scm/bb/tizen-distro.git
10 years agobitbake: toaster: Update API
Belen Barros Pena [Fri, 28 Mar 2014 19:28:45 +0000 (19:28 +0000)]
bitbake: toaster: Update API

Make sure the REST API includes the latest changes
to the database schema.

(Bitbake rev: fb3d1d189f010488a8726872b01313857697751b)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: link task order to right tasks page
David Reyna [Sat, 29 Mar 2014 14:47:43 +0000 (07:47 -0700)]
bitbake: toaster: link task order to right tasks page

When linking from a task order number to the All Tasks
page, automatically display the correct page for that link
anchor.

[YOCTO #5933]

(Bitbake rev: e5b6681432b627eb73bfa766d3154162206f374e)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agositeconfig: Unbreak after sstate changes some time ago
Richard Purdie [Fri, 4 Apr 2014 17:12:58 +0000 (18:12 +0100)]
siteconfig: Unbreak after sstate changes some time ago

Some time ago, the sstate name field was dropped. This code wouldn't have
worked since then. Makes me wonder if we really need it.

Anyhow, my last patch properly breaks it. This fixes the naming so
it works as designed again.

(From OE-Core rev: d282b276aa5e58d306be1c8ef9a985f2267a612b)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agosstate: Remove name sstate parameter
Richard Purdie [Fri, 4 Apr 2014 17:06:32 +0000 (18:06 +0100)]
sstate: Remove name sstate parameter

A while ago we stopped supporting sstate names which were different from the
taskname. This patch finishes cleaning up some code remnants from that
which were causing data duplication and confusion.

(From OE-Core rev: 5a19863e389d28d8db4a86e409c1daa3c6b46eff)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agosstate: Fix an error handling the taskname
Richard Purdie [Fri, 4 Apr 2014 17:05:43 +0000 (18:05 +0100)]
sstate: Fix an error handling the taskname

Looking at the code, its clear 'task' is meant not to have the do_ prefix,
however its also clear it can be left in through some code paths.

One result of this can be files not being cleaned from the sysroot correctly.

Fix this.

(From OE-Core rev: 3c912102fd9ea6b360f1af209da21e8f27b845c0)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoREADME.hardware: tweak edgerouter instructions
Paul Eggleton [Fri, 4 Apr 2014 14:26:26 +0000 (15:26 +0100)]
README.hardware: tweak edgerouter instructions

* You need a rollover cable for the serial port
* Some minor grammar / typo / formatting fixes

(From meta-yocto rev: 79f01604e672ce0eb7af787c27ec561063222796)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoREADME.hardware: drop routerstationpro in favour of edgerouter
Paul Eggleton [Fri, 4 Apr 2014 14:26:25 +0000 (15:26 +0100)]
README.hardware: drop routerstationpro in favour of edgerouter

(From meta-yocto rev: c7f9076e49178e0e0160710e4ca645ae2d73152e)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: cooker: Overwrite IMAGE_BASENAME to default in custom image
Marius Avram [Tue, 1 Apr 2014 10:59:05 +0000 (13:59 +0300)]
bitbake: cooker: Overwrite IMAGE_BASENAME to default in custom image

This solves a problem of custom images which inherit a base
image with IMAGE_BASENAME overwritten in their recipe by a
different value than its default one: ${PN}.

The value of IMAGE_BASE causes a crash when hob will try to
create symbolic links to the resulting images from the deploy
directory, because it will look for names similar to
<original_recipe_name>-edited-timestamp-machine.rootfs.*
which might be different from the actual resulting image.

The solution is to simply overwrite IMAGE_BASENAME in the
custom recipe to the default value in the case IMAGE_BASENAME
is found in the base recipe.

Some recipes which were affected by this problem are those
from meta-fsl-demos (e.g.: fsl-image-test).

[YOCTO #6017]

(Bitbake rev: e42ee93519000f827be49659b6b5fb7717b3d592)

Signed-off-by: Marius Avram <marius.avram@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: hob: fix customization of empty image recipe
Marius Avram [Thu, 3 Apr 2014 12:15:59 +0000 (15:15 +0300)]
bitbake: hob: fix customization of empty image recipe

This fixes a problem which happened when you tried to build an
image by selecting 'Start with an empty image recipe' from the
Image configuration page of hob.

The reason on of the bug was that the name of the resulting image
was threated the same way as ordinary custom images, when in fact
they should use the default name: "hob-image", because they do not
derive from any other recipe.

[YOCTO #6102]

(Bitbake rev: fa4ea3b4b40e7e9e6767e0cd51c6701e0af07135)

Signed-off-by: Marius Avram <marius.avram@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: hob: adding INHERIT += " testimage " affects image recipes list
Cristiana Voicu [Fri, 4 Apr 2014 10:53:40 +0000 (13:53 +0300)]
bitbake: hob: adding INHERIT += " testimage " affects image recipes list

To find if a recipe is for an image, Hob checks if it inherits
image.bbclass. But when you add testimage in local.conf, this will
be added for each recipe, and it pass the test. Adding a "/" before
"image.bbclass", will check only for image.bbclass.

[YOCTO #6117]
(Bitbake rev: be8511c9d474c570f6ca7078e28919c8a5175a42)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agogcc: enable multilib setup for powerpc64 arch
Alexandru-Cezar Sardan [Fri, 4 Apr 2014 11:06:54 +0000 (14:06 +0300)]
gcc: enable multilib setup for powerpc64 arch

(From OE-Core rev: 99fadf2ff9f4379121fde6477383e7574519fba4)

Signed-off-by: Alexandru-Cezar Sardan <alexandru.sardan@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoutil-linux: fix parallel build issue
Robert Yang [Fri, 4 Apr 2014 10:44:09 +0000 (18:44 +0800)]
util-linux: fix parallel build issue

The rule SETARCH_MAN_LINKS is used for the files under the sys-utils
dir, for example:

echo ".so man8/setarch.8" > sys-utils/linux32.8

but it depends on nothing so that the sys-utils dir may not exist, we
can create the sys-utils dir to fix problem.

[YOCTO #6115]

(From OE-Core rev: 4c46bfd703409bd55a781742e4afedf88da1124b)

Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoSDK default deploy directory
David Vincent [Fri, 4 Apr 2014 07:47:42 +0000 (09:47 +0200)]
SDK default deploy directory

If the user changes the DEPLOY_DIR variable to set up a custom deploy
directory for images, packages, SDKs as explained in the documentation,
the variable SDK_DEPLOY does not take it into account and fallback to
TMPDIR as default. Therefore, SDKs were not found in the correct
location.

(From OE-Core rev: 0f043e9a3fdc4b489b55e0605fee01927854205a)

Signed-off-by: David Vincent <freesilicon@gmail.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: fix mis-detection of targets as images
Paul Eggleton [Thu, 3 Apr 2014 10:16:24 +0000 (11:16 +0100)]
bitbake: toaster: fix mis-detection of targets as images

If you are using the testimage class then the old test mistook the
global inheritance of testimage.bbclass with the recipe inheriting
image.bbclass because it was only looking for that at the end of the
string.

Also tidy up the code so you an easily tell what it's doing. (The
original method may have been more "pythonic", but it does nothing for
readability.)

