utils: Optimise looping in base_set_filespath

Calling split on the same expression, once per loop iteration is
inefficent and pointless, particularly in a function called by
every recipe during parsing.

(From OE-Core rev: 566c0e874fc1610f3f97737b5601ef22026c918a)

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

libffi: Use a more standard layout for patch files

(From OE-Core rev: cc2d2abad25a82ce722ac255a631430058f8de30)

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

trace-cmd: Simplify SRC_URI patch entry

(From OE-Core rev: dd4cf48b5a5d181d507356a845fecdbb26f37390)

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

matchbox-panel-2: Use a more standard layout for patch files

(From OE-Core rev: 7e78c5901be4d684669146b2f9ab845507156034)

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

bitbake.conf: Drop obsolete FILESDIR setting

FILESPATH is the preferred way of finding files now. Having a value
for FILESDIR which defaults to paths which will have already been
searched is pointless at best. This is the final step in letting
us drop FILESDIR support entirely from bitbake at some future date.

(From OE-Core rev: d6e5ceafcaef06b8a3f9acc2aa826a40a016f913)

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

x-load: Replace FILESDIR with FILESPATH

(From OE-Core rev: 0f7ca16c01f6206cd03d103861b77ce2b540c834)

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

opkg: Drop unnecessary FILESDIR setting

(From OE-Core rev: e869731f80d8c0059b44a7029543b8943fd07653)

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

qemu_git: Drop unnecessary FILESDIR setting

(From OE-Core rev: 8afa069778aea414856a9f538171d9ab1a2e1dd8)

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

gcc: Use FILESPATH instead of FILESDIR and cleanup/simplify

(From OE-Core rev: 4cb359182da00e661fda11a8b31e3611b0df03cb)

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

build-appliance-image: Simplify fetch/unpack execution

This provides a slightly neater way of ensuring fetch/unpack get
executed (image.bbclass marks them as noexec) since I found the
current approach harder to understand at first glance.

(From OE-Core rev: 84021fd694d0a7bb1e4f49c0f7c46a0fbd178924)

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>

opkg: drop PACKAGE_ARCH setting for update-alternatives-cworth

* there is warning about update-alternatives-cworth ipk being
  overwritten in "all" feed when opkg is built for more architectures
* see https://bugzilla.yoctoproject.org/show_bug.cgi?id=3390 for details

(From OE-Core rev: 969d10d0aa5537a119ed9be9e8321dbd73d160fe)

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>

distutils-common-base: Create staticdev pacakge for static libraries

(From OE-Core rev: 0c4c3ebfedd06b2b8f36747bc0c0d986f05420cf)

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

gnutls: Update to 2.12.21

(From OE-Core rev: d55c435575410e1865748ba1914443c94ef9de27)

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

libksba: Update to 1.3.0

Licensing files changes from 1 COPYING to be split.
The Code is marked LGPLv3 or GPLv2 or both in parallel

(From OE-Core rev: 96a29f346f89641eaccdb740e4d21cc4f732768d)

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

dhcp: Update to 4.2.4-P2

(From OE-Core rev: f3ca792a470437cb7fc91ca10f220a48da23d6f5)

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

libxml2: Update to 2.9.0

(From OE-Core rev: a65ea43e26866ddc60051c52b2eb226507fae346)

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

resolvconf: Update to 1.68

(From OE-Core rev: 5255d0ba318f1a24b3ea3947304f4b798a1562be)

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

mc: Update to 4.8.6

(From OE-Core rev: 27bb231a0a80ee39265ad54047d8dfb0feeeb2bc)

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

man-pages: Update to 3.44

(From OE-Core rev: cdf782fab956f5a3b43b3eb256fa57748175b802)

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

tiff: Update to 4.0.3

(From OE-Core rev: 90ad57fbd72edf44336d0ad2c2e3ec861a641fb3)

