command.py: add new API to get the cpu count on the server

Add a new API in command.py to get the cpu count in order to set the appropriate default BB_NUMBER_THREADS and PARALLEL_MAKE variables.

(Bitbake rev: 335047b2e440e65713e88fabb24b47a9c82f939b)

Signed-off-by: Shane Wang <shane.wang@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: Add client socket info for BitBakeServerConnection

In server/client split model, the client will bind to a specific address
and port. We need to pass the values to BitBakeServerConnection().

(Bitbake rev: c8e19c5c389efc06696084c6f9439ba75472c5b7)

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: add -B option to bind with interface

When start bitbake as a server only process, we need to assign certain
interface to it.

(Bitbake rev: 95b97d2dc6466ea3d99371f5b5bd68f6f3c99074)

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

event.py: Add new events RequestPackageInfo and PackageInfo

RequestPackageInfo is triggered by GUI client to request the available
package information.

PackageInfo event is to pass package information back to GUI.

(Bitbake rev: 9020c2d4476766f63ff7e024bbd99043d06feefc)

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cache: Use configuration's hash value to validate cache

Previously we use the file time stamp to judge if a cache is valid.
Here this commit introduce a new method, which calculates the total
hash value for a certain configuration's key/value paris, and tag
it into cache filename, for example, bb_cache.dat.xxxyyyzzz.

This mechanism also ensures the cache's correctness if user
dynamically setting variables from some frontend GUI, like HOB.

(Bitbake rev: 1c1df03a6c4717bfd5faab144c4f8bbfcbae0b57)

Signed-off-by: Dongxiao Xu <dongxiao.xu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

bitbake: add file and line number to ParseError

Ensure that a file and line number are reported for ParseError where
possible. This helps particularly in the case of inherit and require
which previously did not report either of these upon failure.

(Bitbake rev: f588ba69622a2df35417ced184e56c79ac1b40d5)

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

bitbake: fix parse errors not being reported

Fixes a regression introduced in BitBake rev
c9f58ef6b897d3fa5b0d23734b5f2cb3dabb057a which prevents errors during
parsing from being fully reported because BitBake shuts down before it
can print them. Move the error printing to before the shutdown in order
to fix it.

Also remove a couple of superfluous calls to sys.exit() since these
will never be called (as self.shutdown(clean=False) does this itself).

(Bitbake rev: 356c42ed00541dea2c382c57f768a4f431d22db0)

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

siggen.py: Pass the datacache to signature generator classes

The dataCache exposes much useful information to the singature generation code
for example its possible to figure out whether kernel dependencies are
important or not based on the RRECOMMENDS values.

We therefore pass this information to the handler and trust it to handle
any API changes in the dataCache structure.

For backwards compatibility, OE-Core is already updated to handle
the extra argument when it appears.

(Bitbake rev: 4ebc25155d5f96a7f63de22b52b18d045eb4dec7)

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

texi2html: Fix for multilib

(From OE-Core rev: 55f72b98f606c1554eb6edd151292ffdfddf1384)

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

quicky: Fix for multilib

(From OE-Core rev: 1be7873cec6630ddb485590a8473e04cfb224615)

Signed-off-by: Zhai Edwin <edwin.zhai@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

netbase: remove redundant assignments

There's no need to explicitly set PACKAGE_ARCH = MACHINE_ARCH, base.bbclass
takes care of setting this value for us based on the interfaces for those
machines being an OVERRIDE.

(From OE-Core rev: c3d5800d2850a186f91b5a0db642aa5d1c20156b)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

base.bbclass: check all entries of FILESPATH for MACHINE overrides

The logic which looks for MACHINE overrides in SRC_URI and updates
PACKAGE_ARCH was checking only certain subdirectories of the recipes parent
which, amongst other issues, doesn't account for SRC_URI overrides in layers.

This patch changes the logic such that all FILESPATH entries are checked
for children named for MACHINE.

