dpkg: fix configuration issue for mips64

CQID: 431771

configure cannot determine the proper cpu, os, or
architecture for mips64, and possibly other arch's,
because of faulty code added to Arch.pm in the latest
release from upstream.  We remove that code.

(From OE-Core rev: e56b4af3a325d6d5332c779e6253da9b3cd2fce0)

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

systemd: Upgrade to 206

Add new PACKAGE systemd-rpm-macros, this will hold
the macros which are interesting when rpm is used as
package management backend

Forward port uclibc only patches. Add a new patch
to stub out use of preadv/pwritev in testcases

Delete patches that have been merged upstream in systemd

Remove force export of GPERF variable in environment
this was causing AC_CHECK_TOOL to not populate GPERF
variable as expected

systemd needs kmod to be present on rootfs so add it
to RDEPENDS

some services substitute discovered kmod when the service
file is generated during boot, however the discovered kmod
is from native sysroot and it gets into the service file
with absolute path. So specify the target path of kmod using
KMOD variable so the unit files have correct pointer to kmod
on target

Add a patch to make sure that mknod capability is checked
before the service which excercise mknod, this patch is also
submitted to upstream systemd

(From OE-Core rev: 3f1788f8edf18a292cb5d8e16a2a98a19ec89239)

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>

kmod: Upgrade to version 14

The update is a requirement for systemd-206

(From OE-Core rev: faacfeb7cb7bed9ad5eb387d358309a2b1d40f06)

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>

bitbake: data_smart: allow removal of multiple words at once with _remove

    FOO = "foo bar baz"
    FOO_remove = "foo baz"

(Bitbake rev: 04127dec207d6dfc0ada56c5cc67ec9ad30517a8)

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

bitbake: data_smart: use a split/filter/rejoin for _remove

This is more idiomatic, and from the limited performance testing I did, is
faster as well. See https://gist.github.com/kergoth/6360248 for the naive
benchmark.

(Bitbake rev: 1aa49226d5a2bac911feeb90e3d9f19529bc1a3e)

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

bitbake: serv.py: Fix hang when spawned dynamically with bitbake

The PRServer has the possibility to hang indefinitely blocking on a
semaphore processing a xmlrpc request to send an event back to the
main bitbake instance.  This was observed during a "bitbake -e" on a
heavily loaded machine and the main bitbake instance and cooker exited
before the PRServer emitted its first log.

The stack trace is provided below as to show what happens every time a
logger.info() is executed in the PRServer.  Not only does it write to
the stream handler but it also tries to send the event to the main
event processor.

    self._notempty.acquire()
    self.queue.put(event)
    _ui_handlers[h].event.send(event)
    fire_ui_handlers(event, d)
    fire(record, None)
    self.emit(record)
    hdlr.handle(record)
    self.callHandlers(record)
    self.handle(record)
    self._log(INFO, msg, args, **kwargs)
    (self.dbfile, self.host, self.port, str(os.getpid())))
    self.work_forever()
    pid = self.daemonize()
    self.prserv.start()
    singleton.start()
    self.prhost = prserv.serv.auto_start(self.data)
    cooker.pre_serve()
    bb.cooker.server_main(self.cooker, self.main)
    self.run()
    code = process_obj._bootstrap()
    self._popen = Popen(self)
    self.serverImpl.start()
    server.detach()
    server = start_server(servermodule, configParams, configuration)
    ret = main()

It was never intended for the PRServer to send its logs anywhere but
its own log file.  The event processing is an artifact of how the
PRServer was forked and it inherits the event log handlers.  The
simple fix is to clean up and purge all the log handlers after the
fork() but before doing any of the typical PRServer work or logging.

(Bitbake rev: 972bc43e6d5b1207b944b3baa8f9805adb35dda7)

Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto-rt: add qemumips and qemuppc to COMPATIBLE_MACHINES

3.10-rt boots and has good cyclictest results on qemuppc and qemumips,
so we can now safely add them into COMPATIBLE_MACHINES.

(From OE-Core rev: 9dd21e4baf0d6220c2f751e62f417b73c6474759)

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

gcc-4.8.inc: Allow lto to be configurable

For some platforms its useful to be able to configure LTO so provide a variable
to allow this to happen.

(From OE-Core rev: e4582a51a2500ad3b418e53170f5fb6b2cbd98a5)

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

perf: Ensure we general PIC code to avoid build failures

Without this we see relocation errors on mips with 3.10. This should be
safe to be included in general.

(From OE-Core rev: 9958653b2bf9e43312a39c6b89ff0ca1cc46995c)

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

bitbake: hob: fixes for image combo box

When an image from scratch is selected, and recipes parsing
is canceled, the image shown by the combo box isn't correct.

[YOCTO #5000]
(Bitbake rev: f8166ace0bd9155199166990ce15da24eb2e793b)

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

bitbake: hob: add event handlers filtering in Hob

Create the _evt_list for hob; it is longer than the knotty
uses because it handles more events.

(Bitbake rev: 715aed74f972bb6e9b6a5130ca9ede48d4f79f0a)

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

bitbake: bitbake/event.py: UIhandler filter should work without a mask

The default for the mask will be * (all the handlers)

(Bitbake rev: 4c95e5f46cf2a656100bbf5a0e5a09d506abf9b9)

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

lttng-modules: Update to 2.2.1 based release

(From OE-Core rev: 8792f38de43b391896c2eccb8086538eb3f6c47b)

Signed-off-by: Tom Zanussi <tom.zanussi@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

README.hardware: Replace atom-pc with genericx86

Update README.hardware to reflect the recent replacement of
atom-pc with genericx86.