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

rt-tests/hwlatdetect: Update to 0.85

(From OE-Core rev: 42fc047daa524ce4f1b9b1b2937d75375d796e6f)

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

libdumpvalue-perl: Update to 1.17

LICENSE file updated Copyright year to 2012, no other change

(From OE-Core rev: 41f3104d1993f374e9e289ac74f821261808f833)

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

libproxy: set LIBEXEC_INSTALL_DIR

This change fixes the following QA Warning

WARNING: QA Issue: libproxy: Files/directories were installed but not shipped
  /usr/libexec
  /usr/libexec/pxgsettings
  /usr/libexec/.debug
  /usr/libexec/.debug/pxgsettings

(From OE-Core rev: b11874ef0b80f8d5150681fc2d6a0ca7a2c8fa06)

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

libtasn1: Update to 2.14

(From OE-Core rev: 74200da342a8888230a09e063bea396a8f611400)

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

multilib.bbclass: Drop populate_sdk_base exclusion

With this recently introduced exclusion, <multilib>-meta-toolchain-sdk
throws errors about missing DEPENDS that don't exist since it needs the
PROVIDES/DEPENDS remapping. This patch tweaks the class tests to fix
the errors.

(From OE-Core rev: 9cc18fe12bd8d1c73df291b4057aab6167ef6b16)

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

libcheck: Add missing rdepends on gawk and set path correctly

The new version of libcheck needs gawk and we need to ensure the paths are
not taken from the build system.

(From OE-Core rev: 120bea8043d3a05174ed034e20d9094784402824)

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

bitbake: server/process.py: Change timeout error handling

In normal usage, we never hit the timeout issue. If we do, it becomes obvious
that the current error handling is not good enough. The request may have made it
to the server and the answer will get queued. This means the next command may get
the return value from the previous command with suitably puzzling results.

Without rewriting large sections of code, its not possible to avoid this problem.
It is better to increase the timeout to several seconds giving the server a chance
to respond and if it does timeout, hard exit since recovery is not possible with the
code base today.

I'd be happy to see the structure of this code improved but this quick fix at least
stops corrupted builds from happening which has to be a good thing.

(Bitbake rev: 410c11dd10736873f2dc587fbe9119c38831e693)

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

package.bbclass: Adapt debugsrc code to workdir layout changes

With the change to WORKDIR layout, splutting PN and PV into two directories,
the debugsrc splutting code layout became suboptimal. This changes things to
include the information as it was before. Ideally this code would be written
to more generically support other layouts buts it not clear that the tools
would even support that right now so this is the best immediate fix.

(From OE-Core rev: 432cfbb403f0e864d1fad383c2bbb6f9bdb80770)

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

mesa-dri: add correct RPROVIDES for the -dev packages

The -dev packages were RPROVIDEing the non-dev names, which is clearly wrong.

(From OE-Core rev: 8d224244a016adc889be132d9994d7c517f7eae3)

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>

Revert "gcc: Use FILESPATH instead of FILESDIR and cleanup/simplify"

This reverts commit 90616875b432a932415063b08497266e70c49d75. It was
never meant to be applied as its an incorrect previous development
verison of a patch in progress.

ia32-base: fix typo in XSERVER_IA32_EXT definition

(From OE-Core rev: 2948fcb2e85d71e3e686760f12e1082dc912f418)

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

yocto-bsp: upgdate configs after xf86-video-omapfb changed name

(From meta-yocto rev: 992c2c469861bfcab45a118ac9ab8887e81f1a11)

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

yocto-bsp: update configs after xorg upgrade

There were some changes in the xserver-xorg upstream project that need
to be reflected here too:
 * extmod module was removed completely as it became empty;
 * DRI1, DRI2, DBE (among others) were made built-in;

(From meta-yocto rev: ed681441a2cf06dc55e71035ecbfc637ff83640d)

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

bitbake: knotty: Colorize knotty interactive console output

Add bold color output to log level name and standard color output to log msg
when bitbake is run from an iteractive console.  Color output is only
enabled
if the terminal supports color.

