buildstats: Update to ensure files are closed

python3 doesn't like files being left open. This updates the code style
to ensure file are closed.

(From OE-Core rev: ec74285bd2108f12f33fc8ac0dc1d124ab48be21)

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

clases/lib: Use modern exception syntax

Update older code to use modern exception handling syntax which
is the form accepted by python 3.

(From OE-Core rev: b010501cd089e649a68f683be0cf4d0aac90fbe3)

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

core-image-minimal-initramfs: Remove odd quote characters

The quote characters present trigger python 3 characters type warnings,
we don't need them so replace them with normal quote characters.

(From OE-Core rev: 3c104443506cb89d72944e46096a94a80838a707)

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

xserver-xorg: Add xkbcomp to RRECOMMENDS

OE-Core commit bdcc5e8f1286d288baf410458efc39a59b68d751 removed xkbcomp
from RDEPENDS for xkeyboard-config but X server still needs it otherwise
it fails to start.

(From OE-Core rev: f2330ebc3071d780cbc6d1ddab5c54bfadf8fffc)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-configure-cross.inc: Remove info generation

Some distros have newer makeinfo which does not go well with
anything below gcc 4.8, this fix avoids regenerating info files

(From OE-Core rev: 50ac2ed5299e2b47b2f3fd9c9cde9d733d2f8d9b)

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

bootimg.bbclass: add comment for NOHDD

Add comment for NOHDD which is used for skipping building the HDDIMG if
set to 1

(From OE-Core rev: 9843ad9d783f68b97fedfe5b435528538bb26c1c)

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

gnutls: upgrade to 2.12.23

Importantly, this fixes CVE-2013-1619.

Upstream doesn't use GNU as a host, so update the SRC_URI.

remove-gets.patch isn't required anymore, obsolete_automake_macros.patch is
merged upstream, and gnutls_fix_for_automake_1.12.1.patch doesn't seem to be
needed.  It was merged and reverted upstream, and my build without it succeeded.

(From OE-Core rev: 9a6395076984350b1dd7005453f97233bbb43132)

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

oprofileui(-server): upgraded to GIT revision f168b8bf

'obsolete_automake_macros.patch' is not needed anymore, because it
was integrated in the specified commit.

(From OE-Core rev: 4c7120a24cef2ff6b9df821c69153363c938ac77)

Signed-off-by: Bogdan Marinescu <bogdan.a.marinescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4: remove trailing slash from prefix

The trailing slashs on "Prefix=${prefix}/" and "-prefix ${prefix}/" are passed
through to the generated pkgconfig files and may be joined to paths like
"/include" yielding a final path with a double forward-slash (eg.
"/usr//include"). This may end up in the debugging symbols in other applications
or libraries which depend on qt4 which in turn causes the debugedit program to
fail with the message "canonicalization unexpectedly shrank by one character"
when it tries to replace the double forward-slash with a single forward-slash.
Thus the function split_and_strip_files fails and ultimately do_package fails.

As this slash is removed from the prefix it is added into the regular expression
used to fix up pkgconfig files later in the recipe.

This error was seen in vlc in meta-openembedded and should be solved by this
change in openembedded-core.

(From OE-Core rev: 61a7329a5c1f0b84f447256134f5a77917f07f38)

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gdb: tabs -> spaces

(From OE-Core rev: 8ec07ae2ebfd97aa1413a3c3be0dc612eee7e929)

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gdb: AArch64 lacks lttng support

(From OE-Core rev: 2288c2a9abe828ffb0bbd527105fb633ac818354)

Signed-off-by: Marcin Juszkiewicz <marcin.juszkiewicz@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

craklib:fix testnum and teststr failed

Error log:
...
$ ./testnum
(null).pwd.gz: No such file or directory
PWOpen: No such file or directory

$ ./util/teststr
(null).pwd.gz: No such file or directory
PWOpen: No such file or directory
...

Set DEFAULT_CRACKLIB_DICT as the path of  PWOpen

(From OE-Core rev: 9c78d2ef5291b29fa313f9d4bf6a81c395207d69)

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

cracklib: fix default dictionary should be generated for target endianness

The previous dict files are NOT byte-order independent, in fact they are
probably ARCHITECTURE SPECIFIC.
Create the dict files in big endian, and convert to host endian while
load them. This could fix the endian issue on multiple platform.

