]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
9 years agoqemurunner: Enable timestamps on kernel boot-up
Leonardo Sandoval [Wed, 4 Nov 2015 22:52:11 +0000 (22:52 +0000)]
qemurunner: Enable timestamps on kernel boot-up

Having the timestamps in the log is useful when debugging those cases when
the booting time takes longer than expected. printk.times=1 enables stamping
into every printk Kernel call.

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoopenssl: fix mips64 configure support
Wenzong Fan [Thu, 5 Nov 2015 06:09:11 +0000 (01:09 -0500)]
openssl: fix mips64 configure support

Match target name linux-mips64 as well, all mips64 targets will have
mips(32) userspace.

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoat: modify sources in do_patch
Ross Burton [Wed, 4 Nov 2015 19:45:39 +0000 (19:45 +0000)]
at: modify sources in do_patch

Instead of copying the forked source files into ${S} in do_compile_prepend(), do
it in a do_patch() postfunc.  This ensures that the source archiver sees the
changes for compliance and legal verification.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agounzip: rename patch to reflect CVE fix
Ross Burton [Wed, 4 Nov 2015 11:23:06 +0000 (11:23 +0000)]
unzip: rename patch to reflect CVE fix

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoreadline: rename patch to contain CVE reference
Ross Burton [Wed, 4 Nov 2015 11:17:50 +0000 (11:17 +0000)]
readline: rename patch to contain CVE reference

To help automated scanning of CVEs, put the CVE ID in the filename.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoqemu: upgrade to 2.4.0.1
Ross Burton [Wed, 4 Nov 2015 11:04:06 +0000 (11:04 +0000)]
qemu: upgrade to 2.4.0.1

This release is just security fixes:
- CVE-2015-5225
- CVE-2015-6815
- CVE-2015-5278
- CVE-2015-5279

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agognome-desktop-testing: fix ptest output format
Maxin B. John [Mon, 2 Nov 2015 14:17:19 +0000 (16:17 +0200)]
gnome-desktop-testing: fix ptest output format

Fix ptest output format for glib-2.0 and other packages
which depends on gnome-desktop-testing framework.

[YOCTO #6601]

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agodefault-distrovars: remove less from WHITELIST_GPL-3.0
Hongxu Jia [Mon, 2 Nov 2015 07:12:50 +0000 (15:12 +0800)]
default-distrovars: remove less from WHITELIST_GPL-3.0

As oe-core commit 37c8c2c5f6937520eaf9f5d981f9315e36eba4bb and
fa7bdf96b8eb8b5078db38249c5c60ec511c35c6 said, Mark Nudelman
(author of less) has given permission to utilize a generic 2-clause
BSD, so we remove it from WHITELIST_GPL-3.0 which caused
a QA Warning while INCOMPATIBLE_LICENSE += "GPLv3 LGPLv3 GPLv3+ LGPLv3+"
...
|NOTE: INCLUDING less as buildable despite INCOMPATIBLE_LICENSE
because it has been whitelisted
...

Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolsof: don't export EXTRA_OEMAKE
Andre McCurdy [Thu, 29 Oct 2015 21:01:35 +0000 (14:01 -0700)]
lsof: don't export EXTRA_OEMAKE

EXTRA_OEMAKE is private to OE and shouldn't be exported to
the lsof build.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocmake.bbclass: don't duplicate CMAKE_C_FLAGS in CMAKE_C_FLAGS_RELEASE
Andre McCurdy [Thu, 29 Oct 2015 21:00:22 +0000 (14:00 -0700)]
cmake.bbclass: don't duplicate CMAKE_C_FLAGS in CMAKE_C_FLAGS_RELEASE

The base CMAKE_<LANG>_FLAGS are included for all build types,
therefore the CMAKE_<LANG>_FLAGS_RELEASE should contain only
additional flags or over-rides.

  https://cmake.org/cmake/help/v3.2/variable/CMAKE_LANG_FLAGS.html

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorm_work.bbclass: Exclude do_rootfs stamp removal
Jate Sujjavanich [Fri, 30 Oct 2015 14:56:01 +0000 (10:56 -0400)]
rm_work.bbclass: Exclude do_rootfs stamp removal

This prevents unneeded rootfs rebuilds if the metadata has not changed
while using rm_work.

Signed-off-by: Jate Sujjavanich <jatedev@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocairo: fix license for cairo-script-interpreter
Andre McCurdy [Fri, 30 Oct 2015 23:07:10 +0000 (16:07 -0700)]
cairo: fix license for cairo-script-interpreter

Without an explicit license, cairo-script-interpreter inherits
the default LICENSE and isn't packaged in builds which blacklist
GPLv3.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoopenssh: enable X11Forwarding if distro feature x11 is set
Kai Kang [Tue, 27 Oct 2015 10:00:18 +0000 (18:00 +0800)]
openssh: enable X11Forwarding if distro feature x11 is set

When distro feature x11 is set, it is better enable X11Forwarding for
ssh daemon. For contrast, dropbear enable X11 forward by default.

It does NOT need to modify ${WORKDIR}/sshd_config, so drop the modification.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoacpid: Upgrade to 2.0.25
Aníbal Limón [Mon, 26 Oct 2015 23:13:22 +0000 (17:13 -0600)]
acpid: Upgrade to 2.0.25

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolibidn: 1.30 -> 1.32
Kai Kang [Tue, 20 Oct 2015 09:10:12 +0000 (17:10 +0800)]
libidn: 1.30 -> 1.32

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agosqlite: 3.8.10.2 -> 3.9.0
Kai Kang [Fri, 16 Oct 2015 06:28:11 +0000 (14:28 +0800)]
sqlite: 3.8.10.2 -> 3.9.0

Upgrade sqlite from 3.8.10.2 to 3.9.0.

