]> code.ossystems Code Review - openembedded-core.git/log
openembedded-core.git
9 years agotestimage: Don't test xorg/vnc on qemuarm64
Richard Purdie [Sat, 27 Jun 2015 21:44:23 +0000 (22:44 +0100)]
testimage: Don't test xorg/vnc on qemuarm64

The qemuarm64 machine doesn't have graphics so don't test xorg/vnc
as they won't work.

[YOCTO #7103]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooeqa/parselogs: Whitelist qemuarm64 runtime errors
Richard Purdie [Sat, 27 Jun 2015 21:43:38 +0000 (22:43 +0100)]
oeqa/parselogs: Whitelist qemuarm64 runtime errors

The qemuarm64 machine doesn't have graphics so whitelist the
X server failures to start.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooeqa/recipetool: Fix symlink resolution
Richard Purdie [Sat, 27 Jun 2015 07:10:29 +0000 (08:10 +0100)]
oeqa/recipetool: Fix symlink resolution

Instead of readlink we need to use realpath to resolve any possible
symlink in the file name. I'd got the two confused in the previous
patch. This should really fix selftest failures on the autobuilder.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agogstreamer1.0-plugins-bad.inc: remove duplicate bzip2 dependency
Andre McCurdy [Fri, 26 Jun 2015 21:13:04 +0000 (14:13 -0700)]
gstreamer1.0-plugins-bad.inc: remove duplicate bzip2 dependency

Dependency on bzip2 is handled by the bz2 PACKAGECONFIG option, so
bzip2 does not need to be included in DEPENDS.

Also fix minor indent issue with bluez PACKAGECONFIG option.

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agotiff: Update to 4.0.4
Randy MacLeod [Fri, 26 Jun 2015 20:27:51 +0000 (16:27 -0400)]
tiff: Update to 4.0.4

Update tiff to latest version. None of the local CVE patches
are needed based on reviewing the ChangeLog so remove them.

Signed-off-by: Randy MacLeod <Randy.MacLeod@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agosubversion: Fix subversion-native on Fedora22
Richard Purdie [Fri, 26 Jun 2015 22:40:41 +0000 (23:40 +0100)]
subversion: Fix subversion-native on Fedora22

Similarly to:
http://git.yoctoproject.org/cgit.cgi/poky/commit/?id=9b19d6548a345009a6de79a6820c07a72054d961

we also need to fix the subversion-native case with gcc5 by using
the same fix to the BUILD_CPPFLAGS.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooeqa/parselogs: Fixed a problem in parselogs
Lucian Musat [Fri, 26 Jun 2015 10:36:40 +0000 (13:36 +0300)]
oeqa/parselogs: Fixed a problem in parselogs

When the test passed it returned an error because
grep did not return any error and the variable that
held the results was being referenced but not
assigned any value.

Signed-off-by: Lucian Musat <george.l.musat@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoquota: fix quota do_install errors
Roy Li [Fri, 26 Jun 2015 07:52:54 +0000 (15:52 +0800)]
quota: fix quota do_install errors

ROOTDIR should be defined, otherwise man files will be installed
to host root dir.

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agokmod: upgrade to 21
Chen Qi [Fri, 26 Jun 2015 02:27:49 +0000 (10:27 +0800)]
kmod: upgrade to 21

Upgrade kmod to 21.
Fix cross compilation problem.

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agorecipetool: add appendsrcfile(s) sub-commands
Christopher Larson [Wed, 24 Jun 2015 22:17:47 +0000 (15:17 -0700)]
recipetool: add appendsrcfile(s) sub-commands

This adds the `appendsrcfile` and `appendsrcfiles` sub-commands, which let the
user add or replace one or more files in the recipe sources, either in a path
relative to `S` or `WORKDIR`. Each file gets added to `SRC_URI` as a file://
URI, using the subdir= parameter to specify the destination directory.

Examples:

    # Adds our defconfig as file://defconfig. If it's already in SRC_URI, it
    # won't be duplicated.
    recipetool appendsrcfile --workdir meta-mylayer linux-mel defconfig
    recipetool appendsrcfiles --workdir meta-mylayer linux-mel defconfig

    # Does the same, handling the different local filename
    recipetool appendsrcfile --workdir meta-mylayer linux-mel defconfig.mine defconfig

    # Adds our device tree files to the source tree
    recipetool appendsrcfiles --destdir arch/arm/boot/dts meta-mylayer linux-mel *.dts

Of course, for the latter example to be of use, the new dts files would need
to be added to `KERNEL_DEVICETREE` as well, and depending on the kernel,
`DEFCONFIG` or `KERNEL_DEFCONFIG` may need to be set.

Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agorecipetool.append: use argparse types for validation
Christopher Larson [Wed, 24 Jun 2015 22:17:46 +0000 (15:17 -0700)]
recipetool.append: use argparse types for validation

This validates the arguments early, when argparse is parsing the arguments, in
a consistent way.

Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooe.recipeutils: fix line.split error in bbappend_recipe
Christopher Larson [Wed, 24 Jun 2015 22:17:45 +0000 (15:17 -0700)]
oe.recipeutils: fix line.split error in bbappend_recipe

Cc: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agobuildhistory.bbclass: Create a files-in-<pkg>.txt containg recipe's pkg files
Leonardo Sandoval [Thu, 25 Jun 2015 13:46:47 +0000 (13:46 +0000)]
buildhistory.bbclass: Create a files-in-<pkg>.txt containg recipe's pkg files

A file similar (in format) to files-in-image.txt is created for each
recipe's build package. This would allow to easily see what the
recipe's package may install when including it into an image.

