)]
commands.py: live output logging + result.error encoding fix

Tests that use bitbake("my-test-image") can run for a long time
without any indication to the user of oe-selftest about what's going
on. The test author has to log the bitbake output explicitly,
otherwise it is lost in case of test failures.

Now it is possible to use bitbake("my-test-image",
output_log=self.logger) to get more output both on the console and in
the XML output (when xmlrunner is installed). Example output:

2017-06-23 12:23:14,144 - oe-selftest - INFO - Running tests...
2017-06-23 12:23:14,145 - oe-selftest - INFO - ----------------------------------------------------------------------
2017-06-23 12:23:14,151 - oe-selftest - INFO - Running: bitbake my-test-image
2017-06-23 12:23:16,363 - oe-selftest - INFO - Loading cache...done.
2017-06-23 12:23:17,575 - oe-selftest - INFO - Loaded 3529 entries from dependency cache.
2017-06-23 12:23:18,811 - oe-selftest - INFO - Parsing recipes...done.
2017-06-23 12:23:19,659 - oe-selftest - INFO - Parsing of 2617 .bb files complete (2612 cached, 5 parsed). 3533 targets, 460 skipped, 0 masked, 0 errors.
2017-06-23 12:23:19,659 - oe-selftest - INFO - NOTE: Resolving any missing task queue dependencies

Because the implementation was already using threading, the same is
done to decouple reading and writing the different pipes instead of
trying to multiplex IO in a single thread. Previously the helper
thread waited for command completion, now that is done in the main
thread.

The most common case (no input data, joined stdout/stderr) still uses
one extra thread and a single read(), so performance should be roughly
the same as before.

Probably unintentionally, result.error was left as byte string when
migrating to Python3. OE-core doesn't seem to use runCmd() with split
output at the moment, so changing result.error to be treated the same
as result.output (i.e. decoded to a normal strings) seems like a
relatively safe API change (or rather, implementation fix).

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
merge: wait()
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoslang: add ptest
Stefan Strogin [Tue, 20 Jun 2017 16:21:02 +0000 (19:21 +0300)]
slang: add ptest

This patch adds ptest for slang, it uses `make runtests` to run tests.
The Makefile runs runtests.sh for each *.sl and *.slc tests.
runtests.sh is patched for printing test results in a common format,
Makefile is patched in order not to compile any dependencies.

Changes v2->v3: rebased on the latest master (1d9a88f6).

Signed-off-by: Stefan Strogin <sstrogin@cisco.com>
Cc: xe-linux-external@cisco.com
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agov86d, qemuboot-x86.inc: use KERNEL_MODULE_AUTOLOAD+KERNEL_MODULE_PROBECONF for uvesaf...
Martin Jansa [Thu, 27 Apr 2017 21:49:37 +0000 (23:49 +0200)]
v86d, qemuboot-x86.inc: use KERNEL_MODULE_AUTOLOAD+KERNEL_MODULE_PROBECONF for uvesafb instead of fbsetup init script

* also add UVESA_MODE variable for easier change of resolution and respect it in QB_KERNEL_CMDLINE_APPEND
  as well
* don't use init script just to call modprobe
* I wasn't able to test this all the way with runqemu, because runqemu
  doesn't work on my system, but I've verified that the right params
  appear there and that I can easily change UVESA_MODE from
  conf/local.conf, the modules.d and modprobe.d files look OK:
  OE qemux86@ ~/build/oe-core/tmp-glibc/deploy/images/qemux86/core-image-sato-qemux86-20170427212613.rootfs
  $ cat etc/modules-load.d/uvesafb.conf
  uvesafb

  OE qemux86@ ~/build/oe-core/tmp-glibc/deploy/images/qemux86/core-image-sato-qemux86-20170427212613.rootfs
  $ cat etc/modprobe.d/uvesafb.conf
  options uvesafb mode_option=1600x1200-32

  so I'll be able to drop this KERNEL_MODULE_AUTOLOAD +
  KERNEL_MODULE_PROBECONF from my DISTRO conf.