* update python function to get right SRC_URI
* drop 0001-using-the-dynamic-library.patch which use dynamic library
  that it is done that way in new version

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agort-tests: bump to v0.94
Josh Cartwright [Tue, 22 Sep 2015 12:24:38 +0000 (07:24 -0500)]
rt-tests: bump to v0.94

All of the rt-tests patches that OE has been carrying have been
upstreamed or superceded by changes in the v0.94 release.

Adjust SRC_URI to point to canonical upstream git repo, instead of a
development tree.

There was a notable change upstream that required slight reworking of
the recipe.  rt-tests now joins other kbuild-inspired projects by making
use of a CROSS_COMPILE flag to indicate the compiler prefix.

Previously TOOLCHAIN_OPTIONS were conveyed via $CC directly, however,
this does not work with CROSS_COMPILE.  Workaround this by both
specifying CROSS_COMPILE, and feeding the rt-tests build system the
proper $(HOST_CC_ARCH)$(TOOLCHAIN_OPTIONS) via $CFLAGS.

Signed-off-by: Josh Cartwright <joshc@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agogbm: Fix "configure: error: gbm requires --enable-dri"
Jens Rehsack [Thu, 8 Oct 2015 15:25:13 +0000 (17:25 +0200)]
gbm: Fix "configure: error: gbm requires --enable-dri"

When building mesa without X11 enabled, and disabling dri package config,
the enforced "--enable-gbm" fails building mesa, because this feature
requires "--enable-dri", too - which has been disabled by package config.

Consequently, when gbm requires dri, --disable-dri causes in --disable-gbm,
which makes explicit --disable-gbm in case of --disable-dri redundant.

Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoxinetd: install xinetd supported services configuration
yadi.hu@windriver.com [Mon, 26 Oct 2015 09:17:03 +0000 (17:17 +0800)]
xinetd: install xinetd supported services configuration

install xinetd supported services configuration, then these services
are able to work

Signed-off-by: Hu <yadi.hu@windriver.com>
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocombo-layer: introduce ability to exclude component from mass update
Alexander D. Kanevskiy [Fri, 16 Oct 2015 15:49:21 +0000 (18:49 +0300)]
combo-layer: introduce ability to exclude component from mass update

There is no ability at the moment for situations where users would like
to keep section in combo-layer.conf but don't update it, unless explicitly
specified.

Now, by adding "update = no" to desired section would exclude
that repository from "combo-layer update" command.

It is still possible to explicitly update it by "combo-layer update $section".

By default, all repositories are assumed as "update = yes"

Signed-off-by: Alexander D. Kanevskiy <kad@kad.name>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-dtb.inc: refactor common code to function get_real_dtb_path_in_kernel
Stefan Christ [Mon, 26 Oct 2015 08:20:37 +0000 (09:20 +0100)]
linux-dtb.inc: refactor common code to function get_real_dtb_path_in_kernel

Signed-off-by: Stefan Christ <s.christ@phytec.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-dtb.inc: refactor common code to function normalize_dtb
Stefan Christ [Mon, 26 Oct 2015 08:20:36 +0000 (09:20 +0100)]
linux-dtb.inc: refactor common code to function normalize_dtb

Signed-off-by: Stefan Christ <s.christ@phytec.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-dtb.inc: explicit test for empty string not needed
Stefan Christ [Mon, 26 Oct 2015 08:20:35 +0000 (09:20 +0100)]
linux-dtb.inc: explicit test for empty string not needed

The for loop already handles the case when KERNEL_DEVICETREE is empty.

Signed-off-by: Stefan Christ <s.christ@phytec.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoptest-runner: Allow running of specific tests
Ioan-Adrian Ratiu [Fri, 30 Oct 2015 14:49:54 +0000 (16:49 +0200)]
ptest-runner: Allow running of specific tests

By default ptest-runner executes all ptests found in a system.
With this change, ptest-runner can be given a list of ptest
package names to run (the default is still available).

For example, to run only the zlib and rt-tests ptests:
"ptest-runner zlib rt-tests"

Signed-off-by: Ioan-Adrian Ratiu <adrian.ratiu@ni.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooeqa/testimage: Add support for test folder export.
Lucian Musat [Fri, 9 Oct 2015 17:16:26 +0000 (20:16 +0300)]
oeqa/testimage: Add support for test folder export.

If tests are organized in folders then it will copy the whole
folder to the export location. This should keep the original
directory structure for exported oeqa/runtime.

Signed-off-by: Lucian Musat <george.l.musat@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooeqa/runtime/multilib: run the arch tests on connmand not connman-applet
Ross Burton [Wed, 28 Oct 2015 15:58:03 +0000 (15:58 +0000)]
oeqa/runtime/multilib: run the arch tests on connmand not connman-applet

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooeqa/runtime: remove dmesg test
Ross Burton [Fri, 24 Jul 2015 10:47:51 +0000 (11:47 +0100)]
oeqa/runtime: remove dmesg test

This has been obsoleted by parselogs.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agonfs-utils/statd: fix a segfault
Chen Qi [Wed, 28 Oct 2015 02:30:09 +0000 (10:30 +0800)]
nfs-utils/statd: fix a segfault

Fix the segfault by separating the socket used in statd from the sockets
of RPC core.

Signed-off-by: Shan Hai <shan.hai@windriver.com>
Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoqemu: enable user mode for mips64 and mips64el
Jackie Huang [Wed, 28 Oct 2015 04:34:55 +0000 (12:34 +0800)]
qemu: enable user mode for mips64 and mips64el

- remove mips64 and mips64el from softmmuonly list
  to enable user mode, they have been supported
  since 2012.
- keep the softmmuonly list and for loop although
  there is only one for now in case more supported
  arches added.

Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agognupg: fix find-version for beta checking
Wenzong Fan [Wed, 28 Oct 2015 05:53:37 +0000 (01:53 -0400)]
gnupg: fix find-version for beta checking

find-version always assumes that gnupg is beta if autogen.sh is run
out of git-repo. This doesn't work for users whom just take release
tarball and re-run autoconf in their local build dir.

This fixes runtime issue:

  $gpg --list-sigs
  gpg: NOTE: THIS IS A DEVELOPMENT VERSION!
  gpg: It is only intended for test purposes and should NOT be
  gpg: used in a production environment or with production keys!

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorpm: define EM_AARCH64 for debugedit
Roy Li [Wed, 28 Oct 2015 09:52:43 +0000 (17:52 +0800)]
rpm: define EM_AARCH64 for debugedit

EM_AARCH64 maybe not be defined due to old version elf.h when compile
rpm-native, and lead to that debugedit can not work on aarch64 elf object
files, since there is no other dependence, except these two macro,
define them to make debugedit work on aarch64 elf files.

  debugedit: /bitbake_build/tmp/work/aarch64-wrs-linux/libvirt/1.2.19-r0/
  package/usr/lib64/libvirt/ptest/daemon/libvirtd_admin_la-admin_server.o:
  Unhandled relocation 258 in .debug_info section

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoxserver-xorg: drop empty ${PN}-security-policy package
Joshua Lock [Fri, 16 Oct 2015 13:07:33 +0000 (14:07 +0100)]
xserver-xorg: drop empty ${PN}-security-policy package

The SecurityPolicy hasn't been included since 2007[1] and the
last remnants were removed from the code base in 2010[2].

1. http://cgit.freedesktop.org/xorg/xserver/commit/?id=9d03ca
2. http://cgit.freedesktop.org/xorg/xserver/commit/?id=19d03d

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoxserver-xorg: add Xwayland RRECOMMENDS
Joshua Lock [Tue, 13 Oct 2015 16:21:08 +0000 (17:21 +0100)]
xserver-xorg: add Xwayland RRECOMMENDS

Xwayland has the same runtime dependencies as the standard
xserver so set RRECOMMENDS the same way.

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoweston: add a PACKAGECONFIG option for xwayland support
Joshua Lock [Tue, 13 Oct 2015 16:15:27 +0000 (17:15 +0100)]
weston: add a PACKAGECONFIG option for xwayland support

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agosystemd: make dbus an optional build time dependency
Joshua Lock [Tue, 13 Oct 2015 16:13:13 +0000 (17:13 +0100)]
systemd: make dbus an optional build time dependency

The dbus buildtime dependency is only required in order to build
dbus related tests in modern systemd, make this a PACKAGECONFIG
option.

This changes the default behaviour to no longer build dbus
related tests without explicitly enabling the PACKAGECONFIG.

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoweston: add PACKAGECONFIG to build with systemd-login support
Joshua Lock [Tue, 20 Oct 2015 09:33:05 +0000 (10:33 +0100)]
weston: add PACKAGECONFIG to build with systemd-login support

Automatically enable this support for distributions which have
systemd in DISTRO_FEATURES.

We add additional patches to weston, backported from Weston
git, to support the newer single libsystemd pkg-config file
and to make the enabling of systemd-login support explicit.

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agosystemd: add PACKAGECONFIG to build with compatibility libraries
Joshua Lock [Tue, 13 Oct 2015 16:07:33 +0000 (17:07 +0100)]
systemd: add PACKAGECONFIG to build with compatibility libraries

Systemd < 209 shipped separate pc files for various interfaces, these
can still be generated for compatibility with code using the older
interfaces.

Add a PACKAGECONFIG option to build systemd with the compatibility
pc files.

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoos-release: put double-quotes around variable contents
Craig McQueen [Thu, 15 Oct 2015 05:13:09 +0000 (16:13 +1100)]
os-release: put double-quotes around variable contents

This makes the resulting /etc/os-release file have valid shell
assignment syntax. This makes it loadable by a shell script, using the
'source' command:

    source /etc/os-release

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocpio: fix testcase symlink-bad-lengths [ LIN8-947 ]
Jun Zhang [Mon, 19 Oct 2015 09:41:55 +0000 (17:41 +0800)]
cpio: fix testcase symlink-bad-lengths [ LIN8-947 ]

* fix the output of stderr & stdout to meet latest cpio

Signed-off-by: Jun Zhang <jun.zhang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocpio: Fix symlink-bad-length test for 64-bit [ LIN8-947 ] architectures.
Jun Zhang [Mon, 19 Oct 2015 09:41:54 +0000 (17:41 +0800)]
cpio: Fix symlink-bad-length test for 64-bit [ LIN8-947 ] architectures.

* src/util.c: Return non-zero exit code if EOF is hit prematurely.

Backport commit:
http://git.savannah.gnu.org/cgit/cpio.git/commit/src/util.c?id=f6a8a2cbd2d5ca40ea94900b55b845dd5ca87328

Signed-off-by: Jun Zhang <jun.zhang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agogtk+3: fix ALTERNATIVE_PRIORITY conflict with gtk+
Robert Yang [Sat, 10 Oct 2015 02:51:33 +0000 (19:51 -0700)]
gtk+3: fix ALTERNATIVE_PRIORITY conflict with gtk+

Signed-off-by: Ross Burton <ross.burton@intel.com>
Conflicts:
gtk+3 and gtk+ have the same priority 10 for gtk-update-icon-cache

Turn up gtk+3's priority to fix the conflict.