Used Jason Wessel's recommendation for transparency on verbose, note and
plain.

(Bitbake rev: 2734240da2cc150f811129a6adf6eb4b2161b204)

Signed-off-by: Seth Bollinger <seth.boll@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: hob: warnings during the build should be displayed in the Issues tab

Any issues encountered during the build (fatal or not) is displayed
in the Issues tab, and the total number of issues is changed.

[YOCTO #3376]
(Bitbake rev: 64c662ab7f09c2e867445e8ba21ea63ae014d45b)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: runqueue: Allow partial setscene task coverage

When the setscene code was originally written it was thought that we'd
allow "partial" coverage. For example, if we just want to build the target
"bash:do_populate_sysroot" and its available from sstate, it makes no sense
to install gcc-cross's sstate package as its simply not needed.

Due to various other issues in the codebase, this functionality was
disabled/removed to allow the setscene code and sstate to stabilise and allow
us to concentrate on other problems.

The time has now come to enable "partial" coverage. There are two major changes
in this patch:

a) Creation of an unskippable list. This lists direct dependencies of
   build targets and hence things that cannot be skipped.

b) Addition of a handler which looks at a given setscene target and what depends
   on it and then decides whether its necessary.

(Bitbake rev: 2a937cd6a6c3110030b40bc4d85e349b85cb4db7)

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

bitbake: parse/cache/cooker: Preserve order in the file inclusion list

The data returned by get_file_depends() may me used in contexts like
checksums where order is important. The current usage of sets means
that some of the checksums can change in circumstances they should not.

This patch changes to use lists, thereby removing the problem.

(Bitbake rev: a44285fc4109236ab89f7aad0a1fc9220eec19b6)

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

libcheck: Update to 0.9.9

(From OE-Core rev: 0d856639e8c4fd31c7af222cdb2f37d6a9bec1db)

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

sqlite3: Update to 3.7.14.1

[YOCTO #3283]

(From OE-Core rev: 141ec5567de19d740e147786e25e17dd10e68001)

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

builder: Add password for user

This is needed to allow openssh to work correctly for the Eclipse
plugin to have access to the build appliance to view/modify recipes
and lauch builds

Default password is "builder"

(From OE-Core rev: ccf86771bf65e9620385abf20049f355dca391df)

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

sysstat: Update to 10.1.2

(From OE-Core rev: 6e1a032c6a2642399c5d32028c597a5affbedce1)

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

mesa: default to enabling EGL and GLES

Even on systems where Mesa has no hardware support, building the software
renderers is useful for build testing and limited functionality.

(From OE-Core rev: e79987bc4bac1d739f92790f8e9840cd02f073d3)

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

mesa: add -mesa suffix to GL packages, RPROVIDE their generic names

When Debian-renaming, all packages that provide GL libraries get renamed to the
same name, and it's entirley possible for a feed to have multiple GL libraries
in.  This obviously creates conflicts.

Resolve this for Mesa by forcing the package names to be of the form libgl-mesa,
and RPROVIDE libgl.

(From OE-Core rev: 64c77bf395310e55b4d8e0ec754fa19e9034ab35)

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

xf86-video-omap: add new recipe to follow the maintained repo

This new recipe is needed because the old driver is unmaintained. This
new recipe will follow the new repo.

(From OE-Core rev: a1d93e6383396dc3ff7cd3a4fccf27895e80af8a)

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

xf86-input-synaptics: add mtdev dependency

After upgrading xserver to 1.13, multitouch support is automatically
enabled in xf86-input-synaptics. Hence, the need for mtdev dependency.

(From OE-Core rev: 6b2ffa1637d9ae067753102efeb78d1eb42a0b8a)

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

xf86-video-vmware: Add compat API

Needed after upgrading xserver to 1.13.