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocopyleft_filter.bbclass: restore possiblity to filter on type
André Draszik [Mon, 26 Jun 2017 08:36:19 +0000 (09:36 +0100)]
copyleft_filter.bbclass: restore possiblity to filter on type

Since the changes introduced in ae9102bda398
("copyleft_filter.bbclass: Allow to filter on name"), it is
impossible to filter on the recipe type, all recipes are
treated as though they should be included if the license
matches, irrespective of the COPYLEFT_RECIPE_TYPES
variable.

Fix this.

Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoselftest/archiver: only execute deploy_archives task
André Draszik [Mon, 26 Jun 2017 08:36:18 +0000 (09:36 +0100)]
selftest/archiver: only execute deploy_archives task

There should be no reason to execute a full build, as we're
just interested in the deployment of the archives.

The newly added tests already do the same.

Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoselftest/archiver: add tests for recipe type filtering
André Draszik [Mon, 26 Jun 2017 08:36:17 +0000 (09:36 +0100)]
selftest/archiver: add tests for recipe type filtering

The archiver used to be able to filter based on COPYLEFT_RECIPE_TYPES.

Unfortunately, this got broken with the fix for
  https://bugzilla.yoctoproject.org/show_bug.cgi?id=6929
in commit ae9102bda398 ("copyleft_filter.bbclass: Allow to filter on name")

Add two tests to prevent that from happening again.

Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoconnman: fix nftables dependency
André Draszik [Mon, 26 Jun 2017 10:08:46 +0000 (11:08 +0100)]
connman: fix nftables dependency

When building with nftables support, connman doesn't ever
depend on the nftables command line tool.

connman will depend on libmnl and libnftnl at build and
run time. In addition, the nftables rules it creates
depend on various kernel modules being present.

Update the PACKAGECONFIG to reflect this. We use the
just introduced RRECOMMENDS field so as to make the
build still succeed if those kernel modules have been
linked statically into the kernel, i.e. when the
packages haven't actually been created.

Signed-off-by: André Draszik <adraszik@tycoint.com>
Acked-by: Sylvain Lemieux <slemieux@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agobase.bbclass: extend PACKAGECONFIG to also allow RRECOMMENDS
André Draszik [Mon, 26 Jun 2017 10:08:45 +0000 (11:08 +0100)]
base.bbclass: extend PACKAGECONFIG to also allow RRECOMMENDS

It can be useful to add RRECOMMENDS to packages created, based
on certain PACKAGECONFIGs.

In particular where a package depends on certain linux kernel
infrastructure (kernel modules) which might or might not be
built as a module, being able to RRECOMMENDS instead of
RDEPENDS on the relevant packages avoids build failures in
case those modules are built statically into the kernel, i.e.
in case no package is being created for them.

Add another field to the PACKAGECONFIG syntax to achieve just
that.

Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoiptables: Apply 0001-fix-build-with-musl.patch unconditionally
Khem Raj [Mon, 26 Jun 2017 05:36:19 +0000 (22:36 -0700)]
iptables: Apply 0001-fix-build-with-musl.patch unconditionally

This patch is generic enough, That it can be applied universally
and makes maintainence easier

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoimage.bbclass: use prependVarFlag for postfuncs
Ed Bartosh [Fri, 16 Jun 2017 12:02:07 +0000 (15:02 +0300)]
image.bbclass: use prependVarFlag for postfuncs

It would be possible to achieve any order of calling functions if
prefuncs are added with appendVarFlag and postfuncs with prependVarFlag.
Then image_X.bbclass can add code with either pre/post-funcs or
do_image_x_append or _prepend.