(From OE-Core rev: f2b572072c754048aaafdc4c42b71af87d61d9e7)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

external-csl-toolchain: skip parsing if CSL_VER_MAIN isn't set

(From OE-Core rev: 264d1254e5ac11613af06c0f1c53c736f54e9f54)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

sanity.bbclass: Add ABI version 8, used for transition from the OEBasic to OEBasicHash stamp layout

(From OE-Core rev: bf440fd971dd549ae4e92c54046ea062ef51cd85)

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

bitbake/ui/knotty: fix incorrect exit code

The recent Ctrl+C patch was not fully tested; the variable it was
relying upon is set to non-zero under other circumstances and thus
bitbake was reporting that it was interrupted and returning a non-zero
exit code when it was not. Track this status in a separate variable
in order to fix the issue.

(Bitbake rev: 95a599067650902727ecb4a39d6dd003c5cfedf3)

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

siggen: compare runtaskdeps dictionary even when they have the same size

* otherwise it reports lots of changed checksums just because it compares different tasks

  notice linux-libc-headers_3.1.bb.do_package_write:
  Task dependency hash changed from 42acced29debf54d598802474c5e9cbb to f438a54f995df77620d0727d6f4b4ce5 (for linux-libc-headers_3.1.bb.do_package_write and linux-gta04_git.bb.do_deploy)
  Task dependency hash changed from 61f8babe1d10c6e7fb1423112bb04e1e to 1b3e21ff106ecfcb7ddf76a1e29537bb (for linux-nokia900-meego_git.bb.do_deploy and linux-gta04_git.bb.do_package_write)
  Task dependency hash changed from 512f9d6686d760b318d8b11c8b589226 to 42acced29debf54d598802474c5e9cbb (for linux-nokia900-meego_git.bb.do_package_write and linux-libc-headers_3.1.bb.do_package_write)
  Task dependency hash changed from 153e91dfd1d2053fda7b98cc08d4b802 to 92a293bdd8ed234932b87a66025038c5 (for systemd-serialgetty.bb.do_package_write and systemd-serialgetty.bb.do_package_write)
  Task dependency hash changed from 730abebf9954794bb440c2f3239f79fe to 413eaebaff27a2fd16f5cf68c1f4ff17 (for systemd_git.bb.do_package_write and systemd_git.bb.do_package_write)

  with this patch:
  Dependency on task linux-nokia900-meego_git.bb.do_package_write was added
  Dependency on task linux-nokia900-meego_git.bb.do_deploy was added
  Dependency on task linux-gta04_git.bb.do_deploy was removed
  Dependency on task linux-gta04_git.bb.do_package_write was removed
  Hash for dependent task systemd_git.bb.do_package_write changed from 730abebf9954794bb440c2f3239f79fe to 413eaebaff27a2fd16f5cf68c1f4ff17
  Hash for dependent task systemd-serialgetty.bb.do_package_write changed from 153e91dfd1d2053fda7b98cc08d4b802 to 92a293bdd8ed234932b87a66025038c5

* added test if there is different task with same hash then we don't
  show it as added/removed dependency, because bitbake doesn't care

(Bitbake rev: ca52bf32b479811bd7fed41648bedcc06b00430b)

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

bitbake/knotty: ensure CommandFailed increases error count

CommandFailed already sets the return code, so we print out the error
summary message, however we don't increase the error count so it is
usually zero in this case. As we are actually showing an ERROR message,
increment the error count so that the summary makes sense.

(Bitbake rev: aeb71c3ca65dc42015c29c0e4cac050bfbcb51a5)

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

bitbake/knotty: ensure we return non-zero exit code on Ctrl+C

If the user uses Ctrl+C to exit out of the build, then ensure we return
a non-zero exit code since the entire build was not completed
successfully.

If nothing else, this is useful to avoid continuing in the bitbake
wrapper script in OE-Core if the initial pseudo build is interrupted.

(Bitbake rev: c11eab7fa230f1fd5cc33589f3555b94e95f202d)

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

bitbake: show appropriate warnings for universe target

When building the universe target:
* Show a warning about universe likely producing errors
* Multiprovider errors are now shown as warnings

Also remove an unused "error" variable in runqueue.py.