(Bitbake rev: b05e741cb5fe44b37538f2b727782f80dc9bb8fa)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: fix help texts not showing for most tasks
Paul Eggleton [Thu, 3 Apr 2014 10:16:23 +0000 (11:16 +0100)]
bitbake: toaster: fix help texts not showing for most tasks

These were not being collected properly because we were explicitly
excluding variables defined as functions from being stored in the
database. We don't want these to be shown in the variables list, and in
any case it makes sense for these to be stored elsewhere, so create a
separate model to store these.

Fixes [YOCTO #6050].

(Bitbake rev: 0d76a5461ce4bd554ff70a465064969e53edf0a4)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agogtk+3: set proper FLAGS for native
Robert Yang [Fri, 28 Mar 2014 10:03:59 +0000 (10:03 +0000)]
gtk+3: set proper FLAGS for native

The target gtk+3 does a native build in its "native" directory, we need
unset the target FLAGS for native build, otherwise, there might be build
failures.

(From OE-Core rev: 3f3e249b98182fa50358ade9278e3e26454275b8)

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>
10 years agocoreutils: fix search paths for libstdbuf.so
Chunrong Guo [Fri, 28 Mar 2014 08:47:05 +0000 (08:47 +0000)]
coreutils: fix search paths for libstdbuf.so

       *fix the following error:
        |stdbuf: failed to find 'libstdbuf.so'
        *PKGLIBEXECDIR is the search paths for libstdbuf.so
        |PKGLIBEXECDIR='$(pkglibexecdir)'
        |pkglibexecdir='${libexecdir}/${PACKAGE}'

(From OE-Core rev: a0745234e78d9161d407f2157dc494fed4487d42)

Signed-off-by: Chunrong Guo <B40290-KZfg59tc24xl57MIdRCFDg@public.gmane.org>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolibarchive: Add PACKAGECONFIG for lzo
Paul Barker [Wed, 2 Apr 2014 12:54:49 +0000 (12:54 +0000)]
libarchive: Add PACKAGECONFIG for lzo

This ensures that the dependency on lzo is deterministic rather than floating.

The configure option to libarchive refers to this library as 'lzo2' but it is
just called 'lzo' in OpenEmbedded.

(From OE-Core rev: 09d729a21a2404095279c717c88ac494e2e716d6)

Signed-off-by: Paul Barker <paul@paulbarker.me.uk>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoselftest/sstatetests.py: Added 'populate_lic' to ignore_patterns
Corneliu Stoicescu [Thu, 3 Apr 2014 10:26:07 +0000 (10:26 +0000)]
selftest/sstatetests.py: Added 'populate_lic' to ignore_patterns

Because the sstate-cache-management script does not remove populate_lic
sstate files, we should ignore them when checking for removed files.

(From OE-Core rev: 5debc2af6672841c126cec5d747e2e3c6407c8be)

Signed-off-by: Corneliu Stoicescu <corneliux.stoicescu@intel.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agopython: fix build error with Readline 6.3
Chong Lu [Thu, 3 Apr 2014 07:52:22 +0000 (07:52 +0000)]
python: fix build error with Readline 6.3

Backport two patches from upstream:

use new readline function types (closes #20374)
Issue #20374: Avoid compiler warnings when compiling readline with libedit.

[YOCTO #6107]

(From OE-Core rev: a6b91ae7dec2edebc0eaea0592c42b1c455ad4d7)

Signed-off-by: Chong Lu <Chong.Lu@windriver.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agorootfs.py: introduce USE_DEVFS check
Matthieu Crapet [Thu, 3 Apr 2014 07:05:50 +0000 (07:05 +0000)]
rootfs.py: introduce USE_DEVFS check

Since commit a83144bac8d67704ff66f5dc0fc56f5b63979694 (2014-02-11), USE_DEVFS is not considered anymore.
For compatibility, let's restore USE_DEVFS semantic.
Also add USE_DEVFS to documentation.conf.

(From OE-Core rev: d12a5e38a02abe3feb3db8ae5ffd9a5005124294)

Signed-off-by: Matthieu Crapet <Matthieu.Crapet@ingenico.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agocmake: Remove dependency on ncurses
Mike Crowe [Wed, 2 Apr 2014 16:04:12 +0000 (16:04 +0000)]
cmake: Remove dependency on ncurses

Commit 2adc9a3f1f1db284f7d91193ad77b3524e0e0d2c stopped ccmake being built
and that is the only part of cmake that relies on curses so we might as
well stop depending on it.

(Tested with a poisoned curses.h to prove that it is unused even if
present.)

(From OE-Core rev: 4f650b538924b1736783fec0de661ec16dc590b5)

Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agocmake-native: Stop building ccmake
Mike Crowe [Wed, 2 Apr 2014 16:04:11 +0000 (16:04 +0000)]
cmake-native: Stop building ccmake

The bootstrapped cmake is called in such a way that it will automatically
enable building ccmake if curses is found. This tool isn't particularly
useful to us and it will cause build problems if ncurses-native is built in
parallel with cmake-native so let's just pass -DBUILD_CursesDialog=0 to
disable the feature altogether as the non-native cmake does.

Unfortunately this requires patching the bootstrap script since there
appears to be no way to get this option through.

(From OE-Core rev: 4bee0a93ed985b38c6b4eb605d8e16f5d7c82d51)

Signed-off-by: Mike Crowe <mac@mcrowe.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoRevert "qt-mobility: remove /usr/lib from ld rpath-link option"
Martin Jansa [Wed, 2 Apr 2014 12:25:36 +0000 (12:25 +0000)]
Revert "qt-mobility: remove /usr/lib from ld rpath-link option"

* Basically part of the headers/libs are installed in
  ${D}(${libdir}/${includedir}) instead of
  ${D}(${libdir}/${includedir})/qt4

* http://lists.openembedded.org/pipermail/openembedded-core/2013-October/085815.html
  is related, but doesn't fix the issue completely, so better
  revert the commit which introduced this issue

[YOCTO #5414]

This reverts commit f7409a9fe83ba2535a43f39ed57cd78242a88557.

(From OE-Core rev: 2b6a1cf9c7e05d204df82aa962d3ae09744dd86a)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agosend-error-report: show response
Martin Jansa [Wed, 2 Apr 2014 12:25:37 +0000 (12:25 +0000)]
send-error-report: show response

* useful when debuging why it was refused by server

(From OE-Core rev: 0b8ff2231a36755a71d8bf8c7854364d69ef2df8)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agogcc-target: remove infodir
Martin Jansa [Wed, 2 Apr 2014 12:25:35 +0000 (12:25 +0000)]
gcc-target: remove infodir

* it uses autotools but doesn't call autotools_do_install
* fixes QA warning:
  gcc-4.8.2: The /usr/share/info/dir file is not meant to be shipped in a particular package.

(From OE-Core rev: 0b62f8b012d8a128f34bd85a26aa97ea939cbf64)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoed: remove infodir
Martin Jansa [Wed, 2 Apr 2014 12:25:34 +0000 (12:25 +0000)]
ed: remove infodir

* unlike 0.5 version, this doesn't use autotools.bbclass which removes infodir automatically
* fixes QA warning:
  ed-1.9: The /usr/share/info/dir file is not meant to be shipped in a particular package.