(From OE-Core rev: 6eebd1699f9efba834e51a0772eab253184af914)

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

xserver-xorg: upgrade to 1.13.0

The patch contains several aditional changes:
 * removed one backported patch (included in the new release);
 * changed mips64-compiler.patch to apply properly;
 * licence checksum for COPYING file changed: some copyright years have
   been changed;
 * bump PR in xorg-driver-common.inc so that all input/video drivers
   get rebuilt. That's becaue the ABI changed;

The following external modules are now built-in:
 * DBE
 * DRI2
 * DRI
 * RECORD

The extmod module was completely removed.

(From OE-Core rev: 506da0d139dd470475a1d6b2dd3ae62406c36816)

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

libdrm: upgrade to 2.4.40

Other changes:
 * removed a backported patch;
 * activated libdrm-omap helper layer which is needed by the latest
   xf86-video-omap xorg driver;
 * split libdrm-drivers package into libdrm-radeon, libdrm-nouveau and
   libdrm-omap, libdrm-intel and libdrm-exynos;

(From OE-Core rev: 8b100befe8dcf7523148b6fc14fa2237d07fe556)

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

xkeyboard-config: upgrade to 2.7

A few extra changes:
 * changed the SRC_URI to the new, valid, one
 * added dependency of gettext (do_qa_configure detected is needed)
 * disable runtime dependency checks at configure time

(From OE-Core rev: c67b5e212244f1bac57e8491c6500656786df3a2)

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

guilt: change upstream tgz location

The kernel.org mirror of the guilt tarball has been missing for a while
and the yocto mirrors have been keeping builds working. Switching to a
debian upstream is better than solely relying on the yocto mirrors for
serving the tgz.

(From OE-Core rev: 71f281f40e25bdd3ea052cb673d06c1a250e618f)

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

kern-tools: flexibility and usability enhancements

Updating the SRCREV to import the following changes.

 [updateme: find the board description with the highest score]

   This removes the requirement that a custom linux-yocto .scc file have
   define KTYPE <foo>, where <foo> is typically "standard". The tools can
   now match on a .scc file that only matches the board, but will still
   chose one that matches the board and kernel type, if available.

 [updateme: allow for tabs or spaces in defines]

   define KMACHINE<tab>$MACHINE was missed by the regex.

 [scc/kgit-meta: detect and avoid duplicating patching]

   To allow feature description to be included multiple times, they were
   previously split into -enable and 'patch' descriptions. With this change
   the patches will be detected as already included, and skipped
   automatically. Removing the need to do this split. It also cleans up
   the ability to warn about multiple includes.

 [kconf_check: add "verify" configuration fragment type]

   This adds the ability for a BSP to have a kernel configuration
   fragment that lists options that must be present. If they are not
   present it is a hard error. "required" is a similar fragment, but
   it adds them to the build, and audits them at the end, but does
   not abort the build if they are present. This is a minor distinction,
   but one that is useful when creating flexible, shared kernel config
   structures.

 [kconf_check: improve kernel audit report formatting]
 [kconf_check: perform validity checks on non-hardware options]
 [kconf_check: cleanups and verbose flag]

   The existing output was verbose and not always useful to the reader.
   This change makes the output more compact, audits non-hardware options
   and gives information

     [invalid (54)]: meta/cfg/preempt-rt/common-pc/invalid.cfg
        This BSP sets config options that are not offered anywhere within this kernel

(From OE-Core rev: 2d328dc0f7dd763c45444394b681d2726b4f6c83)

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

kern-tools: kconf_check: fix find warning

Bumping the kern-tools SRCREV to pickup the following change:

[
    kconf_check: fix find warning

    When searching for all available Kconfig files, kconf_check was using
    $meta_dir instead of $META_DIR. This resulted in a truncated path and
    the following warning:

      find: warning: -path $oe-path/linux/ will not match anything because it ends with /.

    Using the proper variable removes the warning and make sure that we
    do actually search all relevant directories.