Fixes [YOCTO #1936]

(Bitbake rev: 293c796e8a3d1f74ca1e51017b6dede261612281)

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

Fix parser hang at 0% with single-core machines

The number of threads is self.num_processes - 1 due to the range(1,
self.num_processes). Fixed by changing to range(0, self.num_processes).

(Bitbake rev: adc041fd9e3def29cdf9c1ae4849c5383bac46e5)

Signed-off-by: Caleb Crome <caleb@signalessence.com>
Signed-off-by: Christopher Larson <kergoth@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

persist_data: survive if sqlite3 couldn't enable the shared cache

(Bitbake rev: e870fd57bb84a17a871a5b5de3bf5202fe9b135f)

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

knotty: shut down more cleanly

This ensures that when we are shutting down, we drain the queue of events from
the server. This ensures that the server never gets hung up on the join of the
feeder thread associated with the event queue, thereby avoiding a bitbake hang
issue.

Also, since the third keyboard interrupt now behaves the same as the second,
removed that message.

(Bitbake rev: ce9e0946b8a412616074809c76fba54f8bff36c4)

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

cooker: roll our own process pool

This fixes the hang issue encountered with parse errors. The underlying issue
seems to have been the pool.terminate(). This sends SIGTERM to each of the
multiprocessing pool's processes, however, a python process terminating in
this fashion can corrupt any queues it's interacting with, causing a number of
problems for us (e.g. the queue that sends events to the UI).

So instead of using multiprocessing's pool, we roll our own, with the ability
to cancel the work. In the very long term, the python concurrent.futures
module introduced in python 3.2 could be used to resolve this as well.

(Bitbake rev: 7c39cfd8e060cca8753ac4114775447b18e13067)

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

iputils: Add base_libdir to VPATH in order to find the crypto library

(From OE-Core rev: d4ffe12ca36bf10be4e0f9565d7c3d8e6f4a265a)

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

consolekit: depend on libx11

* Fixes the following error at configure time:

|checking for TOOLS... no
|configure: error: Package requirements (x11 >= 1.0.0
|  glib-2.0 >= 2.14
|) were not met:
|
|No package 'x11' found

(From OE-Core rev: e5c60517a22997bb966bbe21a2d7cbc99a97b42b)

Signed-off-by: Andreas Oberritter <obi@opendreambox.org>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

udev.inc has a bad link for udev home page; update the link

Fixed bad link in udev.inc to the udev home page / documentation on
kernel.org

(From OE-Core rev: 479b9a52a013d990da296f96bd658db00423f9a6)

Signed-off-by: Robert Cochran<yocto@mindchasers.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Add sanity checks for eglibc locale

Add sanity checks that test for the existance of folders. This allows to use
an external binary toolchain that doesn't have localization support.

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

libsdl: add fix for new _XGetRequest symbol backported to libx11

(From OE-Core rev: d2c96e22de19a649ff7121e5ea0f95797abba93a)

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

eds-tools: Update it to fix the build error

The head file 'libedataserver/e-data-server-module.h' has been moved
to 'libebackend/e-data-server-module.h', just update eds-tools to get
the latest changes for fixing the path of 'e-data-server-module.h'.

Also update its recipe file to reflect it has been converted to git
repo and split all generated binaries to correct packages.

[YOCTO#1786]

(From OE-Core rev: 38d74ee1fe6f28ec4d7db79f550c7d2830d922fb)

Signed-off-by: Wenzong Fan <wenzong.fan@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

buildhistory_analysis: correctly split RDEPENDS/RRECOMMENDS

Split RDEPENDS and RRECOMMENDS correctly (which may contain version
number specifications after each item).

(From OE-Core rev: d974ed4330f33c4f8dfd635a83c61a0d44e688b6)

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

buildhistory_analysis: avoid noise due to reordering

Avoid noise in the output due to reordering of list variables (except
for PACKAGES where we just report that the order changed). Recent
changes to the buildhistory class itself will avoid this reordering
from occurring but this allows us to examine the results before and
after those changes.