(From OE-Core rev: 205621e3b9358a455e73122941fbbdcde9b2f2a3)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agocmake: Add ugly hack from meta-qt5 to prevent cmake trying to detect qt5
Martin Jansa [Wed, 2 Apr 2014 12:25:33 +0000 (12:25 +0000)]
cmake: Add ugly hack from meta-qt5 to prevent cmake trying to detect qt5

* cmake doesn't have dependency on qt4/qt5, so these tests usually fail
  but still can cause undeterministic results or build failures (when
  OE_QMAKE_PATH_EXTERNAL_HOST_BINS is undefined or native qmake removed
  while running the test in cmake)

(From OE-Core rev: 4f2dee0d8c2c3f60cb8f8021343e973cae31fc50)

Signed-off-by: Martin Jansa <Martin.Jansa@gmail.com>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agosanity: testimage needs DISPLAY set only for qemu targets
Stefan Stanacar [Wed, 2 Apr 2014 10:02:31 +0000 (10:02 +0000)]
sanity: testimage needs DISPLAY set only for qemu targets

There's no point in failing if DISPLAY isn't set if we don't
boot a qemu image when using a controller like SimpleRemoteTarget
or GummibootTarget.

(From OE-Core rev: 9fcd3af8626e1b0979b0cde745fe0880ccc50de7)

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>
10 years agosystemd: backport patch to avoid assertion failures
Jonathan Liu [Wed, 2 Apr 2014 02:50:10 +0000 (02:50 +0000)]
systemd: backport patch to avoid assertion failures

(From OE-Core rev: 6c44d8133a205ca78bde33ebadcc18ac9a64236d)

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>
10 years agodbus: backport memory leak patch for error when listing services
Jonathan Liu [Wed, 2 Apr 2014 02:50:09 +0000 (02:50 +0000)]
dbus: backport memory leak patch for error when listing services

(From OE-Core rev: a09a4e60a454b60c6d6e1e28fb4dda88fce9ae7d)

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>
10 years agodbus: backport fix for bus activation under systemd session
Jonathan Liu [Wed, 2 Apr 2014 02:50:08 +0000 (02:50 +0000)]
dbus: backport fix for bus activation under systemd session

(From OE-Core rev: 8f191a37625ba6fe2e109555fa99658c4fcd7a37)

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>
10 years agoalsa-tools: Fix build without x11
Otavio Salvador [Tue, 1 Apr 2014 16:42:04 +0000 (16:42 +0000)]
alsa-tools: Fix build without x11

The patch had some new tool (hdajackretask) missing which were
triggering build failures in O.S. Systems' builder as:

,----[ Build error in a clean tmp, without x11 ]
| checking for GTK3... no
| configure: error: Package requirements (gtk+-3.0) were not met:
|
| No package 'gtk+-3.0' found
|
| Consider adjusting the PKG_CONFIG_PATH environment variable if you
| installed software in a non-standard prefix.
|
| Alternatively, you may set the environment variables GTK3_CFLAGS
| and GTK3_LIBS to avoid the need to call pkg-config.
| See the pkg-config man page for more details.
| make: *** [all] Error 1
`----

(From OE-Core rev: be37edbdfe1e1ad833155154cabdff16f1d4267f)

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: user-manual-metadata.xml: Added new section on mapping functions.
Scott Rifenbark [Wed, 2 Apr 2014 12:26:02 +0000 (12:26 +0000)]
bitbake: user-manual-metadata.xml: Added new section on mapping functions.

Fixes [YOCTO #5472]
Fixes [YOCTO #1973]

Created a new section called "Automatically Mapping Functions
Within the Context of a Class".  This section addresses the
EXPORT_FUNCTIONS "operator", which was the last of the adjustments
to the variables bug against the BB manual (1973).  The related
bug (5472) is a general bug against enhancing the BB manual, which
this change caps off.

The section here was reviewed and approved by Richard Purdie.

(Bitbake rev: cec33d4fdc05db3a41e978f3a1ab977730c443eb)

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoMake ppce300c3 tune hard-float by default
Mats Kärrman [Tue, 1 Apr 2014 16:20:34 +0000 (16:20 +0000)]
Make ppce300c3 tune hard-float by default

The tuning file for PowerPC e300c3 is soft-float. In OE-classic it was hard-
float and it should be as the c3 has an fpu. I have modified the tuning file
to include both a hard-float version (using the existing ppce300c3 name) and
an optional soft-float version (called ppce300c3-nf).

The following patch also passes a "--with-cpu=e300c3" argument to GLIBC.
For this to have any effect the sqrt/sqrtf implementations added by the
"glibc.fix_sqrt2.patch" are required and also an additional "Implies" file
(added to the mentioned patch as a separate patch for eglibc_2.19).

Tested with eglibc 2.19 on PowerPC MPC5125.

(From OE-Core rev: 9d502ca8551fd461f869395b1b7e62d6dcf59a84)

Signed-off-by: Mats Karrman <mats.karrman@tritech.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoeglibc_2.19: Make ppc e300c3 benefit from 603e sqrt optimizations
Mats Kärrman [Tue, 1 Apr 2014 16:20:18 +0000 (16:20 +0000)]
eglibc_2.19: Make ppc e300c3 benefit from 603e sqrt optimizations

Tested on PowerPC MPC5125.

(From OE-Core rev: 263e278be61093509110267596fdd75c225bf06a)

Signed-off-by: Mats Karrman <mats.karrman@tritech.se>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agogdk-pixbuf: pass GDK_PIXBUF_FATAL_LOADER where relevant
Ross Burton [Tue, 1 Apr 2014 16:23:37 +0000 (17:23 +0100)]
gdk-pixbuf: pass GDK_PIXBUF_FATAL_LOADER where relevant

Pass GDK_PIXBUF_FATAL_LOADER to the sstate postinst and intercept so that any
problems are flagged as errors instead of being silently ignored.

(From OE-Core rev: 0a97b1571821848af11d8651c7145ed9592f9e31)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agogdk-pixbuf: add an option so that loader errors are fatal
Ross Burton [Tue, 1 Apr 2014 16:23:36 +0000 (17:23 +0100)]
gdk-pixbuf: add an option so that loader errors are fatal

So that gdk-pixbuf-query-loader failures can be identified as such (and executed
later, or run on the target) add a magic environment variable return loader
failures from main().

(From OE-Core rev: 8995c2cbb7a08c569d3e554b65f2bc3cc1682e2a)

Signed-off-by: Ross Burton <ross.burton@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolibpam: fix multilib packaging issue for pam-plugins
Ming Liu [Fri, 28 Mar 2014 09:43:08 +0000 (17:43 +0800)]
libpam: fix multilib packaging issue for pam-plugins

libpam might miss ABI specific dependencies for pam-plugins-*, for RPM uses
generic names to check the packages depending on it and doesn't consider the
arch, which will lead to packaging issues in mulbilib build.

pam_plugin_hook is added because the plugin packages are dynamically
generated, so we need to manually process multilib names by add baselib to
RPROVIDES/RDEPENDS as ABI specific tag.

(From OE-Core rev: d08e64a98316d7659b0fb56812667c534f66a1a8)

Signed-off-by: Ming Liu <ming.liu@windriver.com>
I worked with Ming Liu on this particular issue.  You may wonder why
this is necessary let me attempt to explain the underlying causes.

In deb/ipk on a multilib package, the package name has specific multilib
references in it.  I.e. the alternative libraries start with something
like lib32-...  This was done primarily because deb/ipk do not allow two
packages with the same name (but different architectures) to be
installed at the same time.  So the name has to be unique.

In RPM however, the names of the packages and matches with the
architectures and if they are not the same we can do these multilib
installs.  This matches the behavior of other RPM based distributions
and in many ways the tools people are used to working with RPM.  For the
most part this works fine in multilib configurations because additional
per-file dependencies are added that capture the shared library
dependencies with ABI specific information.  This unfortunately fails in
a few cases where plugins are dynamically loaded via dlopen -- such as
libpam.