[Bug #4419]

(From OE-Core rev: 075d5a19bb56b0bf492e7dc7f453c2e23139104b)

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

linux-yocto/3.8: udpate to v3.8.11

Update the linux-yocto-3.8 SRCREVs to the korg 3.8.11 -stable release.

(From OE-Core rev: b60d04331f84e06fae6c074acab045c3e003929b)

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

linux-yocto/3.4: update to v3.4.43, ltsi-65311d8 and v3.4.42-rt56

Updating the 3.4 kernel trees to the latest korg, LTSI and -rt stable
releases.

(From OE-Core rev: 98446039e210936a32e70559a379131e8e0d832f)

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

kern-tools: anchor and delimit regexs

Updating the kern-tools SRCREV with the following fix:

    Updateme is responsible for updating an existing meta-series with new patches,
    configs and tree manipulations. To do this, it first checks for an existing
    board description and generates one if required. It then searches for features
    and fragments to be applied for the tree.

    There were two problems:

     - A top level board description is detected via the presence of "define"
       directives that indicate the board name, the arch and kernel type. The
       test for define would match on patches or fragments with 'define' in their
       name, and would incorrectly use that file as the top level board description.

       This is fixed by ensuring that only defines at the start of a line, or preceded
       by whitepace match.

     - When searching for features that were indicated as 'addon' or 'optional', the
       search would find, and apply, any feature with the passed name as substring
       versus an exact match.

       This is fixed by ensuring that the matched feature name is /<feature name>
       versus <feature name>

(From OE-Core rev: 57ae1e412a35d827f84bf9b1f48747bf703f84b7)

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

linux-yocto/3.4: iwlwifi, ext4, brtfs warning and -rt fixes

Updating the yocto 3.4 SRCREVs for the following fixes:

206d4bb powerpc/perf: run irq-work under softirq context in rt
8a969f9 arm/perf: run irq-work under softirq context in rt
79ba946 iwlwifi: fix unused variable warning
59d93fa ext4: remove unused variable in ext4_update_super()
2385eee Btrfs: fix compile warnings in extent_io.c

(From OE-Core rev: 96113b28031583af1dd42fe7660a9d5ced36d1b6)

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

dmidecode: fix fetch path to use BP

This is so the multilib prefix is not used

(From OE-Core rev: 50e26e553152485461d096463db0fb4f4fa826d6)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-configure-target: Add --enable-dependency-tracking to EXTRA_OECONF

Fixes errors seen while compiling gcc 4.8 for target, errors like

| /home/khem/work/angstrom/build/tmp-angstrom_next-eglibc/work-shared/gcc-4.8.0-r0/gcc-4.8.0/libatomic/cas_n.c:125:1: fatal error: opening dependency file .deps/cas_1_.lo.Ppo: No such file or directory

which are due to disabled dependency tracking but dependency files being still generated
while compiling libatomics.

(From OE-Core rev: c83cbc80af078d7be5deec44bba6e0ca29a5ee1a)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

tcmode-default.inc: Pin binutils to 2.23.2

(From OE-Core rev: 27759800b020ad52c2e91b8eea3acf9216860a0a)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

binutils: backport whitespace fix

This is a backport of a whitespace skipping issue: PR14987 and PR14887

(From OE-Core rev: f4312eaeb6eb2b7196be2b968e2e468e705fb2cf)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

binutils: Upgrade 2.23.1 -> 2.23.2

(From OE-Core rev: 202f46b139ab20079e97cdeb147d59e23bcdfe01)

Signed-off-by: Khem Raj <raj.khem@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

guile: Update to 2.0.9

fix_parallel_build_issue.patch is now part of upstream.

(From OE-Core rev: 87cb30d988843ab1159d23789cd334aaf2a77e34)

Signed-off-by: Jonathan Liu <net147@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemu: Ensure kernel nfsd module is enabled

This will ensure that qemu images that include the nfs-server package have the kernel
feature correctly enabled

(From OE-Core rev: 57c718c6288f2a2538173cdd3d401d70f939a40a)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gdb-cross: gdb-cross requires readline-native

gdb depends on readline so gdb-cross needs to depend on readline-native to
build successfully.