(From OE-Core rev: e23c5b01766602c9c86b0a7ba170fb3b1aceb658)

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

classes/buildhistory: fix splitting on + in package list fields

Ensure we do not erroneously split on + in RDEPENDS/RRECOMMENDS e.g.
libstdc++-dev was being split into libstdc and -dev.

(From OE-Core rev: cad533880df42ad4fe6f04d56d3a59cb4a033275)

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

classes/buildhistory: use hostname instead of reading /etc/hostname

For the purposes of querying the hostname to include it in the commit
message, it seems "cat /etc/hostname" does not work on the Yocto Project
autobuilder machines, and it's likely that the hostname command will be
more generally reliable, so use that instead.

(From OE-Core rev: cb939b753c9e7648a38e22e0349c279da785e69d)

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

classes/buildhistory: sort FILELIST in package info

The FILELIST order can vary depending on the order the files were
written which may change between builds with no ill effect, so sort the
list prior to writing it.

(From OE-Core rev: 2e9981000a211a89f88d7728393cc231e466581a)

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

classes/buildhistory: sort and de-dupe dependency graphs

Sort dependencies of each package which sometimes change order and cause
noise in the buildhistory repo, and at the same time remove duplicates
(which seem to be common especially for the RPM package query output).

(From OE-Core rev: 830df6067c1ea4a5aab580b42ba7e1e84fe1bcbf)

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

classes/buildhistory: squash spaces out of image variables

Values of image variables that are lists (e.g. IMAGE_INSTALL) are easier
to read if there are no extraneous spaces in them, so ensure that there
is only one space between each item.

(From OE-Core rev: 200159125eb6bcfc046c45cf5160b2eb340625e3)

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

task-self-hosted: add connman and connman-plugin-ethernet

Without this, the VMware guest running self-hosted-image can't get IP
address automatically.

(From OE-Core rev: 3fa9a7e4dc9eadce22a9838d2067d5af0111b04b)

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mini-x-session: try to use a resolution that's big enough for hob2's window

The patch was tested on VMware Player 4.0.1.

(From OE-Core rev: d33cc3be420fb509d8dddb1f1a752db6d41f50f9)

Signed-off-by: Dexuan Cui <dexuan.cui@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc: Excluding the DATE/SRCDATE from do_patch dependency.

Corrected the error introduced by commit afb400e of dropping
DATE/SRCDATE out of the vardepsexlude list of do_patch.

(From OE-Core rev: 6546096d068b51cc9cf4557cdb35aee6f62838d5)

Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

rpcbind: add DEPENDS on quota for rpcsvc/rquota.h

(From OE-Core rev: 4771f4af5926d724958ba7cf46b937c53babfc7c)

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

u-boot-fw-utils: Add DEPEND on mtd-utils

(From OE-Core rev: 47d8444edb644a92d3bae524c8a2a42a1cdf740c)

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

zlib: remove ldconfig call from install-libs

* it's called from install-libs target and when /etc/ld.so.cache is writeable by user running bitbake
  then it creates invalid cache (in my case libstdc++.so cannot be found after building zlib(-native)
  and I have to call touch */libstdc++.so && /sbin/ldconfig to fix it

(From OE-Core rev: b5c8add7f3ed58451cb460a242b4edd671ba618e)

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

recipes: bump PR to rebuild .la files without libz.la

(From OE-Core rev: 1da6a2dbd5e7aeea6cd45ca05590bdd50b67bf89)

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

scripts/combo-layer: allow updating a specific component(s) only

If you specify one or more components to update immediately following
the "update" command, only these components will be updated as opposed
to the default behaviour of updating all of them.

(From OE-Core rev: 1779afce024b31edcf846329da705ed5aa9fe63e)

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

conf/bitbake.conf: remove Opie definitions

These values are now set via palmtop-defs.bbclass in meta-opie (and are
not needed for anything else.)

(From OE-Core rev: a18de4dc0b988c79964354496a6a93c7ee5a28f6)

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

classes/sdl: remove Opie-specific desktop file/packaging

This code will not work without palmtopdir being defined, and is better
kept in meta-opie in any case (where we already have the OE-Classic
version of this class).