One possible fix is simply to follow the deb/ipk package naming, but
this causes a design advantage of rpm.  When a package has a dependency
on 'bash', we really don't care what bash is installed, only that -a-
bash is installed.  In the deb/ipk case, the lib32- packages would end
up with a lib32-bash dependency and you could potentially end up with
two 'bash' packages being installed.

So the fix I recommended for the issue was to add the baselib path to
the internal dependencies.  Since we know that the libpam installed in
'lib' needs the modules that were compiled to also work with the 'lib'
version of libpam. While the libpam in 'lib64' need the modules to work
with the 'lib64' version of the plugins.

Existing dependencies are preserved so there is no impact in the ipk/deb
case, the RPM case is resolved as the additional dependency information
is now present for the package manager to select the package we really
want.

If anyone else has a suggestion for an alternative fix, we're interested
-- but this is the best answer we could come up with.  (If any of the
above should be added to the commit message, the YP bug, or
documentation, please let me know and I'll make sure it gets added.)

Signed-off-by: Mark Hatle <mark.hatle@windriver.com>
[YOCTO #4532]
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agowayland-native: disable unused macro checks to fix build issue on Centos5.x
Ting Liu [Fri, 28 Mar 2014 09:42:40 +0000 (17:42 +0800)]
wayland-native: disable unused macro checks to fix build issue on Centos5.x

We only build wayland-native for the scanner, so disable the bits we
don't actually need. This avoid build issue on older distro such as
Centos 5.x:
| error: 'O_CLOEXEC' undeclared (first use in this function)
| error: sys/timerfd.h: No such file or directory
| error: 'CLOCK_MONOTONIC' undeclared (first use in this function)
| error: 'TFD_CLOEXEC' undeclared (first use in this function)
| error: 'SFD_CLOEXEC' undeclared (first use in this function)

(From OE-Core rev: f6d5343ccb97913a874b894fd7405abad59746eb)

Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agokmod: fix O_CLOEXEC not supported on old kernel
Robert Yang [Fri, 28 Mar 2014 09:42:41 +0000 (17:42 +0800)]
kmod: fix O_CLOEXEC not supported on old kernel

O_CLOEXEC is introduced from Linux 2.6.23, so old kernel doesn't have
it, we need check before use.

This patch is much more like a workaround, since it may need fcntl() use
FD_CLOEXEC to replace.

(From OE-Core rev: 44c441222002ee0177100be0431adf91984e90d5)

Signed-off-by: Ting Liu <b28495@freescale.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoopenssl: Fix pod2man des.pod error on Ubuntu 12.04
Baogen Shang [Fri, 28 Mar 2014 09:42:45 +0000 (17:42 +0800)]
openssl: Fix pod2man des.pod error on Ubuntu 12.04

This is a formatting fix, '=back' is required before
'=head1' on Ubuntu 12.04.

(From OE-Core rev: 362d20e04e64a7437d7f61761057c721066e805f)

Signed-off-by: Baogen Shang <baogen.shang@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoQemu:Arm:versatilepb: Add memory size checking
Jiang Lu [Fri, 28 Mar 2014 09:42:44 +0000 (17:42 +0800)]
Qemu:Arm:versatilepb: Add memory size checking

The machine can not work with memory over 256M, so add a checking
at startup. If the memory size exceed 256M, just stop emulation then
throw out warning about memory limitation.

(From OE-Core rev: 48ff812a4b649fa7b1c73740ef65e4855640dc39)

Signed-off-by: Jiang Lu <lu.jiang@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoqemu: Add addition environment space to boot loader qemu-system-mips
Jeff Polk [Fri, 28 Mar 2014 09:42:43 +0000 (17:42 +0800)]
qemu: Add addition environment space to boot loader qemu-system-mips

The qemu mips malta base board boot loader uses environment strings
with a max length of 256 bytes which is not long enough to accommodate
a long NFS path in addition to the normal kernel boot command line
arguments.

The solution is to expand the environment string length to 1024 bytes.

(From OE-Core rev: 1e9e26304b77dd11b30ec983b7fa058378b29c47)

Signed-off-by: Jason Wessel <jason.wessel@windriver.com>
Signed-off-by: Roy Li <rongqing.li@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agocore-image-lsb: enforce pam as a needed distro feature
Cristian Iorga [Mon, 31 Mar 2014 09:51:21 +0000 (12:51 +0300)]
core-image-lsb: enforce pam as a needed distro feature

core-image-lsb only gave a warning:
"WARNING: Building libpam but 'pam' isn't in DISTRO_FEATURES,
PAM won't work correctly"
when the proper DISTRO was not set for it.
default choice would be DISTRO = "poky-lsb",
but not necessarily, depending on each custom distro.

This fix will enforce the proper usage of pam
as a distro feature for core-image-lsb by giving
an error instead of just a warning.

Fixes [YOCTO #6073]

(From OE-Core rev: f7cd6b383f50ebc3000a9d9db8be719ab5b3c0bb)

Signed-off-by: Cristian Iorga <cristian.iorga@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoimage/image-prelink/image-mklibs/sanity: Drop pointless EXPORT_FUNCTIONS
Richard Purdie [Tue, 1 Apr 2014 10:08:58 +0000 (11:08 +0100)]
image/image-prelink/image-mklibs/sanity: Drop pointless EXPORT_FUNCTIONS

I'm sick of seeing people adding to EXPORT_FUNCTIONS in these classes
when they clearly have no idea what it does.

Worse, these uses of it are all broken, the naming is incorrect and
they do nothing. Lets remove them and try and preserve any remaining
part of my sanity.

(From OE-Core rev: 05a2fb19f722652c5d13be911b8ed45a264bbb40)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoqemux86-64: re-enable paravirt guest, bring it in sync with qemux86
Stefan Stanacar [Tue, 1 Apr 2014 10:00:30 +0000 (13:00 +0300)]
qemux86-64: re-enable paravirt guest, bring it in sync with qemux86

Seems wrong that qemux86 has this enabled and qemux86-64 doesn't.
Also this will allow people to use kvm with -cpu=host.
Right now, runqemu qemux86-64 kvm uses -cpu=kvm64 because without this
feature you can't use cpu=host on newer host kernels (>= 3.8).

This basically reverts poky e6149ec6c4c14d3a939022cdc29ac94dced836dd /
oe-core 64749308fadabb4aa7c39f360c6395827bc5eb3a
The reason of that commit (which is more than a year old) was that on
the AB running old kernels (2.6.37) we would see occasional shutdown failures.
(but if memory serves me right the fix might have been just a coincidence).