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
]

[YOCTO #3226]

(From OE-Core rev: 5999ccebc7b071737f82709467e2a2ec152240f6)

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

linux-yocto/3.2: update to v3.2.32 and 3.2.32-rt48

The 3.2 kernel was lagging behind on kernel.org -stable and -rt
updates. Even though no 1.3 BSPs directly use this kernel, it should be
updated for those that may use it.

Sanity test on qemu* for -rt and standard builds.

(From OE-Core rev: 7ad1c853e252bea024043dc79d89405178393c09)

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

linux-yocto/3.4: v3.4.17, v3.4.18, -rt and config changes

Bumping the linux-yocto/3.4 SRCREVs to incorporate the following updates:

  - v3.4.17
  - v3.4.18
  - 3.4.18-rt29

Also incorporating the following meta branch config changes:

  5bd6d0d rangeley: update include to use the new intel-dpdk feature
  4b277c2 dpdk: Add feature Intel DPDK
  3905e74 meta: rangeley: Enable Zlib Compression
  194c5f1 meta: Add a new feature for Zlib
  14cb04d meta: rangeley: Enable AES feature
  8e4dbf6 meta: Add new feature for Ciphers
  7e75c1f enable IPv6 Router Preference (RFC 4191) support
  dfd56d1 Create IPv4 and IPv6 IPSec fragments
  0a85061 rangeley: Add smp support
  1190856 rangeley: Add efi support
  b262e38 rangeley: Add PCI features
  80c9084 rangeley: Add uio and hugetlb support

(From OE-Core rev: 7cc39567cc91955eb3014da6fdbafffa5c3148c7)

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

linux-yocto/3.4: bump kver to v3.4.16

The -stable 3.4 kernel has updated versions, so we import 3.4.16 and
make that our new baseline.

(From OE-Core rev: c476046368ed87a400b3a2fd4344fc48aacc0dbc)

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

linux-yocto/3.4: efi/mmc fixes and fri2 updates

Updating the 3.4 SRCREVs to pick up the following two fixes:

 218bd8d efi: Add patch to fix 32bit EFI service mapping (rhbz 726701)
 b6d08f7 mmc: sdhci: Use DBG() instead of pr_warning() on large timeout

And the following meta branch config updates:

 68a635b fri2: Disable GPIO_PCH for preempt-rt
 2ec32d5 fri2: Add fri2-tiny support
 a7b9607 fri2: Required boot config for fri2
 bed2080 fri2: Remove graphics options from the core fri2 description

(From OE-Core rev: dbd49c9157f933fec9147280a48ce3cda7a697eb)

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

linux-yocto/3.4: nfsd, pci, fishriver and rangely config changes

Updating the 3.4 meta branch with the following configuration changes
and additions:

0541ba5 meta: Rangeley Machine Created
9e3bdb7 meta: Add nfsd kernel features
da9b37d CrystalForest: Enable PCI extended config space for CrystalForest Machine.
628cbe9 meta: Add a new feature for PCI devices.
9c3a2b3 meta: fishriver: remove meta-data

(From OE-Core rev: c11bf4359697f654ff38a32bda5eae71b097d3b8)

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

license: We need to run this task before do_build, there is no dependency on do_package

This change means we have more flexibility about when to schedule the license
task and if it changes, we don't repackage everything (which is pointless).

(From OE-Core rev: ee1293446936c5444ece42b60e3ab94189b2fbc3)

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

bitbake.conf/sanity: Separate versions and PN stamp components into separate directories for WORKDIR and STAMP

This means some of the hacks we have to tell where the package name ends and
the version starts in the directory layout becomes obsolete, simplifying the
work of some of the cleanup scripts. It also makes the layout slightly more
intuitive to the user.

It does force a rebuild onto the user but it will reuse sstate successfully.