(From meta-yocto rev: 632d92087cebb11ff9cb0110f9dae01d04bdd83b)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Cc: Ross Burton <ross.burton@intel.com>
Cc: Scott Rifenbark <scott.m.rifenbark@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

linux-yocto: add bc-native dependency, and move to linux-yocto.inc

As reported by Martin Jansa <martin.jansa@gmail.com>, the following error happens
when building in a minimal environment:

   |   BC      kernel/timeconst.h
   | /bin/sh: bc: command not found
   | make[3]: *** [kernel/timeconst.h] Error 127
   | make[2]: *** [kernel] Error 2
   | make[2]: *** Waiting for unfinished jobs....

kernel commit 70730bca [kernel: Replace timeconst.pl with a bc script] added
a kernel dependency on bc. To support the build of linux-yocto recipes in
these configurations, we add bc-native to the common dependencies.

(From OE-Core rev: c888857b060f04b8689f393ec2d77a950da40f5a)

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: fix patch series to git tree validation

Previous changes to the kern-tools improved functionality to ensure that
as a series is considered, it is checked against the tree to confirm that
all patches are really applied.

There was a bug in the subject based detection, such that the first matching
patch was take, and not the last. This change ensures that we start from
the end of a series, not the start.

(From OE-Core rev: 6357657ec5b5687defaf1acdd94c1cf89aa06541)

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>

bc: add bc-native

To support the building of recent kernels in minimal environments, we should
provide bc-native.

(From OE-Core rev: 2138961cc75be2690f2f9b5df8cb2d810dcebc99)

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>

lib/oeqa/runtime: smart: add checks for smart output

Sometimes smart throws:

    Committing transaction...
    Preparing...                    ######################################## [  0%]
       1:Removing psplash-default   ######################################## [100%]
    error: Couldn't fork %postun: Cannot allocate memory

and returns a 0 exit code (it thinks it succesfully removed the package,
when in reality it didn't), so we need to catch those specifically.

Also, sometimes output from download command is:
    Saving cache...http://192.168.7.1:49456/rpm/x86_64_x32/psplash-default-0.1+git0+afd4e228c6-r15.x86_64_x32.rpm
and that tricks our smart download test, so use a regex there.

(From OE-Core rev: 2ac7783e04f5e8e6005f967e1a6dd65d2fc6a19a)

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

lib/oeqa/runtime: multilib: fix typo

The check was obviously wrong and it surfaced
with the recent change in behaviour for skipping tests.

(From OE-Core rev: 4a14535cd493cb2bdd46b2a5f2a1cd2b38161f0a)

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

u-boot: update to 2013.07

(From OE-Core rev: b2d6f08ff4f03079973eab76a790bf4555e32236)

Signed-off-by: Laszlo Papp <lpapp-RoXCvvDuEio@public.gmane.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gettext: update to upstream version 0.18.3.1

(From OE-Core rev: 18170af0d466815491f445ac7ff23906f7474d1a)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: command.py: Call updateCache for all states != running

updateCache handles the logic for shutting down the parsing so we need
to call it for all cases when we're not running.

This fixes hangs if Ctrl+C is pressed during parsing.

(Bitbake rev: 552b8935dd2f9f11e8d5c08a597a7e966b891480)

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

read-only-rootfs-hook.sh: check before bind mounting /var/lib

It's possible that /var/lib is on a separate writable partition. In such
situation, we should not bind mount /var/lib with tmpfs, becasue it's
already writable.

This patch fixes this problem by checking whether /var/lib is already
on a writable partition.

[YOCTO #4888]

(From OE-Core rev: 86ac10995fd08226f82d87e23fda5d4898c3190f)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

checkroot.sh: check for conflicting configurations

In our current boot scripts, two ways are used to determine whether
the rootfs is read-only or not. One by checking the READ_ONLY_ROOTFS
value in /etc/default/rcS, the other by checking the /etc/fstab entry.

>From a normal Linux user's point of view, the way of checking the
/etc/fstab entry is preferred. However, as there are several boot scripts
that need to know whether rootfs is read-only or not, checking /etc/fstab
in each script brings too much overhead. Thus, these boot scripts
use the READ_ONLY_ROOTFS value in /etc/default/rcS.

In normal use cases, there would be no problem, as both /etc/default/rcS and
the /etc/fstab are modified at rootfs time. However, if the user changes
the mount option for rootfs in /etc/fstab to read-write, and he/she forgets
to change the value in /etc/default/rcS, there would be unexpected results.
For example, the udev-cache would still be disabled.

So at a minimal, a check for conflicting configurations between /etc/fstab
and /etc/default/rcS should be added in checkroot.sh so that there would be
reasonable warnings if users have configured the system in a non-consistent
way.

[YOCTO #4880]

(From OE-Core rev: 1565a0c5a3f245703e280ca90cf11d3f9374788a)

Signed-off-by: Chen Qi <Qi.Chen@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

eglibc: fix directory cleanup for 2.17

Commit e0c2dd275827a4b37b8116d0f0119333638461af broke building
eglibc 2.17, which still ships pt_chown under eglibc/ directory.

Fix by only deleting directory when pt_chown is not there.

Cc: Khem Raj <raj.khem@gmail.com>
(From OE-Core rev: 10722a37bb7efa563425a7389100b9322d96492e)

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

SPDX:real-time license scanning and SPDX output.

SPDX integrates real-time license scanning, generates
SPDX standard output and license verification
information during the OE-Core build process. The
existing module includes scanning patched packages
and creating package and file level SPDX documents.

(From OE-Core rev: 7a37cc81fb95d56b5ac5e5ca22a1900e45717911)

Signed-off-by: liangcao <liangcao@unomaha.edu>
Signed-off-by: Elizabeth Flanagan <elizabeth.flanagan@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

classess/testimage: change default test suites

Some new tests were added, safe to have them in the defaults
for sato-sdk and sato. Not all of the new tests are here though,
either because they aren't applicable to default images or take too long.
(like build iptables/cvs/sudoky ones, they can be enabled
in local.conf and a special target on AB setups.). Also reorder them a bit.