I've tested this change with by:
  - run tests on runqemu qemux86-64 (without kvm) (300 seconds)
  - run tests on runqemu qemux86-64 kvm (which uses by default cpu=kvm64) - (tests
took 20 seconds)
  - run tests on runqemu qemux86-64 kvm with cpu=host - (tests took 18 seconds)

[ YOCTO #5956 ]

(From OE-Core rev: 4b09a1869895e4cd18e82b7d190fbfea3c7922af)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoxorg: Fix for CVE-2013-6424
Kai Kang [Tue, 1 Apr 2014 09:09:50 +0000 (17:09 +0800)]
xorg: Fix for CVE-2013-6424

Integer underflow in the xTrapezoidValid macro in render/picture.h in X.Org
allows context-dependent attackers to cause a denial of service (crash) via
a negative bottom value.

http://web.nvd.nist.gov/view/vuln/detail?vulnId=CVE-2013-6424

(From OE-Core rev: 059dc5f4ef9bcf49cb6520f5f2ab1e739f4d42de)

Signed-off-by: Baogen Shang <baogen.shang@windriver.com>
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoimage.bbclass: add function to disable SSH DNS Lookup for Qemu
Saul Wold [Mon, 31 Mar 2014 17:37:38 +0000 (10:37 -0700)]
image.bbclass: add function to disable SSH DNS Lookup for Qemu

This function disables the reverse DNS lookup on QEMU targets to reduce the
delay when using static IP address. By disabling DNS lookup we can save a great
deal of time during automated testing on the autobuilder (on the order of ~400
seconds per ssh tranaction). This is seen when using the testimage, there is a
delay getting logged-in from the server to target.

It's enabled for all qemu imgaes by default and can be overridden by setting
the SSH_DISABLE_DNS_LOOKUP variable.

[YOCTO #5954]

(From OE-Core rev: c93eeecb15c4acac9226a3394c93d7e99a809d6b)

Signed-off-by: Saul Wold <sgw@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agogcc: changed multilib options handling
Alexandru-Cezar Sardan [Mon, 31 Mar 2014 13:16:13 +0000 (16:16 +0300)]
gcc: changed multilib options handling

Duplicate parameters in the tune args are repeated in the
MULTILIB_OPTIONS variable. This leads to incorrect configurations
if the order of the parameters is bad.
(Eg. "mhard-float m32/mhard-float m64" leads to an incorrect config)
This patch finds the common parameters and removes the duplicates.

(From OE-Core rev: 90dc31c24adfa8e916a9c475ae1afc58ad179dfb)

Signed-off-by: Alexandru-Cezar Sardan <alexandru.sardan@freescale.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: bitbake: cooker: mark setFeatures command as read-only
Alexandru DAMIAN [Tue, 1 Apr 2014 11:24:49 +0000 (12:24 +0100)]
bitbake: bitbake: cooker: mark setFeatures command as read-only

This patch makes sure that the setFeatures command is marked
as read-only and that it can only run if the cooker is in
the initial state.

Additionally, remove logging from the XMLRPC module in favor
of sending the exception to the client for easy processing.

[YOCTO #6089]

(Bitbake rev: f0a1a3e24757f7658d272035620465f92a3e4c3c)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: secondary sort key as table's default order
David Reyna [Wed, 26 Mar 2014 23:24:43 +0000 (16:24 -0700)]
bitbake: toaster: secondary sort key as table's default order

Provide for a secondary sort key based on the table's default
ordering when doing sort on alternate columns.

[YOCTO #5920]

(Bitbake rev: 1a0defce1499fdc320bcb27b41e06bea2ca2aef2)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: unbuilt package dependency formats
Dave Lerner [Wed, 26 Mar 2014 20:49:43 +0000 (15:49 -0500)]
bitbake: toaster: unbuilt package dependency formats

[YOCTO 6057]

For a package shown on the package build dependency page, the dependent
packages may be unbuilt packages, as indicated with the dependent
package's size set to -1.  This fix changes the build template to use
the same formatting functions for unbuilt dependent packages as the
include package templates use for unbuilt dependent packages.

(Bitbake rev: b095ab30a827a50f66a06ac9170d33fae2670736)

Signed-off-by: Dave Lerner <dave.lerner@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Update help text in format_vpackage_namehelp tag
Belen Barros Pena [Mon, 17 Mar 2014 00:11:46 +0000 (00:11 +0000)]
bitbake: toaster: Update help text in format_vpackage_namehelp tag

The text now says that the package has not been built.

(Bitbake rev: 4b64f4452aa2da4dcc0491cefa6f1a07f7dd173f)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Match search results form to no results form
Belen Barros Pena [Sun, 16 Mar 2014 16:59:13 +0000 (16:59 +0000)]
bitbake: toaster: Match search results form to no results form

In the search results form, add the btn class to the clear
search button and set its tabindex to -1 so that you don't
accidentally clear the search when you want to search
again.

(Bitbake rev: 1cd01dbf3cd59bac6b62fe91ba2bafa0c62fd7f1)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Increase animation duration
Belen Barros Pena [Sun, 16 Mar 2014 14:58:33 +0000 (14:58 +0000)]
bitbake: toaster: Increase animation duration

Increase the duration of the blue highlight animation
from 7 to 10 seconds.

(Bitbake rev: 0d48cec969a68f9b70e04be6d86b078df8f1ec5b)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Fix the fade out animation
Belen Barros Pena [Fri, 14 Mar 2014 19:32:05 +0000 (19:32 +0000)]
bitbake: toaster: Fix the fade out animation

Apply the animation to any element with the class
.highlight

(Bitbake rev: b228739888cf8bac99da4aada3c040aac40f784d)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Small tweaks to the packages included interface
Belen Barros Pena [Sun, 16 Mar 2014 14:42:28 +0000 (14:42 +0000)]
bitbake: toaster: Small tweaks to the packages included interface

* Capitalise correctly the label "Reverse runtime dependencies"

* Change dependency popover labels to match the rest of the
interface

* Make sure that dependency links go to the initial tab
of the installed package details pages

(Bitbake rev: 80df010c12f7ba19649a7bbda9d788217cabc57a)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Remove trailing spaces from 'name'
Belen Barros Pena [Sun, 16 Mar 2014 14:21:55 +0000 (14:21 +0000)]
bitbake: toaster: Remove trailing spaces from 'name'

Remove all trailing spaces from 'name' because they show up
in the filter headings, which I find incredibly annoying.

(Bitbake rev: 263eae9d2d7acf62240320765c80f60f3553f620)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Show "No builds found" in the builds table
Belen Barros Pena [Sun, 16 Mar 2014 14:14:47 +0000 (14:14 +0000)]
bitbake: toaster: Show "No builds found" in the builds table

Making sure the h1 of build.html shows "No builds found"
instead of "0 builds found" when a search returns no results.
This matches the builds table to all other Toaster tables.

(Bitbake rev: c8495c38b892d22a1f85286f34b0fdbc17febf78)

Signed-off-by: Belen Barros Pena <belen.barros.pena@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: disable configvar sorts for value and files
David Reyna [Wed, 26 Mar 2014 13:43:04 +0000 (06:43 -0700)]
bitbake: toaster: disable configvar sorts for value and files

Disable the sort on files because it sorts on the file array's
first (and invisible) element. Disable the sort on values
because the raw ASCII sort looks wrong to the general user,
especially for values with leading spaces.

[YOCTO #6004]

(Bitbake rev: 800cbddd612c977960aa4dd93b24c22aac4bfae0)

Signed-off-by: David Reyna <David.Reyna@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: use deploy_dir var to obtain the license.manifest path
Cristiana Voicu [Thu, 27 Mar 2014 09:58:02 +0000 (11:58 +0200)]
bitbake: toaster: use deploy_dir var to obtain the license.manifest path

[YOCTO #6051]
(Bitbake rev: 6dd8133b06bbda5cce50de39123f429a6a3f772d)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: Display task description
Ravi Chintakunta [Tue, 18 Mar 2014 19:17:31 +0000 (15:17 -0400)]
bitbake: toaster: Display task description

Display task description as content of help bubble for a task.