[YOCTO #8477]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocoreutils: fix ALTERNATIVE_PRIORITY conflict with procps and mktemp
Robert Yang [Sat, 10 Oct 2015 02:51:33 +0000 (19:51 -0700)]
coreutils: fix ALTERNATIVE_PRIORITY conflict with procps and mktemp

Signed-off-by: Ross Burton <ross.burton@intel.com>
Conflicts:
coreutils and procps have the same priority 100 for kill.1
coreutils and procps have the same priority 100 for uptime.1
coreutils and procps have the same priority 100 for uptime
coreutils and procps have the same priority 100 for kill
coreutils and mktemp have the same priority 100 for mktemp

Turn down coreutils' priority to fix the conflict, the larger set
of tools should use a lower priority.

[YOCTO #8477]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoutil-linux: fix ALTERNATIVE_PRIORITY conflict with ncurses procps and e2fsprogs
Robert Yang [Sat, 10 Oct 2015 02:51:33 +0000 (19:51 -0700)]
util-linux: fix ALTERNATIVE_PRIORITY conflict with ncurses procps and e2fsprogs

Signed-off-by: Ross Burton <ross.burton@intel.com>
Conflicts:
util-linux and ncurses have the same priority 100 for reset
util-linux and procps have the same priority 100 for kill.1
util-linux and procps have the same priority 100 for kill
e2fsprogs and util-linux have the same priority 100 for libblkid.3
e2fsprogs and util-linux have the same priority 100 for blkid.8
e2fsprogs and util-linux have the same priority 100 for findfs.8
e2fsprogs and util-linux have the same priority 100 for fsck.8

Turn down util-linux' priority to fix the conflict, the larger set
of tools should use a lower priority.

[YOCTO #8477]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoconsole-tools: fix ALTERNATIVE_PRIORITY conflict with kbd
Robert Yang [Sat, 10 Oct 2015 02:51:33 +0000 (19:51 -0700)]
console-tools: fix ALTERNATIVE_PRIORITY conflict with kbd

Signed-off-by: Ross Burton <ross.burton@intel.com>
Conflicts:
console-tools and kbd have the same priority 100 for chvt
console-tools and kbd have the same priority 100 for deallocvt
console-tools and kbd have the same priority 100 for fgconsole
console-tools and kbd have the same priority 100 for openvt

Turn down console-tools' priority to fix the conflict, the larger set
of tools should use a lower priority.

[YOCTO #8477]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agodebianutils: fix ALTERNATIVE_PRIORITY conflict with which
Robert Yang [Sat, 10 Oct 2015 02:51:33 +0000 (19:51 -0700)]
debianutils: fix ALTERNATIVE_PRIORITY conflict with which

Signed-off-by: Ross Burton <ross.burton@intel.com>
Conflicts:
debianutils and which have the same priority 100 for which

Turn down debianutils' priority to fix the conflict, the larger set
of tools should use a lower priority.

[YOCTO #8477]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-dtb.inc: use same variable name DTB for all elements of KERNEL_DEVICETREE
Stefan Christ [Mon, 19 Oct 2015 08:15:38 +0000 (10:15 +0200)]
linux-dtb.inc: use same variable name DTB for all elements of KERNEL_DEVICETREE

Signed-off-by: Stefan Christ <s.christ@phytec.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-dtb.inc: remove unneeded 'cd'
Stefan Christ [Mon, 19 Oct 2015 08:15:37 +0000 (10:15 +0200)]
linux-dtb.inc: remove unneeded 'cd'

Signed-off-by: Stefan Christ <s.christ@phytec.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowebkitgtk: Add upstream patch to fix build problem
Pushpal Sidhu [Fri, 23 Oct 2015 21:03:21 +0000 (14:03 -0700)]
webkitgtk: Add upstream patch to fix build problem

From Upstream webkit: fd15a368fa73fe08d91cc1dd7ef05c0c5a3ae851

Webkit will fail to compile under certain circumstances. Please see
https://bugs.webkit.org/show_bug.cgi?id=149172 for more details.

Signed-off-by: Pushpal Sidhu <psidhu@gateworks.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agopython: don't append -D__SOFTFP__ to TARGET_CC_ARCH for armv6/armv7a
Andre McCurdy [Tue, 20 Oct 2015 02:56:12 +0000 (19:56 -0700)]
python: don't append -D__SOFTFP__ to TARGET_CC_ARCH for armv6/armv7a

Remove obsolete hack which seems to date back to (at least) 2009. It's
potentially harmful as python TARGET_CC_ARCH flags can leak through to
other packages via the sysroot _sysconfigdata.py.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoprexport.bbclass: avoid export for native and crosssdk
Robert Yang [Tue, 20 Oct 2015 09:25:51 +0000 (02:25 -0700)]
prexport.bbclass: avoid export for native and crosssdk

* The prservice is only used by do_package, the native or crosssdk
  doesn't have do_package.
* Change WARN to NOTE when no AUTOPR found, the prexport_handler()
  checks all the parsed recipes, but a lot of them may not be built, for
  example, the lower version or broken recipes. So change WARN to NOTE to
  not confuse user.