The execution order would be:

image_X prefuncs
image prefuncs
do_image_X_prepend from image_X.bbclass
do_image_X from image.bbclass
do_image_X_append from image_X.bbclass
image postfuncs
image_X postfuncs

[YOCTO #11372]

Thanks to Ola Nillsson for the idea.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoimage_types.bbclass: get rid of IMAGE_DEPENDS
Ed Bartosh [Fri, 16 Jun 2017 15:21:26 +0000 (18:21 +0300)]
image_types.bbclass: get rid of IMAGE_DEPENDS

The IMAGE_DEPENDS_<type> variables can be set to add dependencies for
individual image types.  Those dependencies are added to the do_rootfs
task, while they really should be added to the specific image type tasks.

These variables are not documented anywhere.

Replaced usage of IMAGE_DEPENDS_<type> with explicitly added
dependencies to do_image_<type> tasks.

[YOCTO #11302]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopinentry: update to 1.0.0
Armin Kuster [Tue, 14 Mar 2017 02:39:48 +0000 (19:39 -0700)]
pinentry: update to 1.0.0

add pkg-config support for libassuan and gpg-error
updated config options

(From OE-Core rev: 032f91850d2df819a0b8f2762f12d952cc64284e)

Signed-off-by: Armin Kuster <akuster@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolinux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl
André Draszik [Fri, 23 Jun 2017 11:42:13 +0000 (12:42 +0100)]
linux-libc-headers: fix duplicate IFF_LOWER_UP DORMANT ECHO on musl

musl _does_ define IFF_LOWER_UP DORMANT ECHO so we should
prevent redefinition of these when on musl.

As per the included patch, this can be triggered by
(from connman 6to4.c):
    include <errno.h>
    include <stdio.h>
    include <stdlib.h>
    include <string.h>
    include <sys/socket.h>
    include <netinet/in.h>
    include <arpa/inet.h>
    include <net/if.h>
    include <linux/ip.h>
    include <linux/if_tunnel.h>
    include <linux/netlink.h>
    include <linux/rtnetlink.h>
    include <sys/ioctl.h>
    include <unistd.h>

In file included from ../git/src/6to4.c:34:0:
.../usr/include/linux/if.h:97:2: error: expected identifier before numeric constant
  IFF_LOWER_UP   = 1<<16, /* __volatile__ */
  ^

This is because at that time, IFF_LOWER_UP has been converted
to 0x10000 already:
enum net_device_flags {
 0x10000 = 1<<16,
 0x20000 = 1<<17,
 0x40000 = 1<<18,

};

Backport a patch that addresses this.

Signed-off-by: André Draszik <adraszik@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agotzcode-native: quote ${CC}
Enrico Scholz [Fri, 23 Jun 2017 11:19:12 +0000 (13:19 +0200)]
tzcode-native: quote ${CC}

build fails else with

| + make -j 8 -l 6 cc=ccache gcc
| make: *** No rule to make target 'gcc'.  Stop.

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoimage_types.bbclass: Prompt error message on missing setting in UBI and UBIFS
Choong YinThong [Thu, 22 Jun 2017 18:09:43 +0000 (11:09 -0700)]
image_types.bbclass: Prompt error message on missing setting in UBI and UBIFS

Prompt error message to guide user add argument
MKUBIFS_ARGS and UBINIZE_ARGS on
every UBI and UBIFS image creation.

[YOCTO #11589]

Signed-off-by: Choong YinThong <yin.thong.choong@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoselftest/seltest.py: Add test to check imports from other layers
Jose Perez Carranza [Thu, 22 Jun 2017 16:16:32 +0000 (09:16 -0700)]
selftest/seltest.py: Add test to check imports from other layers

This tests adds a check of selftest itself to verify if can
add test from other layers.