(From OE-Core rev: e45b0672e28d19cfb66c9e57e153ef15e004dcaf)

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

stat: remove NATIVE_INSTALL_WORKS

This variable is no longer used in OE-Core.

(From OE-Core rev: b70d100047912569d6eb060eb48a753c56968d89)

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

scripts/combo-layer: add branch option to example config

Minor improvements to the example combo-layer config file:
* Add the recently added branch option to the optional options section
* Mention in comments that last_revision gets updated during "update"
* Tidy up some more grammar

(From OE-Core rev: 065798d8dd0552934175f794f7678e0dd24ef152)

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

scripts/combo-layer: avoid saving last revision if unchanged

If we are running an update and the last revision hasn't changed since
the last update, don't write to the configuration file. This avoids
committing the config file with no changes other than spontaneous
reordering of sections, which sometimes occurs due to the behaviour of
the internal dictionary in Python's ConfigParser class. (This can be
fixed properly but the fix is only easy in Python 2.7+ due to the
availability there of the collections.OrderedDict class, and we
currently want to be compatible with 2.6.x as well.)

(From OE-Core rev: 4592f238de3fe6b0384c334774be9fcfc0985e4f)

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

scripts/combo-layer: avoid error when config file is outside repo

Avoid displaying the error from the "git status" command we use to check
the status of the config file if the config file is outside of the
repository (a situation that is already handled).

(From OE-Core rev: 9d3e54057673348e85ba479705c4ffaad56973d6)

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

scripts/bitbake: pass through debug options to initial pseudo build

If you specify -D or -v options to the bitbake wrapper script, and the
initial pseudo build needs to be done, it's useful to pass these
options through to that command so that you can debug the initial
pseudo build as well.

(From OE-Core rev: dcf6f72c35f7c108e447d01c15df86807e26a88b)

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

guile: fix cross configure failure

[YOCTO #1978]
Fixing the cross configure failure by not searching libltdl and libgmp
in includedir and libdir.

(From OE-Core rev: e5ad03093dfc4364d1407183f458df79f347c7a1)

Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

Revert the switch to futures for now

Without it, we get random hangs on parse failure. With it, some folks have
seen hangs even on successful cases. The former is clearly less problematic.
This is temporary, until I can finish investigating the root causes of both
issues.

(Bitbake rev: db689a99beffea1a285cdfc74a58fe73f1666987)

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

siggen.py: Handle cases where the input to symmetric_difference() may be None

[YOCTO #1967]

(Bitbake rev: f6cee29bf853874876923a7d6ae8574e3801ae8d)

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

valgrind: Fixed Added-support-for-PPC-instructions-mfatbu-mfatbl.patch Upstream-Status

(From OE-Core rev: c8f821484939348dfa63e11c5c9c6571708553ab)

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

gcc: Ensure same sstate hashes for shared soruce directory

Ensure do_patch have the same sstate hashes for nativesdk variants of
gcc family recipes, since they share the same source directory.

(From OE-Core rev: afb400e9c3ddac604906bde44df5da29a6c134d9)

Signed-off-by: Lianhao Lu <lianhao.lu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

gcc-4.6: Some small performance tweaks

a) Use the system zlib to avoid building it for each gcc.
b) -initial and -intermediate don't need plugins or lto-plugin
   so disable them.

(From OE-Core rev: a124a0883507caf4ffb7e25be666f6781b30abdd)

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

package.bbclass: Optimise the per file rpm handling

Currently a process was being forked off for each individual file
this class wanted to inspect with rpmdeps. This converts it to use
rpmdeps-oecore which allows batch processing of these dependencies.

For do_package for perl, this reduced the time by about 1 minute (33%).

(From OE-Core rev: 548037acd63bd4859f8de8d23a3d12f36ce9f97f)

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

sstate.bbclass: Optimise sstate_hardcode_path

The sstate_hardcode_path() function triggered large numbers of exec()
calls when processing packages with large numbers of file relocations
(e.g. perl). This patch optimises those calls into longer single commands
which make the code significantly more efficient.

This reduced the do_package time for perl by 2 minutes (from 4.75 minutes)
for me.