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorecipes: add distro_features_check for some packages
Wenzong Fan [Wed, 21 Oct 2015 08:23:55 +0000 (04:23 -0400)]
recipes: add distro_features_check for some packages

* The packages that depends on gtk+3 require any of distro features
  from ${GTK3DISTROFEATURES}

* The packages that depends on virtual/egl, virtual/libgl ... require
  distro feature 'opengl'

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoscons.bbclass: SCons packages don't require do_configure
Andre McCurdy [Tue, 20 Oct 2015 04:06:43 +0000 (21:06 -0700)]
scons.bbclass: SCons packages don't require do_configure

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobusybox: Schedule mdev after mountall
Khem Raj [Wed, 21 Oct 2015 04:45:03 +0000 (21:45 -0700)]
busybox: Schedule mdev after mountall

mdev does automounting of block media, which expects
/run/ to be mounted, so let mounting happen before running mdev service
and mountall is done with at 03 so trigger mdev at 04

This helps in mounting the devices which are plugged on boot

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobusybox: Fix mdev block device automounting
Khem Raj [Wed, 21 Oct 2015 03:40:20 +0000 (20:40 -0700)]
busybox: Fix mdev block device automounting

The directory ${MDEV_AUTOMOUNT_ROOT}/ which is '/run/media' may not
exist yet when the device is found, so lets use -p switch to mkdir cmd
so it create preceding parent dirs if needed.

Remove some bash specific constructs

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolibarchive: rename patch to reflect CVE
Ross Burton [Tue, 3 Nov 2015 16:30:24 +0000 (16:30 +0000)]
libarchive: rename patch to reflect CVE

This patch is a CVE fix, so rename it to help CVE detection tools identify it as
such.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobinutils: Fix XLP / Octeon 3 instruction clash
Mark Hatle [Tue, 3 Nov 2015 22:21:13 +0000 (16:21 -0600)]
binutils: Fix XLP / Octeon 3 instruction clash

Use the value 0x00000080 for INSN_XLP, as the value 0x00000040 has already
been assigned to INSN_OCTEON3 by the binutils project.

Without this change, invalid instructions can be generated for both INSN_XLP
and INSN_OCTEON3.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobinutils: Fix octeon3 disassembly patch
Mark Hatle [Tue, 20 Oct 2015 15:54:33 +0000 (10:54 -0500)]
binutils: Fix octeon3 disassembly patch

The structure has apparently changed, and there was a missing
setting.  This corrects a segfault when disassembling code.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobuild-appliance-image: Update to jethro head revision 2015-10 2015-10-jethro yocto-2.0
Richard Purdie [Tue, 3 Nov 2015 14:02:57 +0000 (14:02 +0000)]
build-appliance-image: Update to jethro head revision

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agobuild-appliance-image: Update to jethro head revision
Richard Purdie [Mon, 2 Nov 2015 15:54:16 +0000 (15:54 +0000)]
build-appliance-image: Update to jethro head revision

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agounzip: CVE-2015-7696, CVE-2015-7697
Tudor Florea [Thu, 29 Oct 2015 00:14:18 +0000 (01:14 +0100)]
unzip: CVE-2015-7696, CVE-2015-7697

CVE-2015-7696: Fixes a heap overflow triggered by unzipping a file with password
CVE-2015-7697: Fixes a denial of service with a file that never finishes unzipping

References:
http://www.openwall.com/lists/oss-security/2015/10/11/5
https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2015-7696
https://cve.mitre.org/cgi-bin/cvename.cgi?name=CVE-2015-7697

Signed-off-by: Tudor Florea <tudor.florea@enea.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agovte: fix DoS from malicious escape sequence (CVE-2012-2738)
Ross Burton [Mon, 2 Nov 2015 11:03:26 +0000 (11:03 +0000)]
vte: fix DoS from malicious escape sequence (CVE-2012-2738)

Backport a fix from upstream to fix a denial of service via a malicious escape
sequence.

[YOCTO #8617]

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agobuild-appliance-image: Update to jethro head revision
Richard Purdie [Thu, 29 Oct 2015 22:20:28 +0000 (22:20 +0000)]
build-appliance-image: Update to jethro head revision

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoglibc: Allow 64 bit atomics for x86
Juro Bystricky [Thu, 29 Oct 2015 20:43:38 +0000 (13:43 -0700)]
glibc: Allow 64 bit atomics for x86

The fix consist of allowing 64bit atomic ops for x86.
This should be safe for i586 and newer CPUs.
It also makes the synchronization more efficient.

[YOCTO #8140]

Signed-off-by: Juro Bystricky <juro.bystricky@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agolocal.conf.sample: Disable image-prelink by default
Mark Hatle [Mon, 26 Oct 2015 15:30:54 +0000 (10:30 -0500)]
local.conf.sample: Disable image-prelink by default

Due to problems with the prelinker itself, we need to disable the
image-prelink by default.  This will hopefully be re-enabled in the near
future.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoperl: Remove errornous extra path-specs for Module::Build based modules
Jens Rehsack [Thu, 8 Oct 2015 14:23:35 +0000 (16:23 +0200)]
perl: Remove errornous extra path-specs for Module::Build based modules

This patch removes errornous extra path-specs from cpan_build.bbclass
because corrected path specs at build time are enough.

* fixes wrong path used when building using Module::Build toolchain

Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoperl: Correct path for vendorlib, vendorarch, sitelib and sitearch
Jens Rehsack [Thu, 8 Oct 2015 14:22:41 +0000 (16:22 +0200)]
perl: Correct path for vendorlib, vendorarch, sitelib and sitearch

This patch corrects the path specifications when building perl
for vendorlib, vendorarch, sitelib and sitearch to allow newer
dual-life module being installed on host to satisfy configure
and build requirements of some CPAN distributions.

Additionally, fix search path order in perl wrappers.

Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoperl: fix Perl5 module builds
Jens Rehsack [Thu, 8 Oct 2015 14:21:21 +0000 (16:21 +0200)]
perl: fix Perl5 module builds

This patch fixes some issues in classes providing cpan module build support:

* add support even for xs modules with more than 3 levels as
  B::Hooks::End::Of::Scope or Math::Random::ISAAC::XS
* correct handling of Module::Build (as far as stolen from pkgsrc
  and my humble knowledge)
* configure to install to vendor_libs as default, even when
  inherited do_install remains unused (overwritten do_install)

Signed-off-by: Jens Rehsack <sno@netbsd.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu-export-rootfs: update location of unfsd binary
Maxin B. John [Tue, 27 Oct 2015 15:54:30 +0000 (17:54 +0200)]
runqemu-export-rootfs: update location of unfsd binary

oe-core commit: 24b80d211f3808a0ffebee426932f11b8d4d46e0 sets
sbindir = "${bindir}" in the nativesdk class.

So, update the location of unfsd binary from "/usr/sbin" to "/usr/bin" in
runqemu-export-rootfs. Also update unfs3-native to install unfsd under
"bin" directory so the binary is always in the same location.

[YOCTO #8315]

Signed-off-by: Maxin B. John <maxin.john@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agogtk-icon-cache: pass the native libdir to the intercept
Ross Burton [Tue, 27 Oct 2015 17:52:47 +0000 (17:52 +0000)]
gtk-icon-cache: pass the native libdir to the intercept

The intercept runs against the native sysroot so we need to pass it
the native libdir instead of the target libdir, as otherwise it will
use target paths (such as lib64) in the native sysroot.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoconnman: Move wired-setup to ${datadir}
Jussi Kukkonen [Tue, 27 Oct 2015 13:16:40 +0000 (15:16 +0200)]
connman: Move wired-setup to ${datadir}

wired-setup script should not be in ${libdir} as it's not arch
dependent.

This also fixes (or works around) a practical issue where a multilib
build installs the wrong version of connman-conf and then connman
can't find the script.

[YOCTO #8550]

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agouseradd-staticids.bbclass: Do not require trailing colons
Peter Kjellerstedt [Sat, 24 Oct 2015 09:50:17 +0000 (11:50 +0200)]
useradd-staticids.bbclass: Do not require trailing colons

Before, the users and groups specified in the passwd file and the
groups file had to have trailing colons to make sure there were enough
elements in the definitions, or bitbake would throw a Python
exception.  After this change one can omit the trailing colons, which
especially simplifies passwd files used only to specify static UIDs.

Signed-off-by: Peter Kjellerstedt <peter.kjellerstedt@axis.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooeqa/utils/decorators: fix missing keyword arguments on decorators
Paul Eggleton [Mon, 26 Oct 2015 13:28:26 +0000 (13:28 +0000)]
oeqa/utils/decorators: fix missing keyword arguments on decorators

We need to handle keyword arguments here or sending a keyword argument
to a decorated function that accepts keyword arguments will trigger an
error. (This showed up when testcase decorators were added to the
recipetool.RecipetoolTests.test_recipetool_appendsrcfiles_basic_subdir
test).

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoclasses/gtk-icon-cache: don't pass STAGING_LIBDIR_NATIVE to intercepts
Ross Burton [Tue, 20 Oct 2015 19:19:27 +0000 (20:19 +0100)]
classes/gtk-icon-cache: don't pass STAGING_LIBDIR_NATIVE to intercepts

The intercept doesn't need STAGING_LIBDIR_NATIVE anymore, so don't pass it.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agointercepts/update_icon_cache: use STAGING_DIR_NATIVE from environment
Ross Burton [Tue, 20 Oct 2015 19:17:58 +0000 (20:17 +0100)]
intercepts/update_icon_cache: use STAGING_DIR_NATIVE from environment

Instead of expecting that the calling postinst has exported
STAGING_LIBDIR_NATIVE (which will get set to the sysroot at package build time
and may not be correct if sstate is used), use the new STAGING_DIR_NATIVE that
is exported by rootfs.py.

[ YOCTO #8547 ]

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolib/oe/rootfs: tell intercepts where the native sysroot is
Ross Burton [Tue, 20 Oct 2015 19:14:12 +0000 (20:14 +0100)]
lib/oe/rootfs: tell intercepts where the native sysroot is

Some intercepts may want to access files in the native sysroot that are not on
$PATH (such as something in $libexecdir) but any use of STAGING_DIR_NATIVE in
the postinst that calls the intercept will be "baked" into the package, so if
sstate is reused it will use paths that may not exist.

Solve this by exporting the location of the native sysroot in the environment so
the postinst and intercept can use an environment variable instead of a bitbake
variable.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agosubversion: add explicit dependency on file-replacement-native for native builds
Ross Burton [Mon, 19 Oct 2015 21:48:08 +0000 (22:48 +0100)]
subversion: add explicit dependency on file-replacement-native for native builds

9 years agorpm: add explicit dependency on file-replacement-native for native builds
Ross Burton [Mon, 19 Oct 2015 21:47:58 +0000 (22:47 +0100)]
rpm: add explicit dependency on file-replacement-native for native builds

9 years agofile: don't replace host file when built natively
Ross Burton [Fri, 16 Oct 2015 22:02:26 +0000 (23:02 +0100)]
file: don't replace host file when built natively

To avoid races over the file binaries being replaced mid-build we made
file-native an assume-provided binary and check for it at startup, but target
file still needs a native file binary of the same version to compile the magic
data.

The least invasive way of doing this is to build a native file don't put it on
PATH, and tell the target build where to find the native binary.  We do however
want the native libmagic to be installed normally (as for example rpm and
subversion need it) so we can't use NATIVE_PACKAGE_PATH_SUFFIX as that change
libdir.