These files reside on the specific buildhistory recipe package (i.e.
buildhistory/packages/i586-poky-linux/busybox/busybox-dev/files-in-busybox-dev.txt
for busybox-dev package)

[YOCTO #5870]

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoxf86-video-vesa: Upgrade 2.3.3 -> 2.3.4
Jussi Kukkonen [Thu, 25 Jun 2015 11:20:03 +0000 (14:20 +0300)]
xf86-video-vesa: Upgrade 2.3.3 -> 2.3.4

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoxorg-xserver: Upgrade 1.17.1 -> 1.17.2
Jussi Kukkonen [Thu, 25 Jun 2015 11:20:02 +0000 (14:20 +0300)]
xorg-xserver: Upgrade 1.17.1 -> 1.17.2

Fixes CVE-2015-3164 in Xwayland.

Remove unnecessary patches:
 * fix_open_max_preprocessor_error.patch is no longer needed:
   OPEN_MAX is defined in every case now
 * xorg-CVE-2013-6424.patch was handled in pixman several releases ago
 * other removed patches are included in 1.17.2

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agomesa: Upgrade 10.5.6 -> 10.5.8
Jussi Kukkonen [Thu, 25 Jun 2015 11:20:01 +0000 (14:20 +0300)]
mesa: Upgrade 10.5.6 -> 10.5.8

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agofreetype: Upgrade 2.5.5 -> 2.6
Jussi Kukkonen [Thu, 25 Jun 2015 11:20:00 +0000 (14:20 +0300)]
freetype: Upgrade 2.5.5 -> 2.6

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agopopulate_sdk_base.bbclass: use pbzip2 to compress SDK tarfiles
Andre McCurdy [Thu, 25 Jun 2015 19:46:47 +0000 (12:46 -0700)]
populate_sdk_base.bbclass: use pbzip2 to compress SDK tarfiles

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoimage_types.bbclass: use pbzip2 for images requiring bz2 compression
Andre McCurdy [Thu, 25 Jun 2015 19:46:46 +0000 (12:46 -0700)]
image_types.bbclass: use pbzip2 for images requiring bz2 compression

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agopbzip2: Fix native dependency on bzip2-native
Richard Purdie [Fri, 26 Jun 2015 22:49:39 +0000 (23:49 +0100)]
pbzip2: Fix native dependency on bzip2-native

bzip2-native is in ASSUME_PROVIDED but we don't just want "bzip2" but
libbz2 here. To do this, we need to DEPEND on bzip2-replacement-native
which hints to bitbake we really do want to build it.

Add the missing dependency to avoid failures on machines which don't
have libbz2-dev installed.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agopbzip2: add initial recipe, pbzip2 v1.1.12
Andre McCurdy [Thu, 25 Jun 2015 19:46:45 +0000 (12:46 -0700)]
pbzip2: add initial recipe, pbzip2 v1.1.12

Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agorun-postinsts: Fix ipk package management detection
Aníbal Limón [Thu, 25 Jun 2015 18:21:16 +0000 (13:21 -0500)]
run-postinsts: Fix ipk package management detection

run-postinsts always mark ipk package management as true, causing
problems when try to execute opkg-cl and isn't present.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooe/rootfs.py: DpkgRootfs/OpkgRootfs add support for dependency handling in postinsts...
Aníbal Limón [Thu, 25 Jun 2015 18:21:15 +0000 (13:21 -0500)]
oe/rootfs.py: DpkgRootfs/OpkgRootfs add support for dependency handling in postinsts scripts.

The old code don't take into account package dependencies causing
undefined execution order in postinsts scripts, in order to fix:

Add DpkgOpkgRootfs class for store common operations in DpkgRootfs
and OpkgRootfs.

Add _get_delayed_postinsts_common method that process Depends from
status file in dpkg/opkg and resolve dependency order causing an
execption if found circular dependencies.

[YOCTO #5318]

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agocmake: Upgrade 2.8.12.2 -> 3.2.2
Moritz Blume [Tue, 23 Jun 2015 15:50:51 +0000 (17:50 +0200)]
cmake: Upgrade 2.8.12.2 -> 3.2.2

Patches "qt4-fail-silent.patch" and "support-oe-qt4-tools-names.patch"
were adapted in order to fit to CMake 3.2.2 (refer to the commit
message in the respective patch for details).
Patch "cmake-2.8.11.2-FindFreetype.patch" was dropped since it was
rejected upstream and its functionality otherwise implemented in the
meantime.

Signed-off-by: Moritz Blume <moritz.blume@bmw-carit.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooe-selftest: timestamp the test runs
Ross Burton [Mon, 22 Jun 2015 16:20:42 +0000 (17:20 +0100)]
oe-selftest: timestamp the test runs

The selftest can take a couple of hours to run, so add a custom result class to
timestamp the output to make it easy to spot any slow tests.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agouclibc: Cope with other mips32 variants
Ramax Lo [Tue, 23 Jun 2015 16:22:52 +0000 (00:22 +0800)]
uclibc: Cope with other mips32 variants

Update uclibc.inc to handle different tunes of mips32 like mips32el,
mips32-nf, mips32el-nf, etc.