(From OE-Core rev: caa18a99ec002e4e87e32cae8a2d28bb0e32c5a6)

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

python-2.7-manifest: Add missing python-ctypes dependency to python-multiprocessing

(From OE-Core rev: 5abf18a7f11ee9e88e0eec1b66cc63427d9097a8)

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

image_types.bbclass: replace genext2fs with populate-extfs.sh

* The benefits:
  - Really support ext4

  - Support the sparse file (we lost the sparse file in the image in the
    past, the sparse file became into the common file)

  - Fix the error reported by fsck: (ext2/ext3)
      Inode 1025, i_size is 16384, should be 17408.

  - Have a uniform code for ext2/3/4 generation

* Comments from Darren Hart:
Basically, genext2fs doesn't support creating ext4 filesystems. It
creates, as I understand it, an ext2 filesystem, then adds a journal,
and sets some bits. It can't support the newer features like extents. So
what we end up with is a bit of a hack for a filesystem.

The ext tools (e2fsprogs) unfortunately don't provide an integrated
solution for generating prepopulated filesystem images as many other
mkfs* tools do. One thing missing was symlink support in libext2fs. I
added that support and demonstrated a script which uses the e2fsprogs
debugfs tool that can populate the newly formatted filesystem from a
directory and without root privileges.

[YOCTO #3848]

(From OE-Core rev: 40c3e18f43b2f074cec97d21aeb8d21f26dd5048)

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>

lib/oeqa: change behaviour for unskippable tests

When a test module wants to be skipped because it doesn't
apply to the image but it was nevertheless a required
test (one in TEST_SUITES), we issued an warning that it
was a required test and went on with running the module.
Usually all tests in the module failed (e.g gcc tests on a non-sdk image),
but this allowed us to know that something went wrong with the image
(some package/feature didn't make it).

However, instead of just issuing an warning and running the tests
it's better to throw an exception. The traceback will tell us what's wrong,
and we don't run every single test method.
Output will look like this:
--snip--
| NOTE: Test modules  ['oeqa.runtime.ping', 'oeqa.runtime.ssh', 'oeqa.runtime.gcc']
| NOTE: Found 5 tests
| test_ping (oeqa.runtime.ping.PingTest) ... ok
| test_ssh (oeqa.runtime.ssh.SshTest) ... ok
| ERROR
|
| ======================================================================
| ERROR: setUpModule (oeqa.runtime.gcc)
| ----------------------------------------------------------------------
| Traceback (most recent call last):
|   File "/mnt/back/yocto/poky/meta/lib/oeqa/runtime/gcc.py", line 8, in setUpModule
|     skipModule("Image doesn't have tools-sdk in IMAGE_FEATURES")
|   File "/mnt/back/yocto/poky/meta/lib/oeqa/oetest.py", line 108, in skipModule
|     "\nor the image really doesn't have the requred feature/package when it should." % (modname, reason))
| Exception:
| Test gcc wants to be skipped.
| Reason is: Image doesn't have tools-sdk in IMAGE_FEATURES
| Test was required in TEST_SUITES, so either the condition for skipping is wrong
| or the image really doesn't have the requred feature/package when it should.
|
| ----------------------------------------------------------------------
| Ran 2 tests in 1.036s
|
| FAILED (errors=1)
| NOTE: Sending SIGTERM to runqemu
--snip--

(From OE-Core rev: fd51cecf8b258d9f839a0ecebde69d09f75dc468)

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

lib/oeqa/runtime: add iptables, cvs and sudoku projects build tests on target

Downloads iptables/cvs/sudoku-savant sources and builds them on target.

(From OE-Core rev: df4568205c3a7e0b20c6299e29f96bd30560146b)

Signed-off-by: Mihai Prica <mihai.prica@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/utils: targetbuild: Add helper class for building packages on target

This class can be used for test cases that configure
and build packages on target.

(From OE-Core rev: 4b15e82c4fcb0c40b0e316ef2050944eee4418ef)

Signed-off-by: Mihai Prica <mihai.prica@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa: make it possible to restart the target

Tweak QemuRunner so we can actually restart the
qemu target in a test (if we want more memory for example).
Also add a restart method to the base test class so that tests
can use it.

(From OE-Core rev: 9de7fe11967576f4a8b24e653c6b9a02e5f6d85b)

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

lib/oeqa/runtime: smart: add new smart tests

Add class to be inherited by smart tests, along with more basic tests and tests
using a rpm repository.

(From OE-Core rev: f6186b4204dcc421b4e616774315c8a2a77fb5c5)

Signed-off-by: Mihai Lindner <mihaix.lindner@linux.intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/utils: decorators: return the decorated method

Decorators should return whatever the decorated methods return.

(From OE-Core rev: c92513d6ff3f8f06d937a5cdf4d94708f27c3850)

Signed-off-by: Mihai Lindner <mihaix.lindner@linux.intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/utils: qemurunner: save host IP address

Save host IP address to host_ip.
Read /proc/PID/cmdline on host to look for IPs of target and host;
instead of running 'ps'.
Also removed some extra empty lines from file.

(From OE-Core rev: 1fcf10db10fa36430e37c95c9fee27197e73f7a5)

Signed-off-by: Mihai Lindner <mihaix.lindner@linux.intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/utils: new file: httpserver.py useful for serving files over HTTP to the target

It can be used by smart repo/channel tests to serve deploy_dir.

(From OE-Core rev: e38e18d6923cc3db50b56fa3fc64081fe4aa8669)

Signed-off-by: Mihai Lindner <mihaix.lindner@linux.intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: rpm: add install and erase tests

Copies to target rpm-doc file from deploy_dir
and tries to install and then remove that package.
rpm-doc was chosen because it's small, it only adds
a few files to target, and it's almost always found in
deploy_dir for images with package-management/rpm.