[ YOCTO #8144 ]

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agosanity: check that the host has file installed
Ross Burton [Fri, 16 Oct 2015 21:49:37 +0000 (22:49 +0100)]
sanity: check that the host has file installed

Now that file-native is ASSUME_PROVIDED, check that it's actually present.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobitbake: add file-native to ASSUME_PROVIDED
Ross Burton [Fri, 16 Oct 2015 21:49:26 +0000 (22:49 +0100)]
bitbake: add file-native to ASSUME_PROVIDED

Various key parts of the core classes (for example, do_package and
do_populate_sysroot) currently require file.  As it's not possible to build a
file-native without invoking do_populate_sysroot mark file-native as
ASSUME_PROVIDED and expect to use the host's binary.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoRevert "runqemu-export-rootfs: update location of unfsd binary"
Ross Burton [Wed, 21 Oct 2015 21:14:49 +0000 (22:14 +0100)]
Revert "runqemu-export-rootfs: update location of unfsd binary"

unfsd appears to be moving around and is in sbin for at least two different
users, so revert the change to expect it in bin.

This reverts commit e56bda210e216251c04d872211081a89ac06dde6.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agopopulate_sdk_base: Ensure PKGDATA_DIR exists
Richard Purdie [Fri, 23 Oct 2015 03:30:36 +0000 (20:30 -0700)]
populate_sdk_base: Ensure PKGDATA_DIR exists

The code assumes that PKG_DATADIR exists and will fail if an image has not been
generated which creates it. This occurs when something like buildtools-tarball
is built which doesn't have target packages, only nativesdk ones.

Since this shouldn't be fatal, workaround this by creating the missing
directory.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoPerl: Use CC version not $Config(gccversion)
Jeremy Puhlman [Fri, 23 Oct 2015 17:18:07 +0000 (10:18 -0700)]
Perl: Use CC version not $Config(gccversion)

Get version data from querying $CC rather then
$Config(gccversion) which comes from running version of
perl. Since perl-native is not likely compiled by gcc
5 at this point, it will never trigger the required
fixes for gcc 5.

[YOCTO #8367]

Signed-off-by: Jeremy Puhlman <jpuhlman@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowic/utils/oe/misc.py: Preserve PATH when running native tools
Randy Witt [Fri, 23 Oct 2015 02:53:56 +0000 (19:53 -0700)]
wic/utils/oe/misc.py: Preserve PATH when running native tools

Previously exec_native_cmd() would remove all items from PATH except for
the native sysroot. This can cause issues for the tools that are created
using create_wrapper().

Now instead of wiping out the PATH, run a sanity check to check if the
command is in the native sysroot.

Signed-off-by: Randy Witt <randy.e.witt@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agomtools_4.0.18.bb: Use create_wrapper() for mcopy
Randy Witt [Fri, 23 Oct 2015 02:46:11 +0000 (19:46 -0700)]
mtools_4.0.18.bb: Use create_wrapper() for mcopy

mcopy uses the IBM850 codepage from gconv. The default install location
for gconv will not match the actual install location since it can be
pulled from sstate. This patch overrides the default location when
running by adding GCONV_PATH to the environment for mcopy.

[YOCTO #7629]

Signed-off-by: Randy Witt <randy.e.witt@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoscripts/oe-pkgdata-util: Fix variable name in error handling
Richard Purdie [Mon, 19 Oct 2015 19:50:39 +0000 (19:50 +0000)]
scripts/oe-pkgdata-util: Fix variable name in error handling

Fix:
    logger.error('Unable to find pkgdata directory %s' % pkgdata_dir)
NameError: global name 'pkgdata_dir' is not defined

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoAdd 850 codepage to uninative-tarball
Randy Witt [Mon, 19 Oct 2015 05:27:02 +0000 (22:27 -0700)]
Add 850 codepage to uninative-tarball

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoe2fsprogs: backport a patch to fix filetype for hardlink
Robert Yang [Fri, 23 Oct 2015 11:22:04 +0000 (04:22 -0700)]
e2fsprogs: backport a patch to fix filetype for hardlink

Backport a patch to fix hardlinks filetype:
IMAGE_INSTALL_append = " e2fsprogs"
$ ./tmp/sysroots/x86_64-linux/sbin/fsck.ext4 tmp/deploy/images/qemux86/core-image-minimal-qemux86.ext4 -f
Pass 1: Checking inodes, blocks, and sizes
Pass 2: Checking directory structure
Setting filetype for entry 'fsck.ext2' in /sbin (80) to 1.
Setting filetype for entry 'mkfs.ext4' in /sbin (80) to 1.
Setting filetype for entry 'fsck.ext4' in /sbin (80) to 1.
Setting filetype for entry 'mkfs.ext4dev' in /sbin (80) to 1.
Setting filetype for entry 'fsck.ext3' in /sbin (80) to 1.
Setting filetype for entry 'mkfs.ext2' in /sbin (80) to 1.
Setting filetype for entry 'mkfs.ext3' in /sbin (80) to 1.
Setting filetype for entry 'e2fsck' in /sbin (80) to 1.
Pass 3: Checking directory connectivity
Pass 4: Checking reference counts
Pass 5: Checking group summary information

test.img: ***** FILE SYSTEM WAS MODIFIED *****
test.img: 799/65536 files (0.1% non-contiguous), 14652/262144 blocks

Now when run it again, we may get:
[snip]
Pass 3A: Optimizing directories
[snip]
test.img: ***** FILE SYSTEM WAS MODIFIED *****
test.img: 799/65536 files (0.1% non-contiguous), 14652/262144 blocks

This is fine since it is optimizing, from "man e2fsck":
e2fsck may sometimes optimize a few directories --- for example, if
directory indexing is enabled and a directory is not indexed and would
benefit from being indexed, or if the index structures are corrupted
and need to be rebuilt.