(From OE-Core rev: 4159ff761cf29a03aeb56c7686a4e7af02b52219)

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

gnupg: Add missing DEPENDS on libgcrypt

Found during single recipe rebuilds

(From OE-Core rev: bb672b1591197966b79fbcf7753f995b8e829e65)

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

sat-solver: Fix build with zlib 1.2.6

(From OE-Core rev: 6064f6bd7c28640db59a216aa578f3cca6219b1f)

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

zlib: Upgrade 1.2.5 -> 1.2.6

Dont use autotools, it really not so autoconf like.
the configure script gets updated with every release of zlib
and we overwrite that. Instead use the upstream provided
configure

copyright year was changed in zlib.h which caused change in
LIC_FILE_CHECKSUM

fix.inverted.LFS.logic.patch is already applied upstream so drop it

Drop the configure.ac and Makefile.am scripts since we do not
autoreconf anymore and do not inherit autotools anymore

Bump PR for depending recipes so a rebuild it ensues so that
they dont depend on .la anymore
and add missing dependencies discovered during incremental
build

(From OE-Core rev: 50ad5230ea9e0982cdfda23fb9fcfccf89d28f29)

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

distro-tracking-fields: update clutter, clutter_git and mx

(From OE-Core rev: e303aeec9a97903479baad836671591b511b3e19)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

clutter: upgrade git recipe to reflect 1.8.4 tag

(From OE-Core rev: 37b47a8f038947c6725dec10845412df7f67e9e4)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mx: ensure gettext and dbus are present

Upstream informs me that D-bus isn't really optional as it's required
to make some of the API work such as the single instance MxApplication.
Therefore ensure we're building with it enabled.

Gettext macros are used in the configure.ac so inherit gettext to
ensure the required dependencies are present.

(From OE-Core rev: fe2c13684e2a0b652007e92a6f11942c6366483e)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

mx: update to 1.4.2

Patch in a copy of introspection.m4 so that we can continue to easily and
cleanly disable introspection.

(From OE-Core rev: 84d62181f80402fa461675e03b20ba08da1991e8)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

clutter-1.8: update to clutter-1.8.4

Added a note about why we're setting S, so I don't have to figure it out again
next time I try and update clutter.

(From OE-Core rev: b752e2d6ae286db5588547fc49f1aa64e112fa32)

Signed-off-by: Joshua Lock <josh@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

powerpc64-linux: Update cached autoconf values for libpcap on powerpc64

Without this fix, libpcap will fail to build on powerpc64

(From OE-Core rev: 01a28e8b04c206e54bcb01ccd7401c08d5c600ce)

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

u-boot-fw-utils: Add for v2011.06

This provides the 'fw_setenv' and 'fw_getenv' programs for the target
so that the U-Boot environment can be modified from Linux.  These
programs are system-agnostic and rely on a config file that a given BSP
would provide and potentially RRECOMMEND this be installed.

(From OE-Core rev: 4a2543c0641f1df6d74c18cfe0c3ac3cfd50e60c)

Signed-off-by: Tom Rini <trini@ti.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

distutils.bbclass: override LDSHARED so we use the linker for this build and not the one used in sstate-cache

Without this fix, when packages are being built using distutils and
the python packages were deployed from sstate-cache is it possible
that the LD command will contain an invalid sysroot override.

We can fix this by always exported LDSHARED, which is the env var
that distutil looks for to override creating shared libraries.

(From OE-Core rev: 3f6b859a29ba7f570b9dae3b5bb7ab4bd7b8cee4)

Signed-off-by: Matthew McClintock <msm@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>

cross.bbclass: add virtclass handler

Allow use of BBCLASSEXTEND with 'cross' and use of virtclass-cross in
recipes.

(From OE-Core rev: 9624a817bb7d2d9fce10704b1c1a047f93260f36)

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

useradd.bbclass: skip processing on virtclass-cross extended packages

(From OE-Core rev: 4308acbbd43e6b8b37123d95df6675233007dae4)

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

external-csl-toolchain: fix mtd-user.h issue

The use of __packed, which isn't defined in userspace, caused busybox build
failures.