(From OE-Core rev: a2d2f2b7b111863d3c50dedded37aab813d9634f)

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

lib/oeqa/runtime: add vncserver for target test

(From OE-Core rev: 56bc5d717b34563ed36c0618305e4ec5080c3a27)

Signed-off-by: Mihai Prica <mihai.prica@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add adjust date and time test

(From OE-Core rev: ede300f44f1770b1e3e5c59f65cf5079379a2bc1)

Signed-off-by: Mihai Prica <mihai.prica@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add new scp test

Copies a 5MB to target using scp, more of an
network test than a scp one.

(From OE-Core rev: 2ec4a0686b9a91e56dfba3fa2e574c0c531508ff)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add new PAM support test

New test which verifies some usual commands functionality with PAM support

(From OE-Core rev: 06c7914e3354ff4e430a6b664f40e5a71e212761)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add new skeletoninit test

New test which verifies the usage of skeleton init script available with meta-skeleton layer

(From OE-Core rev: fcc59cbcdb1550489d372edf9f465efa7165245f)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add new logrotate test

New logrotate test which verifies the log directory change in logrotate.conf.
Needs an image with logrotate installed.

(From OE-Core rev: 186d79a603b5cbf5a93e6f5dbba5f62ed8d4d8d8)

Signed-off-by: Alexandru Palalau <alexandrux.palalau@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add a test for ldd

This test checks that at least one path in RTLDLIST exists.

(From OE-Core rev: d53ffbf21a6b17f0f0537929a95a502958d82a09)

Signed-off-by: Cornel Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add test for perl

This test runs a perl script on the target and checks the output.

(From OE-Core rev: 95abdbf2d715c9e7e3368f5193167feb5efc6261)

Signed-off-by: Cornel Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

lib/oeqa/runtime: add basic test for x32 images

Checks that an x86-64-x32 image has the right binaries.

(From OE-Core rev: da42e38b2bc4a606b717ac302cfc77aefb1a5795)

Signed-off-by: Alexandru Georgescu <alexandru.c.georgescu@intel.com>
Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

connman-gnome: fix DHCP segfault

In networks that don't have a DHCP server configured, ipv4 address
allocation fails and the ipv4 structure doesn't get populated.
The patch checks this case also.

[YOCTO #3945]

(From OE-Core rev: 2e3bff33f4ebeb6ac2272ab377d00416ef1af83f)

Signed-off-by: Mihai Prica <mihai.prica@intel.com>
Signed-off-by: Emilia Ciobanu <emilia.maria.silvia.ciobanu at intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

opkg: add missing initialization of conf->exclude_list

(From OE-Core rev: e4190b601c0212f0009ddb2dfa00d2880bddbc29)

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>

busybox: enable shared memory logging for systemd

The default command-line options for launching busybox's syslogd are
"-C" if using sysvinit; but because we use /etc/default/busybox-syslog
when using systemd which the sysvinit initscript doesn't use, and no
/etc/default/busybox-syslog is installed by default in OE-Core, the
default arguments with systemd were no arguments at all with OE-Core
alone. Effectively merge in the bbappend from meta-oe that adds a
default file in order to set the default options to "-C" for systemd as
well.

(From OE-Core rev: a23aa8e7467cf2b7f4e8ff85a3aa841ff6b508e5)

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>

ccache: update to upstream version 3.1.9

(From OE-Core rev: 9cf157c71453cbf835f237ca3fa05b845005fe51)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

libpam: add PACKAGECONFIG data concerning audit

We do not want libpam to build using audit just
because it happens to be lying around, so we
create PACKAGECONFIG[] data to give us explicit
control.

(From OE-Core rev: 4db6aa2094447f8d2a9c234089a80ddcd78fcbd0)

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

cronie: add PACKAGECONFIG data

We add PACKAGECONFIG[] data for audit and move
the current pam related stuff into it's own entry.

(From OE-Core rev: ada22cab96bb6b7cc98782a14e7c44fbebde77df)

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

e2fsprogs: add populate-extfs.sh

This script is originally from Darren Hart, it will be used for creating
the ext* filesystem from a given directory, which will replace the
genext2fs in image_types.bbclass at the moment, we may use the mke2fs to
replace this script again when it has the initial directory support.

Changes of the script:
* Rename it from mkdebugfs.sh to populate-extfs.sh
* Add a simple usage
* Add checking for the number of the parameters
* Add the "regular empty file" and "fifo" file type
* Set mode, uid and gid for the file
* Save the command lines to a file and batch run them
* Change the error message
* Improve the performance
* Add the support for hardlink

[YOCTO #3848]

(From OE-Core rev: 265f91149aa8c475ebe5b7069044ed94b7857fa9)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
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>

e2fsprogs: properly set up extent header in do_write

do_write doesn't fully set up the first extent header on a new
inode, so if we write a 0-length file, and don't write any data
to the new file, we end up creating something that looks corrupt
to kernelspace:

EXT4-fs error (device loop0): ext4_ext_check_inode:464: inode #12: comm
ls: bad header/extent: invalid magic - magic 0, entries 0, max 0(0),
depth 0(0)

Do something similar to ext4_ext_tree_init() here, and
fill out the first extent header upon creation to avoid this.

[YOCTO #3848]

(From OE-Core rev: 7d1e51681d25f6e6d2c20744825723ad5c83861c)

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>

e2fsprogs: only update the icache for ext2_inode

We only read the cache when:

bufsize == sizeof(struct ext2_inode)

then we should only update the cache in the same condition, otherwise
there would be errors, for example:

cache[0]: cached ino 14 when bufsize = 128 by ext2fs_write_inode_full()
cache[1]: cached ino 14 when bufsize = 156 by ext2fs_read_inode_full()

Then update the cache:
cache[0]: cached ino 15 when bufsize = 156 by ext2fs_read_inode_full()

Then the ino 14 would hit the cache[1] when bufsize = 128 (but it was
cached by bufsize = 156), so there would be errors.