Signed-off-by: Ramax Lo <ramaxlo@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agopiglit: fix floating dependence on freeglut
Roy Li [Thu, 18 Jun 2015 09:18:57 +0000 (17:18 +0800)]
piglit: fix floating dependence on freeglut

cmake can not handle "ifdef" and make floating dependence on freeglut_ext.h
when make_depend, so remove to include freeglut_ext.h from header file if
freegult is not in PACKAGECONFIG to fix this issue

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agogdb: Add missing runtime dependency for python option
Andreas Oberritter [Tue, 23 Jun 2015 05:12:58 +0000 (22:12 -0700)]
gdb: Add missing runtime dependency for python option

With python enabled, gdb refuses to start without core
python modules:

| Could not find platform independent libraries <prefix>
| Could not find platform dependent libraries <exec_prefix>
| Consider setting $PYTHONHOME to <prefix>[:<exec_prefix>]
| ImportError: No module named site

It also complains if python-codecs is missing.

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Andre McCurdy <armccurdy@gmail.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agounzip: fix four CVE defects
Roy Li [Tue, 23 Jun 2015 05:32:06 +0000 (13:32 +0800)]
unzip: fix four CVE defects

Port four patches from unzip_6.0-8+deb7u2.debian.tar.gz to fix:
     cve-2014-8139
     cve-2014-8140
     cve-2014-8141
     cve-2014-9636

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoqemu: backport patches to fix CVE issues
Kai Kang [Thu, 18 Jun 2015 09:02:42 +0000 (17:02 +0800)]
qemu: backport patches to fix CVE issues

Backport patches to fix CVE-2015-4103, CVE-2015-4104, CVE-2015-4105 and
CVE-2015-4106. These patches are from debian, but they are originally
from:

http://git.qemu.org/?p=qemu.git;a=shortlog;h=c25bbf1

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoterminal.py: Allow devshell/menuconfig on recent gnome-terminal
Leonardo Sandoval [Tue, 2 Jun 2015 19:54:27 +0000 (14:54 -0500)]
terminal.py: Allow devshell/menuconfig on recent gnome-terminal

Recent versions of gnome-terminal does not support non-UTF8 charset:
https://bugzilla.gnome.org/show_bug.cgi?id=732127 as a result, devshell and
menuconfig tasks silently hang  (error found on trace log of 'strace -f -v
-s 8192 -e write=2 bitbake -c devshell quilt-native': "Non UTF-8 locale
(ANSI_X3.4-1968) is not supported!"). As a workaround, clearing the LC_ALL
environment variable so it uses the locale. Once fixed on the gnome-terminal
project, this should be removed.

Tested on gnome-terminal versions:

       GNOME Terminal 3.4.1.1
       GNOME Terminal 3.14.2 (Default on Ubuntu 15.04)

[YOCTO #7791]

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobluez5: upgrade to 5.31
Cristian Iorga [Mon, 22 Jun 2015 11:12:37 +0000 (14:12 +0300)]
bluez5: upgrade to 5.31

Mostly a bug fix release,
with the the following improvements:
- support for handling BNEP setup response;
- support for setting GATT database security flags;
- support for setting discovery filters interface;
- support for user controlled advertising interface.

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolinux-firmware: upgrade to commit 3161bfa479d5e9ed4f46b57df9bcecbbc4f8eb3c
Cristian Iorga [Mon, 22 Jun 2015 16:55:15 +0000 (19:55 +0300)]
linux-firmware: upgrade to commit 3161bfa479d5e9ed4f46b57df9bcecbbc4f8eb3c

Updates to firmware binaries from (not exhaustive list):
Broadcom, Atheros, Realtek, Marvell, Intel, AMD, Siano.

- AMD microcode copyright years changed;
- Siano license added;
- Intel SST2 license added;

Fixes [YOCTO #6965].

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoweston: fix libdir/libexecdir confusion
Ross Burton [Tue, 23 Jun 2015 10:43:43 +0000 (11:43 +0100)]
weston: fix libdir/libexecdir confusion

Use $libexecdir and $libdir as appropriate so the files are packaged correctly
when libexecdir != $libdir/$BPN.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobluez5: fix lib/libexecdir confusion
Ross Burton [Mon, 22 Jun 2015 21:16:01 +0000 (22:16 +0100)]
bluez5: fix lib/libexecdir confusion

Use $libexecdir and $libdir as appropriate so the files are packaged correctly
when libexecdir != $libdir/$BPN.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agotestimage: Handle ipk/deb packaging format tests correctly
Richard Purdie [Tue, 23 Jun 2015 11:13:02 +0000 (12:13 +0100)]
testimage: Handle ipk/deb packaging format tests correctly

The default test list only works for rpm packaging. This fixes it for
deb and ipk too.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowic: oe-selftest: Test image compressing
Ed Bartosh [Tue, 23 Jun 2015 08:19:49 +0000 (11:19 +0300)]
wic: oe-selftest: Test image compressing

Added 4 new testcases for 'wic --compress-with <compressor>' functionality.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Make code more pythonic
Ed Bartosh [Mon, 22 Jun 2015 13:07:13 +0000 (16:07 +0300)]
wic: Make code more pythonic

Used more pythonic way to make a list.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Implement image compressing
Ed Bartosh [Mon, 22 Jun 2015 12:45:30 +0000 (15:45 +0300)]
wic: Implement image compressing

Implemented compressing result image with specified compressor.
Updated reporting code to show compressed image.