[YOCTO #5748]

(Bitbake rev: 4ffc380d11cff4e1d32d2bf5fb6c7cda6f7e22e6)

Signed-off-by: Ravi Chintakunta <ravi.chintakunta@timesys.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: fix timezone detection
Alexandru DAMIAN [Fri, 28 Mar 2014 14:43:48 +0000 (14:43 +0000)]
bitbake: toaster: fix timezone detection

This patch replaces faulty timezone detection with a version
that simply reads the TZ environment variable if it is set.

If the TZ is not set, we do a reverse match search among known
timezone definitions and take the first match.

    [YOCTO #5499]

(Bitbake rev: 3a0a556a65368f02635606e4eb707ca08e25007a)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: toaster: update saving sstate task data
Alexandru DAMIAN [Tue, 25 Mar 2014 15:51:07 +0000 (15:51 +0000)]
bitbake: toaster: update saving sstate task data

This is an update on the sstate file saving data.
It saves both found and missed sstate tasks.

(Bitbake rev: 60c577b1080219b795d3c8ab4e149e929cf9ce14)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: runqueue: Address issues with incomplete sstate sets
Richard Purdie [Tue, 1 Apr 2014 08:16:30 +0000 (09:16 +0100)]
bitbake: runqueue: Address issues with incomplete sstate sets

The first part of the sstate code checks en-mass whether given checksums
are available. The next part of the code then either triggers those
setscene tasks either running them or skipping them if they've been
covered by others.

The problems was that this second part would always skip a task if it
was unavailable in the first part, even if it would have otherwise been
covered by other tasks.

This mean the mere presence of an artefact (or lack of presence) could
cause a different build failure.

The issue reproduces if you run a build and populate an sstate feed, then
run a second build off that feed, then run a third build off the sstate
feed of the second build (which is reduced in size).

The fix is rather than immediately skipping tasks if the checksum is
unavailable, create a list of missing tasks, then, if that task cannot
be covered by others we can skip it later. The deferral makes the
behaviour the same even when the cache is "incomplete".

[YOCTO #6081]

(Bitbake rev: 5edb1a3e3f454ba6e65551174d86229db2f99636)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: runqueue: Fix sstate task dependency problems
Richard Purdie [Mon, 31 Mar 2014 22:07:20 +0000 (23:07 +0100)]
bitbake: runqueue: Fix sstate task dependency problems

If a setscene task has [depends], its possible they may still get executed out
of order. The issue is that the dependencies are set to set() for all tasks
involved. This patch adds back in explict dependencies within these chains
to avoid the setscene task failures.

[YOCTO #6069]

(Bitbake rev: 724c889eed3b03d3199810c185086d3973af826c)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoDrop beagleboard, replaced by beaglebone
Richard Purdie [Fri, 28 Mar 2014 18:07:01 +0000 (18:07 +0000)]
Drop beagleboard, replaced by beaglebone

(From meta-yocto rev: b125761e5e31fc506f7b5b87308850cd573ab1b6)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agopoky: add edgerouter information to README.hardware
Yang Wei [Mon, 31 Mar 2014 18:01:59 +0000 (14:01 -0400)]
poky: add edgerouter information to README.hardware

Adding the summary, preparation and boot information for the newly introduced
edgerouter BSP.

(From meta-yocto rev: cc8c7a3c94d2cafe899adc7f8199237d4a72f205)

Signed-off-by: Yang Wei <Wei.Yang@windriver.com>
Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolinux-libc-headers: remove 3.10 recipe
Bruce Ashfield [Mon, 31 Mar 2014 17:56:33 +0000 (13:56 -0400)]
linux-libc-headers: remove 3.10 recipe

3.14 is now the reference for libc-headers. After building and booting 3.x based
BSPs against the 3.14 headers, we can safely remove the old version and patches
that are now part of the mainline kernel.

(From OE-Core rev: ade26bc63fdf89f297bec5f67bfff108e90438fc)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolibc-headers: set TC default to 3.14
Bruce Ashfield [Mon, 31 Mar 2014 17:56:32 +0000 (13:56 -0400)]
libc-headers: set TC default to 3.14

(From OE-Core rev: acd025cbeef533137522e60dd2cc5a6a4300d1b5)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolinux-libc-headers: add 3.14 libc headers
Bruce Ashfield [Mon, 31 Mar 2014 17:56:31 +0000 (13:56 -0400)]
linux-libc-headers: add 3.14 libc headers

Introduce the 3.14 linux-libc-headers recipe, now that the 3.14 kernel is
available, and the default for the qemu reference BSPs.

The three patches which were required for the previous 3.10 libc-headers
are not required for 3.14 and can be ignored.

(From OE-Core rev: 6b25918ecce6d555632d9576e16ad443b5d1780a)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolinux-libc-headers: make compression format configurable
Bruce Ashfield [Mon, 31 Mar 2014 17:56:30 +0000 (13:56 -0400)]
linux-libc-headers: make compression format configurable

As of the 3.13 kernel bz2 compressed tarballs are not available. To support
older header tarballs, and newer ones that require the 'xz' compressed
bundles, we can break out a variable that allows versioned libc headers to
select the archive format that works.

(From OE-Core rev: 116228fceca7fc2a7b557133b1f8f28f41af1ee5)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolinux-yocto/3.14: introduce versioned recipes
Bruce Ashfield [Mon, 31 Mar 2014 17:56:29 +0000 (13:56 -0400)]
linux-yocto/3.14: introduce versioned recipes

The release kernel for Yocto 1.6 is the 3.14 kernel, so we introduce
the versioned recipes here.

(From OE-Core rev: 92776093766d4b0bb2613214274fa28dc59b6126)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoRevert "zisofs-tools-native: Add missing dependency on zlib-native"
Richard Purdie [Mon, 31 Mar 2014 21:54:39 +0000 (22:54 +0100)]
Revert "zisofs-tools-native: Add missing dependency on zlib-native"

This reverts commit c581059df88d5801cc20ab24a096e4a67b737d49.

The same thing was already applied at a slightly different place in the file.

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolib/oeqa: add a test target controller for EFI targets
Stefan Stanacar [Sun, 30 Mar 2014 14:47:35 +0000 (17:47 +0300)]
lib/oeqa: add a test target controller for EFI targets

The purpose of this module is to deploy a test image on a EFI-enabled hardware
and run our runtime tests. A bit of background:
 - testimage.bbclass uses the concept of TEST_TARGET which is a class name
that is responsible for target deploying. A layer can provide
it's own TEST_TARGET. Right now has OE-core has a QemuTarget and a SimpleRemoteTarget
(ssh into an already up and running machine and run tests), the default one being qemu.
 - basically testimage does something like:
    target.deploy()
    try:
target.start()
        runTests()
    finally:
target.stop()

This module assumes a running EFI machine with gummiboot as bootloader and
core-image-testmaster installed (or similar). Also your hardware under test has
to be in a DHCP-enabled network that gives it the same IP for each reboot.