[YOCTO #9770]

Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Jose Perez Carranza <jose.perez.carranza@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocurl: update to 7.54.1
Oleksandr Kravchuk [Mon, 19 Jun 2017 01:20:05 +0000 (03:20 +0200)]
curl: update to 7.54.1

Signed-off-by: Oleksandr Kravchuk <oleksandr.kravchuk@pelagicore.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoexpect: use u-a for mkpasswd
Martin Jansa [Sat, 17 Jun 2017 17:45:30 +0000 (19:45 +0200)]
expect: use u-a for mkpasswd

* when busybox is built with CONFIG_CRYPTPW=y, then it will provide
  alternative for mkpasswd, which will fail in postinst, because
  mkpasswd from expect wasn't using u-a:
  update-alternatives: Error: not linking rootfs/usr/bin/mkpasswd to
  /bin/busybox.nosuid since rootfs/usr/bin/mkpasswd exists and is not
  a link

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agotcl: cleanups
Ross Burton [Thu, 22 Jun 2017 15:33:29 +0000 (16:33 +0100)]
tcl: cleanups

No need to set DEPENDS_class-native explicitly as self-dependencies are ignored
now.

Don't rewrite do_configure, autotools_do_configure works.

Actually invoke the install-private-headers target (autotools_do_install doesn't
take arguments).

Remove redundant cd.

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agometa: Fix return value checks from subprocess.call()'s
Mikko Rapeli [Thu, 22 Jun 2017 13:23:14 +0000 (16:23 +0300)]
meta: Fix return value checks from subprocess.call()'s

Python function subprocess.call() returns the return value of the
executed process. If return values are not checked, errors may
go unnoticed and bad things can happen.

Change all callers of subprocess.call() which do not check for
the return value to use subprocess.check_call() which raises
CalledProcessError if the subprocess returns with non-zero value.

https://docs.python.org/2/library/subprocess.html#using-the-subprocess-module

All users of the function were found with:

$ git grep "subprocess\.call" | \
  egrep -v 'if.*subprocess\.call|=\ +subprocess\.call|return.*subprocess\.call'

Tested similar patch on top of yocto jethro. Only compile tested
core-image-minimal on poky master branch.

Signed-off-by: Mikko Rapeli <mikko.rapeli@bmw.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agox86-base.inc: Don't add live to IMAGE_FSTYPES, default instead
California Sullivan [Wed, 21 Jun 2017 00:14:19 +0000 (17:14 -0700)]
x86-base.inc: Don't add live to IMAGE_FSTYPES, default instead

Using += makes it difficult to remove. Soft-set with ?= instead.

Signed-off-by: California Sullivan <california.l.sullivan@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agostaging.bbclass: Make use of oe.package.strip_execs
Tobias Hagelborn [Tue, 20 Jun 2017 07:42:43 +0000 (09:42 +0200)]
staging.bbclass: Make use of oe.package.strip_execs

Make use of the library function oe.package.strip_execs for stripping
sysroot executables. oe.packge.strip_execs is based on code previously
residing in sysroot_strip.

Signed-off-by: Tobias Hagelborn <tobiasha@axis.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agopackage.py: Add function strip_execs
Tobias Hagelborn [Tue, 20 Jun 2017 07:42:42 +0000 (09:42 +0200)]
package.py: Add function strip_execs

Strip all executables in a directory.
Utility function  placed in oe-package together with run_strip.
strip_execs is based on strip_sysroot from staging.bbclass
Moving out datastore references in favor of function parameters.

Signed-off-by: Tobias Hagelborn <tobiasha@axis.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agocups: don't change permissions of /var/run/cups/certs in do_package
Ross Burton [Mon, 26 Jun 2017 10:08:18 +0000 (11:08 +0100)]
cups: don't change permissions of /var/run/cups/certs in do_package

This directory is deleted by do_install, so luckily the lack of error checking
meant this didn't break.

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agometa/conf/layer.conf: bump layer version for LSB changes
Ross Burton [Wed, 28 Jun 2017 14:23:44 +0000 (15:23 +0100)]
meta/conf/layer.conf: bump layer version for LSB changes

As oe-core has just dropped a number of recipes that were for LSB
conformance, bump the layer version so we can also remove meta-qt4 from
the autobuilder.