[YOCTO #3848]

(From OE-Core rev: ad8452196c5b1a54c14fd00bbf421f68aea65186)

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>

e2fsprogs: let debugfs do sparse copy

Let debugfs do sparse copy when src is a sparse file, just like
"cp --sparse=auto"

This patch has been reviewed by the linux-ext4 mailing list, but isn't
merged atm.

[YOCTO #3848]

(From OE-Core rev: 723adaf8fbba61b7f1adc8e4a13ddf1cfb5c0bcf)

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>

e2fsprogs: the max length of debugfs argument is too short

The max length of debugfs argument is 256 which is too short, the
arguments are two paths, the PATH_MAX is 4096 according to
/usr/include/linux/limits.h, so use BUFSIZ (which is 8192 on Linux
systems), that's also what the ss library uses.

This patch has been reviewed by the linux-ext4 mailing list, but isn't
merged atm.

[YOCTO #3848]

(From OE-Core rev: a916a127768291ca7c614976e05c90153fec2956)

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>

oe.types: add 'path' type

- path normalization ('normalize' flag, defaults to enabled)
- existence verification for paths we know should exist ('mustexist' flag)
- supports clean handling of relative paths ('relativeto' flag)

(From OE-Core rev: a598242197312fa6d43179c283da2d0873de2919)

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

stat: fix the wrong return code

After stat -h is executed, the return code is 1.
This patch changes it to 0.

(From OE-Core rev: c10435025f5c51c4827d71af82a9a517f5f8f6d8)

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

gtk-engines: fix build with automake-1.13

Add patch substitute-tests.patch that works around automake
TESTS limitation.
See http://debbugs.gnu.org/cgi/bugreport.cgi?bug=13771.

(From OE-Core rev: ace40e1dfcaca2aa1eae2daf29c812291063fbff)

Signed-off-by: Marko Lindqvist <cazfi74@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

pulseaudio: Fix up configure options and use gtk+3

(From OE-Core rev: b51bfbc304eba448b9913d36444d70a643564c64)

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

alsa-lib: fix function definitions error for mips

Functions atomic_add(s) and atomic_sub(s) are defined with 'extern
__inline__' that may cause compile fails when cross compile for mips.
The error message looks like:

| pcm/.libs/libpcm.a(pcm_meter.o): In function
`snd_pcm_meter_update_scope':
| .../alsa-lib-1.0.27.2/src/pcm/pcm_meter.c:139: undefined reference to
`atomic_sub'

Replace the 'extern __inline__' with 'static __inline__' to fix this
issue.

(From OE-Core rev: 80b4eb9e12ccf815261f6a67fd9b8d97717e82d5)

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

connman: upgrade to 1.17

connman 1.17 is mainly a bugfix release,
as such, recommended to upgrade to.

(From OE-Core rev: ad98f2bfa2c6ebb53b9c1fa4afad8842ab65062c)

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>

harfbuzz: upgrade to v0.9.19

Remove the disable_graphite.patch because the --without-graphite2
configure option can be used instead.

Split package into harfbuzz and harfbuzz-icu. The ICU support is
now built into a separate library that will be shipped in a
different package.

(From OE-Core rev: c93d147e0889ac91a01e3570f818624f68b3c8e8)

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

gpgme: update to 1.4.3

(From OE-Core rev: 3b4f5c4153c0af0e5205d9c30e4b7265c36079d3)

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>

runqemu-ifup: when tunctl can't be found, say what package builds it

If runqemu is used without actually building any qemu images (i.e. you
downloaded the images) it's likely that qemu-helper-native hasn't been built.
Instead of just saying what command can't be found, tell the user how to solve
their problem.

(From OE-Core rev: 45f80a0c46035dc73818ce4bd818a4f6197d954f)

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>

openssl: avoid NULL pointer dereference in three places

There are three potential NULL pointer dereference in
EVP_DigestInit_ex(), dh_pub_encode() and dsa_pub_encode()
functions.
Fix them by adding proper null pointer check.

[YOCTO #4600]
[ CQID: WIND00373257 ]

(From OE-Core rev: 4779d3c89cf0129763a4f5b7306c1247a0d6d021)

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

openssh: add init.d/sshd status command for LSB compliance

(From OE-Core rev: d1e518b6942b945be7a4d44bc137a1441af8f837)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

at: add init.d/atd status command for LSB compliance

(From OE-Core rev: 9c19592afac6b38df7038da02bcda26dc28b4e4c)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sysvinit: add init.d/bootlogd status command for LSB compliance

(From OE-Core rev: 95180b966adb6ebe6451ab0e91187e7db93f22e4)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

dbus: add init.d/dbus-1 status command for LSB compliance

(From OE-Core rev: 9262d3b06eee7e698c990399d2f10b7d68d17211)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

nfs-utils: add init.d/nfsserver status command for LSB compliance

(From OE-Core rev: d5132337fa08ee74eaebadaeaecb6b4f5b2c5190)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sysklogd: add init.d/syslog status command for LSB compliance

(From OE-Core rev: 7f0314c095032c231e2408c0276fea56b4e68884)

Signed-off-by: Li Wang <li.wang@windriver.com>
Signed-off-by: Jackie Huang <jackie.huang@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

glib-2.0: Specify libiconv when building on uclibc

We use GNU libiconv seprate package on uclibc
specifying this makes the configure clear of not
trying to detect glibc/iconv and then also finding
libiconv in sysroot and confusing itself with errors like

| gconvert.c:66:2: error: #error GNU libiconv not in use but included iconv.h is from libiconv
|  #error GNU libiconv not in use but included iconv.h is from libiconv
|   ^

(From OE-Core rev: 38b6c4df7c215ed7fd6be107fbc2527e66791e2e)

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>

bitbake: data_smart: Add _remove operator

There are long standing complaints about the fact its very difficult
to remove a portion of a variable. The immediate request is for a -=
and =- operator. The trouble is that += and =+ are "immediate"
operators and are applied straight away. Most people would expect
-= and =- to be deferred to have the effect most people desire and
therefore implementing -= and =- would just make the situation more
confusing.