One time setup (master image):
 - build core-image-testmaster with EFI_PROVIDER = "gummiboot"
 - install the image on the target

Test image setup:
 - build your test image, e.g core-image-sato as you usually do, but with these in local.conf:
    IMAGE_FSTYPES += "tar.gz"
 - Now run the tests:
    INHERIT += "testimage"
    TEST_TARGET = "GummibootTarget"
    TEST_TARGET_IP = "192.168.2.3"
    bitbake core-image-sato -c testimage

Other notes:
 - TEST_POWERCONTROL_CMD (togheter with TEST_POWERCONTROL_EXTRA_ARGS) can be a command that runs on the host and does power cycling.
The test code passes one argument to that command: off, on or cycle (off then on). In my case I use something like
TEST_POWERCONTROL_CMD="powercontrol.exp test 10.11.12.1 nuc1" in local.conf.
Basically my expect script does: 'ssh test@10.11.12.1 "pyctl nuc1 <arg>" and runs a python script there that controls power for a label called nuc1'.
The reason why my expect script has to ssh into another machine is because of network topology, and that machine is the one actually connected
to the test rack and the power strip. That's why TEST_POWERCONTROL_CMD and _ARGS need to be customized for one's setup, the only requirement being
that it accepts: on/off/cycle as the last argument.
 - if no command is defined it would use classic reboot. This is fine as long as the machine
actually reboots (as in the ssh test hasn't failed), but it's useful for "simple-setup-with-one-board-on-the-desk" scenario, where
some manual interaction is okay from time to time.

[YOCTO #5614]

(From OE-Core rev: e00f888a88d0851b088c232dec66418e575a2e90)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agorecipes-extended: add master image for testing purposes
Stefan Stanacar [Sun, 30 Mar 2014 14:47:34 +0000 (17:47 +0300)]
recipes-extended: add master image for testing purposes

Add a custom recipe with custom initramfs and installers, that makes
it easy to deploy a master image for testing purposes.
We need a master image running on the target hardware, that should be a
known good build, with a set of utilities installed so that we use it
to deploy the images under test.

This core-image-testmaster recipe isn't a requirement per se, any image can
be used as long as the required conditions are met.
The test code assumes:
 - that the device has a second rootfs labeled as testrootfs
 - it has a properly configured bootloader entry (called test) for the second kernel and rootfs
 - the master image has a /etc/masterimage file so it can differentiate between master
and test images
 - the master image has tar, mount, bash (basically the normal linux utilities not the busybox
ones)

[YOCTO #5614]

(From OE-Core rev: dea237ccd9407288cd3a73e1deca270619dd6d4a)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoinitrdscripts: add install scripts with a second rootfs
Stefan Stanacar [Sun, 30 Mar 2014 14:47:33 +0000 (17:47 +0300)]
initrdscripts: add install scripts with a second rootfs

These are meant to be used by a master image, for a simple
initial setup.
The install scripts are similar to the default ones, but:
- custom partitioning, replaces the swap partiton with a second root filesystem
- adds labels to the partitions
- preconfigures a boot loader entry for the second rootfs

Part of [YOCTO #5614]

(From OE-Core rev: 39fcab00cd3b85d40966689e31b4c7748f630739)

Signed-off-by: Stefan Stanacar <stefanx.stanacar@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoopenssh: build without libbsd
Joe Slater [Sun, 30 Mar 2014 20:48:45 +0000 (13:48 -0700)]
openssh: build without libbsd

We do not DEPEND on libbsd, so we do not want to
build with it just because libutil.h is found by configure.
As noted in the patch, specifying --disable-libutil to
configure does not work, so we provide "cached" configure
variables.

(From OE-Core rev: 103ef2295c728e427acc27bb071e786946c459f2)

Signed-off-by: Joe Slater <jslater@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoclasses/sanity: check if SDKMACHINE setting has taken effect
Paul Eggleton [Mon, 31 Mar 2014 10:29:36 +0000 (11:29 +0100)]
classes/sanity: check if SDKMACHINE setting has taken effect

If you try to set SDKMACHINE in a distro configuration file, it won't
take effect because by the time that is parsed the line in bitbake.conf
which includes the appropriate conf file for SDKMACHINE has already been
parsed. Check that SDK_ARCH has changed from its default value and show
an error if it hasn't in order to catch this misconfiguration.

Fixes [YOCTO #5861].

(From OE-Core rev: 25ba4042ae782016aaf1cb5d3dac09b2a1030a1e)

Signed-off-by: Paul Eggleton <paul.eggleton@linux.intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agopseudo-1.5.1: keep install command directory mode
yanjun.zhu [Mon, 31 Mar 2014 14:41:08 +0000 (22:41 +0800)]
pseudo-1.5.1: keep install command directory mode

When install command sets the created directory mode, pseudo will change
the mode of the directory to 0700 incorrectly. Backport patch to fix it.

(From OE-Core rev: af595b09d570cbd320e4e138651144ac96bfbb83)

Signed-off-by: yanjun.zhu <yanjun.zhu@windriver.com>
Signed-off-by: Kai Kang <kai.kang@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolibomxil-0.9.3: fix configure unrecognised option
Matthieu Crapet [Mon, 31 Mar 2014 16:14:57 +0000 (18:14 +0200)]
libomxil-0.9.3: fix configure unrecognised option

Drop --disable-ffmpegcomponents which is deprecated since libomxil-bellagio-0.9.1
Explicitly disable doc generation to prevent using doxygen from build machine.

Components are external and are available separately here:
http://sourceforge.net/projects/omxil/files/components/

(From OE-Core rev: ff321fec0c5611b69a99901cac74bfd76b409d77)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agotoaster.bbclass: the license.manifest is located in DEPLOY_DIR
Cristiana Voicu [Mon, 31 Mar 2014 16:39:19 +0000 (17:39 +0100)]
toaster.bbclass: the license.manifest is located in DEPLOY_DIR

Replaced DEPLOY_DIR_IMAGE with DEPLOY_DIR

[YOCTO #6051]
(From OE-Core rev: 64c0ea3adccfc8af911b5e25e4c64a0abc578e6c)

Signed-off-by: Cristiana Voicu <cristiana.voicu@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agosstate.bbclass: update missed sstate event
Alexandru DAMIAN [Mon, 31 Mar 2014 16:39:18 +0000 (17:39 +0100)]
sstate.bbclass: update missed sstate event

This is a patch to update the missed sstate event with
info about the sstate files locations that were found.
It's needed as to display the found file in the toaster ui.

Also fixes a bug where a setscene task may have appeared in the
missed list even if it was found in a sstate mirror.

(From OE-Core rev: ad66cd521d3e661dd57c5aa02c204585101984f3)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agopackagegroup-core-tools-testapps: Move GLTOOLS to X11GLTOOLS
Otavio Salvador [Mon, 31 Mar 2014 14:03:09 +0000 (11:03 -0300)]
packagegroup-core-tools-testapps: Move GLTOOLS to X11GLTOOLS

piglit and mesa-demos are not buildable in x11-less distros so we must
to add those only when opengl and x11 DISTRO_FEATURES are available.

(From OE-Core rev: ab0bb02d410e4f0713e75192eb217991b3f672aa)

Signed-off-by: Otavio Salvador <otavio@ossystems.com.br>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolinux-yocto/3.10: fix qemuarm build failure
Bruce Ashfield [Sun, 30 Mar 2014 19:14:46 +0000 (15:14 -0400)]
linux-yocto/3.10: fix qemuarm build failure

The v3.10.24 merge created a merge conflict, which was not properly
resolved. Fixing the merge conflict and fixing the build of qemu arm.

(From OE-Core rev: 2116e326d9d7039aac4ec6c7ae5d2a2bedfb4a74)

Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agosstatesig: Anchor inherits class tests
Richard Purdie [Sun, 30 Mar 2014 22:49:52 +0000 (22:49 +0000)]
sstatesig: Anchor inherits class tests

There was a nasty sstate hash corruption issue occurring where the
fact the testimage bbclass was inherited meant that the checksum
changed due to testimage.bbclass being confused with image.bbclass.