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogrub: fix 0001-grub.d-10_linux.in-add-oe-s-kernel-name.patch missing
Hongxu Jia [Thu, 22 Jun 2017 02:20:09 +0000 (22:20 -0400)]
grub: fix 0001-grub.d-10_linux.in-add-oe-s-kernel-name.patch missing

Since grub upgraded to 2.02, it forgot to aplly
0001-grub.d-10_linux.in-add-oe-s-kernel-name.patch.
...
commit 4f27e4042dfb39caa71c7c6eb0a327de6af4d563
Author: Khem Raj <raj.khem@gmail.com>
Date:   Fri Apr 21 20:36:06 2017 +0000

    grub: Update to 2.02
...

The missing caused grub-mkconfig could not detect kernel bzImage.

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoopenssh: Remove deprecated sshd option
Gary Thomas [Thu, 18 May 2017 09:09:15 +0000 (11:09 +0200)]
openssh: Remove deprecated sshd option

The UsePrivilegeSeparation is no longer supported (recent SSHD always runs
with previlege separation), so remove this option from the default config
file to avoid this warning:
  /etc/ssh/sshd_config line 110: Deprecated option UsePrivilegeSeparation

Signed-off-by: Gary Thomas <gary@mlbassoc.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogdk-pixbuf: Make loader.cache reproducible
Jussi Kukkonen [Thu, 22 Jun 2017 09:52:06 +0000 (12:52 +0300)]
gdk-pixbuf: Make loader.cache reproducible

Make the loader order in the file reliable to enable more reproducible
builds.

[YOCTO #11610]

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibcroco: Upgrade 0.6.11 -> 0.6.12
Fan Xin [Thu, 22 Jun 2017 02:30:17 +0000 (11:30 +0900)]
libcroco: Upgrade 0.6.11 -> 0.6.12

Signed-off-by: Fan Xin <fan.xin@jp.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibsndfile1: Fix CVE-2017-6892
Fan Xin [Thu, 22 Jun 2017 01:44:02 +0000 (10:44 +0900)]
libsndfile1: Fix CVE-2017-6892

Backport upstream patch to fix CVE-2017-6892.

CVE: CVE-2017-6892

Signed-off-by: Fan Xin <fan.xin@jp.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agowayland-protocols: upgrade to 1.8
Denys Dmytriyenko [Wed, 21 Jun 2017 15:36:46 +0000 (11:36 -0400)]
wayland-protocols: upgrade to 1.8

Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibpcre2: 10.22 -> 10.23
Fan Xin [Wed, 21 Jun 2017 08:08:30 +0000 (17:08 +0900)]
libpcre2: 10.22 -> 10.23

1. Upgrade libpcre2 from 10.22 to 10.23

2. Update the checksum of LIC_FILES_CHKSUM
   The copyright time of LICENCE is updated to 2017, the content of LICENCE has no change.

Signed-off-by: Fan Xin <fan.xin@jp.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomesa: Update to 17.1.3 release
Otavio Salvador [Tue, 20 Jun 2017 21:10:28 +0000 (18:10 -0300)]
mesa: Update to 17.1.3 release

17.1.2 release includes fixes for:

  Bug 98833 - [REGRESSION, bisected] Wayland revert commit breaks
              non-Vsync fullscreen frame updates
  Bug 100741 - Chromium - Memory leak
  Bug 100877 - vulkan/tests/block_pool_no_free regression
  Bug 101110 - Build failure in GNOME Continuous

17.1.3 release includes fixes for:

  Bug 100988 - glXGetCurrentDisplay() no longer works for FakeGLX
               contexts?