(From OE-Core rev: 05075cf3138d1c61f5cf4fe0e1a4587acc00c692)

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

image.bbclass: Add missing dependency on do_package data

Since the packaging functions now reference the pkgdata files written out during
do_package, we need to reference this dependency explicitly.

(From OE-Core rev: 1e9c9d164f8d12c8de205e04bf7c1dae3660f12a)

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

sstate: Implement a setscene dependency validation routine to allow skipping of some sstate installation

This is a first attempt at logic to determine when a sstate dependency needs
to be installed and when it does not. Its a start at the logic and errs on the
side of caution, as it gets wider testing, we can refine the logic as needed.

This code should allow a significant performance speedup to certain workflows, for
example "bitbake xxx-image -c rootfs" will not populate the target sysroot.

(From OE-Core rev: b43faba37816817edc5240a139361d16e07c6131)

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

scripts/pybootchart: Allow minimum task length to be configured from the commandline

Rather than hardcode the value of "8", allow the minimum task length to be
configured from the commandline using the -m option. "-m 0" means all
tasks will be graphed.

(From OE-Core rev: 30001153d3ce7dadf8f1ec79e634a638a9994518)

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

scripts/pybootchart: Fix missing entries bug

If two entries have the same start time, the data store used will cause
all but one of the entries to be lost. This patch enhances the data
storage structure to avoid this problem and allow more than one
event to start at the same time.

(From OE-Core rev: 220b071fd8d1cc6bdbca58f75489e3c9b34921ca)

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

flac: fix text relocations

The recent sanity checks were flagging:

 ELF binary '.../libFLAC.so.8.2.0' has relocations in .text

This is caused by hand-written assembler being invoked badly.  Apply a patch
from upstream git that uses PIC instead of relocations.

[ YOCTO: #3461 ]

(From OE-Core rev: 9b5660ee0e507852a02ba5281b571f3e55dffc18)

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

ncurses: update gnu-config files in do_configure()

(From OE-Core rev: 52d4c2cb6cd15f8ebaacc92ddf71274bf7a421d5)

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

kmod: fix git repo URL

(From OE-Core rev: 156e0fca979585f72323041f8d8aeafcbd43dfc3)

Signed-off-by: Eric Bénard <eric@eukrea.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

qt4: remove 4.8.1

(From OE-Core rev: 5f6d7d61d79215ffe38aa6b122ae5ac0af7e859a)

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

qt4: remove negative preference on 4.8.3

(From OE-Core rev: 36e42fa771ddd11e169d92dd31d213ba84538012)

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

gcc: Use FILESPATH instead of FILESDIR and cleanup/simplify

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

yocto-bsp: update qemu machine.conf template

Update the template with the changes from the last commits to meta/conf/machine/qemux86.conf
Building sato image for a machine created using yocto-bsp with qemux86/x86-64 templates fails because nothing RPROVIDES qemugl anymore so remove support from the template as well.
Also drop redundant glibc configure knobs (they are no longer optional and they don't exist in conf/machine/qemux86.conf anymore so for consistency the template shouldn't keep them).

(From meta-yocto rev: 644c201a8fb9e589cdda1f76385a0b41549ea057)

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

pulseaudio: explicitly disable xen, rather than letting it detect

(From OE-Core rev: d109b24ad354382cd40b28e86211e53929a0910f)

Signed-off-by: Gilbert Coville <gilbert_coville@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

opkg: added alternatives-ln patch

Use 'ln -n' to avoid dereferencing links to host files.

(From OE-Core rev: e5aef500e11cbf7d1cd20b588fcea2c5fd6b5d0e)

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

binutils: fixed --enable-targets option

There does not exist an '--enable-target=all' option

(From OE-Core rev: 60fe4e80ca5845a0d03f918b80d6e980c13378b9)

Signed-off-by: Enrico Scholz <enrico.scholz@sigma-chemnitz.de>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cmake.bbclass: use DEPENDS_prepend instead of += for cmake-native

Otherwise when a recipe using DEPENDS=, the cmake-native dependency disappears.