This deferred operation is much more similar to the override syntax
which happens at data store finalisation. The _remove operator is
therefore in keeping with the _append and _prepend operations.

This code is loosely based on a patch from Peter Seebach although it
has been rewritten to be simpler, more efficient and avoid some
potential bugs.

The code currently only works on space delimited variables, which
are by far the most commom type. If bitbake is ehanced to support
types natively in future, we can adjust this code to adapt to that.

(Bitbake rev: 9c91948e10df278dad4832487fa56888cd58d187)

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

bitbake: bitbake: Add ui event handlers filtering

Add functionality to allow UIs to update and change the types of events they
recieve. To do this we need to add a new command and also need to be able
to obtain the current event hander ID. In the case of xmlrpc, this is
straightforward, in the case of the process server we need to save the result
in a multiprocessing.Value() so we can retrive it. An excplit command
was added to the server API to facilitate this.

The same function can also be used to mask or unmask specific log messages,
allowing the UI to optionally differ from the standard set of message
filtering.

Based upon work by Cristiana Voicu <cristiana.voicu@intel.com>

(Bitbake rev: ba5a6c88785d9889d4172ec79937ac2a5555327e)

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

bitbake: server/xmlrpc/prserv: Add sane timeout to default xmlrpc server

The standard python socket connect has long timouts which make sense for remote
connections but not local things like the PR Service. This adds a timeout
parameter to the common xmlrpc server creation function and sets it to a more
reasonable 5 seconds.

Making the PR server instantly exit is a good way to test the effect of this
on bitbake.

We can remove the bodged timeout in the PRServer terminate function which
has the side effect of affecting global scope.

(Bitbake rev: 8c01cff94787abbb64fbdf0c16cd63f8f97a7e03)

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

bitbake: process: Improve exit handling and hangs

It turns out we have a number of different ways the process server termination can
hang. If we call cancel_join_thread() on the event queue, it means that it can be left
containing partial data. This means the reading of the event queue in the terminate()
function can hang, the timeout and block parameters to Queue.get() don't make any
difference.

Equally, if we don't call cancel_join_thread(), the join_thread in terminate()
will hang giving a different deadlock.

The best solution I could find is to loop over the process is_alive() after requesting
it stops,  trying to join the thread and if that fails, try and flush the event
queue again.

It wasn't clear what difference a force option should make in this case, we're
gracefully trying to empty queues and shut down regardless of whether its a SIGTERM
so I've simply removed the force option.

(Bitbake rev: c5c8f33ca4b81877a0115887849881001b745bf0)

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

bitbake: parse: Return IOError when including file with absolute path

* resolve_file was behaving different when relative and absolute
  paths were passed to it

* include relative-path/non-existent-file.inc
  works correctly resolve_file throws IOError, BBHandler.py:handle()
  doesn't catch it, ConfHandler.py:include() catches IOError and shows:
  DEBUG: CONF file 'relative-path/non-existent-file.inc' not found
* include /absolute-path/non-existent-file.inc
  was failing, because resolve_file just returns fn,
  BBHandler.py:handle() calls bb.parse.mark_dependency(d, abs_fn)
  which throws:
  OSError: [Errno 2] No such file or directory: '/absolute-path/non-existent-file.inc'
  and parsing fails.
  Ad isfile() test for absolute fn and throw IOError to make
  resolve_file behavior consistent for both paths.

* I know we had some issues with -b relative-path-to-recipe.bb and
  absolute path, so consider this patch only as RFC and documentation of
  this problem

* Catch OSError too in ConfHandler.py:include() e.g. in case the file exists, but user
  cannot read it or something like that.

(Bitbake rev: b0bbd89a4f0b98fa1ab28b8e0526cd9ddb76fa57)

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

nativesdk-ca-certificates: prepopulate ca-certificates.crt

As postinsts aren't run for nativesdk packages when populating an SDK, we need
to prepopulate up-front.

(From OE-Core rev: 09e768b68b3605e897d422c9c7b3815f3b994d31)

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

ca-certificates: add recipe (version 20130610)

We need this for certain nativesdk recipes, as we can't rely on the
certificate path or bundle path being the same across distros, and it's useful
in many cases on the target as well.

This is based on the 20130119 recipe from meta-oe, with the following changes:

- use the debian git repository to avoid vanishing sources
- obey our target paths
- default to a sysroot relative to the script location (make relocatable)
- define SUMMARY
- don't inherit autotools, this isn't an autotools package
- add MPL-2.0 to LICENSE, as that's the license of the certdata
- install the script man page
- use a native rather than cross recipe, as it's not bound in any way to the
  target system
- add nativesdk to bbclassextend, for use in SDKs

(From OE-Core rev: ad2851cf0abc2ab35e0f60c96d3142c29a07c8fc)

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

xserver-nodm-init: improvements from meta-oe

Bring in some changes from the meta-oe version of this script in order
to bring them closer together. (Unfortunately it does not completely
supersede the recipes in meta-oe - there is still some unification to be
done between x11-common and xserver-common.)