[YOCTO #7593]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Add new argument to direct plugin
Ed Bartosh [Mon, 22 Jun 2015 12:15:22 +0000 (15:15 +0300)]
wic: Add new argument to direct plugin

Added 'compressor' argument to Direct plugin API to
pass a name of compressor utility.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Add new argument to wic_create function
Ed Bartosh [Mon, 22 Jun 2015 11:58:19 +0000 (14:58 +0300)]
wic: Add new argument to wic_create function

Added 'compressor' argument to wic_create to pass a name of
compressor utility to the wic engine.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Add --compress-with command line option
Ed Bartosh [Mon, 22 Jun 2015 11:42:37 +0000 (14:42 +0300)]
wic: Add --compress-with command line option

Added -c/--compress-with command line option to 'wic create'
subcommand. This option is used to specify compressor
utility to compress the image produced by wic. gzip, bzip2 and
xz compressors are supported in this implementation.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agosanity.bbclass: check /bin/sh is dash or bash
Robert Yang [Fri, 26 Jun 2015 06:23:21 +0000 (23:23 -0700)]
sanity.bbclass: check /bin/sh is dash or bash

The build would fail when /bin/sh links to ksh or csh, we only test dash
and bash AFAIK.

* When /bin/sh -> csh:
  $ bitbake quilt-native
$ bitbake quilt-native -cfetch
Illegal variable name.
Illegal variable name.
[snip]
uname: extra operand `2'
Try `uname --help' for more information.

* When /bin/sh -> ksh:
  If there are only a few tasks running, for example,
  "bitbake quilt-native", the build would be OK, but it would fail if we
  run "bitbake world" for a while, there would be a lot of "Broken pipe"
  errors:
Exception: CalledProcessError: Command
'cd /path/to/xx; find . -type d -print | tar -cf - -C /path/to/sysroot-destdir -p --files-from - --no-recursion | tar -xf - -C /path/to/xxx'
returned non-zero exit status 2 with output tar: This does not look like a tar archive
tar: Exiting with failure status due to previous errors
find: `standard output': Broken pipe
find: write error

[YOCTO #7917]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agotoolchain-scripts: fix for multilib
Robert Yang [Fri, 26 Jun 2015 03:06:16 +0000 (20:06 -0700)]
toolchain-scripts: fix for multilib

Use MLPREFIX to fix:
| cat: /path/to/sysroots/lib32-qemux86-64/sysroot-providers/virtual_libc: No such file or directory

[YOCTO #7924]

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooeqa/selftest/recipetool: Fix appendfile_binary test on opensuse131
Richard Purdie [Thu, 25 Jun 2015 15:47:28 +0000 (16:47 +0100)]
oeqa/selftest/recipetool: Fix appendfile_binary test on opensuse131

On OpenSUSE 13.1, /bin/ls is a symlink to /usr/bin/ls. This means the
test doesn't use a binary file an fails on that system.

Ensure we resolve any symlink using readlink to avoid this failure.

======================================================================
FAIL: test_recipetool_appendfile_binary (oeqa.selftest.recipetool.RecipetoolTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-oe-selftest/build/meta/lib/oeqa/selftest/recipetool.py", line 115, in test_recipetool_appendfile_binary
    self.assertIn('WARNING: ', result.output)
AssertionError: 'WARNING: ' not found in 'Parsing recipes..done.\nNOTE: Writing append file /tmp/recipetoolqai0NY8l/recipes-core/coreutils/coreutils_8.23.bbappend\nNOTE: Copying /bin/ls to /tmp/recipetoolqai0NY8l/recipes-core/coreutils/coreutils/ls'

[YOCTO #7920]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agolib/devtool/standard: Fix patch cleanup
Richard Purdie [Thu, 25 Jun 2015 15:09:14 +0000 (16:09 +0100)]
lib/devtool/standard: Fix patch cleanup

If patches fail to apply with git, quilt it used as a fallback. If that
happens, the code in this class is meant to handle cleanup of these patch
files. In the case where ${S} is a subdir of the git tree, the code doesn't
correctly set the patches directory.

This change correctly sets the patches directory (which is different to the
location of the git repository).

[YOCTO #7911]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agolib/oe/patch: Fix git patch application for source in subdirectory
Richard Purdie [Thu, 25 Jun 2015 15:08:26 +0000 (16:08 +0100)]
lib/oe/patch: Fix git patch application for source in subdirectory

Similarly to:
http://git.yoctoproject.org/cgit.cgi/poky/commit/meta/lib/oe/patch.py?id=f205ccaf48ac36f4b26efc4aeb2e9d2939b28646
we need to fix patch application for source which is in a subdirectory.

Passing "." as the git directory or work-dir appears to work (or is ignored)
in some versions of git but does not work in others, probably quite correctly.

Since we have reporoot from the above patch, pass this in directly.

This bug caused this sanity test failure on some machines:

FAIL: test_devtool_modify_git (oeqa.selftest.devtool.DevtoolTests)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/pokybuild/yocto-autobuilder/yocto-worker/nightly-oe-selftest/build/meta/lib/oeqa/selftest/devtool.py", line 390, in test_devtool_modify_git
    self.assertEqual(result.output.strip(), "", 'Created git repo is not clean')
AssertionError: '?? util/mkelfImage/patches/' != '' : Created git repo is not clean

since git apply would fail, it would then fall back to quilt
and the git tree would be left unclean.

[YOCTO #7911]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooe/utils.py: Add support for init/end helper functions in ThreadWorker.
Aníbal Limón [Tue, 23 Jun 2015 16:49:54 +0000 (11:49 -0500)]
oe/utils.py: Add support for init/end helper functions in ThreadWorker.

Add init/end helper functions for ThreadWorker also pass ThreadWorker
as first argument to init/end/func functions this enables per-thread
storage handling.

classes/sstate.bbclass: Add thread_worker argument to checkstatus
function.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agooe/utils.py: Fix thread leakage in ThreadPool
Aníbal Limón [Tue, 23 Jun 2015 16:49:53 +0000 (11:49 -0500)]
oe/utils.py: Fix thread leakage in ThreadPool

In order to fix Thread leakage caused by not call join() in Threads,

Pass num_tasks in ThreadPool for add all the tasks into a Queue this
enable catch of Queue.Empty exception and exit the threads.

classes/sstate.bbclass: Change checkstatus function to match new
ThreadPool operation.