(From OE-Core rev: ee132d68220d7f515760d47db57e00d1d8263a1a)

Signed-off-by: Mike Crowe <mac@mcrowe.com>
Acked-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libgnome-keyring: remove, nothing in oe-core uses this anymore

(From OE-Core rev: d5f985f51b250b88dd910ca3a1fc8f2992271e60)

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

gnome-keyring: remove, nothing uses it in oe-core

(From OE-Core rev: fbefbf60fd18dae135b96d0bfd2db832f568fea9)

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

libsoup: upgrade to 2.42.1

Upgrade to the latest stable release, and fix the dependencies.

gnome-keyring use has been removed from libsoup, and sqlite3 is now a hard
dependency.  The optional libsoup-gnome library is now a small stub, so enable
it by default but allow it to be disabled via PACKAGECONFIG.

(From OE-Core rev: 3f093e24e78cc8a553bebc03ec9f3a32700981e6)

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

packagegroup-core-x11-sato: Add udev-extraconf to maintain compatibility

Since we have removed udev-extraconf as a RRECOMMENDS in udev
add it back here to maintain compatibility

(From OE-Core rev: 8e2df16e9ab7433a868da2a3db4a25441b92d0ed)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

packagegroup-core-lsb: Add udev-extraconf for LSB compatibility

(From OE-Core rev: ce6927b0d662c8929f514c92bb0e96c54bf47446)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev: do not recommend udev-extraconf

same as already sent by Koen Kooi [1]

[1] http://patches.openembedded.org/patch/48183/

(From OE-Core rev: b190d9d692bde12cce1062bd0cba65a5b7fc8188)

Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd: do not recommend udev-extraconf for udev

* adding this extra functionality is an image- or distro-decision
* at least automount-rule/script breaks other automount solutions causing
  misleading kernel messages as 'VFS: could not find a valid V7 on sda'

same as already sent by Koen Kooi [1]

[1] http://patches.openembedded.org/patch/48139/

(From OE-Core rev: 860cd2f9ba5803824ab59bf9c0e84b83f7907dc1)

Signed-off-by: Andreas Müller <schnitzeltony@googlemail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemux86: disable paravirt guest in linux 3.4, causes test failures

The shutdown sanity test that we previously saw on qemux86-64 in 2.6.37 have
started to re-appear reproducably on qemux86 in 3.4 in the Danny series.  Until
this can be root-caused and fixed in the kernel, disable paravirtualisation
across all kernels and all qemu machines.

[ YOCTO #4196 ]

(From OE-Core rev: e20c3bf6ce0d8ab5653739dd02555186b78b8912)

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

archive-*-source.bbclass: handle dependency in anonymous function

Using "before do_rootfs" would have this task added to do_rootfs
dependencies regardless of the filtering options. Instead, add this
dependency in the anonymous python function.

(From OE-Core rev: 2400a74923e51e17ecfa94c2d63908b0b2aac76a)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cronie: Convert crontab to a system user

crontab is meant to be a system user, not a normal user of the system
so lets pass the correct parameters to the system. If we don't do this
it interferes with the setup of normal users, particularly if some
preset list of users is desired.

(From OE-Core rev: ea6b28cad959d9ebbaf0ecbd695d1d72c0c01151)

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

grub-efi-native: Cleanup whitespace

A few of the recent commits have used inconsistent whitespace with
respect to tabs and spaces for indentation and alignment. Correct it.

(From OE-Core rev: 3472c1f7ab409cd91c1d4782d9e00880b84e3ae8)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

neard: Rename the recipe as we are already in 0.10

(From OE-Core rev: 0f26f430d2bcca37eea63740dae51ac1e49bc550)

Signed-off-by: Jukka Rissanen <jukka.rissanen@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

buildhistory: change default BUILDHISTORY_DIR to TOPDIR

* like PERSISTENT_DIR I find BUILDHISTORY_DIR more persistent
  then tmp-eglibc directory
* WARN: people with existing checkouts in TMPDIR/buildhistory
  need to move it to new location first (in order to save history)

(From OE-Core rev: d3ad9a4c573fec08b5606a4c3e86f93cb95e8718)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: fetch2/svn: Do not save authentication

(Bitbake rev: a9b1e7c2ad13ff47acb629bae082d4f96e174f99)

Signed-off-by: Simon Vanveerdeghem <simon.vanveerdeghem@barco.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: bitbake: remove stack trace depth limitation

In case of an error in bitbake, the stack trace is limited to 5
items. This is an endless source of confusion and it makes bugs
reports impractical, since a full stack trace can't be included
in the bug report. This patch simply removes the depth limit.