[YOCTO #8544]

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooeqa/selftest: Added testcase decorators.
Daniel Istrate [Fri, 23 Oct 2015 12:07:51 +0000 (15:07 +0300)]
oeqa/selftest: Added testcase decorators.

Added testcase decorators for testopia integration.

Signed-off-by: Daniel Istrate <daniel.alexandrux.istrate@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu-ifup: Check if the tap interface is set up correctly
Mariano Lopez [Thu, 22 Oct 2015 13:39:20 +0000 (13:39 +0000)]
runqemu-ifup: Check if the tap interface is set up correctly

The process to set up a tap interface is as follows:
    - tap interface is created
    - An IP address is assigned to the tap interface
    - The interface is bring up
    - A route is added to the target using the tap
      interface

Of all the previous steps, only the first one is
check if it was sucessful. The status of the others
are ignored and all of them are required to have
basic connectivity in the target.

This patch adds the checks for the rest of the stpes
needed to set up the tap interface.

[YOCTO #8550]

Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoqemurunner: Show the output of runqemu script
Mariano Lopez [Thu, 22 Oct 2015 13:39:19 +0000 (13:39 +0000)]
qemurunner: Show the output of runqemu script

This change will show runqemu output in the logs
when running the runtime tests. This is helpful
for debugging testimage errors.

[YOCTO #8550]

Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu-internal: Enable support for use virtio devices.
Aníbal Limón [Thu, 22 Oct 2015 15:18:39 +0000 (10:18 -0500)]
runqemu-internal: Enable support for use virtio devices.

Enable virtio usage for default in runqemu also get rid
of duplicate configuration from now all qemu machines uses
virtio.

[YOCTO #8427]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-yocto{, -rt}: Enable support for virtio drivers in qemu machines.
Aníbal Limón [Thu, 22 Oct 2015 15:18:38 +0000 (10:18 -0500)]
linux-yocto{, -rt}: Enable support for virtio drivers in qemu machines.

In order to use virtio devices as default in runqemu script
because these drivers are designed to use in vrit providing
better performance.

[YOCTO #8427]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu: Enable support for kvm without vhost in x86 and x86_64
Aníbal Limón [Thu, 22 Oct 2015 15:18:37 +0000 (10:18 -0500)]
runqemu: Enable support for kvm without vhost in x86 and x86_64

KVM can be used without vhost so add a new option to runqemu for
use kvm with vhost.

Example,
runqemu qemux86 core-image-minimal kvm # kvm without vhost
runqemu qemux86 core-image-minimal kvm-vhost # kvm with vhost

[YOCTO #7443]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoprserv.bbclass: remove it since it is null
Robert Yang [Tue, 20 Oct 2015 09:25:50 +0000 (02:25 -0700)]
prserv.bbclass: remove it since it is null

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoinitscripts/sysfs.sh: Mount devtmpfs on /dev/ if needed
Mike Looijmans [Tue, 20 Oct 2015 06:43:47 +0000 (08:43 +0200)]
initscripts/sysfs.sh: Mount devtmpfs on /dev/ if needed

When booting from an initrd disk, or when the kernel config option
DEVTMPFS_MOUNT isn't provided, /dev/ will not be mounted at boot.

This small addition will check if /dev/ is "useful", and if not, will
mount devtmpfs if the kernel provides it.

With this change, it is possible to set an initscripts style image type
to "cpio.gz" and boot it as initrd. Without this change, the image won't
work properly because of the missing devices.

Signed-off-by: Mike Looijmans <mike.looijmans@topic.nl>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoimage-mklibs.bbclass: update i586 TARGET_ARCH test to i*86
Andre McCurdy [Wed, 21 Oct 2015 06:01:56 +0000 (23:01 -0700)]
image-mklibs.bbclass: update i586 TARGET_ARCH test to i*86

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobase.bbclass: considering multilib when setting LICENSE_EXCLUSION
Jian Liu [Thu, 8 Oct 2015 06:47:51 +0000 (14:47 +0800)]
base.bbclass: considering multilib when setting LICENSE_EXCLUSION

The PACKAGES is not mapped with MLPREFIX when setting LICENSE_EXCLUSION
in base.bbclass. For example,
  For libgcc-dev,
    LICENSE_EXCLUSION-libgcc-dev=1
   but for lib32-libgcc-dev,
    LICENSE_EXCLUSION-libgcc-dev=1
Obviously it is wrong for lib32-libgcc-dev.

Add MLPREFIX before the package name during setting LICENSE_EXCLUSION

Signed-off-by: Jian Liu <jian.liu@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agogcc-target.inc: Add support for executable thats may have a suffix
Mark Hatle [Tue, 20 Oct 2015 16:29:37 +0000 (11:29 -0500)]
gcc-target.inc: Add support for executable thats may have a suffix

In the past GCC has used a wildcard to permit generating executables
that may have a suffix, such as .exe.  This wild card was lost in one
of the updates.  Adding the wild card back in fixes a number of issues
when generating a mingw gcc.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocairo: backport fix for compatibility with OpenGL ES 2.0
Andre McCurdy [Wed, 21 Oct 2015 03:20:35 +0000 (20:20 -0700)]
cairo: backport fix for compatibility with OpenGL ES 2.0

To maintain compatibility with OpenGL ES 2.0, the matrix in
_cairo_gl_shader_bind_matrix() should be manually transposed,
and GL_FALSE passed as the transpose argument to the
glUniformMatrix3fv() call as it is the only valid value for
that parameter in OpenGL ES 2.0.

  http://lists.cairographics.org/archives/cairo/2015-May/026253.html
  http://cgit.freedesktop.org/cairo/commit/?id=f52f0e2feb1ad0a4de23c475a8c020d41a1764a8

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>