(From OE-Core rev: 2b35539d96325d8e687451543d4f52f1a07bf1c6)

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

default-providers: add default provider for make

remake PROVIDES make, so we need a default provider.

(From OE-Core rev: 9af884d433d18582b14977eb340cfdfa4801e7fe)

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

libcgroup/libxkbcommon: Use BPN in SRC_URI

If we don't do this, multilib and other varients using BBCLASSEXTEND
will fail.

(From OE-Core rev: 9a97367038a1e2431bf94211dabbc5aedbbee3bb)

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

Fix the first line typo of adt-installer

[YOCTO #3384]

(From OE-Core rev: 039e119590b2f3e1d912b446fa68b6cf936d21c2)

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

less: Update to 451

LICENSE file was changed to match the BSD-2 Clause

(From OE-Core rev: 4b6a70e60790a32d89e2e5cdded4af83e9d303ae)

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

insane.bbclass: add qa package name check

Check if package names match the [a-z0-9.+-]+ regular
expression

[YOCTO #3139]

(From OE-Core rev: 55dd271be1aee21e36d130359f4f21841623c425)

Signed-off-by: Constantin Musca <constantinx.musca@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate: Drop now unneeded python whitelist entries

(From OE-Core rev: 2a9a3e5e3e9229eb11f20eeabef7929014bccd11)

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

python: Resolve intermediate staging issues

Its bad practise to poke into the sysroot without knowledge of sstate.

This adds a patch to python allowing us to account for cross compiling
and allow it to find the Makefile/pyconfig.h files without needing them
in the sysroot for do_compile/do_install to complete.

Tested on two architectures and compared with buildhistory with no
significant delta.

(From OE-Core rev: 16da4f75a75dc8020803df9ea73a2a7ead88cc5a)

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

packagegroup-self-hosted: add sftp server

Which is needed for integration with Eclispe plugin

(From OE-Core rev: 57127ff6f42145bb1a200c8c3267158df637fbfd)

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

psmisc: Update to 22.20

(From OE-Core rev: e4fc11305e2e09b2883cb455e0772a01e9f6dd4a)

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

kconfig-frontends: Update to 3.6

(From OE-Core rev: 3fb19c1044b46ee7b0d898af4fc6f46bbd957b2f)

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

mx: Update to 1.4.7

Source moved to GitHub

(From OE-Core rev: 1f2f35ba7607b503af2cdf51a459c6de786f554d)

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

pulse: fix Bashism in string test

(From OE-Core rev: 342daf26eaf0d885278b06b8d820db238cbf4d61)

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

make-3.82: Add patch for archive expression expansion issues

bitbake make-native; MACHINE=qemuarm bitbake icu

would fail with:

*** No rule to make target `uconvmsg/libuconvmsg.a(uconvmsg/uconvmsg_dat.ao', needed by `uconvmsg/libuconvmsg.a'.  Stop

which is caused by a bug in make 3.82 which the attached patch fixes.

(From OE-Core rev: 06e64233a3a00a3c60fab7d92cbb18cd9feadc8d)

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

xf86-input-synaptics: add mtdev dependency

After upgrading xserver to 1.13, multitouch support is automatically
enabled in xf86-input-synaptics. Hence, the need for mtdev dependency.

(From OE-Core rev: 03d787efe0d83b20155508811f901b05a910940c)

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

mdadm: upgrade to 3.2.6

(From OE-Core rev: c7aacc09c4d3d68bdd6fa7419a7ea1c2b2a007ae)

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

fontconfig: upgrade to 2.10.1

A couple of changes:
 * licence snippet in fccache.c moved down the file;
 * new files appeared in this version, added them to fontconfig package

(From OE-Core rev: f6ca099d9cbd2ed1c181e8e91cc16d0550701f26)

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

xcb-proto: upgrade to 1.8

(From OE-Core rev: dd378e8bd3419219585968158be1972bffa57247)

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

xf86-input-mouse: upgrade to 1.8.1

Licence chacksum changed because Red Hat added a copyright line.