(Bitbake rev: 02629c42fb09413d9da16cfe43e03338ce7db3ff)

Signed-off-by: Bogdan Marinescu <bogdan.a.marinescu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: data_smart.py: add some internal bitbake variables to configuration hash

Take __BBTASKS, __BBHANDLERS and __BBANONFUNCS into account when
computing the configuration hash.

[YOCTO #4447]

(Bitbake rev: 260ced7452405fc43ce3d9dd6798236aa07cc716)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: prserv/cooker: Handle PRService errors cleanly

Current if the PR Service fails to start, bitbake carries on regardless or
hangs with no error message. This adds an exception and then handles it correctly
so the UIs correctly handle the error and exit cleanly.

[YOCTO #4010]

(Bitbake rev: 949c01228a977c3b92bfc0802f6c71b40d8e05b3)

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

bitbake: goggle: fix invalid error check

runCommand was returning 'error' but checking for 'err'
resulting in an exception.

(Bitbake rev: 263aa44ed47e0cb03e1a97e7ed4a50bd9f828c78)

Signed-off-by: Martin Donnelly <martin.donnelly@ge.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: Update to version 1.19.0 for master

(Bitbake rev: c47088a86fd8ad06c1810d04d9537c4cd01e8bef)

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

ia32-base.inc: remove inapropriate grub dependency

There is no good reason for ia32 machines to have hard dependency on grub,
as there are other bootloaders available for ia32 platforms.

(From OE-Core rev: d03c0c24704c6ab6d2cfcf9bf705f6ace2a247cc)

Signed-off-by: Tomas Frydrych <tomas@sleepfive.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dmidecode: Add dmidecode to dev tools

Dmidecode reports information about your system's hardware as described
in your system BIOS according to the SMBIOS/DMI standard.

This recipe was updated to version 2.12 and cleaned up slightly from the
OE 2.10 version and proposed here for inclusion in oe-core.

(From OE-Core rev: 120495843118b55e865bb8d87933bd585fced992)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-4.7: fix gcc ICE when building with gcc-4.8

Backport fix for internal compiler error when building gcc-4.7.2 with a
gcc-4.8.0 host gcc.

See upstream bug here: http://gcc.gnu.org/bugzilla/show_bug.cgi?id=56308

(From OE-Core rev: b1dc91969f9bb0c2a3a4336f5e9a2f57aabb9f78)

Signed-off-by: Josh Cartwright <josh.cartwright@ni.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

fs-perms.txt: Don't tinker with file modes in /usr/src/debug

Since 6775feb9fe935ab01fd9cae2b2d3fce5824a9a72 our local "copy" of the
debug sources has in fact been hardlinked to ${S} and potentially other
places too.  This means that any modifications we make to these files
might have wider consequences than intended.

Avoid this potential pitfall by telling fixup_perms() to leave the file
modes in this directory alone.  No great harm will result from shipping
debug sources with a mode other than 0644: if the mode was permissive
enough for us to compile the sources in the first place then it must
also be permissive enough for subsequent debugging.

(From OE-Core rev: 91b02c65d83811738d4c0e4b7c454459430c8b9b)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-runtime: Avoid error when trying to remove nonexistent directories

If we didn't build libgomp then we won't have installed anything into
${infodir} or ${libdir}/gcc/${TARGET_SYS}/${BINV}/finclude.  Check
whether those directories exist before trying to remove them, else we
will lose.

(From OE-Core rev: 507e14ecdc5b4ff2ee7f1128d9f30c2948e10d5a)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

archive-*-source.bbclass: have do_dumpdata_create_diff_gz task run before do_rootfs

do_rootfs[cleandirs] contains ${S} and, if do_rootfs task starts before
the do_dumpdata_create_diff_gz is finished, an error will occur in the
process because the directory will be removed while still needed by the
create_diff_gz() function.