Signed-off-by: Aníbal Limón <anibal.limon@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agowpa-supplicant: Revert "Make SystemD D-Bus config conditional"
Otavio Salvador [Wed, 24 Jun 2015 10:43:33 +0000 (11:43 +0100)]
wpa-supplicant: Revert "Make SystemD D-Bus config conditional"

The D-Bus config is not systemd-specific. It is required for the D-Bus
communication to be operational.

This reverts commit e658ee16dc026b96f67a4c9666d3eb7bf7027de3.

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoapt-native: Drop obsolete symlinks
Richard Purdie [Tue, 23 Jun 2015 14:13:35 +0000 (15:13 +0100)]
apt-native: Drop obsolete symlinks

Apt no longer uses GLIBC_VER and LIBSTDCPP_VER variables nor appears
to need these odd symlinks. The variables expand to empty enties
which then create symlinks which loop back on themselves.

To fix, drop the obsolete code.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agometa-environment: Ensure we append to the default TOOLCHAIN_NEED_CONFIGSITE_CACHE
Richard Purdie [Fri, 19 Jun 2015 10:25:12 +0000 (11:25 +0100)]
meta-environment: Ensure we append to the default TOOLCHAIN_NEED_CONFIGSITE_CACHE

The toolchain-scripts class was changed to assign to the variable using ??=
which allows users to override it. We therefore need to _append instead
of += to have the change take effect as intended.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agotoolchain-scrpts: Fix sitecache issues with multilib
Richard Purdie [Fri, 19 Jun 2015 10:24:39 +0000 (11:24 +0100)]
toolchain-scrpts: Fix sitecache issues with multilib

The use of TCLIBC in TOOLCHAIN_NEED_CONFIGSITE_CACHE is problematic
since a multilib may have both uclibc and glibc for different multilibs
yet switching between them doesn't change TCLIBC. This would result
in "lib32-glibc" being attempted when lib32 was actually uclibc.

The fix here is to use the virtual providers which bitbake switches
to point correctly at the right things.

This does mean we need to resolve virtual providers but we can do this using
sysroot-providers.

[YCOTO #7607]

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agostaging: Strip files in sysroot
Richard Purdie [Fri, 19 Jun 2015 09:09:40 +0000 (10:09 +0100)]
staging: Strip files in sysroot

Add functionality to strip binaries/libraries going into the sysroot. Whilst
this does fractionally slow down the build, it also significantly reduces the
size of the sstate cache files.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agometa: Add explict getVar param for (non) expansion
Richard Purdie [Thu, 18 Jun 2015 14:14:16 +0000 (15:14 +0100)]
meta: Add explict getVar param for (non) expansion

Rather than just use d.getVar(X), use the more explict d.getVar(X, False)
since at some point in the future, having the default of expansion would
be nice. This is the first step towards that.

This patch was mostly made using the command:

sed -e 's:\(getVar([^,()]*\)\s*):\1, False):g' -i `grep -ril getVar *`

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
9 years agoneard: remove unused service file
Ross Burton [Mon, 22 Jun 2015 14:44:04 +0000 (15:44 +0100)]
neard: remove unused service file

Upstream ships a systemd service file now, so we don't need this anymore.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorpm: Rebrand rpm custom macro paths to be distro specific
Mark Hatle [Wed, 17 Jun 2015 17:04:41 +0000 (12:04 -0500)]
rpm: Rebrand rpm custom macro paths to be distro specific

RPM was previously hard coded to use the 'poky' directory.  We really want
the directory name to be distribution specific.

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorpm: Fix lua 'print' statement capture
Mark Hatle [Wed, 17 Jun 2015 17:04:40 +0000 (12:04 -0500)]
rpm: Fix lua 'print' statement capture

The print statement should capture the output and send it to the script
processing engine, and not display it directly to the screen.

Note, this is only a bug if 'lua' support has been enabled in the RPM
recipe's PACKAGECONFIG.

This patch is from: http://rpm5.org/cvs/patchset?cn=17671

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorpm: Fix CVE-2013-6435
Leonardo Sandoval [Wed, 10 Jun 2015 14:58:38 +0000 (14:58 +0000)]
rpm: Fix CVE-2013-6435

Backport to fix CVE-2013-6435. Description on [1] and original
patch taken from [2].

[1] https://bugzilla.redhat.com/show_bug.cgi?id=CVE-2013-6435
[2] https://bugzilla.redhat.com/attachment.cgi?id=956207

[YOCTO #7181]

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
9 years agorpm: Fix CVE-2014-8118
Leonardo Sandoval [Wed, 10 Jun 2015 14:55:14 +0000 (14:55 +0000)]
rpm: Fix CVE-2014-8118

Backport patch to fix CVE-2014-8118. Description is on [1] and
original patch taken from [2].

[1] https://bugzilla.redhat.com/show_bug.cgi?id=1168715
[2] https://bugzilla.redhat.com/attachment.cgi?id=962159

[YOCTO #7181]

Signed-off-by: Leonardo Sandoval <leonardo.sandoval.gonzalez@linux.intel.com>
9 years agodistutils-common-base: add to, don't set, FILES_${PN}
Joshua Lock [Wed, 17 Jun 2015 08:33:18 +0000 (09:33 +0100)]
distutils-common-base: add to, don't set, FILES_${PN}

If we set FILES_${PN} and a recipe inherits other classes that
modify FILES_${PN} *before* distutils-common-base is included, any
changes to FILES_${PN} made by those classes are lost.