(From OE-Core rev: 9c44bb6434a1d2347ebbb38bde9671688b0ae447)

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

bitbake-layers: add copyright notice

This ought to have been added earlier. (I consulted with Chris Larson on
the notice covering his work.)

(Bitbake rev: 69ed02b645bdc28c1bf24e436beeed37d226b56b)

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

tcmode-default: add preference for gdb/gdbserver

(From OE-Core rev: 4ef5e70f531f48cef90805402c16ec02ad3f2b92)

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

classes/buildhistory: sort list fields in package info

Sort DEPENDS, RDEPENDS, and RRECOMMENDS in package info files so that
any changes in order (which are not important) are smoothed out in the
change history.

Fixes [YOCTO #1961]

(From OE-Core rev: 06b740d4ca077fb4c89ee6d1065fabb02da45ec6)

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

classes/buildhistory: sort image file list

Sort the contents of files-in-image.txt to avoid unnecessary changes
showing up in the history due to reordering.

(From OE-Core rev: 791ad7a522f6800114febd1605072202c1450356)

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

external-csl-toolchain: split out gdbserver packages

(From OE-Core rev: b6ddd434c47c66fdd6dd27a31e545f46b52bc90f)

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

external-csl-toolchain: correct binary packaging

CS puts optimized executables in usr/lib/bin, not usr/bin.

(From OE-Core rev: 25d22bfb0970f7152b678796f2282dced7230fe1)

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

external-csl-toolchain: set PKG/PKGV for staticdev

(From OE-Core rev: 481f6f32cd414ff11f154adc24f1b596495286df)

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

task-core-tools-profile: fix valgrind for arm and systemtap for mips

(From OE-Core rev: e5f41c2213568c7aef24ce8f666042492054377d)

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

mirrors.bbclass: Add yp.org and oe.org to fallback mirrors

This should improve the fetching errors users see
when upstream src_uri is unavailable

(From OE-Core rev: ca8a3422eb96ea431c322532dfd1be9980d4d48c)

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

uclibc: Upgrade recipes from 0.9.32 -> 0.9.33

Prefer 0.9.33 by default
Delete recipes for 0.9.32

(From OE-Core rev: 5d2c0b051072b3cee510dcad8d20fb26d3843f62)

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

dhcp: move dhcp leases files handling to postinst/postrm

Both server and client needs access to leases files so its creation
and removal are now handled by postinst/postrm scripts.

(From OE-Core rev: c8a9ae7c047051958b8667e77288c5ac295766c6)

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

gcc-cross-testing: Fix evaluation of user and target name

Dont use -q to grep we pipe to /dev/null anyway all we
care is the return status of grep

(From OE-Core rev: 9c575b5f37ee53eaf26bfdc468d4cbb661db703f)

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

classes,recipes: Replace POKY_* with CORE_IMAGE_*

(From OE-Core rev: b14a92f4450bb08b9f11d3fd7a99155315527cdd)

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

pseudo: Wrap renameat and opendir

(From OE-Core rev: f6056cf0e7c76f2f3df650b088ce84df41ec14ca)

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

mtools-3.9.9: Use correct URI for the mirror

(From OE-Core rev: b4206046476dc6ff758bb45cf0c97ab2c57893b6)

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

libx11: Backport _XGetRequest API

This API is relatively new and will be in next release of libX11
this is needed for OE qemu-native which links to libx11-native
but uses libGL from build system and herein lies the problem
if you have bleeding edge distro e.g. ubuntu 12.04(development)
libGL.so expects this API to be present. Backported to 1.4.4
thusly

(From OE-Core rev: e08604800fe1cb8f240f53c147ceb4ee08a29b91)

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

libdrm: Fix build failure unearthed by uclibc

uclibc wants to know value of __USE_GNU for certain
header prototypes to be enabled. So we move the config.h
header up in search list.