This patch will force the do_dumpdata_create_diff_gz task to run before
do_rootfs when the final image is created.

[YOCTO #4310]

(From OE-Core rev: cd90be31571178d6822dba5a94a2795209a3576c)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa-demos: Use PACKAGECONFIG to enable/disable features

This allow selection of following features:

 drm, egl, freetype2, gbm, gles1, gles2, glut, osmesa, vg, wayland, x11

The x11 features is enabled depending on distro features but the
wayland has not been enabled as it does not work with Wayland
1.0. Rest were enabled for a sane default.

(From OE-Core rev: dc2de313febfc817640e4e55c75d0cb21c70c0ba)

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image_types: fix default location of kernel when generating elf images

Generation of elf images fails because kernel images are no longer staged under
${STAGING_DIR_HOST}/kernel, but rather ${STAGING_DIR_HOST}/usr/src/kernel. This
patch fixes the path to point to the correct location.

(From OE-Core rev: 6e57a3231fb29f869d476b8511d6f4393f82651b)

Signed-off-by: Tomas Frydrych <tomas@sleepfive.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

image.bbclass: Don't mark do_rootfs and do_build as nostamp

When image.bbclass was first conceived (as rootfs_ipk.bbclass), oemake
was unable to figure out when the inputs to do_rootfs had changed in
such a way that it would need to be rerun.  Absent any reliable
information of this kind, it was necessary to adopt the conservative
assumption that images always needed rebuilding and, to this end,
do_rootfs and do_build in image recipes were marked as nostamp.  (The
nostamp annotation for do_build was added in oe-classic commit
80d622e0c1dbc284858e01d3eb670303a6d8cdf9 from January 2005; the exact
point at which do_rootfs was made nostamp is obscure, but it predates
oe-classic revision 63fad339e01d4b16105146c32a61f24460397126.)

This situation persisted for the following eight years but, during that
period, oemake evolved into today's sstate-enabled bitbake, oe itself
gained the OEBasicHash algorithm which can reliably detect changes to
the input data for do_rootfs or any other task, and OEBasicHash was made
the default for oe-core in 4199efed48005a62267fa3374c33b13627d85f44
(June 2012).

Given these various changes in the wider landscape, there is no longer
any obvious benefit to having these two tasks marked as nostamp, and
indeed having them so marked causes needless rebuilds which can be
annoying.  Anybody who does genuinely wish to rerun these tasks when
nothing in the input has changed can run "bitbake -f", just like with
any other task.  So, let's remove the nostamp annotations and just let
bitbake figure out when to rerun them.

(From OE-Core rev: 8505008c115efb54d18e5f25441c7a938a32ffaf)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classes/image: allow complementary package globs to be extended

Make it easy for the wildcard specifications for complementary package
features to be extended outside of image.bbclass. For example, to add a
new "foo-pkgs" item that could be added to IMAGE_FEATURES that would
cause *-foo packages to be installed for all packages currently in the
image, you can now use this line at the global level:

COMPLEMENTARY_GLOB[foo-pkgs] = "*-foo"

Implements [YOCTO #4228].

(From OE-Core rev: 4e39c1b9dbba1d2b07ffc2c6a1a252fc8f7680ee)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

path.py: Deal with race issue

The change to use copyhardlinktree in some of the sstate code instead of
copytree exposed a race condition. This is due to cp failing if it finds
a directory doesn't exist yet some other process creates it while cp was
trying to create it itself. tar doesn't error in this case.

To fix this we need to create the directory structure with tar, then
use cp to hardlink the files. Messy but probably worth doing.

I also took the opportunity to remove src_bak since the code is neater
without it.

(From OE-Core rev: 2f954a9a6932f1e6c564e7e7aacaac628a75eed7)

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

mkefidisk: Unmount after partitioning

Some automounters are rather overzealous and like to mount things
immediately after partitioning. This can happen if the disk is being
reused and the partitions align exactly with the existing partitions
which have already been formatted. Move the unmount code into a function
and call it before and after partitioning.