Instead, append the additional directories we want to include in
FILES_${PN}

Signed-off-by: Joshua Lock <joshua.lock@collabora.co.uk>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agopython-2.7/3.4-manifest.inc: add missing dependencies for python-netserver
Henning Heinold [Tue, 16 Jun 2015 18:24:49 +0000 (20:24 +0200)]
python-2.7/3.4-manifest.inc: add missing dependencies for python-netserver

* SimpleHTTPServer imports shutil so we need
  definitly python-shell

* SocketServer can be run without the threading modul,
  but I think we should run the full blown stuff, so
  let us add python-threading

* this work was sponsored by sysmocom GmbH

Signed-off-by: Henning Heinold <henning@itconsulting-heinold.de>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agopython3-ctypes: Fix cross compilation for arm targets
Alejandro Hernandez [Tue, 16 Jun 2015 15:29:13 +0000 (15:29 +0000)]
python3-ctypes: Fix cross compilation for arm targets

When cross compiling for arm targets ctypes compilation fails because
it uses _sysconfigdata from the HOST, this patches makes it use the
one from TARGET fixing compilation of this module

[YOCTO #7873]

Signed-off-by: Alejandro Hernandez <alejandro.hernandez@linux.intel.com>
Signed-off-by: Jonas Göransson <jonas.goransson@qmatic.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowic: removed unused command line option
Ed Bartosh [Wed, 17 Jun 2015 13:49:33 +0000 (16:49 +0300)]
wic: removed unused command line option

Removed -i/--infile wic command line option.
Removed properties_file and properties arguments of
wic_create function.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: code cleanup: superfluous-parens
Ed Bartosh [Wed, 17 Jun 2015 11:47:47 +0000 (14:47 +0300)]
wic: code cleanup: superfluous-parens

Removed unncecessary parents after 'if' 'del' and 'print' keywords.
Fixed pyling warning: Unnecessary parens after 'xxx' keyword

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: oe-selftest: Add 3 tests of 'wic help' command
Ed Bartosh [Tue, 16 Jun 2015 13:26:27 +0000 (16:26 +0300)]
wic: oe-selftest: Add 3 tests of 'wic help' command

Added tests for 'wic test overview', 'wic test plugins' and
'wic test kickstart' commands.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Return error code when wic fails to invoke command
Ed Bartosh [Tue, 16 Jun 2015 11:19:49 +0000 (14:19 +0300)]
wic: Return error code when wic fails to invoke command

Return 1 if command doesn't exist or wic is called without
any commmand.
Return result of invoke_command as wic return code.

Added tests for unsupported command and no command.

Fixed typo in test case test02_createhelp spotted by this fix.

[YOCTO #7856]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agowic: Integrated plugin docstrings into 'wic help plugins' output
Ed Bartosh [Wed, 17 Jun 2015 06:56:44 +0000 (09:56 +0300)]
wic: Integrated plugin docstrings into 'wic help plugins' output

Added mechanism to show docstrings of plugin classes as a part of
plugins help page.
For missing plugins the following warning message is shown:
<class '<plugin class spec>'> is missing docstring.

[YOCTO #7118]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowic: Add missing docstrings to plugin classes
Ed Bartosh [Wed, 17 Jun 2015 08:53:04 +0000 (11:53 +0300)]
wic: Add missing docstrings to plugin classes

Docstings from plugin classes are used as part of
output of 'wic help plugins'. Adding them makes help page
more informative.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowic: Code cleanup: bad-classmethod-argument
Ed Bartosh [Wed, 17 Jun 2015 11:25:23 +0000 (14:25 +0300)]
wic: Code cleanup: bad-classmethod-argument

Fixed wrong name for the first argument in class methods.
Pylint complains about the issue this way:
    Class method should have 'cls' as first argument

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agowic: Document creation of empty partition
Ed Bartosh [Tue, 16 Jun 2015 10:18:35 +0000 (13:18 +0300)]
wic: Document creation of empty partition

Added short explanation of how to create empty pattition
to the 'wic help kickstart' output.

[YOCTO #7131]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agooe-selftest: wic: Fix testcase
Ed Bartosh [Tue, 16 Jun 2015 06:43:17 +0000 (09:43 +0300)]
oe-selftest: wic: Fix testcase

Fixed test05_build_artifacts testcase by using values of MACHINE
and BUILD_SYS bitbake variables in paths to artifacts.

Test was failing because of hardcoded machine(qemux86) and
build_sys(qemux86-poky-linux) in artifact paths.

[YOCTO #7730]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: Properly delete partition table
Ed Bartosh [Tue, 16 Jun 2015 04:56:25 +0000 (07:56 +0300)]
init-install: Properly delete partition table

Fixed deletion of the partition table by increasing
amount of sectors from 2(correct for msdos PT) to 35 as
GPT size is 34 sectors + 1 sector for protective MBR.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: code cleanup: replace /dev/$device -> $device
Ed Bartosh [Mon, 15 Jun 2015 15:47:56 +0000 (18:47 +0300)]
init-install: code cleanup: replace /dev/$device -> $device

Shortened code by including /dev/ prefix into variable.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: code cleanup: Replace tabs with spaces
Ed Bartosh [Mon, 15 Jun 2015 15:18:16 +0000 (18:18 +0300)]
init-install: code cleanup: Replace tabs with spaces

Cleaned up spaces from init-install* shell scripts.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: Specify partition name in parted command line
Ed Bartosh [Sun, 14 Jun 2015 20:05:28 +0000 (23:05 +0300)]
init-install: Specify partition name in parted command line

parted allows to use names for partitions if GPT partition table
is used on the device. msdos partitioning can have only partition
types: 'primary', 'logical' or 'extended'.