The 17.1.3 release includes the "util/rand_xor: add missing include
statements" which is now merged upstream (thanks Nicolas Dechesne);
the referred patch is then dropped from the recipe.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agobtrfs-tools: upgrade to 4.11
Alexander Kanavin [Mon, 19 Jun 2017 11:44:47 +0000 (14:44 +0300)]
btrfs-tools: upgrade to 4.11

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
7 years agolibfakekey: upgrade to 0.3+gitAUTOINC+7ad885912e
Alexander Kanavin [Mon, 19 Jun 2017 11:44:43 +0000 (14:44 +0300)]
libfakekey: upgrade to 0.3+gitAUTOINC+7ad885912e

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
7 years agovala: upgrade to 0.36.3
Alexander Kanavin [Mon, 19 Jun 2017 11:44:34 +0000 (14:44 +0300)]
vala: upgrade to 0.36.3

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>
7 years agoltp: use upstream patch to fix faccessat/fchmodat build warnings
Fathi Boudra [Tue, 20 Jun 2017 15:30:54 +0000 (18:30 +0300)]
ltp: use upstream patch to fix faccessat/fchmodat build warnings

Use 0037-faccessat-and-fchmodat-Fx-build-warnings.patch instead of
0012-fix-faccessat01.c-build-fails-with-security-flags.patch.
The new patch has been merged upstream and will be available in next LTP
release.

Signed-off-by: Fathi Boudra <fathi.boudra@linaro.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoconnman: fix build-time warning with sysvinit
Maxin B. John [Tue, 20 Jun 2017 08:35:40 +0000 (11:35 +0300)]
connman: fix build-time warning with sysvinit

Due to recent modifications related to systemd, sysvinit builds began
to show this warning:

WARNING: connman-1.34-r0 do_package: connman: NOT adding alternative
provide /etc/resolv.conf: /etc/resolv-conf.connman does not exist

Fix this warning by making those updates specific to systemd.

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agofile: 5.30 -> 5.31
Fan Xin [Tue, 20 Jun 2017 04:53:06 +0000 (13:53 +0900)]
file: 5.30 -> 5.31

1. Upgrade file from 5.30 to 5.31

2. Rebase the following patch file.
   debian-742262.patch

Signed-off-by: Fan Xin <fan.xin@jp.fujitsu.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agolibinput: upgrade to 1.7.3
Denys Dmytriyenko [Mon, 19 Jun 2017 17:30:08 +0000 (13:30 -0400)]
libinput: upgrade to 1.7.3

Signed-off-by: Denys Dmytriyenko <denys@ti.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agomesa.inc: replace deprecated configure options
Trevor Woerner [Mon, 19 Jun 2017 14:14:19 +0000 (10:14 -0400)]
mesa.inc: replace deprecated configure options

--with-egl-platforms -> --with-platforms
--enable-gallium-llvm -> --enable-llvm

Signed-off-by: Trevor Woerner <twoerner@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agogdb: fix gdbserver not working in musl/mips context
André Draszik [Mon, 19 Jun 2017 11:49:38 +0000 (12:49 +0100)]
gdb: fix gdbserver not working in musl/mips context

Similar to
  https://bugs.lede-project.org/index.php?do=details&task_id=637&openedfrom=-1%2Bweek

Signed-off-by: André Draszik <adraszik@tycoint.com>
Acked-by: Sylvain Lemieux <slemieux@tycoint.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoopkg-utils: rename recipe to follow versioned releases
Alejandro del Castillo [Fri, 16 Jun 2017 19:48:27 +0000 (14:48 -0500)]
opkg-utils: rename recipe to follow versioned releases

Signed-off-by: Alejandro del Castillo <alejandro.delcastillo@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agodhcp: remove PN from PACKAGES
Ross Burton [Tue, 20 Jun 2017 13:22:51 +0000 (14:22 +0100)]
dhcp: remove PN from PACKAGES

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agoepiphany: remove unused intltool build-dependency
Ross Burton [Fri, 16 Jun 2017 19:19:14 +0000 (20:19 +0100)]
epiphany: remove unused intltool build-dependency

Signed-off-by: Ross Burton <ross.burton@intel.com>
7 years agodmidecode: update to 3.1
Alexander Kanavin [Fri, 16 Jun 2017 11:59:19 +0000 (14:59 +0300)]
dmidecode: update to 3.1

Signed-off-by: Alexander Kanavin <alexander.kanavin@linux.intel.com>