(From OE-Core rev: 00c892acc24860c26fe658ff96cfa002a1c96410)

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

xf86-video-intel: upgrade to 2.20.12

(From OE-Core rev: 31ade27a9d6011ff3a96aa2ef81be662f83c640f)

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

xf86-video-vesa: upgrade to 2.3.2

(From OE-Core rev: efdf6fc1cdbd5427eb94b034b2adc6a45a351b75)

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

python-smartpm: Add basic knowledge of RPMSENSE_MISSINGOK

Currently smart does not support recommend dependencies.  Add the first set
of 'support' for RPMSENSE_MISSINGOK (the flag that makes something a
recommend).  This initial support ends up ignoring the recommendation, but is
written in a way that it will be the basis of eventual support.

(From OE-Core rev: 4f1f6d39803c94cf9ff55f0a4616e7a1703bcef6)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpm: Add additional RPMSENSE values to python module

We add a number of additional RPMSENSE values to the python module to better
support the dependency calculations in SMART.

(From OE-Core rev: 431352d063b353ee0e0eaa5bfe24450962d71d6b)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

python-smartpm: Add smartpm recipe

This is the initial integration, basic functionality such as 'smart query'
has been tested.  Active use of remote feeds and such has not yet been
verified.

Thanks to Paul Eggleton for corrections and bug fixes for the initial
integration.

(From OE-Core rev: 92182ca88aff9cec04b2af5e9babaf33bf61f0af)

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

rpm: Slightly change the way python-rpm is constructed

If python support is enabled we want to make sure that the RPM python support
is packaged properly.

Move the components into the site-packages directory, move the .la files to a
new -dev package.

Add "rpm" as a dependency of python-rpm, otherwise rpm and the associated
libraries won't be available.

Fixup python wrapper to handle automatic relocation, as supported by the
vendor WINDRIVER configuration.  (Based on a patch from Paul Eggleton)

(From OE-Core rev: cd0473a145cec51be736b6141b0b18a82b64d483)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate: Bump version number to deal with layout fixes

The recent preveeding sstate directory layout fixes made the code do what it
was originally intended to do, as can be clearly seen from the code.
Unfortunately this changed the contents and layout of the sstate files
themselves since the bug was leading to a directory prefix being missing.

This is now resulting in chaotic messages on the console since things
are getting confused with the two different layouts. The simplest way to
resolve this is to bump the version number, hence moving the new layout
into its own new namespace.

Its worth noting that whilst the failure messages are scary, the failure
mode is relatively harmless since it will just fall back to building the
data rather than installing from sstate.

Usually I'd give more notice of a change like this but under the
circumstances, I'm just going to push this in to resolve the failures
people are seeing. Initially I thought the problem was limited to
some of the -cross packages and therefore of low impact but that is
clearly not the case.

(From OE-Core rev: b53ea6687b6201c8c5ab5cb0d2a845ef7e7b2abe)

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

bash: fix mkbuiltins build failure

On hosts with FORTIFY_SOURCES, stringize support is required, as it's used by
the macros to wrap functions (e.g. read and open in unistd.h). Those wrappers
use the STRING() macro from unistd.h. A header in the bash sources overrides
the unistd.h macro to 'x' when HAVE_STRINGIZE is not defined, causing the
wrappers to generate calls to 'xread' and 'xopen', which do not exist,
resulting in a failure to link.

Assume we have stringize support when cross-compiling, which works around the
issue.

It may be best for upstream to either give up on supporting compilers without
stringize support, or to not define STRING() at all when FORTIFY_SOURCES is
defined, letting the unistd.h one be used, instead.

(From OE-Core rev: f7a25dd72d1d463eb72d48c6f9dd968d376496c0)

Signed-off-by: Christopher Larson <chris_larson@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>