(From OE-Core rev: f1854e458e5e77806b1fc837033500fa91272261)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Make toolchain output name using SDK_VERSION instead of DISTRO_VERSION to be consistent [yocto #2342]

(From OE-Core rev: 733f1d2a1bcf1c115613e21f7d09f1b4bee216ce)

Signed-off-by: Jessica Zhang <jessica.zhang@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4: Fix sundry erroneous assumptions about ${prefix}

FILES_${QT_BASE_NAME}-tests was hard-coded to "/usr/tests/..." but Qt
actually installs these files into ${prefix}/tests.

Conversely, FILES_${PN}-dbg in qt4.inc was defined in terms of
${exec_prefix}/src, which appears commendable but doesn't actually match
where package.bbclass will put the sources since the latter is
hard-coded to /usr/src.

This fixes a large number of "installed but not shipped" warnings
when ${prefix} is set to something other than /usr.

(From OE-Core rev: c0cfd4f84c8f4843027332cfd6cf99c452c50dbb)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dpkg: Ensure target packages don't reference perlnative

Currently a load of scripts in ${bindir} start with:

which is undesireable, this patch fixes that.

(From OE-Core rev: e3634ec359a71c9858698cb597d23c05b1184d2e)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

adt-installer: do not dereference symlinks when relocating

This was already fixed for meta-toolchain.

[YOCTO #4157]

(From OE-Core rev: 9e9331a4b859c4205c2a0406ab0224ac575e351b)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

netbase: upgrade to 5.1

netbase dir has generic content, hence the name change;

(From OE-Core rev: 6b78673934e1f444991b2743957b362443c6a7cd)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

systemd-serialgetty: change to handle SERIAL_CONSOLES

SERIAL_CONSOLES is now set from SERIAL_CONSOLE if not already set.

This is a little bit gnarly because we have to create a separate service
file for each different baud rate; assume that the first baud rate is
the default, thus preserving the previous behaviour in the event there
is only one baud rate in use.

This change also installs the service file before modifying it in place,
allowing do_install to re-execute properly; additionally the service
file now has the correct permissions (i.e. no execute bit set).

(From OE-Core rev: cd89a71d0c3dce14d59134667bc47fa8210cfa7a)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

shadow-securetty: change to handle SERIAL_CONSOLES

SERIAL_CONSOLES is now set from SERIAL_CONSOLE if not already set. This
change also installs the file before modifying it in place, allowing
do_install to re-execute properly.

(From OE-Core rev: 013a1e304e6845eda97b94d133ad75bbccff0b9a)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sysvinit-inittab: remove specific handling of SERIAL_CONSOLE

We already handle SERIAL_CONSOLES here and this is now set from
SERIAL_CONSOLE if not already set.

(From OE-Core rev: 40acb88afb2ebd2d468bb2fce51c6ce6bf3d8403)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake.conf: set SERIAL_CONSOLES from SERIAL_CONSOLE by default

This is the first step in deprecating SERIAL_CONSOLE without affecting
machine configurations that still use it.

(From OE-Core rev: 3f0d665384e6d1b7aa2854a9cc4f13e0961bacb7)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sysvinit-inittab: ensure unique label for SERIAL_CONSOLES entries

The label field in /etc/inittab entries needs to be unique, and the
numeric label being used for the SERIAL_CONSOLES getty entries was
clashing with the entries added for standard ttyX entries added via
SYSVINIT_ENABLED_GETTYS. Use the part after "tty" in the device name
(which is what the comment further down explicitly says should be done)
as the label rather than a simple incrementing number.

Fixes [YOCTO #4374].

(From OE-Core rev: 28d3202befcec72554885f8ea9cb7985523b89f5)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

xkeyboard-config: Remove xkbcomp from RDEPENDS_${PN}

Nothing in this package uses xkbcomp so depending on it here is not appropriate.
Whatever package(s) is/are invoking that binary should be the ones to depend
on it.

(From OE-Core rev: bdcc5e8f1286d288baf410458efc39a59b68d751)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate.bbclass: make hard links for staging files

Make hard links for staging files instead of copy to save the disk space
(3G will be saved for a core-image-sato build), and it doesn't affect
much on the build time.