Used meaningful partition names in parted command line for GPT
partitioning.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: Specify filesystem type in parted command line
Ed Bartosh [Sun, 14 Jun 2015 15:45:21 +0000 (18:45 +0300)]
init-install: Specify filesystem type in parted command line

Explicitly specified filesystem type for parted mkpart command.
This makes partition table to look more informative.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: Implement UUID support
Ed Bartosh [Sun, 14 Jun 2015 13:26:27 +0000 (16:26 +0300)]
init-install: Implement UUID support

Used partition UUID in kernel command line to specify root partition.
Searched root device by file system uuid in GRUB configuration.
Used partition UUID in /etc/fstab to specify swap partition.
Used filesystem UUID in /etc/fstab to specify boot partition.

[YOCTO #6101]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install: Use GPT table with GRUB 2
Ed Bartosh [Thu, 11 Jun 2015 16:01:35 +0000 (19:01 +0300)]
init-install: Use GPT table with GRUB 2

Changed partition type from 'msdos' to 'gpt'.
Added special partition for grub stage2 bootloader.

NOTE: This is done only for GRUB 2 as legacy GRUB is
rarely used and doesn't support GPT partitions.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinit-install-efi: Implement UUID support
Ed Bartosh [Wed, 10 Jun 2015 14:49:32 +0000 (17:49 +0300)]
init-install-efi: Implement UUID support

Using UUID in favor of device names is more reliable as
UUID names are persistent.

Device names can change as the order of adding device nodes
is arbitrary. This sometimes results in device names switching
on each boot, which can cause system fail to boot.
Persistent naming solves these issues.

Used partition UUID in kernel command line to specify root partition.
Used partition UUID in /etc/fstab to specify swap partition.
Used filesystem UUID in /etc/fstab to specify boot partition.

[YOCTO #6101]

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agobusybox: Enable UUID-related options
Ed Bartosh [Mon, 15 Jun 2015 15:01:47 +0000 (18:01 +0300)]
busybox: Enable UUID-related options

Enabled MOUNT_LABEL and VOLUMEID* features for busybox
mount to understand 'UUID=' syntax in fstab.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoinitramfs-live-install: Add blkid to initramfs
Ed Bartosh [Wed, 10 Jun 2015 14:40:43 +0000 (17:40 +0300)]
initramfs-live-install: Add blkid to initramfs

Added util-linux-blkid to the list of dependencies of
initramfs-live-install and initramfs-live-install-efi.

This is a part of the work to support partiion UUID in installer.
blkid is going to be used to get partition and filesystem UUIDs.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agoimage-live: Set syslinux timeout to 5s
Ed Bartosh [Wed, 10 Jun 2015 13:42:33 +0000 (16:42 +0300)]
image-live: Set syslinux timeout to 5s

Increased syslinux timeout to 5s as default 1s timeout
is not enough to notice syslinux prompt on some devices.

Signed-off-by: Ed Bartosh <ed.bartosh@linux.intel.com>
9 years agort-tests: bump version to v0.92
Maxin B. John [Tue, 16 Jun 2015 21:25:07 +0000 (23:25 +0200)]
rt-tests: bump version to v0.92

Version upgrade to v0.92
Rebased below listed patch to newer source code:
added-missing-dependencies.patch

Signed-off-by: Maxin B. John <maxin.john@enea.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agobabeltrace: Fix invalid pointer free with trace collection
Roy Li [Tue, 16 Jun 2015 02:50:04 +0000 (10:50 +0800)]
babeltrace: Fix invalid pointer free with trace collection

This fixed the bug https://bugs.lttng.org/issues/790

Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agodbus-test: Upgrade 1.8.10 -> 1.8.18
Jussi Kukkonen [Tue, 16 Jun 2015 08:00:59 +0000 (11:00 +0300)]
dbus-test: Upgrade 1.8.10 -> 1.8.18

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
9 years agodbus: Upgrade 1.8.16 -> 1.8.18
Jussi Kukkonen [Mon, 15 Jun 2015 14:01:00 +0000 (17:01 +0300)]
dbus: Upgrade 1.8.16 -> 1.8.18

Signed-off-by: Jussi Kukkonen <jussi.kukkonen@intel.com>
9 years agopackagegroup-core-tools-profile.bb: Remove oprofile.
Philip Balister [Mon, 1 Jun 2015 20:17:22 +0000 (16:17 -0400)]
packagegroup-core-tools-profile.bb: Remove oprofile.

Oprofile functions duplicate perf. Since perf is easier to use drop oprofile
from the core-tools-profile package.

Also, the oprofile-ui package opens a port on the device. removing this
package reduces a possible security issue for devices with profiling tools.

Signed-off-by: Philip Balister <philip@balister.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agopackagegroup-core-sdk.bb: Drop distcc from the core-sdk.
Philip Balister [Mon, 1 Jun 2015 20:17:21 +0000 (16:17 -0400)]
packagegroup-core-sdk.bb: Drop distcc from the core-sdk.

distcc has been used in the past to speed up native compile across
several machines. This is less an issue on modern embedded systems.

Also, improvements in sdk generation have reduced the need for on target
compile of large projects.

Since the distcc packages start a daemon at boot, drop it from the
packagegroup to avoid having some slight runtime impact on systems
installing an sdk.