(From OE-Core rev: 086679cfbc30dba4af2f650984084195e5171b8f)

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

python: Fix build failure of python-elementtree on x86_64

The problem is due to mixing of headers from host system
thusly corrected in setup.py by checkinng if we are cross
compiling

(From OE-Core rev: 25016d158af927e0c9e70c3df4674f53fa54fec8)

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

gconf,consolekit: Avoid polkit when compiling with uclibc

polkit uses netdb functionality from glibc which is not
available in uclibc therefore we avoid polkit when compiling
for uclibc systems.

(From OE-Core rev: 7eae02f47a08587cf5be3d39df0184cf346cbaea)

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

runqemu-internal: qemuarm can handle 256M

Raise the check for max RAM supported for QEMU/arm
its 256M since we patched it to be so in OE

(From OE-Core rev: cfd84a05581e068d705b0804829c4833a06bffa6)

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

Enable options needed for ltp to compile

(From OE-Core rev: 22309029a9599b8f947c1fca664760123c325885)

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

task-core-tools: Divide it into 3 recipes

task-core-tools-debug, task-core-tools-profile, task-core-tools-testapps
otherwise if we choose one through PACKAGE_GROUPS all packages
are built since they are in same recipe.

(From OE-Core rev: 180227d8ff71c42e0d16ad7eb4a5f9b7d6a0c1c6)

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

linux-yocto: fri2: use emgd-1.10

Updating meta for the following change:

[
    fri2: use emgd-1.10

    Use the yocto/emgd-1.10 branch in place of yocto/emgd.
]

(From OE-Core rev: aa6293ffbc7f8f17e55f2105de56384e14591b44)

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

kern-tools: remove explicit 'yocto' references from auto-bsp handling

To adapt to different branch names in a repository, removing explicit
references to 'yocto' in automatically created BSPs ensures that they
are consistent with the merged variants. Existing BSPs and auto BSPs
are not impacted by this change.

(From OE-Core rev: 74613d02aa844cac3be0626cc64139643fe17a4f)

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

linux-yocto: rt compiliation fix

Updating the meta SRCREC to pickup the following changes:

  70e86dc meta/rt: update rt patch cache

Without this change a 3.0 based -rt build will fail to patch.

(From OE-Core rev: e5da5e84e64362f59a62f65b9f260f96af96bc11)

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

linux-yocto: meta updates for sys940x and cleanup

Updating the meta SRCREV to pickup the following changes:

[
  Author: Darren Hart <dvhart@linux.intel.com>
  Date:   Wed Feb 1 09:37:48 2012 -0800

    Remove boot-live cfg duplication from BSPs

    Several BSPs duplicated the boot-live fragment in their BSP
    specific config. Remove the duplication and add CONFIG_RD_GZIP
    and CONFIG_BLK_DEV_SR to the boot-live fragment.

    (From OE-Core rev: a243dded5c84d48728308a769e527302b7539939)

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Acked-by: Tom Zanussi <tzanussi@intel.com>
  Author: Darren Hart <dvhart@linux.intel.com>
  Date:   Tue Jan 31 13:18:17 2012 -0800

    meta: sys940x BSP meta data

    The Inforce SYS940x-ECX Developer-Ready Reference Platform features:
    o Intel Atom E6xx (0.6-1.6 GHz)
    o Up to 1GB on-board DDR2
    o Intel Platform Controller Hub EG20T
    o VGA,LVDS
    o HD Audio
    o SD Card
    o Dual SATA
    o Mini-PCIe

    http://www.inforcecomputing.com/SYS940X_ECX.html

Signed-off-by: Darren Hart <dvhart@linux.intel.com>
Acked-by: Tom Zanussi <tzanussi@intel.com>
Signed-off-by: Darren Hart <dvhart@linux.intel.com>
]

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

linux-yocto: locate and use out of tree features specified on the SRC_URI

In a similar manner to calling the patch.bbclass to locate patches that
were listed on the SRC_URI, it is also useful to query about 'other' items
that are on the SRC_URI. In the case of linux-yocto, it allows us to
know about kernel features that were specific on the URI and then apply
them to the current tree.

(From OE-Core rev: 3e2ce87566124db8c78472f9a4f00ab26410f213)

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