The following directories are affected:
1) The sysroot
2) The DEPLOY_DIR
3) The pkgdata

[YOCTO #4372]

(From OE-Core rev: 5853e0f482b22258c909268fe71673a29e31989b)

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

tune-thumb.inc: Remove, replaced by arm/feature-arm-thumb.inc

(From OE-Core rev: f4b451c8ad8f857b1789d75d68ce8ea8fc73542e)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

neard: add systemd unit file

Add a systemd unit file, and respect the sysvinit feature when installing the
init script.

(thanks to Jukka Rissanen for the unit file)

(From OE-Core rev: e86538db13605953465aacc3f2fb7d719dad919e)

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

sudo: disable SSP for auxiliary build tools mksiglist and mksigname

The do_compile_prepend() fragment ensures that the non-installable build
tools mksiglist and mksigname are built using the BUILD_CC, but if the the
BUILD_CC does not support SSP and the cross compiler does, the build fails
due to the SSP flags set in the Makefile. Ensuring that SSP is not enabled
when building these tools prevents this from happening.

(From OE-Core rev: ad9a9c6e6b4f00bc47449bbe8c85777ea7c89b88)

Signed-off-by: Tomas Frydrych <tomas@sleepfive.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

neard: re-order recipe so the statement order is logical

(From OE-Core rev: 373dc2778f263baa7958c48cc1644008e2a40676)

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

menu-cache: Fix segmentation fault

[YOCTO #4353]

(From OE-Core rev: 8c9c6155e6d83675a94b4eaae4ffc6dfeca399ee)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libpcap: Fix up patches

These patches where tweaking both configure.in and configure, which is
bad since configure changes after an autoconf. If the patch is re-applied
for some reason it would notice that it's already appiled and fails.

[YOCTO #4314]

(From OE-Core rev: c32b425c10566cf2aaf187e98b6e2e29022bea2f)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

alsa-utils: Move alsaconf to its own recipe

18575b082a4042376fd1575465e69562dea04ddc added bash as a dependency of
alsa-utils-alsaconf so that the script interpreter will be available at
run time.  However, this has the undesirable side effect of making bash
be a build dependency for alsa-utils and, for those folks who don't need
alsaconf but do want some other part of alsa-utils, this cure is worse
than the original disease.

Fix this by moving alsaconf to a separate recipe so that the bash
dependency only applies when alsaconf is specifically requested.

(From OE-Core rev: 7317c8055cf3af8912a66badb3074f0a60f75ec2)

Signed-off-by: Phil Blundell <philb@gnu.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman: add PACKAGECONFIG options

- change TIST from being explicitly built, to a PACKAGECONFIG
- move wifi, 3g and bluetooth to PACKAGECONFIG
- change RDEPENDS and RPROVIDES to check PACKAGECONFIG rather
  than DISTRO_FEATURES

(From OE-Core rev: da3f9f7febdc485a356ccb102e0d93cd8f059da7)

Signed-off-by: Jack Mitchell <jack.mitchell@dbbroadcast.co.uk>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

matchbox-panel-2: bump srcrev to fix compile warnings

The older versions of matchbox-panel were naively using -Werror which causes
warnings with gcc 4.6 (which were patched away) and again more with gcc 4.8.
I'd already fixed this upstream so bump the srvrev and drop the patch.

(From OE-Core rev: ae38cac225f0d84e6acecb6bc46df939d37cd031)

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

Pass the mlprefix to postinst_intercept script

This is needed in order to have separate multilib intercept hooks.

(From OE-Core rev: 44eae13c164ca6e15a6013eafeab3eb44f24c8e0)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

scripts/postinst-intercepts: create separete hooks for multilib

When using multilib, the hooks for lib32/lib64 must be different because
the libdir/base_libdir point to different locations. Postinstalls
calling postint_intercept script must pass the mlprefix in the 3rd
argument.

(From OE-Core rev: 2c5c6e3ffcd561c25a34603922b622449f677a34)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Revert "qemu.bbclass: Use the correct qemu binary in multilib cases"

This reverts commit 9f5a6f89d9f4a6c7bed3b163e6eaa764d762f523.