This patch anchors the bbclass names to avoid this confusion.

(From OE-Core rev: 943a75a4f3b6877e4092dae14b59b7afef8cad3d)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: user-manual-execution.xml: Added how BB processes curly braces.
Scott Rifenbark [Fri, 28 Mar 2014 13:12:52 +0000 (13:12 +0000)]
bitbake: user-manual-execution.xml: Added how BB processes curly braces.

(Bitbake rev: f4ebc4de63d64e3b5f87e1d0f51507760c3d82d7)

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: user-manual-execution.xml: Updated "Checksum (Signatures) section.
Scott Rifenbark [Thu, 27 Mar 2014 15:35:43 +0000 (15:35 +0000)]
bitbake: user-manual-execution.xml: Updated "Checksum (Signatures) section.

Fixes [YOCTO #5357]

Added a paragraph at the end of this section to address the fact
that the user can now debug Bitbake's processing of signatures.
This introduces the -S bitbake option and the fact that the
user can provide a couple parameters with it: "none" and
"printdiff".

(Bitbake rev: 4093fa6c96eebe0bbafb93dc27d8a978cca436c4)

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: user-manual-intro.xml: updated bitbake -h output
Scott Rifenbark [Thu, 27 Mar 2014 13:59:40 +0000 (13:59 +0000)]
bitbake: user-manual-intro.xml: updated bitbake -h output

The new -S syntax is in there now.

(Bitbake rev: 5a406a69c14b6ddf179fea2c4bc68cddf9eddeea)

Signed-off-by: Scott Rifenbark <scott.m.rifenbark@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoxkeyboard-config: Add missing dependency on libxslt-native for xsltproc
Richard Purdie [Thu, 27 Mar 2014 17:11:43 +0000 (17:11 +0000)]
xkeyboard-config: Add missing dependency on libxslt-native for xsltproc

(From OE-Core rev: 3366da30f32653ce0dc28878029df81c5ca9bcc5)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: server/xmlrpc: Simlify featureset handling
Richard Purdie [Fri, 28 Mar 2014 17:06:43 +0000 (17:06 +0000)]
bitbake: server/xmlrpc: Simlify featureset handling

Rather than passing featureset around various places where the data doesn't
really belong, run a command at connection time to set the appropriate
features. This is similar to what the process server does.

(Bitbake rev: c3b5cc5691291c74dd315c4439c80e0e4b2b5c1d)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: bin/bitbake: Only try and process an event_queue if it exists
Richard Purdie [Fri, 28 Mar 2014 17:07:21 +0000 (17:07 +0000)]
bitbake: bin/bitbake: Only try and process an event_queue if it exists

The connection may have failed before the event queue has been setup.
Handle this correctly in the exception handler.

(Bitbake rev: db4d80b5c2d32117cdf06333b9627202998b1512)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: bitbake: toaster: do not trap SIGCHLD
Alexandru DAMIAN [Fri, 28 Mar 2014 15:30:35 +0000 (15:30 +0000)]
bitbake: bitbake: toaster: do not trap SIGCHLD

We remove trapping SIGCHLD due to a weird interaction
with the bash version used with Ubuntu 14.04 LTS.

(Bitbake rev: 59f2f33440449c586c23dd3a192698a37aaf0595)

Signed-off-by: Alexandru DAMIAN <alexandru.damian@intel.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: prserv: Fix exit race issues
Richard Purdie [Fri, 28 Mar 2014 12:10:54 +0000 (12:10 +0000)]
bitbake: prserv: Fix exit race issues

We shouldn't immediately remove the pid file when stopping the server, if we do, this
causes a traceback within the server itself which can then hang. Fix this by removing
the stale pid file as the last thing we do.

Also:

* don't printing a new "waiting" line every 0.5 seconds.
* make the loop more granular since the user can 'feel' the 0.5 seconds

[YOCTO #5984]

(Bitbake rev: 81f41a806aeddcc38992163557672e296bcbc967)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: bin/bitbake/cooker: Ensure initial featureset is optimal
Richard Purdie [Fri, 28 Mar 2014 11:09:19 +0000 (11:09 +0000)]
bitbake: bin/bitbake/cooker: Ensure initial featureset is optimal

If the featureset didn't match the defaults, we'd pay the price of two
base configuration parses which showed up adversely in the performance
benchmarks. This also passes the feature set into the cooker creation so
in the common case we don't have to reset the server. This speeds up
both knotty and hob startup. If the featureset doesn't match, the system
will reset as before, this just streamlines the common case.

(Bitbake rev: 1249543c4dbf3edeac033d888497864cfc807a4e)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agobitbake: cooker: Only change self.data if it exists
Richard Purdie [Fri, 28 Mar 2014 11:06:18 +0000 (11:06 +0000)]
bitbake: cooker: Only change self.data if it exists

With the change to more optimal default featureset behaviour, a race was
exposed by hob where the code may try and change self.data before it
exists. This change avoids that.

When the datastore is created, the cooker configuration is used so
data tracking is correctly handled regardless.

(Bitbake rev: 9d8f7efbc39d64124936ccaeb3c47a112e595d78)

Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agozisofs-tools-native: Add missing dependency on zlib-native
Drew Moseley [Thu, 27 Mar 2014 18:18:31 +0000 (14:18 -0400)]
zisofs-tools-native: Add missing dependency on zlib-native

(From OE-Core rev: c581059df88d5801cc20ab24a096e4a67b737d49)

Signed-off-by: Drew Moseley <drew_moseley@mentor.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agoe2fsprogs: Add e2fsprogs-mke2fs and e2fsprogs-e2fsck as recommend packages
Yue Tao [Fri, 28 Mar 2014 09:43:09 +0000 (17:43 +0800)]
e2fsprogs: Add e2fsprogs-mke2fs and e2fsprogs-e2fsck as recommend packages

The mke2fs and e2fsck commands are regular tools of e2fsprogs, so they
should be installed.

(From OE-Core rev: a2b6610dceb0c27484d3958d84d7f491aa759a04)

Signed-off-by: Yue Tao <Yue.Tao@windriver.com>
Signed-off-by: Robert Yang <liezhi.yang@windriver.com>
Signed-off-by: Jeff Polk <jeff.polk@windriver.com>
Signed-off-by: Hongxu Jia <hongxu.jia@windriver.com>
Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
10 years agolttng-ust: add python to lttng-ust's RDEPENDS
Hongxu Jia [Fri, 28 Mar 2014 09:43:07 +0000 (17:43 +0800)]
lttng-ust: add python to lttng-ust's RDEPENDS

| Note: adding Smart RPM DB channel
|
| Note: to be installed:  run-postinsts@x86_64 kernel-modules@qemux86_64 packagegroup-core-boot@qemux86_64 lttng-ust@x86_64
| Loading cache...
| Updating cache...               ######################################## [100%]
|
| Computing transaction...error: Can't install lttng-ust-2:2.3.0-r0.0@x86_64: no package provides /usr/bin/python
|
| Saving cache...
|
| WARNING: exit code 1 from a shell command.
| ERROR: Function failed: do_rootfs (log file is located at tmp/work/qemux86_64-wrs-linux/wrlinux-image-glibc-small/1.0-r1/temp/do_rootfs/log.do_rootfs.13619)

(From OE-Core rev: 0360a49f0c216b3f35c4bffabab3406c31589fd3)

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