* Add systemd support (with extensions for ROOTLESS_X)
* Remove sudo from RDEPENDS_${PN} since this hasn't been used by this
  for a long time
* Rearrange recipe slightly to make the order logical

A couple of additional improvements at the same time:
* Use ROOT_HOME for root's home directory
* Set short description in SUMMARY instead of DESCRIPTION

(From OE-Core rev: d4f82500fb30e9e25b517e65c2bd72b3c2bad070)

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>

classes: Remove references to _remove in function names since this may become a bitbake keyword

There is a good chance we might want to support a bitbake operator
"_remove" which works in a similar way to _append and _prepend. As
such, we can't use those keywords in function or variable names.

(From OE-Core rev: 491fde8cd3fd493f9fec2fd434fe1be547f66148)

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>

linux-yocto: introduce v3.10

Introduce the 3.10 kernel. This is based on the 3.10.9 upstream kernel
and replaces the 3.8 recipes, which will be removed once all reference
boards have been updated.

3.10 also the latest LTSI and will be updated with that content when it
becomes available.

Other features of the 3.10 kernel include:

  - refreshed -rt support
  - refreshed yaffs2, aufs3
  - cryptodev
  - bfs, edf, and OCF staged features
  - scrubbed and updated meta data for v3.10
  - improved tools support for meta data updates and queue maintenance
  - patch carry forward from all previous linux-yocto kernels and
    configuration.

This kernel has been built and boot tested on all qemu machines and architectures.

(From OE-Core rev: 6dc46834c6edaf358c18b26e4304bc9e7413eb60)

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

gst-plugins-good: fix 3.10 libc-headers build failure

kernel commit 24b9f5017 [[media] V4L: Remove deprecated image centering controls]
removed the definitions of V4L2_CID_HCENTER and V4L2_CID_VCENTER after three
years of depreciation.

The ioctl values are still free, and the case statement which processess them
in v4l2 userspace falls through to the proper replacement. So in the short
term, we can explicitly define them using the old absolute values, and everything
will work.

(From OE-Core rev: 1e35d04067357b1c65d9e2c4796acfa02274fd9d)

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

linux-libc-headers: Reinstate scripts-Makefile.headersinst-install-headers-from-sc.patch

The autobuilders and their long paths are still triggering errors during
the headers installation. Reinstate the previous patch for this,
after updating for 3.10.

(From OE-Core rev: fe4428fd740b3937007e0a3f893714ff04c33533)

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

linux-libc-headers: ptrace.h: remove ptrace_peeksiginfo_args

The addition of ptrace_peeksiginfo_args to the uapi in kernel commit
84c751bd [ptrace: add ability to retrieve signals without removing from a queue (v4)]
means that existing applications using glibc versions that define ptrace_peeksiginfo_args
in sys/ptrace.h will get duplicate structure definitions like:

    | In file included from /poky-master/build/tmp/work/i586-poky-linux/strace/4.8-r0/strace-4.8/process.c:66:0:
    | /poky-master/build/tmp/sysroots/qemux86/usr/include/linux/ptrace.h:58:8: error: redefinition of 'struct ptrace_peeksiginfo_args'
    |  struct ptrace_peeksiginfo_args {
    |         ^
    | In file included from /poky-master/build/tmp/work/i586-poky-linux/strace/4.8-r0/strace-4.8/defs.h:159:0,
    |                  from /poky-master/build/tmp/work/i586-poky-linux/strace/4.8-r0/strace-4.8/process.c:37:
    | /poky-master/build/tmp/sysroots/qemux86/usr/include/sys/ptrace.h:191:8: note: originally defined here
    |  struct ptrace_peeksiginfo_args
    |         ^
    | make[2]: *** [process.o] Error 1

Reverting to the previous status of not exporting this structure temporarily
fixes applications, until they can be adjusted to not mix sys/ptrace.h and
linux/ptrace.h includes.

(From OE-Core rev: 7c207e4c3c0e3b575c67a302b97b0dc700a8fc15)

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-libc-headers: update to v3.10

Now that the 3.10 kernel has been released we can bump the libc-headers to
that version and remove the 3.8 variant. Userspace compatibility is
maintained through kernel versions, we also make the single 3.10 version the
toolchain default.

(From OE-Core rev: 4e79a46254e778f85c00efd4b0085cbaeb6e0d4d)

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

guilt: update to latest git version

Uprev guilt to the latest guilt version from its upstream repository.
As part of the uprev all of the previous changes required for the
yocto kernel tools to use git to manipulate series files have been
dropped. These changes were specific to circumventing parts of guilt's
internal santiy checking to allow specific Yocto kernel manipluation
of sub-series files.

Since the kernel tools no longer need guilt, we can use an up to date
and nearly pure upstream version of guilt.

(From OE-Core rev: 595c4469adc36d88ba2403915fc6c1d355014a58)

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

kern-tools: usability, bug fixes and no guilt

Updating the kern-tools SRCREV to pick up the following fixes:

   60a894e kgit-s2q: add proper commit ID handling for mixed am/apply usage
   3b08257 kgit-s2q: delete pruning of path support.
   c5868b4 kgit-s2q: Restore implicit exit status to "git apply" section
   1bd00b9 kgit-scc: mask warnings from cleanup phase 5
   bb75299 kgit-s2q: fix commit warp when running "git am --abort"
   ef9571b kgit-scc: cleanup git rebase-apply dir
   fdb7d21 kgit-scc: ensure treegen stops if a meta series fails
   008987b config: add kconfig cleaning options
   69ff569 kgit-s2q: strip blank lines and comments
   e7b4540 kgit-init: disable garbage collection on a new tree
   417eaed kgit-s2q: delete old LTSI patch dir finding code
   21f2200 kgit-scc: better error checking on resume
   ad5084c kern-tools: use .meta as meta data container
   1deb5d8 kgit-meta: don't push patches without a series file
   eb431a1 kgit-s2q: aid patch reject resolution via helper scripts
   f859c40 kgit-s2q: only use patch annotations when explicitly asked
   333ae18 kgit: speed patch application by batching patches
   bf6991d kgit: teach tools about non-default meta dirs
   bcfc712 kgit-s2q: usability improvements
   cb28803 kgit-s2q: fix patch prefix stripping.
   37f40e1 kgit-s2q: warn/exit with error if patch not in series
   f4704d2 kgit-s2q: consistent rm usage
   e11819c kgit-s2q: standardize on use of git mailinfo
   36a5eda kgit: remove guilt dependency
   c461a4f spp/scc: export mark commands to meta-series
   5311162 updateme: ensure that generated features are only used once
   4f7a263 kgit-checkpoint: clear .gitignore for meta branch
   21ee6f2 updateme: enforce a matching machine
   b08749d kgit-scc: remove -meta files after consruction