Signed-off-by: Philip Balister <philip@balister.org>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu-internal: set mutual exclusiveness for serial and nographic options
Jagadeesh Krishnanjanappa [Tue, 16 Jun 2015 10:48:30 +0000 (16:18 +0530)]
runqemu-internal: set mutual exclusiveness for serial and nographic options

Use "-nographic" option only if "serial" option is not
specified. Otherwise we get below error when
'runqemu <kernel_image> <rootfs_image> serial' is executed,

(snip)
QEMU 2.2.0 monitor - type 'help' for more information
(qemu) qemu-system-aarch64: -serial stdio: cannot use stdio by multiple
character devices
-- CUT --

Signed-off-by: Jagadeesh Krishnanjanappa <jkrishnanjanappa@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu-script: define console for qemuarm NFS booting
Jagadeesh Krishnanjanappa [Tue, 16 Jun 2015 10:48:28 +0000 (16:18 +0530)]
runqemu-script: define console for qemuarm NFS booting

Add console=ttyAMA0,115200 as one of the boot parameters
for qemuarm, in order to print bootlog messages on the
console.

Signed-off-by: Jagadeesh Krishnanjanappa <jkrishnanjanappa@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu-internal: add support to boot arm64 qemu target via NFS
Jagadeesh Krishnanjanappa [Tue, 16 Jun 2015 10:48:29 +0000 (16:18 +0530)]
runqemu-internal: add support to boot arm64 qemu target via NFS

Add required boot parameters inorder to boot arm64 qemu
target via NFS

Signed-off-by: Jagadeesh Krishnanjanappa <jkrishnanjanappa@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agorunqemu: fix MACHINE being detected as qemuarm for qemuarm64 kernel image
Jagadeesh Krishnanjanappa [Tue, 16 Jun 2015 10:48:27 +0000 (16:18 +0530)]
runqemu: fix MACHINE being detected as qemuarm for qemuarm64 kernel image

Basically, runqemu script autodetects MACHINE type based on the
kernel image name; if MACHINE name is not specified. Since 'qemuarm'
string is common in both qemuarm amnd qemuarm64 kernel image names, the
MACHINE type is considered as 'qemuarm' even when qemuarm64 kernel
image name is given in command line.

Signed-off-by: Jagadeesh Krishnanjanappa <jkrishnanjanappa@mvista.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoqt4-4.8.7: restore arm64 patches which are dropped by mistake
Kai Kang [Wed, 17 Jun 2015 01:32:47 +0000 (09:32 +0800)]
qt4-4.8.7: restore arm64 patches which are dropped by mistake

During upgrade qt4 from 4.8.6 to 4.8.7, four arm64 patches are removed.
But three of them are dropped by mistake that they are not merged by
upstream and still needed.

Restore them.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocups: upgrade from 2.0.2 to 2.0.3
Kai Kang [Wed, 17 Jun 2015 07:17:53 +0000 (15:17 +0800)]
cups: upgrade from 2.0.2 to 2.0.3

Upgrade cups from 2.0.2 to 2.0.3. In its release log, it fixes
CERT VU #810572/CVE-2015-1158/CVE-2015-1159.

Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoopenssl: fix building on x32 systems
Cristian Iorga [Mon, 15 Jun 2015 21:32:15 +0000 (00:32 +0300)]
openssl: fix building on x32 systems

Fix build on Fedora 21 i686.

When building on x32 systems where the default type is 32bit,
make sure that 64bit integers can be represented transparently.

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoRevert "distrodata: Remove unnecessary include of package_regex.inc"
Ross Burton [Tue, 16 Jun 2015 09:31:29 +0000 (10:31 +0100)]
Revert "distrodata: Remove unnecessary include of package_regex.inc"

Instead of asking the user to include a number of configuration files, the class
can do this.  Next step is to fix the documentation.

This reverts commit 2a4ee94667d4d356cad2ca6d60a100a30c92737b.

Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agolttng-modules: drop makefile patch
Nathan Lynch [Mon, 15 Jun 2015 18:37:24 +0000 (13:37 -0500)]
lttng-modules: drop makefile patch

The lttng-modules Makefile allows us to specify the kernel we're
building against via the KERNELDIR variable on the build command line.
Use that and drop the patch (which tends to need updating whenever the
upstream Makefile changes).

Signed-off-by: Nathan Lynch <nathan_lynch@mentor.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agoinsane.bbclass: fix false negative in build-deps QA check
Patrick Ohly [Mon, 15 Jun 2015 16:05:13 +0000 (18:05 +0200)]
insane.bbclass: fix false negative in build-deps QA check

When a recipe called 'foobar-test' links against 'foobar' without
listing that in DEPENDS, the build-deps check misses that error
because it looks for 'foobar' in a package string containing (among
others) the 'foobar-test' name, leading to the incorrect conclusion
that the package is listed as dependency.

The 'packages' string needs to be split into individual package names
before the check. Doing that once directly after reading the value is
more efficient than splitting inside package_qa_check_rdepends() because
the caller also needs the individual components.

Also use a set to speed up the 'package in packages' check.

Signed-off-by: Patrick Ohly <patrick.ohly@intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>
9 years agocopyleft_filter.bbclass: Allow to filter on name
Mariano Lopez [Tue, 9 Jun 2015 12:46:48 +0000 (12:46 +0000)]
copyleft_filter.bbclass: Allow to filter on name

The archiver uses a license based filter to provide the source code.
This patch allows to search on name based on two new variables (COPYLEFT_PN_INCLUDE,
COPYLEFT_PN_EXCLUDE). Both variables are empty by default.
The filter by name has higher priority than the license filter.

[YOCTO # 6929]

Signed-off-by: Mariano Lopez <mariano.lopez@linux.intel.com>
Signed-off-by: Ross Burton <ross.burton@intel.com>