The reason for reverting this is:
 * qemuwrapper has now a fallback method;
 * when using multilib, calling qemu_target_binary from recipes would
   always point to the qemu binary corresponding to the machine
   architecture. Hence, postinstalls needing to use qemu would call the
   wrong qemu user emulation binary;

(From OE-Core rev: 15408466515cec7cbb4c394aa203c87b6165f884)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pango: fix postinstall when using multilib

The pango-query-modules binary gets a multilib prefix and the
postinstall has to call the appropriate binary.

(From OE-Core rev: 21ae18ca5e3be0b3e5cb0fdcf19b1476dbd38b0c)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qemuwrapper: use fallback in case the ELF binary is wrong

This wrapper script is called mainly from intercept hooks and allarch
packages postinstalls. When multilib is used, the qemuwrapper script
points to the binary that matches the MACHINE architecture.

For example: if MACHINE=qemux86_64 and we activate multilib, then the
postinstalls for lib32 packages would call qemu-x86_64 with 32 bit
binaries and they would certainly fail.

This patch adds just a fallback method if the exit code of the previous
qemu call corresponds to "Invalid ELF image for this architecture"
error. This will allow us to have all postinstalls run on host.

(From OE-Core rev: 0c6ddb84043f0f917543cdaf4814efc15cd0273f)

Signed-off-by: Laurentiu Palcu <laurentiu.palcu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

build-appliance: Add version info

(From OE-Core rev: 1acc6c60c39a6c8982b228ac8dbc85acc89a0032)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-networking: Update to 2.36

COPYING file: Address and Formating changes only

(From OE-Core rev: 827c4a6e337f6701945bf12a4a93f98011103597)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: Update to 2.36

Remove automake patch that is now correctly supported upstream
Codegen files have moved to glib-2.0, so correct packaging and
removal.

(From OE-Core rev: 0d13d9947262b09cd69bc526ea2738e50c658744)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mesa-demos: Update to 8.1.0

Removed patches that now appear in the upstream.

(From OE-Core rev: 2329297b12e2eade895fff8d3d98722a15e0b7ec)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glew: fix generation of glew.pc needed by mesa-demos

(From OE-Core rev: 0273f31d20f3e3c24e8b309f928595885eb04af2)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libxkbcommon: Update to 0.3.0

(From OE-Core rev: 28e185ba401fa1f2c11fca65bd873a0f5d78b811)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pkgconfig: Move patches in prep for 0.28 update

(From OE-Core rev: 076e1c9895cc4f559fd58956a6eb1098918292b5)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libgcrypt: Update to 1.5.1

Remove automake patch as it is now supported by upstream

(From OE-Core rev: bf4c807aa8479ba475c7484dabe6cdb0def5f4ec)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nspr: Update to 4.9.6

(From OE-Core rev: 16ddc0138f385ac124a843bb1d1256536c40d0f6)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libffi: Update to 3.0.13

New patch is required to fix the location of the libdir defined i
the libffi.la file which caused a bad RPATH (/usr/lib/../lib) due
to the gcc -print-multi-os-directory returning ../lib.

Remove Aarch64 patches as it is now supported upsteam. Other patch
code is also upstreamed

(From OE-Core rev: 6f1caa75b181eb40fdbbd4d9979a5f61c0b9435a)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

file: Update to 5.14

(From OE-Core rev: 13b6bedab766657270a12c7c332914c69ebae085)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dpkg: Update to 1.16.10

(From OE-Core rev: 47e37874372b405e64ed42168b66ef3dccefdb08)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

resolvconf: Update to 1.71

(From OE-Core rev: 2f0f7c34f805b4a5e89ec138107bcc434d06f8d0)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dhcp: Update to 4.2.5-P1

(From OE-Core rev: 0358a0bd39b8a40984bc15312d030f8cfd60805f)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sqlite: Update to 3071602

(From OE-Core rev: d0e0d1322a8b4bc6d1557555bd2a0bb9c5fa37d7)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sysstat: Update to 10.1.5

(From OE-Core rev: 7e59a2ccfcb5cab3529068f90fea19fda28c7261)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lsbinitscripts: Update to 9.46

(From OE-Core rev: 047c55735347051b858dd6e471493bd0780840b9)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

less: Update to 458

(From OE-Core rev: dd4f419413cf6ae1f8a069ae8a8cceae6ce02453)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>