These are bug fixes, usability changes as well as the removal of the
guilt dependency. During the uprev of the guilt package, the amount of
circumvention of the typical guilt workflow and checks meant that using
it as a series -> branch manager was no longer appropriate. As a result
a new tools kgit-s2q (series 2 queue) was created based on git-quiltimport,
git am, and the LTSI tree generation scripts.

The result is better series to branch validation, faster application and
a simpler management model. This tool is backwards compatible with any
tree previously constructed with guilt. We are now "guilt free"

(From OE-Core rev: 983bff587b60fdd0244ad00f238df5ed50cc1e1a)

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

Add memory resident bitbake script

This adds an init script which instead of the standard bitbake, launches
a memory resident bitbake, defaulting to port 12345. It expects a port
number to use as the first option.

Right now this is experimental but I think its probably worth wrapping
up in a form people can more easily experiment with it. There are some
known issues:

a) It throws some debug output due to the lack of a UI which we need
   to clean up
b) It should probably be able to auto select a free port
c) You get a nice backtrace if you specify a build directory but
   not a port number

I'd also highlight there are security issues here if you don't trust
users who can connect into localhost. We might need to look at named
pipes or something similar for something limited to the current user.

(From OE-Core rev: 52c7f8bba86a43b89f24a23d545c99d75b67555f)

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

bitbake: prserv/serv: Fix pid file removal

Mark Hatle spotted there were pid files being left around. This patch
fixes things so the removal function is called correctly, the code
contained a typo.

(Bitbake rev: c696a16c8200c31c52750037eeafe07e065b6517)

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

bitbake: event/msg: Add primitive server side UI log record filtering

Currently one of the bigger bottlenecks in bitbake is passing all the
log messages over IPC to the UI. This is worthwhile if the UI is going
to use them, pointless otherwise. The memory resident bitbake suffers
from this performance issue particularly badly.

This patch filters the log events on the server side with the global
log levels and hence reduces the traffic. This speeds up parsing
(18.5s down to 17s) and bitbake general command overhead is reduced
(7.3s for a NOP to 6.2s).

What isn't added here is general event filtering or the ability to
change the log levels once set. Provision is made for adding this
in a follow up patch though.

(Bitbake rev: 1bf0e88f57ba0bca62532e81d0d62cf88e2abcbb)

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

bitbake: build.py: create symlink for run.do_xxx scripts

The 'courtesy' symlink for log.do_xxx are quite useful when debugging, so
with this commit, we now get similar 'courtesy' symlink for run.do_xxx
scripts.

We only create symlink for tasks, not individual functions.

The symlink is create right before the actual runfile is created, indeed
we cannot create the symlink right after running the task since a failure
or execption can happen, in which case the symlink wouldn't be created,
and symlink are particularely useful when the task failed!

Another option would be create the symlink after the runfile is created,
and before the script is executed, but that means we need to duplicate the
code in case of Shell vs Python task.

(Bitbake rev: a672b39c5d529ba85d72eee8fef4c4273eaa5397)

Signed-off-by: Nicolas Dechesne <nicolas.dechesne@linaro.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sstate: Fix the relative symlink replacement code

ant reported on irc that the sstate absolute to relative symlink creation
code wasn't working in klibc. He was correct although the level of breakage is
rather surprising since it only worked for one level of symlink (usr/include) with
everything else being broken.

The reason is probably that nothing really uses absolute paths, we use relative
paths where at all possible already. Nothing in the target sysroot should use
absolute paths for a start. In this regard, the klibc-dev package is broken and
needs fixing. It will currently break when building for one machine, then switching
to another of the same TUNE_PKGARCH and installing from sstate but that is a
separate issue.

This patch fixes the symlink creation code by firstly passing in the correct
value we need (where the symlink will end up) and seccondly, actually using it.

I've also tweaked the debug message to contain appropriate information and got
right of the double "//" value the existing code created in favour of the form
'./..' which looks neater.

(From OE-Core rev: 9b05c65450526522d7358d0c0901b594de546748)

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

xdg-utils: add runtime dependency on xprop

xprop is called by xdg-utils scripts.

(From OE-Core rev: 58d05c817cf3be071c66c7e4dbbb9942fd7a0752)

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

packagegroup-self-hosted: pcmanfm and xdg-utils integration

PCManFm file manager is integrated in Build Appliance;
xdg-utils is also integrated for file
association support.

(From OE-Core rev: a031523eb072df8f04dbae296a44ad51268acdf4)

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

recipes-graphics/builder: fix wrong user name

Fix wrong inclusion of local username.
The right "builder" user is now used in
builder_hob_start shell script.

(From OE-Core rev: 0b0d48a4b8af310d1f6eabf2c4d459cda1c2b4bb)

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

packagegroup-self-hosted: Add the sato icon theme

Hicolor icon theme does not properly displays icons for
folders in Build Appliance.
Sato icon theme is working correctly.
Also, settings-daemon needs to be added to image in order to
properly display folder icons.

(From OE-Core rev: a5188962c4ce6312fd625c2ab9601c24765255ed)

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